Handbook of Oriental Studies Handbuch der Orientalistik

Homework

profileAlammari

ENG310-BaalbakTheArabicLexicographicalTradition.pdf

 

The Arabic Lexicographical Tradition

 

 

Handbook of Oriental Studies Handbuch der Orientalistik

section one

The Near and Middle East

Edited by

Maribel Fierro (Madrid) M. Şükrü Hanioğlu (Princeton)

Renata Holod (University of Pennsylvania) Kees Versteegh (Nijmegen)

VOLUME 107

The titles published in this series are listed at brill.com/ho1

 

 

The Arabic Lexicographical Tradition

From the 2nd/8th to the 12th/18th Century

By

Ramzi Baalbaki

LEIDEN | BOSTON

 

 

Cover illustration: © MS Landberg MSS 247, Beinecke Rare Book and Manuscript Library of Yale University, 63 v–64 r.

Library of Congress Cataloging-in-Publication Data

Ba’labakki, Ramzi.  The Arabic lexicographical tradition : from the 2nd/8th to the 12th/18th century / by Ramzi Baalbaki.   pages cm. — (Handbook of Oriental studies. Section 1, the Near and Middle East, ISSN 0169-9423 ; volume 107)  Includes bibliographical references and index.  ISBN 978-90-04-27397-9 (hardback : alk. paper) — ISBN 978-90-04-27401-3 (e-book) 1. Arabic language— Lexicography—History. I. Title.

PJ6611.B35 2014  492.703—dc23

2014008094

This publication has been typeset in the multilingual ‘Brill’ typeface. With over 5,100 characters covering Latin, ipa, Greek, and Cyrillic, this typeface is especially suitable for use in the humanities. For more information, please see brill.com/brill-typeface.

issn 0169-9423 isbn 978 90 04 27397 9 (hardback) isbn 978 90 04 27401 3 (e-book)

Copyright 2014 by Koninklijke Brill nv, Leiden, The Netherlands. Koninklijke Brill nv incorporates the imprints Brill, Brill Nijhoff, Global Oriental and Hotei Publishing. All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission from the publisher. Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill nv provided that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive, Suite 910, Danvers, ma 01923, usa. Fees are subject to change.

This book is printed on acid-free paper.

 

 

Contents

Preface  vii

1 Early Lexicographical Activity  1 1 The Background of Linguistic Study  1 2 The Speech of the Bedouins (Aʿrāb)  7 3 The Collection of Data  16 4 The Epochs of Reliable Usage (ʿUṣūr al-Iḥtiǧāǧ)  29 5 The Role of Ġarīb  36 6 The Compilation of Lexica  46 7 Remarks on Contemporary Scholarship and the Originality of Arabic

Lexicography  53

2 Mubawwab (Onomasiological) and Specialized Lexica  62 1 Introduction  62 2 al-Ġarīb (Strange Usage) and al-Nawādir (Rare Usage)  63

2.1 Ġarīb al-Qurʾān  64 2.2 Ġarīb al-Ḥadīṯ  71 2.3 General Ġarīb/Nādir Material  84

3 al-Amṯāl (Proverbs)  100 3.1 Books with no Alphabetical Arrangement  104 3.2 Alphabetically Arranged Books  112 3.3 Books Specialized in a Particular Aspect of Proverbs  120

4 al-Nabāt (Plants), al-Ḥayawān (Animals), Ḫalq al-Insān (Human Body), etc.  132 4.1 Plants  135 4.2 Animals  140 4.3 Human Body  148 4.4 Miscellanea  156

5 al-Muʿarrab (Arabized Words)  161 6 Laḥn al-ʿĀmma (Solecism)  170 7 al-Aḍdād (Words with Two Contradictory Meanings)  188 8 al-Muštarak (Homonyms) and al-Mutarādif (Synonyms)  198 9 al-Ḥurūf/al-Aṣwāt (Particles/Letters)  211 10 al-Abniya (Morphological Patterns)  232

10.1 al-Ištiqāq (Derivation)  234 10.2 al-Muḏakkar wa-l-muʾannaṯ (Masculine and Feminine)  239 10.3 al-Maqṣūr wa-l-mamdūd (Abbreviated and Prolonged

Patterns)  241

 

 

vi contents

10.4 al-Muṯallaṯāt (Triplets)  248 10.5 Nominal Patterns  251 10.6 Faʿala and Af ʿala  254 10.7 Verbal Patterns in General  258 10.8 Nominal and Verbal Patterns  262

11 Multithematic Works  266

3 Muǧannas (Semasiological) Lexica  279 1 Introduction  279 2 The Phonetic-Permutative System  280

2.1 Kitāb al-ʿAyn by al-Ḫalīl b. Aḥmad  282 2.2 al-Bāriʿ fī l-luġa by Abū ʿAlī al-Qālī  303 2.3 Tahḏīb al-luġa by Azharī  311 2.4 al-Muḥīṭ fī l-luġa by al-Ṣāḥib b. ʿAbbād  319 2.5 al-Muḥkam wa-l-muḥīṭ al-aʿẓam by Ibn Sīda  322

3 The Alphabetical System  329 3.1 K. al-Ǧīm by Abū ʿAmr al-Šaybānī  332 3.2 Ǧamharat al-luġa by Ibn Durayd  338 3.3 Maqāyīs al-luġa and Muǧmal al-luġa by Ibn Fāris  347 3.4 Asās al-balāġa by Zamaḫšarī  356

4 The Rhyme System  363 4.1 al-Taqfiya fī l-luġa by Bandanīǧī  370 4.2 Tāǧ al-luġa wa-ṣaḥāḥ al-ʿArabiyya (al-Saḥāḥ/al-Ṣiḥāḥ)

by Ǧawharī  373 4.3 al-ʿUbāb al-zāḫir wa-l-lubāb al-fāḫir by Ṣaġānī  381 4.4 Lisān al-ʿArab by Ibn Manẓūr  385 4.5 al-Qāmūs al-muḥīṭ by Fīrūzābādī  391 4.6 Tāǧ al-ʿarūs min ǧawāhir al-Qāmūs by Zabīdī  397

Epilogue  402

Bibliographical References  419 Primary Sources  419 Secondary Sources  452

Indices  467 Index of Names  467 Index of Books  477

 

 

Preface

Rooted in the Arabic-Islamic culture and firmly linked to other indigenous scholarly interests, particularly Qurʾānic sciences and grammar, the lexico- graphical tradition represents a major aspect of the Arabic linguistic heri- tage. It is quite remarkable that al-Ḫalīl b. Aḥmad’s (d. 175/791) K. al-ʿAyn, and more specifically its introduction which lays the foundations of lexicography, is contemporaneous with the most influential work on grammar, al-Kitāb, authored by his student Sībawayhi (d. 180/796). Both works are closely related to the study of the Qurʾān, heavily rely on the material which was compiled by the philologists during the era of data collection as of the early second/eighth century, and derive attested material or šawāhid primarily from poetry. The early distinction between naḥw – which deals with syntax, morphology/mor- phophonolgy and, to a lesser extent, phonetics – and luġa – a domain that explores meanings of words, dialectal variations, strange usage (ġarīb), etc. – eventually gave rise to two diverse, but related, traditions, namely, grammar and lexicography. Both disciplines had an immense influence on Arabic cul- ture, and this is best demonstrated by the fact that the “classical” period of each extends over a thousand years.

The growing interest in the last few decades in the study of the linguistic tradition of the Arabs has primarily focused on grammar. In spite of the recent publication of several important lexica and of better editions of previously published ones, progress in the study of lexicography has been relatively mod- est. Although J. Haywood’s Arabic Lexicography (first published in 1959; second edition in 1965) has been outdated for quite a long time, it has remained the main book-length general study of the field in western scholarship for more than half a century. There are also a few pioneering works in Arabic, fore- most among which is Ḥ. Naṣṣār’s al-Muʿǧam al-ʿArabī (first published in 1956; second edition in 1968), which has had an immense influence on later stud- ies and remains unsurpassed in many ways. It should be noted, however, that several of the works which Naṣṣār uses in manuscript form have since been published and others have become known. Furthermore, in the onomasio- logical or mubawwab type he ignores a number of important genres, such as proverbs and words with two contradictory meanings (aḍdād), as well as the majority of subjects related to particles (ḥurūf) and letters (aṣwāt). The pres- ent book is an attempt to write as complete a history of the lexicographical tradition up to the twelfth/eighteenth century as our present knowledge of the sources allows. Although setting the twelfth/eighteenth century as a chrono- logical limit carries us beyond the medieval period in the strict sense of the

 

 

viii preface

term, it allows for the inclusion of some of the major works – such as Zabīdī’s (d. 1205/1790) Tāǧ al-ʿarūs – which belong in method and spirit to a much ear- lier period. Lexicography in the modern period is not within the scope of this work and certainly deserves a study in its own right. It is hoped that this book will place lexicography within the wider context of Islamic scholarship and shed light on several aspects of the field that have not hitherto received the attention they merit.

Chapter One examines early lexicographical activity and the factors that gave rise to the compilation of lexica. Due to the reliance of the lexicogra- phers in the early period on the speech of those Bedouins (Aʿrāb) whom they describe as eloquent ( fuṣaḥāʾ), their social and linguistic characteristics as portrayed in the sources are discussed, as are the notions of the epochs of reli- able usage (ʿuṣūr al-iḥtiǧāǧ) and ġarīb, both of which are strongly related to Bedouin speech. Each of the two other chapters is devoted to one of the two types of lexical writing, the onomasiological and the semasiological, or, to use Ibn Sīda’s (d. 458/1066) terminology, the mubawwab and the muǧannas.

The first type includes lexica that deal with a specific topic area or genre as well as multithematic works or thesauri that deal with several topics. One finds in the same genre – such as proverbs and aḍdād – works that are arranged according to form (lafẓ), e.g. alphabetically, and others that do not follow a formal arrangement. Yet both types will be discussed within the same genre in order to present a full picture of the development of that genre and a more meaningful comparison among the works that fall under it. Given the large number of mubawwab works, the survey cannot always be exhaustive, but in each of the ten sections of which Chapter Two consists (apart from a short introduction), all the sources that are deemed representative of the genre at hand are discussed, in particular the ones whose impact on the develop- ment of the genre or influence on subsequent authors is most obvious. In certain cases, the vastness of the genre dictates its division into several sub- genres, as in the section on morphological patterns (abniya), which is divided into eight subsections, e.g. derivation (al-ištiqāq), masculine and feminine (al-muḏakkar wa-l-muʾannaṯ), abbreviated and prolonged patterns (al-maqṣūr wa-l-mamdūd), triplets (al-muṯallaṯāt), etc.

The term semasiological or muǧannas will exclusively refer to the general lexica in which sign leads to meaning and which, rather than being special- ized in a specific topic or several topics, normally aim at listing all the roots or lexical items of the language. Yet the boundaries between the two types, the mubawwab and the muǧannas, are not always clear. Other than the fact that they are contemporaneous and do not represent successive stages in lex- icographical writing, it is not always easy to determine under which type a

 

 

ixpreface

certain work should be discussed. Both also share much of their source mate- rial, but each obviously serves a different purpose, as is evident in the case of authors who compiled lexica of both types. As far as numbers are concerned, muǧannas lexica are considerably fewer than mubawwab ones (hence the rela- tive shortness of Chapter Three in comparison with Chapter Two), although in most cases muǧannas works are more voluminous than mubawwab ones. In the three sections based on the type of arrangement of muǧannas lexica, i.e. the phonetic-permutative, the alphabetical and the rhyme systems, the discus- sion will concentrate not only on the formal arrangement of each lexicon but also on other fundamental issues, such as the scope of its material, the position of its author vis-à-vis strange or rare usage, the internal arrangement of the lemmata, and its reception by subsequent authors.

The decision to undertake the ambitious endeavor of writing a detailed study of the lexicographical tradition came after a lengthy phase of hesitation. The idea has incessantly been on my mind since I prepared my own notes on the subject, part of which notes I used in a series of graduate seminars that have been conducted as of the early 1980s. The difficulties that inevitably surround this project continuously deterred me from embarking on it whenever I seri- ously considered doing so, although my notes were growing, and my percep- tion of the field in general was becoming more defined. When I finally decided that it was high time I did so, I received support and encouragement from a number of colleagues and friends, two of whom deserve special recognition. Kees Versteegh was the first colleague who I seriously consulted about the proj- ect, even suggesting to him many years ago to be co-author. Unfortunately he was then too busy to be able to take up my offer, but he has ever since relent- lessly reminded me of its importance. When I finally showed him an outline, he enthusiastically encouraged me to proceed with the aim of publishing the final product in the prestigious HdO series. The other source of encouragement came from my student, now colleague at the Department of Arabic & N.E.L. at the American University of Beirut, Bilal Orfali. Bilal took it upon himself to mercilessly remind me of the need for a book in English on Arabic lexicogra- phy, and exercised emotional blackmail on me by appealing to my soft spot: the Arabic language, to coerce me to see that this project comes to fruition. His words of encouragement during our lengthy discussions complemented his valuable help in obtaining a large number of sources, mainly in electronic form, that were otherwise inaccessible to me.

I am also indebted to many others for the kind help they extended to me during my work on the book. Samer Traboulsi provided me with copies of several books that I could not obtain, and Maurice Pomerantz kindly brought to my attention al-Ḥasan b. Muḥammad al-Ṣafadī’s manuscript al-Maqāmāt

 

 

x preface

al-Ǧalāliyya, which contains the very interesting al-Maqāma al-Ṣād-sīniyya wa-l-ġayn-rāʾiyya wa-l-ṭāʾ-tāʾiyya. My two graduate assistants Ali Rida and Nermine al-Horr carefully proofread parts of the manuscript. Mrs. Rula Baalbaki meticulously read the whole manuscript and professionally com- mented on stylistic matters. Mrs. Rana Kaidbey, the Jewett Chair administra- tive and research assistant, provided valuable assistance in technical matters, particularly transliteration and typesetting, and acted as liaison with the staff of the Jafet Library at AUB, whom I also thank for their continued profes- sional help.

Ramzi Baalbaki Margaret Weyerhaeuser Jewett Chair of Arabic American University of Beirut February 2014

 

 

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004�740�3_�0�

chapter 1

Early Lexicographical Activity

1 The Background of Linguistic Study

It is virtually impossible in the period which precedes the writing of linguis- tic treatises to separate between philological activity and interest in studying the Qurʾānic text. Indeed, a linguistic component is discernible in the earli- est material we have on Qurʾānic reading, prophetic tradition, jurisprudence and exegesis. Versteegh (1993), for example, has demonstrated that the non- technical terms used by early exegetes such as Muǧāhid b. Ǧabr (d. 104/722), Muḥammad b. al-Sāʾib al-Kalbī (d. 146/763), Muqātil b. Sulaymān (d. 150/767), and Sufyān al-Ṯawrī (d. 161/778) constituted the link between every-day vocab- ulary and the later technical terminology of the traditional grammarians. Obviously, this could not have been the case had there been an early distinc- tion between philology and exegesis. The material available in the sources concerning the early vocalization and dotting of the Qurʾānic text and thus Arabic script in general, irrespective of its trustworthiness,1 provides another example of the link between philology, particularly grammar (naḥw), and serv- ing the Qurʾānic text, at least as presented in the sources. Both the emergence and the refinement of the system of vowels and diacritics are associated with early grammarians, starting with Abū l-Aswad al-Duʾalī (d. 69/688), who is uni- versally credited in the sources with being not only the first to vocalize the text of the Qurʾān, but also the first to lay the foundations of grammar (awwal man waḍaʿa l-ʿArabiyya).2 His pupil, Naṣr b. ʿĀṣim (d. 89/708), also credited

1 It is now almost certain, for example, that dotting letters cannot be attributed to Naṣr b. ʿĀṣim (d. 89/708), or even to Abū l-Aswad al-Duʾalī (d. 69/688) if we are to understand that the naqṭ (lit. dotting) attributed to him refers not merely to vowels – as the sources explain – but also, more generally, to dots. The dates of the two men cannot support such an allegation. To take the earlier of the two, Abū l-Aswad, most sources agree that he died in 69 A.H., and a few (e.g. Zubaydī, Ṭabaqāt 26; Qifṭī, Inbāh I, 55) indicate that he lived eighty-five years. This would place his birth date at about 16 before Higra or 606 A.D. However, we possess an undated document whose archaeological context suggests that it was written between the second quarter of the sixth century and the beginning the seventh century A.D. and which consists of one fully vocalized word (the proper noun Nāyif); cf. al-Ghul (2004: 106). Furthermore, several words in the Zuhayr inscription, dated 24 A.H., are fully vocalized; cf. Ghabban (2008: 211–212).

2 Abū l-Ṭayyib, Marātib 26–29; Sīrāfī, Aḫbār 13–16; Zubaydī, Ṭabaqāt 21.

 

 

2 chapter 1

with being the first to lay the foundations of grammar (awwal man waḍaʿa l-ʿArabiyya),3 is said to have been the first to differentiate between similar sym- bols in script by using dots.4 Other pupils of Abū l-Aswad’s who are reported to have adopted and spread his system of vocalization include grammarians such as Maymūn al-Aqran (d. ?), ʿ Anbasa b. Maʿdān al-Fīl (d. ?), ʿ Abdallāh b. Abī Isḥāq (d. 117/735), and Yaḥyā b. Yaʿmar (d. 129/746).5 More significantly, none other than al-Ḫalīl b. Aḥmad (d. 175/791), who with Sībawayhi (d. 180/796) is the real founder of the Arabic grammatical theory, is said to have introduced further improvements to the Qurʾānic text, including marks for long vowels, hamza, šadda (doubling), and several phonetic phenomena.6 Accordingly, an uninterrupted chain which extends from Abū l-Aswad to Ḫalīl is established in the sources with respect to script improvement in order to minimize error in the recitation of the Qurʾān.

Qurʾānic readings (qirāʾāt) is another domain that reveals the link between linguistic enquiry and the study of the Qurʾān, which the sources highlight. It is thus no coincidence that most of the early scholars interested in luġa (philology) and/or naḥw were Qurʾānic readers (qurrāʾ).7 These include Naṣr b. ʿĀṣim,8 ʿAbdalraḥmān b. Hurmuz (d. 117/735),9 ʿAbdallāh b. Abī Isḥāq,10 ʿĪsā b. ʿUmar (d. 149/766),11 Abū ʿAmr b. al-ʿAlāʾ (d. 154/771),12 and Hārūn b. Mūsā (d. c. 170/786),13 in addition to Kisāʾī (d. 189/805), who, like Abū ʿAmr b. al-ʿAlāʾ, was one of the seven authorized readers.14 Furthermore, early linguistic activ- ity is linked with other domains of religious enquiry, such as Ḥadīṯ (prophetic tradition), fiqh (jurisprudence) and tafsīr (exegesis).15 Accordingly, it is safe to assume that the collection of the material which later formed the basis of Arabic lexicography took place within the wider context of these nascent domains which later developed into independent sciences.

3 Abū l-Ṭayyib, Marātib 32; Sīrāfī, Aḫbār 20; Zubaydī, Ṭabaqāt 27. 4 ʿAskarī, Taṣḥīf 14. 5 Dānī, Muḥkam 6 and Naqṭ 125; Zubaydī, Ṭabaqāt 29. 6 Dānī, Muḥkam 7 and Naqṭ 125. 7 Cf. Baalbaki (2008: 5–6). 8 Sīrāfī, Aḫbār 21. 9 Ibid., 22; Ibn al-Nadīm, Fihrist 45. 10 Ibn al-Anbārī, Nuzha 26; cf. Ibn al-Ǧazarī, Ġāya I, 410. 11 Ibn al-Anbārī, Nuzha 28; cf. Zubaydī, Ṭabaqāt 41 for some of his readings. 12 Sīrāfī, Aḫbār 21; Ibn al-Nadīm, Fihrist 30. 13 Suyūṭī, Buġya II, 321; Ibn al-Ǧazarī, Ġāya II, 348. 14 Ibn al-Anbārī, Nuzha 58. 15 Cf. Baalbaki (2008: 6–9).

 

 

3early lexicographical activity

Within the linguistically oriented sciences, as distinct from but closely related to qirāʾāt, Ḥadīṯ, fiqh and tafsīr, two main branches emerge from the beginning. These are luġa and naḥw. The latter, of course, is the study of gram- mar, which has developed into what is known as the Arabic Grammatical Tradition (AGT) and dealt with matters related to syntax, morphology (includ- ing morphophonology) and, to a lesser extent, phonetics. The branch dealing with luġa is identified with scholars referred to in the sources as luġawiyyūn, i.e. philologists (in the narrow sense of the word, rather than in the sense which is often applied to linguistic study in general) and lexicographers who, as of the second/eighth century, have begun to explore issues related to the collection of linguistic data, word meanings in attested material, and dialectal variations particularly in the realm of ġarīb (strange usage).16 At the early stages of lin- guistic activity in the second/eighth century, the boundaries between the two branches of luġa and naḥw is often blurred, and although the later biographi- cal sources make a sharp distinction between luġawiyyūn and naḥwiyyūn, this distinction is often simplistic and unconvincing. For example, Zubaydī (d. 379/989), in his Ṭabaqāt al-naḥwiyyīn wa-l-luġawiyyīn, lists members of each group separately, but at times lists a scholar – as is the case of Abū ʿAmr b. al-ʿAlāʾ – with both groups.17 On other occasions, his inclusion of a scholar in one group may be challenged, as with Abū l-Ḫaṭṭāb al-Aḫfaš al-Kabīr (d. 177?/793?), whom he lists under naḥwiyyūn, whereas the evidence furnished by Sībawayhi’s al-Kitāb suggests that the opposite is more likely to be true.18

The biographers’ distinction between luġa/luġawiyyūn and naḥw/naḥwiyyūn notwithstanding, the two main branches of linguistic sciences share to a large extent the factors that led to their emergence. Other than their common affin- ity to early Qurʾānic study, laḥn (solecism) seems to have contributed signifi- cantly in both branches to the early awareness of the need to seek reliable data in order to determine proper usage. As far as naḥw is concerned, it is highly significant that both the allegedly first grammarian, Abū l-Aswad al-Duʾalī, and the author of the first and most authoritative book in the history of the tradi- tion, Sībawayhi, reportedly started their grammatical enquiry in response to the occurrence of laḥn. The most famous riwāya or anecdote related to Abū l-Aswad is that his daughter addressed him by saying mā ašaddu l-ḥarri (What hotness is most severe?) when she wanted to exclaim and thus should have

16 Baalbaki (2007: xiii–xiv). 17 Zubaydī, Ṭabaqāt 35, 159. 18 Ibid., 40; cf. Baalbaki (2008: 14).

 

 

4 chapter 1

said mā ašadda l-ḥarra (How hot it is!).19 A similar anecdote interestingly demonstrates that the corrupt reading of the Qurʾān is what prompted Abū l-Aswad to lay the foundations of grammar (waḍaʿa l-naḥw). The anecdote has it that he heard someone recite the verse anna l-Lāha barīʾun min al-mušrikīna wa-rasūluhu (“that God and His Apostle dissolve [treaty] obligations with the Pagans”; Q 9: 3) by using the genitive in wa-rasūlihi, disastrously suggest- ing that God has repudiated his prophet, Muḥammad. As for Sībawayhi, the bibliographical sources20 report the following anecdote: He attended the circle of the Basran muftī and Ḥadīṯ scholar (muḥaddiṯ), Ḥammād b. Salama (d. 167/784), whom he heard relate the ḥadīṯ: laysa min aṣḥābī illā man law šiʾtu la-aḫaḏtu ʿalayhi laysa Abā l-Dardāʾi (There is no companion of mine that I would not find fault with if I wanted except Abū l-Dardāʾ). Having mistakenly thought that laysa is the negative particle which requires its noun to be in the nominative, rather than an exceptive particle that is followed by the accusa- tive, Sībawayhi volunteered to “correct” Ḥammād’s reading to Abū instead of Abā. When he finally realized that it was he who was mistaken, he vowed to seek a discipline (ʿilm) which would safeguard against error, and thus studied naḥw and authored his al-Kitāb.

But laḥn was not restricted to naḥw in the sense of the syntactical relation- ships among the constituents of speech, and in particular the effect of the operants (ʿawāmil) on the elements they govern. Several examples related to semantics show that certain words were not used in their proper context, resulting in constructions which generated ridicule. ʿUbaydallāh b. Ziyād b. Abīhi (d. 67/686), for example, is reported to have inappropriately used the verb fataḥa (to open) instead of salla (to unsheathe) in saying iftaḥū suyūfakum (lit. Open your swords) rather than sullū suyūfakum (Unsheathe your swords), prompting the ridicule of the poet Yazīd b. Mufarriġ al-Ḥimyarī (d. 69/688).21 A similar example is that Ḫālid b. ʿAbdallāh al-Qasrī (d. 126/743) was ridi- culed by poets and was sent a letter of rebuke by Caliph Hišām b. ʿAbdalmalik (d. 125/743) for having said while on stage aṭʿimūnī māʾan (lit. Feed me water).22 Most embarrassing of all was Suwayd b. Manǧūf’s (d. ?) use of ist (anus) with

19 For this and other anecdotes, see Abū l-Ṭayyib, Marātib 26; Sīrāfī, Aḫbār 17–19; Zubaydī, Ṭabaqāt 21–22; Ibn al-Anbārī, Nuzha 19–21; Qifṭī, Inbāh I, 50–51. See also Baalbaki (2008: 3–4).

20 Sīrāfī, Aḫbār 43–44; Zubaydī, Ṭabaqāt 66; Marzubānī, Muqtabas 95; Ibn al-Anbārī, Nuzha 54–55. See also Zaǧǧāǧī, Maǧālis 118 for another ḥadīṯ which Sībawayhi made an error in reading, which incident triggered his interest in grammar.

21 Ǧāḥiẓ, Bayān II, 210–211. 22 Mubarrad, Kāmil I, 31; see text of letter in IV, 121–125.

 

 

5early lexicographical activity

reference to earth when he addressed al-Haṯhāṯ b. Ṯawr (d. ?) by saying iǧlis ʿalā st (= ist) al-arḍ (lit. Sit on the earth’s anus)!23 Furthermore, one gains the impression from the anecdotes which Ǧāḥiẓ (d. 255/869) reports from a period as early as the first few decades of Hiǧra and thereafter that corruption of speech in aspects other than syntax (i.e. the domain of naḥw) was quite wide- spread. He refers to those non-Arabs who speak Arabic with an accent as aṣḥāb al-lakan min al-ʿAǧam and even specifies the accent as lukna Rūmiyya (Greek accent) or lukna Fārisiyya (Persian accent).24 What is more alarming is that foreign accents found their way to the speech of some Arabs who were closely connected with non-Arabs, such as Ṣuhayb b. Sinān (d. 38/659) of the tribe al-Namir b. Qāsiṭ, who is said to have had a lukna Rūmiyya, and ʿUbaydallāh b. Ziyād b. Abīhi, whose accent is described as Fārisiyya. Both men were unable to pronounce pharyngeal ḥ and thus changed it to laryngeal h.25

We have no reason to doubt the authenticity of the above reports, nor most of the other similar material reported by Ǧāḥiẓ. Unlike those anecdotes which may well have been fabricated to explain the emergence of naḥw as a discipline, Ǧāḥiẓ’s anecdotes seem more “natural”, and some of them are sup- ported by written documents such as the previously mentioned letter of Hišām b. ʿAbdalmalik to Ḫālid b. ʿAbdallāh al-Qasrī. What may be dubious, however, is the assertion of some authors, such as Zubaydī (d. 379/989), that laḥn was unknown in the Ǧāhiliyya (pre-Islamic times) and the early part (ṣadr) of Islam. According to this view, it was only after the Islamic conquest, and due to the intermingling of the Arabs with other nations, that corruption of speech took place.26 This view does not take into account the natural develop- ment of language and the fact that usage is constantly subject to change, often resulting in usage that linguists dub as corrupt. Moreover, reports that Prophet Muḥammad and some of his Companions were sharply critical of, even tak- ing punitive measures against those who committed errors in speech,27 are

23 Ǧāḥiẓ, Bayān II, 211. 24 Ibid., I, 71–72. 25 Ibid., I, 72. 26 Zubaydī, Laḥn 4. 27 For example, Prophet Muḥammad, having heard an orator commit laḥn, is said to have

advised that he be guided or instructed properly since he went astray (Abū l-Ṭayyib, Marātib 23; Ibn Ǧinnī, Ḫaṣāʾiṣ II, 8). It is remarkable that the two verbs used in the Prophet’s reported statement aršidū aḫākum fa-qad ḍalla (Guide your brother for he has gone astray) are connected with guidance (rašād) and straying from the right path (ḍalāl), as if to highlight the moral dimension of committing speech errors. Among the Prophet’s Companions, ʿUmar b. al-Ḫaṭṭāb (d. 23/644) reportedly ordered one of his administrators to whip the latter’s writer and delay his pay for a whole year for a mistake he had made in

 

 

6 chapter 1

most likely intended to provide irrefutable justification for the philologists and grammarians alike to combat the spread of laḥn.

Both the naḥwiyyūn and the luġawiyyūn of this early period seem to have been convinced that there was a standard of “good” Arabic against which they could judge usage. Their collection of linguistic data mainly from the Bedouins (see Section 2 below) proves that they actively sought a variety of Arabic which they could describe as free of laḥn. It is interesting to note here that, in both disciplines alike, laḥn is visibly present not only in the earliest reports avail- able in the sources but also in extant works that directly deal with it as a sub- ject worthy of investigation. In naḥw, early interest in operants (ʿawāmil) and their effect on the words they govern28 is obviously related to laḥn since most of the syntactical examples of laḥn cited in the sources have to do with the formal effect of operants on other parts of the construction. These ʿawāmil, of course, feature prominently in Sībawayhi’s syntactical analysis, and mastery of their regimen (ʿamal) is shown to be essential for the successful communica- tion between the speaker and the addressee. In the realm of luġa, treatment of laḥn took a more direct form with monographs which list common mistakes and cite attested material (i.e. šawāhid, pl. of šāhid, lit. evidence, proof) that indicates correct usage.29 The earliest extant monograph of this type is Kisāʾīs (d. 189/805) Mā talḥan fīhi l-ʿāmma, and similar titles are ascribed in the sources to most of his renowned contemporaries, such as Farrāʾ (d. 207/822), Abū ʿUbayda (d. 209/824), Aṣmaʿī (d. 216/831), and Abū ʿUbayd (d. 224/838). In both disciplines, the shifts that took place in the language, mainly following the conquests, sharpened the awareness of scholars for the need to preserve what they considered to be correct usage and to systematize the available material, either by describing the morphological and syntactical characteris- tics of Arabic or by defining the lexical meanings of words as attested in reli- able usage. As we shall see later, it is within this Arabic-Islamic context, rather than through foreign influence, that the roots of Arabic lexicography should be traced.

a letter addressed to ʿUmar (Abū l-Ṭayyib, Marātib 23; Ibn Ǧinnī, Ḫaṣāʾiṣ II, 8). Two other Companions, Ibn ʿAbbās (d. 68/687) and Ibn ʿUmar (d. 73/692), are reported to have had the habit of beating their children in case they committed laḥn (Abū Ṭāhir, Aḫbār 43; Ibn al-Anbārī, Aḍdād 244).

28 Cf. Ibn Sallām, Ṭabaqāt I, 12; Zubaydī, Ṭabaqāt 11; Ibn al-Anbārī, Nuzha 18; Suyūṭī, Aḫbār 163.

29 See Chapter 2, Section 6 below.

 

 

7early lexicographical activity

2 The Speech of the Bedouins (Aʿrāb)

In determining correct usage deemed to be free of laḥn, as well as in trying to understand the meaning of certain words described as ġarīb (strange), whether in the Qurʾān or poetry, the philologists or luġawiyyūn resorted to the Bedouins as the most reliable source for what is considered to be the “purest” form of Arabic. The data collected by the philologists who came into contact with these Bedouins formed much of the raw material of Arabic lexica, and it is therefore imperative to determine who these Bedouins were and what char- acterized them socially and linguistically.

The reliable Bedouins referred to by the philologists are consistently described as fuṣaḥāʾ (pl. of faṣīḥ, eloquent) on the grounds that their language is characterized by purity, clarity, precision and freedom from error. From the manner in which they are portrayed in the sources, the Bedouins are further characterized by a number of traits which are highlighted by the luġawiyyūn. Prominent among these are the following:

1. They were desert dwellers whose areas of residence were not only uninhab- ited by non-Arabs but also far from their “corrupting” presence. The accounts (aḫbār) suggest that this was of prime importance to the philologists and that the Bedouins themselves often boasted the fact that they did not intermingle with non-Arabs. One report has it that Abū ʿAmr b. al-ʿAlāʾ (d. 154/770) met in Mecca a Bedouin (Aʿrābī) from the Nahd clan of Asad who resided in ʿUmān. Obviously impressed by his speech, Abū ʿAmr asked the Bedouin about the secret of his eloquence ( faṣāḥa). The Bedouin responded by saying that his people resided in an area in which the rumbling of waves was not heard (innā sakannā quṭran lā nasmaʿ fīhi nāǧiḫat al-tayyār).30 He then proceeded, in a highly stylish manner, to describe the barren nature of his homeland, but was interrupted by Abū ʿAmr, who was eager to verify that the Bedouins there were entirely dependent on palm trees and camels. The concept of secluded areas is also the basis of a unique text by Abū Naṣr al-Fārābī (d. 339/950) in his K. al-Ḥurūf (also known as K. al-Alfāẓ wa-l-ḥurūf). The text is also preserved in al-Muzhir, and Suyūṭī (d. 911/1505) seems either to have depended on a dif- ferent version of Fārābī’s book or, more likely, to have expanded Fārābī’s text.31 According to the expanded version of the text, most of the material which the philologists recorded and set as a model for emulation (wa-llaḏīna ʿanhum nuqilat al-ʿArabiyya wa-bihim uqtudiya) derived from the tribes of Qays, Tamīm and Asad, and to a lesser extent Huḏayl and parts of Kināna and of Ṭayyiʾ, to

30 Qālī, Amālī (Ḏayl) III, 16. 31 Fārābī, Ḥurūf 147; Suyūṭī, Muzhir I, 211–212. See also Larcher (2006).

 

 

8 chapter 1

the exclusion of the other tribes, settled Arabs, and those tribes whose land boarded that of other nations. Accordingly, those whose dialects were left out include Laḫm and Ǧuḏām due to their proximity to the Egyptians (Qibṭ), and the people of Yemen who intermingled with the Indians and the Abyssinians. Quite significantly, one of the inadmissible dialects specified in Suyūṭī’s ver- sion of Fārābī’s text is that of Azd ʿUmān’s inasmuch as they resided in Bahrain and intermixed with the Indians and the Persians. This stands in sharp con- trast to the case of the Bedouin who resided in ʿUmān and who Abū ʿAmr b. al-ʿAlāʾ met in Mecca, and the conclusion to be drawn is that eloquence is not merely tribe-related but is also conditional upon where the informant resides, since close proximity to non-Arabs is seen as a corrupting factor which ought to be taken into account if the purest form of Arabic is sought after.

There is an obvious discrepancy, however, in the case of Qurayš. Other than the main reason cited for their faṣāḥa, i.e. that God chose Prophet Muḥammad from their midst, the philologists unanimously assert (aǧmaʿa ʿ ulamāʾunā) that the Qurašiyyūn, by virtue of Mecca being the meeting place of the various Arab tribes particularly during the pilgrimage season, were in a position to select and adopt the best dialects and the purest forms of speech of the various other tribes (aḥsan luġātihim wa-aṣfā kalāmihim), and by integrating these into their own intuitive capacities (ilā naḥāʾizihim wa-salāʾiqihim allatī ṭubiʿū ʿalayhā), they surely became the most eloquent Arabs (afṣaḥ al-ʿArab).32 Contrary to the tribes which intermixed with non-Arabs and supposedly lost their faṣāḥa, the faṣāḥa of the Qurašiyyūn is partly ascribed to the very fact that they were not in seclusion from other speech communities. Authors who represented this point of view were indeed aware of the inconsistency in applying their own criterion of seclusion, for Qurayš ought to have acquired from the other Arabs usage that qualifies as faṣīḥ as well as usage that does not. The ingenious but quite farfetched solution which the authors bring forward states that, because of their refined linguistic sense, the Qurašiyyūn took over only the best of usage in the speech and poetry of the Arab tribes, deliberately avoiding “dispraised” dialects (luġāt maḏmūma), such as the ʿanʿana of Tamīm (substitution of ʿayn for hamza; e.g. ʿanna<anna),33 the kaškaša of Asad (substitution of šīn for kāf in the feminine pronominal suffix; e.g. ʿalayši<ʿalayki),34 the faḥfaḥa of Huḏayl

32 Ibn Fāris, Ṣāḥibī 52–53; Suyūṭī, Muzhir I, 210. 33 Ibn Fāris, Ṣāḥibī 53; Suyūṭī, Muzhir I, 221 and Iqtirāḥ 199. 34 Ibn Fāris, Ṣāḥibī 53; Suyūṭī, Muzhir I, 221 and Iqtirāḥ 199; cf. Ibn Durayd, Ǧamhara I, 42.

 

 

9early lexicographical activity

(substitution of ʿayn for ḥāʾ; e.g. ʿattā<ḥattā),35 and the ʿaǧʿaǧa of Quḍāʿa (sub- stitution of doubled ǧīm for doubled yāʾ; Tamīmiǧǧ<Tamīmiyy).36

The religious motives behind the ascription to Qurayš of the highest degree of faṣāḥa among the Arabs are evident. This notwithstanding, it is quite reveal- ing that in Fārābī’s previously cited text, Qurayš is not specifically mentioned among those tribes on whom the philologists depended in the collection of their data. Qurayš is also remarkably absent from the list of tribes and speech communities (e.g. Bakr b. Hawāzin, Banū Kilāb, Banū Hilāl, ʿĀliyat al-Sāfila, and Sāfilat al-ʿĀliya) specified by Abū Zayd al-Anṣārī as the only ones he would cite whenever he used the phrase qālat al-ʿArab (The Arabs say).37 Furthermore, of great interest is the ḥadīṯ: anā afṣaḥu man naṭaqa bi-l-ḍādi bayda annī min Qurayšin (I am the most eloquent among the speakers of Arabic . . . from Qurayš). The lacuna in our translation depends, of course, on the meaning of bayda anna. Any medieval or contemporary grammar book would place bayda among exceptive particles (adawāt al-istiṯnāʾ), but the grammarians insist that in this ḥadīṯ it means “because” (min aǧli) – rather than “except that” (illā) or “despite” (ʿalā) – in which case the Prophet’s eloquence becomes a result of his being a Qurašī.38 But regardless of the true meaning of bayda anna in this case, there is strong reason to conclude that the privileged position assigned to Qurayš is a tribute to its religious status by the philologists and does not faith- fully correspond to their actual dependence on the Bedouins of tribes such as Tamīm, Qays and Asad in the search for and recording of faṣīḥ usage.

2. Their faṣāḥa stems from their salīqa (natural disposition), particularly as opposed to the grammarians whose speech is described as artificial and pre- tentious. This contrast is succinctly expressed in the following unascribed line

35 Suyūṭī, Muzhir I, 222 and Iqtirāḥ 200. 36 Ibn Durayd, Ǧamhara I, 42; Suyūṭī, Muzhir I, 222 and Iqtirāḥ 201. 37 Ibn Rašīq, ʿUmda 89; Suyūṭī, Muzhir I, 151 and Iqtirāḥ 207. 38 Ibn Hišām, Muġnī I, 114; Suyūṭī, Hamʿ I, 232. It should be noted that Ibn Mālik (d. 672/1274)

and other unspecified grammarians say that bayda in this ḥadīṯ retains its exceptive sense. They surely must have realized how farfetched it would be to assign to bayda a meaning that is hardly used by the Arabs. But they still were eager to dismiss the pos- sibility that the ḥadīṯ could mean that the Prophet was the most eloquent Arab despite belonging to Qurayš and having been breast-fed (wa-sturḍiʿtu) in Banū Saʿd b. Bakr, as the rest of the ḥadīṯ indicates. Accordingly, they went as far as claiming that bayda is used in the ḥadīṯ in a stylistic manner similar to that in which ġayra is used when the speaker intends to confirm what precedes it, as in the famous line wa-lā ʿayba fīhim ġayra anna suyūfahum * bihinna fulūlun min qirāʿi l-katāʾibi (And there is no blemish in them, save that their swords have in them notches from the conflicting of the troops); translation adopted from Lane (1863–93: I, 281).

 

 

10 chapter 1

of poetry: wa-lastu bi-naḥwiyyin yalūku lisānahu * wa-lākin salīqiyyūn aqūlu fa-uʿribu39 (I am not a grammarian who rolls his tongue [with words]; rather, I am naturally disposed: [when] I speak, I speak intelligibly). A similar line is kam bayna qawmin qad iḥtālū li-manṭiqihim * wa-bayna qawmin ʿalā iʿrābihim ṭubiʿū (What a difference between folks who contrive to [manage] their speech, and others whose intelligibility is innate). The line is part of a satirical poem in which a certain ʿAmmār al-Kalbī attacks the grammarians for having criti- cized a line of his.40 In another account, al-Aḫfaš al-Awsaṭ (d. 215/830) cited ʿAmmār’s poem after he heard a Bedouin (Aʿrābī) ridicule the grammarians whose jargon he could not comprehend.41

The notion of Bedouin salīqa – or what may be referred to as salīqiyya42 – is supported by anecdotal material in the sources. Other than the stipulation that their areas of residence should be sheltered from the “corrupting” influ- ence of non-Bedouin dialects, trustworthy informants had to fulfill another condition which confirms that their salīqa is likewise “sheltered” from the influence of the nascent linguistic sciences. Ǧāḥiẓ (d. 255/869) reports that al-Rabīʿ b. ʿAbdalraḥmān al-Sulamī asked a Bedouin (Aʿrābī) whether he would pronounce the word Isrāʾīl with a (median) hamza (a-tahmiz Isrāʾīl). Ignorant of the technical term hamz (uttering of hamza) in al-Rabīʿ’s question, the Bedouin responded by saying innī iḏan la-raǧulu sawʾin (I would thus be an evil man) because he only knew the general and nontechnical meaning of hamz, i.e. squeezing or pressuring.43 Furthermore, al-Rabīʿ asked the same Bedouin a-taǧurr Filasṭīn (Would you use the genitive with Filasṭīn [Palestine]?), and he naively said innī iḏan la-qawiyyun (I would thus be a strong man),44 since he did not know the technical meaning of ǧarr (genitive). For his part, Ibn

39 Astarābāḏī, Šarḥ II, 28; Ušmūnī, Šarḥ III, 732; Ibn Manẓūr, Lisān (SLQ). 40 Ibn Ǧinnī, Ḫaṣāʾiṣ I, 239–240; Yāqūt, Muʿǧam IV, 1595. For comments on and translation of

ʿAmmār’s poem, see van Gelder (2011: 253–255). 41 Qifṭī, Inbāh II, 42–43. 42 The term is used by Ibn Ǧinnī in his Ḫaṣāʾiṣ I, 76. See also Abū Misḥal’s (Nawādir I, 342)

distinction between two types of salīqiyya or salīqa, one of which is associated with faṣīḥ speech and the other with speech characterized by laḥn and lack of faṣāḥa. Cf. also Ǧabal (1986a: 13 ff.).

43 Ǧāḥiẓ, Bayān II, 220 and Ḥayawān III, 18; Ibn Qutayba, ʿUyūn II, 157; Ibn Fāris, Ṣāḥibī 35. In another anecdote related to hamz, a Bedouin responds to the question a-tahmiz al-fa⁠ʾra (Would you utter fa⁠ʾra [rat] with a hamza?) by saying al-hirra tahmizuhā (The cat squeezes it); cf. Ibn Qutayba, ʿUyūn II, 157. Hamz, in the sense of uttering a sound force- fully, is also applied in similar accounts to unhamzated words, such as rumḥ (spear) and turs (shield); cf. Ǧāḥiẓ, Bayān II, 221–222.

44 Ǧāḥiẓ, Bayān II, 220 and Ḥayawān III, 18; Ibn Qutayba, ʿUyūn II, 157.

 

 

11early lexicographical activity

Fāris (d. 395/1004) reports that one of the fuṣaḥāʾ responded to the question li-ma naṣabta banī (Why did you use the accusative with banī [sons of]) by wondering what naṣabta means. Ibn Fāris explains that for this individual, the word naṣb means nothing other than isnād (setting up).45 One certainly finds in such accounts a deliberate effort on the philologists’ part to prove that the Bedouins were totally ignorant of linguistic jargon, in order to confirm that they were reliable informants.

In tandem with such accounts, there are several riwāyas which may well have been circulated, if not specifically designed, to prove that the Bedouins were illiterate. This, of course, would lend further support to the notion of salīqiyya which the Bedouins enjoyed since their faṣāḥa could then only be attributed to a natural disposition and not to any acquired knowledge. Trustworthy fuṣaḥāʾ were thus expected not to know the name of the Arabic letters. al-Aḫfāš (prob- ably al-Awsaṭ; d. 215/830) reports that an eloquent Bedouin (Aʿrābī faṣīḥ) was asked to recite a poem with dāl (the letter d) as its rhyme, and his immediate response was to ask what dāl meant.46 Abū Ḥayya al-Numayrī (d. c. 183/800), known for his eloquence,47 did not know what kāf meant when he was asked to recite a poem ending with that letter.48 Another poet, Ḏū l-Rumma (d. 117/735), who was famous for his eloquence and vast knowledge of Arabic – so much that his poetry alone is said to represent one third of the Arabic language49 – was keen to deny any knowledge of writing, obviously in order not to tarnish his reputation as a trustworthy source of faṣīḥ usage. On one occa- sion, he is said to have asked Ḥammād al-Rāwiya (d. 155/772) to write the letter lām. When Ḥammād was amazed to discover that Ḏū l-Rumma knew how to write, the latter immediately asked him not to tell anyone (uktum ʿalayya), and tried to justify himself by claiming that he had met someone who was cog- nizant of writing and frequented his folks in the desert and taught them the letters of the alphabet by tracing them on sand. Ḏū l-Rumma also reassured his converser that although he liked those letters and they stuck to his memory (fa-staḥsantuhā fa-ṯabatat fī qalbī), he never wrote them with his own hand.50 According to another riwāya, Ḏū l-Rumma asked ʿῙsā b. ʿUmar (d. 149/766)

45 Ibn Fāris, Ṣāḥibī 35. 46 Ibid., loc. cit. 47 Iṣfahānī, Aġānī XVI, 236. 48 Ibn Fāris, Ṣāḥibī 35. See also Lisān (QFW), where Ibn Manẓūr cites a slightly different ver-

sion of the anecdote and asserts that Abū Ḥayya’s eloquence more than compensates for his ignorance of the meaning of the letter qāf, for which he substituted kāf.

49 Ibn Diḥya, Muṭrib 206. 50 Ṣūlī, Adab 62; cf. Marzubānī, Muwaššaḥ 280.

 

 

12 chapter 1

to lift the level of a letter he had written, but when he saw that this left ʿῙsā bewildered, he hastened to ask him to keep the matter in confidence because it would be a shameful thing (ʿayb) if others knew that he was acquainted with writing.51 Confirmation of the philologists’ refusal to record data derived from individuals cognizant of writing also comes from Ǧāḥiẓ (d. 255/869), who reported that Abū l-Mufaḍḍal/l-Faḍl al-ʿAnbarī (d. ?) told ʿAlī b. Bašīr (d. ?) that he had found on the road a book and was told that it contained poetry. When ʿAlī showed interest in seeing the book provided the poetry therein was muqa- yyad (dotted and vocalized),52 Abū l-Mufaḍḍal did not recognize the technical meaning of muqayyad – which to him could only mean “bound” – and thus said wa-l-Lāh mā adrī a-muqayyad huwa am maġlūl (I absolutely do not know whether it is bound or cuffed).53 Ǧāḥiẓ skillfully summed up the purport of this anecdote as follows: Had Abū l-Mufaḍḍal known the meaning of taqyīd, he would have been discarded as a source of linguistic data (wa-law ʿarafa l-taqyīd lam yultafat ilā riwāyatihi).

3. They instinctively adhere to their dialectal usage and resist all attempts to persuade them to change it. Understandably, the philologists highlighted this trait because it provided firm theoretical grounds for their initiative for collecting data. Without it, the attribution of any usage to the Bedouins would be open to doubt, at least as far as the adherence of the informant to his dia- lect is concerned. Two of the earliest scholars best known for collecting data from the Aʿrāb, namely, ʿῙsā b. ʿUmar (d. 149/766) and Abū ʿAmr b. al-ʿAlāʾ (d. 154/770), are reported to have personally verified the fact that an Aʿrābī sticks to his dialect and rejects to alter it if asked to. The issue started when Abū ʿAmr was blamed by ʿῙsā for allowing the nominative following the excep- tive particle illā in constructions of the type laysa l-ṭību illā l-misku/a (Perfume is but musk).54 To verify Abū ʿAmr’s claim that no Ḥiǧāzī on earth would use anything but the accusative here (laysa fī l-arḍ Ḥiǧāzī illā wa-huwa yanṣib) and

51 Ibn Qutayba, Šiʿr 438. 52 The translation rests on the meaning of qayyada in the lexica. In Lisān (QYD), Ibn Manẓūr

explains taqyīd al-ḫaṭṭ as the dotting and vocalization of letters (tanqīṭuhu wa-iʿǧāmuhu wa-šakluhu). Another explanation is that muqayyad refers to poems whose lines end with quiescence (sukūn). Although the latter explanation seems appropriate when applied to poetry, it is difficult to see why ʿAlī should be specifically interested in this type of rhyme, particularly that one cannot imagine that a whole book could contain this muqayyad type to the exclusion of the muṭlaq type, which ends with long vowels and makes up the vast majority of Arabic poetry.

53 Ǧāḥiẓ, Bayān I, 163–164; II, 221. 54 Zubaydī, Ṭabaqāt 42–44; Qālī, Amālī (Ḏayl) III, 39; Suyūṭī, Muzhir II, 277–278. Cf. a similar

construction in Ǧāḥiẓ, Ḥayawān V, 309.

 

 

13early lexicographical activity

that, in contrast, no Tamīmī on earth would use anything but the nominative, the two men agreed to send two of their disciples, Ḫalaf al-Aḥmar (d. 180/796) and Yaḥyā b. al-Mubārak al-Yazīdī (d. 202/818), to Abū Mahdiyya (also known as Abū l-Mahdī; d. ?) and al-Muntaǧiʿ al-Tamīmī (d. ?) – two of the Bedouin fuṣaḥāʾ often cited in the sources. While Abū Mahdiyya, representing the Ḥiǧāzī dialect, confirmed the correctness of the accusative, Ḫalaf and Yaḥyā repeatedly used the nominative in various constructions of the same type until their informant impatiently said that this conformed neither to his usage nor to that of his people (laysa hāḏā min laḥnī wa-lā laḥn qawmī). The informant then recited three lines of his own poetry in which he shuns the use of foreign words and asserts that he would never abandon his own dialectal usage for that of others (wa-lā tārikan laḥnī li-uḥsina laḥnahum . . .). For his part, the rep- resentative of the Tamīmī usage, al-Muntaǧiʿ, refused to use the accusative no matter how hard Ḫalaf and Yaḥyā tried to teach him to use it (fa-laqqannāhu l-naṣb wa-ǧahidnā bihi fī ḏālika fa-lam yanṣib wa-abā illā l-rafʿ ). It is particu- larly interesting that this anecdote centers around the main dialectal schism or divide which both grammatical and philological sources cite, that is Ḥiǧāzī versus Tamīmī usage. Confirmation of this schism by Bedouin fuṣaḥāʾ thus gives credibility to the neatly contrastive picture presented in the sources with regard to the two major dialectal areas,55 as in the Tamīmī pronunciation (taḥqīq, nabr) of hamza versus its lightening (taḫfīf, tashīl) by the Ḥiǧāzīs;56 in the use of the nominative in the predicate of mā by the Tamīmīs in opposition to the accusative used by the Ḥiǧāzīs;57 and in the use of kasra (Tamīm) or fatḥa (Ḥiǧāz) in the pronominal prefix of the imperfect.58

In his al-Ḫaṣāʾiṣ, Ibn Ǧinnī (d. 392/1002) often refers to the refined linguis- tic sense of the Aʿrāb, one aspect of which seems to be their resistance to any attempt to make them alter their accustomed speech. He reports an amusing anecdote on the authority of Abū Ḥātim al-Siǧistānī (d. 255/869) who heard an Aʿrābī in Mecca recite the verse ṭūbā lahum wa-ḥusnu ma⁠ʾābin (blessedness for them and a beautiful place of return; Q 13: 29) by saying ṭībā, instead of ṭūbā. Abū Ḥātim repeatedly corrected the Aʿrābī, who stubbornly stuck to his read- ing of ṭībā. In an attempt to show the Aʿrābī where his pronunciation should be

55 See a list of words pronounced differently by the Tamīmīs and Ḥiǧāzīs in Suyūṭī, Muzhir II, 275–277, and a discussion of the main dialectal differences between the two groups in Ṣāliḥ (1970: 72 ff.).

56 Cf. Ibn Manẓūr, Lisān, Introduction 22. 57 The terms mā l-Tamīmiyya and mā l-Ḥiǧāziyya are used to refer to the two types of mā, as

in mā Zaydun qāʾimun/qāʾiman; cf. Ibn Hišām, Muġnī I, 303; Murādī, Ǧanā 322. 58 Cf. Sībawayhi, Kitāb IV, 110–113; Lablī, Buġya 102–103.

 

 

14 chapter 1

changed, Abū Ḥātim isolated the first syllable of the word and repeatedly said ṭū ṭū, only to hear the Aʿrābī repeat after him ṭī ṭī!59 Ibn Ǧinnī commends the Aʿrābī’s adherence to his dialect and refusal to imitate Abū Ḥātim’s pronuncia- tion (a-fa-lā tarā ilā stiʿṣām hāḏā l-Aʿrābī bi-luġatihi wa-tarkihi mutābaʿat Abī Ḥātim). What makes this anecdote even more significant is that it has to do with a Qurʾānic verse. Although the reading of ṭūbā with -ī- instead of -ū- is inadmissible because it contradicts the written form, Ibn Ǧinnī disregards this serious violation and praises the Aʿrābī, who, in spite of giving the onlooker the impression of being rude and uncivil (wa-anta taʿtaqiduhu ǧāfiyan kazzan), has such an innate feel (ṭabʿ, salīqiyya, naǧr)60 for the language that he replaces the heavy wāw (i.e. long vowel -ū-) with the lighter yāʾ (i.e. long vowel -ī-), unin- fluenced by Abū Ḥātim’s instruction (fa-lam yuʾaṯṯir fīhi l-talqīn).

4. The most eloquent Bedouins use an elevated form of Arabic, character- ized by a high degree of precision in the choice of words, an astounding abil- ity to generate rhyme, and an overwhelming disposition to use vocabulary of the ġarīb (strange) type which other speakers hardly ever use. Samples of the speech of such Bedouins, eagerly sought after by the philologists, are scattered in the sources, but a sizable portion of them is preserved by Suyūṭī (d. 911/1505) in the final section or ḫātima of his al-Muzhir,61 with the earlier sources prop- erly acknowledged. It is obvious that most of these samples are not free from, if not entirely the result of artificiality. The Bedouins undoubtedly used or were cognizant of ġarīb material, given that the philologists recorded such material directly from amongst their fuṣaḥāʾ, but lengthy samples which are constituted mostly of words of the ġarīb type, or which are made up of several short rhym- ing phrases, do not seem to have occurred naturally in speech. The philologists were surely keen to demonstrate the superiority of Bedouin speech to any other variety, and they thus tried to surround it with a halo of purity and fluency. In one of the least complicated statements, which hardly contains any ġarīb word, for example, an Aʿrābī speaks of a palm tree in the following fashion: ḥamluhā ġiḏāʾ wa-saʿafuhā ḍiyāʾ wa-ǧiḏʿuhā bināʾ wa-karabuhā ṣilāʾ wa-līfuhā rišāʾ wa-ḫūṣuhā wiʿāʾ wa-qarwuhā ināʾ (Its produce [provides] nourishment, its branches light, its trunk building material, the stalk of its leaves heat [of fire], its fiber rope, its leaves vessels, and its bottom utensils).62 Other types of statements attributed to the Aʿrāb combine rhyme and rarely used words, as

59 Ibn Ǧinnī, Ḫaṣāʾiṣ I, 76, 384; cf. Yāqūt, Muʿǧam IV, 1595; Ibn Manẓūr, Lisān (ṬYB). 60 Ibn Ǧinnī, Ḫaṣāʾiṣ I, 76; see also II, 117 for salīqiyya defined as ṭabīʿa (nature), and cf. the

term naḥīza in Yāqūt, Muʿǧam IV, 1595. 61 Suyūṭī, Muzhir II, 506–550. 62 Qālī, Amālī (Ḏayl) III, 16.

 

 

15early lexicographical activity

in the case of the well-known Umm al-Hayṯam of the Minqar tribe, a contem- porary of Abū ʿUbayda Maʿmar b. al-Muṯannā’s (d. 209/824), who, when asked what her illness was, said: kuntu waḥmā bi-l-dika fa-šahidtu ma⁠ʾduba fa-akaltu ǧubǧuba min ṣafīf hillaʿa fa-ʿtaratnī zullaḫa (I was craving [in pregnancy] for tallow, so when I attended a banquet, I ate the meat-stuffed stomach of a sliced goat and thus had an attack of back pain).63 Her addressee understandably expressed his astonishment to her, but she herself was surprised that he did not understand her words, and wondered whether people use two different types of speech, asserting that she only spoke in distinct Arabic (a-wa-li-l-nās kalāmān? wa-l-Lāh mā kallamtukum illā bi-l-ʿArabī al-faṣīḥ).

Such exaggerated statements attributed by the philologists to the Aʿrāb, how- ever, suggest exactly what Umm al-Hayṯam seems to deny – that is, the exis- tence of a type of elevated speech that is hardly understood by philologists,64 let alone other Arabs. The Bedouins who use this elevated variety can thus be confidently proclaimed by the philologists as the definitive source of faṣīḥ usage and arbiters who could settle linguistic arguments that often erupt among the philologists themselves. The natural linguistic disposition of these Aʿrāb, with which the philologists justify the occurrence of ġarīb, is nowhere more manifest than in statements attributed to minors (ṣiġār).65 Qālī (d. 356/967), for example, cites a brief statement uttered by a lad (ġulām) describing three horses in his father’s courtyard. The statement contains no more than nineteen words, fourteen of which Qālī had to explain and then adduce attested material (šawāhid) to support his explanation in four whole pages.66 Yet irrespective of whether or not the philologists modified, or even invented samples attributed to the Bedouin fuṣaḥāʾ, the data recorded in these samples, like other ġarīb data, constituted a substantial source for authors of lexica. This will become clearer when we discuss the various lexica in Chapters 2 and 3, but suffice it here to cite two examples which demonstrate that in some cases a word need only be attested in the speech of one Bedouin to find its way to some lexica. Ibn Durayd (d. 321/933) reported that an Aʿrābī appealed to an

63 Ibid., III, 69; Suyūṭī, Muzhir II, 539–540, 546. 64 Note that Umm al-Hayṯam’s interlocutor in the above quoted anecdotes is a certain ʿ Umar

b. Ḫālid al-ʿUṯmānī, who seems to have been a disciple of Abū ʿUbayda’s. 65 See the title of Suyūṭī’s final chapter of his Muzhir II, 506 (kalām fuṣaḥāʾ al-ʿArab

wa-nisāʾihim wa-ṣiġārihim wa-imāʾihim). An amazing story recorded by some biographers is that of the lad (ġulām) who frequented the allegedly first grammarian, Abū l-Aswad al-Duʾalī (d. 69/688), and was so confident of himself as to claim that one of the words he uttered was not known by Abū l-Aswad himself; cf. Abū l-Ṭayyib, Marātib 27–29; Zubaydī, Ṭabaqāt 23–24.

66 Qālī, Amālī I, 57–61; cf. Suyūṭī, Muzhir II, 524–525.

 

 

16 chapter 1

Iraqi prince for retaliation by shouting al-qiṣāṣāʾ.67 Suyūṭī (d. 911/1505) men- tions this account and describes the word qiṣāṣāʾ/quṣāṣāʾ as rare and anoma- lous (nādir šāḏḏ) since Sībawayhi does not mention the pattern fuʿālāʾ in his al-Kitāb.68 Yet, in spite of Suyūṭī’s assertion that a word attested in the speech of only one Aʿrābī should be viewed with caution, qiṣāṣāʾ (also cited as quṣāṣāʾ and qaṣāṣāʾ) is recorded in several lexica without mention of the fact that it was attested no more than once.69 The second example is that when Umm al-Hayṯam was shown some seeds by Abū Ḥātim al-Siǧistānī, who wanted to know the Arabic equivalent of their Persian name, asfīyūš, she paused then said that the Arabic name was buḫduq (in Fīrūzābādī’s al-Qāmūs, buḥduq). In spite of Abū Ḥātim’s assertion that he never heard this word from other than Umm al-Hayṯam, the word is recorded in several lexica, either with or without mention of Umm al-Hayṯam as its only source.70

3 The Collection of Data

Having specified the speech of the Bedouins as the variety which represents the “purest” form of Arabic and the one which should accordingly be the model not only to be emulated but also to be used in determining the degree of faṣāḥa and even acceptability of other varieties, it was natural that the early philolo- gists and grammarians should seek to collect linguistic data from Bedouin informants. The list of scholars who either transmitted and commented on dialectal usage or are actually reported to have made the journey to the des- ert (bādiya) to record data as part of the effort known as ǧamʿ al-luġa is quite impressive and includes almost all the major linguists from the early second/ eighth century till about the middle of third/ninth century.71 We thus learn from the sources that ʿAbdallāh b. Abī Isḥāq (d. 117/735) attributed the shift of s to ṣ to the dialect of ʿAmr b. Tamīm;72 that ʿĪsā b. ʿUmar (d. 149/766) preferred the forms summ (poison) and šuhd (honey), used in the dialect of al-ʿĀliya in

67 Ibn Durayd, Ǧamhara III, 1230. 68 Suyūṭī, Muzhir I, 254. 69 Ibn Manẓūr, Lisān (QṢṢ); Fīrūzābādī, Qāmūs (QṢṢ). Zabīdī (Tāǧ: QṢṢ) mentions that the

word is reported by Ibn Durayd and is šāḏḏ. 70 Ibn Durayd, Ǧamhara II, 1116 ; Suyūṭī, Muzhir I, 252. See also Ibn Manẓūr, Lisān (BḪDQ);

Fīrūzābādī, Qāmūs (BḤDQ); Zabīdī, Tāǧ (BḪDQ). 71 Cf. Baalbaki (2008: 24–25). 72 Zubaydī, Ṭabaqāt 32; Qifṭī, Inbāh II, 108.

 

 

17early lexicographical activity

Medina, to samm and šahd which occur in other dialects;73 that Abū ʿAmr b. al-ʿAlā’s (d. 154/770) house was filled to its roof with books containing material he collected from the fuṣaḥāʾ;74 that Ḫalīl b. Aḥmad (d. 175/791) derived his knowledge from the deserts (bawādī) of Ḥiǧāz, Naǧd and Tihāma;75 that Yūnus (d. 182/798) heard (samiʿa) the speech of the Arabs;76 that Kisāʾī (d. 189/805) depleted fifteen bottles of ink by writing data taken from the Arabs;77 that Yazīdī (d. 202/818) was expert in the dialects of the Arabs;78 that al-Naḍr b. Šumayl (d. 203/819) spent forty years in the desert and learnt from fuṣaḥāʾ such as Abū Ḫayra al-Aʿrābī and Abū l-Duqayš;79 that Abū ʿAmr al-Šaybānī (d. 206/821) went to the desert with two large vessels (distīǧān) of ink and did not depart before depleting them in recording data;80 that Farrāʾ (d. 207/822) based his study of language and grammar on the speech of the Arabs;81 that Abū Zayd al-Anṣārī (d. 215/830) extensively transmitted usage by the Bedouins;82 that Aṣmaʿī’s (d. 216/831) vast knowledge of luġa and riwāya was unmatched;83 that Abū ʿUbayd (d. 224/838) spent forty years collecting, for inclusion in his book al-Ġarīb al-muṣannaf, data taken directly from the Arabs;84 and that Ibn al-Aʿrābī (d. 231/845) acquired linguistic material from the Aʿrāb – some of whom, such as al-Ṣaqīl, ʿUǧruma and Abū l-Makārim, are mentioned by name85 – and that he reported usage he had heard from fuṣaḥāʾ al-Aʿrāb.86 Other than the mate- rial available in the general and bibliographical sources, direct evidence in the Kitāb suggests that Bedouin native speakers, to whom Sībawayhi (d. 180/796) often refers, were a valuable source for his linguistic repertoire. Although the sources do not specifically mention any journey Sībawayhi made to the desert

73 Zubaydī, Ṭabaqāt 41. Note also that ʿĪsā blamed Kisāʾī (d. 189/805) for discussing the vari- ous theoretical possibilities of a certain expression and asked him to adduce the speech of the Arabs (kalām al-ʿArab) as the only acceptable proof; cf. ibid., 42; Qifṭī, Inbāh II, 376–377.

74 Qifṭī, Inbāh IV, 133. 75 Ibn al-Anbārī, Nuzha 59; Qifṭī, Inbāh II, 258. 76 Sīrāfī, Aḫbār 33; Ibn al-Anbārī, Nuzha 47; Suyūṭī, Buġya II, 365. 77 Qifṭī, Inbāh II, 258; Yāqūt, Muʿǧam IV, 1738. 78 Qifṭī, Inbāh IV, 32; Suyūṭī, Muzhir II, 340. 79 Ibn al-Anbārī, Nuzha 37; Suyūṭī, Muzhir II, 316. 80 Ibn al-Anbārī, Nuzha 78; Qifṭī, Inbāh I, 259. 81 Zubaydī, Ṭabaqāt 131; Qifṭī, Inbāh IV, 8. 82 Zubaydī, Ṭabaqāt 166; cf. Marzubānī, Muqtabas 105. 83 Ibn al-Anbārī, Nuzha 91. 84 Qifṭī, Inbāh III, 22; Ibn Ḫillikān, Wafayāt IV, 61; cf. below, 74. 85 Abū l-Ṭayyib, Marātib 147. 86 Qifṭī, Inbāh III, 130.

 

 

18 chapter 1

in pursuit of the Aʿrāb, statements such as samiʿnā l-ʿArab al-fuṣaḥāʾ (We heard the eloquent Arabs), samiʿnā l-ʿArab al-mawṯūq bihim (We heard the trustwor- thy Arabs), qawm min al-ʿArab turḍā ʿArabiyyatuhum (Arabs whose Arabic is reliable), and min afwāh al-ʿArab (from the mouths of the Arabs)87 prove that he was in direct contact with Bedouin informants. On one occasion, he identi- fies one of these informants, an Aʿrābī called Abū Murhib.88 Sībawayhi’s Kufan contemporary, Farrāʾ, also reports material he heard directly from the Aʿrāb or through his teacher Kisāʾī.89

As the Aʿrāb realized how much they were sought after by the philologists and grammarians, many of them left the bādiya (desert) and joined the ḥāḍira (settled area; city). One can thus talk of a counter-journey by virtue of which the Aʿrāb brought their linguistic expertise to the cities in which linguistic scholar- ship thrived, most notably Basra and Baghdad, and, to a lesser extent, Kufa. The numbers of these Aʿrāb cannot be accurately determined, but it seems that there were tens or hundreds, judging by the lists available in a few sources. The earliest lists are those compiled by two contemporary authors, Ibn al-Nadīm (d. 380/990) and Marzubānī (d. 384/994). In a section devoted to the famous elo- quent Aʿrāb whose speech was heard by scholars ( fuṣaḥāʾ al-Aʿrāb al-mušharīn allaḏīna samiʿa minhum al-ʿulamāʾ), Ibn al-Nadīm mentions twenty-one names and then adds another forty-nine under different headings.90 For his part, Marzubānī alphabetically lists at the end of his Muʿǧam al-Šuʿarāʾ three hundred and fifty-three names (a few of which are common to both lists), but since these include two categories, namely, unknown poets and unrenowned Bedouins (min al-šuʿarāʾ al-maǧhūlīn wa-l-Aʿrāb al-maġmūrīn),91 it is impossi- ble to determine with any precision the number of Aʿrāb mentioned, although some names (e.g. Abū Salhab al-Fārisī and Abū Saʿd al-Aṣbahānī) can easily be excluded in any enumeration of Bedouin fuṣaḥāʾ. It should be remembered however that Marzubānī’s list is restricted to individuals who are known by their kunya (i.e. form of name beginning with Abū or Umm), and thus the number of the Bedouin fuṣaḥāʾ consulted by the philologists must have been considerably larger. Almost three centuries later, Qifṭī (d. 646/1248) follows in the footsteps of Marzubānī and ends his voluminous biographical dictionary Inbāh al-ruwāt with an extensive list of Aʿrāb who came to the ḥāḍira.92 The list

87 Sībawayhi, Kitāb I, 219; II, 336; I, 182; II, 74 respectively. 88 Ibid., I, 328. 89 Farrāʾ, Maʿānī I, 23, 209; II, 223, 312; III, 229, 246; cf. Kinberg (1996: 469–470). 90 Ibn al-Nadīm, Fihrist 49–55. See also Ḥāǧǧ Ṣāliḥ (2007: 366–369). 91 Marzubānī, Muʿǧam 507–515. 92 Qifṭī, Inbāh IV, 120–190.

 

 

19early lexicographical activity

comprises one hundred and three individuals, but again some of them (such as Abū l-Fatḥ b. al-Muqaddar al-Aṣbahānī al-Naḥwī) are obviously not Aʿrāb. The biographies of several other Bedouin fuṣaḥāʾ are also included before this list.93

Other than individual names, the sources at times mention groups of Aʿrāb who made the counter-journey from the desert to the settled areas and who constituted a mine of linguistic data for the philologists. A resident of Basra and Baghdad, Abū ʿUbayda (d. 209/824), for example, confirms that he fre- quented a number of Bedouins who came from the desert area of Ǧaʿfar b. Ǧaʿfar b. Kilāb and recorded ( fa-naktub ʿanhum) their recitation of poetry pre- dating Umruʾ al-Qays.94 Similarly, Aṣmaʿī (d. 216/831) is reported to have read to Abū ʿAmr b. al-ʿAlāʾ (d. 154/770) the data of ġarīb material he had collected from those informants he met in Mirbad (i.e. of Basra).95 He is also said to have frequented the Aʿrāb in Basra so much that they became familiar with him and knew what his purpose was (aktub ʿanhum kaṯīran ḥattā alifūnī wa-ʿarafū murādī).96 The influence of such Aʿrāb seems to have gone beyond the circles of philologists, and we are told that the famous poet Baššār b. Burd (d. 167/784) used to recite some of his panegyric poetry (most probably for ensuring its linguistic soundness) to a group of eloquent Aʿrāb from the Qays ʿAylān tribe who resided for a while on the outskirts of Basra.97 Held in such high esteem, the Bedouin fuṣaḥāʾ became arbiters in linguistic controversies, as in the pre- viously mentioned anecdote, according to which both ʿῙsā b. ʿUmar and Abū ʿAmr b. al-ʿAlāʾ sent two of their disciples to Abū Mahdiyya and al-Muntaǧiʿ for determining the correct case ending of the noun which follows the exceptive particle in constructions of the type laysa l-ṭību illā l-misku/a.98 A more famous controversy in which the input of the Bedouins was sought is that known as al-masʾala al-zunbūriyya, which reportedly took place between Sībawayhi (d. 180/796) and Kisāʾī (d. 189/805). Having differed on whether to use the nominative pronoun hiya or the accusative pronoun iyyāhā in the construc- tion kuntu aẓunnu l-ʿaqraba ašadda lasʿatan min al-zunbūri fa-iḏā huwa hiya/ iyyāhā (I used to think that the sting of a scorpion was more intense than that of a wasp, but [I discovered that] it was the same), the Aʿrāb who were nearby

93 E.g. Abū l-Baydāʾ (ibid., IV, 102), Abū Ṯuwāba (IV, 104), Abū Ṯarwān (IV, 105), and Abū Ḫayra (IV, 117).

94 Qurašī, Ǧamhara 24. 95 Qālī, Amālī (Ḏayl) III, 182. 96 Suyūṭī, Muzhir II, 307–308. 97 Iṣfahānī, Aġānī III, 200–201. 98 Cf. above, 12.

 

 

20 chapter 1

(ʿalā l-bāb) were consulted to determine the correct usage.99 Irrespective of the outcome of this consultation and of the later accusation leveled at the Ḥuṭama Aʿrāb by Sībawayhi’s Basran compatriots of unfairly siding with Kisāʾī as their alleged patron, it is quite telling that when the attendees suggested that the Aʿrāb be consulted, neither Sībawayhi nor Kisāʾī objected. In one of the versions of this anecdote, Kisāʾī is portrayed as the one who suggested to Sībawayhi that the Arabs be consulted specifically for being easily available: hāḏihi l-ʿArab bi-bābika qad ǧamaʿtahum min kull awb wa-wafadat ʿalayka min kull ṣuqʿ (The Arabs are at your doorstep; you have gathered them from everywhere and they have come to you from all regions). Among those who were readily available to arbitrate in this case were Abū Faqʿas, Abū Diṯār, Abū l-Ǧarrāḥ and Abū Ṯarwān, all of whom are listed by Ibn al-Nadīm100 and others among the Bedouin fuṣaḥāʾ, and some of whom are often cited in the sources in connection with other linguistic matters.101

The data which the philologists recorded on the authority of the Bedouin fuṣaḥāʾ provided much of the raw material for the early monographs that dealt with ġarīb and nawādir or with specific semantic fields. The sought- after fuṣaḥāʾ themselves were also active participants in such early writings, and in fact the boundaries between the philologists who collected linguistic data from their Bedouin ruwāt (informers, narrators) and those ruwāt are often blurred. Ibn al-Nadīm, for example, mentions in his list of Bedouin fuṣaḥāʾ a certain Qurayba Umm al-Buhlūl al-Asadiyya and ascribes to her a book entitled al-Nawādir wa-l-maṣādir.102 Similarly, Abū ʿAlī al-Ḥirmāzī, a Bedouin rāwiya who came to Basra, authored a book entitled Ḫalq al-insān,103 and Abū Ḫayra, a Bedouin of the ʿAdiyy tribe, authored K. al-Ḥašarāt104 and K. al-Ṣifāt.105 One of the extant sources of this type is K. al-Nawādir, authored by Abū Misḥal al-Aʿrābī (d. 231/845), a Bedouin who came to Baghdad106 and is

99 Zubaydī, Ṭabaqāt 68–71; Qifṭī, Inbāh II, 358–359. 100 Ibn al-Nadīm, Fihrist 53, 57. 101 Whereas Abū Faqʿas and Abū Diṯār are mentioned only in connection with the Zunbūriyya

question, Abū Ṯarwān and Abū l-Ǧarrāḥ are often cited in other sources in connec- tion with poetry material and ġarīb data. For Abū Ṯarwān, cf. Farrāʾ, Maʿānī I, 4, 139; Ibn al-Sikkīt, Iṣlāḥ 133, 213; Qālī, Amālī I, 211; and for Abū l-Ǧarrāḥ, cf. Ibn al-Sikkīt, Iṣlāḥ 9, 104, 133, 215; Ṯaʿlab, Maǧālis I, 101, 199. For more on the role of the Bedouins as arbiters in controversial linguistic issues, see Blau (1963: 42–51).

102 Ibn al-Nadīm, Fihrist 53. 103 Ibid., 54. 104 Ibid., 51; Yāqūt, Muʿǧam VI, 2761; Suyūṭī, Buġya II, 317. 105 Qifṭī, Inbāh I, 144. 106 Ibid., II, 218.

 

 

21early lexicographical activity

cited by several philologists, although he himself reports dialectal variants and ġarīb material on the authority of Kisāʾī and Farrāʾ.107 It is therefore not surpris- ing that Zubaydī (d. 379/989) lists in his section devoted to Basran philologists (al-luġawiyyūn al-Baṣriyyūn) three of the well-known Bedouin fuṣaḥāʾ who are often quoted in the sources, namely, al-Muntaǧiʿ al-Aʿrābī, Abū Mahdiyya al-Aʿrābī and Abū Mālik al-Aʿrābī.108 Furthermore, the fact that some Bedouin fuṣaḥāʾ became philologists and scholars in their own right is supported by reports that they became teachers by profession. One such individual is Abū l-Baydāʾ al-Riyāḥī, an Aʿrābī who came to Basra where he taught boys for a fee (wa-kāna yuʿallim al-ṣibyān bi-uǧra) and spent his life disseminating knowl- edge (aqāma bihā ayyām ʿumrihi yuʾḫaḏ ʿanhu l-ʿilm).109 Some philologists are even said to have owed their knowledge and fame to their contact with the Bedouins who came to settled areas. Ibn al-Aʿrābī’s (d. 231/845) scholarship, for example, is said to have been quite modest (lam yazal . . . murmidan fī ʿilmihi) until he acquired ġarīb material from the Aʿrāb of Yamāma who came to Kufa.110 The focus on ġarīb (for which see Section 5 below) by the philologists of the early period demonstrates the central role of the Bedouin fuṣaḥāʾ in the pro- cess of data collection since they were almost exclusively the source of ġarīb material which the philologists dearly treasured.

Yet the counter-journey of the Bedouin fuṣaḥāʾ to the settled areas in which most philologists resided and the intermingling of the two groups as shown above came at a price, for some of these Bedouins were soon accused of losing their faṣāḥa due to their lengthy stay in settled areas away from the bādiya. Already in the first half of the third/ninth century, Ǧāḥiẓ (d. 255/869) asserted that the grammarians (naḥwiyyūn) did not rely on the speech of any Aʿrābī who understood faulty constructions since this means that he got used to hearing them, and this is proof that the length of his stay away from his linguistically sheltered area of origin resulted in corrupting his speech and diminishing his eloquence (tufsid al-luġa wa-tunqiṣ al-bayān).111 From first-hand experience, Ǧāḥiẓ talks about the wide disparity (bawn baʿīd) in the speech of Yazīd b. Kaṯwa from the time he came to Basra to the day he died. Although Ǧāḥiẓ does not specify the duration of Yazīd’s stay in Basra, the context clearly indicates that his stay away from his native bādiya – in a location which Ǧāḥiẓ describes as being on the borderline between the area of faṣāḥa (i.e. desert) and that of

107 Abū Misḥal, Nawādir 52, 86, 128, 144, 154, 247, 251, 489, 504, etc. 108 Zubaydī, Ṭabaqāt 157. 109 Ibn al-Nadīm, Fihrist 49; cf. Ǧāḥiẓ, Bayān I, 252. 110 Zubaydī, Ṭabaqāt 196. 111 Ǧāḥiẓ, Bayān I, 162–163.

 

 

22 chapter 1

ʿuǧma (lit. barbarousness; i.e. settled areas) – was long enough to cause a dis- cernible deterioration in the purity of his speech. One of the Bedouin fuṣaḥāʾ whose name appears in the previously mentioned lists of Ibn al-Nadīm, Marzubānī and Qifṭī,112 namely, Nahšal b. Zayd, better known as Abū Ḫayra al-Aʿrābī, was also accused of having been adversely affected by his lengthy stay in Basra. Asked by Abū ʿAmr b. al-ʿAlāʾ (d. 154/770) whether he would say ista⁠ʾṣala l-Lāhu ʿirqātihim or ʿirqātahum (God utterly destroyed their race), Abū Ḫayra chose the latter. For Abū ʿAmr, that was the wrong answer, and so he said to Abū Ḫayra lāna ǧilduka (lit. Your skin has softened), by which he meant that he had lost the harshness of the Bedouins and with it his faṣāḥa.113 Regardless of the controversy concerning the correctness of ʿirqātahum,114 and consequently of the rightness of Abū ʿAmr’s assumption that Abū Ḫayra made an error, it is significant that Abū ʿAmr made a link between sedentary life and the corruption of speech. More significant, however, if the anecdote is to be accepted as true, is that it took place in the middle of the second/eighth cen- tury at the latest, that is, in the period during which the speech of the Aʿrāb was taken to be trustworthy. As we shall see in Section 4 below, the time framework set by the philologists for acceptable linguistic material was of vital impor- tance for determining the data which they admitted to their lexica.

The charge of acquiring corrupt speech due to lengthy absence from the bādiya did not even spare some poets, most notably those known for their faṣāḥa. One such poet is Ḏū l-Rumma (d. 117/735), whose vast knowledge of Arabic has been referred to earlier.115 His use of the word zawǧa (wife) instead of zawǧ in one of his poems was criticized by Aṣmaʿī (d. 216/813), who claimed that this was due to his recurrent visits to sedentary areas. This was wittily put by Aṣmaʿī by referring to Ḏū l-Rumma as having suffered indigestion from eat- ing tender plants and pickles in grocers’ shops (qad akala l-baql wa-l-mamlūḥ

112 Ibn al-Nadīm, Fihrist 51; Marzubānī, Muʿǧam 509; Qifṭī, Inbāh IV, 117; cf. Ibn al-Anbārī, Nuzha 32.

113 Zaǧǧāǧī, Maǧālis 6; Ibn Ǧinnī, Ḫaṣāʾiṣ I, 384; Qifṭī, Inbāh IV, 117–118. In another anecdote reported by Ibn Ǧinnī (Ḫaṣāʾiṣ III, 305), Abū Ḫayra was even blamed for corrupting the speech of his mother (afsadaki bnuki) who used the expression ġumiya ʿalā l-marīḍi (The sick person fainted) instead of uġmiya ʿalā l-marīḍi. Coming from a rival Bedouin, al-Muntaǧiʿ, such an accusation is most probably a gross exaggeration.

114 Māzinī (d. 249/863), for example, claims that ʿirqātahum is correct, based on the assump- tion that ʿirqāt is singular, in contrast to ʿirqātihim, which would be the plural of ʿirq and hence a sound feminine plural whose accusative takes a kasra; cf. Zaǧǧāǧī, Maǧālis 6. For another interpretation of Abū ʿAmr’s response to Abū Ḫayra, see Ibn Ǧinnī, Ḫaṣāʾiṣ I, 384.

115 See above, 11.

 

 

23early lexicographical activity

fī ḥawānīt al-baqqālīn ḥattā bašima).116 al-Aṣmaʿī was even more critical of another poet famous for his faṣāḥa, namely, the Kufan al-Kumayt al-Asadī (d. 126/744), who he accused of being less like a Bedouin than Ḏū l-Rumma.117 Furthermore, a charge of lack of faṣāḥa is sometimes addressed against a whole group of Aʿrāb who came to reside in urban areas. One such group is the Aʿrāb of Ubulla, a town on the banks of the river Diǧla, close to Basra.118 Kisāʾī’s (d. 189/805) knowledge of Arabic is said to have become corrupt due to inter- mingling with these Aʿrāb although, as an apprentice, he had spent seventeen years with Abū ʿAmr b. al-ʿAlāʾ (d. 154/770), one of the most respectable Basran scholars of the time.119 In a different account, the Ḥuṭama Aʿrāb of Baghdad are alleged to be the reason for Kisāʾī’s unsound knowledge since he acquired their corrupt usage ( fa-aḫaḏa ʿanhum al-fasād min al-ḫaṭa⁠ʾ wa-l-laḥn), and thus spoiled what he had already acquired in Basra at the hands of Abū ʿAmr b. al-ʿAlāʾ, Yūnus b. Ḥabīb (d. 182/798) and ʿIsā b. ʿUmar (d. 149/766).120 Such accounts are not free from partisan motives and are most probably Basran alle- gations against a renowned Kufan scholar such as Kisāʾī. Yet what interests us here is the mere fact that during the period in which the collection of linguistic data was at its peak (i.e. the second half of the second/eighth century, or pos- sibly earlier as Abū ʿAmr’s story with Abū Ḫayra suggests), some scholars seem to have realized that they had to deal critically with the material available to them, and that there were different levels of faṣāḥa, even among the Aʿrāb, with which to contend.

The introduction of isnād, or ascription of data to its ultimate source often by citing a continuous chain of transmitters, is directly linked with the need to demonstrate the trustworthiness of the reported material. Already in the works of authors from the latter part of the second/eighth and first half of the third/ ninth centuries, such as Sībawayhi, Kisāʾī, Abū ʿ Ubayda and Abū Misḥal, data is often, but not regularly, attributed to the direct source person from whom the author heard it. With time, the chain of transmitters naturally grew longer and data’s inclusion in the beginning of the riwāya became more frequent. This is clearly the case, for example, in Ṯaʿlab’s (d. 291/904) al-Maǧālis, where dialectal variants and ġarīb material are often introduced by a chain which includes

116 Zaǧǧāǧī, Maǧālis 150; Marzubānī, Muwaššaḥ 284; Ibn Ǧinnī, Ḫaṣāʾiṣ III, 295. According to Aṣmaʿī, Ḏū l-Rumma mostly frequented Basra and Yamāma (Marzubānī, Muwaššaḥ 272). Cf. also Yāqūt, Muʿǧam II, 815 for another usage by Ḏū l-Rumma criticized by Aṣmaʿī.

117 Marzubānī, Muwaššaḥ 272. 118 Yāqūt, Buldān I, 77. 119 Suyūṭī, Iqtirāḥ 206. 120 Qifṭī, Inbāh II, 274; Yāqūt, Muʿǧam IV, 1743–1744.

 

 

24 chapter 1

several names. The issues related to isnād, such as its validity and develop- ment, do not directly concern us here. It should be noted, however, that its introduction to linguistic sciences is probably a borrowing from the scholars of Ḥadīṯ (prophetic tradition), and this further demonstrates the interconnected- ness of the religious and linguistic domains of study in the formative period of both. In general, the use of isnād was not always considered to be a require- ment by the philologists, contrary to Ḥadīṯ scholars who painstakingly devel- oped methods of criticism in which isnād occupies a central part. Although one sometimes finds linguistic data, as well as anecdotal material related to linguistic issues, introduced by full isnād, much of the data escapes such scru- tiny, and instead, incomplete chains of transmitters, known as isnād munqaṭiʿ, are introduced. The reason for this discrepancy between Ḥadīṯ and philology is certainly due to the different nature of the two domains, since the presence of various motives for fabrication of Ḥadīṯ is hardly applicable to philology. The intention of those who fabricated Ḥadīṯ, for example, was: to astonish the common people in order to receive financial reward; spread false doctrines; promote heretical teachings; or to confront the laxity of their times by invent- ing traditions which would exhort men to live righteously.121 Unlike in Ḥadīṯ, it is difficult to see how the invention of linguistic data or its false ascription to a certain user or tribe could benefit the fabricator, or be the result of politi- cal or religious agendas. Yet, there is evidence that some linguistic material was indeed the result of fabrication, possibly in an attempt by some to claim knowledge or to avoid the embarrassment of admitting ignorance when asked about a certain usage or meaning.

To begin with, it is important to note that charges of fabrication did not spare some of the most authoritative scholars of the second/eighth and third/ ninth centuries. On the authority of Abū Ḥātim al-Siǧistānī (d. 255/869), for example, the renowned Kufan grammarian, Kisāʾī (d. 189/805), used to instruct the Aʿrāb to utter what suited his needs ( yulaqqinuhum mā yurīd), and hence the material he transmitted from those Aʿrāb is described as dis- carded (maṭrūḥa), although he was the most knowledgeable of the Kufans in language and Qurʾān.122 Also, Ibn l-Aʿrābī (d. 231/845), according to Ṯaʿlab (d. 291/904), did not accept a certain word reported by another renowned scholar, Aṣmaʿī (d. 216/831), claiming that he had heard from a thousand Aʿrābī

121 Cf. Robson (1986: 24–25). For a detailed study of Ḥadīṯ fabrication, cf. Juynboll (1983), esp. Chapter 3.

122 Abū l-Ṭayyib, Marātib 120; text also in Yāqūt, Muʿǧam IV, 1746–1747.

 

 

25early lexicographical activity

what would contradict Aṣmaʿī’s report.123 For his part, Sībawayhi (d. 180/796) was not accused of fabrication, but of falling victim to fabricators, one of whom, Abān al-Lāḥiqī (d. 200/815), is said to have told Māzinī (d. 249/863) that when Sībawayhi asked him whether he knew of any attested usage in which the intensive form of the pattern faʿil took a direct object, he invented for him a line of poetry which contained the expression ḥaḏirun umūran (very cautious of things) and which Sībawayhi included in his al-Kitāb.124 In addition to this line, the fifty unascribed lines, known as al-abyāt al-ḫamsūn in Sībawayhi’s al-Kitāb, are considered by some critics to have been part of poetry deliber- ately invented by “post-classical” (muwalladūn) poets to deceive the masters of the field (dassūhā ʿalā l-a⁠ʾimma).125 Although it is difficult to believe that, with all Sībawayhi’s scrupulousness in reporting usage, be it prose or poetry,126 he could have inadvertently included fifty lines which he did not recognize as forgery, the fact remains that third/ninth century scholars such as Abū Ḥātim al-Siǧistānī, Ibn al-Aʿrābī and Māzinī were mindful of the fact that some fabri- cated material might have found its way to the works of the a⁠ʾimma, and subse- quently the later generations who largely depended on their masters.127

Allegations of forgery also spread to the very source which the philolo- gists held in great esteem, namely, the Aʿrāb. In the previously quoted masʾala zunbūriyya,128 for example, the Ḥuṭama Aʿrāb were accused of false testi- mony in order to prove Kisāʾī right in his controversy with Sībawayhi. Another example which could indicate fabrication is the word buḫduq (a kind of seed), whose only source, as mentioned earlier,129 is Umm al-Hayṯam, and which appears in several lexica based solely on her testimony. There certainly is room

123 Yāqūt, Muʿǧam VI, 2531. See also Abū l-Ṭayyib’s (Marātib 84) defense of Aṣmaʿī against allegations of disseminating lies among the Aʿrāb made by his nephew ʿAbdalraḥmān b. ʿAbdallāh (d. ?), better known as Ibn Aḫī l-Aṣmaʿī; cf. Qifṭī, Inbāh II, 161.

124 Suyūṭī, Muzhir I, 180. The line in question is ḥaḏirun umūran lā tuḍīru wa-āminun * mā laysa munǧiyahū min al-aqdāri (Very cautious of things not to be feared, and trusting in that which will not save him from the decrees of destiny; translation from Lane [1863–93: II, 534]); cf. Sībawayhi, Kitāb I, 113.

125 Suyūṭī, Iqtirāḥ 60. 126 Cf. Baalbaki (2008: 25–26; 35–47). See also Ḥadīṯī (1974: 110–112) for a refutation of the

allegations that the Kitāb contains fabricated lines of poetry. 127 One example is Sībawayhi’s student, Muḥammad b. al-Mustanīr (d. 206/821), better known

as Quṭrub. According to Ibn al-Sikkīt (d. 244/858), he recorded on Quṭrub’s authority the equivalent of a bookcase (katabtu ʿan Quṭrub qimaṭran) only to discover that he lied in his linguistic data; cf. Marzubānī, Muqtabas 178; Yāqūt, Muʿǧam VI, 2646; Suyūṭī, Buġya I, 243.

128 Cf. above, 19. 129 Cf. above, 16.

 

 

26 chapter 1

for suspected fabrication here, as Abū Ḥātim reported that Umm al-Hayṯam pondered for a good while before naming the seed which he showed to her ( fa-afkarat sāʿatan ṯumma qālat).130 Other words were sometimes easier to detect as forgery and hence not be admitted to the corpus. The word muḫfanʿil, reported by a man who visited Abū Ḥātim al-Siǧistānī, is one such example. Abū Ḥātim is said to have laughed heartily when that man boasted that Abū Ḥātim did not know the word which described someone who wore a ḫuff (boot) in one foot and a naʿl (sandal) in another, hence muḫfanʿil.131 Accordingly, neither this active participle nor its supposed verbal form *iḫfanʿala appears in any of the lexica. Yet, several anecdotes suggest that some philologists were trapped into providing fictitious meanings to words which were specifically devised to test their readiness to do so. Mubarrad (d. 285/898) is said to have fallen victim to a number of students who were scanning a line of poetry which included the construction fa-stabqi baʿḍanā (so spare some of us) and whose last foot of the pattern mafāʿilu is qibaʿḍanā. From this meaningless blend of -qi and baʿḍ-, they extracted the inexistent word *qibaʿḍ and conspired to ask Mubarrad what it meant. Unwary of their plot, Mubarrad unhesitatingly responsed that qibaʿḍ is a synonym for quṭn (cotton), and even produced a hemistich in which the word is allegedly used.132 A similar anecdote involves Muḥammad b. ʿAbdalwāḥid al-Zāhid, nicknamed Ġulām Ṯaʿlab (d. 345/957). A group of stu- dents on their way to him reportedly passed by a qanṭara (arch) and agreed to spell the word backwards – yielding an inexistent word, *haraṭnaq. When they asked him what it meant, Muḥammad promptly gave them an answer; they secretly laughed with one another and left.133

The preparedness of some scholars and informants to forge words or pro- vide meanings for words they had never heard naturally caused some lexicog- raphers to be cautious about material that was not widely reported or was not supported by actual, and preferably verifiable usage. Ibn Durayd (d. 321/933) in his al-Ǧamhara, for example, very frequently comments on a word by saying laysa bi-ṯabt (not certain or established) or lā aḥuqquhu/ lā adrī mā ṣiḥḥatuhu (I cannot confirm its correctness). On one occasion, Ibn Durayd declared that he was cautious of a word transmitted by Aṣmaʿī and reported that Aṣmaʿī’s nephew also doubted whether his uncle had actually heard it.134 Suyūṭī

130 Ibn Durayd, Ǧamhara II, 1116. 131 Abū l-Ṭayyib, Marātib 132. 132 Ibn al-Anbārī, Nuzha 166; Yāqūt, Muʿǧam VI, 2679. 133 Qifṭī, Inbāh III, 172. 134 Ibn Durayd, Ǧamhara II, 1134; III, 1233, 1279. The word in question is ǧilḥiẓāʾ, which Ibn

Durayd twice explains as an adjective for a treeless land, but a third time as an adjective for a land full of trees.

 

 

27early lexicographical activity

(d. 911/1505) noticed this tendency towards caution and quoted dozens of examples in which such expressions occured in the Ǧamhara and a few other lexica, such as Abū ʿUbayd’s (d. 224/838) al-Ġarīb al-muṣannaf, Ibn Fāris’s (d. 395/1004) al-Muǧmal, and Ǧawharī’s (d. c. 400/1010) al-Ṣaḥāḥ.135 This not- withstanding, the lexicographers did little, compared with Ḥadīṯ scholars, to develop truly rigorous methods of data criticism, and this was most probably due to the fact that, other than prestigiously claiming knowledge or avoiding admission of ignorance, transmitters of linguistic data had very little to gain by forging data.

Forgery of linguistic material is perhaps comparable to forgery of poems in that the motives for both are different from those in Ḥadīṯ forgery, and cer- tainly less varied. This could partly explain why in poetry, as in philology, tools of criticism are not as developed as in Ḥadīṯ, and there is no tradition of ǧarḥ wa-taʿdīl (disparaging and declaring trustworthy), whereby the transmitters’ credentials are checked.136 What the three domains have in common, how- ever, is that forgers normally try to make sure that their product is as close to the “real thing” as possible. In poetry, some early scholars found it extremely difficult to distinguish between forged and authentic material. According to al-Mufaḍḍal al-Ḍabbī (d. 168?/784?), only an expert critic (ʿālim nāqid) – a cat- egory of scholars very difficult to come by (wa-ayna ḏālika) – is able to detect forgeries by someone as knowledgeable about the dialects of the Arabs and their poetry as the well-known Kufan forger, Ḥammād b. Sābūr al-Rāwiya (d. 155/772).137 This seems to have persuaded some to give little importance to whether the poetry at hand is truly that of the poet to whom it is ascribed. In one account, Aṣmaʿī (d. 216/831) chose from a poetry collection twenty poems of the raǧaz meter alleged to be authored by al-Aġlab al-ʿIǧlī (d. 21/642). When Abū Ḥātim al-Siǧistānī (d. 255/869) asked him how he could do that after having said that he knew only two and a half raǧaz poems by al-Aġlab, he defended his choice by saying that if the poems were not al-Aġlab’s, then they must be the work of another trustworthy poet (fa-in lam yakun lahu fa-huwa li-ġayrihi mimman huwa ṯabt aw ṯiqa).138 A similar state of affairs was most probably dominant in philology. With the lack of reliable tools by which the

135 Suyūṭī, Muzhir I, 103–113, 182. 136 In spite of the apparent difference between poetry and Ḥadīṯ as far as the development of

critical methods are concerned, poetry critics, according to some, are wiser (aʿqal) than their Ḥadīṯ counterparts because they are more successful in detecting forgery, and even when they quote forged material, they do admit that it is forged; cf. Qālī, Amālī (Ḏayl) III, 105 (also quoted in Suyūṭī, Muzhir I, 175).

137 Yāqūt, Muʿǧam III, 1204. 138 Abū Ḥātim, Fuḥūla 117–118.

 

 

28 chapter 1

validity of reported usage could be checked, all the lexicographers were left with when confronted with doubtful words was to use expressions such as Ibn Durayd’s laysa bi-ṯabt or lā aḥuqquhu. As we shall see later, some lexicographers eliminated from their lexica words which they considered doubtful. This inci- dentally resulted in an appreciable shrinkage in these lexica, as in Ǧawharī’s (d. c. 400/1010) al-Ṣaḥāḥ. The fact that other lexica are replete with such mate- rial in spite of their authors’ suspicion about its correctness – as is the case with Ibn Durayd in the Ǧamhara – demonstrates the absence of strict crite- ria according to which words could be either admitted to or eliminated from the corpus.

The practice of forgery of linguistic material receives confirmation from two of the earliest texts we possess, namely, Ḫalīl’s (d. 175/791) K. al-ʿAyn and Sībawayhi’s (d. 180/796) al-Kitāb. In his introduction to al-ʿAyn, Ḫalīl refers to an anonymous group of scholars whom he calls naḥārīr (pl. of niḥrīr, skillful or learned) and accuses them of intending to cause confusion and obfuscation (irādat al-labs wa-l-taʿnīt) by creating words which do resemble other words and patterns (ašbaha lafẓahum wa-ta⁠ʾlīfahum), but which are neologisms (muwalladāt; cf. muḥdaṯa, mubtadaʿa) that are not permissible (lā taǧūz) in the speech of the Arabs.139 Ḫalīl warns against words – such as kašaʿṯaǧ, ḫaḍaʿṯaǧ and kašaʿṭaǧ – coined by the naḥārīr and tries to formulate criteria which would readily dismiss them as impermissible neologisms. According to him, rules of word composition in Arabic stipulate that all quadriliterals (rubāʿiyya) and quinqueliterals (ḫumāsiyya) cannot be devoid of liquids or labials, except for some ten anomalous (šawāḏḏ) words.140 Therefore, the above quoted words invented by the naḥārīr might well resemble Arabic words of the same pattern, such as (our examples) safarǧal (quince) and ġaḍanfar (lion), but should be rejected because they not only violate the phonotactic principles which govern quadriliterals and quinqueliterals, but also because they are never used in actual speech. Ḫalīl’s criterion based on the phonologi- cal characteristics of Arabic and the mutual compatibility of Arabic phonemes also holds true in the detection of foreign words, as in the rule that no Arabic word begins with n-r. In his introduction, Ḫalīl is obviously keen to demar- cate the speech of the Arabs from the neologisms of the naḥārīr and foreign words alike, and declares that his target is to encompass the two types which occur in Arabic speech: the clear (wāḍiḥ) and the strange (ġarīb).141 His posi- tion vis-à-vis the naḥārīr is matched in the Kitāb, at the level of grammatical

139 Ḫalīl, ʿAyn I, 52, 53; cf. Baalbaki (2008: 20, 50–51, 224–225). 140 Ibid., I, 53. 141 Ibid., I, 60.

 

 

29early lexicographical activity

constructions, by Sībawayhi’s refutation of the naḥwiyyūn, whom he mentions twenty-one times.142 Prompted by their interest in the structural regularities of Arabic, the naḥwiyyūn were accused by Sībawayhi of using qiyās (analogy) in a purely speculative manner in order to invent forms and utterances which do bear resemblance to actual usage, but which should be rejected because they do not occur in the speech of the Arabs. The fact that both Ḫalīl and Sībawayhi insist on restricting linguistic data to what is attested shows the close proxim- ity between lexicography and grammar in their theoretical bases with regard to the data which qualifies to be included in the lexical and grammatical corpus and which, to the exclusion of forged or analogically created material, deserves to be examined as Arabic speech.

4 The Epochs of Reliable Usage (ʿUṣūr al-Iḥtiǧāǧ)

The issues related to data collection were not restricted to assigning the tribes and geographical areas which represent the “purest” form of Arabic and finding ways to verify the authenticity of the data and guard against fabrication. The philologists also had to tackle the more intricate task of defining the time lim- its which apply to reliable usage. In this respect, two of the four major sources of data, namely, the Qurʾān and prophetic Ḥadīṯ, posed no problem, given that the historical period to which they belong certainly predates the second half of the second/eighth century, i.e. the period during which data collection of kalām al-ʿArab was at its peak.143 But these two sources indeed posed other methodological problems for philologists and grammarians alike, including, in the case of the Qurʾān, the acceptance or rejection of Qurʾānic readings (qirāʾāt), the possible effect of theological beliefs on linguistic issues, and the weight to be assigned to Qurʾānic šawāhid (pl. of šāhid, lit. evidence, proof) in comparison with other sources of data, particularly poetry.144 Discussion of these methodological issues is beyond the present scope, but as far as Ḥadīṯ is concerned, it may be appropriate here to mention the difference between the early philologists and grammarians in availing themselves of its wide cor- pus. Sībawayhi’s (d. 180/796) šawāhid, for example, include about 1050 lines

142 Carter (1972: 76, n. 1); cf. also Talmon (1982: 14–15; 2003: 12 where twenty-eight loci of con- troversy with the naḥwiyyūn are identified in the Kitāb), and Baalbaki (2008: 18–20).

143 We are not concerned here with the question of when the canonization of the Qurʾānic text or the Ḥadīṯ corpus took place. For the early philologists, Qurʾān and Ḥadīṯ belonged to the first few decades of Islam, and no temporal limitation of any kind applied to them.

144 Cf. Baalbaki (2008: 36–39).

 

 

30 chapter 1

of poetry, 447 Qurʾānic verses, 350 speech patterns or idiomatic expressions and 41 proverbs, but only seven or eight ḥadīṯs145 according to Hārūn’s index, or twelve according to another enumeration.146 This obvious discrepancy is largely due to the fact that Ḥadīṯ was not always transmitted verbatim147 (note that Sībawayhi’s narration of the ḥadīṯs he cites does not exactly match their wording in the later canonical sources), and that some of its transmitters were not even native speakers of Arabic.148 Accordingly, the use of Ḥadīṯ as linguis- tic evidence in the study of syntax, at least according to the early grammarians, is inappropriate due to the changes that might have been introduced to the original riwāya. In contrast, early lexicographers had no qualms about citing Ḥadīṯ to prove the existence of a certain word or clarify its meaning, prob- ably on the assumption that for this specific purpose it makes little difference whether the Ḥadīṯ was transmitted verbatim or as paraphrased. In other words, the lexicographers might have assumed that even if Ḥadīṯ was not transmitted verbatim, change would primarily affect its structure; whereas lexical items – particularly those which constitute the šawāhid – are likely to be retained by the transmitter. Whatever the case may be, the abundance of the šawāhid of Ḥadīṯ and Aṯar (i.e. sayings of the Prophet’s Companions or Successors) in the lexica is quite remarkable. In the oldest muǧannas lexicon, K. al-ʿAyn, the num- ber of ḥadīṯs and aṯars cited is 428, in sharp contrast to Sībawayhi’s seven or eight citations of Ḥadīṯ. Continuing this approach, Bandanīǧī (d. 284/897) cites 121 ḥadīṯs and aṯars in al-Taqfiya, and Ibn Durayd (d. 321/933) 508 in Ǧamharat al-luġa. Eventually, as we shall see later,149 several lexical works of the ġarīb genre were devoted to the lexemes of Ḥadīṯ, such as Ibn al-Aṯīr’s (d. 606/1210) al-Niyāha fī ġarīb al-Ḥadīṯ, one of the five works which Ibn Manẓūr (d. 711/1311) adopted as sources in compiling his famous Lisān al-ʿArab.

The time limits imposed on linguistic data, and the so-called ʿ uṣūr al-iḥtiǧāǧ or epochs during which usage may be considered as reliable, thus solely apply to the two other sources of the philologists’ corpus, i.e. prose and poetry. Prose is usually referred to by the expression kalām al-ʿArab (the speech of the Arabs), and is specifically the speech of those Aʿrāb whose dialects were con- sidered faṣīḥ, in addition to a sizable body of proverbs and speech patterns or idiomatic expressions. Indeed, the material which the philologists amassed as

145 One of these ḥadīṯs is quoted in two different versions, hence the two different possible enumerations. See Hārūn’s indices in Kitāb V, 32; cf. Baalbaki (2008: 9, n. 36).

146 Ḥadīṯī (1981: 50–78, esp. 77). 147 Cf. Suyūṭī, Iqtirāḥ 54; Baġdādī, Ḫizāna I, 9. 148 Suyūṭī, Iqtirāḥ 53–54; Baġdādī, Ḫizāna I, 11–12. 149 Cf. below, 71 ff.

 

 

31early lexicographical activity

a result of their effort at collecting data during the second/eighth and early third/ninth centuries forms the backbone of the prose corpus. Shortly fol- lowing that period, however, the growing doubts as to whether the Aʿrāb still retained their faṣāḥa and could thus still be considered a reliable source of data became serious enough to warrant discussion in order to decide on a time frame when the corpus should be closed.150 In very broad terms, the corpus of prose was open approximately until no later than the end of the second/ eighth century in the urban areas (amṣār), after which time the speech of sed- entary people was generally viewed as corrupt, witness the large number of previously quoted accounts which contrast sedentary speech with the faṣāḥa of the Aʿrāb. In the case of the Aʿrāb themselves, the corpus remained open until about the end of the fourth/tenth century. Confirmation of this time limi- tation comes from Ibn Ǧinnī (d. 392/1002), who himself frequently consulted Bedouin informants.151 He reported that he became to a large extent trustful of a man who claimed to possess Bedouin faṣāḥa until he heard him one day recite a poem of his in which he used the words ašʾa⁠ʾuhā and adʾa⁠ʾuhā, instead of ašʾāhā (I outstrip it) and adʾāhā (I beguile it [i.e. the prey]). These two forms in which two hamzas occur in succession, according to Ibn Ǧinnī, are imper- missible both in qiyās (analogical formation) and samāʿ (attested data).152 On another occasion, the same man used in his poetry the expression ka⁠ʾanna fāya (as if my mouth), instead of ka⁠ʾanna fiyya. Consequently, Ibn Ǧinnī was strongly persuaded that this man was far from being eloquent ( fa-qawiya fī nafsī bi-ḏālika buʿduhu ʿan al-faṣāḥa), although he regrettably noted that he was one of the best informants he knew (min amṯal man ra⁠ʾaynāhu mimman ǧāʾanā maǧī ʾahu). Based on this experience, Ibn Ǧinnī came to the conclu- sion that in his time (fī waqtinā hāḏā) there was hardly any faṣīḥ Bedouin left, and even if a certain Bedouin gave the initial impression that his speech was characterized by faṣāḥa, one would almost surely find reason to change that impression when one realized, as he himself did, how much faulty usage tar- nished the purity of that Bedouin’s faṣāḥa.

Among the last lexicographers from whom we have evidence of directly transmitting data they heard from the Aʿrāb are two of Ibn Ǧinnī’s contem- poraries, namely, Azharī (d. 370/981) and Ǧawharī (d. c. 400/1010). In the introduction to his Tahḏīb al-luġa, Azharī tells us that he was imprisoned for a period of time by a group of Aʿrāb mainly from the tribe of Hawāzin,

150 Cf. Baalbaki (2008: 40–41; 2011: 102). 151 Ibn Ǧinnī, Ḫaṣāʾiṣ I, 76, 78, 240–242, 250. 152 Ibid., II, 5–8.

 

 

32 chapter 1

which also included a few Tamīmīs and Asadīs.153 He describes their speech as almost free of error (wa-lā yakād yaqaʿ fī manṭiqihim laḥn aw ḫaṭa⁠ʾ fāḥiš) and confirms that he collected from their speech a vast number of words (alfāẓ ǧamma) which he recorded in his lexicon. Indeed, the Tahḏīb abounds with material which Azharī ascribes to the Aʿrāb, whose tribes he often specifies.154 For his part, Ǧawharī reports a conversation which he initiated in Naǧd with an Aʿrābī of the Tamīm tribe to establish whether the small piece of wood inserted in the groove of a pulley is called niḫās or niḥās.155 After the fourth/ tenth century, there have been only sporadic references by the lexicographers to usage they directly heard from the Aʿrāb. One example is that Zamaḫšarī (d. 538/1144) adduces a statement he heard in Mecca from a woman beggar from the region of Sarāt (Sarawiyya mustaǧdiya), in support of his interpreta- tion of the Qurʾānic word nāẓira (looking towards; Q 75: 23).156

In the case of poetry, considerably stricter criteria for admission of mate- rial to the corpus were applied. The distinction between prose and poetry has a strong methodological basis in the grammatical tradition, starting with Sībawayhi, who recurrently highlights the inherent differences between kalām and šiʿr and devotes to poetic license, in the introductory part of his al-Kitāb, known as the Risāla, a chapter entitled mā yaḥtamil al-šiʿr.157 For Sībawayhi, poetic usage is often inadmissible (lā yaǧūz) or weak (ḍaʿīf ) in actual speech, and poetry is replete with šawāhid that do not conform to the norms of kalām and thus require special interpretation and justification.158 In both gram- mar and lexicography, poetry was the subject of great interest due to the vast wealth of the linguistic material it contained, particularly in the realm of ġarīb. In other areas, such as genealogies, geographical works, biographical sources, and Ayyām al-ʿArab (Battle-days of the Arabs), poetry was regarded as the reg- ister (dīwān) of the Arabs; or as Marzūqī (d. 421/1030) puts it, a “depository” of

153 Azharī, Tahḏīb I, 7; cf. below, 312–313. 154 See, for example, the following instances in which he refers in the first volume of the

Tahḏīb to direct samāʿ from the Arabs: ʿQL (I, 241); NQʿ (I, 263); ʿFQ (I, 267); ʿKM (I, 328); ǦRʿ (I, 361); ʿRŠ (I, 414). For examples of reference to specific tribes, see RĠL (VIII, 98; Banū Kilāb); KNZ (X, 98; al-Baḥrāniyyūn); RṮʾ (XV, 124; Banū Muḍarris).

155 Ǧawharī, Ṣaḥāḥ (NḪS); cf. Suyūṭī, Muzhir II, 311–312. See also below, 376. 156 Zamaḫšarī, Kaššāf IV, 192; cf. Zamaḫšarī, Asās (NḪS). 157 Sībawayhi, Kitāb I, 26–32. Note that Sībawayhi also devotes whole chapters to certain phe-

nomena which are unique to poetic usage, such as the use of euphonic elision (tarḫīm) in non-vocative contexts, the use of the independent accusative pronoun iyyā, and the vari- ous changes to which words are subjected as a result of rhyme (qāfiya); ibid., I, 269–274, 362; IV, 204–216; cf. Baalbaki (2008: 43–44).

158 E.g. Kitāb I, 26, 48, 72, 101, 209, 361.

 

 

33early lexicographical activity

knowledge and the equivalent of books to other nations.159 It is telling that Ibn Fāris (d. 395/1004) justifies the linguistic dimension of the saying al-šiʿr dīwān al-ʿArab (Poetry is the register of the Arabs)160 by saying that poetry is the chan- nel through which language was comprehended (wa-minhu tuʿullimat al-luġa) and that it is the authoritative source (ḥuǧǧa) in the problematic ġarīb in the Qurʾān and the Ḥadīṯ of the Prophet, his Companions (Ṣaḥāba) and those next in the order of time (Tābiʿūn).161

The peculiarities of poetry naturally contributed to the conviction of the grammarians and philologists that in many respects – including the issue of time limitations – it should be treated differently from prose. Another con- tributing factor to this conviction seems to have been the broad agreement between linguists and poetry critics162 that the available corpus of poetry could not be treated en bloc and had to be subject to some kind of classifica- tion. Sources as of the third/ninth century have clearly proposed an epoch- based approach to poetry. Perhaps “temporal” is a better word to describe their classification than “chronological” since the latter term implies a more accurate or detailed division of the epochs involved than is actually justified in the sources.163 One of the earliest authors to adopt this temporal classification is Ibn Sallām al-Ǧumaḥī (d. 232/846), who broadly classifies poets into pre- Islamic (Ǧāhilī), Islamic, and those described as muḫaḍramūn because they straddle the two periods of Ǧāhiliyya and Islam.164 Ibn Qutayba (d. 276/889) for his part asserts that precedence in time should not be a factor in the ranking of poets.165 However, he does not in reality diverge from the traditional temporal classification given that he defends, for his book, the inclusion only of famous poets (al-mašhūrūn min al-šuʿarāʾ), whom he defines as those who are quoted in connection with ġarīb, grammar, Qurʾān and prophetic tradition,166 that is, Ǧāhilī and early Islamic poets. Further evidence for the temporal classification

159 Ibn Sallām, Ṭabaqāt I, 25; Ibn Fāris, Ṣāḥibī 275; Marzūqī, Šarḥ I, 3. 160 The expression also appears in the opening verse of a poem in Abū Firās al-Ḥamdānī’s

(d. 357/968) Dīwān 22. 161 Ibn Fāris, Ṣāḥibī 275. 162 The distinction between these two groups is only meant in a very general sense, and

although it may be obvious in the case of critics like Ibn Sallām and Ibn Rašīq who do not have any significant contribution to linguistic sciences, it may not be applicable in the case of an author like Ibn Qutayba whose contribution includes poetry collection and evaluation as well as philological and grammatical works; cf. Baalbaki (2011: 113, n. 31).

163 Discussion of this classification is largely based on Baalbaki (2011: 102–104). 164 Ibn Sallām, Ṭabaqāt I, 23–24. 165 Ibn Qutayba, Šiʿr 10–11. 166 Ibid., 7.

 

 

34 chapter 1

of poetry is found, among others, in Marzubānī’s (d. 384/994) arrangement of al-Muwaššaḥ and, much later, in al-Šarīf al-Ǧurǧānī’s (d. 816/1413) commentary on Zamaḫšarī’s (d. 538/1144) al-Kaššāf.167

The poetry critics’ temporal classification of poetry is paralleled in the lin- guistic sources. The classification cited by Baġdādī (d. 1093/1682), and which sums up the traditional stance of the philologists and grammarians vis-à-vis the permissibility, or otherwise, of citing poetry as testimony to acceptable usage,168 is fully congruent to Ibn Rašīq’s (d. 463/1071) classification which is based on literary considerations.169 Both authors acknowledge four categories of poets: the Ǧāhiliyyūn, the muḫaḍramūn, the Islāmiyyūn or mutaqaddimūn, and the muwalladūn or muḥdaṯūn, who loosely include those who are not of “pure” Arab origin or who roughly belong to the “post-classical” era, around the end of the second/eighth century. From the earliest extant grammatical source, Sībawayhi’s al-Kitāb, we know that there was an obvious exclusion of muwallad poetry from the body of šawāhid.170 Given Sībawayhi’s influence on the Arabic linguistic tradition as a whole, and the grammatical tradition in particular, the exclusion of the muwallad from his poetry šawāhid must have strongly influenced subsequent grammarians who generally also kept muwal- lad outside the sphere of acceptable linguistic testimony, that is, and in tem- poral terms, outside ʿuṣūr al-iḥtiǧāǧ. In fact, all the poets who are referred to as the last ones who belonged to these ʿuṣūr and are dubbed sāqat al-šuʿarāʾ (lit. the poets of the rear) by Aṣmaʿī171 – including Ruʾba (d. 145/762), Ibn Mayyāda (d. 149/766), al-Ḥakam al-Ḫuḍrī (d. 150/767), Makīn al-ʿUḏrī (d. c. 160/777), and Ibn Harma (d. 176/792) – are Sībawayhi’s contemporaries. There is evidence however that scholars prior to both Sībawayhi and the poets who are generally considered as muwalladūn, had an even stricter approach to the notion of muwallad. Abū ʿAmr b. al-ʿAlāʾ (d. 154/770), for example, refers to the poetry of Ǧarīr (d. 110/728) and Farazdaq (d. 110/728) also as muwal- lad. Although Ibn Rašīq apologetically says that Abū ʿAmr meant that the poetry of Ǧarīr and Farazdaq is muwallad in comparison with the poetry of the Ǧāhiliyyūn and muḫaḍramūn,172 other accounts reveal that Abū ʿAmr was deeply cautious of the linguistic value of all Islamic poetry. For example, he

167 Ǧurǧānī, Ḥāšiya I, 220–221. 168 Baġdādī, Ḫizāna I, 5–6. 169 As in Ibn Rašīq’s (ʿUmda I, 113) expressions aġmaḍ maslakan, araqq ḥāšiyatan, ḥalāwat

lafẓihi, rašāqat maʿnāhu, ṭilāwa, labāqa, etc. 170 For a more detailed discussion of this issue, cf. Baalbaki (2008: 42–43). 171 Ibn Qutayba, Šiʿr 639; cf. Iṣfahānī, Aġānī IV, 375; V, 238. 172 Ibn Rašīq, ʿUmda I, 90.

 

 

35early lexicographical activity

is reported to have said that had Aḫṭal (d. 90/708) lived even for one day in Ǧāhiliyya, he would not have given precedence to any other poet over him.173 Precedence here most probably refers to the linguistic, rather than the artistic value of Aḫṭal’s poetry, given Abū ʿAmr’s predominantly philological interests. Moreover, Aṣmaʿī (d. 216/831) claims that he accompanied Abū ʿAmr for eight years (ten years in another manuscript) but never heard him adduce as testi- mony ( yaḥtaǧǧ) even one line of poetry from the Islamic period.174 The general preference of Ǧāhilī poetry to Islamic poetry in relation to linguistic testimony explains why Baššār b. Burd (d. 167/784) is said to have blamed aḏān (i.e. call to prayers, hence the Islamic period) for the disparagement of his poetry (azrā bi-siʿrī l-aḏān).175

In spite of Abū ʿAmr’s position, numerous šawāhid are attributed to Aḫṭal, Ǧarīr and, above all, Farazdaq in both grammatical sources and lexica.176 But even if Abū ʿAmr’s extreme position is not taken into account, the general con- sensus, which left the corpus open for prose or ordinary speech till the end of the fourth/tenth century with regard to the Aʿrāb, was to discontinue the contribution of poetry to the linguistic corpus more than two centuries prior to that. As the grammarians disagreed with the lexicographers on the citation of Ḥadīṯ, they also differed with them, albeit to a lesser degree, concerning the permissibility of citing poets from the period following ʿuṣūr al-iḥtiǧāǧ. Thus, there are hardly any grammatical šawāhid that are attributed to some of the most famous poets, such as Abū Nuwās (d. 198/814), Abū l-ʿAtāhiya (d. 211/826), Abū Tammām (d. 231/846), Ibn al-Rūmī (d. 283/896), Buḥturī (d. 284/898), Mutanabbī (d. 354/965), and Abū l-ʿAlāʾ al-Maʿarrī (d. 449/1057).177 In the lexica, the picture is somewhat different. For example, in Lisān al-ʿArab, Ibn Manẓūr

173 Iṣfahānī, Aġānī VIII, 284. A similar account is Aṣmaʿī’s assertion that had Baššār b. Burd lived earlier, he would have preferred him to many a poet; cf. Iṣfahānī, Aġānī III, 137.

174 Ibn Rašīq, ʿUmda I, 90 and n. 2 on the same page. 175 Iṣfahānī, Aġānī III, 136. 176 Cf. Hārūn’s indices to Sībawayhi’s Kitāb (V, 181, 183, 192), and Abū l-Hayǧāʾ & ʿAmāyira’s

indices to Ibn Manẓūr’s Lisān (1987: III, 725–726, 752–755, 846–848). 177 According to the index prepared by Ḥaddād (1984: 805 ff.) based on 3,800 lines of poetry

šawāhid, only one line is attributed to Abū Nuwās, two to Abū l-ʿAtāhiya, and none to any of the other five poets mentioned above. Note that Hārūn’s indices of šawāhid (1972–73: II, 586 ff.) contain several references to these poets, but this is largely due to his inclusion of certain literary, i.e. not strictly grammatical works in the list of his sources. In a few cases, the grammarians did cite muwallad poetry, but for its meaning rather than form or structure, as Ibn Ǧinnī (Ḫaṣāʾiṣ I, 24) explains when he cites Mutanabbī and mentions Mubarrad’s citing of Abū Tammām. Note also the apologetic position of Zamaḫšarī (Kaššāf I, 220–221) when he cites a line by Abū Tammām, and that of Ibn Ḫillikān (Wafayāt II, 81)

 

 

36 chapter 1

(d.711/1311) attributes three šawāhid to Abū Nuwās, eight to Abū Tammām, one to Buḥturī, eleven to Mutanabbī, three to Maʿarrī, and none to Abū l-ʿAtāhiya or Ibn al-Rūmī.178 These numbers are dwarfed by the tens or hundreds of šawāhid which Ibn Manẓūr attributes to Ǧāhilī and early Islamic poets such as Aʿšā, Umruʾ al-Qays, Aws b. Ḥaǧar, Ḏū l-Rumma, Ruʾba b. al-ʿAǧǧāǧ, Zuhayr b. Abī Sulmā, al-Šammāḫ b. Ḍirār, Ṭarafa, al-Ṭirimmāḥ b. Ḥakīm, al-ʿAǧǧāǧ b. Ruʾba, al-Kumayt al-Asadī, al-Nābiġa al-Ǧaʿdī, al-Nābiġa al-Ḏubyānī, and Abū l-Naǧm al-ʿIǧlī.179 The presence in the lexica of at least some šawāhid attributed to poets of the third/ninth century onward proves that the lexicographers were more tolerant of citing certain lexemes in relatively late poetry – as they were vis-à-vis Ḥadīṯ and Aṯar – than were the grammarians in relying on the tes- timony of this poetry in syntactical matters. Yet, by and large, the relatively short period of ʿuṣūr al-iḥtiǧāǧ assigned by the grammarians, and to a lesser extent by the lexicographers, for poetry as compared to ordinary speech deprived both the study of grammar and the corpus of lexical items from a source which had the potential for enriching the available data, particularly since some of the above mentioned poets of the third/ninth century onward, such as Mutanabbī, were highly esteemed for their linguistic competence and grammatical knowledge180 or were even considered, as was Maʿarrī by some authors, to be philologists in their own right.181 It is indeed strange that Maʿarrī is quoted twenty-eight times in the Lisān as a philologist and narrator, and only three times as a poet. In view of this attitude towards ʿuṣūr al-iḥtiǧāǧ, the later lexicographers normally repeat the same šawāhid cited by earlier authors and take little notice of contemporary poetry and the semantic development which it could attest to in comparison with earlier usage.

5 The Role of Ġarīb

In recording kalām al-ʿArab, particularly from Bedouin sources, the philologists were greatly interested in what they called ġarīb: strange or uncommon/rare

as he explains away Abū ʿAlī al-Fārisī’s citation of Abū Tammām’s poetry on one occasion (cf. Fārisī, Šiʿr I, 315).

178 Abū l-Hayǧāʾ & ʿAmāyira (1987: III, 721 ff.); cf. Ḥamza (2011: 60–62); Ǧabal (1986a: 109 ff.; 1986b: 25–29, 54–56).

179 Names are listed in their order of appearance in the indices of Abū l-Hayǧāʾ and ʿAmāyira (but note that we have modified the number of citations based on al-Lisān’s text); cf. also Ayyūbī (1980: 90, 290, 370, 433).

180 For Mutanabbī’s vast knowledge of philological and grammatical issues, see Ibn Ḫillikān, Wafayāt I, 120–121 and ʿAbbāsī, Maʿāhid I, 30.

181 See his lengthy biography in Qifṭī, Inbāh I, 81–118.

 

 

37early lexicographical activity

usage. Their interest in ġarīb material, however, was not restricted to prose and poetry, but included the two other sources of linguistic data, namely, Qurʾān and Ḥadīṯ. We shall later examine the two genres of writing known as ġarīb al-Qurʾān and ġarīb al-Ḥadīṯ,182 but first we need here to define what ġarīb really means and try to understand how it came to be such an important com- ponent of early lexical works.

In addition to ġarīb, the philologists used several terms to express the notion of strange usage, notably nādir (pl. nawādir, rare), but also ḥūšī or waḥšī (unfamiliar, uncouth, barbarous), and even šāḏḏ (anomalous). The philologists nowhere offer meaningful distinctions among the purports of these terms,183 and one indeed finds no clear difference in the content of books whose titles refer either to ġarīb or nawādir.184 Accordingly, ġarīb, nādir and other similar terms may be safely considered as synonymous. To take Abū Misḥal al-Aʿrābī’s (d. 231/845) K. al-Nawādir as an example of an early lexical work composed of a large body of material of the ġarīb or nādir type, most of its content may be classified in the following general areas:

1. Words that are rarely used in any of the four sources of linguistic data; e.g. raǧulun ǧardabān or ǧurdubān or ǧardabīl (one who eats with his right hand), and kurbuǧ/kurbaǧ or kurbuq/kurbaq (shop).185 In certain cases, a word may be cited because it is less common than another widely used word from the same root; e.g. šarafa, a synonym of šaraf (honor).186

2. Collections of synonyms, all or most of which are only rarely attested; e.g. iʿranfaza, iḥra⁠ʾabba, iǧra⁠ʾanna, ǧasa⁠ʾa and taraza (to stiffen or die of cold), and mā lahu min ḏālika ḥawīl wa-lā zawīl wa-lā maḥīṣ wa-lā mafīṣ wa-lā nawīṣ (to have no escape or alternative).187

3. Words that are pronounced differently in various dialects; e.g. ġilẓa, ġulẓa and ġalẓa (roughness or rudeness), and āhulu bihi, āhilu bihi and āhalu bihi (I like his company).188

4. Anomalous forms that represent an exception to a generally applicable norm; e.g. the two forms ruʾya (seeing) and ǧurda (denudement), said to

182 See below, 64, 71. 183 See below, 84. 184 Šalqānī (1971: 89–100) artificially distinguishes between ġarīb and nādir. He rightly argues

that ġarīb words are those whose meaning is obscure (ġāmiḍ), but his claim that nawādir represent a miscellanea (tafārīq) of ġarīb implies an admission that the two terms may well be synonymous.

185 Abū Misḥal, Nawādir I, 136, 328. 186 Ibid., II, 498. 187 Ibid., I, 96, 165 (cf. II, 487). 188 Ibid., I, 241, 256.

 

 

38 chapter 1

be the only verbal nouns expressing manner (i.e. maṣdar nawʿ or hayʾa) which are not of the pattern fiʿla.189

5. Idiomatic expressions that do not contain any ġarīb words; e.g. iḫtalaṭa l-laylu bi-l-turābi (lit. The night mingled with the sand), which means that darkness intensified or that the affair was confusing, and ǧāʾa yanfuḍu yadayhi (He came shaking his hands [as if to remove what stuck to them]), which means he came empty-handed.190

6. Rare syntactic phenomena; e.g. the use of the relative pronoun mā before lā l-nāhiya (lā of prohibition), as in hāḏā mā lā turidhu (lit. This is what I prohibit you from wanting, i.e. this is what you should not want), and the use of the preposition bi- instead of ilā in the expression aḥsantu bi-fulānin (I was charitable to him; I did good to him).191

Although the above types mostly represent uncommon or rarely attested words, expressions, etc., the criteria for labeling usage as ġarīb are not always clear. Whereas words such as ǧardabīl, kurbuǧ and iʿranzafa may well have been obscure, if not totally incomprehensible to most Arabs including the philolo- gists, there are several entries which do not give a similar impression.192 One thus wonders why Abū Misḥal’s book contains entries such as istaḫartu l-Lāha (I asked God for proper guidance), šatta amru l-qawmi yašittu šattan (The peo- ple dispersed [followed by the imperfect and verbal noun]), and raǧulun waḍī ʿ or wasīṭ or rafī ʿ fī qawmihi (a man of lowly, median, or eminent status among his folks), in which each of the three adjectives is then followed by the appro- priate verbal noun.193 Similarly, most of the substantives cited in Abū Misḥal’s lengthy list of the type raǧulun bayyinu l-ruǧūlati (lit. a man of apparent man- hood, i.e. a true man) are obviously quite common and widely known, such as fatāʾ/futuwwa (youth), ubuwwa (fatherhood), bunuwwa (sonship), ʿumūma (uncleship), ǧiwār (neighborhood), šaǧāʿa (courage), ẓarf (wittiness), karam (generosity), wasāma (prettiness), ṭūl (tallness), qiṣar (shortness), ǧamāl

189 Ibid., II, 504. 190 Ibid., I, 332. 191 Ibid., I, 241; II, 520. 192 Note that Ibn Hišām tries to assign approximate numerical values to nādir and other

terms. If we take the number twenty-three, for example, he says that the terms muṭṭarid (uninterrupted), ġālib (dominant), kaṯīr (abundant), qalīl (scarce), and nādir (rare) may be represented by twenty-three, twenty, fifteen, three, and one respectively; cf. Suyūṭī, Muzhir I, 234. Although this is a fair representation of the general meaning of these terms, it does not mean that they are used in the sources with any precision based on objective criteria.

193 Abū Misḥal, Nawādir I, 223, 304; II, 490.

 

 

39early lexicographical activity

(beauty), etc.194 Perhaps the reason for the inclusion of these substantives is to illustrate the vastness of the material that can be included in exemplifying the typical expression “adjective + bayyin + verbal noun”. A similar explanation may be given to Abū Misḥal’s inclusion in his material of proverbs which do not seem to be particularly ġarīb,195 but which demonstrate the remarkable potential for expression of meaning based on stereotyped expressions perti- nent to specific situations.

Interest in ġarīb material is often associated in the sources with the very early period of Arabic philological activity. Nowhere is this association clearer than in the reports which trace this interest back to the famous ṣaḥābī (Prophet’s Companion), Ibn ʿAbbās (d. 68/687). According to one report, people blocked the road leading to Ibn ʿAbbās’s house, so he asked his attendant to admit first those who wanted to ask him about the Qurʾān and its ḥurūf (words, or pos- sibly modes of reading), followed respectively by those who sought answers to questions related to the interpretation of the Qurʾān, the lawful and the forbid- den (al-ḥalāl wa-l-ḥarām), jurisprudence ( fiqh), religious obligations ( farāʾiḍ), and finally ʿArabiyya, poetry and ġarīb. The account which Abū Nuʿaym al-Iṣfahānī (d. 430/1038) reports based on a long chain of authority (isnād) twice refers to the pride which Qurayš should take in this maǧlis (assembly).196 But irrespective of the authenticity of this account and the obvious attempt at praising Qurayš, it demonstrates the perception that Ibn ʿAbbās was equally competent in religious and linguistic issues. This is essential for understanding another account on which are based the masāʾil (questions) about which Nāfiʿ b. al-Azraq (d. 65/685) supposedly asked Ibn ʿAbbās, and which highlight Ibn Abbās’s vast knowledge of the ġarīb in Qurʾān and poetry alike. According to this account, the two Kharijite leaders Nāfiʿ b. al-Azraq and Naǧda b. ʿUwaymir (d. 69/688) met in Mecca Ibn ʿAbbās who was answering questions related to Qurʾānic interpretation. Amazed by how daring Ibn ʿAbbās was to deal with such matters, Nāfiʿ himself began asking him questions which typically start by a Qurʾānic quotation containing a word that requires explanation. When Ibn ʿAbbās gave his explanation, Nāfiʿ would typically ask whether the Arabs were cognizant of it. Ibn ʿAbbās would then cite a line (or more) of poetry in which the word in question appeared.197 The staggering number of passages

194 Ibid., I, 320–325; cf. Ibn Durayd, Ǧamhara III, 1290. 195 See the index of proverbs (ibid., II, 709–710), and in particular those which start with the

comparative pattern afʿalu min (I, 258–262). 196 Iṣfahānī, Ḥilya I, 320–321. 197 Ibn ʿAbbās, Masāʾil 33 ff.

 

 

40 chapter 1

presented in this manner – two hundred and eighty-seven in all198 – shows that for early philologists, poetry was the medium through which Qurʾānic ġarīb was best comprehended. Of course, this further means that poetry itself was replete with ġarīb material. Ibn ʿAbbās himself is reported to have advised those who asked him about ġarīb in the Qurʾān to seek their answers in poetry since it was the dīwān of the Arabs.199 A variant of this report mentions ġarīb al-luġa, that is, ġarīb in general, rather than merely Qurʾānic ġarīb, as mate- rial to be understood by reference to poetry.200 Also worth mentioning is that in his K. al-Zīna, Abū Ḥātim al-Rāzī (d. 322/934) argues that were it not for people’s need for poetry to know the language of the Arabs and elucidate the ġarīb in the Qurʾān and Ḥadīṯ, poetry would have become obsolete and poets would have no longer been remembered.201

As in the general type of ġarīb or nādir represented by Abū Misḥal’s book, not all the material presented as Qurʾānic ġarīb in the passages ascribed to Ibn ʿAbbās (or in the list of words also ascribed to him and arranged accord- ing to the order of the sūras of the Qurʾān)202 could have been unfamiliar or of obscure meaning. Words like šuwāẓ (smokeless flame; Q 55: 35), ḥaraḍ (fatigued; Q 12: 85), dihāq (brimful; Q 78: 34) and ḥanīḏ (roasted with heated stones; Q 11: 69), and an expression like wa-ltaffat al-sāqu bi-l-sāqi (and one leg joined with another; Q 75: 29)203 are most likely to have required explanation, but this can hardly be the case in words such as ḥatm (ordinance; Q 19: 71), aflaḥa (to prosper; Q 23: 1), ḫālidūna (immortal; Q 2: 25), alīm (painful; Q 2: 10), alfaynā (we found; Q 2: 170), and naḥs (misfortune; Q 54: 19).204 One possible explanation is that certain words were included not because they were rarely used but in order to clarify their meaning in a Qurʾānic context. For example, aflaḥa might have been included to specify what the notion of falāḥ entails for the believers in the verse qad aflaḥa l-muʾminūna (The believers indeed pros- per; Q 23: 1). Similarly, the verbs fāzū (they won) and saʿidū (they were happy) used by Ibn ʿAbbās in his explanation elucidate the religious meaning of falāḥ

198 This number is based on A. M. Dālī’s edition, which includes 52 passages from manu- script 3849 of the Ẓāhiriyya collection, 156 out of the 190 passages recorded by Suyūṭī in his Itqān, 18 out of the 50 passages in Ibn al-Anbārī’s Īḍāḥ, one out of the 7 passages in Mubarrad’s Kāmil, and 60 passages taken from I. Sāmarrāʾī’s edition of Suʾālāt Nāfiʿ b. al-Azraq (Baghdad: Maṭbaʿat al-Maʿārif, 1968).

199 Qurṭubī, Ǧāmiʿ I, 21. 200 Zarkašī, Burhān I, 293. 201 Rāzī, Zīna 123. 202 Suyūṭī, Itqān II, 6–54. 203 Ibn ʿAbbās, Masāʾil 35, 52, 70, 143, 51 respectively. 204 Ibid., 67, 94, 95, 101, 104, 179.

 

 

41early lexicographical activity

by describing the state of believers who reap their everlasting rewards.205 It is perhaps this aspect of the explanation of Qurʾānic ġarīb that was most contro- versial in the view of some philologists, since it touched on matters related to faith. As mentioned earlier, the motivation behind Nāfiʿ’s questions that were allegedly addressed to Ibn ʿ Abbās was Nāfiʿ’s conviction that the latter was bold enough (cf. mā aǧra⁠ʾaka “How bold you are”!)206 to explain Qurʾānic words and expressions.

Several of the philologists of the second/eighth and third/ninth centuries also expressed strong reservations against Qurʾānic interpretation by fellow phi- lologists. The sharpest criticism was directed against Abū ʿ Ubayda (d. 209/824), author of Maǧāz al-Qurʾān. According to Abū Ḥātim al-Siǧistānī (d. 255/869), it is impermissible for anyone to write such a book, and to be whipped is less painful than to read it.207 The reason for Abū Ḥātim’s unrelenting position may be clarified from two other accounts, the first of which is related by Abū Ḥātim himself on the authority of Abū ʿUmar al-Ǧarmī (d. 225/840), who blamed Abū ʿUbayda because his interpretation differed from that of the fuqahāʾ (jurists).208 The second account is that Aṣmaʿī (d. 216/831) censured Abū ʿUbayda for inter- preting the Qurʾān based on his own views ( yufassir ḏālika bi-ra⁠ʾyihi).209 The unanimous position of Abū Ḥātim, Abū ʿ Umar and Aṣmaʿī against Abū ʿ Ubayda obviously rests on religious grounds (cf. the expression mā yaḥill li-aḥad an yaktubahu [it is not licit for anyone to write it]), for they seem to have been convinced that Qurʾānic interpretation is the realm of the fuqahāʾ, and that anyone who ventures to interpret scripture will commit unforgivable mistakes, as Abū ʿUbayda did.210 Abū ʿUbayda’s position, on the other hand, acknowl- edges an independent role for the philologists in dealing with Qurʾānic inter- pretation. Having heard Aṣmaʿī’s criticism of Maǧāz al-Qurʾān, he is reported to have headed to his maǧlis and asked him about the meaning of the word

205 Ibid., 94. 206 Ibid., 23. 207 Zubaydī, Ṭabaqāt 176. 208 Ibid., loc. cit. 209 Ibn al-Anbārī, Nuzha 87–88; Yāqūt, Muʿǧam VI, 2707. 210 In contrast to Qurʾānic interpretation, the explanation of ġarīb in poetry caused little

controversy concerning its permissibility. This is highlighted in an account which relates that when the early philologist ʿAbdallāh b. Abī Isḥāq (d. 117/735) heard that the famous Basran jurist Ibn Sīrīn (d. 110/729) blamed him for interpreting poetry, his answer was that, in poetry interpretation, no legal judgment related to what is lawful or forbidden is involved (inna l-fatwā fī l-šiʿr lā tuḥill ḥarāman wa-lā tuḥarrim ḥalālan), the implication, of course, being that interpreting the Qurʾān does raise such concerns for interpreters; cf. Qifṭī, Inbāh II, 106.

 

 

42 chapter 1

ḫubz in the verse innī arānī aḥmilu fawqa ra⁠ʾsī ḫubzan (I see myself carrying bread on my head; Q 12: 36). When Aṣmaʿī explained that it referred to the stuff that is baked and eaten, Abū ʿUbayda hastened to accuse him of interpreting the Qurʾān based on his own view ( fassarta kitāb Allāh bi-ra⁠ʾyika), a charge vehemently denied by Aṣmaʿī, who asserted that his explanation was founded on what he confidently knew (hāḏā šayʾ bāna lī fa-qultuhu), and not on his own opinion. Abū ʿUbayda then ended the conversation by maintaining that he too did not resort to his opinions, but to his firm knowledge (i.e. of Arabic).211 Even more telling is Abū ʿ Ubayda’s shameless answer to Abū ʿ Umar’s accusation that his interpretation of the Qurʾān contrasted with that of the fuqahāʾ. His state- ment hāḏā tafsīr al-Aʿrāb al-bawwālīn ʿalā aʿqābihim fa-in šiʾta fa-ḫuḏhu wa-in šiʾta fa-ḏarhu (This is the interpretation of the Bedouins who urinate on their heels, and you are free to take it or leave it)212 clearly demonstrates his con- viction that the philologists’ approach to Qurʾānic interpretation characteristi- cally relies on the speech of the fuṣaḥāʾ to understand the meaning of words and expressions and is thus unlike any other type of tafsīr.

Fortunately for the history of philology and lexicography, Abū ʿUbayda’s method was triumphant, and the linguistically oriented approach to Qurʾānic interpretation, among whose earliest extant sources are Farrāʾ’s (d. 207/822) Maʿānī l-Qurʾān and al-Aḫfaš al-Awsaṭ’s (d. 215/830) book of the same title, persisted in the tradition. Surely, some of the material in these sources was incorporated into the lexica, which often include direct references to them. Yet it should be noted that philologists who tackled Qurʾānic issues, such as ġarīb, generally took an apologetic stance to demonstrate their so-called “pious” approach to the Holy Book. This might explain the eager attempt to establish an uninterrupted tradition of using poetry in explaining Qurʾānic usage, even before the time of Ibn ʿ Abbās. In this context, there are accounts that date back to the Prophet’s Companions (e.g. Caliph ʿUmar; d. 23/644) and typically por- tray poetry as the key to understanding Qurʾānic usage.213 Even an author such as Abū Zayd al-Qurašī (fl. in the third/ninth century?) whose Ǧamharat ašʿār al-ʿArab is basically composed of meanings of words which occur in poetry, with only sporadic references to the Qurʾān,214 sets out in his introduction to prove that the Qurʾān and the speech of the Arabs share the various meanings of words and metaphors (wa-fī l-Qurʾān miṯl mā fī kalām al-ʿArab min al-lafẓ

211 Yāqūt, Muʿǧam VI, 2707. 212 Zubaydī, Ṭabaqāt 176. 213 See the anecdote cited by Zamaḫšarī (Kaššāf II, 205) and Qurṭubī (Ǧāmiʿ X, 110–111).

Similar early riwāyas are also cited by Ibn al-Anbārī, Īḍāḥ I, 61–99. 214 Qurašī, Ǧamhara 66, 70, 105, etc.

 

 

43early lexicographical activity

al-muḫtalif wa-maǧāz al-maʿānī).215 Abū Zayd’s view implies that Qurʾān provides evidence needed for the interpretation of poetry and vice versa. It goes without saying that in most cases the testimony of poetry was sought for understanding Qurʾānic ġarīb, rather than the other way around.216

Poetry being the genre which preserves the vast majority of ġarīb material is evident in the sources; equally evident is the fondness of both grammar- ians and philologists of rare words, forms and constructions that occur therein. For example, poetry provides the largest portion of the body of the grammar- ians’ šawāhid or attested material brought as the main evidence of uncommon usage.217 It also features strongly in works that cite a closed set of words to which a certain phenomenon is confined, such as words of the patterns afanʿal and afʿalān.218 It is also common for philologists to adduce poetry material almost exclusively as evidence of ġarīb usage in connection with the so-called “dispraised” dialects (luġāt maḏmūma).219 But perhaps the most striking expression of the link which philologists establish between poetry and those forms which are ġarīb in the sense that they typify extremely rare phenomena is Ibn Ḫālawayhi’s (d. 370/980) book Laysa fī kalām al-ʿArab. It is replete with expressions of the following type: The dual of waḥdahu (alone; i.e. waḥīdayni) appears only in a line of poetry by ʿUmāra (i.e. b. ʿAqīl; d. 239/853); the word

215 Ibid., 4. 216 See, for example, Iṣfahānī, Aġānī VI, 70 for an account in which Ḥammād al-Rāwiya

(d. 155/772) reportedly explained a word in Ibn Muzāḥim’s poetry by resorting to a Qurʾānic verse.

217 Note, for example, that the number of poetry šawāhid in Sībawayhi’s Kitāb exceeds the number of šawāhid drawn from all other genres put together. Also, the term šawāhid is used by some later grammarians to refer to poetic šawāhid exclusively, as witnessed by numerous works which are devoted solely to poetic šawāhid yet whose titles sim- ply mention šawāhid without any further specification. Examples include Ibn Hišām’s (d. 761/1360) Taḫlīṣ al-šawāhid wa-talḫīṣ al-fawāʾid and Suyūṭī’s (d. 911/1505) Šarḥ šawāhid al-Muġnī; cf. Baalbaki (2008: 44).

218 See Sībawayhi’s poetry evidence for alandad (strongly hostile; Kitāb IV, 247) and arwanān (a grievous day or night; IV, 248); cf. below, 94. The context in which Sībawayhi cites izmawl (swift in running, said of a mountain-goat; IV, 246) may also be another instance of this phenomenon. Note also that Sībawayhi (IV, 247) mentions the adjective inqaḥl (very aged) without specifying that it is the only example of the pattern infaʿl, and that Ibn Ǧinnī (Ḫaṣāʾiṣ I, 229) confirms its uniqueness and cites a line of poetry in which it appears. For other instances of words which are known solely from poetry, see Ḫaṣāʾiṣ II, 21–24. Cf. also Baalbaki (2011: 114, n. 47).

219 Ibn Durayd, Ǧamhara I, 42–44 and Ibn Fāris, Ṣāḥibī 53–55. Ibn Ǧinnī (Ḫaṣāʾiṣ II, 11–12) even allows the use of such dialects to be analogically extended if the need arises in poetry or rhymed prose (saǧʿ).

 

 

44 chapter 1

ṣaydan (fox) appears only in Kuṯayyir’s (d. 105/723) poetry; the pattern yafanʿal appears only in Ṭufayl al-Ġanawī’s (d. c. 610 A.D.) poetry (i.e. Yabambam, place name); and ʿiyām as the plural formation of ʿaymān (longing for milk) appears only in one line of poetry (ascribed elsewhere to al-Nābiġa al-Ǧaʿdī; d. 50/670).220 To be sure, both grammarians and philologists were criticized for their approach. Ǧāḥiẓ (d. 255/869), for example, notes that the grammar- ians are interested in poetry only from the perspective of iʿrāb (case endings) and that poetry narrators – who surely include those whom we refer to as phi- lologists and lexicographers – are interested only in strange usage and diffi- cult meanings (wa-lam ara ġāyat al-naḥwiyyīn illā kull šiʿr fīhi iʿrāb wa-lam ara ġāyat ruwāt al-ašʿār illā kull šiʿr fīhi ġarīb aw maʿnā ṣaʿb).221 This approach did not go unnoticed even by poets who complained of the philologists’ prefer- ence for what is uncommon and pompous. Abū Nuwās (d. 198/814), for exam- ple, is quoted as saying that had all of his poetry been grandiloquent, no other poet would have been preferred to him (law kāna šiʿrī kulluhu yamla⁠ʾ al-fam mā taqaddamanī aḥad).222

More specifically, raǧaz is the meter which is most closely identified with ġarīb. The present author has previously argued that raǧaz cannot be regarded as merely one of sixteen meters of classical Arabic poetry since its special lin- guistic characteristics set it apart from all other meters.223 Its metric structure, made up of six mustafʿilun feet, is extremely flexible since this particular foot admits a large number of possible changes. Yet its unique character stems rather from the nature of its subjects, which are closely associated with desert life, and from the uncommon words and structures that are often used in connection with these subjects.224 Given their fondness of uncommon and irregular usage, the philologists not surprisingly devoted great attention to the formal aspects of raǧaz as the embodiment of the Bedouins’ linguistic purity. In fact, when Aṣmaʿī (d. 216/831) was asked about the vast amount of raǧaz which he memo- rized, he unequivocally declared that raǧaz was what we were most after and

220 Ibn Ḫālawayhi, Laysa 231, 294, 375, 380 respectively. 221 Ǧāḥiẓ, Bayān IV, 24. Note also the statement by al-Rāġib al-Iṣfahānī (d. 502/1108) in which

he uses mustaġrab, a synonym of ġarīb, to describe the type of iʿrāb in which the gram- marians were particularly interested: wa-kaṯīr min al-naḥwiyyīn lā yamīlūn min al-šiʿr [illā] ilā mā fīhi iʿrāb mustaġrab wa-maʿnā mustaṣʿab (and most grammarians, as far as poetry is concerned, care only for what contains unfamiliar usage of case endings or com- plex meaning); cf. Rāġīb, Muḥāḍarāt I, 94. See also Baalbaki (2008: 44–45; 2011: 104–107).

222 Marzubānī, Muwaššaḥ 409. 223 Baalbaki (2011: 106–107). 224 See Kaššāš (1995: 172–195) and ʿUbaydī (1970: 134–143) for a discussion of the various char-

acteristics of raǧaz, both in form and content.

 

 

45early lexicographical activity

what we most cared for (innahu kāna hammanā wa-sadamanā).225 Another good example of the early lexicographers’ identification of raǧaz with ġarīb is Abū Zayd al-Anṣārī’s (d. 215/830) al-Nawādir fī l-luġa, whose title indicates that it is devoted to rare and unfamiliar usage. Out of the fifteen chapters of which the book consists, two are on šiʿr (i.e. poetry other than the raǧaz meter), seven on raǧaz, and six on nawādir which themselves contain a large amount of raǧaz. The book’s indices show that the number of šawāhid in the raǧaz meter is about two-thirds of the šawāhid in all other fifteen meters put together.226 Similarly, there is a disproportionately high ratio of raǧaz compared to šiʿr in Ibn Durayd’s (d. 321/933) Ǧamharat al-luġa,227 which is famous for its inclusion of unfamiliar usage in spite of its author’s claim to the contrary.228 The raǧaz contained in these and similar books was also a major source of the dialec- tal idiosyncrasies with which grammatical and philological sources abound. In grammatical controversies, for example, raǧaz features prominently in the šawāhid which illustrate dialectal usage in conflict with the norm.229 Similarly, the majority of the šawāhid which relate to “unorthodox” phonological usage are embedded in raǧaz material.230 It should be noted finally that although raǧaz preserved a massive body of actually used dialectal forms which found their way to the lexica, it might have contained invented forms which are incon- sistent with the methods of the genuine (aqḥāḥ) native speakers, as Ibn Ǧinnī (d. 392/1002) puts it. He specifically mentions two of the most famous ruǧǧāz, ʿAǧǧāǧ (d. c. 90/708) and his son Ruʾba (d. 145/762), who were so versed in raǧaz (li-īġālihimā fī l-raǧaz) that they had, due to the shortness of its hemistichs, to resort to invention.231 If Ibn Ǧinnī is right in his accusation, then some ruǧǧāz could be compared to the previously mentioned naḥwiyyūn of Sībawayhi and naḥārīr of Ḫalīl,232 although the available material is too scarce to allow one to assess their influence on the data available in the sources.

225 Abū l-Ṭayyib, Marātib 95. 226 Abū Zayd, Nawādir’s indices, 648–738. 227 Ǧamhara’s indices, III, 1381–1508; cf. below, 344. 228 See my introduction to the Ǧamhara I, 25–27. 229 The Kufans, for example, reportedly allowed the construction yā Allāhumma (O Lord!)

which retains both the vocative particle and the suffix -umma, contrary to standard usage in which only one of them is used (i.e. yā Allāh or Allāhumma). Not unexpectedly, all three šawāhid which the Kufans adduce in support of their position are of the raǧaz meter; cf. Ibn al-Anbārī, Inṣāf I, 341–343.

230 Cf. the šawāhid cited by Ibn Durayd, Ǧamhara I, 42–44 and the sources cited in the foot- notes therein.

231 Ibn Ǧinnī, Ḫaṣāʾiṣ III, 298; cf. I, 360–361, 369. 232 See above, 28.

 

 

46 chapter 1

6 The Compilation of Lexica

One of the problematic issues in the early period of lexicography is how to figure out the relationship between the collection of data and the actual com- pilation of lexica. A logically tempting proposal is the one advanced by Aḥmad Amīn, who distinguishes between three stages (marāḥil) in the early period of lexicographical activity.233 The first stage is that during which the philolo- gists collected the data from the Bedouins but did not classify them. The col- lected items would thus include a word related to rain, followed by another related to sword, then a third related to plants and so on. The second stage witnessed the classification of these words according to subject, resulting in monographs that dealt with a particular topic. Examples include Abū Zayd al-Anṣārī’s (d. 215/830) K. al-Maṭar (The Book of Rain) and Aṣmaʿī’s (d. 216/831) numerous short books on subjects such as cattle, camels, horses, palm trees, etc. Later authors put together these single-theme books in the form of mul- tithematic works such as Abū ʿUbayd’s (224/838) al-Ġarīb al-muṣannaf (our example) and Ibn Sīda’s (d. 458/1066) al-Muḫaṣṣaṣ. In the final stage, authors devised ways to arrange all Arabic words in lexica which would allow the user to find any word in a predictable place based not on its meaning, but on unified principles of arrangement of lafẓ, as is the case for instance in Ḫalīl’s (d. 175/791) K. al-ʿAyn.

Amīn’s seemingly logical sequence, however, hardly corresponds to the chronology one infers from the sources. First of all, the process of data collec- tion which started in the middle of the second/eighth century with scholars such as Abū ʿAmr b. al-ʿAlāʾ (d. 154/770), or even earlier if we are to include figures like ʿAbdallāh b. Abī Isḥāq (d. 117/735),234 continued to the middle of the third/ninth century with scholars like Māzinī (d. 249/863), Abū Ḥātim al-Siǧistānī (d. 255/869), and Riyāšī (d. 257/871), that is, well after the appear- ance of books of the alleged second and third stages.235 Furthermore, the sources mention early books ascribed to contemporaries of Ḫalīl’s from whom

233 Amīn (1938: II, 263 ff.). Note that in a lecture delivered at a conference in Rabat in 1940 and published in 1961, W. Marçais reproduces, almost verbatim, Amīn’s proposal of these stages, but without acknowledgment, perhaps because the article was published in its original form without footnotes; cf. Marçais (1961: 149–151). A similar view is advanced by Ḥumaydī in his introduction to Ibn Qutayba, Ǧarāṯīm I, 28–29, and Diyāb in his introduc- tion to Iskāfī, Mabādiʾ 15–16; see also Abū Šarīfa et al. (1989: 116–118); Ġurāb (2005: 48–50).

234 See above, 16. 235 In support of his belief that the third stage chronologically follows the first two, Amīn

argues that Aṣmaʿī and Abū Zayd were contemporaries of Ḫalīl’s but outlived him consid- erably, and hence the ʿAyn may well have been authored during their lifetime. But even

 

 

47early lexicographical activity

he gathered linguistic data. One such individual is Abū Ḫayra al-Aʿrābī (d. ?), whom Ḫalīl is said to have frequented and used as informant,236 and who is actually quoted several times in the ʿAyn.237 To him is ascribed, by none other than Ibn al-Nadīm, a book entitled K. al-Ḥašarāt,238 whose title indicates that it dealt with vocabulary related to insects, most probably in a fashion similar to that of the extant single-topic monographs authored by Abū Zayd al-Anṣārī, Aṣmaʿī and Abū Ḥātim al-Siǧistānī. Another scholar mentioned along with Abū Ḫayra, among others, as an informant of Ḫalīl’s is ʿAmr b. Kirkira (d. ?),239 to whom three titles are ascribed, namely, K. Ḫalq al-insān,240 K. al-Ḫayl,241 and K. al-Nawādir.242 Also, to Abū ʿAmr b. al-ʿAlāʾ, one of the earliest scholars rec- ognized for data collection, is ascribed K. al-Nawādir.243 As far as Amīn’s third stage is concerned, the mere existence of K. al-ʿAyn is proof that the chronolog- ical order of the three stages is incorrect. If only for the sake of argument one concedes that not the whole lexicon was authored by Ḫalīl, it could not have been completed long after his death in 175/791, and it thus predates the conclu- sion of the first two stages. In fact, al-Layṯ b. al-Muẓaffar (d. 190/805), to whom the book is ascribed by some early scholars, reportedly took it upon himself to finish the ʿAyn,244 which his teacher Ḫalīl, according to another account, used to dictate to him.245 Accordingly, it can be safely concluded that data collec- tion, authoring single-topic monographs – or even multithematic books (cf. al-Ġarīb al-muṣannaf ) – and compiling lexica which are arranged based on lafẓ represent concurrent types of activity, and not consecutive stages in the history of Arabic lexicography.

On a larger scale, one of the main features of Arabic lexicography is the coexistence for several centuries of two types of dictionaries, the mubaw- wab and the muǧannas – two terms which we have adopted from Ibn Sīda’s (d. 458/1066) introduction to his al-Muḫaṣṣaṣ, in which he most clearly contrasts the two types and gives examples for each of them from previous

if this were the case, the dates of Māzinī, Siǧistānī and Riyāšī prove that data collection continued well after Ḫalīl’s death.

236 Abū l-Ṭayyib, Marātib 71. 237 Ḫalīl, ʿAyn II, 250; III, 316; V, 35, 234, 276; VIII, 368. 238 Ibn al-Nadīm, Fihrist 51. 239 Abū l-Ṭayyib, Marātib 71. 240 Zubaydī, Ṭabaqāt 157; Ibn al-Nadīm, Fihrist 49. 241 Ibn al-Nadīm, Fihrist 49. 242 Abū l-Ṭayyib, Marātib 71. 243 Ibn al-Nadīm, Fihrist 96. 244 Azharī, Tahḏīb I, 41. 245 Ibn al-Anbārī, Nuzha 45.

 

 

48 chapter 1

works.246 The mubawwab type refers to the onomasiological lexica or thesauri in which meaning leads to sign since they deal with specific topic areas. Several specialized works, however, are arranged according to form (lafẓ), although in the same genre one finds other works which do not follow such a method of arrangement. For practical reasons, books which adopt either method will be discussed within the same genre they represent. In the other type, which is called muǧannas or semasiological, sign leads to meaning and material is thus uniformly arranged according to principles related to form (lafẓ) and not meaning (maʿnā). The term muǧannas will be exclusively used for the general lexica which normally aim at lising all lexical items of the language and are thus not specialized dictionaries. That each of the two types serves a differ- ent purpose is strongly supported by the fact that some scholars authored works of both types. Ibn Durayd (d. 321/933), for example, is famous for his al-Ǧamhara, of the muǧannas type, but also actively contributed to writing in the thematic fashion (witness the host of chapters appended to that lexi- con). Following in Ibn Durayd’s footsteps, Ibn Sīda, and as late as the fifth/ eleventh century, authored two major lexica, one of which, the Muḫaṣṣaṣ is of the mubawwab type, and the other, the Muḥkam, is of the muǧannas type. He explicitly says that because the Muḥkam is a muǧannas lexicon which is intended to guide the searcher to the exact place in which a word is recorded (li-adull al-bāḥiṯ ʿalā maẓinnat al-kalima al-maṭlūba), he wanted to match it (aradtu an aʿdil bihi) with a lexicon of the mubawwab type, i.e. the Muḫaṣṣaṣ. He describes the latter as more appropriate for the orator (ḫaṭīb) and poet (šāʿir), since it gives them the opportunity to choose from among the various synonyms and adjectives listed the mot juste – the one that exactly suits their needs and fits the required rhyme in prose and poetry (saǧʿ aw qāfiya).247 It is also significant that some authors of muǧannas works derive their material from both types of lexica. For example, Ibn Fāris (d. 395/1004) in the introduc- tion of Maqāyīs al-luġa names five major sources for his lexicon, three of which are mubawwab (i.e. Abū ʿUbayd’s Ġarīb al-Ḥadīṯ and Muṣannaf al-ġarīb [i.e. al-Ġarīb al-muṣannaf ], and Ibn al-Sikkīt’s K. [Iṣlāḥ] al-manṭiq) and the other two muǧannas (i.e. Ḫalīl’s al-ʿAyn and Ibn Durayd’s al-Ǧamhara).248

The fact that each of the two types of lexica, the mubawwab and the muǧannas, serves a specific purpose not shared by the other is the main rea- son why they coexisted for so long in the Arabic tradition. Had this not been the case, one would have expected the earlier mubawwab type – which unlike

246 Ibn Sīda, Muḫaṣṣaṣ I, 10, 12. 247 Ibid., I, 10. 248 Ibn Fāris, Maqāyīs I, 4–5.

 

 

49early lexicographical activity

its counterpart is of little help for one who intends to search for the mean- ing of a particular word – to be discontinued in favor of the muǧannas type soon after the ʿAyn was authored, or at least soon after it came into circulation or other muǧannas type lexica, such as Ibn Durayd’s Ǧamhara, appeared. It seems that, with the expansion of Arabic as a spoken and literary language in vast areas of the Islamic world (Persia being a notable exception) in the wake of the conquests, and with Arabic gaining importance as the language of administration following the Arabization of the official registers (dawāwīn) in the reign of ʿAbdalmalik b. Marwān (d. 86/705; r. 65/685–86/705), scholars of the period responded to the growing need to study Arabic by focusing not only on the phonological, morphological and syntactical characteristics of the language – hence grammatical study – but also on words and their mean- ings. It is quite strange, however, that early lexica of both types, the mubaw- wab and the muǧannas, focused more on rare, even obsolete words than on more commonly used words in spoken or literary language. Eventually, some authors became aware of the need for lexica which focus on what is actually used rather than on obscure or uncommon words. Ibn Durayd, for example, declares in the introduction of the Ǧamhara that he chose its title to reflect the fact that he selected the bulk or “multitude” (ǧumhūr) of the speech of the Arabs and avoided the uncouth or barbarous (waḥšī) and the objectionable (mustankar).249 The “fact” which Ibn Durayd refers to, however, is more of a myth since the Ǧamhara is replete with words that are either rare, obsolete or of dubious meaning. Yet, his above quoted statement reflects an obvious aware- ness of the needs of users who want to learn about words that are in circula- tion and not those which sporadically occur (or rather occurred) in the speech of some Bedouin informant or in an obscure line of poetry. In this respect, Ibn Fāris was more faithful to his promise than Ibn Durayd. His Muǧmal al-luġa, as he promised it would be in his introduction,250 is largely free from obscure or uncommon words since he designed it more for the educated user than for scholars of Arabic; this was not the case with the ʿAyn and the Ǧamhara, as his veiled criticism of both implies. But irrespective of such differences among individual authors, the great variety in the lexicographical tradition, embod- ied mainly in the coexistence of the mubawwab and the muǧannas lexica, and also in the striking diversity of the topics examined in the former type, was in part a response to the growing demand for lexical works which, regardless of whether they are arranged according to topic or lafẓ, define meaning and adduce attested material which confirms correct usage. The complementarity

249 Ibn Durayd, Ǧamhara I, 41. 250 Ibn Fāris, Muǧmal 75–76.

 

 

50 chapter 1

of this branch of linguistic investigation and grammatical study will be exam- ined in Section 7 below.

It may be appropriate at this point to discuss the term muʿǧam (pl. maʿāǧim or muʿǧamāt),251 which the Arabs use for “lexicon”, particularly the muǧannas type. To begin with, the term qāmūs, which has largely replaced muʿǧām in modern usage, was introduced to lexicography only in the early ninth/ fifteenth century following the title that Fīrūzābādī (d. 817/1415) gave to his famous lexicon, al-Qāmūs al-muḥīṭ. It seems that the wide popularity which this lexicon enjoyed resulted in borrowing the word qāmūs (originally from Greek Ωκɛανός), a synonym for sea or, more specifically, its deepest part or its main body of water, for “lexicon” in general, usually with the connotation of completeness or exhaustiveness. How the term muʿǧam itself came to denote “lexicon” is less clear. The root of the word, ʿ ǦM, as Ibn Ǧinnī (d. 392/1002) notes by citing a number of its derivatives, has the basic meaning of causing confu- sion and obscurity (li-l-ibhām wa-l-iḫfāʾ) and is thus the antonym of distinct articulation and clear exposition (al-bayān wa-l-ifṣāḥ).252 The use of muʿǧam to denote “lexicon” raises the question of the semantic relationship between the general sense of the root ʿ ǦM and the notion of lexicon. Ibn Ǧinnī convincingly argues that the fourth form of the verb, i.e., aʿǧama whose passive participle is muʿǧam, indicates the removal of ʿuǧma (want of clearness or perspicuity). In other words, the hamza of the fourth form has the function of salb, that is negation or deprivation, as in aškā, which means to remove someone’s šakwā or cause of complaint.253 Furthermore, Ibn Ǧinnī notes that a synonym of the fourth form aʿǧama is the second form ʿaǧǧama, which also indicates salb, as does marraḍa, which can have a privative sense, that is, to remove someone’s illness or disease.254 This meaning of ʿaǧǧama interestingly occurs in the ʿAyn in relation with dotting. Ḫalīl explains that the taʿǧīm of a book means its dot- ting in order that its ʿuǧma turns to clarity and that it be set right (wa-taʿǧīm al-kitāb tanqīṭuhu kay tastabīn ʿuǧmatuhu wa-yaṣiḥḥ).255 Ḫalīl’s observation

251 Although some contemporary Arab scholars regard maʿāǧim as unacceptable on morpho- logical grounds, and insist that muʿǧamāt is the only correct plural of muʿǧam, Asad (1969: 76–88) persuasively argues that maʿāǧim is a perfectly acceptable form that conforms to the rules of plural formation.

252 Ibn Ǧinnī, Sirr I, 36. 253 Ibid., I, 37–39; cf. Astarābāḏī, Šarḥ I, 83, 91; Suyūṭī, Hamʿ II, 161. 254 For other examples of fourth form verbs which indicate salb, see Astarābāḏī, Šarḥ I, 92, 94;

Suyūṭī, Hamʿ II, 161. 255 Ḫalīl, ʿAyn (ʿǦM; I, 238). Note that the expression tastabīn ʿuǧmatuhu can only mean that

its ʿuǧma is reversed or dispelled, and cannot possibly mean that it becomes apparent or manifest (in which case taʿǧīm would not indicate salb).

 

 

51early lexicographical activity

provides an essential clue for the link we need to make between the priva- tive sense of the verb aʿǧama and the technical term muʿǧam. One possibil- ity we can suggest is that, as passive participle, muʿǧam originally referred to the letter of the alphabet which, through dotting or lack thereof, is manifestly distinguished from other letters of similar shape, as in dāl and ḏāl or ǧīm, hāʾ and ḫāʾ. Of course, the absence of a dot in, say, ḥāʾ and dāl (hence the desig- nation muhmal, lit. ignored) is in itself a marker since it distinguishes them from dotted letters that bear resemblance to their shape. In all cases, the use of muʿǧam as an adjective for a letter or a word (e.g. ḥarfun muʿǧamun; kalima- tun muʿǧamatun) might have been extended to the book in which these letters and words are assembled and arranged according to a particular order, hence kitābun muʿǧamun. The word was then used substantivally to refer to the book itself, as happened with so many other adjectives such as maʿrūf (known) and munkar (disapproved), which are passive participles that came to be used sub- stantivally to signify a good or an evil deed, as in al-maʿrūf wa-l-munkar.

The other possible explanation is the one expounded by Ibn Ǧinnī, who bases his argument on the meaning of the expression ḥurūf al-muʿǧam (let- ters of the alphabet). He argues that muʿǧam cannot be an adjective for ḥurūf, primarily because the former is indefinite and the latter definite. He further argues that, had it been an adjective, one would have expected it to be femi- nine, as in the expression taʿallamtu l-ḥurūf al-muʿǧama (I learnt the letters of the alphabet). After establishing that the two words have a construct/genitive relationship, he supports Mubarrad’s (d. 285/898) view that muʿǧam is a ver- bal noun which has the same status as iʿǧām (al-muʿǧam maṣdar bi-manzilat al-iʿǧām), just as mudḫal (letting in) and muḫraǧ (letting out) are verbal nouns which can replace idḫāl and iḫrāǧ, as in adḫaltuhu mudḫalan and aḫraǧtuhu muḫraǧan.256 The rarity of this morphological phenomenon257 however undermines Mubarrad’s and hence Ibn Ǧinnī’s explanation, motivated in the first place by a desire to avoid the assumption of the elision of a word from ḥurūf al-muʿǧam, which would otherwise have to be interpreted as a shortened form of ḥurūf al-kalām /al-lafẓ al-muʿǧam.

Morphological controversies aside, the term muʿǧam started to appear in the titles of books in fields other than lexicography as of the third/ninth century at the latest. We owe our earliest example to Ibn al-Nadīm (d. 380/990) who

256 Ibn Ǧinnī, Sirr 33–36. 257 Note that for his third example, Ibn Ǧinnī had to resort to the unorthodox reading of

mukram (i.e. verbal noun similar to ikrām, honoring) instead of the regular reading of mukrim (one who honors) in the verse wa-man yuhini l-Lāhu fa-mā lahu min mukrimin (and such as God shall disgrace, none can raise to honor; Q 22: 18).

 

 

52 chapter 1

ascribes to Buzurǧ al-ʿArūḍī, a contemporary of al-Faḍl b. Yaḥyā l-Barmakī’s (d. 193/808), a book on prosody entitled K. Maʿānī l-ʿarūḍ ʿalā ḥurūf al-muʿǧam.258 For his part, Yāqūt tells us that Ḥubayš b. Mūsā l-Ṣīnī (or l-Ḍabbī) authored for the Abbasid Caliph al-Mutawakkil (d. 247/861; r. 232/847–247/861) a book entitled K. al-Aġānī ʿalā ḥurūf al-muʿǧam.259 Naṣṣār260 rightly observes that such titles were reckoned to be too long and were shortened either by omitting the word muʿǧam, as in Qurayṣ al-Muġannī’s (d. 324/936) book K. Ṣināʿat al-ġināʾ wa-aḫbār al-muġannīn wa-ḏikr al-aṣwāt allatī ġunniya bihā ʿalā l-ḥurūf 261 (apparently with little effect on such a long title!), or by drop- ping ḥurūf and starting with muʿǧam, hence Muʿǧam al-Šuʿarāʾ, Muʿǧam al-Buldān, etc. Among the earliest books whose titles start with muʿǧam are Muʿǧam al-Ṣaḥāba by Abū Yaʿlā l-Mawṣilī (d. 307/919), al-Muʿǧam al-kabīr and al-Muʿǧam al-ṣaġīr (both on Ṣaḥāba) by Abū l-Qāsim al-Baġawī, known as Ibn Bint Manīʿ (d. 317/929; but note that his early date of birth 213/828 indicates that his books could have predated Abū Yaʿlā’s), al-Muʿǧam al-kabīr, al-Muʿǧam al-awsaṭ and al-Muʿǧam al-ṣaġīr (all on qurrāʾ and their qirāʾāt) by Abū Bakr al-Naqqāš (d. 351/962), and Muʿǧam al-Šuyūḫ by Ibn Marzūq al-Baġdādī (d. 351/962).262 Thus, by the fourth/tenth century, the use of muʿǧam in book titles was widespread. In lexicography, however, this seems not to have been the case as most of the lexica of the period did not include the term in their titles, e.g. Ǧamharat al-luġa by Ibn Durayd (d. 321/933), Tahḏīb al-luġa by Azharī (d. 370/981), and al-Ṣaḥāḥ by Ǧawharī (d. c. 400/1010). The word kitāb was often appended to these titles, hence Kitāb al-ʿAyn and Kitāb Ǧamharat al-luġa etc., a notable exception being Ibn Fāris’s (d. 395/1004) Maqāyīs al-luġa, which was referred to both as Kitāb Maqāyīs al-luġa and Muʿǧam Maqāyīs al-luġa.263 Even in later periods, muʿǧam was rarely appended to titles of lexica, hence neither Ibn Manẓūr’s (d. 711/1311) lexicon was normally referred to as Muʿǧam Lisān al-ʿArab, nor Zabīdī’s (d. 1205/1790) as Muʿǧam Tāǧ al-ʿarūs.

258 Ibn al-Nadīm, Fihrist 78–79. 259 Yāqūt, Muʿǧam II, 805. 260 Naṣṣār (1968: I, 13); cf. Āl Yāsīn (1980: 220). 261 Ibn al-Nadīm, Fihrist 173. 262 Cf. these, and other titles, in Naṣṣār (1968: I, 13–14); Āl Yāsīn (1980: 221); A. ʿAṭṭār (1990:

53–54). 263 Ibrāhīm al-Sāmarrāʾī’s claim, reported by his student M. Ḥ. Āl Yāsīn (1980: 222), that no

lexicon included the word muʿǧam in its title before Maqāyīs denies the possibility that the sources might have failed to attach muʿǧam to titles of other lexica as they often did with the word kitāb.

 

 

53early lexicographical activity

7 Remarks on Contemporary Scholarship and the Originality of Arabic Lexicography

The last few decades witnessed remarkable interest in the study of the Arabic Linguistic Tradition and a surge in the number of articles and books dealing with various aspects of this tradition. Since the 1960s, scholarly activity pri- marily focused on the study of Arabic grammar. In contrast, relatively modest progress took place in the study of lexicography despite the recent publication of essential texts and better editions of previously published ones. Up to the present day, the main book-length general study of the field in western schol- arship is J. Haywood’s Arabic Lexicography (first published in 1959; 2nd ed. 1965), now in many ways outdated. S. Wild (1965) is still useful as an introduc- tion to the field although it deals mainly with K. al-ʿAyn. As far as articles are concerned, there are a few early ones, e.g. F. Krenkow (1924) and E. Braünlich (1926), but these are now largely obsolete. Among the useful ones which sur- vey the history of lexicography or deal with general issues related to the field are J. Kraemer (1953), F. Rundgren (1973), H. Fleisch (1984), H. Gätje (1985), M. G. Carter (1990), A. Sanni (1992), and T. Seidensticker (2002, 2008). The vast- ness of the lexicographical tradition is demonstrated by two invaluable tools at the disposal of researchers, namely, F. Sezgin’s Geschichte des arabischen Schrifttums (vol. 8, 1982) and R. Weipert’s Classical Arabic Philology & Poetry: A Bibliographical Handbook of Important Editions from 1960 to 2000 (2002). R. Weipert (1989) also includes useful bibliographical material. The main refer- ence in Arabic is Naṣṣār’s al-Muʿǧam al-ʿArabī (1968; first published in 1956), a pioneering work which examines the various lexica in great detail and is thus an indispensable reference, although many of the sources which the author uses in manuscript form have since been published and others have become known. Moreover, it is to be noted that the book does not deal with a number of important genres, such as amṯāl (proverbs), aḍdād (words with two contra- dictory meanings), and the vast majority of subjects related to ḥurūf (particles) and aṣwāt (letters). There are as well several other books in Arabic (see bibli- ography), such as those written by M. Abū l-Faraǧ (1966), ʿU. Daqqāq (1968), ʿA. M. Aḥmad (1969), E. Yaʿqūb (1981), M. ʿUryān (1984), ʿA. Ṣūfī (1986), R. Qāsim (1987), B. ʿAṭṭār (1990), Y. ʿAbdallāh (1991), A. Bātilī (1992), ʿA. ʿAbdalǧalīl (1999), A. Rabīʿī (2001), M. Muḥammad (2001), ʿĪ. Barhūma (2005), ʿA. Ḥ. Ġurāb (2005), Ḫ. Nuʿmān (2007), and A. Maʿtūq (2008). But most of these are textbooks which merely provide an overview of the field and are thus of little scholarly value, or they draw heavily on Naṣṣār (some even plagiarize him264 or contain gross

264 Cf., in particular, Kišlī (1996a, b, c); Barhūma (2005). Parts of ʿ Īd (1992) are also plagiarized; cf., for example, the chapter on Qālī’s al-Bāriʿ (111 ff.) with Wadġīrī (1984: 20 ff.).

 

 

54 chapter 1

factual errors).265 In contrast, there are a few well-written books in Arabic, nota- bly ʿA. Darwīš’s pioneering, but now outdated book al-Maʿāǧim al-ʿArabiyya (1956), Ḥ. Naṣṣār’s study of several single-topic lexica (1984), and A. M. ʿUmar’s impressive study of one type of lexica, namely, those based on morphologi- cal patterns (1995). Also worthy of mention among recommended studies are the relevant parts of M. Ḥ. Āl Yāsīn (1980), R. ʿAbdaltawwāb (1980), ʿA. Wadġīrī (1984, 2008), M. R. Ḥamzāwī (1991), Ḥ. Ḫalīl (1997), and ʿA. Ǧubūrī (2004). A. Iqbāl’s Muʿǧam al-maʿāǧim (2011) includes useful bibliographical informa- tion on extant lexica as well as on those that are mentioned in the sources but did not reach us. The list by Āl Yāsīn (1981) is limited to authors whose dates of death do not go beyond 300/912.

Several of the above mentioned works discuss the question of foreign influ- ence on Arabic lexicography, and although I do not intend to deal extensively with a subject that hardly rests on any solid evidence, a few remarks may be in order before I try to demonstrate that the roots of Arabic lexicography are firmly based in the Arabic-Islamic culture, as is particularly evident in its rela- tionship with the other indigenous scholarly activity, the study of grammar.

Claims of foreign influence primarily invoke Indian and Greek parallels.266 The possibility of the Indian influence on Arabic lexicography stems from suggested parallels in the Arabic and Indian phonetic traditions, and draws support from Ḫalīl’s arrangement of the ʿAyn according to maḫāriǧ (pl. of maḫraǧ, point or place of articulation). Yet, proponents of the dependence of the Arabic phonetic teachings upon the Indians admit that it is “difficult to trace any direct influence (our italics)”.267 Closer examination of the system- atic parallels between the ancient Indian and the Arabic phonetic traditions, particularly that they both adopt an articulatory basis for the arrangement of letters and proceed at the source of the airstream and work outward, reveals that these parallels reflect a superficial resemblance and that the fundamental concepts which underlie the two approaches are substantially different.268 For instance, the notions of maḫraǧ and ḥayyiz (area of articulation) are static, but flexible, and hence are in contrast with the dynamic, but in its application rigid,

265 One example is that ʿAbdallāh (1991) lists Fārābī’s (d. 350/961) Dīwān al-adab with lexica that are arranged according to the first radical of the root (p.103), whereas in fact it follows the rhyme system as the author herself rightly mentions elsewhere (p. 173)!

266 Cf. Wild (1962; 1965); ʿUmar (1972, esp. 133–134, 142–153); Rundgren (1973); Danecki (1978; 1985); Law (1990); see also Weninger (1994).

267 Danecki (1985: 134); cf. Law (1990: 216). 268 Baalbaki (2007: xxiii).

 

 

55early lexicographical activity

opposition of sthàna (stationary place of articulation)269 and karaṇa (mov- ing articulation). As Law notes, “the borrowing of a system without any of the details on which it rests is almost unknown. Yet this is precisely what is alleged in recent works on the question of Indian influence in Arabic phonetics . . .”.270 Furthermore, the existence in Sanskrit of classified vocabularies similar to those in Arabic271 is no proof of influence. In fact, these do exist in other tra- ditions, including the Chinese,272 and reflect common interests of early phi- lologists rather than borrowing or imitation. This notwithstanding, Haywood astonishingly claims that, despite Ḫalīl’s originality, “it is too much to believe that his phonetic ideas were his own”, adding that they were possibly based on Sanskrit traditions!273 Such insistence on foreign influence is totally unjusti- fied since, as will be demonstrated shortly, there is strong reason to believe that Ḫalīl’s views are the result of his own probing intellect and experimentation.

As far as the Greek tradition is concerned, Rundgren suggests the possibility of direct influence based on the dominance of poetry in Greek linguistics and of poetry šawāhid in Arabic lexicography.274 This astonishing position in effect means that the indigenous roots of Arabic lexicography can only be proven if the Arabs did not use their vast corpus of poetry, to which they referred as dīwān al-ʿArab and the “depository” of their knowledge,275 in establishing the meaning of words or providing proof of actual usage! One certainly does not have to go, as did Rundgren, as far back as Aristotle to find possible origins for Ḫalīl’s terms ḏawāq (sampling; lit. tasting) and ǧars (ring) in γεΰσις and ψόφος respectively,276 when the introduction to the ʿAyn provides a much simpler and more natural explanation. In the first few lines of this introduction, we are told how Ḫalīl used to open his mouth with an alif and then pronounce the letters in the following fashion: ab, at, aḥ, aʿ, aġ, etc.277 This is a most natural way of discovering the exact points or places of articulation, and its experi- mental origins, which reveal close reliance on morphological considerations, are preserved in Ḫalīl’s own lively discussions with his pupils as reported in detail by Sībawayhi.278 Similarly, Ḫalīl’s comment that the ʿayn and qāf are

269 Law (1990: 223). 270 Ibid., 216. 271 Cf. ʿUmar (1972: 133–134). 272 Haywood (1965: 6). 273 Ibid., 26; cf. 37. 274 Rundgren (1973: 149–150). 275 See above, 32–33. 276 Rundgren (1973: 152). 277 Ḫalīl, ʿAyn I, 47. 278 Sībawayhi, Kitāb III, 320–321.

 

 

56 chapter 1

“the freest of letters and with the greatest ring” (aṭlaq al-ḥurūf wa-aḍḫamuhā ǧarsan)279 – which contains the only occurrence of ǧars in al-ʿAyn’s intro- duction – is simply a description of the effect those two sounds have on the bināʾ (structure; here word) of which they are part. This effect is expressed by a native speaker (assumed here to be Ḫalīl, but may also be a contempo- rary of his), who intuitively felt that the presence of ʿayn or qāf compensates for the absence of al-ḥurūf al-ḏulq (letters of fluency; i.e. rāʾ, lām and nūn) in words such as the quadriliterals ʿasǧad (gold) and qudāḥis (brave and strong).280 Furthermore, would one find Greek origins for the notion of ṭalāqa (freedom), ḍaḫāma (greatness), ṣalāba (stiffness), kazāza (brittleness), and ḫufūt (soft- ness), all which occur in the same location as ǧars and describe certain letters including ʿayn and qāf ? Finally on Greek influence, a curious anecdote quoted by Zubaydī (d. 379/989) has it that the Byzantine emperor (malik al-Yūnāniyya) sent Ḫalīl a letter written in Greek. After spending a whole month decipher- ing the letter, he was able to understand it ( fahimahu), having worked on the assumption that it must have started with the expression bi-smi l-Lāhi or the like.281 It is indeed strange that some tried to use this anecdote to prove that Ḫalīl mastered Greek and thus must have been influenced by the Greek linguistic tradition.282 As Haywood observes, “the whole story does not merit serious attention – though it would be most convenient to believe that al-Khalīl really learned Greek, and therefore could have had first-hand knowledge of Greek lexicography”.283 The basic flaw in this account is the claim that Ḫalīl was able to understand the text, and not merely decipher its characters. This is impossible to be the case, for even if one accepts that he was able to deter- mine the equivalence between Arabic and Greek characters, the claim that he understood the text can only lead to either of two ridiculous conclusions: that the text was Arabic transcribed in Greek, or that Ḫalīl was already cognizant of Greek but not of Greek characters!

It was also suggested that the rhyme system in which words are typically arranged according to the radicals of their roots in a rhyme order, that is, start- ing with the last consonant of the root, might have been influenced by a Hebrew dictionary known as the Agron.284 Authored by Saadia Gaon, also known as

279 Ḫalīl, ʿAyn I, 53; translation of this expression is adopted from Sara (1991: 14). 280 Cf. ibid., II, 315; III, 323. 281 Zubaydī, Ṭabaqāt 51; Ibn Nubāta, Sarḥ 268. 282 See a discussion of this view in Maḫzūmī (1986: 68–70); cf. Bustānī (1876–1900: VII, 461). 283 Haywood (1965: 21); cf. Talmon (1997: 47–48). 284 Haywood (1965: 68–69, 120–121); cf. ʿUmar (1972: 149); Āl Yāsīn (1980: 90); Fenton (1995:

661).

 

 

57early lexicographical activity

Saʿīd b. Yūsuf al-Fayyūmī (d. 331/942), it consists of two parts, the first of which is arranged according to initial letters, whereas the second is arranged accord- ing to final letters in order to help poets in their rhymes.285 The assumption was that this lexicon predates the first Arabic lexicon which follows the rhyme sys- tem, Dīwān al-adab, whose author, Fārābī (d. 350/961), was the maternal uncle of Ǧawharī’s (d. c. 400/1010), author of the famous al-Ṣaḥāḥ (itself long thought to have been the first lexicon to adopt the rhyme order). But we now know that Dīwān al-adab and al-Ṣaḥāḥ were preceded by Bandanīǧī’s (d. 284/897) K. al-Taqfiya (lit. Book of Rhyme), which is arranged according to the last radical of the root, but without any further arrangement based on the other radicals. Even Bandanīǧī is preceded by Ibn al-Sikkīt (d. 244/858), who arranged lexi- cal items in a lengthy chapter of Iṣlāḥ al-manṭiq based on their final letters.286 Saadia was born in 279/892, thirty-five years following Ibn al-Sikkīt’s death, and was a boy of five when Bandanīǧī died in 284/897. Accordingly, if the existence of influence had at all been possible (although no direct evidence exists for this assumption), then he must have been on the receiving end.

Proponents of foreign influence on Arabic lexicography in general and on K. al-ʿAyn in particular are totally oblivious of the fact that the Arabic sources are as completely silent about the presence of any foreign influence in lexi- cography as they are about any such influence in grammar. Even if there had been a conspiracy among early philologists (and faithfully continued by later scholars) to be silent about any foreign influence, one would have expected at least the faintest allusion to such a major scheme. To be sure, indirect influ- ence is always theoretically possible, but the firm evidence needed to prove direct influence on Arabic lexicography is totally lacking. More importantly, that Arabic lexicography is the fruit of indigenous endeavors is strongly sup- ported by the host of information about early data collection and the distinc- tiveness of the Arab philologists’ interest in rare and uncommon words, more than in “ordinary” usage. This is witnessed in their interest in interpreting Qurʾānic ġarīb and providing šawāhid derived from the speech of the Bedouin fuṣaḥāʾ. But perhaps the most compelling evidence of the independent ori- gin of Arabic lexicography comes from the available early texts, specifically al-ʿAyn’s introduction, and from the close affinity between lexicography and grammar, in particular as recent researchers have largely given up on theories of foreign influence on grammar, in favor of tracing its origin to

285 For more on Gaon as a philologist and grammarian, see Malter (1921) and Skoss (1955). 286 See below, 365.

 

 

58 chapter 1

other native scholarly activity, such as Qurʾānic commentaries, as Versteegh demonstrated.287

The introduction of the ʿAyn, whether its bulk was actually written by Ḫalīl himself or by a student of his who reported his views, strongly suggests that it is the result of an effort at finding ways to describe Arabic phonetics in order to arrange the letters of the alphabet according to some principle; determine the number of radicals in Arabic words; arrange radicals in a manner which helps the user find the word he needs to check; and make sure that this arrange- ment can exhaust all the words used by Arabic speakers. The methods adopted by Ḫalīl to achieve each of these objectives is expounded in a manner which betrays his innovative approach to issues not previously discussed, at least not with the aim of drawing the plan for an exhaustive lexicon. This is clear, for example, in statements describing how be reversed (qallaba) the order of alif, bāʾ, tāʾ, ṯāʾ, etc. to yield an arrangement which starts from the throat (fa-waḍaʿahā ʿ alā qadr maḫraǧihā min al-ḥalq), how he classified words into bil- iterals, triliterals, quadriliterals and quinqueliterals with the intention of sepa- rating entries based on morphological considerations, and how he exhausted the various possible arrangements of the radicals of these biliterals, triliter- als, etc. with the aim of including their permutations under one heading.288 In fact, the first few lines of the introduction explicitly mention how Ḫalīl put his mind to work ( fa-aʿmala fikrahu)289 in order to plan for his lexicon. Neither such statements nor the previously mentioned text of the Kitāb which preserves the lively and exploratory discussion that took place between Ḫalīl and his students concerning the sampling (ḏawāq) of letters (ḥurūf ) and the morphological justification for isolating letters to be sampled is suggestive of a readily available model to be emulated or carried over into Arabic. In this respect, the ʿAyn is quite similar to its counterpart in grammar, the Kitāb, whose first few chapters (known as the Risāla) in particular are obviously the result of an effort to lay the basis of a grammatical theory and probe the characteristics of some of the essential components of kalām, such as the parts of speech and the case endings, rather than the adoption of a foreign model of linguistic study to be applied to Arabic. Just as one can trace in the Kitāb the develop- ment to a technical term of a word that is still used in its original or general

287 Versteegh (1990; 1993). For a discussion of recent scholarship on foreign influences on grammar, see Baalbaki (2007: xx–xxvii).

288 Ḫalīl, ʿAyn I, 48, 59. 289 Ibid., I, 47.

 

 

59early lexicographical activity

meaning,290 it is manifest in the ʿAyn that many of the words used to describe the sounds of Arabic may have already achieved the status of technical terms, but they still preserved their general, nontechnical meanings. The choice of these terms is the result of an intuitive approach that rests on a refined sense of discovering phonetic traits and not on an available system that merely awaits to be Arabized. As pointed out earlier, notions such as ṭalāqa, ḍaḫāma, ṣalāba, kazāza and ḫufūt (cf. also buḥḥa “hoarseness”, hatta/hahha “tenseness”, līn “softness”, hašāša “delicacy”, and naṣāʿa “clarity”) are certainly not the result of translating foreign terms (if these exist anywhere else to begin with), but the product of a fresh attempt that mainly relies on “intuition” and Sprachgefühl in studying Arabic sounds as the necessary step towards exhausting all the theo- retically possible roots, and proceeds from there to explain words that do occur in the speech of the Arabs.

The close affinity between early lexicography and grammar is a reflection of the fact that both fields of linguistic investigation share much of the com- mon background expounded in the previous sections. Both fields are closely related to the study of the Qurʾān and, to a lesser extent, the Ḥadīṯ, and share a common corpus which draws heavily on poetry material and deems the usage of the Bedouin fuṣaḥāʾ as the “purest” form of Arabic speech. These two fields have also witnessed almost concurrently in the second half of the second/eighth century the authoring of a major authoritative work which set the trend for later scholars. Interestingly enough, the first such work in gram- mar is authored by Sībawayhi but is heavily influenced by Ḫalīl’s views,291 and its counterpart in lexicography, the ʿAyn, also bears the mark of Ḫalīl as we shall see later.292 In many respects, the introduction to the ʿAyn resembles the Kitāb’s Risāla since both discuss basic concepts essential for the whole work, and Ḫalīl’s naḥārīr bear strong resemblance to Sībawayhi’s naḥwiyyūn. More broadly, the two works are similar in that their authors are equally keen to determine the characteristics of Arabic usage in order to isolate artificially cre- ated words or constructions as well as foreign borrowings in phonetics and vocabulary. In the Kitāb, this intention is mostly implied, but is sometimes explicitly expressed in Sībawayhi’s position vis-à-vis the neologisms of the naḥwiyyūn and the chapters or parts thereof in which he discusses foreign

290 An example is the word muḍāriʿ which is used both in the general sense of “similar to” or “resembling” – e.g. al-afʿāl al-muḍāriʿa li-asmāʾ al-fāʿilīn (verbs which resemble active participles) – and as a technical term which refers to the imperfect – e.g. al-muḍāriʿ min al-afʿāl (imperfect verbs); cf. Sībawayhi, Kitāb I, 13–14.

291 For Ḫalīl’s influence on Sībawayhi’s Kitāb, cf. Baalbaki (2008: 16–18). 292 See below, 282 ff.

 

 

60 chapter 1

words, phonetic changes in Arabized Persian words, and the seven sounds (ḥurūf ) – most probably of foreign origin – which he describes as inappropri- ate (ġayr mustaḥsana) in Qurʾānic reading and poetry.293 Ḫalīl more directly refers to the distinction between what is Arabic and what is not. For example, he tries to establish general phonotactic principles which characterize Arabic and whose absence from a word reveals its alien origin.294 On one occasion, he explicitly states that the examples he cites are artificially created in order that the authentic Arabic structures be distinguished from the foreign ones (li-yuʿraf ṣaḥīḥ bināʾ kalām al-ʿArab min al-daḫīl).295

Throughout the long history of the grammatical and lexicographical tra- ditions, the two fields, visibly, had a complementary relationship. One of the prime objectives of the present book is to bring into light the complementarity of the two fields as well as the effect of early grammatical works on the content of mubawwab lexica in particular. For example, lexical works which deal with the meaning of ġarīb words make up for Sībawayhi’s almost total disregard for the meaning of ġarīb, or even extremely rare words which he cites as part of his discussion of appending (ilḥāq);296 e.g. qahbalis (huge woman), ǧaḥmariš (old woman), ṣahṣaliq (vehement voice), bulaʿbīs (wonder, marvel), duraḫmīl (calamity), qarṭabūs (calamity), hammariš (old woman), and hummaqiʿ (fruit of a thorny tree).297 Furthermore, much of Sībawayhi’s morphological mate- rial is adopted and expanded by authors of mubawwab lexica dealing with subjects such as muḏakkar and muʾannaṯ (masculine and feminine), maqṣūr and mamdūd (abbreviated and prolonged patterns), and verbs of the first and fourth patterns faʿala and afʿala, as well as by authors of lexica which are arranged according to morphological patterns, such as Saraqusṭī’s (d. after 400/1010) K. al-Afʿāl. In fact, the two traditions of grammar and lexicography are best understood when studied as part of the wider framework of Arabic linguistic activity, rather than isolated fields, particularly because they share a common background and much of their data and subject matter, and because several authors were both grammarians and lexicographers.

293 Sībawayhi, Kitāb VI, 303–304, 305–307, 432 respectively. 294 See above, 28. 295 Ḫalīl, ʿAyn I, 54. 296 Ilḥāq is an analytical morphophonological process through which an expanded form is

derived on the basis of another existing pattern (wazn). In the case of Arabized words, forms are said to be of the same pattern as their Arabic counterparts, such as bahraǧ, whose pattern is the same as salhab (Sībawayhi, Kitāb IV, 303). For a detailed study of ilḥāq in the grammatical tradition, see Baalbaki (2001–2002: 1–25).

297 Sībawayhi, Kitāb IV, 302–303, 330.

 

 

61early lexicographical activity

The next two chapters survey the development of the various genres of lexical writing. Chapter 2 is devoted to the mubawwab type, and to each of its genres is assigned a special section which typically traces its historical devel- opment and highlights the major works which represent it. The relationship between the genre and grammatical study and the former’s contribution to the muǧannas type will also be examined, wherever possible. The last section in this chapter deals with multithematic works, such as Abū ʿUbayd’s (d. 224/838) al-Ġarīb al-muṣannaf and Ibn Sīda’s (d. 458/1066) al-Muḫaṣṣaṣ. Chapter 3, on the other hand, is a study of the muǧannas lexica, beginning with Ḫalīl’s (d. 175/791) al-ʿAyn and ending with Zabīdī’s (d. 1205/1790) Tāǧ al-ʿarūs. It is divided into three parts, based on the various systems of word arrangement, namely, the phonetic-permutative system, the alphabetical system and the rhyme system. The number of lexica examined in this chapter is far less than in the previous one, but because they are general, unspecialized lexica, the material included in most of them substantially exceeds the material which the mubawwab type normally includes. Each lexicon will be discussed sepa- rately, with special reference to the overall arrangement of its material, the internal arrangement of the lemmata, its aim, scope of attested material, and influence on later works. Whereas in Chapter 2 we shall rarely go beyond the tenth/sixteenth century, we shall exceed that limitation in Chapter 3 so as to include one of the major lexica in the tradition, Zabīdī’s Tāǧ al-ʿarūs. Strictly speaking, the latter does not belong to the medieval era, but it will be included not only because of its massive size and wide reputation, but also because its faithfulness to the approach of the earlier authors firmly places it in the medi- eval tradition. On the other hand, certain genres that are not philologically ori- ented will be excluded despite their arrangement in the form of lexica. These include works on buldān (cities, regions, etc.) – such as Yāqūt’s (d. 626/1229) Muʿǧam al-Buldān – and some works on ḥayawān (animals) – such as Damīrī’s (d. 808/1405) Ḥayāt al-ḥayawān al-kubrā – as well as lexica of philosophical terms, medical terms, names of drugs, etc. and the numerous biographical works which are either restricted to a specific group – such as poets, judges, Ḥanābila, Ṣūfīs, physicians, etc. – or are less specialized – such as Ibn Ḫillikān’s (d. 681/1282) Wafayāt al-aʿyān and Ṣafadī’s (d. 764/1363) al-Wāfī bi-l-Wafayāt. Finally, bilingual lexica and glossaries are by no means common and do not form an integral part of the tradition. Yet mention will be made of those that clearly belong to a certain genre, as in the case of Zamaḫšarī’s (d. 538/1144) Arabic-Persian lexicon Muqaddimat al-adab, which deals with nominal and verbal patterns.298

298 See below, 265.

 

 

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004�740�3_�03

chapter 2

Mubawwab (Onomasiological) and Specialized Lexica

1 Introduction

In the preceding chapter we discussed early philological activity and the fac- tors that led to the emergence of lexical writing. We also discussed the process of data collection and concluded that the philologists regarded the speech of the Bedouin fuṣaḥāʾ as the “purest” form of Arabic and that much of their effort concentrated on strange and uncommon usage. In this and the following chap- ter we shall trace the historical development of the two types of lexica which began to appear in the second/eighth century, namely, those of the mubawwab (onomasiological) type and the muǧannas (semasiological) type. The present chapter is devoted to the first of these two types, which we shall generically refer to as mubawwab. It embraces lexica and thesauri in which meaning leads to sign, and includes specialized dictionaries, such as those that deal with Arabized words, solecism, morphological patterns, etc. Each of the next nine sections of the chapter will deal with a particular genre of single-topic lexica, whereas Section 11 discusses multithematic works, some of which are compre- hensive thesauri. As we have seen above, each of the two types, the mubaw- wab and the muǧannas, serves a different purpose, and the appearance of the muǧannas type shortly after mubawwab lexica did not result in their discon- tinuation; accordingly both types coexisted throughout the tradition. Indeed, some works which are devoted to particular subjects – such as Qurʾānic ġarīb or proverbs – were themselves arranged according to form (lafẓ, e.g. alphabeti- cally), and these, along with books of the same genre which are not formally arranged, will be considered in their respective sections in this chapter. In con- trast with muǧannas lexica which are relatively few in number, there is a vast literature of mubawwab writings, though a sizable portion of which is made up of short risālas restricted to specific topics. The vastness of this literature dictates that our survey cannot always be exhaustive (in spite of the fact that this chapter is considerably longer than the one on muǧannas lexica). The dis- cussion shall include all the sources that we believe are representative of each genre, particularly those that had the greatest impact on that genre and on subsequent authors.

 

 

63MUBAWWAB AND SPECIALIZED LEXICA

2 al-Ġarīb (Strange Usage) and al-Nawādir (Rare Usage)

As mentioned earlier, there was clear interest in ġarīb and nādir material at a very early stage of philological activity. It should be remembered that the phi- lologists and grammarians alike cast their net very wide indeed so as to include in the corpus of what they call ʿArabiyya a large variety of dialects which natu- rally comprise a great deal of irreconcilable characteristics and idiosyncrasies. For example, Sībawayhi (d. 180/796) refers to a dozen or so dialects,1 most nota- ble among which are the Ḥiǧāzī and the Tamīmī ones, each of which covers substantial areas of the peninsula and most probably comprises a number of sub-dialects.2 Other dialects cited by Sībawayhi include those of Asad, Bakr b. Wāʾil, Fazāra, Ġaniyy, Ḫaṯʿam, Huḏayl, Kaʿb, Qays, Rabīʿa, Saʿd, Sulaym and Ṭayyiʾ. Similarly, lexical works of the mubawwab type include a huge body of dialectal data which constitutes a large portion of the material on ġarīb and nādir. This is due to the fact that the philologists who collected the data from the tribes were keen to record dialectal variants and lexemes peculiar to spe- cific tribes and not known to most others. In this respect, the philologists drew most of their data from poetry, in particular raǧaz. The conservative nature of poetry must have contributed to its status as the ultimate source of strange or uncommon usage, and the philologists and grammarians expressed their awareness of this special status not only through the concept of poetic license (iḍṭirār, ḍarūra),3 but also by trying to interpret and justify aberrant material which poetry, in contrast to other genres, preserves.4

1 Cf. Baalbaki (2008: 38–40). 2 Note also that the dialects of Ḥiǧāz and Tamīm can agree on one aspect of a larger issue

on which they differ. Thus, the pattern faʿāli in the Ḥiǧāzī dialect always ends with a kasra, whereas it is normally treated as a diptote in the Tamīmī dialect, except in the case of a final -r, which takes a kasra throughout as in Ḥiǧāzī usage. In some cases, even a final ḍamma is attested in words of this pattern ending with -r, as in a line by Aʿšā. See Sībawayhi, Kitāb III, 277–279; cf. Baalbaki (1990: 21, 25).

3 See, for example, Sībawayhi’s chapter on poetic license, entitled bāb mā yaḥtamil al-šiʿr (Kitāb I, 26–32). Cf. above, 32.

4 Ibn Ǧinnī (d. 392/1002), for example, cites the occurrence, mainly in poetry, of anomalous forms, such as ḍaninū instead of ḍannū (they grudged) and aṭwalti instead of aṭalti (you pro- longed), and formulates the principle that they are an indication (manbaha) of the original forms which preceded the introduction of change peculiar to their class of words (Ḫaṣāʾiṣ I, 160–161, 256–264). For further examples and for the role of the dialects in anomalous forms, see Baalbaki (2005: 89–95).

 

 

64 chapter 2

Given the early interest in strange and uncommon usage, it is not surprising that the lexicographical tradition abounds with works related to this phenom- enon. We can distinguish three genres of writing which express the lexicogra- phers’ concern for studying what they considered to be ġarīb, namely, ġarīb al-Qurʾān, ġarīb al-Ḥadīṯ, and the general ġarīb/nādir material. This section will deal with the three genres in this order.

2.1 Ġarīb al-Qurʾān The Qurʾānic text is one of the four major sources from which the philologists and grammarians derive their corpus of attested usage, and we have already established that the earliest stages of linguistic enquiry were closely related to the study of the Qurʾān. Yet the ascription of works which deal with ġarīb al-Qurʾān to an early figure such as Ibn ʿAbbās (d. 68/687) should be treated with extreme caution. This certainly applies to material related by ʿAlī b. Abī Ṭalḥa (d. 143/760) on the authority of Ibn ʿAbbās, quoted by a much later source, Suyūṭī (d. 911/1505),5 as well as to the previously quoted masāʾil about which Nāfiʿ b. al-Azraq (d. 65/685) supposedly asked Ibn ʿAbbās.6 The authen- ticity of the ascription of such material to Ibn ʿAbbās is questionable, given that the accounts of the encounter between him and those who questioned him on a number of Qurʾānic, jurisprudential and philological issues were obvi- ously meant by later authors to ascribe a special status to Qurayš, as is clear in Iṣfahānī’s (d. 430/1038) reference to the pride which Qurayš should take in Ibn ʿAbbās’s maǧālis.7 Furthermore, it is inconceivable at this very early stage that there was a systematic attempt to exhaust ġarīb al-Qurʾān to the extent reported by later scholars. If Ibn ʿAbbās had anything to do with works such as Masāʾil Nāfiʿ b. al-Azraq, Ġarīb al-Qurʾān and al-Luġāt fī l-Qurʾān (for the latter, see Section 5 below), it would much more likely be that later authors – such as Ibn Ḥasnūn (d. 386/996) in the case of the book on luġāt – assembled under these titles material on Qurʾānic ġarīb, some of which they were able to trace back to Ibn ʿAbbās. Some other accounts may well have been intentionally ascribed to him, being the earliest recognized authority on the subject of ġarīb. For similar reasons, the ascription to the Imām of the Zaydiyya, Zayd b. ʿAlī (d. 122/740), of Tafsīr ġarīb al-Qurʾān,8 is dubious, in spite of the claim of its edi- tor that its ascription to Zayd is genuine and that it is the first complete book

5 Suyūṭī, Itqān II, 5 ff. 6 See above, 39. 7 Iṣfahānī, Ḥilya I, 320–321. 8 Cf. Madelung (1965: 57).

 

 

65MUBAWWAB AND SPECIALIZED LEXICA

on ġarīb that has reached us.9 Books ascribed to Ibn ʿAbbās and Zayd b. ʿAlī aside, the first possibly authentic book that we know of is al-Ġarīb fī l-Qurʾān by the Kufan Abān b. Taġlib al-Bakrī al-Ǧurayrī (d. 141/758). Although it did not reach us, the book is briefly described by Yāqūt (d. 626/1229), who says that it contains poetry šawāhid and that ʿAbdalraḥmān b. Muḥammad al-Azdī al-Kūfī (d. ?) combined it with other books in one book of his own.10 We also have no reason to doubt the ascription of books on Qurʾānic ġarīb to several prolific authors of the second half of the second/eighth century or early third/ninth century, although they are lost to us. Among these authors, some of whose other works are extant, are Kisāʾī (d. 189/805), Muʾarriǧ al-Sadūsī (d. 195/810), Yaḥyā b. al-Mubārak al-Yazīdī (d. 202/818), al-Naḍr b. Šumayl (d. 203/819), Abū ʿUbayda Maʿmar b. al-Muṯannā (d. 209/824), Abū ʿUbayd al-Qāsim b. Sallām (d. 224/838), and Muḥammad b. Sallām al-Ǧumaḥī (d. 232/846). To Aṣmaʿī (d. 216/831) is also ascribed a book entitled Ġarīb al-Qurʾān, but this contrasts with the previously cited riwāyas which suggest that he was vehemently opposed to attempts by other philologists to explain the meaning of Qurʾānic words.11 If to books titled Ġarīb al-Qurʾān up to the end of the third/ninth century are added those that deal with maʿānī/maǧāz al-Qurʾān in the same period, we end up with about thirty titles, only a handful of which are extant.

In this same period during which lexical works dealing with Qurʾānic ġarīb visibly prospered, the grammarians examined a large body of Qurʾānic šawāhid and meticulously commented on their syntactic peculiarities. Although this type of study is not traditionally identified with the study of ġarīb, it certainly focused on those syntactic characteristics of the Qurʾān which required expla- nation or interpretation. This approach is best represented in Sībawayhi’s al-Kitāb, as well as in a number of linguistically oriented exegetical works, most notably Farrāʾ’s (d. 207/822) Maʿānī l-Qurʾān, Abū ʿUbayda’s Maǧāz al-Qurʾān, and al-Aḫfaš al-Awsaṭ’s (d. 215/830) Maʿānī l-Qurʾān. The obvious difference in subject matter between these works and lexical works on Qurʾānic ġarīb dem- onstrates the complementarity of the grammatical and lexicographical tradi- tions – a theme to be highlighted in upcoming sections.

The most important extant sources devoted to Qurʾānic ġarīb as of the third/ ninth century are the following:12

9 Zayd b. ʿAlī, Tafsīr 14. 10 Yāqūt, Muʿǧam I, 38. 11 See above, 41–42. 12 The book titled Ġarīb al-Qurʾān by al-Qāḍī ʿIyāḍ (d. 544/1149) is not included in this list

because it is not an original work; rather it is a collection of ʿIyāḍ’s views on ġarīb derived from his other works and arranged alphabetically by the editor; cf. Rabābiʿa (2010).

 

 

66 chapter 2

1. Ġarīb al-Qurʾān wa-tafsīruhu by ʿAbdallāh b. Yaḥyā b. al-Mubārak al-Yazīdī (d. 237/851).

2. Tafsīr ġarīb al-Qurʾān by Ibn Qutayba (d. 276/889). 3. Ġarīb al-Qurʾān or Nuzhat al-qulūb by Muḥammad b. ʿUzayr (or ʿUzayz)

al-Siǧistānī (d. 330/941). 4. Yāqūtat al-ṣirāṭ fī tafsīr ġarīb al-Qurʾān by Abū ʿUmar al-Zāhid, better

known as Ġulām Ṯaʿlab (d. 345/957). 5. K. al-Ġarībayn fī l-Qurʾān wa-l-Ḥadīṯ by Abū ʿ Ubayd al-Harawī (d. 401/1011). 6. al-ʿUmda fī ġarīb al-Qurʾān and Tafsīr al-muškil min ġarīb al-Qurʾān

al-ʿaẓīm ʿ alā l-īǧāz wa-l-iḫtiṣār by Makkī b. Abī Ṭālib al-Qaysī (d. 437/1045). 7. al-Mufradāt fī ġarīb al-Qurʾān by al-Rāġib al-Iṣfahānī (d. c. 502/1108). 8. Nafas al-ṣabāḥ fī ġarīb al-Qurʾān wa-nāsiḫihi wa-mansūḫihi by Abū Ǧaʿfar

al-Ḫazraǧī (d. 582/1187). 9. Taḏkirat al-arīb fī tafsīr al-ġarīb by Ibn al-Ǧawzī (d. 597/1201). 10. Tuḥfat al-arīb bi-mā fī l-Qurʾān min al-ġarīb by Abū Ḥayyān al-Andalusī

(d. 745/1344).

The list begins with Yazīdī’s book based on the assumption that its author is Abū ʿAbdalraḥmān ʿAbdallāh b. Yaḥyā b. al-Mubārak al-Yazīdī, and not Abū ʿAbdallāh Muḥammad b. al-ʿAbbās al-Yazīdī (d. 310/922 or 313/925). This is supported by evidence provided by Ibn al-Nadīm (d. 380/990) and Qifṭī (d. 646/1248), both of whom ascribe to Abū ʿAbdalraḥmān a book entitled Ġarīb al-Qurʾān.13 As far as the arrangement of material in the above listed sources is concerned, Yazīdī’s book as well as Ibn Qutayba’s follow the order of the Qurʾānic sūras and verses, and thus serve the user who reads the Qurʾān in sequence or knows the exact place in which the word occurs in the Qurʾān, more than the one who needs to check the meaning of a certain word irre- spective of its place in the Qurʾān. In this respect both books resemble the lin- guistically oriented exegetical sources (such as Farrāʾ’s and al-Aḫfaš al-Awsaṭ’s Maʿānī l-Qurʾān) which also follow the arrangement of the Qurʾānic text, but for a better reason since they primarily discuss morphological and syntacti- cal issues which are difficult to be arranged alphabetically, and only occa- sionally give the meanings of individual words. This method of arrangement persisted in the genre of Qurʾānic ġarīb even after earlier authors of the genre had attempted to follow some sort of organization based on root. Thus the above-cited titles by Makkī b. Abī Ṭālib, Ḫazraǧī and Ibn al-Ǧawzī are arranged

13 Ibn al-Nadīm, Fihrist 37, 56; Qifṭī, Inbāh II, 151. See other arguments in support of the book’s ascription to Abū ʿAbdalraḥmān in M. S. al-Ḥāǧǧ’s introduction to his edition of the book, 18–21.

 

 

67MUBAWWAB AND SPECIALIZED LEXICA

according to the order of the Qurʾānic sūras and verses although Muḥammad b. ʿUzayr al-Siǧistānī had introduced early in the fourth/tenth century a different arrangement, albeit not fully alphabetical. According to this arrangement, only the first letter of the word is taken into consideration, and the words are listed in the form they are used in the Qurʾān without shedding their morphologi- cal augments (e.g. yaʿba⁠ʾu, yahīmūna, yastaṣriḫuhu, ya⁠ʾtamirūna, yakfulūnahu, yarbū, etc. are all listed under yāʾ).14 Moreover, Siǧistānī separates words which begin with the same letter according to the vowel of that letter, starting with fatḥa, followed by ḍamma and kasra. Accordingly, akama, aḥassa, aḥṣanna, azlām, aḥbār, anbāʾ, akinna, asāṭīr etc. (which are listed in this order since only the a- is considered) occur before ummiyyūna, ušribū, uǧūrahunna, ulāti, uffin, umm al-qurā, etc. and these before ihdinā, ihbiṭū, istaǧāba, iʿṣār, ilḥāf, inǧil, etc.15 He also lists words which begin with lām alif  in a separate chapter. It is note- worthy that about five centuries later, Ibn al-Hāʾim (d. 815/1412) rearranged Siǧistānī’s book according to the order of the sūras of the Qurʾān. With only few additions, as he explains in the introduction,16 his book is titled al-Tibyān fī tafsīr ġarīb al-Qurʾān.

Full alphabetical order on the basis of the roots of words was used for the first time in this genre, at least in the works we know of, by Abū ʿUbayd al-Harawī who combines in his K. al-Ġarībayn, as its title indicates, both ġarīb al-Qurʾān and ġarīb al-Ḥadīṯ.17 In his introduction, Harawī indicates that he adopted this arrangement in order to facilitate the use of his book.18 No separation in his chapters is made between the words in the Qurʾān and those in the Ḥadīṯ, the only criterion being the alphabetical order of the roots of words in both. al-Rāġib al-Iṣfahānī also arranges the Qurʾānic words in his al-Mufradāt based on the same criterion, although some of the defective, gem- inate and hamzated roots are not always listed based on unified norms.19 But it seems that, as the arrangement of the roots according to their final radicals became more widespread in muǧannas dictionaries, the system adopted by Harawī and Rāġib, which starts with the first radical, was abandoned by some authors in favor of this rival system. We are told that Muḥammad b. Abī Bakr al-Rāzī (d. after 666/1268) arranged his book on Qurʾānic ġarīb according to

14 Siǧistānī, Ġarīb 281. 15 Ibid., 8–10, 33–37, 38–41. 16 Ibn al-Hāʾim, Tibyān 43. 17 For other authors who are reported to have written books on both kinds of ġarīb, see Iqbāl

(2011: 35–36). 18 Harawī, Ġarībayn 35. 19 Cf. Naṣṣār (1968: I, 44).

 

 

68 chapter 2

Ǧawharī’s (d. c. 400/1010) rhyme system in al-Ṣaḥāḥ,20 that is, according to the final radical. Having himself abridged the Ṣaḥāḥ in the famous Muḫtār al-Ṣaḥāḥ, it was probably natural that Rāzī would make that choice. Finally, it is rather strange that as late as the eighth/fourteenth century, long after the use of the full alphabetical order became the norm in this genre and in muǧannas lexica (irrespective of whether the arrangement is by the first or last radical), Abū Ḥayyān opted to use a two-thirds alphabetical system which takes into consideration only the first radical, followed by the last. In the chapter of qāf, for instance, triliteral roots ending with rāʾ are arranged in the following order: qtr, qṭr, qṣr, qrr, qsr, qdr, and qbr,21 whereas their proper order in a fully devel- oped alphabetical system would be qbr, qtr, qdr, qrr, qsr, qṣr and qṭr.

The above sources on Qurʾānic ġarīb differ less in explaining words than in their order of arrangement. The works of Yazīdī, Siǧistānī, Zāhid, Makkī b. Abī Ṭālib and Abū Ḥayyān are relatively concise, and they generally contain very few quotations from earlier scholars, although Zāhid frequently quotes his teacher Ṯaʿlab. In particular, Abū Ḥayyān explains words very briefly and only rarely comments on their morphology or cites their derivatives. But when it comes to brevity of explanation, Makkī b. Abī Ṭālib is unmatched, since more often than not he explains a word merely by citing another single word, in effect making the ʿUmda a book of synonyms. In his other book, Tafsīr al-muškil, Makkī explains words more elaborately although he often explains a word by merely citing its synonym. Harawī often quotes earlier scholars with full isnād and delves into semantic issues largely ignored by other authors of the genre, such as his detailed study of the various connotations of the prepo- sition bi-.22 This has certainly contributed to the large size of his book which is already a combination of the ġarīb of the Qurʾān and the Ḥadīṯ together. Both Ḫazraǧī and Ibn al-Ǧawzī include material on nāsiḫ and mansūḫ (the abrogat- ing and the abrogated), and the latter explains that his book is the only one in the genre which deals with what is ġarīb related to both form and meaning as opposed to earlier works that are restricted to what is ġarīb in form.23

Each of the two books by Ibn Qutayba and Rāġib has specific characteris- tics in its method of explanation to justify distinguishing it from other works. Ibn Qutayba’s method differs from that of other authors in that he cites a siz- able number of šawāhid from poetry, proverbs and the speech of the Arabs.24

20 Ḥāǧǧī Ḫalīfa, Kašf II, 1208. 21 Abū Ḥayyān, Tuḥfa 256–258. 22 Harawī, Ġarībayn I, 238–241. 23 Ibn al-Ǧawzī, Taḏkira 13. 24 See the indices of Ibn Qutayba’s Tafsīr, 559–574.

 

 

69MUBAWWAB AND SPECIALIZED LEXICA

What makes his book even more distinctive are the two chapters which pre- cede the main body of the book in which the sūras are arranged according to their order in the Qurʾān. In the first of these two chapters, Ibn Qutayba deals with the morphology and meaning of the names (asmāʾ) and attributes (ṣifāt) of God, and in the second, he explains words which frequently occur in the Qurʾān and are thus not specific to any particular sūra, such as ǧinn (genii), ins (mankind), malāʾika (angels), širk (polytheism), sulṭān (authority), etc.25 Although grouping such words in a special chapter relieves Ibn Qutayba from mentioning them in the various locations in which they occur in the sūras, it complicates the task of the reader for whom the book is devised, that is, the one who follows the order of the sūras and verses in searching for the meaning of a specific word. It is also worth mentioning that in the introduction he says that he does not intend to deal with problematic issues in the Qurʾān (dūna ta⁠ʾwīl muškilihi) since he has devoted a whole book for that purpose.26 In that book, Ta⁠ʾwīl muškil al-Qurʾān, the meanings of words are often cited, but only as part of a general theme in which the author argues against claims that the Qurʾān includes a lot of inconsistencies, contradictions, linguistic errors etc. In many ways, however, the linguistic evidence in both books is remarkably similar, often with the very same words in both to explain meanings of certain lexemes. As for Rāġib’s al-Mufradāt, it is by far the most comprehensive of the sources on Qurʾānic ġarīb known to us. Having arranged the book alphabeti- cally, Rāġib was able to discuss the meaning of words by examining under one heading the different contexts in which they are used in various places in the Qurʾān. The lemmata, which normally begin with the general meaning of the root (often referred to as aṣl), are thus full-fledged essays which discuss the semantic range of a certain word and its derivatives. In the lemma ḤML, for example, Rāġib gives the general meaning of the root and then proceeds to the various contexts in which it is used, such as carrying a burden on one’s back, carrying a child inside a woman’s belly, bearing of fruit by trees, shoul- dering responsibility, carrying of water by clouds, etc.27 Rāġib deliberately uses this survey of the semantic range of the derivatives of a certain root in order to demonstrate how words are used metaphorically. In fact, he explicitly says in his introduction that he intends to show the semantic relationships (munāsabāt) between the derivatives of a root and the metaphorical uses of

25 Ibid., 6–20, 21–37. 26 Ibid., 3. 27 Rāġib, Mufradāt 131–132; for the term aṣl, see for example 41, 99, 124, 127, 157, 159, 163, 186,

211, etc.

 

 

70 chapter 2

words.28 As a result of Rāġib’s discussion of different derivatives under one lemma and of his interest in structure as a major factor in understanding metaphor, the Mufradāt, unlike most other books in the genre, abounds with comments related to both morphology and syntax.

Another aspect of ġarīb in the Qurʾān is tackled by Zamaḫšarī (d. 538/1144) in Nukat al-Aʿrāb fī ġarīb al-iʿrāb fī l-Qurʾān al-karīm and Abū l-Barakāt b. al-Anbārī (d. 577/1181) in al-Bayān fī ġarīb iʿrāb al-Qurʾān. Both books are primarily a syntactical study of the various possibilities of iʿrāb in Qurʾānic verses, that is, of the syntactical function of the components of structure and the impact which this function can have on meaning. Chapters are arranged according to the order of Qurʾānic sūras, and their titles in al-Bayān begin with the expression ġarāʾib iʿrāb sūrat al-Fātiḥa/al-Baqara/etc.29 It is to be noted, however, that the issues raised by Ibn al-Anbārī and Zamaḫšarī do not differ from those discussed in traditional grammatical works or books on the iʿrāb of the Qurʾān. Their insistence on representing their material as ġarīb thus raises the possibility that the scope of Qurʾānic ġarīb is significantly broader than one gathers from the sources discussed above.

One issue that is relevant to all sources on Qurʾānic ġarīb relates to the cri- teria used to identify words as ġarīb. Most of the above mentioned authors do not specify these criteria, but Abū Ḥayyān in his very short introduction to Tuḥfat al-arīb briefly mentions that there are two types of Qurʾānic words, the first of which is known both to the laymen (ʿāmma) as well as to the elite (ḫāṣṣa), such as samāʾ (sky), arḍ (earth), fawq (up), and taḥt (down), whereas the second is known only to those who are cognizant of and well-versed in Arabic (man lahu ṭṭilāʿ wa-tabaḥḥur fī l-luġa al-ʿArabiyya).30 The second type, he says, is the one on which authors have written books and is called ġarīb al-Qurʾān. We have earlier touched on this subject in relation with the words about whose meanings Nāfiʿ b. al-Azraq questioned Ibn ʿAbbās.31 Although these words which warrant explanation from Ibn ʿAbbās are supposedly ġarīb, some of them hardly qualify for inclusion with ġarīb material. Indeed, all other sources on Qur’ānic ġarīb that we have discussed include words of this type. Examples include ra⁠ʾūf (merciful), ḥasra (sorrow), anša⁠ʾnā (we started),

28 Ibid., 4. 29 Ibn al-Anbārī, Bayān I, 31, 43. The word ġarīb also appears in the titles of all chapters in

Zamaḫšarī’s Nukat al-Aʿrāb, but they seem to be added by the editor, who does not explain the reason for this addition.

30 Abū Ḥayyān, Tuḥfa 40. 31 See above, 39.

 

 

71MUBAWWAB AND SPECIALIZED LEXICA

baġtatan (suddenly), matīn (tough),32 ḫizy (disgrace), ukul (fruit), ṣayd al-baḥr (catch of fish), nafaq (tunnel), naba⁠ʾ (news),33 balāʾ (affliction), fiʾa (faction), sabīl (way), ṯuʿbān (serpent), and qiṭaʿ (pieces).34 As suggested earlier, some words such as these may have been included not because they were uncom- monly used or considered difficult for most Arabs, but in order to specify their exact meaning within a Qurʾānic context. Among the more obvious examples is ṣabr (patience) in the verse wa-staʿīnū bi-l-ṣabri wa-l-ṣalāti (Seek [God’s] help with patience and prayer; Q 2: 45). Ibn Qutayba seems to have included ṣabr among the words he explains because he believes that in this context it means fasting (ṣawm).35 Similarly, a word as common as uǧūrahunna (lit. their pay) is listed by Siǧistānī and explained as muhūrahunna because in a Qurʾānic con- text it refers to the dowers given to women (e.g. Q 4: 24).36 Yet this explanation cannot be applied in all cases, and there are in all works of the genre words that most probably neither belong to the realm of ġarīb nor have a specific con- notation in Qurʾānic usage (cf. ṯuʿbān and qiṭaʿ, cited above). It is interesting to note that Ibn Qutayba promises his reader not to cite linguistic evidence for commonly used words (wa-an lā nastašhid ʿalā l-lafẓ al-mubtaḏal),37 but does not explain why such words should be included in the first place! As the genre developed with Rāġib’s al-Mufradāt, which examines various derivatives of the same root, the notion of ġarīb was certainly broadened to include a large num- ber of more commonly used words. In the lemma ḤǦR, for example, a dozen or so derivatives are mentioned, some of which are commonly used words such as ḥaǧar/ḥiǧāra (stone/stones) and taḥaǧǧara (to petrify), while others may have required explanation, such as ḥiǧr (mare) and ḥaǧǧūra (a game for boys).38

2.2 Ġarīb al-Ḥadīṯ The second genre of ġarīb books deals with the meanings of words in the pro- phetic tradition (Ḥadīṯ). As one of the four major sources of data on attested usage, the use of Ḥadīṯ as linguistic evidence posed a methodological prob- lem related to its transmission. As previously mentioned, Ḥadīṯ was not always transmitted verbatim, and some of its transmitters were not even native

32 Yazīdī, Ġarīb 83, 86, 134, 136, 154. 33 Ibn Qutayba, Tafsīr 61, 97, 147, 153, 154. 34 Makkī, ʿUmda 75, 96, 120, 136, 152. 35 Ibn Qutayba, Tafsīr 47. 36 Siǧistānī, Ġarīb 34. Cf. below, 202 for the meaning of Qurʾānic tarǧūna in Mubarrad’s Mā

ttafaqa lafẓuhu wa-ḫtalafa maʿnāhu min al-Qurʾān al-maǧīd. 37 Ibn Qutayba, Tafsīr 3. 38 Rāġib, Mufradāt 107.

 

 

72 chapter 2

speakers of Arabic. The grammarians and lexicographers obviously did not see eye to eye on whether this state of affairs affected the value of Ḥadīṯ as linguis- tic evidence, for whereas early grammarians such as Sībawayhi (d. 180/796) only sparingly quoted Ḥadīṯ, lexicographers as early as Ḫalīl (d. 175/791) cited it much more frequently. Thus, in contrast to Sībawayhi’s seven or eight cita- tions, the number of ḥadīṯs and aṯars by Ḫalīl in the ʿAyn is 428.39 For lack of a better explanation, it may be suggested that this discrepancy is probably due to the different perspectives of both groups in dealing with Ḥadīṯ. In their study of syntax, early grammarians seem to have generally avoided Ḥadīṯ on the assumption that its structure would likely be altered if it was not transmitted verbatim. For their part, the lexicographers largely disregarded the transmis- sion issue, perhaps because they thought that it would have little impact on lexemes, since individual words may have been more easily preserved than the whole structure of the cited ḥadīṯ. In all cases, the existence of a whole genre on ġarīb al-Ḥadīṯ proves that the lexicographers generally did not share the early grammarians’ concerns regarding the way Ḥadīṯ was transmitted. Eventually, the grammarians themselves seem to have overcome this issue since most later grammarians had no qualms about using Ḥadīṯ as linguistic testimony.40

Unlike Qurʾānic ġarīb, there are no riwāyas attributed to the Prophet’s Companions which later transmitters assembled into books of the genre ġarīb al-Ḥadīṯ. According to Ibn al-Aṯīr (d. 606/1210), himself the author of a volu- minous book on ġarīb al-Ḥadīṯ, credit should go to Abū ʿUbayda Maʿmar b. al-Muṯannā (d. 209/824) for being the first to assemble some material related to this genre (awwal man ǧamaʿa fī hāḏā l-fann šayʾan), albeit in a short mono- graph of a few pages (awrāq maʿdūdāt).41 Reference to Abū ʿUbayda’s book is also made by al-Ḫaṭīb al-Baġdādī (d. 463/1072), who also mentions another early short monograph on the same subject authored by the Basran grammar- ian Abū ʿAdnān ʿAbdalraḥmān b. ʿAbdalaʿlā al-Sulamī.42 Naṣṣār suggests that this monograph may well have predated Abū ʿUbayda’s since Abū ʿAdnān was a rāwiya (relater, reciter) of Abū l-Baydāʾ al-Riyāḥī, a contemporary of Yūnus b. Ḥabīb’s (d. 182/798).43 It is in fact impossible to verify which of the two was authored first, but it is noteworthy that both belong to the same period in which most of the earliest books on Qurʾānic ġarīb were reportedly written, excluding, of course, those that are ascribed to Ibn ʿAbbās. If the ascription of

39 See above, 30. 40 Baġdādī, Ḫizāna I, 9–15; cf. Ḥadīṯī (1981: 13–29); Faǧǧāl (1997: 99–136). 41 Ibn al-Aṯīr, Nihāya I, 5; cf. below, 80. 42 Baġdādī, Tārīḫ XII, 405; cf. Ibn al-Nadīm, Fihrist 51. 43 Naṣṣār (1968: I, 50).

 

 

73MUBAWWAB AND SPECIALIZED LEXICA

al-Ġarīb fī l-Qurʾān to Abān b. Taġlib al-Bakrī al-Ǧurayrī (d. 141/758) is correct,44 then the genre of ġarīb al-Qurʾān predates its counterpart in Ḥadīṯ by more than half a century. It is not surprising that this could be the case, given the fact that the centrality of the Qurʾān to early linguistic sciences unmistakably exceeds that of Ḥadīṯ.

Many books on ġarīb al-Ḥadīṯ reportedly authored in the third/ninth and fourth/tenth centuries (this based on the authors’ death dates) are lost to us. Among the authors to whom the bibliographical and biographical sources attribute such books are al-Naḍr b. Šumayl (d. 203/819) – who died before Abū ʿUbayda, but whose book was written after his, according to Ibn al-Aṯīr45 – Abū ʿAmr al-Šaybānī (d. 206/821), Quṭrub (d. 206/821), Abū Zayd al-Anṣārī (d. 215/830), al-Aḫfaš al-Awsaṭ (d. 215/830), Aṣmaʿī (d. 216/831), Ibn al-Aʿrābī (d. 231/845), Muḥammad b. Ḥabīb (d. 245/860), Šamir b. Ḥamdawayhi (d. 255/869), Salama b. ʿĀṣim (d. after 270/883), Mubarrad (d. 285/898), Ṯaʿlab (d. 291/904), Abū Muḥammad al-Qāsim b. Muḥammad al-Anbārī (d. 304/917), Abū Mūsā l-Ḥāmiḍ (d. 305/918), Ibn Durayd (d. 321/933), Abū Bakr Muḥammad b. al-Qāsim al-Anbārī (d. 328/940), and Ibn Durustawayhi (d. 347/958). Fortunately, three books from the third/ninth century and two from the fourth/tenth have survived, and these are included in the following list of extant sources which will be discussed according to their chronological order (except for Ḫaṭṭābī’s and Saraqusṭī’s):

1. Ġarīb al-Ḥadīṯ by Abū ʿUbayd al-Qāsim b. Sallām (d. 224/838). 2. Ġarīb al-Ḥadīṯ by Ibn Qutayba (d. 276/889); also al-Masāʾil wa-l-aǧwiba fī

l-Ḥadīṯ wa-l-tafsīr and Ta⁠ʾwīl muḫtalif al-Ḥadīṯ by the same author. 3. Ġarīb al-Ḥadīṯ by Abū Isḥāq al-Ḥarbī (d. 285/898). 4. al-Dalāʾil fī ġarīb al-Ḥadīṯ by Saraqusṭī (d. 302/915). 5. Ġarīb al-Ḥadīṯ by Abū Sulaymān al-Ḫaṭṭābī (d. 388/998). 6. K. al-Ġarībayn fī l-Qurʾān wa-l-Ḥadīṯ by Abū ʿ Ubayd al-Harawī (d. 401/1011);

also mentioned earlier among sources on ġarīb al-Qurʾān. 7. al-Fāʾiq fī ġarīb al-Ḥadīṯ by Zamaḫšarī (d. 538/1144). 8. Mašāriq al-anwār ʿalā ṣiḥāḥ al-āṯār by al-Qāḍī ʿIyāḍ (d. 544/1149). 9. Ġarīb al-Ḥadīṯ by Ibn al-Ǧawzī (d. 597/1201). 10. al-Nihāya fī ġarīb al-Ḥadīṯ wa-l-Aṯar by Ibn al-Aṯīr (d. 606/1210). 11. al-Muǧarrad li-luġat al-Ḥadīṯ by Muwaffaq al-Dīn ʿAbdallaṭīf al-Baġdādī

(d. 629/1231).

44 See above, 65. 45 Ibn al-Aṯīr, Nihāya I, 5–6.

 

 

74 chapter 2

The first of these extant sources, Abū ʿUbayd’s Ġarīb al-Ḥadīṯ is held in high esteem by later authors, and Ḫaṭṭābī even considers it to be the first book of the genre (obviously disregarding the earlier short works by Abū ʿUbayda and Abū ʿAdnān) and the model (imām) on which later authors depended.46 For his part, Ibn al-Aṯīr mentions earlier works but argues that even if Abū ʿ Ubayd’s came at a later date (wa-in kāna aḫīran), it eventually became foremost (ṣāra . . . awwa- lan) due to its rich and useful material.47 Ibn al-Aṯīr also claims that Abū ʿUbayd reportedly claimed to have spent forty years collecting material for his book, but it seems that Ibn al-Aṯīr mistakenly applied this statement to Ġarīb al-Ḥadīṯ whereas it actually refers to Abū ʿUbayd’s other famous book with an analogous title, al-Ġarīb al-muṣannaf.48 More importantly, both Ḫaṭṭābī and Ibn al-Aṯīr agree that the two books of Abū ʿ Ubayd’s and Ibn Qutayba’s form the basis on which all later authors of this genre depended. In particular, Ḫaṭṭābī asserts that these two books make all books prior to his own time dispensable since they together contain the bulk of explanations and interpretations (tafsīr wa-ta⁠ʾwīl) related to Ḥadīṯ, although they might have missed a tiny fraction of material which is insignificant anyway (al-nabḏ al-yasīr allaḏī lā yuʿtadd bihi wa-lā yuʾbah lahu).49 Concerning arrangement, Abū ʿUbayd adopts the sys- tem of isnād, beginning with prophetic Ḥadīṯ, followed by that of the Ṣaḥāba (Companions) and Tābiʿūn (Successors) and a number of ḥadīṯs which cannot be attributed to any particular individual. There is no further internal arrange- ment within this general division. In this respect, Abū ʿ Ubayd set a trend which some later authors emulated, including Ibn Qutayba and Ḫaṭṭābī.

Ibn Qutayba himself expresses his admiration of Abū ʿUbayd’s book, and in his own book Ġarīb al-Ḥadīṯ, follows Abū ʿ Ubayd’s arrangement of the material, but adds five introductory chapters. He ends his introduction with the hope that his book and Abū ʿUbayd’s will leave nothing (i.e. for future scholars) to be said about ġarīb al-Ḥadīṯ.50 Yet, Ibn Qutayba claims that, after further scrutiny, he gave up his initial view that Abū ʿUbayd’s book explains all ġarīb words in the Ḥadīṯ, since he discovered that the material which the latter had omitted

46 Ḫaṭṭābī, Ġarīb I, 48. 47 Ibn al-Aṯīr, Nihāya I, 6. 48 Abū l-Ṭayyib (Marātib 148–149) was the first to mention such a statement in connection

with al-Ġarīb al-muṣannaf although the period he specified was thirty years. Qifṭī (Inbāh III, 22) also quotes a similar statement attributed to Abū ʿUbayd with regard to his al-Ġarīb al-muṣannaf on which he spent forty years. Cf. also Yāqūt, Muʿǧam V, 2199 and Ibn Ḫillikān, Wafayāt IV, 61.

49 Ḫaṭṭābī, Ġarīb I, 50–51. 50 Ibn Qutayba, Ġarīb I, 152.

 

 

75MUBAWWAB AND SPECIALIZED LEXICA

was equivalent to, or more than the material he mentioned.51 Furthermore, he wrote a separate book, Iṣlāḥ ġalaṭ Abī ʿUbayd fī Ġarīb al-Ḥadīṯ, in which he, in spite of an apologetic introduction,52 objected to Abū ʿUbayd’s explana- tion of fifty-three ḥadīṯs. In his al-Ġarīb, Ibn Qutayba does not normally repeat what Abū ʿ Ubayd included in his book, but adds what he missed and adopts his method of isnād. There is, however, a basic difference in approach by the two authors, for whereas Abū ʿUbayd briefly explains the words he considers to be ġarīb, Ibn Qutayba – as he himself declares in his introduction – elaborates on matters related to derivation (ištiqāq), extensively quotes poetry šawāhid, and cites relevant anecdotes and proverbs in order not to restrict his book to ġarīb. Reminiscent of his Tafsīr ġarīb al-Qurʾān (which he begins with two chapters that differ from the main body of the book)53 are the above mentioned five introductory chapters which are distinct from the rest of the book and thus are not based on Abū ʿUbayd’s book. These chapters deal with the derivation and meanings of widely used terms and words related to jurisprudence (fiqh) and judicial decisions (aḥkām), such as tayammum (washing with sand etc. instead of water), ṣiyām (fasting), and istilām al-ḥaǧar (touching the Black Stone).54 They also include words which occur in the Prophet’s ḥadīṯs concerning the Qurʾān, relevant Qurʾānic words, and the names of certain sects.

In another book, al-Masāʾil wa-l-aǧwiba fī l-Ḥadīṯ wa-l-tafsīr, Ibn Qutayba discusses 190 masʾalas related to Ḥadīṯ, most of which focus on ġarīb words.55 He obviously wrote this book after his al-Ġarīb since masʾala 21 contains sev- eral ḥadīṯs which his supposed interlocutor tells him were not included in his earlier book.56 As for his Ta⁠ʾwīl muḫtalif al-Ḥadīṯ, it is the counterpart of his previously mentioned book Ta⁠ʾwīl muškil al-Qurʾān57 in that it is a refutation of certain groups, such as the Ḫawāriǧ, the Murǧiʾa, the Rāfiḍa and the Muʿtazila, who interpret Ḥadīṯ in a manner that suits their ends, in addition to refuting, often on linguistic grounds, allegations of inconsistencies and contradictions in Ḥadīṯ.

Since the books of Saraqusṭī and Ḫaṭṭābī are similar in their arrangement to Abū ʿUbayd’s and Ibn Qutayba’s, it would make better sense – at the risk of

51 Ibid., I, 150. 52 Ibn Qutayba, Iṣlāḥ 41–47. 53 See above, 69. 54 Ibn Qutayba, Ġarīb I, 153, 217, 221. 55 See the editors’ classification of these masʾalas according to content; Ibn Qutayba, Masāʾil

18–20. 56 Ibid., 69–85. 57 See above, 69.

 

 

76 chapter 2

upsetting the chronological order – to bring them in at this point, particularly because all of the other five sources follow a completely different arrangement. Saraqusṭī’s al-Dalāʾil fī ġarīb al-Ḥadīṯ,58 described by some as the best book of this genre in Andalus,59 is based on the books by Abū ʿUbayd and Ibn Qutayba. Unfortunately, its introduction is lost, but it is clear from the book’s content that the author complements the two earlier books and often amends some of the errors they contain. He also explains the meanings of ġarīb words in the quoted ḥadīṯs. His task must have been facilitated by the fact that he preserved the same arrangement followed by his predecessors.

Ḫaṭṭābī explicitly says that he adopted the arrangement of both Abū ʿUbayd and Ibn Qutayba, and he normally does not repeat the ḥadīṯs which they explain unless he intends to add some clarification or draw attention to a dis- crepancy between his two predecessors in explaining a certain word. His mate- rial thus consists mainly of ġarīb words which do not occur in the two earlier books (alfāẓ ġarība lā aṣl lahā fī l-kitābayn)60 and which he explains by exten- sively quoting šawāhid, particularly from poetry.61 Furthermore, he assembles towards the end of his book words which were corruptly transmitted in Ḥadīṯ and gives their correct forms.62 In his introduction, he proudly mentions this contribution which he says is badly needed by scholars.63 But in spite of his additions to Abū ʿUbayd and Ibn Qutayba, their two books and his are funda- mentally similar, and, as Ibn al-Aṯīr notes, what they have in common is the wide popularity they enjoyed among scholars of the various regions (ʿulamāʾ al-amṣār) and the lack of any arrangement of words to help the user find the ḥadīṯ which he seeks.64

Among the extant sources, Ḥarbī’s Ġarīb al-Ḥadīṯ is the first that is arranged, not only according to isnād but also formal (lafẓī) criteria. It is highly prob- able that no other author adopted form as a basis of arrangement in this genre before him, as may be concluded from Ibn al-Aṯīr’s critical survey of third/ninth century sources.65 Unfortunately, only the fifth and last volume (muǧallada) of the book has survived, and so we have no way of knowing whether or not Ḥarbī

58 For other titles which the book is given in some sources, see the editor’s introduction to Saraqusṭī, Dalāʾil I, 95.

59 Cf. Zubaydī, Ṭabaqāt 285; see also Ḥumaydī, Ǧaḏwa 331–332. 60 Ḫaṭṭābī, Ġarīb I, 48. 61 See the indices of Ḫaṭṭābī’s Ġarīb III, 535–614. 62 Ibid., III, 219–257. 63 Ibid., I, 49. 64 Ibn al-Aṯīr, Nihāya I, 8. 65 Ibid., loc. cit.

 

 

77MUBAWWAB AND SPECIALIZED LEXICA

wrote an introduction in which he surveyed earlier works or set his plan for the book. From the sheer size of the fifth muǧallada published in three volumes, however, one gathers an idea about the voluminosity of the book. It also seems to be replete with poetry šawāhid, since their number in the published part alone is 1060, some of which are made up of several lines.66 But the most strik- ing feature of the book is that it adopts the permutative system, that is the inclusion in each lemma not only of the root of a word which appears in a specific ḥadīṯ but also the other possible permutations of its radicals which occur in other ḥadīṯs. Accordingly, with the lemma QMR are also included the roots QRM, MRQ, RMQ, RQM and MQR.67 Being one of the earliest authors in the history of Arabic lexicography to adopt this system, Ḥarbī most probably borrowed it directly from Ḫalīl (d. 175/791), who was the first to use it in his al-ʿAyn, although Ḥarbī arranged the roots in no specific order. It is also worth noting that Ḥarbī is one 0f only a handful of authors of lexica of the mubaw- wab type to use this system. As far as consistency in the application of the sys- tem is concerned, there are certain discrepancies in Ḥarbī’s book, particularly with regard to weak, hamzated, geminate and reduplicated roots. For example, under the lemma RM one find words of the roots RMY, RMM, RMRM, MRY, RYM, ʾRM and RʾM.68 Also, the following roots are consecutively arranged: WʾR, WRY, WRʾ, RWʾ, RʾY, RYʾ, RYY, ʾYR, RWY and ʾRY.69 In more ways than one, Ḥarbī’s book reflects his interest in linguistic matters even at the risk of going beyond the limits of his immediate subject, the Ḥadīṯ. For example, many of the roots he mentions do not contain a single citation of Ḥadīṯ and theoreti- cally should not have been included. In such lemmata, he instead explains the meanings of words which do not occur in Ḥadīṯ.70 Furthermore, he frequently digresses from his discussion of ġarīb words in Ḥadīṯ to explain words (includ- ing synonyms) which belong to a particular semantic field and are often the subject of single-topic works or part of multithematic ones – e.g. words related to gold and silver, palm trees, horses, love, the human eye, old age, and wine.71

Ḥarbī’s adoption of permutations in the arrangement of roots certainly does not facilitate the search for words in his book although it represents a tangible improvement on the earlier method of arranging ḥadīṯs according to

66 See the editor’s introduction to Ḥarbī’s Ġarīb I, 118. 67 Ibid., II, 372, 376, 380, 383, 385, 387. 68 Ibid., I, 66–78. 69 Ibid., II, 751, 754, 759, 762, 764, 771–772, 774, 779, 785. 70 See, for example, the lemmata KŠM (II, 541), ḤLṬ (II, 635), ʿRD (II, 697), MŠR (III, 953), and

RʿQ (III, 1032). 71 Ibid., I, 53; II, 432–433, 446–448, 649–650, 653–654, 703–704; III, 1005 respectively.

 

 

78 chapter 2

isnād only. The difficulty of using Ḥarbī’s book is duly noted by Ibn al-Aṯīr, who says that if one had to search in it for a particular ḥadīṯ, one would only find it after a lot of fatigue and trouble (taʿab wa-ʿanāʾ).72 It is therefore not surprising that none of the subsequent authors used Ḥarbī’s method of arrangement. The first author in the genre to use the conventional alphabetical order without the permutations is Harawī, whose book embraces the ġarīb of both Qurʾān and Ḥadīṯ as previously mentioned. It is quite remarkable that his arrange- ment was applied systematically by later authors of the genre, whereas their counterparts who authored muǧannas lexica were generally more inclined towards the rhyme system.

The remaining five authors in our list (i.e. Zamaḫšarī, ʿIyāḍ, Ibn al-Ǧawzī, Ibn al-Aṯīr and Baġdādī) unanimously arranged their books alphabetically and did not adopt Ḥarbī’s system based on permutations. Zamaḫšarī’s al-Fāʾiq fī ġarīb al-Ḥadīṯ73 is among the most comprehensive works of the genre. It is alphabetically arranged, but on the basis of the first and second radicals only. For example, in the chapter of qāf and tāʾ, the following roots occur in this order: QTR, QTN, QTT, QTR, QTT (both repeated), QTL, QTM, QTB and QTL.74 Coming from the same author of another dictionary, Asās al-balāġa, which follows full alphabetical order, this two-thirds alphabetical system is quite strange. Ibn al-Aṯīr in his survey of earlier works does not mention this incom- plete arrangement, but points out a shortcoming of another type, namely, that words in the Fāʾiq often occur, not in the lemmata of their own roots, but in those of the roots of other words which occur in the same ḥadīṯ.75 Zamaḫšarī himself was aware of this complexity and thus at the end of each chapter (e.g. ʾB, ʾT, ʾṮ, etc.) refers the reader to the other lemmata which contain words of the same root.

ʿIyāḍ’s Mašāriq al-anwār, in spite of its extensive material which occupies about 800 pages, has a narrow scope since it is restricted to what the author believes are errors that occur in three of the most authoritative collections of Ḥadīṯ, namely, Muwaṭṭa⁠ʾ Mālik, Ṣaḥīḥ al-Buḫārī and Ṣaḥīḥ Muslim. As ʿIyāḍ explains in the introduction, the roots of words are listed in full alphabetical order for ease of use.76 The lemmata are frequently followed by comments on differences among scholars in reading or explaining certain words. True, these comments, which are normally introduced by the terms iḫtilāf (disagreement)

72 Ibn al-Aṯīr, Nihāya I, 8. 73 For more detailed studies of al-Fāʾiq, see ʿUbaydī (2001b) and Sulaymān (2011). 74 Zamaḫšarī, Fāʾiq III, 156–158. 75 Ibn al-Aṯīr, Nihāya I, 9. 76 ʿIyāḍ, Mašāriq I, 6. For more details about the book, see Wadġīrī (2008: 86–91).

 

 

79MUBAWWAB AND SPECIALIZED LEXICA

and wahm (misinterpretation), disrupt the flow of the lemmata,77 but they serve as a useful reminder of the controversy that surrounds a large portion of Ḥadīṯ material.

In his introduction to Ġarīb al-Ḥadīṯ, Ibn al-Ǧawzī mentions a number of his predecessors who wrote books on the subject. He criticizes Abū ʿUbayd for having left out a large body of material – a charge he also raises, albeit to a lesser degree, against both Ḥarbī and Ibn Qutayba, who claim to have made up for what Abū ʿUbayd had missed. He also accuses Ḫaṭṭābī of having only partially made up for what Ḥarbī and Ibn Qutayba had left out.78 His harshest criticism, however, was addressed against Harawī, not only for claiming that he left out no ġarīb material at all, while in fact he only settled for what Azharī (d. 370/981) had mentioned in Tahḏīb al-luġa, but also for including words which are not ġarīb and thus need no explanation.79 Although Ibn al-Ǧawzī promises to make up for the omissions of his predecessors, he does not seem to have convinced Ibn al-Aṯīr that he has achieved his mission successfully. Ibn al-Aṯīr particularly refutes Ibn al-Ǧawzī’s position vis-à-vis Harawī, and asserts that he slavishly followed Harawī’s method in K. al-Ġarībayn, and abridged that book’s content, only sporadically adding words not mentioned by Harawī.80 Indeed, comparison between the two works (which both follow full alphabeti- cal order) proves Ibn al-Aṯīr generally right concerning the number of cited traditions (as, for example, in words beginning with the letter ṯ, of which each book contains roughly one hundred ḥadīṯs and aṯars). As for abridgement, Ibn al-Ǧawzī himself tells us that he preferred brevity in explaining words and chose to relieve the user from reading details related to conjugation and derivation (al-taṣrīf wa-l-ištiqāq).81 His observation that such morphological matters are more appropriately discussed in books on language (kutub al-luġa) betrays his belief that the sole function of a lexicon is to give the meanings of words, not to delve into other aspects of linguistic enquiry.

The last major lexicon of the genre ġarīb al-Ḥadīṯ is Ibn al-Aṯīr’s al-Nihāya fī ġarīb al-Ḥadīṯ wa-l-Aṯar, whose title is deliberately chosen by the author to indicate that the material he assembled in it is more comprehensive than that of all earlier lexica. This fact is highlighted by Ibn al-Aṯīr is his introduction, which presents a useful survey of the genre since the third/ninth century and which demonstrates how ġarīb material in Ḥadīṯ significantly expanded as

77 E.g. ibid., I, 12, 17, 19–20, 22, etc. 78 Ibn al-Ǧawzī, Ġarīb I, 2–3. 79 Ibid., I, 4. 80 Ibn al-Aṯīr, Nihāya I, 9–10. 81 Ibn al-Ǧawzī, Ġarīb I, 4.

 

 

80 chapter 2

authors continued to discover that their predecessors, in spite of their alle- gation of exhaustiveness, did leave out enough ġarīb words to justify further attempts at completeness. Ibn al-Aṯīr is no exception, and he explicitly states that he initially wanted to combine in his book the material related to Ḥadīṯ in Harawī’s K. al-Ġarībayn and Abū Musā l-Iṣfahānī’s (d. 581/1185) al-Muġīṯ fī ġarīb al-Qurʾān wa-l-Ḥadīṯ, on the assumption that these two voluminous books complement each other in the sense that if one does not find a certain word in one of them he should find it in the other.82 Following further scrutiny, however, Ibn al-Aṯīr discovered that even these two sources combined do not contain certain ġarīb words which occur in ḥadīṯs cited in the most famous collections, such as Ṣaḥīḥ al-Buḫārī and Ṣaḥīḥ Muslim. Ibn al-Aṯīr thus decided to cast his net wide and add to his sources books on ġarīb al-Ḥadīṯ, which he did not specify but described as belonging to the early, middle and later or contemporary periods (kutub al-Ḥadīṯ al-mudawwana al-muṣannafa fī awwal al-zamān wa-awsaṭihi wa-āḫirihi). One may safely conclude that he must have consulted the bulk of earlier works on Ḥadīṯ, particularly because he con- firmed that he even searched in other sources, including various books on lan- guage (kutub al-luġa ʿalā ḫtilāfihā), for ġarīb words in Ḥadīṯ. Consequently, the Nihāya is the most comprehensive book of its genre. Compared with the previ- ously mentioned figure of a hundred ḥadīṯs beginning with the letter ṯ in each of Harawī’s and Ibn al-Ǧawzī’s books, for example, the Nihāya contains about one hundred and sixty. Another example is that Abū ʿUbayd in his al-Ġarīb cites about twenty-five ḥadīṯs with words beginning with the letter t, compared to more than a hundred and ninety in the Nihāya. Another contributing factor to the increase in the number of ḥadīṯs cited in the Nihāya as compared with earlier sources has to do with the fact that knowledge of Arabic in the time of Ibn al-Aṯīr (d. 606/1210) had significantly and gradually deteriorated from what it was about four centuries earlier when the first books of the genre were authored. Ibn al-Aṯīr alludes to this in the beginning of his survey of earlier works of the genre. He ascribes the brevity of Abū ʿUbayda’s (d. 209/824) work, which is said to have consisted of only a few pages (awrāq maʿdūdāt), not only to its being the first book of its kind, but also to the fact that people at the time at least had some knowledge of Arabic (kāna fīhim baqiyya wa-ʿindahum maʿrifa),83 in contrast to his own days in which even students (ṭalaba) of ġarīb al-Ḥadīṯ could hardly differentiate between the original radicals and prefixes (lā yakādūna yafriqūna bayna l-aṣlī wa-l-zāʾid).84 Based on that, he says, he had

82 Ibn al-Aṯīr, Nihāya I, 10; cf. Ḫalīl (1997: 430–431). 83 Ibid., I, 5; cf. above, 72. 84 Ibid., I, 11.

 

 

81MUBAWWAB AND SPECIALIZED LEXICA

to list alphabetically some words which began with prefixes, and not under their roots – e.g. timrāḥa (energetic) and muʿtāṭ (a she-goat not conceiving due to abundance of fat) under t and m, and not under their respective roots, mrḥ and ʿyṭ.85 Other than these instances, roots are cited and arranged in full alphabetical order.

The last book in the above list, Baġdādī’s al-Muǧarrad li-luġat al-Ḥadīṯ, is a lexicon of words in Ḥadīṯ arranged by the full alphabetical order of their roots. Baġdādī mentions in a brief introduction an earlier book of his on ġarīb al-Ḥadīṯ (now lost to us), which he arranged alphabetically and in which he cited the Ḥadīṯ of the Prophet, followed by that of the Ṣaḥāba (Companions) and Tābiʿūn (Successors).86 In al-Muǧarrad, he briefly explains the meanings of the words without citing the ḥadīṯs themselves or any supporting šawāhid.

In spite of the strong link between works on ġarīb al-Qurʾān and those on ġarīb al-Ḥadīṯ, and given that both deal with a well-defined linguistic corpus of a religious dimension and that some authors, such as Harawī in K. al-Ġarībayn and Iṣfahānī in al-Muġīṯ (mentioned by Ibn al-Aṯīr), combined them under one title, there is a basic difference between the two genres as far as the nature of words included in each is concerned. It was mentioned earlier87 that works on Qurʾānic ġarīb uniformly contain commonly used words which one would not have expected to be classified as ġarīb. In contrast, this phenomenon is practically absent from sources on ġarīb al-Ḥadīṯ as these contain almost exclusively uncommon words which require explanation (at least as far as a modern educated native speaker of Arabic is able to judge). A cursory look at words mentioned in any of the sources discussed above would readily reveal this reality. We have previously suggested that the reason for the inclusion of non-ġarīb words in works on Qurʾānic ġarīb might be to explain their mean- ings in a Qurʾānic context since that may differ significantly from an ordinary context. Authors of books on ġarīb al-Ḥadīṯ might thus have found it unneces- sary to do the same with words in Ḥadīṯ since this would have resulted in the repetition of words already discussed in a related genre. But since this can only be true for words which are common to the two corpuses and does not explain why ġarīb al-Ḥadīṯ should be the genre that is affected by the content of ġarīb al-Qurʾān and not the other way round even if occasionally, it is more probable that authors of ġarīb al-Ḥadīṯ did not find significant differences in connota- tions of words used in Ḥadīṯ and in ordinary prose, and consequently did not include non-ġarīb words in their data. Another possible explanation has to do

85 Ibid., I, 196–197; IV, 341. 86 Baġdādī, Muǧarrad I, 94. 87 See above, 40, 71.

 

 

82 chapter 2

with the size of the corpus of each genre. Ḥadīṯ obviously being the larger of the two (witness the size of some ṣaḥīḥ works and some of the works which deal with its ġarīb), it was exempt from the inclusion of the more commonly used words in the study of its ġarīb, as this would have resulted in expanding an already sizable corpus. One notable exception, however, is the explanation of common words when used metaphorically, as in raqaba (neck) which in ka⁠ʾannamā aʿtaqa raqabatan (as if he has freed a serf) refers to the individual as a whole.88 It is perhaps examples of this type that Ḫaṭṭābī had in a mind when he accused some of his predecessors of including in their books ḥadīṯs which are almost free from semantic difficulty (lā yakād yuškil minhā šayʾ).89 Ḫaṭṭābī’s accusation cannot be substantiated from the sources available to us and is thus either a gross overstatement or directed against authors whose works did not reach us.

The vast corpus of words which the books on ġarīb al-Ḥadīṯ contain naturally led authors of the muǧannas lexica to draw heavily on them. Several of these authors mention books of this genre among their sources. For example, Azharī (d. 370/981) specifically mentions as sources for his al-Tahḏīb the books of Abū ʿUbayd, al-Naḍr b. Šumayl and Ibn Qutayba,90 and Ibn Sīda (d. 458/1066) men- tions šurūḥ al-Ḥadīṯ (commentaries on Ḥadīṯ), which are obviously books on ġarīb, among his sources, but does not specify further.91 More significantly, Ibn Fāris (d. 395/1004) declares that his Maqāyīs al-luġa is basically derived from five books (fa-hāḏihi l-kutub al-ḫamsa muʿtamadunā fī mā stanbaṭnāhu), one of which is Abū ʿUbayd’s Ġarīb al-Ḥadīṯ.92 For his part, Ibn Manẓūr (d. 711/1311) has fully incorporated into Lisān al-ʿArab, one of the most famous and compre- hensive muǧannas lexica, Ibn al-Aṯīr’s al-Nihāya, which he says is one of only five sources which he combined in authoring his lexicon.93

There are also lexica which deal with words that are used in jurisprudence (fiqh), and although these are not exactly works on ġarīb, some of them rely heavily on Ḥadīṯ and the lexica that deal with its ġarīb. Among the earliest of these is al-Zāhir fī ġarīb alfāẓ al-Imām al-Šāfiʿī, authored by Azharī (d. 370/981)

88 Ibn al-Aṯīr, Nihāya II, 249. 89 Ḫaṭṭābī, Ġarīb I, 50. 90 Azharī, Tahḏīb I, 14–15, 18, 20, 30–32. 91 Ibn Sīda, Muḥkam I, 47. Based on comparing the Muḥkam with sources on ġarīb al-Ḥadīṯ,

Ǧubūrī in his introduction to Ibn Qutayba’s Ġarīb al-Ḥadīṯ (I, 30) concludes that the books of Abū ʿUbayd, Ibn Qutayba and Harawī are among Ibn Sīda’s sources.

92 Ibn Fāris, Maqāyīs I, 4–5. 93 Ibn Manẓūr, Lisān I, 8.

 

 

83MUBAWWAB AND SPECIALIZED LEXICA

and also mentioned in his al-Tahḏīb.94 As its title indicates, the book con- tains ġarīb words used by the famous imām Muḥammad b. Idrīs al-Šāfiʿī (d. 204/820). Azharī gathers these words, which he arranges thematically, from Ismāʿīl b. Yaḥyā l-Muzanī’s (d. 264/878) K. al-Ǧāmiʿ, itself an abridge- ment of Šāfiʿī’s K. al-Umm. Also worthy of mentioning are two works from the seventh/thirteenth century, the first of which is Muṭarrazī’s (d. 610/1213) al-Muġrib fī tartīb al-Muʿrib, in which he alphabetically arranges words and expressions included in his own book al-Muʿrib. In al-Muġrib’s introduction, Muṭarrazī points out that he adopted in his lexicon Harawī’s arrangement of K. al-Ġarībayn,95 an alphabetical arrangement based on the roots of words. He does admit, however, that in words of more than three radicals he took into consideration the first, second and last radicals only. Accordingly, DḪRṢ occurs after DḪS, and QRṬQ after QRF, whereas the reverse would have been expected had a fully alphabetical order been followed.96 The second lexicon is Nawawī’s (d. 676/1277) Tahḏīb al-asmāʾ wa-l-luġāt, whose title indicates that he included in it proper nouns (e.g. names of the Prophet’s Companions), followed by ordi- nary words. These words are arranged alphabetically on the basis of their roots, although a few of them are listed with their affixes since their roots may not be readily recognizable to some readers.97 Following each chapter (i.e. alif, bāʾ, tā⁠ʾ, etc.) is an addendum exclusively for place names listed alphabetically and, of course, “roots” in this case are not taken into consideration. As for his sources, Nawawī mentions books on jurisprudence and theology, as well as muǧannas and mubawwab lexica, among which are five works on ġarīb al-Ḥadīṯ, namely, the books of Abū ʿUbayda, Abū ʿUbayd, Ibn Qutayba, Ḫaṭṭābī and Harawī.98 In the eighth/fourteenth century, Fayyūmī (d. c. 770/1368) wrote his famous lexi- con entitled al-Miṣbāḥ al-munīr fī ġarīb al-Šarḥ al-kabīr, an abridgement of an earlier work of his.99 In this abridgement he explains ġarīb words and expres- sions encountered in ʿAbdalkarīm b. Muḥammad al-Rāfiʿī’s (d. 623/1226) Fatḥ al-ʿAzīz ʿalā Kitāb al-Waǧīz, a commentary on Ġazālī’s (d. 505/1111) al-Waǧīz which deals with the fiqh of the Šāfiʿī order. The words are arranged in a full

94 See Azharī’s reference to his book Tafsīr ḥurūf al-Muḫtaṣar (another title for al-Zāhir) in Tahḏīb III, 239 and XV, 427–428; cf. the editor’s introduction to the Zāhir 57.

95 Muṭarrazī, Muġrib 16. 96 Ibid., 161, 379. 97 Nawawī, Tahḏīb I, 7. 98 Note also that at times he justifies the inclusion of a word under its first letter by pointing

out that this letter is an original radical, and not an augment; see, for example, TRǦM and MTRS (ibid., III, 41; IV, 133).

99 For a detailed study of Fayyūmī’s book, cf. Aḥmad (1969: 153–165).

 

 

84 chapter 2

alphabetical order based on roots, with special attention given to morphologi- cal and syntactical issues.

2.3 General Ġarīb/Nādir Material Less specialized than lexica which deal exclusively either with ġarīb in the Qurʾān or the Ḥadīṯ are those which more generally include ġarīb material in poetry and ordinary speech. As has been pointed out earlier,100 there are sev- eral terms other than ġarīb which also express the notion of strange usage. Among these are nādir (pl. nawādir, rare), ḥūšī or waḥšī (unfamiliar, uncouth), and šāḏḏ (pl. šawāḏḏ, anomalous). Suyūṭī (d. 911/1505) devotes chapter thir- teen of his al-Muzhir to the ḥūšī, ġarāʾib, šawāḏḏ and nawādir.101 To these four terms which appear in the title of the chapter he adds the terms šawārid (pl. of šārid, erratic) and muškil (problematic). Although he tries to explain some of these terms separately, he starts his chapter by saying that the four terms mentioned in the title of the chapter are close in meaning (mutaqāriba), and are all antonyms of faṣīḥ. From the context one concludes that faṣīḥ, which normally refers to eloquent or elevated usage, also means that a word does not sound detestable (nafara ʿan al-samʿ) or is only known to the eminent scholar (al-ʿālim al-mubarriz) or the pure Bedouin (al-Aʿrābī al-quḥḥ). Later in the chapter, Suyūṭī practically admits that ġarīb, ḥūšī, šārid and šāḏḏ are synonymous,102 although in the illustrations which he then cites from earlier sources he lists nawādir, šawārid and ġarāʾib separately. It is interesting to note that, among these terms, the only one which the philologists found appropri- ate for use with Qurʾān or Ḥadīṯ is ġarīb. Thus, they speak of ġarīb al-Qurʾān/ al-Ḥadīṯ, but not its nādir or šārid, and definitely not its ḥūšī or sāḏḏ. The latter term indeed appears in a Qurʾānic context, but only to describe the so-called anomalous qirāʾāt, as in the title of Ibn Ḫālawayhi’s (d. 370/980) book Muḫtaṣar fī šawāḏḏ al-Qurʾān, where šawāḏḏ refers to unorthodox readings ascribed in each case to (a) specific reader(s).

It is extremely difficult to confine the list of sources in the genre of ġarīb to a manageable number. To begin with, the unmistakable interest of scholars in ġarīb material meant that it permeated the various philological and grammati-

100 See above, 37. 101 Suyūṭī, Muzhir I, 233–239. The term ābida is also cited as a synonym of waḥšī and ġarīb;

see Kurāʿ, Muǧarrad I, 35; cf. Ibn Manẓūr, Lisān (ʾBD). Note also the terms ʿuqmī and ġāmiḍ, which Azharī uses in Tahḏīb (ĠRB; VIII, 115) to explain what is meant by ġarīb min al-kalām; cf. Ḫalīl, ʿAyn (ĠRB; IV, 411), where ġāmiḍ is also used.

102 Ibid., I, 234 (wa-l-ġarāʾib ǧamʿ ġarība wa-hiya bi-maʿnā l-ḥūšī wa-l-šawārid ǧamʿ šārida wa-hiya ayḍan bi-maʿnāhā . . . al-nādira wa-hiya bi-maʿnā l-šawārid).

 

 

85MUBAWWAB AND SPECIALIZED LEXICA

cal sources, as well as books on literature, poetry, aḫbār, etc. One illuminating example of the difficulty one encounters in listing works under this genre is al-Ġarīb al-muṣannaf by Abū ʿUbayd (d. 224/838). As a multithematic work, it will be discussed later with its counterparts in Section 11.103 Yet the book’s title clearly indicates its author’s focus on ġarīb. More specifically, some of the titles of its chapters contain the term nawādir, as in nawādir al-asmāʾ (rare nouns) and nawādir al-afʿāl (rare verbs).104 The same two titles appear also in the mul- tithematic work ascribed to Ibn Qutayba (d. 276/889), K. al-Ǧarāṯīm.105 Similar to Abū ʿUbayd’s book, Rabaʿī’s (d. 480/1087) Niẓām al-ġarīb could easily be clas- sified as a book of the ġarīb genre although we have opted to include it with other multithematic sources to be discussed later.106 In muǧannas lexica one also finds certain chapters that are devoted to rare usage, as in the latter parts of the Ǧamhara in which Ibn Durayd (d. 321/933) includes several topics of the nawādir type.107 Their mere inclusion demonstrates the loose boundaries, at least in al-Ǧamhara’s case, between the muǧannas and mubawwab types, both of which abound with ġarīb and nādir material.

Even more indicative than the Ǧamhara of the difficulty of the distinction between the ġarīb/nādir genre and some early muǧannas lexica is Abū ʿAmr al-Šaybānī’s (d. 206/821) K. al-Ǧīm. For reasons that primarily have to do with demonstrating the evolution of muǧannas lexica, and in particular those that use the alphabetical order beginning with the first radical of the root, it is natural that the Ǧīm be examined with similarly arranged lexica in the next chapter. Classification considerations notwithstanding, the Ǧīm essentially belongs to the ġarīb/nādir genre since it basically consists of unfamiliar words which occur in the various Arabic dialects. This description is corroborated by Qifṭī (d. 646/1248) who quotes Abū ʿAbdallāh Muḥammad b. al-Ḥusayn al-Yamanī (d. 400/1010) as having said in his Ṭabaqāt al-nuḥāt wa-l-luġawiyyīn (which is lost to us) that Abū ʿAmr’s aim in the Ǧīm was to collect unfamiliar words and not those that are commonly used (ǧamaʿa fīhi l-ḥūšī wa-lam yaqṣid al-mustaʿmal).108 Abū ʿAmr’s typically Kufan interest in ġarīb/nādir is further demonstrated in his K. al-Nawādir, some of whose content is preserved to us in the criticism of its author’s errors by Abū l-Qāsim ʿAlī b. Ḥamza (d. 375/985)

103 See below, 268. 104 Abū ʿUbayd, Ġarīb I, 348, 360. Suyūṭī (Muzhir I, 234) specifically mentions these two

chapters as examples of nawādir material scattered in the sources. 105 Ibn Qutayba, Ǧarāṯīm II, 301, 311. 106 See below, 275. 107 Ibn Durayd, Ǧamhara II, 1086–1109; III, 1274–1337. 108 Qifṭī, Inbāh I, 261.

 

 

86 chapter 2

in al-Tanbīhāt ʿalā mā fī Nawādir Abī ʿAmr al-Šaybānī min aġālīṭ, and in quota- tions by later authors, such as Suyūṭī (d. 911/1505).109

Other than the above mentioned works, certain books that deal with mor- phological patterns (abniya) or homonyms (muštarak), and which will be discussed later, also qualify as books on ġarīb. Examples include some books on muṯallaṯāt, Abū ʿUmar al-Zāhid’s (i.e. Ġulām Ṯaʿlab; d. 345/957) al-ʿAšarāt fī ġarīb al-luġa, and al-Qazzāz al-Qayrawānī’s (d. 412/1021) al-ʿAšarāt fī l-luġa.110 Furthermore, it should be noted that a huge body of ġarīb/nādir material in prose and poetry is found in most sources in which dictation and discussion sessions (amālī/maǧālis) are recorded. Although these works vary in their focus, from the literary to the philological to the grammatical, they share an interest in strange words, expressions, dialects, etc. – an interest that can be partially explained by the existence of controversies which often arise among scholars as to the credibility and permissibility of some ġarīb data. Among the most famous of such works are Ṯaʿlab’s (d. 291/904) Maǧālis, Ibn al-Muzarraʿ’s (d. 304/916) Amālī, Yazīdī’s (d. 310/922) Amālī, Zaǧǧāǧī’s (d. 337/949) Amālī and Maǧālis al-ʿulamāʾ, and the Amālī of Qālī (d. 356/967), Marzūqī (d. 421/1030), al-Šarīf al-Murtaḍā (d. 436/1044), Ibn al-Šaǧarī (d. 542/1148), Suhaylī (d. 581/1185), and Ibn al-Ḥāǧib (d. 646/1249).

Before discussing the extant sources of the genre, it should be pointed out that books entitled al-Nawādir are among the earliest mentioned in the sources. The first book bearing this title is attributed to one of the earliest scholars, Abū ʿAmr b. al-ʿAlāʾ (d. 154/770).111 In addition to this, Ibn al-Nadīm mentions twenty-one books entitled al-Nawādir, mostly by authors of the third/ninth century.112 Other works mention several books belonging to the same period and carrying the same title, two of which only have survived (nos. 1 and 2 in the list below). There are, however, some quotations from the lost titles in a number of lexica and other sources. For example, material from al-Aḫfaš al-Awsaṭ’s (d. 215/830) al-Nawādir and Ibn al-Aʿrābī’s (d. 231/845) al-Nawādir is quoted by several later authors.113 Moreover, Suyūṭī’s quotations preserved to

109 Suyūṭī, Muzhir I, 546, 549; II, 198–199, 271, 290–291. 110 See below, 248 ff., 203–204. 111 Ibn al-Nadīm, Fihrist 96. 112 Ibid., loc. cit. Among these books is K. al-Nawādir by Liḥyānī (fl. in the second half of

the second/eighth century), reconstructed by Ḥ. Ǧ. Ḥaddād from surviving quotations which, in his judgment, belong to the realm of ġarīb. The book will not be included in the discussion since it is not an original work and since the quotations, except for a small number of cases (e.g. 37, 38, 43, 49, 50, 69, 146, 189, etc.), do not explicitly mention that Liḥyānī’s views are derived from his al-Nawādir; cf. Ḥaddād (2007).

113 Sezgin (1982: 80, 271–272).

 

 

87MUBAWWAB AND SPECIALIZED LEXICA

us some linguistic data found in the Nawādir of Yūns b. Ḥabīb (d. 182/798), Abū ʿAmr al-Šaybānī (d. 206/821), and Ibn al-Aʿrābī.114 Other than books entitled al-Nawādir, some of the books that are lost to us include the terms ġarīb, waḥšī and šawāḏḏ in their titles, such as Abū Misḥal’s (d. 231/845) al-Ġarīb al-waḥšī and Ṯaʿlab’s (d. 291/904) al-Šawāḏḏ.115 Of the same genre are books entitled al-Luġāt, which, as we know from surviving quotations and from material on luġāt scattered in the extant sources, primarily include ġarīb dialectal usage. Among the authors to whom a book of this title is attributed by Ibn al-Nadīm are Yūnus b. Ḥabīb, Abū ʿAmr al-Šaybānī, Farrāʾ (d. 207/822), Abū ʿUbayda (d. 209/824), Abū Zayd al-Anṣārī (d. 215/830),116 Aṣmaʿī (d. 216/831), and Ibn Durayd (d. 321/933).117

The following list contains some of the most important extant sources of the genre of general ġarīb/nādir:118

1. al-Nawādir fī l-luġa by Abū Zayd al-Anṣārī (d. 215/830). 2. al-Nawādir by Abū Misḥal al-Aʿrābī (d. 231/845). 3. al-Muntaḫab min ġarīb kalām al-ʿArab and al-Muǧarrad fī ġarīb kalām

al-ʿArab wa-luġātihā by Kurāʿ al-Naml (d. 310/922). 4. Ġarīb al-luġa by Abū Bakr b. al-Anbārī (d. 328/940). 5. al-Mudāḫal fī l-luġa by Abū ʿUmar al-Zāhid, better known as Ġulām

Ṯaʿlab (d. 345/957). 6. Šaǧar al-durr fī tadāḫul al-kalām bi-l-maʿānī l-muḫtalifa by Abū l-Ṭayyib

al-Luġawī (d. 351/962). 7. Laysa fī kalām al-ʿArab by Ibn Ḫālawayhi (d. 370/980). 8. al-Musalsal fī ġarīb luġat al-ʿArab by Muḥammad b. Yūsuf al-Tamīmī

(d. 538/1143). 9. Section 2 of Ittifāq al-mabānī wa-ftirāq al-maʿānī by Sulaymān b. Banīn

al-Daqīqī (d. 614/1217). 10. al-Šawārid fī l-luġa by Ṣaġānī (d. 650/1252).

114 Suyūṭī, Muzhir I, 453; II, 275–276, 289 (Yūnus). For Abū ʿAmr, see n. 109 above, and for Ibn al-Aʿrābī, who is quoted nineteen times, see Muzhir’s index (II, 650).

115 Ibn al-Nadīm, Fihrist 52, 81. 116 It is highly probable that two of Ibn Durayd’s chapters in the latter part of Ǧamharat

al-luġa (III, 1294–1295, 1306–1311), both of which contain material on luġāt ascribed to Abū Zayd, are ultimately derived from this book.

117 Ibid., 48, 75, 73, 59, 60, 61, 67 respectively. 118 Parts of Abū ʿAlī al-Haǧarī’s (fl. in the second half of the third/ninth century) al-Taʿlīqāt

wa-l-nawādir have survived, but the book is mainly an anthology of poetry with only brief linguistic material and ample geographical and genealogical information. Cf. Ǧāsir (1992) for a study of the book and selections from its material.

 

 

88 chapter 2

Titles 5, 6, 8 and 9 form a separate group within the genre and will be discussed together. First, though, the other titles will be discussed in the chronological order in which they are listed above.

The title al-Nawādir is common to the first two books, authored by Abū Zayd and Abū Misḥal. Both books belong to the early period of data collec- tion or ǧamʿ al-luġa and have several features in common. We have already discussed the content of Abū Misḥal’s book in an attempt to determine what is meant by ġarīb or nādir.119 In this respect, Abū Zayd’s book is quite similar to Abū Misḥal’s as it records, for example, words that are rarely used in any of the four sources of linguistic data, dialectal differences, anomalous forms and expressions, as well as data on synonyms, homonyms, aḍdād (words with two contradictory meanings), ibdāl (substitution of consonants and semivow- els), elision of hamza, etc.120 Furthermore, both books lack organization, in particular Abū Misḥal’s, which consists of a large body of reported data listed as consecutive items not arranged according to any discernible criterion. As for Abū Zayd’s book, it is divided into fifteen chapters, two of which are on šiʿr (that is, poetry of all meters other than raǧaz), seven on raǧaz and six on nawādir. Beyond this division into chapters – which themselves are presented randomly and not arranged according to their three types or lengths121 – the book, like Abū Misḥal’s, lacks internal arrangement. In a typical chapter on raǧaz, Abū Zayd quotes lines of raǧaz and explains the ġarīb material which they contain. Chapters on šiʿr and nawādir, on the other hand, deal with poetry and prose material respectively. Yet, the division into three types is not decisive as chapters on raǧaz contain šiʿr and vice versa, in addition to other šawāhid derived from the Qurʾān (and to a lesser extent Ḥadīṯ), proverbs and idiomatic expressions. The chapters on nawādir are the closest to Abū Misḥal’s book and consist of a host of attested material normally introduced in both works by expressions such as yuqāl (it is said), qāla (he said), qālū (they said), taqūl (one says), etc. Whether the material introduced by such expressions is attributed to a particular tribe or not (the latter being the norm), it obviously is the result of data collection discussed in Section 3 of the previous chapter.

119 See above, 37–38. 120 For a list of dialectal material which contains numerous examples of the phenomena

mentioned above, see the index of Abū Zayd’s Nawādir 672–673. 121 For example, the two chapters on šiʿr are nos. 1 and 10, and several chapters on raǧaz

(nos. 2–4 and 6–8) are arranged consecutively without any clear difference between one chapter and the other. As far as length is concerned, chapters widely range from seventy- seven pages in the case of a raǧaz chapter (no. 2, pp. 161–237) to four pages in another raǧaz chapter (no. 6, pp. 321–324).

 

 

89MUBAWWAB AND SPECIALIZED LEXICA

Indeed, Abū Zayd is reported to have said that he heard the šiʿr which his book contains from al-Mufaḍḍal al-Ḍabbī (d. 168?/784?), whereas he heard the raǧaz and the dialectal material (luġāt) directly from the Arabs.122 The Kufan scholar al-Mufaḍḍal himself was well-known for his interest in ġarīb,123 and Abū Zayd, a Basran, clearly had no problem in relying heavily on his riwāya. Accordingly, the data in his book may be regarded as having both Basran and Kufan ori- gins, with no indication of any partisan preference. In the case of Abū Misḥal, it cannot be ruled out that he was the source of some data in his own book, al-Nawādir, since he is known both as an author and as one of the Bedouin fuṣaḥāʾ who were considered to be trustworthy ruwāt (informers, narrators).124

Compared with Abū Misḥal’s, Abū Zayd’s book gained more reputation and was quoted more in the sources. Its popularity may be due in part to the fact that it is in a sense similar to the well-known anthologies or iḫtiyārāt (such as the Mufaḍḍaliyyāt, Aṣmaʿiyyāt and Ḥamāsa), although Abū Zayd’s criteria for inclusion of both šiʿr and raǧaz in his work are purely linguistic, and more spe- cifically have to do with ġarīb. Rather than start with šiʿr or raǧaz, Abū Misḥal records usage and then adduces his šawāhid from either. This notwithstand- ing, the ġarīb material in both works is frequently quoted in the muǧannas lexica. Azharī (d. 370/981), for example, in his introduction to Tahḏīb al-luġa specifically mentions Abū Zayd’s al-Nawādir among the sources he relied on in authoring his dictionary, and praises it for its ġarīb and nādir material.125

As of the early fourth century, works on ġarīb naturally no longer followed the model adopted a century earlier by Abū Zayd and Abū Misḥal. With the completion of data collection, authors sought various ways of arranging their ġarīb material in a meaningful manner. The earliest sources from the period which have come into our possession are two works by Abū l-Ḥasan ʿAlī b. al-Ḥasan al-Hunāʾī (d. 310/922), better known as Kurāʿ al-Naml (lit. ant’s foot, due to his shortness or ugliness), or simply Kurāʿ. These are al-Muntaḫab min ġarīb kalām al-ʿArab and al-Muǧarrad fī ġarīb kalām al-ʿArab wa-luġātihā (see no. 3 in the above list). The first of the two, al-Muntaḫab, is in three parts and is similar in content to multithematic lexica, such as Abū ʿUbayd’s (d. 224/838)

122 Contrary to this statement by Abū Ḥātim al-Siǧistānī (d. 255/869), Tawwazī (d. 233/847) claims that Abū Zayd reports the raǧaz in his book on the authority of al-Mufaḍḍal al-Ḍabbī, and that he directly heard from the Arabs his data on šiʿr and luġāt; see Abū Zayd, Nawādir 42.

123 Cf. Qifṭī, Inbāh III, 304; Suyūṭī, Buġya II, 297. 124 See above, 20–21. 125 Azharī, Tahḏīb I, 12; cf. the title of Azharī’s chapter (I, 8): Ḏikr al-a⁠ʾimma allaḏīna ʿtimādī

ʿalayhim fī mā ǧamaʿtu fī hāḏā l-kitāb.

 

 

90 chapter 2

al-Ġarīb al-muṣannaf and Ibn Qutayba’s (d. 276/889) Adab al-kātib. The first and major part126 consists of various themes, such as human body, animals, qualities of appearance and behavior, metals, earth, sun, moon, food, bribery, plants, clothes, illnesses, etc. The second part127 is devoted to dialectal vari- ants in verbs and nouns, and is strikingly similar in content to Ibn al-Sikkīt’s (d. 244/858) Iṣlāḥ al-manṭiq. The third part128 includes various chapters which either resemble the first part or deal with specific linguistic phenomena such as rare usage, aḍdād (words with two contradictory meanings), qalb (metath- esis), ištiqāq (derivation), zawāʾid (augments), ḥaḏf (elision), Arabized words (muʿarrab), etc. The ġarīb material thus lies in the second part and some chap- ters in the third. But whereas ġarīb occurs mostly as supporting material for phenomena discussed in the third part, the second part is primarily concerned with ġarīb, and in fact provides the strongest justification for the inclusion of the term ġarīb in the book’s title. It is arranged into chapters according to nom- inal and verbal patterns, and each chapter includes words which have two or more variants in different dialects. Although Kurāʿ does not normally specify the form which he regards as less familiar, it is often easy to determine which of the cited forms may be described as ġarīb. For example, yaquẓ (vigilant) and sahur (wakeful) are clearly meant to be shown as ġarīb in comparison to yaqiẓ and sahir.129 In other cases, Kurāʿ cites examples of patterns which but rarely occur. Among these are the patterns finnaʿl (as in šinnaḫf “tall”), fuʿʿalla (as in kummahda “glans”), and funʿalla (as in ǧunbaqṯa “huge woman”).130 Also indic- ative of interest in rare and uncommon usage is material of the laysa type, typi- cally formulated as statements which begin with laysa (there is not) followed by a closed set of rare forms which are the only examples of a certain phe- nomenon. For example, no noun begins with two consecutive t’s other than tatful (fox, or a kind of plant or tree), and there is no word of the pattern fuʿyal other than ʿUlyab (place name).131 Later in the same century, Ibn Ḫālawayhi (d. 370/980), as we shall see later in this survey, devoted a whole book to this type of material which became largely associated with his name.

Kurāʿ’s other book, al-Muǧarrad, deals exclusively with ġarīb. In fact, Ibn al-Nadīm (d. 380/990) refers to it as Muǧarrad al-ġarīb and quotes its author’s introductory statement in which he says that his book deals with strange usage

126 Kurāʿ, Muntaḫab I, 46-II, 509. 127 Ibid., II, 509–557. 128 Ibid., II, 557–779. 129 Ibid., II, 509. 130 Ibid., II, 564–565. 131 Ibid., II, 559.

 

 

91MUBAWWAB AND SPECIALIZED LEXICA

and dialectal variants (ġarīb kalām al-ʿArab wa-luġātihā).132 The editor of the book admits adding this statement to the title of the book,133 although it is normally referred to in the sources simply as al-Muǧarrad. In all cases, Kurāʿ’s mentioning of luġāt with ġarīb in his introductory statement ties in well with the fact that the ġarīb material, which primarily occupies the second part of his al-Muntaḫab, is presented in the form of dialectal variants, as we have noted above. It should also be noted that the title of Kurāʿ’s book, al-Muǧarrad (lit. the one that is “devoid of” or “stripped of”) is due to its being a shortened ver- sion of a larger work, al-Munaḍḍad, which did not reach us and whose šawāhid and names of narrators were left out to make al-Muǧarrad more accessible to users.

The most interesting feature in al-Muǧarrad is certainly its arrangement, since it is one of the earliest works that are alphabetically arranged. Were it not restricted to ġarīb and thus placed with the general lexica of the muǧannas type, it would be chronologically second only to Abū ʿAmr al-Šaybānī’s (d. 206/821) K. al-Ǧīm since it precedes – at least based on death dates – Ibn Durayd’s (d. 321/933) Ǧamharat al-luġa. Kurāʿ in fact asserts that the alphabeti- cal arrangement of his lexicon should help its user find the word he looks for without much difficulty.134 He thus divides the lexicon into twenty-eight sec- tions or abwāb (i.e. ʾ, b, t, ṯ, ǧ, etc.), each of which is further divided into twenty- eight chapters or fuṣūl following the same order. By considering the first two radicals only, the chapters amount to 784 (i.e. 28 x 28), of which he says 631 are used (mustaʿmal) and 153 unused (muhmal). In each chapter he either lists the mustaʿmal or indicates that the two consonants at hand are muhmal, that is, they yield no biliteral, triliteral or other roots. In the section on ṯāʾ, for example, the muhmal chapters are ṯt, ṯṯ, ṯḥ, ṯḏ, ṯz, ṯs, ṯš, ṯṣ, ṯḍ and ṯẓ.135 Ibn al-Nadīm’s com- ment that al-Muǧarrad is modeled on K. al-ʿAyn but differs from it in arrange- ment (ʿalā miṯāl K. al-ʿAyn wa-ʿalā ġayr tartībihi)136 is to be understood as a reference to Kurāʿʾs distinction in each chapter between the mustaʿmal and the muhmal, as in al-ʿAyn, although the latter is phonetically and not alphabeti- cally arranged. Concerning the internal arrangement of each chapter however Kurāʿ does not strictly abide by the alphabetical order as determined by what follows the first and second letters. An example also drawn from the letter ṯāʾ is the chapter ṯr in which the roots of the cited words appear in the following

132 Ibn al-Nadīm, Fihrist 92; cf. Kurāʿ, Muǧarrad I, 31. 133 See his introduction to Kurāʿ’s Muǧarrad I, 15. 134 Ibid., I, 31. 135 Ibid., I, 357–362. 136 Ibn al-Nadīm, Fihrist 92.

 

 

92 chapter 2

order: ṯrʾ ṯrr, ṯrmṭ, ṯrmd, ṯrml, ṯrd, ṯrb, ṯrṭʾ, ṯrṭm, ṯrʿl, ṯrʿṭ and ṯrtm.137 It should also be noted that Kurāʿ, contrary to most lexicographers, takes augments (zawāʾid) into consideration in his arrangement. Hence, ṯābara, ṯāfana, ṯāmir, ṯāqib and ṯāqaftu are placed in the chapter of ṯāʾ followed by alif,138 and not under ṯbr, ṯfn, ṯmr, ṯqb and ṯqf respectively. As a result, the bāb of alif is the longest in his lexicon because it includes not only words which begin with an original hamza (e.g. uḥāḥ “thirst”)139 but also words which begin with a prefixed hamza (e.g. adlam “black”).140 Also included under alif are verbs (and their derivatives) of patterns such as istafʿala, iftaʿala, if ʿalalla and infaʿala, as in istaġraba (to guffaw), iftāqa (to become poor), isbaġalla (to become wet), and intiǧāʿ (migra- tion). The first and third of these verbs appear in the chapter of alif and sīn; the second in that of alif and fāʾ; and the fourth in that of alif and nūn.141 Similarly, verbs of the pattern tafaʿʿala are listed under tāʾ, as in tabaǧǧaḥa (to boast), which occurs in the chapter of tāʾ and bāʾ.142

Kurāʿ’s ġarīb material was a source for authors of mubawwab and muǧannas lexica alike. Ibn Sīda (d. 458/1066) for example mentions Kurāʿ’s books among the sources he used in his mubawwab lexicon, al-Muḫaṣṣaṣ,143 as well as in his muǧannas one, al-Muḥkam.144 Furthermore, Ibn Manẓūr (d. 711/1311) in his Lisān al-ʿArab quotes Kurāʿ in several hundred entries,145 and Zabīdī (d. 1205/1790) in his Tāǧ al-ʿarūs specifically mentions al-Muǧarrad in his sur- vey of earlier lexica.146

The four books discussed above, i.e. Abū Zayd’s al-Nawādir, Abū Misḥal’s al-Nawādir and Kurāʿ’s al-Muntaḫab and al-Muǧarrad, are the major works in the genre of general ġarīb. Later works are, on the whole, more specialized and thus are shorter and less comprehensive. Having placed nos. 5, 6, 8, and 9 in a special category, we are left with the books authored by Ibn al-Anbārī, Ibn Ḫālawayhi and Ṣaġānī.

137 Kurāʿ, Muǧarrad I, 359–361. Note that Kurāʿ makes no distinction between alif and hamza, hence ṯrʾ is placed in the beginning of the chapter although the verb ṯarā (to become numerous) ends with an alif, which is normally considered by philologists to be originally a wāw or yāʾ.

138 Ibid., I, 355–356. 139 Ibid., I, 72. 140 Ibid., I, 94. 141 Ibid., I, 127, 174, 123, 216 respectively. 142 Ibid., I, 310. 143 Ibn Sīda, Muḫaṣṣaṣ I, 12. 144 Ibn Sīda, Muḥkam I, 47–48. 145 See Lisān’s indices by Abū l-Hayǧāʾ & ʿAmāyira (1987: III, 483–488). 146 Zabīdī, Tāǧ I, 12.

 

 

93MUBAWWAB AND SPECIALIZED LEXICA

Abū Bakr b. al-Anbārī’s (d. 328/940) Ġarīb al-luġa is unique among the above listed sources of general ġarīb because it is made up of about a hundred lines of the author’s own poetry of the shortened (maǧzūʾ) raǧaz meter. Each line nor- mally consists of four ġarīb words which the author briefly explains (with very few šawāhid) before proceeding to the next line. Due to constraints of rhyme and meter, the book naturally lacks any alphabetical or thematic organization, and it seems that Ibn al-Anbārī made this sacrifice, so to speak, because he was first and foremost interested in the didactic aspect of his book, given that pupils are more likely to memorize poetry than prose.

No less unique than Ibn al-Anbārī’s book is Ibn Ḫālawayhi’s (d. 370/980) Laysa fī kalām al-ʿArab, already alluded to in our discussion of Kurāʿ’s al-Muǧarrad. The book deals almost exclusively with morphological patterns and phe- nomena which are so rare that words which represent them can be specified within a closed set. Chapters typically begin with the expression laysa fī kalām al-ʿArab . . . illā (There is not in the speech of the Arabs . . . except), hence the title of the book. In Chapter 81, for example, Ibn Ḫālawayhi states that in the speech of the Arabs, there are only three examples of the pattern fiʿla whose plural may be both of the patterns fiʿal and fuʿal. These are ḥilya (ornament), liḥya (beard), and ǧizya (tribute), whose plurals are ḥilā/ḥulā, liḥā/luḥā, and ǧizā/ǧuzā respectively.147 Similarly, he identifies in Chapter 146 only two words of the pattern faʿyal, namely, ḍahyad (robust) and Ṣahyad (place name).148 It is important to note that this type of formulation owes its origin to lexicogra- phers and grammarians as early as Ḫalīl (d. 175/791) and Sībawayhi (d. 180/796). Both authors were keen on establishing criteria for the morphological struc- ture of Arabic words in order to distinguish them from non-Arabic words and from neologisms created by those whom Ḫalīl calls naḥārīr.149 Ḫalīl, for exam- ple, asserts that no quadriliteral or quinqueliteral word can be devoid of liq- uids or labials except for about ten words (e.g. ʿasǧad “gold” and qudāḥis “brave and strong”), which he describes as anomalous (sawāḏḏ).150 More pertinently, he uses the expression laysa fī kalām al-ʿArab to deny the Arabic origin of the words duʿšūqa (ladybug) and ǧulāhiq (bullets) and any word which begins with n -r.151 For his part, Sībawayhi denies the existence in kalām (speech) of a number of patterns, such as fiʿul; if ʿul; fāʿul; fāʿayl, fāʿīl, fāʿawl, fāʿalāʾ; fiʿālā;

147 Ibn Ḫālawayhi, Laysa 162. 148 Ibid., 293. 149 Cf. above, 28. 150 Ḫalīl, ʿAyn I, 53; cf. II, 315; III, 323; cf. above, 56. 151 Ibid., I, 53; cf. II, 268; V, 243.

 

 

94 chapter 2

yafʿāl, yufʿūl, yufʿul; fiʿʿayl; fuʿlun, faʿlan; fiʿawwal, fuʿawwal, etc.152 Statements to this effect are normally introduced in the Kitāb by either of two expressions: laysa fī l-kalām and lā naʿlam fī l-kalām (We do not know in speech of . . .). At times, Sībawayhi specifies a closed set which includes the only examples of a certain pattern. For example, the only words of the pattern afanʿal are alanǧaǧ (aloes wood), Abanbam (place name), and alandad (strongly hostile).153 It is such statements, which enumerate words of a closed set, that Ibn Ḫālawayhi’s Laysa consists of, and although he obviously wanted his book to be as com- prehensive as possible,154 he seems to have left out a number of examples in several sets.155 Yet, as far as we know, the book is the only attempt to assemble, in one work, material of this ġarīb type which earlier scholars often mention in their books, often as necessary proof of their mastery of Arabic usage.

Ṣaġānī’s (d. 650/1252) book al-Šawārid fī l-luġa shows that interest in ġarīb continued for several centuries after the earliest books on ġarīb were authored during the period of data collection (second/eighth and third/ninth centu- ries). The title of the book is itself reminiscent of Abū ʿUbayda’s (d. 209/824) K. al-Šawārid156 which did not reach us, but whose title strongly indicates that it dealt with ġarīb. Suyūṭī (d. 911/1505) in fact equates šawārid with ġarāʾib and mentions Ṣaġānī’s book as an example of this genre of books.157 The period to which Ṣaġānī’s book belongs, however, differs significantly from the period of data collection since there were no longer any fuṣaḥāʾ from whom authors could receive linguistic data. This fact is very well reflected in the content of al-Šawārid, which is divided into four parts, all of which are fully dependent on material available in earlier sources. The first part is interestingly devoted to anomalous Qurʾānic readings (al-šawāḏḏ min al-qirāʾāt) since the author correctly realized that these readings contained a large body of ġarīb mate- rial, particularly as some of the individuals he quotes were not established Qurʾānic readers, but poets (specifically known for their raǧaz), such as Ruʾba b. al-ʿAǧǧāǧ (d. 145/762) and Abū Ḥayya al-Numayrī (d. 183/200), both of whom were known for using ġarīb in their raǧaz. In an obvious attempt to highlight

152 Sībawayhi, Kitāb IV, 244, 245, 249, 250, 255, 265–266, 268, 270, 274 respectively (semicolons in text indicate change of page); cf. Baalbaki (2008: 225).

153 Ibid., IV, 247; cf. above, 43, n. 218. See also Ibn Ḫālawayhi, Laysa 169–170 and a list of laysa- type statements used by Mubarrad (d. 285/898) in the indices of his Muqtaḍab IV, 205.

154 Cf. his introductory statement (Laysa 27) in which he declares that the book contains all the material of which he is cognizant (mā aḥāṭa bihi ʿilmī).

155 Cf. ʿAṭṭār’s introduction to Laysa 17–18. 156 Ibn al-Nadīm, Fihrist 59. 157 Suyūṭī, Muzhir I, 234.

 

 

95MUBAWWAB AND SPECIALIZED LEXICA

the ġarīb aspect of his book, Ṣaġānī begins each entry (according to the order of Qurʾānic sūras) by introducing a ġarīb usage, to be followed by the qirāʾa which exemplifies it. For example, he says that tahlika (destruction) is a vari- ant (luġa) of tahluka – which is obviously the standard form – and as proof ascribes to Ḫalīl b. Aḥmad the reading wa-lā tulqū bi-aydīkum ilā l-tahlika (and make not your own hands contribute to your destruction; Q 2: 195).158 Similarly, tabūt and tābūh are rare variants for tābūt, and both were read in Q: 248 by Zayd b. Ṯābit (d. 45/665) and Ubayy b. Kaʿb (d. 21/642).159

The other three parts of Ṣaġānī’s book include material on general ġarīb/ nādir drawn from earlier sources. Fortunately enough, the entire second and third parts preserved material from sources which have not survived. The sec- ond part lists 160 items which one of Sībawayhi’s teachers, Yūnus b. Ḥabīb (d. 182/798), mentions in his K. al-Luġāt, and which Ṣaġānī believes repre- sent usage reported only by Yūnus.160 Some of these items are also ascribed by Ṣaġānī in his al-ʿUbāb to Yūnus.161 The third part is much smaller than the second and consists only of thirty items which Ṣaġānī chooses from Abū Ḥātim al-Siǧistānī’s (d. 255/869) Taqwīm al-mufsad wa-l-muzāl ʿan ǧihatihi min kalām al-ʿArab, based on his belief that they were not reported by any other scholar.162 Unlike the second and third parts, the fourth is derived from several sources, some of which Ṣaġānī mentions by name, such as Aṣmaʿī’s (d. 216/831) al-Maqṣūr wa-l-mamdūd, Ibn al-Sarrāǧ’s (d. 316/929) Maʿānī l-šiʿr, Abū Bakr b. al-Anbārī’s (d. 328/940) al-Muḏakkar wa-l-muʾannaṯ, and Ǧawharī’s (d. c. 400/1010) al-Ṣaḥāḥ.163 In other cases, Ṣaġānī does not specify his sources, but it is often easy to identify them, as in the case of Abū ʿAmr al-Šaybānī’s (d. 206/821) K. al-Ǧīm and Ibn Ḫālawayhi’s (d. 370/980) Laysa.164 This part of the book is by far the largest165 and contains nearly 1,100 entries, some of which include poetry šawāhid. What is interesting is that the first and last few pages of this part are randomly arranged, but the remaining material, which forms more than three-quarters of the material,166 is alphabetically arranged according only to the first radical of the word. Words starting with the letter

158 Ṣaġānī, Šawārid 139–140. 159 Ibid., 142–143. 160 Ibid., 175–197. 161 See, for example, Ṣaġānī, ʿUbāb (SWS, FRĠ); cf. Ṣaġānī, Šawārid 180, 185. Cf. above, n. 114. 162 Ṣaġānī, Šawārid 201–205. 163 Ibid., 224, 226, 230, 354 respectively. 164 See examples in the editor’s introduction, 101–102. 165 Ibid., 209–366. 166 Ibid., 230–352.

 

 

96 chapter 2

ḍād, for example, appear in the following order (cited here in root form): ḌHB, ḌMĠ, ḌRS, ḌʾL, ḌYF, ḌBN, ḌǦʿ, ḌLL, ḌRR, ḌMD, ḌWD, and ḌRʾ.167 One pos- sible explanation is that Ṣaġānī did not have the chance to incorporate the first and last sections of this part into the bulk of his material. Yet a much better explanation is that the alphabetically arranged section is derived from a single source, most probably Šaybānī’s K. al-Ǧīm. The latter possibility is supported by the fact that most of the alphabetically arranged entries faithfully follow their arrangement in al-Ǧīm. Obviously, this can hardly be the result of chance, given that words in al-Ǧīm are arranged based on their first radicals only.

Within the genre of ġarīb/nādir is a type known as mušaǧǧar168 (branched), or mudāḫal (intermixed, intertwined), or musalsal (serialized), represented in the above list by items 5, 6, 8 and 9. The inventor of this type is most prob- ably Abū ʿUmar al-Zāhid (also known as al-Muṭarriz; d. 345/957), who was an apprentice of the Kufan philologist Ṯaʿlab (d. 291/904), and is hence usually referred to as Ġulām Ṯaʿlab. Neither he nor any of the subsequent authors of this type refers to a work earlier than his, although much of his material in al-Mudāḫal fī l-luġa is reported on the authority of earlier scholars, particularly Kufans, who include, other than Ṯaʿlab, Abū ʿAmr al-Šaybānī (d. 206/821), Farrāʾ (d. 207/822), Ibn al-Aʿrābī (d. 231/845), and Salama b. ʿĀṣim (d. after 270/883). Given the Kufan reputation for vastness of riwāya based on their contact with Bedouin fuṣaḥāʾ, it is not surprising that there is a Kufan connection in the establishment of this type of writing which primarily aims at demonstrating the extensive vocabulary of Arabic, particularly in the domain of ġarīb. Zāhid’s book is made up of thirty-one short chapters, each of which begins with a word which is explained by another word, itself then explained by a third, and so on. For example, the third chapter begins with the following words which the author heard from Ṯaʿlab on the authority of Ibn al-Aʿrābī: al-firsika: al-ḫawḫa; wa-l-ḫawḫa: al-ṯawb al-aḥmar; wa-l-aḥmar: allaḏī lā silāḥa maʿahu; wa-l-silāḥ: šaḥm al-ibil; wa-l-šaḥm: al-bayāḍ; wa-l-bayāḍ: al-laban; wa-l-laban: waǧaʿ al-ʿunuq min al-wisāda; wa-l-ʿunuq: al-ǧamāʿa min al-nās, etc.169 (al- firsika: peach; al-ḫawḫa: red garment; al-aḥmar: an unarmed person; al-silāḥ: camel’s fat; al-šaḥm: whiteness; al-bayāḍ: milk; al-laban: neck stiffness caused by a pillow; al-ʿunuq: group of people, etc.). As this specimen shows, words are cited for those meanings which might provide continuity for the chain of ġarīb usage at hand. Thus, ġarīb words in the beginning of each chapter are usually

167 Ibid., 301–303. 168 This term is chosen by Suyūṭī in introducing the thirty-first type of the linguistic sciences

(ʿulūm al-luġa) in his Muzhir I, 454–459. 169 Zāhid, Mudāḫal 30.

 

 

97MUBAWWAB AND SPECIALIZED LEXICA

explained by familiar ones, which in turn have ġarīb meanings that maintain the chain further.170

The second of the mušaǧǧar books carries the title Šaǧar al-durr fī tadāḫul al-kalām bi-l-maʿānī l-muḫtalifa, and is authored by the well-known philolo- gist, Abū l-Ṭayyib al-Luġawī (d. 351/962). Although Abū l-Ṭayyib does not refer in his book to Abū ʿUmar al-Zāhid, who was one of his teachers,171 it is quite clear that Šaǧar al-durr’s basic idea is derived from al-Mudāḫal. Other than the title of the book which includes the word tadāḫul, Abū l-Ṭayyib uses the term muḏāhala – reminiscent of the title of Zāhid’s book, al-Mudāḫal – in the first sentence in which he introduces his book in order to explain its struc- ture. He says that the book intermixes lexemes of different meanings (hāḏā kitāb mudāḫalat al-kalām bi-l-maʿānī l-muḫtalifa).172 Yet Abū l-Ṭayyib adopts a much more sophisticated approach than his predecessor, and his book is devised in what may be described as a geometric fashion. It is divided into chapters, each of which he calls šaǧara (tree) and includes a hundred words and ten poetry šawāhid. Each šaǧara (except for the sixth) is further divided into several furūʿ (pl. of farʿ, branch), each of which contains ten words, and two poetry šawāhid.173 Although the last šaǧara contains no furūʿ and only one poetry šāhid placed at its very end, it is more extensive than the first five since it includes five hundred words instead of one hundred. Abū l-Ṭayyib defends the choice of title for his book by alerting his reader to the fact that the ety- mology of šaǧara (and derivatives such as tašāǧara, ištiǧār, šiǧār and mišǧar) is closely related the notion of tadāḫul (cf. Zāhid’s title). The initial word in each chapter is the base from which the rest of the šaǧara and its furūʿ branch out. In the fourth šaǧara, for example, the initial word is ʿayn and the result- ing sequence begins as follows: al-ʿayn: ʿayn al-waǧh; wa-l-waǧh: al-qaṣd; wa-l- qaṣd: al-kasr; wa-l-kasr: ǧānib al-ḫibāʾ; wa-l-ḫibāʾ: maṣdar ḫāba⁠ʾtu l-raǧula iḏā ḫaba⁠ʾta lahu ḫabʾan wa-ḫaba⁠ʾa laka miṯlahu; wa-l-ḫabʾ: al-saḥāb, etc. (al-ʿayn: eye; al-waǧh: aim, intention; al-qaṣd: breaking; al-kasr: edge, lower hem of the tent; al-ḫibāʾ: verbal noun of ḫāba⁠ʾa, to mutually hide something from some- one; al-ḫabʾ: clouds, etc.).174 After the end of the sequence which consists of

170 Chapter 22 (pp. 73–75), contrary to most other chapters, is replete with ġarīb words that are explained by other ġarīb words. This is intended, however, to highlight the existence, in the dialect of the Aʿrābī faṣīḥ on whose authority these words are reported, unlike other chapters, of an extraordinary amount of ġarīb material.

171 Suyūṭī, Muzhir II, 120. 172 Abū l-Ṭayyib, Šaǧar 61. 173 Ibid., 61–62 174 Ibid., 161.

 

 

98 chapter 2

a hundred words, there are eight furūʿ, each of which starts with one partic- ular meaning of the same word, in this case, ʿayn, and comprises ten words arranged according to the same principle as in the šaǧara. The branching here is based on homonymy (see Section 8 below) as eight different mean- ings of ʿayn are cited, namely, ʿayn al-šams (the disc of the sun), naqd (spe- cie), mawḍiʿ infiǧār al-māʾ (fountainhead), ʿayn al-mīzān (tilt of the scales), maṭar lā yuqliʿ ayyāman (rain that continues for days), ra⁠ʾīs al-qawm (chief), nafs al-šayʾ (selfsame), and ḏahab (gold).175 It should be noted that, compared with the words which Zāhid’s Mudāḫal contains, those listed in both the ašǧār and the furūʿ of Šaǧar al-durr seem to be generally less ġarīb or nādir (at least as far as a contemporary observer is likely to determine). This may be due in part to the difference in size between the two books, as Abū l-Ṭayyib may not have been able, had he restricted himself to extremely strange or rare usage, to make his book as extensive as it is. Moreover, he might have wanted to depart in this respect from his teacher and prove that this genre is best served by bringing together two types of words, one of which is the truly ġarīb type, whereas the other does require explanation but is not as rare or unfamiliar as the first type.

Four of Abū l-Ṭayyib’s ašǧār and their furūʿ form most of the second section of Sulaymān b. Banīn al-Daqīqī’s (d. 614/1217) Ittifāq al-mabānī wa-ftirāq al-maʿānī (no. 9 in our list, mentioned here before no 8, Tamīmī’s al-Musalsal). The four ašǧār appear in the following order in Daqīqī’s book: (1) šaǧarat al-ʿayn (eye); (2) šagarat al-ṯawr (ox); (3) šaǧarat al-hilāl (crescent); and (4) šagarat al-ruʾba (need).176 The order in which these appear in Šaǧar al-durr is 4, 3, 2 and 5 respectively, and the difference in order between the two books may be due to differences in the manuscripts of Šaǧar al-durr, as the editor points out.177 The difference in order may not be intentional and, by itself, is not enough proof that Daqīqī tried to conceal the source of his material. Yet, other evidence sug- gests that he was indeed trying to do that. First of all, he mentions Abū l-Ṭayyib only once, in the beginning of the first šaǧara.178 Had the three other ašǧār been cited without interruption, the reader might think that all four ašǧār and their furūʿ are ascribed to Abū l-Ṭayyib based on that single mention, in spite of the length of the quoted material. But Daqīqī inserts, following each šaǧara and its furūʿ, a number of chapters on homonyms, including aḍdād (i.e. words

175 Ibid., 171–172, 174, 177, 181, 183, 188, 190 respectively. 176 Daqīqī, Ittifāq 107, 147, 163, 181. 177 Ibid., 56. 178 Ibid., 108.

 

 

99MUBAWWAB AND SPECIALIZED LEXICA

with two contradictory meanings),179 and thus interrupts the material adopted from Abū l-Ṭayyib without mentioning him as his source when he resumes quoting him. Moreover, Daqīqī provides his own introduction to the first, sec- ond and fourth ašǧār, in which he mentions a number of meanings of the word after which the šaǧara is named, further giving the impression that he is the author of the material that follows. Irrespective of the validity of these doubts, the section of his book which deals with mušaǧǧar is merely a replicum of the best part of Abū l-Ṭayyib’s book.

Also of the mušaǧǧar type is al-Musalsal fī ġarīb luġat al-ʿArab by the Andalusian scholar Abū l-Ṭāhir Muḥammad b. Yūsuf al-Tamīmī (d. 538/1143). Based on his introduction, it seems that he only knew of Zāhid’s al-Mudāḫal,180 since he nowhere mentions Abū l-Ṭayyib’s Šaǧar al-durr. He expresses the view that al-Mudāḫal is too short (istanzartuhu) and incomplete (lam yustawfa tamāmuhu), and even accuses its author of improvisation and haste (irtaǧalahu rtiǧālan wa-ǧarat rakāʾibuhu fīhi ʿiǧālan).181 Accordingly, Tamīmī’s book is meant to be much more comprehensive than his predecessor’s, and indeed is about three times longer. It is, like other mušaǧǧar works, based on a succession of lexemes in each chapter, but the innovation introduced by Tamīmī is that in each of his fifty chapters he was keen to start with a line of poetry, one of whose words serves as the first in a series, which also ends with a line of poetry. Coupled with the abundance of poetry šawāhid in the book (more than four hundred in all), Tamīmī’s method of beginning and ending all chapters with poetry šawāhid is a reminder of the close affinity between poetry and ġarīb material. It is quite telling that the book contains twenty-five words, the specific meanings assigned to which by Tamīmī are not found in the muǧannas lexica.182 Examples include ziyāda (in the sense of maqbiḍ al-qaws “bow’s handgrip”), qāmūs (in the sense of nammām “talebearer”), bāzī (in the sense of ʿatīq “ancient”), and saḥīq (in the sense of al-fatīt min al-misk “crum- bled musk”).183 This shows that, in spite of the arduous effort of the authors of muǧannas lexica to embrace in their works ġarīb material scattered in various mubawwab lexica, the latter do report meanings which the muǧannas lexica fail to cite for some words included therein.

179 See, for example, ibid., 124, 138, 143–144, 159, 179. 180 Tamīmī, Musalsal 34. 181 Ibid., 34–35. 182 See editor’s introduction to Musalsal, 8–9. 183 Ibid., 78, 131, 215, 307.

 

 

100 chapter 2

3 al-Amṯāl (Proverbs)

Arab scholars were interested in proverbs from various perspectives, includ- ing the historical, the cultural and the philological. Many proverbs form an essential part of fables and of narratives such as Ayyām al-ʿArab (Battle-days of the Arabs), and many others were considered worthy of being recorded in tandem with expressions representative of ḥikma (wisdom), and even shown to have parallels in Qurʾān and Ḥadīṯ. Yet, the dominant aspect in the study of amṯāl is philological, and it is not surprising that a large number of extant and lost works of the genre were authored by famous philologists, as we shall see. In this respect, the complementarity between grammar (naḥw) and phi- lology (luġa) is evident. In both disciplines, proverbs were viewed as a special type of kalām al-ʿArab, which has its own idiosyncrasies and is thus worthy of being recorded (i.e. as part of the process of data collection), and of being used as testimony for faṣīḥ usage. But whereas the grammarians focused on the syntactical aspects of proverbs, the philologists were more interested in explaining their meanings and metaphorical significance, relating them to specific individuals or incidents, and eventually arranging them in the form of lexica. The boundaries between the two groups of scholars are, of course, difficult to draw at times because of the diverse interests of many authors. For example, Zamaḫšarī (d. 538/1144), who is a most renowned grammarian, authored al-Mustaqṣā fī amṯāl al-ʿArab, which is one of the largest and most well-organized works of the genre. In this book, Zamaḫšarī is closer to philolo- gists than to grammarians, given that he concentrates on the interpretation of the meanings of proverbs and not their syntactical characteristics.

Before discussing the various books of proverbs, it may be useful to briefly point out the typical interest of grammarians in this part of their data. For this purpose, we shall consider the most ancient and creative work in the history of grammar, namely, Sībawayhi’s (d. 180/796) al-Kitāb, which contains some forty proverbs.184 These are adduced as šawāhid derived from kalām al-ʿArab along with stereotype expressions which are not normally considered proverbs.185 Several observations concerning the use of proverbs in the Kitāb can be made. First, they are linked to the notions of kaṯra (frequency) and saʿat al-kalām (latitude of speech). Thus, the proverb kilayhimā wa-tamran (Both of these

184 See the indices of ʿUḍayma (1975: 893–895) and of Hārūn’s edition of the Kitāb V, 33–34. 185 See a list of such expressions in Hārūn’s indices of the Kitāb V, 35–43. Examples include

raǧaʿa adrāǧahu (He retraced his steps; ibid., I, 415), kallamtuhu fāhu ilā fiyya (I spoke to him, his mouth being near to mine; ibid., I, 377, 391–392), and lā abā laka (May you have no father; ibid., II, 206, 279).

 

 

101MUBAWWAB AND SPECIALIZED LEXICA

and dates) is quoted as an example of frequent usage which results in elision (here, the verb aʿṭinī “give me”).186 As for saʿat al-kalām, it is explicitly linked with commonly used proverbs (cf. Sībawayhi’s term al-maṯal al-ǧārī), as in the proverb aṭirrī innaki nāʿilatun wa-ǧmaʿī (Act with boldness for you are wear- ing sandals and gather [the cattle to the sides of the valley]), which was origi- nally addressed to a woman, but may be used – without changing the feminine verb – to address a man.187 The second observation is that the foremost reason for citing proverbs in the Kitāb is to demonstrate brevity of expression, usually by the speaker’s elision of a part of the structure and its subsequent taqdīr (suppletive insertion) by Sībawayhi.188 Proverbs thus attest to the validity of a major analytical tool (i.e. taqdīr) on which the grammatical theory is depen- dent. The third observation is that Sībawayhi often cites proverbs because they preserve rare usage. The proverb ʿasā l-ġuwayru abʾusā (Perhaps the little cave is calamities) is thus brought as a šāhid for the rare phenomenon of using a noun, and not a verb, as the predicate of laysa.189 Similarly, inna l-fukāhata la-maqwadatun ilā l-aḏā (Joking is conducive to hurt) provides a rare example of a word (i.e. maqwada, rather than *maqāda) which is not subject to rules of vowel mutation.190 Later grammarians largely adopt Sībawayhi’s approach and even use as šawāhid his corpus of proverbs, as does Mubarrad (d. 285/898) for example in his al-Muqtaḍab.191

The above observations clearly demonstrate the special status which the grammarians give for proverbs in their study of structure. Above all, their view that the exact wording of proverbs may not be changed is a major factor in establishing this special status. Sībawayhi was the first to allude to the fact that proverbs should be quoted verbatim. In his discussion of the previously quoted saying aṭirrī fa-innaki nāʿilatun wa-ǧmaʿī, he explains that when it is used in addressing a man, the intention of the speaker is to tell the male addressee that he is of the same status (manzila) as the woman who was first addressed with this construction.192 Borrowing Sībawayhi’s exact words about the use of this proverb, Mubarrad cites a similar proverb addressed to a woman, al-ṣayfa

186 Ibid., I, 280–281; cf. IV, 44 for another instance of a proverb linked to kaṯra. 187 Ibid., I, 292–293. 188 E.g. ibid., I, 255–256, 260–261, 268, 272–273, 275, 280, 282; II, 231, 391, etc. 189 Ibid., I, 51, 159; III, 158. 190 Ibid., IV, 350. 191 See Mubarrad, Muqtaḍab II, 145; III, 70, 72, and compare the list of proverbs in Muqtaḍab

(indices, IV, 265–266) with that in the Kitāb (indices, V, 33–34). For other examples of proverbs which the grammarians use as syntactical šawāhid, see Qaṭāmiš (1988: 230 ff.).

192 Sībawayhi, Kitāb I, 292–293.

 

 

102 chapter 2

ḍayyaʿti l-labana (In summer you wasted the milk), and in no unclear terms establishes the principle that proverbs are to be quoted verbatim (ʿalā mā ǧarā fī l-aṣl). This principle is adopted by many lexicographers, such as Ibn Durayd (d. 321/933), who goes as far as saying that even if solecism appears in proverbs, the Arabs quote them without any change.193 Authors of books on proverbs also uphold this principle. ʿAskarī (d. after 395/1005) for example succinctly formulates it by saying al-amṯāl tuḥkā (Proverbs are quoted ver- batim) and lā tuġayyar ṣīġatuhā (Their structure is not subject to change).194 Similarly, Zamaḫšarī (d. 538/1144) states that al-amṯāl yutakallam bihā kamā hiya (Proverbs are quoted as is).195 A number of similar statements on the impermissibility of changing proverbs are quoted by Suyūṭī in Chapter 35 of his al-Muzhir.196

At a scale larger than the affinity between grammarians and philologists in dealing with proverbs, it should be noted that there is a strong bond between the study of proverbs and the material classified as ġarīb. Suffice it to point out that ʿAskarī stipulates that knowledge of ġarīb is a necessary skill for anyone who applies himself to the study of what he calls the unique type of knowledge (nawʿ min al-ʿilm munfarid bi-nafsihi)197 which proverbs represent, and that Ḫuwayyī (d. c. 549/1154) considers the study of proverbs to be one of the components of Arabic linguistics (ʿilm al-ʿArabiyya),198 in which ġarīb, of course, has a central role. Given the lexicographers’ interest in explaining ġarīb words as well as providing šawāhid for the words which they explain, it is not surprising that proverbs are frequently quoted in muǧannas lexica. In the earliest such lexicon, Ḫalīl’s (d. 175/791) al-ʿAyn, there are 70 proverbs, and 315 in Ibn Durayd’s (d. 321/933) al-Ǧamhara. In later and more comprehensive lexica, the number is considerably higher, as in Ibn Manẓūr’s (d. 711/1311) Lisān al-ʿArab, which contains 2,278 proverbs.199 The books on amṯāl were obviously the ultimate source of many, if not most of these proverbs. Azharī (d. 370/981) for example mentions that he read Abū ʿUbayd’s K. al-Amṯāl to his teacher (aqra⁠ʾanīhi; qara⁠ʾtu ʿalayhi), Abū l-Faḍl al-Munḏirī (d. 329/939), and that it is

193 Ibn Durayd, Ǧamhara I, 373. 194 ʿAskarī, Ǧamhara I, 7. 195 Zamaḫšarī, Mustaqṣā, Introduction I, h. 196 Suyūṭī, Muzhir I, 487–488. 197 ʿAskarī, Ǧamhara I, 5. 198 Ḫuwayyī, Farāʾid 18–19. 199 In one instance, Ibn Manẓūr quotes fourteen proverbs related to naʿāma (ostrich) in a

single entry; see Lisān (NʿM).

 

 

103MUBAWWAB AND SPECIALIZED LEXICA

the source of the proverbs quoted in his Tahḏīb al-luġa.200 This demonstrates, as in other genres, that the material of mubawwab lexica was one of the major building blocks of the muǧannas ones.

The genre of writing which deals with proverbs was among the earliest to appear in the history of the mubawwab type of lexica. Bibliographical and bio- graphical sources mention three early scholars to each of whom they ascribe a book on proverbs. These are Ṣuḥār b. ʿAyyāš (or ʿAbbās or ʿĀbis),201 ʿIlāqa b. Karīm (or Kuršum or Kursum)202 al-Kilābī, and ʿAbīd b. Šarya (or Sarya or Sāriya)203 al-Ǧurhumī. Although we do not have firm dates for these three paro- emiographers, they probably flourished towards the middle of the first/seventh century. Ṣuḥār, for example, was a ṣaḥābī (Companion of the Prophet),204 and ʿAbīd, who is said to have met Muʿāwiya b. Abī Sufyān (d. 60/680) and told him that he was 120 years old, was still alive when ʿAbdalmalik b. Marwān became Caliph in 65/685.205 In spite of such early dates we have reason to believe that these scholars were interested in proverbs and probably collected some. The ascription by Ibn al-Nadīm (d. 380/990) to each of them of a book entitled K. al-Amṯāl, however, most probably means that the content of each had been orally preserved by a succession of later scholars, but was then recorded with proper acknowledgment of the source. With regard to ʿIlāqa, Ibn al-Nadīm explicitly mentions that he saw his book which consists of about fifty pages.206 Furthermore, Abū ʿUbayd al-Bakrī (d. 487/1094), whose commentary on Abū ʿUbayd al-Qāsim b. Sallām’s (d. 224/838) K. al-Amṯāl is well-known (see infra), quotes all three scholars in his book,207 although it is not clear whether his quotations are taken directly from their so-called books or from riwāyas attributed to them orally in earlier sources. ʿAbīd’s book is also mentioned by Maydānī (d. 518/1124) in the introduction to his Maǧmaʿ al-amṯāl,208 but it cannot be ascertained whether Maydānī meant to include it among the fifty

200 Azharī, Tahḏīb I, 13, 20; cf. I, 26. 201 All three possibilities are mentioned by Ibn Ḥaǧar in Iṣāba II, 176. 202 Karīm and Kuršum appear in different manuscripts of Ibn al-Nadīm’s Fihrist 102, and

Kursum is used by Yāqūt, Muʿǧam IV, 1630. 203 Yāqūt (Muʿǧam IV, 1581) mentions the three possibilities, whereas Ibn al-Nadīm (Fihrist

102) mentions Šarya only. 204 Ibn Ḥaǧar, Iṣāba II, 176–178 (no. 4041). 205 Yāqūt, Muʿǧam IV, 1583. 206 Ibn al-Nadīm, Fihrist 102 (also quoted in Yāqūt, Muʿǧam IV, 1630). 207 Bakrī, Faṣl 388 (Ṣuḥār); 42, 79, 110, 364, 415, 426 (ʿIlāqa); 69, 79, 103, 210, 295, 319–320 (ʿAbīd).

See also Sellheim (1954: 29). 208 Maydānī, Maǧmaʿ I, 4.

 

 

104 chapter 2

sources which he says that he consulted in authoring his book, although the text allows for such an interpretation.

Apart from these three early scholars, books that deal with proverbs are attributed in the bibliographical and biographical sources, as well in later books on proverbs, to a number of the most famous philologists of the second half of the second/eighth century and the first half of the third/ninth century. Among these are Basran and Kufan scholars, including Abū ʿAmr b. al-ʿAlāʾ (d. 154/770), Šarqī (or al-Walīd) b. al-Quṭāmī (d. 155/772), Yūnus b. Ḥabīb (d. 182/798), Abū l-Ḥasan al-Liḥyānī (fl. in the second half of the second/eighth century), al-Naḍr b. Šumayl (d. 203/819), Abū ʿAmr al-Šaybānī (d. 206/821), Abū ʿUbayda Maʿmar b. al-Muṯannā (d. 209/824), Aṣmaʿī (d. 216/831),209 Ibn al-Aʿrābī (d. 231/845), Tawwazī (d. 233/847), Ibn al-Sikkīt (d. 244/858), and Ǧāḥiẓ (d. 251/869). The abundance of books on proverbs authored by these philologists demonstrates their keen interest in the study of this genre of kalām al-ʿArab. Although some of these works are quoted by subsequent authors of books on proverbs, none of them has survived. There is, however, a relatively large number of extant sources of the genre, and these will be surveyed in the rest of this section.

In order to facilitate the discussion, we shall classify the extant sources in three categories. The first includes books that are not arranged alphabeti- cally; the second those that are arranged alphabetically either partially or fully; and the third those that specialize in a particular aspect of proverbs (such as proverbs of the pattern afʿal or those that are derived from poetry, Ḥadīṯ, etc.), irrespective of their arrangement.

3.1 Books with no Alphabetical Arrangement It is natural that this category of books be chronologically prior to the other two categories. The dates of its authors range from about 168/784 to 328/940, whereas the first author to arrange proverbs alphabetically is Ḥamza b. al-Ḥasan al-Iṣfahānī (d. after 351/962). In fact, six of the first earliest seven books of the whole genre belong to this category, the exception being al-Ḥusayn b. al-Faḍl’s (d. 282/895) K. al-Amṯāl al-kāmina fī l-Qurʾān.210 These seven books are:

1. Amṯāl al-ʿArab by al-Mufaḍḍal b. Muḥammad al-Ḍabbī (d. 168?/784?). 2. al-Amṯāl by Muʾarriǧ al-Sadūsī (d. between 193/808 and 198/813).

209 Proverbial material attributed to Aṣmaʿī in the sources has been collected by Iyād ʿAbdalmaǧīd Ibrāhīm (Muʾassasat al-Warrāq li-l-Ṭibāʿa wa-l-Našr, Amman 2002) and misleadingly published in the form of an alphabetically arranged lexicon under the title K. al-Amṯāl!

210 See below, 127.

 

 

105MUBAWWAB AND SPECIALIZED LEXICA

3. al-Amṯāl by Abū Zayd al-Anṣārī (d. 215/830). The book itself is lost to us, but a very short abridgement by a certain ʿAlī b. Muḥammad al-Nīlī has survived and was published under the title of al-Muḫtār min K. al-Amṯāl.

4. al-Amṯāl by Abū ʿUbayd al-Qāsim b. Sallām (d. 224/838). Based on it are Abū ʿUbayd al-Bakrī’s (d. 487/1094) Faṣl al-maqāl fī šarḥ K. al-Amṯāl and Sulaymān b. Mūsā l-Kalāʿī’s (d. 634/1237) Nuktat al-amṯāl wa-nafṯat al-siḥr al-ḥalāl.

5. al-Amṯāl by Abū ʿIkrima al-Ḍabbī (d. 250/864). 6. al-Fāḫir by al-Mufaḍḍal b. Salama (d. 290/903 or 300/913). 7. al-Zāhir fī maʿānī kalimāt al-nās by Abū Bakr b. al-Anbārī (d. 328/940).

al-Mufaḍḍal’s book reached us through the riwāya of his student Abū l-Ḥasan al-Ṭūsī (d. ?), also a student of Ibn al-Aʿrābī’s (d. 231/845), himself an author of a book on proverbs, as previously mentioned. As the editor notes, the text includes comments by Ṭūsī, Ibn al-Aʿrābī and Kisāʾī (d. 189/805), which indi- cates that Ḍabbī’s text was not preserved in its original form.211 The collection consists of eighty-eight stories mostly related to Ayyām al-ʿArab (Battle-days of the Arabs, such as the wars of Dāḥis and al-Basūs)212 and to a number of indi- viduals of the Ǧāhiliyya (pre-Islamic period), such as Luqmān, al-Nuʿmān b. al-Munḏir and al-Sulayk b. al-Sulaka.213 As for the proverbs themselves (more than two hundred in all), they are mentioned within the stories according to which the book is arranged. In a few cases, Mufaḍḍal explains some ġarīb words which occur in these proverbs,214 as well as in some of the poetry which the stories contain.215

A significant shift in the arrangement of material took place in later works, whereby proverbs were quoted consecutively followed by the explanation of their meanings and, at times, the occasion on which they were first used. The first extant book to adopt this method in which proverbs were not sub- sidiary to the stories with which they are connected is K. al-Amṯāl by Muʾarriǧ al-Sadūsī (d. between 193/808 and 198/813), a Bedouin who settled in Basra and is often referred to as Abū Fayd. The book consists of 104 items which seem to be randomly arranged. It should be noted, however, that a few of these items consist of one word and are not truly proverbs, and it is not clear why they are included in the first place. If, on the other hand, these words have anything

211 Ḍabbī, Amṯāl 38–39; cf. Sellheim (1954: 46). 212 Ibid., 81, 108 (Dāḥis); 130 (al-Basūs). 213 Ibid., 151–163 (Luqmān); 50–52, 59–61, 113–116 (al-Nuʿmān); 61–64 (al-Sulayk). 214 Ibid., 51, 57, 61, 63, 66, 73, etc. 215 Ibid., 48, 60, 70–72, 80–81, etc.

 

 

106 chapter 2

to do with proverbs, they might have been quoted because they were parts of proverbs which are unknown to us. What is noteworthy is that these words mostly belong to the category of ġarīb, as in the case of ǧumma (a company demanding a bloodwit or some other purpose), waqb/waqba (hole, cavity), and suwāf (mortal disease of camels).216 Also indicative of Muʾarriǧ’s interest in the philological aspect of proverbs is the relatively large number of šawāhid (ninety-five lines of šiʿr and eighty-two of raǧaz)217 which are cited in such a short monograph as his. Moreover, his priority after citing a proverb is nor- mally to explain the meaning of some of its words which may not be easily understood by the reader. It is also telling that Muʾarriǧ quotes in his book sev- eral Aʿrāb from whom he derives some of his proverbs, šawāhid and meaning of ġarīb words.218 In short, his book is much more philologically oriented that his predecessor’s, al-Mufaḍḍal.

Although the third book in the above list, K. al-Amṯāl by Abū Zayd al-Anṣārī (d. 215/830), is lost to us, we know about some of its characteristics from a very short abridgement by ʿAlī b. Muḥammad al-Nīlī, of whom we have no informa- tion from the sources. The manuscript of the abridgement consists of a single page, which begins with the expression hāḏihi amṯāl muntaḫaba min Amṯāl Abī Zayd Saʿīd b. Aws al-Anṣārī (These are proverbs chosen from the Book of Proverbs by Abū Zayd . . .).219 Accordingly, the editor gave it the title al-Muḫtār min K. al-Amṯāl. Although the sources do not mention this abridgement, they confirm the existence of a book on amṯāl by Abū Zayd. Maydānī (d. 518/1124) for example mentions it as one of the sources he used in his Maǧmaʿ al-amṯāl,220 and attributes to Abū Zayd material which is probably derived from it.221 The book is also mentioned among Abū Zayd’s works listed by Yāqūt (d. 626/1229) and Ṣafadī (d. 764/1363),222 and Ibn Manẓūr quotes one of its proverbs.223 Based on the abridgement, the proverbs in the book are not alphabetically arranged, but consecutively listed, as in Muʾarriǧ’s al-Amṯāl; that is, they are not mentioned within stories, as in Mufaḍḍal’s book. The book is also similar to Muʾarriǧ’s in that the proverbs are followed by their meanings and often by the occasion of their first occurrence.

216 Muʾarriǧ, Amṯāl 57, 60, 97. 217 See the editor’s introduction 22. 218 Ibid., 53 (Abū Ḫālid al-Kilābī); 81 (Abū l-Ḫansāʾ); 62, 66, 68, 80 (Abū l-Duqayš). 219 Nīlī, Muḫtār 80. 220 Maydānī, Maǧmaʿ I, 4. 221 E.g. ibid., II, 369 (proverb 4387). See also Takrītī (1974: 12). 222 Yāqūt, Muʿǧam III, 1362; Ṣafadī, Wāfī XV, 201–202. 223 Ibn Manẓūr, Lisān (ĠRR). The proverb in question is anā ġarīruka min hāḏā l-amri (I am

knowledgeable enough in this matter that I do not have to prepare my answer).

 

 

107MUBAWWAB AND SPECIALIZED LEXICA

Abū ʿUbayd al-Qāsim b. Sallām’s (d. 224/838) K. al-Amṯāl represents a major development in the genre. As far as content is concerned, it includes 1,386 prov- erbs, and is thus much larger than the extant works discussed above. Whether similar works that are attributed in the sources to philologists contemporary with Abū ʿUbayd – such as Abū ʿUbayda, Aṣmaʿī and Ibn al-Aʿrābī – were of comparable size is something that we are unable to determine. But although Abū ʿ Ubayd might not be an innovator when it comes to the size of his book, his arrangement of its material is most probably unprecedented. It is divided into major themes that are further divided into 270 chapters which include its 1,386 proverbs. Among the major themes are speech, family relations, noble man- ners, glory, generosity, property, knowledge, sound opinion, injustice, disgrace, error, stinginess, cowardice, and calamities. Under the theme of stinginess (buḫl), for example, there are twelve chapters (comprising forty proverbs) that deal with more specific matters related to the main theme, such as the charac- teristics of the miser, stinginess of the wealthy, the miser who instructs others to be miserly, the miser who is coerced into giving out money, the previously wealthy miser who becomes poverty-stricken, etc.224 That this is an innovative manner of introducing proverbs is supported by a statement attributed to Ibn Durustawayhi (d. 347/958), in which he mentions several Basran and Kufan scholars who authored books on proverbs before Abū ʿUbayd, but asserts that the latter’s contribution was to collect their riwāyas and to classify the material of his own book and present it in an organized manner (ǧamaʿa riwāyātihim fī kitābihi wa-bawwabahu abwāban wa-aḥsana ta⁠ʾlīfahu).225 It is noteworthy that Abū ʿUbayd’s al-Ġarīb al-muṣannaf is as well organized as his al-Amṯāl, and is similarly divided into major themes which are in turn subdivided into more specific subjects. Moreover, both books lack any alphabetical arrange- ment, and it is remarkable, in al-Amṯāl, that this was not considered an option even in a chapter which readily lends itself to such an arrangement. This is the chapter entitled bāb ḏikr al-amṯāl fī muntahā l-tašbīh wa-ġāyatihi (Chapter on proverbs which imply the utmost degree of comparison).226 Later authors nor- mally refer to these proverbs, of which Abū ʿUbayd mentions seventy-three, as those of the pattern afʿal 227 since they all begin with comparative adjectives of this pattern. Had Abū ʿUbayd considered the introduction of an alphabetical

224 Abū ʿUbayd, Amṯāl 306–315. 225 Baġdādī, Tārīḫ XII, 404–405; cf. Qifṭī, Inbāh III, 14. 226 Abū ʿUbayd, Amṯāl 360–375. 227 Note that Abū Hilāl al-ʿAskarī (d. after 395/1005) in his Ǧamharat al-amṯāl (I, 6) refers

to these proverbs as the ones of the afʿal pattern (al-amṯāl ʿalā afʿal min kaḏā) as well as those which indicate the utmost degree of comparison (tanāhī) and exaggeration (mubālaġa).

 

 

108 chapter 2

arrangement to any one section of his book contrary to the other sections, this would have been the most likely candidate. Obviously, however, Abū ʿUbayd was interested solely in the thematic arrangement of his material. Despite the fact that Abū ʿUbayd’s book was held in high esteem by later authors, his method of organization was largely abandoned in favor of the alphabetical arrangement as this survey will demonstrate.

Other than organization, the book is characterized by the abundance of its philological material as Abū ʿUbayd often gives the meaning of words that he deems to be ġarīb or not commonly used, or explains his proverbs in a man- ner that reveals the meaning of such words. He also sporadically mentions the occasion on which a proverb was first used, or assigns its first user. As for sources, he normally begins his chapters by specifying the authority (or authorities) from whom he derives the proverbs that follow. In certain cases, it is indicated that he himself is the source of the proverbs cited (cf. the expres- sion: qāla Abū ʿUbayd).228

Many of the scholars after Abū ʿUbayd wrote commentaries on his al-Amṯāl or expanded its material, as is clearly indicated in a copy owned by Abū Bakr b. al-Anbārī (d. 328/940).229 According to this copy, among these scholars are Salama b. ʿĀṣim (d. after 270/883) and al-Zubayr b. Bakkār (d. 256/870), both of whose comments, according to Abū ʿUbayd’s scribe (kātib), ʿAlī b. ʿAbdalʿazīz b. al-Marzubān al-Baġawī (d. 286/899), were incorporated into the text and duly ascribed to them.230 Other additions to the book, which are said to be many times its size, are ascribed by Azharī (d. 370/981) in his Tahḏīb al-luġa to Abū l-Faḍl al-Munḏirī (d. 329/940),231 although no reference is made to him in the manuscript used in the published text. It was in the Andalus that the book made its largest impact. An abridgement of its content, faithful to its thematic arrangement, is incorporated by Ibn ʿAbd Rabbihi (d. 328/940) into his al-ʿIqd al-farīd.232 Furthermore, a well-known Andalusian work associated with Abū ʿUbayd’s book is Abū ʿUbayd al-Bakrī’s (d. 487/1094) Faṣl al-maqāl fī šarḥ K. al-Amṯāl. In his introduction, Bakrī justifies authoring the book by accusing Abū ʿUbayd of not explaining many of his proverbs, or of neglecting

228 E.g. Abū ʿ Ubayd, Amṯāl 41–42, 46–48, 52–53, 55–56, 59, 61, 66, etc. For a more comprehensive study of the characteristics of Abū ʿUbayd’s book, and its influence on later authors, cf. Sellheim (1954: 78 ff.). See also Kassis (1999: 15).

229 Ibid., 33–34. 230 Ibid., 37, 59, 68, 95, 176 (Salama); 37, 42, 53–56, 59, 62, 64, 68, 71–72, etc. (al-Zubayr). 231 Azharī, Tahḏīb I, 20. 232 Ibn ʿAbd Rabbihi, ʿIqd III, 82–132. This abridgement is the largest part of what Ibn ʿAbd

Rabbihi calls K. al-Ǧawhara fī l-amṯāl (ʿIqd III, 66–134).

 

 

109MUBAWWAB AND SPECIALIZED LEXICA

many of the anecdotal material (aḫbār) related to them. Other than dealing with these two aspects, he says that he intends to attribute to the relevant poets much of the poetry which his predecessor left unascribed; to add a large number of proverbs; and to explain a number of ġarīb words.233 Indeed, Bakrī does just that, except that he only sporadically adds to his commentary on his predecessor’s text a few proverbs or lines of poetry whose meanings are comparable to the proverbs in question.234 The book generally follows Abū ʿUbayd’s arrangement of his al-Amṯāl, although Bakrī at times slightly modifies the order of the chapters, particularly in the final section of the book.235 Its twenty themes are divided into 238 chapters or subcategories, compared with 270 in Abū ʿUbayd’s book. It is noteworthy that the book abounds in criticism of Abū ʿUbayd’s alleged mistakes,236 but its real value is that its author pre- served to us many of the philological and historical comments and elucida- tions made by earlier scholars in connection with K. al-Amṯāl.237

Nearly all of the proverbs in Abū ʿUbayd’s book form the basis which another Andalusian scholar, Sulaymān b. Mūsā l-Kalāʿī (d. 634/1237), used in authoring Nuktat al-amṯāl wa-nafṯat al-siḥr al-ḥalāl. This is a literary text of a predicatory nature, in which Kalāʿī’s own words are interwoven primarily with Abū ʿUbayd’s proverbs, with occasional additions from the Qurʾān, Ḥadīṯ and poetry. Its arrangement is based on Abū ʿUbayd’s work, but as Kalāʿī mentions in his introduction,238 he had at times to introduce slight changes to the order of Abū ʿUbayd’s proverbs. For example, in the previously mentioned chapter that deals with proverbs which imply the utmost degree of comparison (i.e. proverbs beginning with the pattern afʿal),239 the order in which several prov- erbs appear differs from their order in al-Amṯāl, as in the first three proverbs (nos. 1,210–1,212), the third of which is placed before the second in Kalāʿī’s text. Moreover, one of Abū ʿUbayd’s proverbs in this chapter, amsaḫu min laḥmi l-ḥuwāri (tasteless [lit. more tasteless than] as the flesh of a new-born camel), is transferred by Kalāʿī to a different chapter altogether, namely, the chapter that deals with types of cowardice (al-amṯāl fī ṣunūf al-ǧubn wa-anwāʿihi).240

233 Bakrī, Faṣl 3. 234 Ibid., 58–60, 86, 109, 121, 136, etc. 235 Cf. Abū ʿUbayd, Amṯāl 375–395 and Bakrī, Faṣl 507–518. In the latter, Bakrī merges ten of

Abū ʿUbayd’s themes into a single theme. 236 Bakrī, Faṣl 130, 238, 291, 336, 356, etc. 237 For a detailed study of Bakrī’s riwāya of Abū ʿ Ubayd’s book and of his sources, see Sellheim

(1954: 95–102). 238 Kalāʿī, Nukta 2. 239 Abū ʿUbayd, Amṯāl 360–375; Kalāʿī, Nukta 225–234. 240 Kalāʿī, Nukta 199.

 

 

110 chapter 2

At times, he also changes the word order of a certain proverb so that its last word would match a neighboring word. Thus, abṣaru min ġurābin (more sharp- sighted than a crow) is cited in the following context: wa-in kunta min ġurābin abṣara, where abṣara rhymes with taḥḏara (that you be cautious).241 These changes notwithstanding, Kalāʿī’s book is a unique attempt to incorporate into a literary text almost the whole content of an earlier book of the amṯāl genre.

The three remaining books in the above list may be viewed as a subcategory within their group due to the similarity of their content. All of them contain, in addition to proverbs, proverbial phrases which are frequently used in speech, but which require explanation, particularly as their users are often ignorant of their meanings. The earliest of these works is K. al-Amṯāl by Abū ʿIkrima al-Ḍabbī (d. 250/864), chronologically the first extant source after Abū ʿ Ubayd’s book. In his short introduction, Abū ʿIkrima says that he derived his material from the current speech of the Arabs (kalām al-ʿArab al-sāʾir), which he believes needs to be explained because it is frequently used (yuḥtāǧ ilā tafsīrihi li-kaṯrat istiʿmālihi).242 This implies that speakers often use stereotyped expressions without knowing their exact meanings. The first expression in the book is the commonly used salutation ḥayyāka l-Lāhu wa-bayyāka. Abū ʿIkrima considers various possibilities of interpreting the two verbs ḥayyāka (e.g. May God make you have dominion; or May God prolong your life) and bayyāka (e.g. May God bring you near [unto Himself]; or May God prepare for you an abode; or May God make you laugh).243 For this purpose, he cites several scholars and a dozen poetry šawāhid. The whole book is made up of a total of 111 proverbs and pro- verbial phrases, some of which are derived from Ḥadīṯ.244 Frequently occur- ring in current speech, these expressions – which are neither alphabetically nor thematically arranged – generally do not contain ġarīb usage, contrary to some of the material cited in above mentioned books, such as Muʾarriǧ’s and Abū ʿUbayd’s.

al-Fāḫir by al-Mufaḍḍal b. Salama (d. 290/903 or 300/913) is similar in con- tent to Abū ʿIkrima’s book, but Mufaḍḍal, more explicitly so than his prede- cessor, states that the ʿāmma (common people, generality of people) who use the expressions which make up his book do not actually know their mean- ings (wa-hum lā yadrūna maʿnā mā yatakallmūna bihi min ḏālika).245 That the author sets out to examine the speech of the ʿāmma is supported by the fact

241 Ibid., 225. 242 Abū ʿIkrima, Amṯāl 23. 243 Ibid., 24–30. 244 Ibid., 80, 103–104, 109. 245 Mufaḍḍal, Fāḫir 1.

 

 

111MUBAWWAB AND SPECIALIZED LEXICA

that in certain cases, the book is close to the genre of laḥn al-ʿāmma (sole- cism; see Section 6 below). He notes for example that they wrongly use the forms azdarayhi and ġumār in the two expressions ǧāʾa yaḍribu bi-aṣdarayhi (He came empty-handed) and daḫala fī ḫumāri l-nāsi (He entered among the crowd so as not to be seen) respectively.246 As for proverbs, they form less than half of the 521 entries of which the book consists.247 Like Muʾarriǧ’s book, some of these entries are made up of a single word, whereas others of frequently used short constructions. Examples include iḥtalaṭa (to become exceedingly angry), ḫaḍaʿa lahu (He succumbed to him), šāʿa l-ḫabaru (The news spread), and al-ṭāmma (calamity).248 But most of these expressions, unlike Muʾarriǧ’s book, are not of the ġarīb type, and the reason for their inclusion is apparently to show the etymology of the word in question. Thus, the original meaning of ḫaḍaʿa is said to be related to the stooping of one’s head when calamity strikes, and that of šāʿa to the continuous and strong ejection of a camel’s urine. For such philological comments, Mufaḍḍal is heavily dependent on earlier schol- ars, among whom are fellow Kufans, such as Abū ʿAmr al-Šaybānī, Farrāʾ, Abū ʿUbayd and Ibn al-Aʿrābī, as well as Basrans, such as Yūnus b. Ḥabīb, al-Naḍr b. Šumayl, Abū ʿUbayda, Abū Zayd and Aṣmaʿī.

The title of the third book in this subcategory, al-Zāhir fī maʿānī kalimāt al-nās, reveals its author’s interest in current usage. In its introduction, Ibn al-Anbārī (d. 328/940) says that people use in their prayers and in their invoca- tion and glorification of God kalām (expressions, phrases, etc.) whose mean- ing they do not know (wa-hum ġayr ʿālimīn bi-maʿnā mā yatakallamūna bihi), and that the same phenomenon also occurs in the proverbs and conversa- tions (muḥāwarāt) of the common people or ʿawāmm.249 The similarity with Mufaḍḍal’s book, as far as content and purpose are concerned, is unmistakable, except that al-Zāhir generally offers more detailed explanations and contains a larger number of entries (845, compared with 521 in al-Fāḫir), in which the number of proverbial expressions largely exceeds that of proverbs. But despite the similarity between the two books (both of which, moreover, lack any alpha- betical or thematic arrangement), Ṣūlī’s (d. 335/946) claim, reported in the introductory remarks preceding al-Fāḫir’s text,250 that Ibn al-Anbārī’s book is taken from al-Fāḫir is certainly an exaggeration, given Ibn al-Anbārī’s extensive

246 Ibid., 246. 247 Sellheim (1954: 119); cf. Āl Yāsīn (1980: 178). 248 Mufaḍḍal, Fāḫir 114, 117, 204, 318. 249 Ibn al-Anbārī, Zāhir I, 3. 250 Mufaḍḍal, Fāḫir 1.

 

 

112 chapter 2

šawāhid, and, more importantly, all his detailed philological comments.251 In fact, he highlights his philological interest from the outset by saying that he will not exclude from his book material related to naḥw, ġarīb and luġa.252 This interest is also reflected in the many entries which discuss the etymology of commonly used words, such as muʾmin (believer) ʿābid (worshiper), ḥakīm (wise), ʿāqil (sane), ẓarīf (witty), ẓālim (unjust), miskīn (poor), yatīm (orphan), fāǧir (dissolute), mulḥid (atheist),253 etc. Furthermore, he often points out – as do authors of the laḥn al-ʿāmma genre – mistakes made by the ʿāmma or ʿawāmm (common people) as in the forms ǧaddan, instead of ǧiddan (exceed- ingly), and šaḥḥāṯ, instead of šaḥḥāḏ (importunate beggar).254 In very general terms, comparison between the three books by Abū ʿIkrima, Mufaḍḍal and Ibn al-Anbārī reveals a gradual increase both in detailed textual and philological issues and in the number of proverbial expressions which are commonly used but widely misunderstood.

3.2 Alphabetically Arranged Books Almost two centuries after the earliest books of the genre, the first author to arrange proverbs alphabetically, albeit on a partial basis, is Ḥamza b. al-Ḥasan al-Iṣfahānī (d. after 351/962) in his al-Durra al-fāḫira fī l-amṯāl al-sāʾira. The formal (lafẓī) criteria in the arrangement of proverbs were thus adopted con- siderably later than in the genre ġarīb al-Ḥadīṯ, although the first author to do so in that genre, Abū Isḥāq al-Ḥarbī (d. 285/898), chose to include within the alphabetical arrangement a system of permutation of radicals.255 But the alphabetical arrangement of proverbs is roughly contemporaneous with that in the genre of ġarīb al-Qurʾān as Muḥammad b. ʿUzayr (or ʿUzayz) al-Siǧistānī (d. 330/941) is the first to introduce a partial alphabetical arrangement to his Ġarīb al-Qurʾān or Nuzhat al-qulūb.256 Since Ḥamza’s book includes only prov- erbs of the afʿal type, it will be discussed under the third category, but it should be noted here that Ḥamza clearly indicates in the introduction that the rea- son for adopting the alphabetical arrangement is to facilitate the use of his book (li-yashul tanāwul mā yurād minhu ʿalā multamisihi).257 The existence of several muǧannas lexica by the middle of the fourth/tenth century most

251 Cf. Sellheim (1954: 125). 252 Ibn al Anbārī, Zāhir I, 3. 253 Ibid., I, 105, 107, 109, 111–112, 116, 127, 129, 142–143 respectively. 254 Ibid., I, 23, 412. 255 See above, 76. 256 See above, 67. 257 Iṣfahānī, Durra I, 56.

 

 

113MUBAWWAB AND SPECIALIZED LEXICA

probably contributed to the realization of the advantage which an arrange- ment based on a certain formal criterion offers with respect to ease of use. Following Ḥamza’s example, most subsequent authors adopted an alphabeti- cal system of arrangement, be it on partial or full basis.

Other than Ḥamza’s, the most important books of the genre which are alphabetically arranged are the following:

1. Ǧamharat al-amṯāl by Abū Hilāl al-ʿAskarī (d. after 395/1005). 2. al-Amṯāl by Zayd b. ʿAbdallāh b. Masʿūd b. Rifāʿa (d. after 400/1010). 3. al-Wasīṭ fī l-amṯāl attributed to Abū l-Ḥasan al-Wāḥidī (d. 468/1076). 4. Maǧmaʿ al-amṯāl by Maydānī (d. 518/1124). 5. al-Mustaqṣā fī amṯāl al-ʿArab by Zamaḫšarī (d. 538/1144). 6. Farāʾid al-ḫarāʾid fī l-amṯāl by Abū Yaʿqūb al-Ḫuwayyī (d. c. 549/1154). 7. Nuzhat al-anfus wa-rawḍat al-maǧlis by Abū Saʿīd al-ʿIrāqī (d. 561/1166). 8. Ġurar al-amṯāl wa-durar al-aqwāl by Bayhaqī (d. 565/1170). 9. Timṯāl al-amṯāl by Abū l-Maḥāsin al-Šaybī (d. 837/1433).

Besides alphabetical arrangement, most of these sources have in common their authors’ pursuit of exhaustiveness. This fact is highlighted in the very titles of Ǧamharat al-amṯāl (The Corpus of Proverbs) and Maǧmaʿ al-amṯāl (The Collection of Proverbs). Even in Zamaḫšarī’s title, al-Mustaqṣā fī amṯāl al-ʿArab (The Investigation of the Proverbs of the Arabs), the notion of istiqṣāʾ clearly indicates the author’s uttermost effort to exhaustively gather Arabic proverbs in his book.258 ʿAskarī is not only the first author to point out this quality in the title of his book but also the first among authors whose books are not restricted to proverbs of the pattern afʿal to arrange his material alphabetically. There is strong reason to believe that he borrowed this system from Ḥamza al-Iṣfahānī’s al-Durra al-fāḫira, which is the only book he mentions in his introduction. By comparing the two books, it becomes quite clear that ʿAskarī devised his book after the pattern of Ḥamza’s. In fact, he himself acknowledges his reliance on Ḥamza for the proverbs of the pattern afʿal (which constitute more than a third of the whole book), although he says that he did not include Ḥamza’s muwallad proverbs,259 which are not purely Arabic or belong to the post-classical era. Not only does ʿAskarī emulate Ḥamza in giving the list of the afʿal proverbs which each chapter contains before individually explaining them, but he also gener- alizes this to the other proverbs. Thus, each of his twenty-nine chapters (except

258 Cf. below, 121, n. 308. 259 ʿAskarī, Ǧamhara I, 6.

 

 

114 chapter 2

for Chapter twenty-eight)260 begins with two lists, the first of which includes all the proverbs in that chapter other than those of the afʿal pattern, and the second all the afʿal proverbs to be cited, largely based on Ḥamza’s list. The alphabetical arrangement of al-Ǧamhara is thus a direct result of its author’s adoption of Ḥamza’s lists and of his introduction of similar lists which include non-afʿal proverbs. ʿAskarī in his arrangement, like his predecessor, takes into account only the first letter of the alphabet. In the case of the afʿal proverbs, of course, that first letter is not the prefixed hamza common to all such prov- erbs, but the next letter, which is also the first radical of the root. As far as content is concerned, al-Ǧamhara consists of 1,972 entries, the vast majority of which are proverbs, with only very few proverbial phrases.261 The number of proverbs, it should be noted, is larger than the number of entries, since some entries include several proverbs (in one instance, more than eighty).262 The book is thus substantially different from Abū ʿIkrima’s al-Amṯāl, Mufaḍḍal’s al-Fāḫir and Ibn al-Anbārī’s al-Zāhir, in all of which proverbial phrases feature strongly. Most of the proverbs, moreover, are briefly explained, although in some cases ʿAskarī includes lengthy citations of poetry,263 detailed philologi- cal comments,264 and whole stories related to some proverbs.265

The relatively little-known book of Ibn Rifāʿa (d. after 400/1010),266 al-Amṯāl, is similar to ʿAskarī’s in that its alphabetical arrangement does not go beyond the first letter of each proverb. Yet, its arrangement is rather peculiar because in several letters the author groups together a number of proverbs based on the pattern, word, particle, etc. with which they begin. This phenomenon features most prominently in the letter alif, which begins with an alphabetical arrange- ment of afʿal proverbs based, in addition to the alif, on the next letter, which

260 This is the chapter of lām alif, which naturally includes no proverbs of the afʿal pattern since any such proverb is listed in the chapter of lām.

261 E.g. the phrases bayḍat al-balad (an abandoned ostrich egg; said of one who is alone or has no aid) and al-hiyāṭ wa-l-miyāṭ (raising a clamor, engaging in conflict); ibid., I, 231; II, 361.

262 See item 882 (ibid., I, 493–495). Item 1318 (II, 91–92) contains about 25 proverbs. ʿAskarī himself mentions this phenomenon in his introduction (I, 6).

263 E.g. ibid., I, 57, 127, 136; II, 62, 185, 244. 264 As, for example, the tens of expressions beginning with ibn or bint, which he cites as part

of his explanation of proverb 21 (I, 35–48). 265 E.g. ibid., I, 141–142, 145–147, 207–211, 424–426; II 93–97, 194–196, 241–242, 413–418. 266 Sellheim (1954: 143) questions the attribution of the book to Ibn Rifāʿa, as a quotation

ascribed to him by Maydānī (Maǧmaʿ I, 164) does not appear in it. This argument however is obviously not a conclusive proof of faulty attribution. Regardless of that, Ṣafadī (Wafayāt XV, 49) does ascribe to Ibn Rifāʿa a book entitled K. al-Amṯāl.

 

 

115MUBAWWAB AND SPECIALIZED LEXICA

represents the first radical of the word’s root. Following that is a number of sections within the bāb of alif,267 each of which includes a group of proverbs which start with (1) an imperative verb, (2) the interrogative particles hamza or ayy, (3) inna or innamā, (4) an, (5) in, (6) a past tense verb, (7) iḏā (8) anta, (9) anā, (10) ist (anus), (11) iyyāka, (12) innahu, and (13) the definite article al-. Obviously, these sections cannot include all proverbs beginning with alif, and the author accordingly had to add other sections to allow for the inclusion of the rest of these proverbs. These latter sections appear intermittently in a man- ner which disrupts the sequence in which the above mentioned groups occur.268 To add to the confusion, proverbs often appear in sections to which they do not belong, as in the case of a proverb which begins with inna but is listed with proverbs which begin with al-.269 Furthermore, some sections are confus- ingly divided into several headings, each called faṣl (chapter) with no apparent reason. Proverbs beginning with al-, for example, appear under three different headings.270 Ibn Rifāʿa more sparingly uses his technique of grouping proverbs together in other letters. Examples include, under the letter r, proverbs begin- ning with rubba,271 under l those beginning with laysa, law and laqītuhu/ laqītu,272 and under h those beginning with hāḏā/hāḏihi, huwa and hal.273 Although such groups can facilitate the search for a particular proverb, this advantage is limited to a few chapters and certainly does not make up for the absence of a system of arrangement that goes beyond the first letter. The con- tent of the book is made up of 1,458 proverbs which are very briefly explained (at times even cited without any explanation) and include, according to the editor,274 twenty proverbs which do not appear in any other source.

As in the two previous books, the author of al-Wasīṭ fī l-amṯāl takes into consideration in its alphabetical arrangement only the first letter. The book’s editor erroneously ascribes it to Abū l-Ḥasan al-Wāḥidī (d. 468/1076), but Sellheim rightly asserts that it is mainly an abridgement by Abū Saʿīd al-ʿIrāqī (d. 561/1166) of his own Nuzhat al-anfus wa-rawḍat al-maǧlis (for which see below).275 In his introduction, the author of al-Wasīṭ indicates that his book

267 Ibn Rifāʿa, Amṯāl 4–97. 268 E.g. ibid., 68, 74–76. 269 Ibid., 86 (no. 410). 270 Ibid., 77, 88, 97. 271 Ibid., 133–137. 272 Ibid., 203–206, 206–208, 208–212. 273 Ibid., 265–268, 268–269, 270. 274 Ibid., Introduction ḥ. 275 Sellheim (1988a: 353–359; 1991: 824). Note that Nuzhat al-anfus is mentioned by the author

of al-Wasīṭ 42, 46, 64–65, 79.

 

 

116 chapter 2

consists of twenty-seven chapters;276 yet it includes at least twenty-eight chapters (the last of which is that of lām alif ), and possibly a twenty-ninth for the letter yāʾ, which we do not possess due to the fact that the manuscript is incomplete.277 The total number of proverbs is 186, many of which are ascribed to their first users. The author tells us that his book, as far as size is concerned, stands in the middle between his two other books on proverbs, al-Basīṭ and al-Waǧīz (wāsiṭa baynahumā).278 The book’s title, al-Wasīṭ (i.e. middle-sized), seems to be deliberately chosen to reflect this fact, and the use of the term nubḏa (extract, excerpt)279 further reveals the selective nature of the work. But despite its brevity, the book abounds with anecdotal material and poetry šawāhid which the author frequently includes to explain the circumstances which led to the use of his proverbs.280

Maydānī’s (d. 518/1124) Maǧmaʿ al-amṯāl is at once the most famous book of the whole genre and the one that contains the largest number of proverbs (with the exception of books that list proverbs in poetry; see third category below). It consists of 4,765 proverbs including those of the afʿal type, in addi- tion to about 900 muwallad proverbs. A typical chapter begins with non-afʿal proverbs, followed by afʿal ones and then the muwallad. Since there is no chap- ter for lām alif, there are 28 chapters of this type, which are followed by two more chapters, the first of which is not related to proverbs but lists 132 names of the Battle-days of the Arabs (Ayyām al-ʿArab) in pre-Islamic times and 93 in Islam.281 The other chapter cites 228 expressions attributed to the Prophet, his Companions and other famous figures.282 These expressions are obviously chosen because they resemble proverbs in form and content. The total num- ber of proverbs is thus close to 6,000, which is consistent with Maydānī’s own estimate.283 But, exhaustive as the book may be, Maydānī admits that it is only human not to be able to fully exhaust the material at hand.284 For sources, he mainly relies on Ḥamza al-Iṣfahānī’s (d. after 351/962) al-Durra al-fāḫira in the

276 Wāḥidī, Wasīṭ 31. 277 Ibid., editor’s introduction 23. 278 Ibid., 31. That al-Waǧīz is the shortest of the three books is clear not only because of its

name, but also because its author refers the reader four times in his book (ibid., 41, 98, 170, 199) to al-Basīṭ for further elaboration.

279 See the title of the first chapter (Nubḏa min amṯāl al-ʿArab); ibid., 3. 280 E.g. ibid., 37, 40–41, 47, 50–52, 70–71, etc. 281 Maydānī, Maǧmaʿ II, 430–444, 444–448. For the general characteristics of the book, see

Ḫalīl (1983: 161–178). 282 Ibid., II, 448–461. 283 Maydānī’s expression (ibid., I, 5) is wa-huwa sittat ālāf wa-nayyif (It is in excess of 6000). 284 Ibid., loc. cit.

 

 

117MUBAWWAB AND SPECIALIZED LEXICA

sections of afʿal proverbs, and acknowledges this reliance by saying that he incorporated the content of Ḥamza’s book into his own (wa-naqaltu mā fī kitāb Ḥamza b. al-Ḥasan ilā hāḏā l-kitāb).285 Yet, he left out some three hundred of Ḥamza’s proverbs,286 added only a few from other sources, and for no obvious reason (such as stricter alphabetical ordering) regularly changed the order of the ones he copied from Ḥamza. Other than al-Durra, Maydānī mentions hav- ing painstakingly scrutinized more than fifty books of proverbs,287 many of whose authors are indeed acknowledged in the text. The abundance of these sources resulted not only in a huge corpus of proverbs but also in the inclusion of some of the most extensive accounts of historical events recorded in any book of its genre, amounting in the case of Dāḥis and Ġabrāʾ (two horses said to have ignited a forty-year war in Ǧāhiliyya) to about eleven pages.288

The fact that Maydānī’s book is very rich in the number of its entries but its alphabetical arrangement does not go beyond the first letter makes it frustrat- ingly difficult to use. In other words, one has to use an index to be able to quickly locate a certain proverb in this lexicon because its alphabetical arrangement is only partial. The first to overcome this drawback is the famous lexicogra- pher, grammarian and exegete Zamaḫšarī (d. 538/1144) in his al-Mustaqṣā fī amṯāl al-ʿArab, which he completed in 499/1106289 at the age of thirty-one. In his introduction, Zamaḫšarī explains that in arranging his proverbs he went beyond the first letter and took into consideration the awāsiṭ and awāḫir of the words (lit. their middle and final letters), that is, all the successive letters.290 He so meticulously observed this alphabetical arrangement as to consider the letters of the second word etc. if two or more proverbs start with one or more common words. Accordingly, afʿal proverbs no longer had to be listed in a separate section, but appeared as the alphabetical order dictated (incidentally, contributing to the swelling of the chapter of alif, the number of whose entries exceeds the number of entries in all other letters put together). Zamaḫšarī’s strictness in alphabetical ordering is no reason for surprise, given that he is also author of the muǧannas lexicon Asās al-balāġa, which adopts full alphabetical arrangement.291 Compared with Maǧmaʿ al-amṯāl, Zamaḫšarī’s book contains no more than 3,461 proverbs and is much briefer in explaining the proverbs

285 Ibid., I, 4. 286 Cf. Sellheim (1954: 150). 287 Maydānī, Maǧmaʿ I, 4. 288 Ibid., II, 110–121. 289 Ḥāǧǧī Ḫalīfa, Kašf II, 1675. 290 Zamaḫšarī, Mustaqṣā, Introduction ǧ, d. 291 See below, 357–358.

 

 

118 chapter 2

and commenting on their origin. The book’s brevity – which Zamaḫšarī points out in his introduction – might be one reason why subsequent authors, in spite of its full alphabetical arrangement and hence ease of use, preferred to it Maydānī’s opus.

Also in the sixth/twelfth century, Abū Yaʿqūb al-Ḫuwayyī (d. c. 549/1154) authored Farāʾid al-ḫarāʾid fī l-amṯāl. A student of Maydānī’s, he was highly dependent on the content of Maǧmaʿ al-amṯāl for the 1,720 proverbs292 which he included in his own book and which he, like his teacher, arranged alpha- betically based on the first letter only. This notwithstanding, al-Farāʾid is not merely an abridgement of Maydānī’s book, as the internal arrangement of its chapters readily reveals. A typical chapter such as that of alif, bāʾ or tāʾ, starts with the non-afʿal proverbs which begin with that letter, followed by apho- risms (ḥikam), widely used lines of poetry (abyāt sāʾira), afʿal proverbs and muwallad proverbs. Since the second and third sections (i.e. aphorisms and lines of poetry) are not part of Maǧmaʿ al-amṯāl, Ḫuwayyī’s reliance on his teacher is restricted to the other sections. With few exceptions,293 the prov- erbs in these other sections are taken entirely from Maydānī, although they are uniformly listed (perhaps deliberately) in a different order.294 Furthermore, Ḫuwayyī often differs with his teacher in the wording and explanation of prov- erbs and in the supporting šawāhid,295 and adds a final chapter for words of wisdom and preachment ascribed to the early pious Muslims known as al-salaf al-ṣāliḥ. He also devotes a separate chapter to lām alif, unlike Maydānī who includes it in the lām chapter, although the latter groups together proverbs beginning with the negative particle lā within that chapter. It is interesting to note that Ḫuwayyī, perhaps influenced by this group of proverbs which Maydānī lists in a separate section of the chapter of lām, places the lām alif chapter right after the lām chapter and not in its habitual place immediately preceding yāʾ. In short, it seems that Ḫuwayyī, in spite of relying heavily on Maydānī’s proverbs, is keen to demonstrate that his book has its own character. Indeed, in his introduction he sharply criticizes Maydānī for including in his book proverbs that are not in use (uhmila) as well as anecdotes which relate

292 This number includes only afʿal and non-afʿal proverbs to the exclusion of aphorisms, lines of poetry and muwallad proverbs.

293 Among Ḫuwayyī’s proverbs which are not in Maydānī’s book are araqqu min dīn al-Qarāmiṭa (weaker than the faith of the Karmatians) and aṭwalu min layl al-ḍarīr (longer than the night of the blind); see Ḫuwayyī, Farāʾid 259–342.

294 For example, the first proverb under the letter tāʾ in al-Farāʾid (p. 109) corresponds to the sixty-eighth proverb under tāʾ in Maǧmaʿ al-amṯāl (I, 136, no. 677).

295 See the editor’s introduction to al-Farāʾid 6–7.

 

 

119MUBAWWAB AND SPECIALIZED LEXICA

the false deeds (abāṭīl) of those whom he describes as uncivilized and foolish (al-aǧlāf al-maǧāhīl).296 The latter morally charged comment is quite surpris- ing since Ḫuwayyī himself, immediately preceding that comment, asserts the role of naql (transmission, tradition) in the study of proverbs – a role which Arab philologists hardly subject to moral criteria, be it in relation to proverbs or other linguistic forms of expression.

ʿIrāqī’s Nuzhat al-anfus wa-rawḍat al-maǧlis has already been mentioned with al-Wasīṭ fī l-amṯāl. It consists of some 900 entries, in full alphabetical arrangement. The author is obviously interested in ascribing the proverbs to their first users wherever possible, and often includes lengthy stories that are associated with certain proverbs. In his introduction, he mentions prov- erbs as one of several components of his book, which, he says, contains anec- dotes, goodly narratives, literary material, wisdom sayings, etc.297 Other than proverbs, the entries include numerous commonly used expressions which ʿIrāqī explains extensively at times. For example, out of twenty-two entries under the letter tāʾ, there are thirteen such expressions, among which are taǧabbara l-raǧulu (He became arrogant), tašāǧarnā fī kaḏā (We quarreled over so and so), tawassamtu l-ḫayra fī waǧhihi (I saw promising signs in his face).298

Abū l-Ḥasan al-Bayhaqī (d. 565/1170), a student of Maydānī’s, authored Ġurar al-amṯāl wa-durar al-aqwāl. This work, as yet unpublished, comprises 2,900 proverbs,299 including muwallad ones, and is alphabetically arranged according to the first letter.

The last book in the above list, Abū l-Maḥāsin al-Šaybī’s (d. 837/1433) Timṯāl al-amṯāl, was authored about three centuries later than its two main sources, Maǧmaʿ al-amṯāl and al-Mustaqṣā. After citing a proverb, Šaybī almost regu- larly mentions whether it is found in either, both or none of these two books. The other sources from which he derives some proverbs or complements the material of al-Maǧmaʿ or al-Mustaqṣā include Mubarrad’s (d. 285/898) al-Kāmil, Iṣfahānī’s (d. 356/967) al-Aġānī, Ǧawharī’s (d. c. 400/1010) al-Ṣaḥāḥ, and Ḥarīrī’s (d. 516/1122) Durrat al-ġawwāṣ.300 The total number of proverbs in the book is 441, and it is obvious that Šaybī chose only those proverbs which admit more than one riwāya (hence his discussion of the difference among authors in citing them) and those which lend themselves to lengthy historical or philological comments (hence his reliance on both literary and linguistic

296 Ibid., 19. 297 ʿIrāqī, Nuzha 3. See also Sellheim (1988: 82–94). 298 Ibid., 140, 146, 149. 299 The chapter of hamza includes 436 proverbs, edited by El-Saghir (1984: 1–183). 300 Šaybī, Timṯāl II, 524; I, 284; I, 185; II, 578 respectively.

 

 

120 chapter 2

sources). As far as alphabetical order is concerned, Šaybī followed Zamaḫšarī’s strict system which takes all the letters of a word into consideration and which even goes to the subsequent words in order to ensure full alphabetical order- ing. Unfortunately, his reliance on earlier authors has precluded the inclusion of any new material that could have enhanced our knowledge of proverbs, par- ticularly muwallad ones that might have arisen since the time of Maydānī and Zamaḫšarī (i.e. sixth/twelfth century).

3.3 Books Specialized in a Particular Aspect of Proverbs In the previous two categories, we surveyed books of proverbs based on their arrangement (mostly alphabetical, but also thematic as in Abū ʿUbayd’s K. al-Amṯāl) or lack thereof. But, irrespective of arrangement, a number of works were written in which proverbs of a specific type or genre are grouped together. These works vary considerably, but can be divided, based on content, into four types: those that deal with (1) afʿal proverbs, (2) proverbs in poetry, (3) Qurʾān and Ḥadīṯ, or, more generally, wise sayings from various genres, and (4) muwallad (post-classical) and ʿāmmī (vernacular) proverbs. The inclusion by several paroemiographers of the second and third types under the genre of proverbs obviously broadens the notion of proverb to comprise expressions from outside the sphere of prose or kalām al-ʿArab (as opposed in particular to poetry and Qurʾān). It also blurs the distinction between proverbs, as say- ings which embody collective wisdom often expressing the experience of “common” people, and the sayings of prophets, sages, poets, etc. as preserved in the wider wisdom and religious literature. It is certain that the absence in philological sources of a clear definition of proverbs has contributed to their confusion with other expressions that share some of their characteristics in form and content.

To the first of the four types mentioned above belongs one of the most impressive works of the whole genre, Ḥamza al-Iṣfahānī’s (d, after 351/962) al-Durra al-fāḫira fī l-amṯāl al-sāʾira.301 The abundance of proverbs which begin with comparative adjectives of the pattern afʿal and their uniformity of form seems to have attracted the attention of early scholars, some of whom collected them in monographs. Ḥamza in fact mentions in his introduction several of his predecessors who collected these proverbs, including Aṣmaʿī (d. 216/831) and Liḥyānī (fl. in the second half of the second/eighth century),

301 To be noted is that, based on the various ways in which this book was referred to in the sources and manuscripts, the editor pieced together the book’s title cited above (see his introduction 33–34). The book was also published under another title, Sawāʾir al-amṯāl ʿalā afʿal (ed. by Fahmī Saʿd, ʿĀlam al-Kutub, Beirut 1988).

 

 

121MUBAWWAB AND SPECIALIZED LEXICA

whose monographs are in no more than ten pages each.302 He also mentions the chapter which Abū ʿUbayd devoted for such proverbs in his K. al-Amṯāl.303 Muḥammad b. Ḥabīb (d. 245/860), according to Ḥamza, took over the proverbs cited by his three predecessors and added to them many others, bringing the total number of items to no more than 390.304 Obviously, Ḥamza was keen on exhausting afʿal proverbs to the fullest possible extent. In his introduction he proudly announces that his book contains more than 1,200 proverbs305 in addition to more than 500 muwallad ones, with a grand total of over 1,800.306 Moreover, at the end of his twenty-ninth chapter, which concludes the muwallad proverbs, he repeats that the total number of proverbs in the book exceeds 1,800.307 He also uses the terms istiqṣāʾ and istīfāʾ,308 both of which indicate his pursuit of exhaustiveness. He includes proverbs which he ascribes to earlier authors but admits that he either is ignorant of their meaning or their circumstances.309 This further proves his exhaustive attempt, even at the risk of merely citing a proverb with little additional information.

As far as arrangement is concerned, it has already been pointed out that Ḥamza’s book is the first in the amṯāl genre to be arranged alphabetically, albeit on the basis of the first letter only.310 The book furthermore is quite well organized. Its introduction includes useful syntactic and semantic observa- tions related to the pattern afʿal as well as an analysis of certain aspects of afʿal proverbs in general, such as their inclusion of a good deal of material related to animals, and at a later stage to certain individuals,311 as well as the connected- ness between some proverbs and specific tribes or cities. Following the intro- duction, each of the twenty-eight chapters begins with a list of the proverbs it contains and their exact number. Chapter twenty-nine includes seven groups

302 Iṣfahānī, Durra I, 55. 303 See above, 107. 304 For seven proverbs of the afʿal type attributed to Muḥammad b. Ḥabīb in K. al-Amṯāl, see

Ḥamīdullāh (1956: 44–45). 305 Note that the editor puts this number at 1,293 (Durra’s introduction, 40, n. 2). 306 Iṣfahānī, Durra I, 56. 307 Ibid., II, 469. 308 Ibid., II, 438. See also our previous comment (p. 113) on the title of Zamaḫšarī’s

al-Mustaqṣā. 309 Ibid., I, 193, 231–232, 253. 310 See above, 112. 311 Note Ḥamza’s use of the conjunction ṯumma in the expression ṯumma ḍarabū baʿḍ hāḏihi

l-amṯāl bi-l-riǧāl (and then they used some of these proverbs in relation with individuals; ibid., 64); that is, the use of afʿal proverbs for people came at a stage later than its use for animals.

 

 

122 chapter 2

of muwallad proverbs, in some of which afʿal occurs as the first word, but else- where in others.312 Most of these proverbs, of which several occur in poetry, are not explained, obviously because their meaning is clear. The last chapter of the book deals with what the author describes as rare expressions of a proverbial nature (nawādir min al-kalām ǧāriya maǧrā l-amṯāl).313 These include more than five hundred epithets (kunā) which begin with abū, umm, ibn, bint, ibnā, banū and banāt (such as Abū Zayd for old age and bint al-šafa, lit. daughter of the lip, for word). The proverbial nature of such metaphors is obviously doubt- ful, and this is even more the case in the lengthy list of words in which the dual refers to two objects of different nature (such as al-qamarān, lit. the two moons, for the moon and the sun), as well as in the thirty fables (ḫurāfāt) and the collection of expressions used by the Arabs in enchantment (ruqya) with which the book is concluded.

The material collected by Ḥamza was a major source for later authors of books on proverbs, and we have already discussed his influence on ʿAskarī’s Ǧamharat al-amṯāl.314 Among the authors of the major subsequent works, Maydānī and Zamaḫšarī in particular are indebted to him. Maydānī con- firms that he incorporated Ḥamza’s proverbs into his own book,315 whereas Zamaḫšarī’s proverbs of the afʿal type largely agree with those of Ḥamza’s and may well have been taken directly from him or indirectly through ʿAskarī’s book, although Zamaḫšarī does not specify his sources.316 This notwithstand- ing, Ḥamza’s method of restricting the content of his book to the afʿal type was largely discontinued. Among the extant sources, only Afʿal min kaḏā by a con- temporary of Ḥamza’s, Abū ʿAlī al-Qālī (d. 356/967), is of similar content. The fact that this monograph includes a mere 358 proverbs makes it highly unlikely that Abū ʿAlī was cognizant of Ḥamza’s much larger work when he authored his book. Of the earlier works devoted to afʿal proverbs, Qālī seems to have made use of Muḥammad b. Ḥabīb’s and Aṣmaʿī’s.317 As for arrangement, the book is neither alphabetically nor thematically arranged. Instead, Qālī divides his material into ninety-five chapters, each of which (except for the last) contains between one and thirty-four proverbs which begin with a word of the com- parative pattern afʿal. The last chapter lists thirty-four proverbs which begin with different words. In several cases, words of opposite meanings are placed

312 Ibid., II, 443–469. 313 Ibid., II, 471. 314 See above, 113–114. 315 Maydānī, Maǧmaʿ I, 4; cf. above, 117. 316 Cf. Iṣfahānī, Durra, editor’s introduction 42. 317 Qālī, Afʿal 5, 6, 7, 9, 14, 22, 41, 47, 69 (Ibn Ḥabīb); 12, 22 (Aṣmaʿī).

 

 

123MUBAWWAB AND SPECIALIZED LEXICA

in consecutive chapters, as in the case of aḥsan and aqbaḥ (better/worse), aṯqal and aḫaff (heavier/lighter), amarr and aʿḏab (more bitter/sweeter), and aʿṭaš/arwā (thirstier/more sated with liquid).318 All proverbs are either very briefly explained or not explained at all, and in some cases, muwallad proverbs – typically introduced by the expression yaqūl al-ḥaḍar/ahl al-ḥaḍar (Urban people say)319 – are quoted towards the end of the relevant chapter.

Although lines of poetry are occasionally cited by some authors (e.g. Ḥamza in al-Durra al-fāḫira, as noted above), most of the sources of the genre exclude poetry as a source of proverbs and proverbial expressions. Poetry is normally cited either as evidence for a certain usage or as part of the histori- cal circumstances with which certain proverbs are connected. To some early authors, however, are ascribed books devoted to lines of poetry which con- tain proverbs. The first of these is al-Abyāt al-sāʾira by Abū l-Minhāl ʿUyayna b. al-Minhāl (also referred to as ʿUyayna b. ʿAbdalraḥmān and Abū ʿUyayna b. al-Minhāl; d. ?),320 who was a student of Ḫalīl’s (d. 175/791), according to Ibn al-Nadīm.321 Two other books by the same title are also ascribed by Ibn al-Nadīm to Abū l-ʿAmayṯal (d. 240/854) and Sukkarī (d. 275/888).322 We hardly know anything about the content of these books and cannot be sure whether Ḥamza al-Iṣfahānī, author of one of the major works on proverbs which occur in poetry, K. al-Amṯāl al-ṣādira ʿan buyūt al-šiʿr, knew of them. But it is note- worthy that Ḥamza, who does acknowledge his sources in al-Durra al-fāḫira, is silent about these (and other) books in his al-Amṯāl al-ṣādira.

Regardless of its sources, al-Amṯāl al-ṣādira is the first extant book in which proverbs are included on the strength that they occur in poetry. In his intro- duction, Ḥamza clarifies that his book includes not only lines of poetry which contain old proverbs (amṯāl qadīma), but also lines whose words were later used as proverbs.323 It is most probably the latter type that Ḥamza had in mind when he attributed to poetry the abundance of proverbs used by the Arabs in comparison with other nations. In this regard, he says that Persian prov- erbs hardly come up to a tenth (lam taʿšur) of those of the Arabs because, as a nation, the Arabs are unique in their large stock of proverbs due to the contin- uous contribution of poetry to this stock.324 Compared with al-Durra al-fāḫira,

318 Ibid., 3–5, 36–37, 66–67, 68–69. 319 Ibid., 5, 25, 27, 36, etc. 320 Qifṭī, Inbāh II, 384; IV, 173. 321 Ibn al-Nadīm, Fihrist 120. 322 Ibid., 55, 86. 323 Iṣfahānī, al-Amṯāl al-ṣādira 56. 324 Ibid., 65.

 

 

124 chapter 2

which includes about 1,800 proverbs, the number of hemistichs and lines in al-Amṯāl al-ṣādira is more than three times as many – a staggering 5,519 items,325 none of which is ascribed. Even if we were not to individually count each line in stanzas that have two or more lines, as the editor did, the number of proverbs (i.e. not hemistichs and lines) would still be significantly more than double that in al-Durra. Ḥamza’s organizational skills are even more evident in al-Amṯāl al-ṣādira than in his other book. Hence, the majority of the items (5,050 in all) are arranged in the first five sections according to length. The first section includes single hemistichs, the second single lines, the third double hemistichs,326 the fourth double lines, and the fifth lines that are translated from Persian, also arranged according to length. In the beginning of each sec- tion, Ḥamza specifies the number of proverbs it includes, as he does in al-Durra. The internal arrangement of the first four sections rests on the initial word. For example, the first section is divided into fifteen chapters based on whether the line begins with (1) qad, hal, law and rubba; (2) yā, ʿan, min, man, mā and lā; (3) alā, ammā and lammā; (4) lam, lan and laysa; (5) lākinna, lākinnamā and lākin; (6) innamā, inna and in; (7) laʿalla, ʿasā and biʾsa; (8) ʿalā, ilā and fī; (9) ʿinda, bayna and iḏā; (10) ayy, kayfa, kam and matā; (11) kafā, kull, baʿḍ and ḏū; (12) huwa, hiya, ḏāka and tilka; (13) different (i.e. unspecified) words; (14) a particular letter, from alif to yāʾ; and (15) an affix, such as kāf, lām and bāʾ. In this subjective and original arrangement – which is largely followed in the next three chapters as well – Ḥamza mostly groups together particles which share their etymology (no. 5), government of other elements of the structure (ʿamal; no. 8), grammatical “category” (adverbs and interrogative particles in 9 and 10 respectively), etc. In the fourteenth chapter (and its counterparts in sections 2, 3, and 4), the lines are arranged based on the first letter only. But curiously enough, lines beginning with alif in that chapter of sections 1 and 2 seem to follow full alphabetical arrangement.327 It is highly unlikely that the order in which the twenty hemistichs in the first section and the twenty-three in the second appear is the result of pure chance. Yet it remains unclear why Ḥamza gave up that full alphabetical ordering in subsequent letters. The last two sec- tions of the book, similar to the latter parts of al-Durra, include miscellaneous

325 The last proverb in the book carries this number (ibid., 768), but the editor counts 1,674 hemistichs and 3,790 full lines, with a total of 5,464 items (ibid., 48).

326 Most of the double hemistichs are of the raǧaz meter, in which each hemistich constitutes a single line. In other than raǧaz, however, Ḥamza considers, contrary to tradition, two rhyming hemistichs as two lines, rather than a single line, hence their inclusion in the third section.

327 See items 912–931 (ibid., 210–213) and 2554–2576 (ibid., 439–442).

 

 

125MUBAWWAB AND SPECIALIZED LEXICA

topics, some of which differ from the rest of the book. Among these are prov- erbs which occur in prose, wise sayings of Persian sages and kings,328 and vari- ous anecdotal material in both poetry and prose.329

The other major work on proverbial material in poetry is Ibn Aydamur’s (d. 710/1310) al-Durr al-farīd. Before discussing it however we shall examine three relatively short works, the first of which goes back to the fourth/tenth century. This is al-Amṯāl al-sāʾira min šiʿr al-Mutanabbī by Abū l-Qāsim Ismāʿīl b. ʿAbbād al-Ṭālaqānī, better known as al-Ṣāḥib b. ʿAbbād (d. 385/995). As its title indicates, the book is a collection of Abū l-Ṭayyib al-Mutanabbī’s poetry (a total of about 370 lines) which contains expressions that came to be used as proverbs. The material – which includes single lines as well as several con- secutive lines from the same poem – is neither alphabetically nor themati- cally arranged. In order to justify restricting his book to one poet, al-Ṣāḥib says that, in addition to Abū l-Ṭayyib’s unmatched skills as a poet, he is particu- larly ahead of his peers in his use of expressions that soon gain currency and assume the characteristics of proverbs (lahu fī l-amṯāl ḫuṣūṣan maḏhab sabaqa bihi amṯālahu).330 He even mentions that he often heard his Sultan, Faḫr al- Dawla b. Buwayh (d. 387/997), citing as proverbs (wa-samiʿtuhu . . . yatamaṯṯal kaṯīran) some of the choicest of Abū l-Ṭayyib’s poetry. Yet, in a defensive man- ner, al-Ṣāḥib expresses his readiness, subject to the wishes of the Sultan, to dictate a book on proverbs which occur in pre-Islamic, muḫaḍram (strad- dling Ǧāhiliyya and Islam) and Islamic poetry since no scholar convincingly and exhaustively collected this material in a book (fa-mā aǧid man ʿamila fī ḏālika min al-udabāʾ kitāban muqniʿan aw ǧamʿan mušbiʿan).331 This comment strongly suggests that al-Ṣāḥib did not know of Ḥamza’s arduous effort in pro- ducing al-Amṯāl al-ṣādira.

The second book is Muḥammad b. Abī Bakr al-Rāzī’s (d. after 666/1268) K. al-Amṯāl wa-l-ḥikam, a collection of 605 lines of poetry and 293 hemistichs, most of which are without ascription. In his introduction, Rāzī explains that his book is a short collection (muḫtaṣar) of lines and hemistichs that are so much used in writing and conversation that they have become proverbs (ṣārat amṯālan sāʾira).332 The book is divided into two sections, the first of which is for single lines and the second for hemistichs. The first section includes ten chapters, and the second eight, and each of these has a specific theme.

328 Ibid., 739–752. 329 Ibid., 755–809. 330 Ṣāḥib, Amṯāl 22. 331 Ibid., 23. 332 Rāzī, Amṯāl 11.

 

 

126 chapter 2

In general, the themes are of two types, the first of which has a strong moral component, such as dependence on God, asceticism (zuhd), solace or consola- tion, and good manners; whereas the second includes notions such as court- ship, praise, thankfulness, admonition, complaint, satire, reprimand, threat, apology, etc.

The third book is Šams al-Dīn al-Nawāǧī’s (d. 859/1455) Zahr al-rabīʿ fī l-maṯal al-badīʿ, which, as its compiler explains in his introduction, is confined to the proverbs that are current among people (al-amṯāl al-mutadāwala bayna l-nās) and are extracted from his other book, which he refers to as Tuḥfat al-adīb.333 It includes 423 entries, the vast majority of which are single lines. All items are not ascribed and are arranged, according to their rhyme, from alif to yāʾ.

Ibn Aydamur’s (d. 710/1310) al-Durr al-farīd wa-bayt al-qaṣīd is certainly the largest collection of lines of poetry that are used as proverbs (sawāʾir masīr al-amṯāl) and are cited on various occasions (yustašhad bihā fī kaṯīr min al-aḥwāl).334 Although Ibn Aydamur acknowledges earlier efforts in this domain, he accuses his predecessors of failure to collect but very few lines of poetry (qalīla ǧiddan maʿdūma maʿdūda ʿaddan),335 and asserts that his book contains twenty thousand lines336 which he collected over a long period of time. This number dwarfs both Ḥamza’s 5,591 lines in al-Abyāt al-ṣādira and Maydānī’s 6,000 proverbs, although comparison with the latter may not be fully justified due to the difference in the corpus from which Maydānī and Ibn Aydamur derive their material. As far as arrangement is concerned, Ibn Aydamur adopts a fully alphabetical order based on the first word, rather than rhyme. In case the first word in two or more lines is the same, then the follow- ing word (or words) is taken into consideration. Ibn Aydamur clarifies that he adopted this system from the lexica of language, Ḥadīṯ, medicine and history.337 There are, however, certain lines that are arranged in a manner which violates the strict arrangement adopted by the author (e.g. fa-in anta lam . . ., fa-in antum lam . . ., and fa-in anta lāqayta).338 Furthermore, Ibn Aydamur deliber- ately violates the alphabetical order by starting the collection with lines which begin with al-ḥamdu li-l-Lāhi (Praise be to God), followed by those which

333 Nawāǧī, Zahr 79. 334 Ibn Aydamur, Durr I, 174. 335 Ibid., I, 181. 336 Based on a rough calculation of the number of pages in which the lines are listed,

multiplied by the constant number of ten lines in each page, the total number is closer to 18,150.

337 Ibid., I, 182. 338 Ibid., IV, 141–142.

 

 

127MUBAWWAB AND SPECIALIZED LEXICA

begin with Allāh,339 and appropriately ends it with lines commencing with astaġfiru l-Lāha (I ask God’s forgiveness).340 The book’s manuscript (which was reproduced in five volumes) is full of other lines of poetry which either resemble those in Ibn Aydamur’s text or are derived from the same poems. It also often includes the names of the poets whose lines are quoted as well as their biographies.

The third type of books that are specialized in a particular aspect of prov- erbs derives its material from three sources, namely, Qurʾān, Ḥadīṯ and wisdom literature. Unlike poetry which, as Ḥamza notes in al-Amṯāl al-ṣādira, contains either old proverbs or expressions which became so widespread that they were often used as proverbs, these three sources fit into the latter category exclu- sively. As far as Qurʾānic proverbs are concerned, the bibliographical and bio- graphical sources ascribe books that deal with them to authors from the late third/ninth and early fourth/tenth centuries, such as al-Ǧunayd (d. 298/910) and Nifṭawayhi (d. 323/935), to each of whom is ascribed a book entitled K. Amṯāl al-Qurʾān.341 Most probably to the same period belongs an extant work which may well be similar in content to earlier works of this genre. This is al-Amṯāl al-kāmina fī l-Qurʾān by al-Ḥusayn (or al-Ḥasan) b. al-Faḍl, who, according to the editor,342 is most likely to be the Kufan Qurʾānic scholar by the name of al-Ḥusayn b. al-Faḍl who settled in Nishapur and died in 282/895. The book consists of thirty-six short paragraphs of no apparent arrangement, each including a response by al-Ḥusayn on whether a certain proverb has a parallel in the Qurʾān. al-Ḥusayn typically adduces one or more such parallels without further comment. For example, the verse wa-fīkum sammāʿūna lahum (Some among you would have listened to them; Q 9: 47) is cited as equivalent to the proverb li-l-ḥīṭāni āḏānun (Walls have ears),343 and the verse wa-lakum fī l-qiṣāṣi ḥayātun yā ulī l-albābi (In retaliation you have a safeguard for your lives O men of understanding; Q 2: 179) is given as a parallel to al-qatlu anfā li-l-qatli (Killing is more preclusive of killing).344 Even if these Qurʾānic expres- sions are not proverbs in the strict sense, their comparison with proverbs is most probably based either on their currency in writing and speech, or on their semantic component which, at least theoretically, disposes them to be

339 Ibid., I, 185–195. 340 Ibid., V, 535–536. 341 Ibn al-Nadīm, Fihrist 41, 238 (al-Ǧunayd), and Yāqūt, Muʿǧam I, 122 (Nifṭawayhi). See also

Sellheim (1954: 20). 342 al-Ḥusayn b. al-Faḍl, Amṯāl 306–307. 343 Ibid., 318. 344 Ibid., 321.

 

 

128 chapter 2

used as proverbs. In a chapter entitled Fī amṯāl al-Qurʾān, Suyūṭī (d. 911/1505) in al-Itqān fī ʿ ulūm al-Qurʾān quotes thirteen examples from al-Ḥusayn b. al-Faḍl’s book.345 He explains that such proverbs are described as kāmina (latent; cf. the title al-Amṯāl al-kāmina fī l-Qurʾān) because they do not explicitly contain the word maṯal, as in maṯaluhum ka-maṯali l-laḏī stawqada nāran (They are like one who kindled a fire; Q 2: 17). Expressions of this latter type also occur in books that deal with proverbs in Ḥadīṯ as well (see next paragraph). Suyūṭī also quotes, in two of his books,346 thirty Qurʾānic expressions that are proverbially used (ǧāriya maǧrā l-maṯal). As he himself mentions, these are derived from a chapter in K. al-Ādāb by Maǧd al-Mulk Ǧaʿfar b. Šams al-Ḫilāfa (d. 622/1225).347

Several extant sources also collect ḥadīṯs that are used as proverbs. A fourth/ tenth century book of this type is Amṯāl al-Ḥadīṯ al-marwiyya ʿan al-nabī by al-Ḥasan b. ʿAbdalraḥmān b. Ḫallād al-Rāmahurmuzī (d. c. 360/971). The ḥadīṯs which are quoted by the author and are classified as proverbs in other books on amṯāl, however, are rather few. One such ḥadīṯ is iyyākum wa-ḫaḍrāʾa l-dimani (Beware of the beautiful woman who is of bad origin), which is also quoted by Maydānī and Zamaḫšarī.348 The majority of the other ḥadīṯs are not prov- erbs in the strict sense; rather, they are similes which involve tamṯīl (illustra- tion) and uniformly use the expression maṯalu X is maṯalu/ka-maṯali Y,349 such as maṯalu l-muʾmini l-qawiyyi maṯalu l-naḫlati wa-maṯalu l-muʾmini l-ḍaʿīfi ka-ḫāmati l-zarʿi (The example of the strong believer is that of a palm tree; and the example of the weak believer is that of a tender plant), and maṯalu ummatī maṯalu l-maṭari lā yudrā awwaluhu ḫayrun am āḫiruhu (The example of my community is that of rain – one never knows whether its first showers are better or its last).350 Rāmahurmuzī often extensively explains uncommon words in his ḥadīṯs and adduces poetry šawāhid for that purpose, but follows no alphabetical order. Moreover, the last two chapters of the book are devoted to rare usage (nawādir) in a number of ḥadīṯs.351

Another book that deals with proverbs in Ḥadīṯ is al-Amṯāl fī l-Ḥadīṯ al-nabawī, by Rāmahurmuzīʾs contemporary, Abū l-Šayḫ al-Iṣfahānī (d. 369/ 979). The first volume includes 123 ḥadīṯs which have become proverbs (ṣāra

345 Suyūṭī, Itqān IV, 41–43. 346 Ibid., IV, 43–45; Suyūṭī; Muʿtarak I, 470–471. 347 Cf. Ǧaʿfar, Ādāb 61–63. 348 Rāmahurmuzī, Amṯāl 121; cf. Maydānī, Maǧmaʿ I, 32; Zamaḫšarī, Mustaqṣā I, 451. 349 A large body of such expressions in Qurʾān, Ḥadīṯ and wisdom literature was collected

earlier by al-Ḥakīm al-Tirmiḏī (d. c. 320/932) in al-Amṯāl min al-Kitāb wa-l-Sunna. 350 Rāmahurmuzī, Amṯāl 79, 105. 351 Ibid., 159–164.

 

 

129MUBAWWAB AND SPECIALIZED LEXICA

maṯalan),352 whereas the second lists expressions of the tamṯīl type mentioned above. This distinction in subject matter is important because it betrays the conviction that similes of the type maṯalu X is maṯalu Y are not truly proverbs and should not be mixed with commonly used ḥadīṯs that are quoted as prov- erbs. The material in both sections is not arranged according to any criterion, and the author often cites various chains of authority (isnād) and riwāyas of the same ḥadīṯ although he does not explain the meaning of the cited ḥadīṯs. Towards the end of the book, he departs from its theme and lists sayings attrib- uted to prophets other than Muḥammad and to other Muslim dignitaries such as ʿĀʾiša and ʿAlī. Most surprising, however, is the inclusion of several proverbs attributed to the Ǧāhilī sage, Akṯam b. Ṣayfī.353

In the fifth/eleventh century, Abū ʿAbdallāh al-Quḍāʿī (d. 454/1062) col- lected in his Šihāb al-aḫbār fī l-ḥikam wa-l-amṯāl wa-l-ādāb al-marwiyya ʿan al-rasūl al-muḫtār about 1,200 ḥadīṯs, according to his introduction.354 As the book’s title indicates, proverbs are only part of the collected items, and Quḍāʿī further explains in the introduction that the ḥadīṯs he cites involve wisdom sayings (ḥikma) which he further specifies as waṣāyā (precepts), ādāb (rules of disciplines), mawāʿiẓ (sermons), and amṯāl. The ḥadīṯs are consecutively listed without any explanation, but are divided over fifteen untitled chapters which group together ḥadīṯs whose first words share a common feature in form or which begin with a specific word or particle. Among these chapters, for exam- ple, are those whose ḥadīṯs begin with a noun, the conditional particle man (itself considered by grammarians to be a noun, ism), a past or present verb, an imperative verb, the negative particle mā, the emphatic particle inna, etc.355 Chapter eleven, it should be noted, is devoted to expressions of the previously mentioned tamṯīl type (i.e. maṯalu X is maṯalu Y).356

Proverbs attributed to the fourth caliph and Prophet’s cousin, ʿAlī b. Abī Ṭālib (d. 40/661), were also compiled by an anonymous author. Two hundred and eighty-two of these proverbs are simply listed alphabetically in Amṯāl sayyidinā ʿAlī from alif to yāʾ, based only on the first letter.

Two more books are worthy of mention. The first, K. al-Amṯāl al-musammā bi-l-Farāʾid wa-l-qalāʾid, is erroneously attributed to Ṯaʿālibī (d. 429/1039), but is most probably authored by his contemporary Abū l-Ḥasan al-Ahwāzī

352 Abū l-Šayḫ, Amṯāl I, 21. 353 Ibid., II, 271–282. 354 Quḍāʿī, Šihāb 2. 355 Ibid., 2–64, 64–96, 97–107, 108–138, 138–147, 168–184. 356 Ibid., 213–218.

 

 

130 chapter 2

(d. 428/1037).357 This is a collection of short sayings that are thematically arranged in eight chapters (e.g. knowledge and reason, asceticism, high moral standards, etc.). What is interesting is that these sayings, as the author men- tions in his introduction, are of his own composition and are intended to be brief and proverb-like (wa-qad ǧamaʿnā min inšāʾinā fī kitābinā hāḏā alfāẓan waǧīza aǧraynāhā muǧrā l-amṯāl).358 This astounding remark demonstrates how the notion of proverb developed to include statements that are deliber- ately composed with the intention of making them available (especially for writers) to be used as proverbs, contrary to the “natural” way in which proverbs ordinarily come into usage. The second book is ʿAlī b. Muḥammad al-Māwardī’s (d. 450/1058) al-Amṯāl wa-l-ḥikam. It consists of ten chapters, each of which contains thirty ḥadīṯs, thirty wisdom sayings (referred to as amṯāl al-ḥukamāʾ, “proverbs of the wise”), and thirty lines of poetry.359 This brings the total num- ber of items to nine hundred. The division of the material into ten chapters however does not rest on any thematic grounds. It should also be noted that although some of the items which the author assembled in a number of chap- ters entitled amṯāl al-ḥukamāʾ360 are proverbs which occur in earlier books on amṯāl (e.g. Ḥamza al-Iṣfahānī’s al-Durra al-fāḫira), the vast majority of the items are wisdom sayings (ḥikma) which occur in more general literary sources, including Ǧāḥiẓ’s (d. 255/ 869) al-Bayān wa-l-tabyīn and Ibn Qutayba’s (d. 276/889) works, such as ʿUyūn al-aḫbār and Faḍl al-ʿArab. Furthermore, some of the items are ascribed to non-Arabs, such as Anūširwān, Buzarǧimahr, and Greek and Indian sages.361 It seems that as the boundaries between amṯāl and ḥikam became increasingly blurred, some authors chose to derive their material not only from earlier works on amṯāl, but also from the wider spec- trum of literary and philological sources which abound with wisdom sayings.

The fourth type of books that specialize in a particular aspect of proverbs deals with the muwallad (post-classical) and ʿ āmmī (vernacular). It thus largely falls beyond the scope of the present study, and will be only briefly mentioned for the sake of giving the full picture of the whole genre. As far as muwallad proverbs are concerned, some of the previously mentioned authors do include them in their works although they belong to the post-classical era. As previ- ously noted, muwallad proverbs form part of al-Durra al-fāḫira and Maǧmaʿ al-amṯāl, but are listed in separate sections as if to highlight their special

357 Cf. Orfali (2009: 301). 358 Ahwāzī, Farāʾid 3. 359 Māwardī, Amṯāl 30. 360 See, for example, ibid., 172 (items 767, 769–770, 772–774, 777). 361 Ibid., 121–124.

 

 

131MUBAWWAB AND SPECIALIZED LEXICA

status. For his part, ʿAskarī (d. after 395/1005), in his introduction to Ǧamharat al-amṯāl,362 contrasts proverbs that are ʿArabī and ṣaḥīḥ (correct) with those that are muwallad and saqīm (weak or faulty). According to this distinction, what is muwallad is not considered to be purely Arabic and hence should be excluded. Yet, in the fourth/tenth century, Abū Bakr al-Ḫuwārizmī (d. 383/993) devotes his al-Amṯāl al-muwallada to these proverbs, and sets in his intro- duction a more detailed criterion than ʿAskarī’s for what a muwallad proverb is. He explains that his proverbs were more recently invented (istaḥdaṯahā) by people of the ʿAbbāsid period (abnāʾ al-dawla al-ʿAbbāsiyya), mainly in Baghdad and Damascus, and not by the early Arabs (al-ʿArab al-uwal) to whom he generically refers as Ḥiǧāziyyūn.363 These proverbs, he further clari- fies, are not to be found in the works of earlier lexicographers, such as Abū ʿUbayd, al-Mufaḍḍal al-Ḍabbī and Aṣmaʿī; rather, he had to collect them from the speech of those whom he describes as deceitful (šuṭṭār) and mischievous (ʿayyārūn), as well as from singers (muġannūn), comedians (muḍḥikūn), beg- gars (suʾʾāl), travelers (sābila), humorous individuals (ẓurafāʾ), and mystics (ṣūfiyya).364 He also mentions among his sources professional writers, mer- chants, poets, kings and servants.365 Because his materiel is directly taken from such sources (min afwāh . . .), Ḫuwārizmī felt that he had to apologize to his reader (wa-naḥnu naʿtaḏir ilayka) for having to conform to the standards of his times during which it was common to complain of the complexity of grammar and language and to ridicule pre-Islamic poetry as old-fashioned and forsaken.366 In collecting his material, Ḫuwārizmī naturally had to depart from the traditional method of collecting proverbs based on riwāya or on ear- lier books. It is remarkable however that apart from a number of words that are clearly muwallad,367 the proverbs he cites do not reveal any significant departure from Fuṣḥā. It is therefore quite likely that the author intervened to conceal some of the non-standard linguistic characteristics of his proverbs, contrary to Ṭālaqānī and Zaǧǧālī (see below), in order to allow that they be read in conformity to Fuṣḥā. The book also contains many proverbs that are derived from poetry and Qurʾān. It consists of a total of 2,119 items which are arranged in twenty-one chapters based either on a certain theme (e.g. cursing

362 ʿAskarī, Ǧamhara I, 6. 363 Ḫuwārizmī, Amṯāl 71–72. 364 Ibid., 66. 365 Ibid., 71–72. 366 Ibid., 68. 367 See the book’s indices, 529–530.

 

 

132 chapter 2

or imprecating, praise and sympathy, impudence, etc.) or on formal criteria (e.g. proverbs beginning with afʿal or with ka⁠ʾanna/ka⁠ʾannamā).

Unlike Ḫuwārizmī, Abū l-Ḥasan al-Ṭālaqānī, who compiled his Risālat al-amṯāl al-Baġdādiyya allatī taǧrī bayna l-ʿāmma in 421/1030, and Abū Yaḥyā al-Zaǧǧālī (d. 694/1295), compiler of Amṯāl al-ʿawāmm fī l-Andalus, cite most of their proverbs in their ʿāmmī (vernacular) form, hence the term ʿāmma/ ʿawāmm, rather than muwallad, in the titles of their books. Examples include ḥaǧǧ wa-l-nās rāǧiʿīn (He went on pilgrimage as people were coming back [from Mecca]), where the oblique form is used in rāǧiʿīn,368 and bi-l-isbināḫ tistaġnū ʿ an al-afrāḫ (By eating spinach, you can dispense with chicken), where the plural verb appears without a final nūn.369 In several cases, both authors cite proverbs which exactly match the wording of non-ʿāmmī proverbs cited by other authors,370 or which are effectively the ʿāmmī versions of those prov- erbs.371 Both works are arranged alphabetically according to the first letter only; the number of proverbs in Ṭālaqānī’s is about 600 and in Zaǧǧālī’s 2,157. Proverbs of the ʿ āmma in Andalus are also listed by Muḥammad b. Muḥammad Ibn ʿĀṣim al-Ġarnāṭī (d. 829/1294) in a chapter of his Ḥadāʾiq al-azhār.372 Their total number is 851, and they do not differ in arrangement from Zaǧǧālī’s book.

4 al-Nabāt (Plants), al-Ḥayawān (Animals), Ḫalq al-Insān (Human Body), etc.

Scholars of the late second/eighth and early third/ninth centuries authored a large number of short monographs (and a few extensive works) which contain vocabulary related to a narrow semantic field. Although most of these works are lost, those that remain give a clear picture of this branch of lexicography which contributed not only to multithematic works but also to lexica of the muǧannas type. It is obvious that these monographs are the direct result of the process of data collection (ǧamʿ al-luġa) started by philologists who made the journey to the desert in order to record the faṣīḥ usage of the Aʿrāb. In fact,

368 Ṭālaqānī, Risāla 13. 369 Zaǧǧālī, Amṯāl II, 27. 370 E.g. rubba aklatin tamnaʿu akalāt (Many a time one meal prevents several); Ṭālaqānī,

Risāla 17 [vocalization ours]; cf. Zamaḫšarī, Mustaqṣā II, 93. 371 E.g. uqil li-l-baġl man huwa wildak qāl al-faras ḫālī (Asked who his father was, the mule

answered: “My materanl uncle is the horse”); Zaǧǧālī, Amṯāl II, 21; cf. qīla li-l-baġli man abūka qāla l-farasu ḫālī (Maydānī, Maǧmaʿ II, 110).

372 The chapter is published by Ahwānī (1962: 297–364).

 

 

133MUBAWWAB AND SPECIALIZED LEXICA

most of the philologists who went on such journeys are themselves authors of short monographs of specialized vocabulary. These include al-Naḍr b. Šumayl (d. 203/819), Abū ʿAmr al-Šaybānī (d. 206/821), Farrāʾ (d. 207/822), Abū Zayd al-Anṣārī (d. 215/830), Aṣmaʿī (d. 216/831), Abū ʿUbayd al-Qāsim b. Sallām (d. 224/838), and Ibn al-Aʿrābī (d. 231/845). Monographs by authors of this period form the vast majority of this type of writing, and although some later scholars – as late as the tenth/sixteenth and the eleventh/seventeenth centu- ries in some cases – still authored such monographs, they had in most cases precious little to add to the early material which the second/eighth and third/ ninth century philologists had recorded and preserved from being lost forever due to the rapid change in lifestyle in the early Islamic period.

The existence of the short thematic monographs in the same period as that of the first grammatical works and grammatically oriented exegetical works demonstrates the early distinction between luġa and naḥw.373 Semantics was not among the primary linguistic interests of the grammarians, and although they do cite rare or ġarīb lexemes, they normally examine their patterns from a morphophonological perspective with very little interest in their meanings.374 This domain was almost entirely left for the philologists (luġawiyyūn), who, in spite of the fact that they did not develop a theory of semantics that can even come close to the grammatical theory, should be credited for having assembled in their short monographs a huge corpus of linguistic material from various semantic fields and cited related šawāhid of attested usage, mainly from poetry. Within this corpus, the ġarīb occupies a central position. On many occasions, for example, one discovers that even some philologists of that period never heard of or did not know the meanings of certain lexemes cited in these monographs. Thus, Abū Ḥātim al-Siǧistānī (d. 255/869) is reported not to have heard of a plant known as duʿāʿ, and Abū l-Ḥasan al-Aḫfaš al-Awsaṭ (d. 215/830) admits that he did not know the meaning of baʿāʿa (or perhaps luʿāʿa or nuʿāʿa), a soft kind of herb.375 Recurring reference to dialectal variants and idiosyncrasies provides further evidence for the interest of the authors of these monographs in ġarīb since dialectal material is a major source of rare or strange usage. To take the genre of nabāt (plants) as an example, one finds in it references to the dialects of Ḥiǧāz, Naǧd, Madīna, Yamāma, Naǧrān, Tihāma,

373 Cf. above, 4. 374 Sībawayhi’s total disregard for the meaning of ġarīb words which he cites in his study of

appending (ilḥāq) has been already mentioned above (cf. p. 60). 375 Aṣmaʿī, Nabāt 6; cf. Azharī, Tahḏīb I, 114.

 

 

134 chapter 2

Baḥrayn, Baṣra, Kūfa, Ḥīra, Šām, Tamīm, Balḥāriṯ b. Kaʿb, etc.376 There are also references to individual Bedouin fuṣaḥāʾ known for their use of ġarīb, such as Umm al-Hayṯam and Ibnat al-Ḫuss.377 Reference to the foreign origin of cer- tain lexemes378 might also be indicative of ġarīb in certain cases.

It is not surprising, given that single-theme works mostly belong to the sec- ond/eighth and third/ninth centuries, that their vast majority are not alpha- betically arranged. With a few exceptions, even later authors of similar works follow their predecessors in not arranging their material alphabetically. The early period in which most of these works were authored, on the other hand, meant that they provided much of the raw material for later muǧannas lex- ica. In this respect, the study of the sources cited in Ibn Manẓūr’s (d. 711/1311) Lisān al-ʿArab should give a good idea about the contribution of single-topic thesauri to muǧannas dictionaries since the Lisān itself is derived from five earlier muǧannas dictionaries, as we shall see later.379 From the genre of nabāt are mentioned Abū Zayd’s (d. 215/830) al-Kala⁠ʾ wa-l-šaǧar,380 Dīnawarī’s (d. 282/895) al-Nabāt,381 and ʿAlī b. Ḥamza’s (d. 375/985) al-Nabāt,382 and from the genre of ḥayawān al-Naḍr b. Šumayl’s (d. 203/819) al-Ḫayl and al-Wuḥūš,383 Abū ʿUbayda’s (d. 209/824) al-Ḫayl, Ṣifāt al-ḫayl and al-Šiyāt,384 Abū Zayd al-Anṣārī’s (d. 215/830) al-Ġanam,385 Aṣmaʿī’s (d. 216/831) al-Ḫayl and al-Faras,386 Abū ʿUbayd’s (d. 224/838) al-Ḫayl,387 Abū Ḥātim al-Siǧistānī’s (d. 255/869)

376 Examples are: Aṣmaʿī, Naḫl 65, 67, 71; Abū Ḥātim, Naḫla 138; Dīnawarī, Nabāt V, 68, 153, 204 (Ḥiǧāz); Aṣmaʿī, Naḫl 65, 66, 69, 71; Abū Ḥātim, Naḫla 134, 138, 162; Dīnawarī, Nabāt V, 68, 153 (Naǧd); Aṣmaʿī, Naḫl 68, 69; Abū Ḥatīm, Naḫla 149, 162; Dīnawarī, Nabāt V, 204 (Madīna); Abū Zayd, Šaǧar 184; Aṣmaʿī, Naḫl 72; Abū Ḥātim, Naḫla 130, 143, 144 (Yamāma); ibid., 144 (Naǧrān); Abū Zayd, Šaǧar 162; Dīnawarī, Nabāt V, 154 (Tihāma); ibid., V, 153, 194 (Baḥrayn); Abū Ḥātim, Naḫla 146; Dīnawarī, Nabāt V, 71, 153, 194 (Baṣra); Abū Ḥātim, Naḫla 141 (Kūfa); Dīnawarī, Nabāt V, 178 (Ḥīra); ibid., V, 41, 106, 166, 174 (Šām); Abū Zayd, Šaǧar 184 (Tamīm); Aṣmaʿī, Naḫl 68 (Balḥāriṯ b. Kaʿb).

377 Abū Ḥātim, Naḫla 119, 126; Dīnawarī, Nabāt III & V, 23, 25, 76; V, 31, 154. 378 E.g. Abū Ḥātim, Karm 89; Abū Ḥātim, Naḫla 134, 144, 158, 167; Dīnawarī, Nabāt V, 74

(Fārisiyya); Abū Ḥātim, Naḫla 134; Dīnawarī, Nabāt V, 74 (Nabaṭiyya). 379 See below, 386. 380 Ibn Manẓūr, Lisān (ʿḌḌ). 381 Ibid. (ŠBB, NḌǦ, BQR, etc.). 382 Ibid. (ǦʾṮ, BSL). 383 Ibid. (FṢṢ, RḌF [al-Ḫayl]; DMY [al-Wuḥūš]). 384 Ibid. (ḌBḤ, RṢʿ, ḤFR, etc. [al-Ḫayl]; MḤṢ [Ṣifāt al-ḫayl]; QṬʿ [al-Šiyāt]). 385 Ibid. (ḤDʾ, ǦHW). 386 Ibid. (LMʿ [al-Ḫayl]; ʿWǦ, ḪRF, FYL [al-Faras]). 387 Ibid. (ḪNB, MLK, NZW).

 

 

135MUBAWWAB AND SPECIALIZED LEXICA

al-Ṭayr,388 and Šamir b. Ḥamdawayhi’s (d. 255/869) al-Ḥayyāt.389 Material is also cited from monographs on various other topics, including Abū Zayd’s al-Maṭar,390 Abū ʿUbayd’s al-Laban,391 Abū Ḥātim’s al-Šams wa-l-qamar,392 and Šamir’s al-Silāḥ.393 It is interesting to note the existence of a reverse phe- nomenon since some of the mutlithematic thesauri, such as the voluminous al-Muḫaṣṣaṣ by Ibn Sīda al-Andalusī (d. 458/1066), derive their material, in addition to single-topic works, from muǧannas lexica.394 In the section on ʿiḍāh (great thorny trees), for example, Ibn Sīda cites not only Abū Zayd, Abū ʿUbayd, Ibn al-Aʿrābī, Ibn al-Sikkīt (d. 244/858) and Dīnawarī as authors of earlier mubawwab works, but also Ḫalīl b. Aḥmad (d. 175/791) and Ibn Durayd (d. 321/933) as authors of the muǧannas lexica al-ʿAyn and Ǧamharat al-luġa respectively.395 Although the main trend was for authors of muǧannas lexica to borrow material from mubawwab ones, the reverse phenomenon demon- strates that since the writing of the mubawwab type continued long after the widespread use of the alphabetical arrangement of lexical items, authors of these lexica availed themselves of the huge corpus which the muǧannas type readily offered.

The following discussion of this branch of lexicography shall deal with the extant works according to their themes, namely, plants, animals, human body, and miscellaneous topics.

4.1 Plants Based on the bibliographical and biographical sources, most of the mono- graphs that deal with plants go back to the first half of the third/ninth century (according to the death dates of their authors). Other than the extant sources discussed below, titles of this theme cited in Ibn al-Nadīm’s (d. 380/990) al- Fihrist and Yāqūt’s (d. 626/1229) Muʿǧam al-Udabāʾ include the following: al-Naḍr b. Šumayl’s (d. 203/819) al-Ṣifāt, parts of whose fifth section deal with zarʿ (planting), karm (vineyard), ʿinab (grapes), asmāʾ al-buqūl (names of herbs), and šaǧar (trees);396 Abū ʿAmr al-Šaybānī’s (d. 206/821) al-Naḫla;397

388 Ibid. (ĠRN). 389 Ibid. (ŠǦʿ, FʿW). 390 Ibid. (ṮWR, ĠRL). 391 Ibid. (ṢBR). 392 Ibid. (YWḤ). 393 Ibid. (HLL, ʿLM, ʿLH). 394 See below, 276–277. 395 Ibn Sīda, Muḫaṣṣaṣ XI, 181–183. 396 Ibn al-Nadīm, Fihrist 57–58; Yāqūt, Muʿǧam VI, 2761; cf. Azharī, Tahḏīb I, 17. 397 Ibn al-Nadīm, Fihrist 75.

 

 

136 chapter 2

Abū ʿUbayda’s (d. 209/824) al-Zarʿ;398 Abū Zayd’s (d. 215/830) al-Tamr;399 al-Madāʾinī’s (d. 225/840) al-Naḫl;400 Abū Naṣr al-Bāhilī’s (d. 231/846) al-Šaǧar wa-l-nabāt and al-Zarʿ wa-l-naḫl;401 Ibn al-Aʿrābī’s (d. 231/845) Ṣifat al-naḫl, Ṣifat al-zarʿ, al-Nabāt and al-Nabt wa-l-baql;402 Ibn al-Sikkīt’s (d. 244/858) al-Šaǧar wa-l-nabāt;403 Muḥammad b. Ḥabīb’s (d. 245/860) al-Nabāt;404 Abū Ḥātim al-Siǧistānī’s (d. 255/869) al-Nabāt, al-Zarʿ, and al-ʿUšb wa-l-baql;405 and Ǧāḥiẓ’s (d. 255/869) al-Zarʿ wa-l-naḫl.406 The writing of similar monographs continued throughout the fourth/tenth century, but there also appeared a few voluminous works, most notably Dīnawarī’s (d. 282/895) al-Nabāt (for which see below) and Ibn Waḥšiyya’s (fl. in the third/ninth century) al-Filāḥa al-Nabaṭiyya. The latter is an impressive encyclopedia of plants and meth- ods of their cultivation, but it is not within our scope since it was, as stated in the author’s introduction, translated from the language of the Chaldeans into Arabic.407 Ibn Waḥšiyya cites the original title of the book in Nabatean (i.e. Syriac) as Iflāḥ al-arḍ wa-iṣlāḥ al-zarʿ wa-l-šagar wa-l-ṯimār wa-dafʿ al-āfāt ʿanhā and says that its translation took place in 291/903–904.

Works of the genre that have survived either in whole or in part are the following:

1. al-Šaǧar wa-l-kala⁠ʾ by Abū Zayd al-Anṣārī (d. 215/830). 2. al-Nabāt and al-Naḫl by Aṣmaʿī (d. 216/831). 3. al-Karm and al-Naḫla by Abū Ḥātim al-Siǧistānī (d. 255/869). 4. al-Nabāt by Abū Ḥanīfa al-Dīnawarī (d. 282/895).

The first book was originally published in 1909 by S. Nagelberg, who mistakenly attributed it to Ibn Ḫālawayhi (d. 370/980). A later edition by A. Abū Suwaylim and M. Šawābika however rightly attributes it to its true author. The various citations from the book in later sources suggest that the author is a much ear- lier scholar, Abū Zayd al-Anṣārī, one of the most active philologists in the era

398 Ibid., 59; Yāqūt, Muʿǧam VI, 2708. 399 Ibn al-Nadīm, Fihrist 60; Yāqūt, Muʿǧam III, 1361. 400 Yāqūt, Muʿǧam IV, 1858; cf. Ibn al-Nadīm, Fihrist 117 where the title is undotted. 401 Ibn al-Nadīm, Fihrist 61; Yāqūt, Muʿǧam I, 227. 402 Ibn al-Nadīm, Fihrist 76; Yāqūt, Muʿǧam VI, 2533. 403 Ibn al-Nadīm, Fihrist 79; Yāqūt, Muʿǧam V, 1922 (al-Nabāt); cf. Ibn Sīda, Muḫaṣṣaṣ I, 11. 404 Yāqūt, Muʿǧam VI, 2482. 405 Ibn al-Nadīm, Fihrist 64; Yāqūt, Muʿǧam III, 1407–1408 (al-Šaǧar wa-l-nabāt). 406 Ibn al-Nadīm, Fihrist 210; Yāqūt, Muʿǧam V, 2118. 407 Ibn Waḥšiyya, Filāḥa I, 5. For the controversy surrounding the book’s translation, see Fahd

(1986: 964).

 

 

137MUBAWWAB AND SPECIALIZED LEXICA

of data collection.408 The book consists of two sections, each of which might have been an independent monograph since it is referred to as kitāb in Ibn Ḫālawayhi’s riwāya. These are Asmāʾ al-šaǧar (Names of Trees) and Asmāʾ al-kala⁠ʾ (Names of Herbages). The former is arranged according to the type of trees under discussion whereas the latter is not as clearly arranged. The author, for example, divides šaǧar in the first kitāb into two types, namely, the ʿiḍāh (great thorny trees) – which in turn are divided into al-ʿiḍāh al-ḫāliṣ (genuine ʿiḍāh) and ʿiḍāh al-qiyās (ʿiḍāh of bows) – and non-ʿiḍāh which include small thorny trees (ʿiḍḍ and širs).409 It is obvious that Abū Zayd’s purpose is primarily to record the names of trees and herbages (usually by citing the generic noun, followed by the singular form; e.g. ṭalḥ/ṭalḥa “acacia”), and he only makes brief reference to their characteristics, use and habitat and cites only a handful of supportive poetry šawāhid.

Aṣmaʿī’s al-Nabāt resembles Abū Zayd’s book in that it is basically a list of names of plants loosely arranged according to their types. Grouped together, for example, are plants which belong to a certain type, such as aḥrār al-baql (herbs that are eaten uncooked), ḏukūr al-baql (thick and rough herbs), and ḥamḍ/ḫulla (salty/non-salty plants).410 Some sections are arranged according to habitat rather than type, such as the trees of Ḥiǧāz and the mountains of Naǧd.411 Unlike Abū Zayd, Aṣmaʿī does not usually give the singular form of the names, but his poetry šawāhid are more numerous. As for al-Naḫl, its attri- bution to Aṣmaʿī is in some doubt since a number of quotations of it in later sources do not agree with the published text, itself based on an unattributed manuscript. This notwithstanding, much of the material of the monograph appears in Abū ʿUbayd’s (d. 224/838) al-Ġarīb al-muṣannaf and is attributed to Aṣmaʿī.412 Naṣṣār rightly suggests that al-Naḫl’s text is derived from Abū ʿUbayd’s book (which is largely dependent on Aṣmaʿī) by a scholar – possibly Ibn Qutayba (d. 276/889) – who omitted from it the names of Abū ʿUbayd’s narrators (ruwāt) and the poetry šawāhid.413 In all cases, al-Naḫl shows bet- ter arrangement of its content than al-Nabāt. Its author divides the text into short sections which deal with specific aspects of palm trees, such as their fruits, branches, heights, types, defects, etc. This arrangement becomes easier

408 See the editors’ comments in Abū Zayd, Šaǧar 139–140. 409 Ibid., 150, 159, 162–163. 410 Aṣmaʿī, Nabāt 13, 15, 17–18. 411 Ibid., 23–24. 412 See, for example, Aṣmaʿī, Naḫl 69–71 and Abū ʿUbayd, Ġarīb II, 486–490. 413 Cf. Naṣṣār (1985: 50).

 

 

138 chapter 2

to explain if one subscribes to the idea that the text is based on al-Ġarīb al-muṣannaf, which is well-known for the accuracy of its internal arrangement.

From a slightly later period, we are in possession of two texts by Abū Ḥātim al-Siǧistānī, al-Karm and al-Naḫla. Both are considerably longer than the three previous monographs, and it is likely that Abū Ḥātim benefited from works authored a few decades earlier. The attribution of the first text by its editors, A. Haffner and L. Cheikho, to Aṣmaʿī (based on the fact that its manuscript was found with K. al-Naḫl, which they also ascribe to Aṣmaʿī) is most probably incorrect. The real author seems to be Abū Ḥātim, as indicated in the manu- script itself (cf. K. al-Karm ʿan Abī Ḥātim al-Siǧistānī).414 Indeed, Ibn al-Nadīm and Yāqūt ascribe to Abū Ḥātim, but not to Aṣmaʿī, a book by this title.415 For his material, Abū Ḥātim interestingly relies on specific informants, most nota- ble of whom is the one he refers to as Ṭāʾifī along with a group of Ṭāʾifiyyūn. He also relies on a certain Abū ʿAlī al-Ǧaʿdī and on someone whom he calls Abū l-Ḫaṭṭāb without any mention of his tribe or abode. Accordingly, much of the text directly reports the speech of these individuals, hence the occur- rence of some forms in the first person (e.g. nusammīhi or sammaynāhu; “we call it”).416 Terms related to vineyards and grapes are either embedded in the description of the various stages of the plant’s cycle, or consecutively listed as in the various types of grapes. The final parts also include vocabulary related to the making of ḫamr (wine), ḫall (vinegar) and rubb (thickened juice of fruit).417 As for the second title, al-Naḫla, its text is considerably longer than that of al-Naḫl discussed above and shows marked development as far as arrange- ment is concerned. It is divided into two main parts, each of which starts with the traditional basmala and blessing of the Prophet as if it were an indepen- dent treatise.418 The first part focuses on the prestigious status of the palm tree, its mention in the Qurʾān and Ḥadīṯ, and its presence in the lands of Muslims, but not those of the polytheists. In the second part lie most of the terms related to palm trees and dates, both of which are surveyed with great detail. The text also reveals Abū Ḥātim’s interest in dialects and Arabized words and is replete with poetry šawāhid, unlike the earlier monographs discussed above.

Abū Ḥanīfa al-Dīnawarī’s al-Nabāt is certainly the most important landmark in the history of the genre. Authored only a few decades after the monographs

414 Abū Ḥātim, Karm 73. 415 Ibn al-Nadīm, Fihrist 64; Yāqūt, Muʿǧam III, 1408. 416 Abū Ḥātim, Karm 81. Note that Abū Ḥātim’s text is included in its entirety in K. al-Ǧarāṯīm,

which is erroneously attributed to Ibn Qutayba; cf. below, 271–272. 417 Ibid., 90–94. 418 Abū Ḥātim, Naḫla 112, 123.

 

 

139MUBAWWAB AND SPECIALIZED LEXICA

mentioned above, it is unprecedented both in size and arrangement. As far as size is concerned, ʿAbdalqādir al-Baġdādī (d. 1093/1682) mentions that the book is in six large volumes (muǧalladāt kibār sitta),419 whereas Sīrāfī’s (d. 368/979) division is different, and the second part of the fifth volume (published in 1953) corresponds to his eighth.420 The book thus represents a significant departure from the monograph type of texts which form the overwhelming majority of this branch of lexicography. Based on earlier written works and on informa- tion provided by the Bedouins, the book (of which only the third and fifth volumes have survived) is truly a botanical encyclopedia which contains not only detailed descriptions of plants and their use for various purposes, such as kindling and bow-making, but also of the different aspects of Bedouin tradi- tion related to plants. The other innovation which the book brings to the genre is the way in which it is structured. It is from Dīnawarī’s own reference to the structure of his book at the end of the first part of the fifth volume421 that we know that it is in two major parts, the first of which includes a general study of plants and the characteristics of each of their various genera. In the second part, Dīnawarī alphabetically lists the plants and individually discusses each, without having to repeat in each the characteristics of the genus to which it belongs. The alphabetical arrangement (i.e. ʾ, b, t, ṯ, etc.), which he says is the same as that of the generality of people (ʿāmma), is based solely on the first letters of words and makes no distinction between the original letters and pre- fixes (hence islīḥ and iġrīḍ are under alif ). In a rare instance, Dīnawarī reflects on the merits and demerits of adopting alphabetical arrangement of the sec- ond part of his book. He admits that this formal criterion results in the citing of plants without any distinction between large and small trees or between trees, herbs, and shrubs. Yet, he opts for its adoption to facilitate the book’s use, based on the fact that the desired distinction was achieved in the first part of the book. The most interesting part of Dīnawarī’s reflective statement is the assertion that he prefers the arrangement of the lexical items with respect to their first, rather than last, letters (wa-taṣnīfuhā ʿalā ḥurūf awāʾilihā aḥabb ilayya min taṣnīfihā ʿalā ḥurūf awāḫirihā). This is a striking statement in an era from which we know of no lexica that are arranged based on final letters, except for the first lexicon of this kind, al-Taqfiya,422 by Dīnawarī’s contem- porary, Abū Bišr al-Bandanīǧī, who died only two years after him in 284/897. Dīnawarī’s statement implies that alphabetical arrangement based on the final

419 Baġdādī, Ḫizāna I, 25. 420 Dīnawarī, Nabāt V, 5. 421 Ibid., III & V, 397. 422 For K. al-Taqfiya ascribed to Ibn Qutayba (d. 276/889), see below, 370.

 

 

140 chapter 2

letter of the word was sufficiently common to rival the one based on the first letter as a model to be adopted by lexicographers. Admittedly, this challenges our present knowledge of the first attempts to classify words in accordance with their final letters and hence of the first lexicon to adopt this model which later became the most dominant in Arabic lexicography.

It is remarkable that interest in authoring monographs or books on plants waned significantly as of the fourth/tenth century, contrary to many other genres. It should be remembered, however, that the practice of including plants among the topics of multithematic works continued beyond that date. The first chapters on plants in such works go back to Abū ʿUbayd’s (d. 224/838) al-Ġarīb al-muṣannaf which includes a chapter titled al-Šaǧar wa-l-nabāt and another titled al-Naḫl.423 Similar chapters are found in al-Muntaḫab min ġarīb kalām al-ʿArab by Kurāʿ al-Naml (d. 310/922),424 in addition to one that deals with karm, in K. al-Ǧarāṯīm, which is attributed to Ibn Qutayba (d. 276/889).425 But the most comprehensive chapters of this type are those in al-Muḫaṣṣaṣ by Ibn Sīda (d. 458/1066), who seems to have derived a large portion of his mate- rial from Abū ʿUbayd’s al-Ġarīb al-muṣannaf.

4.2 Animals Works which deal with animals are much more numerous than those that deal with plants and have continued to be authored until a much later period. This notwithstanding, both genres are alike in that in their majority they are not alphabetically arranged and primarily aim at collecting the pertinent vocabulary with only little interest in the internal arrangement of the mate- rial. Furthermore, most of the authors of the third/ninth century to whom are ascribed monographs on plants are also reported to have authored mono- graphs on various aspects related to animals. These authors are mostly lexicog- raphers who were active in the process of data collection, and to try to track the names of those of them to whom bibliographical and biographical sources attribute monographs on animals would certainly yield a list that is quite simi- lar to the one found at the beginning of the previous section devoted to plants.

Monographs and books dealing with ḫayl (horses) form a subcategory within the genre, and the numerous extant texts that fall under it will be dis- cussed separately. Other than ḫayl, the following extant sources deal either with animals in general or with a specific animal or group of animals:

423 Abū ʿUbayd, Ġarīb II, 419–438; 479–492. 424 Kurāʿ, Muntaḫab II, 455–470. 425 Ibn Qutayba, Ǧarāṯīm II, 55–114.

 

 

141MUBAWWAB AND SPECIALIZED LEXICA

1. al-Ibil, al-Šāʾ, and al-Wuḥūš, all by Aṣmaʿī (d. 216/831). 2. al-Ḥayawān by Ǧāḥiẓ (d. 255/869). 3. al-ʿAsal wa-l-naḥl by Abū Ḥanīfa al-Dīnawarī (d. 282/895). 4. Asmāʾ al-asad by Ibn Ḫālawayhi (d. 370/980). 5. Fī asāmī l-ḏiʾb wa-kunāhu by Ṣaġānī (d. 650/1252). 6. Ḥayāt al-ḥayawān al-kubrā by Damīrī (d. 808/1405).

These works vary tremendously in size. Ṣaġānī wrote a text a couple of pages long; Ǧāḥiẓ’s text is in more than six volumes, whereas Damīrī wrote two large volumes. The two books of Ǧāḥiẓ and Damīrī do not exactly fall within our scope. Although the former includes philological comments and a large num- ber of šawāhid,426 it is essentially a literary source which includes anecdotal material related to animals and details of their behavior as well as digressions into sociological, theological and metaphysical matters. Damīrī’s book – which follows full alphabetical arrangement – also contains some philological infor- mation, particularly as relates to the correct vocalization of the names of ani- mals and to the etymology of these names. The philological aspect however forms only a minor part of this encyclopedic work which abounds with anecdotal material related to animals and with digressions, one of which is a lengthy survey of the history of the Caliphs.427 Accordingly, we are left only with relatively short monographs, of which three of the earliest ones have for- tuitously survived and can give us some idea about the content of works from the formative period on three different topics, ibil (camels), šāʾ (sheep), and wuḥūš (wild animals).

K. al-Ibil survived in two versions, which differ both in length428 and arrange- ment of material. Each of the two versions represents a different riwāya, the shorter of which begins with a chain of authority, contrary to the longer one. The existence of the two riwāyas ties in well with the practice by some authors of dictating the same book several times. In fact, Tibrīzī (d. 502/1109) reports that Aṣmaʿī dictated his Ḫalq al-insān fifteen times and that each of the resulting versions differs from the others in the presence or absence of some material.429 The contents of the monograph, in their order of appearance in

426 For Ǧāḥiẓ not having been commonly recognized by his contemporaries or by later authors as a philologist in the strict sense of the word, see Baalbaki (2009: 91–110).

427 Damīrī, Ḥayāt I, 44–96. 428 In Haffner’s edition, the first is in 71 pages and the second in 21. See al-Kanz al-luġawī

fī l-lasan al-ʿArabī (al-Maṭbaʿa al-Kāṯūlīkiyya, Beirut 1903), 66–136, 137–157. For Ḍāmin’s edition, see bibliography.

429 Tibrīzī, Šarḥ I, 76.

 

 

142 chapter 2

the longer version are: the she-camel’s procreation (ḥaml), abundance/defi- ciency in milk (ġazāra/bakʾ), names of groups of camels (i.e. based on their numbers), illnesses, types of walking, colors, periods between two drinkings (aẓmāʾ), types of branding (mawāsim), taznīm (cutting of ear and leaving it hanging), sounds, and adjectives relating to speed. Aṣmaʿī’s two other mono- graphs are comparable in size to the shorter version of al-Ibil. The subjects of al-Šāʾ include procreation and names of the young and adjectives relating to teeth, milk, illnesses, horns, fodder, etc. As for al-Wuḥūš, the wild or wild varieties of animals it deals with are wild asses (ḥamīr), wild cows (baqar), antelopes (ẓibāʾ), ibexes (wuʿūl), ostriches (naʿām), lions (usūd), wolves (ḏiʾāb), hyenas (ḍibāʿ), foxes (ṯaʿālib), and hares (arānib).430 Recorded are the names of the males, females and young of these animals and vocabulary related to their voices, colors, dwelling places, etc. All three monographs are basically lists of words (many of which are of the ġarīb type) that are briefly explained and at times supported by poetry šawāhid.

Mention should also be made of a book titled al-Naʿam wa-l-bahāʾim wa-l- waḥš wa-l-sibāʿ wa-l-ṭayr wa-l-hawāmm wa-ḥašarāt al-arḍ published by M. Bouyges in 1908. It is in fact not an independent book, but part of the multithe- matic work attributed to Ibn Qutayba (d. 276/889), K. al-Ǧarāṯīm.431

K. al-ʿAsal wa-l-naḥl (Honey and Bees) belongs to a topic most of whose monographs have been lost, but some of which are mentioned, among oth- ers, by Ibn al-Nadīm. These include not only works on the very same sub- ject, such as al-Naḥl wa-l-ʿasal by Abū Ḥātim al-Siǧistānī (d. 255/869),432 but also on insects in general or on specific insects, such as al-Ḥašarāt (Insects) by Ibn al-Sikkīt (d. 244/858),433 al-Ḏubāb (Flies) by Ibn al-Aʿrābī (d. 231/845),434 al-Ǧarād (Locusts) by Aḥmad b. Ḥātim al-Bāhilī (d. 231/846) and ʿAlī b. Sulaymān al-Aḫfaš al-Ṣaġīr (d. 315/927),435 and al-ʿAqārib (Scorpions) by Abū ʿUbayda (d. 209/824).436 The manuscript that survived of al-ʿAsal wa-l-naḥl attributes it to Abū ʿUmar al-Zāhid or Ġulām Ṯaʿlab (d. 345/957), but based on its comparison with texts in Ibn Sīda’s al-Muḫaṣṣaṣ and Ibn Manẓūr’s Lisān al-ʿArab, the editor convincingly arrives at the conclusion that its author is

430 Vocabulary related to wuḥūš is also found in Quṭrub’s (d. 206/821) K. al-Farq, for which see below, 149.

431 Ibn Qutayba, Ǧarāṯīm II, 171–290. 432 Ibn al-Nadīm, Fihrist 64. 433 Ibid., 79. 434 Ibid., 76. 435 Ibid., 61, 91. 436 Ibid., 59 (al-ʿUqāb), but see correct reading in Yāqūt, Muʿǧam VI, 2708.

 

 

143MUBAWWAB AND SPECIALIZED LEXICA

Abū Ḥanīfa al-Dīnawarī.437 Further support of this conclusion is given by the fact that, in spite of its inclusion of vocabulary related to honey and bees and of several poetry šawāhid, the monograph is not primarily as philological in nature as one would expect from an author like Ġulām Ṯaʿlab; rather, it focuses on subjects such as the nature of bees, the plants they feed on, the production of honey, illnesses that afflict beehives, etc.

Items 4 and 5 in the above list are similar in that each pertains to a particular animal. The first is Asmāʾ al-asad by Ibn Ḫālawayhi, who begins his monograph by the phrase laysa fī ǧamīʿ kalām al-ʿArab . . . illā.438 This is reminiscent of his style in Laysa fī kalām al-ʿArab, of which it may have constituted a part.439 To kalām al-ʿArab Ibn Ḫālawayhi adds here the books of philology (kutub al-luġa) and asserts that the five hundred or so names and adjectives which his mono- graph contains exhaust all the relevant material to be found in the speech of the Arabs and in philological works. As for Fī asāmī l-ḏiʾb wa-kunāhu, Ṣaġānī lists in it 121 names (asāmī), nineteen epithets or bynames (kunā), ten names and two epithets of the female wolf – a total of 152 lexemes. Despite the few- ness of these lexemes, Ṣaġānī lists them according to their last letters, making this lexicon one of the shortest to be arranged in this manner, and perhaps in any other manner of alphabetical arrangement. The observance of final letters in this short lexicon comes as no surprise from the author of two major lexica that are similarly arranged, namely, al-ʿUbāb al-zāḫir wa-l-lubāb al-fāḫir and al-Takmila wa-l-ḏayl wa-l-ṣila.440

Vocabulary related to animals is also present in most of the multithematic works, such as the above mentioned K. al-Ǧarāṯīm attributed to Ibn Qutayba (d. 276/889) and al-Muntaḫab by Kurāʿ al-Naml (d. 310/922).441 Most worthy of mention, however, are the two works which contain the largest material on animals. The first of these is the earliest extant multithematic work, Abū ʿUbayd’s (d. 224/838) al-Ġarīb al-muṣannaf, which contains the following rel- evant chapters: al-Ḫayl, al-Ṭuyūr wa-l-hawāmm, al-Ibil wa-nuʿūtuhā, al-Ġanam wa-nuʿūtuhā, al-Waḥš, and al-Sibāʿ.442 The same subjects are treated in even more detail by Ibn Sīda (d. 458/1066) in almost two and a half volumes (middle

437 Dīnawarī, ʿAsal 115–118. 438 Ibn Ḫālawayhi, Asmāʾ 8. Note that the word ǧamīʿ does not occur in the introductory

phrases used in Laysa. 439 See above, 93–94. 440 See below, 378, 381. Note however that Ṣaġānī arranges alphabetically other works of his,

namely, Asmā l-ġāda fī asmāʾ al-ʿāda, Aḍdād and Yafʿūl; see below, 160, 197, 253. 441 Kurāʿ, Muntaḫab I, 326–330. 442 Abū ʿ Ubayd, Ġarīb I, 281–291, 319–335; III, 832–894, 895–905, 906–911, 912–923 respectively.

 

 

144 chapter 2

of the sixth till the end of the eighth) of his seventeen-volume encyclopedia, al-Muḫaṣṣaṣ. The chapter on birds (K. al-Ṭayr)443 is particularly interesting because it preserves material that is largely lost to us due to the loss of earlier monographs of the same title.444

Monographs and books on ḫayl are among the most numerous in the lexi- cographical tradition and span the whole length of its history.445 There are also chapters on ḫayl in most multithematic books. As for the first works titled K. al-Ḫayl, they are attributed in the sources to philologists of the period of data collection, that is, authors whose dates of death are mostly in the first half of the third/ninth century. The following five works of that period have survived, and we propose to discuss them before briefly surveying the numerous works of the later period:

1. Ansāb al-ḫayl by Hišām b. Muḥammad Ibn al-Kalbī (d. 204/819 or 206/821). 2. al-Ḫayl by Abū ʿUbayda (d. 209/824). 3. al-Ḫayl by Aṣmaʿī (d. 216/831). 4. Asmāʾ ḫayl al-ʿArab wa-fursānihā by Ibn al-Aʿrābī (d. 231/845). 5. ʿAwn ahl al-ǧihād min al-umarāʾ wa-l-aǧnād by Muḥammad b. Yaʿqūb

al-Ḫuttulī (or al-Ḫuttalī), also known as Ibn Aḫī Ḥizām (d. 250/864).

Among these works, Abū ʿUbayda’s and Aṣmaʿī’s are the most philologically oriented and thus the most representative of the lost works attributed in the sources to the philologists of the early period. Abū ʿUbayda begins his book by citing poetry and Ḥadīṯ which attest to the merits of horses. The philological material immediately follows, with the vocabulary spread over several topics, starting with the names of the various parts of the horse’s body. Other topics include the parts of the body which coincide with names of ṭayr (here, birds or flying insects),446 horse-calls, innate and acquired blemishes, good quali- ties, differences between male and female horses, horses that need no ḍamr (preparation for racing, etc. by making the horse light of flesh), names of some famous horses, qualities which the Arabs like in their horses, colors, spots,

443 Ibn Sīda, Muḫaṣṣaṣ VIII, 124–176. 444 E.g. K. al-Ṭayr by Aḥmad b. Ḥātim al-Bāhilī (d. 231/846) and Abū Ḥātim al-Siǧistānī

(d. 255/869); cf. Ibn al-Nadīm, Fihrist 61, 64. See also Ibn Qutayba’s chapter (Adab 165–169) on hawāmm, ḏubāb and ṣiġār al-ṭayr.

445 On types of horse descriptions and Arabic horse terminology, see Kutasi (2010). 446 E.g. ʿuṣfūr (bird) for a prominent bone in the temple of the horse, and ḏubāb (common

fly) for a black spot in the dark part of its eye; cf. Abū ʿ Ubayda, Ḫayl 46. Note that Haywood (1965: 43) wrongly refers to these as birds which follow horses!

 

 

145MUBAWWAB AND SPECIALIZED LEXICA

gait, types of activity, neighing and finally a selection of poems which describe horses. Although the same topic is at times discussed in various places, there is in the book a clear attempt at internal organization. Abū ʿUbayda often lists at the beginning of his sections – particularly those related to colors447 – the vocabulary each contains and then proceeds to give their meanings. As for Aṣmaʿī’s book, it is considerably smaller than Abū ʿUbayda’s, but largely similar to it in topics. It contains fewer poetry šawāhid, but gives a detailed description of the various parts of the horse’s body from ear to hoof. In general, Aṣmaʿī is more attentive to linguistic details and, unlike his contemporary, had firsthand knowledge of the anatomy of the horse. In a revealing anecdote, Aṣmaʿī reports that in the presence of al-Faḍl b. al-Rabīʿ (d. 208/824), Abū ʿUbayda boasted about his book on ḫayl which consists of fifty volumes (ǧild), whereas Aṣmaʿī’s is one volume only. Regardless of this exaggeration, al-Faḍl ordered both books and a horse to be fetched, and asked each of the two authors to read his book and put his hand at the relevant place in the horse’s body. Abū ʿ Ubayda refused, saying that he was no veterinarian (laysa anā bayṭār [sic]) and that his book contained material he had heard from the Arabs. In contrast, Aṣmaʿī reports that he bared his forearms and legs and energetically reached each part of the horse’s body, named it and recited (i.e. poetry šawāhid) about it (wa-unšid fīhi).448 Aṣmaʿī’s performance won him the horse and he would often ride it and visit Abū ʿUbayda to annoy him. Obviously, Abū ʿUbayda’s breadth of knowledge was no match for Aṣmaʿī’s practical knowledge, considered in this early period to be a valuable asset which gives credence to philological accounts.

The three other books in the above mentioned list differ significantly from Abū ʿUbayda’s and Aṣmaʿī’s. As its title indicates, Ibn al-Kalbī’s Ansāb al-ḫayl deals with pedigrees. The author gives the names of individual horses (with supporting poetry material), starting with Zād al-rākib, said to belong to the Prophets Dāwūd and his son Sulaymān and to be the horse from which all other Arab horses descended.449 Ibn al-Aʿrābī’s Asmāʾ ḫayl al-ʿArab wa-fursānihā also gives the names of famous horses, arranged mainly according to Arab tribes (e.g. ḫayl Qurayš, ḫayl banī Asad), but in some cases based on a particular group (ḫayl al-Anṣār) or place (ḫayl al-Yaman). The names of the owners of these horses are also given and Ibn al-Aʿrābī documents his material by citing poetry which refers to his horses and their owners. As for Ḫuttulī’s (or Ḫuttalī’s) ʿAwn

447 Ibid., 103 ff. 448 Baġdādī, Tārīḫ Baġdād X, 415; cf. XIII, 255–256 and Yāqūt, Muʿǧam VI, 2707–2708. The text

is also in Ibn al-Anbārī, Nuzha 97–98; Qifṭī, Inbāh II, 202. See also Gruendler’s (2011: 20 ff.) interesting comments on this incident and its relevance to book culture before print.

449 Ibn al-Kalbī, Ansāb 14–15.

 

 

146 chapter 2

ahl al-ǧihād min al-umarāʾ wa-l-aǧnād, it includes philological material, partic- ularly the vocabulary related to parts of the body and to color and spots.450 It is not clear whether Ḫuttulī derived this material from earlier sources which he did not mention, or depended on his own knowledge of horses and their anat- omy. Be that as it may, it is clear that Ḫuttulī’s primary interest does not lie in the philological aspects related to his topic (witness the obvious paucity of his poetry šawāhid).451 As an established veterinarian and overseer of the caliphal stables of al-Muʿtaṣim (r. 218/833–227/842) he was an expert in practical mat- ters pertaining to horses, hence his inclusion of topics such as the dressing, feeding and shoeing of horses as well as horsemanship and detailed accounts of the treatment of horses’ illnesses. The veterinary aspect of the work is high- lighted by Ḫuttulī in his introduction as he puts his book on the same footing (yuwāzī) as K. al-Ṣināʿa al-kabīra, Galen’s book on medicine.452

Broadly speaking, works on ḫayl in the first half of the third/ninth century can be divided into those that are purely philological (i.e. Abū ʿUbayda’s and Aṣmaʿī’s), those that focus on individual horses and the poetry in which they are mentioned (i.e. Ibn al-Kalbī’s and Ibn al-Aʿrābī’s), and Ḫuttulī’s book which straddles philology and practical topics, particularly veterinarian ones. Later works are no less varied in their focus and it should be sufficient to make gen- eral observations on the following list which contains the relevant titles, with the exception of those that deal exclusively with veterinary topics:

1. Asmāʾ ḫayl al-ʿArab wa-ansābuhā wa-ḏikr fursānihā by al-Aswad al-Ġundiǧānī (d. 430/1038).

2. Tārīḫ al-ḫuyūl al-ʿArabiyya aw Šarḥ Urǧūza fī ṣifāt al-ḫayl wa-alwānihā wa-mā yuḥmad minhā wa-mā yuḏamm by Aḥmad b. ʿAbdallāh b. Ḥamza al-Ḥamzī (d. 656/1258).

3. al-Ḥalba fī asmāʾ al-ḫayl al-mašhūra fī l-Ǧāhiliyya wa-l-Islām by al-Ṣāḥibī al-Tāǧī (d. after 677/1278).453

4. Faḍl al-ḫayl wa-aǧnāsuhā wa-awṣāfuhā wa-maḥāsinuhā wa-faḍāʾiluhā by ʿAbdalmuʾmin b. Ḫalaf al-Dimyāṭī (d. 705/1306).

5. Ḥilyat al-fursān wa-šiʿār al-šuǧʿān by Ibn Huḏayl al-Andalusī (d. after 763/1361).

450 Ḫuttulī, ʿAwn 43–59, 71–85. 451 Ibid., Index, 190–191. 452 Ibid., 27. 453 See also Ḍāmin (1991b), Mā lam yunšar min al-Ḥalba, and in particular p. 546, where the

editor corrects an earlier mistake concerning Ṣāḥibī’s date of death.

 

 

147MUBAWWAB AND SPECIALIZED LEXICA

6. al-aqwāl al-kāfiya wa-l-fuṣūl al-šāfiya fī l-ḫayl by al-Malik al-Muǧāhid al-Rasūlī (d. 764/1363).

7. al-Ḫayl, Maṭlaʿ al-yumn wa-l-iqbāl fī ntiqāʾ Kitāb al-Iḥtifāl by Ibn Ǧuzayy al-Kalbī al-Ġarnāṭī (eighth/fourteenth century).

8. Qaṭr al-sayl fī amr al-ḫayl by Bulqīnī (d. 805/1403). 9. Maǧrā l-sawābiq by Ibn Ḥiǧǧa al-Ḥamawī (d. 837/1433). 10. Ǧarr al-ḏayl fī ʿilm al-ḫayl by Suyūṭī (d. 911/1505). 11. Fawāʾid al-nayl bi-faḍāʾil al-ḫayl by ʿAlī b. ʿAbdalqādir al-Ṭabarī

(d. 1070/1660).

Some of these works expand the material present in earlier works. For exam- ple, al-Aswad al-Ġundiǧānī’s list of names of horses amounts to 575, com- pared with about two hundred in Ibn al-Aʿrābī’s book, and Dimyāṭī devotes almost his entire book to the prophetic ḥadīṯs which relate to horses. On the other hand, hardly any new topics were introduced, one exception being Ibn Huḏayl’s introduction of the related topic of weapons to his book.454 Material on this topic accounts for about a third of the book, but its six chapters which are placed at its very end, following poetry on ḫayl – with which several books on ḫayl are concluded – obviously stand out as an entity which is only indi- rectly related to the subject of horses. As for arrangement, only al-Aswad al-Ġundiǧānī’s and al-Ṣāḥibī al-Tāǧī’s are alphabetically arranged (based on the first letter only). The fact that these two books list names of horses and do not deal with a variety of subjects probably contributed to this type of arrange- ment. The other authors arrange their books thematically, and even when they assign chapters for names of horses or assemble the vocabulary related to the horse’s anatomy, they do not arrange their material alphabetically. Finally, it comes as no surprise that Suyūṭī’s Ǧarr al-ḏayl is among the most philologi- cally oriented sources of the genre. Not only does it include the traditional topics related to body parts, neighing, colors, blemishes, gait, etc. as well as the mention of ḫayl in Qurʾān, Ḥadīṯ and poetry, but also a chapter on the amṯāl related to horses; an urǧūza of Suyūṭī’s own composition in which he includes in 24 hemistichs the names of birds and insects which are also used for the horse’s body parts; and a maqāma on horses and camels by Badr al-Dīn b. Ḥabīb al-Ḥalabī (d. 779/1377).455 Suyūṭī’s urǧūza seems to be part of a wider tradition,

454 The six chapters are on suyūf (swords), rimāḥ (spears), qisiyy wa-nabl (bows and arrows), durūʿ (coats of defense), tirasa (shields), and weapons and equipment (ʿudda) in general; cf. Ibn Huḏayl, Ḥilya 185–240. See also below, 159.

455 Suyūṭī, Ǧarr 99–102, 80–81, 129–134 respectively; cf. Suyūṭī, Muzhir I, 377–381 for parts of the horse’s body which are names of birds and insects.

 

 

148 chapter 2

which includes, for example, ʿAbdallāh b. Ḥamza b. Sulaymānʾs (d. 614/1217) Urǧūza fī ṣifāt al-ḫayl wa-alwānihā wa-mā yuḥmad minhā wa-mā yuḏamm, on which we have a commentary by the author’s son, Aḥmad b. ʿAbdallāh b. Ḥamza al-Ḥamzī (d. 656/1258).456 Furthermore, Suyūṭī incorporates into his book other material drawn from earlier authors, including literary texts on horses, such as Maǧrā l-sawābiq by Ibn Ḥiǧǧa and a chapter from a risāla by Ibn Nubāta (d. 768/1366).457 Also included is a chapter in which Suyūṭī lists words which relate to horses and which are mentioned by Fīrūzābādī (d. 817/1415) in al-Qāmūs al-muḥīṭ.458

4.3 Human Body This genre is strongly related to the one on animals, and most evidently so in monographs which deal with the difference between human beings and animals and are usually titled K. al-Farq. The other monographs of the genre mostly carry the title K. Ḫalq al-insān and focus on human body parts, but also include material pertaining to some of man’s physical (and at times behav- ioral) aspects.459 The earliest authors of both types of monographs are phi- lologists of the first half of the third/ninth century, several of whose works did survive (see list below). Other philologists of the same era whose works did not reach us include al-Naḍr b. Šumayl (d. 203/819), Abū ʿAmr al-Šaybānī (d. 206/821), Abū ʿUbayda (d. 209/824), Abū Zayd al-Anṣārī (d. 215/830), Saʿdān b. al-Mubārak al-Ḍarīr (d. 220/835), Ibn al- Aʿrābī (d. 231/845), and Ibn al-Sikkīt (d. 244/858). The fact that most of these authors were active in collecting lin- guistic data from the Bedouins explains, judging by the extant works, the high rate of ġarīb words in this genre, as is also the case in monographs on plants and animals. Moreover, one of the Bedouin fuṣaḥāʾ, Abū Mālik ʿAmr b. Kirkira (d. ?), on whose authority ġarīb material is reported by some early philolo- gists, is himself the author of a book titled Ḫalq al-insān,460 and is frequently quoted by Ṯābit b. Abī Ṯābit in his book of the same title.461 Also noteworthy is that in works which deal with farq, it is implied that the existence of special- ized terms used exclusively for human beings or for particular animals (such as horses, camels, beasts of prey, birds, etc.) is a reflection of the accuracy of Arabic vocabulary – a general characteristic of the speech of the Arabs often

456 Ḥamzī, Tārīḫ 33–61. 457 Suyūṭī, Ǧarr 135–147, 150–154. 458 Ibid., 119–121. 459 More details pertaining to the genre are found in Saṭl (1976, 1998) and Ṣāliḥ (1989). 460 Ibn al-Nadīm, Fihrist 49. 461 Ṯābit, Ḫalq 46–48, 51, 70, 81, 92, etc.

 

 

149MUBAWWAB AND SPECIALIZED LEXICA

highlighted by later philologists.462 A good example is that of the terms which mean “nose” in various creatures (see the discussion of Quṭrub’s book below). Authors of the early period surely wanted to preserve these specialized terms and provide šawāhid (mainly from poetry) which testify to their accurate meanings in actual usage.

We shall begin by discussing the following extant works of the formative period and then trace the developments that took place in works from the fourth/tenth century onward:

1. al-Farq (also known as Mā ḫālafa fīhi l-insān al-bahīma min farqihi/ qarnihi ilā qadamihi) by Quṭrub (d. 206/821).

2. al-Farq and Ḫalq al-insān by Aṣmaʿī (d. 216/831). 3. Ḫalq al-insān fī l-luġa by Muḥammad b. Ḥabīb (d. 245/860). 4. al-Farq by Abū Ḥātim al-Siǧistānī (d. 255/869). 5. al-Farq and Ḫalq al-insān by Ṯābit b. Abī Ṯābit (d. around the middle of

the third/ninth century).

Quṭrub deals with the following six themes: (1) parts of the body, secre- tions (e.g. luʿāb “saliva”), certain functions (e.g. ǧulūs “sitting”) and sexual drive (šahwa); (2) pregnancy, procreation and names of offspring; (3) groups (ǧamāʿa); (4) voice; (5) crying out to urge, restrain, chide, etc. (zaǧr); and (6) death. A typical theme (or section in a theme) begins by citing the words that are used for human beings, followed consecutively by those used for ani- mals with ḥāfir (hoof; e.g. horses), ḫuff (foot [of camels]), ẓilf (cloven hoof; e.g. sheep), barāṯin (claws; e.g. dogs), and ǧanāḥ (wing; e.g. birds and flies). For example, the terms corresponding to anf, ḫaṭm, ḫurṭūm, ʿirnīn and marsin used for the human nose are a donkey’s nuḫara, a camel’s ḫurṭūm and ḫaṭm, a cow’s and a pig’s finṭīsa, a dog’s harṯama and ʿawtama, and a pigeon’s qirṭimatān.463 In certain cases, further divisions are made of one group of animals based on specific terminology, such as the division of cloven-hoofed animals into ḍa⁠ʾn (sheep), ʿanz (she-goats), šāʾ al-waḥš (wild cows, antelopes and ibexes), etc.464

462 See, for instance, Ibn Ǧinnī’s chapter on the subtle difference in meaning between words whose forms are only slightly different (Ḫaṣāʾiṣ II, 145–152) and Ibn Fāris’s chapter in which he cites words which are correctly used only if certain semantic conditions are met (Ṣāḥibī 98–99). On a larger scale, ʿAskarī’s al-Furūq al-luġawiyya, for example, highlights subtle semantic differences between words that are often used interchangeably (cf. below, 210).

463 Quṭrub, Farq 47–48. 464 Ibid., 104–113.

 

 

150 chapter 2

In his K. al-Farq, Aṣmaʿī treats the same themes as in Quṭrub’s book, except for the sixth, i.e. death. Both authors often include poetry šawāhid and cite certain morphological forms, such as imperfect verbs, verbal nouns, feminine and plural nouns, etc. Yet, Aṣmaʿī’s book is considerably shorter than Quṭrub’s (about a third of it) and lacks its internal arrangement, that is, the specific order which Quṭrub observes to compare words that apply to human beings with those pertaining first to animals with ḥāfir, then those with ḫuff, and so on.

The two other works on farq in the above list are Abū Ḥātim’s and Ṯābit’s. In fact, the former is almost certainly a variant riwāya of Aṣmaʿī’s book. It consists of the same twenty-eight chapters in Aṣmaʿī’s, in exactly the same order. The minor differences between the two texts, as shown in the comparative lists prepared by M. Ḥ. ʿAwwād,465 are due to the existence of two slightly differ- ent riwāyas and thus invalidate the arguments presented by the book’s editor, Ḥ. Ṣ al-Ḍāmin, concerning its attribution.466 As for Ṯābit’s book, much of its material is ascribed to Aṣmaʿī, and its themes largely correspond to his. Yet, it is not merely a riwāya of Aṣmaʿī’s book since the themes in each are differently arranged and since Ṯābit includes lexemes and šawāhid not cited by Aṣmaʿī as well as a few themes – e.g. death (mawt) and speeding and running (surʿa and ʿadw)467 – which Aṣmaʿī ignores.

The first extant work on ḫalq al-insān is Aṣmaʿī’s. The fact that Aṣmaʿī (and Ṯabīt a few decades later) authored two separate books, one on this subject and another on the related subject of farq, indicates that the tradition of separat- ing the two domains goes back to the earliest stages of lexical writing. On the other hand, it should be noted that at times the two types share some of their material: books on farq necessarily contain lexical items related to humans, particularly concerning body parts, and books on ḫalq al-insān contain spo- radic reference to differences in terminology which apply to man and beast. Aṣmaʿī for example indicates in Ḫalq al-insān that maḫāḍ (parturition) is used for both, but ṭalq (throes of childbirth) only for humans.468 As far as content is concerned, Aṣmaʿī’s book may be divided into three sections: an introduc- tory one, another that details body parts, and a concluding chapter. The intro- ductory section deals with pregnancy, procreation and the different stages of human life from birth to old age, and the concluding chapter lists words related to various bodily and behavioral traits. The bulk of the book is devoted to body parts, and the first chapter in this section includes terminology related

465 ʿAwwād (1992: 350–372). 466 Abū Ḥātim, Farq 11–13. 467 Ṯābit, Farq 93–102. 468 Aṣmaʿī, Ḫalq 229.

 

 

151MUBAWWAB AND SPECIALIZED LEXICA

to the body as a whole, such as šaḫṣ, ṭalal and āl (figure).469 Each of the chap- ters that follow focuses on a single organ or group of organs, starting with the head, scalp, ears, hair, beard, face, forehead, eyebrows, etc., and ending with genital organs, hips, thighs, knees, legs and feet. As in his previously mentioned K. al-Ḫayl, Aṣmaʿī demonstrates impressive knowledge of the vocabulary related to body parts, a large portion of which may be ascribed to material he collected from the Bedouin fuṣaḥāʾ.

Unlike Aṣmaʿī, Muḥammad b. Ḥabīb adopts in his short monograph Ḫalq al-insān fī l-luġa an alphabetical arrangement which does not go beyond the first letter. This arrangement (probably unprecedented in this genre) is cer- tainly useful for the one who seeks the meaning of a specific word, but it natu- rally abolishes the concept of listing together related lexical items, as is the case in Aṣmaʿī’s thematic classification. In most cases, the items are either cited without explanation or are very briefly explained. More on Ibn Ḥabīb’s book will follow in this section when we discuss a much later work which is based on it, namely, Ġazzī’s (d. 984/1577) Ḏikr aʿḍāʾ al-insān.

Ṯābit b. Abī Ṯābit’s Ḫalq al-insān, as far as we know, is the most complete work of the genre. In the introduction, Ṯābit mentions that his material is derived from Abū ʿUbayd, ʿAlī b. Muġīra al-Aṯram (d. 232/846), Salama b. ʿĀṣim (d. after 270/883), Abū Naṣr al-Bāhilī (d. 231/846), Ibn al-Aʿrābī, Aṣmaʿī and Abū Zayd al-Anṣārī. In his own book, Ṯābit puts together and summarizes (ǧamaʿnā ḏālika wa-laḫḫaṣnāhu) the material available in his predecessors’ works and complements it with new material.470 Indeed, his work is lengthier and includes more poetry šawāhid than Aṣmaʿī’s, but is quite similar to it in themes and their arrangement. The fact that Ṯābit quotes Aṣmaʿī more than any other of his predecessors,471 as do other authors as late as Ġazzī in Ḏikr aʿḍāʾ al-insān,472 demonstrates that Aṣmaʿī’s influence on the genre exceeds that of any of his contemporaries.

Material on ḫalq al-insān is also found in a multithematic book of the third/ ninth century, Abū ʿUbayd’s (d. 224/838) al-Ġarīb al-muṣannaf. Although Abū ʿUbayd does not systematically list body parts as does Aṣmaʿī and other authors, his first chapter includes the basic vocabulary related to human organs, but only in a cursory manner.473 His second chapter lists several adjectives (mostly

469 Ibid., 163. 470 Ṯābit, Ḫalq 1. 471 Cf. ibid., Index, 367–373. 472 Ġazzī, Ḏikr, Index, 168–169. 473 Abū ʿUbayd, Ġarīb I, 29–46.

 

 

152 chapter 2

of the pattern afʿal, e.g. afḥaǧ “bowlegged”) which relate to ḫalq al-insān.474 Most of the book’s other chapters on human beings, however, focus on vocabu- lary related to mental, social and behavioral aspects (e.g. intelligence, courage, insanity, gluttony, beauty and ugliness, social groups, guests, servants, etc.). This vocabulary was part of the raw material of which later authors of muǧannas lexica availed themselves in collecting their data. Abundant material related to ḫalq al-insān is also found in Kurāʿ al-Naml’s (d. 310/922) al-Muntaḫab.475

After the third/ninth century, philologists continued to compile books on the subjects of farq and ḫalq al-insān, with a clear inclination towards the lat- ter. In fact, of the former type we have only one extant work, K. al-Farq by Ibn Fāris (d. 395/1004). Although it is almost two centuries later than books of the same title discussed above, it adds little to their material (e.g. names of male and female animals),476 and does not even match Quṭrub’s book as far as con- sistency and arrangement are concerned. We are thus left with the following works on ḫalq al-insān:

1. Ḫalq al-insān by Zaǧǧāǧ (d. 311/923). 2. Ġarāʾib ḫalq al-insān by Ibn Ḫālawayhi (d. 370/980). 3. Maqāla fī asmāʾ aʿḍāʾ al-insān and Istiʿārat aʿḍāʾ al-insān by Ibn Fāris

(d. 395/1004). 4. Kitāb fīhi ḏikr šayʾ min al-ḥilā by al-Qazzāz al-Qayrawānī (d. 412/1021). 5. Ḫalq al-insān by Iskāfī (d. 420/1029). 6. Ḫalq al-insān by Saʿīd b. Hibatallāh b. al-Ḥusayn (d. 495/1101). 7. Ḫalq al-insān fī l-luġa by al-Ḥasan b. Aḥmad b. ʿAbdalraḥmān (sixth/

twelfth century?). 8. Ġāyat al-iḥsān fī ḫalq al-insān by Suyūṭī (d. 911/1505). 9. Ḏikr aʿḍāʾ al-insān by Badr al-Dīn al-Ġazzī (d. 984/1577).

The method established by philologists such as Aṣmaʿī with regard to ḫalq al-insān clearly influenced later authors. For example, the two books by Zaǧǧāǧ and Iskāfī greatly resemble Aṣmaʿī’s with minor differences in content and order of themes. Thus, Zaǧǧāǧ includes a chapter on ist (anus)477 – which does not feature in Aṣmaʿī’s book – but unlike Aṣmaʿī, he does not deal with preg- nancy and procreation. As far as Iskāfī is concerned, his chapter on pregnancy

474 Ibid., I, 46–52. 475 Kurāʿ, Muntaḫab I, 46 ff. 476 Ibn Fāris, Farq 94–98. 477 Zaǧǧāǧ, Ḫalq 46.

 

 

153MUBAWWAB AND SPECIALIZED LEXICA

and procreation is his last, i.e. forty-ninth chapter,478 whereas it is the first in Aṣmaʿī. Suyūṭī also arranges his book according to body parts (including a chapter on ist),479 but innovatively divides most of his chapters into two sec- tions, one for names (asmāʾ) and another for adjectives (ṣifāt). Later authors did occasionally add to the corpus certain lexemes that had been ignored by their predecessors, but were heavily dependent on them for the bulk of their vocabulary. This applies not only to authors such as Zaǧǧāǧ, Iskāfī and Suyūṭī, whose approach resembles that of the philologists of the formative period, but also to those who introduced certain changes to the earlier approach.

From a developmental point of view, one can trace in the above listed works several attempts, either at focusing on a specific aspect of the vocab- ulary related to ḫalq al-insān, or at approaching the subject matter from a new perspective. Accordingly, Ibn Ḫālawayhi focuses in his short monograph Ġarāʾib ḫalq al-insān on strange and rare usage (ġarāʾib ḫalq al-insān wa-ṭarāʾif al-alfāẓ)480 – a telling example of the continuous concern for ġarīb, nādir, etc. well beyond the era of data collection. His list includes several lexical items not mentioned in Aṣmaʿī’s book, such as kuṯḥuma (thick, said of a beard), ḥafallaǧ (bowlegged), ġunduba (hard flesh around the fauces), qimʿāla (head of the penis), waġf (poor sight), ǧuʿšūš (tall), etc. The focus in another mono- graph, Ibn Fāris’s Istiʿārat aʿḍa⁠ʾ al-insān, is on the figurative use of words used originally for about a hundred body parts. Examples include the use of ǧabha (forehead) to mean horses or a group of people, of anf (nose) to refer to the projecting part of a mountain or to the most vehement cold, and of ẓufr (nail) to apply to the end of the bow (beyond the place where the string is tied).481 Ibn Fāris’s attempt is obviously related to his method of grouping together in his muǧannas dictionary, Maqāyīs al-luġa, words that belong to the same aṣl within the same root.482 Indeed, all the figurative usages in the three examples mentioned above appear in the Maqāyīs, and in each case Ibn Fāris argues that they belong to the same aṣl which also embraces the meaning of the relevant body part.483

Several authors introduced new dimensions to the genre by adopting an innovative approach to its huge corpus of lexemes. Among these is Ibn Fāris

478 Iskāfī, Ḫalq 173–176. 479 Suyūṭī, Ġāya 291–293. 480 Ibn Ḫālawayhi, Ġarāʾib 143. The monograph is part of the author’s Laysa but its text does

not feature in ʿAṭṭār’s edition of the latter. 481 Ibn Fāris, Istiʿāra 88, 91, 99. 482 See below, 350 ff. 483 Ibn Fāris, Maqāyīs I, 503, 147; III, 466 respectively.

 

 

154 chapter 2

in his short monograph titled Maqāla fī asmāʾ aʿḍāʾ al-insān, in which he views with dismay the fact that some of those who thoroughly study strange or unfa- miliar usage (man taʿammaqa fī ġarīb al-kalām wa-waḥšiyyihi) are not compe- tent enough to name their own body parts if they want to indicate a feeling of pain and thus would point these parts by hand instead of uttering their names.484 This observation demonstrates the bookish nature of ġarīb mate- rial and its remoteness from the practical needs of the speaker. Ibn Fāris lists the names of the major body parts and their descriptive epithets, from the head downwards as in earlier sources, but evidently avoids ġarīb. He is thus compelled to ignore a large number of lexemes, primarily those which refer to the minor parts that only feature in a detailed study of human anatomy or are highly specialized adjectives pertaining to things such as colors, illnesses, defects, etc. The practical aspect which prompted Ibn Fāris to write his Maqāla was also the motive behind al-Qazzāz al-Qayrawānīʾs authoring of a text titled Kitāb fīhi ḏikr šayʾ min al-ḥilā (ḥilā being the plural of ḥilya, a synonym of ḫilqa “physical constitution”). Dissatisfied with writers’ conventional description of slaves (ʿādat al-kuttāb fī l-ḥilya fī l-raqīq), al-Qazzāz proposes a model descrip- tion to be emulated by traders (mutabāyiʿūn) and soldiers (ǧund).485 For a fuller description (kamāl al-taḥliya), he starts with the more general traits, such as sex (e.g. ġulām/ǧāriya “male/female slave”), ethnic origin, and char- acteristics regarding color, age and build. Following this introductory part is the main body of the text, which consists of eighteen brief sections in each of which are adjectives related to a particular part of the body, from head (ra⁠ʾs) to feet (qadamān). The last two chapters deal with innate and newly existing defects (al-ʿuyūb al-lāzima/al-ḥādiṯa).486 Although the lexical items mentioned in these two chapters are found in previous works, grouping them together, rather than citing them separately in the relevant sections on body parts, is unprecedented as far as we know. Finally, a strong medical component is introduced to the study of ḫalq al-insān by the famous Baghdadi physician Saʿīd b. Hibatallāh b. al-Ḥusayn.487 His book, however, does not strictly belong to the genre although it includes the names of the major body parts as well as some interesting technical terms related to them – such as iftiḍāḍ (devirgin- ation), ālāt (internal organs), and hayʾa (dissection) – or to certain illnesses – such as istisqāʾ (dropsy), riʿda (chill), ṣarʿ (epilepsy), and kalaf (freckles).488 It is

484 Ibn Fāris, Maqāla 110. 485 Qazzāz, Ḥilā 3. 486 Ibid., 20–22. 487 For Saʿīd’s biography, see Ibn Abī Uṣaybiʿa, ʿUyūn II, 324 and Ṣafadī, Wāfī XV, 268–269. 488 Saʿīd b. Hibatallāh, Ḫalq, Index, 253–258.

 

 

155MUBAWWAB AND SPECIALIZED LEXICA

essentially a medical text of fifty chapters,489 which primarily deals with sexual organs, intercourse, pregnancy, fetal characteristics, nursing, raising children, etc. as well as with some philosophical and metaphysical issues, such as the difference between rūḥ and nafs (both interchangeably used for soul, spirit) and reincarnation (tanāsuḫ).

Judging by the extant works, the thematic arrangement of material in the genre of ḫalq al-insān is considerably more dominant than the alphabeti- cal model. Among the works discussed above, only Muḥammad b. Ḥabīb’s is alphabetically arranged. The same model, which does not go beyond the first letter of the word, is also adopted by two authors from a much later period, namely, al-Ḥasan b. Aḥmad b. ʿAbdalraḥmān in Ḫalq al-insān fī l-luġa and Ġazzī in Ḏikr aʿḍāʾ al-insān. The former explains in his introduction that his lexicon includes names of body organs and other components of the body (such as blood and brain) but not epithets in which people vary (e.g. tallness and short- ness, and nose shape) or symptoms (e.g. blindness, lameness and baldness).490 His philological interest is reflected in the inclusion in his plan of the citing of morphological patterns (e.g. masculine and feminine forms), syntactical issues related to the function of words in structure (i.e. iʿrāb), and poetry šawāhid. For this material, he seems to have at least partially depended on Ṯābit b. Abī Ṯābit, as the editor’s textual comparison shows,491 although al-Ḥasan himself does not reveal his sources. It is quite interesting that a note appended to Ṯābit’s published manuscript by its copyist (and possibly preserved in other manu- scripts) might have caught al-Ḥasan’s attention, in which case it would fur- nish further evidence for his familiarity with Ṯābit’s book. The note mentions that the number of body organs which begin with the letter kāf is thirteen, and lists them with an additional term which applies to women only (namely, kaʿṯab, “a woman with large pubes or pudendum”).492 In the first few lines of the introduction,493 al-Ḥasan explains that the inability of some learned individuals (muta⁠ʾaddibūn) to cite these words which begin with kāf if asked to do so prompted him to author his lexicon and to arrange it alphabetically. An innovative aspect of the arrangement of al-Ḥasan’s book is that, wherever applicable, a chapter starts by lexemes which apply to men and women alike, followed by two, also alphabetically arranged sections, which list lexemes per- taining either exclusively to men or to women. Also noteworthy is that roots

489 See the author’s own listing of the titles of these chapters; ibid., 15–18. 490 al-Ḥasan b. Aḥmad, Ḫalq 47–48. 491 Ibid., 34–38. 492 Ṯābit, Ḫalq 331. 493 al-Ḥasan b. Aḥmad, Ḫalq 47.

 

 

156 chapter 2

have no role in the arrangement, hence aǧlād, taǧālīd and muḥayyā appear under alif, tāʾ and mīm respectively.

Ġazzī’s book is interesting because it is based on Muḥammad b. Ḥabīb’s which precedes it by more than seven centuries and which is the first alpha- betically arranged work of the genre that we know. Each chapter (except for yāʾ which Ibn Ḥabīb’s book does not include)494 starts with the lexical items men- tioned by Ibn Ḥabīb, followed by those which Ġazzī supplements. The num- ber of Ibn Ḥabīb’s lexical items is 266, to which Ġazzī adds 497, bringing the total to 763.495 To explain both groups, Ġazzī mostly depends on two famous muǧannas lexica, namely, Ǧawharī’s (d. c. 400/1010) al-Ṣaḥāḥ and Fīrūzābādī’s (d. 817/1415) al-Qāmūs.496 With the exception of Ṣaġānī’s (d. 650/1252) Ḫalq al-insān, which he cites only once,497 Ġazzī mentions no earlier works of the genre, be they thematically or alphabetically arranged.

4.4 Miscellanea Words related to various other topics were recorded with their šawāhid as part of the data collection which took place as of the second half of the sec- ond/eighth century. Fortunately, a few monographs on some of these topics have come into our possession, although in the case of several topics only one specimen has survived out of many that are mentioned in the bibliographical and biographical sources. Monographs on various subjects related to natural phenomena feature prominently in this miscellany. We shall therefore discuss these first before turning to the other topics.

Within the general topic of natural phenomena, it is possible to distinguish between works which focus primarily, if not exclusively, on philological mat- ters and those which straddle philology, aḫbār (anecdotal material), and the science of the age. The following titles belong to the first group:

1. al-azmina wa-talbiyat al-Ǧāhiliyya by Quṭrub (d. 206/821). 2. al-Ayyām wa-l-layālī wa-l-šuhūr by Farrāʾ (d. 207/822). 3. al-Maṭar by Abū Zayd al-Anṣārī (d. 215/830). 4. Ṣifat al-saḥāb wa-l-ġayṯ wa-aḫbār al-ruwwād wa-mā ḥamidū min al-kala⁠ʾ

by Ibn Durayd (d. 321/933). 5. Risāla fī asmāʾ al-rīḥ by Ibn Ḫālawayhi (d. 370/980).

494 Ġazzī, Ḏikr 152 (wa-lam yataʿarraḍ Ibn Ḥabīb li-l-yāʾ bi-l-kulliyya); cf. Ibn Ḥabīb, Ḫalq 60, where wāw is the book’s last chapter.

495 Ibid., 13. According to my enumeration, Ibn Ḥabīb’s items are 270. 496 Ibid., Index, 174. 497 Ibid., 129.

 

 

157MUBAWWAB AND SPECIALIZED LEXICA

The first three authors are among the most famous in the formative period and among those most quoted in the muǧannas lexica. Their writings on natural phenomena may be viewed as part of the general effort to document kalām al-ʿArab, particularly in the speech of the Bedouin fuṣaḥāʾ and in Ǧāhilī and early Islamic poetry. Quṭrub’s book, for example, surveys and explains lexical items related to the sun, moon, nights of the month, stars, day and night, heat and cold, shade, etc., in addition to names of days, months and years. It also includes a chapter in which the talbiya – i.e., the crying of labbayka (lit. here I am! Or at your service!) particularly in pilgrimage – of many tribes is cited.498 Although the material of this chapter is unrelated to the book’s theme, its inclusion immediately following the mention of the month of Ḏū l-ḥiǧǧa (during which pilgrimage is performed) is obviously the reason behind this digression.499 Quṭrub’s philological approach is evident in his frequent com- ments related to masculine, feminine and plural patterns, morphophonologi- cal processes, aḍdād (words with two contradictory meanings), etc., but above all to etymology, as for instance in explaining the names of months. His text is replete with poetry šawāhid (64 from poetry and 54 from raǧaz), in addition to 27 Qurʾānic verses.500 The same philological matters, such as the names of days and months and epithets related to day and night, years, the sun and the moon, etc., are also discussed in a shorter text by Farrāʾ. In line with his interest in Qurʾānic exegesis as we know from Maʿānī l-Qurʾān, Farrāʾ’s book includes a short chapter in which he explains several verses related to days and months.501 As for Abū Zayd’s monograph, K. al-Maṭar, it is a collection of lexical items related to rain, seasons, heat, thunder, lightning, clouds, rivers, water, and wells. Its šawāhid are fewer than those in Quṭrub and Farrāʾ, but it is equally philological in nature and seems to be based on actual usage which Abū Zayd heard from tribal sources, particularly Qaysiyyūn, ʿAnbariyyūn and Kilābiyyūn.502

From the fourth/tenth century we have two texts of purely philological ori- entation. The first, by Ibn Durayd, is a monograph on cloud and rain, and is unique in that it is essentially a collection of quotations derived from seek- ers of herbage (ruwwād), each of which is followed by an explanation of its ġarīb material.503 With the exception of one quotation ascribed to Prophet

498 Quṭrub, Azmina 54–61. 499 Note that the book’s title reflects the fact that the talbiya is distinct from the main theme

of azmina. 500 Ibid., Introduction 14. 501 Farrāʾ, Ayyām 52–54. 502 Abū Zayd, Maṭar 100, 103, 111. 503 Note the expression tafsīr ġarībihi; Ibn Durayd, Saḥāb 24.

 

 

158 chapter 2

Muḥammad and another to the daughter of the Yemeni poet Muʿaqqir b. Ḥimār (d. 45 before Hiǧra?/580?), all other quotations are attributed to Aʿrāb, and the vast majority of these go back to Aṣmaʿī on the authority either of his nephew ʿAbdalraḥmān or of Abū Ḥātim, or both.504 In this respect, Aṣmaʿī’s influence on later authors in connection to the riwāya of ġarīb is apparent, as is the case in other genres of writing, such as plants and animals. The second text, also a short monograph, is Ibn Ḫālawayhi’s Risāla fī asmāʾ al-rīḥ. In the first part of the risāla,505 which serves as an introduction, the author explores the various meanings of the word rīḥ and elucidates its meaning in several Qurʾānic sūras and ḥadīṯs. The second part consists of consecutively cited lexical items used as names for winds (asmāʾ al-riyāḥ), and although most of these belong to the realm of ġarīb, Ibn Ḫālawayhi only rarely gives their meaning.

Mention can also be made of a short literary text by Ibn Fāris (d. 395/1004) titled al-Layl wa-l-nahār. It is in the form of a conversation between two inter- locutors, one of whom exposes the merits and virtues of night, and the other of day. As one would expect from a philologist such as Ibn Fāris, he often intro- duces philology to the arguments which the interlocutors forward by citing the Arabs’ use of lexical items related to day and night.506

Several other sources approach natural phenomena from a wider perspec- tive which is not restricted to philology. These works, which comprise those on anwāʾ (i.e. the acronychal setting and helical rising of stars and constella- tions and their influence on the winds and the rains, etc.),507 generally include rather detailed information about the sun, the mansions of the moon, stars, and constellations, and at times refer to the contribution of other nations, such as the Rūm, the Suryāniyyūn, the ʿIbrāniyyūn, the ʿAǧam and the Qibṭ,508 to this domain. They even delve into philosophical issues, such as the essence of time.509 To various degrees, these works do share the material available in the philologically oriented monographs and books, as most of their authors, such as Ibn Qutayba and Ibn al-Aǧdābī, made significant contributions to phi- lology. Most representative of this group of works are:510

504 Texts ascribed to Bedouin ruwwād are also reported on Aṣmaʿī’s authority in Marzūqī’s Azmina II, 127–131. Note also the famous text on ġayṯ attributed to ʿĀmir b. Šarāḥīl al-Šaʿbī (d. 103/721); for works which quote it or explain its content, see Faḥḥām (1983: 31, n. 2).

505 Ibn Ḫālawayhi, Rīḥ 291–296. 506 Ibn Fāris, Layl 23–24, 33. 507 Cf. Pellat (1986a: 523). 508 Ibn al-Aǧdābī, Azmina, Index, 166–167. 509 Marzūqī, Azmina I, 126–132. 510 For works on anwāʾ which have not survived, see the list of Iqbāl (2011: 120–126).

 

 

159MUBAWWAB AND SPECIALIZED LEXICA

1. al-Anwāʾ fī mawāsim al-ʿArab by Ibn Qutayba (d. 276/889). 2. al-Maṭar wa-l-raʿd wa-l-barq wa-l-rīḥ by Ibn Abī l-Dunyā (d. 281/894). 3. al-Anwāʾ wa-l-azmina wa-maʿrifat aʿyān al-kawākib fī l-nuǧūm by Ibn

ʿĀṣim (d. 403/1013). 4. al-Azmina wa-l-amkina by Marzūqī (d. 421/1030). 5. al-Azmina wa-l-anwāʾ by Ibn al-Aǧdābī (d. c. 600/1252).

Apart from natural phenomena, there are several works (mostly short mono- graphs) which deal with a variety of topics. Most of these go back to the first half of the third/ninth century, after which the writing of monographs on spe- cific themes largely retreated in favor of multithematic works. Among these early works is K. al-Dawāhī whose text, in a rare instance in the whole tradi- tion, incorporates, by the copyist’s own assertion,511 two books. One of these is by Abū ʿUbayda (d. 209/824) and the other is by Muḥammad b. al-Ḥasan b. Dīnār al-Hāšimī, known as al-Aḥwal (d. between 281/894 and 291/904). The resulting text includes 188 names and epithets of dawāhī,512 pl. of dāhiya (calamity, misfortune), a number falling short of the four hundred lexical items which are claimed to be used for dawāhī and which gave rise to the satirical saying takāṯur asmāʾ al-dawāhī min al-dawāhī (The enormity of the names of calamities is in itself a calamity).513 The short monograph by Abū Zayd al-Anṣārī (d. 215/830) titled K. al-Liba⁠ʾ wa-l-laban (i.e beestings and milk) explains lexical items related to a she-camel’s first milk at the time of giving birth and to other types of milk and their descriptive epithets. On the subject of weapons, we possess K. al-Silāḥ by Aṣmaʿī (d. 216/831). As for Abū ʿUbayd’s (d. 224/838) independently published monograph K. al-Silāḥ, it is in fact part of his al-Ġarīb al-muṣannaf.514 Although there are sections unique to each (e.g. Aṣmaʿī’s chapter on katāʾib “battalions”515 and Abū ʿ Ubayd’s on sikkīn “knife”),516 the two works have most of their material and order of subjects in common. Both, for example, deal with suyūf (swords), rimāḥ (spears), qisiyy (bows),

511 Abū ʿUbayda & Ibn Dīnār, Dawāhī 67. The copyist admits that he started by copying Ibn Dīnār’s book thinking that it was Abū ʿUbayda’s, hence his use of the latter text to check the content of the former. He later discovered that the two texts were different and decided to combine the bulk of both in one book.

512 This enumeration excludes the lexical items listed under the three headings ʿaǧab (wonder; 4 items), sabb (cursing; 7 items), and kaḏib (lying; 37 items); cf. ibid., 60–66. On dawāhī, see also Weipert (2004).

513 Cf. Suyūṭī, Muzhir I, 325. 514 Abū ʿUbayd, Ġarīb I, 293–318. 515 Aṣmaʿī, Silāḥ 109–112. 516 Abū ʿUbayd, Silāḥ 39–40 (= Ġarīb I, 317–318).

 

 

160 chapter 2

sihām (arrows), durūʿ (coats of defense), tirasa (shields), ǧiʿāb (quivers), etc. Finally, Ibn al-Aʿrābī (d. 231/845) gathers in his K. al-Biʾr vocabulary related to wells, in particular their digging and drawing water from them, as well as their names and descriptive epithets.

From the second part of the third/ninth century, we have several works that are either attributed to Ibn Qutayba (d. 276/889) or are genuinely his. One of these, al-Raḥl wa-l-manzil, deals with a camel’s saddle and with abodes, but its attribution to Ibn Qutayba is incorrect since the text is a chapter in K. al-Ǧarāṯīm, itself erroneously attributed to him.517 Of doubtful attribution to him also, based on comparison with quotations in later sources from his own K. Ālat al-kuttāb,518 is Risālat al-ḫaṭṭ wa-l-qalam, which includes vocabulary on writing tools – e.g. qalam (pen), dawāt (inkwell), līqa (tuft of threads inserted in inkwell), midād (ink), etc. – and on books, writing style, paper, reading, etc. As for K. al-Ašriba wa-ḏikr iḫtilāf al-nās fīhā and K. al-Maysir wa-l-qidāḥ, they are both authored by Ibn Qutayba. The former, apart from few philological comments (e.g. the meaning of ḫamr “grape-wine” and nabīḏ “beverage made of raisins or dates”),519 deals with other matters such as the arguments for the legalization or prohibition of wine and the benefits and drawbacks of drinking alcohol. In contrast, vocabulary has a more central part in the book on may- sir (gambling) and qidāḥ (rods used in gambling). Thus, in addition to details concerning the game of maysir, the book for the most part explains the lexical items related to gambling, the ten rods, and the shares into which the ǧazūr (slaughtered camel) is divided for the game.

From the fourth/tenth century we have a unique monograph devoted to vocabulary related to the saddle and bridle. This is Ibn Durayd’s (d. 321/933) Ṣifat al-sarǧ wa-l-liǧām. We know from other works, such as Aṣmaʿī’s K. al-Silāḥ,520 that this is one of the themes in which philologists were inter- ested as of the period of data collection, but Ibn Durayd’s monograph is the only surviving work of the genre. It abounds with philological comments on morphological forms and patterns, morphophonological change, derivation, usage by the ʿāmma (common people), etc.

As mentioned earlier, the tradition of writing monographs on specific themes, particularly in the third/ninth century, was largely abandoned in favor of multithematic works. Yet, one still finds such monographs from as late as the seventh/thirteenth century. One such monograph is Asmā l-ġāda fī asmāʾ

517 See below, 271. 518 Ibn Qutayba, Ḫaṭṭ, Introduction 3–8. 519 Ibn Qutayba, Ašriba 29–34. 520 Aṣmaʿī, Silāḥ 113–116 (sarǧ); 116–118 (liǧām).

 

 

161MUBAWWAB AND SPECIALIZED LEXICA

al-ʿāda in which Ṣaġānī (d. 650/1252) alphabetically lists 118 words which refer to habit, such as ǧibilla, ḥarīka, da⁠ʾb, daydan, saǧiyya, šakīma, ʿ irāq, qarīḥa, naǧr and hiǧǧīr.521 It is noteworthy that in this work, unlike his other monograph, Fī asāmī l-ḏiʾb wa-kunāhu,522 words are arranged according to their first and not last letters. An example of a book-length work which deals with one theme in an even later period is Šams al-Dīn al-Nawāǧī’s (d. 859/1455) treatise on wine, titled Ḥalbat al-kumayt. Although it is a literary work which deals with subjects such as the origin of wine, its characteristics, benefits and drawbacks, men- tion in poetry, and drinking companions, it does include some philological material, most notably the first chapter, devoted to the various names of wine ( fī asmāʾ al-ḫamra).523

Most of the miscellaneous topics treated by the monographs mentioned above correspond to chapters in multithematic works, such as al-Ġarīb al-muṣannaf by Abū ʿUbayd (d. 224/838), al-Ǧarāṯīm attributed to Ibn Qutayba (d. 276/889), al-Muntaḫab by Kurāʿ al-Naml (d. 310/922), and al-Muḫaṣṣaṣ by Ibn Sīda (d. 458/1066). Kurāʿ’s book, as an example, contains, in addition to several chapters on natural phenomena such as those discussed above, chap- ters that deal with dawāhī (calamities), laban (milk), ḫamr (wine), ʿ āda (habit), biʾr (well), silāḥ (weapons), and maysir (gambling).524 It is almost an ines- capable conclusion that various other topics (e.g. sarāb “mirage” and malāhī “playthings and musical instruments”) which feature as chapters in such mul- tithematic works were also treated in separate monographs that did not reach us, but some of which are mentioned by Ibn al-Nadīm (d. 380/990), among other authors.525

5 al-Muʿarrab (Arabized Words)

Interest in Arabized words goes back to the first muǧannas dictionary, Ḫalīl’s (d. 175/791) K. al-ʿAyn, and the first book on grammar, Sībawayhi’s (d. 180/796)

521 Note that words containing augments are placed according to their roots, hence the inclusion of taǧālīd and umdūd under the letters ǧ and m respectively; cf. Ṣaġānī, ʿĀda 151, 156.

522 See above, 143. 523 Nawāǧī, Ḥalba 13–22. 524 Kurāʿ, Muntaḫab I, 349–350, 382–385, 385–386, 403; II, 439–441, 490–509, 762–764

respectively. 525 Ibid., I, 374–375 (sarāb); Abū ʿUbayd, Ġarīb III, 702–704 (malāhī); cf. Ibn al-Nadīm, Fihrist

120, 80.

 

 

162 chapter 2

al-Kitāb. In the latter, there are two short chapters which deal with such words, mainly from two perspectives: whether ilḥāq (appending)526 applies to them or not, and the phonemic changes which some Arabic loanwords from Persian undergo.527 Yet, most grammarians after Sībawayhi exclude muʿarrab from their works, possibly on the assumption that it belongs to the realm of luġa, rather than naḥw. Concerning Ḫalīl, his keen interest in separating Arabic words from foreign (daḫīl) ones, mainly by applying phonotactic principles, has already been noted.528 This separation was crucial for him not only as one in a series of steps to achieve the aim of exhausting correct Arabic usage, regardless of whether the meaning of a certain word is clear or strange (ḥattā nastawʿib kalām al-ʿArab al-wāḍiḥ wa-l-ġarīb),529 but also to detect those forms which the naḥārīr invented and are not permissible in the speech of the Arabs. Later authors often quote the principles and criteria expounded by Ḫalīl and Sībawayhi,530 but hardly expand their theoretical bases. Suyūṭī (d. 911/1505), as late as the tenth/sixteenth century (if one goes by the date of his death), repeats, on the authority of Abū Ḥayyān al-Ġarnāṭī (d. 745/1344), almost exactly the same criteria set by Ḫalīl and Sībawayhi for identifying Arabized words.531 As far as terminology is concerned, muʿarrab is synonymous with daḫīl, as Suyūṭī asserts (wa-yuṭlaq ʿ alā l-muʿarrab daḫīl).532 The term muwallad, however, is usually reserved for post-classical neologisms. Suyūṭī, for example, discusses muwallad in a separate chapter which immediately follows a chapter on muʿarrab and another on neologisms of the Islamic era.533 It is therefore not surprising that muwallad material only modestly features in muʿarrab works (apart perhaps from the books of Ḫafāǧī and Muḥibbī; see below), and is nor- mally quoted in books on solecism (laḥn al-ʿāmma).

The question of whether the Qurʾān contains muʿarrab words is certainly linked to the early interest in identifying such words, and it demonstrates the influence of religious matters on philological study. Unlike most other mubawwab genres of lexicography, muʿarrab is quite late in appearance. The first book devoted to muʿarrab words in general may well be Ǧawālīqī’s

526 For ilḥāq, see above, 60, n. 296. 527 Sībawayhi, Kitāb IV, 303–307. 528 See above, 28, 60. 529 Ḫalīl, ʿAyn I, 60. 530 See, for example, Ǧawālīqī, Muʿarrab 6–12; Ibn Sīda, Muḫaṣṣaṣ XIV, 39–40. 531 Suyūṭī, Muzhir I, 270. 532 Ibid., I, 269. See also Ǧawālīqī, Muʿarrab 11, where muʿarraba (l. 2) and daḫīl (l. 10) are used

synonymously. 533 Suyūṭī, Muzhir I, 268, 294, 304.

 

 

163MUBAWWAB AND SPECIALIZED LEXICA

(d. 540/1145) al-Muʿarrab min al-kalām al-aʿǧamī ʿalā ḥurūf al-muʿǧam. Yet works on loanwords in the Qurʾān predate the more general works such as Ǧawālīqī’s. Judging by the one extant work, al-Luġāt fī l-Qurʾān, attributed to Ibn ʿAbbās (d. 68/687), works of the same title (or of the title Luġāt al-Qurʾān) which are attributed in the sources to relatively early authors – such as Hišām b. Muḥammad Ibn al-Kalbī (d. 204/819), Farrāʾ (d. 207/822), Abū Zayd al-Anṣārī (d. 215/830), and Ibn Durayd (d. 321/933)534 – probably identified a number of Qurʾānic words that are either used by specific tribes or are of foreign origin. It should be remembered however that the attribution of al-Luġāt fī l-Qurʾān to Ibn ʿAbbās is as doubtful as the ascription to him of material related to ġarīb al-Qurʾān.535 In fact, its riwāya, based on the isnād reported by Ibn Ḥasnūn (d. 386/996), is about three centuries later than the date of Ibn ʿAbbās’s death.536 Irrespective of this, the monograph includes 325 Qurʾānic words, the vast majority of which are said to be specific to certain tribes (mostly Qurayš, fol- lowed by Huḏayl, Kināna, Ḥimyar, Ǧurhum and Tamīm), and twenty eight of which, as the author puts it, are consistent with (wāfaqat/tuwāfiq) their coun- terparts in other languages (i.e. not Arabized).537 For example, sariyy (rivulet; Q 19: 24), miškāt (niche; Q 24: 35), and alīm (painful; Q 33: 8) are said to be the same as in Suryāniyya, Ḥabašiyya and ʿIbriyya respectively.538 This inter- pretation, as we know from a similar view attributed in later works to Abū ʿUbayda (d. 209/824),539 rests on the denial of the existence of loanwords in the Qurʾān, and its proponents universally cite verses from the Qurʾān which describe it as ʿArabī (e.g. Q 26: 195; 43: 3). According to Abū ʿUbayda, to say that the Qurʾān contains any foreign element is tantamount to committing a grievous sin against God (fa-qad aʿẓama ʿalā l-Lāhi l-qawl).540 Other scholars, strangely including Ibn ʿAbbās, are said to have acknowledged the existence in the Qurʾān of a large number of words which are not Arabic (bi-luġāt al-ʿaǧam; min ġayr lisān al-ʿArab).541 For his part, Abū ʿUbayd (d. 224/839) tries to bridge the gap between these two conflicting views by suggesting that the words in question are originally ( fī l-aṣl) not Arabic, but have become, following their

534 Sezgin (1982: 120, 125, 79, 105). 535 See above, 64. 536 Cf. Rippin (1981: 19) for the isnād of the various versions of the work. 537 See the editor’s enumeration in Ibn ʿAbbās, Luġāt 7, and Rippin (1981: 21). 538 Ibn ʿAbbās, Luġāt 34, 36, 38. 539 Ibn Fāris, Ṣāḥibī 59–61; Ǧawālīqī, Muʿarrab 4–5; Suyūṭī, Muzhir I, 268–269; cf. Ibn Qutayba,

Adab 384. 540 Ibn Fāris, Ṣāḥibī 59 (also 62: fa-qad aʿẓama wa-akbara); Ǧawālīqī, Muʿarrab 4. 541 Ibn Fāris, Ṣāḥibī 60; Ǧawālīqī, Muʿarrab 5.

 

 

164 chapter 2

Arabization, an integral part of Arabic (wa-ḥawwalathā ʿan alfāẓ al-ʿaǧam ilā alfāẓihā fa-ṣārat ʿArabiyya).542

The study of foreign words in the Qurʾān seems to have developed very little since the time of the scholars whose views were mentioned above. Confirmation of this fact comes from Suyūṭī’s (d. 911/1505) contribution to the subject, which includes, in addition to two chapters in al-Itqān, two monographs, namely, al-Muhaḏḏab fī mā waqaʿa fī l-Qurʾān min al-muʿarrab (an abridgement of which occurs in al-Itqān) and al-Mutawakkilī, both of which deal exclusively with loanwords to the exclusion of Arabic dialects. In all three, Suyūṭī merely reports the views of earlier scholars, such as Ibn ʿAbbās, Abū ʿUbayda, Abū ʿUbayd and Ibn Fāris (d. 395/1004). Although he quotes several works on philol- ogy, Qurʾānic readings, and exegesis, it seems that his sources include only very few works on the subject of Qurʾānic loanwords, the most notable of which is entitled Luġāt al-Qurʾān or al-Luġāt allatī nazala bihā l-Qurʾān by a certain Abū l-Qāsim,543 presumed to be Muḥammad b. ʿAbdallāh b. al-Ǧidd al-Fihrī al-Lablī (d. 515/1121).544 This apparent scarcity of works dealing with Qurʾānic loanwords can partially explain why, other than the monograph attributed to Ibn ʿAbbās, there are no extant works on the subject before Suyūṭī’s, unless one assumes that all earlier works have been lost. As for subject matter, Suyūṭī’s quotations from Abū l-Qāsim’s work in both al-Muhaḏḏab and al-Mutawakkilī indicate that it included loanwords in the Qurʾān, whereas his quotations in al-Itqān prove that it also included Qurʾānic words which were used by specific tribes. Finally, Suyūṭī uses in his two monographs two different methods of arranging his material. Both methods depart from al-Luġāt fī l-Qurʾān attrib- uted to Ibn ʿAbbās, which is arranged based on the order of the sūras of the Qurʾān. In al-Muhaḏḏab, he follows a strict alphabetical order, whereas he arranges al-Mutawakkilī based on the languages from which the loanwords are borrowed. Abū l-Qāsim’s material on specific dialects, judging by quotations in al-Itqān, seems to have been arranged according to tribes (e.g. Kināna, Huḏayl, Ḥimyar, Ǧurhum, etc.) with no alphabetical order. Quotations in al-Muhaḏḏab and al-Mutawakkilī, however, provide no clue as to the arrangement of loan- words in Abū l-Qāsim’s book.

Other works of the genre do not differentiate between Qurʾānic and non- Qurʾānic loanwords, and several multithematic works prior to Ǧawālīqī’s

542 Ibn Fāris, Ṣāḥibī 61; cf. Ǧawālīqī’s expression (Muʿarrab 5): fa-ṣāra ʿArabiyyan bi-taʿrībihā iyyāhu.

543 Suyūṭī, Itqān II, 91–102; Muhaḏḏab 40, 42, 46, 53, 55, 57, 72, 74, 79, 88–89, 93–94; Mutawakkilī 65, 96, 136, 150.

544 See the editor’s observations in Mutawakkilī 14–16.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

165MUBAWWAB AND SPECIALIZED LEXICA

al-Muʿarrab include short chapters which contain both types.545 The first such book is al-Ġarīb al-muṣannaf, in which Abū ʿ Ubayd (d. 224/838) attributes some of his data to Abū ʿUbayda (d. 209/824) and Aṣmaʿī (d. 216/831).546 Another third/ninth century author, Ibn Qutayba (d. 276/889), writes a larger chapter in Adab al-kātib, and also quotes Abū ʿUbayda and Aṣmaʿī (and in one case Ibn ʿAbbās).547 Kurāʿ al-Naml’s (d. 310/922) al-Muntaḫab lists several Arabized words, but nowhere mentions his sources.548 Much of Ibn Qutayba’s material and šawāhid are shared by Ibn Durayd (d. 321/933), who includes in the appen- dices of Ǧamharat al-luġa two sections on loan words, the latter of which is restricted to four words, namely, sawḏaq, sawḏanīq and sūḏāniq (all meaning falcon), and zindīq (a believer in the eternity of the present world).549 In the main chapter, Ibn Durayd tries to separate words according to their origin, be it Fārisī, Rūmī, Nabaṭī, or Suryānī, but often ends one of his subdivisions only to resume it elsewhere in the chapter.550 In the fifth century, Ṯaʿālibī (d. 429/1039) devotes Chapter twenty-nine of his Fiqh al-luġa551 to what he calls a compari- son (muwāzana) between Arabic and Persian. He begins by a list of Arabic words that have no Persian equivalents, but then moves to what he depicts as words whose forms are the same (ʿalā lafẓ wāḥid) in Arabic and Persian, and words which the Arabs had to borrow from Persian because they lacked their equivalents in their own language. He divides the latter type according to semantic fields, such as awānī (utensils), malābis (clothes), ṭabīḫ (cooked food), etc. but cites no more than a few words in each. The chapter then ends with words originating from Rūmiyya. Also in the fifth century, Ibn Sīda’s (d. 458/1066) al-Muḫaṣṣaṣ includes two chapters on muʿarrab, the second of which is restricted to a few proper nouns described as rare foreign words (nādir al-aʿǧamī).552 In the other chapter, Ibn Sīda starts by quoting the whole text of Sībawayhi’s two chapters on Arabization and then lists several Arabized words, often acknowledging earlier scholars such as Ḫalīl, Abū ʿUbayd, Ibn al-Sikkīt (d. 244/858), and Ibn Durayd.553

545 Note that mention of Arabized words also takes place in more general works – e.g. Ǧāḥiẓ’s Bayān I, 19–20 – but these are outside the scope of the present study.

546 Abū ʿUbayd, Ġarīb III, 668–672. 547 Ibn Qutayba, Adab 383–390. 548 Kurāʿ, Muntaḫab II, 600–603. 549 Ibn Durayd, Ǧamhara III, 1329. 550 Ibid., III, 1323–1327. 551 Ṯaʿālibī, Fiqh 314–318. 552 Ibn Sīda, Muḫaṣṣaṣ XVI, 9. 553 Ibid., XIV, 39–44.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

166 chapter 2

Book length works typically do not differentiate between Qurʾānic and non- Qurʾānic words and were quite late to appear, as previously mentioned. The most important of these are:

1. al-Muʿarrab min al-kalām al-aʿǧamī ʿalā ḥurūf al-muʿǧam by Ǧawālīqī (d. 540/1145).

2. Fī l-taʿrīb wa-l-muʿarrab, also known as Ḥāšiyat Ibn Barrī (d. 582/1187) ʿalā Kitāb al-Muʿarrab.

3. al-Taḏyīl wa-l-takmīl li-mā stuʿmila min al-lafẓ al-daḫīl by Bišbīšī (d. 820/1417).

4. Risāla fī taḥqīq taʿrīb al-kalima al-aʿǧamiyya by Ibn Kamāl Bāšā (d. 940/1534).

5. Šifāʾ al-ġalīl fī mā fī kalām al-ʿArab min al-daḫīl by Ḫafāǧī (d. 1069/1659). 6. Qaṣd al-sabīl fī mā fī l-luġa al-ʿArabiyya min al-daḫīl by Muḥibbī

(d. 1111/1699).

Generally speaking, these relatively late authors were as dependent as authors of multithematic works on earlier scholars. Suffice it to say that the latest in the above list, Muḥibbī, still quotes in his introductory section the better part of Sībawayhi’s two chapters on Arabization.554 Much of the material which Ibn Qutayba and Ibn Durayd, among others, cited in their works (along with the supporting šawāhid) is quoted in the later sources. On another issue, it should be noted that the approach to language study by the Arab lexicogra- phers and grammarians is overwhelmingly synchronic. Yet, there is reason to believe that there was early interest in comparing Arabic to other languages. Some of the comments in this respect are quite amazing, such as Ḫalīl’s obser- vation that the Kanʿāniyyūn (Canaanites) spoke a language that resembled Arabic (wa-kānū yatakallamūna bi-luġa tuḍāriʿ al-ʿArabiyya).555 There are also other comments and anecdotal material which suggest a primitive knowledge by philologists and other scholars of certain words in, or facts related to, other Semitic languages.556 Persian, of course, was familiar to many philologists, and in a few cases was their native language (Sībawayhi being the most obvious example). Some authors of the muʿarrab genre obviously knew Persian very well – such as Ibn Kamāl Bāšā, witness his recurrent citing of Persian lines

554 Muḥibbī, Qaṣd I, 113–115. 555 Ḫalīl, ʿAyn I, 205. 556 See a discussion of this issue in Baalbaki (1983a: 117–127); see also Baalbaki (1983b:

317–318).

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

167MUBAWWAB AND SPECIALIZED LEXICA

of poetry.557 However, there were others who probably had a sketchy knowl- edge of it or even totally depended on earlier scholars in determining that a certain word was borrowed from Persian. This led in many cases to the ascrip- tion of a certain word either to Persian or another language, as in ǧūdiyāʾ (gar- ment), which Ǧawālīqī says is either Nabaṭī or Fārisī, and ḫandarīs (old wine), which he reports as either Rūmī or Fārisī.558 Indeed, the same phenomenon occurs with other languages; hence, according to Ǧawālīqī, barḫ (something plentiful and cheap) is either ʿIbrī or Suryānī, durāqin (peach) either Suryānī or Rūmī, and firdaws (garden) either Rūmī or Nabaṭī.559 Furthermore, authors often merely describe words noncommittally with expressions such as aʿǧamī (foreign), aʿǧamī muʿarrab, muʿarrab, laysa bi-ʿArabī/laysa min kalām al-ʿArab (not Arabic), laysa bi-ʿArabī maḥḍ (not purely Arabic), lā aḥsabuhā ʿArabiyya ṣaḥīḥa ( I do not reckon it as genuinely Arabic), etc.560 There even are words in the repertoire of muʿarrab whose very Arabization is disputed, such as Saqar (fire of the world to come), said by some to be aʿǧamī, but derived by others from a supposedly Arabic root SQR, used in the expression saqarathu l-šamsu (The sun melted it).561

Now for a few observations on individual books in the above list. The most famous of these is Ǧawālīqī’s al-Muʿarrab, which constitutes most probably the first attempt to produce a book-length collection of Arabized words, with an introduction in which are assembled the views of earlier scholars on the presence of loanwords in the Qurʾān and their comments on changes which words undergo in the process of Arabization. Indeed, Ǧawālīqī, who nor- mally acknowledges his sources, does not mention any earlier work devoted to muʿarrab. His book embraces 743 words spread over twenty-six chapters (since there is no chapter for ḍād or ẓāʾ), and muwallad words are normally not included.562 Its alphabetical arrangement is based only on the word’s first letter.

Ǧawālīqī’s work gave rise to several commentaries, among which are the second and third titles mentioned above. Ibn Barrī comments on a number of words in al-Muʿarrab and normally follows their original order (hence only the first letter is taken into consideration in arrangement). Bišbīšī on the other

557 Ibn Kamāl Bāšā, Risāla 86, 104, 109–111, 138–141. 558 Ǧawālīqī, Muʿarrab 111, 124–125. 559 Ibid., 81–143, 240–241. 560 E.g. ibid., 58, 143; 76, 87; 79, 145; 52, 59; 96; 23, 64, 69; 66 respectively. 561 Ibid., 198. See also Ibn Barrī, Ḥāšiya 81 for whether ḫurram (plentiful; said of lifestyle) is

Arabized or Arabic. 562 Exceptions include zalābiya (pancake; 175) and ṭaraš (deafness; 224).

 

amgad
Highlight
amgad
Highlight
amgad
Highlight

 

168 chapter 2

hand arranges his words in full alphabetical order and, aside from commenting on Ǧawālīqī’s text, adds material described by a certain copyist of one of al-Muʿarrab’s manuscripts as exceeding the original several times.563 Common to both authors, however, are the types of their comments, which mostly involve correction of forms or meanings of words mentioned by Ǧawālīqī, correction and proper ascription of his šawāhid, and, less often, discussion of whether a certain word is Arabized or not.

Ibn Kamāl Bāšā’s short treatise is different from other works on muʿarrab in that it is founded on the changes, or lack thereof, in words borrowed from foreign languages. The author puts together the observations of earlier schol- ars and clearly identifies four types of loanwords, based on whether a change in consonants or vowels takes place or not, and whether ilḥāq (appending) is applied or not.564 For example, Ḫurāsān represents a word that has been nei- ther changed nor appended (mulḥaqa) to Arabic patterns; ḫurram (plentiful; said of lifestyle) one that has not been changed but appended; āǧurr (baked bricks) one that has been changed but not appended; and dirham (silver coin; dram) one that has been both changed and appended. The title of the Risāla (Treatise on the Verification of Arabized Foreign Words) indicates that its author is interested in guidelines related to identifying Arabized words – although in the relatively few examples which he cites, compared to other authors, the discussion hardly reflects this interest – rather than in listing the largest pos- sible number of such words. Given that his material is divided according to the above mentioned types,565 Ibn Kamāl Bāšā seems to have found little reason for the alphabetical ordering of the few examples cited in each type.

Contrary to Ibn Kamāl Bāšā, both Ḫafāǧī and Muḥibbī try to be as exhaus- tive as possible in collecting Arabized words. In his introduction, Ḫafāǧī describes Ǧawālīqī’s al-Muʿarrab as the best in its genre (aǧall mā ṣunnifa fī hāḏā l-bāb),566 but asserts that his own book is more comprehensive and that it is unprecedented in including muwallad material. In fact, a large number of his 1,389 entries belong to muwallad rather than muʿarrab, including expres- sions used by common people (ʿāmma), such as ḫālī l-ġurfa (lit. empty-roomed; i.e. empty-headed) in Baghdadi usage and sakrān ṭīna (heavily drunk; i.e. as if having fallen in ṭīn “clay”).567 Strangely enough in the eleventh/seventeenth century, the entries are not arranged beyond the first letter, perhaps following

563 See the introduction of al-Muʿarrab’s editor, 14. 564 Ibn Kamāl Bāšā, Risāla 46–48; cf. Ḫafāǧī, Šifāʾ 43–44. See also above, 60, n. 296. 565 Ibid., 67, 115, 125. 566 Ḫafāǧī, Šifāʾ 32. 567 Ibid., 138, 180.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

169MUBAWWAB AND SPECIALIZED LEXICA

Ǧawālīqī’s example, although Ḫafāǧī does not follow the order in which words appear in his predecessor’s book. Muḥibbī’s arrangement of material, on the other hand, is fully alphabetical, but, like Ḫafāǧī, he declares his intention to surpass his predecessors (Ḫafāǧī himself included) in the amount of material cited.568 To achieve this, he not only includes material which he describes as muwallad or laḥn (solecism), much of which is probably taken from Ḫafāǧī,569 but also a host of words and expressions which he depicts as neither muwallad nor laḥn, and which are certainly not muʿarrab. Totally unjustified, for exam- ple, is the inclusion of names of certain sects or groups, such as Ṯumāmiyya, Ǧāḥiẓiyya, Ǧārūdiyya, Ḫawāriǧ, Šīʿa, Ṣifātiyya, etc.570 Even a word such as Ǧāhiliyya or an expression such as al-ṣaḥīḥ min al-Ḥadīṯ (correct or reliable prophetic traditions)571 are listed as entries, in a clear attempt at inflating the book’s content.

As hinted above, authors of works on muʿarrab derived from earlier scholars most of their comments on the various changes that foreign words are subject to in adapting them to the Arabic noun system and added precious little to their observations. Dependence on earlier scholarship, particularly from the formative period of the second/eighth and third/ninth centuries, surely marks many other genres of writing, but the fact that the first book on muʿarrab most probably did not appear before the first part of the sixth/twelfth century must have contributed to the exceptionally heavy reliance of the genre on previous comments and observations. Although later authors did enlarge the repertoire of Arabized words, they did little to develop the criteria set by some of the ear- liest philologists concerning phonemic change, phonotactic principles, ilḥāq, etc., let alone come up with a clear theoretical basis for Arabization beyond the four types elaborated on by Ibn Kamāl Bāšā. And neither did they extensively classify and discuss loanwords from the perspective of the field of material culture to which they belong. Accordingly, Ḥamzāwī’s suggestion that Ǧawālīqī was the first author to adopt a theoretical approach (muqāraba naẓariyya) to language interaction (al-tadāḫul al-luġawī) is obviously an unfounded exaggeration.572

568 Muḥibbī, Qaṣd I, 103–104. 569 E.g. ʿ Ayša for ʿ Āʾiša, and ṭibāʿ, pl. of ṭabʿ (disposition), instead of the unchanged plural form

ṭabʿ; Ibid., II, 250, 307; cf. Ḫafāǧī, Šifāʾ 207, 212. 570 Muḥibbī, Qaṣd I, 360, 362–363, 467–468; II, 216, 227–229. 571 Ibid., I, 367; II, 221. 572 Ḥamzāwī (1991: 266 ff.).

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

170 chapter 2

6 Laḥn al-ʿĀmma (Solecism)

Linguistic error is a subject that is closely related, at least according to the view presented mainly in the biographical sources, to the beginning of grammatical activity. As previously noted, both Abū l-Aswad al-Duʾalī (d. 69/688), alleged to be the first grammarian, and Sībawayhi (d. 180/796), the most influential figure in the grammatical tradition, are said to have been prompted to investigate grammatical issues following the occurrence of laḥn, either in the speech of others, as in Abū l-Aswad’s case, or in one’s own, as in Sībawayhi’s.573 The gen- eral term denoting linguistic error, laḥn, is frequently used in its grammatical sense in biographical works and other sources which try to explain the emer- gence of grammatical study as a reaction to the dramatic increase in language corruption.574 More authentically, it occurs on two occasions in the Kitāb, the first of which in the description of the phrase *yā aḫūnā (O our brother), pro- posed by Sībawayhi as an impermissible alternative for the correct yā aḫānā, and the second in connection with the Qurʾānic reading hāʾulāʾi banātī hunna aṭhara lakum (Here are my daughters: they are purer for you; Q 11: 78), instead of aṭharu.575 A synonym of laḥn in the Kitāb is ġalaṭ, which Sībawayhi uses to describe attested usage that he deems to be impermissible, such as nuwayb, diminutive of nāb (canine tooth), instead of nuyayb.576 Sībawayhi also uses the term ḫaṭa⁠ʾ as a synonym for laḥn, but with reference to unattested usage assumed by him or by other grammarians.577 Both terms, as well as wahm/ awhām and saqṭa/saqaṭāt, are used in later sources to express the same notion as laḥn.578

Another type of laḥn that is reported to have taken place as of the first/ seventh century has to do with phonology, morphology and semantics. Examples of this type are widely quoted in the sources, most notably in al-Bayān wa-l-tabyīn by Ǧāḥiẓ (d. 255/869).579 In fact, most of the material cited by authors of the genre is related to this type, and not to the previous

573 See above, 3. 574 Cf. Ibn Sallām, Ṭabaqāt I, 12; Zubaydī, Ṭabaqāt 11–12; Ibn al-Anbārī, Nuzha 201–21; Suyūṭī,

Aḫbār 162–164. See also Baalbaki (1995: 124–125) for the close connection between laḥn and the grammatical topics which are said to have captured the attention of the early grammarians.

575 Sībawayhi, Kitāb II, 184, 396–397. 576 Ibid., III, 462. 577 Ibid., II, 226; III, 312, 472. For a more detailed discussion of ġalaṭ and ḫaṭa⁠ʾ and their

comparison with tawahhum, see Baalbaki (1982: 239–240). 578 For a general study of the development of the genre laḥn and the relationship between

laḥn and linguistic change, see Maṭar (1966) and ʿAbdaltawwāb (2000). 579 Ǧāḥiẓ, Bayān I, 69–74; II, 210–224; cf. above, 5–6.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

171MUBAWWAB AND SPECIALIZED LEXICA

type which pertains to syntax and operants. Common to both types, however, is a puristic approach which regards correctness and faṣāḥa to be exclusive to a variety of ʿArabiyya whose norms are derived from pre-Islamic poetry, the Qurʾān and the speech of the Aʿrāb. Given their efforts to standardize usage, it is not surprising that grammarians and philologists alike readily consider any departure from these norms as corruption or decay of speech, rather than evidence of inescapable linguistic evolution. Even when they defend ʿāmmī usage (as Ibn Hišām al-Laḫmī does in his al-Madḫal; see below), they do so on the basis of its similarity to attested dialectal material from earlier periods. In line with this approach is the fact that most philologists normally blame the ʿāmma/ʿawāmm (common people; generality of people) for laḥn, as if to corroborate their view that deviation from the norm is by definition a form of speech corruption that takes place on a wide scale. But although the ʿāmma were the most obvious target for blame in this respect, there is evidence which suggests that the term ʿāmma/ʿawāmm was not uniformly used to refer to the lower orders of society which can be readily accused of ignorance and hence of linguistic error. Ǧāḥiẓ, for example, explains that by ʿawāmm he does not refer to the peasants ( fallāḥūn), the lowest of mankind (ḥušwa), the handi- craftsmen (ṣunnāʿ ), or the tradesmen (bāʿa), nor to the Kurds of the mountains or the dwellers of the islands in the seas; rather, he means the ʿawāmm who possess intelligence and morals superior to those of barbarians such as the Ṭaylasān, Ǧīlān and Zanǧ, but who still do not attain the level of the ḫāṣṣa (elite).580 On the subject of laḥn, Ǧāḥiẓ devotes a chapter to those whom he depicts as laḥḥānūn bulaġāʾ, that is, eloquent people – including two jurists, Ḫālid b. al-Ḥāriṯ (d. 186/802) and Bišr b. al-Mufaḍḍal (d. 187/803) – who err fre- quently.581 This chapter clearly demonstrates that as early as the time of Ǧāḥiẓ, laḥn was hardly restricted to the ʿāmma. As we shall see below, some authors did refer to the ḫāṣṣa in their works, and some even wrote monographs which deal exclusively with the laḥn of certain highly educated groups such as trans- mitters of Ḥadīṯ, jurists and even grammarians.

To several scholars of the second/eighth and third/ninth centuries are attributed monographs on laḥn al-ʿāmma. The content of these and later monographs was part of the raw material which authors of muǧannas lexica included in their corpus.582 The earliest work of the genre is Kisāʾī’s (d. 189/805),

580 Ibid., I, 137. 581 Ibid., II, 221. 582 See, for example, index 13 of Ibn Durayd’s Ǧamhara (III, 1733–1734) for usage attributed

to the ʿāmma. In Lisān al-ʿArab, Ibn Manẓūr mentions several books of the genre, such as Ṯaʿlab’s al-Faṣīḥ, Zubaydī’s Laḥn al-ʿawāmm, Ḥarīrī’s Durrat al-ġawwāṣ, and Ǧawālīqī’s al-Takmila; cf. the relevant parts in Abū l-Hayǧāʾ & ʿAmāyira (1987, vol. II).

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

172 chapter 2

for which see below. He and those of his contemporaries who reportedly con- tributed to the genre were among the most active in the era of data collection (ǧamʿ al-luġa) and must have noted the discrepancy between the speech of the Bedouin fuṣaḥāʾ, which they considered to be representative of the “per- fect” form of Arabic, and the everyday speech of people mainly in urban areas. By comparing the two varieties, they seem to have concluded that the latter is a corruption of the former and that deviations by reference to the norm should be recorded, denounced and amended. Monographs up to the middle of the third/ninth century583 – all of which contain the term ʿāmma in their titles – include Farrāʾ’s (d. 207/822) al-Bahī/al-Bahāʾ fī mā talḥan fīhi l-ʿāmma,584 Abū ʿUbayda’s (d. 209/824) Mā talḥan fīhi l-ʿāmma,585 Aṣmaʿī’s (d. 216/831) Fī mā yalḥan fīhi l-ʿāmma,586 Abū ʿUbayd’s (d. 224/838) Mā ḫālafat fīhi l-ʿāmma luġāt al-ʿArab,587 Abū Naṣr al-Bāhilī’s (d. 231/846) Mā yalḥan fīhi l-ʿāmma,588 Māzinī’s (d. 249/863) Mā yalḥan fīhi l-ʿāmma,589 and Abū Ḥātim al-Siǧistānī’s (d. 255/869) Mā talḥan fīhi l-ʿāmma.590 Although none of these works has sur- vived, we do possess the following titles from the second/eighth and third/ ninth centuries:

1. Mā talḥan fīhi l-ʿāmma by Kisāʾī (d. 189/805). 2. al-Ḥurūf allatī yutakallam bihā fī ġayr mawḍiʿihā and Iṣlāḥ al-manṭiq by

Ibn al-Sikkīt (d. 244/858).

583 See a fuller list from this and other periods, as well as a discussion of earlier lists, in ʿAbdaltawwāb (2000: 72–94).

584 Ibn al-Nadīm, Fihrist 73; Yāqūt, Muʿǧam II, 539; VI, 2815; Suyūṭī, Buġya II, 333. Ibn Ḫillikān (Wafayāt VI, 181) reports having seen Farrāʾ’s book and asserts that it is the origin from which Ṯaʿlab derived his own book, al-Faṣīḥ.

585 Ibn al-Nadīm, Fihrist 60; Yāqūt, Muʿǧam VI, 2708; Suyūṭī, Buġya II, 295. 586 Mentioned by Ibn al-Ǧawzī, Taqwīm 74 (see also 175 for a possible quotation from it) and

quoted by Ibn Yaʿīš, Šarḥ I, 8. 587 Quoted twice by Ibn Manẓūr, Lisān (QZZ, QQZ). The book which Ibn Manẓūr refers to,

however, is almost certainly a chapter of the same title (with the addition of min al-kalām at the end) in Abū ʿUbayd’s Ġarīb (III, 672–673). This very short chapter remarkably contains the material quoted by Ibn Manẓūr.

588 Ibn al-Nadīm, Fihrist 61; Yāqūt, Muʿǧam I, 227; Suyūṭī, Buġya I, 301. 589 Ibn al-Nadīm, Fihrist 63; Yāqūt, Muʿǧam II, 763; Suyūṭī, Buġya I, 465; also quoted in Ibn

Makkī, Taṯqīf 324. 590 Ibn al-Nadīm, Fihrist 64; Yāqūt, Muʿǧam III, 1407. The shorter title, Laḥn al-ʿāmma, appears

in Zubaydī, Laḥn 5 and Suyūṭī, Buġya I, 606. The book is most probably the same as the one referred to by Azharī (Tahḏīb I, 22–23) and quoted by Ibn Manẓūr (Lisān: ʾHL, SDM).

 

amgad
Highlight
amgad
Highlight
amgad
Highlight

 

173MUBAWWAB AND SPECIALIZED LEXICA

3. al-Faṣīḥ by Ṯaʿlab (d. 291/904). 4. al-Fāḫir by al-Mufaḍḍal b. Salama (d. 290/903 or 300/913).

The attribution of the first title to Kisāʾī is controversial, particularly because the sources do not mention it among his works and because it includes five words whose forms do not agree with what other works attribute to him.591 Yet, its attribution to Kisāʾī is quite likely, given that its material often agrees with riwāyas ascribed to him in several lexica, as its editor, R. ʿAbdaltawwāb, demonstrates.592 In all cases, the fact that the material of the text lacks any arrangement suggests an early date. It comprises 107 brief entries (with a total of about 250 words) which mostly begin with the expression taqūl (You shall say) or yuqāl (It is said), followed by the correct form. The incorrect form – normally introduced by lā taqūl/lā yuqāl – is not always mentioned, and thus one often has to resort to other sources to determine what the incorrect form could have been. For example, ḫiṭmī (preparation for washing the head), with an initial kasra, is cited as the correct form of the word,593 but since the incor- rect form is not cited, one wonders whether it has an initial ḍamma or fatḥa. We learn from Zubaydī (d. 379/989), however, that the form used by the ʿāmma is ḫaṭmī, with a fatḥa.594 Most of the monograph’s material has to do with similar changes in the vowels of nouns and verbs and, to a lesser extent, with morphological patterns, particularly those of masculine and feminine nouns.595

The monograph al-Ḥurūf allatī yutakallam bihā fī ġayr mawḍiʿihā by the Kufan philologist Ibn al-Sikkīt differs significantly from Kisāʾī’s, both in its arrangement and content. Unlike his predecessor, Ibn al-Sikkīt arranges his material in four distinct parts. In the first, he mentions a number of metaphors which he apparently believes are unjustified. One example is the expression ġalīẓ al-mašāfir (thick-lipped) in which mašāfir, pl. of mišfar (lip of a camel), is used with reference to a man.596 The second part includes a number of lines of poetry in which words are used permissively or incorrectly (ǧawwazathā l-ʿArab aw ġaliṭat fīhā). In the first three lines, for example, the proper noun

591 For a discussion of the authenticity of the work’s attribution to Kisāʾī, see Pellat (1986: 607) and ʿAbdaltawwāb’s introduction to Kisāʾī, Mā talḥan 69 ff.

592 Introduction to Kisāʾī, Mā talḥan 75–76. The editor also mentions (ibid., 3–4) a manuscript entitled al-Ifhām fī mā talḥan fīhi l-ʿawāmm which is similarly attributed to Kisāʾī but contains, in addition to material taken from Mā talḥan fīhi l-ʿāmma, material from later works on laḥn.

593 Ibid., 116–117. 594 Zubaydī, Laḥn 270. 595 Masculine and feminine patterns are mostly discussed in items 56–59. 596 Ibn al-Sikkīt, Ḥurūf 36.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

174 chapter 2

Sulaymān is changed either to Sallām or Sulaym.597 The third part is devoted to duals in which one of two nouns designates both of them, such as al-Aqraʿān, used for al-Aqraʿ b. Ḥābis and his brother Firās (a phenomenon which phi- lologists call taġlīb “attribution of predominance”), and duals which refer to a trait shared by two objects, such as al-Aḥmarān (the two red things) for meat and wine.598 Finally, the fourth part includes a few words and expressions in which the plural is used instead of the dual or the singular (obviously if one goes by the literal meaning). Hence, in the above mentioned expression ġalīẓ al-mašāfir, as well as in waqaʿa fī lahawāt al-layṯ (He fell into the lion’s uvulas), the plural mašāfir and lahawāt occur in lieu of the dual al-mišfarayn and the singular lahāt.599

A much more important work by Ibn al-Sikkīt is Iṣlāḥ al-manṭiq – a book that was commented on, expanded, or abridged by a large number of scholars.600 Although only ten of its chapters specifically mention the ʿāmma, the other chapters are replete with the introductory expression taqūl/yuqāl, which obviously introduces what is considered correct usage, and at times this is contrasted with the incorrect forms (e.g. wa-yuqāl fulān aʿsaru yasarun [X is ambidextrous] . . . wa-lā yuqāl aʿsaru aysaru).601 The ten chapters in which the ʿāmma are mentioned do not differ in style from the rest of the book but deal with three general types of incorrect usage, namely, mistakes in vowels, consonants and meanings. One chapter, for example, lists nouns with an initial kasra which the ʿāmma change either to fatḥa (e.g. ṣannāra < ṣinnāra “head of a spindle”) or ḍamma (e.g. šuhrīz < šihrīz “type of dates”).602 Consonantal change includes a chapter on the incorrect omission of hamza (e.g. fiyām < fiʾām “company of men”) and another on the incorrect shifting of ṣād to sīn and vice versa (e.g. basaqa < baṣaqa “to spit” and šamūṣ < šamūs “a horse etc. that refuses to be mounted”).603 As for incorrect change in meaning or, more gener- ally, improper use of a word in context, it is introduced in two chapters by the expression mimmā taḍaʿuhu l-ʿāmma fī ġayr mawḍiʿihi (example of misuse of words by common people).604 One instance is the use of tanazzaha for going

597 Ibid., 40–41. Note that Ibn Durayd’s chapter in Ǧamhara III, 1327–1329 is on the same theme and includes some of the šawāhid cited by Ibn al-Sikkīt. See also Ibn Ǧinnī, Ḫaṣāʾiṣ III, 273 ff.

598 Ibn al-Sikkīt, Ḥurūf 47, 52. 599 Ibid., 54, 53; cf. below, 238. 600 Cf. Sezgin (1982: 131–132); Iqbāl (2011: 77–79). 601 Ibn al-Sikkīt, Iṣlāḥ 294. 602 Ibid., 173, 175. 603 Ibid., 146, 184, 185. 604 Ibid., 284, 287.

 

amgad
Highlight

 

175MUBAWWAB AND SPECIALIZED LEXICA

out to the gardens, whereas its correct meaning is keeping away from bodies of water (miyāh) and lands that are close to water (aryāf ).605 It is also note- worthy that Ibn al-Sikkīt’s examples often include dialectal material and that although he uniformly accepts dialectal variants, he reserves the term faṣīḥ for what he regards as the representative of the norm, but which has less accept- able variants.606

Of no lesser impact on the philological tradition than Iṣlāḥ al-manṭiq is a considerably shorter book by Ṯaʿlab, al-Faṣīḥ, which was also the subject of a large number of commentaries, expansions and critical writings.607 Its thirty chapters deal almost exclusively with morphological patterns of verbs and nouns and resemble some of the chapters in al-Iṣlāḥ. Among the chapters that deal with verbs are ones devoted to the patterns faʿaltu, faʿiltu, fuʿila and afʿala as well as finally hamzated verbs. Chapters on nouns, on the other hand, include those which deal with verbal nouns that are adjectivally used (mā ǧāʾa waṣfan min al-maṣādir),608 nouns whose first vowel is a fatḥa, kasra or ḍamma, and nouns which contain a doubled consonant, a hamza, or a final -h that is not a feminine marker. In a number of chapters on verbs and nouns, Ṯaʿlab consid- ers differences in meaning arising from the various patterns under consider- ation, as in naqiha (to comprehend) and naqaha (to convalesce), and ḥumūla (burdens) and ḥamūla (camels, horses, etc. that bear burdens).609 Although Ṯaʿlab mostly cites the forms which he considers to be correct, and only rarely points out incorrect forms,610 his book manifestly belongs to the genre of laḥn, and it is clear in the epilogue that Ṯaʿlab regards it as such (allafnāhu ʿalā naḥw mā allafa l-nās wa-nasabūhu ilā mā talḥan fīhi l-ʿawāmm).611 The term laḥn, it should be noted, refers here not only to errors that occur in the every- day speech of the generality of people, but also in writing (cf. the expression mimmā yaǧrī fī kalām al-nās wa-katbihim).612 In other words, laḥn, according to Ṯaʿlab, is not restricted to the commoners but also occurs in written texts, that is in the usage of those referred to by some later authors as ḫāṣṣa. The final

605 Ibid., 287; cf. Tibrīzī, Tahḏīb II, 113. Note also that Ibn Qutayba (Adab 34) cites this example without mentioning Ibn al-Sikkīt and rejects considering it as laḥn.

606 See the two chapters on faʿaltu and faʿiltu (ibid., 188–190, 208–210). 607 Cf. Sezgin (1982: 142–144); ʿAbdaltawwāb (2000: 187–190); Iqbāl (2011: 79–86). 608 E.g. the verbal noun ḫaṣm in the expression huwa/hiya/humā, etc. ḫaṣm (He/she etc. is

an antagonist or ligitant); Ṯaʿlab, Faṣīḥ 41. See other examples cited by Ibn Fāris in his additions to Ṯaʿlab’s material (Tamām 26–27).

609 Ṯaʿlab, Faṣīḥ 17, 63. 610 E.g. ibid., 92, 93, 96, 99. 611 Ibid., 104. 612 Ibid., 2.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

176 chapter 2

word in the above expression is unvocalized and may well be read alternatively wa-kutubihim (and their books), which would be an even more serious accusa- tion against the ḫāṣṣa. Based on Ibn Ḫillikān’s comparison between al-Faṣīḥ and Farrāʾ’s al-Bahī and his conclusion that the former is basically a modified version of the latter and that Ṯaʿlab merely rearranged his predecessor’s mate- rial and provided only few additions (ziyāda yasīra) of his own,613 one cannot ignore the possibility that the inclusion of laḥn material derived from the so- called ḫāṣṣa goes back to the earliest stage of this genre of writing.

al-Mufaḍḍal b. Salama’s al-Fāḫir, previously discussed with the books of amṯāl, does include material of the laḥn al-ʿāmma type.614 It does not, however, directly belong to the genre since the main purpose of its author is to cite and elucidate expressions which the ʿāmma use without knowing their meanings, but which do not necessarily contain laḥn. From a slightly later period, Abū Bakr b. al-Anbārī (d. 328/940) in al-Zāhir fī maʿānī kalimāt al-nās also cites laḥn attributed to the ʿāmma,615 although, like al-Mufaḍḍal, he mostly deals with phrases and expressions which they use but widely miscomprehend.

Material on laḥn al-ʿāmma is also found in multithematic works of the third/ninth century. Abū ʿUbayd’s (d. 224/838) al-Ġarīb al-muṣannaf contains a very short chapter on the topic,616 and it seems that Ibn Manẓūr (d. 711/1311) mistook it for a separate book.617 The chapter comprises thirty words, remark- ably ten of which are muʿarrab. For his part, Ibn Qutayba (d. 276/889) devotes one of the four sections of his famous Adab al-kātib to what he tellingly calls Taqwīm al-lisān (Amendment of Speech).618 The issues discussed in this sec- tion resemble to a great extent those discussed by Ibn al-Sikkīt and Ṯaʿlab and largely focus on vocalic and consonantal change, and verbal and nominal pat- terns. Other sections of the book sporadically contain material on laḥn, most notably its first chapter. The title of this chapter, Bāb maʿrifat mā yaḍaʿuhu l-nās fī ġayr mawḍiʿihi,619 is reminiscent of the two previously mentioned chapters in al-Iṣlāḥ with a similar title and likewise deals with usage that has obviously undergone semantic change, but which Ibn Qutayba considers to be incorrect. It is also worth mentioning that laḥn may have contributed to the emergence of several types of works on abniya (see Section 10 below), such as muḏakkar

613 Ibn Ḫillikān, Wafayāt VI, 181; cf. above, 172, n. 584. 614 See above, 110–111. 615 Ibn al-Anbārī, Zāhir 19, 23, 412; cf. above, 111–112. 616 Abū ʿUbayd, Ġarīb III, 672–673. 617 See above, 172, n. 587. 618 Ibn Qutayba, Adab 238–332. 619 Ibid., 17–36; cf. Baṭalyawsī, Iqtiḍāb II, 5 ff.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

177MUBAWWAB AND SPECIALIZED LEXICA

and muʾannaṯ (masculine and feminine), maqṣūr and mamdūd (abbreviated and prolonged patterns), and the verbal patterns faʿala and afʿala, given that users frequently commit laḥn which involves these topics.

Ǧāḥiẓ’s illustrations of laḥn and his clarification of what he means by the term ʿawāmm have been mentioned earlier in this section, but it may be appropriate at this point to examine his allusion to a different meaning of the word laḥn. To begin with, it is essential to point out that Ǧāḥiẓ’s views on laḥn significantly differ from those of most other authors in that he deems it admissible, even desirable, in certain contexts. For example, he notes that it is more easily acceptable if it occurs in the speech of charming slave girls or pretty young women (al-ǧawārī l-ẓirāf . . . al-šawābb al-milāḥ) and that it may well represent a natural disposition in a speech community (ʿalā saǧiyyat sukkān ahl al-balad).620 Furthermore, he contends that, in reporting anec- dotal material, iʿrāb (case-endings) is appropriate for the speech of the Aʿrāb (Bedouins), but not for the speech of the muwalladūn (post-classical speak- ers) and the ʿawāmm or the baladiyyūn (commoners).621 Such permissive views are obviously contradictory to the intolerant approach to laḥn of most grammarians and philologists,622 but they reveal Ǧāḥiẓ’s keen perception of the intricate sociolinguistic issues related to speech habits. In light of this, it is not surprising that he cites as evidence of the notion istimlāḥ al-laḥn (deem- ing laḥn as pleasing to the ear) Mālik b. Asmāʾ’s (d. c. 100/718) line manṭiqun ṣāʾibun wa-talḥanu aḥyā/nan wa-aḥlā l-ḥadīṯi mā kāna laḥnā ([she has] correct speech, but also she sometimes commits laḥn; and the sweetest utterance is the one with laḥn). According to ʿAlī b. Yaḥyā l-Munaǧǧim (d. 275/888), Ǧāḥiẓ later realized that he had misinterpreted the line by failing to see that laḥn meant “allusion” (taʿrīḍ; cf. li-tastur maʿnāhu wa-tuwarrī ʿ anhu “to veil its mean- ing and disguise it”), rather than “error”.623 Irrespective of ʿAlī’s claim, the use of the term laḥn to indicate error of speech seems to have developed from an original sense of “manner of speaking”. This would readily explain the use of the term to indicate a specific dialect, as in the previously cited hemistich

620 Ǧāḥiẓ, Bayān I, 146. 621 Ibid., I, 145–146; Ǧāḥiẓ, Ḥayawān I, 282; III, 39. 622 See Baalbaki (2009: 91–110) for a discussion of Ǧāḥiẓ’s philological views which firmly

placed him outside the mainstream philological tradition. For laḥn in particular, see pp. 100–102.

623 Iṣfahānī, Aġānī XVII, 164; cf. Marzubānī, Muʿǧam 266; Baġdādī, Tārīḫ XII, 214–215; Yāqūt, Muʿǧam V, 2110; Ibn Manẓūr, Lisān (LḤN). For arguments against Yaḥyā’s claim that Ǧāḥiẓ reconsidered his interpretation of Asmāʾ’s line, see Baalbaki (2009: 101–102).

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

178 chapter 2

wa-lā tārikan laḥnī li-uḥsina laḥnahum,624 or even singing (ġināʾ), given that it indicates a specific manner of performance of a given tune.625 In their largely normative approach, the philologists and grammarians regarded as incorrect any use that deviated from the standards which they set for acceptable usage, and their uninterrupted use of the term laḥn to express this deviation certainly contributed to the fact that the term eventually became so closely identified with linguistic error that occurrences which conveyed a different sense were the source of controversy, as in the case of Mālik’s line.

Works from the fourth/tenth century onward significantly vary in size, con- tent, arrangement, attitude toward linguistic error, and even regional pecu- liarities. To facilitate the survey and discussion of these works, we propose to divide them into two groups: non-alphabetically arranged works and alphabet- ically arranged ones.626 At the end of the first group, mention will be made of non-alphabetically arranged works that are devoted to laḥn of a specific group of scholars.

The most important extant titles of the non-alphabetically arranged type are:

1. Laḥn al-ʿawāmm by Zubaydī (d. 379/989). 2. Taṯqīf al-lisān wa-talqīḥ al-ǧanān by Ibn Makkī al-Ṣiqillī (d. 501/1107). 3. Durrat al-ġawwāṣ fī awhām al-ḫawāṣṣ by Ḥarīrī (d. 516/1122). 4. Takmilat iṣlāḥ mā taġlaṭ fīhi l-ʿāmma by Ǧawālīqī (d. 540/1145). 5. al-Madḫal ilā taqwīm al-lisān wa-taʿlīm al-bayān by Ibn Hišām al-Laḫmī

(d. 577/1181). 6. Laḥn al-ʿawāmm fī mā yataʿallaq bi-ʿilm al-kalām by Sakūnī (d. 717/1317). 7. al-Ǧumāna fī izālat al-raṭāna by Ibn al-Imām (ninth/fifteenth century). 8. Sahm al-alḥāẓ fī wahm al-alfāẓ; ʿIqd al-ḫalāṣ fī naqd kalām al-ḫawāṣṣ;

and Baḥr al-ʿawwām fī mā aṣāba fīhi l-ʿawāmm by Ibn al-Ḥanbalī (d. 971/1563).

624 See above, 13; cf. Ibn al-Anbārī, Aḍdād 240 (bi-laḥn al-Yaman ay bi-luġatihim). 625 Bakrī, Simṭ I, 21. For the semantic development of laḥn, see Fück (1955: 195–205); Pellat

(1986a: 606); ʿAbdaltawwāb (2000: 13–34); Ayoub (2007: 628–629); Sanni (2010: 1–19). 626 In both groups are listed works up to the tenth/sixteenth century. For later works, as well

as for lost works as of the second half of the third/ninth century (lost works of an earlier date were mentioned above, 172), the most comprehensive and reliable list is that by ʿAbdaltawwāb (2000: 105–108).

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

179MUBAWWAB AND SPECIALIZED LEXICA

Zubaydī’s Laḥn al-ʿawāmm or Iṣlāḥ laḥn al-ʿāmma bi-l-Andalus627 is the first book in the genre by an Andalusian scholar. It is, in particular its introduction, interesting in more than one respect. To begin with, Zubaydī tries to define the speakers from whom he derives his material. For this purpose, he says that his book contains errors of the ʿāmma, which have infiltrated the speech of most of the elite (al-kaṯra min al-ḫāṣṣa) so much that one encounters such errors in poems and in the writings of eminent scribes and letters of high-ranking functionaries. Accordingly, he declares that his book deals only with the errors of the ḫāṣṣa and that the corrupt usage of the dahmāʾ (masses, multitudes) and suqqāṭ (lowest of people) should not be part of it to ensure that it does not become too long.628 But almost immediately following that, he defends the inclusion in his book of vulgar speech (al-kalām al-sūqī) and commonly used speech (al-lafẓ al-mustaʿmal al-ʿāmmī) by asserting that change takes place only in what is frequently used, rather than in what he terms waḥšī (unfamil- iar, uncouth).629 The picture is further confused by the fact that Zubaydī refers to the ʿāmma, rather than the ḫāṣṣa, not only in the book’s title but also in the titles of its three major divisions, the first of which deals with changes in vowels and morphological forms and patterns, and the other two with seman- tic change.630 Pellat suggests that the use of ʿāmma in the titles of books of the genre is a pure euphemism which is “designed to disguise the truth and spare the feelings of the khāṣṣa” and which puts the blame on the ʿāmma for linguistic deviations.631 This might explain the titles of some books and chap- ters therein, but it remains that Zubaydī, in spite of the apparent confusion in the manner he defines the focus of his book, clearly distinguishes between the ʿāmma and the ḫāṣṣa, not only in his introduction but also in several of its entries. In fact, there are instances in which he mentions the general error, the one committed by the ʿāmma, and then specifically cites its occurrence in the speech of those whom he obviously considers to be among the ḫāṣṣa since he refers to them as scholars (ahl al-ʿilm), eminent writers (mutaqaddimū l-kuttāb) and respectable men of letters and orators (aǧillat al-udabāʾ; ǧillat

627 The second title appears in Ibn Šuhayd’s (d. 426/1035) Tahḏīb 26. It is noteworthy that Ibn Ḫayr al-Išbīlī (Fahrasa 346) reports that there are two compilations (al-ta⁠ʾlīf al-awwal wa-l-ṯānī) by Zubaydī of his Laḥn al-ʿāmma. It seems that Zubaydī added as many as 104 entries to the earlier book which contains 352 entries (Tahḏīb, Introduction 10). Ibn Šuhayd explains that, in his own book, he arranges alphabetically both compilations of Zubaydī’s.

628 Zubaydī, Laḥn 7–8. 629 Ibid., 9. 630 Ibid., 11, 206, 240 ( the latter is a pronominal reference: yūqiʿūnahu, i.e. the ʿāmma). 631 Pellat (1986b: 606).

 

amgad
Highlight
amgad
Highlight
amgad
Highlight

 

180 chapter 2

al-ḫuṭabāʾ).632 From another perspective, Zubaydī’s material is original, in the sense that he does not derive it from earlier authors. This fact is evident in the introduction since Zubaydī mentions the ascription of Abū Ḥātim al-Siǧistānī (d. 255/869), to people of the Orient (ahl al-Mašriq), certain errors which, according to Zubaydī, do not feature in the speech of his own ʿāmma, i.e. in Andalus.633 Alternatively, he notes that errors which are current in his own lin- guistic community (fī zamāninā wa-ufuqinā) are not mentioned by Abū Ḥātim or other authors (i.e. of the Mašriq). Furthermore, he often compares errors of his community with errors that occur in the Mašriq, as in nayfaq (part of trousers through which the waistband passes), which is erroneously changed to nāfiq in the West and nīfaq in the East.634

The difference between the Mašriq and the Maġrib in laḥn is also discussed by Ibn Makkī in the introduction to his Taṯqīf al-lisān. He cites several words that are erroneously pronounced by ahl al-Mašriq and ahl al-Andalus, but not by his own countrymen (ahl baladinā), i.e. speakers of the Maġribī dialect of Sicily.635 He also notes that there are instances in which the two latter groups agree or differ in their laḥn. Accordingly, he records the errors which he has directly heard from people (al-aġālīṭ allatī samiʿtuhā min al-nās) and which do not feature in earlier books (kutub al-mutaqaddimīn). Ibn Makkī, however, makes it clear that the errors which he includes in his book occur in the speech of people from the various levels of society (ʿalā ḫtilāf ṭabaqātihim). Indeed, his description of the general degradation and adulteration of language leaves no doubt about his conviction that laḥn is prevalent in the speech of both the ʿāmma and ḫāṣṣa. He specifically mentions some errors of ḫāṣṣāt al-nās in his introduction,636 and even devotes four chapters to a comparison between the usage of the ʿāmma and ḫāṣṣa.637 In one of these chapters, he finds both groups to be mistaken, whereas in the other three he sides with the ʿ āmma who are either wrongly accused of laḥn by the ḫāṣṣa or use correct forms which the ḫāṣṣa themselves use incorrectly. Ibn Makkī’s material is further proof that, as a genre, laḥn al-ʿāmma often includes errors found in the speech of edu- cated people and even established scholars. Five of his chapters, moreover, deal with laḥn that occurs in the speech and writing of five specific elitist groups, namely, Qurʾānic readers, scholars of Ḥadīṯ, jurists, writers of official

632 Zubaydī, Laḥn 78, 117, 202. 633 Ibid., 6–7. Cf. some of the words he mentions with Kisāʾī, Laḥn 101, 106, 111, 133. 634 Ibid., 125–126. For other similar occurrences, see 156, 181. 635 Ibn Makkī, Taṯqīf 46–47. For Ibn Makkī’s book, see also Rizzitano (1956). 636 Ibid., 44–45. 637 Chapters 31–34, pp. 275–301.

 

amgad
Highlight

 

181MUBAWWAB AND SPECIALIZED LEXICA

documents (ahl al-waṯāʾiq), and physicians.638 As a whole, the book consists of fifty chapters, most of which deal with various phonological, morphologi- cal and semantic changes pertaining to laḥn. Unlike other chapters, the first lists mistakes that arise from altering diacritical points (taṣḥīf ), although in certain cases it may well be that Ibn Makkī’s examples are influenced by the manner in which words are actually pronounced, as in tār<ṯa⁠ʾr (vengeance) and ǧadīda<ǧaḏīḏa (meal of parched barley or wheat).639 Only in this and the second chapter – which deals with tabdīl, i.e. substitution of consonants and semivowels640 – does Ibn Makkī arrange his material based on alphabetical order, irrespective of where taṣḥīf or tabdīl occur within words. The forty-eight other chapters completely lack arrangement.

Ḥarīrī’s Durrat al-ġawwāṣ fī awhām al-ḫawāṣṣ is one of the most famous works of the genre.641 The inclusion of the term ḫawāṣṣ (pl. of ḫāṣṣa) in the title is a reflection of its author’s assertion in his introduction that men of let- ters and eminent scribes rival (ḍāhaw) the ʿāmma in errors that occur in their speech and writing.642 Some of the books that did not survive precede Ḥarīrī’s in specifying the ḫāṣṣa/ḫawāṣṣ in their titles; e.g. Mā laḥana fīhi l-ḫawāṣṣ min al-ʿulamāʾ by Abū Aḥmad al-ʿAskarī (d. 382/993)643 and Laḥn al-ḫāṣṣa by Abū Hilāl al-ʿAskarī (d. after 395/1005).644 Such titles of books are more faithful to their content than, say, Zubaydī’s laḥn al-ʿawāmm, which, as noted earlier, includes material derived from both groups. Ḥarīrī includes in his book 222 items which, strangely enough, are neither thematically nor alphabetically arranged. About two centuries later, Ibn Manẓūr (d. 711/1311) arranged its items in alphabetical order in a book entitled Tahḏīb al-ḥawāṣṣ min Durrat al-ġawwāṣ,645 not surprisingly starting with the last then followed by the first radical as in his Lisān al-ʿArab.

Ǧawālīqī’s al-Takmila is also known as Laḥn al-ʿāmma, Mā talḥan fīhi l-ʿāmma, and Tatimmat Durrat al-ġawwāṣ.646 As the last title indicates, it is a supplement to Ḥarīrī’s book, although some of its material is mentioned

638 Chapters 35–39, pp. 302–336. A sixth group discussed in a separate chapter is ahl al-samāʿ (singers; 337–346).

639 Ibid., 53, 65. 640 Ibid., 84–115. 641 For a detailed study, see Ayoub (2001: 67–141). 642 Ḥarīrī, Durra 3; see also the book’s epilogue (283), where the term aʿyān (dignitaries) is

mentioned. 643 Qifṭī, Inbāh I, 346. 644 Ibn al-Ǧawzī, Taqwīm 75; Suyūṭī, Buġya I, 506. 645 ʿAbdaltawwāb (2000: 299). 646 Qifṭī, Inbāh III, 335; Suyūṭī, Buġya II, 308; Baġdādī, Ḫizāna I, 27, 50; VII, 65 (see also V, 132).

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

182 chapter 2

by Ḥarīrī.647 The material is divided into two parts without further arrange- ment. The first part includes words which have undergone semantic change in the usage of the ʿāmma and are introduced by the familiar phrase mimmā taḍaʿuhu l-ʿāmma ġayr mawḍiʿihi.648 The second part is longer and deals with words which exhibit phonological and morphological change, including augmentation, shedding of vowels or consonants, and substituting one pat- tern for another.649

The fifth item in the above list, Ibn Hišām al-Laḫmī’s al-Madḫal, represents the first serious attempt to defend the ʿāmma against allegations of laḥn by earlier authors. Its first chapter includes sixty-five items taken from Zubaydī’s Laḥn and its second sixty-two from Ibn Makkī’s Taṯqīf.650 Ibn Hišām rejects the views of both authors and argues that the usage of the ʿāmma is permis- sible based on the opinion of renowned scholars, such as Ḫalīl, Sībawayhi, Ibn al-Aʿrābī, Farrāʾ, Ṯaʿlab, Ibn Durayd, Ibn Ǧinnī, and Ibn Fāris. His main argument, which is incidentally mentioned in one of the items taken from Zubaydī,651 is that the ʿāmma should not be accused of laḥn as long as their usage agrees with an attested dialect (luġa masmūʿa), even if that dialect is not faṣīḥ (eloquent). In support of his argument, he cites, among others, Ḫalīl’s view that Arabic is too wide to allow for accusing a speaker of laḥn (luġat al-ʿArab akbar min an yulaḥḥan [ fīhā] mutakallim).652 Yet, Ibn Hišām finds in the language of the ʿāmma of his own time in Spain errors which defy justifica- tion (lā yaḥtamil al-ta⁠ʾwīl) and cannot be supported by the speech of the Arabs (wa-lā ʿalayhi min lisān al-ʿArab dalīl).653 Other than citing 417 such errors,654 he lists 830 instances of semantic errors committed by the ʿāmma655 and 112 expressions borrowed from poetry, which they either corrupt or are igno- rant of the verses from which they are extracted.656 Ibn Hišām’s material was arranged by another Maġribī author, Muḥammad b. Hāniʾ al-Laḫmī al-Sabtī (d. 733/1332), whose Inšād al-ḍawāll wa-iršād al-suʾʾāl was in turn abridged by

647 Cf., for example, items 141 and 157 in Ḥarīrī, Durra and their equivalents in Ǧawālīqī, Takmila 124–125, 88 respectively.

648 Ǧawālīqī, Takmila 5. 649 Ibid., 25–62. 650 Ibn Hišām, Madḫal 11–45, 46–71. 651 Item 26, p. 28. 652 Ibid., loc. cit. The editor wrongly has akṯar and yulḥan. 653 Ibid., 111; cf. 71. 654 Ibid., 111–146. 655 Ibid., 147–227. 656 Ibid., 228–253.

 

 

183MUBAWWAB AND SPECIALIZED LEXICA

a third Maġribī, Abū Ǧaʿfar Aḥmad b. ʿAlī b. Muḥammad Ibn Ḫātima al-Anṣārī (d. 724/1324), in Īrād al-la⁠ʾāl min Inšād al-ḍawāll. . . .657

The next two works also belong to the Maġribī tradition. Laḥn al-ʿawāmm fī mā yataʿallaq bi-ʿilm al-kalām by the Ašʿarī scholar, ʿUmar b. Muḥammad al-Sakūnī, is worth mentioning only because its title can wrongly suggest that it belongs to the genre under discussion. In fact, it contains some two hundred current expressions which are objectionable to the author on religious, rather than linguistic, grounds. One example is yā sākin al-samāʾ (O Dweller of the Sky), which he rejects because it conflicts with a Qurʾānic verse and a ḥadīṯ.658 Several examples are drawn from poetry, including that of Ḥallāǧ, Mutanabbī, Maʿarrī, Abū Nuwās and Maġribī poets.659 As for al-Ǧumāna – whose editor attributes it with no supporting evidence to the little-known ninth/fifteenth century author, Ibn al-Imām – it is a short treatise divided into fifteen chap- ters, most of which deal with phonological and morphophonological changes introduced by the ʿāmma. The final chapter, however, is devoted to semantic change.660 Much of the material cited is derived from earlier authors, such as Ibn Qutayba and Zubaydī.661

Each of Ibn al-Ḥanbalī’s three books contributes to the study of laḥn from a different perspective. Sahm al-alḥāẓ is a short supplement to Ḥarīrī’s al-Durra, and embraces 128 entries of no specific order. Errors are amended mostly by cit- ing earlier sources, including lexica such as Ǧawharī’s (d. c. 400/1010) al-Ṣaḥāḥ and Fīrūzābādī’s (d. 817/1415) al-Qāmūs.662 Ibn al-Ḥanbalī does not mention whether these errors occur in the speech of the ʿāmma or ḫāṣṣa, but at such a late date, it was increasingly difficult to make this distinction given the wide gap between current usage at any level and the norms of ʿArabiyya that apply to the classical era. A longer work, ʿIqd al-ḫalāṣ fī naqd kalām al-ḫawāṣṣ – an expansion of an earlier book by the same author, entitled al-Durr al-multaqaṭ fī tabyīn al-ġalaṭ663 – is also related to the Durra. The reference to the speech of the ḫawāṣṣ in its title reflects the fact that its content is based on Ḥarīrī’s mate- rial, which is mostly related to these ḫawāṣṣ. In the first of its two chapters,664

657 Ibn al-Imām, Ǧumāna, editor’s introduction, ṭ; ʿAbdaltawwāb (2000: 313); Wadġīrī (2008: 244); Gilliot (2012: 202–204).

658 Sakūnī, Laḥn 138–139. 659 Ibid., 146 ff. 660 Ibn al-Imām, Ǧumāna 35–40. 661 E.g. ibid., 7, 13, 16, 19, 21, etc. See also Wadġīrī (2008: 247–249). 662 Ibn al-Ḥanbalī, Sahm 18–21, 23–29, etc. See also Ḍāmin (1991c) for an additional 17

previously unpublished items. 663 Ibn al-Ḥanablī, ʿIqd 170. 664 Ibid., 171–261.

 

 

184 chapter 2

Ibn al-Ḥanbalī argues for the permissibility of usage rejected by Ḥarīrī and cites earlier lexicographers, grammarians and poets in support of his argu- ment. By contrast, he quotes in the second chapter665 some of Ḥarīrī’s laḥn material in which he agrees with his view, and provides further arguments in support of the rejection of the cited usage. In his third work, Baḥr al-ʿawwām fī mā aṣāba fīhi l-ʿawāmm, Ibn al-Ḥanbalī adduces arguments in defense of the ʿāmma in usage dismissed by earlier authors as laḥn. Unlike the previous book, the material (223 entries in all) is not restricted to Ḥarīrī’s al-Durra but is often derived from other works, most notably Ibn al-Sikkīt’s Iṣlāḥ al-manṭiq and Ibn Qutayba’s Adab al-kātib.666 Furthermore, some of Ibn al-Ḥanbalī’s entries are not attributed to earlier works, and it is specifically in these entries that an unprecedented tolerance of laḥn emerges. One of the most stunning examples is the one in which he justifies the omission of the final nūn from the indic- ative forms of what the grammarians call al-afʿāl al-ḫamsa (the five verbs), such as yafʿalū<yafʿalūna and tafʿalī<tafʿalīna.667 Other similarly astounding examples include the justification of using the jussive instead of the indica- tive, as in ya⁠ʾkul<ya⁠ʾkulu, the omission of hamza from the first person singu- lar independent pronoun after the conjunction wāw, i.e. wanā<wa-anā, and the use of the plural verb instead of the dual (e.g. fulānun wa-fulānun ǧāʾūnī, instead of ǧāʾānī).668 Such examples show the dramatic increase in the level of tolerance of laḥn, so to speak, ever since Ibn Hišām’s attempt to justify some of the material considered by earlier authors as laḥn, and may be regarded as the practical application of a motto attributed to one of the earliest grammarians, Abū l-Ḫaṭṭāb al-Aḫfaš al-Kabīr (d. 177/793), namely, that the most proficient grammarian is the one who never describes anyone’s speech as laḥn (anḥā l-nās man lam yulaḥḥin aḥadā).669 It is extremely doubtful however whether Aḫfaš had foreseen the type of changes which Ibn al-Ḥanbalī defended follow- ing eight centuries of linguistic development!

Before turning to the books that are alphabetically arranged, note should be made of works which deal with a particular category of scholars. Such works clearly address a specific branch of ḫāṣṣa which the authors are interested in. This is in contrast to more general works which use the term rather vaguely to refer to educated people in general, or to eminent writers, men of letters, functionaries, etc. It has already been pointed out that Ibn Makkī includes in

665 Ibid., 263–359. 666 Ibn al-Ḥanbalī, Baḥr, Index, 301–302, 306. 667 Ibid., 133–134. 668 Ibid., 138–139, 148–149. 669 Ibn Hišām, Madḫal 28.

 

 

185MUBAWWAB AND SPECIALIZED LEXICA

his Taṯqīf al-lisān five chapters, each of which deals with errors that are spe- cific to a particular group of scholars or educated people. Ibn Makkī, however, seems to continue an earlier tradition rather than establish a new line of inves- tigation. To ʿUmar b. Šabba (d. 262/876) is attributed al-Naḥw wa-man kāna yalḥan min al-naḥwiyyīn670 – an early work which suggests that masters of the field that was reportedly established to combat laḥn soon became victims of it. The book did not survive but its title suggests that it was most probably not alphabetically arranged and that the errors of individual grammarians were surveyed in separate sections, perhaps in the manner we know from Abū Aḥmad al-ʿAskarī’s (d. 382/993) Šarḥ mā yaqaʿ fīhi l-taṣḥīf wa-l-taḥrīf. Another work that did not survive and of which we know very little is Mā laḥana fīhi l-ḫawāṣṣ min al-ʿulamāʾ, also by Abū Aḥmad al-ʿAskarī.671 We are however in possession of Ḫaṭṭābī’s (d. 388/998) Iṣlāḥ ġalaṭ al-muḥaddiṯīn and Ibn Barrī’s (d. 582/1187) Ġalaṭ al-ḍuʿafāʾ min al-fuqahāʾ – two short monographs which are not alphabetically arranged. Ḫaṭṭābī’s consists of 143 entries containing words which most scholars of Ḥadīṯ transmit in a distorted manner (malḥūna wa-muḥarrafa) or by adopting the weaker of two possible readings.672 As for Ibn Barrī’s, it embraces ninety-eight words which are erroneously used by those jurists whom he describes as weak (ḍuʿafāʾ).673 Although Ibn Barrī does not acknowledge his sources, the book’s editor identifies about fifty words taken from Ibn Makkī’s Taṯqīf al-lisān, thirty from Zubaydī’s Laḥn al-ʿawāmm, and ten from Ḥarīrī’s al-Durra.674

Alphabetically arranged works in this genre are significantly fewer than those that are not alphabetically arranged. The main reason for this most probably stems from the fact that most authors were interested in highlighting the types of changes embedded in laḥn material. This obviously applies to Ibn Makkī, most of whose fifty chapters deal with specific types of change – such as vowel omission, insertion or shift – as well as changes that pertain to certain morphological patterns and categories – such as active and passive participles, masculine and feminine, singular and plural, diminutive, gentilic adjective (nisba), etc. Any alphabetical arrangement would have certainly obliterated the common features among the host of examples included in any of these chapters. Another reason for the relative scarcity of alphabetically arranged

670 Yāqūt, Muʿǧam V, 2903; Suyūṭī, Buġya II, 219. The title given by Ibn al-Nadīm in Fihrist 125 is al-Istiʿẓām li-l-naḥw wa-man kāna . . .

671 Qifṭī, Inbāh I, 346; cf. above, 181. 672 Ḫaṭṭābī, Iṣlāḥ 19. 673 Ibn Barrī, Ġalaṭ 14. 674 See the editor’s introduction, 9.

 

amgad
Highlight

 

186 chapter 2

works is that several authors wrote supplements to or amendments of earlier works and must have found it feasible to follow the same arrangement as the original authors, such as Ibn Makkī and Ḥarīrī. This notwithstanding, some authors gave preference to the ease of use of alphabetically arranged mate- rial over the thematic type of arrangement. Only a few decades after Zubaydī (d. 379/989), Ibn Šuhayd (d. 426/1035), as noted earlier,675 alphabetically rearranged his Laḥn al-ʿawāmm based on the first letter of each word. Ibn Makkī himself also deemed it advantageous in his first two chapters to list his material in partial alphabetical order.676 The following four books are also alphabetically arranged:

1. Taqwīm al-lisān by Ibn al-Ǧawzī (d. 597/1201). 2. Taṣḥīḥ al-taṣḥīf wa-taḥrīr al-taḥrīf by Ṣafadī (d. 764/1363). 3. al-Tanbīh ʿalā ġalaṭ al-ǧāhil wa-l-nabīh by Ibn Kamāl Bāšā (d. 940/1534). 4. Ḫayr al-kalām fī l-taqaṣṣī ʿan aġlāṭ al-ʿawāmm by Ibn Bālī al-Qusṭanṭīnī

(d. 992/1584).

The title of Ibn al-Ǧawzī’s book is reminiscent of Ibn Hišām’s and ultimately goes back to Ibn Qutayba’s chapter by the same title in Adab al-kātib. In the introduction, Ibn al-Ǧawzī declares that he had intended to arrange his mate- rial according to types of error, but then decided to adopt an alphabetical order so as to facilitate the book’s use.677 He also opted to list as entries the correct, rather than incorrect, forms as this would also contribute in his view to ease of reference. It should be noted however that this option presupposes that the user knows the correct form. Such a presupposition can often contribute to the difficulty one encounters in searching for words whose correct forms one does not know, particularly when the correct and incorrect forms do not share the same sequence of consonants (let alone the fact that the arrange- ment of words does not go beyond their first letters). The roots of words, fur- thermore, are ignored, hence adlaǧa (to journey at night), tarquwa (clavicle), and miftāḥ are listed under alif, tāʾ and mīm respectively.678 This often pres- ents another difficulty for the user, who, for example, is supposed to know that al-šāt taǧtarr (The ewe ruminates)679 – corrupted by the ʿāmma to taštarr – is placed under tāʾ because the author chose the form taǧtarr, which starts

675 See above, 179, n. 627. 676 See above, 181. 677 Ibn al-Ǧawzī, Taqwīm 74. 678 Ibid., 79, 105, 182. 679 Ibid., 104.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

187MUBAWWAB AND SPECIALIZED LEXICA

with the pronominal prefix ta-, rather than, say, yaǧtarr had the agent been a masculine noun (e.g. al-ṯawr yaǧtarr “The ox ruminates”)! As far as content is concerned, Ibn al-Ǧawzī himself admits in the introduction that he chose it from works of earlier authors of books on laḥn, such as Farrāʾ, Aṣmaʿī, Abū ʿUbayd, Abū Ḥātim, Ibn al-Sikkīt, etc. and that his own contribution is lim- ited to arrangement and abridgement (al-tartīb wa-l-iḫtiṣār).680 Like some of his predecessors, he generally uses the term ʿāmma in spite of the fact that he asserts that many of those whom he describes as educated (al-muntasibūn ilā l-ʿilm), i.e. the ḫāṣṣa, have adopted the inferior (marḏūl) usage of the ʿ awāmm.681

The lengthy introduction of Ṣafadī’s book is almost fully devoted to errors that visually arise from taṣḥīf (altering diacritical points) and includes a host of interesting examples which demonstrate the pitfalls that can be caused by defective writing. The more general term, taḥrīf (altering word forms), is often used by Ṣafadī in conjunction with taṣḥīf and with no apparent difference in meaning.682 Yet the majority of the 1,985 entries in the book belong to the realm of laḥn and have nothing to do with visually based errors. Ṣafadī’s rev- elation of his sources in the introduction, and his reference to them in every entry by letters that stand for the authors or their books, show that the mate- rial related to taṣḥīf/taḥrīf is mostly drawn from several works that particularly deal with this phenomenon, and at least three of his nine sources fall into this category.683 These are Mā ṣaḥḥafa fīhi l-Kūfiyyūn by Ṣūlī (d. 335/946), al-Tanbīh ʿalā ḥudūṯ al-taṣḥīf by Ḥamza al-Iṣfahānī (d. after 351/962), and al-Taṣḥīf 684 by Abū Aḥmad al-ʿAskarī (d. 382/993). Of the other six sources, five deal with laḥn proper and these have been discussed above, namely, the books of Zubaydī, Ibn Makkī, Ḥarīrī, Ǧawālīqī and Ibn al-Ǧawzī. For the sixth work, attributed to a certain al-Ḍiyāʾ Mūsā l-Nāsiḫ al-Ašrafī, Ṣafadī gives no title, but his refer- ences to it indicate that it dealt with laḥn, rather than taṣḥīf.685 Ṣafadī, unlike Ibn al-Ǧawzī, arranges his entries alphabetically on the basis of the erroneous

680 Ibid., 74–75. 681 Ibid., 73. 682 See, for example, Ṣafadī, Taṣḥīḥ 27 (mimmā yuṣaḥḥaf wa-yuḥarraf ) and the various

possibilities of reading certain words based on the diacritical points chosen by the reader. See, however, ibid., 53 where taḥrīf refers specifically to vocalic change in the proper noun Ḥirāʾ (mountain in Mecca).

683 Ibid., 60–65; cf. Gilliot (1989: 275–276). 684 Comparison between Ṣafadī’s Taṣḥīḥ (e.g. 207, 326, 377) and Abū Aḥmad al-ʿAskarī’s Šarḥ

mā yaqaʿ fīhi l-taṣḥīf wa-l-taḥrīf (163, 105, 108 respectively) proves that it is this particular book by ʿAskarī that Suyūṭī refers to, and not one of his two other books, Taṣḥīfāt al-muḥaddiṯīn and Aḫbār al-muṣaḥḥifīn.

685 Ṣafadī, Taṣḥīḥ 164, 169, 188, 206, 209, 219, 234, etc.

 

 

188 chapter 2

forms, which is more convenient for the reader. His system takes into consid- eration all the letters of the words (i.e. not roots), although he often does not adhere to it strictly (e.g. zumurrud, ṭarfa and muṯnā occur before zummuǧ, ṭarsūs and maṯbūt respectively).686

In the introduction to his short treatise, al-Tanbīh, Ibn Kamāl Bāšā explains that he does not go beyond the second letter of the word in his alphabetical arrangement in order not to increase the number of sections and chapters.687 This statement, however, is of little practical value since the relatively few words which the book contains (only a little more than one hundred in his own estimation)688 are spread over twenty-three chapters and the user can hardly miss the required word. Arrangement apart, the book is interesting because at such a late stage in the tradition, the author primarily relies on material he had personally recorded from the speech of his own companions and fellows (aṣḥāb, iḫwān).689 Moreover, he asserts that some of the errors he collected are specifically used by the ḫāṣṣa whereas others occur only in the speech of the ʿāmma.690 To the exclusion of words that are considered by some philologists, but not all, as impermissible and those that philologists unanimously reject but are widespread among writers or authors (ahl al-taṣnīf ), Ibn Kamāl Bāšā’s alphabetical list is restricted to words which he describes as saqaṭ (evil speech; lit. refuse) and for which he finds no grounds for justification.691

Finally, Ibn Bālī al-Qusṭanṭīnī’s Ḫayr al-kalām is a collection of 223 words whose forms or meanings are distorted by the ʿawāmm. It is arranged alpha- betically on the basis of the first letter of the word regardless of its root, and almost in every entry, Ibn Bālī acknowledges the author(s) of books on laḥn and/or the lexicographer(s) from whom he derives his information.

7 al-Aḍdād (Words with Two Contradictory Meanings)

Arab philologists refer to words which have two contradictory (i.e. opposed or mutually exclusive) meanings as aḍdād (pl. of ḍidd). They regard such words

686 Ibid., 296, 364, 465. 687 Ibn Kamāl Bāšā, Tanbīh 54. 688 Ibid., 53. The actual number of words, apart from illustrations in the introduction, is

ninety-six. 689 Ibid., 49. 690 Ibid., 54. 691 Ibid., 53, 59.

 

amgad
Highlight
amgad
Highlight

 

189MUBAWWAB AND SPECIALIZED LEXICA

as a particular type of homonymous polysemic words (nawʿ min al-muštarak)692 since their meanings are not merely different (e.g. ʿamm, which means either “paternal uncle” or “company of men”)693 but contradictory (e.g. nāhil, which means either “thirsty” or “sated”, and ǧudd, which means either “a well abound- ing with water” or “one containing little water”).694 Unlike several other top- ics of investigation in muǧannas lexica, the study of aḍdād almost exclusively belongs to the realm of luġa, rather than naḥw. Early grammatical works almost completely ignore aḍdād, as well as words which represent other semantic notions, such as synonymous (mutarādif ) and homonymous poly- semic (muštarak) words of which aḍdād are a special branch.695 Sībawayhi’s (d. 180/796) position in this respect seems to have had a great impact on later grammarians. In the introductory part of his book, known as the Risāla, he briefly points out three types of semantic relationships, which the reader later discovers are not among his analytical tools and do not feature elsewhere in the grammatical theory he adopts. The mention of these three types in the intro- duction is thus most probably intended to demarcate the boundaries between them and the issues – essentially naḥw (syntax) and ṣarf (morphology) – with which grammar is concerned. Without using any of the technical terms which later came to designate the second and third types (i.e. tarāduf/mutarādif and ištirāk/muštarak), he explains that the speech of the Arabs includes divergence of form and meaning (iḫtilāf al-lafẓayn li-ḫtilāf al-maʿnayayn), divergence of form and coincidence of meaning (iḫtilāf al-lafẓayn wa-l-maʿnā wāḥid), and coincidence of form and divergence of meaning (ittifāq al-lafẓayn wa-ḫtilāf al-maʿnayayn).696 Sībawayhi’s example for the third type is waǧada, which can mean either “to be angry with” (e.g. waǧadtu ʿalayhi “I was angry with him”) or “to find a lost thing” (e.g. waǧadtu l-ḍāllata “I found my object of pursuit”). Another prominent grammarian, Mubarrad (d. 285/898), adopts Sībawayhi’s division and even cites the same example (i.e. waǧada) for the third type.697 Mubarrad, however, adds for that type several examples which demonstrate that two words which agree in lafẓ can express two contradictory meanings. The examples he cites (e.g. ǧawn for “black” and “white”) are among those cited by several authors of the genre. It is interesting to note however that Mubarrad’s

692 Suyūṭī, Muzhir I, 387. 693 Ibn Durayd, Ǧamhara I, 157; cf. Suyūṭī, Muzhir I, 37o. 694 Ibn al-Anbārī, Aḍdād 116, 206. 695 Works which deal with mutarādif and muštarak will be discussed in Section 8 of this

chapter. 696 Sībawayhi, Kitāb I, 24. 697 Mubarrad, Mā ttafaqa 3.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

190 chapter 2

reference to the three types of semantic relationships, including his examples on aḍdād, provides an essential introduction to his short monograph entitled Mā ttafaqa lafẓuhu wa-ḫtalafa maʿnāhu min al-Qurʾān al-maǧīd, which is not a work on naḥw. In contrast, its occurrence in the chapter of biliteral words in al-Muqtaḍab, his major work on naḥw, is merely accidental and digressive. Commenting on the fact that ʿalā can be either a verb or a preposition, he cites in one of al-Muqtaḍab’s chapters the three types (with no reference to aḍdād) in order to show that ʿalā belongs to the third (i.e. ittifāq al-lafẓayn wa-ḫtilāf al-maʿnayayn), and then clarifies that the chapter will be resumed following this digression (fa-hāḏā ʿāriḍ fī l-kitāb ṯumma naʿūd ilā l-bāb).698

Early interest in aḍdād is certainly related to Qurʾānic exegesis, as is evident in some of the earliest books of the genre. Quṭrub (d. 206/821) for example begins his monograph on aḍdād by demonstrating how, in a Qurʾānic context, certain words have to be understood as having a specific meaning. One such word is ẓanna, which in the verse innī ẓanantu annī mulāqin ḥisābiyah (I did really know that I should come to my account; Q 69: 20) does not have the (more common) meaning of “to suppose; to presume; to opine” since this would indi- cate doubt (šakk) on the part of the believer and thus lead to disbelief (kufr). To avert such an interpretation, ẓanantu here has to be understood as indicating certainty ( yaqīn),699 hence its inclusion in aḍdād. Another author, Abū Ḥātim al-Siǧistānī (d. 255/869), quotes the same verse in the first few lines of his K. al-Aḍdād and asserts that the reason for authoring his book (ḥamalanā ʿalā ta⁠ʾlīfihi) is to inform the one who does not master Arabic usage that a verb such as ẓanna in the Qurʾān can mean ayqana (to know for certain).700 Dogmatic reasons of this kind prompted some philologists to consider certain words as aḍdād (referred to by Cohen as aḍdād théologiques),701 thus contributing to a remarkable increase in the number of words that are claimed to be aḍdād.

From a cultural point of view, the Šuʿūbiyyūn (i.e. those who exalt mainly Persian culture over Arabic culture) claimed that the existence of aḍdād in Arabic has led to obscurity and confusion. This gave rise to intense discussion in some books of the genre and other philological works as well. To begin with, it should be noted that Arab philologists differ concerning the very existence of aḍdād. The first reference to a word with two contradictory meanings (although without the term aḍdād) is probably the one mentioned in K. al-ʿAyn. Ḫalīl notes that šaʿb can mean either tafarruq (separation) or taǧammuʿ (gathering)

698 Mubarrad, Muqtaḍab I, 46. 699 Quṭrub, Aḍdād 244. 700 Abū Ḥātim, Aḍdād 128. 701 Cohen (1961: 23); for further Qurʾānic examples, see 19–23.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

191MUBAWWAB AND SPECIALIZED LEXICA

and considers the phenomenon to be one of the astonishing features of speech (min ʿaǧāʾib al-kalām) and indicative of the extensiveness (wusʿ) of Arabic.702 This accommodating position may well be an indirect refutation of the Šuʿūbī claims concerning the occurrence of aḍdād in Arabic – a tempting possibil- ity given that Ḫalīl’s comment occurs only a few lines after his explanation of the meaning of the term Šuʿūbī. Other scholars who acknowledge their existence include Mubarrad (quoted above) and Ibn Fāris (d. 395/1004), who believes that the use of one noun to express two contradictory mean- ings is one of the norms of the speech of the Arabs (wa-min sunan al-ʿArab fī l-asmāʾ an yusammū l-mutaḍāddayn bi-sm wāḥid).703 The position of Ibn Durayd (d. 321/933) is more qualified, for whereas he cites several aḍdād in al-Ǧamhara,704 he denies the possibility that a word which has two contradic- tory meanings, not occurring in one and the same dialect can be included in aḍdād.705 On another occasion, he expresses the wish not to comment on a word which was claimed to be an example of aḍdād in the Qurʾān.706 There are scholars, however, who vehemently deny the existence of aḍdād. Most promi- nent among these is Ibn Durustawayhi (d. 347/958) who points out in his com- mentary on Ṯaʿlab’s al-Faṣīḥ, known as Taṣḥīḥ al-Faṣīḥ, that he has authored a book entitled Ibṭāl al-aḍdād,707 in which he argues against those who admit the existence of aḍdād. Ibn Sīda (d. 458/1066) also reports the denial by one of Abū ʿAlī al-Fārisī’s (d. 377/987) teachers of the aḍdād which philologists often discuss.708 For his part, Ǧawālīqī (d. 540/1145) reports a similar view by Ṯaʿlab (d. 291/904) and obviously sides with him and other proponents of this view by referring to them as the truly expert scholars of Arabic (al-muḥaqqiqūn min ʿulamāʾ al-ʿArabiyya) and by citing several examples of words which are claimed to be aḍdād but which can be interpreted differently.709 We also learn from Ṣaġānī (d. 650/1252) of the existence of a number of books (kutub muṣannafa) whose authors refute the ones who believe in the existence of aḍdād.710 Invalidating aḍdād in such works was most probably prompted by an Arab reaction to the claims of the Šuʿūbiyyūn, although it is interesting

702 Ḫalīl, ʿAyn I, 263. 703 Ibn Fāris, Ṣāḥibī 97; quoted also in Suyūṭī, Muzhir I, 387. 704 The total number of aḍdād in al-Ǧamhara is thirty-three; see Index I (III, 1735). 705 Ibid., I, 343; cf. Suyūṭī, Muzhir I, 396. 706 Ibn Durayd, Ǧamhara I, 457. 707 Ibn Durustawayhi, Taṣḥīḥ I, 359. 708 Ibn Sīda, Muḫaṣṣaṣ XIII, 259. 709 Ǧawālīqī, Šarḥ 251. 710 Ṣaġānī, Aḍdād 222.

 

 

192 chapter 2

to note that the author of Ibṭāl al-aḍdād, Ibn Durustawayhi, was himself of Iranian descent. It may well be, as Blachère suggests, that Ibn Durustawayhi’s case is one of “hyper-arabisme” in which Persian authors vigorously defend what was later called ʿurūba.711

A more subtle approach in the defense of Arabic against Šuʿūbī claims was the attempt by some scholars to justify or qualify the existence of aḍdād. The clearest text in this respect is Abū Bakr b. al-Anbārī’s (d. 328/940) introduction to his K. al-Aḍdād, which starts with laying down the claim of those who dis- parage the Arabs (ahl . . . al-izrāʾ ʿ alā l-ʿArab) that the presence of aḍdād in their language is due to their lack of wisdom (nuqṣān ḥikmatihim), deficiency of elo- quence (qillat balāġatihim), and abundance of ambiguity in their dialogues (kaṯrat al-iltibās fī muḥāwarātihim).712 In refutation of this claim, Ibn al-Anbārī adduces several arguments, foremost among which is that context eliminates any ambiguity since no sensible or judicious person would fail to know which of the two meanings of ǧalal (i.e. a formidable or a contemptible matter) is meant by the speaker.713 He also quotes other scholars who assert that if a word has two contradictory meanings, one of them represents the aṣl (origin), whereas the other is the result of ittisāʿ (latitude/extension of speech) – a notion that features prominently in grammatical works, particularly Sībawayhi’s al-Kitāb.714 A third argument justifies the existence of some aḍdād by ascribing each of the two directly opposed meanings to a different tribe (ḥayy), before interaction caused the occurrence of both in one dialect.715 In contrast to words which differ in form and meaning and those which express the same meanings (i.e. synonyms), words with more than one meaning (i.e. homonymous polysemic words, including aḍdād) are, according to Ibn al-Anbārī, quite few though interesting (al-qalīl al-ẓarīf ) in the speech of the Arabs.716 Yet Ibn al-Anbārī includes among his 357 items a large number of words which are not truly aḍdād. Indeed, this phenomenon is so widespread that some authors – e.g. Abū l-Ṭayyib al-Luġawī (d. 351/962)717 – bitterly complain that earlier authors

711 Blachère (1967: 400). 712 Ibn al-Anbārī, Aḍdād 1. For further discussion of the claims of the Šuʿūbiyyūn with respect

to aḍdād, see Versteegh (1981: 412–413). 713 Ibid., 2–3. 714 Note that the absence of ambiguity (labs) is a necessary condition for any construction

which Sībawayhi considers to be an instance of what he calls saʿat al-kalām; cf. Baalbaki (2008: 204–205) for several examples.

715 Ibn al-Anbārī, Aḍdād 11–12. 716 Ibid., 6; cf. Quṭrub’s expression (Aḍdād 244): wa-innamā ḫaṣaṣnāhu bi-l-iḫbār ʿanhu

li-qillatihi fī kalāmihim wa-ẓarāfatihi. 717 Abū l-Ṭayyib, Aḍdād I, 2.

 

 

193MUBAWWAB AND SPECIALIZED LEXICA

of the genre have included in their material words that do not belong to aḍdād. Ṣaġānī succinctly formulates this issue by saying that such material is unnatu- ral (yanbū l-ṭabʿ ʿanhu) because it does not fit the definition of aḍdād (li-buʿdihi ʿan ḥadd al-aḍdād).718 Although Ṣaġānī does not explain the elements which form the definition of aḍdād, it is clear that many authors extend the notion of aḍdād too far indeed and that many of the words they cite are, to use Cohen’s expression, faux aḍdād.719 Without going into details,720 a sizeable number of the so-called aḍdād can be dismissed on various grounds, including morpho- logical pattern, type of preposition used with verbs, alteration of diacritical points (taṣḥīf ), as well as stylistic reasons such as the ironic or euphemistic use of words.

Several books on aḍdād that are mentioned in bibliographical and bio- graphical sources have not been preserved.721 Among these in the third/ninth and fourth/tenth centuries are works by Abū ʿUbayda (d. 209/824),722 Ṯaʿlab (d. 291/904),723 Āmidī (d. 370/980),724 and Ibn Fāris (d. 395/1004).725 There are also chapters on aḍdād in multithematic works, such as Abū ʿUbayd’s (d. 224/838) al-Ġarīb al-muṣannaf (but see discussion of the third item below), Ibn Qutayba’s (d. 276/889) Adab al-kātib, in which twenty-seven aḍdād are mentioned,726 and Kurāʿ al-Naml’s (d. 310/922) al-Muntaḫab min ġarīb kalām al-ʿArab, which includes 183 aḍdād.727 Ibn Sīda (d. 458/1066) lists about a

718 Ṣaġānī, Aḍdād 222. 719 Cf. Cohen (1961: 7). 720 For details, see Cohen (1961: 7 ff.); Weil (1979: 184); ʿAbdaltawwāb (1980: 339–357). For a

book-length study of aḍdād, see Āl Yāsīn (1974). 721 Note that Sezgin (1982: 99) mentions a K. al-Aḍdād by Abū ʿAlī ʿUbayd b. Ḏakwān, a

contemporary of Mubarrad’s (d. 285/898), based on Ibn al-Nadīm’s Fihrist. But it is clear in the Fihrist (p. 65) that the book referred to by Ibn al-Nadīm is Tawwazī’s and that it was narrated by Abū Ḏakwān al-Qāsim b. Ismāʿīl. Also noteworthy is that Abū ʿAlī’s name in Yāqūt’s Muʿǧam (IV, 1622) appears as ʿAsal b. Ḏakwān and that Yāqūt mentions the same two titles attributed to him by Ibn al-Nadīm (i.e. al-Ǧawāb al-muskit and Aqsām al-ʿArabiyya) but not a book on aḍdād.

722 Ibn al-Nadīm, Fihrist 59; Yāqūt, Muʿǧam VI, 2708; Ibn Ḫillikān, Wafayāt V, 239; Qifṭī, Inbāh III, 286.

723 Ibn Ḫayr, Fahrasa 381 (ǧuzʾ fīhi l-aḍdād). 724 Yāqūt, Muʿǧam II, 851; Ṣafadī, Wāfī XI, 409 (in both, K. al-Ḥurūf min al-uṣūl fī l-aḍdād);

Suyūṭī, Buġya I, 501. 725 This book is not mentioned in the sources, but Ibn Fāris himself refers to it in Ṣāḥibī 98. 726 Ibn Qutayba, Adab 177–181. 727 Kurāʿ, Muntaḫab II, 584–593.

 

amgad
Highlight
amgad
Highlight

 

194 chapter 2

hundred aḍdād and acknowledges his indebtedness in particular to Abū ʿAlī al-Fārisī (d. 377/987), but also quotes several earlier authors of the genre.728

The following works on aḍdād are preserved:

1. al-Aḍdād by Quṭrub (d. 206/821). 2. al-Aḍdād by Aṣmaʿī (d. 216/831). 3. al-Aḍdād by Abū ʿUbayd (d. 224/838). 4. al-Aḍdād by Tawwazī (d. 233/847). 5. al-Aḍdād by Ibn al-Sikkīt (d. 244/858). 6. al-Aḍdād (full title K. al-Maqlūb lafẓuhu fī kalām al-ʿArab wa-l-muzāl ʿan

ǧihatihi wa-l-aḍdād) by Abū Ḥātim al-Siǧistānī (d. 255/869). 7. al-Aḍdād by Abū Bakr b. al-Anbārī (d. 328/940). 8. al-Aḍdād fī kalām al-ʿArab by Abū l-Ṭayyib al-Luġawī (d. 351/962). 9. al-Aḍdād fī l-luġa by Ibn al-Dahhān (d. 569/1174). 10. al-Aḍdād by Ṣaġānī (d. 650/1252). 11. Risālat al-aḍdād by Munšī (d. 1001/1592).729

Quṭrub, a famous student of Sībawayhi’s, begins his book with the same dis- tinction as in the Kitāb between three types of semantic relationships, the third of which is coincidence of form and divergence of meaning.730 Other than homonymous polysemic words to which Sībawayhi refers, Quṭrub explains that this third type includes words which have two contradictory meanings. Of these aḍdād, Quṭrub lists 218 for which he frequently adduces textual evidence (šawāhid) mostly from Qurʾān and poetry.

To Aṣmaʿī is attributed K. al-Aḍdād in several sources,731 but there is strong reason to believe that the text published by Haffner and attributed to Aṣmaʿī is not his. In fact, Haffner himself noted the great similarity between Aṣmaʿī’s book and that by Ibn al-Sikkīt (no. 5 in the above list) and concluded that the latter is merely another riwāya of the first.732 But the fact that the text is replete with names of scholars who are rarely quoted by Aṣmaʿī in his other works, in particular his two adversaries, Abū ʿUbayda (d. 209/824) and Ibn al-Aʿrābī

728 Ibn Sīda, Muḫaṣṣaṣ XIII, 258–266. 729 For books beyond this date, see Aḥmad’s introduction to Abū Ḥātim, Aḍdād 86–90; Āl

Yāsīn’s introduction to the three books on aḍdād by Abū ʿ Ubayd, Tawwazī and Munšī 7–11; Iqbāl (2011: 293–301).

730 Quṭrub, Aḍdād 243–244. 731 E.g. Ibn al-Nadīm, Fihrist 61; Ibn Ḫillikān, Wafayāt III, 176. 732 See Haffner’s footnote in Ibn al-Sikkīt, Aḍdād 163.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

195MUBAWWAB AND SPECIALIZED LEXICA

(d. 231/845),733 makes it much more likely that the book ascribed by Haffner to Aṣmaʿī is another riwāya of Ibn al-Sikkīt’s book, and not the other way round. The fact that the book ascribed to Aṣmaʿī starts with the expression qāla l-Aṣmaʿī 734 probably prompted the copyist to attribute the whole book to him. The same item which follows that expression however is also the first item in Ibn al-Sikkīt’s book and is attributed to Abū Saʿīd (i.e. Aṣmaʿī) following the opening statement qāla . . . Ibn al-Sikkīt.735 This strongly argues in favor of attributing the work to Ibn al-Sikkīt and not to Aṣmaʿī. The number of words is 105 in Aṣmaʿī’s text and ninety-four in Ibn al-Sikkīt’s, and the šawāhid from Qurʾān and poetry are almost identical in both.

Abū ʿUbayd’s al-Aḍdād is in fact the chapter entitled Kitāb al-Aḍdād in his al-Ġarīb al-muṣannaf,736 but is often erroneously referred to by researchers as an independent work,737 probably following Suyūṭī who ascribes to Abū ʿ Ubayd a book on aḍdād, unlike all other sources.738 The number of aḍdād items in this chapter is thirty-eight, and among the earlier scholars from whom Abū ʿUbayd derives his material and šawāhid are Kisāʾī (d. 189/805), Abū ʿUbayda, Abū Zayd al-Anṣārī (d. 215/830) and Aṣmaʿī. The last three scholars are also frequently quoted in Tawwazī’s book, which contains eighty items and many of the šawāhid of his predecessors. The book is cited by Tawwāzī’s student, Mubarrad (d. 285/898), who quotes two of his aḍdād.739

As previously mentioned, Abū Ḥātim clearly cites the proper understand- ing of Qurʾānic words as the reason for authoring his book. He lists 170 items, some of which are derived from earlier scholars (again, most notably, Abū ʿUbayda, Abū Zayd and Aṣmaʿī), and frequently assigns the dialects in which they appear.740 On several occasions, however, he cites words which have two contradictory meanings but expresses his doubt whether these words are actually used or not (a-tuqāl am lā).741 Abū Ḥātim most probably doubts such words, either because they were not reported by a reliable authority, or because he has not heard them being used by the Arabs. It is noteworthy that the full title of the book is K. al-Maqlūb lafẓuhu fī kalām al-ʿArab wa-l-muzāl ʿan

733 Cf. ʿAbdaltawwāb (1980: 238–239); Šalqānī (1971: 352–354). 734 Aṣmaʿī, Aḍdād 5. 735 Ibn al-Sikkīt, Aḍdād 163. 736 Abū ʿUbayd, Ġarīb II, 622–634. 737 See Weil (1979: 184); Aḥmad’s introduction to Abū Ḥātim, Aḍdād 86; Bettini (2006: 627). 738 Suyūṭī, Muzhir I, 581; II, 249; cf. Āl Yāsīn’s introduction to Abū ʿUbayd’s Aḍdād 27–29. 739 Mubarrad, Kāmil I, 255; III, 228; cf. Tawwazī, Aḍdād 100–101, 80. 740 Abū Ḥātim, Aḍdād 142, 172, 243, etc. 741 Ibid., 240, 244–245, 247.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight

 

196 chapter 2

ǧihatihi wa-l-aḍdād. What is meant by the first two terms becomes clear when one compares Abū Ḥātim’s material with the last chapter of Abū l-Ṭayyib’s K. al-Aḍdād, entitled Bāb takallamat bihi l-ʿArab maqlūb al-maʿnā muzālan ʿan ǧihatihi fa-ḫuliṭa bi-l-aḍdād wa-laysa minhā.742 In this chapter, Abū l-Ṭayyib demonstrates that aḍdād can be confused with words whose syntactic posi- tions or semantic relationships are reversed, as in the hemistich ilā an taġība l-šamsu min ḥayṯu taṭluʿu (until the sun sets from the place of sunrise), by which is meant taṭluʿ al-šams min ḥayṯu taġīb.743 Such “pseudo-aḍdād” occur mostly in constructions744 which both authors describe as maqlūb lafẓuhu/ maqlūb al-maʿnā (of reversed form or meaning) or muzāl ʿan ǧihatihi (altered from its origin), and both terms must be understood as referring to a single phenomenon rather than two different ones. The link made by the two authors between aḍdād and the notion of reversal or alteration from origin strongly indicates that, to them, the occurrence of aḍdād is essentially an anomalous phenomenon which deviates from an origin (aṣl). This origin is most probably related to muštarak, that is, homonymous polysemic words which do have two or more meanings but which are not contradictory to each other.

Ibn al-Anbārī’s 357 aḍdād items – the largest collection in the genre – are extensively explained and supported by numerous šawāhid. A significant num- ber of these items, however, are farfetched and belong to the previously men- tioned faux aḍdād, which can be explained away on various grounds. Ironic as it may seem, Ibn al-Anbārī intended, as part of his defense of the Arabs against the Šuʿūbiyyūn, to list as many aḍdād as he possibly could. The main argument of the Šuʿūbiyyūn, as noted earlier, is that the occurrence of aḍdād in the speech of the Arabs is indicative of their lack of wisdom, deficiency of eloquence, and abundance of ambiguity in dialogue. Therefore, Ibn al-Anbārī must have wanted to demonstrate, through lengthy explanations in most cases, that the applicability to a large body of material of the justifications for aḍdād, mentioned in his introduction, is in itself a powerful counterargument which invalidates Šuʿūbī claims against the Arabs. As part of his attempt to increase

742 Abū l-Ṭayyib, Aḍdād II, 720–732. 743 Ibid., II, 723. The construction nāʾa bī l-ḥimlu, which is said to be the maqlūb of nuʾtu

bi-l-ḥimli (I rose with my burden with difficulty) is mentioned by both authors (Abū Ḥātim, Aḍdād 247; Abū l-Ṭayyib, Aḍdād II, 720). Cf. also examples of maqlūb with the construction adḫaltu fī ra⁠ʾsī l-qalansuwata (I entered the bonnet into my head), instead of adḫaltu fī l-qalansuwati ra⁠ʾsī, which Sībawayhi (Kitāb I, 181) interprets as an instance of saʿat al-kalām (latitude of speech).

744 See, however, Abū Ḥātim, Aḍdād 184 and Abū l-Ṭayyib, Aḍdād II, 721, where rakūb (riding animal) and irkabī (Ride! [in addressing horses]) are single words which represent this phenomenon.

 

amgad
Highlight

 

197MUBAWWAB AND SPECIALIZED LEXICA

his corpus, Ibn al-Anbārī cites several words which, he says, resemble aḍdād,745 and others which qualify as aḍdād in the context of Qurʾān or poetry746 (cf. his argument that context eliminates the possibility of ambiguity). This notwith- standing, some of the items are specifically mentioned in order to demonstrate that they do not satisfy certain conditions for inclusion in aḍdād. Accordingly, he dismisses Quṭrub’s claim that tariba (to become poor) and atraba (to become rich),747 and that ṯalaltu ʿaršahu (I demolished his throne) and aṯlaltu ʿaršahu (I repaired his throne)748 are examples of aḍdād since in each case the verb belongs to a different conjugation.

All seven books discussed above apparently have their items haphazardly listed. In contrast, the four remaining ones follow an alphabetical arrange- ment, but, unlike the case of laḥn al-ʿāmma where this arrangement shat- ters the types into which laḥn might be grouped, its introduction (probably for the first time by Abū l-Ṭayyib) has no such effect since earlier authors did not arrange their items based on any thematic principle that they might have recognized. Abū l-Ṭayyib apologetically explains in his short introduction that he opted for an alphabetical arrangement because people of his own time preferred it and their ardor was confined to it (iḏ kānat himam ahl zamāninā maqṣūra ʿalayhi wa-qulūbuhum māʾila ilayhi).749 Yet, his arrangement is only partial and does not go beyond the first letter of the roots of words. The book contains 300 items, most of which are extensively discussed and uniformly supported by šawāhid.

In his short monograph, Ibn al-Dahhān very briefly mentions the two con- tradictory meanings of each of his 291 items. These are arranged, just like Abū l-Ṭayyib’s book, based only on the first letter of the root of the word. Ṣaġānī also writes very brief explanations of the 337 items which his book embraces. He ends his list by apologizing for having included in it words which he col- lected from earlier scholars, such as Ibn al-Anbārī, but which he believes are not true aḍdād.750 Unlike Abū l-Ṭayyib and Ibn al-Dahhān, Ṣaġānī takes into consideration in his alphabetical arrangement all the letters of the root. Finally, Munšī is the only author who arranges his aḍdād – 225 briefly explained

745 Ibn al-Anbārī, Aḍdād 136, 257, 258, 289. 746 Ibid., 286, 297, 299, 386, 388. 747 Ibid., 380; cf. Quṭrub, Aḍdād 267. 748 Ibid., 387; cf. Quṭrub, Aḍdād 268. 749 Abū l-Ṭayyib, Aḍdād I, 2. 750 Ṣaġānī, Aḍdād 248.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

198 chapter 2

items – according to the final letter of the root, followed by the first (e.g. the roots ṯʾṯʾ, ǧf ʾ, ḫǧʾ, drʾ, ḏrʾ, nwʾ and qrʾ).751

8 al-Muštarak (Homonyms) and al-Mutarādif (Synonyms)

In the previous section, we discussed a specific type of muštarak, namely, aḍdād. The more general type, however, refers to homonymous polysemic words which do not indicate contradictory meanings, such as ʿamm, which means either “paternal uncle” or “company of men”.752 It is this type which Sībawayhi (d. 180/796) refers to in the introductory part of his book by the expression ittifāq al-lafẓayn wa-ḫtilāf al-maʿnayayn, i.e. coincidence of form and divergence of meaning. For synonyms Sībawayhi uses the expression iḫtilāf al-lafẓayn wa-l-maʿnā wāḥid, i.e. divergence of form and coincidence of meaning.753 As we pointed out earlier, the various types of semantic rela- tionships that Sībawayhi briefly mentions do not form part of his main gram- matical theory or of the Arabic grammatical theory in general. In both types, muštarak and mutarādif, the lexicographical tradition complements works on grammar, and this section is devoted to the two genres of writing which deal with both types.

As in the case of aḍdād, some authors deny the very existence of ištirāk (homonymy). Commenting on Sībawayhi’s example waǧada, which can mean either “to be angry with” or “to find a lost thing”, Ibn Durustawayhi (d. 347/958), for example, insists that a more profound analysis reveals that the two mean- ings are one and the same, namely, “to hit up something, be it good or bad” (iṣābat al-šayʾ ḫayran kāna aw šarran).754 This argument is in line with his posi- tion vis-à-vis aḍdād since he argues against their existence in a book titled Ibṭāl al-aḍdād.755 It is also evident that Ibn Durustawayhi takes issue with tarāduf (synonymy) as is often exemplified by philologists. For example, he insists that because of the difference in form between the two verbal patterns faʿala and afʿala – subject of many lexica (see Section 10.6 below) – they cannot be con- sidered synonymous unless they occur in two different dialects ( fī luġatayn

751 Munšī, Aḍdād. Note that the arrangement is not always exact and that the root qrʾ should have occurred before nwʾ (140), and fǧǧ before ntǧ (145), etc.

752 See above, 189. 753 Sībawayhi, Kitāb I, 24; cf. above, 189. 754 Ibn Durustawayhi, Taṣḥīḥ I, 363–364; cf. Suyūṭī, Muzhir I, 384. 755 Ibid., I, 359; cf. above, 191.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

199MUBAWWAB AND SPECIALIZED LEXICA

muḫtalifatayn),756 as this would conflict with the concept of ḥikma (wisdom).757 In al-Muzhir, Suyūṭī (d. 911/1505) reports views of some scholars who deny the existence of muštarak or mutarādif, as well as views of others who confirm their existence.758 The traditional legal theorists (uṣūliyyūn) and logicians also contribute to this general debate, but this largely falls beyond the scope of the present work.759

Several multithematic works contain material on muštarak and mutarādif. The earliest of these, Abū ʿUbayd’s (d. 224/838) al-Ġarīb al-muṣannaf, for example, includes a chapter titled iḫtilāf al-afʿāl bi-ttifāq al-maʿnā, in which are listed synonymous verbs, such as qašā and qašara (to peel) and aḥamma and aǧamma (to be imminent).760 Several of the book’s short chapters under the general heading Kitāb al-Aǧnās761 also include words of the muštarak type, such as naʿāma (group of people, horse’s name, darkness, etc.) and lawā (to delay paying one’s debt, to twist, to wait).762 As for Ibn Sīda’s (d. 458/1066) al-Muḫaṣṣaṣ, the largest thesaurus of its kind, synonyms are a fixed feature of the majority of its thematically arranged chapters. In fact, Ibn Sīda mentions al-asmāʾ al-mutarādifa and al-asmāʾ al-muštaraka in his introduction and indi- cates that they will be mentioned in their proper places of his book.763 It is also noteworthy that some scholars have authored monographs which essen- tially include synonymous nouns in a specific semantic field. Among these are two previously mentioned titles under the genre which deals with animals, namely, Ibn Ḫālawayhi’s (d. 370/980) Asmāʾ al-asad and Ṣaġānī’s (d. 650/1252) Fī asāmī l-ḏiʾb wa-kunāhu.764 This tradition seems to have persisted until as late as the ninth/fifteenth century, as Suyūṭī (d. 911/1505) attributes to Fīrūzābādī (d. 817/1415) al-Rawḍ al-maslūf fī mā lahu smāni ilā ulūf,765 whose title indi- cates (if one overlooks the exaggeration in claiming the existence of thou- sands of synonyms in some cases) that it contains lists of synonymous nouns, of which Suyūṭī unfortunately includes no quotations. From a more general perspective, material on synonyms permeates books which primarily belong

756 Ibid., I, 165–167; cf. Suyūṭī, Muzhir I, 384–385 and n. 1095 below. Cf. also Ibn Durustawayhi’s text with ʿAskarī, Furūq 12.

757 For ḥikma as the rational factor which underlies Arabic usage, see Baalbaki (2008: 67–68). 758 Suyūṭī, Muzhir I, 369–386, 402–413. 759 For details, see Kafrāwī (2002) and Bettini (2008 a,b). 760 Abū ʿUbayd, Ġarīb II, 616–618. 761 Ibid., III, 924–1010. 762 Ibid., III, 971, 999. 763 Ibn Sīda, Muḫaṣṣaṣ I, 3. 764 See above, 143. 765 Suyūṭī, Muzhir I, 407.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

200 chapter 2

to the genre of ġarīb, such as Abū Misḥal’s (d. 231/845) al-Nawādir and Kurāʿ al-Naml’s (d. 310/922) al-Muntaḫab.

The following works belong to the genre of muštarak:

1. al-Aǧnās min kalām al-ʿArab wa-mā štabaha fī l-lafẓ wa-ḫtalafa fī l-maʿnā, claimed by an anonymous author to be compiled from Abū ʿUbayd’s (d. 224/838) Ġarīb al-Ḥadīṯ.

2. Mā ttafaqa lafẓuhu wa-ḫtalafa maʿnāhu by Abū l-ʿAmayṯal (d. 240/854). 3. Mā ttafaqa lafẓuhu wa-ḫtalafa maʿnāhu min al-Qurʾān al-maǧīd by

Mubarrad (d. 285/898). 4. al-Munaǧǧad fī l-luġa by Kurāʿ al-Naml (d. 310/922). 5. al-Malāḥin by Ibn Durayd (d. 321/933). 6. al-ʿAšarāt fī ġarīb al-luġa by Abū ʿUmar al-Zāhid, better known as Ġulām

Ṯaʿlab (d. 345/957). 7. al-ʿAšarāt fī l-luġa by al-Qazzāz al-Qayrawānī (d. 412/1021). 8. Mā ttafaqa lafẓuhu wa-ḫtalafa maʿnāhu by Ibn al-Šaǧarī (d. 542/1148).

The anonymous al-Aǧnās min kalām al-ʿArab is a collection of 149 words of the muštarak type which the author claims to be derived from Abū ʿUbayd’s Ġarīb al-Ḥadīṯ. For example, ṣayāṣī (sing. ṣīṣiya) is said to have three different meanings, namely, “horn (of a bull or cow)”, “fortress”, and “a weaver’s šawka, i.e. an instrument with which he makes the warp and woof even”.766 Indeed, all three meanings are mentioned by Abū ʿUbayd, though in somewhat different wording.767 However, in most cases, the words cited in al-Aǧnās either do not feature in Abū ʿUbayd’s Ġarīb al-Ḥadīṯ,768 or are given meanings which only partially correspond to those cited in the latter.769 It is theoretically possible that the material is derived from riwāyas of al-Ġarīb which significantly dif- fer from the ones we know, but, given the large magnitude of the differences between the two texts, it is much more likely that the material is erroneously attributed to Abū ʿ Ubayd. Yet it should be noted that books on aǧnās did appear in the same period as Abū ʿUbayd’s, since a book titled K. al-Aǧnās is attributed to Aṣmaʿī (d. 216/831).770 From a reference to it by Ibn Fāris (d. 395/1004), we are given to understand that Aṣmaʿī was the first to use the term aǧnās to refer to

766 Anonymous, Aǧnās 3. 767 Abū ʿUbayd, Ḥadīṯ II, 84. 768 E.g. bayẓ, katūm, barbar, saḫām, ṣaqaʿ, nāǧir, salīḥ, etc.; cf. Aǧnās 1–4. 769 E.g. šawā, ṭaḫāʾ, kalb and ṣadā; cf. Aǧnās 2–4 and Abū ʿUbayd, Ḥadīṯ III, 367; IV, 421; III, 197;

IV, 492; II, 25, 168–169; I, 26–28 respectively. 770 Azharī, Tahḏīb I, 32; Ibn al-Nadīm, Fihrist 61; Yāqūt, Muʿǧam VI, 2844 (al-Aǧnās al-akbar).

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

201MUBAWWAB AND SPECIALIZED LEXICA

homonymous polysemic words.771 Other third/ninth century authors to whom are attributed books of the same title include Bāhilī (d. 231/846)772 and Ibn al-Sikkīt (d. 244/858).773

The other books in the above list may be divided into two types. Those authored by Abū l-ʿAmayṯal, Kurāʿ al-Naml and Ibn al-Šaǧarī represent the more general type, which contains lists of muštarak words, whereas the books of Mubarrad, Ibn Durayd, Ġulām Ṯaʿlab and al-Qazzāz al-Qayrawānī belong to the other type, which examines a specific aspect of muštarak.

Abū l-ʿAmayṯal’s book, also known as al-Ma⁠ʾṯūr min al-luġa, is a collection of 273 muštarak words which are listed in no specific order, and whose vari- ous meanings are briefly explained and range from two (e.g. atān) to fourteen (e.g. qarn).774

Kurāʿ al-Naml’s book, on the other hand, is arranged primarily on thematic bases. It contains the following six chapters: human body parts, animals, birds, weapons, the sky, and the earth and what it contains. Based on their various meanings, muštarak words in such an onomasiological arrangement can be theoretically placed in more than one chapter, but Kurāʿ places each word only under its most common meaning and lists with it all the other meanings. The last chapter, which includes 669 words out of a total of 884,775 is the only one that is alphabetically arranged from alif to yā (based on the first two letters of the word, and not its root).776 The only plausible explanation for this discrep- ancy is that Kurāʿ did not find it necessary to arrange alphabetically the first five chapters, given the relatively few words they contain. On another note, some of the book’s material might have been a source of inspiration to later authors. In this respect, it is hard not to notice the similarity between Kurāʿ’s first chapter and Ibn Fāris’s (d. 395/1004) Istiʿārat aʿḍāʾ al-insān, which is essen- tially a book on muštarak.777 In fact, the use of istiʿāra in its title points to the role of metaphor in the formation of muštarak words. Furthermore, some of the words which appear in Kurāʿ’s chapter on birds and are related to the body parts of a horse (e.g. farḫ “young one of a bird” and “forepart of the brain of a

771 Ibn Fāris, Maqāyīs I, 486. 772 Ibn Ḫayr, Fahrasa 381, 398. 773 Ibn al-Nadīm, Fihrist 79; Yāqūt, Muʿǧam VI, 2841 (al-Aǧnās al-kabīr). 774 Abū l-ʿAmayṯal, Mā ttafaqa 34, 32. 775 Cf. Omer (1993: 5–6). 776 For example, ta’wīl, taballud, taḥayyur and tadaṯṯur are placed under tāʾ (Kurāʿ, Munaǧǧad

148–150; but cf. tafrīṭ, under fāʾ; ibid., 285), and ma⁠ʾtam, mubīn, maǧāʿa and maḥdūd under mīm (ibid., 324–325).

777 See above, 153.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

202 chapter 2

horse”; ḏubāb “common fly” and “a black spot in the horse’s eye”)778 are also collected by later authors, such as Suyūṭī (d. 911/1505), in Ǧarr al-ḏayl fī ʿilm al-ḫayl.779

By the time of Ibn al-Šaǧarī in the sixth/twelfth century, the genre had obvi- ously developed as far as the size of the material is concerned, since his book contains 1,670 muštarak words. Unlike Kurāʿ’s book, it is not divided accord- ing to themes, but is alphabetically arranged throughout, albeit on the basis of only the first letter of the word, irrespective of its root.

Authors of books of the second type are interested in a specific aspect of the phenomenon of ištirāk. Mubarrad’s short monograph, for example, specifi- cally deals with Qurʾānic words which can be interpreted as having a meaning that differs from the more common one. An example is the verse mā lakum lā tarǧūna li-l-Lāhi waqāran (Why are you not afraid of God’s greatness [accord- ing to Mubarrad’s interpretation]; Q 71: 13), where tarǧūna in this context has the meaning of “to be afraid of” (taḫāfūna), rather than the usual meaning of “to hope for”.780 Mubarrad’s introduction, as previously shown,781 is similar to Sībawayhi’s short chapter on semantic relationships, but Mubarrad adds examples which demonstrate words that have contradictory meanings (i.e. aḍdād). The meanings of the examples he cites uniformly depend on context, hence his recurrent reference to maǧāz or figurative expressions, and to ḥaḏf or the omission of certain parts of the construction.782 From a wider perspec- tive, Mubarrad’s book may be regarded as part of a larger tradition established by exegetes who cite various meanings for the same Qurʾānic word.783 The two earliest books which we possess in this tradition are Muqātil b. Sulaymān’s (d. 150/767) al-Ašbāh wa-l-naẓāʾir fī l-Qurʾān al-karīm and Hārūn b. Mūsā l-Aʿwar’s (d. c. 170/786) al-Wuǧūh wa-l-naẓāʾir fī l-Qurʾān al-karīm. Both authors give the various meanings a Qurʾānic word might have in different verses. For example, Muqātil cites seventeen different meanings (or shades of meanings) for the word hudā, thirteen for sabīl, eleven for ḥaqq, seven for ḫalq, and two for āya.784 Various context-dependent meanings of particles, such as hal, aw,

778 Kurāʿ, Munaǧǧad 95–96 and Suyūṭī, Ǧarr 77–78; cf. Zaǧǧāǧī, Aḫbār 212–213. 779 See above, 147. 780 Mubarrad, Mā ttafaqa 8. 781 See above, 189. 782 Mubarrad, Mā ttafaqa 16, 22, 24–26, 29, 31 (maǧāz); 29–31, 36 (ḥaḏf ). 783 Note that some books in this genre are alphabetically arranged, such as Abū Hilāl

al-ʿAskarī’s (d. after 395/1005) al-Wuǧūh wa-l-naẓāʾir fī l-Qurʾān al-karīm, Ismāʿīl b. Aḥmad al-Ḥīrī’s (d. after 430/1039) Wuǧūh al-Qurʾān, and Ibn al-Ǧawzī’s (d. 597/1201) Nuzhat al-aʿyun al-nawāẓir fī ʿilm al-wuǧūh wa-l-naẓāʾir.

784 Muqātil, Ašbāh 89, 185,175, 261, 300.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

203MUBAWWAB AND SPECIALIZED LEXICA

am, fī and ḥattā, are also cited,785 and there exist unmistakable parallels with later grammar books, particularly those which deal with maʿānī l-ḥurūf (see Section 9 below). The approach which Muqātil and Hārūn represent, however, was challenged by al-Ḥakīm al-Tirmiḏī (d. c. 320/932) who, in Taḥṣīl naẓāʾir al-Qurʾān, like Ibn Durustawayhi, denies the existence of ištirāk, and focuses solely on the Qurʾānic text. In his introduction, Tirmiḏī refers to an earlier book on naẓāʾir (probably Muqātil’s due to the similarity in material, particularly in the early part of Tirmiḏī’s book) and argues that the various meanings assigned by its author to each word go back to one general common meaning, which he tries to trace in each case.786 In short, Mubarrad’s book belongs to a tradition in which philology and exegesis are inextricably linked.787

Ibn Durayd’s al-Malāḥin deals with a very specific type of muštarak. It con- tains 184 words, each of which has, in addition to its usual meaning, another which is much less common. This quality allows the speaker who has to resort to paronomasia when he is unjustly put under oath (al-muǧbar al-muḍṭahad ʿalā l-yamīn al-mukrah ʿalayhā)788 to utter a construction which suggests that the usual meaning is intended, whereas he has the less common meaning in mind. For example, if someone is asked under oath whether he has entered the house (bayt) of someone else, he can deny this by saying wa-l-Lāhi mā daḫaltu li-fulānin baytan wa-lā ra⁠ʾaytu lahu baytan (I swear by God that I have not entered nor seen that person’s bayt).789 Accordingly, the listener would think that bayt has the prevalent meaning of “house”, whereas the speaker has in mind the rare meaning of “tomb”.

Material on muštarak is also present in books of the type known as ʿ ašarāt.790 These are divided into chapters which normally contain a group of ten words that share a common feature, or one word which has numerous meanings. In many cases, the words belong to the realm of ġarīb. The first book of this type is authored by Abū ʿUmar al-Zāhid, known as Ġulām Ṯaʿlab, and is divided into sixty chapters, some of which contain words that rhyme and share the same

785 Ibid., 151, 189, 213, 214, 269. 786 For example, he traces all the meanings of hudā (a total of eighteen; cf. Muqātil, Ašbāh

89–95, where seventeen are mentioned) to one word (kalima wāḥida faqaṭ), mayl (inclination), and tries to demonstrate that this meaning is present in the various verses in which it occurs; cf. Tirmiḏī, Taḥṣīl 19–24.

787 For book-length studies of muštarak in the Qurʾān, see Makram (1996), Munaǧǧid (1998), and ʿArār (2012).

788 Ibn Durayd, Malāḥin 63. 789 Ibid., 82. 790 To Ibn Ḫālawayhi (d. 370/980) is also attributed a book entitled al-ʿAšarāt; cf. Ḥāǧǧī Ḫalīfa,

Kašf II, 1439.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

204 chapter 2

morphological pattern. The chapter entitled tarīṣ (firm, solid), for example, contains the following nine words: tarīṣ, ǧanīṣ, raḫīṣ, kaṣīṣ, farīṣ, qaṣīṣ, ḫarīṣ, baṣīṣ and raṣīṣ.791 Only in the case of farīṣ in this chapter does Zāhid cite more than one meaning. Other chapters begin with the various meanings of a cer- tain word (e.g. bard, for which seven meanings are given) and then list other rhyming words, e.g. ǧard and ḥard.792 In the chapter on ḫinḏīḏ, nine different meanings of the word are mentioned, followed by only one other word, miḏmīḏ (liar).793 As for al-Qazzāz al-Qayrawānī, he begins his book on ʿašarāt by con- siderably expanding one of his predecessor’s chapters.794 The rest of the book is divided into 167 alphabetically arranged words, to each of which are ascribed ten different meanings, although in some cases the number exceeds ten (as in qarn, for which twenty-eight meanings are listed).795 Accordingly, the book is almost totally devoted to muštarak, unlike Zāhid’s. It is interesting to note that Ṣafadī (d. 764/1363) quotes al-Qazzāz as saying at the end of his book that he intends to write K. al-Miʾāt.796 Although the printed version does not contain a statement to this effect, its mention by Ṣafadī may well indicate that al-Qazzāz wanted to test the limits of this genre of writing by greatly increasing the num- ber of words which form a separate group. Mention can also be made of a section entitled Mā ttafaqa lafẓuhu wa-ḫtalafa maʿnāhu in Sulaymān b. Banīn al-Daqīqī’s (d. 614/1217) Ittifāq al-mabānī wa-ftirāq al-maʿānī. But although this section begins with a number of meanings of the word ʿayn and includes several chapters on homonyms, the best part of its material is not related to muštarak, but is made up of four ašǧār derived from Abū l-Ṭayyib al-Luġawī’s (d. 351/962) Šaǧar al-durr, which in turn contains material on muštarak, as pre- viously mentioned.797

On mutarādif, we are in possession of the following works:

1. Mā ḫtalafat alfāẓuhu wa-ttafaqat maʿānīhi by Aṣmaʿī (d. 216/831). 2. al-Alfāẓ by Ibn al-Sikkīt (d. 244/858). 3. al-Alfāẓ al-kitābiyya by Hamaḏānī (d. c. 320/932). 4. al-Alfāẓ; al-kitāba wa-l-taʿbīr by Ibn al-Marzubān al-Bāḥiṯ (d. c. 330/941).

791 Zāhid, ʿAšarāt 28. 792 Ibid., 114. 793 Ibid., 108. 794 The title of the chapter is al-maṯʿ (a hideous manner of walking); see Zāhid, ʿAšarāt 82 and

Qazzāz, ʿAšarāt 27–37. 795 Qazzāz, ʿAšarāt 247–250; see also Ḍāmin (1991a). 796 Ṣafadī, Wāfī II, 305. 797 The two books by Abū l-Ṭayyib and Daqīqī were discussed earlier; cf. above, 97–99.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

205MUBAWWAB AND SPECIALIZED LEXICA

5. Ǧawāhir al-alfāẓ by Qudāma b. Ǧaʿfar (d. 337/948). 6. al-Alfāẓ al-mutarādifa al-mutaqāribat al-maʿnā by Rummānī (d. 384/994). 7. Mutaḫayyar al-alfāẓ by Ibn Fāris (d. 395/1004). 8. Section 3 of Ittifāq al-mabānī wa-ftirāq al-maʿānī by Sulaymān b. Banīn

al-Daqīqī (d. 614/1217). 9. al-Alfāẓ al-muḫtalifa fī l-maʿānī l-muʾtalifa by Ibn Mālik (d. 672/1274).

That books on mutarādif go back to the earliest period of lexical writing is quite significant, given that lexica of this mubawwab or onomasiological type resemble the contemporary lexica of the muǧannas or semasiological type – e.g. Ḫalīl’s (d. 175/716) K. al-ʿAyn – in giving synonyms of lexemes. Yet, the two types are different in that in the mubawwab type, several synonyms are nor- mally given for each word, whereas muǧannas lexica usually cite only one, or even explain the word in question without mention of any synonym. Aṣmaʿī’s monograph is the first extant work in the mubawwab type, and its material is neither thematically nor alphabetically arranged. In fact, lack of alphabeti- cal, though not thematic, arrangement is a feature that is common to all other works on mutarādif listed above. Aṣmaʿī lists nominal as well as verbal syn- onyms, such as kušāḥa, qumāma, ḫumāma, kunāsa and kibā (sweepings), and ǧarana, marana and ṭābaqa (to become accustomed to).798 Some of the cited words, however, are not strictly mutarādifāt, such as arbaʿa (to have offspring in the prime of one’s manhood) and aṣāfa (to have offspring at an advanced age), and wakr/wakn, ʿušš, ufḥūṣ and udḥiyy, each of which refers to a specific type of a bird’s nest (e.g. an ufḥūṣ refers to one on the ground, and an udḥiyy is one made by an ostrich).799

Authored only a few decades after Aṣmaʿī, Ibn al-Sikkīt’s K. al-Alfāẓ, as pre- served in the commentary by Tibrīzī (d. 502/1109) known as Tahḏīb al-Alfāẓ, represents a major development in the genre. Not only is its material much more extensive than Aṣmaʿī’s, but it is thematically arranged into 148 mostly short chapters in which are listed words and expressions belonging to a spe- cific semantic field, such as wealth and fertility, poverty and barrenness, groups of individuals, battalions, stinginess, illness, wine, color, tallness, shortness, old women, sun, moon, calamities, death, love, heaviness, water, etc. The book can thus be classified also under multithematic works (cf. Section 11 below), although it has a strong component of mutarādif words.800 Indeed, some chapters are almost fully restricted to synonyms – as in the various adjectives

798 Aṣmaʿī, Mā ḫtalafat 36, 46–47. 799 Ibid., 57, 66. 800 See also below, 268.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

206 chapter 2

denoting the miserly (e.g. šaḥīḥ, naḥīḥ, ḍanīn, ḥiṣrim, ṣāmir, masīk, anūḥ, ḍirizz, ḥātir, qātir, etc.),801 and the verbs which indicate lying (e.g. kaḏaba, walaʿa, māna, tasaddaǧa, zaʿafa, ibtašaka, bašaka, saraǧa, ḫadaba, ʿabaṭa, etc.)802 – but there are chapters which contain material not related to synonymity. For example, the chapter on miyāh (water) – which does include some synonyms, e.g. ʿaḏb, nuqāḫ, zulāl, etc. for fresh water – mostly deals with words which express various states or aspects of water (such as freshness, saltiness, con- tamination, abundance, color, etc.).803 Similarly, the chapter on laḥm (meat) contains very few synonyms, but lists words related primarily to the cutting and cooking of meat.804

al-Alfāẓ al-kitābiyya, also known as Alfāẓ al-ašbāh wa-l-naẓāʾir, by Hama- ḏānī contains 365 short chapters which cover a large variety of semantic fields. Although it is similar to Ibn al-Sikkīt’s al-Alfāẓ, it differs from it in that Hamaḏānī further expands the concept of synonymity to include, in an almost systematic manner, phrases which can be alternatively used to express a cer- tain meaning. For example, in the chapter on aǧnās al-ʿaṭaš (types of thirst),805 he lists synonyms of ʿ ataš (e.g. ġulla, ġalīl, ẓama⁠ʾ, ṣadā, ḥirra, nahal, ǧuwād, etc.) and of ʿ aṭšān “thirsty” ( e.g. haymān, ẓamʾān, ṣādin, nāhil, hāʾim, ḥāʾim, etc.), but then cites several idiomatic expressions which share the meaning of “to quest someone’s thirst for revenge”, such as šafaytu ṣadra fulānin min ʿaduwwihi, barradtu ġalīlahu, naqaʿtu ġullatahu, šafaytu ḥurqatahu, arwaytu ḥirratahu, qaṣaʿtu ṣārratahu, etc. As Hamaḏānī explains in the introduction, he collected his phrases from the works of professional writers and erudite authors in order to provide the writer with a pool of idiomatic expressions from which he can choose and thus be able, for example, to replace aṣlaḥa l-fāsida (to rectify a state of affairs) with synonymous expressions such as lamma l-šaʿaṯa, ḍamma l-našara and rataqa l-fatqa.806

Idiomatic expressions are perhaps nowhere given prominence in the genre as much as in al-Alfāẓ; al-kitāba wa-l-taʿbīr by Ibn al-Marzubān al-Bāḥiṯ. The use of taʿbīr in the title reflects the author’s interest, not only in single lex- emes but also in the style of expression. The book is unique in its organiza- tion; it is divided into twenty-eight themes (e.g. dependence on God and his Prophet, introduction of books, congratulations, intercessions, condolences,

801 Ibn al-Sikkīt, Alfāẓ 69–76. 802 Ibid., 258–262. 803 Ibid., 557–562. 804 Ibid., 605–613. 805 Hamaḏānī, Alfāẓ 76–77. 806 Ibid., viii; see also p. 1.

 

amgad
Highlight
amgad
Highlight

 

207MUBAWWAB AND SPECIALIZED LEXICA

thanks, excuses, etc.), each theme including a model text which consists of a number of expressions. Each expression is followed by phrases which can replace it. At times, alternatives are suggested even for individual words within an expression. In the theme of condolences, for example, the following ele- ments form the model text:807

balaġanī l-ḫabaru l-faẓīʿu bi-ḥādiṯi qaḍāʾi l-Lāhi wa-qadarihi bi-fulānin fa-ḫašaʿa fī ṭarfī fa-innā li-l-Lāhi wa-innā ilayhi rāǧiʿūna istislāman li-amrihi wa-raḥima l-Lāhu fulānan wa-aʿẓama lahu l-maṯwabata wa-l-aǧra wa-uqsimu bi-l-Lāhi l-ʿālimi bi-muḍmarāti l-qulūbi anna mā nālanī min al-qalaqi. . . . man ṣafat laka niyyatuhu . . .

(The terrible news came to me about God’s inevitable predestination concerning X; so it cast my sight towards the ground. [I iterate]: “To God we belong and to Him is our return” [Q 2: 156], in succumbing to His command. May He have mercy on the soul of X and may He increase his reward and recompense. I swear by God, who knows what hearts conceal, that the worry which befell me . . . He whose interest towards you is pure . . .).

Following the first expression, for example, the following alternative expres- sions are listed: ittaṣala bī l-naba⁠ʾu l-muʾlimu; warada ʿalayya l-ḫabaru l-fādiḥu; tanāhā ilayya ḫabaru muṣābika; taqāḏafa ilayya l-naba⁠ʾu l-mukribu; tasāqaṭa ilayya l-ḫabaru l-bāhiẓu; ruqqiya ilayya naba⁠ʾu l-raziyyati l-bāhirati; and numiya ilayya ḫabaru l-muṣībati l-muqawwiḍati. For the practical purpose of writing a text, these expressions may be considered synonymous, although they dif- fer in the exact meaning of some of their components (cf., for example, faẓīʿ “terrible”, muʾlim “painful” and muqawwiḍa “devastating”). In other cases, the alternative expressions differ more significantly in meaning, as in raḥima l-Lāhu fulānan (May God have mercy on the soul of X), which can be replaced by naḍḍara l-Lāhu waǧhahu (May God comfort his face), ḍāʿafa ḥasanātihi/ ġafara sayyiʾātihi (May He double his good actions/forgive him his evil deeds), alḥaqahu bi-l-abrāri min salafihi (May He join him to the pious among his

807 Ibn al-Marzubān, Alfāẓ 113–118.

 

 

208 chapter 2

ancestors), etc. An example of a single word for which synonyms are suggested is rāǧiʿūna, hence ṣāʾirūna, munqalibūna, ʿāʾidūna, ʿāṭifūna and āyibūna. It should also be noted that the last two expressions quoted above are left incom- plete so as to allow more flexibility in the proposed text. As for the number of alternative expressions, it ranges from five (in the case of rāǧiʿūna) to seventy- five (in the case of wa-aʿẓama lahu l-maṯwabata wa-l-aǧra). Accordingly, the model text can be regenerated in an extremely large variety of ways so that no two resulting texts would read exactly the same.

Ǧawāhir al-alfāẓ is authored by one of the most prominent rhetoricians, Qudāma b. Ǧaʿfar. The book’s content faithfully reflects its author’s interest in eloquent style, which, according to him, is characterized primarily by the rhyming of words (saǧʿ) and parallelism between components of consecutive phrases (tarṣīʿ). In the introduction, Qudāma gives an illustrative example on the authority of an earlier author who consecutively cites the phrases aṣlaḥa l-fāsida, ḍamma l-našara, sadda l-ṯalma, and asā l-kalma (to rectify a state of affairs).808 Instead, Qudāma suggests that aṣlaḥa l-fāsida be coupled, not with ḍamma l-našara, with which it differs in pattern and rhyme, but rather with allafa l-šārida and saddada l-ʿānida. Similarly, he argues that aṣlaḥa mā fasada and ṣallaḥa fāsidahu are best coupled with qawwama l-awada and raǧǧaʿa šāridahu respectively. Throughout the book – which consists of 372 themes in which single words normally occur before phrases809 – the principle of group- ing together rhyming words and parallel phrases is strictly observed. As a result, the alternatives proposed to the user, unlike most other works in the genre, are arranged in a manner which should help him produce a text that abounds with rhyme and parallelism, in line with the prevalent tradition of elevated style in the author’s time. By comparing the only chapter which proposes a text in a manner similar to that of Ibn al-Marzubān’s with the corresponding chap- ter in the latter’s book (i.e. the chapter on condolences),810 it is evident that Qudāma’s text is much more artificially crafted due to his preoccupation with saǧʿ and tarṣīʿ. Both works, however, are extremely useful mubawwab lexica for the writer who looks for synonyms which muǧannas lexica do not normally cite in one place. Yūsuf b. ʿAbdallāh al-Zaǧǧāǧī (d. 514/1024) in ʿUmdat al-kuttāb follows Qudāma’s order of themes but alters the original text by amending cer- tain expressions and adding others.

808 Qudāma, Ǧawāhir 2–3. The author referred to by Qudāma is most probably Hamaḏānī, whose very first chapter contains the terms which Qudāma cites; cf. Alfāẓ 1–2.

809 Cf. Qudāma, Ǧawāhir, Chapters 37 and 121, pp. 61, 121–123. 810 Ibid., 400–422; cf. Ibn al-Marzubān, Alfāẓ 113–118.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

209MUBAWWAB AND SPECIALIZED LEXICA

The last four works in the above list demonstrate that the genre of mutarādif has hardly developed at all following Ibn al-Sikkīt, Hamaḏānī, Ibn al-Marzubān and Qudāma. As far as the size of the material is concerned, the books by Rummānī, Ibn Fāris and Ibn Mālik, as well as the relevant chapter in Daqīqī’s, are considerably shorter than earlier works, with the possible excep- tion of Aṣmāʿī’s monograph (itself longer than Ḍaqīqī’s chapter). Furthermore, arrangement by theme seems to have persevered, since we do not know of any attempt at alphabetical arrangement. It can be argued that an alphabeti- cal ordering would have necessitated repetition of synonyms under several entries, but we know that in the case of the muštarak genre, Kurāʿ al-Naml, who adopted an onomasiological arrangement in al-Munaǧǧad, listed the vari- ous meanings of a word only in one chapter, namely, the one whose theme is the closest to that word’s most common meaning. In other words, an alpha- betical arrangement of mutarādif words and phrases is not theoretically inconceivable, although it could limit the number of entries considerably if Kurāʿ’s system were adopted, and perhaps would require a great deal of cross- referencing. Concerning specific features in the last four works, Rummānī’s title, al-Alfāẓ al-mutarādifa al-mutaqāribat al-maʿnā is interesting since it highlights the author’s conviction that the term mutarādif may well be applied to words or phrases which are close, but not identical, in meaning. The vast majority of its 142 short, thematically arranged sections include single words, rather than phrases.811 Ibn Fāris’s book contains 114 mostly short sections and is characterized by a heavy reliance on poetry šawāhid. Daqīqī’s third section of Ittifāq al-mabānī occupies no more than a few pages of the book.812 The author begins by arbitrarily citing various synonyms and divides the rest of the material into seventeen themes. For his part, Ibn Mālik divides his book into 210 extremely short sections, which mostly list synonyms that are single words, with a few notable exceptions of phrasal material.813

Some of the scholars who deny the existence of muštarak have also authored works to express their disapproval of the predominant view that two or more words can have the same meaning. The period in which this issue was most hotly debated is the fourth/tenth century – the same period from which we pos- sess five of the books on mutarādif discussed above. It is quite telling that books titled al-Furūq primarily belong to this period as well. The earliest we know of is al-Furūq wa-manʿ al-tarāduf by al-Ḥakīm al-Tirmiḏī (d. c. 320/932), whose

811 See Rummānī, Alfāẓ 66, 80–81 (items 55, 118–120) for several phrases. 812 Daqīqī, Ittifāq 245–261. 813 Ibn Mālik, Alfāẓ 63, 72, 107, 128–138.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight

 

210 chapter 2

denial of muštarak in Taḥṣīl naẓāʾir al-Qurʾān has already been mentioned.814 In the introduction to al-Furūq, he briefly explains that actions are performed either out of good or bad faith.815 To this discrepancy he ascribes the apparent synonymity in 156 pairs of words,816 which he then explains in a manner that demonstrates, in each case, that the words in question express two different concepts. One clear example of his is the differentiation between muḥāǧǧa (argumentation) and muǧādala (disputation), for whereas the purpose of the former is to establish the truth and defend it, the latter is performed merely out of interest in defeating one’s opponent.817

The second book we know of, al-Furūq, by Abū l-Ṭayyib al-Luġawī (d. 351/962), is unfortunately lost, except for a somewhat lengthy quotation from it preserved by Suyūṭī.818 In this quotation, Abū l-Ṭayyib draws a dis- tinction between thirty-nine different adjectives which describe the hand on which is left a trace of what it touched. For example, ġamira or nadila, waḍira, laṯiqa, nasima, ḫamiṭa, radiġa, naḍiḫa and šaḥiṭa are the precise adjectives for the hand which came in contact with meat, milk, honey, cheese, vinegar, clay, watermelon and blood. Yet, it is not clear whether Abū l-Ṭayyib in this book expresses a clear position vis-à-vis tarāduf. It should also be noted that Abū l-Ṭayyib’s text is quoted by Suyūṭī in the context of his discussion of words which have a specific connotation (wuḍiʿa li-maʿnā ḫāṣṣ),819 and not in the chapter on tarāduf.

Abū Hilāl al-ʿAskarī (d. after 395/1005) divides his al-Furūq al-luġawiyya into thirty chapters, in the first of which he argues that difference in expressions (ʿibārāt) and nouns (asmāʾ) necessarily results in a difference in meaning – an argument which rests on the usual ascription of wisdom (ḥikma) to the pre- sumed inventor of language (wāḍiʿ al-luġa).820 In the case of faʿala and afʿala, for example, he asserts that it is impossible (muḥāl) that they have the same meaning, unless they occur in two different dialects.821 ʿAskarī also discusses criteria which help in the distinction between two words that may be errone- ously described as synonyms. These criteria are either semantic (e.g. the mean- ing of the opposite “naqīḍ” of each of the two words), syntactic (e.g. the type of

814 See above, 203. 815 Tirmiḏī, Furūq 61. 816 Note that Tirmiḏī says that the total number of pairs in his book is 164 (ibid., 64). 817 Ibid., 87. 818 Suyūṭī, Muzhir I, 447. 819 Ibid., I, 435. 820 ʿAskarī, Furūq 11. 821 Ibid., 12.

 

amgad
Highlight

 

211MUBAWWAB AND SPECIALIZED LEXICA

prepositions which are normally used with the word), or morphological (e.g. the root from which the word is derived).822 Each of the twenty-nine other chapters is devoted to a broad theme (e.g. grammatical terms, lying, knowl- edge, life, action, origin, wealth, increase, protection, arrogance, etc.), under which are listed sets of two (or more) words whose nuances are clarified.

9 al-Ḥurūf/al-Aṣwāt (Particles/Letters)

The genre of writing which investigates the “letters of the alphabet” is strongly related to the three main domains of naḥw, namely, syntax, morphology and phonology (cited in the order in which they are usually studied in grammatical works). The term ḥarf (pl. ḥurūf ), which is the most widely used term to refer to a phonological unit, is also used in morphology to designate a morphologi- cal component of a word (as in the terms ḥarf aṣlī “radical” and ḥurūf al-ziyāda “letters of augmentation”), and in syntax to designate the part of speech which is neither a verb nor a noun (e.g. ḥarf ǧarr “preposition” and ḥarf istifhām “inter- rogative particle”). The term ṣawt (pl. aṣwāt)823 refers specifically to a “sound” and hence excludes the scriptural dimension which ḥarf confusingly has (as in referring to the long vowel in qāla as a ḥarf, obviously under the influence of script). In this section, we shall mainly discuss mubawwab works in the realm of phonology, but since this realm is inextricably linked to morphology and syntax in these works, it is inevitable to refer to morphological/morphopho- nological and, to a lesser extent, grammatical issues in the discussion to fol- low. Accordingly, the title of this section, al-ḥurūf/al-aṣwāt (particles/letters), is intended to reflect this inevitability.824

One of the sources which most clearly demonstrate the use of ḥarf in all three main domains of naḥw is Ibn Ǧinnīʾs (d. 392/1002) Sirr ṣināʿat al-iʿrāb. In each chapter, Ibn Ǧinnī examines a particular ḥarf by briefly determining

822 Ibid., 14–16. 823 For the relationship between ḥarf and ṣawt, see Fleisch (1986: 597). 824 Note that some scholars of Qurʾānic taǧwīd (i.e. the art of reciting the Qurʾān) describe

the points of articulation (maḫāriǧ) and the phonological characteristics (ṣifāt) of consonants, as well as matters realted to pause (waqf ), imāla (fronting and raising of ā towards ī), etc. Since we shall not deal with taǧwīd material, suffice it here to cite but one typical short monograph which deals with maḫāriǧ and ṣifāt, namely, Ibn al-Ṭaḥḥān’s (d. after 560/1165) Maḫāriǧ al-ḥurūf wa-ṣifātuhā. As indicated in its introduction (p. 78), the monograph aims at describing the ḥurūf which form the basis of Qurʾānic reading (qirāʾa, tilāwa), that is, all the consonants of Arabic as well as the alif of imāla and some specific types of certain consonants (e.g. al-nūn al-muḫfāt “concealed nūn”).

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

212 chapter 2

whether it is voiced (maǧhūr) or voiceless (mahmūs) and then extensively investigating its morphological and syntactical characteristics. The chapter on ḥarf al-fāʾ,825 for example, begins by defining fāʾ as a voiceless consonant which occurs as a radical (aṣl, as in faḥm “carbon”) or a substitute for another consonant (as in fumma for ṯumma “then”). The phonological and morphologi- cal features are supplemented by a detailed study of some syntactical aspects of fāʾ, such as its use as a conjunction or an otiose particle (zāʾida).826 In all three levels, the fāʾ is referred to as ḥarf, as are indeed the other letters of the alphabet discussed in separate chapters.

Several books entitled K. al-Aṣwāt are ascribed in the sources almost exclu- sively to authors from the second/eighth and third/ninth centuries, most notably Quṭrub (d. 206/821),827 al-Aḫfaš al-Awsaṭ (d. 215/830),828 Aṣmaʿī (d. 216/831),829 and Ibn al-Sikkīt (d. 244/858).830 None of these has survived, but Suyūṭī (d. 911/1505) briefly quotes four passages from Ibn al-Sikkīt’s book in al-Muzhir and possibly one from Aṣmaʿī’s. These quotations neither yield a clear picture about the content of the two books nor inform us exactly how the genre of aṣwāt differs from that of ḥurūf.831 In contrast, several books that deal with ḥurūf have survived, and although a number of them deal primarily with syntactical aspects, their inclusion in the following list is based on the fact that they are specialized lexica, a number of which are alphabetically arranged:

825 Ibn Ǧinnī, Sirr I, 247–276. 826 Ibid., I, 251 ff. 827 Ibn al-Nadīm, Fihrist 58; Yāqūt, Muʿǧam VI, 2647; Ibn Ḫillikān, Wafayāt IV, 312; Qifṭī, Inbāh

III, 220. 828 Ibn al-Nadīm, Fihrist 58; Yāqūt, Muʿǧam III, 1376; Ibn Ḫillikān, Wafayāt II, 381; Qifṭī, Inbāh

II, 42. 829 Ibn al-Nadīm, Fihrist 61. 830 Ibid., 79; Ibn Sīda, Muḫaṣṣaṣ I, 12; Ibn Ḫillikān, Wafayāt VI, 400; Qifṭī, Inbāh IV, 61. 831 Suyūṭī (Muzhir I, 52) does not specifically mention K. al-Aṣwāt by Aṣmaʿī, but quotes three

of the words which Aṣmaʿī explains and which are related to the aṣwāt of horses. These words, it should be noted, are not found in Aṣmaʿī’s K. al-Ḫayl. As for Ibn al-Sikkīt, three of the quotations on his authority (ibid., I, 559, 566; II, 90) are related to aṣwāt, but the fourth (II, 205) is not. The first of the three relevant quotations is that both ṣaranqaḥ and ṣalanqaḥ are applied to a man vehement of voice. The two words might have been included by Ibn al-Sikkīt either as part of a list of words related to aṣwāt (in the sense of “voices”) or as an example of substitution of consonants (ibdāl). The second quotation is related to three types of lisps and may well be part of a fuller discussion of similar phenomena. The third is realted to the crying out to cattle, camels, etc. (zaǧr), but it is impossible to tell whether this is included as part of a study of sounds of this type or as an example of dropping the hamza in hāyaytu>ha⁠ʾha⁠ʾtu (I cried out hī hī to the camels), which is quoted in the passage.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

213MUBAWWAB AND SPECIALIZED LEXICA

1. al-Ḥurūf, attributed to al-Ḫalīl b. Aḥmad (d. 175/791). 2. Risāla fī l-ḥurūf al-ʿArabiyya, attributed to al-Naḍr b. Šumayl (d. 203/819). 3. Ḥurūf al-hiǧāʾ by Abū l-Ḥasan al-Muzanī (d. early fourth/tenth century). 4. Ḥurūf al-maʿānī by Zaǧǧāǧī (d. 337/949). 5. Maʿānī l-ḥurūf (also known as Manāzil al-ḥurūf ) by Rummānī (d. 384/994). 6. al-Uzhiyya fī ʿilm al-ḥurūf by Abū l-Ḥasan al-Harawī (d. c. 415/1025). 7. al-Ḥurūf by Aḥmad b. Muḥammad al-Rāzī (d. after 638/1240). 8. Raṣf al-mabānī fī šarḥ ḥurūf al-maʿānī by Mālaqī (d. 702/1302). 9. Ǧawāhir al-adab fī maʿrifat kalām al-ʿArab by Irbilī (d. 741/1340). 10. al-Ǧanā l-dānī fī ḥurūf al-maʿānī by Murādī (d. 749/1348). 11. Maṣābiḥ al-maġānī fī ḥurūf al-maʿānī by Ibn Nūr al-Dīn al-Mawzaʿī

(d. 825/1422).

The short tractate attributed to al-Ḫalīl b. Aḥmad is almost certainly not his, and is nowhere mentioned in the bibliographical and biographical sources. As its editor notes,832 the first reference to it in the genre is made by Rāzī (d. after 638/1240), who summarizes its content in two different accounts in the seventh chapter of his book.833 A more substantial argument against its attribution to Ḫalīl, however, is not mentioned by the editor. This is the fact that none of the book’s twenty-nine šawāhid (i.e. one for each of the twenty- nine letters of the alphabet, including lām alif ) is found in Ḫalīl’s K. al-ʿAyn (on the assumption that he is indeed the author of al-ʿAyn) or is linked to Ḫalīl in later works. As far as the author’s method is concerned, he alphabetically lists the letters (i.e. alif, bāʾ, tāʾ, ṯāʾ, ǧīm, etc.) and gives the meaning of each (e.g. alif means “a wretched, weak man”, and dāl “a fat woman”),834 followed by a poetry šāhid.

No less doubtful is the attribution to al-Naḍr b. Šumayl of Risāla fī l-ḥurūf al-ʿArabiyya, mainly because no such title is ascribed to al-Naḍr in any source. This short monograph seems to serve a didactic purpose. It lists the letters of the alphabet in the same order as in the book attributed to Ḫalīl and sketch- ily specifies their various awǧuh/wuǧūh (types of usage), which straddle the phonological, morphological and syntactical levels. The letter kāf, for example, has five awǧuh, including being a radical of an assumed root (e.g. kafara “to disbelieve”), a substitute for qāf (e.g. kahara instead of qahara “to subdue”), and an otiose particle (zāʾida; e.g. laysa ka-miṯlihi šayʾun “There is nothing like

832 See ʿAbdaltawwāb’s introduction to Ḫalīl, Ḥurūf 12. 833 Rāzī, Ḥurūf 141–143. 834 Ḫalīl, Ḥurūf 34, 37.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

214 chapter 2

Him”; Q 42: 11).835 The number of awǧuh ranges from one (as in ṯāʾ and ḥāʾ) to twenty-two (as in alif ).

With one exception, all the above listed works from the fourth/tenth century onward are primarily syntactical in nature as they deal with ḥurūf al-maʿānī (lit. letters with meanings; i.e. particles), so named to differentiate them from ḥurūf al-hiǧāʾ or letters of the alphabet. These include mostly, but not exclu- sively, particles which govern the verb (e.g. subjunctive and jussive particles) and the noun (e.g. prepositions and inna), as well as particles which govern neither (e.g. conjunctions and interrogative particles). Other than particles, some verbs and nouns (according to the traditional classification) are men- tioned; e.g. kāna and iḏā. Before briefly examining these works, however, men- tion can be made of the one exception noted above, namely, Rāzī’s K. al-Ḥurūf, which is a short monograph containing a miscellany of topics divided into fifteen chapters. Among these topics are the manner and points of articula- tion of the letters of the alphabet (Chapters two and five); some of the wuǧūh in which each letter of the alphabet is used, above all in the substitution of one consonant by another (Chapters three and thirteen); anecdotes related to abǧad, hawwaz, etc. (Chapter four); a summary of K. al-Ḥurūf which Rāzī attri- butes to Ḫalīl (Chapter seven); various meanings of the word ḥarf (Chapter ten); dotted and undotted letters (Chapter fourteen); the letters which occur in the beginning of some sūras (Chapter fifteen); and a variety of anecdotes and lines of poetry in which certain letters (e.g. nūn and ṣād) are mentioned. The book lacks focus in general, not only due to the incongruent nature of its contents, but also because some of its material is repeated in different chapters (e.g. Chapters two and five, and Chapters three and thirteen).

Based on arrangement, or lack thereof, the remaining eight works can be divided into three groups. The first lacks any arrangement and includes Zaǧǧāǧī’s Ḥurūf al-maʿānī and Harawī’s al-Uzhiyya. Zaǧǧāǧī mentions 137 par- ticles, followed by a number of prepositions which can replace other prepo- sitions. Some of his entries are not normally recognized as particles, such as qurābataka (near you), aʿlā/asfal (above/below), and they even include expres- sions such as al-taḥiyyātu li-l-Lāhi (All sovereignty belongs to God).836 Harawī’s book is considerably larger, but includes only thirty-one particles, which are dis- cussed extensively and supported by numerous šawāhid. Like Zaǧǧāǧī’s book, a chapter is devoted following the list of relevant particles to the substitution

835 Naḍr, Ḥurūf 165. 836 Zaǧǧāǧī, Ḥurūf 18, 23, 22 respectively.

 

 

215MUBAWWAB AND SPECIALIZED LEXICA

that takes place among prepositions, after which Harawī discusses allaḏī and other relative pronouns.837

The second group includes works in which particles are alphabetically arranged regardless of the number of their radicals (in case these vary in num- ber). The earliest of these is K. Ḥurūf al-hiǧāʾ by Abū l-Ḥasan al-Muzanī. Its material is restricted to the letters of the alphabet, which are described at the phonological, morphological and syntactical levels. As in al-Naḍr’s Risāla, the particles that are included are merely those which are made up of only one letter, such as the preposition bi- and the conjunction fa-.838 The other two works in this group, Mālaqī’s Raṣf al-mabānī and Ibn Nūr al-Dīn’s Maṣābīḥ al-maġānī, are quite comprehensive. Mālaqī follows full alphabetical order based on the Maġribī arrangement of the letters of the alphabet (hence parti- cles beginning with ʿayn, fāʾ, qāf and sīn occur after those beginning with nūn).839 In contrast, Ibn Nūr al-Dīn does not go beyond the first letter of the word in his alphabetical arrangement, probably influenced by Ibn Hišām’s (d. 761/1360) Muġnī l-labīb ʿan kutub al-Aʿārīb, a famous work on grammar whose first part is an extensive treatise on ḥurūf al-maʿānī.840 Probably also based on the Muġnī,841 Ibn Nūr al-Dīn includes several particles which do not feature in many other works, including Mālaqī’s.

The third group includes three books in which particles are divided accord- ing to the number of their letters (i.e. from one to five) and each group is alpha- betically arranged. These are Rummānī’s Maʿānī l-ḥurūf – which precedes the other two by almost four centuries – Irbilī’s Ǧawāhir al-adab, and Murādī’s al-Ǧanā l-dānī. In all three books, alphabetical arrangement is ignored at times, as in the section on quadriliterals in Rummānī’s book.842 In grouping particles, long vowels and doubled consonants are taken into account, hence the classification of wā and rubba with biliterals and triliterals respectively. Quinqueliterals include only lākinna in Irbilī’s work,843 to which Murādī adds

837 Harawī, Uzhiyya 277–300, 301–316. 838 Muzanī, Ḥurūf I, 514; II, 53. 839 Mālaqī, Raṣf 366–398. 840 Ibn Hišām, Muġnī I, 9-II, 374. 841 The editor of Maṣābīḥ al-maġānī (p. 25) counts more than eighty passages in which Ibn

Nūr al-Dīn obviously relies on Ibn Hišām, although he acknowledges him only once. 842 Rummānī, Maʿānī 118–134. 843 Irbilī, Ǧawāhir 430. That lākinna is quinqueliteral is based on its alleged derivation from

lā + ka + inna. The total number of its ḥurūf however should have been six according to the author’s classification of its three components in their respective chapters (2+1+3), but it seems that he did not wish to exceed the maximum number of radicals in nouns (i.e. five) where only consonants are counted.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

216 chapter 2

antumā and antunna when used as pronouns of separation ( faṣl).844 Since Rummānī discusses lākinna under lākin,845 he does not assign a separate chap- ter for quinqueliterals.

That several works on ḥurūf al-maʿānī follow an alphabetical arrange- ment and are thus true lexica is almost unique in the grammatical tradition. Obviously, these particles lend themselves to such an arrangement, contrary to the various other topics which grammatical works normally discuss. Even an author such as Rummānī, known for the complexity of his approach to grammar and for combining logic with grammar,846 must have found it use- ful to adopt an alphabetical arrangement of particles, possibly for didac- tic purposes. Another lexicon related to grammar but on a different topic is K. Šarḥ abniyat Sībawayhi, in which Ibn al-Dahhān (d. 569/1164) explains the meanings of apparently difficult words which Sībawayhi cites to illustrate morphological patterns. The author arranges this closed set of words in full alphabetical order so as to facilitate the use of his book, as he notes in his introduction.847 A similar work is Ǧawālīqī’s (d. 540/1145) abridgement of Šarḥ amṯilat Sībawayhi848 by Muḥammad b. ʿĪsā b. ʿUṯmān al-ʿAṭṭār, whose dates are not known, but who was a student of Sīrāfī’s (d. 368/979).849 Its alphabetical arrangement, however, does not go beyond the first letter of each word.

Other than works which deal with particles, there is a large number of works which deal with specific letters of the alphabet. These include hamza/alif, lām, wāw, hāʾ and yāʾ, as well as several pairs of letters, most notably ḍād and ẓāʾ (at times with other letters), sīn and ṣād, sīn and šīn and dāl and ḏāl.850 We shall survey these in this order and then discuss works on iddiġām, ibdāl and itbāʿ.

Due to orthographic reasons, philologists often use hamza and alif interchangeably, as in the term hamzat/alif al-waṣl (conjunctive hamza/alif ).

844 Murādī, Ǧanā 620. 845 Rummānī, Maʿānī 133–134. 846 Cf. Ibn al-Anbārī, Nuzha 234; Suyūṭī, Buġya II, 181. 847 Ibn al-Dahhān, Šarḥ 27. A similar, but earlier work by Abū Ḥātim al-Siǧistānī (d. 255/869),

Tafsīr ġarīb mā fī Kitāb Sībawayhi min al-abniya, is not alphabetically arranged. 848 Note that Ǧawālīqī’s abridgement is published under the same title (i.e. Šarḥ amṯilat

Sībawayhi), and not as Muḫtaṣar Kitāb . . . or the like. 849 Suyūṭī, Buġya I, 206. 850 See below, 224–225 for ġayn and rāʾ and ṭāʾ and tāʾ. Note also that Ibn al-Dahhān men-

tions in his Aḍdād (p. 6) a book of his own entitled K. al-Ġayn wa-l-rāʾ, which he says is alphabetically arranged according to the roots of words. This may have included words in which rāʾ is changed to ġayn due to luṯġa (lisp). Cf. Yāqūt, Muʿǧam III, 1371, where the title is cited as K. Izālat al-mirāʾ fī l-ġayn wa-l-rāʾ. For this and other types of luṯġa, cf. Suyūṭī, Muzhir I, 556–566.

 

amgad
Highlight

 

217MUBAWWAB AND SPECIALIZED LEXICA

The dialectal variations related to pronouncing the hamza (taḥqīq, nabr) and lightening it (taḫfīf, tashīl),851 as well as its substitution (ibdāl) by wāw and yāʾ, are frequently referred to in grammatical and multithematic works.852 They also feature in books on qirāʾāt since they give rise to various Qurʾānic readings.853 Issues related to hamza and alif however are the exclusive subject of works attributed to philologists as early as ʿAbdallāh b. Abī Isḥāq al-Ḥaḍramī (d. 117/735).854 The following extant sources which deal with these issues are mostly short monographs:

1. al-Hamz by Abū Zayd al-Anṣārī (d. 215/830). It consists of twenty-nine short and untitled chapters which include hamzated words arranged according to the position which the hamza occupies (i.e. initial, medial or final), or accord- ing to its vowel or the vowel that precedes it. An attempt at alphabetical arrangement based on the initial letter of words is sometimes noted,855 but is not systematically applied in all chapters.

2. Šarḥ al-alifāt856 by Abū Bakr b. al-Anbārī (d. 328/940). Its material, which is restricted to initial hamza, is divided into three types, based on whether the word in question is a verb, noun, or particle (including demonstratives, such as ulāʾikum). In the case of verbs, for example, six varieties of hamza (each of which is referred to as alif ) are recognized, some of which do not belong to the morphological level. These are alif al-aṣl (i.e. a radical), alif al-qaṭʿ (i.e. disjunctive hamza), alif al-waṣl (i.e. conjunctive hamza), alif al-istifhām (inter- rogative hamza), alif al-muḫbir ʿan nafsihi (hamza of the first person singular in the imperfect, as in aʿbudu “I worship”), and alif mā lam yusamma fāʿiluhu (i.e. the hamza which occurs in passive forms of the perfect, as in the patterns ufʿila and uftuʿila).857

3. al-Alifāt by Ibn Ḫālawayhi (d. 370/980). The author mentions in his intro- duction eighty-nine types of alif, most of which, by his own admission, are

851 Cf. above, 13. 852 Cf. Sībawayhi, Kitāb III, 541–556; Mubarrad, Muqtaḍab I, 62–63, 155–167; Abū ʿUbayd,

Ġarīb III, 680–684. 853 E.g. Ibn Muǧāhid, Sabʿa 136–144; Makkī, Kašf I, 70–121; Ibn al-Ǧazarī, Našr I, 362–491. 854 To ʿAbdallāh is attributed K. al-Hamz in Abū l-Ṭayyib, Marātib 31. For the competition

between him and Abū ʿAmr b. al-ʿAlāʾ (d. 154/770) in mastering matters related to hamz, see Sīrāfī, Aḫbār 26; Zubaydī, Ṭabaqāt 31; Qifṭī, Inbāh II, 106.

855 E.g. ḫf ʾ is palced before ḥlʾ, but then the arrangement becomes systematic, hence ḍbʾ, ṭʾṭʾ, ẓmʾ, ʿbʾ, f ʾw, qnʾ, lbʾ and msʾ; cf. Abū Zayd, Hamz 843–849.

856 Also published under the title Muḫtaṣar fī ḏikr al-alifāt, ed. by Ḥasan Šāḏilī Farhūd, Dār al-Turāṯ, Cairo 1980.

857 Ibn al-Anbārī, Alifāt 283, 284–285, 448, 450 (for the fifth and sixth types) respectively.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

218 chapter 2

marginal ( furūʿ)858. These include, for example, the alif of nouns in the accusa- tive and the one that is part of the feminine plural ending -āt. The text of the book, however, is restricted to alif al-waṣl in verbs and nouns, alif al-aṣl, alif al-faṣl (i.e. the alif written at the end of masculine plural perfect verbs), and alif al-qaṭʿ.

4. Mā yaḥtāǧ ilayhi l-kātib min mahmūz wa-maqṣūr wa-mamdūd mimmā yuktab bi-l-alif wa-l-yāʾ ʿalā ḥurūf al-muʿǧam859 by Ibn Ǧinnī (d. 392/1002). This contains a small collection of words of finally hamzated verbs (e.g. bada⁠ʾa, ǧaba⁠ʾa, ḥaša⁠ʾa, ḫaba⁠ʾa, etc.) which are simply listed (i.e. with no explanation or šawāhid), mostly in full alphabetical order. Following this is a very short sec- tion on maqṣūr and mamdūd.860

5. ʿUqūd al-hamz wa-ḫawāṣṣ amṯilat al-fiʿl, also by Ibn Ǧinnī, is a very short risāla on the writing of hamza, on the basis of its position in the word and its vocalic environment.

6. al-Alifāt wa-maʿrifat uṣūlihā by Dānī (d. 444/1053). This monograph resembles Ibn al-Anbārī’s (see no. 2 above), not only in its division of hamza into the same three types but also in the varieties listed under each type (e.g. the same six terms are used with verbs as in Ibn al-Anbārī’s book).

7. Bāb al-hiǧāʾ by Ibn al-Dahhān (d. 569/1174). Much of the material of this monograph is related to the writing of both alif and hamza.

8. Šarḥ al-Naẓm al-awǧaz fī mā yuhmaz wa-mā lā yuhmaz by Ibn Mālik (d. 672/1274). This consists of 218 lines of poetry (of the ṭawīl meter, hence not an urǧūza). Each line, save for the introduction and line 189, includes two or more sets, each consisting of two words that are attested with or without a hamza. The first 188 lines, which form the first section, embrace words whose meanings differ according to whether they are hamzated or not (e.g. qara⁠ʾa “to read” and qarā “to give hospitality to”).861 The second section includes nineteen lines in which are mentioned words whose meanings are the same regardless of whether they are hamzated or not (e.g. arǧa⁠ʾtu and arǧaytu “I postponed”).862

The following works deal with the letter lām:

858 Ibn Ḫālawayhi, Alifāt 18. 859 This risāla is also known as al-Alfāẓ al-mahmūza and was published under this title by

Ṣalāḥ al-Dīn al-Munaǧǧid in Ṯalāṯ rasāʾil fī l-luġa, Dār al-Kitāb al-Ǧadīd, Beirut 1981, and by Māzin al-Mubārak in Risālatān li-Ibn Ǧinnī, Dār al-Fikr al-Muʿāṣir, Beirut & Dār al-Fikr, Damascus 1988.

860 See also below, 247. 861 Ibn Mālik, Naẓm 31. 862 Ibid., 122.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

219MUBAWWAB AND SPECIALIZED LEXICA

1. al-Lāmāt by Zaǧǧāǧī (d. 337/949). The author identifies thirty-one types of lām, the vast majority of which belong to the syntactical level – such as lām al-ibtidāʾ (inceptive la-) and lām al-amr (li- of command)863 – although some belong to morphology – e.g. the otiose lām suffixed to ʿabd (male slave), yield- ing ʿabdal, which also means “male slave”.864

2. Risāla fī l-lāmāt by Abū Ǧaʿfar al-Naḥḥās (d. 338/950). Eleven types of lām are identified, all of which occur in syntax, with each type supported by numer- ous šawāhid, almost exclusively from the Qurʾān. The risāla thus belongs to works that are devoted to Qurʾānic lāmāt, of which Ibn al-Nadīm (d. 380/990) mentions a few.865

3. al-Lāmāt by Ibn Fāris (d. 395/1004). As is clear in its introduction, this monograph deals exclusively with Qurʾānic lāmāt. Ibn Fāris notes that some authors reduce the types of lām to eight, while others increase this number (kaṯṯara) to more than thirty.866 For his part, he mentions twelve types, all of which are related to syntax, beginning with five whose vowel is fatḥa, hence la-, followed by seven whose vowel is kasra, hence li-.

4. al-Lāmāt by Abū l-Ḥasan al-Harawī (d. 415/1025). This is the most exten- sive work in the genre and includes thirty-four types of lām, mostly in syntax. In all these types, according to Harawī, the lām is affixed (zāʾida), in contrast to the lām referred to as aṣliyya, that is, the lām that constitutes one of the radicals of a word’s root.867

On the syntactical use of wāw, there is a lengthy work by Ḫalīl b. Kaykaldī al-ʿAlāʾī (d. 761/1359) entitled al-Fuṣūl al-mufīda fī l-wāw al-mazīda. Among the major types discussed are al-wāw al-ʿāṭifa (copulative wāw), wāw al-ḥāl (wāw of circumstance), the wāw which precedes the mafʿūl maʿahu (concomitate object), and the wāw which precedes the imperfect subjunctive verb.868

Both hāʾ and yāʾ are part of a short monograph by Abū Bakr b. al- Anbārī (d. 328/940) entitled Ḏikr mā yūqaf ʿalayhi bi-l-hāʾ wa-l-yāʾ fī Kitāb Allāh. Ibn al-Anbārī illustrates by a handful of words that the feminine marker in Qurʾānic reading becomes hāʾ (i.e. -a) in pause. In each word, such as raḥma (mercy), he says that it always ends in hāʾ, except in a certain number of verses, which he then quotes. In these verses (seven in the case of raḥma), the feminine word is almost uniformly in the construct state. Following a chapter on the writing

863 Zaǧǧāǧī, Lāmāt 69, 88. 864 Ibid., 143. 865 Ibn al-Nadīm, Fihrist 38. 866 Ibn Fāris, Lāmāt 771. 867 Harawī, Lāmāt 29. 868 Ibn Kaykaldī, Fuṣūl 55, 155, 188, 207.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

220 chapter 2

of some compound particles and words as one unit or two separate units (e.g. allā/an+lā, mimmā/min+mā, biʾsamā/biʾsa+mā), Ibn al-Anbārī illustrates the omission of the first person singular suffix -ya in Qurʾānic words of the type rabbi<rabbī (Oh my Lord) and ʿibādi<ʿibādī (My servants). By describing yāʾ as sāqiṭ (omitted) Ibn al-Anbārī obviously refers to orthography since the long -ī is merely shortened in pronunciation to -i.

Besides works which deal with single letters, some deal with a pair of let- ters that need to be differentiated. Works on ḍād and ẓā ʾ are by far the most numerous in the tradition. This should come as no surprise given the very early riwāyas which demonstrate that these two letters were often confused by speakers. The Caliph ʿUmar (d. 23/644), for example, reportedly asked a man who used the word ḍaby (antelope) why he did not use (the more frequent) ẓaby. The man responded by asserting that the former is a dialectal variant (luġa).869 Furthermore, one of the anecdotes cited to justify the alleged found- ing (waḍʿ) of naḥw by Abū l-Aswad al-Duʾalī (d. 69/688) has it that he heard a Persian man say farasī ḍāliʿ (My horse is lame), instead of ẓāliʿ.870 Several authors of the genre point out that the two letters are the source of confusion not only in pronunciation but also in writing (iltibās ḥaqīqat kitābatihimā).871 It should be noted that the inclusion by authors of words that are extremely common (e.g. ẓalla “to continue” and ḍalla “to stray”; ḥaẓara “to prohibit” and ḥaḍara “to attend”)872 demonstrates that the mix-up in ḍād and ẓāʾ permeated the whole vocabulary as a dialectal phenomenon and that its occurrence was not restricted to literary or less commonly used words.

Works on ḍād and ẓāʾ range from short monographs to lengthy books, as the following list shows:873

1. Urǧūzat al-ẓāʾ wa-l-ḍād, wrongly attributed to Ibn Qutayba (d. 276/889). Its real author is most probably Abū Naṣr Muḥammad b. Aḥmad al-Farrūḫī (d. 557/1162).874 The text includes forty-five lines, each of which (as of the

869 Qālī, Amālī (Ḏayl) III, 142; cf. Suyūṭī, Muzhir II, 562–563, where the man is reported to have used the expression ẓaḥḥā bi-ḍabyin (i.e. ḍaḥḥā bi-ẓabyin “sacrificed an antelope”), in which ẓ replaces ḍ and vice versa.

870 Sīrāfī, Aḫbār 18; Zubaydī, Ṭabaqāt 22. 871 Ṣāḥib, Farq 3; cf. Zanǧānī, Farq 19. 872 Both examples are derived from Ibn Qutayba, Urǧūza 462–463. 873 For a fuller list of thirty works, including lost and unpublished titles, see ʿAbdaltawwāb’s

introduction to Ibn al-Anbārī, Zīna 23–35. In his introduction to Qaysī, Maʿrifa 9–10, Ḍāmin adds ten more titles to ʿAbdaltawwāb’s list. See also Ḍāmin’s introduction to Ibn Mālik, Iʿtimād 6–12.

874 Cf. ʿAbdaltawwāb’s introduction to Anbārī, Zīna 28–29. See also Muḥsin (2008).

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

221MUBAWWAB AND SPECIALIZED LEXICA

ninth) begins with a word that includes ẓāʾ, followed by its counterpart with ḍād, and the meaning of each.

2. al-Farq bayna l-ḍād wa-l-ẓāʾ by Abū l-Qāsim Ismāʿīl b. ʿAbbād al-Ṭālaqānī, better known as al-Ṣāḥib b. ʿAbbād (d. 385/995). The author lists and explains twenty-seven instances of words (normally including several derivatives from the same root) whose meanings differ depending on whether they are pro- nounced with a ẓāʾ or a ḍād (in this order).875 These are followed by fifty-eight words (and their derivatives) that include a ẓāʾ, but have no counterparts with a ḍād.876

3. al-Rawḥa by Muḥammad b. al-Ḥasan al-Ǧarbāḏaqānī (fl. in the second half of the fourth/tenth century). This is an impressive work which extensively explains words that include either a ḍād or a ẓāʾ, at times in more detail and with more šawāhid than some of the most comprehensive muǧannas lexica. It also presents a rare example of the phonetic system of arrangement used out- side the sphere of lexica which cover the whole language. The author begins each letter by explaining the words in which it is combined with ḍād, and then with ẓāʾ. Accordingly, the first few chapters are ʿ +ḍ,ʿ+ẓ, ḥ+ḍ, ḥ+ẓ, ġ+ḍ, ġ+ẓ, ḫ+ḍ, ḫ+ẓ877 (note that the order of letters differs from that in Ḫalīl’s al-ʿAyn). The quadriliterals are normally mentioned at the end of the chapters, and some roots are followed by their taqālīb or permutations (e.g. ḥḍr and ḥrḍ).878

4. al-Ḍād wa-l-ẓāʾ by Abū l-Faraǧ b. Suhayl al-Naḥwī (d. fifth/eleventh cen- tury). It consists of two sections, one for ḍād and another for ẓāʾ. The first section is then divided into smaller units, each of which includes words that consist of one of the letters of the alphabet (from alif to yāʾ) and ḍād. The words within each unit are not alphabetically arranged. The same process is then repeated for ẓāʾ in the second section. Accordingly, the first unit of the first section is for alif and ḍād (and includes the words arḍ, aġḍaytu, aḍaǧǧa, aqaḍḍa, aḍarra, iḍṭalaʿa, afḍā, awmaḍa, armaḍa, etc.),879 the second for bāʾ and ḍād, the third for tāʾ and ḍād, and so on.

5. Four lines of poetry by Abū l-ʿAbbās Aḥmad b. ʿAmmār al-Muqriʾ (d. c. 440/1048), in which he assembles twenty-nine words which contain ẓāʾ and are used in the Qurʾān. These lines are preserved in a commentary by Ismāʿīl b. Aḥmad al-Tuǧībī al-Barqī (d. c. 445/1053).880

875 Ṣāḥib, Farq 4–26. 876 Ibid., 26–42. 877 Ǧarbāḏaqānī, Rawḥa 2, 37, 48, 70, 100, 116, 121, 148. 878 Ibid., 53, 57. 879 Ibn Suhayl, Ḍād 290–292. 880 Tuǧībī, Šarḥ 33–126; see p. 35 for Ibn ʿAmmār’s lines.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

222 chapter 2

6. al-Farq bayna l-ḍād wa-l-ẓāʾ fī Kitāb Allāh ʿazza wa-ǧalla wa-fī l-mašhūr min al-kalām by Abū ʿAmr al-Dānī (d. 444/1053). Being the fewer of the two, the book is restricted to words with ẓāʾ, as the author explains.881 Following a chap- ter on the phonetic characteristics of both letters, the book consists of thirty- two chapters in each of which are cited mostly Qurʾānic words which contain ẓāʾ and are of the same root. The first chapter, for example, includes words of the root ẓnn, such as ẓann, ẓunūn, ẓanīn, ẓanūn, taẓannī, etc.882 Following these are fifty-four words (some with other derivatives from the same root), which do not occur in the Qurʾān.883

7. Fī maʿrifat al-ḍād wa-l-ẓāʾ by ʿAlī b. Abī l-Faraǧ al-Qaysī al-Ṣiqillī (d. after 470/1077). Words with ḍād and those with ẓāʾ are listed in two separate sections which are followed by two short sections for words which are pronounced with ḍād or ẓāʾ, and either have the same or different meanings.

8. al-Farq bayna l-ḍād wa-l-ẓāʾ by Abū l-Qāsim al-Zanǧānī (d. 471/1078). It includes twenty-nine sets, of two words each, in which the one with ḍād differs in meaning from that with ẓāʾ (e.g. ʿaẓm “bone” and ʿaḍm “handle of a bow”; ġayẓ “wrath” and ġayḍ “sinking of water”).884 In the discussion are included derivatives from the same roots as those of the two representative words in the relevant set.

9. Ḥaṣr ḥarf al-ẓāʾ by Abū l-Ḥasan al-Ḫawlānī (d. after 485/1092). As its title indicates, this short tractate is restricted to words with ẓāʾ. Under each letter of the alphabet in its Maġribī arrangement are listed, in no specific order, words which include ẓāʾ. The total number of these words is ninety-three.

10. al-Farq bayna l-ḍād wa-l-ẓāʾ by Ḥarīrī (d. 516/1122), author of the maqāmāt. The author explains in his introduction that his book lists, in alpha- betical order, words with ẓāʾ so that they be differentiated from those with ḍād. Ḥarīrī is also author of nineteen lines of poetry on words with ẓāʾ as part of his al-Maqāma al-Ḥalabiyya.885

11. Zīnat al-fuḍalāʾ fī l-farq bayna l-ḍād wa-l-ẓāʾ by Abū l-Barakāt b. al-Anbārī (d. 577/1181). The first section of the book lists words with ḍād only, the second words with ẓāʾ only, and the third words whose meanings change depending on whether they are pronounced with ḍād or ẓāʾ. None of these sections is alphabetically arranged.

881 Dānī, Farq 30. 882 Ibid., 36–40. 883 Ibid., 89–99. 884 Zanǧānī, Farq 24, 31. 885 Ḥarīrī, Maqāmāt 504–507.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

223MUBAWWAB AND SPECIALIZED LEXICA

12. Ẓāʾāt al-Qurʾān by Abū l-Rabīʿ al-Saraqūsī (d. before 591/1194). The text begins with three lines of poetry composed by the author himself, in which he cites twenty-one words which represent all the roots with ẓāʾ in the Qurʾān. In explaining these words, he cites other Qurʾānic words derived from the same roots and at times their counterparts with a ḍād, also occurring in the Qurʾān.

13. Muḫtaṣar fī l-farq bayna l-ḍād wa-l-ẓāʾ by Muḥammad b. Našwān al-Ḥimyarī (d. 610/1213). Unlike earlier works, the sections of this book are arranged according to the number of radicals of the cited words and the posi- tion (i.e. initial, medial and final) of the ḍād and ẓāʾ. Accordingly, it starts with biliterals in which either letter occurs initially, followed by those in which it occurs medially. Triliterals then follow, arranged according to initial, medial and final positions. The final section, which includes words with ẓāʾ only, is also arranged based on the number of roots of the word, and includes quad- riliterals. Full alphabetical order is observed in each section. It is noteworthy that Ibn Našwān, by adopting a division based on morphological grounds, was probably influenced by the system which his father, Našwān b. Saʿīd al-Ḥimyarī (d. 573/1178), had adopted in his famous Šams al-ʿulūm.886

Later authors chiefly repeat the same material which the above mentioned works include. Among these authors are Aḥmad b. Ḥammād al-Ḥarrānī (d. after 618/1221) in al-Miṣbāḥ fī l-farq bayna l-ḍād wa-l-ẓāʾ fī l-Qurʾān al-ʿazīz naẓman wa-naṯran; Muḥammad b. Aḥmad b. al-Ṣābūnī (d. 634/1237) in Maʿrifat al-farq bayna l-ẓāʾ wa-l-ḍād; Yūsuf b. Ismāʿīl al-Maqdisī (d. 637/1240) in al-Ẓāʾ; Ibn Mālik, author of the famous Alfiyya (d. 672/1274), in several works, including al-Iʿtimād fī naẓāʾir al-ẓāʾ wa-l-ḍād, al-Iʿtiḍād fī l-farq bayna l-ẓāʾ wa-l-ḍād, and Urǧūza fī l-farq bayna l-ḍād wa-l-ẓāʾ; Abū Ḥayyān al-Naḥwī al-Andalusī (d. 745/ 1344) in al-Irtiḍāʾ fī l-farq bayna l-ḍād wa-l-ẓāʾ; Muḥammad b. Abī Bakr al-Šaṭṭī (d. 749/1348) in Kāšif maḥāsin al-ġurra li-ṭālib manāfiʿ al-Durra (i.e. Rasʿanī’s [d. 661/1263] poem of thirty-two verses entitled Durrat al-qāriʾ fī l-farq bayna l-ḍād wa-l-ẓa⁠ʾ); and ʿAbdallāh b. ʿAlī al-Mawṣilī (d. 797/1395) in al-Farq bayna l-ḍād wa-l-ẓāʾ.

To ḍād and ẓāʾ is added the letter ḏāl in al-Iqtiḍāʾ li-l-farq bayna l-ḏāl wa-l- ḍād wa-l-ẓāʾ by Abū ʿAbdallāh Muḥammad b. Aḥmad al-Dānī (d. 470?/1077?). The book is divided into several sections, none of which is alphabetically arranged. The first section deals with words that have ḍād, ẓāʾ or ḏāl, but with different meanings (e.g. ḥaḍḍ “urging”, ḥaẓẓ “fortune”, and ḥaḏḏ “cutting off”).887 The other sections include pairs of words which have either ḍād and ẓāʾ, ẓāʾ

886 See below, 265. 887 Dānī, Iqtiḍāʾ 39–40.

 

amgad
Highlight
amgad
Highlight

 

224 chapter 2

and ḏāl, or ḍād and ḏāl. Also noted are some words which may be written with ẓāʾ, ḍād or ḏāl, such as ḥanẓā, ḥanḍā and ḥanḏā (to curse).888

The title of Ibn al-Sīd al-Baṭalyawsī’s (d. 521/1127) al-Farq bayna l-ḥurūf al-ḫamsa (i.e. ẓāʾ, ḍād, ḏāl, ṣād and sīn) is rather confusing since it could imply that words with any of these letters are discussed as one group. In fact, the book – which is one of the most comprehensive in the genre of ḍād and ẓāʾ – is in two parts, neither of which is alphabetically arranged. The first part is devoted to words with ẓāʾ, ḍād and ḏāl. Sections in this part greatly resemble those in Dānī’s al-Iqtiḍāʾ, and the two books also share a large number of their examples. The second part examines words with ṣād and sīn (e.g. ṣāʿid “ascend- ing” and sāʿid “forearm”, but ṣalhab and salhab, which both mean “tall”).889 As the author notes, this second part is larger than the first.890 In discussing these five letters under one title, Baṭalyawsī seems to continue an earlier tra- dition, since we know that Abū l-Fahd al-Naḥwī al-Baṣrī, a contemporary of Mubarrad’s (d. 285/898) and teacher of Zaǧǧāǧ’s (d. 311/923),891 authored a book entitled K. al-Ẓāʾ wa-l-ḍād wa-l-ḏāl wa-l-sīn wa-l-ṣād.892 It has also been pointed out earlier that Ibn al-Sikkīt’s (d. 244/858) Iṣlāḥ al-manṭiq includes a chapter on the incorrect use of ṣād instead of sīn and vice versa.893

On sīn and ṣād as well, we are in possession of Muḥammad b. Aḥmad al-Anṣārī’s (d. c. 470/1077) al-Tabyīn wa-l-iqtiṣād fī l-farq bayna l-sīn wa-l-ṣād. It follows no alphabetical arrangement in any of its three parts, the first of which explains words with sīn or ṣād that differ in meaning. The second part is restricted to words with ṣād only, and the third to words that can be writ- ten either with sīn or ṣād. Mention should also be made of the fifth maqāma of al-Maqāmāt al-Ǧalāliyya (al-Ṣafadiyya) by al-Ḥasan b. Abī Muḥammad al-Ṣafadī (fl. in the first quarter of the eighth/fourteenth century). It is called al-Maqāma al-ṣād-sīniyya, but to this title is added in the margin wa-l-ġayn- rāʾiyya wa-l-ṭāʾ-tāʾiyya. It includes eighty pairs of words, each with ṣād and sīn. The vast majority of these pairs have different meanings (e.g. ṣayf “sum- mer” and sayf “sword”), but a few have the same meaning (e.g. ṣirāṭ and sirāṭ “way”). These are followed by thirty-five pairs of words, in most of which ġayn occupies the same position as its counterpart (e.g. šaġaf “passion” and šaraf

888 Ibid., 179. 889 Baṭalyawsī, Farq 377, 703. 890 Ibid., 371. 891 Zubaydī, Ṭabaqāt 119; Ibn al-Nadīm, Fihrist 93; Qifṭī, Inbāh IV, 158; cf. Fīrūzābādī, Bulġa 185;

Suyūṭī, Buġya II, 249. 892 Ibn Ḫayr, Fahrasa 363. 893 See above, 174.

 

amgad
Highlight

 

225MUBAWWAB AND SPECIALIZED LEXICA

“honor”), although at times both letters occur in two words in a different order (e.g. ġārib “withers” and rāġib “willing”). These peculiar pairs with ġayn and rāʾ, particularly of the first type, may have been listed because they were confused due to the defective pronunciation of rāʾ.894 At the end of the maqāma are listed in the margin fifteen pairs with ṭāʾ and tāʾ (e.g. ṭīn “clay” and tīn “fig”), but these seem to be a later addition to the text.

Words which are attested both with sīn and šīn without change in meaning (e.g. tafassa⁠ʾa and tafašša⁠ʾa, both meaning “to spread”)895 are alphabetically listed by Fīrūzābādī (d. 837/1415) in his Taḥbīr al-muwaššīn fī l-taʿbīr bi-l-sīn wa-l-šīn. Unlike his famous al-Qāmūs al-muḥīṭ, the arrangement is not based on the final consonant of the root of the word, but on the first, and augments are at times taken into consideration in the arrangement (e.g. taḥasḥus, tasar- rum and taʿakkus are placed under tāʾ).896 The number of entries is 142 (two of which appear twice), and it is not clear whether Fīrūzābādī’s other book, al-Taḥbīr al-kabīr, to which he alludes,897 contained – as its title could suggest – a larger number of entries.

The final pair of letters discussed by philologists is dāl and ḏāl. On this, we are in possession of Muqaddima fī kalimāt ittafaqat fīhā l-dāl wa-l-ḏāl ḫaṭṭan wa-ḫtalafat maʿnan by Murādī (d. 749/1348), author of the previously men- tioned work on ḥurūf, al-Ǧanā l-dānī. It consists of twenty-five verses, each of which (except for the first and the last) includes two words of different mean- ings, one with dāl and the other with ḏāl (e.g. dabba “to creep” and ḏabba “to drive away”).898 Murādī briefly explains the meaning of the pair of words in each of his verses.

Characteristics of ḥurūf further feature in the study of iddiġām, ibdāl and itbāʿ. Iddiġām (also idġām) refers to the assimilatory process resulting in the gemination of consonants or their change (e.g. yaẓlimūnnī< yaẓlimūnanī “they treat me unjustly”; battuʾṯirūna<bal tuʾṯirūna “yet, you prefer”; Q 87: 16).899 It is mostly discussed within works on naḥw and ṣarf, as in the last few chapters of Sībawayhi’s (d. 180/796) al-Kitāb900 and in a lengthy chapter in Ibn ʿUṣfūr’s (d. 669/1271) al-Mumtiʿ.901 Many of the intricate details of iddiġām also occur

894 See above, 216, n. 850. 895 Fīrūzābādī, Taḥbīr 49. 896 Ibid., 44, 46, 47. 897 Ibid., 69. 898 Murādī, Muqaddima 141. 899 Sībawayhi, Kitāb IV, 438, 459. 900 Ibid., IV, 431–485. 901 Ibn ʿUṣfūr, Mumtiʿ II, 631–727.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

226 chapter 2

in Sīrāfī’s (d. 368/979) commentary on Sībawayhi’s al-Kitāb.902 In addition, Sīrāfī authored an independent monograph on iddiġām, entitled Mā ḏakarahu l-Kūfiyyūn min al-iddiġām. It consists of eleven controversial issues (masāʾil), in which Sīrāfī refutes the Kufan views on matters related to iddiġām and gener- ally characterizes their approach to it as incomplete and not comprehensive.903

In contrast, several independent works deal with lexical ibdāl. The term ibdāl (substitution, replacement) in the tradition comprises two types, the first of which, grammatical ibdāl (ibdāl naḥwī), deals with certain morphopho- nological changes in nouns, verbs and particles, and is restricted to a num- ber of consonants (eleven, according to Sībawayhi),904 usually mentioned in grammatical sources. For example, the tāʾ of the pattern iftaʿala is substituted by ṭāʾ when the first radical of the root is ḍād, as in iḍṭahada<*iḍtahada (to oppress).905 Lexical ibdāl (ibdāl luġawī), on the other hand, comprises exam- ples from practically all consonants, and receives little attention in the gram- matical theory.906 It is however discussed in independent monographs, among the oldest of which are Abū ʿUbayda’s (d. 209/824) K. al-Ibdāl 907 and Aṣmaʿī’s (d. 216/831) K. al-Qalb wa-l-ibdāl,908 both lost to us.

From extant works of the genre and other philological sources, we know of several synonyms for ibdāl, such as qalb, muʿāqaba, taʿāqub, iʿtiqāb, naẓāʾir and ištiqāq akbar. The examples quoted in the sources are not restricted to doublets (or, less often, triplets) which differ in phonemes that have adjacent points of articulation, such as kaṯab and kaṯam (nearness)909 or quḥḥ and kuḥḥ (pure),910 but also include some doublets whose corresponding phonemes in question are quite dissimilar. Many of these are most probably the result of taṣḥīf (alter- ation of diacritical points), as in ʿataša and ʿanaša (to bend), and zuḥlūfa and

902 The section on iddiġām in Sīrāfī’s Šarḥ Kitāb Sībawayhi was published independently under the title K. al-Iddiġām (see primary sources).

903 Sīrāfī, Mā ḏakarahu 132 (wa-maḏhab al-Kūfiyyīn fī l-iddiġām qalīl laysa bi-ʿāmm mustawʿib li-l-ḥurūf wa-l-kalām ʿalayhā).

904 Kitāb IV, 237. The eleven letters are hamza, alif, hāʾ, yāʾ, tāʾ, dāl, ṭāʾ, mīm, ǧīm, nūn and wāw. Some grammarians add further letters to Sībawayhi’s list; see, for example, Ibn ʿUṣfūr, Mumtiʿ I, 410, where sīn, ṣād, zāy, ʿayn, kāf, fāʾ and šīn are mentioned.

905 Ibid., 239; cf. Ibn Ǧinnī, Munṣif II, 327–330. 906 For a detailed study on lexical ibdāl and works that deal with it, see Hämeen-Anttila

(1993). 907 Yāqūt, Muʿǧam VI, 2708. The authenticity of the attribution to Abū ʿUbayda of a

monograph on ibdāl, however, is questioned by Hämeen-Anttila (1993: 12–13). 908 Ibn al-Nadīm, Fihrist 61; Qifṭī, Inbāh II, 203; Suyūṭī, Buġya II, 113. 909 Ibn al-Sikkīt, Ibdāl 73; Abū l-Ṭayyib, Ibdāl I, 49. 910 Ibn al-Sikkīt, Ibdāl 113; Abū l-Ṭayyib, Ibdāl II, 357.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

227MUBAWWAB AND SPECIALIZED LEXICA

zuḥlūqa (a sloping slide),911 where the orthographic similarity between tāʾ and nūn and between fāʾ and qāf seems to have resulted in two different readings in each case. Certain doublets might also be the result of luṯġa (lisp), such as ǧinṯ and ǧins (origin).912 Indeed, Ǧawharī (d. c. 400/1010) indicates that he is not sure whether the former is a dialectal variant (luġa) or the result of lisping (luṯġa).913 It is obvious that he believes that dialectal differences can lead to ibdāl, and in this he is in agreement with Abū l-Ṭayyib al-Luġawī (d. 351/962), who asserts that ibdāl does not result from an intentional substitution by the Arabs of one letter by the other, but from dialectal differences which affect one letter of a certain word, resulting in two words that are pronounced differ- ently but have the same meaning.914 Yet even sameness of meaning is at times disregarded as a necessary condition for ibdāl material. For example, the two variants of one of the frequently cited doublets, ǧaṯā and ǧaḏā,915 seem to dif- fer in meaning, for whereas the former indicates sitting upon one’s knees, the latter refers to standing upon the extremities of one’s toes, as both Ibn al-Aʿrābī (d. 231/845) and Ṯaʿlab (d. 291/904) explain.916

Unlike most other genres of the mubawwab lexica, the writing of indepen- dent works on ibdāl seems to have been discontinued following the fourth/ tenth century. It did, however, persist in some multithematic works in continu- ation of a tradition which we know from Abū ʿUbayd’s (d. 224/838) al-Ġarīb al-muṣannaf,917 Ibn Qutayba’s (d. 276/889) Adab al-kātib,918 and Kurāʿ al-Naml’s (d. 310/922) al-Muntaḫab min ġarīb kalām al-ʿArab.919 For example, a large por- tion of ibdāl material is cited by Abū ʿAlī al-Qālī (d. 356/967) in al-Amālī 920 and Ibn Sīda (d. 458/1066) in al-Muḫaṣṣaṣ.921 Mention has already been made ear- lier in this section of Ibn Ǧinnī’s (d. 392/1002) Sirr ṣināʿat al-iʿrāb,922 where the

911 Abū l-Ṭayyib, Ibdāl I, 147; II, 337. 912 Ibid., I, 174. 913 Ǧawharī, Ṣaḥāḥ (ǦNṮ ); cf. Abū l-Ṭayyib, Ibdāl I, 174. 914 Abū l-Ṭayyib’s text is part of the lost introduction to his K. al-Ibdāl, and is preserved in

Suyūṭī, Muzhir I, 460. 915 Ibn al-Sikkīt, Ibdāl 108; Zaǧǧāǧī, Ibdāl 47; Abū l-Ṭayyib, Ibdāl I, 160; cf. Abū ʿUbayd, Ġarīb

III, 655. 916 Ibn Manẓūr, Lisān (ǦḎW ); cf. below, 352 and Ḫalīl, ʿAyn VI, 171. 917 Abū ʿUbayd, Ġarīb III, 654–656, where eighteen examples are cited, in addition to five

which involve geminate roots (e.g. qaṣṣaytu and qaṣaṣtu “I cut”). 918 Ibn Qutayba, Adab 374–383, 459–461. 919 Kurāʿ, Muntaḫab II, 657–661. 920 Qālī, Amālī, II, 22–23, 34–35, 41–44, 52–54, etc. 921 Ibn Sīda, Muḫaṣṣaṣ XIII, 267–290; XIV, 11–26. 922 See above, 211–212.

 

 

228 chapter 2

grammatical and lexical types of ibdāl are uniformly discussed in each letter. Ibdāl material is also frequently cited in muǧannas lexica and is mainly drawn from authors of independent books on ibdāl.923 Those which have survived are the following:

1. al-Ibdāl (or al-Qalb wa-l-ibdāl) by Ibn al-Sikkīt (d. 244/858). 2. A reconstruction of al-Iʿtiqāb by Abū Turāb Isḥāq b. Faraǧ (d. late third/

ninth century). 3. al-Ibdāl wa-l-muʿāqaba wa-l-naẓāʾir by Zaǧǧāǧī (d. 337/949). 4. al-Ibdāl by Abū l-Ṭayyib al-Luġawī (d. 351/962).

Ibn al-Sikkīt’s book is referred to in the sources as al-Qalb wa-l-ibdāl or simply al-Ibdāl. The text published by A. Haffner924 under the former title is in fact a redaction of Ibn al-Sikkītʾs book by Abū Yaʿqūb Yūsuf b. Yaʿqūb al-Naǧīramī (d. 423/1031), as noted by Ḥ. Šaraf, who published Ibn al-Sikkīt’s book under the title al-Ibdāl.925 Šaraf ’s assumption that the published text is only the latter part of Ibn al-Sikkīt’s book (hence the title K. al-Ibdāl, rather than K. al-Qalb wa-l-ibdāl) is difficult to ascertain. If the term qalb in Ibn al-Sikkīt’s title, as Hämeen-Anttila notes,926 simply means “changing”, then it might have been loosely used as a synonym for ibdāl. Yet the term also means “metathesis” and Šaraf seems to have assumed that the first part of Ibn al-Sikkīt’s work was devoted to this notion. This assumption cannot be dismissed, since the term qalb in the sense of metathesis was used before Ibn al-Sikkīt (e.g. in Abū ʿUbayd’s al-Ġarīb).927 The book is divided into thirty-six chapters, each of which (except for the last three) is devoted to a pair of consonants which are subject to ibdāl, e.g. nūn and lām, bāʾ and mīm, mīm and nūn, etc. The other three chapters deal with ibdāl involving several other less frequently occur- ring pairs, and with the suffixation of mīm or nūn to some nouns (e.g. zurqum

923 Cf. Hämeen-Anttila (1993: 44–48) for a list of quotations from Ibn al-Sikkīt in Azharī’s Tahḏīb, Ǧawharī’s Ṣaḥāḥ, and Ibn Manẓūr’s Lisān.

924 Published in al-Kanz al-luġawī fī l-lasan al-ʿArabī (pp. 3–65), al-Maṭbaʿa al-Kāṯūlīkiyya, Beirut 1903.

925 See Šaraf ’s introduction to Ibn al-Sikkīt, Ibdāl 9–10. 926 Hämeen-Anttila (1993: 34–35). 927 Cf. bāb al-maqlūb in Abū ʿUbayd, Ġarīb III, 647–654. Examples of chapters on qalb in

later sources include the ones in Kurāʿ al-Naml, Muntaḫab II, 594–598, Ibn Sīda, Muḫaṣṣaṣ XIV, 27–28, and Suyūṭī, Muzhir I, 476–481. Note also that Suyūṭī (Muzhir I, 476) specifies metathesis as the subject of Ibn al-Sikkīt’s book.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

229MUBAWWAB AND SPECIALIZED LEXICA

“intensely blue-eyed” and ḍayfan “one who comes with a guest”, obviously related to azraq “blue” and ḍayf “guest” respectively).928

Abū Turāb’s K. al-Iʿtiqāb was probably larger than Ibn al-Sikkīt’s book.929 Its reconstruction contains 357 fragments, mostly taken from Azharī’s (d. 370/981) Tahḏīb al-luġa, itself one of the sources of Ibn Manẓūr’s (d. 711/1311) Lisān al-ʿArab. Ibdāl pairs, as Azharī’s quotations indicate, are listed in separate chapters,930 but there is no clue as to their arrangement. In addition to dou- blets, Abū Turāb mentions some triplets which contain ibdāl (e.g. baḫaza, baḫasa and baḫaṣa “to put out someone’s eye”).931 Also noteworthy is the fact that he collected some of his data directly from the Bedouins who were the paraphernalia of the Ṭāhirid court in Nishapur.932

Zaǧǧāǧī’s first statement in his very brief introduction to his book indicates that the three terms which its title includes are synonyms.933 Muʿāqaba, how- ever, has a narrower sense in some sources, since it is restricted to the inter- changeability between wāw and yāʾ. Ibn Sīda (d. 458/1066), for example, uses the term exclusively with these two letters.934 Zaǧǧāǧī, unlike Ibn al-Sikkīt, does include a chapter on wāw and yā,935 but it is not clear whether he reserves the term muʿāqaba for this type of substitution. The book’s thirty-four chapters are not alphabetically arranged, but there is an attempt at grouping together some chapters on phonetic grounds, as in four consecutive chapters on tāʾ and other dentals (tāʾ, dāl and ṭāʾ; tāʾ and dāl; dāl and ṭāʾ; tāʾ and ṭāʾ).936 The total number of items amounts to 290 (including some examples of grammatical ibdāl), only sixty-seven of which are common with Ibn al-Sikkīt’s items.937

Two significant developments in the genre took place in Abū l-Ṭayyib’s K. al-Ibdāl. The first is the arrangement of the chapters in strict alphabetical order, hence bāʾ and ḏāl, bāʾ and rāʾ, bāʾ and zāy, etc. As far as the internal arrangement of material is concerned, Abū l-Ṭayyib normally cites in one block examples taken from Ibn al-Sikkīt, followed by other examples mostly cited in

928 Ibn al-Sikkīt, Ibdāl 147, 149. 929 Note that Qifṭī (Inbāh IV, 103) describes Abū Turāb’s book as kabīr (large). 930 Cf. Hämeen-Anttila (1993: 62). 931 Item 104, p. 15; cf. Azharī, Tahḏīb (BḪZ; VII, 213); Ibn Manẓūr, Lisān (BḪZ). 932 Cf. Hämeen-Anttila (1993: 15). 933 Zaǧǧāǧī, Ibdāl 1: yuqāl li-hāḏihi l-ḥurūf al-ibdāl wa-l-muʿāqaba wa-l-naẓāʾir. 934 Ibn Sīda, Muḫaṣṣaṣ XIV, 19–27. 935 Zaǧǧāǧī, Ibdāl 20–28. 936 Ibid., 40, 42, 43, 44. 937 Hämeen-Anttila (1993: 52).

 

 

230 chapter 2

anagrammatic order.938 The second development is the clear attempt on the part of Abū l-Ṭayyib at producing a most comprehensive list of ibdāl mate- rial in each chapter. This is not restricted to cases in which a certain letter is substituted by another on phonological, or even orthographical grounds, but applies also to its alleged substitution by totally unrelated letters. Among the letters that are said to substitute sīn, for example, are ṭāʾ, ġayn, qāf, kāf, lām and hāʾ, resulting in most improbable pairs such as fasa⁠ʾa and faṭa⁠ʾa (to beat with a stick), israndā and iġrandā (to overcome someone and beat him), ma⁠ʾlūs and ma⁠ʾlūq (weak-minded), summ and kumm (concern), musaġsaġ and mulaġlaġ (seasoned with grease), and salata and halata (to scrape off).939 As a result of this approach, for each letter of the alphabet are listed, under separate head- ings, several letters that have an ibdāl relationship with it. The only exception is ẓāʾ, but it does appear in doublets cited under other letters.940 Accordingly, Abū l-Ṭayyib’s book may be described as the ultimate lexicon on ibdāl in the tradition.

The alliterative phenomena of itbāʿ also received the attention of philolo- gists, although they differed significantly in defining it. In commenting on con- structions such as ḥārr yārr (very hot) and ʿaṭšān naṭšān (very thirsty), Abū ʿUbayd (d. 224/838) says that the second word is never used alone, but corrobo- rates the previous one (tābiʿa li-l-ūlā ʿ alā waǧh al-tawkīd lahā).941 In contrast, he says that bayyāka in the well-known construction ḥayyāka l-Lāhu wa-bayyāka (May God prolong your life) is preceded by the conjunction wāw and hence is not an instance of itbāʿ, which hardly ever occurs with wāw (lā yakād yakūn bi-l-wāw). This view is not shared by Abū l-Ṭayyib al-Luġawī (d. 351/962), who argues that the presence or absence of the wāw is not to be taken into account (al-iʿtibār laysa bi-l-wāw) in determining what itbāʿ is. He cites the construc- tions ǧāʾiʿ nāʾiʿ (very hungry) and ǧūʿan wa-nūʿan (May you be inflicted with extreme hunger), both of which he considers to be examples of itbāʿ although the latter contains the conjunction.942 Scholars also often differ as to whether in itbāʿ constructions, the second word (or tābiʿ) can have a meaning of its own or not943 – in other words, whether it exists independently of the first word (or matbūʿ) and can thus claim independent existence in the Arabic lexicon.

938 See a detailed example taken from the chapter on lām and nūn (Abū l-Ṭayyib, Ibdāl II, 382–414) in Hämeen-Anttila (1993: 106–119).

939 Abū l-Ṭayyib, Ibdāl II, 198, 200, 204, 205, 208, 214. 940 E.g. ǧīm and ẓāʾ (I, 235), ḏāl and ẓāʾ (II, 20–21), zāy and ẓāʾ (II, 140), sīn and ẓāʾ (II, 198), etc. 941 Abū ʿUbayd, Ḥadīṯ II, 279; text also in Suyūṭī, Muzhir I, 414–415. 942 Abū l-Ṭayyib, Itbāʿ 3. 943 Cf. Suyūṭī, Muzhir I, 415–416.

 

amgad
Highlight

 

231MUBAWWAB AND SPECIALIZED LEXICA

An example is innahu la-ḏū ǧūd wa-sūd (he is very generous), where sūd is said by some to have no meaning of its own since it is the result of a deliberate change of the first radical of the previous word, whereas others interpret it as a variant of suʾdad (eminence).944 According to the first interpretation, the construction is an example of itbāʿ, whereas in the second interpretation, sūd has a meaning of its own and cannot be the result of itbāʿ. Constructions in which the second word has an independent meaning are normally referred to as tawkīd (corroboration) or muzāwaǧa (coupling).

The earliest list of itbāʿ constructions is found in Abū ʿUbayd’s (d. 224/838) al-Ġarīb al-muṣannaf, which includes about fifty doublets (e.g. ḥasan basan “very good”).945 Other lists in multithematic works or muǧannas lexica are found in Ṯaʿlab’s (d. 291/904) al-Maǧālis,946 Kurāʿ al-Naml’s (d. 310/922) al-Muntaḫab,947 Hamaḏānī’s (d. 320/933) al-Alfāẓ al-kitābiyya,948 Ibn Durayd’s (d. 321/933) Ǧamharat al-luġa,949 Qālī’s (d. 356/967) al-Amālī,950 and Ibn Sīda’s (d. 458/1066) al-Muḫaṣṣaṣ.951 There are also two independent lexica on itbāʿ, namely, al-Itbāʿ by Abū l-Ṭayyib al-Luġawī (d. 351/962) and al-Itbāʿ wa-l- muzāwaǧa by Ibn Fāris (d. 395/1004). The former contains 220 itbāʿ construc- tions arranged in chapters from alif to yāʾ, based only on the first letter of each word in which itbāʿ occurs (i.e. second word in doublets and third in triplets). Each chapter is further divided into two sections, one for itbāʿ, followed by another for tawkīd.952 Under itbāʿ are listed constructions whose second and third words do not exist independently of the first, whereas the sections on tawkīd include those constructions whose second or third items may be used independently since they have a meaning of their own. As for Ibn Fāris, his book embraces 334 itbāʿ constructions arranged, like Abū l-Ṭayyib’s, in chap- ters from alif to yāʾ. But, unlike his predecessor, Ibn Fāris arranges words based

944 Abū l-Ṭayyib, Itbāʿ 51. 945 Abū ʿUbayd, Ġarīb III, 657–658. 946 Ṯaʿlab, Maǧālis I, 202, 205–206. 947 Kurāʿ, Muntaḫab II, 598–600. 948 Hamaḏānī, Alfāẓ 295–296. 949 Ibn Durayd, Ǧamhara III, 1253–1254. 950 Qālī, Amālī II, 208–218. 951 Ibn Sīda, Muḫaṣṣaṣ XIV, 28–38. 952 Note that some scholars do not reserve the term tawkīd for constructions whose second

and third words can be independently used. Abū Hilāl al-ʿAskarī (d. after 395/1005), for example, believes that in constructions like ʿaṭšān naṭšān, the second word (which does not exist independently of the first) is tawkīd since it serves the same purpose as the second word in the supposed original construction *ʿaṭšān ʿaṭšān, which the Arabs supposedly changed by altering the first letter of its second word (Ṣināʿatayn 194).

 

amgad
Highlight
amgad
Highlight
amgad
Highlight

 

232 chapter 2

on their last letter (without taking other letters into consideration). Although one might wonder why Ibn Fāris – who in his two muǧannas lexica, Maqāyīs al-luġa and Muǧmal al-luġa, alphabetically arranges entries strating with the first letter – should opt for such an arrangement in his itbāʿ lexicon, there is a very good reason for his choice, namely, that one of the conditions for itbāʿ is that both the tābiʿ and the matbūʿ should end with the same letter. Accordingly, this type of arrangement groups together, in each chapter, words which share the same final letter since it is not normally subject to change (but cf. ǧūʿan lahu wa-ǧūdan wa-ǧūsan, where the last two words, as Ibn Fāris notes, do not rhyme with the first).953 In contrast, Abū l-Ṭayyib’s arrangement revolves around the substitute letter, normally in the first radical of the tābiʿ. In connection with constructions whose tābiʿ has an independent meaning, Ibn Fāris uses the terms muzāwaǧ and tazwīg.954

10 al-Abniya (Morphological Patterns)

Most works which deal with abniya (sg. bināʾ) and related morphological issues owe their notions and much of their subject matter to the grammati- cal tradition. The most obvious example is the part of Sībawayhi’s (d. 180/796) al-Kitāb which deals with morphology, since many of the issues examined in it were further discussed and expanded in independent works authored by subsequent philologists. Other than Sībawayhi’s discussion of issues related, for example, to muḏakkar and muʾannaṯ and to maqṣūr and mamdūd (see nos. 2 and 3 below), the Kitāb embraces the vast majority of Arabic nomi- nal and verbal patterns. The total number of these patterns is 342 (308 for nouns and 34 for verbs)955 and, as Zubaydī (d. 379/989) notes, grammarians believe that, except for three patterns which he overlooked (mā ḫalā ṯalāṯat abniya šaḏḏat ʿan ǧamīʿihi), Sībawayhi’s list exhausts all Arabic patterns.956 This notwithstanding, it has already been pointed out that Sībawayhi did not normally provide any explanation of the meanings of words which he

953 Ibn Fāris, Itbāʿ 97. Ibn Fāris’s interest in rhyme is also clear in his inclusion of rhyming expressions (saǧʿ) which fall outside the sphere of itbāʿ (ibid., 59, 72, 74).

954 Ibid., 46, 48, 50, etc. 955 See a detailed count of their types in ʿ Umar (1995: 12–13, 69–75); cf. Ibn al-Qaṭṭāʿ, Abniya 89. 956 Zubaydī, Istidrāk 1. Note that Zubaydī mentions in his introduction that he found out

the number of abniya ignored by Sībawayhi to be about eighty. Cf. ʿUmar (1995: 70), who specifies this number as eighty-eight. See also a list of the abniya supposedly ignored by Sībawayhi in Sīrāfī, Fawāʾit 67–99.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

233MUBAWWAB AND SPECIALIZED LEXICA

cited – particularly words which illustrate certain rare patterns – and that some later authors wrote independent monographs to explain these words.957 The lack of interest in this aspect of semantics persisted in other major books on grammar, such as Mubarrad’s (d. 285/898) al-Muqtaḍab and, to a lesser extent, Ibn al-Sarrāǧʾ’s (d. 316/929) al-Uṣūl fī l-naḥw,958 but later authors, such as Ibn Yaʿīš (d. 643/1245) and Suyūṭī (d. 911/1505), more regularly gave the meanings of ġarīb words cited in their study of abniya.959 Most of the lexica which deal with abniya and which will be discussed below complement the early works of naḥw in this respect since they normally explain cited words, particularly ġarīb or less frequently used ones. This is yet another instance of the complementar- ity of the two traditions of naḥw and luġa (including the lexicographical tradi- tion) as defined in the previous chapter.960

Books which deal with morphology and morphophonology in general nor- mally examine a number of issues related to some of the genres of writing dis- cussed below (e.g. 10.1–3, 5–7). The earliest extant book of this type is Māzinī’s (d. 249/863) K. al-Taṣrīf, famous for the commentary Ibn Ǧinnī (d. 392/1002) wrote on it under the title al-Munṣif. Among the most important sources after al-Taṣrīf are Muʾaddib’s (d. c. 338/949) Daqāʾiq al-taṣrīf, Maydānī’s (d. 518/1124) Nuzhat al-ṭarf fī ʿilm al-ṣarf, Ibn ʿUṣfūr’s (d. 669/1271) al-Mumtiʿ fī l-taṣrīf and its abbreviation al-Mubdiʿ fī l-taṣrīf by Abū Ḥayyān al-Naḥwī al-Andalusī (d. 745/1344), and Astarābāḏī’s (d. c. 686/1287) Šarḥ Šāfiyat Ibn al-Ḥāǧib. Of particular significance for our study is the fact that these works contain a large body of material that can be described as ġarīb (strange) or šāḏḏ (anomalous).961 This matches a similar phenomenon in works on naḥw, which generally devote a great deal of attention to this type of usage in syntax. Accordingly, interest in strange and uncommon/rare usage in both mubawwab and muǧannas lexica should be placed within this wider context, particularly as much of the data used by both grammarians and lexicographers owes its origin to the same pro- cess of data collection which took place in the second/eighth and third/ninth centuries.962

957 See above, 60, 216. 958 See, for example, Mubarrad, Muqtaḍab I, 66–68; II, 107–109; Ibn al-Sarrāǧ, Uṣūl III, 184–210. 959 Ibn Yaʿīš, Šarḥ VI, 136–143; Suyūṭī, Hamʿ II, 158–160. 960 See above, 3. 961 For example, see in Ibn Ǧinnī’s Munṣif words which illustrate quadriliterals and

quinqueliterals (I, 24–34), ilḥāq (I, 34–53), and other irregular verbal and nominal patterns (I, 275–279). For statements which express the notion of laysa fī kalām al-ʿArab . . . illā (cf. above, 43–44, 93), see ibid., I, 48 and Ibn ʿUṣfūr, Mumtiʿ I, 118, 133, 135, 300.

962 See above, 16–29.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

234 chapter 2

Multithematic works also contain material related to the genres of writing which follow. The oldest extant work of this type, Abū ʿUbayd’s (d. 224/838) al-Ġarīb al-muṣannaf, for example, deals with the morphological patterns (abniya) of nouns (asmāʾ), verbs (afʿāl) and verbal nouns (maṣādir),963 as well as with issues related to muḏakkar and muʾannaṯ.964 Ibn al-Sikkīt’s (d. 244/858) Iṣlāḥ al-manṭiq and Kurāʿ al-Naml’s (d. 310/922) al-Muntaḫab min ġarīb kalām al-ʿArab also contain material related to various nominal and verbal patterns.965 Ibn Sīda (d. 458/1066) in al-Muḫaṣṣaṣ devotes extremely lengthy chapters to various types of abniya, maqṣūr and mamdūd, and muḏakkar and muʾannaṯ.966 The content of his chapters exceeds that of several independent works of these three genres. Also noteworthy is the lengthy chapter designated for nominal patterns in Ibn Durayd’s (d. 321/933) muǧannas lexicon, Ǧamharat al-luġa.967

Due to the very large number of works under the genre of abniya (in the broad sense discussed above), we shall divide them into the following eight types:

10.1 al-Ištiqāq (Derivation) Philologists distinguish between several types of ištiqāq although their termi- nology does not always faithfully reflect this distinction. The most general type (i.e. derivation of a word from another which shares its root) is normally dis- cussed in books on ṣarf/taṣrīf. It is simply referred to as ištiqāq, but sometimes also as al-ištiqāq al-ṣaġīr/al-aṣġar in order to distinguish it from al-ištiqāq al-kabīr and al-ištiqāq al-akbar.968 al-Ištiqāq al-kabīr usually refers to metathe- sis, which involves a change in the order of succession of the root’s radicals but not in its meaning (e.g. ǧaḏaba and ǧabaḏa “to pull” and bakala and labaka “to mix”).969 al-Ištiqāq al-akbar was coined by Ibn Ǧinnī to refer to the semantic relationship which, according to him, is retained by the various combinations (tarākīb) of a triliteral root, irrespective of the order of succession of its radi- cals (e.g. ǦBR, ǦRB, BǦR, BRǦ, RǦB and RBǦ, all of which, he argues, share the sense of quwwa and šidda “strength and sternness”).970

963 Abū ʿUbayd, Ġarīb II, 513–566, 567–621; III, 684–689 respectively. 964 Ibid., III, 659–661. 965 Ibn al-Sikkīt, Iṣlāḥ 3–223; Kurāʿ, Muntaḫab II, 509–583. 966 Ibn Sīda, Muḫaṣṣaṣ XIV, 122-XV, 94; XV, 95-XVI, 79; XVI, 79-XVII, 128 respectively. 967 Ibn Durayd, Ǧamhara II, 1162-III, 1253. 968 For inconsistencies in using these terms, see Tarazī (1968: 17). 969 Ibn Fāris, Ṣāḥibī 202; cf. Suyūṭī, Muzhir I, 476. 970 Ibn Ǧinnī, Ḫaṣāʾiṣ I, 5–17; II, 133–139.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

235MUBAWWAB AND SPECIALIZED LEXICA

Several books on ištiqāq are ascribed to philologists as of the third/ninth century.971 Judging by the few extant sources and by the titles of others, it seems that most of these works do not deal with issues related to ištiqāq and discussed in books of morphology and morphophonology, such as Māzinī’s al-Taṣrīf or Ibn ʿUṣfūr’s al-Mumtiʿ, but with the derivation of proper nouns. The earliest such work, Quṭrub’s (d. 206/821) K. al-Ištiqāq, is lost but we do know that it dealt with the derivation of proper nouns since Zubaydī (d. 379/989) mentions that Abū l-Walīd al-Mahrī (d. 253/867) wrote a book on ištiqāq al-asmāʾ in which he included material not cited by Quṭrub.972 To Abū Naṣr al-Bāhilī (d. 231/845) is also ascribed a book titled Ištiqāq al-asmāʾ.973 One can- not establish with any degree of certainty, however, whether other lost books titled al-Ištiqāq – such as those ascribed to al-Aḫfaš al-Awsaṭ (d. 215/830), Mubarrad (d. 285/898), al-Mufaḍḍal b. Salama (d. 290/903 or 300/913), and Zaǧǧāǧ (d. 311/923) – dealt exclusively with proper nouns or were similar in content to Ibn al-Sarrāǧ’s (d. 316/929) Risālat al-ištiqāq. In this short tractate, Ibn al-Sarrāǧ discusses various issues related to ištiqāq, including its defini- tion; whether it applies to all words which share the order of succession of their radicals; methods of establishing which of two (or more) words is the aṣl (origin) of ištiqāq and which is the farʿ (branch, i.e. derived word); the pur- pose of ištiqāq and why it is needed; and guarding against pitfalls in identify- ing derived words.974 Also included are certain phonotactic principles which characterize Arabic words,975 most probably in order to draw clear boundaries between them and foreign words, since the latter, as Ibn al-Sarrāǧ explains in another chapter, should not be interpreted as derivatives of Arabic roots.976

On the ištiqāq of proper nouns, we possess one of the earliest works, Aṣmaʿī’s (d. 216/831) K. Ištiqāq al-asmāʾ, also known by the shorter title of K. al-Ištiqāq. The proper nouns it includes are primarily those of individuals and tribes, in addition to a few place names. These amount to 133 names which are neither alphabetically nor thematically arranged. Aṣmaʿī typically gives the etymology and meaning of the noun under discussion and cites other words which are derived from the same root, often including poetry šawāhid. It is also worth noting that proper nouns and place names constitute most of Kurāʿ al-Naml’s (d. 310/922) examples in his chapter on ištiqāq in al-Muntaḫab min ġarīb kalām

971 See a full list in the introduction of the editors of Aṣmaʿī’s Ištiqāq 46–52. 972 Zubaydī, Ṭabaqāt 230. 973 Ibn al-Nadīm, Fihrist 60; Yāqūt, Muʿǧam I, 227. 974 Ibn al-Sarrāǧ, Ištiqāq 19–33. 975 Ibid., 34–38. 976 Ibid., 31; cf. Suyūṭī, Muzhir I, 287.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

236 chapter 2

al-ʿArab.977 The development in this genre of writing can be demonstrated by comparing Aṣmaʿī’s book with another relatively early extant work, namely, Ibn Durayd’s (d. 321/933) K. al-Ištiqāq, which is also referred to as K. Ištiqāq al-asmāʾ 978 and K. Ištiqāq asmāʾ al-qabāʾil.979 Like Aṣmaʿī, Ibn Durayd explains the meanings of the proper nouns which he cites and frequently supports his arguments by poetry šawāhid. Yet, his material is much more comprehensive than Aṣmaʿī’s and is arranged based on the views of the Arab genealogists. He begins with the etymology of the Prophet’s name, Muḥammad, followed by the names of his ancestry and the various Arab tribes. Unlike Aṣmaʿī, Ibn Durayd mentions some of the most prominent individuals of each tribe and often cites anecdotes which aim to explain the origin of certain names. His interest in the etymology of proper nouns – also evident in his muǧannas lexicon, Ǧamharat al-luġa, several hundreds of whose entries include proper nouns980 – obviously stems from his attempt to refute the claims of the Šuʿūbiyyūn, whom he quotes as alleging that the Arabs often use proper nouns which have no origin (aṣl) in their language or whose meanings are unknown to them.981 This link between philological study and the refutation of the Šuʿūbiyyūn is not unique in the tra- dition, and its role in the study of aḍdād has been pointed out earlier.982

Related to the ištiqāq of proper nouns are works which deal with the asmāʾ (names, proper nouns) and ṣifāt (attributes) of God. The earliest work of this type mentioned in the sources, Abū Gaʿfar al-Naḥḥās’s (d. 338/950) K. al-Ištiqāq li-asmāʾ Allāh,983 is lost but we do possess K. Ištiqāq asmāʾ Allāh by Abū Ǧaʿfar’s contemporary, the famous grammarian, Abū l-Qāsim al-Zaǧǧāǧī (d. 337/949). The main part of the book consists of the ninety-nine names and attributes listed in the order of their appearance in the Qurʾānic text. Zaǧǧāǧī typically gives the etymology and meaning of each name or attribute and discusses at length other derivatives from the same root as well as matters related to the interpretation and qirāʾāt of some verses. At times he also mentions differ- ences among philologists in determining the etymology of a certain word, most notably in explaining the word Allāh.984 The latter part of the book includes a discussion of some issues of a more general nature, such as the difference

977 Kurāʿ, Muntaḫab II, 661–678. 978 Azharī, Tahḏīb I, 31. 979 Yāqūt, Muʿǧam VI, 2495. 980 See the indices of Ibn Durayd’s Ǧamhara III, 1699–1722; cf. below, 345. 981 Ibn Durayd, Ištiqāq 4. 982 See above, 190. 983 Yāqūt, Muʿǧam I, 469. 984 Zaǧǧāǧī, Ištiqāq 225–296.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

237MUBAWWAB AND SPECIALIZED LEXICA

between ism and ṣifa vis-à-vis ištiqāq and a refutation of the views of some philologists who deny the very existence of ištiqāq.985 Mention can also be made of a lengthy chapter titled Ištiqāq asmāʾ Allāh in Ibn Sīda’s (d. 458/1066) al-Muḫaṣṣaṣ.986

Also related to ištiqāq are works that deal with diminutives or with the sin- gular, dual and plural. On diminutives, a book titled K. al-Taṣġīr is ascribed to each of the Kufan scholars Ruʾāsī (d. 187/803) and Ṯaʿlab (d. 291/904), but neither has survived.987 As for singular, dual and plural, the following titles are mentioned by Ibn al-Nadīm (d. 380/990) and Yāqūt (d. 626/1229): Farrāʾ’s (d. 207/822) al-Ǧamʿ wa-l-taṯniya fī l-Qurʾān,988 Abū ʿUbayda’s (d. 209/824) al-Ǧamʿ wa-l-taṯniya,989 Abū Zayd al-Anṣārī’s (d. 215/830) al-Wāḥid and al-Ǧamʿ wa-l-taṯniya,990 Ǧarmī’s (d. 225/840) al-Taṯniya wa-l-ǧamʿ,991 and al-Aḫfaš al-Ṣaġīr’s (d. 315/927) al-Taṯniya wa-l-ǧamʿ.992 Since none of these titles has survived, it is not possible to determine their content and whether they belong to the realm of luġa or naḥw. Yet, we are able to form an idea about the content of some books that deal with the dual, and to a lesser extent, the plural, since there are rather lengthy quotations in Ibn Sīda’s (d. 458/1066) al-Muḫaṣṣaṣ and Suyūṭī’s (d. 911/1505) al-Muzhir from Ibn al-Sikkīt’s (d. 244/858) al-Muṯannā wa-l-mukannā wa-l-mubannā wa-l-muʾāḫā wa-l-mušbah wa-l-munḥal, and since we do possess an extant work on the dual, namely, Abū l-Ṭayyib al-Luġawī’s (d. 351/962) al-Muṯannā. One subject which Ibn al-Sikkīt is reported to have discussed with regard to plural nouns has to do with the notion of prevalence (ġalaba, taġlīb), as exemplified by al-Maʿāwil and al-Qutaybāt, two clans who were named after their ancestors, Miʿwala b. Šams and Qutayba respective- ly.993 Other quotations from the book show that the notion of prevalence in dual nouns was also a major theme pursued by Ibn al-Sikkīt.994 Much of Abū l-Ṭayyib’s book also deals with this theme, particularly in four of its ten chap- ters devoted to nouns.995 Examples common to both works include al-ʿUmarān

985 Ibid., 268–274, 277–292. 986 Ibn Sīda, Muḫaṣṣaṣ XVII, 134–160. 987 Ibn al-Nadīm, Fihrist 71, 81; Yāqūt, Muʿǧam II, 553; VI, 2488, 2572; Qifṭī, Inbāh I, 136; IV, 107. 988 Ibn al-Nadīm, Fihrist 73; Yāqūt, Muʿǧam VI, 2815. 989 Ibn al-Nadīm, Fihrist 60; Yāqūt, Muʿǧam VI, 2708. 990 Ibn al-Nadīm, Fihrist 60; Yāqūt, Muʿǧam III, 1361. 991 Ibn al-Nadīm, Fihrist 62; Yāqūt, Muʿǧam III, 1035; IV, 1444. 992 Ibn al-Nadīm, Fihrist 91; Yāqūt, Muʿǧam IV, 1771. 993 Suyūṭī, Muzhir II, 204. 994 Ibn Sīda, Muḫaṣṣaṣ XIII, 223–226; Suyūṭī, Muzhir II, 173–182, 185–189. 995 Abū l-Ṭayyib, Muṯannā 4, 27, 35, 53.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

238 chapter 2

(for Abū Bakr and ʿUmar) and al-aswadān (for dates and water),996 where the dual of one of a pair of items denotes both. Dual nouns and proper nouns of this type are also the subject of an extant short monograph by one of Ibn al-Sikkīt’s contemporaries, Ibn Ḥabīb (d. 245/860), and is titled K. Mā ǧāʾa smāni aḥaduhumā ašhar min ṣāḥibihi fa-summiyā bihi. Furthermore, Ibn al-Sikkīt, as Suyūṭī reports, allocates a chapter for plural nouns that occur where the dual or singular are expected, as in alqāhu fī lahawāt al-layṯ (He threw him into the lion’s uvulas).997 Similarly, one of Abū l-Ṭayyib’s chapters deals with dual nouns which occur instead of the singular, such as māta ḥatfa anfayhi (He died on his bed; i.e. he was not slain), where anfayhi may be used instead of the singular anfihi.998 Other material common to the two books includes dual nouns which differ in their ištiqāq from each of their two components, such as al-malawān for al-layl (night) and al-nahār,999 and dual proper nouns whose singulars are identical, such as al-ʿĀmirān, by which are meant both ʿĀmir b. Mālik b. Ǧaʿfar and ʿĀmir b. al-Ṭufayl b. Mālik b. Ǧaʿfar.1000 Based on the above similarities, it is probable that the two books are generally representative of the largely lost tradition of writing specialized monographs on muṯannā.

Finally on ištiqāq, the notion of naḥt (blending, i.e. coining a new word from two or more independent words) is worthy of mention. Although it is discussed in some general works on philology,1001 there does not seem to have been a tra- dition of writing independent monographs to deal with it, perhaps due to the relative rarity of manḥūt words in Arabic. One exception is Tanbīh al-bāriʿīn ʿ alā l-manḥūt min kalām al-ʿArab, ascribed by Yāqūt to Abū ʿAlī al-Ḥasan b. al-Ḫaṭīr al-Fārisī, known as al-Ẓahīr (d. 598/1201). Yāqūt explains that Abū ʿAlī authored this book in response to a query by Abū l-Fatḥ ʿUṯmān b. ʿIsā l-Naḥwī al-Balaṭī (or al-Bulayṭī or al-Malaṭī; d. 599/1202) about the word šaqaḥṭab (a ram with two hideous horns or four horns), said to be a blend of šiqq (half of, division of) and ḥaṭab (firewood).1002 In muǧannas lexica, Ibn Fāris’s (d. 395/1004) Maqāyīs al-luġa includes more references to naḥt than any other lexicon.1003

996 Ibid., 4, 27; Ibn Sīda, Muḫaṣṣaṣ XIII, 223; Suyūṭī, Muzhir II, 186, 173. 997 Suyūṭī, Muzhir II, 191. A similar expression is cited in Ibn al-Sikkīt, Ḥurūf 53 (cf. above, 174)

and Kurāʿ, Muntaḫab II, 641. 998 Abū l-Ṭayyib, Muṯannā 63. 999 Ibn Sīda, Muḫaṣṣaṣ XIII, 223; Suyūṭī, Muzhir II, 173; Abū l-Ṭayyib, Muṯannā 56. 1000 Ibn Sīda, Muḫaṣṣaṣ XIII, 229; Suyūṭī, Muzhir II, 187; see also Abū l-Ṭayyib, Muṯannā 17 (for

ʿĀmir b. Ṣaʿṣaʿa and ʿĀmir b. Rabīʿa b. ʿĀmir b. Ṣaʿṣaʿa). 1001 E.g. Ibn Fāris, Ṣāḥibī 271; Suyūṭī, Muzhir I, 482–485. 1002 Yāqūt, Muʿǧam II, 858; cf. Suyūṭī, Muzhir I, 482–483 and Zabīdī, Tāǧ (ŠQḤṬB). 1003 See below, 353–354.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

239MUBAWWAB AND SPECIALIZED LEXICA

10.2 al-Muḏakkar wa-l-muʾannaṯ (Masculine and Feminine) Issues related to masculine and feminine were discussed in more than thirty independent monographs, most of which are lost.1004 Following is a list of the most important extant works of the genre, followed by a few observations on their content and arrangement:

1. al-Muḏakkar wa-l-muʾannaṯ by Farrāʾ (d. 207/822). 2. al-Muḏakkar wa-l-muʾannaṯ by Abū Ḥātim al-Siǧistānī (d. 255/869). 3. al-Muḏakkar wa-l-muʾannaṯ by Mubarrad (d. 285/898). 4. Muḫtaṣar al-Muḏakkar wa-l-muʾannaṯ by al-Mufaḍḍal b. Salama

(d. 290/903 or 300/913). 5. Mā yuḏakkar wa-mā yuʾannaṯ min al-insān wa-l-libās by Abū Mūsā

l-Ḥāmiḍ (d. 305/918). 6. al-Muḏakkar wa-l-muʾannaṯ by Nifṭawayhi (d. 323/935). 7. al-Muḏakkar wa-l-muʾannaṯ by Abū Bakr b. al-Anbārī (d. 328/940). 8. al-Muḏakkar wa-l-muʾannaṯ by Ibn al-Tustarī (d. 361/972). 9. al-Muḏakkar wa-l-muʾannaṯ by Ibn Ǧinnī (d. 392/1002). 10. al-Muḏakkar wa-l-muʾannaṯ by Ibn Fāris (d. 395/1004). 11. al-Bulġa fī l-farq bayna l-muḏakkar wa-l-muʾannaṯ by Abū l-Barakāt b.

al-Anbārī (d. 577/1181). 12. Fī l-asmāʾ al-muʾannaṯa al-samāʿiyya by Rāzī (d. after 666/1268). 13. Tadmīṯ al-taḏkīr fī l-ta⁠ʾnīṯ wa-l-taḏkīr by Ibrāhīm b. ʿUmar al-Ǧaʿbarī

(d. 732/1332).

Since the majority of subjects related to masculine and feminine were dis- cussed by authors of the third/ninth century, we shall examine the first four books in the above list before tracing later developments in the genre. The four books share much of their material and deal with a large variety of topics, such as feminine markers (i.e. -h, -āʾ, -ā); the suffix -h in masculine nouns (e.g. ʿallāma “a very learned person”); patterns which are used with both masculine and feminine, such as faʿīl which indicates the passive participle (e.g. qatīl “mur- dered”) and faʿūl which indicates the active participle (e.g. ṣabūr “patient”); words which can be treated as both masculine and feminine (e.g. ʿunuq “neck” and lisān “tongue”); words that are known as muʾannaṯ samāʿī, that is, words which have no feminine ending but are either feminine (e.g. fa⁠ʾs “ax”), or in a few cases both masculine and feminine (e.g. ʿunuq “neck”); adjectives that are specifically used with the feminine and hence have no marker (e.g. ṭāliq “divorced”); the plurals and diminutives of feminine words; suffixation of -h

1004 See ʿAbdaltawwāb’s introduction to Mufaḍḍal, Muḫtaṣar 23–31 and Iqbāl (2011: 264–271).

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

240 chapter 2

to generic nouns for indicating the singular (e.g. ǧarād “locust” and ǧarāda “a locust”) and the use of the two forms to indicate masculine and feminine respectively (i.e. ǧarād for masculine singular and ǧarāda for female sin- gular); feminine nouns that are diptotes; and the use of the feminine with numerals and with the names of Qurʾānic sūras, tribes, winds, months, let- ters of the alphabet, etc. At times, even syntactical issues related to mascu- line and feminine are mentioned, such as the use of the masculine with the verb when it is separated from its feminine agent, as in the expression ḥaḍara l-qāḍiya mra⁠ʾatun (A woman appeared before the judge), which goes back to Sībawayhi.1005 Most of the topics listed above are discussed in grammatical works,1006 but are scattered over several chapters, which include, other than chapters that directly deal with masculine and feminine, those that deal with plurals, diptotes, numerals and specific patterns.

The content of many later works – such as those by Nifṭawayhi, Ibn Fāris and Abū l-Barakāt b. al-Anbārī – is largely similar to that of the four books from the third/ninth century. In the case of Abū Bakr b. al-Anbārī’s book, how- ever, there is a marked development in size, since it is several-fold larger than any of the previous works. The book does include a few topics that were not discussed by some or all of the four earlier authors, such as feminine nouns and adjectives of the pattern faʿāli, the vocative forms of certain masculine and feminine nouns, feminine adjectives of the patterns fuʿul, fuʿlūl, fiʿlil, faʿlal, fiʿl, fuʿl and faʿl, and the conjugation of the imperative forms hāti (give me), taʿāla/halumma (come on), etc.1007 Yet the huge size of the book is primarily due to the extensive discussion of topics only briefly treated by earlier authors and to the numerous examples and šawāhid cited by Ibn al-Anbārī. This is clearly demonstrated by the fact that chapters assigned to words which refer to human body parts and objects (ašyāʾ) and which can be either masculine, feminine, or both, extend over 170 pages.1008

Other extant works represent further developments in the genre. One such development is that some monographs were devoted entirely to a specific topic related to masculine and feminine. Thus, al-Ḥāmiḍ deals exclusively with human body parts and clothing, whereas Rāzī cites, in thirteen verses, sixty words that are muʾannaṯ samāʿī. In resorting to poetry, Rāzī seems to continue

1005 Abū Ḥātim, Muḏakkar 85; cf. Sībawayhi, Kitāb II, 83 and Mubarrad, Muqtaḍab II, 338. 1006 For Sībawayhi’s Kitāb, see ʿ Uḍayma (1975: 274–280); cf. the indices of Mubarrad, Muqtaḍab

IV, 107–110. 1007 Ibn al-Anbārī, Muḏakkar 599–606, 611–615, 684–701 (see also 721–725 for the nominal

pattern faʿalā), 726–731 respectively. 1008 Ibid., 261–430.

 

 

241MUBAWWAB AND SPECIALIZED LEXICA

an earlier tradition since Ibn Durayd’s (d. 321/933) dīwān contains fifteen verses in which are listed human body parts that are either masculine, femi- nine, or both.1009 Also in poetry form is Tadmīṯ al-taḏkīr fī l-ta⁠ʾnīṯ wa-l-taḏkīr by Ibrāhīm b. ʿ Umar al-Ǧaʿbarī (d. 732/1332). It consists of about 280 verses, most of which, however, deal with grammatical issues related to gender. Furthermore, works that are alphabetically arranged were not late to appear. Ibn al-Tustarī’s is the first extant work of this type and contains, following a brief discussion of issues related to masculine and feminine, an alphabetically arranged list of words which he describes as muškil (difficult, intricate).1010 These include not only muʾannaṯ samāʿī, but also other words which are either only masculine or both masculine and feminine. Apparently, some of these words were subject to laḥn by speakers as were other specific words, such as baʿḍ (some), miṯl (like), and the relative pronouns ayy and man,1011 hence their description as muškil. Ibn Ǧinnī’s monograph contains a similarly brief list, most of whose entries are also cited by Ibn al-Tustarī. As far as arrangement is concerned, Ibn Ǧinnī takes only the first letter into consideration, whereas Ibn al-Tustarī mostly fol- lows alphabetical order, but not in all chapters (cf. alif, bāʾ and ṯāʾ, which are not alphabetically arranged).

10.3 al-Maqṣūr wa-l-mamdūd (Abbreviated and Prolonged Patterns) This genre is closely related to that of muḏakkar and muʾannaṯ, given that both alif maqṣūra (-ā) and alif mamdūda (-āʾ), with which maqṣūr (abbrevi- ated) and mamdūd (prolonged) words respectively end, are among feminine markers. Maqṣūr and mamdūd words which end with either suffix (e.g. quṣwā “farthest” and ǧawzāʾ “Gemini”) are thus feminine. As for maqṣūr and mamdūd words which are masculine, their final -ā or -āʾ is not a suffix, but part of the root, as in fatā (young man) from the root ftw/y (but cf. the feminine word raḥā “quern”, from the root rḥw/y) and ṯuġāʾ (bleating of the goat and sheep) from the root ṯġw. Because of the affinity between issues related to muḏakkar and muʾannaṯ, and to maqṣūr and mamdūd (cf. the discussion in both genres of feminine words with respect to their treatment as diptotes and to their dual and plural forms), it comes as no surprise that a number of authors wrote mono- graphs on both subjects. Of the thirteen authors of extant works on muḏakkar and muʾannaṯ listed in the previous section, nine also wrote on maqṣūr and mamdūd, namely, Farrāʾ, Abū Ḥātim al-Siǧistānī, Mubarrad, al-Mufaḍḍal b.

1009 Ibn Durayd, Dīwān 143–144. Note that each of the three types is distributed over five verses which differ in meter from the other two types.

1010 Ibn al-Tustarī, Muḏakkar 56. 1011 Ibid., 64, 103, 61–62, 104 respectively.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight

 

242 chapter 2

Salama, Nifṭawayhi, Abū Bakr b. al-Anbārī, Ibn al-Tustarī, Ibn Ǧinnī and Abū l-Barakāt b. al-Anbārī. Other authors who are reported in the sources to have written on both subjects include Aṣmaʿī (d. 216/831), Abū ʿUbayd al-Qāsim b. Sallām (d. 224/838), Ibn al-Sikkīt (d. 244/858), Abū ʿAṣīda Aḥmad b. ʿUbayd b. Nāṣiḥ (d. 273/886), Abū Ǧaʿfar al-Ṭabarī (d. after 304/916), Zaǧǧāǧ (d. 311/923), Ibn Šuqayr (d. 317/929), Ibn Kaysān (d. 320/932), Muḥammad b. ʿ Uṯmān al-Ǧaʿd al-Šaybānī (d. after 320/932), ʿAbdallāh b. Muḥammad al-Ḫazzāz (d. 325/937), al-Waššāʾ (d. 325/937), Ibn Durustawayhi (d. 347/958), Ibn Miqsam al-ʿAṭṭār (d. 354/965), Ibn Ḫālawayhi (d. 370/980), Abū l-Ḥasan al-Šimšāṭī (d. after 377/987), Abū l-Ǧūd al-ʿAǧlānī (d. c. 400/1010), and Ibn Sīda (d. 458/1066). As with muḏakkar and muʾannaṯ, it is evident that the occurrence of laḥn was among the reasons which prompted philologists to author monographs on maqṣūr and mamdūd. The confusion between words of both types partially stems from the widespread elision in certain dialects of hamza, which has often led to the coalescence into one form of two words with different meanings. For example, the elision of the final hamza in ǧadāʾ (profit) and ruġāʾ (grumbling of the camel) has resulted in two forms which are identical with the maqṣūr words ǧadā (gift) and ruġā (pl. of riġwa “froth”) respectively.1012 Another source of confusion has to do with the way maqṣūr words and mamdūd ones (whose final hamza is elided) are written. Several authors of the genre constantly refer to this matter when they cite words that may be confused, or even designate one or more chapters in order to differentiate between words that are written with alif maqṣūra (usually referred to as yāʾ) or with an ordinary alif.1013

In the grammatical tradition, maqṣūr and mamdūd are usually discussed under ṣarf in works which deal with both naḥw and ṣarf and feature promi- nently in works that deal exclusively with ṣarf. To give but one example, Sībawayhi (d. 180/796) raises several issues related to both types. In the case of maqṣūr – which he often refers to as manqūṣ1014 – he gives its definition;1015 cites some of its regular or qiyāsī patterns as well as examples which can only be known through attestation (samāʿ);1016 and discusses various rules pertaining

1012 Cf. Ibn Wallād, Maqṣūr 26, 53. 1013 See, for example, Farrāʾ, Maqṣūr 16–29; Waššāʾ, Mamdūd 52–55; Ibn Wallād, Maqṣūr

162–167. 1014 For referring to maqṣūr as manqūṣ, see also Ibn Wallād, Maqṣūr 3. 1015 Sībawayhi, Kitāb III, 536. 1016 Ibid., III, 536–539. Qiyāsī forms include verbal nouns of final weak verbs whose perfect

and imperfect are of the patterns faʿila and yafʿalu respectively, and whose adjectives are of the pattern afʿal; e.g. ʿamā (blindness), since its verb is ʿamiya/yaʿmā and its adjective is aʿmā. Examples of samāʿī words are qafā (back) and raḥā (quern).

 

amgad
Highlight
amgad
Highlight

 

243MUBAWWAB AND SPECIALIZED LEXICA

to maqṣūr words with respect to dual, plural, annexation of first person sin- gular suffix, diminutive, gentilic adjective (nisba), nunation (i.e. treatment as triptote or diptote), and imāla (fronting and raising of ā towards ī).1017 Similar issues also feature in his study of mamdūd.1018

Independent monographs on maqṣūr and mamdūd comprise most of the material available in grammatical works, but focus more on the patterns of maqṣūr and mamdūd words and on the distinction between the two types in form, meaning and writing. Out of the fifty or so works referred to in the sources,1019 about a dozen have survived. We shall first examine four of the earliest monographs and then discuss the development of the genre in subse- quent works. These four monographs share the title al-Maqṣūr wa-l-mamdūd 1020 (though in reverse order in Waššāʾ’s) and are authored by Farrāʾ (d. 207/822), Ibn al-Sikkīt (d. 244/858), Nifṭawayhi (d. 323/935), and Waššāʾ (d. 325/937).

The first two chapters in Farrāʾ’s book are devoted to the regular or qiyāsī patterns of maqṣūr and mamdūd words and the distinction between words of both types whose writing can be confusing (yuškil).1021 Most of the rest of the book is divided into chapters based on the vowel of the first letter of the word. Material on mamdūd, for instance, is divided into words which begin with kasra (e.g. ridāʾ “dress”), fatḥa (e.g. ṯanāʾ “praise”), and ḍamma (e.g. ḥudāʾ “cameleer’s song”).1022 A number of chapters are also devoted to words whose first vowels determines whether they are maqṣūr or mamdūd. One such group is that of words which begin with fatḥa and are mamdūd, but become maqṣūr if they have an initial kasra (e.g. balāʾ and bilā “wear, attrition”; qarāʾ and qirā “hospitality”).1023 The longest chapter, however, is reserved for maqṣūr words which have no mamdūd counterparts.1024 Throughout the book, Farrāʾ is keen

1017 Ibid., III, 386–390 (dual); III, 390–391, 394, 609, 617 (plural); III, 413–414 (annexation); III, 471 (diminutive); III, 342–343, 352–355 (gentilic adjective); III, 205–206, 210–213, 219 (nunation); IV, 119 (imāla).

1018 Ibid., III, 539 (definition); III, 539–541 (regular patterns and samāʿ); III, 391–392 (dual); III, 609, 617–618, 644, 647 (plural); III, 420–421 (diminutive); III, 349–352, 355–357 (gentilic adjective); III, 206, 213–215 (nunation).

1019 The most comprehensive list is the one by Harīdī in his introduction to Qālī, Maqṣūr 63–77. See also ʿAbdaltawwāb’s list in his introduction to Waššāʾ, Mamdūd 15–23; cf. Iqbāl (2011: 272–279).

1020 Note, however, that Farrāʾ’s book is also published under the title al-Manqūṣ wa-l-mamdūd (ed. by ʿAbdalʿazīz al-Maymanī al-Rāǧakūtī, Dār al-Maʿārif, Cairo 1967).

1021 Farrāʾ, Maqṣūr 7–29. 1022 Ibid., 82–90. 1023 Ibid., 30–31. 1024 Ibid., 54–81.

 

 

244 chapter 2

on indicating the type of alif used in writing the words he cites, probably in response to widespread error in this respect.

The three other books share some of the features in Farrāʾ’s, in particular the allocation of certain chapters to words which begin with a specific vowel and the distinction between maqṣūr and mamdūd words in meaning and (to a lesser extent than Farrāʾ) in writing. Most of Ibn al-Sikkīt’s chapters contain lists of words that have a specific pattern, such as fiʿʿīlā (e.g. ḫillīfā “caliphate”) and fuʿal (e.g. ḫuṭā “steps”) for maqṣūr;1025 and fāʿilāʾ (e.g. sāfiyāʾ “dust blown by the wind”) and mafʿūlāʾ (e.g. mašyūḫāʾ “elderly men”) for mamdūd.1026 Nifṭawayhi’s book is characterized by its author’s special attention to whether the cited words of both types are triptotes (munṣarif ) or diptotes (ġayr munṣarif ), as the titles of the vast majority of its chapters indicate. Accordingly, mamdūd words which begin with fatḥa, for example, are listed in two chapters, the first of which contains triptotes (e.g. hawāʾ “air”; sanāʾ “highness”) and the other diptotes (e.g. ḍarrāʾ “distress”; ramḍāʾ “swelter”).1027 As for Waššāʾ, he indicates that his book is a synopsis (muḫtaṣar) intended for the beginner (mutaʿallim).1028 It basically consists of the regular patterns of maqṣūr and mamdūd, and of sev- eral short chapters in which are grouped together words of both types, which begin with a specific vowel but whose meanings are different.

Specific features of the other extant works are briefly noted below:1029

1. Šarḥ al-Maqṣūr wa-l-mamdūd by Ibn Durayd (d. 321/933). This is a brief expla- nation of Ibn Durayd’s own qaṣīda (poem) on maqṣūr and mamdūd, the first of its kind which we possess.1030 The qaṣīda consists of fifty-seven verses divided into seven sections which range from one verse to thirty-three. As in earlier works, the sections are divided according to initial vowels and to whether the meanings of the maqṣūr and mamdūd words are different or not.

2. al-Maqṣūr wa-l-mamdūd by Ibn Wallād (d. 332/944). Among extant works, this is the first that is alphabetically arranged (according only to the first letter of the word). In the introduction, Ibn Wallād makes an interesting compari- son between Ḫalīl’s arrangement and his own, and explains that, for reasons which have to do with ease of use, his lexicon differs from al-ʿAyn in starting

1025 Ibn al-Sikkīt, Maqṣūr 48, 50. 1026 Ibid., 70, 72. 1027 Nifṭawayhi, Maqṣūr 33, 37. 1028 Waššāʾ, Mamdūd 56. 1029 For several Maġribī commentaries on maqṣūr and mamdūd works, see Wadġīrī (2008:

224–227). 1030 Ibn Durayd, Dīwān 138, 142.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight

 

245MUBAWWAB AND SPECIALIZED LEXICA

with alif, i.e. hamza (although he admits that some readers might blame him because it is a ḥarf muʿtall “weak letter”), and in listing words based on the order of their letters and not under their roots.1031 A typical chapter (e.g. alif, bāʾ, tā⁠ʾ etc.) consists of the four following parts: maqṣūr words whose meanings differ from those of their mamdūd counterparts; maqṣūr and mamdūd words which have the same meaning; words that are exclusively maqṣūr; and words that are exclusively mamdūd. The last two chapters are further divided into sections based on whether the first letter of the cited word is fatḥa, ḍamma or kasra, with special headings designated for words whose roots consist of more than three consonants, and for hamzated words. Appended to the lexicon are several chapters which discuss the regular patterns of maqṣūr and mamdūd words, the dual and plural of both types, and issues related to their written forms.1032 The total number of words cited is 1,168 (see 5 below).

3. al-Maqṣūr wa-l-mamdūd, attributed to Abū ʿUmar al-Zāhid, better known as Ġulām Ṯaʿlab (d. 345/957). This short risāla contains a list of 235 maqṣūr and mamdūd words, some of which are very briefly explained while others are merely cited without explanation. Its sections (none of which is alphabeti- cally arranged) are divided on the basis of the written form of the word, its initial vowel, and the difference in meaning between maqṣūr words and their mamdūd counterparts.

4. Šarḥ mā yuktab bi-l-yāʾ min al-asmāʾ al-maqṣūra wa-l-afʿāl by Ibn Durustawayhi (d. 347/958). As its title indicates, this short lexicon is restricted to maqṣūr nouns and verbs (e.g. ḥaṣā “pebbles”, sabā “to capture”).1033 Words are listed in full alphabetical order in most chapters (cf. however the chapters of šīn and ʿayn, where this principle is not strictly applied).1034

5. al-Maqṣūr wa-l-mamdūd by Abū ʿAlī al-Qālī (d. 356/967). This is the larg- est work in the genre and embraces 1,544 words, a few hundred more than the second largest work, Ibn Wallād’s, which contains 1,168 words.1035 It includes more extensive comments on the cited words than in earlier works and its two major sections, one for maqṣūr and another for mamdūd, are meticulously arranged. In each section, Qālī starts by the qiyāsī patterns and some mor- phological issues (e.g. dual forms) and then divides the bulk of his material into words which begin with fatḥa, kasra and ḍamma. Each section is further divided according to attested patterns, which amount to fifty-seven. Of these,

1031 Ibn Wallād, Maqṣūr 2–3; cf. below, 281. 1032 Ibid., 135, 167. For alleged errors in Ibn Wallād’s book, see ʿAlī b. Ḥamza, Tanbīhāt 325–354. 1033 Ibn Durustawayhi, Šarḥ 76, 82. 1034 Ibid., Šarḥ 83, 86. 1035 See the editor’s introduction to Qālī, Maqṣūr, 97.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

246 chapter 2

as he states in the introduction, thirty-five are exclusively nouns (asmāʾ; e.g. fāʿillā, as in bāqillā “beans”), nine are exclusively adjectives (ṣifāt; e.g. faʿallalā, as in ḍabaġṭarā “strong and feebleminded”), and thirteen are either nouns or adjectives (e.g. fuʿlā, as in the noun ḥummā “fever” and the adjective ḥublā “pregnant”).1036 The rare and anomalous forms (nawādir wa-šawāḏḏ) are listed in a separate section for each of the maqṣūr and mamdūd,1037 and are mostly taken by Qālī from Ḫalīl’s al-ʿAyn or directly from his own teacher, Ibn Durayd. Furthermore, the section on maqṣūr contains two other chapters, one for hamzated forms, and another for words which can be either maqṣūr or mamdūd with no change in meaning.1038 With regard to the arrangement of letters, Qālī, as in his muǧannas dictionary al-Bāriʿ, adopts a phonetic arrange- ment which differs from that of Ḫalīl’s and is closer to Sībawayhi’s.1039 The order of the letters is as follows: ʾ, h, ʿ, ḥ, ġ, ḫ, q, k, ḍ, ǧ, š, y, l, r, n, ṭ, d, t, ṣ, z, s, ẓ, ḏ, ṯ, f, b, m, w, and only the first letter of the word is taken into consideration in the arrangement. Unlike Ibn Wallād, who defends his adoption of the alpha- betical system on the basis of its ease although it begins with a ḥarf muʿtall (cf. 2 above), and unlike his own teacher Ibn Durayd, who previously cited the same reason for choosing that system,1040 Qālī does not comment on the issue of ease, but simply states that he deems it appropriate (ra⁠ʾaynā) to adopt a system which progresses from the farthest point of articulation (i.e. hamza) to the nearest (i.e. wāw).1041 Also noteworthy is the fact that Ibn Sīda (d. 458/1066) adopts in his discussion of maqṣūr and mamdūd in al-Muḫaṣṣaṣ Qālī’s method of dividing the material according to patterns, but rearranges these so as to bring together patterns which differ only in their initial vowels.1042 More sig- nificantly perhaps, he modifies Qālī’s phonetic ordering of the letters so as to fit Ḫalīl’s arrangement, which Ibn Sīda adopts in his own al-Muḥkam.

6. Maqāyīs al-maqṣūr wa-l-mamdūd by Abū ʿAlī al-Fārisī (d. 377/987). The book consists of two sections, one for maqṣūr and another for mamdūd. Various types of both are discussed, with special interest in certain patterns (e.g.

1036 Qālī, Maqṣūr 7–10. 1037 Ibid., 293, 494. 1038 Ibid., 267, 283. 1039 Sībawayhi, Kitāb IV, 431; see also below, 305. 1040 Ibn Durayd, Ǧamhara I, 40. 1041 Qālī, Maqṣūr 11. 1042 For example, the patterns faʿallā, fiʿillā, fuʿallā and fuʿullā, which are discussed by Qālī in

three different parts of his book (Maqṣūr 154–155, 200–202, 258–260), are consecutively examined by Ibn Sīda (Muḫaṣṣaṣ XV, 205–207).

 

amgad
Highlight
amgad
Highlight
amgad
Highlight

 

247MUBAWWAB AND SPECIALIZED LEXICA

verbal nouns and passive participles of maqṣūr).1043 One theme which recurs throughout the book is ilḥāq1044 as applied to maqṣūr and mamdūd forms.1045

7. Mā yaḥtāǧ ilayhi l-kātib min mahmūz wa-maqṣūr wa-mamdūd mimmā yuk- tab bi-l-alif wa-l-yāʾ ʿalā ḥurūf al-muʿǧam by Ibn Ǧinnī (d. 392/1002). Following an alphabetically arranged list of hamzated words,1046 the author briefly men- tions a few rules for writing the final alif of maqṣūr nouns and verbs.

8. Ḥilyat al-ʿuqūd fī l-farq bayna l-maqṣūr wa-l-mamdūd by Abū l-Barakāt b. al-Anbārī (d. 577/1181). The section on maqṣūr begins with its definition and its qiyāsī patterns and samāʿī forms, followed by a list of maqṣūr words whose initial vowel is fatḥa, kasra or ḍamma. The same topics are repeated with the mamdūd. The topics and the arrangement of material based on initial vowels, as well as lack of alphabetical order in word lists, bring Ibn al-Anbārī’s book close to that of the four authors discussed earlier in this section. In fact, this demonstrates how little this genre has developed over several centuries, apart from the introduction of alphabetical order, possibly for the first time by Ibn Wallād (although he does not differ much from earlier authors in other criteria of arrangement and in the topics of discussion), and certainly apart from Qālī, whose approach is markedly different from that of his predecessors.

9. Tuḥfat al-mawdūd fī l-maqṣūr wa-l-mamdūd by Ibn Mālik (d. 672/1274). Like Ibn Durayd (cf. 1 above), Ibn Mālik explains his own qaṣīda on maqṣūr and mamdūd, which consists of 151 verses (in addition to ten in the introduc- tion and five in the epilogue), almost three times the size of Ibn Durayd’s. Each verse of the qaṣīda contains two maqṣūr words and their mamdūd counter- parts, and in the commentary Ibn Mālik explains the meanings of the words and for that purpose often quotes poetry šawāhid. The qaṣīda is divided into sections according to the same principles applied by Ibn Durayd, but Ibn Mālik’s division is more elaborate, with sixteen sections instead of only seven.

10. Manẓūmat al-maqṣūr wa-l-mamdūd by Ibn Ǧābir al-Andalusī (d. 780/1378). This qaṣīda greatly resembles Ibn Mālik’s despite its author’s criticism in its introductory part (line 10) of his predecessor on the basis of the complexity of his poetic composition.1047 It is made up of 153 lines (very close to the 151 lines in Ibn Mālik’s qaṣīda), in addition to fourteen in the introduction, seventeen in the epilogue and sixteen for the titles of sections. The author faithfully follows Ibn Mālik in the division of his sections and even in the examples he cites.

1043 Fārisī, Maqāyīs 35–36. 1044 Cf. above, 60, n. 296. 1045 Fārisī, Maqāyīs 28, 65, 72, 96. 1046 Cf. above, 218. 1047 Ibn Ǧābir, Manẓūma 14.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

248 chapter 2

10.4 al-Muṯallaṯāt (Triplets) The term muṯallaṯ (pl. muṯallaṯāt) refers to a group of three words which are identical in root and pattern but one of whose radicals has fatḥa in one, kasra in another, and ḍamma in the third. Words in some triplets are identical in meaning – such as ṯanāʾ, ṯināʾ and ṯunāʾ (praise)1048 – but in most triplets words have different meanings – such as salām (greeting), silām (stones), and sulām (bones of the outer side of the hand and foot).1049 In a few cases, the triplets may differ in more than one vowel, as in the reduplicated forms samsam (wolf or a place name), simsim (sesame), and sumsum (a swift man).1050 Although one of Sībawayhi’s (d. 180/796) students, Quṭrub (see below), is generally cred- ited with being the first to author a book on muṯallaṯāt, neither the Kitāb nor subsequent grammatical works deal with this phenomenon. Whether Quṭrub was the first to note the phenomenon or merely the first to deal with it in an independent work is indeed difficult to establish.

We shall examine the following extant works of this genre:

1. al-Muṯallaṯāt by Quṭrub (d. 206/821). 2. A part of K. al-Muṯallaṯ by al-Qazzāz al-Qayrawānī (d. 412/1021). 3. al-Muṯallaṯ by Ibn al-Sīd al-Baṭalyawsī (d. 521/1127). 4. al-Iʿlām bi-muṯallaṯ al-kalām and Ikmāl al-iʿlām bi-taṯlīṯ al-kalām by Ibn

Mālik (d. 672/1274). 5. al-Durar al-mubaṯṯaṯa fī l-ġurar al-muṯallaṯa by Fīrūzābādī (d. 817/1415).

It should come as no surprise that only one book from the third/ninth cen- tury has been preserved, given that it is one of only two books on muṯallaṯāt mentioned from that period in the sources, the other being Abū Zayd al-Anṣārī’s K. al-Taṯlīṯ.1051 In fact, sources also mention only two other books by fourth/tenth century scholars, namely, K. al-Muṯallaṯ by Waššāʾ (d. 325/937)1052 and K. al-Muṯallaṯ al-ṣaḥīḥ by Abū l-Ḥasan al-Šimšāṭī (d. after 377/987).1053 The authoring of books on muṯallaṯāt, however, seems to have somewhat picked up as of the fifth/eleventh century. From that period until the early ninth/fifteenth century, the number of works mentioned in the sources amounts to twelve.1054

1048 Baṭalyawsī, Muṯallaṯ I, 387–388. 1049 Quṭrub, Muṯallaṯāt 32–33. 1050 Baṭalyawsī, Muṯallaṯ II, 432–433. 1051 Yāqūt, Muʿǧam III, 1361; Suyūṭī, Buġya I, 583. 1052 Ibn al-Nadīm, Fihrist 93; Yāqūt, Muʿǧam V, 2304; Qifṭī, Inbāh III, 62; Suyūṭī, Buġya I, 18. 1053 Ibn al-Nadīm, Fihrist 172; Yāqūt, Muʿǧam IV, 1908. 1054 See a full list in Farṭūsī’s introduction to Baṭalyawsī, Muṯallaṯ I, 50–62.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

249MUBAWWAB AND SPECIALIZED LEXICA

Quṭrub’s book comprises thirty-two muṯallaṯāt (listed in no order), thirty of which consist of nouns which differ in meaning. The other two are made up of verbs which also have different meanings, e.g. ḥalama (to dream), ḥalima (to become spoilt or perforated; said of hide or skin), and ḥaluma (to become forbearing).1055 The words of each triplet are briefly explained, often with sup- porting poetry šawāhid. The content of the book strongly suggests that the notion of muṯallaṯāt in Quṭrub’s time was restricted to triplets whose words have different meanings and that the expansion of this notion to include trip- lets whose words share the same meaning was a later innovation. Several com- mentaries and versified versions of Quṭrub’s book have also been preserved.1056

al-Qazzāz al-Qayrawānī’s book is only partially preserved and it is not possi- ble to determine whether it included any triplets whose words share the same meaning. According to the editor of its last chapter (which exclusively deals with verbs), the number of triplets in the preserved parts of the book amounts to 220, arranged according to the Maġribī alphabet and on the basis of the first letter of the word.1057

The first complete book which we have after Quṭrub’s is the one by Ibn al-Sīd al-Baṭalyawsī. The contrast between the two books in size and content is staggering. In fact, Baṭalyawsī himself criticizes Quṭrub’s book not only for its conciseness, but also for its inclusion of material that does not truly belong to muṯallaṯāt.1058 Indeed, the number of triplets collected by Baṭalyawsī – which, according to his own enumeration, is 695 of triplets with words of dif- ferent meanings and 138 of those with words of the same meaning1059 – dwarfs Quṭrub’s thirty-two triplets. Concerning Quṭrub’s inclusion of material that does not fit the criteria of muṯallaṯāt, Baṭalyawsī cites kala⁠ʾ (herbage), kulā (pl. of kulya, kidney), and kilā (protection),1060 and asserts that the three words do not form a muṯallaṯ since the first ends with a hamza, the second is not hamzated, and the third is mamdūd (i.e. originally kilāʾ). As for triplets whose words share the same meaning, Baṭalyawsī’s book, as far as we are able to determine, is the first to include them. It seems, however, that they are mostly dialectal variants of the same word; e.g. ǧaḏwa, ǧiḏwa and ǧuḏwa (firebrand),

1055 Quṭrub, Muṯallaṯāt 34. 1056 See a list of commentaries and versifications in Farṭūsī’s introduction to Baṭalyawsī,

Muṯallaṯ I, 51–58. See also Quṭrub, Muṯallaṯāt 31–68 for the text in urǧūza form; cf. Wadġīrī (2008: 212–222).

1057 See the editor’s introduction to Qazzāz, Muṯallaṯ 302–304; cf. Wadġīrī (2008: 211–212). 1058 Baṭalyawsī, Muṯallaṯ I, 297. 1059 Ibid., I, 298, 299. 1060 Quṭrub, Muṯallaṯāt 49; Baṭalyawsī, Muṯallaṯ I, 297–298.

 

 

250 chapter 2

rašwa, rišwa and rušwa (bribe), ʿafw, ʿifw and ʿufw (young mule), etc.1061 and are closely related to dialectal material found in books on ġarīb. Like al-Qazzāz, Baṭalyawsī adopts the Maġribī order of the alphabet and arranges words in each chapter according the fist letter of the word. In each letter, he begins with triplets whose words agree in meaning (except for ṯ, ḏ, ẓ and ḍ where no exam- ples are found), followed by triplets whose words have different meanings.

Ibn Mālik’s contribution to muṯallaṯāt is quite extensive. His al-Iʿlām bi-muṯallaṯ al-kalām is an urǧūza of 247 verses, and seems to be more elaborate than a prose text, with the same title, which he most probably authored earlier.1062 The urǧūza begins with triplets in which words share the same meaning. These are divided into nouns and verbs, and it is obvious that most of the cited words are dialectal variants which at times do not even fit the definition of triplets, particularly when they exceed three in number, as in the following seven forms iṣbaʿ, uṣbaʿ, iṣbiʿ, uṣbuʿ, aṣbaʿ, iṣbuʿ and uṣbūʿ, all of which mean “finger”.1063 The rest of the book, which is devoted to triplets whose words have different mean- ings, is divided into chapters from alif to yāʾ. In each chapter, words appear according to full alphabetical order. The number of triplets cited in the urǧūza is 1,364, of which 159 belong to the type whose words agree in meaning (com- pared in the prose text to 793, out of which 96 are of that type).1064 In spite of this large number, Ibn Mālik wanted to author a book which would include an even larger number of triplets – hence the name Ikmāl (lit. completion, sup- plementation) in its title. As he explains in his introduction, he did not know of Baṭalyawsī’s book when he wrote al-Iʿlām, and thus wanted to make use, in his new book, of its material which he had not included in al-Iʿlām, as well as of material in Azharī’s (d. 370/981) Tahḏīb al-luġa and Ibn Durayd’s (d. 321/933) Ǧamharat al-luġa, among others.1065 Accordingly, the number of triplets in the new book is 2,294, of which only 163 are of the type whose words do not vary in meaning. As in al-Iʿlām, Ibn Mālik begins with the latter type (but note that words are divided here on the basis of the position of the varying vowel), and then divides the rest of the book into chapters from alif to yāʾ, listing words in full alphabetical order.

The last book in our list, Fīrūzābādī’s al-Durar al-mubaṯṯaṯa, is restricted to triplets whose words have the same meaning. Their total number in the chapters from alif to yāʾ is 453, and they are arranged in full alphabetical order,

1061 Baṭalyawsī, Muṯallaṯ I, 393; II, 29, 251. 1062 Cf. Ġāmidī’s introduction to Ibn Mālik, Ikmāl I, 63–65. 1063 Ibn Mālik, Iʿlām 13. Uṣbūʿ is described as the prolonged (mušbaʿ) version of uṣbuʿ. 1064 Ġāmidī’s introduction to Ikmāl I, 83–84. 1065 Ibn Mālik, Ikmāl I, 4–6.

 

amgad
Highlight
amgad
Highlight

 

251MUBAWWAB AND SPECIALIZED LEXICA

starting with the first consonant of the word and not the last, and with no consideration for roots,1066 unlike his al-Qāmūs al-muḥīṭ. Fīrūzābādī mentions in his introduction a number of maṯallaṯāt books which he consulted, and cor- rectly notes that his own contains more triplets of this type than any of them.1067

Finally on muṯallaṯāt, Ibn Fāris’s (d. 395/1004) K. al-Ṯalāṯa is worthy of mention since it deals with a special type of triplets which differs from those encountered in other works. Words in these triplets share the same pattern (with no difference in vowels) as well as the radicals of their roots, but in reverse order. Among the examples cited are ḥalīm (forbearing), ḥamīl (of suspected origin) and laḥīm (murdered), and ʿuqāb (large banner), qubāʿ (feebleminded) and buʿāq (sound of vehement rain).1068 Although permutations of the radicals are restricted to three, they are reminiscent of the six permutations of triliteral verbs cited in earlier muǧannas lexica, such as Ḫalīl’s (d. 175/791) al-ʿAyn and Ibn Durayd’s (d. 321/933) Ǧamharat al-luġa. The thirty-four triplets cited by Ibn Fāris are all nouns, and are listed in no specific order.

10.5 Nominal Patterns As noted earlier, nominal patterns were discussed in the grammatical tradi- tion and in multithematic works, and even formed part of the material of some muǧannas lexica.1069 Other than the nominal patterns referred to under ištiqāq (e.g. taṣġīr, taṯniya, etc; see 10.1 above), muḏakkar and muʾannaṯ, and maqṣūr and mamdūd (see 10.2 and 10.3 above), a number of monographs deal with specific nominal patterns. In particular, verbal nouns seem to have been the focus of the interest of early lexicographers, as several books titled al-Maṣādir are attributed to them in the sources.1070 Third/ninth and fourth/ tenth century scholars to whom is attributed a book with such a title include Kisāʾī (d. 189/805), al-Naḍr b. Šumayl (d. 203/819), Abū ʿUbayda (d. 209/824), Abū Zayd al-Anṣārī (d. 215/830), Nifṭawayhi (d. 323/935), Fārābī (d. 350/961), and one of the Bedouin fuṣaḥāʾ, Qurayba Umm al-Buhlūl al-Asadiyya (fl. in the

1066 For example, tafāwut (disparity) and ma⁠ʾdaba (banquet) are placed under t and m respectively; cf. Fīrūzābādī, Durar 84, 181. There even are entries under yāʾ (ibid., 215–231) which are verbs in the imperfect, such as yaqlā (to hate), yanbaʿ (to pour out), and yanhaq (to bray). For reading aids in the Durar, see below, 396.

1067 Ibid., 48–49. 1068 Ibn Fāris, Ṯalāṯa 31, 45–46. 1069 See above, 232–234. 1070 Verbal nouns are also discussed in the more general philological works, such as Ṯaʿlab’s

(d. 291/904) al-Faṣīḥ (cf. above, 175) and Ibn Fāris’s (d. 395/1004) Tamām Faṣīḥ al-kalām 22 ff., and in multithematic works as early as Abū ʿUbayd’s (d. 224/838) al-Ġarīb al-muṣannaf III, 684–689.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

252 chapter 2

era of data collection, probably in the first half of the third/ninth century).1071 Books entitled Maṣādir al-Qurʾān or al-Maṣādir fī l-Qurʾān are also attributed to Farrāʾ (d. 207/822), Aṣmaʿī (d. 216/831), and Ibrāhīm b. Yaḥyā b. al-Mubārak al-Yazīdī (d. 225/840).1072 No book of either genre has survived, and hence it is not possible to tell whether the approach of authors of the first genre differs from that of the grammarians,1073 or what aspect of verbal nouns in the Qurʾān was discussed by authors of the second genre.

A relatively small number of works which exclusively deal with nominal patterns have survived (see also 10.8 below for nominal patterns in more gen- eral lexica). Among these is Ibn Ǧinnī’s (d. 392/1002) monograph, al-Muqtaḍab min kalām al-ʿArab, in which he deals with the mafʿūl (i.e. passive participle) pattern from triliteral hollow verbs, such as mašūb (admixed) and mašīd (con- structed), from šwb and šyd respectively.1074 Following a short introduction on qiyāsī and samāʿī participles of the mafʿūl pattern is a lexicon in which exam- ples of this pattern are listed in chapters from alif to hāʾ. In each chapter, forms derived from verbs whose second radical is wāw are cited before those whose second radical is yāʾ. Words are arranged in full alphabetical order based on the third radical and then the first. As for the rules that apply to doubly weak roots (e.g. qww, šwy and ḥyy), they are discussed after the chapter on hāʾ.1075

In an equally short tractate titled K. al-Nayrūz, Ibn Fāris (d. 395/1004) explains the meaning of the word Nayrūz (Persian New Year’s Day) and then lists and explains forty-three other words of the same pattern, fayʿūl, including those which denote place names (e.g. Bayrūt) and proper nouns (e.g. Ayyūb).1076 Some of his material is drawn from Ibn Durayd’s (d. 321/933) list of twenty- eight words of the pattern fayʿūl in Ǧamharat al-luġa.1077

1071 Ibn al-Nadīm, Fihrist 72, 58, 59, 60, 90, 53; Yāqūt, Muʿǧam IV, 1752; VI, 2761; I, 122; VI, 2805; Qifṭī, Inbāh III, 352, 286; II, 53; IV, 121 respectively.

1072 Ibn al-Nadīm, Fihrist 73, 61, 56; Yāqūt, Muʿǧam VI, 2815; I, 160; Qifṭī, Inbāh IV, 22; II, 203; I, 225, 226 respectively.

1073 Most of Sībawayhi’s material on verbal nouns is discussed under several headings of his lengthy chapter on the patterns of verbs and verbal nouns (Kitāb IV, 5–97). Topics include the meanings of certain patterns (e.g. the use of faʿal to indicate hunger or thirst, and of fuʿla to indicate color; IV, 21, 25); examples representative of specific patterns (e.g. faʿūl and mafʿūl; IV, 42, 43); and regular patterns connected to specific types of verbs (e.g. quadriliterals and weak verbs which begin with wāw; IV, 85, 92).

1074 Ibn Ǧinnī, Muqtaḍab 11, 15. 1075 Ibid., 33–35. 1076 Ibn Fāris, Nayrūz, 18, 24. 1077 Ibn Durayd, Ǧamhara II, 1204–1205.

 

amgad
Highlight
amgad
Highlight

 

253MUBAWWAB AND SPECIALIZED LEXICA

A very short text on the pattern tif ʿāl by the famous poet and scholar Abū l-ʿAlāʾ al-Maʿarrī (d. 449/1057) is also preserved. The material was dictated in one of Maʿarrī’s maǧālis and recorded by his student, al-Ḫaṭīb al-Ṭibrīzī (d. 502/1109). The text includes twenty-six randomly arranged words (com- pared with seventeen in Ǧamharat al-luġa),1078 some of which are substan- tives (i.e. asmāʾ, rather than maṣādir), such as timsāḥ (crocodile), and others are verbal nouns (maṣādir), such as tibyān (manifestation).1079

From the seventh/thirteenth century we have four monographs on nominal patterns, all of which are by Ṣaġānī (d. 650/1252). These are Naqʿat al-ṣadyān fī mā ǧāʾa ʿalā l-faʿalān,1080 al-Infiʿāl, Yafʿūl, and Mā banathu l-ʿArab ʿalā faʿāli. The first includes 211 words which are verbal nouns of the pattern faʿalān, e.g. ḏawabān (melting), waṯabān (jumping), and daraǧān (walking).1081 The sec- ond embraces 437 words of the reflexive pattern infaʿala and its verbal noun infiʿāl, e.g. insadaḥa (to lie down) and inhiǧās (moving backward).1082 In the introduction, Ṣaġānī proudly asserts that he is the first to author a book on the infaʿala pattern.1083 The book is divided into sections based on the type of the triliteral root, e.g. sound (sālim), reduplicated (muḍāʿaf ), hollow (aǧwaf ), defective (nāqiṣ), and hamzated (mahmūz). Of the pattern yafʿūl, Ṣaġānī col- lects forty-one words, which are either substantives, e.g. yaʿfūr (gazelle) and yanbūʿ (fountainhead), or place names, e.g. Yasnūm and Yankūb.1084 In the fourth monograph, Ṣaġānī lists 130 words of the pattern faʿāli, followed by seven words which also end with kasra, but whose roots are either redupli- cated, e.g. ʿarʿāri (a game for boys), or quadriliterals, e.g. dahdāʿi (a crying at she-goats).1085 Suyūṭī (d. 911/1505) quotes in his al-Muzhir the bulk of Ṣaġānī’s words, but without the šawāhid or detailed explanations, and rearranges them into three groups.1086 Words in the first group indicate the imperative or what

1078 Ibid., II, 1205. 1079 Maʿarrī, Tifʿāl 7, 9. 1080 It is likely that the correct title of the book is simply K. al-Faʿalān and that Naqʿat al-ṣadyān,

which is the title of another book by Ṣaġānī, was erroneously added to it. The book titled Naqʿat al-ṣadyān deals with the names of muḥaddiṯūn (transmitters of prophetic Ḥadīṯ) and was published by Aḥmad Ḫān in al-Mawrid 16/2 (1987), 151–186.

1081 Ṣaġānī, Faʿalān 22, 24, 26. 1082 Ṣaġānī, Infiʿāl 12, 28. Note also that Ibn al-Qaṭṭāʿ in his lists of verbs often cites the verbal

noun, rather than the verb; cf. Afʿāl I, 372–380; II, 401–407, etc. 1083 Ibid., 1. 1084 Ṣaġānī, Yafʿūl 25, 31, 22, 32. For a list of words of the pattern yafʿūl in Lisān al-ʿArab, see

ʿAmāyira (2007: 42–46). 1085 Ṣaġānī, Faʿāli 100, 102–103. 1086 Suyūṭī, Muzhir II, 131–132.

 

amgad
Highlight

 

254 chapter 2

the grammarians call ism fiʿl (lit. proper name of the verb; e.g. tarāki “leave”!). In the second group are listed place names (e.g. Ḥadābi), and in the third ani- mals (e.g. ǧaʿāri “she-hyena”), inanimate objects (e.g. barāḥi “sun” and ǧadāʿi “a year of drought”), female proper nouns (e.g. Saǧāḥi), curses addressed to women (e.g. yā ḫabāṯi “O wicked woman”), etc. In all four of his books, Ṣaġānī seems to be keen on exhausting the words which are of the pattern at hand. This is clearly spelled out in the case of al-Infiʿāl,1087 and can also be confirmed by the fact that, compared with the 211 words of the pattern faʿalān men- tioned by Ṣaġānī, Ibn Qutayba (d. 276/889) lists merely seventeen,1088 and Ibn Durayd (d. 321/933) no more than twenty-four.1089 In the case of yafʿūl, Ibn Durayd mentions thirty-two examples,1090 whereas Ṣaġānī lists forty-one. As for arrangement, Ṣaġānī arranges words according to the last radical of their roots, followed by the first, in three of his four books, as he does in his two muǧannas dictionaries, al-ʿUbāb al-zāḫir and al-Takmila wa-l-ḏayl wa-l-ṣila. For no obvious reason, however, he arranges words in his book on yafʿūl according to the first radical of their roots.

10.6 Faʿala and Afʿala Of the thirty-four verbal patterns of Arabic,1091 the ones that were the focus of the philologists’ attention as early as the third/ninth century are faʿala and afʿala (i.e. first and fourth forms). With the exception of infaʿala and its verbal noun infiʿāl, on which Ṣaġānī wrote a much later monograph (see 10.5 above), none of the other patterns seems to have been discussed in a separate work.1092 The reason for the interest in faʿala (and its two counterparts faʿila and faʿula) and afʿala most likely stems from the fact that they can either have the same meaning (e.g. baraqa and abraqa, both of which mean “to threaten”) or two different meanings (e.g. ṯāba “to return”, but aṯāba “to reward”).1093 As Ibn Sīda (d. 458/1066) notes in his lengthy discussion of faʿala and afʿala,1094 this

1087 Ṣaġānī, Infiʿāl 1–2. 1088 Ibn Qutayba, Adab 466. 1089 Ibn Durayd, Ǧamhara III, 1236–1237. 1090 Ibid., II, 1200–1201. 1091 Cf. above, 232. 1092 The ascription to Ṣaġānī of a book titled al-Iftiʿāl is doubtful and is most probably due

to the distortion of the written form of infiʿāl. The book is not mentioned in the sources contemporary with Ṣaġānī and appears in much later works, such as Ḥāǧǧī Ḫalīfa, Kašf II, 1394 and Laknawī, Fawāʾid 63.

1093 Zaǧǧāǧ, Faʿaltu 6, 14–15. 1094 Ibn Sīda, Muḫaṣṣaṣ XIV, 166–173; XV, 2–57.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

255MUBAWWAB AND SPECIALIZED LEXICA

phenomenon is closely related to dialectal differences1095 – more reason to arouse the interest of philologists. The resulting confusion between the two patterns by speakers seems to have prompted some scholars to try to amend the situation. This is readily demonstrable in the earliest extant monograph on laḥn, i.e. Mā talḥan fīhi l-ʿāmma by Kisāʾī (d. 189/805), where several items deal with material related to faʿala and afʿala,1096 and in the lengthy section of Ibn al-Sikkīt’s (d. 244/858) Iṣlāḥ al-manṭiq in which are listed verbs of the faʿala and afʿala patterns that are erroneously changed by the ʿāmma (com- mon people, generality of people) to afʿala and faʿala respectively.1097

As in many other genres, the writing of monographs on faʿala and afʿala goes back to the third/ninth century. Ibn al-Nadīm (d. 380/990) attributes a monograph titled Faʿala wa-afʿala or Faʿaltu wa-afʿaltu to eight philolo- gists of this period, namely, Quṭrub (d. 206/821), Farrāʾ (d. 207/822), Abū ʿUbayda (d. 209/824), Abū Zayd al-Anṣārī (d. 215/830), Aṣmaʿī (d. 216/831), Tawwazī (d. 233/847), Ibn al-Sikkīt (d. 244/858), and Aḥwal (d. after 250/864).1098 Although none of these has survived, the following few that have are mostly from the third/ninth and fourth/tenth centuries:

1. Faʿaltu wa-afʿaltu by Abū Ḥātim al-Siǧistānī (d. 255/869). 2. Faʿaltu wa-afʿaltu by Zaǧǧāǧ (d. 311/923). 3. al-Afʿāl by Ibn al-Qūṭiyya (d. 367/977). 4. Part of Tamām Faṣīḥ al-kalām by Ibn Fāris (d. 395/1004). 5. Mā ǧāʾa ʿalā faʿaltu wa-afʿaltu bi-maʿnā wāḥid by Ǧawālīqī (d. 540/1145).

The absence of mention of Abū Ḥātim’s book in the biographical and bib- liographical sources notwithstanding, there is little doubt that he is the real author, given the many quotations from it in later works which agree

1095 Ibid., XIV, 171. Note also that Ibn Durustawayhi (d. 347/958) insists that faʿala and afʿala can have the same meaning only in two different dialects (Taṣḥīḥ I, 165–167; also quoted in Suyūṭī, Muzhir I, 384–385).

1096 See Kisāʾī, Mā talḥan 110 (šaġala/ašġala; waʿada/awʿada), 119 (šakala/aškala; ḥarama/ aḥrama), 121 (ġalaqa/aġlaqa), 128 (ʿayiya/aʿyā), 129 (ḥadda/aḥadda), 130 (ṣaḥā/aṣḥā), 133 (kabata/akbata; ḫaṣā/aḫṣā), 134 (malaka/amlaka), 135 (ṣadaqa/aṣdaqa), 136 (qabasa/ aqbasa; ḥāṭa/aḥāṭa), 137 (dafaqa/adfaqa; haraqa/ahraqa).

1097 Ibn al-Sikkīt, Iṣlāḥ 225–227, 227–280. See also above, 175, where mention is made of material on verbs in Ṯaʿlab’s (d. 291/904) al-Faṣīḥ.

1098 Ibn al-Nadīm, Fihrist 58, 74, 59, 60, 61, 63, 79, 87 respectively. See also a list of works on faʿala and afʿala attributed to authors from the third/ninth to the seventh/thirteenth century in the editor’s introduction to Abū Ḥātim, Faʿaltu 67–70.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

256 chapter 2

with its content and which are attributed to Abū Ḥātim.1099 The same text is erroneously attributed to Aṣmaʿī,1100 probably because Abū Ḥātim begins the book by acknowledging Aṣmaʿī, with whom he checked, word for word, examples of faʿala and afʿala which have the same meaning (hāḏā bāb faʿaltu wa-afʿaltu bi-maʿnā wāḥid ʿan . . . al-Aṣmaʿī sa⁠ʾaltuhu ʿanhu ḥarfan ḥarfan).1101 This statement might suggest that the text is restricted to the first type in which the two patterns have the same meaning or that another section of the book contains examples of the second type in which they differ in meaning. But in reality there are no divisions in the book, and it does contain examples of both types,1102 as well as examples in which only faʿala or afʿala are attest- ed.1103 It is interesting to note, however, that in an earlier multithematic work, Abū ʿUbayd (d. 224/838) does separate the two types,1104 as does one of Abū Ḥātim’s contemporaries, Ibn Qutayba (d. 276/889).1105 Similarly, Ibn Durayd (d. 321/933) includes, in the latter parts of his muǧannas lexicon, Ǧamharat al-luġa, a chapter on faʿala and afʿala in which their two types are clearly distinguished.1106

Compared with earlier authors, Zaǧǧāǧ not only separates the two types of faʿala and afʿala, but organizes the material alphabetically based only on the first letter of the verb’s root (though we do not know for certain whether he was the first to do so). The major section of the book1107 is devoted to the two types of faʿala and afʿala and is divided into chapters from bāʾ to yāʾ, with hamza placed just before yāʾ, obviously in order not to be confused with the hamza of the pattern afʿala, which is not part of the root. In each chapter, examples of the first type are listed before those of the second, but with no further alpha- betical arrangement. In two shorter sections1108 he lists – also in alphabetical order based on the first letter only – examples in which one of the two patterns is either the only correct form or the choicer (mā ḫtīra fīhi).

1099 Editor’s introduction to Abū Ḥātim, Faʿaltu 46–51. 1100 Ed. by ʿAbdalkarīm Ibrāhīm al-ʿAzbāwī in Maǧallat al-Baḥṯ al-ʿIlmī wa-l-Turāṯ al-Islāmī, 4

(1981), 467–527. 1101 Abū Ḥātim, Faʿaltu 82. 1102 An example of the second type is ǧabara (to splint broken bones) and aǧbara (to compel);

ibid., 96. 1103 E.g. ǧāḥa (exterminate), but not *aǧāḥa; adǧana (to become cloudy), but not *daǧana;

ibid., 101–117. 1104 Abū ʿUbayd, Ġarīb II, 567–576, 576–585. 1105 Ibn Qutayba, Adab 333 ff. 1106 Ibn Durayd, Ǧamhara III, 1257–1265, 1265–1267. 1107 Zaǧǧāǧ, Faʿaltu 5–103. 1108 Ibid., 107–123, 127–146.

 

amgad
Highlight
amgad
Highlight

 

257MUBAWWAB AND SPECIALIZED LEXICA

Ibn al-Qūṭiyya’s K. al-Afʿāl is considerably more extensive than Abū Ḥātim’s or Zaǧǧaǧ’s. It is extremely well organized, but it is specifically this feature that makes it hard to use, in addition to its unique alphabetical arrangement. Following an introduction on the verbal patterns faʿala/faʿula/faʿila and afʿala, and on the various patterns of their verbal nouns,1109 the book is divided in a manner that resembles Zaǧǧāǧ’s, that is, into the following three sections, although the last two are in reverse order: faʿala and afʿala of both types, afʿala (referred to as rubāʿī), and faʿala (referred to as ṯulāṯī).1110 In the first section – which is divided into chapters from hamza to yāʾ – verbs of the first type in which both patterns share the same meaning precede those of the second, as in Zaǧǧāǧ’s book. Further internal divisions (as in the letter lām for example)1111 are based on the type of the roots of verbs in the following order: reduplicated (muḍāʿaf ), sound (ṣaḥīḥ), hamzated (mahmūz), and unsound with either wāw or yāʾ (muʿtall). Verbs are further divided, based on the vowel of their second radical and whether they have one form or more, as follows: faʿala; faʿila; faʿala, fuʿila and faʿila; faʿala and faʿula; faʿala, faʿula and faʿila; faʿala and faʿila; faʿila and faʿula. As for verbs with wāw and yāʾ, they are divided into the following subcategories: hollow (aǧwaf ) verbs with wāw, yāʾ or both; defective (nāqiṣ) verbs of the same three types; and defective verbs whose final radical is vocal- ized (e.g. laġiya ‘to talk nonsense’). The same principle of arrangement is also applied in the second and third sections. Accordingly, in order for the user to find a certain verb, it is not only necessary to specify its pattern and type of root, but also to know whether it has only one pattern or more. The book is thus more of a study of the morphology of verbs than an easily accessible lexicon. What makes its use even more difficult is its hybrid alphabetical arrangement which is based on the phonetic system, but with several letters of similarly written forms placed consecutively. The order of letters is as follows: ʾ, h, ʿ, ġ, ḫ, ḥ, ǧ, q, k, s, š, ṣ, ḍ, l, r, n, ṭ, ẓ, ḏ, d, b, t, ṯ, z, f, m, w, y (note that neither r and z nor f and q are placed adjacently in spite of similarity in form). The fact that only the first letter is taken into consideration in arrangement is another contributing factor to the painstaking effort needed to allocate a specific verb.

Material on faʿala and afʿala forms almost half of Ibn Fāris’s short risāla titled Tamām Faṣīḥ al-kalām, a supplement to Ṯaʿlab’s (d. 291/904) al-Faṣīḥ.1112

1109 Ibn al-Qūṭiyya, Afʿāl 1–9. 1110 Ibid., 1–162, 163–175, 176–304. 1111 Ibid., 90–96. 1112 Cf. above, 175, n. 608; 251, n. 1070. The reason for including Ibn Fāris’s book in this section

(rather than under Section 8 below) although it does contain material related to nouns, is that its section on verbs is restricted to faʿala and afʿala, contrary, say, to Fārābī’s Dīwān al-adab, in which other verbal patterns (e.g. faʿʿala, fāʿala, iftaʿala, etc.) are included.

 

 

258 chapter 2

Although Ibn Fāris designates a short section for faʿala and afʿala which differ in meaning,1113 his main focus is on the unaugmented (muǧarrad) verb, i.e. faʿala/faʿila/faʿula. Examples of the following types are listed, without inter- nal alphabetical order: faʿala (intransitive); faʿila; faʿala (transitive); fuʿila; faʿila and faʿala with different meanings; afʿala; verbs which are followed by a prepo- sition; hamzated verbs of both faʿala and afʿala patterns; and verbs which can be either transitive or intransitive. The risāla also includes some patterns of nouns (asmāʾ) and verbal nouns (maṣādir).

Ǧawālīqī, in the sixth/twelfth century, seems to be heavily dependent on Aṣmaʿī, given that on the margin of the first page of the book’s manuscript, it is noted that words ascribed to Aṣmaʿī are marked with the letter ṣ.1114 This is an unmistakable example of how dependent late philologists were in certain genres on much earlier authors, particularly from the period of data collection. The book is restricted to the type in which faʿala and afʿala agree in mean- ing. Its alphabetical arrangement does not go beyond the first letter, and as in Zaǧǧāǧ’s book, hamza is placed after hāʾ, but before yāʾ.

10.7 Verbal Patterns in General Ibn al-Qūṭiyya’s (d. 367/977) K. al-Afʿāl, which deals exclusively with the first and fourth patterns faʿala and afʿala (see 10.6 above), gave rise to two com- prehensive lexica which include the various verbal patterns, and not merely faʿala (or its counterparts faʿila and faʿula) and afʿala. Both lexica are titled K. al-Afʿāl and their authors are also Andalusian scholars, namely, Saraqusṭī (d. after 400/1010) and Ibn al-Qaṭṭāʿ (d. 515/1121). Verbal patterns other than faʿala and af ʿala are sporadically discussed in earlier mubawwab lexica, such as Abū ʿ Ubayd’s (d. 224/838) al-Ġarīb al-muṣannaf, which includes two short chap- ters on fāʿala,1115 and Ibn Qutayba’s (d. 276/889) Adab al-kātib, which includes a chapter on the reflexive pattern infaʿala and several others on the meanings associated with verbal patterns.1116 From a later period, Ibn Sīda’s (d. 458/1066) al-Muḫaṣṣaṣ includes extensive material on several verbal patterns.1117

Saraqusṭī begins his book by praising Ibn al-Qūṭiyya’s work as the best in its genre.1118 Yet he points out that the brevity of the work and the absence of the explanation of many of its verbs made its use uniquely difficult. Moreover, he

1113 Ibn Fāris, Tamām 18–20. 1114 Ǧawālīqī, Faʿaltu 27 (cf. editor’s introduction, 12). 1115 Abū ʿUbayd, Ġarīb II, 600–601. 1116 Ibn Qutayba, Adab 352–353, 354–363. 1117 Ibn Sīda, Muḫaṣṣaṣ XIV, 173–184. 1118 Saraqusṭī, Afʿāl I, 52.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

259MUBAWWAB AND SPECIALIZED LEXICA

finds fault in Ibn al-Qūṭiyya’s restriction of the material to faʿala and afʿala, to the exclusion of other verbal patterns. With the inclusion of these patterns, Saraqusṭī’s own book represents a major step forward in lexica devoted to verbs, although there are earlier works which include these patterns but which are not restricted to verbs, such as Fārābī’s (d. 350/961) Dīwān al-adab (see 10.8 below). As far as arrangement is concerned, however, Saraqusṭī’s book is as difficult to use as Ibn al-Qūṭiyya’s, although he differs with him on certain aspects. Instead of dividing the book into major divisions based on the pat- terns (as did Ibn al-Qūṭiyya), Saraqusṭī divides it into twenty-eight chapters in line with the letters of the alphabet, and lists patterns within each chapter. The order of the letters follows the phonetic principle, but instead of adopting Ibn al-Qūṭiyya’s unique order, Saraqusṭī follows an order which largely resembles that of Sībawayhi’s, i.e. ʾ, h, ʿ, ḥ, ḫ, ġ, q, k, ḍ, ǧ, š, l, r, n, ṭ, d, t, ṣ, z, s, ẓ, ḏ, ṯ, f, b, m, w, y.1119 Each chapter is then divided into the following four sections: faʿala and afʿala with the same meaning; faʿala and afʿala with different meanings; faʿala (i.e. with no afʿala counterpart); and afʿala (with no faʿala counterpart), followed by the rest of the verbal patterns. In the first three sections, verbs of the faʿala pattern are further divided, as in Ibn al-Qūṭiyya’s book, into muḍāʿaf, ṣaḥīḥ, mahmūz and muʿtall (i.e. unsound, comprising both aǧwaf and nāqiṣ), and are arranged according to the vowel of their second radical and whether they have one or more forms (e.g. faʿala; faʿala and faʿila; faʿala, faʿula and faʿila; faʿala and faʿula; faʿula; and faʿila).1120 In the fourth section the patterns are normally arranged in the following order: afʿala, faʿlala, tafaʿlala, faʿʿala, tafaʿʿala, if ʿalalla, if ʿanlala, if ʿawʿala, faʿwala, fayʿala, istafʿala, infaʿala, iftaʿala, fāʿala, fawʿala, tafawʿala, tafāʿala, if ʿanlā, and ifwaʿʿala, with further subdivi- sions for the muḍāʿaf, mahmūz and muʿtall. Although Saraqusṭī has expanded Ibn al-Qūṭiyya⁠ʾs material on faʿala and afʿala – either by adding new verbs or by providing more explanation and šawāhid for already existing ones – and incorporated all other verbal patterns, it remains that his book, like his prede- cessor’s, is extremely difficult to use, particularly as its alphabetical arrange- ment likewise does not go beyond the first letter, and that some verbs are cited under several headings.1121

More than a century after Saraquṣṭī, Ibn al-Qaṭṭāʿ also authored his K. al-Afʿāl on the basis of Ibn al-Qūṭiyya’s book. In recognition of Ibn al-Qūṭiyya, he refers to his own book in the introduction as Tahḏīb Kitāb al-Afʿāl, and praises his

1119 Ibid., I, 54; cf. Sībawayhi, Kitāb IV, 431. 1120 See this sequence in the chapter on ġayn (I, 258, 276, 294, 297, 298, 299). 1121 For example, Wadġīrī (1984: 107) lists thirty-six triliteral verbs which begin with rāʾ and are

cited in more than one location of the lexicon.

 

 

260 chapter 2

predecessor for having surpassed all previous scholars who wrote on verbal patterns.1122 This notwithstanding, he criticizes Ibn al-Qūṭiyya’s book for the complexity of its arrangement, which resulted in the dispersal of the same verb throughout several chapters, and for the confinement of its material to faʿala and afʿala. In order to diminish the complexity of arrangement, Ibn al-Qaṭṭāʿ introduces several modifications to Ibn al-Qūṭiyya’s system. For instance, he annulls his predecessor’s division of the book into three sections (i.e. one for faʿala and afʿala of both types, a second for afʿala, and a third for faʿala), as well as the internal divisions based on the vowel of the second radical of the verb and on other considerations (as in faʿala, faʿila and faʿula, and hollow verbs with a medial wāw or yāʾ). He further lists in one location the various forms of each verb, instead of leaving them scattered as in Ibn al-Qūṭiyya’s book. He pre- serves, however, the distinction between ṣaḥīḥ, muḍāʿaf, mahmūz and muʿtall (though in this specific order in which ṣaḥīḥ verbs are listed first). He also sup- plements Ibn al-Qūṭiyya’s book by including reduplicated quadriliterals (e.g. dahdaha “to roll a stone down”),1123 and all the augmented forms. In each letter, these additions follow the material on faʿala and afʿala. Ibn al-Qaṭṭāʿ’s supple- mentary material is easily recognizable since he marks it with the letter ʿayn, whereas Ibn al-Qūṭiyya’s words are marked with a qāf.1124 His additions are obviously meant to be exhaustive in all sections and patterns, as he explains in the introduction (ʿalā l-tamām wa-l-kamāl),1125 although it is noted that these additions do not contain any šawāhid, unlike Saraqusṭī’s. Another major step towards simplification (which Ibn al-Qaṭṭāʿ strangely does not mention in the introduction) is that he abandons Ibn al-Qūṭiyya’s phonetic system in favor of an alphabetical arrangement, but based only on the first letter of the word.

Two other books, also authored by Andalusian scholars, deal with specific aspects of verbs. The vowels and forms of the imperfect verb are discussed by Abū Ǧaʿfar al-Lablī (d. 691/1291) in Buġyat al-āmāl fī maʿrifat mustaqbalāt al-afʿāl. Reference to the vowels of the second radical of the imperfect of tri- literals (e.g. ʿayn al-fiʿl) is sporadically made in earlier works on afʿāl,1126 and a number of multithematic works, such as Abū ʿUbayd’s (d. 224/838) al-Ġarīb al-muṣannaf and Ibn Qutayba’s (d. 276/889) Adab al-kātib, contain short chapters in which imperfect triliteral verbs are cited primarily in order to

1122 Ibn al-Qaṭṭāʿ, Afʿāl I, 3. 1123 Ibid., I, 374. 1124 Note that no such distinction is needed in the sections on patterns other than faʿala and

afʿala since these do not feature in Ibn al-Qūṭiyya’s book. 1125 Ibid., I, 5. 1126 E.g. Ibn al-Qūṭiyya, Afʿāl 1–3; Saraqusṭī, Afʿāl I, 57–61; Ibn al-Qaṭṭāʿ, Afʿāl I, 6–11.

 

 

261MUBAWWAB AND SPECIALIZED LEXICA

determine the vowel of their second radical.1127 As far as we are able to determine, however, Lablī’s book is the only one which is exclusively devoted to the imperfect. It is divided into two parts, one for triliteral verbs, and the other for augmented patterns. Although ʿayn al-fiʿl is regular (qiyāsī) in a few triliteral types, it mostly cannot be determined in triliterals by reference to any rule (hence samāʿī). Lablī often gives criteria which are helpful in deter- mining the vowel of the ʿayn in certain types of verbs that are not fully regu- lar (e.g. the occurrence of faṭha with verbs whose second or third radical is a guttural, and the distinction between transitive and intransitive muḍāʿaf verbs, normally by a ḍamma in the former and a kasra in the latter).1128 In augmented forms, Lablī notes that their imperfect is fully regular, and one might feel that this part of the book is quite redundant, were it not for the examples which the author cites for some rare patterns, such as faʿlala and fanʿala (in the perfect),1129 and which can be useful for students to whom the book is primarily addressed. Finally, it is noteworthy that Zabīdī (d. 1205/1790) makes use of Lablī’s book in his two muǧannas lexica, Tāǧ al-ʿarūs and al- Takmila wa-l-ḏayl wa-l-ṣila.1130

The other book is al-Ḥilya fī mā li-kull fiʿl min taṣrīf wa-binya, authored by Yūsuf b. Muḥammad b. ʿAntara in 660/1262.1131 It consists of fifty chapters, each of which deals with a specific type of verbs. These types are peculiarly arranged based on the number of their letters (i.e. root and augments, if any), e.g. ʿaḍḍa (two letters), waʿada (three letters), daḥraǧa and awdaʿa (four let- ters), izdaǧara (five letters), and istaġfara (6 letters).1132 Verbs of the last type, for example, occupy six chapters since six of their types are cited, as repre- sented by the verbs istaġfara, istawdaʿa, ista⁠ʾǧara, istaʿāna, istawlā and ista⁠ʾnā.1133 In each chapter Ibn ʿAntara typically gives the forms of the imperfect and imperative (both with the various pronouns), verbal noun(s), active partici- ple, and passive participle of the one verb which represents the type under discussion.

1127 Abū ʿUbayd, Ġarīb II, 601–609; Ibn Qutayba, Adab 367–373. 1128 Lablī, Buġya 33–36, 71–72. 1129 Ibid., 92–95. 1130 Zabīdī, Tāǧ I, 4, 25 and Takmila, Introduction I, 15. 1131 Ibn ʿ Antara (Ḥilya II, 365) mentions at the end of his book that he completed it in 660 A.H.

at the age of twenty-seven; hence his date of birth is 633/1235. 1132 Ibid., II, 65, 125, 289, 291, 325, 348. 1133 Ibid., II, 348–365.

 

 

262 chapter 2

10.8 Nominal and Verbal Patterns The following lexica embrace both nominal and verbal patterns:

1. Dīwān al-adab by Isḥāq b. Ibrāhīm al-Fārābī (d. 350/961). 2. al-Mūʿab by Ibn al-Tayyānī (d. 436/1044). 3. Abniyat al-asmāʾ wa-l-afʿāl wa-l-maṣādir by Ibn al-Qaṭṭāʿ (d. 515/1121). 4. Tāǧ al-maṣādir by Aḥmad b. ʿAlī al-Bayhaqī, known as Bū Ǧaʿfarak

(d. 544/1150). 5. Šams al-ʿulūm wa-dawāʾ kalām al-ʿArab min al-kulūm by Našwān b. Saʿīd

al-Ḥimyarī (d. 573/1178).

In his introduction, Fārābī proudly declares that the content and arrange- ment of his book are unprecedented (muštamilan ʿalā ta⁠ʾlīf lam usbaq ilayhi wa-sābiqan bi-taṣnīf lam uzāḥam ʿalayhi).1134 Indeed, he is the first author to combine nominal and verbal patterns in one book. As for his reference to taṣnīf, it can apply either to the subtle arrangement of the material (including the various patterns) or to the type of alphabetical arrangement based on the final letter (or perhaps to both). On the first count, Fārābī is a true innovator as far as our knowledge goes. As for arrangement, the rhyme system he adopts is the most interesting aspect of his lexicon, since it is a model previously used by Bandanīǧī (d. 284/897) in al-Taqfiya, but adopted later by many lexicographers for centuries to come.1135 Fārābī’s arrangement, however, is more developed than Bandanīǧī’s since it also takes the first and then the second radical into consideration. It is quite telling that the first author to adopt this system without any consideration for patterns is Fārābī’s own nephew, Ǧawharī (d. c. 400/1010) in al-Ṣaḥāḥ.1136 Fārābī’s lexicon is meticulously divided into sections and sub- sections in accordance with strict criteria which the author applies. Following an introduction in which he discusses various morphological aspects of nouns and verbs and explains that he opted for the alphabetical system, not Ḫalīl’s phonetic system, due to the former’s ease and currency among the ḫāṣṣa (elite or highly educated people) and the ʿāmma (common people, generality of people),1137 he divides the book into six sections, each of which is called kitāb. These sections are: sālim (sound; synonym of ṣaḥīḥ), muḍāʿaf (reduplicated), miṯāl (having wāw or yāʾ as first radical), ḏawāt al-ṯalāṯa (his term for “hollow”),

1134 Fārābī, Dīwān I, 72–73. 1135 See below, 362–364, 370–372 (including our comments on Ibn al-Sikkīt’s Iṣlāḥ al-manṭiq). 1136 For the difference between Fārābī and Ǧawharī in arranging verbs with final weak

radicals, see below, 375. 1137 Fārābī, Dīwān I, 87.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

263MUBAWWAB AND SPECIALIZED LEXICA

ḏawāt al-arbaʿa (his term for finally weak forms), and mahmūz (hamzated). Each kitāb is then divided into a subsection for nouns, followed by one for verbs (note that verbal nouns are only mentioned if they are irregular).1138 In turn, each of the two subsections is divided into chapters which are alphabeti- cally arranged according to the final radical of the root, and in which triliteral forms precede augmented ones. In the case of sound verbs, for example, the patterns are arranged in the following order: faʿala, afʿala, faʿʿala, fāʿala, iftaʿala, infaʿala, istafʿala, tafaʿʿala, tafāʿala, if ʿalla, if ʿālla, and quadriliterals and forms that are appended to them (mulḥaq).1139 The same order is strictly adhered to in other types of verbs (e.g. muḍāʿaf, etc.). Even vowels and augments are taken into consideration in the arrangement of words, as in faʿala, faʿula and faʿila (i.e. fatḥa, ḍamma and kasra in this order), faʿl and faʿal (i.e sukūn before vowel), faʿlā and faʿlāʾ (i.e. -ā before -āʾ in feminine markers), and faʿlāʾ and faʿlān (i.e. -āʾ before -ān).1140 Other subsidiary rules of arrangement are applied in several parts of the book, such as those which deal with miṯāl and mahmūz.1141

Little is known about the second book in the above list, Ibn al-Tayyānī’s al-Mūʿab (also often referred to as Tanqīḥ/Talqīḥ al-ʿAyn). In a text preserved by Suyūṭī (d. 911/1505), Abū l-Ḥasan al-Šārrī (d. 649/1251) explains that, in spite of its valuableness, the lexicon was not truly appreciated and thus only few copies of it existed.1142 In the modern period, one copy was acquired by Père Anastase-Marie al-Karmalī, who published a short specimen of it in 1914.1143 Unfortunately, the fate of the manuscript is not known,1144 and, apart from a few references to it in later lexica,1145 our knowledge of its content is restricted to the sample published by Karmalī. Although very concise, this sample is crucial for correcting a widespread misconception about the book, originat- ing from Šārrī’s text quoted in al-Muzhir. Since Šārrī mentioned that al-Mūʿab was based on Ḫalīl’s (d. 175/791) al-ʿAyn, it was generally assumed that it fol- lowed Ḫalīl’s arrangement, based on points of articulation and permutations

1138 Ibid., I, 90. 1139 Ibid., II, 98, 279, 338, 381, 394, 421, 428, 437, 466, 473, 475–476; cf. Fārābī’s introduction I,

77–78. For ilḥāq, see above, 60, n. 296. 1140 Ibid., I, 87. 1141 For a more detailed description of arrangement, see ʿUmar (1995: 49–54); cf. ʿUmar’s

introduction to Fārābī, Dīwān I, 10–16. 1142 Suyūṭī, Muzhir I, 88–89. 1143 Karmalī (1914: 5–14). 1144 Qazzāz (1981: 30); Wadġīrī (1984: 64). 1145 Ǧubūrī (2004: 77–80).

 

 

264 chapter 2

of roots.1146 In fact, Ibn al-Tayyānī adopts an alphabetical arrangement based on the final letters of the roots and takes no account of root permutation. As far as content is concerned, the book is divided according to the various verbal and nominal patterns. The sample published by Karmalī deals with verbs of the pattern faʿala whose imperfect is of the pattern yafʿilu, and lists the following roots consecutively: tbb, ḥbb, dbb, zbb, šbb, ḍbb, ġbb, nbb, hbb, btt, štt, ktt, dṯṯ, rṯṯ, ġṯṯ, mṯṯ, nṯṯ, ṯǧǧ, dǧǧ, šǧǧ, etc. According to Karmalī, the book also contains lexi- cal items and poetry šawāhid which are not found in some of the most compre- hensive lexica. Furthermore, Šārrī mentions that Ibn al-Tayyānī included in his book the trustworthy material of al-ʿAyn, to the exclusion of doubtful šawāhid, words and patterns, as well as material taken from Ibn Durayd’s (d. 321/933) al-Ǧamhara. It is not possible however from the available sample to establish this fact, and there may be reason to believe that the manuscript acquired by Karmalī is incomplete since it consists of no more than 124 pages.

Ibn al-Qaṭṭāʿ, whose K. al-Afʿāl was discussed in 10.7 above, is also author of a more general lexicon, K. Abniyat al-asmāʾ wa-l-afʿāl wa-l-maṣādir. Although the book combines nominal patterns (i.e. nouns and verbal nouns) and verbal ones, its bulk consists of nouns (asmāʾ). It does not contain as many subdi- visions as Fārābī’s book (which is mentioned among his sources), but rather chapters which do not feature in it, such as those on particles (ḥurūf ), active and passive participles (asmāʾ al-fāʿilīn wa-l-mafʿūlīn), plurals, and even Arabized words.1147 It begins with a useful introduction on nominal and verbal patterns, particularly with respect to the minimum and maximum number of letters in both types as well as in particles. Also included in the introduction is a discussion of augments (zawāʾid) and substitution (ibdāl). Nouns are divided into biliterals, triliterals, quadriliterals and quinqueliterals. In these, augments are not counted, hence ḫiṣṣīṣāʾ (particularity) and ṭurṭubba (a woman having long breasts) are considered biliteral and triliteral respectively.1148 Verbs on the other hand are divided into three triliteral patterns (i.e. faʿala, faʿila and faʿula); their augmented derivatives (i.e. faʿʿala, tafāʿala, tafaʿʿala, infaʿala, iftaʿala, etc.); one quadriliteral pattern (i.e. faʿlala); and its augmented derivatives (i.e. tafaʿlala, if ʿanlala, if ʿalalla, if ʿallala, etc.).1149 The advantage of this division is that it limits to four the number of basic patterns by assuming that the other patterns are subsumed under them. This is essentially the same as Fārābī’s

1146 Cf. Wadġīrī (1984: 62–64), where he points out that he and earlier researchers fell victim to this error.

1147 Ibn al-Qaṭṭāʿ, Abniya 98, 116, 170, 199, 260, 315, 319, 320. 1148 Ibid., 110, 133. 1149 Ibid., 324 ff., 339 ff.

 

 

265MUBAWWAB AND SPECIALIZED LEXICA

division, but the latter’s is somewhat obscured by the numerous subdivisions which he observes. As for verbal nouns, Ibn al-Qaṭṭāʿ lists them in the same order as that of the corresponding verbs.1150 Since only very few examples of each verbal and nominal pattern are cited (one to three in most cases), they are not arranged alphabetically.

The same division of verbs is adopted by Bū Ǧaʿfarak in Tāǧ al-maṣādir. Although the book deals primarily with verbs, its author’s focus on ver- bal nouns, as highlighted in the title, is evident. Its introduction, moreover, includes several nominal patterns, such as active and passive participles, nouns of intensiveness (mubālaġa), nouns of time and place, instrumen- tal nouns, etc. A comparison between this book and Fārābī’s Dīwān al-adab shows Bū Ǧaʿfarak’s great dependence on his predecessor in both material and arrangement.1151 Mention can also be made of a book by Bū Ǧaʿfarak’s contem- porary, the famous Muʿtazilite scholar and philologist, Zamaḫšarī (d. 538/1144). The book’s title is Muqaddimat al-adab, and its Arabic-Persian text contains three sections: nouns, verbs and particles.1152 Only in the section on verbs does Zamaḫšarī arrange his material on the basis of abniya. The main divisions, i.e. triliterals, augmented triliterals, and quadriliterals, are subdivided into sections based on the verb’s vowels (e.g. faʿala/yafʿilu; faʿala/yafʿulu) or type (e.g. sound, reduplicated, etc.), and words are arranged on the basis of the last radical, as in Dīwān al-adab.

Našwān b. Saʿīd al-Ḥimyarī’s voluminous book, Šams al-ʿulūm, is the most extensive work of its kind.1153 It begins with a general introduction which deals mostly with morphological and morphophonological issues, such as augments (zawāʾid), substitution (ibdāl), omission (ḥaḏf ), the minimum and maximum number of letters in the three parts of speech, etc. Also included in the intro- duction is a chapter on the points of articulation. The lexical material is then divided in a manner that differs in part from that of previous books on pat- terns. In the introduction, Našwān explains that he intended to arrange his lexicon in a manner that would guard it against error in diacritics as well as in vowels. He therefore divided it into sections (kutub) from alif to yāʾ according to the first radical, and then arranged each kitāb (other than in the muḍāʿaf )

1150 Ibid., 370–383; see also 152. 1151 See the comparison made by ʿUmar (1995: 184–185) between the two books. Bū Ǧaʿfarak

also authored an Arabic-Persian lexicon, of which there is an Indian lithographic edition, 1320 A.H. This carries the same title as the purely Arabic version and agrees with it in arrangement and material; cf. ʿUmar (1995: 185–186).

1152 Zamaḫšarī, Muqaddima 3–85, 86–286, 287–288; cf. ʿUmar (1995: 195–200). 1153 On Našwān and his contribution to philological and grammatical study, see Nāǧī (2004).

 

 

266 chapter 2

into chapters (abwāb) which are based on the second radical, but which also take the other radicals into account in the arrangement.1154 According to him, this would ensure that words are listed under their correct letters, and it is implied that were the main divisions of the book based on patterns, rather than letters of the alphabet, diacritical marks would more likely be confused. Each kitāb is divided into two subsections, one for muḍāʿaf (e.g. ʾbb, ʾdd, etc.) and another for other patterns, that is, patterns with no reduplicated radicals (e.g. words whose first two radicals are hamza and bāʾ, then hamza and tāʾ, etc.). These subsections are further divided into two parts, one for nouns, fol- lowed by another for verbs, and, in each part, the method of arranging the material is largely similar to that in Fārābī’s book (e.g. unaugmented forms pre- cede augmented ones, and a special sequence is observed in both nominal and verbal patterns). In alphabetical arrangement, Našwān differs from Fārābī, for whereas the latter adopts the rhyme system, Našwān applies full alphabetical order to roots of words in each pattern. Moreover, the book, unlike any other in its genre, is replete with information not directly related to its theme of abniya, with an obvious bias towards Yemeni culture. In fact, Našwān tells his reader in the introduction that he intends to include in the book information related to kings, dignitaries, usefulness of trees, constitution of stones, Qurʾānic sciences and exegesis, anecdotal material, genealogies, calculation, judicial decisions, interpretation of dreams, and astronomy.1155

11 Multithematic Works

In almost all previous sections of this chapter, reference was made to the contribution of multithematic works to mubawwab lexica and vice versa. Accordingly, much of the material that belongs to this section has already been discussed earlier. We shall therefore herein briefly examine the content of the most important multithematic sources, including those which have not been frequently cited before.

Contrary to the assumption that monographs which deal with a single nar- rowly defined topic represent a stage prior to that of multithematic works,1156 both types appeared during the same period, i.e. as of the second half of the second/eighth century, and coexisted for several centuries. Both types were also among the sources from which authors of muǧannas lexica derived their

1154 Našwān, Šams I, 34. 1155 Ibid., I, 36–38. 1156 Amīn (1938: II, 263 ff.); cf. above, 46.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

267MUBAWWAB AND SPECIALIZED LEXICA

material. Books titled al-Ṣifāt are attributed to some of the earliest scholars in the period of data collection. The use of the plural form ṣifāt may well be a reflection of the fact that these books contained various topics, given that in the titles of some single-topic works the singular ṣifa is used, as in Ibn al-Aʿrābī’s (d. 231/845) Ṣifat al-naḫl and Ṣifat al-zarʿ (but cf. Ṣifāt al-ġanam by al-Aḫfaš al-Awsaṭ [d. 215/830]).1157 The earliest book by the title al-Ṣifāt is the one ascribed to Nahšal b. Zayd, better known as Abū Ḫayra al-Aʿrābī, a con- temporary of Abū ʿAmr b. al-ʿAlāʾ’s (d. 154/770).1158 Another early work by the same title is attributed to al-Naḍr b. Šumayl (d. 203/819), and, thanks to the description of its content by Ibn al-Nadīm (d. 380/990),1159 we know that it consists of five parts which cover a variety of topics, such as human body, gen- erosity, attributes of women, abodes, camels, clouds, birds, sun, moon, wells, wine, plants and rain. Other early authors who reportedly authored a book titled al-Ṣifāt include Quṭrub (d. 206/821), Abū Zayd al-Anṣārī (d. 215/830), and Aṣmaʿī (d. 216/831).1160 In the fourth/tenth century, both Luġda al-Iṣfahānī (d. 310/922)1161 and al-Qāsim b. Muḥammad al-Daymartī (d. c. 375/985)1162 are reported to have authored a book by the same title. The term muṣannaf was also used in the titles of multithematic books to indicate that they comprise various topics. Other than Abū ʿUbayd’s (d. 224/838) al-Ġarīb al-muṣannaf (for which see below), two early books by the same title are ascribed to al-Qāsim b. Maʿn (d. 175/791)1163 and Abū ʿAmr al-Šaybānī (d. 206/821).1164 As for al-Muṣannaf al-ġarīb, which Yāqūt ascribes to Quṭrub,1165 a comparison between the latter’s list of Quṭrub’s work and that of Ibn al-Nadīm’s1166 strongly suggests that it is no other than K. al-Ṣifāt mentioned by Ibn al-Nadīm.

Before listing the most important extant works of the genre, it should be noted that some of the books mentioned in earlier sections do qualify for inclusion under multithematic works as well. One example is Kurāʿ al-Naml’s (d. 310/922) al-Munaǧǧad, which was discussed with works on muštarak.1167 The thematic division of the book brings it close to multithematic works,

1157 Ibn al-Nadīm, Fihrist 76, 58; Yāqūt, Muʿǧam VI, 2533; III, 1376. 1158 Azharī, Tahḏīb I, 33; Qifṭī, Inbāh I, 144. 1159 Ibn al-Nadīm, Fihrist 57–58. 1160 Ibid., 58, 61; Azharī, Tahḏīb I, 12, 32; Qifṭī, Inbāh I, 143; II, 35; III, 220. 1161 Ibn al-Nadīm, Fihrist 89; ʿAskarī, Talḫīṣ I, 3; Qifṭī, Inbāh III, 43; Yāqūt, Muʿǧam II, 874. 1162 ʿAskarī, Talḫīṣ I, 3; Yāqūt, Muʿǧam V, 2230. 1163 Yāqūt, Muʿǧam V, 2231; Suyūṭī, Buġya II, 263. 1164 Ibn al-Nadīm, Fihrist 75; Yāqūt, Muʿǧam II, 627; Suyūṭī, Buġya I, 440. 1165 Yāqūt, Muʿǧam VI, 2647. 1166 Ibn al-Nadīm, Fihrist 58. 1167 See above, 201.

 

 

268 chapter 2

but since its author’s primary occupation is with homonymous polysemic words, its inclusion under that genre makes better sense. Another example is Ibn al-Sikkīt’s (d. 244/858) al-Alfāẓ, which has a more elaborate thematic arrangement than al-Munaǧǧad, but which, because of the prominence of its mutarādif material, was discussed earlier with works on synonyms.1168 As far as scope is concerned, the works to be discussed below vary considerably, for whereas the authors of some of the more comprehensive thesauri try to cover all Arabic vocabulary, others settle for the much more modest target of includ- ing sample chapters or words related to them. Moreover, there are multithe- matic works – such as Ibn Qutayba’s (d. 276/889) ʿUyūn al-aḫbār and al-Maʿānī l-kabīr, and Abū Hilāl al-ʿAskarī’s (d. after 395/1005) Dīwān al-maʿānī – which are essentially works on literature or collections of literary šawāhid and thus fall beyond the present scope. Hence, we are left with the following titles:

1. al-Ġarīb al-muṣannaf by Abū ʿUbayd al-Qāsim b. Sallām (d. 224/838). 2. Adab al-kātib by Ibn Qutayba (d. 276/889) and al-Ǧarāṯīm, which is attrib-

uted to him. 3. al-Muntaḫab min ġarīb kalām al-ʿArab by Kurāʿ al-Naml (d. 310/922). 4. al-Zīna fī l-kalimāt al-Islāmiyya al-ʿArabiyya by Abū Ḥātim al-Rāzī

(d. 322/934). 5. Ṣināʿat al-kuttāb by Abū Ǧaʿfar al-Naḥḥās (d. 338/950). 6. al-Samāʾ wa-l-ʿālam by Abū ʿAbdallāh Muḥammad b. Abān b. Sayyid b.

Abān al-Laḫmī (d. 354/965). 7. al-Talḫīṣ fī maʿrifat asmāʾ al-ašyāʾ by Abū Hilāl al-ʿAskarī (d. after 395/1005). 8. Mabādiʾ al-luġa by al-Ḫaṭīb al-Iskāfī (d. 420/1029). 9. Fiqh al-luġa by Ṯaʿālibī (d. 429/1039). 10. al-Muḫaṣṣaṣ by Ibn Sīda (d. 458/1066). 11. Niẓām al-ġarīb by ʿĪsā b. Ibrāhīm al-Rabaʿī (d. 480/1087). 12. Kifāyat al-mutaḥaffiẓ wa-nihāyat al-mutalaffiẓ by Ibn al-Aǧdābī

(d. c. 650/1252).

Although al-Ġarīb al-muṣannaf is the first extant work in the genre, it is obvious that its author draws on material cited by earlier or contemporary scholars. Among those whom he acknowledges are Kisāʾī (d. 189/805), ʿAlī b. al-Mubārak al-Aḥmar (d. 194/810), Yaḥyā b. al-Mubārak al-Yazīdī (d. 202/818), ʿAbdallāh b. Saʿīd al-Umawī (d. c. 203/819), Abū ʿAmr al-Šaybānī (d. 206/821), Farrāʾ (d. 207/822), Abū ʿUbayda (d. 209/824), Abū Zayd al-Anṣārī (d. 215/ 830), Aṣmaʿī (d. 216/831), and Ibn al-Aʿrābī (d. 231/845). In fact, there is a striking

1168 See above, 205.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

269MUBAWWAB AND SPECIALIZED LEXICA

resemblance between certain sections of Abū ʿUbayd’s book and monographs authored by some of these scholars. We have pointed out, for example, the great affinity between K. al-Naḫl attributed to Aṣmaʿī, and Abū ʿUbayd’s mate- rial on the subject.1169 Another example is that, in the section on human body,1170 Abū ʿUbayd makes frequent reference to several scholars who have authored monographs of the same genre, including Šaybānī, Abū ʿUbayda, Abū Zayd, Aṣmaʿī and Ibn al-Aʿrābī, although he mentions none of their books by title. Furthermore, much of the material of al-Ġarīb al-muṣannaf is derived from the Aʿrāb, or the Bedouin fuṣaḥāʾ who were usually consulted by early philolo- gists, and who in some cases were authors in their own right. These include Abū l-Ǧarrāḥ al-ʿUqaylī, Abū l-Ḥasan al-Aʿrābī, Abū Ziyād al-Kilābī, Abū Šanbal al-Aʿrābī, Abū ʿAlqama al-Ṯaqafī, Abū Faqʿas al-Aʿrābī, Abū l-Qaʿqāʿ al-Yaškurī, Abū l-Walīd al-Kilābī, and al-ʿAdabbas al-Kinānī. Abū ʿUbayd’s reliance on these fuṣaḥāʾ and his frequent citing of dialectal material (cf. the report that he spent forty years collecting data directly from the Arabs for inclusion in his book)1171 are certainly related to the strength of the component of ġarīb in his book, as is also reflected in the inclusion of the term ġarīb in its title. This corroborates our view that some multithematic works can be classified with books on ġarīb as well.1172

al-Ġarīb al-muṣannaf consists of twenty-six sections, each of which is called kitāb.1173 These are elaborately divided into about nine hundred chapters, of which 179 occur in the final kitāb (which contains a miscellany of data – hence its title al-Aǧnās) and are mostly very short.1174 The sections are of two types, the first of which may be referred to as the general type, and contains classified vocabulary related to themes such as human body (including human traits, groups of people, kinship, etc.), women, clothes, food, sicknesses, wine, abodes, horses, weapons, birds and pests, utensils, mountains, plants, water, camels, sheep, and wild animals. The internal arrangement of these sections, the chap- ters included under them, and the words cited therein, is largely arbitrary. The second type deals with purely philological topics not related to themes of the previous general type, such as phonological issues, nominal patterns,

1169 See Aṣmaʿī, Naḫl 69–71 and Abū ʿUbayd, Ġarīb II, 486–490; cf. above, 137. 1170 Abū ʿUbayd, Ġarīb I, 29–133. 1171 Cf. above, 17, 74. 1172 Cf. above, 85. 1173 Note that ʿAbdaltawwāb’s (1980: 259) list cites the titles of twenty-five sections to the

exclusion of the one on ḫamr (wine); cf. Abū ʿUbayd, Ġarīb I, 241–260. See also a book- length study of Abū ʿUbayd’s book in ʿAbdaltawwāb (1962).

1174 Abū ʿUbayd, Ġarīb III, 924–1011.

 

 

270 chapter 2

verbal patterns, aḍdād and Arabized words. Following the section on aḍdād is a variety of chapters which belong to either of the two types.1175 Among the philological topics in these chapters are maqlūb (words formed by metathesis), itbāʿ/ittibāʿ (alliteration), masculine and feminine, words with dialectal vari- ants, Arabized words, hamz, and verbal nouns. The whole book, moreover, is characterized by brevity in explaining words and relative scarcity of šawāhid in comparison to the vast number of words it contains (a total of 17,970, accord- ing to Zubaydī).1176 The comprehensive nature of the book and its elaborate arrangement commanded great respect from later scholars, and several com- mentaries on it are mentioned in the sources.1177 In this respect, it is probable that a number of sections of the book were circulated as independent mono- graphs, given that they were referred to as such by other authors.1178

Ibn Qutayba’s Adab al-kātib is one of the most famous works in the philo- logical tradition as a whole, and numerous scholars have written commentar- ies on it.1179 It resembles al-Ġarīb al-muṣannaf in that it includes, in addition to vocabulary related to general themes, chapters which are devoted to a number of philological topics. Yet its arrangement is different since it is divided into four sections or kutub, each of which includes several related topics and addresses a specific aspect of the skills which writers are supposed to master. The first section, Kitāb al-Maʿrifa (Book of Knowledge), begins with a number of chap- ters related to words and expressions that are often misused by writers, but its bulk consists of several thematically arranged chapters which deal with topics such as meanings of proper nouns, sky, stars, plants, horses, human body, food, instruments, clothes, weapons, insects, metals, etc. In contrast, the other sec- tions do not include themes of this general type. Accordingly, the second sec- tion, Kitāb Taqwīm al-yad (Book of Amendment of Script), discusses matters related to handwriting, such as conjunctive and disjunctive alif (alif al-waṣl/ al-qaṭʿ), the writing of mā enclitically or as a separate word (e.g. ḥayṯumā vs. ḥayṯu mā “wherever”), Qurʾānic words written with a wāw instead of alif (e.g. ṣalāt “prayer”), and words written with alif maqṣūra (normally referred to as yāʾ)

1175 Ibid., III, 647–831. 1176 Zubaydī, Ṭabaqāt 202; Qifṭī, Inbāh III, 21; Suyūṭī, Buġya II, 254 (where the number is

17,770)! 1177 Cf. Sezgin (1982: 83–84). Note, however, that Abū ʿUbayd was accused of error by some,

most notably ʿAlī b. Ḥamza (d. 375/985) in Tanbīhāt 189–273. 1178 For example, Suyūṭī (Muzhir I, 581; II, 249) mentions K. al-Aḍdād (cf. above, 195), and Ibn

Manẓūr (Lisān: QZZ, QQZ) refers to K. Mā ḫālafat fīhi l-ʿāmma luġāt al-ʿArab (cf. above, 172, n. 587). For K. al-Silāḥ, see above, 159. See also Sezgin (1982: 83).

1179 Cf. Sezgin (1984: 156–157).

 

 

271MUBAWWAB AND SPECIALIZED LEXICA

or alif mamdūda (i.e. ordinary alif ). This part of Ibn Qutayba’s book probably influenced later authors who discussed similar orthographical issues.1180 The third section, Kitāb Taqwīm al-lisān (Book of Amendment of Speech), essen- tially belongs to the genre laḥn al-ʿāmma, and has been mentioned earlier,1181 whereas the fourth, Kitāb al-Abniya (Book of Patterns), deals mostly with ver- bal and nominal patterns (abniyat al-afʿāl/al-asmāʾ), but includes a few other topics, such as metathesis and Arabized words.1182

As for K. al-Ǧarāṯīm, its attribution to Ibn Qutayba on the title page of the manuscript is almost certainly erroneous since the biographical and biblio- graphical sources do not mention such a title among Ibn Qutayba’s works. Furthermore, the real author quotes Ibn Qutayba on one occasion, and the quotation indeed occurs in both Adab al-kātib and ʿUyūn al-aḫbār.1183 The book’s author seems to be a certain Anas, whose name is mentioned seven times in the text.1184 All we know about this figure is that he was a contempo- rary of Nifṭawayhi’s (d. 323/935), since he himself mentions an encounter with the latter.1185 The fact that the book is strikingly similar in content to al-Ġarīb al-muṣannaf yet with better organized themes suggests that the author rear- ranged some of Abū ʿUbayd’s material by bringing together chapters which occur in disparate parts of al-Ġarīb in spite of the similarity of their content.1186 Moreover, the author reduced the number of Abū ʿUbayd’s šawāhid and omit- ted the names of the scholars to whom the material is ascribed. He also omit- ted the numerous chapters of the second type referred to above (i.e. purely

1180 As, for example, in the previously mentioned Ḏikr mā yūqaf ʿ alayhi bi-l-hāʾ wa-l-yāʾ fī Kitāb Allāh by Abū Bakr b. al-Anbārī (d. 328/940), ʿUqūd al-hamz by Ibn Ǧinnī (d. 392/1002), and Bāb al-hiǧāʾ by Ibn al-Dahhān (d. 569/1174); cf. above, 218–219. Also to be noted is that some grammatical works include (normally at the end) a chapter on orthography, called al-ḫaṭṭ or al-hiǧāʾ, which deals with matters similar to those discussed by Ibn Qutayba. See, for example, Astarābāḏī (d. c. 686/1287), Šarḥ III, 312–333 and Suyūṭī (d. 911/1505), Hamʿ II, 231–244. Ibn Mālik’s (d. 672/1274) last chapter of Tashīl al-fawāʾid (332–338) also deals with orthography. Expanded by Abū Ḥayyān al-Andalusī (d. 745/1344) in al-Taḏyīl wa-l-takmīl, this chapter has also been published by Turkī b. Sahw b. Nazzāl al-ʿUtaybī in an independent volume titled al-Hiǧāʾ: Āḫir abwāb al-Taḏyīl wa-l-takmīl.

1181 See above, 176. 1182 Ibn Qutayba, Adab 381–383, 383–390. 1183 Ibn Qutayba, Ǧarāṯīm II, 284; Adab 167; ʿUyūn II, 98. 1184 Ibn Qutayba, Ǧarāṯīm I, 264, 325; II, 86, 99, 138, 285, 327. 1185 Ibid., II, 99; see also the editor’s introduction I, 37–57. 1186 See examples of the rearrangement of Abū ʿUbayd’s material in the editor’s introduction

to Ibn Qutayba, Ǧarāṯīm I, 45–46.

 

 

272 chapter 2

philological topics), except for two chapters on rare verbs and rare nouns,1187 and a third which deals with rhyme, poetry and prosody1188 – all of which are placed at the end of the book, separate from the general themes. On the other hand, the author supplements Abū ʿUbayd’s material in certain cases. For example, he adds material on animals derived from Ǧāḥiẓ’s (d. 255/869) K. al-Ḥayawān,1189 and incorporates into his book Abū Ḥātim al-Siǧistānī’s (d. 255/869) K. al-Karm (including the section on ḫamr “wine”) in its entirety.1190

In line with its title and its author’s interest in ġarīb (cf. also his al-Muǧarrad fī ġarīb kalām al-ʿArab wa-luġātihā), Kurāʿ al-Naml’s al-Muntaḫab min ġarīb kalām al-ʿArab was discussed with works on ġarīb.1191 But, as previously noted, it is essentially a multithematic lexicon of the type discussed in this section. It consists of 335 chapters which cover, as does al-Ġarīb al-muṣannaf and al-Ǧarāṯīm, general themes as well as philological topics. The latter type occurs in the second part which deals with nominal and verbal patterns, as well as in most of the third. Kurāʿ obviously incorporated into his book material from earlier thesauri, most notably Ibn al-Sikkīt’s (d. 244/858) Iṣlāḥ al-manṭiq. Yet it is evident that he omitted much of their šawāhid, particularly in the first two parts of the book, which occupy more than two-thirds of it and hardly contain any šawāhid.

K. al-Zīna fī l-kalimāt al-Islāmiyya al-ʿArabiyya by Rāzī is significantly differ- ent from other multithematic works since it is devoted to what he calls Islamic words; that is, words which occur in the Qurʾān or have a specific meaning in an Islamic context. Its inclusion with multithematic works, however, is supported by the arrangement of its material into various topics, which Rāzī details in his introduction.1192 Following two chapters in which are highlighted the merits of Arabic language and poetry, the third sets the pace for the rest of the book. It alerts the reader to the need of knowing the meanings of Islamic words and attributes of God, introduces him to foreign words in the Qurʾān, and cites expressions which the author believes were used for the first time in the Qurʾān.1193 The rest of the published parts of the book (excluding parts which deal with Islamic sects) consists of the two types of chapters. In the first

1187 Ibid., II, 301–322. 1188 Ibid., II, 323–337. 1189 Cf. ibid., II, 245–251 and Ǧāḥiẓ, Ḥayawān VII, 71, 86, 123–124, 128–129, 137–138, 192–196,

226–232, 241–243, 250–251. 1190 Ibn Qutayba, Ǧarāṯīm II, 83–114 and Abū Ḥātim, Karm 73–94. 1191 See above, 89–90. 1192 Rāzī, Zīna 67–69. 1193 Ibid., 134–157.

 

 

273MUBAWWAB AND SPECIALIZED LEXICA

Rāzī focuses on a specific attribute of God – such as al-ṣamad (the Eternal), al-muṣawwir (the Modeller), al-ʿazīz (the Mighty), al-ḥakīm (the Wise), etc. – and discusses its various meanings, based on its derivation or morphological pattern and the contexts in which it is used. The other type deals with specific Islamic notions – such as ḫalq (creation), qadar (destiny), ʿarš (throne), ǧanna (paradise), ṯawāb (retribution) and qiyāma (resurrection).

Similar to Adab al-kātib by Ibn Qutayba, Ṣināʿat al-kuttāb by Abū Ǧaʿfar al-Naḥḥās focuses on the skills which the writer should master. Following a lengthy introduction,1194 the book consists of eleven sections, each of which begins with a synopsis of its content and is devoted to a specific topic. The main topics are: the expression bi-sm Allāh al-raḥmān al-raḥīm (In the name of God, the Compassionate, the Merciful), letters of the alphabet, names of days and months, writing tools and conventions, selected issues in grammar and rhetoric, and famous orations. The tenth section embraces several largely unrelated topics, such as faʿala and afʿala, other verbal patterns, adjectives of each of the first ten horses in a race, types of food, wounds, illnesses, teeth, day and night, etc.

The book titled al-Samāʾ wa-l-ʿālam is lost to us except for its third section (al-sifr al-ṯāliṯ). It is referred to in several sources as K. al-ʿĀlam and is ascribed to Abū l-Qāsim Aḥmad b. Abān b. Sayyid b. Abān al-Laḫmī (d. 382/992).1195 The real author, however, is almost certainly Abū l-Qāsim’s brother, Abū ʿAbdallāh Muḥammad b. Abān (d. 354/965).1196 First of all, both the book and the full name of its author (i.e. Abū ʿAbdallāh Muḥammad, and not Abū l-Qāsim Aḥmad) are mentioned by Lablī (d. 691/1291) in his Buġyat al-āmāl, a book on imperfect verbs which we have previously mentioned.1197 More importantly, Muḥammad is unmistakably cited as author on the title page of the manuscript of the surviving third section, and the expression qāla Abū ʿAbdallāh often occurs in the text itself. As far as size and content are concerned, the sources mention that it is voluminous1198 and that its material is successively arranged

1194 Naḥḥās, Ṣināʿa 25–45. 1195 Yāqūt, Muʿǧam I, 164; Qifṭī, Inbāh I, 66; Suyūṭī, Buġya I, 291. Note that the book is also

erroneously attributed to Ibn Sīda, probably due to the similarity between his name and Sayyid (i.e. Abū l-Qāsim’s grandfather); cf. Yāqūt, Muʿǧam IV, 1648–1649; cf. Ibn Ḥazm, Rasāʾil II, 182.

1196 For Abū ʿAbdallāh’s biography, see Ibn al-Faraḍī, Tārīḫ II, 67a; Ṣafadī, Wāfī I, 334; Suyūṭī, Buġya I, 7.

1197 Lablī, Buġya 102; cf. above, 260. 1198 Cf. the expression fī kitābihi l-kabīr in Lablī, Buġya 102. Yāqūt (Muʿǧam I, 164) puts the

number of its “volumes” at a hundred (miʾat muǧallad), whereas Qifṭī (Inbāh I, 66) mentions the same number, but forty elsewhere (ibid., II, 334).

 

 

274 chapter 2

starting with the sky and ending with the atom (bada⁠ʾa bi-l-falak wa-ḫatama bi-l-ḏarra).1199 According to Qifṭī (d. 646/1248), the book is astounding (ġarīb ʿaǧīb), and no one who loves Arabic should ignore it.1200 The surviving sec- tion discusses a few organs of the human body (e.g. baṭn “belly” and qadam “foot”) and several human traits (awṣāf ), such as sounds, walking, shortness and length, sleep and wakefulness, food and drink, etc. It also includes lots of anecdotal material and aḫbār, unlike most other multithematic works. A com- parison between the surviving section of the book and Ibn Sīda’s al-Muḫaṣṣaṣ reveals that the latter considerably resembles it in material, but less so in arrangement.1201 It is difficult, however, to determine whether resemblance in material is due to Ibn Sīda’s dependence on his predecessor (albeit without any acknowledgement) or to common sources used by both men. But given Ibn Sīda’s keenness on listing his sources in the introduction of al-Muḫaṣṣaṣ (as he also does in al-Muḥkam) and continuously acknowledging them in its text, there is no reason to suppose that he deliberately made an exception in the case of Muḥammad’s book.

K. al-Talḫīṣ by Abū Hilāl al-ʿAskarī is the first extant book of the genre which consists purely of the first type of themes referred to earlier. Its title – which literally means “summary” or “synopsis” – is most probably a reference to the fact that it does not deal with patterns and includes relatively few šawāhid, as Abū Hilāl indicates in his introduction.1202 This notwithstanding, the material is described in the introduction as more comprehensive (aǧmaʿ) than in earlier works1203 and is properly organized into forty sections, which are subdivided into more specific topics. For example, the section on weapons and war con- sists of the following topics: swords, spears, coats of mail, bows, strings, arrows, whips, location of war, adjectives for armies, and names (asmāʾ) of war.1204 It is noteworthy that the book does not contain a special heading under which are listed words that express remainders of things (baqiyya/baqāyā), although a short section of this type is found in a few earlier multithematic works, namely, al-Ġarīb al-muṣannaf, al-Ǧarāṯīm and al-Muntaḫab.1205 The texts in these works are visibly alike, and are almost restricted to remainder of debt and food. Abū Hilāl, however, wrote a comprehensive monograph on the subject,

1199 Yāqūt, Muʿǧam I, 64; Qifṭī, Inbāh IV, 192. 1200 Qifṭī, Inbāh II, 334. 1201 Wadġīrī (1984: 81–86). 1202 ʿAskarī, Talḫīṣ I, 1. 1203 Ibid., I, 3. 1204 Ibid., II, 523–541. 1205 Abū ʿUbayd, Ġarīb III, 803–804; Ibn Qutayba, Ǧarāṯīm I, 331–332; Kurāʿ, Muntaḫab I,

356–357.

 

 

275MUBAWWAB AND SPECIALIZED LEXICA

al-Muʿǧam fī baqiyyat al-ašyāʾ, and arranged the words alphabetically, based on the first letter only.1206

Multithematic works from the fifth/eleventh century onward contain essentially the same topics treated in earlier works, but differ amongst them in arrangement, scope, and inclusion (or lack thereof) of chapters which deal with issues related to phonology, morphology, aḍdād, Arabized words, etc. Apart from Ibn Sīda’s huge opus, al-Muḫaṣṣaṣ, four titles are cited in the list above. Iskāfī’s relatively short book, Mabādiʾ al-luġa, consists of sixty chapters in which the author includes but few šawāhid and almost entirely omits the names of scholars from whom his material is derived. Ṯaʿālibī’s Fiqh al-luġa,1207 on the other hand, is divided into thirty chapters, which are further subdivided into about six hundred topics in which the author is heavily dependent on Abū ʿUbayd in his al-Ġarīb al-muṣannaf, as well as on other authorities whom he often cites. As for Niẓām al-ġarīb by the Yemeni scholar ʿĪsā b. Ibrāhīm al-Rabaʿī, it contains words which, as the author explains, are current in spite of being ġarīb (al-mustaʿmal min ġarīb al-luġa).1208 It is made up of 104 short chapters which contain many šawāhid. In contrast, Ibn al-Aǧdābī’s Kifāyat al-mutaḥaffiẓ contains no šawāhid at all, and is obviously addressed to the beginner.1209 What all four books have in common is that – with the exception of a chapter on foreign words and their Arabic equivalents in Ṯaʿālibī’s book1210 and a brief reference to dual nouns and passive verbs at the end of Rabaʿī’s1211 – their material is made up solely of the general type of themes, most of which are common to all four.

Ibn Sīda’s al-Muḫaṣṣaṣ is by far the most extensive extant multithe- matic work, and our numerous references to it in several genres discussed above is a reflection of the vastness of its material and diversity of its topics. What remains to be said here are a few comments on the raison d’être of the lexicon, its sources, and the arrangement of its contents.1212

1206 Abū Hilāl points out in the introduction (Baqiyya 63) that he arranged his material alphabetically. It should be noted that the editors of an earlier edition intervened in the text by rearranging words in full alphabetical order; see the edition of I. al-Abyārī and ʿA. Šalabī, Dār al-Kutub al-Miṣriyya, Cairo 1934, 24–25.

1207 For comments on the contents of this book, see ʿAbdaltawwāb (1994: 63–70) and Ḫalīl (1997: 336–348).

1208 Rabaʿī, Niẓām 20. 1209 For a list of commentaries on Ibn al-Aǧdābī’s book, see Wadġīrī (2008: 62–66). 1210 Ṯaʿālibī, Fiqh 314–319. 1211 Rabaʿī, Niẓām 274, 275. 1212 For more detailed studies, see Ṭālibī (1956), Haywood (1962), Cabanelas Rodriguez (1966),

Muṭlaq (1967: 360–371), Ḫalīl (1997: 349–374), Zīrǧāwī (2010).

 

amgad
Highlight

 

276 chapter 2

As previously pointed out, we have adopted the terms mubawwab and muǧannas from Ibn Sīda’s introduction to his book in order to refer consecu- tively to onomasiological and semasiological lexica.1213 It is evident that Ibn Sīda wanted to justify authoring a lexicon of the first type (i.e. al-Muḫaṣṣaṣ) having also authored, probably simultaneously,1214 a lexicon of the latter type (i.e. al-Muḥkam). He explains that by virtue of its arrangement, al-Muḥkam guides its user to the exact place of the word which he seeks. Accordingly, he intended to author a lexicon which would match it (aradtu an aʿdil bihi), and which would better suit the orator (ḫaṭīb) and poet (šāʿir) who do not seek the meaning of a specific word, but need to choose the most suitable word from among numerous nouns (asmāʾ) and adjectives (awṣāf ) that are related to a certain semantic field. In other words, this type of arrangement enables the user to find words which he might not know – a function which muǧannas lexica hardly perform. Ibn Sīda’s statement is the clearest in the tradition as far as the relationship between mubawwab and muǧannas lexica is concerned, and it provides the best explanation for the fact that the two types coexisted for several centuries.

Although Ibn Sīda arrogantly praises his book in the introduction, he makes no secret of his heavy reliance on earlier authors. His impressive list of sources includes both mubawwab and muǧannas lexica as well as books on grammar and philology.1215 Among the mubawwab lexica are Abū ʿUbayd’s al-Muṣannaf (i.e. al-Ġarīb al-muṣannaf ) and Ġarīb al-Ḥadīṯ, all the books of Ibn al-Sikkīt, Ṯaʿlab’s al-Faṣīḥ and al-Nawādir, Abū Ḥanīfa al-Dīnawarī’s al-Anwāʾ and al-Nabāt, as well as books by Farrāʾ, Aṣmaʿī, Abū Zayd, Abū Ḥātim, Mubarrad, Kurāʿ al-Naml, al-Naḍr b. Šumayl, Ibn al-Aʿrābī, Liḥyānī, and Ibn Qutayba. The muǧannas sources which he specifies are Ibn Durayd’s al-Ǧamhara, Ḫalīl’s al-ʿAyn, and Qālī’s al-Bāriʿ. As for grammar and philology, he mentions Sībawayhi’s al-Kitāb and some commentaries on it, several books by Abū ʿAlī al-Fārisī, Ibn Ǧinnī and Rummānī, and Abū Bakr b. al-Anbārī’s al-Zāhir. Moreover, the text of al-Muḫaṣṣaṣ is replete with the names of the scholars on whom Ibn Sīda relies for data, and the frequency of the recurrence of most of these names is obviously due to his method of acknowledging his source(s) almost in every detail or word in the majority of his chapters. Among the sources he acknowledges, Abū ʿUbayd’s al-Ġarīb al-muṣannaf is probably the one from which Ibn Sīda benefited the most, given that he derives much of his material directly from Abū ʿUbayd and that the two books are generally similar

1213 Ibn Sīda, Muḫaṣṣaṣ I, 10, 12; cf. above, 47–48. 1214 Cf. below, 323. 1215 Ibn Sīda, Muḫaṣṣaṣ I, 11–13.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

277MUBAWWAB AND SPECIALIZED LEXICA

in arrangement and topics, in spite of some differences between them in these two aspects. Interestingly enough, Yāqūt (d. 626/1229) reports that Ibn Sīda recited al-Ġarīb al-muṣannaf by heart from beginning to end to Abū ʿUmar al-Ṭalamankī (d. 429/1038) in Murcia.1216

Yet Ibn Sīda was keen on authoring a much more comprehensive book than Abū ʿUbayd’s or any of the works he enumerates as his sources. He thus clearly states in the introduction that he did not find among the earlier books any which embraces most, let alone all, of the data related to this diverse and vast language (hāḏihi l-luġa al-mutašaʿʿiba al-fasīḥa).1217 As for his own book, he describes it as mustawʿib tāmm (comprehensive and complete) and mustaqṣā (fully investigated), and asserts his claim of comprehensiveness by using the expression (iddaʿaytu l-iḥāṭa).1218 In fact, the book contains the full content of some monographs on plants, human body, horses, camels, cattle, natural phenomena, etc., as well as philological topics, such as aḍdād, hamz, ḥurūf, abniya, muḏakkar and muʾannaṯ, etc. It also has a huge number of šawāhid, particularly from poetry (whose šawāhid alone occupy 274 pages of the gen- eral index).1219 In short, al-Muḫaṣṣaṣ may be regarded as the counterpart of Ibn Manẓūr’s (d. 711/1311) muǧannas lexicon, Lisān al-ʿArab, as both authors bring together, for the sake of comprehensiveness, a number of earlier works, although Ibn Manẓūr’s sources (five in all) are much fewer than Ibn Sīda’s. It should be noted, however, that one of Ibn Manẓūr’s sources is al-Muḥkam, itself derived by Ibn Sīda from a large number of sources that he also used in al-Muḫaṣṣaṣ.

Other than praising his book for comprehensiveness, Ibn Sīda boasts its unique arrangement according to which the general (aʿamm) precedes the specific (aḫaṣṣ), universals (kulliyyāt) precede particulars (ǧuzʾiyyāt), and substance (ǧawhar) precedes accident (ʿaraḍ).1220 As an example, he cites the section on human beings, titled ḫalq al-insān, in which he begins with the stages of creation and then proceeds from the general topics to specific ones. Based on the two criteria of comprehensiveness and logical arrangement, Ibn Sīda announces that his book is unmatched by any other work on philology (infaṣala hāḏā l-kitāb min ǧamīʿ kutub al-luġa).1221 This notwithstanding, the actual arrangement of the book does not strictly reflect the plan which Ibn

1216 Yāqūt, Muʿǧam IV, 1649; cf. Ibn Ḫillikān, Wafayāt III, 330. 1217 Ibn Sīda, Muḫaṣṣaṣ I, 7. 1218 Ibid., 3, 12–13. 1219 See Hārūn (1991: 149–422). 1220 Ibn Sīda, Muḫaṣṣaṣ I, 10. 1221 Ibid., I, 14.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight

 

278 chapter 2

Sīda expounds in his introduction, for although he divides its material into sec- tions (kutub) which are further divided into chapters (abwāb), the principles behind the division are not always clear. For example, the first kitāb, which deals with ḫalq al-insān, includes 122 topics, but for no obvious reason, only three of these are identified as bāb.1222 Furthermore, within the same kitāb, a number of bābs are at times collectively referred to as kitāb.1223 As for the internal arrangement of the chapters, one cannot discern any adherence to the principles stated by Ibn Sīda in his introduction, as the cited words seem to be arranged either randomly or on the basis of the sources from which they are derived.

On a wider level, al-Muḫaṣṣaṣ is divided into two major parts, the first of which consists of the general themes, whereas the other comprises the topics related to specific philological issues. The first of the two occupies the best part of the book (almost thirteen volumes out of a total of seventeen), and, based on Ibn Sīda’s arrangement of the material, Ṭālibī identifies the following four major areas into which the general themes may be classified:1224 human beings and their environment (procreation, body parts, instincts, women, clothes, food, illnesses, sleep, sexual relations, abodes, weapons, and fighting); animals (horses, battalions, banners, camels, cattle, savage beasts, beasts of prey, hunt- ing and its tools, small creatures, pests, insects, birds, and bird hunting and its tools); nature and plants (anwāʾ,1225 natural phenomena and man’s utilization of them, palm trees, other trees, sweet-smelling plants, meadows, and metals); and man in society (e.g. travel, character, social relations, play, beliefs, etc.). As for the second major part, its main constituents are: surnames which begin with ab, umm, ibn, etc., duals, negation, words with contradictory meanings, substitution of letters, hamz, particles, indeclinable words, diminutives, anom- alous plurals, verbs, verbal nouns, abbreviated and prolonged patterns, mascu- line and feminine, and numbers. This comprehensiveness of al-Muḫaṣṣaṣ is most probably the reason why extensive works in the genre practically came to an end relatively early, compared with other genres. Later authors must have found very little to add to its material, and obviously lacked any innovative approach to arrange the available data on an altogether different basis.

1222 Ibid., I, 15-II, 148. For the three bābs in question, see I, 17, 74; II, 112. 1223 Cf. the expression Kitāb al-Aṣwāt (ibid., II, 148) with reference to a number of bābs within

Kitāb Ḫalq al-insān. 1224 Ṭālibī (1956: 70–72). 1225 Cf. above, 158.

 

 

© koninklijke brill nv, leiden, ���4 | doi ��.��63/9789004�740�3_�04

chapter 3

Muǧannas (Semasiological) Lexica

1 Introduction

The muǧannas type of lexica (i.e. general, unspecialized works in which sign leads to meaning) appeared only shortly after the first mubawwab lexica, and it has already been argued that the two types practically belong to the same stage of philological activity and in no way represent two consecutive stages of lexicography as some have claimed.1 In their earliest stages, both types were closely linked to interest in the philological study of the Qurʾān and the desire to preserve correct usage and systematize the available data, particularly due to the linguistic shifts that took place with the expansion of Arabic fol- lowing the conquests and the Arabization of the official registers, or dawāwīn, in the second half of the first/seventh century. But in spite of their common background, there was an early awareness that each of the two types served a purpose which the other could not achieve, for whereas the muǧannas type proceeded from sign to meaning, the mubawwab proceeded from meaning to sign. As previously mentioned, Ibn Sīda’s (d. 458/1066) succinct definition of the purpose of each type, as represented in his al-Muḥkam and al-Muḫaṣṣaṣ, clearly justifies the coexistence of the two types for several centuries.2

Unlike mubawwab lexica, most muǧannas ones aim at exhausting the lin- guistic corpus of Arabic, i.e. the inclusion of all the roots of the language, and, in some of the more comprehensive works, the greatest part of the deriva- tives of these roots. This naturally made the authoring of a muǧannas lexicon a much more arduous task than writing a mubawwab one, perhaps with the exception of the few voluminous and comprehensive multithematic works which are best exemplified by Ibn Sīda’s al-Muḫaṣṣaṣ. It is therefore not sur- prising that the number of lexica to be discussed in this chapter is much less than the number of those encountered in the previous one. But in spite of the relative scarcity of muǧannas lexica, they exhibit a noticeable variety not only in their arrangement but also in content. The criterion of the arrange- ment of material has traditionally been adopted in classifying these lexica, and we shall make no exception to this practice. Accordingly, our classification includes three major types: the phonetic-permutative, the alphabetical, and

1 Cf. above, 46–47. 2 Cf. above, 47–48, 276.

 

amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Sticky Note
Cancelled set by amgad
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Highlight

 

280 chapter 3

the rhyme systems. Yet, one should be aware of the fact that this classification is not without some overlaps. Ibn Durayd’s (d. 321/933) Ǧamharat al-luġa, for example, is alphabetically arranged but lists together the various permutations of the radicals. Moreover, the rhyme system adopts an alphabetical arrange- ment, except that roots are arranged according to their last letter, and not their first. Above all, the classification of muǧannas lexica based on the type of arrangement should not divert one from the fact that the three types men- tioned above are not separate stages in the development of the lexicographical tradition. Although the rhyme system, which was fully developed by Ǧawharī (d. c. 400/1010) in al-Ṣaḥāḥ, eventually became dominant, there were major muǧannas lexica which continued to adopt the phonetic-permutative and the alphabetical systems well into the fifth/eleventh and sixth/twelfth centuries respectively. Following our discussion of the lexica of all three types, the epi- logue shall examine, among other subjects, some of the issues which pertain to all of them regardless of their type of arrangement.

2 The Phonetic-Permutative System

This system of arranging Arabic roots is most probably the oldest3 and goes back to the first muǧannas lexicon in the tradition, al-Ḫalīl b. Aḥmad’s (d. 175/791) K. al-ʿAyn. What is exclusive to the lexica which adopt this system is the phonetic aspect, that is, the arrangement of letters based on their place of articulation. The method of including in each lemma all the permutations (taqālīb) of the radicals is however not exclusive to this group of lexica since Ibn Durayd (d. 321/933) adopts this method but still arranges the roots alpha- betically in his Ǧamharat al-luġa. The originator of the phonetic alphabet (i.e. ʿ, ḥ, h, ḫ, ġ, etc. with minor modifications at times) is almost certainly Ḫalīl, and the reasons for originating this system will be discussed as part of our discus- sion of K. al-ʿAyn.

From a wider perspective, phonetic arrangement is not restricted to muǧannas lexica but also features in some mubawwab ones. Among the exam- ples encountered in the previous chapter are Qālī’s (d. 356/967) al-Maqṣūr wa-l-mamdūd, Ibn al-Qūṭiyya’s (d. 367/977) al-Afʿāl and Saraqusṭī’s (d. after 400/1010) al-Afʿāl.4 Permutations were also adopted by a number of authors of mubawwab lexica, the earliest of whom is most probably Ḥarbī (d. 285/898)

3 For chronological comparison with K. al-Ǧīm, cf. below, 332–333. 4 Cf. above, 245, 258–260.

 

amgad
Highlight
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Highlight
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline

 

281Muǧannas (Semasiological) Lexica

in his Ġarīb al-Ḥadīṯ.5 Ǧarbāḏaqānī (fl. in the second half of the fourth/tenth century) also introduced permutations in some entries of his phonetically arranged lexicon, al-Rawḥa, which deals with ḍād and ẓāʾ.6 Furthermore, the notion of permutations seems to have inspired Ibn Ǧinnī (392/1002) with his ingenious theory (albeit farfetched and demonstrable in no more than a few roots) known as al-ištiqāq al-akbar.7 According to this theory, the six possible combinations resulting from the change of order in the radicals of a triliteral root can share a general meaning, as in ǦBR, ǦRB, BǦR, BRǦ, RǦB and RBǦ, all of which are claimed to express the notion of strength or intensity. Furthermore, Ibn Fāris (d. 395/1004) was probably influenced in his K. al-Ṯalāṯa by permuta- tions of roots, although he restricted them to three, as in ḥalīm (forbearing), ḥamīl (of suspected origin), and laḥīm (murdered).8

The phonetic system of arranging roots is undoubtedly the most difficult and impractical in the lexicographical tradition. Several lexicographers have pointed out this fact in justifying their choice of arrangement of their lexica. Ibn Durayd, for example, quite tellingly preserves the permutations adopted in al-ʿAyn but declares that he chose the alphabetical system (i.e. not the phonetic one) because its arrangement of letters is easier for the user, given that it is known both to the ʿāmma (common people, generality of people) and the ḫāṣṣa (elite or highly educated people).9 In other words, the phonetic arrangement of letters, with which the latter group may have been familiar, was hardly known to the former, and whereas al-ʿAyn was the first attempt at writing a muǧannas lexicon and hence was not addressed to the needs of the ʿāmma, Ibn Durayd was keen for his own lexicon to be more widely cir- culated, hence his abandonment of the phonetic arrangement. Moreover, Ibn Wallād (d. 332/944) in the introduction of al-Maqṣūr wa-l-mamdūd (which is alphabetically arranged based on the first letter only) explains that, for ease of use, his lexicon differs from al-ʿAyn in starting with alif, i.e. hamza. He then interestingly notes that some readers might blame him for this because alif is ḥarf muʿtall (weak letter).10 This strongly reminds one of the reason cited by Ḫalīl for avoiding to begin with alif,11 and may be interpreted as indicating that, in spite of the difficulty which characterizes Ḫalīl’s phonetic arrangement

5 Cf. above, 77. 6 Cf. above, 221. 7 Ibn Ǧinnī, Ḫaṣāʾiṣ II, 133–139. 8 Cf. above, 251. 9 Ibn Durayd, Ǧamhara I, 40. 10 Ibn Wallād, Maqṣūr 2–3; cf. above, 244. 11 Ḫalīl, ʿAyn I, 47; cf. Iskāfī, Muḫtaṣar I, 73–74.

 

amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline

 

282 chapter 3

in comparison with the alphabetical one, proponents of the former were con- vinced of its superiority, at least with respect to the fact that it did not start with a weak letter that was often subject to elision or change. It should also be noted that to al-Mufaḍḍal b. Salama (d. 290/903 or 300/913) is attributed a book entitled al-Bāriʿ ( fī ʿilm al-luġa) of which, according to Ibn al-Nadīm (d. 380/990), he only completed the letters ’, h, ʿ, ḥ, ġ and ḫ.12 Unless Ibn Nadīm had randomly arranged these letters (and there is no reason why he should have done so), Mufaḍḍal’s lexicon may be an example of a relatively early pho- netically arranged lexicon whose author did not share Ḫalīl’s concern about starting with alif and thus adopted a phonetic arrangement similar to that of Sībawayhi’s.13

The following lexica are arranged according to the phonetic-permutative system:

1. Kitāb al-ʿAyn by al-Ḫalīl b. Aḥmad al-Farāhīdī (d. 175/791). 2. al-Bāriʿ fī l-luġa by Abū ʿAlī al-Qālī (d. 356/967). 3. Tahḏīb al-luġa by Azharī (d. 370/981). 4. al-Muḥīṭ fī l-luġa by al-Ṣāḥib b. ʿAbbād (d. 385/995). 5. al-Muḥkam wa-l-muḥīṭ al-aʿẓam by Ibn Sīda (d. 458/1066).14

Several of the works that comment on, amend or expand the content of some of these lexica naturally follow their arrangement. The more significant of such works will be mentioned within their relevant sections.

2.1 Kitāb al-ʿAyn by al-Ḫalīl b. Aḥmad (d. 175/791) The study of the method and content of K. al-ʿAyn is closely related to two controversial issues, namely, its originality and the identity of its author. As far as originality is concerned, we have already argued that al-ʿAyn is the result of Ḫalīl’s probing intellect (most notably in devising a plan for the lexicon) and his intuitive approach to the phonetic traits of the language as reflected, for example, in the lively discussions which took place between him and his disciples concerning the sampling (ḏawāq) of letters based on his own

12 Ibn al-Nadīm, Fihrist 80. Note that the book is a refutation of Ḫalīl’s al-ʿAyn; cf. Qifṭī, Inbāh III, 97, 306; IV, 68.

13 See Kitāb IV, 433–434. 14 To an Andalusian contemporary of Ibn Sīda’s, Abū ʿAbdallāh Muḥammad b. Yūnus

al-Ḥiǧāzī (d. 462/1070), is attributed a lexicon entitled al-Mubarriz fī l-luġa which prob- ably followed the phonetic-permutative system since Ibn Ḫayr (Fahrasa 357) likens it to Ibn Sīda’s al-Muḥkam (kitāb miṯl al-Muḥkam).

 

amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Underline

 

283Muǧannas (Semasiological) Lexica

Sprachgefühl.15 As in the case of the Arabic grammatical theory, early lexico- graphical activity, as exemplified by al-ʿAyn, is basically free from direct foreign influence, firmly based in the Arabic-Islamic culture and inextricably linked to other areas of inquiry pursued by Arab scholars, particularly the study of the Qurʾānic text. Both fields, as previously argued, are the result of indigenous investigation and speculation, rather than the borrowing of a foreign system that was readily available and merely needed to be Arabized. Had this been the case, it would be strange indeed that the Arabic sources do not say a single word about the presence of foreign influence on the inception of grammatical and lexicographical activity. Although indirect influences in both fields cannot be totally ruled out, proponents of foreign origins have so far failed to present any evidence of direct influence, whereas the view which supports indigenous origins is strongly supported by textual evidence and by the interrelationship among several types of linguistic and Qurʾānic fields of study, such as exegesis (tafsīr), jurisprudence ( fiqh), and Qurʾānic readings (qirāʾāt).

The question of the identity of the author of al-ʿAyn has been the subject of interest of Arab scholars almost since the lexicon was authored or, in the least, circulated. It has also been discussed, at times extensively, by numer- ous contemporary scholars, among whom are Bräunlich (1926), Darwīš (1956), Naṣṣār (1968; first published 1956), Wild (1965), ʿUryān (1984), Maḫzūmī (1986), Kišlī (1996a), Talmon (1997), Mullā (2001), Ǧubūrī (2004), and Schoeler (2006).16 Accordingly, we need only briefly point out the main arguments available in the medieval sources concerning this issue. It is essential however to assert at this point that, despite the significance of determining the book’s real author in establishing historical facts, the issue has little bearing on the importance of al-ʿAyn as the first Arabic lexicon which is the product of a sophisticated plan that is detailed in its introduction. Even if al-Ḫalīl were not its author in full or in part, it must have been authored only a few years or a couple of decades following his death, that is, by the time of his disciple al-Layṯ b. al-Muẓaffar (d. 190/805), to whom al-ʿAyn is attributed in several sources (see below). In all cases, the purpose of the discussion concerning the author of al-ʿAyn is primar- ily to shed light not only on the reception of the book in learned circles but also on matters related to its content and organization.

From Suyūṭī’s (d. 911/1505) numerous quotations of earlier authors,17 and from other citations in the sources, it is clear that, despite some reports which

15 Cf. above, 58–59; see also Sībawayhi, Kitāb III, 320–321. 16 See a detailed list of works which deal with the issue of the attribution of al-ʿAyn to Ḫalīl

in Talmon (1997: 91–92). 17 Suyūṭī, Muzhir I, 76–92.

 

amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline

 

284 chapter 3

ascribe to one scholar two partially or fully contradictory opinions, there are three basic views regarding the issue of the identity of al-ʿAyn’s author. The first denies that Ḫalīl had anything to do with authoring the lexicon; the second asserts that he was the author of both its introduction and text; and the third ascribes to him only a partial role, although there is no consensus as to the nature and extent of this role.

The most substantial arguments against the attribution of al-ʿAyn to Ḫalīl are forwarded by Zubaydī (d. 379/989) in the introduction to his Istidrāk al-ġalaṭ al-wāqiʿ fī K. al-ʿAyn,18 a book in which he amends some of the errors with which he believes al-ʿAyn is replete. Irrespective of Zubaydī’s timid admis- sion that the book’s origin goes back to Ḫalīl (sabbaba aṣlahu)19 and that other scholars completed it after his death, the arguments he produces against its attribution to Ḫalīl better reflect his real position. These are that the book is fraught with errors which could hardly be attributable to a highly esteemed scholar such as Ḫalīl;20 that some of its riwāyas are ascribed to philologists who are predated by Ḫalīl (e.g. Abū ʿUbayd who was born in 154/770 and who was either sixteen or twenty-one years old when Ḫalīl died);21 that it includes šawāhid from poets who were contemporary with Ḫalīl and who are not nor- mally cited by the early philologists; that Ḫalīl’s contemporaries do not quote al-ʿAyn in their works; and that all grammatical views in the lexicon are Kufan, rather than Basran (i.e. as should be expected from a Basran like Ḫalīl, were he the real author of the lexicon).

As counterarguments, it should be remembered that several early texts, other than al-ʿAyn, were subject to errors committed by copyists or were augmented by riwāyas which students added to their own copies and which became indis- tinguishable from the original text.22 One example is Bandanīǧī’s (d. 284/897) K. al-Taqfiya23 into whose text were incorporated additions by Aḥmad b. ʿAbdallāh b. Muslim al-Dīnawarī (d. 322/934), Ġulām Ṯaʿlab (d. 345/957), and Ibn Ḫālawayhi (d. 370/980).24 Another example is that when Aṣmaʿī (d. 216/831) was brought a copy of his own K. al-Nawādir, he realized that parts

18 Zubaydī, Istidrāk al-ġalaṭ 42–50; text also quoted almost in its entirety in Suyūṭī, Muzhir I, 79–86.

19 Zubaydī, Istidrāk al-ġalaṭ 45; cf. Zubaydī, Muḫtaṣar 42. 20 See Badrī (1999: 33 ff.) for instances of alleged taṣḥīf in al-ʿAyn. 21 See also examples of citations in al-ʿAyn of scholars who outlived Ḫalīl in Ǧubūrī (2004:

49–50). 22 For glosses contained in the edited text of the Kitāb, see Humbert (1997). 23 See below, 370. 24 See the editor’s introduction to Bandanīǧī, Taqfiya 25–26.

 

amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline

 

285Muǧannas (Semasiological) Lexica

of the text were added to the original. When he marked these additions, they amounted, according to Salama b. ʿĀṣim (d. after 270/883), to more than one third of the whole book.25 Additions to original texts certainly included poetry šawāhid, and it is thus natural that lines of poetry ascribed to poets either con- temporary with or even later than the relevant author should occur in certain texts. As for the numerous errors which, according to Zubaydī, occur in al-ʿAyn, it should be noted that it was perhaps the norm, rather than the exception, that lexicographers and grammarians were criticized or accused of error by their contemporaries or later scholars. Mubarrad’s (d. 285/898) refutation of Sībawayhi (d. 180/796)26 and Nifṭawayhi’s (d. 233/935) bitter attack on Ibn Durayd (d. 321/933)27 are well documented. In fact, Zubaydī himself authored, in an indirect accusation of Sībawayhi for not exhausting morphological pat- terns, a book titled K. al-Istidrāk ʿalā Sībawayhi fī kitāb al-abniya wa-l-ziyādāt ʿalā mā awradahu fīhi muhaḏḏaban.

As for the alleged absence of quotations from al-ʿAyn in the works of early scholars (the latest among whom to be mentioned by Zubaydī28 is Abū Ḥātim al-Siǧistānī [d. 255/869]), it has already been noted that, during the early stages of philological activity, attribution of data to its ultimate source was by no means a regular practice.29 For example, Abū ʿAmr al-Šaybānī (d. 206/821) practically ignores mention of any philologist of the period in his K. al-Ǧīm although the book is replete with dialectal material which he does ascribe to the relevant tribes or to Bedouin informants.30 Starting with Ibn Durayd’s al-Ǧamhara, followed by lexica such as Azharī’s (d. 370/981) al-Tahḏīb and Ibn Fāris’s (d. 395/1004) al-Maqāyīs and al-Muǧmal, lexicographical material derived from al-ʿAyn and attributed to Ḫalīl becomes extremely common. Outside philological sources, a very early quotation from Ḫalīl, consistent with al-ʿAyn, occurs in Ibn Hišām’s (d. 213/828) famous al-Sīra al-nabawiyya.31

25 Azharī, Tahḏīb I, 5. 26 Cf. Bernards (1997). 27 Yāqūt, Muʿǧam VI, 2496; Suyūṭī, Muzhir I, 94. 28 Zubaydī, Istidrāk al-ġalaṭ 48. 29 Cf. above, 23. 30 Cf. below, 335. 31 Ibn Hišām (Sīra I, 612) quotes Ḫalīl as saying that al-ʿayhab means “a man who is too

weak to take vengeance for himself” (al-raǧul al-ḍaʿīf ʿan idrāk watrihi); cf. Ḫalīl, ʿAyn I, 109: al-ʿayhab: al-balīd min al-riǧāl al-ḍaʿīf ʿan idrāk watrihi. One, however, cannot exclude the possibility that the quotation from Ḫalīl may be a later addition to the text of the Sīra since it does not occur in some of its manuscripts. Dīnawarī’s (282/895) K. al-Nabāt also contains some quotations from al-ʿAyn, although these are not ascribed to Ḫalīl; see Schoeler (2006: 163).

 

 

286 chapter 3

A related argument intended to deny Ḫalīl’s role in al-ʿAyn is that there is no riwāya of the lexicon which goes back to Ḫalīl.32 This is contradicted, however, by other reports, such as the one in which Ibn Fāris lists in the introduction of al-Maqāyīs the scholars on whose authority his riwāya of al-ʿAyn is based.33

Zubaydī’s claim that all the grammatical views expressed in al-ʿAyn are Kufan is a gross exaggeration. In fact, Zubaydī himself in the text of Istidrāk al-ġalaṭ ignores this (largely anachronistic) argument, except on one occasion in which he alleges that Ḫalīl classifies the root DḪDḪ, as the Kufans do, under triliter- als, rather than reduplicated quadriliterals, as the Basrans do.34 By comparing Zubaydī’s text with al-ʿAyn,35 however, it becomes clear that Ḫalīl compares the meaning of DḪDḪ with that of DWḪ and does not refer to its derivation. Furthermore, Talmon’s comparison between al-ʿAyn and Sībawayhi’s al-Kitāb demonstrates that, in spite of how equivocal the testimony of some items may be and of certain “shades of dissimilarity” between the two texts in concep- tion, terminology and extensiveness, the items in which Ḫalīl is expressly men- tioned in both texts “justify crediting the grammatical material in K. al-ʿAyn with a Ḫalīlian descent”,36 hence with a Basran character according to the traditional distinction by later grammarians between two trends: one Basran and the other Kufan. Points of disagreement on grammatical issues between al-ʿAyn and al-Kitāb, Talmon concludes, are mainly the result of the difference between the two texts in scope and potential readership. It should also be noted that a number of grammatical views that are assumed to be Kufan and occur in early works other than al-Kitāb do agree with the grammatical teach- ings of K. al-ʿAyn.37 But one certainly doubts whether these views, as well as those which occur in al-ʿAyn and are claimed to be Kufan, were restrictively Kufan at such an early date of grammatical activity.

One of the main reasons for the rejection by certain scholars of the attri- bution of al-ʿAyn to Ḫalīl is their conviction that it embraces errors which do not become a scholar of his stature. Yet most of these scholars were also con- vinced that it was none other than Ḫalīl who initiated the idea of devising

32 Ibn al-Nadīm, Fihrist 48. 33 Ibn Fāris, Maqāyīs I, 3–4. See also Schoeler (2006: 153). 34 Zubaydī, Istidrāk al-ġalaṭ 94; cf. 48. Among the other alleged morphological errors which

Zubaydī (p. 49) claims to be found in al-ʿAyn is that Ḫalīl did not distinguish between quadriliterals and quinqueliterals. This allegation is not true and is quite puzzling given that Zubaydī follows in his own book Ḫalīl’s suit in distinguishing between the two types.

35 Ḫalīl, ʿAyn, DḪ (IV, 138); cf. DWḪ (IV, 295). 36 Talmon (1997: 256–259). 37 Ibid., 277.

 

amgad
Underline

 

287Muǧannas (Semasiological) Lexica

the first Arabic lexicon. It is thus not surprising that they tried to come up with a solution which would give Ḫalīl some credit in al-ʿAyn but would at the same time allow them to blame its shortcomings on someone else. Indeed, several authors explicitly suggest such an interpretation, among the earliest of whom is Ṯaʿlab (d. 291/904), who asserts that because Ḫalīl was unrivalled (lam yura miṯluhu), he could have only planned the lexicon (rasamahu) but not provided its actual material (ḥašāhu), or else he would not have left any blemishes (mā baqqā fīhi šayʾan). According to Ṯaʿlab, al-ʿAyn became deficient (iḫtalla) because philologists other than Ḫalīl supplied its material, after which the lexicon spread not through riwāya, but through copied manuscripts.38 A few decades later, Zubaydī (d. 379/989) reports that, independently of Ṯaʿlab’s text which he neither saw nor heard of, he came to a similar conclusion when he realized that Ḫalīl initiated the work (sabbaba aṣlahu), but died before finishing it.39 Ibn Ǧinnī (d. 392/1002) goes one step further when he explains that the confusion, deficiency and corruption (al-taḫlīṭ wa-l-ḫalal wa-l-fasād) which characterize al-ʿAyn could not be ascribed to Ḫalīl’s most insignificant disciples (aṣġar atbāʿ al-Ḫalīl), let alone Ḫalīl himself.40 He concludes that, if Ḫalīl had anything to do with the lexicon, he must have merely suggested that it be authored (awma⁠ʾa ilā ʿamal al-kitāb īmāʾan).

Medieval scholars who attribute to Ḫalīl some involvement in authoring al-ʿAyn differ as to the extent of his contribution to it. The predominant view is that, following Ḫalīl’s death, his student al-Layṯ b. al-Muẓaffar (d. 190/805) took it upon himself to complete it. Some riwāyas attribute to Ḫalīl, other than setting the plan, the completion of the letter ʿayn,41 whereas an unattributed riwāya mentioned by Suyūṭī claims that he reached the letter ġayn (i.e. the fifth letter in al-ʿAyn’s arrangement).42 Other scholars deny that Ḫalīl wrote any of the entries and restrict his contribution to providing the lexicon’s concept and plan. Sīrāfī (d. 368/979), for example, attributes to Ḫalīl the first part of the book43 – most probably a reference to its introduction. Moreover, Azharī (d. 370/981), ascribes to him the concise foundation (al-ta⁠ʾsīs al-muǧmal) in the

38 Abū l-Ṭayyib, Marātib 57; text also in Suyūṭī, Muzhir I, 78. 39 Zubaydī, Istidrāk al-ġalaṭ 45. 40 Ibn Ǧinnī, Ḫaṣāʾiṣ III, 288. 41 Abū l-Ṭayyib, Marātib 57; Yāqūt, Muʿǧam V, 2254. 42 Suyūṭī, Muzhir I, 77. 43 Sīrāfī’s exact words are: wa-ʿamila awwal K. al-ʿAyn al-maʿrūf al-mašhūr allaḏī bihi

yatahayya⁠ʾ ḍabṭ al-luġa (Aḫbār 38). I am inclined to take bihi yatahayya⁠ʾ ḍabṭ al-luġa (through which language can be verified) as an adjective of awwal, rather than K. al-ʿAyn, since it is this awwal (i.e. the introduction), and not the lemmata, which sets the phonetic and morphological rules upon which the work rests and which includes several norms

 

 

288 chapter 3

beginning of the lexicon44 – an obvious reference to its introduction. A more active role is ascribed to Layṯ by Ibn al-Muʿtazz (d. 296/909) who cites a lengthy story, according to which Layṯ’s wife burnt her husband’s copy of al-ʿAyn in revenge for his infidelity. Since there was no other copy of the book, and since Layṯ had memorized half of it, he wrote down that half and collaborated with contemporary scholars to write the other half.45 Other accounts which credit Layṯ with a significant contribution to authoring al-ʿAyn claim that Ḫalīl used to dictate the lexicon’s material to Layṯ and ask him to check the correctness of anything which he (i.e. Ḫalīl) was in doubt of. Having accomplished this task, according to one account, Layṯ declared that he finished the book (ilā an ʿamiltu l-kitāb).46 In fact, there may be textual evidence from al-ʿAyn in sup- port of reports that Layṯ checked the use of certain lexical items. In the entry ʿKŠ, for example, Ḫalīl’s interlocutor, assumed to be Layṯ, asked him why he claimed that this root was unused (muhmal, lit. neglected) although the Arabs do use ʿUkāša as a proper noun. Ḫalīl responded by saying that proper nouns do not follow the norm (qiyās). Immediately following this, Layṯ mentions that he asked Abū l-Duqayš, one of the most famous eloquent Aʿrāb, about the meaning of his name. When the latter confirmed that he had never heard an explanation of it, Layṯ astonishingly asked him how he could have a surname of whose meaning he was ignorant. The conversation is concluded when Abū l-Duqayš, like Ḫalīl, asserts that proper nouns are markers (ʿalāmāt) which people use as they choose and which do not conform to qiyās or any binding principle (ḥatm).47 Another example is that, in the entry QNḤ, Layṯ reports (i.e. Ḫalīl’s view) that qanḥ is the fastening with a stick of the side-post of a door, but then cites a different opinion attributed to someone other than Ḫalīl (qāla ġayr al-Ḫalīl). This undisclosed authority denies knowing any meaning for the word other than sipping one’s drink.48 Moreover, expressions of the type qāla ġayr al-Ḫalīl (Someone other than Ḫalīl said), qāla ġayruhu (Someone else

and criteria for the acceptance of reported usage and for determining whether a certain lexical item is Arabic or not.

44 Azharī, Tahḏīb I, 41. 45 Ibn al-Muʿtazz, Ṭabaqāt 96–98; text partially also in Yāqūt, Muʿǧam V, 2254–2255. 46 Ibn al-Nadīm, Fihrist 48. 47 Ḫalīl, ʿAyn, ʿKŠ (I, 190). 48 Ibid., QNḤ (III, 50). See also ʿHB (I, 109), where ʿayhab is said to mean “a sluggish man

who is too weak to take vengeance for himself”. Layṯ however reports that, according to a certain Abū Saʿīd, the correct word is ġayhab, although he admits the possibility that ʿayn and ġayn may be interchangeable in this word.

 

amgad
Underline

 

289Muǧannas (Semasiological) Lexica

said), qāla āḫarūn (Others said), etc.49 may well be indicative of Layṯ’s efforts at establishing the correctness of the use or of the meanings of certain lexi- cal items and at adding material which he heard directly through the riwāya of philologists other than Ḫalīl or directly from the Aʿrāb, some of whom are mentioned by name.50

The most important part of al-ʿAyn is undeniably its introduction. This valuable document contains the main phonetic and morphological ideas of Ḫalīl’s which lay the foundations of the lexicon. In fact, the introduction bears the mark of Ḫalīl in more than one respect, and although it is apparently a riwāya by one of his disciples – i.e. it is not written by Ḫalīl himself unlike, say, al-Kitāb which was undoubtedly written by Sībawayhi himself – there is no reason to suppose that a disciple of his intentionally ascribed his own ideas to him. Even the riwāyas which claim that Layṯ loved Ḫalīl (min ḥubbihi li-l-Ḫalīl) to the extent that he ascribed al-ʿAyn to him51 primarily refer to Layṯ’s contribution to the lemmata rather than the introduction. We have also noted previously that some of the harshest critics of the lexicon do admit that Ḫalīl set its plan, and these at times specifically credit him with the introduction. But apart from the conflicting reports in the sources, the attribution of the introduction to Ḫalīl makes perfect sense from what we know about him. The profile which one builds of Ḫalīl from the various sources is that of a man of great talent and sharp intellect. In Zubaydī’s words, no one, other than the prophets, had a mind sharper than Ḫalīl’s (adaqq ḏihnan min al-Ḫalīl).52 He is generally portrayed as an outstanding innovator who, among other things, laid the foundations of Arabic metrics (ʿarūḍ) and was a highly proficient math- ematician and a contributor to musicology.53 It is interesting in this respect that when Ibn al-Muqaffaʿ (d. 142/759) met Ḫalīl, he was reported to have been impressed more by his intellect than his knowledge (ʿaqluhu akṯar min ʿilmihi), whereas Ḫalīl was impressed more by Ibn al-Muqaffaʿ’s knowledge than his intellect (ʿilmuhu akṯar min ʿaqlihi).54 We also know that the bulk of technical terms and analytical tools of the Arabic Grammatical Theory can be traced,

49 Some of these occurrences in the third volume, for example, are found on pp. 58, 92, 381; cf. 26 (qultu li-l-Ḫalīl); 38, 309, 316, 411 (qāla Abū l-Duqayš); 277 (sa⁠ʾaltu Abā l-Duqayš).

50 Cf. Wild (1965: 17–19); Naṣṣār (1968: I, 292); Talmon (1997: 116–117). 51 Abūl-Ṭayyib, Marātib 58. According to Abū l-Ṭayyib, Layṯ nicknamed his tongue “Ḫalīl”

and hence wherever the phrase qāla l-Ḫalīl (but not qāla l-Ḫalīl b. Aḥmad) occurs, it intro- duces material authored by Layṯ! Cf. also Azharī, Tahḏīb I, 28–29.

52 Zubaydī, Istidrāk al-ġalaṭ 45. 53 For more on Ḫalīl’s talent and innovations, see Talmon (1997: 44–51); Neubauer (1998). 54 Zubaydī, Ṭabaqāt 49; Qifṭī, Inbāh I, 380.

 

amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline

 

290 chapter 3

in Sībawayhi’s al-Kitāb, to Ḫalīl.55 In short, the ascription to Ḫalīl of laying the foundations of the first lexicon is certainly consistent with his other achieve- ments. Indeed, from what we know about the philologists of the early period, the one to whom the ascription of proposing the essentials of lexicography would be least surprising is none other than Ḫalīl.

Furthermore, the lexicon itself, as well as its introduction, contain material which clearly bears Ḫalīl’s mark. For example, many of the grammatical views which Sībawayhi attributes to Ḫalīl do occur in al-ʿAyn. One particularly telling example occurs both in the introduction and a relevant lemma. This has to do with biliteral particles such as law and hal whose second radical is geminated when they are treated as nouns, hence lawwun and hallun.56 Sībawayhi’s heavy dependence on Ḫalīl, however, does not negate his intellectual independence given that, on several occasions, he differs with his teacher or argues against a view of his.57 As far as al-ʿAyn is concerned, the most important difference between it and the Kitāb relates to points of articulation and arrangement of letters. Thus, several points of articulation or “locales” (each called ḥayyiz) specified by Ḫalīl (e.g. lahāh “uvula” and ʿakada “root of the tongue”) are not recognized by Sībawayhi,58 and the latter identifies sixteen points of articula- tion, whereas Ḫalīl’s division includes only nine.59 As for the phonetic order of the letters, the arrangement in al-ʿAyn and al-Kitāb is as follows:

al-ʿAyn: ʿ, ḥ, h, ḫ, ġ, q, k, ǧ, š, ḍ, ṣ, s, z, ṭ, d, t, ẓ, ḏ, ṯ, r, l, n, f, b, m, w, alif, y, hamza60 al-Kitāb: hamza, alif, h, ʿ, ḥ, ġ, ḫ, q, k, ḍ, ǧ, š, y, l, r, n, ṭ, d, t, ṣ, z, s, ẓ, ḏ, ṯ, f, b, m, w61

55 Cf. Baalbaki (2008: 16–17). For technical terms in particular, see the long list compiled by ʿAbābina (1984: 158–176).

56 Ḫalīl, ʿAyn I, 50; HL (III, 352); Sībawayhi, Kitāb III, 260–261. This type of gemination also takes place in words which the grammarians traditionally classify as nouns, rather than particles. One such word is the relative pronoun ḏū which Sībawayhi reports, on Ḫalīl’s authority, that as a noun it becomes ḏawwun (ibid., III, 263).

57 Cf. Baalbaki (2008: 17). 58 Ḫalīl, ʿAyn I, 52, 58. Other points of difference between Sībawayhi and Ḫalīl on phonetic

matters are discussed by Talmon (1997: 283–287); see also al-Nassir (1993: 12–17) and Sara (2013: 524).

59 Ḫalīl, ʿAyn I, 58; Sībawayhi, Kitāb IV, 433; cf. Talmon (1997: 284); Carter (2004: 124). 60 Ḫalīl, ʿAyn I, 48, 58. 61 Sībawayhi, Kitāb IV, 431. Note that the above arrangement is cited by Sībawayhi before

the one in which he describes the points of articulation (IV, 433) and is slightly different from it.

 

amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline

 

291Muǧannas (Semasiological) Lexica

If Sībawayhi drew his arrangement from Ḫalīl – as is most likely to be the case, given Sībawayhi’s heavy reliance in his book on his teacher – differ- ences between his arrangement and what we assume to be Ḫalīl’s in al-ʿAyn would be an example of Sībawayhi’s critical approach to some of Ḫalīl’s views. Furthermore, if we were to suppose, solely for the sake of argument, that Sībawayhi’s arrangement was taken from Ḫalīl without introducing any modification to it, this still would not mean that the one in al-ʿAyn’s introduc- tion was not Ḫalīl’s (and hence that the introduction was not his). In fact, the arrangement in al-ʿAyn’s introduction itself contains several inconsistencies. For example, Ḫalīl places ǧ with q and k on one occasion, but with š and ḍ on another.62 He also places d before t and ṯ before ḏ in his initial list, but in the reverse order when he describes the points of articulation.63 Furthermore, his ordering of the three weak letters, alif, wāw and yāʾ, as well as the glottal stop, is not consistent since they occur in four different arrangements: w, alif, y, hamza; w, y, alif, hamza; alif, w, y, hamza; and y, w, alif, hamza.64 Based on al-ʿAyn’s intro- duction, it seems that Ḫalīl did not have a decisive view with respect to the arrangement of letters. To be sure, however, he (or whoever authored the lem- mata) eventually had to adopt a specific order by which to arrange the letters in the lexicon itself. Naṣṣār is probably right to suggest that Ḫalīl’s introduction demonstrates that he continuously pondered the points of articulation and the arrangement of letters, and that Layṯ faithfully reported the various stages of development of his teacher’s views.65

The abundance of material on meters and rhyme in al-ʿAyn also lends sup- port to the view that Ḫalīl is the author of the lexicon. The thirty-eight loci which contain prosodic matters66 are significantly spread over the whole book, and do not only occur in the first few letters which some, as previously noted, claim are the only parts authored by Ḫalīl. Given the scarcity in al-ʿAyn of technical terms from other fields of study (a notable exception being music, one of Ḫalīl’s other areas of interest, since the lexicon contains eleven refer- ences to musical notions and terms),67 the presence of thirty-eight references

62 Ḫalīl, ʿAyn I, 52, 58. 63 Ibid., 48, 58. 64 Ibid., 48, 57, 58 (l. 4–5), 58 (l. 13). 65 Naṣṣār (1968: I, 245). 66 Cf. Talmon (1997: 117). Metrical matters occur in I, 119, 212, 214, 245, 251, 275, 299; III, 67, 139,

233, 345, 384, 385; IV, 72, 79, 158, 208, 260, 279; V, 53, 222, 283, 328, 415; VI, 12, 64, 163, 241; VII, 51, 127, 130, 218, 223, 229, 468; VIII, 17, 264.

67 Ḫalīl, ʿAyn I, 89, 359; III, 229, 230; V, 306; VI, 3, 48; VII, 41, 51, 146; VIII, 243; cf. Talmon (1997: 118).

 

amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline

 

292 chapter 3

to prosodic matters is indeed remarkable. Other than the more general terms (including names of meters) such as ʿarūḍ, qarīḍ, šiʿr, qāfiya, wazn, rawiyy, taṣrīʿ, ziḥāf, maǧzūʾ, mašṭūr, hazaǧ, munsariḥ, raǧaz, muǧtaṯṯ, basīṭ and madīd, highly technical terms of the science of ʿarūḍ, which Ḫalīl is generally cred- ited with being its founder, include muḫallaʿ, iḍǧāʿ, mutašaʿʿiṯ, mawqūf, ḫuzla, aḫram, maḫbūn, makfūf, mutadārak, ikfāʾ, muḍamman, fāṣila, fāḍila, muṣallam, musammaṭ, manhūk, sinād, īṭāʾ and muraffal. From a wider perspective, it is essential to note the great affinity between Ḫalīl’s system of permutations and the five circles (dawāʾir) which form an essential part of his theory of ʿarūḍ. As an example, we can take the fourth circle called dāʾirat al-muǧtalab. This can be illustrated by a circle which includes, based on which point among several in the circle one chooses to start, nine theoretically possible meters. Of these, only six are actually used (mustaʿmal), namely, sarī ʿ, munsariḥ, ḫafīf, muḍāriʿ, muqtaḍab and muǧtaṯṯ, whereas the other three are not used (muhmal).68 This is strikingly similar to Ḫalīl’s permutations of roots. If one imagines a circle which has, say in the case of triliterals, three starting points, each representing a letter, such as ḥ, ṭ and l, the various combinations become obvious and one discovers – as in the case of meters – that some are mustaʿmal while others are muhmal or mulġā.69 In the case of ḥ, ṭ and l,70 ṬLḤ, ṬḤL, LṬḤ and ḤLṬ are used and thus correspond to sarīʿ, munsariḥ, etc. in metrics. On the other hand, LḤṬ and ḤṬL are not used, and hence correspond to the three unused meters of the muǧtalab circle.

Ḫalīl must have realized that the method of collecting lexical items related to a specific field, as is the practice in mubawwab lexica, is essentially different from that which is required in a lexicon which could first embrace all the roots of the language, be they used or not, and then list the various derivatives of each used root. Indeed, the first of these two steps is not only more basic but would also result in an exact enumeration of the theoretically possible roots (see below), whereas the listing of derivatives of used roots is not meant to be exhaustive, at least not in every entry. Ḫalīl’s main contribution surely lies in the system which he devised so that no Arabic root would escape the notice of any philologist who might intend to author a lexicon that exhausts all roots. This system rests on three basic principles, each of which embraces a finite set of elements and is reflected in the structure of the lexicon. The first principle is that the letters of the alphabet form a closed set. Ḫalīl determines that this set

68 Tibrīzī, Wāfī 175–179 (cf. introduction 14–16); Sakkākī, Miftāḥ 521–522; cf. Zamaḫšarī, Qisṭās 155.

69 See Ibn Durayd’s representation of radicals as points on a circle in Ǧamhara III, 1338. 70 Ḫalīl, ʿAyn III, 169–171.

 

amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Underline

 

293Muǧannas (Semasiological) Lexica

is made up of twenty-nine elements (i.e not twenty-eight as is almost univer- sally acknowledged later, although some authors introduce the lām alif as the twenty-ninth element), given the distinction he makes between the glottal stop (hamza) and the long vowel ā or “soft” alif (al-alif al-layyina).71 These letters obviously have to be arranged in a specific order so that the roots – of which they form the building blocks – can be accordingly arranged.72 In this respect, Ḫalīl chose not to follow the two arrangements which were known in his time. The first, ʾ , b, ǧ, d, h, w, z, etc., is of foreign origin and is used in calculation (ḥisāb al-ǧummal), whereas the second, ʾ, b, t, ṯ, ǧ, ḥ, ḫ, etc., is derived from the first and arranged according to the forms of letters.73 One cannot determine with certainty the specific reason for which Ḫalīl did not choose either system, but he most probably found that both lacked any linguistic justification for them to suit his lexicon, in contrast to his phonetic arrangement which clearly has a solid linguistic basis. His decision to devise a phonetic arrangement despite the presence of at least two alternatives is yet another proof of his innovative approach and independent thought and is the fruit of the phonetic views he formed as a result of examining the various points of articulation and charac- teristics of letters. The lexicon itself is divided into chapters based on Ḫalīl’s arrangement, although it should be noted that w, alif, y and hamza are placed together in one chapter at the end of the book.

The second principle is that the number of radicals (i.e. consonants) in Arabic words ranges from two to five. Hence there is a closed set which consists of four elements based on the number of radicals, namely, the biliteral (e.g. qad and lam), the triliteral (e.g. ḍaraba and ǧamal), the quadriliteral (e.g. daḥraǧa and ʿaqrab), and the quinqueliteral (e.g. iqšaʿarra and safarǧal).74 Ḫalīl’s examples demonstrate that the biliterals are exclusively particles, whereas the triliterals, quadriliterals and quinqueliterals include both verbs and nouns. Accordingly, all three parts of speech recognized in the grammatical theory fall within the parameters of the four constituents of this closed set. This mor- phological feature is also reflected in the division of the lexicon, as each bāb (i.e. ʿ, ḥ, h, ḫ, etc.) is divided into four sections, which begin with the biliteral

71 Ibid., I, 57. 72 For the views of Mubarrad and al-Ṣāḥib b. ʿAbbād as to why al-ʿAyn begins with the letter

ʿayn, see below, 320. 73 The Arab origin of this arrangement is noted in some sources, such as Dānī, Muḥkam 27

ff. See also p. 30, where the difference between the Eastern and Western arrangements is noted.

74 Ḫalīl, ʿAyn I, 48–49.

 

amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline

 

294 chapter 3

and end with the quinqueliteral (if available).75 Also based on morphological grounds is the internal division of the sections of triliterals into ṣaḥīḥ (sound), muʿtall (unsound) and lafīf (containing two weak radicals). For example, following the ṣaḥīḥ roots of the letter ǧ, such as the roots which contain ǧ, š, and r (e.g. ŠǦR and ǦRŠ),76 are chapters which list the muʿtall and the lafīf. The former refers to roots that include w, alif, y or hamza, such as ŠǦW, WŠǦ, ǦYŠ and ǦŠʾ, all of which have š and ǧ in common.77 Thus the muʿtall is a general notion which comprises several types of roots, including the defective (nāqiṣ, here ŠǦW), the initially weak (miṯāl, here WŠǦ), the hollow (aǧwaf, here ǦYŠ), and the hamzated (mahmūz, here ǦŠʾ). As for the lafīf, it is used in a broader sense than is traditionally the case in morphological sources because it not only includes roots which have w or y, both presumed to have shifted from “soft” alif (e.g. ǦWW, ǦWY, WǦY and WYǦ), but also those which either have one or more hamzas (e.g. ǦʾW, ʾǦʾ and ǦYʾ, the latter two of which are normally referred to by morphologists as mahmūz) or, more strangely, those which have only one w, alif, y or hamza (e.g. WǦǦ, ʾǦǦ and ǦʾǦ, the first two of which are normally called muḍaʿʿaf/muḍāʿaf “reduplicated”).78

Having established that roots are finite in number, and given that the letters from which they are formed are also finite, as expounded in the lexicon’s intro- duction, Ḫalīl proceeds to determine how those two closed sets can contribute to the purpose of exhausting all possible roots which Arabic embraces. The notion of taṣarruf (which occurs in its verbal form, i.e. tataṣarraf )79 is what he finally proposes to achieve this purpose. This refers to the permutations which in later works are normally called taqālīb.80 Of these, there are two in biliterals (al-kalima al-ṯunāʾiyya tataṣarraf ʿalā waǧhayn), as in qad and daq or šad and daš. In triliterals, quadriliterals and quinqueliterals, the number of possible permutations is six, twenty-four, and one hundred twenty respectively. Ḫalīl notes that some of the resulting roots are not actually used and that, particu- larly in quinqueliterals, the majority of combinations are abolished ( yustaʿmal aqalluhu wa-yulġā akṯaruhu). The lemmata, needless to say, are restricted to

75 See, for example, VIII, 370, 404, 410, 420, 436, where it is noted that no quadriliterals or quinqueliterals are found in the case of lām, nūn, fāʾ, bāʾ and mīm respectively.

76 Ibid., VI, 30–35. 77 Ibid., VI, 156–159. 78 See the examples cited above in the chapter of lafīf under the letter ǧ in ʿAyn VI, 196–199. 79 Ibid., I, 59. 80 Cf. Ibn Durayd, Ǧamhara III, 1338, where qalb is used, and Ibn Ǧinnī, Ḫaṣāʾiṣ II, 133–139,

where the six taqālīb which result from three radicals are discussed as part of the theory of al-ištiqāq al-akbar (cf. above, 281).

 

amgad
Underline
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Highlight
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Highlight

 

295Muǧannas (Semasiological) Lexica

roots that are in use, but it should be noted that the author often mentions, in the beginning of the lemmata of biliterals and triliterals, the unused per- mutations. But this is not so in quadriliterals and quinqueliterals where only used permutations are mentioned since unused ones are much more numer- ous. Also noteworthy is the fact that, due to the application of the principle of taqālīb, the early letters include a much larger number of roots than later letters. Since any given root is listed in the chapter of the radical whose alpha- betical (i.e. phonetic) order precedes that of the other radicals in that root, the number of roots that remain for successive letters becomes increasingly smaller. For example, the chapter of f, b and m – i.e. the last three consonants placed just before w, alif, y and hamza – occupy no more than thirty-two pages of the eighth volume,81 whereas the first three letters (i.e. ʿ, ḥ and h) occupy almost 1,200 pages in more than three volumes.

Based on the above three principles, Ḫalīl accomplished his mission of exhausting all possible Arabic roots – a mission which is highlighted in the first few lines of the introduction where Layṯ (or another disciple of Ḫalīl’s) explains that Ḫalīl set out to exhaust all possible combinations of consonants so that none would escape his count (fa-lā yaḫruǧ minhā ʿanhu šayʾ; fa-lā yašiḏḏ ʿanhu šayʾ).82 All subsequent muǧannas lexica are indebted to Ḫalīl for devising a scheme that achieves exhaustiveness, and in spite of the various differences between al-ʿAyn and many of these lexica in the arrangement of letters, the division of the material based on the number of radicals of roots, and the observance of the taqālīb, none of their authors had to invent – as Ḫalīl did – on his own, a complete plan without recourse to an earlier one. In other words, al-ʿAyn represents the discovery stage in Arabic lexicography, after which there merely were modifications in certain aspects – such as alphabeti- cal arrangement, internal divisions, extent of šawāhid, etc. – but no alterna- tive scheme which had to be invented, from scratch, in all its details. We have already argued that Ḫalīl’s painstaking efforts at establishing the principles upon which his lexicon rests are reflected in certain phrases or terms which appear in al-ʿAyn’s introduction.83 The phrase fa-aʿmala fikrahu (He put his mind to work) and the term ḏawāq (sampling of letters)84 clearly demonstrate a process of an innovative nature, as do the technical terms which are used in

81 Ḫalīl, ʿAyn VIII, 405–436. 82 Ibid., I, 47. Note that some authors attribute to Ḫalīl specifying the exact number of used

and unused roots in biliterals, triliterals, quadriliterals and quinqueliterals; cf. Suyūṭī, Muzhir I, 74. See also Ibn Durayd, Ǧamhara III, 1338–1339 and Zubaydī, Muḫtaṣar II, 457.

83 Cf. above, 58. 84 Ḫalīl, ʿAyn I, 47.

 

amgad
Underline

 

296 chapter 3

the introduction but which still preserve their general, non-technical mean- ing. Further proof of a fresh attempt at coining terms is apparent in Ḫalīl’s use of either of two terms which express the same notion, as in hatta, which he replaced once by hahha (wa-qāla marratan hahha), and šafawiyya, substi- tuted once by šafahiyya (wa-qāla marratan šafahiyya).85 Finally, it should be pointed out that, according to a quotation by Layṯ, Ḫalīl used the expression fī muʾallafinā (in our composed work). If the quotation is accurate, it could indi- cate that Ḫalīl “had begun to write a proper book for readers, more particularly for dictionary users”86 – a huge step in the history of Arabic scholarship.

The purpose of exhausting all possible roots, which lies at the heart of the system devised by Ḫalīl, seems to have prompted him to be cautious of allow- ing into the linguistic corpus lexical items that are not Arabic. Indeed, in the first few lines of the introduction, it is clearly indicated that Ḫalīl was inter- ested in exhausting the speech of the Arabs (kalām al-ʿArab),87 and in the final few lines he is quoted as saying that his lexicon would embrace the speech of the Arabs, be it clear or strange (in meaning) (ḥattā nastawʿib kalām al-ʿArab al-wāḍiḥ wa-l-ġarīb).88 He also recognizes the importance of discovering the criteria for differentiating between Arabic and borrowed words (li-yuʿraf ṣaḥīḥ bināʾ kalām al-ʿArab min al-daḫīl).89 The insistence on specifying Arabic usage may seem to be a case of stating the obvious, but it should be viewed within a wider framework which the introduction itself provides, namely, the intention of the so-called naḥārīr (pl. of niḥrīr, skillful or learned) to cause confusion and obfuscation (irādat al-labs wa-l-taʿnīt) by creating neologisms (muwalladāt) which, despite their resemblance to other words and patterns (ašbaha lafẓahum wa-ta⁠ʾlīfahum) are impermissible in the speech of the Arabs.90 In order to detect such material coined by the naḥārīr, Ḫalīl exam- ined the rules of word composition in Arabic. These, for example, stipulate that all quadriliterals and quinqueliterals must not be devoid of liquids or labi- als, except for about ten anomalous (šawāḏḏ) words, of which he mentions six (e.g. ʿasǧad “gold” and qudāḥis “brave and strong”).91 Accordingly, words such as kašaʿṯaǧ, ḫaḍaʿṯaǧ and kašaʿṭaǧ are neologisms that are not to be consid- ered Arabic. Furthermore, Ḫalīl notes the restrictions on the co-occurrence of

85 Ibid., I, 57–58. 86 Ibid., I, 60. Cf. Schoeler (2006: 148, 151). 87 Ibid., I, 47. 88 Ibid., I, 60. 89 Ibid., I, 54. 90 Ibid., I, 52–53; cf. above, 28, 59, 162. 91 Ibid., I, 53; cf. II, 286.

 

 

297Muǧannas (Semasiological) Lexica

certain consonants in the same root and observes that no Arabic root begins with n-r. The validity of such rules is also supported by samāʿ or attested data since Ḫalīl asserts, for example, that words similar to kašaʿṯaǧ are nowhere to be heard in the speech of the Arabs (lasta wāǧidan man yasmaʿ min kalām al-ʿArab . . .).92 This confident statement of Ḫalīl’s strongly implies extensive knowledge of the various Arabic dialects either through his own initiative of deriving linguistic data from the deserts (bawādī) of Ḥiǧāz, Naǧd and Tihāma93 or, more probably, through the collective effort of philologists in the process known as ǧamʿ al-luġa (collection of linguistic data) discussed in Chapter 1.94 Given that ġarīb or strange/rare usage was eagerly sought after by philologists during that process, it is not at all surprising that Ḫalīl puts this type of mate- rial on an equal footing with more common or widespread usage to which he refers, in the above-quoted statement at the end of his introduction, as wāḍiḥ. The criterion for inclusion of lexical items in the lexicon is thus clear: not only do they have to resemble other lexical items used by the Arabs but also have to be actually used, irrespective of whether they are classified as wāḍiḥ (hence widely used) or ġarīb (hence less commonly used). It is interesting to note that, as evidence of actual usage, Ḫalīl at times specifically mentions that a lexical item is used in his own native city of Basra but not by the Bedouins,95 or that it has come in his own time to acquire a specific meaning associated with a group of speakers.96

In spite of occasional reference to specific dialects in al-ʿAyn, most of the material is not ascribed to any group of users. Several informants about whom we know very little (if at all) from the sources are also cited97 and are most likely to be among the eloquent Aʿrāb who often supplied data to philologists. The other sources obviously include poetry, Qurʾān, Ḥadīṯ and proverbs,98 as

92 Ibid., I, 52. 93 Sīrāfī, Aḫbār 33; Ibn al-Anbārī, Nuzha 47; Suyūṭī, Buġya II, 365; cf. above, 17. 94 Cf. above, 16–29. 95 For example, nawā l-ʿaqūq (easily broken date stones) is said to be used by the Basrans,

but not the Aʿrāb in their deserts (I, 63), and ḥarrāqāt (ships from which fire is thrown at the enemy) is described as Basran usage (III, 44).

96 Miḥrāb, for example, is said to refer “among the common people nowadays” (ʿinda l-ʿāmma al-yawma) to the chamber of the imām in a mosque (III, 214). Similarly, skill- ful Iraqi money-changers (ṣayārifat al-ʿIrāq) are reported to have become fond of using (lahiǧa bihā) the originally Ḥīrī ʿIbādī (Ḥimyariyya ʿIbādiyya; the first word obviously cor- ruption of Ḥīriyya) word šašqala for weighing dinars (V, 245).

97 E.g. Abū Aḥmad, Ḥammās, Mūsā and Šuǧāʿ. Cf. the indices of al-ʿAyn prepared by M. Āl ʿUṣfūr (see secondary sources); cf. also Wild (1965: 17–18); Naṣṣār (1968: I, 292–293).

98 Cf. Wild (1965: 42–57).

 

 

298 chapter 3

is the case of early mubawwab lexica discussed in the previous chapter or in the major grammatical work of the period, Sībawayhi’s al-Kitāb. It should be noted, however, that many of al-ʿAyn’s poetry šawāhid strangely do not occur in any later lexicon99 – perhaps due to the previously mentioned allegation that it contains šawāhid that are attributed to poets whom early philologists do not normally cite.100 As far as Ḥadīṯ is concerned, it has already been pointed out that, perhaps due to the different perspectives which the naḥwiyyūn and luġawiyyūn had, the former group, at least in the early stages of the tradition, were much more cautious than the latter in citing it.101 This is evident in the fact that, compared to the seven or eight citations of Ḥadīṯ in Sībawayhi’s al-Kitāb, al-ʿAyn includes 428 ḥadīṯs and aṯars.

In contrast to the exhaustiveness of the roots in al-ʿAyn as previously explained, the derivatives which are listed under these roots are most certainly not exhaustive. Accordingly, later works cite a large number of lexical items – including proper nouns and place names – which do not occur in al-ʿAyn. Moreover, the explanation of the lexical items is mostly characterized by brev- ity, and the lemmata contain very few digressions, unlike several later lexica. Some lemmata, however, contain morphological and syntactical comments which can be relatively lengthy in comparison with the brief definitions of words.102 The root derivatives that are normally mentioned are the verb in its perfect and imperfect forms, the infinitive, the active participle and the assimilate adjective (ṣifa mušabbaha), in addition to plural forms. The verb is frequently placed in the beginning (intransitive usually before transitive, unaugmented usually before augmented), followed by the infinitive(s) and the participle, as in ḫaḏiya l-ḥimāru yaḫḏā ḫaḏan fa-huwa aḫdā and ḫalā yaḫlū ḫalāʾan fa-huwa ḫālin.103 But this is by no means a standard template in all lemmata. Often the infinitive is placed before the verb, as in al-baḏḫ . . . baḏaḫa yabḏaḫu baḏḫan wa-buḏūḫan wa-raǧulun bāḏiḫun wa-baḏḏāḫun,104 or is cited without the verb, as in the root ḪḎF.105 At times there is also an attempt at

99 Cf. Baalbaki (1998: 57). Note also that al-ʿAyn’s editors frequently mention in the footnotes that a certain šāhid is unique to this lexicon. See, however, Baalbaki (1986: 87–96) for errors in the ascription of some of al-ʿAyn’s šawāhid by its editors.

100 See above, 284. 101 See above, 29–30, 72. 102 See, for example, the comments on the patterns of ašyāʾ (VI, 295–296) and fam (VIII, 406).

See also the list of grammatical comments in al-ʿAyn extracted by Talmon (1997: 127–214). 103 Ḫalīl, ʿAyn IV, 297, 306. 104 Ibid., IV, 246. 105 Ibid., IV, 245, where ḫaḏf is cited, but not ḫaḏafa yaḫḏifu. Note also expressions such as

wa-lam nasmaʿ fīhi fiʿlan and lā yuštaqq minhu fiʿl (III, 22; IV, 246), which indicate that a certain noun has no corresponding verb.

 

 

299Muǧannas (Semasiological) Lexica

showing the semantic relationships between derivatives of the same root, as in qaṯāmi (hyena) which is allegedly so named because it gets tarnished by its own excrement – an act expressed by the infinitive qaṯm.106

The absence in al-ʿAyn of a rigorous system for the internal arrangement of the lemmata as well as for the inclusion of all used derivatives of the same root obviously left its mark on the lexicographical tradition as a whole. Despite some improvements in certain lexica from this perspective, the user often has to read almost the whole lemma in order to find the required word. Furthermore, authors of muǧannas lexica after al-ʿAyn largely retained some of its other main features. For example, they kept taking into consideration in their arrange- ment only the root of the word without any affixes (although some mubawwab lexica did take affixes into account in their ordering)107 and adopted the same corpora as al-ʿAyn for deriving their data, with little interest in newly coined words and technical terms or new meanings of old words. More specifically, Ḫalīl’s method of phonetic arrangement and permutations was almost entirely adopted in the fourth/tenth and fifth/eleventh centuries by Qālī (d. 356/967) in al-Bāriʿ, Azharī (d. 370/981) in al-Tahḏīb, al-Ṣāḥib b. ʿAbbād (d. 385/995) in al-Muḥīṭ, and Ibn Sīda (d. 458/1066) in al-Muḥkam. Other fourth/tenth century authors, on the other hand, gave up Ḫalīl’s phonetic arrangement in favor of the alphabetical one, but were still influenced by him either in the method of taqālīb – i.e. Ibn Durayd (d. 321/933) in al-Ǧamhara – or in dividing the mate- rial into chapters according to the number of radicals in words – i.e. Ibn Fāris (d. 395/1004) in al-Maqāyīs and al-Muǧmal.108

The influence of al-ʿAyn on subsequent authors is also evident in the large number of commentaries, abridgements or expansions to which it gave rise. Most of these have not reached us,109 although in certain cases information is available from quotations in other works. For example, Azharī’s harsh attack on Aḥmad b. Muḥammad al-Buštī al-Ḫārzanǧī (d. 348/959) contains relatively lengthy quotations from the latter’s supplement to al-ʿAyn, which is entitled Takmilat K. al-ʿAyn and follows al-ʿAyn’s arrangement.110 Judging by the numer- ous quotations made by al-Ṣāḥib b. ʿAbbād (d. 385/995) in al-Muḥīṭ fī l-luġa

106 Ibid., V, 140. 107 E.g. Kurāʿ al-Naml (d. 310/922) in al-Muǧarrad (cf. above, 92) and Siǧistānī (d. 330/941) in

Ġarīb al-Qurʾān (cf. above, 67). 108 For more details on al-ʿAyn’s influence on subsequent authors, see Wild (1965: 58–91);

Fleisch (1984: 173–188); Ǧubūrī (2004: 84–88). 109 For a list of such works, see Wild (1965: 21–25); Naṣṣār (1968: I, 296–312); Sezgin (1982: VIII,

53–56). 110 Azharī, Tahḏīb I, 32–40.

 

 

300 chapter 3

from al-Takmila,111 the latter must have contained vast material not found in al-ʿAyn. Azharī also devastatingly attacks Abū l-Azhar al-Buḫārī’s (fl. in the second half of the third/ninth century) K. al-Ḥaṣāʾil, in which he included material not mentioned in al-ʿAyn.112 In this case, Azharī says that the book is worse than Ḫārzanǧī’s and decides not to quote any of its content. This notwithstanding, Qifṭī (d. 646/1248) – who examined the first volume of al-Ḥaṣāʾil – vindicates its author against Azharī’s claims and asserts that it contains material ignored by Ḫalīl, mostly in the letter ʿayn, and that it does not repeat al-ʿAyn’s text except where necessary.113 Also based on al-ʿAyn, but arranged alphabetically according to the final letters of roots, is Ibn al-Tayyānī’s (d. 436/1044) al-Mūʿab, which was discussed earlier with works that deal with nominal and verbal patterns.114

Among the extant sources related to al-ʿAyn are two of Zubaydī’s works, namely, Istidrāk al-ġalaṭ al-wāqiʿ fī K. al-ʿAyn and Muḫtaṣar al-ʿAyn. In the for- mer’s introduction,115 Zubaydī sets forth his arguments against the attribu- tion of al-ʿAyn to Ḫalīl, as previously noted. He also explains that in an earlier book of his in which he amended the mistakes of his teacher, Abū ʿAlī al-Qālī (d. 356/967), in al-Bāriʿ, he mentioned more mistakes of al-ʿAyn than in his present book.116 Zubaydī follows al-ʿAyn’s arrangement of letters and sec- tions into which each letter (i.e. chapter) is divided, and the errors he cites mostly have to do with the alteration of diacritical points (taṣḥīf) and with morphology – a fact already noted by Suyūṭī (d. 911/1505).117 An example of taṣḥīf is that saḥb (vehemence of eating or drinking) in al-ʿAyn, according to Zubaydī, is a corruption of saḥt.118 Morphological errors, on the other hand, pertain mostly to words which do not represent any known pattern – such as ihbayyaḫa (to strut), whose pattern if ʿayyala, according to Zubaydī, is not used in verbs119 – and words which are classified under the wrong section based on their etymology – such as walwala (ululation), which Zubaydī says is

111 Cf. below, 321. 112 Ibid., I, 40. 113 Qifṭī, Inbāh IV, 99. 114 See above, 263. 115 This introduction was first edited separately by Wadġīrī (1984: 139–146). Most of it is also

preserved by Suyūṭī (Muzhir I, 79–86), who also quotes some of the book’s material (ibid., II, 381–390).

116 Zubaydī, Istidrāk al-ġalaṭ I, 49–50. 117 Suyūṭī, Muzhir I, 86. 118 Zubaydī, Istidrāk al-ġalaṭ 64–65; cf. Ḫalīl, ʿAyn III, 151. 119 Ibid., 82; cf. Ḫalīl, ʿAyn III, 359.

 

 

301Muǧannas (Semasiological) Lexica

reduplicated and not derived from the triliteral root WYL as in al-ʿAyn,120 and sarandā (brave man), which Zubaydī believes to be of the triliteral root SRD, whereas it is listed in al-ʿAyn under quadriliterals.121 To be noted, however, is that the published text of al-ʿAyn often contrasts with what Zubaydī attributes to its author. For example, Zubaydī claims that dalanẓā (huge camel) is listed under quadriliterals in al-ʿAyn, whereas he asserts that it is derived from the triliteral root DLẒ by the affixation of nūn.122 In fact, not only is it placed with triliterals in al-ʿAyn but it is specifically said to be derived from DLẒ.123 This dis- crepancy is most probably due to difference in manuscripts, particularly since we know from Baṭalyawsī (d. 521/1127) that Zubaydī’s copy of al-ʿAyn can differ from all other known copies.124

Authored before Istidrāk al-ġalaṭ, the other book by Zubaydī is an abridge- ment of al-ʿAyn called Muḫtaṣar al-ʿAyn. It seems to have been quite popular for its brevity and is said to have been one of four abridgements which were generally preferred to their original texts.125 Zubaydī maintains in his abridge- ment what he describes as the core of al-ʿAyn (tuʾḫaḏ ʿuyūnuhu)126 and thus omits the regular (qiyāsī) forms of the infinitive and imperfect,127 as well as poetry šawāhid, although he preserves a few from Qurʾān and Ḥadīṯ. As far as arrangement is concerned, Zubaydī follows al-ʿAyn’s phonetic order of letters, but introduces a few changes to the divisions of each letter, mainly on morpho- logical grounds. For example, he designates a section, which follows the one on sound triliterals, to unsound reduplicated biliterals (al-ṯunāʾī al-muḍāʿaf min al-muʿtall). These normally occur in al-ʿAyn under the lafīf in triliterals. For example, ḫawḫ (peach) and waḫwaḫa (imitation of the sounds of certain birds), which are listed in al-ʿAyn under the triliteral lafīf of the letter ḫ,128 are moved by Zubaydī to the section devoted to unsound reduplicated biliterals, along with words containing ḫ and ʾ or ḫ and y.129 Zubaydī also reorganizes the sections of unsound triliterals (al-ṯulāṯī al-muʿtall) by successively listing roots with hamza, yāʾ and wāw, whereas they are listed in al-ʿAyn in no specific order.

120 Ibid., 80; cf. Ḫalīl, ʿAyn III, 343. 121 Ibid., 185; cf. Ḫalīl, ʿAyn VII, 340. 122 Ibid., 201. 123 Ḫalīl, ʿAyn VIII, 18. 124 Baṭalyawsī, Muṯallaṯ II, 337. 125 Suyūṭī, Muzhir I, 87. 126 Zubaydī, Muḫtaṣar I, 41. 127 Ibid., II, 457. 128 Ḫalīl, ʿAyn IV, 317–318. 129 Zubaydī, Muḫtaṣar I, 460.

 

 

302 chapter 3

Furthermore, he often supplements al-ʿAyn’s text with lexical items ignored by its author.130

Another book with the title Muḫtaṣar al-ʿAyn is authored by Abū l-Ḥasan ʿAlī b. al-Qāsim al-Ḫawāfī, a scholar of the third/ninth century.131 The book fol- lows al-ʿAyn not only in its phonetic ordering of letters but also in the internal arrangement of each chapter. Yet it differs significantly from al-ʿAyn in con- tent. The author explains in the introduction132 that he is interested in what he calls al-ḍarb al-awsaṭ (lit. middle type), that is, lexical material which is neither strange and obscure (al-ġarīb al-ġāmiḍ) nor commonplace and easily under- stood by the generality of people (al-qarīb al-sahl allaḏī tafhamuhu l-ʿāmma). Accordingly, he deletes from the origin a substantial number of roots none of whose derivatives he seems to have found consistent with his definition of al-ḍarb al-awsaṭ. More importantly, Ḫawāfī’s purpose is more than to simply abridge al-ʿAyn, and he makes it clear in the introduction that he wants his book to include the ġarīb material of the Quʾrān and Ḥadīṯ as well as nawādir (i.e. rare usage – hardly compatible with his al-ḍarb al-awsaṭ!) and proverbs. The book abounds with lexical items that are related to jurisprudence ( fiqh) and ġarīb al-Qurʾān/al-Ḥadīṯ. Unlike Zubaydī, Ḫawāfī includes in his text a large number of šawāhid which are not found in al-ʿAyn. These are not restricted to his main focus (i.e. Qurʾān and Ḥadīṯ) but also include a number of poetry šawāhid. In short, the work greatly resembles al-ʿAyn in form and arrangement, and its author’s abridgement of some material is compensated for by sizable additions, in line with his assertion in the introduction that he wants his book to be exhaustive of most of what a scholar might need (ḥattā yakūn kitāban ǧāmiʿan li-ǧull mā yaḥtāǧ ilayhi man yumāris al-ʿilm).133

We are also in possession of al-Ḫaṭīb al-Iskāfī’s (d. 420/1029) Muḫtaṣar K. al-ʿAyn. It begins with an introduction that summarizes some of the essential points in Ḫalīl’s introduction (e.g. biliterals, triliterals, etc., sound and unsound roots, and phonetic ordering of letters),134 but which does not indicate the pur- pose or method of the abridged version. Like Ḫawāfī, Iskāfī follows al-ʿAyn’s phonetic arrangement as well as the internal division of its chapters. Other than the omission of morphological and syntactical comments, his abridge- ment rests mainly on the reduction of the number of šawāhid, particularly

130 According to the editor of al-Muḫtaṣar, these lexical items amount to 240; cf. I, 22 and the detailed list in II, 570–575.

131 See the editor’s introduction to Ḫawāfī, Muḫtaṣar I, 1–2. 132 Ḫawāfī, Muḫtaṣar I, 1. 133 Ibid., loc. cit. 134 Iskāfī, Muḫtaṣar I, 71–74.

 

 

303Muǧannas (Semasiological) Lexica

from poetry, and the abbreviation of the definitions of lexical items. The importance of the work, however, truly lies in its inclusion of material which does not occur in the known version of al-ʿAyn. Given that Iskāfī’s own addi- tions (which he marked in the text) are very few, the editor convincingly argues that the considerable amount of material which al-Muḫtaṣar contains and which occur neither in the extant manuscripts of al-ʿAyn nor in its published text goes back to an early copy of al-ʿAyn.135 Whether or not that copy is the one which Ḫalīl wrote (or dictated to Layṯ), as the author suggests, is obviously impossible to substantiate.

2.2 al-Bāriʿ fī l-luġa by Abū ʿAlī al-Qālī (d. 356/967) Authored almost two centuries after al-ʿAyn, the first muǧannas lexicon in Arab Spain, al-Bāriʿ, was heavily influenced by Ḫalīl’s system.136 According to Zubaydī (d. 379/989), al-ʿAyn was introduced to Andalus by Qāsim b. Ṯābit al-Saraqusṭī (d. 302/915), author of al-Dalāʾil fī ġarīb al-Ḥadīṯ,137 and his father Ṯābit b. ʿAbdalʿazīz (also known as Ṯābit b. Ḥazm; d. 313/925), who outlived his son.138 By the reign in Andalus of al-Mustanṣir al-Umawī (350/961 – 366/976), there were many manuscripts of al-ʿAyn, and it is reported that al-Mustanṣir asked a number of scholars, including Qālī, to compare between those manu- scripts, some of which were full of mistakes.139 But even before al-Mustanṣir’s request, Qālī had already spent several years in authoring al-Bāriʿ. Qifṭī (d. 646/1248) mentions a statement written by Qālī’s son in which he specified the month of Raǧab of the year 339 A.H. as the date when his father had started to write his lexicon.140 Following an unspecified period of difficulties which distracted Qālī from continuing the work he had started, he resumed writing al-Bāriʿ in 349 A.H., and when he was through with the draft in the month of Šawwāl in 355 A.H., he started copying it, but died in the following year. The rest of the draft according to Ibn al-Abbār (d. 658/1260) was then copied by two scribes, Muḥammad b. al-Ḥusayn al-Fihrī (d. ?) and Muḥammad b. Maʿmar al-Ǧayyānī (d. 377/987).141 The two scribes presented the finished manuscript to al-Mustanṣir who wanted to find out by how many words it exceeded the

135 Introduction to Iskāfī, Muḫtaṣar I, 34–35. 136 For a general study on Qālī and his influence on linguistic and literary studies in Andalus,

see Wadġīrī (1983). 137 Cf. above, 76. 138 Zubaydī, Ṭabaqāt 284. 139 Ḥumaydī, Ǧaḏwa 51–52; cf. Qifṭī, Inbāh III, 71–72. 140 Qifṭī, Inbāh I, 244. 141 Ibn al-Abbār, Takmila I, 106; cf. Marrākušī, Ḏayl VI, 175.

 

amgad
Highlight
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline

 

304 chapter 3

manuscript of al-ʿAyn which Qālī and his colleagues had decided was the best one available. Given Qālī’s profound knowledge of al-ʿAyn and its various man- uscripts, al-Mustanṣir’s query implies that his lexicon was based on al-ʿAyn and that it would be easy to compare the two works to determine the number of new words introduced by Qālī. In Ibn al-Abbār’s riwāya, the exact answer given to al-Mustanṣir was 5,683 words. Further evidence related to the comparison between the two lexica comes from Ibn Ḫayr (d. 575/1179), who tells us that Qālī’s lexicon was larger than al-ʿAyn by more than 400 pages and that Qālī’s additions were of two types: supplying the words (or roots) which Ḫalīl had left out, and broadly expanding the material which, in al-ʿAyn, was either too brief or not supported by šawāhid.142

Qālī’s knowledge of al-ʿAyn certainly goes back to the period which he spent in Baghdad between 305/918 and 328/940 before he left to Andalus.143 His fore- most teacher during that period was none other than Ibn Durayd (d. 321/933), author of the famous Ǧamharat al-luġa. Ibn Ḫayr mentions tens of books which Qālī studied with his teacher, including al-Ǧamhara,144 itself influenced by al-ʿAyn. But despite the similarity between al-ʿAyn and al-Ǧamhara, there were enough differences between the two to offer Qālī two distinct models of arrangement, at least in two major aspects: Ḫalīl’s phonetic arrangement and his division of each chapter (i.e. letter) into biliterals, triliterals, etc., and Ibn Durayd’s alphabetical arrangement and his division of the whole lexicon and not individual letters according to root length. In both aspects, Qālī fol- lowed Ḫalīl, and although the choice of adopting either Ḫalīl’s or Ibn Durayd’s method of structuring individual chapters or the whole book based on root length would have had little effect on the user of al-Bāriʿ, his choice of adher- ing to Ḫalīl’s phonetic arrangement is rather surprising, given his teacher’s conviction145 – as well as the conviction of other contemporary scholars, such as Ibn Wallād (d. 332/944)146 – that the alphabetical system is easier and more accessible to users. Yet it seems that the authoritative status of al-ʿAyn made it a more appealing model to emulate than the relatively recent al-Ǧamhara. In this respect, it should be noted that Qālī was not deterred by the doubts raised concerning Ḫalīl’s authorship of al-ʿAyn – doubts which Qālī himself, accord- ing to his student Zubaydī, attributed to Abū Ḥātim al-Siǧistānī (d. 255/869)

142 Ibn Ḫayr, Fahrasa 354–355. 143 Zubaydī, Ṭabaqāt 186. 144 Ibn Ḫayr, Fahrasa 348–349, 361, 366, 371–372, 375, 380, 383–384, 389, 392–400. 145 Ibn Durayd, Ǧamhara I, 40. 146 Cf. above, 245, 281.

 

amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline

 

305Muǧannas (Semasiological) Lexica

and his contemporaries when al-ʿAyn was imported from Ḫurāsān.147 Ḫalīl is regularly mentioned by name in al-Bāriʿ, unlike Qālī’s other book, al-Maqṣūr wa-l-mamdūd,148 where the expression ṣāḥib al-ʿAyn (al-ʿAyn’s author) is reg- ularly mentioned.149 It is probable that Qālī, having drawn so heavily from al-ʿAyn’s material in al-Bāriʿ, found it inappropriate to leave any doubt sur- rounding the identity of its author. In contrast, the use of the expression ṣāḥib al-ʿAyn – if it is indeed meant to indicate uncertainty about who its real author was – would not seem out of place in al-Maqṣūr wa-l-mamdūd since the book deals merely with a restricted subject and is not a comprehensive lexicon of Arabic usage. Qālī’s position on the identity of al-ʿAyn’s author, however, has little bearing on the fact that al-Bāriʿ emulates al-ʿAyn in its essential features although it has some characteristics of its own, as we shall demonstrate.

Qālī’s phonetic arrangement of letters differs from Ḫalīl’s in some respects. The problem with determining with precision the order of letters in al-Bāriʿ is that its extant parts represent a fraction of the original, more specifically, only parts of the letters h, ḫ, ġ, q, ǧ, ṭ, d and t (listed in the order of their occurrence in the published text). The introduction, in which Qālī might have specified the order of letters, as he did in al-Maqṣūr wa-l-mamdūd, is also missing. In fact, compared with the 4,446 pages of the original manuscript,150 the extant parts are about 140 pages only.151 According to C. S. Fulton, who reproduced in facsimile the British Museum manuscript of al-Bāriʿ (London, 1933), the pho- netic order of the lexicon is as follows:

h, ḥ, ʿ, ḫ, ġ, q, k, ḍ, ǧ, š, l, r, n, ṭ, d, t, ṣ, z, s, ẓ, ḏ, ṯ, f, b, m, w, alif, y

Yet it would be strange indeed for Qālī to choose in al-Bāriʿ an order which differs from the one he adopted in al-Maqṣūr wa-l-mamdūd and which he described in its introduction as proceeding from the farthest point of articula- tion to the nearest. This order is as follows:

ʾ, h, ʿ, ḥ, ġ, ḫ, q, k, ḍ, ǧ, š, y, l, r, n, ṭ, d, t, ṣ, z, s, ẓ, ḏ, ṯ, f, b, m, w152

147 Istidrāk al-ġalaṭ 47; cf. Suyūṭī, Muzhir I, 83–84. 148 Cf. above, 245. 149 Qālī, Maqṣūr 35, 124, 293–295, 320, 324, etc. 150 Ibn Ḫayr, Fahrasa 355; cf. Zubaydī, Ṭabaqāt 186 and Yāqūt, Muʿǧam II, 730 where the num-

ber of pages is said to be 5,000, and Ibn al-Ḥanbalī, Šaḏarāt, where the number is 3,000. 151 Wadġīrī (1984: 20). 152 Qālī, Maqṣūr 11; cf. above, 246.

 

amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline
amgad
Underline

 

306 chapter 3

The most noteworthy difference between Fulton’s order and the one described in al-Maqṣūr wa-l-mamdūd is that the former does not include hamza and places alif towards the end, whereas Qālī places the hamza first and explains that he dropped the alif because no word can begin with it. Furthermore, in the previously mentioned statement by Qālī’s son concerning his father’s lexi- con, it is clearly indicated that the first three chapters which his father had copied before he died were kitāb al-hamz, kitāb al-hāʾ and kitāb al-ʿayn.153 This agrees with the order of the first three letters as specified by Qālī in al-Maqṣūr wa-l-mamdūd and strongly suggests that he adopted the same arrangement in both books. As noted earlier, Qālī’s arrangement is closer to Sībawayhi’s than to Ḫalīl’s154 and demonstrates that he did not slavishly follow al-ʿAyn in all aspects.

The arrangement of the material within each chapter also shows certain modifications introduced by Qālī to Ḫalīl’s system. The chapters are succes- sively divided into biliterals, sound triliterals, unsound triliterals, ḥawāšī, awšāb, quadriliterals and quinqueliterals. In the section of biliterals, Qālī only includes roots whose second radical is geminated, that is, roots that are bilit- eral in their written form but triliteral in reality (al-ṯunāʾī fī l-ḫaṭṭ wa-l-ṯulāṯī fī l-ḥaqīqa li-tašaddud aḥad ḥarfayhi);155 e.g. ǦLL and LǦǦ, ǦDD and DǦǦ, ǦZZ and ZǦǦ,156 etc. Other biliterals which are not geminated but contain an unsound letter, such as the interjections īh and wāh,157 are mentioned with what he calls ḥawāšī (lit. marginal forms). Ḥawāšī also include triliter- als that contain two unsound letters and thus correspond to Ḫalīl’s lafīf; e.g. hayy bin bayy (a person who is unknown and whose father is also unknown).158 Onomatopoeic biliterals and reduplicated biliterals – such as ṭaq (sound of stones), qaq-qaq (ha-ha), and tiġin tiġin (ha-ha)159 – are listed under awšāb (lit. rabble, social misfits),160 which he defines as a collection of words which are

153 Qifṭī, Inbāh I, 244. 154 Cf. above, 246. 155 Qālī, Bāriʿ 568, 579–580, 584, 586, etc. Note that the extant parts of the lexicon do not

contain biliterals that are not geminated (except those that contain an unsound letter). Such biliterals (e.g. the particles qad and min) are placed in al-ʿAyn in the same section as other biliterals (cf. Ḫalīl, ʿAyn V, 16; VIII, 375). Qālī, on the other hand, either ignores them altogether or, more likely, places them in sections of which we have no extant examples.

156 See these and other roots in Qālī, Bāriʿ 564–602. 157 Ibid., 170–171. 158 Ibid., 174. 159 Ibid., 447, 523. 160 This translation of awšāb is taken from Carter (1990: 111), who suggests that Qālī’s term is

perhaps a result of equating the perfect language with the perfect society.

 

amgad
Underline
amgad
Underline
amgad
Underline

 

307Muǧannas (Semasiological) Lexica

either onomatopoeic (ḥikāyāt/aṣwāt), or cries by which one chides or urges an animal (zaǧr), or are defective in form (manqūṣāt).161 Under awšāb are also listed onomatopoeic words which are either triliteral, such as waġā (noise of war), or quadriliteral, such as ġaṭġaṭa (cry of a sand grouse or sound of a boil- ing cooking-pot).162 Unfortunately, the extant parts of the lexicon contain but few examples of awšāb, and we cannot form a full picture of this type, which is unique to al-Bāriʿ among muǧannas lexica. Yet we know from the awšāb sections of the letters ġ and q163 that, in spite of the elaborate division of their material into biliterals, triliterals and quadriliterals, they contain several incon- sistencies in arrangement. For example, the root WĠY is listed with unsound triliterals (where awāġī “place through which water flows” is mentioned), as well as with awšāb (where waġā is cited because of its onomatopoeic nature).164 Furthermore, the quadriliteral qaṭqaṭa is listed under awšāb when it refers to the cry of a sand grouse, but is mentioned as a non-onomatopoeic word with the meaning of “weak rain” in the section of quadriliterals.165 It is also note- worthy that the sections on ḥawāšī and awšāb, unlike other sections of the lexicon, violate the principle of taqālīb, which is strictly adhered to in al-ʿAyn and the other lexica that follow the phonetic-permutative system.166 As for quinqueliterals, they do not occur in the extant parts of the lexicon, but since, in these parts, none is listed with quadriliterals, they were most probably listed in independent sections.

The changes introduced by Qālī to the system of al-ʿAyn in the order of letters and division of sections are minor in comparison with the additions he introduced to its material. This is largely due to Qālī’s method of collecting data from a large number of sources and ascribing lexical items as well as specific meanings of certain words to the philologists who reported them. In describ- ing al-Bāriʿ, Zubaydī highlights the fact that Qālī assembled in it philological works (kutub al-luġa) and ascribed each word to the philologist who reported

161 Qālī, Bāriʿ 446. 162 Ibid., 447–448. 163 Ibid., 447–448, 523–527. 164 Ibid., 446, 447. 165 Ibid., 525, 560. Quadriliterals other than onomatopoeic ones occur at times in two places

as well; e.g. qarqama/qurqum (552, 557), qafandar (554, 556), and funduq/fundāq (557, 560).

166 Qālī seems to ignore taqālīb intentionally in the sections on ḥawāšī and awšāb perhaps because the occurrence of taqālīb in these types of words is rare. Even when two words such as ṭaqṭaqa and qaṭqaṭa offer the chance of being considered as maqlūb, he lists them under separate divisions of awšāb (Bāriʿ 523, 525).

 

 

308 chapter 3

it.167 It should be borne in mind that, by Qālī’s time, philologists had but lim- ited access to trustworthy Aʿrāb from whom to derive their data.168 Given that the process of data collection known as ǧamʿ al-luġa had already been accom- plished, Qālī and most other lexicographers made up for lack of direct hear- ing (samāʿ) from the Aʿrāb by citing earlier philologists who collected their data from the trustworthy Bedouins of their time. In al-Bāriʿ, Ḫalīl is cited in almost every page, often more than once, followed (in order of frequency) by Abū Zayd al-Anṣārī (d. 215/830), Ibn al-Sikkīt (d. 244/858), Aṣmaʿī (d. 216/831), Abū Ḥātim al-Siǧistānī (d. 255/869), Abū ʿUbayda (d. 209/824), Ibn Durayd (d. 321/933), Abū ʿAmr al-Šaybānī (d. 206/821), Abū ʿUbayd (d. 224/838), Farrāʾ (d. 207/822), Kisāʾī (d. 189/805), and Ibn Kaysān (d. 320/932).169 These scholars include Basrans as well as Kufans – a reminder that the lexicographical tradi- tion is largely free from the partisan divide which by the fourth/tenth century was evident in the grammatical tradition. Also noteworthy are the relatively few times in which Qālī’s principal teacher, Ibn Durayd, is mentioned.170 This stands in sharp contrast with Qālī’s K. al-Amālī, a miscellany of literary mate- rial and anecdotes of which Ibn Durayd is one of the main sources of riwāya. It may well be that in spite of Qālī’s extensive citing of his teacher in poetry and anecdotal material in al-Amālī, he opted to cite him sparingly in lexicographi- cal material in al-Bāriʿ due to the doubts expressed by some contemporary scholars as to the authenticity of some of al-Ǧamhara’s lexical items and the accusation that it is a plagiarized version of al-ʿAyn.171

To demonstrate Qālī’s method of deriving data from various sources, we can examine the lemma WǦH. Ḫalīl’s text is cited at the end of the lemma, as is often the case, and is in seven lines in each of al-ʿAyn and al-Bāriʿ.172 The rest of the material is in about forty lines and is mainly derived from Abū ʿUbayda (mentioned three times) and Abū Zayd (mentioned twice), in addition to one unspecified source (wa-qāla ġayruhu). Other than the lexical items which occur in the cited text of al-ʿAyn, the lemma contains material ignored by Ḫalīl. This material consists of lexical items (e.g. tawǧīh “proximity of a horse’s hooves”), new meanings of words cited in al-ʿAyn (e.g. tawaǧǧaha, in the sense of “to face the wind as one voids his excrement”), idiomatic expressions (e.g.

167 Zubaydī, Ṭabaqāt 186. 168 Cf. above, 30–32. 169 See al-Bāriʿ’s indices, 745–752. 170 According to the indices of al-Bāriʿ 747, Ibn Durayd is mentioned sixty-four times; cf. Wild

(1965: 67) and Fleisch (1984: 180). 171 See below 312, 315, 339, 345. 172 Ḫalīl, ʿAyn IV, 66; Qālī, Bāriʿ 92, 95.

 

 

309Muǧannas (Semasiological) Lexica

ʿabd al-waǧh, ḥurr al-waǧh and sahl al-waǧh), and, strangely enough, a techni- cal term related to metrics (tawǧīh). The lemma also contains seven poetry šawāhid as well as dialectal material attributed by Abū Zayd to the Kilābiyyūn, a Numayrī, and an Aʿrābī. Based on such differences in many other lemmata, Ṭaʿʿān’s claim that al-Bāriʿ is identical with al-ʿAyn and that it is thus the old- est copy of al-ʿAyn in our possession173 is certainly erroneous. Although Qālī emulates Ḫalīl’s system (with some modifications in the order of letters and division of chapters), the content of his lexicon, which incorporates al-ʿAyn, is also derived from a large number of sources, including mubawwab lexica, some of which are mentioned in the text (e.g. Abū Zayd’s al-Luġāt, Aṣmaʿī’s al-Ḫayl, and Ibn al-Sikkīt’s al-Alfāẓ).174 By diversifying his sources, Qālī was able to check the correctness of some of the lexical items which he cited. It is quite significant in this respect that, on one occasion, he expresses doubt as to the correctness of a lexical item (namely, aḍmaġa “to slaver excessively”) because he found it in no other source than al-ʿAyn.175 Qālī’s efforts at scruti- nizing the data and diversifying his sources made al-Bāriʿ, according to Suyūṭī, one of the most trustworthy lexica (min aṣaḥḥ mā ullifa fī l-luġa ʿalā ḥurūf al-muʿǧam).176 Yet Suyūṭī further notes that, in spite of this, it did not enjoy much popularity (wa-lam yuʿarriǧū . . . ʿalā Bāriʿ Abī ʿAlī).177

The lemma WǦH also demonstrates two other features of al-Bāriʿ, the first of which is its author’s interest in dialectal usage.178 Qālī derives his dialectal material either from al-ʿAyn or, as in the case of WǦH, from his other sources. He also shows special interest in Arabized words, and, according to Zubaydī in al-Mustadrak (for which see below), words of Persian origin, including those which Qālī mentioned in his lemmata, form a separate chapter at the end of al-Bāriʿ.179 In this, Qālī apparently followed in the footsteps of his teacher, Ibn Durayd, who appended to his al-Ǧamhara several chapters on a miscellany of topics,