249 33 11MB
English Pages [262] Year 2000
The Development of Exegesis in Early Islam
CURZON STUDIES IN THE QUR'AN Editor: Andrew Rippin, University of Calgary In its examination of critical issues in the scholarly study of the Qur'an and its commentaries, this series targets the disciplines of archaeology, history, textual history, anthropology, theology and literary criticism. The contemporary relevance of the Qur'an in the Muslim world, its role in politics and in legal debates are also dealt with, as are debates surrounding Qur'anic studies in the Muslim world.
LITERARY STRUCTURES OF RELIGIOUS MEANING IN THE QUR'AN Edited lYy Issa J Boullata THE DEVELOPMENT OF EXEGESIS IN EARLY ISLAM The Authenticity of Muslim Literature from the Formative Period Herbert Berg
The Development of Exegesis in Early Islam The Authenticity of Muslim Literature from the Formative Period
Herbert Berg
~ l RoutledgeCurzon ~~
Taylor & Francis Group
LONDON AND NEW YORK
First Published in 2000 by RoutledgeCurzon Richmond, Surrey http://www.routledgecurzon.co. uk Transferred to Digital Printing 2005
© 2000 Herbert Berg Typeset in New Baskerville by LaserScript Ltd, Mitcham, Surrey All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Catawguing in Publication Data A catalogue record of this book is available from the British Library Library of Congress Cataloguing in Publication Data A catalogue record for this book has been requested
ISBN 0-7007-1224-0
To A.S., M.S., K.D., and M.P. and to all who try, in their way, to be free
Contents
Acknowledgements
XI
1 INTRODUCTION
1
An Experiment Using Exegetical/fadzths
1
Assumptions, Conclusions, and Qualifications
3
2 lfAD]TH CRITICISM
6
Traditional Sunni Muslim Account
6
Early Western Scepticism I. Goldziher and the Advocacy of Scepticism J. Schacht and Fictitious Legal/fadzths E. Stetter and Topoi and Schemata in lfadfths
8 9 12 17
Reaction Against Scepticism N. Abbott and the Early Continuous Written Tradition F. Sezgin and the Cataloguing of Early Texts M. M. Azami and the Critique of Schacht
18 18 21 23
The Search for Middle Ground G. H. A. Juynboll and the Refinement of Schacht's Methods F. Rahman and an Attempt to Save the Sunna G. Schoeler and the Oral/Written Distinction H. Motzki and the Implausibility of Fabrication J. Horovitz, J. W. Fiick, J. Robson, N.J. Coulson, and U. Rubin
26 26 32 34 36 38
Renewed scepticism M. Cook and the Spread of Isnads N. Calder and the Common Link as the Locus of Controversy
42 42 45
Analyses and Conclusions
48
VII
The Development of Exegesis in Early Islam
3 EXEGETICAL lfADlTHS AND THE ORIGINS OF TAFSlR
65
Defending Isnads and Reconstructing Early Tafs'irs H. Horst and Counting Isnads Birkeland, Abbott, Sezgin, and Goldfeld G. Stauth and Comparing Transmissions from Mujahid
65 66 69 73
The Scepticism and Literary Analysis of J. Wansbrough, A. Rippin, et al.
78
Middle Ground C.H.M. Versteegh, M. Muranyi, New Manuscripts and Reconstructions C. Gilliot and the Function of Isnads in the Imaginaire F. Leemhuis' Intermediate Position Other Scholars Who Have Entered the Fray
83
Implications for the Study of Early "tafszr(s)"
92
84 88 89 91
Excursus: THE AUTHENTICITY OF THE HISTORICAL lfADlTHS 106 4 METHODOLOGY: ISNADS AND EXEGETICAL DEVICES
112
Content Versus Style An Alternative
113 118
Parameters: al-'fabari and Ibn cAbbas Al-Tabari Al-Tabarfs Tafs'ir as a Text Al-'fabari as a Compiler Al-'fabari as an Exegete Ibn cAbbas The Mythic Ibn cAbbas The Students and Works Attributed to Ibn cAbbas
120 120 122 124 125 129 131 135
Hypotheses for Possible Outcomes
137
Database of Isnads and Exegetical Devices Construction of the Database Selecting the Sample Set of Informants and Students Exegetical Devices Wansbrough's Twelve Additional Categories Determining the Exegetical Device(s) for a lfadzth
141 143 145 148 148 155 156
5 DATA AND ANALYSIS: THE AUTHENTICITY OF IBN CABBAs'S lfADlTHS IN AL-TABARl'S TAFSlR Analysis 1: Ibn cAbbas versus his Students and the Informants of al-Tabari viii
173 174
Contents
Analysis of Individual Devices Patterns and their Implications
175 187
Analysis 2: The Students of Ibn cAbbas versus the Informants of al-Tabari Analysis of Individual Devices Patterns and their Implications
191 192 205
The Provenance and Chronology of Exegetical lfadzths
208
6 CONCLUSIONS
219
The Impasse
219
Resolving the Impasse: History versus Literature
226
Bibliography Index
232 247
lX
Acknowledgements
wish to acknowledge my indebtedness to all those friends, scholars, and institutions without whom I would not have been able to complete this study. Much of the early research was funded by the Social Sciences and Humanities Research Council of Canada. I would also like to thank the Centre of the Study of Religion at the University of Toronto, the Department of Religion at Middlebury College, the Department of Religion at the University of Vermont, the Department of Near Eastern Studies at Cornell University, and the Department of Philosophy and Religion at the University of North Carolina at Wilmington for their financial support. I am particularly grateful to Professors Daniel Sahas, Willard G. Oxtoby, B. Todd Lawson, and Jane D. McAuliffe. Daniel Sahas showed me that there was more to academia than mathematics and computers by introducing me to the study of Islam and religions. Willard Oxtoby continued where Daniel Sahas left off. From the beginning of my graduate studies he has been my mentor, guiding both my studies and career. To Todd Lawson, who first encouraged me to read John Wansbrough's work, goes the credit for igniting my interest in the Islamic-origins debate that eventually led to this study. More recently, Jane McAuliffe has provided much needed and invaluable guidance. I am most grateful to her for her gentle but firm encouragement to complete my dissertation out of which this current study grew. I would also like to thank Professors Linda Northrup, Willard G. Oxtoby, Jane D. McAuliffe, Kevin Reinhart, and Andrew Rippin, particularly the latter, for their numerous helpful comments on earlier versions of the manuscript. I greatly appreciate the unpublished works furnished to me by Professors Andrew Rippin, G.H.A. Juynboll, Harald Motzki, and Willi Braun. Professors Donald Wiebe, Neil McMullin, Russell McCutcheon, Darlene Juschka, William E. Arnal, and especially Willi Braun, though they may not realize it, each had an
I
xi
The Development of Exegesis in Early Islam
enormous impact on this study through their efforts to inculcate in me some methodological and theoretical sophistication. Finally, I would like to thank my parents, Alfred and Frida Berg, and my friends, Shirley, Steve, Derek, Jason, Bill, Darlene, Helen May, Russell, Willi, Paula, Ellis, and Kate. Without their financial support, love, encouragement, advice, and/or patience, this study (and much else) would not have been possible.
xii
Chapter 1
Introduction
he study of Islamic origins, whether focussing on the Qur 3 an, its exegesis ( tajsfr), the life of the Prophet ( sfra), jurisprudence (fiqh), or even Arabic grammar (naiJ,w), has been largely confounded by a growing scepticism. While many scholars of Islam still have no qualms about the authenticity of purportedly early Islamic texts, many others have serious reservations. At the crux of this debate is the value scholars assign to the chain of transmitters (the isnarl), which is intended to demonstrate the authenticity and indicate the provenance of the tradition (IJ,adrth) or book of which it is a part. The isniid is seen either as reasonably reliable guarantee of the historicity of its adjoining text (the matn), or as a complete fabrication designed to insinuate chronological priority and hence authority into a later matn. 1 Scholars holding the latter view look, not to the isniid, but often to the matn itself to provide answers to the questions of its own chronology and provenance. These two approaches to isnads are indicative of the approaches to early Islamic texts in general. If the consequences of this disagreement amounted to a simple debate about whether a given IJ,adfth or text is authentic, they would not be particularly noteworthy. Unfortunately, the consequences are far more grave. The scholars on each side of the debate have produced descriptions of Islamic origins that are radically different and mutually exclusive. So different are the methods used and the conclusions reached by sceptical scholars from those of more sanguine scholars that they have come to an impasse.
T
AN EXPERIMENT USING EXEGETICAL
I:IADiTHS
Not surprisingly therefore, the first of two purposes of this study is to help resolve this impasse. To do so I will draw on the methods of both groups of scholars. More specifically, I will determine if there is a correlation 1
The Development of Exegesis in Early Islam
between the results of an isnad-analysis and of a literary analysis of matns, in order to evaluate the historical value of isnads. The initial question to be answered is, of course, where to begin? Isnads are ubiquitous in the literature of early Islam. lfadzthll, composed of matns and isniids, are particularly prominent in the legal, historical, and exegetical texts. Exegetical }Jadzths are an especially practical and valuable subset of the }Jadzths with which to delve into the questions of the authenticity of isnads and, hence, }Jadzths as a whole. Not only has this genre been somewhat neglected, but it also: (1) involves }Jadzths concerned with the most important text oflslam, the Qur 3 an; (2) contains a vast number of l],adzths, many of which have isnads which cite the important transmitters of the }Jadithll of the other genres; (3) encompasses the historical, and legal genres insofar as they relate to the Qur 3 an; and ( 4) adduces hadzthll which initially seem to be immune from the sectarian and legal debates and competition which may have affected the reliability of other }Jadzths. These considerations make exegetical }Jadzths ideal for testing the reliability of the isnad system. AbiiJacfar MuQ.ammad ibnJarir al-'fabari (d. 311/923) is one oflslam's most respected early historians and exegetes, and his Jamie al-bayan .fi ta,wfl ay al-Qur,an, also known as the Tafszr of al-Tabart, is the first substantial compilation of the opinions of the earliest quranic exegetes. Though the Tafszr is not exclusively a compilation, it does contain a vast number of exegetical }Jadzths purporting to provide the exegetical opinions of the earliest generations of Muslims, the Companions of the Prophet, their Successors, and the Successors of the Successors. Thus, it may serve as a valuable repository of Muslim thought and practice during Islam's formative period. This is not to say that I am assuming, like most Muslim scholars and many Western scholars, 2 that the Taftzrcontains the authentic words of the earlier exegetes, and most, if not all, of the material available to al-Tabart. Such assumptions may be simply untenable given the doubts raised by John Wansbrough about the authenticity of all early Muslim written sources in general 3 and those raised by Ignaz Goldziher and Joseph Schacht about the authenticity of both the isnads and the matns of the }Jadzth literature in particular. 4 Though these doubts were raised with respect to the prophetic }Jadzths of the legal realm, their similarity in form and content with the exegetical }Jadzths suggest the same doubts may apply. 5 Consequently, to make such assumptions would be tantamount to assuming what I hope to prove. Analysizing each of the 38,397 }Jadzths6 (many with multiple isniids) in the Tafslr is impractical. Therefore it is necessary to restrict the number of }Jadzths to be examined. There are several ways to do this: random selection of }Jadzths, selecting }Jadtths from several suras or specific verses, or selecting a particular exegete. While all three methods may, in various ways, be non-representative of the }Jadtths in al-Tabarl as a whole, the first 2
Introduction
two methods are more likely to be so because many isnads are not distributed uniformly over the entire Qur~an and some are completely absent from large sections in the Tajszr. However, if an appropriate exegete were chosen, one who appears throughout the Tafszr and in various different isnads, this problem could be overcome. Fortunately, the Tafs'ir contains an excellent candidate and one who is worthy of study in his own right. He is seen by some scholars of quranic exegesis as the earliest and most respected quranic exegete, one who cannot be gainsaid, and by others as the (fictional or mythic) personification of consensus of the early Sunni community. In either case, there is no other quranic exegete of his stature. In many ways, he is to exegetical l}adtths what Mul).ammad is to legal l}adtths. This exegete is, of course, cAbd Allah ibn cAbbas (d. 67-8/ 686-8), the Prophet's paternal cousin and ancestor of the cAbbasid dynasty. By employing a database of these isnads (that is, of all l}adtths whose isnads name Ibn cAbbas as either the exegete or as a transmitter) and the exegetical type(s) of their respective matns, my experiment will be able to trace the "stylistic fingerprint" of Ibn cAbbas's exegetical methods (or more accurately, the exegetical methods of a large number of l}adtths ascribed to Ibn cAbbas) through several overlapping lines of transmission. These statistical analyses will suggest that these particular isnads are not a reliable source of historical information. This conclusion in turn supports the suggestion that a mythic status developed around Ibn cAbbas as an exegete and encourages an exploration of how this status was invoked and manipulated. And because Ibn cAbbas can serve as a paradigm for understanding exegesis associated with other early figures, this study will shed light on the development of exegesis in Islam. Furthermore, since all of these results are intimately connected with the questions of historiography and authenticity, my work aims at a greater understanding of not only the genesis of quranic exegesis, but also Islamic origins in general.
ASSUMPTIONS, CONCLUSIONS, AND QUALIFICATIONS Definitively resolving the aforementioned impasse is an ambitious goal. Obviously, generalizing from exegeticall}adfths to other genres of l}adzths is not likely to be convincing to all scholars. Questions raised about al-Tabari's role in collecting, editing, and tendential manipulation of the material in his Tafsirand Ibn cAbbas's mythic status in the realm of quranic exegesis are problematic, but not overly so. There is a more serious stumbling block. The most important reason for my experiment's inability to convince some scholars lies in the failing that scholars on both sides of this debate - both the sceptical and the sanguine - share: the results of their work is dictated by their presuppositions. 3
The Development of Exegesis in Early Islam
This statement is not self-evident. All of the participants in this debate feel that their conclusions have been based solid and thorough scholarship. Yet, the debate remains. Therefore, in chapters 2 and 3 I examine and discuss the arguments of various scholars on the issue of the authenticity of }Jadfths in general and of exegetical }Jadfths in particular. This survey is a necessary introduction to my experiment whose hypothesis and parameters are described in chapter 4 and whose results are given in chapter 5. The survey will also allow me to determine which, if any, of the various theories regarding the emergence of }Jadzths, exegetical or otherwise, best fit the data. However, the more important goal of chapters 2 and 3, which is also the second purpose of this study, is not simply to highlight the differences between the two major positions in order to justifY by own study. Rather, I will demonstrate that the stance each scholar takes determines his or her results and that the arguments are often circular. I will argue that those who doubt the authenticity of }Jadzths presuppose the widespread fabrication of isniids. By treating the supposedly historical information provided by the isniids as a deliberate attempt to mask the true origin of their matns, these scholars are forced to conclude that }Jadzths, and texts for that matter, must have been produced much later than they purport to. In other words, they are not authentic. On the other hand, those who defend the authenticity of }Jadzths presuppose the veracity of isniids. In some cases, this is explicit; they argue on the basis of isniids and the symbiotic biographical literature for an early, continuous, and widespread practice of written and oral transmission of }Jadzths. In other cases, especially with those who seek a middle ground, this is implicit; most often they argue that similar extant matns allow them to reconstruct earlier or original forms of the reports. Their comparisons and reconstructions are only possible, however, when they accept the information provided in the isniids - doubting only those isniids that do not fit their reconstructions. Thus, if one compares only presuppositions of those who trust isniids and those who search for a middle ground, one finds very little to distinguish them. And it is hardly surprising that the results of the latter vary only slightly from those of the former as well. In both cases, their defence of the isniid-system is predicated on the overall reliability of that system. In the end, even my own attempt at an empirical experiment requires that I make some presuppositions, and it will be argued by critics, that my results are no less immune from my presuppositions. Therefore, though my study takes on the guise of an experiment with an hypothesis and a statistical test, it too will fail to convince those who disagree with my presuppositions. This is not surprising. And it is unlikely, therefore, that the impasse will be resolved.
4
Introduction
NOTES 1 I shall use the term l],adzth to refer to any material presented in the isnad-matn format, regardless of whether the isnads end with MuJ:iammad. I will distinguish between the l],ad'iths which comprise the sunna, tafsir, and sira by referring to the legal, exegetical, and historical genres of IJ,ad'iths. I do so for the sake of clarity, realizing that all of these genres overlap to some extent. The reader should bear in mind that other scholars may use the terms l],ad'ith and tradition (as well as sunna and tafsir) in slightly different ways. When these differences are important to the discussion, their connotations will be explained. 2 The distinction between "Muslim" and 'Western" is becoming more blurred. There are Muslims who work within the 'Western" tradition, and Western scholars (geographically speaking) who operate within a Muslim framework. However, such a distinction remains useful when dealing with basic approaches to early Islamic texts, one of which is confessionally motivated and one of which is not (at least not explicitly). 3 See Chapter 2, page 42, and Chapter 3, pages 78-83. 4 See Chapter 2, pages 9-12 and 12-6, respectively. 5 Not all scholars believe that such an assumption is justified. Further to this, see Chapter 3, pages 66-78. 6 Al-Tabari, Tafsir al-Tabari al-musammii Jamie al-bayan ft ta"wil al-Qy,r"an (Beirut: Dar al-Kutub al511miyya, 1992), 12:753. All references to the Tafsircome from this edition. See Chapter 4, pages 159-60, note 42.
5
Chapter 2
lfadith Criticism
T
he assumption of most Muslim scholars has been that the IJ,adfth material, at least that contained in the classical canonical collections, is authentic. Canonical status is conferred upon al-Jamic al-!jal],fl], of Abu cAbd Allah Mul)ammad ibn Ismacrl al-Bukhari (d. 256/870) and al-Jamic al-!jaiJ,ziJ, of Abu al-I:Iusayn Muslim ibn al-I:Iajjaj (d. 261/875), and, to a lesser degree, upon the Kitab al-sunan of Sulayman ibn al-Ashcath Abu Da"ud al-Sijistani (d. 275/889), al-Jamic al-!jaiJ,fiJ of Abu elsa Mul)ammad al-Tirmidhi (d. 279/892-3), the Kitab al-sunan of Al)mad ibn Shucayb al-Nasa"i (d. 303/915), and the Kitab al-sunan of Abu cAbd Allah Mul)ammad ibn Yazid al-Rabaci al-Qazwini ibn Majah (d. 273/887). To these six collections are occasionally added other works, most notably the Musnad of Al)mad ibn I:Ianbal (d. 241/855), but these others have not quite achieved the same degree of authority. And, while these six works are not immune from criticism, it is generally believed that among them they contain an authentic, authoritative, and fairly complete record of the words and deeds of Mul)ammad. Although, many Western scholars have not been as generous in their assessment of the material in them, most Muslims continue to feel that the rigorous analysis to which the transmitters of it were subjected by these collectors assures its authenticity.
TRADITIONAL SUNNI MUSLIM ACCOUNT 1 For Muslims, transmitting the words and deeds of Mul)ammad is as old as Islam itself. The Qur"an orders Muslims to follow the example of the Messenger2 and so from the very beginning the Companions (!fa!Jaba) concerned themselves with following the sunna (conduct or custom) of the Prophet, which was embodied in IJ,adfths (reports or anecdotes) narrating his words and deeds. Mul)ammad is thought to have taken some pains to ensure the use and dissemination of his sunna. 3 6
I:Iadith Criticism
Generally, the Umayyad caliph cumar II (d. 101/720) is credited with having ordered the first collection of l],adzth material in an official manner, fearing that some of it might be lost. 4 Abu Bakr ibn MuQ.ammad ibn I:Iazm (d. 120/737) and MuQ.ammad ibn Muslim ibn Shihab al-Zuhri (d. 124/ 742), known simply as Ibn Shihab or al-Zuhri, are among those who compiled l],adzths at cumar II's behest. This delay, nearly a century, in having the l],adzths recorded resulted from reservations expressed by MuQ.ammad and especially the first four Caliphs to commit to written form anything other than the Qur"an, lest it be confused with the Qur"an. cumar I is the primary locus for many accounts about l],adfth collection. He is portrayed as desiring to initiate this project but as unwilling to do so, fearing that Muslims might then neglect the Qur"an. The movement to finally record l],adzths initiated by cumar II and Ibn Shihab, though begun somewhat haphazardly, culminated with the six canonical collections after having received impetus from the establishment of the sunna as the second source of law in Islam, particularly through the efforts of the famous jurist MuQ.ammad ibn ldris al-Shaficr (d. 204/820). The actual text of the l],adfth is known as the matn. For a matn to be recognized as an authentic record of one of MuQ.ammad's acts or sayings, it needs to have attached to it the list of the people who were transmitters ( mul],addiths) of the matn. This isniid, or chain of authorities, provides the name of the eyewitness of the actual event, the person to whom s/he related the event, the person to whom this mul],addith related the matn, and so forth until the l],adfth was recorded. The isniid portion of the l],adzth was an early standard practice as well, according to Muslim interpretation of the sources. To MuQ.ammad ibn Sirin (d. 110/728) is attributed the following remark: They (sc. the traditionalists) were not used to inquiring after the isnad, but when the fitna (= civil war) occurred they said: Name us your informants. Thus if these were ahl as-sunna (= the people of the catholic Muslim community) their traditions were accepted, but if they were ahl al-bidac, their traditions were not accepted. 5 For most Muslim scholars, this fitna is the one following the assassination of the third caliph, cUthman (d. 35/656). And so, the regular use of isnads for l],adzths is thought to have begun shortly after 35 A.H. (656 C.E.). This date then also marks the beginning of l],adzth study as a science in the Muslim community. The implication of Ibn Srrrn's statement is that well-meaning but misguided or even unscrupulous people fabricated or altered l],adfths for political, dogmatic, or personal reasons. Muslims freely admit this. But, according to traditional accounts, these vast numbers of obviously false and doubtful l],adzths were eliminated in the painstaking process of producing the classical collections. In the third/ninth century, the sifting 7
The Development of Exegesis in Early Islam
out of these spurious l],adfths focussed largely on the isnad. That is to say, the compilers systematically examined each of the transmitters of every l],adfth (though often the examination ignored the first level of transmitters, the Companions, who were thought of as being above charges of falsification). Analysis focussed on the transmitter's date and place of birth, familial connections, teachers, students, journeys, moral behaviour, religious beliefs, literary output, and date of death. This allowed compilers to determine not only reliability (thiqat), but also the contemporaneity and geographical proximity of transmitters juxtaposed within the isnad, in an attempt to ascertain whether they could have come in contact. In addition to this biographical analysis ( cilm al-rijaf), the cohesion (itti~af) of the isnad was examined. The continuity of the isnad was evaluated for missing or unknown mul],addiths or for not reaching back to Mul)ammad and stopping at a Companion or Successor. In addition, the number of simultaneous transmitters was tallied. A l],adfth with numerous transmitters at every level of the isnad ( mutawatir) was deemed to be beyond doubt of forgery, while one with three or more at each level ( mashhur), one with just one at a particular level (gharib), or one with one transmitter at each level (jard) was considered binding but with less weight. On these three bases, a particular l],adfth would be classified as ~al],fl], (sound or authentic), l],asan (good), fj,acif (weak), or saqfm (spurious). 6 So Muslims have never suggested that forgery of l],adfths was not a problem in early Islam. What they do claim is that the forgeries have been eliminated and that that which has been preserved is, on the whole, the actual words and deeds of the Prophet Mul)ammad. 7 Moreover, since criticism of l],adfths in the Muslim world has focussed on their isnads, 8 dating of a particular l],adfth is done by ascription. That is, a prophetic l],adfth came into circulation during the life of Mul)ammad, and one that terminates with a Companion was probably born in the first few decades after the death of the Prophet. The method of criticism and the conclusions it has reached have not changed significantly since the third/ninth century. Even much of modern Muslim scholarship, while continuing to debate the validity or authenticity of individual l],adfths or perhaps the l],adfths of a particular transmitter, employs the same methods and biographical (or rijaf) materials. 9
EARLY WESTERN SCEPTICISM Many Western scholars have accepted, with some reservations, these assumptions and used l],adfths as fairly reliable historical sources, but for many others the authenticity and the date of origin of the l],adfth material are issues that have produced, and continue to produce, heated debate. In 8
I:Iadith Criticism
1848 Gustav Weil, after noting that al-Bukhari deemed only 4,000 of his original 600,000 J;,adfths to be authentic, suggests that a European critic is further required to reject without hesitation at least half of these 4,000. 10 He was soon followed by Aloys Sprenger, who also suggests that many of the J;,adfths cannot be considered authentic.l 1 However, that there is a debate about the authenticity of J;,aduh material in the West is largely due to the innovative theories of lgnaz Goldziher. The subsequent direction this debate has taken, a direction which has focussed on the role of J;,adfths in the origin and development of early Muslim jurisprudence, is largely due to the work of Joseph Schacht.
I. Goldziher and the Advocacy of Scepticism While others had expressed some doubt about the authenticity of }J,adfths before Goldziher, 12 it was he who in the second volume of his Muhammedanische Studien first clearly articulated this scepticism. Familiarity with the vast number of J;,adfths in the canonical collections induced "sceptical caution rather than optimistic trust." 13 Goldziher concluded that these J;,adfths could "not serve as a document for the history of the infancy of Islam, but [served] rather as a reflection of the tendencies which appeared in the community during the maturer stages of its development. "14 Goldziher's suspicions about the authenticity of J;,adfths sprang from several observations. The material found in later collections makes no references to earlier written collections and uses terms in the isnads which imply oral transmission, not written sources. 15 Moreover, the ubiquitous contradictory traditions, the apparent proliferation of J;,adfths in later collections not attested to in earlier ones, and the fact that younger Companions of Mul).ammad seem to have known more about him (that is, they transmitted more J;,adfths) than the older Companions who presumably knew the Prophet for a greater length of time, suggested to Goldizher that large-scale fabrication of J;,adfths took place. As a result, Goldziher provides a significantly different version of the origin and development of J;,adfth literature. Goldziher has no trouble accepting that the Companions preserved the words and deeds of their prophet after his death, and that these might have been recorded in written form in ~al;,ifas. In this way he remains very close to the Muslim interpretation of the development of J;,adfth literature. He not only presumes that the Companions tried to preserve the sayings and judgments of Mul).ammad, but also that some of them likely did so in written form (that is, in ~al;,ifas). And, when these Companions passed on what they had heard and recorded to the next generation of Muslims, the use of the isnad began. 16 But for Goldziher, the invention of and interpolation into J;,adfths also began very early, for both political 17 and 9
The Droelopment of Exegesis in Early Islam
paraenetic 18 reasons. And so mutually exclusive }fadzths proliferated; "it is not surprising that, among the hotly debated controversial issues of Islam, whether political or doctrinal, there is not one in which the champions of the various views are unable to cite a number of traditions, all equipped with imposing isniids. " 19 With the rise of the cAbbasids the situation changed significantly, according to Goldziher. cAbbasid rule was more theocratic than the more secular "Arab paganism" of the Umayyads. 2 Consequently, the new dynasty encouraged the development of the shari:ca and even employed court theologians to advise the caliphs, some of whom themselves studied and participated in theological debates. This attempt to give public life a more religious character also involved giving official recognition to the sunna. The rise of the sunna had begun during the Umayyad period in part in opposition to the perceived wickedness of the time, but its supporters remained relatively ineffective until the advent of the cAbbasid revolution. The report that the Umayyad caliph cumar II commissioned the first collection of }fadfths must be dismissed as untrustworthy because of the number of contradictions in the account and the absence of references to Abii Bakr ibn I:Iazm's work in later literature. For Goldziher, this claim is hagiographic, that is, "nothing but an expression of the good opinion that people had of the pious caliph and his love for the sunna."2 1 Goldziher maintains that, while reliance on the sunna to regulate the empire was favoured, there was still in these early years of Islam insufficient material going back to Mul).ammad himself. Scholars sought to fill the gaps left by the Qur 3 an and the sunna with material from other sources. Some borrowed from Roman law. Others attempted to fill these lacunae with their own opinions ( ra,y). This latter option came under a concerted attack by those who believed that all legal and ethical questions (not addressed by the Qur 3 an) must be referred back to the Prophet himself, that is, must be rooted in }fadzths. These supporters of }fadfths ( ahl al-}fadfth) were extremely successful in establishing }fadfths as a primary source of law and in discrediting ra,y. But in many ways it was a Pyrrhic victory. The various legal madhhabs were loath to sacrifice their doctrines and so they found it more expedient to fabricate }fadfths or adapt existing }fadfths in their support. Even the advocates of ra,y were eventually persuaded or cajoled into accepting the authority of }fadfths and so they too "found" }fadfths which substantiated their doctrines that had hitherto been based upon the opinions of their schools' founders and teachers. 22
°
The insistence of the advocates of }fadfths that the only opinions of any value were those which could appeal to the authority of the Prophet resulted in the situation that ''where no traditional matter was to be had, men speedily began to fabricate it. The greater the demand, the busier was invention with the manufacture of apocryphal traditions in support of the respective theses. "23 The (alab journeys which followed, during which the 10
I:IadHh Criticism
travellers sought to collect l}adfths from the various centres of the Islamic empire, helped construct a more uniform corpus of extant l}adiths out of the various disparate local collections. Eventually, however there were reactions to this widespread fabrication of l}adiths. Goldziher traces three such reactions to this phenomenon. Ironically, fabricated l}adiths began to circulate in which Mul).ammad is made to condemn those who would fabricate l}adiths about him. Others simply rejected the whole corpus of l}adiths and referred only to the Qur"an. The third reaction was the one which arose among the traditionalists themselves and came eventually to dominate. They developed a means by which to evaluate the authenticity of any l}adith. This method focussed not on the actual contents of the l}adith ( matn) but on the transmitters of the matn, that is, on the isnad. Goldziher seems to suggest that this critique was in nascent form already around 150 A.H. Even with this type of examination, forgeries continued to be made through the manipulation of the isnad in somewhat more subtle ways. According to Goldziher, l}adiths, which originally had isnads ending with Companions or Successors, were often extended back to the Prophet. That is al}adith mawqufa were transformed into al}adfth marfuca by tacking on the Prophet and any other necessary names to the end of the isnad. Isnads were also "tampered" with by the mucammarfn- the long-lived ones. For Goldziher these were persons who pretended to have had direct contact with Mul).ammad even though this might mean that they would have to be well over a hundred years old (and at times hundreds of years old). 24 As stated earlier, Goldziher questions the traditional date at which the formal collection of l}adfths began. It was not in the time of cumar II, but with the Muwa((aJ of Malik ibn Anas (d. 179/795) . that the process started. 25 That is, Goldziher believes l}adfth gathering began only towards the end of the second century A.H. (late eighth or early ninth C.E.) with jiqh works being the precursors to proper l}adfth works. 26 These latter works came soon after, as a more a systematic arrangement of the l}adith material became necessary. As the insistence that legal and religious practice be rooted in l}adfths had grown, so too had the available material. This arrangement took two forms: the musnad (arranged according to the isnads) and the muJannaf (arranged according to topic). The muJannaft came to predominate, but the musnads continued to be compiled. 27 An example of a musnad is the compilation of Al).mad ibn f:Ianbal. The first muJannaf that gained prevalence was the compilation of al-Bukhari. It, unlike the Muwat(aJ, is a work of l}adfths with al-Bukhari's contribution being one of selection and arrangement. Mter the compilation of the six canonical collections in the middle of the third century A.H. (second half of the ninth century C.E.), Goldziher feels, there was a decline in l}adfth literature in the sense that, instead of being compilers of new material, 11
The Droelopment of Exegesis in Early Islam
J;,adzth scholars became copyists and editors producing mukhtaJars, or "abridged versions." In summary, Goldziher sees in J;,adzths "a battlefield of the political and dynastic conflicts of the first few centuries of Islam; it is a mirror of the aspirations of various parties, each of which wants to make the Prophet himself their witness and authority." 28 Likewise, Every stream and counter-stream of thought in Islam has found its expression in the form of a Q.adith, and there is no difference in this respect between the various contrasting opinions in whatever field. What we learnt about political parties holds true too for differences regarding religious law, dogmatic points of difference etc. Every raJy or hawii, every sunna and bidca has sought and found expression in the form· of Q.adith. 29 And even though Muslim traditionalists developed elaborate means to scrutinize the mass of traditions that were then extant in the Muslim lands, they were "able to exclude only part of the most obvious falsifications from the Q.adith material. "30 Goldziher, for all his scepticism, accepted that the practice of preserving J;,adzths was authentic and that some J;,adzths were likely to be authentic. 31 However, having said that, Goldziher is adamant in maintaining that: In the absence of authentic evidence it would indeed be rash to attempt to express the most tentative opinions as to which parts of the Q.adith are the oldest material, or even as to which of them date back to the generation immediately following the Prophet's death. Closer acquaintance with the vast stock of Q.adiths induces sceptical caution rather than optimistic trust regarding the material brought together in the carefully compiled collections. 32 And so it is in his advocacy of scepticism that Goldziher made his great impact on the course of J;,adzth studies in the West. 33 Goldziher never went much beyond this simple scepticism about the authenticity of the bulk of the J;,adfth material to advance a more practical theory for determining the chronology and provenance of any specific }j,adfth. He limited his dating of }j,adzths to the general comments like "maturer stages of its development" or "first few centuries of Islam". Although he hesitated to date the traditions, the scholars who continued his work expended considerable effort in that very endeavour.
J.
Schacht and Fictitious Legal l:ladiths
Of the next generation of scholars, Joseph Schacht most prominently carried on Goldziher's tradition of scepticism 34 and his Origins of Muhammadan jurisprudence in turn serves, along side Goldziher's Muslim
12
I:Iadith Criticism
Studies, as the foundation or at least a point of departure for almost all other studies on IJ,adtths in the West. 35 In Origins Schacht's main concern is the origin of Islamic law, the sharica, and particularly the role of al-Shafici in its development. This traditionist and legal theorist is thought to be responsible for championing the sunna - sunna specifically understood as the model behaviour of Mul).ammad as opposed to the 'living tradition' of the Muslim community which might or might not claim to have such a direct connection to Mul).ammad. 36 In so doing, Schacht discusses the process of development of IJ,adtth material (and hence its authenticity and chronology). Schacht asserts that IJ,adtths, particularly from Mul).ammad, did not form, together with the Qur 3 an, the original bases of Islamic law and jurisprudence as is traditionally assumed. Rather, IJ,adtths were an innovation begun after some of the legal foundation had already been built. "The ancient schools of law shared the old concept of sunna or 'living tradition' as the ideal practice of the community, expressed in the accepted doctrine of the school. "37 And this ideal practice was embodied in various forms, but certainly not exclusively in the IJ,adtths from the Prophet. Schacht argues that it was not until al-Shafici that 'sunna' was exclusively identified with the contents of IJ,adtths from the Prophet to which he gave, not for the first time, but for the first time consistently, overriding authority. Al-Shafici argued that even a single, isolated IJ,adtth going back to Mul).ammad, assuming its isniid is not suspect, takes precedence over the opinions and arguments of any and all Companions, Successors, and later authorities. Schacht notes that: Two generations before Shafici reference to traditions from Companions and Successors was the rule, to traditions from the Prophet himself the exception, and it was left to Shafici to make the exception the principle. We shall have to conclude that, generally and broadly speaking, traditions from Companions and Successors are earlier than those from the Prophet. 38 Based on these conclusions, Schacht offers the following schema of the growth of legal IJ,adtths. The ancient schools of law had a 'living tradition' (sunna) which was largely based on individual reasoning (ra,y). Later this sunna came to be associated with and attributed to the earlier generations of the Successors and Companions. Later still, IJ,adtths with isniids extending back to Mul).ammad came into circulation by traditionists towards the middle of the second century. Finally, the efforts of al-Shafici and other traditionists secured for these IJ,adtths from the Prophet supreme authority. 39 However, the development of prophetic tradition did not cease at this point. In fact, as a result of the new authority conferred upon them, Schacht suggests that a large number of the IJ,adtths preserved in the classical collections originated both during and after 13
The Devewpment of Exegesis in Early Islam
al-Shaficr's time. That is, most Prophetic }Jadzths in the collections of Bukhari, Muslim, and the others originated, not with Mul).ammad, but circa the middle of the second century A.H., while }Jadzths citing the opinions of Companions and other authorities originated somewhat earlier. In one of his most emphatic statements, Schacht concludes that "... every legal tradition from the Prophet, until the contrary is proved, must be .taken not as an authentic or essentially authentic, even if slightly obscured, statement valid for his time or of the time of the Companions, but is the fictitious expression of a legal doctrine formulated at a later date. "40 Schacht therefore dismisses Muslim scholarship on }Jadzths, which itself is based on the study and criticism of isnads as "irrelevant for the purpose of historical analysis." 41 Although Schacht offers a far more refined argument than Goldziher, he has not yet gone far beyond him in his theories. This Schacht does, however, in the methods for determining the provenance of specific }Jadzths which he develops. His unique contribution lies in his alternative to the "irrelevant" methods of Muslims; he suggests that the date of a }Jadzth can be ascertained from its first appearance in the legal discussion, from its relative position in the history of the problem with which it is concerned, and from certain indications in the text and the isntid. What is meant by 'its first appearance in the legal discussion' is obvious. If a particular }Jadzth is adduced in one text but is not to be found in an earlier text in which that same }Jadzth would have been of crucial importance, then it is safe to assume that the }Jadzth was not yet extant and was invented sometime after the writing of the earlier one. This is essentially an argument from silence, but quite a compelling one. By 'its relative position in the history of the problem' Schacht means to suggest that }Jadzths were frequently fabricated in a polemical context. That is to say, they were designed specifically to refute certain pre-existing doctrines or practices. A new }Jadzth or set of }Jadzths would then provoke the supporters or practitioners of the attacked doctrine or practice to manufacture }Jadzths to both defend it and to undermine the refuting }Jadzths. Their opponents would then respond with more and usually more elaborate }Jadzths. Thus, by juxtaposing various parallel or related }Jadfths and comparing their matns, one may be able to reconstruct the chronology of the }Jadzths surrounding a particular controversy. The doctrine or practice being attacked is, of course, chronologically prior to the }Jadzth countering it. A }Jadzth defending the practice or doctrine is likely to be after the counter-}Jadzth. 42 'Indications in the text' means looking at the authority cited in a }Jadfth. In the course of polemical discussions, each group was forced to project its doctrine to increasingly higher authorities. That is, teachings once ascribed to Successors become those of Companions, and the latter in turn become the words of the Prophet himself. Schacht argues that: 14
f.ladith Criticism
Whenever we find, as frequently happens, alleged opm10ns of Successors, alleged decisions of the Companions, and alleged traditions from the Prophet side by side, we must, as a rule and until the contrary is proved, consider the opinions of the Successors as the starting point, and the traditions from the Companions and from the Prophet as secondary developments, intended to provide higher authority for the doctrine in question. 43 Closely related to these textual indications are the 'indications in the isnad,' by which Schacht means his the backward growth of isnads. This theory is summed up in his famous dictum: 'The more perfect the isnad, the later the tradition." 44 Thus, Schacht sees the isnads as the most arbitrary part of the IJ,adzths, but because their fabrication and development follows certain patterns, they nevertheless allow the J;,adzths to be dated in many cases. As the depth of the isnads grew (that is, backward growth), so too did their breadth grow. This 'spread of isnads' occurred because additional isnads were created to support a particular IJ,adzth and in this way obviated the charge that the IJ,adzth was 'isolated.' Thus mutawatir IJ,adzths have no more claim to authenticity than do other IJ,adzths. Schacht argues that: Any typical representative of the group whose doctrine was to be projected back on to an ancient authority, could be chosen at random and put into the isnad. We find therefore an number of alternative names in otherwise identical isnads, where other considerations exclude the possibility of the transmission of a genuine old doctrine by several persons. 45 Because of the arbitrariness of this isnad manufacture, Schacht feels that it would be pointless to attempt to reconstruct the opinions and doctrinal positions of the Companions. "[T]hey are the products of schools of thought which put their doctrines under the authority of the Companions.''46 He also dismisses the claim to genuineness of the IJ,adzths with family isnads (that is, those that were transmitted exclusively within several generations of one family). For Schacht "the existence of a family isnad [is] not an indication of authenticity but only a device for securing its appearance.''47 Schacht observed another phenomenon that he feels can be employed in determining the provenance of report. He notes that in many cases, the isnads of J;,adzths with similar or related contents often contain the same transmitter somewhere in the middle of the isnad. Schacht's own example (see Diagram 1) of this phenomenon shows cAmr ibn Ab1 cAmr as the common link or common transmitter for three instances of the same matn. For Schacht, this is a case where a report has been put into circulation by a traditionist or by someone using his name. As the report 15
The Droeloprrumt of Exegesis in Early Islam
Diagram I Prophet
I Jibir
I
man of the Banii Salama
Prophet
I Jabir
I
Munalib
Prophet
I
Jabir
I
Munalib
cAmr ibn Abl cAmr
cAbd al-cAzlz ibn Mul,lammad
Ibrahim ibn Mul,lammad
Sulayman ibn Bilal
Shaficl
Shaficl
anonymous
I I
Shafici 48
was passed on to others, represented in the diagram by the names below cAmr's, the "real part of the isniid" would branch out into several strands. The isniid would not terminate with the one who put the matn in circulation, for he would have fabricated an isniid reaching back to an authority such as a Companion or the Prophet. This is represented in the diagram by the names above cAmr's and is for Schacht the "fictitious part of the isniid." It would often acquire additional branches to improve its authority. The existence of this common link, Schacht suggests, would be a strong indication in favour of its having originated in or after the time of the aforementioned traditionist, in this case cAmr's, and so fix a terminus a quo for the appearance of the l).adith. 49 Schacht's examination of the development of isniids is also premised on his hypothesis that legal }Jadzths go back only as far as 100 A.H., that is, in the last years of the Umayyad rule- when, according to him, Islamic legal thought began. 5° This concurs with the statement, attributed to Ibn Sirin, that interest in isniids began from the time of the fitna (strife) after which people could not be trusted to give non-partisan reports. For Schacht, the fitna began with the murder of the Caliph Walid ibn Yazid (d. 126/744). Schacht points out that Ibn Sirin died in llO A.H. (728 C.E.), so the tradition is obviously spurious, but nevertheless accepts the dating implied because he sees no evidence of the regular use of isnads prior to 100 A.H.5l 16
I:Iadith Criticism
E. Stetter and Topoi and Schemata in l:fadiths In his 1965 dissertation Eckart Stetter examines the topoi and schemata in l],adzths using a representative segment from Bukhaii's $al],zl],. By a topos, Stetter means a narrative cliche which provides circumstantial details, such as the exact place and/ or time Mul).ammad is to have said or done something, that imply intimate, personal contact. For example, l],adzths often mention that the first transmitter heard Mul).ammad while he was speaking from the minbar or describe what he was wearing. These typified situations and stereotypical figures of speech are superfluous comments, found almost exclusively at the beginning of the matn. These topoi do not just exist to facilitate the flow of the narrative; rather, the motivation behind them was to provide "authentic" detaii. 52 The schemata are also narrative forms that serve to fill lacunae and to connect, associate and organize materials. These forms include the repetition (often in threes) of phrases, the use of parallelism (both of form and content), the use of assonance and rhyme, and so forth. This schematization may also serve a mnemonic purpose, useful, no doubt, in oral transmission. 53 Stetter credits unknown redactors (unknown both individually and collectively) with the influx of these topoi and schemata into preliterary materials. Whether such elements could have entered the preliterary material as oral formulae used by storytellers is a possibility left unaddressed by Stetter. The touches of authenticity, such as circumstantial details about Mul).ammad, would certainly be necessary for any fabricator of l],adzths, and the ready-made topoi and schemata would provide them in both the content and form of the matns. Stetter does not draw on the works of Goldziher or Schacht directly, and in that sense his thesis stands somewhat apart from the authenticity debate. Nevertheless his observations about the presence of these narrative motifs certainly raise questions about the authenticity or, at the very least, the reliability of the l],adzth literature in a manner independent of, though not uncongenial to, the doubts raised by Goldziher and Schacht. 54 Goldziher introduced scepticism about l],adzths. Schacht and Stetter suggest plausible mechanisms for the creation of false l],adzths. Because this scepticism strikes at the very foundation of early Islamic literature, the rest of the edifice begins to crumble. The very piece of evidence that is meant to guarantee genuineness of the matn, the isnad, is being summarily dismissed as a fabrication. Therefore, to non-sceptics the conclusions of Goldziher and Schacht are wrong because they are based on a misunderstanding of the transmission system. In other words, their (false) assumptions about the nature of l],adzths has led then to (false) conclusions. And so, their arguments seem contrived, circular, and contrary to reason to those who disagree with them. 17
The Deuelapment of Exegesis in Early Islam
REACTION AGAINST SCEPTICISM Unlike Stetter's work, the theories and methodologies advanced by Goldziher and Schacht have inspired much comment, commendation, and criticism, and in so doing have determined both the nature and direction of the debate concerning l],adzths. Their work has been attacked on two fronts. 55 The first attempts to correct the aforementioned "misunderstandings" by insisting that l],ad'iths were committed to written form very shortly after the death of MuQ.ammad, or even during his lifetime. And as a further guarantee of their reliable transmission, l],ad'iths were then maintained in written form until they were finally compiled in the classical collections. 56 The second consists of asserting the early and reliable use of the isniid to counter Schacht, for whom its presence, particularly in complete form, is already an indication that a l],ad'ith has been fabricated or at least manipulated. N. Abbott and the Early Continuous Written Tradition Nabia Abbott tries to argue that there was an early and continuous practice of writing l],ad'iths in Islam. By "early" she means that the Companions of the Prophet themselves kept written records of l],ad'iths and by "continuous" that most l],ad'iths were transmitted in written form (alongside the oral transmission) until the time they were compiled in the canonical collections. For her, then, it is this written transmission of l],ad'iths that serves as the guarantee of their authenticity. Abbott suggests that literacy was not uncommon among Arabs even in pre-Islamic times and that reports about MuQ.ammad were already being written during his lifetime. 57 The problem for Abbott, given this suggestion, is the obvious lack of any early attempt to standardize all these reports about MuQ.ammad and, more tacitly, the lack of extant manuscripts from this period. Her solution to this conundrum is to lay the blame squarely on the shoulders of the second caliph, cumar I (d. 23/644). Because ofthe lack of familiarity with the Qur~an in the newly conquered lands outside Arabia, the caliph feared "a development in Islam, parallel to that in Judaism and Christianity, but particularly in the latter, of a body of sacred literature that 58 could compete with, if not distort or challenge the Qur~an." So he destroyed the manuscripts of l],ad'iths he discovered and punished those who had possessed them. Many Companions avoided (at least publicly) the use of written and even oral l],adfths lest they incur the caliph's wrath (even though they did not necessarily concur with him on this issue). However, the real basis for the later collections of l],ad'iths was the relatively few Companions, such as cAbd Allah ibn cAmr ibn al_c~ (d. 65/684), Abu Hurayra (d. 58/678), Ibn cAbbas (d. 67-8/686-8), and Anas ibn Malik (d. 94/712), who continued to collect, record, and transmit them. 59 18
l:ladrth Criticism With the death of cumar and the successful promulgation of the cUthmanic recension of the Qur"an, the two major fears regarding the use of lfadzths were significantly diminished. According to Abbott, the use of lfadzths then flourished in the second half of the first century and even those early Muslims, who like cumar, eschewed the use of lfadzths in written form, succumbed to preserving their knowledge thus. lfadzths were taught in the major centres of Islam, particularly Medina and Mecca, for legal, paraenetic, and entertainment purposes, not only by jurists and judges, but also by teachers, preachers, and storytellers. Abbott recognizes that Western scholars, such as Goldziher and Schacht, question the veracity of the later reports of literary activities during this early period. She states that she herself shared these same doubts but now believes them to be largely unjustified, for the description of this period is relatively consistent and well-attested. Abbott adds: For not only was there a remarkable degree of unanimity among the admiring students and followers of these men and like-minded traditionists concerning the overall literary activity, but reluctant and at times censorious testimony by the opposition bears witness to this literary activity. Furthermore ... there are literally dozens of their contemporaries scattered across the vast empire who were engaged in similar activities but who for one reason or another never received marked public attention. 60 In an attempt to counter Goldziher's suggestion of the secular nature of Umayyad rule, Abbott argues that the Umayyad caliphs Mucawiya (d. 60/ 680), Marwan (d. 65/684) and cAbd al-Malik (d. 86/705), for example, all took an active interest in transmitting and/or recording lfadzths. 61 cumar II is particularly associated with lfadzth literature. Abbott accepts the report (found in the recension of Shaybanl (d. 189/805) of Malik ibn Anas's Muwatta" 62 that this Umayyad caliph commissioned Abu Bakr ibn Mul).ammad ibn cAmr ibn l;lazm (d. 120/738) to record lfadzths and sunna. 63 Abbott argues that he was only one of many the caliph contacted in order to secure authentic lfadzths, and that Ibn Shihab al-Zuhrl was ordered to collate these numerous lfadzths from the various regions of the empire. Abbott further assumes that al-Zuhrl finished this enormous task and that these daftars (manuscripts) were distributed, but that because of resistance in the provinces and the untimely death of cUmar II, they never received much attention 64 (though the work of al-Zuhrilived on through his many noteworthy students). Thus, Abbott has attempted to remedy this "oversight" by Goldziher and to give the Umayyads their due by stressing their role in encouraging the written transmission of the lfadzth material. With this form of transmission of lfadzths, Abbott is also able to provide the following explanation for the appearance of a rapid expansion in the number of lfadzths- perhaps to counter Schacht's spread-of-isnads theory. 19
The Droelopment of Exegesis in Early Islam
Manuscripts, particularly those preserved by succeeding generations of the same family, which were lengthy documents, were divided into separate sections and given the isnild of the original document. From one such document could come hundreds of }Jadfths. "If not fully comprehended, this process would give the impression of a sudden huge increase in the number of traditions ... "65 Furthermore, Abbott argues that: The development of the family isnild and continuous written transmission lead to the ... inescapable conclusion ... that the bulk of the }Jadfth[s] and sunna as they had developed by about the end of the first century was already written down by someone somewhere, even though comparatively small numbers of memorized traditions were being recited orally. 66 That is, she not only accepts the bulk of family isnilds as genuine (unlike Schacht), but also credits them for guaranteeing the authenticity of }Jadfths in general. And, these parallel oral and written transmissions each served to safeguard the other and so prevented the large-scale fabrication of }Jadfths. Therefore, Abbott can conclude that the content of the sunna was more or less fixed by the time of al-Zuhri. Abbott sees in the ri}Jlas (the journeys in search of knowledge and usually associated with oral tradition), in the use of the warrilqiln (stationer-copyists), and in the average memory of the average traditionist evidence for the continued use and production of manuscripts of }Jadfths. In fact, the oral transmission has been overemphasized according to her because Western scholars have generally failed to grasp }Jadfth semantics properly. Arabic terminology for writing materials and in isnilds has also been misunderstood. An example of the former is the word ~a}Jifa. It is normally translated as "sheet (of writing material)" but can refer to anything from a single sheet to a large daftar (manuscript). An example of the latter are the words }Jaddatha (to relate) and akhbara (to tell), which seem to connote oral transmission, but which Abbott states were also used for written transmission. Also, statements in the sources which imply a certain traditionist did not use written materials may simply mean that he did not use them publicly. This is not to say that Abbott equates oral transmission with fabrication and written transmission with authenticity: It would, of course, be absurd to equate oral transmission with excessive fluidity of either form . or content, with the usually accompanying implication of conscious or unconscious fabrication, and it would be equally absurd to equate literary record with complete fixity of form and content implying thereby the exclusion of the probability of fabrication. But it would likewise be absurd not to concede that oral transmission is indeed more conducive to fabrication than is literary fixity. 67
20
I:IadHh Criticism
However, because of her resolute defence of the authenticity of J;,adzths, Abbott places herself in the awkward position of having to explain why Muslims themselves, particularly the early collectors, recognized that there were many more unsound J;,adzths than there were sound ones. Does that not imply fabrication of J;,adfths on a massive scale? Not for Abbott. She suggests that it was the isnads that proliferated, not the matns. With each generation the numbers of J;,adfths (counting any variation of matn or isnad as a separate J;,adzth) increased geometrically until there were literally hundreds of thousands of J;,adzths during the time of Bukhari and Muslim. And since criticism of J;,adfths was directed at their isnads, the high number of unsound J;,adzths merely meant that the isniids were less than perfect while the matns may well have been "sound." This hypothesis also allows her to conclude that fairly exhaustive collections of J;,adfths were possible using only a fraction of the vast material available. 68 That does not mean for Abbott that the J;,adfths found in Bukhari and Muslim are entirely authentic and complete, but that they preserve a "genuine core of the sayings and deeds of Mul).ammad together with a genuine core of sayings and deeds of the Companions and Successors", 69 which had already been more or less fixed by the time of al-Zuhri. While Abbott has for the most part constructed a coherent argument, her reading of the sources seems naive to some scholars. As G. H. A. Juynboll has pointed out, "Abbott seems to rely too heavily on much of the information given in isniids and in books about isnads concerning the three oldest tabaqat."70 Yet it is isnads and their authenticity that lie at the heart of the debate over the authenticity of J;,adfths. Her conclusions come as no surprise once she accepts the historicity of the information contained in isnads, for she has simply developed a circular argument.
F. Sezgin and the Cataloguing of Early Texts Fuat Sezgin also argues that there was an early, continuous written tradition in his Geschichte des arabischen Schriftums, Band I: Qur,iin Wissenschaften, Hadith, Geschichte, Fiqh, Dor;matik, Mystik bis ca. 430 H In many ways his evidence resembles that of Abbott and so it need not be fully restated. However, his argument differs from hers in that it is a much more focussed and concerted attempt to undermine the implications of Goldziher's sceptical approach to the J;,adfth literature. Sezgin realizes that Goldziher did not have all the currently available sources, and in this respect cannot be unduly faulted. However, in a harsher critique, Sezgin devotes considerable energy to trying to demonstrate that Goldziher misunderstood some key terms related to the transmission of J;,adfths. Sezgin lists eight ways in which transmission took place: samac, qira,a, and wijada. Sezgin states that ijaza, munawala, kitaba, iclam al-rawf, wa~fya, only the first two (listening and recitation, respectively) involved
21
The Development of Exegesis in Early Islam .
-J
memorization. The others, and often in practice even sama/ and qzra a, involved written materials. Furthermore, written transmission was as customary as oral transmission. 71 Sezgin concludes from this: Now we must come closer to the fact that these transmission methods reach back in part to the beginning of Islam and that they demonstrate, with the help from references and preserved materials, that from the start exclusively written foundations for the transmission were involved and that the names of the authors are contained in the isnads. 72 Clearly Sezgin (like Abbott) has no doubts as to the authenticity of the isnads. Moreover, he is willing to suggest that from these authorities can be gleaned authors of actual texts. 73 Sezgin traces a very different history for J;,adith literature than that provided by Goldziher. That is, he, like Abbott, tries to make a case for the intense literary activity of the Companions and Successors, reiterating that the terms used by them for the transmission of J;adiths indicate primarily transmission in written form (a fact, Sezgin repeats, that Goldziher misinterpreted). 74 The first stage involved simple books ( ~al;ifas or juzJs) produced by the Companions and the earliest Successors (and in this regard, Goldziher and he agree). But whereas Goldziher puts the collection of l;adiths after the development of fiqh literature, Sezgin envisions the development of l;adith literature to be an independent and continuous practice that began with the ~al;ifas. In the last quarter of the first century and in the first quarter of the next, the scattered l;adiths began to be collected. The first proper collection was made by al-ZuhrL And so the first mu~annaf dates from circa 125 A.H. (742 C.E.), thus antedating the musnads of the end of the second century A.H. - the time 75 during which Goldziher placed the appearance of mu~annafo. Sezgin, true to his claim that texts can be reconstructed, catalogues all the various "texts" that must have at one time been extant. His criterion for listing a text is simply ascription. As for doubting whether one can trust the isnads provided by later writers, Sezgin states very clearly that, "in order to establish the first sources of Islamic literature, one must first of all discard the old presupposition that the isnad was first introduced in the second and third centuries A.H. and that the transmitters names were invented. "76 Sezgin offers no real argument for why one should give up the old "prejudice" against the authenticity of the isnads; he simply says it must be so. However, asJuynboll points out, "unearthing and cataloguing material, as Sezgin has done, is something altogether different from establishing authenticity ... he presents the bulk of them as if he has no qualms as to their genuineness. "77 Sezgin is not exactly presenting theories as facts, but he proceeds from the assumption that transmission took primarily a
22
I:Iadith Criticism
written form and that isnads are a reliable record of that transmission. While the argument is quite consistent, it is also quite circular.
M. M. Azami and the Critique of Schacht Muhammad Mustafa Azami in his two major works, Studies in Early Hadith Literature and On Schacht's Origins of Muhammadan Jurisprudence, 78 has attempted to rectify the perceived inadequacies of Western scholarship on IJ,adzth literature and in particular to refute the theories of Schacht. His methods are not altogether different from Abbott's and Sezgin's, but his focus is. He is not only out to reclaim the authenticity of the IJ,adzth material in the classical collections, but also out to prove the reliability of the isniids which support them. Azami argues (like Abbott and Sezgin) that there was already intense literary activity during the time of the Prophet, which he himself had strongly encouraged. This continued in both the secular and religious realms during the reign of the Umayyads. With this Azami has set the stage for his argument that IJ,adzths were written down even in the time of Mul).ammad. He then proceeds to list the hundreds of Companions, Successors, and scholars from the first 150 years oflslam who, according to him, wrote down IJ,adzths, along with the names of their students who received IJ,adzths from them in written form. 79 That is, he states there has been a continuous, early written tradition, implying thereby that authenticity of the IJ,adzth material is more assured. And again, the argument (such that it is) is one which relies on ascription and treating isniids and the rijiilliterature as independent, but mutually corroborating sources. Azami's far more original and valuable contribution to the study of IJ,adzths comes with his defence of the isniid. Schacht contends that, while the isniid system may be authentic for IJ,adzths whose isniids end in secondcentury scholars, they are certainly not for those which end with the Prophet or the Companions. Azami breaks down Schacht's contention into six main points and addresses each in turn. First Azami addresses the claim made by Schacht and others that the isniid system began in the early second century or perhaps the late first. He cites both Horovitz and Robson 80 to support his claim that the use of isniids, like the use of written records, was very early. Azami adduces the report ascribed to Ibn Srrrn that the use of isniids was demanded after the .fitna, and for Azami, the .fitna refers to the civil war between cAll and Mucawiya (36 A.H.). Furthermore, Azami argues that this report states that it was only after that time that isniids were "demanded," implying that they must have been in use prior to that time, albeit perhaps less stringently. Second, Schacht's statement that the isniid is the most arbitrary part of a IJ,adfth is attacked by Azami. Azami finds ample evidence for rejecting this
23
The Development of Exegesis in Early Islam
assertion in his earlier attempt to show that the isnad system was born in the life of the Prophet and developed into a proper science by the end of the first century. Furthermore, the "numbers of transmitters of one tradition and their different localities make it difficult to imagine the theory of 'projecting back'. "81 In fact, Azami calls it "absurd" and "almost impossible" to fabricate isnads on such an enormous scale. Perhaps even more compelling is Azami's question, why would scholars, already willing to fabricate an isnad, not simply choose the most respected figures? Why have they chosen weak links for their isnads? This is illogical to Azami. 82 Third, on the charge that isnads were gradually improved through fabrication and alteration, Azami admits that faulty isnads exist in J;,adzth material, but rejects the notion that they indicate anything significant about the development of J;,adzths. He points out that al-ShaHr, for example, openly admitted that his faulty memory had caused him to forget parts of isnads. Other tradents, for the sake of brevity, might have chosen to give incomplete isnads. Hence, one cannot necessarily conclude that an incomplete isnad was "improved", when it appears in a more complete form in a later work. Fourth, Schacht claimed that additional authorities were created in al-Shaficl's time to obviate the criticism of a J;,adzth being "isolated" (that is, his spread-of-isnads theory). Azami criticizes Schacht for using primarily an e silentio argument. Just because other transmissions of a J;,adzth were not recorded until later does not mean they did not exist at the time the "isolated" one was recorded. Perhaps at that time adducing a single J;,adzth was considered sufficient and repetitions superfluous. Fifth, as for family isnads and Schacht's assertion that they are spurious, Azami cites Robson's argument that even if there were fabricated family isnads, they must have been modelled on genuine family isnads. 83 Azami suggests it would be more appropriate to say, "All the 'family isnads' are not genuine, and all the 'family isnads' are not spurious."84 Sixth, on the issue of the common-link theory, Azami shows that the one example used by Schacht is not in fact a case of a common link. He points out that a close examination of the al-Shaficl text reveals that there is but one chain from the Prophet to cAmr, who then transmitted it to three of his students. Compare Diagram 1, the one given by Schacht, and Diagram 2, the one given by Azami for the same set of isnads. 85 Azami points out quite rightly that the Prophet:Jabir links are obviously just one link. In addition, he concludes that cAbd al_cAzlz erred in making cAmr's authority a man of Bam1 Salama and so the isnad should look like that given in Diagram 2. Therefore, the one example adduced by Schacht is not in fact an example of a common link. Azami does not seem to realize that the isnads linking the Prophet and cAmr need not fan out in order for cAmr to be the common link. That it fans out after him is the criterion which indicates
24
I:Iadith Criticism
Diagram 2 Prophet
I
Jabir
I
Munalib cAmr
cAbd al-cAziz
Ibrahim
Sulayman 86
that he is a common link. However, Azami, antiCipating the counterargument that would simply adduce other examples of genuine common links, provides an explanation for their occurrence and in so doing he makes a better criticism of the common-link theory. He presumes, given the nature of J;,adfths, that one should expect common transmitters to appear in many chains, since it is possible for a transmitter to have unique information or because a particular transmitter "published" his knowledge gleaned from several sources and so became the point of reference to all later transmitters. Mter having addressed each of these six points, Azami is able to conclude that "[t]here is no reason to reject the isnild system. It is proved that it has every element which can command the acceptance of the system as a whole. "87 And, having vindicated the isnads, Azami turns to the question of the authenticity of J;,adzths in general. He focusses once again on Schacht, who had asserted: "The best way of proving that a tradition did not exist at a certain time is to show that it was not used as a legal argument in a discussion which would have made reference to it imperative, if it had existed. "88 Once again, Azami attacks the examples adduced by Schacht to exemplifY this assertion. He also points out that one should not expect these early Muslim scholars to have been aware of all the extant J;,adfths of their time. Rather snidely, Azami asks why should these scholars make references to these "later" prophetic J;,adzths when Schacht himself states that reference to such J;,adfths was the exception, not the rule. Azami confidently concludes, "So even if mistakes in isnads and a}J,adzth exist, Schacht has produced no evidence that would cause us to impugn the good faith of the majority of the transmitters or abandon the J;,adfth literature."89 While this may be overstating the case somewhat, Azami has
25
The Development of Exegesis in Early Islam
indeed significantly undermined many of Schacht's conclusions. Certainly many of the examples adduced by Schacht were done so incorrectly or inappropriately. Yet, the dismissal of a few examples does not necessarily weaken the overarching patterns suggested by Schacht, especially since many of Azami's conclusions, like those of Abbott and Sezgin, rest on complete faith in the historicity of the source material. This faith, of course, is the problem for sceptics. The arguments of Abbott, Sezgin, and Azami rely on biographical materials that were produced symbiotically with the isnads they seek to defend. These sources are not independent. And so their arguments seem no less contrived, circular, and contrary to reason as those of the sceptics seem to their opponents. As a result, we are left with two seemingly diametrically opposed theories for the origin and development of }Jadzths and, hence, of early Islam itself.
THE SEARCH FOR MIDDLE GROUND Many scholars have found merit in the arguments and theories of Goldziher and Schacht, and in those of Abbott, Sezgin, and Azami. While the scepticism of the former two seems largely justified, these other scholars are loath to accept the full implications of the doubts raised. They are not willing to resign themselves to such uncertainty. Nor are they willing to accept what appears at times to be the seemingly naive position of the latter three. The use of simple ascription is historically untenable to them. And so these other scholars have tried to find an intermediate position between belief and unbelief in the historicity and authenticity of the }Jadzth literature.
G. H. A. Juynboll and the Refinement of Schacht's Methods Gautier H. A. Juynboll, like Azami, has delved deeply into the issues of origin and authenticity of the }Jadzth material as raised by Schacht. But unlike Azami, Juynboll embraces Schacht's work and is in many respects his successor, even though he differs from him on several significant points. That is, Juynboll defends and considerably refines Schacht's theories, but he also retreats from his complete scepticism about the authenticity of }Jadzths. 90 On the whole, he is just as distrustful of the historical value of isnads, but pushes the date for their appearance to not earlier than the end of the first century, which is several very significant decades earlier than Schacht places it. Juynboll sees himself in the line of Goldziher and Schacht, not Abbott, Sezgin, and Azami. The former two are referred to as his predecessors and, while he castigates Schacht's tone and style, he openly acknowledges his debt to his theories. 91 The latter three do not fare as well. Juynboll says: 26
l:ladith Criticism
Something which always struck me in the work of Sezgin, Azmi and also in that of Abbott ... is that they do not seem to realize that, even if a manuscript or a papyrus is unearthed with an allegedly ancient text, this text could easily have been forged by an authority who lived at a time later than the supposedly oldest authority given in its isnad. Isnad fabrication occurred . . . on just as vast a scale as matn fabrication. 92 From the above it may seem that Juynboll is hardly seeking "middle ground." However, he believes that the early reports regarding the origins of J;,adzth material, while obviously not all true, do, when taken as a whole, converge on a fairly reliable and historically accurate description. He adds, "I think that a generous lacing of open-mindedness, which dour sceptics might describe as naivete, is an asset in the historian of early Islamic society rather than a shortcoming to be overcome and suppressed at all costs." 93 In addition to this qualified credulity ofhis,Juynboll finds middle ground in many of his conclusions about the origins and authenticity of the J;,adzth material and sciences, which he largely bases on the awaJil literature - anecdotes about who was the first to do something or when and where certain institutions were first established. 94 According to the awaJil sources, after Mul).ammad's death the first to spread stories about him (in a deliberate manner) were the storytellers ( qu!f!ftl!f), who told stories of an edifying nature. Isnads proper were not attached to these "prophetic" utterances. It is reported that Shacbr (d. 103-10/721-8) was the first person to question someone about an authority and that Shucba ibn al-I:Iajjaj (d. 160/777) was the first to examine every isnad. And so, systematic rijal criticism began about 130/ 747. Hence, isnads did not appear as early as many Muslim scholars believe. For Juynboll, the fitna to which Ibn Sirln alluded was the war between the Umayyads and Zubayrids. 95 This scenario, which places the origin of the isnad around the year 70/690 (as opposed to 35/656), makes the awaJil account of the first isnad critics much more plausible. Juynboll outlines his tentative chronology of the growth of J;,adzths in the following manner. He does not dispute that Muslims began to record things about their prophet during his lifetime, but there is nothing to suggest that this was practised on a significant scale. His examination of the awa'il evidence on the introduction of J;,adzths to various parts of the Islamic world and on the collection of such material indicates a relatively late growth. He suggests that "the earliest origins of standardized J;,adzth[s] cannot be traced back earlier than, at most, to the seventies or eighties of the first century. What had preceded this was ... still unstructured and still unstandardized material of edifying contents ... or with a political slant. .. ". 96 Juynboll further supports this conclusion through an examination of the limited use of J;,ad'iths in the initial years after Mul).ammad's
27
The Development of Exegesis in Early Islam
death and the relatively late development of the concept of prophetic sunna and l]adfth centres. Juynboll notes that the first three caliphs relied on their own judgement and rarely invoked the example of the Prophet. The concept of sunna as sunna of the Prophet (only), developed toward the end of the first century, though a more vague concept of sunna that included Mul_lammad and his most respected Companions predates this more specific sunna. He credits cUmar II with the first organized attempt to apply the sunna of the Prophet. In the I:Iijaz, Egypt, Syria, and Iraq interest increased in l]adfths in the last decades of the first century. During this period isnads were localized and only in the first few decades of the second century do isnads of 'mixed' origin appear. Within this historical framework, Juynboll attempts to answer in a general way the question of where and when l]adfths originated and who brought them into circulation. Because the Successor-Companion link in an isnad is the hardest to establish and because of the regional character of the first few transmitters below the Companion in the isnad, Juynboll concludes that the point of origin is likely the region where the transmitter mentioned at the Successor level resided. 97 Furthermore, since some of the Companions are credited with such incredibly large numbers of l]adfths (many of which are obvious fabrications) and since most were dead when the use of isnads became mandatory (near the end of the first century), it is unlikely that they would be responsible for the l]adfth. So it is the Successor named in the isnad who is the earliest candidate for bringing a tradition into circulation, but since the first major growth of l]adfths occurred several decades after the first century, the Successors to the Successors are the more likely candidates. Nor is it, Juynboll adds, necessarily the case that the Successor (or whoever first circulated a l]adfth) is responsible for having raised the isnad to the level of the Prophet. This may well be a later modification. Juynboll believes that there was large-scale fabrication of matns, some clearly using the preserved memories of what Mul_lammad had said and done as a basis, some clearly in the quranic spirit, and many others not so. The isnad system, born in the 70's, did not develop into a full-fledged science for another half-century. By that time it was too late to evaluate adequately the growing l]adfth literature: sound isnads could be invented in their entirety, and no method had been developed for evaluating the matns. Yet,Juynboll maintains, as a whole the l]adfths do reflect reasonably accurately Mul_lammad's words and deeds: [I] t seems likely that at least part of the prophetic tradition listed in one or more canonical - or even non-canonical - collections deserves to be considered as a fair representation of what the prophet oflslam did or said, or might have done or said, but surely it is unlikely that we will ever find even a moderately successful method
28
l:ladlth Criticism
of proving with incontrovertible certainty the historicity of the ascription of such to the prophet but in a few isolated instances. 98 And so Juynboll, while seemingly as sceptical as Schacht, offers at least the possibility of genuine material being present in the canonical collections. Having discussed the chronology and provenance of }Jadfths in general, Juynboll spends considerable effort refining and employing Schacht's common-link theory in order to the same for individual }Jadfths. In Muslim Tradition, he merely gives what seems to be a defence of the theory against Azami's critique. He explains the non-universal nature of the commonlink phenomenon by suggesting that during the early stages of }Jadfth evolution, its frequency must have been much higher. Juynboll explains: It is because of insertions, interpolations, deletions and simplifications in the matns that additional isnads supporting these alterations became so complex and variegated that the initial isnad or protoisnad, clearly showing up a common-link, supporting the l).adith without accretions was no longer separately discernible. 99
That is, the common link of many }Jadfths has been irrevocably obscured by the sheer number and complexity of fabricated isnads. However, in his articles, Juynboll is less tentative in his use 100 and elaboration 101 of the common-link theory. One of Juynboll's most interesting contributions comes in his distinction between the common-link isnad, in which there is a single strand of three to five transmitters and then a branching out of the chains at the common link, and the inverted common link, in which the common link stands at the end of several chains of transmitters beginning with different eyewitnesses and continues from the common link along a single chain. Compare Diagrams 3 and 4. These two patterns correspond to those found in legal }Jadfths and historical }Jadfths respectively. While Juynboll suggests that the common link of the former likely invented the single strand from himself to the Prophet or Companion, the common link of the latter did not. The single strand from the cl [common link] down to the prophet does not represent the transmission path taken by a prophetic saying, a path which has a claim to (a measure of) historicity, but is a path invented by the cl in order to lend a certain saying more prestige by means of the first and foremost authentication device of his days: the isnad marfu/.1° 2 One would expect, especially for important }Jadfths, that the isnads would begin to fan out after the Prophet, or perhaps the Companion, but not after four or more generations. On the other hand,
29
The Droelopment of Exegesis in Early Islam
Diagram 3 Prophet
I
Companion
I
Successor (Successor)
I
COMMON LINK
transmitter
transmitter
transmitter
transmitter
collection collection collection collection collection collection collection collection
Diagram 4 eyewitness
I
eyewitness
eyewitness
I
I
eyewitness
I
transmitter
transmitter
transmitter
transmitter
transmitter
transmitter
transmitter
transmitter
INVERTED COMMON LINK
I
transmitter transmitter transmitter collection 103
30
J:Iadith Criticism
as a rule the (i)cl [inverted common link] did NOT invent the multiple strands down to various eyewitnesses, he did NOT invent the contents or the gist of the report, and even if it is conceded that he edited several different accounts of the same event and moulded them into one narrative, the gist of the historical event is not the product of his own imagination. 104 It is for this reason that he feels that the authenticity of legal l}adzths, which display the common-link pattern (represented in Diagram 3), are suspect, while historical l}adzths (or akhbar), which display the inverted commonlink pattern (represented in Diagram 4) are to be accepted as reliable. Another important phenomenon detected by Juynboll is the one in which a relatively late transmitter seems to have an independent isnad for a l}adzth that bypasses the common link and merges with the other isnads at the Successor or Companion level. Juynboll refers to this as "diving under the common link," because in his diagrams the Prophet appears at the bottom. (I have inverted his diagrams for the sake of consistency). He suggests that the practice of circumventing the common links is a comparatively late phenomenon, originating towards the endo of the second century A.H. But these dives are useful for dating to: the higher the bypass above the common link, later the origin of that particular strand. 105 Juynboll observes that in the canonicall}aduh literature, l}adzths displaying common links are in fact relatively rare -just a few hundred. On the other hand there are thousands of traditions, which, when their isnads are charted, display a spider pattern. That is, it first appears that there is an early common link, the Prophet, a Companion, or a Successor, but upon closer observation almost all the fanning out occurs in single strands - no transmitter having more than one or two alleged students. See Diagram 5. Juynboll suggests that these spiders should be interpreted as having developed not downwards, but upwards: "the later transmitters/ collectors invented single strands bridging the time gap between themselves and a suitably early, fictitious or historical, [person] ." 106 For these spiders,Juynboll notes, it is impossible to draw conclusions about their chronology, provenance, or authorship. There is one major problem with the conclusions Juynboll draws from the spider pattern. If a report from the Prophet were in fact genuine and faithfully transmitted, its transmission pattern might well resemble the spider pattern. Clearly, one's assumptions on the nature of isnads can dictate how one interprets this pattern. In elaborating and refining Schacht's methods and theories, Juynboll has made considerable advances in determining the chronology, provenance, and authorship of specific l}adzths. He has found a way to salvage historical information from at least part of the isnad. 107 Yet the methods developed by Juynboll allow only relatively few l}adzths to be fully 31
The Development of Exegesis in Early Islam
Diagram 5
alteTative Companion
Fulan
I I
Fulan
Companion
Fulan 3
Successor
I Fulan 2
I
Fulan 1
I
Fulan
I
I
I
Fulan
Fulan
Fulan
Fulan
I
I
I
I
Fulan 4
Fulan
I
Fulan
Fulan
I Fulan
I Fulan
Fulan
Fulan
Fulan
Fulan
Fulan
I
I
I
I
I
I
I
I
I
Fulan 5
I
I
I
Fulan
I
I
Collector 7 Collector 6 Collector 5 Collector 4 Collector 3 Collector 2 Collector ! 108
analyzed in this manner. For the others, perhaps the majority, he, like Goldziher and Schacht before him, simply resigns himself to being uncertain. F. Rahman and an Attempt to Save the Surma Fazlur Rahman stands in a somewhat unique position. On the one hand, he has accepted some of the general conclusions reached by Goldziher: IJadUhs are by and large not historical. On the other hand, as a Muslim, he hesitates to dismiss the IJ,adzths in the canonical collections and the Prophet's sunna as spurious. His theory on the origin and development of IJ,adfths suggests that, while the isnad of a IJ,adzth may well be fabricated and perhaps even the wording of the matn, the gist of the matn is still prophetic and therefore normative for Muslims. In his book, Islam, Rahman begins his chapter on IJ,adzths and the sunna by examining the work of Goldziher, Margoliouth, Lammens and Schacht. He credits Goldziher with seeing the difference between the normative conduct of the community and the actual practice of the community: the former is the sunna and the latter the actual state of affairs. It is this distinction according to Rahman that was overlooked by the other three
32
I:Iadith Criticism
and leads them to a logical contradiction. They combine the two and define sunna to be the normative practice of the Muslim community. Rahman asks, "what sense does it make to say that the normative quality was sought to be conferred on the actual practice by makingit the Sunna of the Prophet?" 109 Rahman's goal in highlighting what he perceives to be a contradiction is to undermine the claim of Margoliouth, Lammens, and Schacht that sunna (the practice of the community) preceded by nearly a century its embodiment in }Jadzths. At least according to his reading of Goldziher, he and Goldziher see the two as having a common origin and being consubstantial. Rahman then criticizes Schacht's suggestion - that }Jadzths with isnads reaching back to the Prophet originated not before the middle of the second century - as being too simple and causing insoluble problems. For Rahman, the sunna varied from place to place. Al-Shaficr simply introduced the concept of the sunna of the Prophet into Muslim jurisprudence in a systematic way. He cannot be used as evidence that such was not the case from the very beginning of Islam in at least some places or in some less systematic way. On the other hand, Rahman has words of praise for Schacht's method of comparing different versions of }Jadzths and for his conclusions that some }Jadzths did not exist in the early period and that later versions tend to contain more information than earlier ones. Like Azami, he cautions that one should be careful with this method since it is possible that earlier reports of a }Jadzth might be less complete simply because the full details only became available with wider contact with Companions and Successors. With this critique of the Western position, Rahman then moves to his own theory. In terms of being sceptical about the literal authenticity of }Jadzths, he is not much less so than Goldziher and Schacht and certainly not significantly less so than Juynboll. He states: Prophetic Sunna, outside the fundamental matters touching the religious and the social and moral life of the Community, could not have been very large, let alone being of such titanic inclusiveness of all the details of daily life as medieval law and I:Iadrth literature make out to be the case.l 10 Mul).ammad had made pronouncements in an ad hoc manner, not in a systematic way as suggested by the canonical collections. Yet that does not mean Rahman doubts that the activities and sayings of Mul).ammad were not preserved in some manner as the Qur"an itself and other documents were, especially since these words and actions were (as the Qur"an itself attests) considered normative. Hence, an informal tradition can be assumed during the lifetime of the Mul).ammad.m A slightly more formal tradition developed after the death of Mul).ammad in that new Muslims would naturally enquire about the 33
The Development of Exegesis in Early Islam
words and deeds of Mu}:lammad. Isnads, which are for Rahman a sign of the formalized discipline of IJ,adzth, appeared much later, around the turn of the first century (though the informal use of IJ,adzths began somewhere between the years 60 and 80) _1 12 A unique contribution from Rahman is his theory of the "silent" transmission of Prophetic sunna. That is, many early Muslims simply lived out the words and acts of Mu}:lammad. And this silent, living tradition, the tradition of what Muslims actually did, is the sunna. And so Rahman states, "that the Sunna and I:Iadrth were coeval and consubstantial in the earliest phase after Mu}:lammad and that both were directed towards and drew their normativity from him. "113 However, as time passed the succeeding generations of Muslims made additions to both the sunna and the IJ,adfth and this led to a disharmony between them (though in general, they were uniform). Eventually this situation led to a need to standardize practices, and the manner in which this was accomplished was the codification of IJ,adzths. At first IJ,adzths were referred back only to Companions because they embodied the words and acts of Mu}:lammad and because the discipline of the isnad had not yet fully developed. Rahman credits al-ShaHr for the place achieved by IJ,adzths in the hearts of Muslims. But the resultant codification and attempt to bring the sunna under the aegis of the sunna of the Prophet (read: prophetic IJ,adzths) led to a massive fabrication of IJ,adzths. And according to Rahman, Muslims were largely successful in bringing the whole of the living tradition (sunna) into the sunna. And thus IJ,adzths (sunna of the Prophet) and sunna (the silent living tradition also rooted in the acts and words of the Prophet) were consubstantial in content once again. 114 As a result, the sunna as currently embodied in IJ,adzths remains normative (though somewhat more flexible). Clearly his goal is to save the sunna, not to devise a new method of isnad criticism. For Rahman the very charge made by the Western sceptics, that IJ,adfths are merely an attempt to give the actual practice ofthe Community prophetic authority, is irrelevant. The actual practice of the Community was already prophetic, at least in spirit if not always in detail. So while the isnads may well be fabricated and in some sense the matns as well, the IJ,adzths nevertheless remain prophetic. 115
G. Schoeler and the Oral/Written Distinction Despite the attempts by Abbott and Sezgin to lay the issues of oral versus written transmission of early Islamic "texts" to rest, the debate continues. Their arguments, while perhaps quite convincing to some, are not consistent with the evidence supplied by these texts as we have received them. The fact is that there are significant, and at times even startling, variances between different recensions of the teachings of a particular 34
I:Iadith Criticism
early Muslim authority. Certainly Sezgin's optimism that earlier, original texts could be reconstructed from later compilations seems unjustified. But does this fact also mean that the written transmission argument put forward by Abbott and Sezgin has been refuted? No, says Gregor Schoeler. In a series of four articles he presents an alternate conception of the mode of transmission of knowledge in the various branches of the Islamic sciences. By suggesting a mixed mode of oral and written transmission, he attempts to preserve the authenticity of the material as it exists today, while still accounting for the observed variances. 116 Schoeler begins with the question of whether or not written texts of !;adzths were prevalent prior to the collections of al-Bukharl and Muslim. Goldziher would say not, Sezgin yes. Schoeler argues against Goldziher's interpretation that encomia, such as mii raJaytu fi yadi-hi kitiiban qattu (I never saw a book/something written in his hands), of early scholars meant that they shunned the use of written materials. Just because they did not employ written notes during their public lectures, does not mean that they did not have recourse to such written materials privately. Even if the shaykh did not use notes, his students likely did, and so the ways in which they preserved and further transmitted the work contributed to the development of different recensions. The students either recorded the content of the lecture in written form during its presentation or later (when they themselves wished to transmit it) from memory or according to another source, such as that of an exemplary copy of the teachings from the shaykh's circle of students. (Of course, if one student recorded the teachings using another student's copy, the former would feel no obligation to cite the author of that copy in his own isniid.) Furthermore, a shaykh, over the years of lecturing, might well present the material differently at different times, thus providing another point of departure for the existence of differing transmissions or recensions of a work. This process of diversification of an authority's teaching was further aided by several factors, according to Schoeler. For a time many scholars did not write and publish their own works, but preferred to leave that task to their students. The MuwattaJ of Malik ibn Anas is a good example of this practice. Therefore, it is not always possible to distinguish between author and transmitter during the early centuries of Islam. Thus Schoeler has argued for a different understanding of the transmission of early Islamic material. He has tried to mitigate the strict distinction between oral and written transmission of materials and thus, "seen correctly, it appears that here writing and orality are more complimentary than mutally exclusive." 117 This theory accounts for the variation observed in different recensions of the teachings of a particular authority. Yet at the same time the authenticity of each recension is maintained. Moreover, not only does the variation itself preclude any attempt to reconstruct the "original" (in the manner advocated by 35
The Droelopment of Exegesis in Early Islam
Sezgin), but so too does the possibility that no original text may have ever existed or served as a prototype for the various recensionsY 8
H. Motzki and the Implausibility of Fabrication Harald Motzki is another scholar who attempts to rectify what he feels are at times the extravagant and unsubstantiated claims made by Schacht. The method he employs looks at both the contents of the isnad and the matn to determine the plausibility of fabrication.II 9 Motzki focusses on the Mu$annaf of cAbd al-Razzaq al-~ancani (d. 211/ 826). This work contains composite riwayat (transmissions), but ninety percent of the IJ,adzths go back to a single transmitter, a common link as it were, which, for Motzki, implies a written text. 120 A statistical analysis of cAbd al-Razzaq's informants shows a divergent pattern in their informants, which Motzki states is inconsistent with the arbitrary manufacture of these IJ,adzths by cAbd al-Razzaq. Furthermore, cAbd al-Razzaq occasionally expresses doubts about his sources and provides anonymous transmissions. This too seems unlikely if these IJ,adzths were fabricated. Having shown to his satisfaction that cAbd al-Razzaq did not forge IJ,adzths, Motzki asks the same question ofcAbd al-Razzaq's sources by focussing on a representative selection of IJ,ad'iths going back to Ibn Jurayj (d. 150/767). These IJ,adzths, when analysed statistically, demonstrate an uneven and sporadic use of many earlier authorities. This strange distribution of authorities and his obvious willingness to express his own opinion without reference to earlier authorities (that is, his use of ra"y) belie the assumption that Ibn Jura)j forged IJ,adzths. This is confirmed by an examination of Ibn Jura)j's sources, which reveal much diversity: variance in content (for example, the use of ra"y is unevenly distributed); the variance in the use of pupil/teacher, son/father, and mawla/patron transmissions (though for any given transmitter, the use of such transmissions is consistent and occasionally even exclusive); variance in the proportions of IJ,ad'iths from the Prophet, Companions, and Successors; variance in the use of isnads; and the variance in the terminology of transmission (for example, the use of can, "from," versus the use of samictu, "I heard"). Each source seems to have an individual character. Motzki argues: Such a diversity can hardly be the result of systematic forgery, but, rather, must have developed over the course of time. We must therefore - until the contrary is proven - start from the assumption that the traditions for which Ibn Jura)j expressly states a person as his source really came from that informant, and thus Ibn Jurayj's transmission, in my opinion, should be regarded as authentic. 121 Motzki does not stop there, however. He attempts to go to yet another level, focussing on the most frequently cited of Ibn Jurayj's sources, cAra, 36
I:Iadi:th Criticism
ibn Abi Rahal) (d. 115/733). Once again he finds evidence against Schacht's theory of the systematic backwards growth of isnads. This evidence consists of two types. What Motzki describes as extrinsic evidence consists of the variance in the genres (that is, responsa and dicta of cAt;l 3 ) , variance in the types of questions in the responsa (direct, indirect, and anonymous), and variance in the actual postitions taken by Ibn Jura)j and his teacher cA!a 3 • That which is described as intrinsic refers to: lbnjurayj's willingness to give his own raJy without projecting it back (that is, Ibn Jura)j feels he is under no obligation to do so); Ibn Jurar.j's commentary on cAta 3 's comments (that is, it seems rather implausible that he first invented texts, and then commented upon them); his use of third-person, indirect transmission from cAta 3 (that is, why would he bother if he were given to forgery?); his occasional uncertainly about the cAt;l 3 's wording; his supplying of variant traditions; and his "hints of deficiencies of cA!a3 • " 122 Motzki goes back yet another level, but cAt;l 3 's ~adrths primarily employ raJy (eighty percent). This implies that he either did not rely on ~adrths or that he did not know many }Jadrths (perhaps because during his time few were in circulation). Nevertheless, when he examines cA!a3 's transmission from Ibn cAbbas, he again finds the usual variance in what few data there are. Motzki sees in cAt;l 3 's infrequent use of earlier authorities, his citations of Ibn cAb bas in various ways, his willingness to contradict Ibn cAbbas, and his variations in style and content evidence for the authenticity for these traditions. Motzki abandons his "variances" argument when he turns to the final level of transmission. There are just too few prophetic ~adrths from either Ibn cAbbas or cAta3 to draw many conclusions. Motzki turns instead to a specific }Jadfth and concludes that there is ample evidence to suggest that cAbd al-Razzaq knew the ~adrth. He argues that since cAbd al-Razzaq knew of it, it can be dated to at least the second half of the first century A.H., which undermines Schacht's assertion that it is from the second quarter of the second century and his general assertion that the more complete isnads are the later ones. 123 Moreover, he argues that "since there is only a generation between cA!a3 and Mul)ammad, these texts are very close to the time and the people they report about, and their authenticity cannot be ruled out a priori - as Schacht has done. "124 In a similar study Motzki also addresses claims made by Schacht when he looks at the jurisprudence of Ibn Shibab al-Zuhri (d. 124/742). Though no doubt some ~adrths which contain al-Zuhri as transmitter are authentic, Schacht would find it impossible to regard them as authentic unless they could be positively shown to be so. 125 However, when Motzki compares the four different lines of transmission ending in al-Zuhri (one from each of the two main recensions of the MuwatfaJ and two from the Mu~annaj), using a few specific examples where some or all of the texts contain related or parallel ~adfths, he concludes that these different versions of al-Zuhri's 37
The Development of Exegesis in Early Islam
jiqh concur in content in most cases and are identical in some. Only rarely do they disagree with one another. Motzki sees in these results strong evidence for trusting the }Jadfths that end in al-Zuhri.l 26 Motzki has argued, at least in the case of the Mu~annaf of cAbd al-Razzaq, that both the matns and the isniids which support them can largely be trusted. That is not to say that he does not concede that }Jadfths were forged. Rather, he suggests, "the mere fact that a}Jiidfth and asiinfd were forged must not lead us to conclude that all of them are fictitious or that the genuine and the spurious cannot be distinguished with some degree of certainty." 127 However, Motzki's comparison on the basis of isniids do seem to preclude systematic fabrication. But it is precisely the isniids that sceptics would say have been fabricated and so should not be the basis of any comparison. Moreover, Motzki's observed "consistent individual character" could be a product of separate fabrications: systemic fabrication need not be systematic fabrication. 128
J. Horovitz, J. W. Fiick, J. Robson, N.J. Coulson, and U. Rubin There are of course many other scholars who have contributed to this debate. Generally speaking, they have felt that the claims made by Goldziher and Schacht, though in part convincing, needed to be tempered. Although they are not major participants in the debate, ]. Horovitz,]. W. Fiick,]. Robson, N.J. Coulson and U. Rubin certainly merit some brief attention. Josef Horovitz addresses the problem raised by the use of isniids (or rather the lack thereof) in the extant writings of curwa ibn Zubayr (d. 92-101/711-20) and by the limited use of them by Ibn lsl_laq (d. 151/ 768). The latter historian did not provide isniids, or provided defective ones, in his biography of Mul_lammad. Caetani had concluded from this that the origin of the isniid must be placed between these two scholars and its perfection still much later in the third century. 129 To this Horovitz responds by pointing out that Ibn lsl_laq's teacher, al-Zuhrl, was already using composite isniids (Sammelisnad) and so argues that "since the use of the isniid in this composite form hardly allows one to imagine that the use of the simple isniid was not already customary for some time." 130 And so for Horovitz the isniid is at least older than al-Zuhrl- though how much so is less certain. As for curwa, who does not seem to use isniids consistently, 131 Horovitz argues that this absence does not mean he did not use isniids (and that the isniids ascribed to him in later historical works are spurious). The material used to make this assessment was a letter written in answer to a request by the Umayyad caliph cAbd al-Malik. Horovitz asserts that the way one writes a letter is different than the way one writes for scholarly purposes. And so Horovitz has no qualms about giving the first use of the isniid in }Jadfth literature as not later than the last 38
I:Iad!th Criticism
third of the first century A.H. 132 Of course, whether the close connection between historical material and other genres of }Jadzth material (on which both sides of the argument rely) can be made, remains debatable.l 33 Johann Flick in discussing the role of }Jadzths in Islam makes an insightful observation. The Companions most frequently cited in isnads as authorities are the younger ones. For example, more }Jadzths cite Abu Hurayra and Ibn CAbbas than cite Abu Bakr and cuthman. This fact has been noted before and used as evidence for the spurious nature of the isnads, for the older Companions should have had more to say about Mul).ammad. Flick reaches the opposite conclusion. He argues that if all isnads were spurious, then it would be more likely for the older Companions to be cited more frequently. In other words, if one is going to the trouble of inventing an isnad, why not simply attach it to an older, more respected Companion? Since the transmitters have not done that, then perhaps the isnads are genuine.l 34 James Robson in a series of articles began to look at the origin of the isnad. 135 While he largely agrees with Schacht's conclusions, Robson would like to confine them primarily to the legal realm, "a sphere where his argument may apply more closely than elsewhere, as changing conditions and the development of legal thought must have demanded new regulations; but one wonders whether the argument is not too sweeping. "136 Robson particularly exempts historical }Jadzths from this scepticism about authenticity.l 37 He offers little evidence in support of this, except the argument that it seems logical that Mul).ammad's followers, because of the impression his personality must have made on them, preserved a genuine core within the }Jadzths. 138 Robson then turns to the question of isnads: even if there is a genuine core, that does not mean the isnads attached to them are genuine. And it is in the discussion of this that Robson makes his own contribution. Once again he offers no textual evidence to support his claim, just the following scenario: During the time of the Companions, the words and deeds of Mul).ammad were simply discussed amongst themselves; no demand for authority was necessary. Mter the death of many of the Companions, Muslims would continue to speak about Mul).ammad, but now, gradually, they might be asked to cite their authority. Thus Robson feels that the middle years of the first century saw the first use of isnads, albeit in a very informal manner. The growth to a formal isnad system was gradual: even Ibn Isl).aq did not feel obligated to use them. AnticipatingJuynboll, Robson accepts the report about Ibn Sirin concerning the beginning of the use of isnads (but places the fitna referred to not in 126 A.H. as Schacht had, but in 64 or 72 A.H., that is, when cAbd Allah ibn Zubayr set up the countercaliphate in Mecca) and so concludes that the use of isnads stems from the last third of the first century. From examples drawn from how Ibn lsl).aq's isnads were preserved in later works, Robson concludes that the matn can 39
The Droelopment of Exegesis in Early Islam
be altered even if an isniid has been attached and remains unaltered, though the kernel of information remains the same. 139 Drawing on Fiick and Horovitz for support, he also insists that despite the spurious material, there remains "some genuine early material." 140 In an incidental point, Robson criticizes Schacht's complete rejection of family isniids as simply a method to attempt to authenticate spurious }Jadzths. Robson points out that the idea to use family isnads in this manner likely resulted from genuine family isniids that served as paradigms for the spurious ones. N.J. Coulson is another scholar who takes issue with some of Schacht's conclusions. Despite praising Schacht for having "formulated a thesis of the origins of Shar(a law which is irrefutable in its broad essentials", 141 Coulson has debated with Schacht over the dating of }Jadzths.l 42 Coulson asserts that Schacht "translates the negative proposition that the evidence of the lfadzth does not take us back beyond the second century of Islam into the positive statement that legal development began only in late Umayyad times." 143 That is, development in law has been too closely identified with the development of }Jadzths by Schacht. Coulson, on the other hand, maintains that there must have been legal activity prior to the turn of the first century surrounding, at the very least, issues broached by the Qur 3 an itself. Furthermore, Coulson admits that, though an isniid may be fictitious, this does not mean the substance of the matn has also been fabricated - it may well represent a precept of Mu}:lammad. Coulson explains: This is not to suggest that the chain of transmission, or the isniid, of this Tradition or that is authentic, for this is, in the great majority of cases, demonstrably not so; but it is suggested that the substance of many Traditions, particularly those which deal with the obvious dayto-day problems arising from the Qur 3 anic laws, may well represent at least an approximation to a decision of the Prophet which had been preserved initially by general oral tradition. If this practical premise is accepted then it is a reasonable principle of historical enquiry that an alleged ruling of the Prophet should be tentatively accepted as such unless some reason can be adduced as to why it should be regarded as fictitious. 144 While certainly not accepting the traditional Islamic view of }Jadzths, Coulson takes issue with Schacht's proposition that each and every legal}Jadzth said to come from the Prophet must be assumed to be wholly unauthentic until proven otherwise, and must be thought of as "the fictitious expression of a legal doctrine formulated at a later date". 145 Coulson states quite bluntly that "the truth lies somewhere between traditional Islamic legal theory and the rigorous historical approach of Schacht."146 Uri Rubin stands somewhere between the sceptics and those who search for middle ground. Concurring with the sceptics, at least with regard to historical }Jadzths, he writes 40
I:Jadith Criticism
The bulk of the texts about the Prophet embody the literary product of Islamic religious devotion, and therefore they will be treated ... not as a door opening onto the "historical" events which are described in them, but rather as a mirror reflecting the state of mind of the believers among whom these texts were created, preserved, and circulated through the ages. 147 The image of Mu}:lammad in the Muslim sources, Rubin argues, is the reflection of the community's self-image. Local Arabian and Qur"anic models, coloured by tensions with medieval Islamic society, were combined to "provide Mu}:lammad with a proper prophetic vita." 148 From the standpoint of the authenticity debate, Rubin's more noteworthy contribution are is conclusions about isnads. Although he recognizes that isnads were fabricated, their presence was always designed to make the reports to which they were attached appear authentic. Efforts, therefore, to determine their authenticity would seem to be futile. 149 However, Rubin's "systematic textual analysis" cuts at the very heart of the claims first made by Schacht and developed by others such asJuynboll, Cook, and Calder. Their assumption that there was "backward growth" of isnads, which simply stated claims "the shorter isnad, the earlier the matn," is wrong. Rubin argues that most of the prophetic and Companion isnads could have first circulated in the first century A.H., while the Companions yet lived. Rubin bases this claim by demonstrating that the names of Successors do not recur in the Companion isnads, regardless of whether they are prophetic or not. In other traditions, Mu}:lammad and the Companions are part of the "original hard core. "150 In yet other traditions, the isnads of prophetic J;,adzths have nothing in common with the nonprophetic versions. All these observations lead Rubin to conclude that there was no backward growth of isnads. That is not to say that Rubin believes that the isnads guarantee authenticity, rather the evidence oflack of backward growth of isnads deprives Schacht of one of its basic dating tools. From what we have seen, traditions with complete isnads, including a Companion, could have come into being as early as the generation of the Companion himself. In later generations, the higher parts of the isnads either remained static or shrank, sometimes with the introduction of changes in the matn.l 51 Horovitz, Fiick, Robson, Coulson, and Rubin (to a lesser extent) represent the movement, of which Juynboll, Rahman, Schoeler, and Motzki are the vanguard, to try to find some historical relevance in the vast l}adzth material by (at least for these four scholars) gradually picking away at the arguments that serve as the foundation for the conclusions of Goldziher and Schacht. 152 In their stated approach to early Islamic texts, those who search for a middle ground have been influenced by sceptical scholars such as 41
The Development of Exegesis in Early Islam
Goldziher and Schacht. However, their final conclusions largely resemble those of sanguine scholars, such as Sezgin and Abbott. This is not too surprising. Most of the arguments are based on similar assumptions: the rijal material and other purportedly early sources are trusted. Since they start with the same assumptions, they reach the same conclusions though tempered somewhat. It should also be noted that many of the scholars who are sanguine about isnads and those who search for a middle ground often cross-reference each other. This might give the impression of an emerging consensus or an overabundance of evidence against the sceptical approach. However, as I have tried to demonstrate, most of these scholars are presenting essentially one argument, over and over again.
RENEWED SCEPTICISM With the attack of Abbott, Sezgin and Azami, the debate might have stagnated into one involving the interpretation of certain words and a few reports which speak of the beginning of the use of isnads. However, a far more sceptical attitude than even that of Schacht's has emerged. John Wansbrough, also picking up where Goldziher had left off, has attempted to demonstrate that all early Muslim texts, whether historical, legal, exegetical, or grammatical, emerged much later and in a much different context than was previously thought. 153 Wansbrough seems to find the acceptance of "isnad apparatus" by Abbott and Sezgin to be rather naive. 154 They are all interdependent and no one particular genre stands as an objective, independent source from which to evaluate the claims of another. Michael Cook and Norman Calder also follow the more sceptical path from Goldziher and Schacht akin to that of Wansbrough. Neither Cook nor Calder deals with issues of authenticity, chronology, and provenance of IJ,adtths in general. Their interests lie in early Muslim theology and jurisprudence respectively, 155 and so each has had to deal with questions of authenticity and the dating of IJ,adtths, whether dogmatic, historical, or legal in nature. The specific issue both scholars have addressed is the usefulness of Schacht's common-link theory for dating IJadtths and argued that it should be subsumed under his theory of the spread of isnads.
M. Cook and the Spread of Isnads In challenging some of the conclusions reached by Joseph van Ess 156 regarding the origin and development of the predestinationist/Qadarite controversy in early Islam, Michael Cook prefaces his analysis of specific texts with a more general discussion of the dating of 1Jadtths. 157 While recognizing that van Ess's method is clearly an orientalist one- that is, one which does not date IJ,adtths by mere ascription - Cook questions the
42
I:Iadith Criticism
validity of his method. Van Ess, like Schacht, recognizes that isnads grew backwards and he accepts the common-link theory. That is, when the isniids of parallel matns seem to converge on a particular transmitter, that transmitter may serve as a terminus ante quem. 158 Cook, in critiquing this method, draws and expands upon another one of Schacht's insights: the spread of isniids - that is, the creation of additional authorities or transmitters for the same matn. This spread of isniids can occur in several ways: (1) Skipping a contemporary transmitter. Hypothetically speaking, assume that Ibn Jubayr transmitted a matn from Ibn cAb bas to Ibn Jura)j. If Ibn Jurayj is scrupulous in his transmitting, he will include lbnJubayr in the isniidwhen he in turn passes on the matn (as illustrated in Diagram 6). This he might not want to do, however. First, he might not want to be seen transmitting from a mere contemporary. Second, "an elegant isniid is a short one. Ideally one should have a saying direct from the mouth of the sayer; and failing that, the fewer the intervening links the better. "159 So Ibn Jurayj might "skip" Ibn Jubayr, by claiming that he heard the matn directly from Ibn cAbbas (and the resulting isniids would look like those in Diagram 7). (2) Ascribing a saying to a different teacher. Hypothetically speaking once again, let us assume that cAbdallah tells Ibn Sacld a matn he has heard from IbnJubayr who has it from Ibn cAbbas. Ibn Sacld, instead of claiming to have heard it from Ibn Jubayr (resulting in the isniid shown in Diagram 8), ascribes the tradition from Ibn cAbbas to his own teacher Ibn Jurayj (resulting in the isniids
Diagram 6
Mul:lammad
I
Ibn cAbbas
I
lbnjubayr
I
lbnjura)j
Diagram 7
Mul:lammad Ibn cAbbas
lbnjubayr
lbnjura)j
43
The Development of Exegesis in Early Islam
shown in Diagram 9). This might be done because it is known that he never met Ibnjubayr or that lbnjubayr is not recognized as an acceptable transmitter in Ibn Sacld's own school. Thus he appropriates the matn in a way that avoids these problems. (3) Obviating the "isolated" charge. Because a well-attested IJ,adfth carries more weight, there would be a strong motivation to "discover" other isnads. 160 All of these methods of creating new isnads, particularly (1) and (2), create the appearance of a common link. In Diagrams 8 and 10, it is from Ibn cAbbas that the isnads appear to fan out, and so it appears that he is the originator of the IJ,adfth. As shown by these hypothetical examples, this need not be the case at all. Thus all three of these methods, if they took place, affect the "common link" and its ability to provide any historically useful information.l61 Cook points out however, that it has yet to be determined whether "the spread of isnads was a process operative on a historically significant scale, or just an ingenious idea of Schacht's."162 With respect to van Ess, Cook points out that the spread-of-isnads theory undermines his chronology of the predestinationist debate - or at least points out a significant vulnerability. This scenario, if it was practiced on a large scale,
DiagramS Mul).ammad Ibn cAbbas IbnJubayr
I
n~i.stcnt
not
Scenario Ib
likely authentic:
i~nads
Comsistent
Scenario II a
(Orrect in some form
iwads
lll)t (Onsi.stcn t
Scenario Ilh
S\stematic:allv fabricated
lsnad~
c:o>nsi~tent
Scenario lila
hapha7arrily fahricated
ilmld;;
not consistent
Scenario IIIb
authentic
HnlUb likeh
consistent
Scenario IVa
fabricated
i.mad.' likdv
con~iMent
not
---
Scenario IVb
as yet unknown
i}nad~
likdv fabricated
inconsistent with Ihn 'Ahha~
iHwili correct in some form
with
Scenario lV
inconsistent with Ibn f lhn aQ.J:iak ibn MuzaQ.im al-HilaH al-KhurasanL see Ibn I:Iajar, Tahdhib, 4:417-8. The transmission from al-I;>aQ.Q.ak is considered weak since he never met Ibn cAbbas, but received l}adiths from Saci:d ibn Jubayr. 112 For Abu MuQ.ammad cAp., ibn Abi: RabaQ. al-Makki: see Ibn I:Iajar, Tahdhib, 7:174-7, Ibn Sacd, al-Tabaqat, 2:386--7, and Schacht, "cAp., b. Abi: RabaQ.," 1:730. 113 For Abu al-Khanab Qatada ibn Dicama ibn Qatada al-Sadusi: al-Ba~ri: see Ibn I:Iajar, Tahdhib, 8:306-10, and Pellat, "I_(atada b. Dicama," 4:748. 114 Fi:ruzabadi:, Tanwir al-miqbas min tafsir Ibn cAbbas. There are numerous earlier editions of this work. It is sometimes referred to as a work of pseudo-al-KalbL 115 Brockelmann, Supplement, 1:331 and 2:235. 116 Sezgin, Geschichte, 1:27 and 1:34-5. 117 Rippin, "Tafsir Ibn cAbbas," pp. 4-47. See also his earlier, but less developed, conclusions in "Al-Zuhri, naskh al-Q}tr"an and the Problem of Early tafsir Texts," pp. 23-4. 118 Rippin concludes: Elements such as the variability in citation of variant readings would seem to be evidence of this. To me, al-Di:nawari:'s introduction is likely to be an accurate record of how the tafsir came into being: it is some sort of distillation of knowledge written for the purpose of introducing the complexity of the Qur 0 an to budding students, done with reference to al-Kalbi:'s tafsir yet probably stemming not from al-Di:nawarl himself but from one of his teachers or from within his circle. That we have anyway of extracting what would be the opinion of al-Kalbi: from the rest of the text is quite plainly not the case; the opinions expressed can be taken only as an expression of the later author's opinion and learning. Rippin, "Tafsir Ibn cAbbas," p. 71. See also Wansbrough, Q}tranic Studies, p. 133 and Patricia Crone, 'Jahill and Jewish Law," p. 174, n. 111, for evidence suggesting that the elliptical form regarding references in the Tafsir Ibn cAbbas presupposes both later masoretic and haggadic materials. 119 Rippin, "Tafsir Ibn cAbbas," pp. 71-4. This is also the opinion of Gilliot. In speaking of these popular "Tafsirs" of Ibn cAbbas, Gilliot says: their existence confirms the highly mythic functions of Ibn cAb bas in the genesis and development of Muslim exegesis. However, these "popular" versions of the exegesis of Ibn cAbbas must not be confused with the views of this same exegete which one finds in the great classical commentaries like that of al-Tabari:, with chains of well-known and above all wellrespected guarantors, which does not mean, for all that, that they permit the reconstruction of a so-called Tafsir of Ibn cAbbas. It is appropriate however to be prudent in this domain when one makes the distinction between popular tradition and scholarly tradition. Gilliot, "Les debuts," p. 85. (my translation)
120 121 122 123 124
An example of these reconstructions is the Tafsir Ibn cAbbas reconstructed by Rashid cAbd al-Muncim al-Rajjal. Versteegh, Arabic Grammar, pp. 61-2. Sezgin, Geschichte, 1:27. Sezgin, Geschichte, 1:27-8. Rippin, "Ibn cAbbas's Gharib al-Q}tr"an," pp. 33. Rippin, "Ibn cAbbas's al-Lughat .ft "l-Q}tr"an," p. 16. See also Wansbrough, Q:tranic Studies, pp. 216--7. 165
The Development of Exegesis in Early Islam
125 Rippin explains: The point of this extended summary of tendencies in this text [al-Lughiit 'l-Qur'an] is this: all these factors, the nature of the definitions, the technical terminology and the methodologies, may well not be exceptional in themselves; however, to find them all combined in one text is, I would suggest, evidence that we are dealing with a work whose genesis is after each of the individual procedures had become established processes and components within the exegetical canon. It is perhaps not a very surprising conclusion to state that this text, like Masii'il and Gharib, although attributed to Ibn cAbbas, can hardly stem from him. Rippin, "Ibn cAbbas's al-Lughiit ft 'l-Qur'an," p. 25.
ft
126 127 128 129 130 131
132 133 134 135 136 137 138 139 140
On the basis of another manuscript, Neuwirth argues that there is a Masii'i? nucleus that dates from the early second century. However, even that date makes the text too late to have been penned by Ibn cAbbas himself. Angelika Neuwirth, "Die Masa'il Niific b. al-Azraq" pp. 233-55. Boullata after examining several extant manuscripts, concludes that the masii'il is a "composite text whose historical fluidity permitted continual growth." Boullata, "Poetry Citation," p. 34. Nevertheless, on the basis of "Arab proclivity to cite proverbs or poetic verses orally", he suggests that the some of the materials "may very well be authentic." Boullata, "Poetry Citation," pp. 38 and 40. For a less romanticized view, see Gilliot, "Textes arabes anciens," pp. 317-19. Sezgin, Geschichte, 1:28. See page 121. See pages 130-l. Yaqut, Irshad, 6:426. Al-Tabari, Annates, 3:2295-561. The underlying assumption is that the informants of al-Tabar1 are actual people who instructed al-Tabar1 in either written or oral form. This assumption seems a valid one for two reasons. First, some of these informants were still alive when al-Tabar1 began writing and would have been able to challenge him if he deviated from their teaching. Second, al-Tabar1 had some opponents (the nascent f.lanbalis) and if they believe he had himself fabricated matns or isniids, they would not have hesitated to publicize such forgeries. See page 121. Compare with Motzki's assertion that variance implies "individual character." See Chapter 2, pages 36-8. It is also possible that the "transmitter" simply attached isniids to pre-existing material. Since my goal is to determine the value of the isniids, the actual origin of the matns is of lesser concern. See Chapter 3, page 96, note 26. Horst states these isnads occur 1,560, 970, and 560 times respectively. Horst, "Zur Uberlieferung," pp. 293-4. The term 'an is translated here as "from." However, it is frequently translated as "on the authority of' and thus need not imply direct audition. 180 times according to Horst, "Zur Uberlieferung," pp. 299-300. See pages 148-57. See page 167, note 148. This attempt to avoid the Muslim framework is not meant to deny the existence of the people named in fabricated isnads, but to suggest their presence in isnads and the rijiil material does not guarantee their existence nor their status as a tradent. For example, there are persons named Abu Salil~
166
Methodology: Isnads and Exegetical Devices
and cAbd Allah ibn SaliQ., as pointed out above. That these two names refer to the same individual for al-Tabarf is evident in the interchangeability of these names in otherwise similar isnads and in the fact that he himself combines the two names at times. If al-Tabarr had not made this identification, the biographical material could provide other possibilities, not definite answers. See discussion page 167, note 148. 141 This seems possible since the former is listed in Ibn I:Iajar, Tahdhzb, 9:214-5 and the latter is not. The probability of a simple typographical error having occurred is further supported by the fact that the isniid in which MuQ.ammad ibn cAbd Allah ibn Zarfc appears is later adduced with the name of MuQ.ammad ibn cAbd Allah ibn Baz( instead. Al-Tabarf, Tafs!r, 2:84-5 and 4:181. Recourse to the original manuscript would be necessary. 142 Juynboll, "Mucancan," 7:260. pp. 293-4 andJuynboll, 143 Juynboll, trans. "Muslim's Introduction to His $a~!~," "Mucancan," 7:260. 144 Juynboll, trans. "Muslim's Introduciton to His $a~!~," pp. 294-302. p. 351, who groups 145 See, for example, cAtar, Manhaj al-naqdfz culiim al-~adfth, the mucancan isniids with the mutta~il (uninterrupted) isniids. 146 There is, however, a need to distinguish passive constructions such as ~uddithtu, "I was told," because they imply that the name of the source is unknown or not recorded. Such a isnad is not classified as mutta~il. 147 For example, al-Tabari, Tafs!r, 1:98. 148 It might be useful do demonstrate with this particular isniid the information that the rijal material can provide. Musa ibn Harlin al-Hamdanf is unidentifiable. See Horst, "Zur Uberlieferung," p. 302, n. 3, and al-Tabarf, The History of al-Tabari Volume I, p. 206, n. 273. cAmr ibn I:Iammad al-Qannad is cAmr ibn I:Iammad ibn TalQ.a ibn al-Qannad, Abu MuQ.ammad al-Kufi but Miisa ibn (d. 222/837). Among those from whom he transmits is Asba~ Harlin is not listed among those who transmit from him. He is considered reliable. See Ibn I:Iajar, Tahdhlb, 8:20. Asba~ is Asba~ ibn N~r al-Hamdani, Abu Yusuf, called Abu Na~r (death date not given). He transmits from Ismacrl al-Suddr, among others, and to cAmr ibn I:Iammad al-Qannad, among others. He is considered reliable. See Ibn I:Iajar, Tahdhfb, 1:192. Ismacrl ibn cAbd al-RaQ.man al-Suddr is Ismacrl ibn cAbd al-RaQ.man ibn Abr Karfma al-Suddr, Abu MuQ.ammad al-QurashL (d. 127 /744-5). He transmits from Ibn cAbbas, Abu SaliQ., among others. Abu Malik is not listed as one of his informants, nor is Asba~ listed as one of those who transmit from him. He is generally considered reliable. See Ibn I:Iajar, Tahdhlb, 1:282-4. Abu Malik is Ghazwan Abu Malik al-Ghifarf al-Kufi (death date not given). Among those from whom he transmits are Ibn cAbbas and a man from the Companions of the Prophet. Ismacrl al-Suddr is one of those who transmit from him. He is considered reliable. See Ibn I:Iajar, Tahdhlb, 8:213. Abu SaliQ. is Badham who is called Abu SaliQ., mawlii ofUmm Hani" the daughter of Abu Talib (death date not given). He transmits from Ibn cAbbas, among others, and to Ismacrl al-Suddr, among others. He is considered unreliable and a liar by some and it is stated that he did not hear ~adfths from Ibn cAbbas, even though he transmitted them. See Ibn l:lajar, Tahdhlb, 1:379-80. Murra al-Hamdanf is Murra ibn Sharahrl al-Hamdanf al-Bakrlr, Abu Isma"rl al-Kufi (d. 76/695-6). Among those from whom he transmits is Ibn Masciid and among those who transmit from him is Ismacrl al-SuddL He is considered reliable. See Ibn I:Iajar, Tahdhlb, 10:81. Ibn Masciid is cAbd Allah ibn Masciid ibn Ghafil ibn Habib ibn Shamj ibn Makhzum ibn Sajira ibn Kahil ibn al-l:larith ibn Tamfm ibn Sacd ibn Hudhayl
167
The Development of Exegesis in Early Islam
149
150 151 152
153
154 155 156
ibn Mudraka ibn Ilyas, Abu cAbd al-Ra}:!man al-Hudhalf (d. 32-3/652-3). He transmits from the Prophet, and Ibn cAbbas and many others transmit from him. As a Companion, his reliability is not questioned. See Ibn I;Iajar, Tahdh!b, 6:26-7. Given this information, the aforementioned complex isnad is easily understood. It is composed of the isnads: Musa - cAmr - Asba~ - al-Suddr - Abu Malik - Ibn cAbbas Musa - cAmr - Asba~ - al-Suddr - Abu Sali}:! - Ibn cAbbas - al-Suddr - Murra - Ibn Mascud and other ComMusa - cAmr - Asba~ panions However, it is still possible that Ibn cAbbas is not a source, but a transmitter, since he too transmits from Ibn Mascud. Rosenthal, trans., History of al-Tabari Volume I, p. 206, assumes that he does not. The larger problem with this data is that it is incomplete and posterior to the isnads. The dates of death are often missing as are names of the people to whom and from whom the tradents transmitted. Contemporaneity cannot be established. Furthermore, it seems that the supplied lists of people to whom and from whom they transmitted which are compiled from examining isnads, not from some outside source of knowledge. In other words, the lists of students and informants is tautologous. As such, the information provided by this rijal material is of little practical value. For example Hannad [ibn al-Sari (d. 243/857), see Ibn I;Iajar, Tahdh!b, 11:62] - cAbda [ibn Sulayman (d. 187-8/803-4), see Ibn I;Iajar, Tahdh!b, 6:399-400] - Sacid ibnJubayr [(d. 95/714), see above, page 177, note 108]- Ibn cAbbas is clearly missing several links. Al-Tabari, Tafszr, 2:142. Shakir and Shakir suggest that the isnad should be like the one which precedes it. Therefore it should read: Hannad- cAbda- Sacid ibn Abi cAruba [(c.155/722), see Ibn I;Iajar, Tahdh!b, 4:56-9] - Qatada [ibn Dicama, see page 177, note 113] - cAzra [ibn cAbd al-Ra}:!man, see Ibn I;Iajar, Tahdhfb, 7:168-9] - Sacfd ibn Jubayr- Ibn cAbbas. Al-Tabari, Tafszr, eds. Shakir and Shakir, 3:428, n. 1. For another example, see the commentary for Qur'an 41:44 for which al-Tabari adduces a variant reading without an isnad. He simply introduces it with "it is related on the authority of Ibn cAbbas." Al-Tabarf, Tafs!r, 11:119. One isnad of the sample set had 6 branches, two had 3 branches, 60 had 2 branches and the other 834 were single-branched. They are also counted as a separate isnad. Another student of Ibn cAbbas who cited commonly is cA~'. Unfortunately, whether this refers fo cA~' ibn Abi Raba}:! orcA~' al-Khurasani is uncertain. Both are students of Ibn cAb bas, and both transmit to many of the same people as well. Because it is so difficult to distinguish between the two transmitters, these Ibn cAbbas-cA~'-~adfths cannot be used to construct a stylistic profile. al-Muthanna ibn Ibrahim al-Amuli (d. after 240/854) - see Horst, "Zur Uberlieferung," p. 293, n. 2; and Rosenthal, "General Introduction," pp. 17 and 190, n. 179. See also Sezgin, Geschichte, 1:20, 27, 35, and 41; and Gilliot, "La Formation intellectuelle de Tabari," pp. 205-6. Abu Kurayb Mu}:!ammad ibn al-cAla' (d. 247 or 248/861-2) -see Ibn I;Iajar, Tahdhfb, 9:333-4, and Rosenthal "General Introduction," pp. 20-1, and 176-7, n. 77. Sufyan ibn Wakfc ibnJarra}:! (d. 247/861)- see Ibn I;Iajar, Tahdhzb, 4:111-2, and Horst, "Zur Uberlieferung," p. 296, n. 7. Abu cAbd Allah Mu}:!ammad ibn I;Iumayd al-Razf (d. 248/862) - see Kha~b al-Baghdadi, Ta'nkh Baghdad, 2:259-64; Ibn I;Iajar al-cAsqalanf, Tahdhfb 168
Methodology: Isnads and Exegetical Devices
157 158 159
160 161 162
163 164
165
al-tahdhzb, 9:108-11; Horst, "Zur Uberlieferung," p. 296, n. 3; and Rosenthal, "General Introduction," pp. 17-9 and 172, n. 26. See also Sezgin, Geschichte, 1:29, 30, 79, 242, and 253. al-Qasim ibn al-l:lasan ibn Yazid al-Hamadhanl (d. 272/885) - see Khapb al-Baghdadi, Ta'nkh, 12:432-3, and Horst, "Zur Uberlieferung," p. 295, n. 1. Mu}:!ammad Abu Musa al-Zamin ibn al-Muthanna (d. 250-2/864-6) - see Ibn I:Iajar, Tahdhfb, 9:368-9, and Khapb al-Baghdadi, Ta'nkh, 3:283-6. Mu}:!ammad ibn Bashshar (also known as Bundar) (d. 252/866) -see Kha~ib al-Baghdadi, Ta'nkh, 2:101-5, Ibn I:Iajar, Tahdhfb, 9:58-60, Horst, "Zur Uberlieferung," p. 296, n. 4, and Rosenthal, "General Introduction," pp. 20-1, and 174, n. 44. Yacqub ibn Ibrahim ibn Kathlr al-Dawraql (d. 252/866) - see Ibn I:Iajar, Tahdhfb, 11:332, and Kha~ib al-Baghdadl, Ta'nkh, 14:277-80. See page 164, notes 108-10, for biographical information on these students of Ibn cAbbas. The distribution of the sample over the quranic material is worth noting, since the type of the material in the Qur'an varies with its location. That is, longer, more prosaic material appears in the longer silras at the beginning of the Qur'an. This variation might affect the exegetical device employed. The isniids that I have selected are distributed across the whole of the Qur'an, with each juz' having at least some IJ,adfths in the sample set. However, the distribution of isniids is not uniform; the first few ajzii' and the last one are over represented. Wansbrough, Qyranic Studies, p. 121. See Jeffery, Materials for the History of the Text of the Qyr'iin: the Old Codices, who collects the variant readings from numerous texts and groups them according to the Companion listed in the isniids. As his title suggests, he hopes thereby to reconstruct the quranic codices which were in circulation prior to their destruction with the promulgation of the cUthmanic text. Wansbrough argues:
From the material assembled by ... Jeffery, it could well be asked to what extent any of the variants, or variant codices (?), may be said to represent traditions genuinely independent of the cUthmanic recension. The infinitesimal differences are not such as would seem to have necessitated suppression of the non-cUthmanic versions, the more so since a minimal standard deviation from the canon was accommodated by the several interpretations of the alfruf doctrine. The choice between interpreting those codices as conscious (i.e. exegetical) variations upon the cUthmanic recension, or as having in common with that recension an earlier Vorlage, is not an easy one. Wansbrough, Quranic Studies, pp. 44-5.
166 167 168 169 170 171 172 173 174
Wansbrough decides that the cUthmanic recension traditions are a fiction and that the variant readings are not the product of an Urtext. Wansbrough, Qyranic Studies, pp. 203-4. Al-Tabarl, Tafsir, 1:94-5. Al-Tabarl, Tafs!r, 11:576. Al-Tabarl, Tafsir, 1:94. Al-Taban, Tafszr, 10:166. Al-Taban, Tafsir, 11:576. Al-Taban, Tafszr, 1:155. Al-Taban, Tafsir, 12:327. Al-Suyup, Itqiin, 1:119. 169
The Development of Exegesis in Early Islam
175 This reputation is built on the contents of the Masa'il text ascribed to him (see pages 135-6), which has Ibn cAbbas providing examples of poetry for 190 quranic passages it discusses. The IJadzths ascribed to him in al-Tabart's Tajsfr do not evidence any such proficiency with or reliance on pre-Islamic or early Islamic poetry to explicate quranic lexica. See also Wansbrough, Q}tranic Studies, p. 217. 176 Al-Bukhan:, fia/Jf/J, 4:403. 177 See al-Farra", Maciini al-QJtr'an, passim. Wansbrough also suggests that the use of poetry has a terminus a quo in the recension oflbn Hisham (d. 218/833) of Ibn Isqaq's Sfra. The redactor shows masoretic influences in his lexical, as opposed to the author's haggadic, approach to quranic vocabulary. Wansbrough, Q}tranic Studies, pp. 216-8. 178 AI-Tabar! frequently adduces a few verses of poetry in support of his personal tajsfr of a particular quranic locution. For example, for Qur"an 13:11 he cites the pre-Islamic poets Tarafa and Labid, without stating his sources, and Imru" al-Qays, citing "some of the grammarians from Ba~ra." AI-Tabar! provides only one IJadzth which adduces some of Labid's poetry to interpret a different part of the same verse. Al-Tabari, Tajsfr, 8:357-8. 179 Al-Tabarl, Tajsfr, 12:429. 180 See pages 152-3. 181 Al-Tabarr, Tajsfr, 8:389 and 1:482, respectively. 182 Al-Farra", Maciinf al-QJtr'an, passim. Other early lexical texts are those of al-Farra"'s teacher al-Kisa"l (d. 189/804) and ofMuqatil (d. 150/767). For the former, see al-Kisa"l, Mutashiibihiit al-Q;tr'iin and for the latter, see Muqatil ibn Sulayman, Al-Ashbiih wa-l-na'{.ii'ir ft al-QJtr'iin al-kanm. 183 See pages 154, 151-2, and 155, respectively. 184 Al-Tabarl, Tafs!r, 5:498. 185 That is not to say that simple glosses or paraphrastic equivalences are not lexical explanations. They certainly are in haggadic tajsfr, but need not be in other contexts. Therefore, for the sake of consistency, these have been classified separately. See page 155. 186 See pages 151-2. 187 Al-Tabarl, Tafs!r, 4:468. 188 See page 162, note 64. 189 Rippin, 'Tafslr," 14:239. 190 Ibn Qutayba, Ta'wfl mushkil al-Q;tr'iin. See also Wansbrough, QJtranic Studies, pp. 228-31. 191 For the earlier usage, see page 152. 192 Wansbrough, Quranic Studies, p. 228. 193 Wansbrough, Quranic Studies, p. 229. See also Jurjanl, Kitab asriir al-balagha. 194 Al-Tabarl, Tajsfr, 1:29-30. 195 See pages 127-8 and Gilliot, "Langue et Coran," pp. 79-106. 196 Wansbrough, Q;tranic Studies, p. 223. 197 Al-Zamakhsharl, al-Knshshiif, 4:54. Technically this example is not grammatical emendation, but semantic. Moreover it is also a variant reading and a textual analogy (see pages 152-3). See also Wansbrough, Q;tranic Studies, p. 222. 198 Al-Zamakhsharl, al-Kashshiif, 1:582. 199 "Here, it may be provisionally proposed that 'textual restoration', conventionally represented by the expression taqdzr, was also (and earlier) called majiiz, and further, that the latter term evolved from the vague designation of an exegetical practice to the closely reasoned description of several rhetorical phenomena found in scripture as well as in profane literature." Wansbrough, 170
Methodology: Isnads and Exegetical Devices "Majaz al-Qp,r"an: Periphrastic Exegesis." p. 248. See also Wansbrough, Qp,ranic Studies, pp. 219-20, and page 151. 200 Al-Tabarl, Tafszr, 10:146. 201 Wansbrough, Qp,ranic Studies, p. 166. 202 This is of course simplifying the use of qiyas somewhat, for "qiyaswas employed
203 204 205
206 207
208 209 210
211 212 213 214 215 216 217 218 219 220
both for extrapolation of fresh principles from existing premisses and for interpretation, as well as for establishment, of the scriptural text". Wansbrough, Qp,ranic Studies, p. 166. Al-Tabarl, Tafszr, 7:563. Other terms used include mithla, ka-qawlihi, and wa-hiya bi-manzila qawlihi which are used much like na~zr by al-Farra', Macan! al-Qp,r,an, 2:137, 157, and 5, respectively. See also Wansbrough, Qp,ranic Studies, p. 212. This too is a simplification of a term whose meaning evolved over time and came to incorporate more than just analogies drawn from quranic passages that shared a particular word. Wansbrough suggests that a work ascribed to Muqatil ibn Sulayman is a good, early example of this collation of various quranic passages for the purposes of a lexical discussion. See Wansbrough, Qp,ranic Studies, pp. 208-9. As Wansbrough points out, with al-Farra' and al-Zamakhshart, na~ir also refers to syntactical and grammatical analogies from scripture. See al-Farra', al-Macani al-Qp,r"an, 1:355 and 2:141-2, al-Zamakhshart, al-Kashshaf, 1:532, and Wansbrough, Qp,ranic Studies, p. 212. Muqatil, Tafsir, 2:588. The implications here are more theological than lexical. Mary was not sent a spirit, which might be too close for some Muslim scholars to the Christian teaching that Jesus was conceived by the Holy Spirit, but sent a mercy or blessing. Rather Jesus was sent to Mary, and he was a mercy or blessing to those who followed him. Al-Tabarl, Tafsir, 4:374. Al-Tabarl, Tafsir, 2:151. See pages 155-6. Other techniques include tak~iJ (specification) and tafsir (corroboration). The former is used to argue that one quranic verse states a general principle and another states a more specific or extended example of the same principle. For example, Qur'an 5:3 prohibits carrion and Qur'an 5:96 is thought to extend that prohibition to carrion from the sea. The latter is used to argue that the two verses do not contradict each other, but corroborate each other. Such is the method used by some Muslim scholars for Qur'an 9:5 and 4 7:4 which deal with war and prisoners of war. Wansbrough, Qp,ranic Studies, pp. 191-2. For a discussion of the verses dealing with widows, see Burton, The Sources of Islamic Law, pp. 57-80. Al-Tabarl, Tafsir, 3:605. The use of mu!Jkam to mean nasikh is not uncommon in early texts like those of Muqatil and pseudo-al-KalbL Wansbrough, Qp,ranic Studies, pp. 149-50. Al-Tabarl, Tafsir, 4:221. Al-Tabarl, Tafsir, 1:530. Muslim, Sa/JilJ, 17/18:119-30. See also Wansbrough, Quranic Studies, pp. 180-l. See pages 154-5. See page 155. Rippin, "Function of asbab al-nuzu~" p. 6. Al-Tabart, Tafsir, 1:103-15. See page 155.
171
The Droelopment of Exegesis in Early Islam
221 Al-Tabarl, Tafsir, 9:96. 222 Al-Tabarl, Tafsir, 12:651. 223 As Rippin points out one of the functions of asbab al-nuzill is to incorporate glosses. Rippin, 'The Function of asbiib al-nuzill," p. 7. It is possible therefore for these glosses to be extracted from their larger narrative context and appear as individual l}adzths. 224 Al-Tabari, Tafsir, 12:305. The rather unlikely link between these two verses goes back to an incident narrated in the Sira about al-Walid ibn al-Mughi:ra's attempt to slander Mul:lammad and his revelations to the pilgrims coming to Mecca. See Ibn lsl:laq, The Life of Muhammad, pp. 121-2. The citing of other quranic verses can occur in many contexts. For example, Muqatil discusses the polysemy or aspects (wujuh) of the quranic term wal}y (normally translated as "revelation" or "inspiration") by listing several passages of the Qur'an to exemplify them. Muqatil, Al-Ashbiih, pp. 168-9. See also Wansbrough, Quranic Studies, pp. 209-10. 225 Al-Tabari:, Tafsir, 4:8. 226 Al-Tabari:, Tafsir, 3:160. 227 The need for consistency is paramount. Even if the exegetical devices outlined above do not exactly match those of Wansbrough, as long as the l}adzths are categorized consistently, the various stylistic profiles can be constructed and compared.
172
Chapter 5
Data and Analysis: The Authenticity of Ibn cAbbas's lfadiths in AI-Tabari's Tafsir
ince individual exegetes would have used particular exegetical devices, the comparison of the frequency of use of these various devices in }Jadzths with the isniids attached to those }Jadzths, should make it possible to ascertain whether the isniids contain reliable information about the earliest generations of quranic exegetes and tradents. In order to perform this comparison, a sample set of }Jadfths has been selected from al-Tabarr's Tafszr on the basis of their isniids, each of which met two conditions. First, each isniid had to contain the name of Ibn cAbbas. That is to say, Ibn cAbbas is either the exegetical authority or one of the transmitters of a }Jadzth from Mul).ammad or some other authority, such as a more senior Companion. There are }Jadzths in the Tafszr in which Ibn cAbbas is mentioned, but in which he serves as neither transmitter nor exegete. These }Jadzths were not included. Second, each isniid had to contain one of the following students oflbn cAbbas: Sacld ibnJubayr, clkrima, or Mujahid ibnJabr; and one of the following informants ofal-Tabarl: Abu Kurayb ibn al-cAla", Sufyan ibn Waklc, Mul).ammad ibn I:Iumayd, Mul).ammad ibn Bashshar, Mul).ammad ibn al-Muthanna, al-Muthanna ibn Ibrahim, Yacqub ibn Ibrahim, or al-Qasim ibn al-I:Iasan. These students and informants were selected precisely because of the high frequency with which each of the students appears along with each of the informants in isniids that include the name of Ibn cAb bas. Each of the matns in the sample set was then classified according to the twelve exegetical devices listed by Wansbrough, plus three additional devices not included in his list, for a total of fifteen: simple lexical identification or paraphrase, anecdote, prophetic tradition, identification, circumstances of revelation, abrogation, analogy, periphrasis, grammatical explanation, lexical explanation, poetic loci probantes, varie lectiones, quranic loci, rhetorical analysis, and non-exegetical comment. It is possible, indeed quite common, for a single }Jadzth to employs several of these devices simultaneously.
S
173
The Development of Exegesis in Early Islam
ANALYSIS 1: IBN cABBAs VERSUS HIS STUDENTS AND THE INFORMANTS OF AL-TABARl The first phase of analysis 1 requires the number of occurrences of each exegetical device in the IJ,adfths for Ibn cAbbas (that is, for all the IJ,adfths in the sample) to be counted. Their occurrences also need to be counted separately in the IJ,adfths transmitted by each of his students and by each of al-Taban's informants. Table 5.1 contains the number of times each of the fifteen exegetical devices is used by Ibn cAb bas and by each of his students. The first column contains the results for all 997 IJ,adzths of the main sample. Since Ibn cAbbas appears in all of these IJ,adzths, these are the totals for Ibn cAbbas. That is, the first column lists the number of times Ibn cAbbas is cited providing a simple gloss or paraphrase, anecdote, and so forth. The other four columns of Table 5.1 contain the summations for the devices used by each of the students of Ibn cAb bas. More precisely, the summations for the main sample, as found in column 1, are separated according to the students oflbn cAb bas: Ibn Jubayr, clkrima, and Mujahid. The fifth column contains data for the }J,adfths with Ibn Jubayr aw clkrima-isnad, which, Table 5.1: Number of occurrences of exegetical devices in lfadzths from Ibn cAb bas and his students Students of Ibn cAbbiis CAbd Allah Sa"id ibn ibn cAbbiis Jubayr simple gloss or paraphrase anecdote prophetic tradition identification circumstances of revelation abrogation analogy periphrasis grammatical explanation lexical explanation poetic loci probantes variae lectiones
quranic loci rhetorical analysis non-exegetical total number of lfadzths
clkrima
IbnJubayr aw~a
Mujlihid ibnJabr
317
134
82
41
60
395 156 448 146
183 57 176 39
91 38 113 32
87 53 108 68
34 8 51 7
14 3 4 4
8 3 2 2
4 0 1 1
0 0 1 0
2 0 0 1
34 9 18 52 0 10
13 1 11 30 0 7
16 4 4 14 0 1
0 1 0 3 0 0
5 3 3 5 0 2
997
429
264
146
128
174
Data and Analysis
because of the uncertainty of the "aw" ("or"), can be treated neither as IJ,adzths with Ibn Jubayr - Ibn cAbbas-isnads nor as ones with clkrima - Ibn cAbbas-isnads. Note that the numbers for any row (that is, for any exegetical device) from the last four columns will add up to the number in the same row in the first column (that is, the number for the main sample). The numbers in the columns do not add up to the total number of IJ,adzths given at the bottom of each column because many IJ,adzths use several exegetical devices in tandem. Table 5.2 is similarly designed. It contains the number of times each of the fifteen exegetical devices is used Ibn cAbbas and by each of al-Tabarl's informants. The first column contains the results for Ibn cAbbas,just as in Table 5.1. In the last eleven columns (three to thirteen) the summations for the main sample, as found in column 1, are separated according to the informants of al-Tabari: Abu Kurayb, Sufyan, Ibn I:Iumayd, Ibn Bashshar, Ibn al-Muthanna, al-Muthanna, Yacqub, and al-Qasim. The fourth, sixth, and eighth columns contain data for IJ,adzths with the Ibn Jubayr aw clkrima-isnad ending with Abu Kurayb, Sufyan, and Ibn I:Iumayd, respectively. Unfortunately, the IJ,adzths with the SufYan - ... - Ibn Jubayr aw clkrima-isnads will be of limited statistical value, since there are only two such IJ,ad'iths. And once again, the numbers for any row (that is, for any exegetical device) from the last eleven columns add up to the number in the same row in the first column (that is, the number for the main sample). Tables 5.1 and 5.2 do show that there are some trends. For instance, most of the IJ,adzths use devices such as simple gloss or paraphrase, anecdote, and identification very frequently and other devices much less frequently. However, it is difficult to discern how significant these trends are with the data in their present form. They need to be normalized so that one can readily compare the columns with each other (that is, Ibn cAbbas with the students and with the informants) and the rows (that is, the devices) with each other. I will normalize the numbers by expressing them in terms of percentages. In other words, instead of presenting the number of times a particular exegetical device is employed by an exegete or transmitter, I will provide the percentage of its use. This is a simple matter of dividing each number in Tables 5.1 and 5.2 by the number at the bottom of its respective column, the total number of IJ,adzths examined, and multiplying by 100. The results of this process for Table 5.1 are given in Table 5.3 and for Table 5.2 in Table 5.4.
Analysis of Individual Devices The main concern of this first analysis is to compare the use of exegetical devices by Ibn cAbbas, his students, and al-Tabari's informants. Therefore, I shall compare Ibn cAb bas and these various transmitters in terms of their 175
total numher of hadiths
quranic lvci rhetorical analr~is nun-exegetical
variaP I.Frtinnf.l
simple glo~ or paraphrase anccrlote prophetic tradition iden t.ification circumstance ol revelation abrogation analogy periphrasis grammatical explanation lexical explanation poetic uxi prolxwles
997
0 10
[l~
34 9 18
I
317 395 1!'>G '"'8 146 11 3 4
Ibn
76 34 0 0 1 0 0 0 0 1 0 0
~-'
I
3 0 0
I
24 13 I 20 3 0 1 0 0 3 2
Yacqiib
,IH
2 1 5 0 0
~
18 19 3 16 3 2 0 0 0
ai-Qasim
Data and Analysis Table 5.3: Percentage of }J,adzths employing the exegetical devices in the IJ,adzths from Ibn cAbbas and his students Students of Ibn cAbblis CAbdAllah Sa"id ibn clkrima IbnJubayr Mujahid ibn cAbblis Jubayr aw~a ibnJabr simple gloss or paraphrase anecdote prophetic tradition identification circumstances of revelation abrogation analogy periphrasis grammatical explanation lexical explanation poetic loci probantes variae lectiones quranic loci rhetorical analysis non-exegetical
31.8 39.6 15.7 44.9 14.7 1.4 0.3 0.4 0.4 3.4 0.9 1.8 5.2 0.0 1.0
31.3 42.7 13.3 41.0 9.1 1.9 0.7 0.5 0.5 3.0 0.2 2.6 7.0 0.0 1.6
31.1 34.5 14.4 42.8 12.1 1.5 0.0 0.4 0.4 6.1 1.5 1.5 5.3 0.0 0.4
28.1 59.6 36.3 74.0 46.6 0.0 0.0 0.7 0.0 0.0 0.7 0.0 2.1 0.0 0.0
46.9 26.6 6.3 39.8 5.5 1.6 0.0 0.0 0.8 3.9 2.3 2.3 3.9 0.0 1.6
use of each of these devices individually. For the sake of convenience, the data in Tables 5.3 and 5.4 are reiterated in a series of smaller tables, one for each exegetical device. (1) Simple gloss or paraphrase: This device is employed in 31.8 percent of the l}adzths of the main sample. As is evident in Table 5.5, there is little difference between this value and those for Ibn Jubayr, clkrima, and Ibn Jubayr aw clkrima. With percentages of 31.3, 31.1, and 28.1, respectively, the difference is less than 1.0 percent for the first two and less than 4.0 percent for the latter. However, the differences between Mujahid and the main sample and between him and the others are significant. At 46.9 percent, the device of simple gloss or paraphrase occurs in 15.0 percent more of his l}adzths than in the l}adzths of the others. For the informants of al-Tabari, the variation is even greater, from a low of 10.3 percent for the l}adzths with the Abu Kurayb - ... - Ibn Jubayr aw clkrima-isnad to a high of 44.4 percent for the l}adzths transmitted by Yacqub. 1 However, both of these percentages are somewhat anomalous. The other eight percentages are much closer, ranging from 27.4 percent to 39.4 percent. Abu Kurayb and Sufyan seem to use this device almost equally (27.4 percent and 27.7 percent). The same is true for Ibn Bashshar, Ibn al-Muthanna, and al-Qasim (37.1 percent, 37.5, and 37.5 percent). These three informants, along with Ibn f.lumayd (39.4 percent) and particularly Yacqub (44.4 percent), transmit l}adzths using simple gloss 177
simple gloss or paraphrase anecdote prophetic tradition ideo tification circumstance of revelation abrogation analogy periphrasis grammatical explanation lexical explanation poetic loci pmhantes variae lectirme; quranie wei rhetorical analysis non-exegetical
31.8 39.6 15.7 44.9 14.7 1.4 0.3 0.4 0.4 3.4 0.9 1.8 5.2 0.0 1.0
Ibn CAhbiis
6.2 0.0 1.7
l.l
27.4 47.5 27.0 41.3 19.7 2.8 0.6 0.0 0.0 3.9 0.0
10.3 87.2 69.2 79.5 84.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.6 0.0 0.0
aw
Kurayb
Abu
27.7 40.6 12.9 45.2 10.3 0.0 0.6 0.6 1.3 3.2 0.0 3.2 3.2 0.0 1.3
50.0 50.0 50.0 50.0 50.0 0.0 0.0 0.0 0.0 0.0 50.0 0.0 50.0 0.0 0.0
ibn Walu""" aw
Sufyan
39.4 35.2 7.7 33.1 4.2 0.0 0.0 1.4 0.0 4.2 1.4 0.7 7.0 0.0 1.4
34.3 49.5 23.8 72.4 32.4 0.0 0.0 1.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0
Ibn l;lumayd aw
37.1 24.7 4.1 48.4 5.2 2.1 0.0 0.0 0.0 7.2 1.0 1.0 4.1 0.0 1.0
37.3 37.5 8.8 30.0 5.0 2.5 0.0 0.0 1.3 5.0 1.3 3.8 8.8 0.0 2.5
30.3 36.4 13.6 63.6 9.1 4.5 0.0 0.0 1.5 0.0 0.0 1.5 6.1 0.0 0.0
alIbn alIbn Bashshar Mut:banna Mut:banna
Informants of al-Tabari
Table 5.4: Percentage of /JadUhs employing the exegetical dniees in the !Jadiths from the informants of al-Tabari
44.4 24.1 1.9 37.0 5.6 0.0 1.9 0.0 0.0 5.6 3.7 7.4 5.6 0.0 0.0
-~-·
Yacqub
·~·-······
37.5 39.6 6.3 33.3 6.3 4.2 0.0 0.0 0.0 4.2 4.2 2.1 10.4 0.0 20.0
·-.-·~
al-Qiisim
-~---
Data and Analysis
or paraphrase significantly more frequently than those for Ibn cAbbas (31.8 percent). Only al-Muthanna is close to this value (30.3 percent). Also of note is the 10.3 percent of the !Jadzths with the Abu Kurayb- ... -Ibn Jubayr aw clkrima-isnads. This seems to indicate a significantly different approach to exegesis in these !Jadzths. Oddly, the percentage for the !Jadzths with the Ibn l;lumayd - . . . - Ibn Jubayr aw clkrima-isnads (34.3 percent) is reasonably close to that of Ibn cAbbas. These figures for the use of this exegetical device seem to suggest that there is considerable consistency between Ibn cAbbas and his students (with the very notable exception of Mujahid) and much greater variation between Ibn cAbbas and the informants of al-Tabari. It is premature to draw any general conclusions on the basis of this observation alone. (2) Anecdote: Anecdote is the second most common exegetical device used in the sample (39.6 percent) and there is significant variation in its use according to Table 5.6. Ibn Jubayr is only slightly higher (at 42.7 percent) than Ibn cAbbas and clkrima slightly lower (34.5 percent). However the Ibn Jubayr aw clkrima !Jadzths use anecdote much more frequently (59.6 percent), while only Mujahid's !Jadzths use anecdote much less frequently (26.6 percent). Variation is also evident with the informants. Some informants, such as al-Qasim, Sufyan, Ibn al-Muthanna, and perhaps also al-Muthanna and Ibn l;lumayd (39.6 percent, 40.6 percent, 37.5 percent, 36.4 percent and 35.2 percent respectively) are reasonably close to the 39.6 percent of Ibn cAbbas. However, Ibn Bashshar and Yacqub are much lower (24.7 percent and 24.1 percent). The most dramatic variation is in the !Jadzths with the Abu Kurayb - ... - Ibn Jubayr aw clkrima-isnads and the Ibn l;lumayd - ... - Ibn Jubayr aw clkrimaisnads. The former !Jadtths use anecdote a remarkable 87.2 percent of the time and the latter 49.5 percent. This deviation is significant enough to suggest that at least the IJadzths with the Abu Kurayb - ... - Ibn Jubayr aw clkrima-isnads are somehow very different than the others. This use of the anecdote suggests a mixed conclusion. Some the students of Ibn cAbbas and informants of al-Tabari are consistent with the Ibn cAb bas of the whole sample, but others are significantly inconsistent. (3) Prophetic tradition: The use of prophetic tradition is significantly less than that of anecdote (15.7 percent overall), but the pattern of use is quite similar. From Table 5.7 it is evident that both lbnJubayr and clkrima use prophetic tradition as frequently as Ibn cAbbas (13.3 percent and 14.4 percent). However, Ibn Jubayr aw clkrima is again significantly higher (36.3 percent) and Mujahid significantly lower (6.3 percent). Of the informants only Sufyan and al-Muthanna (12.9 percent and 13.6 percent) can be said to be similar to Ibn cAbbas. And once again, Ibn Bashshar and Yacqub have much lower values (4.6 percent and 1.9 percent) than the overall value. Ibn l;lumayd, Ibn al-Muthanna, and al-Qasim also have lower values (7.7 percent, 8.8 percent, and 6.3 percent). Both sets of the IbnJubayr 179
~1.3
Ibn Jubayr
31.1
"'krim.a
lbnJ.
28.1
aw "lk.
46.9
Mujiihid 27.4
42.7
39.8
34.5