PAR 246: Introduction to Islam
Sanguine Scholars: Hadith Criticism

from The Development of Exegesis in Early Islam
Richmond: Curzon Press, 2000.
by Herbert Berg

 The assumption of most Muslim scholars has been that the hadith material, at least that contained in the classical canonical collections, is authentic. Canonical status is conferred upon al-Jami` al-sahih of Abu `Abd Allah Muhammad ibn Isma`il al-Bukhari (d. 256/870) and al-Jami` al-sahih of Abu al-Husayn Muslim ibn al-Hajjaj (d. 261/875), and, to a lesser degree, upon the Kitab al-sunan of Sulayman ibn al-Ash`ath Abu Da'ud al-Sijistani (d. 275/889), al-Jami` al-sahih of Abu `Isa Muhammad al-Tirmidhi (d. 279/892-3), the Kitab al-sunan of Ahmad ibn Shu`ayb al-Nasa'i (d. 303/915), and the Kitab al-sunan of Abu `Abd Allah Muhammad ibn Yazid al-Raba`i al-Qazwini ibn Majah (d. 273/887). To these six collections are occasionally added other works, most notably the Musnad of Ahmad ibn Hanbal (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 Muhammad. 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

For Muslims, transmitting the words and deeds of Muhammad is as old as Islam itself. The Qur'an orders Muslims to follow the example of the Messenger and so from the very beginning the Companions (sahaba) concerned themselves with following the sunna (conduct or custom) of the Prophet, which was embodied in hadiths (reports or anecdotes) narrating his words and deeds. Muhammad is thought to have taken some pains to ensure the use and dissemination of his sunna.

        Generally, the Umayyad caliph `Umar II (d. 101/720) is credited with having ordered the first collection of hadith material in an official manner, fearing that some of it might be lost. Abu Bakr ibn Muhammad ibn Hazm (d. 120/737) and Muhammad ibn Muslim ibn Shihab al-Zuhri (d. 124/742), known simply as Ibn Shihab or al-Zuhri, are among those who compiled hadiths at `Umar II’s behest. This delay, nearly a century, in having the hadiths recorded resulted from reservations expressed by Muhammad 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. `Umar I is the primary locus for many accounts about hadith 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 hadiths initiated by `Umar 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 Muhammad ibn Idris al-Shafi`i (d. 204/820).

        The actual text of the hadith is known as the matn. For a matn to be recognized as an authentic record of one of Muhammad’s acts or sayings, it needs to have attached to it the list of the people who were transmitters (muhaddiths) of the matn. This isnad, 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 muhaddith related the matn, and so forth until the hadith was recorded.

        The isnad portion of the hadith was an early standard practice as well, according to Muslim interpretation of the sources. To Muhammad 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-bida`, their traditions were not accepted.

For most Muslim scholars, this fitna is the one following the assassination of the third caliph, `Uthman (d. 35/656). And so, the regular use of isnads for hadiths is thought to have begun shortly after 35 A.H. (656 C.E.). This date then also marks the beginning of hadith study as a science in the Muslim community.

        The implication of Ibn Sirin’s statement is that well-meaning but misguided or even unscrupulous people fabricated or altered hadiths for political, dogmatic, or personal reasons. Muslims freely admit this. But, according to traditional accounts, these vast numbers of obviously false and doubtful hadiths were eliminated in the painstaking process of producing the classical collections. In the third/ninth century, the sifting out of these spurious hadiths focussed largely on the isnad. That is to say, the compilers systematically examined each of the transmitters of every hadith (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 (`ilm al-rijal), the cohesion (ittisal) of the isnad was examined. The continuity of the isnad was evaluated for missing or unknown muhaddiths or for not reaching back to Muhammad and stopping at a Companion or Successor. In addition, the number of simultaneous transmitters was tallied. A hadith 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 (fard) was considered binding but with less weight. On these three bases, a particular hadith would be classified as sahih (sound or authentic), hasan (good), da`if (weak), or saqim (spurious).

            So Muslims have never suggested that forgery of hadiths 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 Muhammad. Moreover, since criticism of hadiths in the Muslim world has focussed on their isnads, dating of a particular hadith is done by ascription. That is, a prophetic hadith came into circulation during the life of Muhammad, 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 hadiths or perhaps the hadiths of a particular transmitter, employs the same methods and biographical (or rijal) materials.

Early Western Scepticism

Many Western scholars have accepted, with some reservations, these assumptions and used hadiths as fairly reliable historical sources, but for many others the authenticity and the date of origin of the hadith material are issues that have produced, and continue to produce, heated debate. In 1848 Gustav Weil, after noting that al-Bukhari deemed only 4,000 of his original 600,000 hadiths to be authentic, suggests that a European critic is further required to reject without hesitation at least half of these 4,000. He was soon followed by Aloys Sprenger, who also suggests that many of the hadiths cannot be considered authentic. However, that there is a debate about the authenticity of hadith material in the West is largely due to the innovative theories of Ignaz Goldziher. The subsequent direction this debate has taken, a direction which has focussed on the role of hadiths 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 hadiths before Goldziher, it was he who in the second volume of his Muhammedanische Studien first clearly articulated this scepticism. Familiarity with the vast number of hadiths in the canonical collections induced "sceptical caution rather than optimistic trust." Goldziher concluded that these hadiths 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."

        Goldziher’s suspicions about the authenticity of hadiths 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. Moreover, the ubiquitous contradictory traditions, the apparent proliferation of hadiths in later collections not attested to in earlier ones, and the fact that younger Companions of Muhammad seem to have known more about him (that is, they transmitted more hadiths) than the older Companions who presumably knew the Prophet for a greater length of time, suggested to Goldizher that large-scale fabrication of hadiths took place.

        As a result, Goldziher provides a significantly different version of the origin and development of hadith 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 sahifas. In this way he remains very close to the Muslim interpretation of the development of hadith literature. He not only presumes that the Companions tried to preserve the sayings and judgments of Muhammad, but also that some of them likely did so in written form (that is, in sahifas). And, when these Companions passed on what they had heard and recorded to the next generation of Muslims, the use of the isnad began. But for Goldziher, the invention of and interpolation into hadiths also began very early, for both political and paraenetic reasons. And so mutually exclusive hadiths 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 isnads."

        With the rise of the `Abbasids the situation changed significantly, according to Goldziher. `Abbasid rule was more theocratic than the more secular "Arab paganism" of the Umayyads. Consequently, the new dynasty encouraged the development of the shari`a 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 `Abbasid revolution. The report that the Umayyad caliph `Umar II commissioned the first collection of hadiths must be dismissed as untrustworthy because of the number of contradictions in the account and the absence of references to Abu Bakr ibn Hazm’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."

        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 Muhammad himself. Scholars sought to fill the gaps left by the Qur'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'an) must be referred back to the Prophet himself, that is, must be rooted in hadiths. These supporters of hadiths (ahl al-hadith) were extremely successful in establishing hadiths 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 hadiths or adapt existing hadiths in their support. Even the advocates of ra'y were eventually persuaded or cajoled into accepting the authority of hadiths and so they too "found" hadiths which substantiated their doctrines that had hitherto been based upon the opinions of their schools’ founders and teachers. The insistence of the advocates of hadiths 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." The talab journeys which followed, during which the travellers sought to collect hadiths from the various centres of the Islamic empire, helped construct a more uniform corpus of extant hadiths out of the various disparate local collections.

        Eventually, however there were reactions to this widespread fabrication of hadiths. Goldziher traces three such reactions to this phenomenon. Ironically, fabricated hadiths began to circulate in which Muhammad is made to condemn those who would fabricate hadiths about him. Others simply rejected the whole corpus of hadiths 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 hadith. This method focussed not on the actual contents of the hadith (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, hadiths, which originally had isnads ending with Companions or Successors, were often extended back to the Prophet. That is ahadith mawqufa were transformed into ahadith marfu`a by tacking on the Prophet and any other necessary names to the end of the isnad. Isnads were also "tampered" with by the mu`ammarin—the long-lived ones. For Goldziher these were persons who pretended to have had direct contact with Muhammad even though this might mean that they would have to be well over a hundred years old (and at times hundreds of years old).

        As stated earlier, Goldziher questions the traditional date at which the formal collection of hadiths began. It was not in the time of `Umar II, but with the Muwatta' of Malik ibn Anas (d. 179/795) that the process started. That is, Goldziher believes hadith gathering began only towards the end of the second century A.H. (late eighth or early ninth C.E.) with fiqh works being the precursors to proper hadith works. These latter works came soon after, as a more a systematic arrangement of the hadith material became necessary. As the insistence that legal and religious practice be rooted in hadiths had grown, so too had the available material. This arrangement took two forms: the musnad (arranged according to the isnads) and the musannaf (arranged according to topic). The musannafs came to predominate, but the musnads continued to be compiled. An example of a musnad is the compilation of Ahmad ibn Hanbal. The first musannaf that gained prevalence was the compilation of al-Bukhari. It, unlike the Muwatta', is a work of hadiths with al-Bukhari’s contribution being one of selection and arrangement. After 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 hadith literature in the sense that, instead of being compilers of new material, hadith scholars became copyists and editors producing mukhtasars, or "abridged versions."

        In summary, Goldziher sees in hadiths "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." Likewise,

Every stream and counter-stream of thought in Islam has found its expression in the form of a hadith, 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 ra'y or hawa, every sunna and bid`a has sought and found expression in the form of hadith.

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 hadith material." Goldziher, for all his scepticism, accepted that the practice of preserving hadiths was authentic and that some hadiths were likely to be authentic. 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 hadith 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 hadiths induces sceptical caution rather than optimistic trust regarding the material brought together in the carefully compiled collections.

And so it is in his advocacy of scepticism that Goldziher made his great impact on the course of hadith studies in the West.

        Goldziher never went much beyond this simple scepticism about the authenticity of the bulk of the hadith material to advance a more practical theory for determining the chronology and provenance of any specific hadith. He limited his dating of hadiths 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 Hadiths

Of the next generation of scholars, Joseph Schacht most prominently carried on Goldziher’s tradition of scepticism and his Origins of Muhammadan Jurisprudence in turn serves, along side Goldziher’s Muslim Studies, as the foundation or at least a point of departure for almost all other studies on hadiths in the West.

        In Origins Schacht’s main concern is the origin of Islamic law, the shari`a, and particularly the role of al-Shafi`i 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 Muhammad as opposed to the ‘living tradition’ of the Muslim community which might or might not claim to have such a direct connection to Muhammad. In so doing, Schacht discusses the process of development of hadith material (and hence its authenticity and chronology).

        Schacht asserts that hadiths, particularly from Muhammad, did not form, together with the Qur'an, the original bases of Islamic law and jurisprudence as is traditionally assumed. Rather, hadiths 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." And this ideal practice was embodied in various forms, but certainly not exclusively in the hadiths from the Prophet. Schacht argues that it was not until al-Shafi`i that ‘sunna’ was exclusively identified with the contents of hadiths from the Prophet to which he gave, not for the first time, but for the first time consistently, overriding authority. Al-Shafi`i argued that even a single, isolated hadith going back to Muhammad, assuming its isnad 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 Shafi`i reference to traditions from Companions and Successors was the rule, to traditions from the Prophet himself the exception, and it was left to Shafi`i 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.

        Based on these conclusions, Schacht offers the following schema of the growth of legal hadiths. 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, hadiths with isnads extending back to Muhammad came into circulation by traditionists towards the middle of the second century. Finally, the efforts of al-Shafi`i and other traditionists secured for these hadiths from the Prophet supreme authority. 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 hadiths preserved in the classical collections originated both during and after al-Shafi`i’s time. That is, most Prophetic hadiths in the collections of Bukhari, Muslim, and the others originated, not with Muhammad, but circa the middle of the second century A.H., while hadiths 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." Schacht therefore dismisses Muslim scholarship on hadiths, which itself is based on the study and criticism of isnads as "irrelevant for the purpose of historical analysis."

        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 hadiths which he develops. His unique contribution lies in his alternative to the "irrelevant" methods of Muslims; he suggests that the date of a hadith 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 isnad.

        What is meant by ‘its first appearance in the legal discussion’ is obvious. If a particular hadith is adduced in one text but is not to be found in an earlier text in which that same hadith would have been of crucial importance, then it is safe to assume that the hadith 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 hadiths 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 hadith or set of hadiths would then provoke the supporters or practitioners of the attacked doctrine or practice to manufacture hadiths to both defend it and to undermine the refuting hadiths. Their opponents would then respond with more and usually more elaborate hadiths. Thus, by juxtaposing various parallel or related hadiths and comparing their matns, one may be able to reconstruct the chronology of the hadiths surrounding a particular controversy. The doctrine or practice being attacked is, of course, chronologically prior to the hadith countering it. A hadith defending the practice or doctrine is likely to be after the counter-hadith.

        ‘Indications in the text’ means looking at the authority cited in a hadith. 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:

Whenever we find, as frequently happens, alleged opinions 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.

        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." Thus, Schacht sees the isnads as the most arbitrary part of the hadiths, but because their fabrication and development follows certain patterns, they nevertheless allow the hadiths 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 hadith and in this way obviated the charge that the hadith was ‘isolated.’ Thus mutawatir hadiths have no more claim to authenticity than do other hadiths. 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.

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." He also dismisses the claim to genuineness of the hadiths 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."

        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 hadiths 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 `Amr ibn Abi `Amr 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 was passed on to others, represented in the diagram by the names below `Amr’s, the "real part of the isnad" would branch out into several strands. The isnad would not terminate with the one who put the matn in circulation, for he would have fabricated an isnad reaching back to an authority such as a Companion or the Prophet. This is represented in the diagram by the names above `Amr’s and is for Schacht the "fictitious part of the isnad." 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 `Amr’s, and so fix a terminus a quo for the appearance of the hadith.

Diagram 1

Prophet  Prophet Prophet
5 5 5
Jabir Jabir Jabir
5 5 5
man of the Banu Salama Muttalib  Muttalib
9444444444444444444447  5 644444444444444444448
  `Amr ibn Abi `Amr  
644444444444444444448 5  94444444444444444447
`Abd al-`Aziz ibn Muhammad Ibrahim ibn Muhammad

Sulayman ibn Bilal

5   5   5
Shafi`i  Shafi`i  anonymous

        Schacht’s examination of the development of isnads is also premised on his hypothesis that legal hadiths 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. This concurs with the statement, attributed to Ibn Sirin, that interest in isnads 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 110 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.

E. Stetter and Topoi and Schemata in Hadiths

In his 1965 dissertation Eckart Stetter examines the topoi and schemata in hadiths using a representative segment from Bukhari’s Sahih. By a topos, Stetter means a narrative cliché which provides circumstantial details, such as the exact place and/or time Muhammad is to have said or done something, that imply intimate, personal contact. For example, hadiths often mention that the first transmitter heard Muhammad 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" detail.

        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.

        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 Muhammad, would certainly be necessary for any fabricator of hadiths, 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 hadith literature in a manner independent of, though not uncongenial to, the doubts raised by Goldziher and Schacht.

        Goldziher introduced scepticism about hadiths. Schacht and Stetter suggest plausible mechanisms for the creation of false hadiths. 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 hadiths has led then to (false) conclusions. And so, their arguments seem contrived, circular, and contrary to reason to those who disagree with them.

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 hadiths. Their work has been attacked on two fronts. The first attempts to correct the aforementioned "misunderstandings" by insisting that hadiths were committed to written form very shortly after the death of Muhammad, or even during his lifetime. And as a further guarantee of their reliable transmission, hadiths were then maintained in written form until they were finally compiled in the classical collections. The second consists of asserting the early and reliable use of the isnad to counter Schacht, for whom its presence, particularly in complete form, is already an indication that a hadith 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 hadiths in Islam. By "early" she means that the Companions of the Prophet themselves kept written records of hadiths and by "continuous" that most hadiths 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 hadiths 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 Muhammad were already being written during his lifetime. The problem for Abbott, given this suggestion, is the obvious lack of any early attempt to standardize all these reports about Muhammad 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, `Umar I (d. 23/644). Because of the 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 could compete with, if not distort or challenge the Qur'an." So he destroyed the manuscripts of hadiths he discovered and punished those who had possessed them. Many Companions avoided (at least publicly) the use of written and even oral hadiths 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 hadiths was the relatively few Companions, such as `Abd Allah ibn `Amr ibn al-`Šs (d. 65/684), Abu Hurayra (d. 58/678), Ibn `Abbas (d. 67–8/686–8), and Anas ibn Malik (d. 94/712), who continued to collect, record, and transmit them.

        With the death of `Umar and the successful promulgation of the `Uthmanic recension of the Qur'an, the two major fears regarding the use of hadiths were significantly diminished. According to Abbott, the use of hadiths then flourished in the second half of the first century and even those early Muslims, who like `Umar, eschewed the use of hadiths in written form, succumbed to preserving their knowledge thus. Hadiths 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.

        In an attempt to counter Goldziher’s suggestion of the secular nature of Umayyad rule, Abbott argues that the Umayyad caliphs Mu`awiya (d. 60/680), Marwan (d. 65/684) and `Abd al-Malik (d. 86/705), for example, all took an active interest in transmitting and/or recording hadiths. `Umar II is particularly associated with hadith literature. Abbott accepts the report (found in the recension of Shaybani (d. 189/805) of Malik ibn Anas’s Muwatta' that this Umayyad caliph commissioned Abu Bakr ibn Muhammad ibn `Amr ibn Hazm (d. 120/738) to record hadiths and sunna. Abbott argues that he was only one of many the caliph contacted in order to secure authentic hadiths, and that Ibn Shihab al-Zuhri was ordered to collate these numerous hadiths from the various regions of the empire. Abbott further assumes that al-Zuhri finished this enormous task and that these daftars (manuscripts) were distributed, but that because of resistance in the provinces and the untimely death of `Umar II, they never received much attention (though the work of al-Zuhri lived 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 hadith material.

        With this form of transmission of hadiths, Abbott is also able to provide the following explanation for the appearance of a rapid expansion in the number of hadiths—perhaps to counter Schacht’s spread-of-isnads theory. Manuscripts, particularly those preserved by succeeding generations of the same family, which were lengthy documents, were divided into separate sections and given the isnad of the original document. From one such document could come hundreds of hadiths. "If not fully comprehended, this process would give the impression of a sudden huge increase in the number of traditions . . ." Furthermore, Abbott argues that:

The development of the family isnad and continuous written transmission lead to the . . . inescapable conclusion . . . that the bulk of the hadith[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.

That is, she not only accepts the bulk of family isnads as genuine (unlike Schacht), but also credits them for guaranteeing the authenticity of hadiths in general. And, these parallel oral and written transmissions each served to safeguard the other and so prevented the large-scale fabrication of hadiths. 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 rihlas (the journeys in search of knowledge and usually associated with oral tradition), in the use of the warraqun (stationer-copyists), and in the average memory of the average traditionist evidence for the continued use and production of manuscripts of hadiths. In fact, the oral transmission has been overemphasized according to her because Western scholars have generally failed to grasp hadith semantics properly. Arabic terminology for writing materials and in isnads has also been misunderstood. An example of the former is the word sahifa. 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 haddatha (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.

        However, because of her resolute defence of the authenticity of hadiths, 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 hadiths than there were sound ones. Does that not imply fabrication of hadiths 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 hadiths (counting any variation of matn or isnad as a separate hadith) increased geometrically until there were literally hundreds of thousands of hadiths during the time of Bukhari and Muslim. And since criticism of hadiths was directed at their isnads, the high number of unsound hadiths merely meant that the isnads were less than perfect while the matns may well have been "sound." This hypothesis also allows her to conclude that fairly exhaustive collections of hadiths were possible using only a fraction of the vast material available. That does not mean for Abbott that the hadiths found in Bukhari and Muslim are entirely authentic and complete, but that they preserve a "genuine core of the sayings and deeds of Muhammad together with a genuine core of sayings and deeds of the Companions and Successors", 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 naïve to some scholars. As G. H. A. Juynboll has pointed out, "Abbott seems to rely too heavily on much of the information given in isnads and in books about isnads concerning the three oldest tabaqat." Yet it is isnads and their authenticity that lie at the heart of the debate over the authenticity of hadiths. 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'an Wissenschaften, Hadith, Geschichte, Fiqh, Dogmatik, 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 hadith 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 hadiths.

        Sezgin lists eight ways in which transmission took place: sama`, qira'a, ijaza, munawala, kitaba, i`lam al-rawi, wasiya, and wijada. Sezgin states that only the first two (listening and recitation, respectively) involved memorization. The others, and often in practice even sama` and qira'a, involved written materials. Furthermore, written transmission was as customary as oral transmission. 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.

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.

        Sezgin traces a very different history for hadith 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 hadiths indicate primarily transmission in written form (a fact, Sezgin repeats, that Goldziher misinterpreted). The first stage involved simple books (sahifas or juz's) produced by the Companions and the earliest Successors (and in this regard, Goldziher and he agree). But whereas Goldziher puts the collection of hadiths after the development of fiqh literature, Sezgin envisions the development of hadith literature to be an independent and continuous practice that began with the sahifas. In the last quarter of the first century and in the first quarter of the next, the scattered hadiths began to be collected. The first proper collection was made by al-Zuhri. And so the first musannaf dates from circa 125 A.H. (742 C.E.), thus antedating the musnads of the end of the second century A.H.—the time during which Goldziher placed the appearance of musannafs.

        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."

        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, as Juynboll 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." Sezgin is not exactly presenting theories as facts, but he proceeds from the assumption that transmission took primarily a 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, has attempted to rectify the perceived inadequacies of Western scholarship on hadith 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 hadith material in the classical collections, but also out to prove the reliability of the isnads 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 hadiths were written down even in the time of Muhammad. He then proceeds to list the hundreds of Companions, Successors, and scholars from the first 150 years of Islam who, according to him, wrote down hadiths, along with the names of their students who received hadiths from them in written form. That is, he states there has been a continuous, early written tradition, implying thereby that authenticity of the hadith material is more assured. And again, the argument (such that it is) is one which relies on ascription and treating isnads and the rijal literature as independent, but mutually corroborating sources.

        Azami’s far more original and valuable contribution to the study of hadiths comes with his defence of the isnad. Schacht contends that, while the isnad system may be authentic for hadiths whose isnads end in second-century 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 isnad system began in the early second century or perhaps the late first. He cites both Horovitz and Robson to support his claim that the use of isnads, like the use of written records, was very early. Azami adduces the report ascribed to Ibn Sirin that the use of isnads was demanded after the fitna, and for Azami, the fitna refers to the civil war between `Ali and Mu`awiya (36 A.H.). Furthermore, Azami argues that this report states that it was only after that time that isnads were "demanded," implying that they must have been in use prior to that time, albeit perhaps less stringently.

        Second, Schacht’s statement that the isnad is the most arbitrary part of a hadith is attacked by Azami. Azami finds ample evidence for rejecting this 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’." 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.

        Third, on the charge that isnads were gradually improved through fabrication and alteration, Azami admits that faulty isnads exist in hadith material, but rejects the notion that they indicate anything significant about the development of hadiths. He points out that al-Shafi`i, 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-Shafi`i’s time to obviate the criticism of a hadith 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 hadith 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 hadith 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. Azami suggests it would be more appropriate to say, "All the ‘family isnads’ are not genuine, and all the ‘family isnads’ are not spurious."

        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-Shafi`i text reveals that there is but one chain from the Prophet to `Amr, 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. Azami points out quite rightly that the Prophet-Jabir links are obviously just one link. In addition, he concludes that `Abd al-`Aziz erred in making `Amr’s authority a man of Banu Salama and so the isnad should look like that given in Diagram 2.

Diagram 2









`Abd al-`Aziz                      Ibrahim                         Sulayman

        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 `Amr need not fan out in order for `Amr to be the common link. That it fans out after him is the criterion which indicates that he is a common link. However, Azami, anticipating the counter-argument 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 hadiths, 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.

        After having addressed each of these six points, Azami is able to conclude that "[t]here is no reason to reject the isnad system. It is proved that it has every element which can command the acceptance of the system as a whole."

        And, having vindicated the isnads, Azami turns to the question of the authenticity of hadiths 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." 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 hadiths of their time. Rather snidely, Azami asks why should these scholars make references to these "later" prophetic hadiths when Schacht himself states that reference to such hadiths was the exception, not the rule.

        Azami confidently concludes, "So even if mistakes in isnads and ahadith exist, Schacht has produced no evidence that would cause us to impugn the good faith of the majority of the transmitters or abandon the hadith literature." While this may be overstating the case somewhat, Azami has 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 hadiths 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 naïve 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 hadith 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 hadith 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 hadiths. 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. The latter three do not fare as well. Juynboll says:

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.

        From the above it may seem that Juynboll is hardly seeking "middle ground." However, he believes that the early reports regarding the origins of hadith 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 naïveté, is an asset in the historian of early Islamic society rather than a shortcoming to be overcome and suppressed at all costs." In addition to this qualified credulity of his, Juynboll finds middle ground in many of his conclusions about the origins and authenticity of the hadith material and sciences, which he largely bases on the awa'il literature—anecdotes about who was the first to do something or when and where certain institutions were first established.

        According to the awa'il sources, after Muhammad’s death the first to spread stories about him (in a deliberate manner) were the storytellers (qussas), who told stories of an edifying nature. Isnads proper were not attached to these "prophetic" utterances. It is reported that Sha`bi (d. 103–10/721–8) was the first person to question someone about an authority and that Shu`ba ibn al-Hajjaj (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 Sirin alluded was the war between the Umayyads and Zubayrids. This scenario, which places the origin of the isnad around the year 70/690 (as opposed to 35/656), makes the awa'il account of the first isnad critics much more plausible.

        Juynboll outlines his tentative chronology of the growth of hadiths 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 hadiths 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 hadith[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 . . ." Juynboll further supports this conclusion through an examination of the limited use of hadiths in the initial years after Muhammad’s death and the relatively late development of the concept of prophetic sunna and hadith 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 Muhammad and his most respected Companions predates this more specific sunna. He credits `Umar II with the first organized attempt to apply the sunna of the Prophet. In the Hijaz, Egypt, Syria, and Iraq interest increased in hadiths 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 hadiths 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. Furthermore, since some of the Companions are credited with such incredibly large numbers of hadiths (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 hadith. 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 hadiths 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 hadith) 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 Muhammad 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 hadith 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 hadiths do reflect reasonably accurately Muhammad’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 of Islam did or said, or might have done or said, but surely it is unlikely that we will ever find even a moderately successful method of proving with incontrovertible certainty the historicity of the ascription of such to the prophet but in a few isolated instances.

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 hadiths in general, Juynboll spends considerable effort refining and employing Schacht’s common-link theory in order to the same for individual hadiths. 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 common-link phenomenon by suggesting that during the early stages of hadith 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 proto-isnad, clearly showing up a common-link, supporting the hadith without accretions was no longer separately discernible.

That is, the common link of many hadiths has been irrevocably obscured by the sheer number and complexity of fabricated isnads. However, in his articles, Juynboll is less tentative in his use and elaboration 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. 


Diagram 3

5 5 9444444444444447
transmitter                     6448 9447                     transmitter
9447      transmitter      transmitter           6448 9447
collection   collection    6448 9447 6448 9447  collection   collection
collection     collection    collection

Diagram 4

eyewitness    eyewitness    eyewitness    eyewitness
5        5                5
transmitter   transmitter   transmitter   transmitter
5       5       5
transmitter   transmitter   transmitter   transmitter
9447 6448 6444448

These two patterns correspond to those found in legal hadiths and historical hadiths 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`.

One would expect, especially for important hadiths, 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,

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.

It is for this reason that he feels that the authenticity of legal hadiths, which display the common-link pattern (represented in Diagram 3), are suspect, while historical hadiths (or akhbar), which display the inverted common-link 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 hadith 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.

        Juynboll observes that in the canonical hadith literature, hadiths 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]." 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.

Diagram 5

alternative Companion                                                Companion
*            +))))))))))))))))))))-*

Fulan                       Successor                                              Fulan 3
                 +))))))-.)))))),                               *
Fulan           Fulan 2                            Fulan 1                                Fulan
*         +))))))-.)))))),          *
Fulan           Fulan             Fulan                                 Fulan                 Fulan
*         *                             *                       *
Fulan           Fulan               Fulan 4                           Fulan 5                Fulan
                  *     +)))-.))),      +)))-.))),              *
Fulan           Fulan        Fulan             Fulan         Fulan            Fulan          Fulan
  *     *         *      *          *             *
 Collector 7   Collector 6   Collector 5   Collector 4   Collector 3   Collector 2   Collector 1

        In elaborating and refining Schacht’s methods and theories, Juynboll has made considerable advances in determining the chronology, provenance, and authorship of specific hadiths. He has found a way to salvage historical information from at least part of the isnad. Yet the methods developed by Juynboll allow only relatively few hadiths to be fully 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 Sunna

Fazlur Rahman stands in a somewhat unique position. On the one hand, he has accepted some of the general conclusions reached by Goldziher: hadiths are by and large not historical. On the other hand, as a Muslim, he hesitates to dismiss the hadiths in the canonical collections and the Prophet’s sunna as spurious. His theory on the origin and development of hadiths suggests that, while the isnad of a hadith 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 hadiths 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 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 making it the Sunna of the Prophet?" 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 hadiths. 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 hadiths 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-Shafi`i 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 hadiths and for his conclusions that some hadiths 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 hadith 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 hadiths, 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 Hadith literature make out to be the case.

Muhammad 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 Muhammad 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 Muhammad.

        A slightly more formal tradition developed after the death of Muhammad in that new Muslims would naturally enquire about the words and deeds of Muhammad. Isnads, which are for Rahman a sign of the formalized discipline of hadith, appeared much later, around the turn of the first century (though the informal use of hadiths began somewhere between the years 60 and 80).

        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 Muhammad. And this silent, living tradition, the tradition of what Muslims actually did, is the sunna. And so Rahman states, "that the Sunna and Hadith were coeval and consubstantial in the earliest phase after Muhammad and that both were directed towards and drew their normativity from him." However, as time passed the succeeding generations of Muslims made additions to both the sunna and the hadith 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 hadiths. At first hadiths were referred back only to Companions because they embodied the words and acts of Muhammad and because the discipline of the isnad had not yet fully developed. Rahman credits al-Shafi`i for the place achieved by hadiths 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 hadiths) led to a massive fabrication of hadiths. And according to Rahman, Muslims were largely successful in bringing the whole of the living tradition (sunna) into the sunna. And thus hadiths (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. As a result, the sunna as currently embodied in hadiths 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 hadiths are merely an attempt to give the actual practice of the 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 hadiths nevertheless remain prophetic.

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 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.

        Schoeler begins with the question of whether or not written texts of hadiths were prevalent prior to the collections of al-Bukhari and Muslim. Goldziher would say not, Sezgin yes. Schoeler argues against Goldziher’s interpretation that encomia, such as ma ra'aytu fi yadi-hi kitaban 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 isnad.) 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 Muwatta' 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." 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 Sezgin), but so too does the possibility that no original text may have ever existed or served as a prototype for the various recensions.

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.

        Motzki focusses on the Musannaf of `Abd al-Razzaq al-San`ani (d. 211/826). This work contains composite riwayat (transmissions), but ninety percent of the hadiths go back to a single transmitter, a common link as it were, which, for Motzki, implies a written text. A statistical analysis of `Abd al-Razzaq’s informants shows a divergent pattern in their informants, which Motzki states is inconsistent with the arbitrary manufacture of these hadiths by `Abd al-Razzaq. Furthermore, `Abd al-Razzaq occasionally expresses doubts about his sources and provides anonymous transmissions. This too seems unlikely if these hadiths were fabricated.

        Having shown to his satisfaction that `Abd al-Razzaq did not forge hadiths, Motzki asks the same question of `Abd al-Razzaq’s sources by focussing on a representative selection of hadiths going back to Ibn Jurayj (d. 150/767). These hadiths, 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 Jurayj forged hadiths. This is confirmed by an examination of Ibn Jurayj’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 hadiths 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 `an, "from," versus the use of sami`tu, "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 Jurayj 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.

        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, `Ata' ibn Abi Rabah (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 `Ata'), variance in the types of questions in the responsa (direct, indirect, and anonymous), and variance in the actual postitions taken by Ibn Jurayj and his teacher `Ata'. That which is described as intrinsic refers to: Ibn Jurayj’s willingness to give his own ra'y without projecting it back (that is, Ibn Jurayj feels he is under no obligation to do so); Ibn Jurayj’s commentary on `Ata'’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 `Ata' (that is, why would he bother if he were given to forgery?); his occasional uncertainly about the `Ata'’s wording; his supplying of variant traditions; and his "hints of deficiencies of `Ata'."

        Motzki goes back yet another level, but `Ata'’s hadiths primarily employ ra'y (eighty percent). This implies that he either did not rely on hadiths or that he did not know many hadiths (perhaps because during his time few were in circulation). Nevertheless, when he examines `Ata'’s transmission from Ibn `Abbas, he again finds the usual variance in what few data there are. Motzki sees in `Ata'’s infrequent use of earlier authorities, his citations of Ibn `Abbas in various ways, his willingness to contradict Ibn `Abbas, 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 hadiths from either Ibn `Abbas or `Ata' to draw many conclusions. Motzki turns instead to a specific hadith and concludes that there is ample evidence to suggest that `Abd al-Razzaq knew the hadith. He argues that since `Abd 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. Moreover, he argues that "since there is only a generation between `Ata' and Muhammad, 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."

        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 hadiths 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. However, when Motzki compares the four different lines of transmission ending in al-Zuhri (one from each of the two main recensions of the Muwatta' and two from the Musannaf), using a few specific examples where some or all of the texts contain related or parallel hadiths, he concludes that these different versions of al-Zuhri’s fiqh 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 hadiths that end in al-Zuhri.

        Motzki has argued, at least in the case of the Musannaf of `Abd al-Razzaq, that both the matns and the isnads which support them can largely be trusted. That is not to say that he does not concede that hadiths were forged. Rather, he suggests, "the mere fact that ahadith and asanid 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." However, Motzki’s comparison on the basis of isnads do seem to preclude systematic fabrication. But it is precisely the isnads 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 is need not be systematic fabrication.

J. Horovitz, J. W. Fück, 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, J. Horovitz, J. W. Fück, J. Robson, N. J. Coulson and U. Rubin certainly merit some brief attention.

        Josef Horovitz addresses the problem raised by the use of isnads (or rather the lack thereof) in the extant writings of `Urwa ibn Zubayr (d. 92–101/711–20) and by the limited use of them by Ibn Ishaq (d. 151/768). The latter historian did not provide isnads, or provided defective ones, in his biography of Muhammad. Caetani had concluded from this that the origin of the isnad must be placed between these two scholars and its perfection still much later in the third century. To this Horovitz responds by pointing out that Ibn Ishaq’s teacher, al-Zuhri, was already using composite isnads (Sammelisnad) and so argues that "since the use of the isnad in this composite form hardly allows one to imagine that the use of the simple isnad was not already customary for some time." And so for Horovitz the isnad is at least older than al-Zuhri—though how much so is less certain. As for `Urwa, who does not seem to use isnads consistently, Horovitz argues that this absence does not mean he did not use isnads (and that the isnads 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 `Abd 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 isnad in hadith literature as not later than the last third of the first century A.H. Of course, whether the close connection between historical material and other genres of hadith material (on which both sides of the argument rely) can be made, remains debatable.

        Johann Fück in discussing the role of hadiths in Islam makes an insightful observation. The Companions most frequently cited in isnads as authorities are the younger ones. For example, more hadiths cite Abu Hurayra and Ibn `Abbas than cite Abu Bakr and `Uthman. 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 Muhammad. Fück 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.

        James Robson in a series of articles began to look at the origin of the isnad. 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." Robson particularly exempts historical hadiths from this scepticism about authenticity. He offers little evidence in support of this, except the argument that it seems logical that Muhammad’s followers, because of the impression his personality must have made on them, preserved a genuine core within the hadiths.

        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 Muhammad were simply discussed amongst themselves; no demand for authority was necessary. After the death of many of the Companions, Muslims would continue to speak about Muhammad, 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 Ishaq did not feel obligated to use them. Anticipating Juynboll, 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 `Abd Allah ibn Zubayr set up the counter-caliphate 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 Ishaq’s isnads were preserved in later works, Robson concludes that the matn can be altered even if an isnad has been attached and remains unaltered, though the kernel of information remains the same. Drawing on Fück and Horovitz for support, he also insists that despite the spurious material, there remains "some genuine early material." In an incidental point, Robson criticizes Schacht’s complete rejection of family isnads as simply a method to attempt to authenticate spurious hadiths. Robson points out that the idea to use family isnads in this manner likely resulted from genuine family isnads 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 Shari`a law which is irrefutable in its broad essentials", Coulson has debated with Schacht over the dating of hadiths. Coulson asserts that Schacht "translates the negative proposition that the evidence of the Hadith does not take us back beyond the second century of Islam into the positive statement that legal development began only in late Umayyad times." That is, development in law has been too closely identified with the development of hadiths 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'an itself. Furthermore, Coulson admits that, though an isnad may be fictitious, this does not mean the substance of the matn has also been fabricated—it may well represent a precept of Muhammad. Coulson explains:

This is not to suggest that the chain of transmission, or the isnad, 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 day-to-day problems arising from the Qur'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.

While certainly not accepting the traditional Islamic view of hadiths, Coulson takes issue with Schacht’s proposition that each and every legal hadith 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". Coulson states quite bluntly that "the truth lies somewhere between traditional Islamic legal theory and the rigorous historical approach of Schacht."

        Uri Rubin stands somewhere between the sceptics and those who search for middle ground. Concurring with the sceptics, at least with regard to historical hadiths, he writes

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.

The image of Muhammad 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 Muhammad with a proper prophetic vita." 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. However, Rubin’s "systematic textual analysis" cuts at the very heart of the claims first made by Schacht and developed by others such as Juynboll, 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, Muhammad and the Companions are part of the "original hard core." In yet other traditions, the isnads of prophetic hadiths have nothing in common with the non-prophetic 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 of lack 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.

        Horovitz, Fück, 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 hadith 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.

        In their stated approach to early Islamic texts, those who search for a middle ground have been influenced by sceptical scholars such as 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, 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. Wansbrough seems to find the acceptance of "isnad apparatus" by Abbott and Sezgin to be rather naïve. 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 hadiths in general. Their interests lie in early Muslim theology and jurisprudence respectively, and so each has had to deal with questions of authenticity and the dating of hadiths, 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 hadiths 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 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 hadiths. While recognizing that van Ess’s method is clearly an orientalist one—that is, one which does not date hadiths by mere ascription—Cook questions the validity of his method. Van Ess, like Schacht, recognizes that isnads grew backwards and he accepts the common-link theory. That is, when the isnads of parallel matns seem to converge on a particular transmitter, that transmitter may serve as a terminus ante quem.

        Cook, in critiquing this method, draws and expands upon another one of Schacht’s insights: the spread of isnads—that is, the creation of additional authorities or transmitters for the same matn. This spread of isnads can occur in several ways: (1) Skipping a contemporary transmitter. Hypothetically speaking, assume that Ibn Jubayr transmitted a matn from Ibn `Abbas to Ibn Jurayj. If Ibn Jurayj is scrupulous in his transmitting, he will include Ibn Jubayr in the isnad when 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 isnad 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." So Ibn Jurayj might "skip" Ibn Jubayr, by claiming that he heard the matn directly from Ibn `Abbas (and the resulting isnads would look like those in Diagram 7).

Diagram 6


Ibn `Abbas

Ibn Jubayr

Ibn Jurayj

Diagram 7


Ibn `Abbas
Ibn Jubayr                  Ibn Jurayj

(2) Ascribing a saying to a different teacher. Hypothetically speaking once again, let us assume that `Abdallah tells Ibn Sa`id a matn he has heard from Ibn Jubayr who has it from Ibn `Abbas. Ibn Sa`id, instead of claiming to have heard it from Ibn Jubayr (resulting in the isnad shown in Diagram 8), ascribes the tradition from Ibn `Abbas to his own teacher Ibn Jurayj (resulting in the isnads shown in Diagram 9). This might be done because it is known that he never met Ibn Jubayr or that Ibn Jubayr is not recognized as an acceptable transmitter in Ibn Sa`id’s own school. Thus he appropriates the matn in a way that avoids these problems. (3) Obviating the "isolated" charge. Because a well-attested hadith carries more weight, there would be a strong motivation to "discover" other isnads.

Diagram 8


Ibn `Abbas

Ibn Jubayr

Ibn Sa`id

Diagram 9


Ibn `Abbas
Ibn Jubayr          Ibn Jurayj
`Abdallah              Ibn Sa`id

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 `Abbas that the isnads appear to fan out, and so it appears that he is the originator of the hadith. 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.

        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." 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, undermines any attempt to use isnads in this manner or in the manner of Juynboll.

        Cook is also aware of Azami’s polemic against Schacht’s scepticism about the authenticity of the hadith corpus in the course of which Azami pointed out that since some hadiths are transmitted by so many Muslims from every generation and in numerous regions in the Islamic empire, "that each of them should put about the same fabricated tradition presupposes a level of conspiratorial action which is historically quite implausible." Cook feels Azami’s argument against the massive fabrication (which is essentially the traditional mutawatir argument) of hadiths to be irrefutable unless one accepts that isnad spread did occur on a significant scale. Cook insists that "any Orientalist attempt to investigate the chronology of traditions in the manner of van Ess is simply a mistake. On this issue the choice is between Schacht and Shafi`i; there is no methodological middle ground."

        Cook is also aware of the limitation of the theory he is suggesting. To accept the notion of the spread of isnads does not necessarily provide a new method of isnad criticism: this theory destroys information, it does not provide information—a destruction Cook considers "irreversible."

N. Calder and the Common Link as the Locus of Controversy

Norman Calder is also sceptical about the value of the common-link theory and the historical information it can convey. But to understand the full import of his critique of it, it is first necessary to discuss his understanding of the process that produced written texts and the material, including the hadiths, contained therein.

        Calder’s Studies in Early Muslim Jurisprudence is primarily a close literary analysis of several major fiqh texts, followed by a more general discussion of the emergence of Muslim jurisprudence in the early centuries of Islam. In the course of evaluating the literary form of these works during this emergence, Calder suggests (as Schoeler also does) that the traditional dichotomy between oral and written transmission should not be drawn as sharply as it has.

        The juristic texts, in the language used to describe transmission (haddatha and qala, for example), attest to significant oral activity (that is, creativity and transmission). Additionally, Calder emphasizes the importance of recognizing that the written texts as they exist today not only represent just a fraction of the written corpus of the time, but also a much smaller fraction of the oral activity. The Arab Muslim milieu until the early decades of the third century A.H. was one that was productive of both oral and written literature. The former, at least in the juristic context, was often the product of a discursive process. And the latter attests to this oral activity (and it seems they also attempt to recreate it). Books remained secondary to the oral transmission of knowledge and existed at first in private notebooks. Calder explains:

the owner of a notebook controlled its contents. Interested parties wrote down their sama` from so-and-so. There is no reason to suppose that one man’s sama` would be the same as another’s from the same master. It is perfectly possible that two listeners at the same majlis would take notes on different dicta; they might record them more or less in their own words; they might even read them back to the master—these different notes—and get his approval. In the end they would be preserving and transmitting their own material, not a ‘book’ by the master. It is the transmitters’ authority or artistry that might eventually precipitate a real fixed test attributed to a named master. Between a first notebook and an achieved fixed text there might be many stages.

And so these notebooks were produced by students not teachers. They might be circulated, even destroyed, but their contents copied to other (note)books—though not without the strong likelihood of both their form and content undergoing some alteration in the process. (Eventually, a system of control was developed in which a teacher would compose and publish his own books by employing secretaries or by requiring his students to compare formally their notes with his own book.) Thus in the third century there was a shift from a predominantly oral milieu to a written one, through the mediation of these notebooks. As Calder points out,

[t]he scholar’s notebook and the institutional redaction are then the two basic types of early juristic literature. Of these former there may be not extant examples. Their existence, however, is securely inferred from the forms and sometimes from the express words of the institutional redactions, and it is confirmed by numerous references in biographical and other literature.

        It is this theory of the organic growth of oral and written juristic literature that serves as background to Calder’s attempt to demonstrate that parallel hadiths, which have a common link, are not the result of a hadith’s having been put into circulation by the common link (or some of his immediate followers), but the result of a very different scenario. Two processes make up this scenario. First, when the matn of a hadith came into circulation and attained acceptance in several (possibly rival) groups, each group would then provide the matn with an isnad which would contain transmitters reflecting their group. The common links occurred in part simply because nearly all these groups (rivals or not) recognized (that is, claimed for themselves) the Successors and the Companions. And so isnads naturally start to converge at the Successor level. Second, these groups were often in competition. That is is to say, "they engage[d] in a mutual process of isnad-criticism, which, again because they share[d] a common respect for the generation of the Companions and the Successors, tends to focus on ousting a hadith by destroying the third or fourth link."

        Modifying Calder’s own example (see Diagram 10), let us say that Group X justifies a law based on a hadith with the following isnad: C—B—A—`Ikrima—Ibn `Abbas—Prophet. Rival Group Y disagrees with the law, but also respects hadiths as a source of law. It cannot simply dismiss the hadith; it must discredit it. To do that, it can only attack the isnad. No group would dare disparage the Prophet or Ibn `Abbas (a well-respected Companion). Even `Ikrima (a Successor) would not normally be open to slander. So Group Y must attack A instead, perhaps by suggesting that A is a liar or has a faulty memory. Group X would respond to this attack on its law by strengthening the hadith by ‘discovering’ another isnad, say F—E—D—`Ikrima—Ibn `Abbas—Prophet. Group Y might then attack D, E, and/or F as it attacked A. However, it might invoke another technique: it discovers its own isnad for the very same tradition, say I—H—G—`Ikrima—Ibn `Abbas—Prophet, but with one very significant difference: one of G, H, or I is recognized by all as a notorious heretic. And hence, any hadith transmitted by him must be false or at least suspicious by association. And so Group X continues to repair the isnad and Group Y continues to attack it. Eventually Group X may simply abandon the Ibn `Abbas— `Ikrima branch of the isnad and invent a completely new one (such as L—K—J—Ibrahim—Ibn Mas`ud—Prophet in Diagram 10).

Diagram 10


Ibn `Abbas              Ibn Mas`ud
`Ikrimah                Ibrahim
5 9447                   5     
A         D         G                   J             
         5    5        5     
B         E         H                  K           
         5    5        5     
C         F          I                   L           

        According to Calder, "The to-and-fro of isnad criticism focuses on weak links which are characteristically the third or fourth links in an isnad. This leads to the common-link phenomenon, which reflects nothing whatsoever about the origins of the matn of the hadith; it reflects isnad criticism and competition in or after the second half of the third century." However, he does not think that the common-link theory is entirely devoid of value. Although the common link is not in fact the inventor or propagator of the hadith, he is simply the figure preceding the figure who became the focus of dispute in the mutual isnad criticism. At least in the field of jurisprudence, Calder states, "the dispute took place in the second half of the third century. It was during this period that all of these isnads were discovered (or invented)."

        Both Cook and Calder extend beyond the scepticism of Schacht and Goldziher. In fact, they press that scepticism to its logical conclusion: if one doubts the isnad system, then one has to doubt all the early texts which employ them to authenticate themselves. As a result, Cook points out that knowledge is "destroyed." Or stated less dramatically, things become less certain than they were. From the less sceptical perspective, Cook and Calder have not pressed anything to its logical conclusion. Their hyper-scepticism results from a need to defend an approach to this early material which is simply wrong: hyper-scepticism allows them then to dismiss any evidence contrary to their sceptical approach.

Analyses and Conclusions

Goldziher and Schacht provide primarily an approach or attitude to hadith literature. Goldziher, except for a few generalizations about isnads growing backwards, does not supply any explicit methodological tools. Schacht has provided, along with his important assertions (such as the need to presume inauthenticity and the negative correlation between the degree of isnad perfection and the chronological appearance of the matn), tools for analyzing and dating hadiths. In addition to his infamous common-link theory, which has been labelled invalid in its assumptions from two opposing camps and yet championed by Juynboll, Schacht has outlined how a hadith might be analyzed with respect to its appearance and relative position in the legal discussion and to indications within its text and isnad, so as to give a fair indication of its authorship, chronology, and provenance. In direct opposition to Goldziher and Schacht (who themselves do not fully agree) stand Abbott, Sezgin, and Azami. Their position is based largely on the assumption that isnads are historically reliable. Azami has attempted to undermine the validity of the methods suggested by Schacht, but provided no alternative, other than that already endorsed by Abbott and Sezgin. Their method for examining hadiths is simply the one of ascription.

        In between these endpoints there seems to be a continuum of views. Rahman’s theory, while very interesting, is an attempt to acknowledge the conclusions of Goldziher—that much of the hadith material is later and reflects the practice of the community—while still dismissing his claim as essentially irrelevant (at least for Muslims). Rubin too acknowledges the later influence of Muslims on purportedly early texts. Juynboll and Motzki have endeavoured in very different ways to refute, but in other ways temper, expand, and apply the methods and theories of Schacht. However, Cook and Calder, who have endeavoured to apply the sceptical stance of Goldziher, Schacht, and Wansbrough, have in the process "destroyed" the very information Juynboll and Motzki seek to create by gleaning the isnads. That is to say, the patterns in isnads from which Juynboll and Motzki seek to extract historical information about transmitters and about the chronology and provenance of hadiths, are dismissed by Cook and Calder as simply evidence of wholesale fabrication. And so, these "patterns" reveal only that no historical information can be reliably found in that particular nexus of isnads. At times, even Juynboll recognizes this possibility. In addition, Schoeler and Calder have both taken issue with the theory of an early and continuous written tradition in Islam as propounded by Abbott, Sezgin, and Azami. Both argue for an intertwined oral-written tradition. For the former, the various recensions are in some sense all authentic, though no "original" is recoverable. For the latter, the organic growth of the texts as we have received them preclude any assumption of authenticity and recoverability.

        Perhaps another way of grouping the positions of these scholars is to look at their respective stances toward the authenticity of matns and isnads. Goldziher and Schacht both place little stock in either, though Goldziher is marginally less sceptical of the possibility of authenticity of matns. They are supported in this scepticism by Cook and Calder. Again Abbott, Sezgin, and Azami are united in their support for the reliability of both matns and isnads. In a qualified manner, Motzki, Horovitz, and Fück support their basic acceptance of the material. As for Juynboll, Rahman, Robson, and Coulson, they seem willing to jettison any notion of reliability when it comes to the isnads. However, each scholar in his own way argues that the matns nevertheless contain the gist, spirit, or essence of what they claim to report. Rubin argues just the opposite: he doubts the authenticity of the matns, but not so much the isnads. For Schoeler, given the nature of the transmission of this material, it is a moot point.

        Each of these models is internally consistent, though many are mutually exclusive. Those who are most sceptical tend, like Goldziher, to assume that only the matn has historically "useful" information and the isnad is of very limited historical value. Muslim scholars and the less sceptical (suspicious?) Western scholars continue to view the isnad as historically useful. So, despite all the attempts to find middle ground, there are in reality only two positions. Cook said the choices were between al-Shafi`i and Schacht. Leaving aside the alliteration, I suggest that the dichotomy is one of Goldziher and Sezgin—with Schacht, Cook, and Calder aligned with the former; Abbott, Azami, Motzki, Horovitz, and Fück with the latter; and Juynboll, Rahman, Robson, and Coulson vacillating between them, though in terms of matns or genres of hadiths other than the legal, they are decidedly closer to the Sezgin camp. Only Rubin seems to be the odd man out.

        The question is, which of the two major positions best fits the data? Or perhaps a less biased formulation of the problem would ask, does either of the models fit the data? Both have arguments that seem consistent, at least to themselves and their like-minded colleagues. I have tried to show that is a product of shared assumptions. When assumptions produce conclusions which in turn justify the assumptions, we have a circular argument. And such an argument is not likely to convince anyone who does not share those assumptions. Another question needs to be asked: can exegetical hadiths (or for that matter, historical hadiths) be treated in the same manner as legal hadiths? Several scholars have already argued that historical hadiths are a separate genre from legal hadiths and so generalizations about the value of the information contained in the isnads of the latter do not necessarily apply to the former. Exegetical hadiths differ most obviously from other hadiths in that they have for the most part not been raised to the Prophet, but generally end with a Companion or Successor. Other scholars argue that the doubts surrounding other genres of hadiths do not apply to exegetical hadiths. If that is the case, perhaps even the approaches of Abbott, Sezgin, and Azami cannot be so readily dismissed. Chapter 3 will show that the study of early quranic exegesis is even more markedly divided between the (at times, qualified) ascriptionists and the (always, unqualified) sceptics.

Back to Schedule

Last modified on December 22, 2002 by Herbert Berg