VIII ON THE AUTHORITATIVENESS OF SUNNI CONSENSUS

I

Sunni Islam recognizes four sources from and through which the laws governing its conduct are derived. These are the Qur°an, the Sunna of the Prophet, the consensus of the community and its scholars,¹ and qiyas, the juridico- logical method of inference.² The first two sources provide the jurist with the material from which he is to extract through qiyas and ijtihad (the disciplined exercise of mental faculty) the law which he believes to the best of his knowledge to be that decreed by God.

This theory of consensus was the product of a relatively late period, when legal theory, u$ul al-fiqh, reached maturity. But in first-century Islam, and for that matter in pre-Islamic Arabia, consensus, as we shall see, existed primarily as a

© 1986 Cambridge University Press social and secular force.⁴ With the emerging political struggles and the sub­sequent growth of the religious movement during the second/eighth century, consensus acquired a religious character and became one of the most remarkable doctrinal features which distinguished ahi al-sunna wal-jama^za from other sects and movements. The important role of consensus as a sanctioning instrument and as a source of law of the “middle nation” was bound to generate extensive discussion and criticism from within and without Sunnism.⁵ The crucial problem which posed itself was the foundation and validity of consensus as a religious doctrine. In other words, classical Muslim jurists, who undertook the task of developing usul al-fiqh, had to prove that consensus rests on the strength of the two primary sources, the Qur³an and/or the Sunna. For, according to the fundamental Islamic tenet, nothing can be regarded as valid or binding if it is not somehow grounded in these sources. The implication of accepting consensus as a sanctioning force and, moreover, as a source of law without basing it on the sources was grave. It meant that those jurists who participated in the formula­tion of consensus were themselves the legislators, the law-givers of the com­munity.

II

Before we proceed with our discussion of the problem known in Arabic religious sources as hujjiyyat al-ijma^c, i.e., the authoritativeness of consensus, let us briefly examine the contribution of modern Western scholarship to this question. This will be done with a view to sorting out those opinions and conclusions which this essay attempts to reconsider or complement. The most pertinent are three views expressed by C. Snouck Hurgronje, N. J. Coulson, and George Hourani.⁶ Hurgronje maintained, and rightly so, that at a relatively early period, when the doctrine of the infallibility of the community was formulated, “people still dared to give greater convincing force to obvious truths by putting them into the mouth of the Prophet.” Thus,

several traditions, in various forms, make him say that “the whole of his community can never be in error”.... In the earliest period, people found it sufficient that the Prophet in his last words should have urged the believers most emphatically to “hold fast to the Book.” When they [i.e., Muslims] arrived at the conviction that the Muslim way of life was not based on the text of the revelation alone, it was inevitable that the Prophet in his farewell speech should have mentioned his sunna as the second foundation of religious duties; finally, when the infallibility of the community had become an essential dogma, this too had to be supported by his words.⁷

But Hurgronje also argued that when Muslims faced the difficulty of interpreting the texts and distinguishing the authentic traditions from the spurious, the infallible authority of the community was called for.

Authoritativeness of Sunni Consensus 429 laws and institutions which came into being after Muhammad’s death, but the whole practice of the Muslim community is based on its own infallibility.”⁸ Elsewhere, he explicitly argued that

attempts have been made to demonstrate it [i.e., consensus] by the Qur^an and the sunna, but it is impossible to hide the fact that this is to go round in a vicious circle. Only the infallible community can explain the sunna and Qur^an accurately; it is then completely idle to claim to establish the infallibility of the community by the authority of the Qur'an and the sunna.⁹

In sum, Hurgronje’s contention seems to be that in the final analysis consensus was the ultimate authority on interpreting, understanding, and authenticating the Qur°an and the Sunna, and that the “foundation of foundations” of Islam was the self-proclaimed infallibility of the community.¹⁰ Hurgronje’s conclusion, therefore, was that consensus rested on a petitio principal

This view was also expounded, though less unequivocally, by Coulson who argued that “although the validity of the principle is formally expressed in a Tradition from the Prophet which states: ‘My community will never agree upon an error,’ it is the ijma^Q itself which guarantees the validity of the Tradition.”¹² A similar view was also held by Joseph Schacht.¹³

George Hourani has studied the question of the authoritativeness of consensus and has reconsidered the statements made by Hurgronje.¹⁴ His findings were published in a lengthy article in 1964. In that article Hourani showed that consensus does not determine the interpretation of the texts on which consensus rests. He also attempted to show that consensus does not rest on a petitio principii. On the first point Hourani argued persuasively, but on the second, on the issue of circularity, his answer was brief and did not reflect the centrality and the true weight of this issue in medieval discussions.

(a) Any sound basis for the authority of an infallible, cumulative consensus, as a definite institution of Islam, must be found in a text of Islamic scripture. On this principle the medieval Muslim scholars were right.

(b) No such basis can be found in the Qur^an, or in any Tradition that can be regarded with confidence as authentic. This is what modern scholarship leads us to think.

Hourani then went on to say that “Using these two conclusions as premises, we arrive at a third: There is no sound basis for the traditional doctrine of consensus in Islam.”¹⁶

Thus Hourani used the results of modern scholarship to judge whether or not a medieval system of belief had rested on a “sound basis.” The question here is not whether we believe the traditions to be authentic, but whether or not Muslims throughout classical and medieval Islam accepted them as such. Insofar as Muslims were concerned, “the entire religion is taken only from God... and then from the language of the Prophet.”¹⁷ The Sunna was undoubtedly con­sidered as part of the wahy, God’s revelation.¹⁸ Believing what the Prophet said was a requisite for ImanS* The Sunna had the important function of clarifying the ambiguities, and if necessary, abrogating verses, in the Qur^Dan.²⁰ Its rejection was tantamount to heresy.

Goldziher’s and Schacht’s results about the spuriousness of hadith literature, though quite illuminating and indispensable for a more accurate understanding of the early developments in Islam, have no bearing whatsoever on whether consensus was authoritative or baseless. Muslims considered (and the majority still do) the Sunna to be a pillar of Islam and as such we must address it in the present inquiry. It must be immediately added that the question of the authori­tativeness of consensus must be dealt with as a subject of intellectual history. All propositions designated by the great majority of Muslim scholars as premises to their arguments must be treated as true premises.

in

The question of the authoritativeness of consensus arose only when it became necessary for Muslims to decide once and for all whether they themselves were the ultimate source of authority or whether that authority was God. In the Jahiliyya, the answer to this question was obvious; the sunna and the consensus of the tribe or the community constituted the law. Consensus, al-amr al-mujtama^{c c}alayh, created and defined the sunna.²² With the advent of Islam, the Qur³an encroached, to the extent dictated by its meagre content of law, on the estab­lished secular sunna. Now, with the help of the Qur°an, God had only begun to partake in determining man’s affairs. The spirit and the letter of the Qur³an and the Prophet’s sfra had undoubtedly left their mark on that society, but not, strictly speaking, on its law. With the growth of Islam, the ancient notion of sunna came to be associated with a new notion, viz. the Sunna of the Prophet. This is not to say that the Prophetic Sunna existed alongside a non-Prophetic sunna. Rather, the Prophetic Sunna was created by what Schacht termed “the living tradition,” the sunna as fixed by the consensus of one particular living generation. Thus the community, by associating its living sunna with the Sunna of the Prophet, granted itself the prerogative of determining the Divine Sunna, which was soon to become the most substantive source of law. In effect this meant that whatever the community decided would prevail. Only with this background can one fully appreciate the later doctrine which prescribed that the Sunna interprets and, when necessary, abrogates the Qur°an. In fact, this is a modified version of a second-century dominant view that it is the Sunna which decides upon the Qur°an and not the reverse.²³ The Prophetic Sunna, the sacrosanctified expression of the “living tradition,” along with the Qur³an, became the ultimate sources of the law. It is only in such an environment that Muhammad b. Hasan al-Shaybam (d. 189/804) could declare on the authority of the Prophet that “Whatever the Muslims see as good is good in the eyes of God, and whatever they see as bad is bad in the eyes of God.”²⁴

The process of creating the Sunna of the Prophet out of the “living tradition” began to be interrupted when the Prophetic Sunna was recorded in the form of hadith. Hadith gave permanence to the Sunna, thereby defining and determining its content once and for all. When the literature of hadith exhausted more or less the content of the Sunna, the full authority to determine the law was transferred from the hands of Muslims to those of God and his Messenger. For it was obvious that the process by which the Sunna was vested with divine authority had already come full circle. The transference of this authority was not sudden but gradual. It seems to have ended sometime in the beginning of the third Islamic century (ca. a.d. 830). However, the intellectual justification of the paramount position of revelation in relation to consensus had just begun.²⁵ The first signs of this process of justification had already appeared in Ibn IdrTs al-Shafi^cT (d. 204/820), who, to our knowledge, was the first jurist to system­atically lay down the jurisprudential principles of substantive law.

Shafi^cI, whose theory represents the almost complete victory of the authority of God over that of man, elevates the Qur³an and the Sunna of the Prophet to a supreme position, and relegates consensus to a subsidiary source of law. He distinguishes between two types of consensus. The first is the consensus of the community and the second that of the ulama, the legal specialists. The consensus of the community is again divided into two types. The first is that which is based on the Qur°an and on a Sunna transmitted “from the generality to the gener­ality,” and the second is not based on revelation,²⁶ and serves as a working consensus (mufid lil-^zamal, as later jurists would put it) which does not lead to certainty (ihata). The consensus of the ulama stands on an even lower level because it has the force of a solitary tradition, khabar al-khassa, which cannot lead to certain knowledge.²⁷ ShafiT does not consider this type of consensus authoritative. Only the first type, the consensus which is reached on the basis of the Qur°an and a Sunna transmitted by the generality of Muslims, can yield certain knowledge of the law decreed by God. ShafiT asserted that “It is impossible that the community should agree on something contrary to the words of the Prophet,”²⁸ and, a fortiori, contrary to the injunctions of the Qur°an. He went as far as to say that “the generality [of Muslims] cannot agree on... an error,”²⁹ thus anticipating the later Prophetic tradition “My community does not agree on an error.” But ShafiT’s acceptance of the consensus of the community even when it is not in custody of a text of revelation seems to have required justification, as he himself held that the Qur^Dan and the Sunna are the ultimate sources on matters of law.³⁰ In other words, ShafiT felt bound, in order to be consistent, to base this type of consensus on the two primary sources. This very fact manifests the triumph of revelation in ShafiT’s jurisprudence, where all things must either be based squarely on revelation or derived therefrom.³¹

In addition to Qur^an IV: 115 (And whoso makes a breach with the Messenger after the guidance had been manifested unto him, and followed other than the believers’ way, we appoint for him what unto which he himself had turned, and expose him unto hell), ShafiT quotes two traditions in justification of consensus, the first of which is: “Three things the heart of the Muslim shall not hate: sincerity of action for the sake of God, good advice for [fellow] Muslims and adherence to their community”;³² the second: “he who wishes to reside in Paradise must hold fast to the community, for Satan comes close to one person but not as close to two.”³³ It must be noted that in his discussions of consensus in the Risala, nowhere does ShafiT adduce the Quranic verse IV.115. The traditions, on the other hand, appear in the Risala on more than one occasion when consensus is discussed. We are therefore justified in assuming that he did not think of this verse as an important argument.³⁴

From the two traditions ShafiT concludes that the Prophet enjoined Muslims to adhere to the community, because the Muslim community cannot agree on an error. In answer to his interlocutor who inquired about the meaning of the traditions, ShafiT says that adherence to the community enjoined by the said traditions meant not physical adherence but adherence in belief. The Prophet, he maintained against his opponent, could not have meant that Muslims should physically hold to the body of the community since they reside in far-flung provinces and non-Muslims live amongst them. The Prophet could have only meant that the duty is to hold fast to the doctrines and actions as known and practiced by the community.³⁵

Although the meaning of these traditions is sufficiently clear, the argument as a whole gives rise to a serious objection. By the same epistemological standards set forth by ShafiT himself, it can be maintained that his argument for authori­tative consensus is not well knit. Though he does not state that these traditions are solitary (akhbar al-khassa), there is no doubt that he considered them as such. As we have seen, ShafiT deemed everything that was based on traditions transmitted through channels fewer than those of the “generality to the gener­ality” to lead only to probable knowledge. The consensus of the community on the basis of a text leads, in ShafiT’s opinion, to ihata, a conclusive and irrevocable knowledge. The weakness of his argument thus stems from the fact that he proves the universal validity of a principle which leads to certain knowledge by evidence which yields mere probability. It cannot be maintained, however, that ShafiT fell in the trap of circularity because the authenticity of the traditions adduced is not claimed to be guaranteed by the community, and their meaning is determined by the standard rules of the Arabic language. What ShafiT needed in order to perfect his theory of consensus was either an unambiguous Qur³anic verse guaranteeing the infallibility of the community or an explicit statement from the Prophet transmitted from the generality to the generality to the same effect. ShafiT would have fallen in the trap of circularity had he held the view that a consensus on a case of law, which is in turn based on a solitary tradition, leads to certainty. Had this been the case, he would have been obliged to justify consensus by the most conclusive and decisive textual evidence, because assigning to consensus the role of transferring rulings from the domain of probability to that of certainty means that consensus is a source of law. To admit, as it were, that consensus can be conclusive only when based on highly reliable texts, absolved consensus of its responsibility as a legislative source, and absolved Shafi^cT of the responsibility of proving that it was one. The certainty of consensus is thus a derivative of the certainty of revelation and not one which is self-constituted. This is in perfect consonance with the status which Shafi^cT accorded to consensus as a general legal principle.

After the death of Shafi^cT, the issue of grounding consensus in revelation became even more problematic. Unlike the theory of Shafi ^CT, which recognized as infallible only the consensus of the community when based on a reliable text, the great majority of later jurists held that infallible consensus is the consensus of the mujtahids (the qualified jurists who are authorized to discover the law) on any case of law, including a case solved by means of qiyas which may take its premises from ahadi traditions.³⁶ The authority of consensus to render probable things certain called for serious attempts at proving, on textual grounds, that consensus is a legitimate source of law (as! min usul al-shar^c).

IV

Unfortunately, there are no extant works from the third/ninth century which can shed light on the subject under discussion. The earliest period from which some writings have reached us is around the middle of the fourth/tenth century. From the work of the HanafT jurist Abu Bakr al-Jassas (d. 370/980)³⁷ and from that of his ShT^cT adversary al-QadT al-Nu^cman (fl. mid. fourth/tenth century), we can gain insight into how the issue of the authoritativeness of Sunni consensus was treated during the first half of the fourth/tenth century.

By the middle of that century textual evidence in support of consensus had become abundant. Jassas cites Qur³an 11:143, “Thus we have appointed you a middle nation, that you may be witnesses against mankind, and that the messenger may be a witness against you.” By equating “middle nation” with uprightness and rectitude, he concludes that the community should be described as infallible and as such it is appointed to witness the deeds of other people just as the Prophet witnessed his community’s deeds. He also cites IV: 115, already quoted by Shafi^cT, and argues that the Qur^an bids adherence to the call of the Prophet and threatens with punishment those who swerve from the path of Islam. The call, for unity behind the Prophet thus gives legitimacy to the agreement of Muslims. IX: 16 also supports the same theme: “Or did you suppose you would be left [in peace] when God knows not as yet those of you who strive, choosing for familiar none save God and his messenger and the believers?” From this he infers that opposing the believers is as serious a sin as opposing the Prophet himself. Jassas produces two more verses which are as inconclusive as the previous ones. They are 111:110 “You are the best community that has been raised up for mankind. You enjoin good and forbid evil,” and XXXI: 15 “And follow the path of him who turns to me.” In the first of the two verses Jassas interprets the praise for the community in terms which indicate that the community cannot agree on an error. In the second verse the reference is, of course, to the Prophet: to follow the Prophet is to save the community from going astray.

Realizing that all these verses have no definite meaning which can establish, once and for all, the authoritativeness of consensus, Jassas resorts to the Sunna to enhance his argument. In addition to the two traditions which Shafi^cT had already introduced, he brings forth five more traditions, all of which are admittedly ahadis.³⁸

In his critique of Sunni legal theory, QadT al-Nu^cman records the Quranic verses and traditions used by the Sunni jurists to justify the authoritativeness of consensus. All but three Quranic verses (VIII: 181, XXII:78, LVII:19) are mentioned by Jassas. The general meaning of these three is identical to that of the aforementioned 11:143 and III: HO.³⁹ It is to be observed here that the number of verses and traditions adduced by Jassas, and obviously by his contemporaries, has increased considerably since the time of Shafi^ci. This increase was, as we shall see, of great importance for later legal theorists.

v

One of the earliest writers in whose works we find a comprehensive treatment of this issue is the Mu^ctazilT jurist and theologian ^cAbd al-Jabbar al-Asadabadi (d. 415/1024). In his definition, consensus is not only the agreement of the qualified jurists but also of the community as a whole.⁴⁰ It constitutes an authority whose proof must derive from the revealed sources. Reason is of no use in establishing the authoritativeness of consensus, because there exists no rational evidence which can prove that a group of people or a community is not liable to error.⁴¹ It is scriptural evidence, he states, which is the only acceptable evidence. ^cAbd al-Jabbar starts by citing Qur°an IV: 115, already quoted by Jassas and allegedly by Shafi^cT. Using the same polemical style of exposition used by the latter in the Risala, ^cAbd al-Jabbar undertakes a lengthy explanation of the meaning of this verse. He argues that God’s warning with regard to dissenting from the lines of the Prophet and the believers necessarily means that following their lead is the right thing to do. The path which the Muslims tread, he reasons, must be, in accordance with the verse, the most righteous and therefore authoritative. Had such authority not been inherent in them they would not have been endowed with the quality of righteousness.⁴² Upon the suggestion of the interlocutor that the warning against breaking away from the path of the Prophet and the believers does not necessarily mean that such path should be followed, ^cAbd al-Jabbar retorts that the meaning is clear: If dis­senting from something is prohibited then it follows that embracing a different thing is also prohibited. From this it also follows that the only choice is to adhere to the path of the Prophet and the believers.

Having conceded to ^cAbd al-Jabbar this interpretation, the interlocutor then adroitly argues that if the authoritativeness of the community’s consensus is contingent upon the condition of following the path of believers, then belief

Authoritativeness of Sunni Consensus 435 (iman), is the criterion for authoritativeness. This being so, the iman of one person would make him an authority, and therefore, his position on a particular issue becomes authoritative in much the same way that the position of the whole community becomes authoritative. Accordingly, the interlocutor says, the accep­tance of consensus on the basis of iman necessitates upholding the decision of an individual Muslim as authoritative. Clearly, the interlocutor is purporting by this reductio ad absurdum argument to undermine the interpretation given to the verse by ^cAbd al-Jabbar. Since the linchpin of this interpretation is the phrase the “believers’ path” (or “believers’ way”), ^cAbd al-Jabbar remarks that the “believers’ path” is expressed in the deeds, utterances, and creed of Muslims as based on evidence from revelation (or reason!).⁴³ It cannot be contended, he says, that the path becomes defined only after Muslims reach a consensus on what constitutes iman.

The interlocutor then points to the plural form of the word “al-mu^mimn” (the believers) and argues that the apparent meaning of the verse is that a consensus can be authoritative if three believers reach it—since three is the smallest number in the plural. ^cAbd al-Jabbar refers his opponent to the chapter in his book entitled “On the General” (al-^Qamm) where he explains that when the definite article al· is prefixed to a noun in the plural, as is the case here, the noun will indicate a class of things, which is in this case all the believers. The opponent then goes on to debate the meaning and definition of the term “believers” and, again, ^cAbd al-Jabbar does not fail to assert his position. After some discussion, the latter concludes that the believers of the community constitute the consensus­making body, and since God has not specified in the verse a certain time for the validity of consensus, he also concludes that the consensus of each and every generation of Muslims is an authority.⁴⁴

^cAbd al-Jabbar dismisses XXXI: 15 “And follow the path of him who turns to me”—adduced earlier by Jassas—on the grounds that the context in which it was revealed makes it evident that the reference is not to any specific group of people, but to mankind in general. Moreover, this injunction, as shown in the lines which open the ay a, refers to the respect man owes his parents.⁴⁵ Similarly, ^cAbd al-Jabbar dismisses 111:103, 111:110, IV:59, VIL181, and XI.118, adduced by other scholars in justification of consensus, on grounds that they bear various meanings not necessarily relevant to consensus.⁴⁶

The second verse produced in justification of consensus is 11:143 (“Thus we have appointed you a middle nation, that you may be witnesses against man­kind, and that the messenger may be a witness against you”) which, ^cAbd al-Jabbar notes, had been adduced earlier by Abu ^CA1T al-Jubba³T (d. 303/915). Like Jassas, Jubba³! and ^cAbd al-Jabbar understand “middle nation” as a metaphor indicating the quality of justice and rectitude inherent in the com­munity of believers. This naturally explains why they were appointed as wit­nesses against other communities. Had their testimony been dubious in the least, such a responsibility would not have been entrusted to them. God’s choice of appointing these witnesses, ^cAbd al-Jabbar insists, leads us to think that God accepts their word as reliable and truthful. Furthermore, in the second part of the verse, the Prophet is also said to be a witness because he is just and his utterances are authoritative. From all this it is concluded that the collective decision of the community constitutes an authority.⁴⁷

Expectedly, the interlocutor attempts to undermine the relevance of the verse to consensus by pointing out that the apparent meaning of the verse does not correspond to reality. While the community as a whole appears to be described as just, it is obvious that amongst the community there are in reality many dissolute and impious persons. Therefore, he says, one must not use this verse in support of consensus. ^cAbd al-Jabbar responds “community” means those who qualify as believers and not necessarily all members of the community. In support of his interpretation he cites other Qur³anic verses in which collective nouns are used in a restrictive sense. Unsatisfied with this, the interlocutor contends that it is the companions of the Prophet who are addressed in the verse rather than the community at large. The phrase “We appointed you” pre­supposes an addressee, who can only be the community of the companions. Accordingly, it is the consensus of the companions which is authoritative, and not that of the entire community of believers. ^cAbd al-Jabbar retorts swiftly and decisively by citing a number of Quranic verses which make it abundantly clear that those who are appointed witnesses are the Muslims as a whole. Illustrative of these verses is XXIII:78, “And strive for Allah with the endeavor which is his right.... He had named you Muslims of old time and in this [Book], that the messenger may be a witness against you, and that you may be witnesses against mankind.”⁴⁸

Having discussed the Qur³anic evidence, ^cAbd al-Jabbar goes on to note that Jubba³? based consensus on Prophetic traditions such as “My community shall never agree on an error.” The former remarks that traditions with regard to the authority of consensus are plentiful and well known, although some are better known than others. He lists about 10 of these traditions saying that many of them were reported in different versions.⁴⁹ The companions of the Prophet knew these traditions and acted upon them. They are so famous, ^cAbd al-Jabbar says, that we need not trace back the wording of their transmission, just as we find no need to trace down the textual evidence upon which many forms of prayer and rituals are based. They are taken to constitute a proof of the authoritativeness of consensus although the form of their transmission is not a mutawatir. The purpose of scrutinizing the authenticity of these traditions is to reach an understanding of their true meaning, but if their meaning is already known on the basis of their multiple non-mutawatir transmission, then such a scrutiny becomes superfluous. ^cAbd al-Jabbar, in answer to his opponent, remarks that it is possible that these ahadi traditions may have existed in the mutawatir fashion at the time of the companions and their successors, but that their transmission has been neglected in later generations. Here, the opponent asks why these traditions, if they were indeed the basis of an authoritative consensus, had been so neglected. ^cAbd al-Jabbar replies that such a happening was possible because all believers acted upon these traditions and their practice of reaching consensus on points of law was so undisputed that the textual basis authorizing such practice was taken for granted. It has been generally known, he says, that some mutawatir traditions have become ahadi due to the neglect of people to transmit them. An excellent case in point is the undisputed zakat which was grounded in traditions from the mutawatir category, traditions which have now become ahadis.⁵⁰ It is on these grounds that the ahadi traditions, relative to consensus, are taken to be authoritative.

In all this the interlocutor sees a contradiction (tanaqud). If the companions were in agreement on accepting the traditions concerning consensus, this would mean, the interlocutor insists, that the companions reached a consensus on the traditions which were in turn used to justify consensus itself. Such a contradic­tion can indeed undermine ^cAbd al-Jabbar’s argument. ^cAbd al-Jabbar, however, explains that the validity of the traditions is not determined by consensus but by custom (al-^cada al-jariya). The companions could not have accepted so unani­mously these traditions unless they were certain of their authenticity. It was inconceivable that they all would agree on a tradition which was dubious, as each of them was free to exercise his own independent ijtihad. One knows from the custom in evidence (al-^ada al-zahira) that when a tradition was not proven authentic, some of the companions would accept it and some would not. Any tradition which is unanimously accepted must, pursuant to the law of custom, be authentic. It is in this order that ^cAbd al-Jabbar wishes to drive his argument: authenticity leads to unanimity—consensus does not authenticate traditions but is the result of their authenticity.⁵¹

The authenticity of traditions is thus guaranteed by custom. Custom, a theological concept already known to Ash^cari (d. 330/941), is described as the repeated actions of God which seem to us as a continuous course of events. But in contradistinction with nature, which is always continuous, custom may allow a break in its continuum. From custom we know that the sun will rise tomorrow morning and that there is an inevitable change of seasons throughout the year. Knowledge of many things in this world is derived from custom.⁵² For instance, if we are able to observe the sun for an extended period of time we know that after every night there is a day and from this we come to a conclusive knowledge that after this night there will be a day, unless, of course, God suddenly decides that He will change his custom of running things in this manner in the world. This change of custom is an irregularity and is known as kharq al·^c a da or, as ^cAbd al-Jabbar calls it, ikhtilaf al-^Q a da. The authenticity of a tradition is known by tracing the history of its transmission and finding that there is no ikhtilaf in custom. The occurrence of any irregularity in custom will point to the deficiency of the tradition. The most important factor which determines a sound custom is the repeated transmission of the tradition under normal circumstances (al-takrar ma^c salamat al-ahwal).⁵³ Custom asserts itself every time a transmitter relays the tradition to an audience, and the more the acts of transmission are repeated the more entrenched and certain our knowledge of the authenticity of the tradition under question will be. Continuity guarantees the knowledge of the Prophetic past.⁵⁴ Conversely, knowledge of this tradition ceases once its transmission is halted, a course of action which constitutes kharq al-^cada.⁵⁵

^cAbd al-Jabbar does not set forth the epistemological principles upon which his argument for authoritativeness is based. He is well aware that consensus is conclusively authoritative (hujja qat^Qiyya), but he does not state whether or not the evidence he advances for its justification is equally certain. His argument is based on Qur³anic verses and traditions which, although admittedly ahadis, he claims them to have been mutawatir at the time of the companions. Although he goes to great lengths to prove that they are mutawatir through the concept of ^cada, ^cAbd al-Jabbar does not take the hadith evidence to be as important or as strong as that of the Qur³an. He observes that the strongest among all the pieces of evidence is IV: 115, the first verse discussed by him.⁵⁶ Thus he finds the Qur³an to be a stronger basis of authority for consensus than the traditions. In light of the principle, which he must have known, that ambiguous Qur³anic verses do not lead to certainty, he seems to have considered IV: 115 as an unambiguous verse which leads to conclusive knowledge. However, by the epistemological standards prevalent at his own times and even afterwards, the Qur³anic verses which he cites are not unambiguous, as is evident from the different interpreta­tions given to them by ^cAbd al-Jabbar and his adversary. The innumerable objections of the opponent to ^cAbd al-Jabbar’s interpretation demonstrate that the verses are each open to more than one interpretation. An unambiguous verse, as defined by Muslim jurists themselves, is a verse the meaning of which is so clear that it would be incapable of another interpretation. Such a verse yields certain knowledge. On the other hand, ^cAbd al-Jabbar makes recourse to the consensus of the companions to prove that the ahadi traditions which support an authoritative consensus were mutawatir in the first Islamic generation. Believing himself to have succeeded in doing so, he uses the allegedly mutawatir traditions with the Quranic verses as the basis of consensus. But if ^cAbd al-Jabbar was totally convinced that he has succeeded in establishing the tawatur of these traditions, why does he insist that IV:115, a verse commonly regarded as ambiguous, is the strongest piece of evidence which can be adduced in justifica­tion of consensus? To this question I can find no answer in ^cAbd al-Jabbar’s writings.

Nor is it possible to find an answer in Abu Husayn al-Basri’s treatise al-Mu^ctamad ft usul al-fiqh, which is based on ^cAbd al-Jabbar’s usul work al-^Ahd.^ Basri follows ^cAbd al-Jabbar in accepting some Quranic verses (11:143, 111:109, IV: 115) while rejecting others (111:103, IV:59). He adopts the argument based on ahadi traditions, but without claiming that at the time of the companions they were mutawatir. The fact of their being ahadis is acknowl­edged, and Basri does not attempt to change their status. He argues, however, that if one is to base consensus on these traditions only, such consensus would not yield absolute certainty, but will merely be authoritative insofar as action (^camal) is concerned.⁵⁸ Commenting on what seems to have been a contem­porary argument that although the aforementioned traditions are ahadis their meaning has been transmitted in the mutawatir fashion (wa^in nuqila bil-ahadi fahnna ma^Qnahu manqulun bil-tawatur), he says that some of these traditions are not genuine, thus their sum total cannot yield absolute certainty.

Rather characteristic of the arguments of ^cAbd al-Jabbar and Basri is the absence of clear-cut epistemological distinctions with regard to the evidence produced in justification of consensus. Such distinctions were always present in the expositions of Badri’s younger contemporaries and their successors. The

Authoritativeness of Sunni Consensus 439 problem that posed itself once these distinctions came to be included in the stated arguments is that consensus was considered conclusively authoritative (hujja qafiyya), but the textual pieces of evidence relative to its authority were, when taken individually, inconclusive. How, in light of this difficulty, later jurists grappled with the problem of authoritativeness will be shown through analyzing, among other things, the writings of JuwaynT (d. 478/1085), SarakhsT (d. 483/1101), Ghazal! (d. 505/1111), AmidT (d. 631/1233), and ShatibT (d. 790/1388).

VI

JuwaynT insists that nothing short of conclusive and certain evidence is acceptable in establishing the authoritativeness of consensus, as consensus itself yields the highest degree of certainty. A principle which leads to certainty must be based on equally certain evidence. In the case of consensus, such evidence must be derived from revelation; rational evidence cannot serve in religious matters.⁵⁹ JuwaynT quotes Qur°an IV: 115 and argues that one of the meanings of “makes a breach with the messenger” is secession from Islam altogether. He remarks that the verse is ambiguous and is capable of more than one interpreta­tion. Similarly, he discards the major tradition “My community shall never agree on an error” on the grounds that it is an ahadi and, like the aforementioned verse, may be subject to at least two interpretations. The first of these interpre­tations is the infallibility of consensus, and the other is apostasy, viz. that the community shall never agree to apostate until the Day of Judgment.⁶⁰ By the same token, he dismisses all other traditions thus far introduced.⁶¹ Against the argument that there is a kind of universal agreement on the acceptability of these ahadi traditions, he cautions, and rightly so, that such an argument would be circular. Consensus, he insists, cannot rest on another consensus. The only admissible evidence is either mutawatir traditions (hadithun maqtu^Qun bihi naqlan) or unambiguous, decisive Quranic verse, and these, he acknowledges, are lacking.⁶²

Since consensus could not be justified by arguments based on rational (non­textual) proof, and since there appeared to be no satisfactory textual evidence, how then could consensus, “the foundation ifimad) of the Shari^ca,” be justified? JuwaynT answers that when consensus is reached on a point of law, it takes one of two forms. The consensus of the first form occurs when the qualified scholars in the far-flung Islamic provinces agree on a zanni legal judgment. Such an agreement of opinions could not be reached on the basis of the jurists’ rational (non-textual) reasoning because, although theoretically possible, custom (mustaqarr al- ^cada) has proved the actual occurrence of such an agreement to be impossible. A consensus on a shar^QT matter on which reason cannot decide cannot be reached without a common cause (sabab jami^z) which leads those who partake in the said consensus to hold the same opinion. Therefore, a total agreement is possible only when the original decision is based on a conclusive text, a text which cannot be the subject of disagreement. For this reason, JuwaynT concludes, when a consensus appears to have been reached on a zanni legal judgment, it must be presumed that such consensus is grounded in indubitable texts.⁶³

The consensus of the second form, also reached on a zanni judgment, is authoritative, JuwaynT says, on the grounds that

we have found that past generations and extinct nations (al-umam al-munqarida) agreed on reproaching those who violate the consensus of the ulama, the ulama of time immemorial. They have charged all those who diverged (from the mainstream) with disloyalty, apostacy and disobedience. Deviation was never considered a trivial matter. [These nations^ viewed the disagreement with the ulama an outright heresy (dalal). Thus, consensus on this [i.e., reproaching divergence] amounts to certain knowledge based on rational grounds. If such [consensus] were attached to the consensus reached on an uncertain (maznUn) legal judgment, then the legal judgment is rendered irrevocably certain (qat^c). Therefore, let consensus on reproaching and reprimanding those who diverge be a legal precept which yields certainty (^n/^c).⁶⁴

Now, by “past generations and extinct nations” JuwaynT was not referring to the ancient Muslims only but to all “nations” which preceded Islam. And by “ulama” he meant all the sages and scholars who decided in the religious affairs of these nations. But having said this, JuwaynT seems to have immediately realized that basing the infallibility of consensus on another consensus, however ancient and glorified the earlier consensus may be, does not really solve the problem. For as he already stated, consensus must be grounded in conclusive textual evidence. Grounding consensus in a non-textually based consensus leaves his argument assailable on the grounds that it leads to circularity. JuwaynT had just criticized those who use traditions which are authenticated by consensus to justify consensus. In order to avoid this impasse he hastily adds that it is probable that the consensus of the first Muslim generations was based on the aforementioned traditions (which he just discarded), but that such consensus was continually used and taken so much for granted that Muslims neglected to transmit the traditions which once constituted its unshakable basis of authority.⁶⁵

A close look at this argument reveals that for JuwaynT consensus is not universally authoritative. In both forms of consensus which he distinguishes there is obviously no evident textual basis except for that according to which the legal case is solved and which is presumed to have been originally conclusive. By the law of custom, the principles of which ^cAbd al-Jabbar had already expounded, JuwaynT claims that the consensus on what seems to be a zanni legal judgment must have been reached on the basis of a presumably qat^ci text. This alone can explain the existence of an agreement on a shar^ci rule. In the second form, there is an implicit assumption that the zanni textual basis of the case was originally qat^ci. But since JuwaynT has already advanced this thesis when dealing with the first form, he goes on to prove that it is in the custom of nations and communities to condemn divergent views and fissiparous tendencies, and thus such a universal consensus must confirm any agreement on a shar^ci rule. JuwaynT admits that he has no textual evidence to prove this (customary) universal consensus, although he allows for the possibility that traditions from the Prophet to this effect were once in existence but are now lost. His ultimate proof for the validity of both forms of consensus is the principle of custom.

Anything else in his argument is merely presumed. Even if one accepts the validity of custom, it follows that in JuwaynT’s theory consensus is not authori­tative as a universal principle. “The truth,” JuwaynT acknowledges, “w that consensus in itself is not an authority. It is inconceivable that those who take part in consensus (mujmUin) can decide a rule [of a case] independently. It must be presumed that they discovered something which brought about their,, 66

consensus.

JuwaynT’s concept of consensus is limited in scope. His arguments do not reflect the need, apparent in the writings of his predecessors and contemporaries, to establish the infallibility of the community, the universal guarantee for an authoritative consensus. The infallibility of the community as a general principle is not recognized by JuwaynT.⁶⁷ Consensus is valid only when the specific case on which it is reached derives from revelation.

In al-Munkhul min Ta^cliqdt al-Usui, which he wrote while still under the influence of his teacher JuwaynT, GhazalT reiterates in his own way the argument as presented by JuwaynT. However, he articulates the first part of the argument to make it somewhat more convincing. He says that if the jurists reach a consensus on what seems to be a zanni legal judgment, this judgment becomes certain because the number of those who participate in consensus is at least equal to the number of transmitters required in the mutawatir transmission. If such consensus is doubted, the reliability of mutawatir traditions, which lead to the highest degree of certitude in law and which are universally accepted, must be doubted as well.⁶⁸ Although with this point GhazalT makes his argument easier to digest, the larger problem of authoritativeness has not yet been solved. What GhazalT succeeds in proving here is that a consensus on a textually based case of law is authoritative only insofar as that particular case is concerned.

In al-Mustasfa, which he wrote during the last few years of his life, GhazalT produced a rather thorough statement concerning consensus. His atti­tude toward textual evidence now differs drastically from that which he adopted when he was under the influence of JuwaynT. While still dismissing the QuPanic evidence as unsatisfactory,⁶⁹ he now accepts the traditions which were used by his predecessors to justify consensus. He enumerates 11 traditions, among which is the tradition “My community shall never agree on an error.” The rest of the traditions, which are all ahadis, were taken to enjoin Muslims to hold fast to the community, to forewarn fissiparous tendencies, and to restate the infallibility of the community.⁷⁰

But GhazalT is well aware that ahadi traditions do not lead to the certainty required in consensus. He therefore introduces a principle which, though it was known to his immediate predecessors, he did not use earlier in al-Mankhul. GhazalT maintains that these traditions, though they are not transmitted arith­metically in the mutawatir fashion, lead to ^Qilm daruri (immediate knowledge)⁷¹ because they are mutawatir in meaning (bil-ma^Qna). This means that the transmission of these traditions occurs through a multitude of ahadi channels, and all traditions, though different in wording, convey the same meaning. The Prophet, GhazalT says, has declared the infallibility of this community by a multitude of explicit statements and intimations which make it necessary for the mind to be certain of such knowledge.⁷² Certain knowledge of the authori­tativeness of consensus immediately occurs in the mind when the mind becomes acquainted with these traditions. Just as the generosity of the reputable Hatim is known without any doubt through the transmission of non-mutawatir reports, so is the authoritativeness of consensus. Ghazal! argues that each of these traditions alone can probably be dubious, but the same cannot be said of all of them as an aggregate. The support which each tradition gains from the others makes it improbable that all of them together are doubtful. Thus, like the mutawatir, these traditions as a whole lead to certainty.⁷³

Ghazal! then goes on to establish the meaning of the tradition “My com­munity shall never agree on an error.” He lists three objections made by the opponent concerning the interpretation of the tradition and then he refutes them. While it is the contention of the opponent that “error” (dalala, khata³) means disbelief hnd heresy, thus depleting the tradition by confining the role of consensus to a mere safeguard against heresy, Ghazal! interprets “error” in the general sense of a mistake.⁷⁴ He refers to two Quranic verses in which the term dalal was used in this latter sense. From the vast bulk of textual evidence, Ghazal! insists, one necessarily understands (fuhima daruratan) that because individual Muslims are fallible, God has bestowed on his community as a whole the quality of infallibility (^cisma), which the Prophet had also enjoyed. Other opponents concede that the meaning of khata³ and dalal is a mistake, but they argue that the verse is not universal in meaning and does not guarantee infallibility in all spheres of religious life. What is meant here, they say, is that the community does not err in such matters as testifying on the Day of Judgment and in transmitting the mutawatir traditions. But Ghazal! rejects this interpreta­tion on the grounds that no one in the community has held such exegetical views; the limitation of the meaning of “mistake” is unjustifiable.

The third and final objection concerns the meaning of “community.” The opponent claims that by “community” the Prophet meant all the generations of his “nation” from his time until the Day of Judgment. Against this Ghazal! retorts that just as insane people and small children are not counted among the (capacitated) members of the religious community, people who died and who are yet unborn cannot be considered as part of the (present) community. By community, therefore, is meant the living generation of Muslims. In short, Ghazal!, like any other Sunni jurist, accepted the standard meaning of the tradition as guaranteeing the infallibility of the community. It is significant to note here that no comprehensive interpretation of the traditions concerned was attempted by any Sunni jurist. Most of the usulists who dealt with the problem seem to have assumed that the meaning of these traditions is sufficiently clear so as to require no thoroughgoing interpretation. In establishing the meaning of the tradition “My community shall never agree on an error” Ghazal! was chiefly responding to the onslaught of the adversary. It may have been the case that since these traditions differed only slightly in the manner they were expressed it was assumed that the elucidation of the fundamental terms employed in the major tradition commented upon by Ghazal! should suffice to establish the general meaning of the others. In any case, the agreement among scholars upon

Authoritativeness of Sunni Consensus 443 the meaning of the traditions must not be taken to imply that such agreement constitutes in effect a consensus on the infallibility of consensus. Usulists make it abundantly clear that the meaning of legal propositions is determined by the standard rules of the Arabic language, rules which are normally expounded in the opening chapters of works on usul al-fiqh.

In addition to these arguments GhazalT goes about proving the authori­tativeness of consensus by means of what he calls rational demonstration (istidlal). Referring to the traditions which he quoted, he says that they have been famous since the time of the companions and except for Nazzam, whose opinion does not count, no one has cast doubt with regard to using them to prove the authoritativeness of consensus. It is inconceivable that many con­secutive generations of Muslims should accept a principle if it has no sound basis, especially that Muslims differed among themselves on many issues, par­ticularly those issues based on ahadi traditions. In other words, what GhazalT is saying here is that if the principle of consensus were based on less than satisfactory traditions it would have been swept away by the opposition of Muslims. But the fact that consensus was so important a doctrine and that it was met with overwhelming approval is evidence that it must have been based on solid grounds.⁷⁵ One knows of the existence of such approval from the lack of contrary evidence; that is, from the absence of opposition. GhazalT maintains that when Nazzam, the first figure in Islam to question and reject the authori­tativeness of consensus, voiced his opposition the community immediately knew of it, and that if there had been other voices, especially among those who were qualified enough to be taken into account, they would have been heard. GhazalT thinks that if Nazzam, whom the community did not regard with high esteem, was noticed, other important figures would have been certainly noticed had they cast doubt on the authoritativeness of consensus.

Against the accusation that the authority of consensus is based on traditions which are authenticated by consensus, GhazalT retorts that the authenticity of traditions is known by the absence of opposition to them, something which is in turn known through ^cada, custom. As applied to this argument, GhazalT’s concept of custom does not differ from that of ^cAbd al-Jabbar. GhazalT insists that through ^cada one knows, for instance, that the opposition to the Qur^Dan did not succeed and that the report about the Prophet appointing ^CA1T as caliph is spurious. The course of events would have been different had the opposition to the Qur³an succeeded or had the appointment of ^CA1T been genuine. The acceptability of the traditions in question became, as GhazalT conceived of it, a “habit” of nature and any break in this “habit” would constitute a break in the ongoing custom (kharq al-^cada). This is why he viewed the opposition of Nazzam to the textual basis of consensus to be so significant, because had Nazzam’s opinion been admissible in Sunni Islam, his opposition would have constituted a breach of what Muslims conceived to be a consistent course of events.

Like JuwaynT and particularly ^cAbd al-Jabbar, GhazalT utilizes the concept of custom to prove the existence of (presumed) decisive textual evidence in favor of an authoritative consensus in the early period of Islam. He says that when the companions agree on a certain issue, their agreement must be presumed to have been based on a conclusive text, for through the law of custom one knows that such unanimity on a dubious text cannot take place. Thus, the disapproval of the companions and their successors on the matter of divergence from the com­munity must have been based on a revelation which was certain.

The change or development of GhazalT’s views as shown in al-Mankhul and al-Mustasfa is quite significant. Though his concept of custom and the use made thereof in the latter work are somewhat different from that presented in the former, the fundamental idea remains the same. In fact, the discourse on the notion of custom in ^cAbd al-Jabbar, JuwaynT, and GhazalT are variations on the same theme. The more significant point in GhazalT is his acceptance of the notion of al-tawatur bil-ma^cna, which seems to have become by his time one of the most central arguments advanced in support of an authoritative consensus. The centrality of this notion appears most clearly in the Usui of SarakhsT.

SharakhsT’s argument is based on the Qur³an and the Sunna. From the Qur^an he cites 11:143, 111:110, and IV: 115. Like ^cAbd al-Jabbar, he does not point out the weight of the verses in the overall proof for establishing the authoritativeness of consensus. After giving the standard interpretation of the verses he moves on to the Sunna, which, he says, contains abundant and well- known traditions {mustafida mashhura). He lists a number of traditions con­cerning the infallibility of the community and argues that traditions supporting this point are numerous and mutawatir. The claim for tawatur, he says, derives from the fact that each transmitter heard a tradition to this effect and related it without anyone doubting the truthfulness of the tradition. A multiplicity of traditions transmitted under these conditions must be considered to be on par with the mutawatir (fadhalika bimanzilati al-tawatur)™ Although SarakhsT does not mention the expression “tawatur bil-ma^cna,” it is abundantly clear that he meant this type of tawatur rather than al-tawatur al-lafzi. A rather similar argu­ment is also advanced by al-KhatTb al-BaghdadT (d. 463/1070), who explicitly states that al-tawatur min tarTq al-ma^cna leads not only to certainty in law in general but also to a conclusively authoritative consensus.⁷⁷

VII

* Toward the end of the sixth/twelfth century or the beginning of the seventh/thirteenth, Sayf al-DTn al-AmidT (d. 631/1233), in a rather exhaustive statement on the problem of authoritativeness, elaborated and recapitulated the arguments as they had been developed by Sunni jurists up to his time. In his exposition we find a well-arranged reformulation of the totality of arguments which were presented in the earlier works in a fragmented manner. AmidT divides his arguments into three types: the first is an argument from the Book, the second from the Sunna, and the third is a rational argument.

In support of the authoritativeness of consensus, AmidT first cites five Qur^Danic ’ verses (11:143, 111:103, 111:110, IV:59, IV: 115), all of which have been quoted

Authoritativeness of Sunni Consensus 445 above. He devotes a dozen pages to the interpretation of these verses and to the refutation of the adversary’s claim that the meaning of these verses does not point to the infallibility of the community.⁷⁸ But after Amid! had undertaken the arduous and prolonged task of showing what he believed to be the true meaning of the verses,⁷⁹ he remarks that the totality of these verses does not lead to certain knowledge which is required in consensus. But he soon adds that the Sunna is the surest way to prove beyond¹ doubt the conclusive authoritativeness of consensus.⁸⁰ He introduces 16 traditions,⁸¹ and although he admits that they are all ahadis, he argues, like GhazalT, that they lead to certain knowledge because they are mutawatira bil-ma^cna.⁸² With an ahadi tradition, certain knowledge occurs when the tradition is supported by qartr'in, that is, circum­stantial evidence including other traditions handed down through various and unconnected channels of transmission or various types of Qur^Danic verses. AmidT gives a non-textual example of qara^Din: If a man looks often at a beautiful woman we would think (nazunn) that the man is in love with this woman. But if we also know that he stays in her company we gain more, but still inconclusive knowledge (zada zannund) about his love for her. If we know other facts, such as the fact that he buys her many gifts, treats her kindly, and that psychological and behavioral changes occur to him, etc., we conclude that he is certainly in love with her.⁸³ By the same token, we gain certain and immediate knowledge (film daruri)^ on the basis of a multitude of traditions which are mutually enhanced, that consensus is authoritative.⁸⁴ This is despite the fact that each tradition can separately be false (yatatarraqu ilayhi al-kadhib). However, AmidT adds, the totality of these traditions yields immediate knowledge in the mind of any man of sound reason.

As to the rest of his argument concerning the hadith basis of consensus, AmidT follows in the footsteps of GhazalT. He argues that had consensus been based on questionable grounds it would not have been so unanimously approved and supported throughout numerous generations, and had there been any opposition to these traditions or to consensus itself, it would have been immediately known, just as the objections of the infamous Nazzam became known. AmidT also follows the same outlines of GhazalT’s argument with regard to custom as the means by which traditions are authenticated, thus avoiding a circular argument.

What AmidT designates as a rational argument is essentially an elaboration of JuwaynT’s argument which he based on the notion of custom and which GhazalT subsequently refined. The main point advanced is that when we observe that the scholars of an earlier generation are unanimously agreed on a legal judgment, that judgment must be presumed to have been based on a conclusive textual evidence. This is so because the number of those who agree in the community is equal to, or larger than, the number of transmitters of a mutawatir tradition. From the law of custom, as previously noted, we know that it is impossible for a vast (mutawatir) number to agree on a falsehood or a forgery. However, AmidT hastens to add that in order to consider such a consensus authoritative, some scholars require the number of those who participate in consensus to be at least equal to that coAstituting the mutawatir.

The question that may be asked here is why Amid! introduces this rational argument and, before that, the indecisive Qur³anic evidence, if he knows that both arguments are inconclusive? An answer to this question requires reference to the logical principle which he, and GhazalT before him, brought into play when discussing the question of al-tawatur bil-ma^cna. Though the Qur^Danic evidence and the rational argument do not conclusively prove the authori­tativeness of consensus, they nonetheless function as qara^in, that is, as circum­stantial pieces of evidence which, as an aggregate of testimonies, enhance his argument based on traditions. As we are able to gather from later sources (some of which will be discussed below), AmidT seemingly saw the Quranic and rational evidence, notwithstanding its weakness, as a supportive proof, just like the kind of proof a single ahadi tradition would offer.

This principle of cumulative or aggregate evidence, still nebulous in the writings of jurists of the fifth/eleventh and sixth/twelfth centuries,⁸⁵ was articu­lated to a large extent by the illustrious Abu Ishaq al-ShatibT (d. 790/1388). In his work al-Muwafaqat fl usul al~ahkam, he explicitly states that the “roots” of law, including consensus, are decisive, certain (qat^ciyya), and infallible (ma^Qsumd)™ because they are founded on conclusive premises (muqaddimat qat^ciyya). The most important source of knowledge which he reckons to be the font of law is that which is based on tawatur, both lafzi and ma^cnawi, as well as induction from the religious texts, al-istiqra ^J fi mawarid al-shari^ra.^ ShatibT argues that induction leads to certainty because in the context of the Shari^ca one can obtain a complete enumeration of the instances pertaining to a certain issue. In other words, it is possible, ShatibT insists, to have a complete induction (istiqra³ tdmm or kulli), because the subject matter of the Qur³an and the Sunna is finite and can thus be exhausted. Because earlier jurists misunderstood this fundamental logical principle, ShatibT argues, some jurists thought that consen­sus itself was zanni. But this is not true, he says, because the cumulative evidence from the two primary sources render consensus certain.⁸⁸ Most legal (textual) evidence does not lead to certainty if it is not supported by other pieces of evidence. If the complete aggregation of this evidence is pronounced unaccept­able, it would then be impossible to arrive at certainty in law. What ShatibT is saying here is that the nature of the Shari^ca requires a reliance on such inductive principles, for it is the only available means to reach a kind of certainty.⁸⁹

The recognized evidence here [i.e., in law] is that which is reached through induction from a totality of uncertain (zanni) pieces of evidence which in their multitude support one meaning so much that they result in certainty (qat^ci); for confluence brings about a strength which disjunction lacks. For this reason the tawatur results in certainty, for it is a variety of induction... and this (tawatur) resembles al-tawatur al-ma^cnawi.⁹⁰

The authoritativeness of consensus, ShatibT concludes, squarely rests on this principle.

It may be relevant here to note that a century before ShatibT, Shihab al-DTn al-QarafT (d. 684/1285) set forth a number of Qur³anic verses and traditions and argued, albeit rather succinctly, that on the basis of the stated textual evi-

Authoritativeness of Sunni Consensus 447 dence and through the application of the method of complete induction it is known beyond any doubt that the consensus of the umma is authoritative and infallible.⁹¹

After an examination of the theories of GhazalT, AmidT, and ShatibT, there remains little doubt that it was primarily the concept of custom and the method of complete induction on the basis of ahadi traditions and equivocal Qur’anic verses (and rational arguments thereof) which were responsible for establishing the authoritativeness of consensus and for saving the Sunni jurists from falling into the trap of circularity. While the great majority of later usulists formulated and presented their arguments each in his own way, the major points which were argued remained essentially the same as those presented by GhazalT, AmidT, and ShatibT. Of those jurists who dealt with this subject, we need only mention IsnawT (d. 772/1370),⁹² Ibn al-Humam (d. 861/1456) and his commentator Ibn AmTr al-Hajj (d. 879/1474),⁹³ and Ansari (d. 1119/1707) and his commentator Ibn ^cAbd al-Shakur (d. 1225/1810).⁹⁴ ’

To illustrate how later jurists dealt with the problem of authoritativeness it must suffice here to discuss briefly the arguments of Ansari and Ibn ^cAbd al-Shakur. They first state that consensus is definitely authoritative and leads to certain knowledge (al~^cilm al-jazim). The basis of its authority is the mutawatir texts whose authenticity is known through custom.⁹⁵ The law of custom would have made it impossible for all the companions of the Prophet to agree on any matter including the reprehension of those who diverged from their beliefs and practices, unless they knew about a conclusive and unambiguous text (nass qati^z) prohibiting dissent and enjoining solidarity. The unqualified acceptance of consensus as a conclusively authoritative principle was brought about by the fact that the number of the companions and the successors (al-tabi^cun) who must have adopted consensus on textual grounds was at least equal to the number required in the mutawatir category.⁹⁶ Such an acceptance was also brought about by concrete mutawatir texts. But before dealing with these texts Ansari and Ibn ^cAbd al-Shakur cite Qur³an IV:115 and 11:143 and conclude that although these verses can be adduced in justification of consensus, they are indecisive and thus cannot yield certainty. The texts which bring about certainty are the traditions of the Prophet to the effect that the community is infallible (ma^csuma). These traditions are mutawatir bil-ma^cna; the number and the fashion in which all these traditions were transmitted is identical to that of the mutawatir. The difference, however, is in the text of the traditions. The theme of the infallibility of the community was expressed in each of these traditions in different words, whereas in a mutawatir lafzi tradition the language is the same in all channels of transmission. The question as to how one can know that these traditions lead to a certain knowledge of the infallibility of the community is answered by Ibn ^cAbd al-Shakur when he says that the common meaning of their sum total (al-qadr al-mushtarak), known by means of an inductive survey of the historical reports and events (mutala^cat al-akhbar wal-waqa³i^c) con­cerning the infallibility of the consensus of the community, leads to such certain knowledge. An inductive survey can yield corroborative evidence, such as the practice of the companions (on the basis of a presumed text) of reprehending those who would diverge from the accepted norm, which further proves the authoritativeness of consensus.⁹⁷

VIII

While formulating their theories the Sunni jurists were acutely aware of the danger of falling into logical contradictions. Most usulists writing after the fourth/tenth century consciously strove to avoid what they called dawr, that is, petitio principii. As we have seen in the theories of Amid! and Ghazal!, and as later emphasized by Ibn ai-Humam, Ansari, and others, dawr was avoided by arguing that traditions, the major scriptural basis which had been advanced in justification of consensus, were authenticated by ^cada rather than by consensus. By the same token, they paid careful attention not to fall into a trap when discussing the doctrine of abrogation (naskh). Insofar as this doctrine pertained to consensus, it was held by all Sunni jurists that consensus cannot be abrogated by the Qur^an and the Sunna nor can it abrogate either of them.⁹⁸ The reason for this is that consensus was recognized to be an operative doctrine only after the death of the Prophet when revelation had ceased (ba^cda inqita^Qi al-wahy). During the life of the Prophet consensus was of course unnecessary, because the Prophet was the ultimate source of revelation and he was not in need of the approval of his companions. But after his death, consensus, claimed the jurists, became indispensable because it was the only means through which certainty about any matter could be reached. Thus, consensus could not abrogate the Qur^Dan and the Sunna because it became operative only after the revelation was completed, and any act of abrogation by consensus would have undoubtedly meant that consensus was concluded on a judgment (hukm) contrary to, or to say the least, different from, that which was decreed in the two divine sources. It is partly for this reason and partly because consensus chronologically occurred after revelation that the Qur³an and the Sunna cannot abrogate consensus.⁹⁹

This doctrine of abrogation not only is in consonance with the arguments adduced to establish the authoritativeness of consensus but is also used by such later scholars as Mulla Khusraw (d. 885/1480) and Izmir! (d. 1166/1752) to further argue, ex post facto, on rational grounds in favor of authoritativeness. They maintain that consensus, which rests on mutawatira bil-ma^cna traditions which in turn are authenticated by custom, is infallible and thus irrepealable, whereas even the strongest Qur°anic verses or traditions can be subject to abrogation by each other. This universally acknowledged doctrine proves that consensus can take precedence over the two sources in the sense that consensus leads to more certainty and is less susceptible to revocation than these two sources. Because of this fact, they reason, the authoritativeness of consensus must be recognized.¹⁰⁰ It is to be noted that Mulla Khusraw and Izmir! are aware of the fact that consensus cannot repeal verses from the Qur°an or hadiths from the Sunna. Their conclusion that consensus takes precedence over the texts must be understood only in relation to the degree of certitude yielded rather than to the competence to abrogate.

IX

From the foregoing discussion a certain pattern of development in the arguments adduced in justification of consensus can be detected. Shafi^cT, viewing consensus as a sanctioning instrument rather than a shar^ci source of law, found it sufficient to argue on grounds of a Quranic verse and two Prophetic tradi­tions. Later jurists, however, elevated consensus to the status of a source of law, a source which has the authority of clothing legal rules with the mantle of certainty. This new role of consensus called for a set of arguments which were different from, and perhaps superior to, the arguments adduced by Shafi^cT. Obviously, it was impossible for jurists to produce mutawatir traditions or decisive Qur³anic verses to this effect. The problem that posed itself now was related to the type of evidence that must be produced in justification of a principle which leads to certain knowledge.

Muslim jurists utilized three concepts in their endeavor to establish the conclusive authoritativeness of consensus. These were the concept of al-tawatur bil-ma^cna, induction, and the originally theological concept of custom. In the early period, the use of one concept did not chronologically coincide with or entail the use of the others. It seems that the notion of custom came to the fore sometime before al-tawatur bil-ma^cna, as is clear from the writings of ^cAbd al-Jabbar and Abu Husayn al-Basri. The former knew of al-tawatur bil-ma^cna as not an altogether well-defined and important concept; one which he con­sidered to be less important than Qur³an IV: 115. That Basri rejected this type of tawatur and instead relied, as his predecessors did, on the Qur³an and custom to prove the authoritativeness of consensus is indicative of the fact that al-tawatur bil-ma^cna had not yet been established as a concept of central importance. This fact is also confirmed by the absence of this concept in the theory of JuwaynT and the young GhazalT.

Later, however, GhazalT argued for authoritativeness on the basis of mutawatir bil-ma^cna traditions which are authenticated by custom. This line of argumenta­tion came to be a standard formulation in the expositions of jurists. Two or three centuries after GhazalT, jurists, such as ShatibT and QarafT, introduced inductive methods of reasoning, thereby giving the argument for consensus a decisive edge. But in GhazalT’s time the role of induction does not seem to have been fully appreciated. This can be seen in GhazalT’s manner of treating the potential evidence in favor of an authoritative consensus. There is nothing in what he said to indicate that his argument by “rational demonstration” was intended to give inductive support to the argument of al-tawatur bil-ma^cna. Nor was the QuPanic evidence used for such purposes. AmidT, on the other hand, was somewhat aware of the importance of the multiplicity of evidence. This is why he expounded virtually every argument and all the pieces of evidence of which he was aware. In this he may be said to have anticipated the articulate theories of QarafT and ShatibT. The latter maintained that all pieces of evidence, whether traditions, verses, or other types of argument, must be taken into account because each additional piece of evidence enhances, by means of inductive support, the argument for authoritativeness. He argued that through the complete enumeration of such evidence certain knowledge can be reached with regard to consensus as an authoritative doctrine.¹⁰¹

With the theory of ShatibT, the Sunni argumentation for authoritative con­sensus seems to have come full circle. Jurists after ShatibT seem to have confined themselves to reformulating and rediscussing what their predecessors had said, with, however, one addition. This is an argument (made by Mulla Khusraw and Izmir!, among others) which proves from the point of view of certitude the superiority of consensus to the Qur°an and the Sunna. It cannot be contended, however, that this argument leads to a petitio principii or that it makes consensus appear absurd for being superior to the sources from which it derives its own authority because this argument is advanced with the knowledge that (1) the authoritativeness of consensus must, as a prerequisite, be established by the texts, and (2) that consensus can by no means abrogate either the Qur³an or the Sunna.

Given that we accept as true premises whatever the mainstream Muslim scholarship designates as propositions—and we have no right to do otherwise— there is nothing in the., theories of jurists after JuwaynT to indicate that their arguments for proving the authoritativeness of consensus were less than con­vincing, whether we view them from the standpoint of logic, law, or theology.

NOTES

Author’s note: This is a revised version of a paper presented at the Eighteenth Annual Meeting of the Middle East Studies Association, San Francisco, November 28-December 1, 1984.

Writers on legal theory (u$ul al-fiqh) disagreed upon the question of who counts in consensus. A commonly held view, however, was that on general issues the community, the umma, can reach a consensus, but on technical and specialized points of law only qualified jurists, the mujtahids, count. See Abd Hdmid al-Ghazali, al-Musta^fa min ^cilm al-u^ul, 2 vols. (Cairo, 1322 H.), vol. I, p. 181 ff.; al-Khatlb al-BaghdadT, al-Faqih wal-mutafaqqih, 2 vols. (Beirut, 1975), vol. I, p. 172.

²The definition of qiyas differs from one jurist to another. Thus one may find that in addition to analogy, qiyas encompasses reductio ad absurdum, a fortiori arguments, deduction, and induction. On the role of deduction in law see Hasan Abdel-Rahman, “La place du syllogisme juridique dans la methode ex6g£tique chez Gazali,” in H. Hubien, ed., Le raisonnement juridique (Brussels, 1971), pp. 185-194. On the use of these arguments see my forthcoming articles, “The Logic of Legal Reasoning in Religious and Non-Religious Cultures: The Case of Islamic Law and Common Law,” Cleveland State Law Review, and “Non-Analogical Arguments in Sunni Juridical Qiyas" Also: Abu H5mid al-GhazaiT, Mi^cyar al-^cilm flfann al-mantiq (Beirut, 1964), pp. 32-33, 113-137, 154, 182 ff.; Abu al-Husayn al-Basri, Kitab al-qiyas al-shar^ci, printed with al-Mu^ctamad fi usul al-fiqh, ed. Muhammad Hamidullah et al., 2 vols. (Damas, 1964-65), vol. II, p. 1031; Muhammad b. ^CA1T al-Shawkani, Irshad al-fuhul ila tahqTq al-haqq min ^cilm al-u^ul (Cairo, 1909), p. 198.

³Abu Ishaq al-ShirazT, al-Luma^c fi usul al-fiqh (Cairo, 1326 H.), pp. 56-58; Abd al-Walid Muhammad Ibn Rushd, Bidayat al-mujtahid, 2 vols. (Cairo, 1974), vol. 1, p. 6. The great majority of jurists upheld the infallibility of the community and, consequently, the conclusive certainty of consensus. See Jamal al-DTn al-IsnawT, Nihayat al-sul ft sharh minhaj al-wuful, a commentary on BaydawTs Minhaj al-wuful ila ^cilm al-usul, 3 vols. (Cairo, 1316 H.), vol. II, pp. 166-167; Muhammad b. ^CA1T al~Shawkani, Irshad al-fuhul, pp. 78-79.

⁴See, for instance, the argument of M. M. Bravmann, The Spiritual Background of Early Islam (Leiden, 1972), pp. 194-198. See further below.

⁵Most noteworthy of the sectarian critics of consensus are the Ràfìda (Imàmiya) of the early ShT^ca, the Isma^ciliyya and the Mu^ctazilf scholar Ibrahim al-Nazzam (d. between 220/835 and 231/845). See Imàm al-Haramayn al-JuwaynT, al-Bur han fl usui al-fiqh, ed. ^cAbd al-^cAzfm DTb, 2 vols. (Cairo, 1979), vol. I, p. 676; Baghdadi, Faqlh, vol. I, p. 154; Sayf al-Dln al-Amidl, al-fykam fl usui al-ahkàm, 3 vols. (Cairo, 1968), vol. I, p. 166. For the criticism of the later Fatimi scholar al-Qadl al-Nu^cman, see his Ikhtilàf u$ùl al-madhàhib, ed. S. T. Lokhandwalla (Simla, 1972), pp. 56-115. Cf. Marie Bernand, “L’igma^c chez ^cAbd al-Gabbàr et l’objection d’an-Nazzàm,” Studia Islamica, 30 (1969), 34 ff. who questions the later accounts attributing to Nazzàm his position against consensus.

⁶Cf. Camille Mansour, L’autorite dans la pensee musulmane (Paris, 1975), pp. 60 ff.

⁷C. Snouck Hurgronje, “The ‘Foundations’ of Islamic Law,” in Selected Works, eds. G.-H. Bousquet and J. Schacht (Leiden, 1975), p. 275.

⁸Ibid., p. 276.

’Hurgronje, “Sur la nature du ‘droit’ Musulmane,” in Selected Works, ibid., p. 227; trans, cited from G. Hourani, “The Basis of Authority of Sunni Consensus,” Studia Islamica, 21 (1964), 51-52.

¹⁰Hurgronje, “The ‘Foundations’,” p. 289.

‘‘idem, “Nature du ‘droit’,” p. 226.

^,2N. J. Coulson, A History of Islamic Law (Edinburgh, 1964), p. 77.

’’Joseph Schacht, The Origins of Muhammadan Jurisprudence (Oxford, 1950), pp. 2, 94, 135. ^l4Hourani, “The Basis of Authority,” pp. 13-60, especially pp. 49-54.

¹⁵Ibid., pp. 14-15: “Thus no assumption is being made at all that consensus is an essential part of Islam, nor even that Traditions of the Prophet are so, though they will be considered seriously.” (Italics are mine.)

^,6lbid„ p. 59.

¹⁷Ibn Hazm, al-Ihkàm fl usui al-ahkàm, 2 vols. (Cairo, 1978), vol. I, p. 12. See also Juwayni, Burhàn, vol. II, p. 152.

^,8Ibn Hazm, Ihkàm, vol. I, pp. 11, 108. On Qur’anic verses and hadith asserting that the Sunna is part of revelation see Baghdadi, Faqlh, vol. I, pp. 90 ff., 94; Ibn Idris al-Shàfi^cT, al-Risàia, ed. M. S. KllanI (Cairo, 1969), par. 236-257; Abu ^CA1I al-FàrisI, Jawàhir al-usùl fl ^cilm hadith al-rasul (Madina, 1973), p. 4.

”lbn Batta, Kitàb al-sharb wal-ibàna ^calà usui al-Sunna wal-diyàna,ed. and trans. H. Laoust (Damascus, 1958), pp. 56 ff. (Arabie text).

²⁰Ibn Hazm, Ihkàm, vol. I, p. 318.

²‘The mutawàtir traditions are transmitted by countless persons who hear or see the Prophet say or do a certain thing and, at each stage of transmission, the tradition is also transmitted by countless persons. The large number of transmitters makes it inconceivable that the witnesses or the transmitters could have agreed on falsifying the report. Because of the authenticity of such a report and the certitude surrounding its transmission, the mutawatir traditions are said to lead to certain knowledge of what they contain. Traditions transmitted by fewer people than those who have witnessed and transmitted the mutawatir traditions are considered ahadi, which do not lead to certain knowledge of the information they convey. Later jurists and theologians distinguished a third type of tradition which they termed mustafld. This type stands in the middle between the mutawatir and t)ie ahadi. See Ibn Khalaf ai-Bàjr, Kitàb al-hudùd fl al-usùl, ed. N. Hammàd (Beirut, 1973), p. 62; ^cAbd al-Qahir al-Baghdàdl, Usui al-dln (Beirut, 1928), p. 12.

²²I. Goldziher, Muslim Studies, ed. S. M. Stern, trans. C. R. Barber and S. M. Stern (New York, 1971), vol. II, p. 25; Bravmann, Spiritual Background, p. 197; Fazlur Rahman, Islamic Methodology in History (Karachi, 1965), pp. 18, 19.

²³Rahman, Islamic Methodology, p. 19.

²⁴The Arabie: “Afa ra^ùhu al-mu^minùna hasanan fahuwa ^cinda Allàhi hasanun wamà ra^ùhu al-muslimùna qablhan fahuwa ^cinda Allàhi qablh," This is also in line with the statement attributed to Ibn ^cAbbàs: “If you hear from me a saying in the name of the Prophet and you find that it does not agree with the book of God or is not liked by people (falam tajidùhu hasanan ^cinda al-nàsi), know that 1 have reported a lie about the Prophet.” Goldziher, Muslim Studies, p. 133; Schacht, Origins, p. 86.

²⁵For the intellectual formulations of the later period (sixth-seventh/twelfth-thirteenth century) see Bernard Weiss, "The Primacy of Revelation in Classical Islamic Legal Theory as Expounded by Sayf al-Din al-Amidi,” Studia Islamica, 59 (1984), 79-109.

VIII

²⁶Norman Calder, “Ikhtilaf and Ijma^c in Shafi^ci’s Risala,” Studia Islamica, 58 (1984), 76-77.

²⁷Ibid., pp. 77-78; Shafi^cI, Risala, par. 1328-1332, especially 1331.

²⁸Cited in Schacht, Origins, p. 92, from Kitab ikhtilaf Malik wal-ShafTi.

²⁹Shafi^cT, Risala, par. 1312.

³⁰Ibid., par. 252-257, 266, 1329, passim.

^3llbid., par. 266.

³²Ibid., par. 1102.

³³lbid., par. 1315.

³⁴Shafi^cT’s reference to this verse is mentioned in the much later works of Ahmad b. al-Husayn al-Bayhaqi (d. 458/1065), Ahkam al-Qur³an, 2 vols. (Beirut, 1975), vol. I, p. 39, also p. 32, and the biographical work of Taj al-DTn al-SubkT (d. 771/1369), fabaqat al-ShafT^ciyya al-Kubra, 6 vols. (Cairo, 1906), vol. II, p. 19. See also 1. Goldziher, Introduction to Islamic Theology and Law, trans. Andras and Ruth Hamori (Princeton, 1981), p. 51. Schacht, Origins, pp. 90, 91, does not seem to have found the verse in any of Shafi^cT’s discussions on the authority of consensus. Further on this point see Abdel-Magid Turki, “L’ijma^c Ummat al-mu^minin entre la doctrine et 1’histoire,” Studia Islamica, 59 (1984), 60 (n. l)-63, 70 (n. 1)—71.

³⁵Shafi^cT, Risala, par. 1319-1320.

^{3sha3izun walqfzun mustafld). These traditions, JuwaynT argues, are merely ahadis, for they are not known to have been transmitted on a wide scale (tawatur) in the community (pp. 22-23). Such ahadi traditions, however they are transmitted, cannot lead to certainty, and thus cannot be used in such serious matters of government as designating a caliph. JuwaynT immediately goes on to say that caliphs are appointed by ikhtiyar (election), and the authoritative basis of such election is the infallible consensus of the community (p. 34). Obviously, when JuwaynT proceeded to establish the authoritativeness of consensus he could not use the principle of tawatur bil-macna, since he had just discarded it as incapable of inducing certainty.}

“JuwaynT, Burhan, vol. I, pp. 680-681; idem, Ghiyath, pp. 36, 40-41.

“JuwaynT, Burhan, vol. I, p. 682.

“ibid.

⁶⁶JuwaynT, Ghiyath, p. 41 at 11. 8-10. In effect, he states the same thing but in a less direct manner in Burhan, vol. I, p. 683 at 11. 8-9.

“JuwaynT, Burhan, vol. I, p. 683 at 11. 8-12.

“GhazalT, al-Mankhul min ta^zllqat al-usul, ed. M. H. Haytu (Damascus, 1970), pp. 305-309, especially p. 307.

⁶⁹GhazalT, Mustasfa, vol. I, p. 175 at 1. 1.

⁷⁰In addition to the tradition about Satan quoted by Shafi^cT and the traditions nos. 1-4 quoted by Jassas (note 38 above) GhazSlT cites the following: (1) “God would not allow my community to agree on an error”; (2) “I have asked God, exalted may he be, not to allow my community to agree on an error, and He answered my request”; (3) “God does not pay attention to those who diverge”; (4) “They [i.e., the community of Muslims] will not be harmed by the disagreement of those who differ with them”; (5) “If a person leaves the community and dies, he will be considered to have died in the days of Jahiliyya.” See GhazalT, Mustasfa, vol. I, p. 175.

⁷¹ Immediate knowledge is that which is forced on the mind through means of sense perception and divine revelation, without inference and reasoning. Such knowledge cannot be doubted. Mediate or discursive knowledge (film nazarl or muktasab), on the other hand, is that which occurs to the mind by reflection and inference. BajT, Hudud, pp. 25-27; BaghdadT, Usulal-dln, p. 8.

⁷²GhazalT, Mustasfa, vol. I, p. 177.

⁷³See also similar arguments in ShTrSzT, Luma^z, p. 47; Ibn Taymiyya, al-Musawwada fl usulal-fiqh (Cairo, 1964), p. 319; Ibn NizSm al-DTn al-Ansari, Fawatih al-rahamut, printed with GhazSlT’s Mustasfa, 2 vols. (Cairo, 1324 H.), vol. II, pp. 119-120. For the opinions of ShirbTnT, Fanari and Taj al-DTn al-SubkT see Jam^z al-jawami^z, 2 vols. (Bombay, 1970), vol. II, p. 130.

⁷⁴GhazSlT, Mustatfa, vol. I, pp. 177-178.

⁷⁵Ibid., p. 176.

⁷⁶Abu Bakr al-SarakhsT, U?ul, ed. Abu al-Wafa al-AfghanT, 2 vols. (Cairo, 1372 H.), vol. I, pp. 296-299.

⁷⁷Al-KhatTb al-BaghdadT, Faqih, vol. I, pp. 167-168.

⁷⁸AmidT, Ihkam, vol. I, pp. 150-162.

⁷⁹For AmidT’s interpretation of the verses see Bernard Weiss, “Al-Amidi on the Basis of Authority of Consensus,” in Donald P. Little, Essays on Islamic Civilization (Leiden, 1976), pp. 344-351.

*°It is of interest to note here that a number of later usulists claimed that AmidT held the authoritativeness of consensus to be zanni rather than qat^ci, i.e., that consensus, when it is reached on a case of law, does not render the said case conclusive and irrevocable. Fakhr al-DTn al-RazT (d. 606/1209) was said to have held the same view. In al-Ihkam there is nothing to substantiate such claims about AmidT. However, it is possible that AmidT held such views in his other works on legal theory or kalam. See the commentary of Sh. Muhammad Ism£^cTl on TahdhTb shark al-isnawi ^cala minhaj al-wufUl ila ^cilm al-utfil, 3 vols. (Cairo, 1976), vol. II, p. 263.

’’See Weiss, “Al-Amidi,” pp. 351 ff. Of the sixteen traditions AmidT quotes, two have not so far been listed here. These are: (1) “Stay with the great majority”; and (2) The Prophet said: “My people will be divided into more than seventy sects, all of which will go to the fire [of Hell] except one.” The Prophet was then asked: “Which sect is it?” and he replied: “My community.” AmidT, Ihkam, vol. I, p. 163.

⁸²AmidT, Ihkam, vol. I, pp. 163, 234.

^8JIbid., pp. 237-238.

⁸⁴Ibid., pp. 164, 238.

⁸⁵TaqT al-DTn Ibn Taymiyya (d. 728/1327) also advanced an argument similar to that of AmidT. See al-Musawwada, pp. 317-319; on al-mutawatir bil-ma^cna traditions, pp. 235, 243.

⁸⁶Abii IshSq al-Shatibi, al-Muwafaqat fi usul al-ahkam, ed. M. D. ^cAbd al-Hamid, 4 vols. (Cairo, 1969), vol. I, p. 10; vol. II, p. 40.

⁸⁷Ibid., vol. I, p. 12.

⁸⁸Ibid.,pp. 14, 16-17.

⁸⁹Ibid.,p. 15.

"ibid., p. 13.

^9,Shihab al-DTn al-QarafT, Sharh tanqth al-fusul, ed. T. Sa^cd (Cairo, 1973), pp. 324-325; on the principle of al-tawatur bil-ma^cna, see pp. 356-357.

⁹²Jamal al-DTn al-IsnawT, Nihayat al-stilfi sharh minhaj al-wusiil, 3 vols. (Cairo, 1317 H.), vol. II, pp. 153-167.

⁹³Kamal al-DTn Ibn al-Humam, al-Tahrir, with the commentary al-TaqrTr wal-TahbTr by Ibn AmTr al-tfSjj, 3 vols. (Cairo, 1317 H.), vol. Ill, pp. 83-86.

⁹⁴Ansari, Fawatih, vol. II, pp. 213-217.

"ibn ^cAbd al-Shakur, Musallam, vol. II, pp. 119-120, 213.

"ibid., p. 214.

⁹⁷Ibid., p. 216.

⁹⁸BaghdadT, Faqih, vol. I, p. 123; PazdawT, U$ul, with the commentary Kashf al-as rar by ^cAla³ al-DTn ^cAbd al-^cAzTz al-Bukhari, 4 vols. (Beirut, 1308 H.), vol. Ill, p. 176; GhazalT, Mustasfa, vol. I, p. 126; Ibn ^CA1T al-FarisT, Jawahir, p. 41; IzmTri, Hashiya ^Qala sharh mirqat al-wu^ul, a commentary on Mirqat al-wutfil by Mulla Khusraw, 2 vols. (Istanbul, 1891), vol. II, p. 181; QarSfT, Sharh, p. 314; SubkT, Jam^z, vol. II, p. 76; Ansari, Fawatih, vol. II, p. 81.

"Baghdad!, Faqih, vol. II, p. 181; QarafT, Sharh, p. 314; SubkT, Jam^c, vol. II, p. 76.

^,00See IzmTri, Elashiya, vol. II, pp. 255-256.

^,0,The great majority of modern Arabic authors on usul al-fiqh reproduce the medieval arguments as thus far outlined. See, e.g., Muhammad al-Khudari, U$ul al-fiqh, 4th ed. (Cairo, 1962), pp. 314— 316. The emphasis on induction as a means of establishing the authoritativeness of consensus also appears in some of these works. See, e.g., ^cAbd al-Fattah HusaynT, al-Ijma^c ma$dar thalith min mafadir al-tashrl^c al-IslamI (Cairo, 1979), p. 280.

ADDENDA

p. 439,11. 4-8: since the writing and publication of the article, several works of usul al-fiqh have been published, including Abu al-Walld al-BajT’s Ihkam al-Fusul fi Ahkam al-Usul (Beirut: Dar al-Gharb al-Islami, 1986), 435­80; Abu Ishaq al-ShTrazT’s Sharh al-Luma, ed. ‘Abd al-Majld TurkT, 2 vols. (Beirut: Dar al-Gharb al-Islami, 1988), II, 665-89; Ahmad b. ‘All Ibn Barhan’s al-Wusul ila al-Usul, ed. ‘Abd al-HamTd Abu Zunayd, 2 vols. (Riyad: Maktabat al-Ma‘arif, 1984), II, 72-6. Generally, the views of these authors seem to be in agreement with those of the mainstream usulists discussed in the article. Fakhr al-DTn al-RazT, however, deviates significantly from them, both in his general view of consensus and in the epistemic status assigned to it as a sanctioning instrument (see his al- Mahsul fi ‘Ibn Usui al-Fiqh, 2 vols. [Beirut: Dar al-Kutub al-‘Ilmiyya, 1988], II, 8-47). In the view of RazI, consensus is not knowable because it is impossible to determine that all the mujtahids throughout the vast stretches of Muslim territories could agree on a particular issue. The only consensus that can be determined is that of the Companions by virtue of the fact that they were relatively few in number. Being indeterminable, consensus loses its significance in RazT’s theory. He argues that none of the Quranic verses or Prophetic traditions (severally or collectively) proves its authoritativeness (hujjiyya) with certainty. And the rational argument does not fare any better. All the available body of textual evidence indicates is that the authoritativeness of consensus is probable (zanni) and this, he maintains, must be sufficient to make of it an instru­ment which sanctions practice (‘amal). In Razi, the issue of circularity necessarily does not arise.

In his Sharh Mukhtasar al-Rawda (ed. ‘Abd Allah al-TurkT, 3 vols. [Beirut: Mu’assasat al-Risala, 1989], III, 14-30), Najm al-DTn al-TufT also insists that neither the Quran nor the Sunna can be brought to bear upon the conclusive authoritativeness of consensus. Yet, unlike RazT, he holds the authoritativeness of consensus to be certain. His argument seems purely rational, based on a reductio ad absurdum. He advances two premises from which he derives a conclusion (III, 25-6): (1) that the epistemological superiority of consensus over any conclusive Quranic or Sunnaic text is established by consensus; and (2) that consensus dictates that nothing other than consensus can override such texts. Thus, if the authoritativeness of consensus were less than certain, these two forms of consensus would be contradictory. But TufT, without explaining why, argues that contradiction here is inconceivable. Again, for an altogether different reason, the issue of circularity does not arise because consensus does not guarantee the very texts by which its own authoritativeness is proven. The argument, however, may be considered circular because the impossibility of contradiction is predicated on the premise that the authoritativeness of consensus is certain, which is the proposition to be proven. Be that as it may, it is noteworthy and significant that Tuff’s legitimation of consensus on rational, non-textual grounds amounts to nothing short of sanctioning the collective voice of the community as an independent source of legislation. For the community, on his view, would be seen as having arrogated to itself the power to determine which legal rules are to be accepted as binding and conclusive and which are not.

pp. 444-6: for a detailed account of AmidT’s exposition of the subject based on both of his works Ihkam and Muntahd, see Bernard Weiss, The Search for God's Law: Islamic Jurisprudence in the Writings of Sayf al-DTn al-Amidl (Salt Lake City: University of Utah Press, 1992), pp. 195 ff., 640. See also my critique of Weiss’s analysis in the International Journal of Middle East Studies, 26, 1 (1994): 152-4.