52 ARGUMENTS FOR AND AGAINST TRADITIONS

veracious transmitter; others say: through two, because Allah demands two trustworthy witnesses; others say: through three, because the Koran says (ix. 122): "a troop of every division of them", and the smallest number to which the term troop can be applied, is three; others say: through four, because Allah demands four witnesses [in the case of adultery]; others: through twelve, because the Koran says (v. 12): "We raised up of them twelve wardens"; others: through twenty, because the Koran says (viii. 65): "If there be of you twenty patient men"; others: through seventy, because the Koran says (vii. 154): "And Moses chose from his people seventy men"' (Ibn Qutaiba, 78 f.). The most commonly held opinion demanded twenty transmitters in each generation.¹

According to Shafi'i, the khabar al-wahid, if related by a trustworthy transmitter, is sufficient to establish the sunna of the Prophet; it cannot be refuted by conclusions drawn from the Koran or from another tradition which is capable of several interpretations; and it does not matter that it is transmitted by only one person (Tr. III, 10). It can be invalidated only by a greater number of traditions to the contrary (Ikh. 165; Ris. 40). Shafi'i devotes three long passages to a detailed argument for the khabar al-wahid.² He even claims a consensus of the scholars, past and present, in its favour;³ but this claim is belied by the strength of the opposition. His only concession is that the khabar al-wahid is weaker than a unanimously recognized sunna and does not produce absolute knowledge, although it must serve as a basis for action.⁴

The later theory on the khabar al-wahid did not go as far as Shafi'i's doctrine.⁵ Among the authors of collections of traditions, Bukhārī (Kitāb akhbār al-āņād) repeats Shafi'ī's essential arguments, Muslim (Bab sihhat al-ihtijāj bil-hadith al-mu'an'an) takes the acceptance of the khabar al-wahid as common ground, Tirmidhi (at the end) includes it in his category of gharib ('strange') traditions, thus setting it apart, and Dāraquțnī (p. 361) accepts it only with certain qualifications.