AND PERSONAL OPINION 123

are two separate aspects: the tradition is always followed unquestioningly, whether it agrees with qiyās or not; if it does not agree, ittibā' becomes the opposite of qiyās; there are also cases where one set of circumstances falls under both rules.'

Shāfi'ī gives the following example. The Prophet decided that the buyer can either keep a muṣarrāt, that is, an animal which the seller has not milked for some time before the sale so as to make its yield of milk appear greater, or return it together with one sā' of dates if he has milked it; he also gave the ruling that 'profit follows responsibility' (al-kharāj bil-ḍamān).¹ In cases to which this rule applies there is no [ideal] part of the price corresponding to the profit [which accrues after the sale in the possession of the buyer], and this rule is extended by qiyās to all parallel cases. In the case of the muṣarrāt, the decision of the Prophet is followed and not extended by qiyās, the Prophet having fixed the unknown quantity of milk in the animal, which has an [ideal] part of the price corresponding to it. Now if someone buys an animal which turns out to be a musarrāt and decides to keep it nevertheless, but after a month finds another hidden fault for which he decides to return it, he can do so, and the milk which has accrued to him during the month belongs to him according to the rule of al-kharāj bil-ḍamān; but he must also give one sā' of dates for the milk which was in the muṣarrāt [at the time of sale]. This detail is decided according to the tradition, and the ownership of the milk which has accrued during the month by analogy with the general rule.

Qiyās is, however, used as a criterion for choosing between conflicting traditions.² Moreover, in Tr. III, 23, Shāfi'ī confirms by analogical reasoning his rejection of a tradition, although he does not call his argument qiyās but 'the decisive proof in our opinion' (al-ḥujja al-thābita 'indanā). These are survivals of the earlier use of systematic reasoning for criticizing traditions.³

The consensus of the Muslims decides which qiyās is right and which is wrong (Ris. 72). The consensus supersedes an analogy based on a tradition from the Prophet (Tr. III, 129).⁴ But qiyās supersedes the 'practice' which may have been introduced only by some Successor (Tr. VIII, 14).

Shāfi'ī's most important methodical rule regarding the use of