as a defect, and favored the version he narrated from Abū ‘Ubayda, from his father, although Abū ‘Ubayda had not heard from his father. Al-Bukhārī, however, judged it authentic through another route, because Abū Ishāq would have the same hadith from several narrators, relating it sometimes from one and sometimes from another. Likewise, al-Zuhrī would sometimes narrate a hadith from Sa‘īd ibn al-Musayyib, sometimes from Abū Salama, and sometimes from both together. Someone unfamiliar with this practice might suppose that narrating it sometimes from one and sometimes from the other was an error, although both versions are authentic. This subject would require a lengthy discussion.

As for the scholars before al-Tirmidhī, this threefold classification is not known to have been used by them. Rather, they divided hadiths into authentic and weak. For them, weak hadiths were of two kinds:

Those whose weakness did not preclude acting upon them, resembling what al-Tirmidhī termed good; and those whose weakness required their rejection—the very weak reports. This is comparable to a sick person’s illness: it may be life-threatening, so that his gifts are charged against one third of his estate, or it may not be life-threatening. This distinction is found in the statements of Imam Ahmad and others.

Hence they say, “This is somewhat weak,” or, “There is weakness in it.” Such descriptions occur in their assessments of hadith.