his entire body of narrations must be accepted; but if he is not reliable, we must (1) discard his reports altogether.’

Such is the method of those with scant learning. It is a faulty approach (2) whose unsoundness is unanimously recognised by the ḥadīth scholars. They cite as proof from a man’s ḥadīth only what others corroborate (3) and what is supported by additional chains and wordings, while they set aside that very man’s report when he narrates something that contradicts the scholars, or when he alone transmits (4) a report that no one else affirms. A mistake in one place does not entail mistakes everywhere; accuracy in part—or even in (5) most—of a narration does not confer immunity from error in the rest, especially when numerous mistakes are already known from such a transmitter. If, after that, he relates something that conflicts with the people and no one corroborates him—whether one merely suspects error or is certain of it—his narration is put aside.

Section

Here (6) two kinds of error befall anyone whose critical discernment and taste (7) fall short of the leading authorities in this field. We shall point them out, for guarding against them is of immense benefit.