Thus, if, for example (1), a report happens to be found in Muslim’s collection and is well-known yet still below the level of tawātur, but some contextual indicator accompanies it so that it conveys certain knowledge, it is given priority over a report that Bukhārī records (2) when the latter stands alone—absolutely (3).
Likewise, should a ḥadīth that neither of them included come through a chain that has been described as the soundest of all isnāds—such as Mālik, from Nāfiʿ, from Ibn ʿUmar—then it takes precedence over what either of the two compilers narrates in isolation, especially if there is in that isnād someone about whom there is criticism (4).

[al-ḥasan li-dhātihi and al-ḥasan li-ghayrihi]
“When precision is slightly reduced,” that is, it has diminished—the Arabs say, “khaffa al-qawmu khufūfan,” meaning “the people became few” (5)—what is meant is that the remaining conditions previously laid down for a ṣaḥīḥ report are still met. In that case the narration is “ḥasan in and of itself,” not because of some external factor. Its soundness results from corroboration, as in the report of a narrator whose inner state is unknown (6) when it is supported by multiple chains. By requiring the rest of the qualifying attributes, the weak category is thereby excluded.