‘ḥasan,’ with no additional qualifier; for in some reports he says: “ḥasan,” in others: “ṣaḥīḥ,” in others: “gharīb,” in still others: “ḥasan ṣaḥīḥ,” in others: “ḥasan gharīb,” in others: “ṣaḥīḥ gharīb” (1), and in yet others: “ḥasan ṣaḥīḥ gharīb.”
His definition applies only to the first of these, and his own wording makes that clear, for at the close of his work (2) he says: “Whenever we say in our book, ‘a ḥadīth is ḥasan,’ what we intend thereby is that its isnād is sound in our view (3): every (4) report whose narrator is not suspected of fabrication, that is related through more than one route of similar strength, and that is not anomalous—such a report, in our estimation, is a ḥasan ḥadīth.” (5)
From this it is evident that he defined only the category to which he applies the single term “ḥasan.” As for that to which he says, “ḥasan ṣaḥīḥ” or “ḥasan gharīb,”
(1) MSS A, B, C, H, T, Y, and K read: “wa-fi ba‘diha sahih gharib, wa-fi ba‘diha hasan gharib” (in some copies it is sound yet rare, and in others good yet rare), with the two clauses transposed.
(2) MSS D, Z, Y, K, and L read: “awaakhir” (the latter portions).
(3) MS B has the verb hasuna (ḥ with fatḥa, s with ḍamma) and isnaduhu in the nominative; MS K records both the reading with fatḥa on the ḥ and ḍamma on the s, and that with fatḥa on the s and its sukun, while isnaduhu appears in the nominative and the genitive. The accepted wording follows MSS D, H, W, Z, H, T, Y, and L. Al-Qari, in Sharh Sharh al-Nukhbah, p. 307, notes: “The word is recorded (i) with fatḥa on the ḥ and s, as a participial adjective, with the final n tanwinned; (ii) with ḍamma on the s and fatḥa on the n, as a past-tense verb—in both cases isnaduhu is nominative because it is the doer; or (iii) with ḍamma on the ḥ and sukun on the s, as a verbal noun in the accusative, annexed to isnadihi.”
(4) MSS B, Y, and K read: “fa-kull” (so each); MS W: “wa-kull” (and each); the margin of W: “idh” (since).
(5) See: al-‘Ilal, printed at the end of al-Jami‘, 6/481.