As for transmitting a report by meaning rather than by its exact wording, the disagreement on this question is well known (1), and the majority likewise regard it as permissible (2). Among their strongest arguments (3) is the unanimous agreement that it is permissible to explain the Sacred Law to non-Arabs in their own language, provided the explainer is proficient in that tongue; if substitution in an entirely different language is allowed, then allowing it within Arabic itself is even more appropriate (4).
It has been said (5) that such substitution is permissible (6) only with individual words, not with full phrases.
It has also been said (7) that it is permissible (8) only for someone who keeps the original wording present in his mind (9), enabling him to modify it deliberately.
And it has been said (10) that it is permissible (11) for a person who had memorized the hadith, then forgot its precise wording while its meaning still remains clearly (12) imprinted in his mind; in order to obtain the legal ruling from it, he may relate it by meaning (13)—in contrast to one who still retains it.
(1) MS A reads “yatamayyazu” (is distinguished), whereas MS Z has “mushtahir” (well-known).
(2) The word “aydan” (also) is missing from MS T. Ibn al-Salah states in his Muqaddima (p. 214): “That is exactly what the conduct of the Companions and the earliest salaf attests to; they often transmitted one and the same meaning on a single matter in different wordings, because their reliance was on the meaning rather than the wording.” See also al-Khatib, al-Kifaya, p. 198.
(3) MS K gives “hujjatuhum” (their proof), while a marginal note in another copy has “hujajuhum” (their proofs).
(4) Al-Khatib alludes to this argument in al-Kifaya, p. 201.
(5) This is the position preferred by al-Khatib in al-Kifaya, where he adds an additional condition (p. 198).
(6) MSS H and W read “tajuz” (is permissible); in Z, H, and M the word is left unpointed.
(7) See al-Khatib, al-Kifaya, p. 199.
(8) MSS H and W again read “tajuz”; in A, Z, and H it is unpointed.
(9) MS D reads “istahdara” (he brought to mind).
(10) This is also the opinion adopted by al-Mawardi; al-Hawi al-Kabir 16:97.
(11) MSS H, W, and Z read “tajuz”; in A, H, and M the word is unpointed.
(12) I.e., engraved. See al-Qari, Sharh Sharh al-Nukhba, p. 499.
(13) MS K reads “limanfaʿa” (for a benefit).