Previous · All chapters · Next

B9 · باب الكلام في قبول أخبار الآحاد والعمل بها · Single reports: binding for action, not certainty; where they bind; which narrators; how they are filtered (v1/321–v2/26)

1. B9a.01 · حجية خبر الواحد · A single report from an upright narrator binds action in religion but does not yield certain knowledge (v1/321–322)
2. B9a.02 · أدلة حجيته · The proofs: the duty to convey implies the duty to accept; the "group" that goes to learn can be one man; the Prophet's envoys; eating what one is given (v1/322–329)
3. B9a.03 · خبر الواحد لا يوجب العلم · Those who say single reports yield knowledge confuse reassurance with certainty; creedal reports bind the heart, a trial like the trial of action or weightier (v1/329–331)
4. B9a.04 · اشتراط العدد · Number is not required in reports; the Companions' requests for a second witness were caution; ⚑ whether ʿUmar would have accepted Abū Mūsā alone (v1/331–333)
5. B9a.05 · أقسام ما يكون خبر الواحد فيه حجة · Four domains: religious rulings (⚑ ḥudūd), binding claims between people, non-binding dealings, and the partly binding (v1/333–338)
6. B9b.01 · أقسام الرواة · Narrators are known or unknown; the known are faqīh or merely reliable (v1/338)
7. B9b.02 · خبر الفقيه مقدم على القياس · A faqīh Companion's report beats qiyās, agreeing or not (v1/338–339)
8. B9b.03 · رواية المعروف غير الفقيه إذا خالفت القياس · A reliable non-faqīh's report yields to sound qiyās only if the umma did not accept it and raʾy has no room (v1/339–342)
9. B9b.04 · رواية المجهول · The "unknown" Companion-narrator: five cases, decided by how the early jurists treated him (v1/342–345)
10. B9c.01 · شرائط الراوي · Four conditions: reason, accuracy, uprightness and Islam; an unbeliever's report is refused for an added motive to lie, not for unbelief itself (v1/345–346)
11. B9c.02 · العقل · Reason is a light in the breast; the law fixes full reason at puberty without defect; a child may hear and later relate (v1/346–348)
12. B9c.03 · الضبط · Accuracy is outer (the wording and its sense in language) and inner (its legal meaning, i.e. fiqh); hence the faqīh's report prevails; Abū Ḥanīfa narrated little for this reason (v1/348–350)
13. B9c.04 · العدالة · Uprightness is shunning what one believes forbidden: no major sins and no persistence in minor ones; the heretic who shuns sins is accepted; the unknown of the first three generations is upright, his report proof only as settled in B9b.04 (v1/350–352)
14. B9c.05 · الإسلام · Islam is outer (birth and upbringing) and inner (assent); a general description of the faith suffices, against detailed interrogation (v1/352–353)
15. B9c.06 · العبد والمرأة والأعمى والمحدود في القذف · Slaves, women and the blind narrate like others; ⚑ the one punished for qadhf, after repentance: ẓāhir al-madhhab vs al-Ḥasan's report (v1/353–355)
16. B9c.07 · النقل بالمعنى · Keeping the wording is better; relating by meaning is allowed after good accuracy, by type of text; the Prophet's "comprehensive words" must be kept verbatim, on Sarakhsī's view (v1/355–357)
17. B9c.08 · الضبط بالكتابة · Writing as a reminder is fine; ⚑ relying on handwriting without remembering: Abū Ḥanīfa never; Abū Yūsuf in the register and ḥadīth, and in a deed only when it is safe in the witness's hands; Muḥammad in all three; Sarakhsī adopts the concession (v1/357–359)
18. B9d.S1 · الانقطاع صورة: مرسل الصحابي · Break in form: the Companions' mursal is proof by agreement (v1/359)
19. B9d.S2 · مراسيل القرون الثلاثة · The mursal of the second and third generations is proof, against al-Shāfiʿī (on the preferred view, unless the narrator is known to relate from the non-upright, B9d.S3); ʿĪsā b. Abān: even stronger than a musnad (v1/360–363)
20. B9d.S3 · مراسيل من بعد القرون الثلاثة · ⚑ After the three generations: al-Karkhī accepts all; ʿĪsā b. Abān accepts the famous carriers of knowledge; al-Rāzī's view, the soundest: only from one known to relate solely from the upright (v1/363–364)
21. B9d.01 · الانقطاع معنى · Break in meaning: a report is cut off by a contrary proof or by a defective narrator (v1/364, 369–370)
22. B9d.02 · مخالفة الكتاب · Filter 1: a single report contradicting the Kitāb is not acted on (v1/364–366)
23. B9d.03 · مخالفة السنة المشهورة · Filter 2: a gharīb report contradicting the well-known Sunna is set aside (v1/366–368)
24. B9d.04 · عموم البلوى · Filter 3: a lone report on a matter of general need is suspect; the act's legal grade (wujūb) may still rest on a single report (v1/368–369)
25. B9d.05 · إعراض الصحابة · Filter 4: a report the Companions ignored while disputing the very question is suspect (v1/369–370)
26. B9d.06 · المستور والفاسق · The unexamined narrator's report is not relied on "in our time"; the sinner's report is not proof in religion, though accepted in dealings of necessity (v1/370–371)
27. B9d.07 · الكافر والصبي والمعتوه · The unbeliever's report counts not at all in religion; the child and imbecile are like the unbeliever, not like the sinner (v1/371–373)
28. B9d.08 · صاحب الهوى · The heretic's testimony is accepted but his narration in religion is not, since partisanship may drive him to corrupt the way of truth (v1/373–374)
29. B9e.01 · أقسام الأخبار · Four kinds of report: certainly true, certainly false, equally balanced, and weighted; the upright narrator's report is the weighted-true kind (v1/374–375)
30. B9e.02 · السماع · Hearing: the teacher reading or the student reading back are true hearing; writing and messenger are near it; permission (ijāza) is valid only when the content is known to the recipient (v1/375–379)
31. B9e.03 · الحفظ والأداء والتدليس · Memory and delivery have strict and concessionary forms; dropping a link is not blameworthy tadlīs unless done to fake a higher chain (v1/379–380)
32. B9e.04 · قول الصحابي أمرنا والسنة كذا · A Companion's "we were commanded" or "the sunna is" does not by itself mean the Prophet, against al-Shāfiʿī's old view (v1/380–381)
33. B9f.01 · إنكار الراوي · ⚑ When the original narrator denies or forgets a report: al-Shāfiʿī acts on it; "our scholars" did not; within the school, Abū Ḥanīfa and Abū Yūsuf against Muḥammad (v2/3–5)
34. B9f.02 · مخالفة الراوي لما روى · If the narrator rules or acts against his report after relating it, the report is no longer proof, being read as his knowledge of abrogation; his interpreting it does not bind others (v2/5–7)
35. B9f.03 · الصحابة والأئمة يخالفون الحديث · A leading Companion's contrary practice voids a report he could not have missed; not one he may have missed; ḥadīth critics' vague criticism is not jarḥ (v2/7–11)
36. B9g.01 · حقيقة التعارض · Revealed proofs do not truly conflict; conflict is in our knowledge (unknown dates); it needs two equal proofs, one time and one subject, each able to abrogate the other (v2/12–13)
37. B9g.02 · حكم التعارض · Conflicting verses of unknown date send you to the Sunna; undated conflicting sunnas to Companion opinion or qiyās; between two qiyās the mujtahid chooses only when neither can be preferred, one being right and one wrong (v2/13–17)
38. B9g.03 · المخلص من التعارض · Resolve apparent conflict in order: by the proofs themselves, by the rulings, by circumstances, by an explicit date, then by an inferred date (v2/18–21)
39. B9g.04 · المثبت والنافي · ⚑ Affirming vs denying reports: al-Karkhī prefers the affirming, ʿĪsā b. Abān sees a conflict; Sarakhsī: a denial based on evidence conflicts, a denial from mere absence of knowledge does not (v2/21–24)
40. B9g.05 · الترجيح بالعدد والزيادة · No preference by number of narrators (Abū Ḥanīfa and Abū Yūsuf, against Muḥammad); an addition from the same narrator is taken; two narrators give two reports; preference is by proof, not by the rank of men (v2/24–26)

B9a.01 · حجية خبر الواحد · A single report from an upright narrator binds action in religion but does not yield certain knowledge (v1/321–322)

• Position: «خبر الْوَاحِد الْعدْل حجَّة للْعَمَل بِهِ فِي أَمر الدّين وَلَا يثبت بِهِ علم الْيَقِين» — "The single report of an upright narrator is a proof for acting on in religion, but certain knowledge is not established by it." This is the view of the jurists of the major cities. (v1/321)
• Who: (1) "some whose view is not counted": no proof in religion at all; (2) some ḥadīth scholars: it yields certain knowledge; among them, some require two narrators, some four. (v1/321)
• The deniers' case: Q 17:36, 4:171, 43:86, 10:36 forbid following what is not known; a report that may be false is not "the truth". Dealings (muʿāmalāt) are an exception for human need; God needs no such concession. Testimony is accepted only by text and for people's rights. Accepting a fallible reporter without a miracle would rank him above the Prophet, whose word was first accepted through miracles. And the roots of religion need certainty. (v1/321–322)

Back to the list

B9a.02 · أدلة حجيته · The proofs: the duty to convey implies the duty to accept; the "group" that goes to learn can be one man; the Prophet's envoys; eating what one is given (v1/322–329)

• Kitāb: "those who conceal what We sent down" (Q 2:159) and the covenant "to make it clear to the people" (Q 3:187) bind each one to convey, and that implies that the hearer must accept, or the command would be pointless. The sinner is excluded by "verify" (Q 49:6). (v1/322)
• "Why should not a group (ṭāʾifa) from every party go out to learn… that they may warn their people" (Q 9:122). A party (firqa) is three or more, so the group is one or two. The early authorities differ (one, two, three, ten), but none said more than ten, and ten do not remove the possibility of lying. The sounder is one (Muḥammad b. Kaʿb; Qatāda on Q 24:2; Q 49:9 "two groups" were two men). "That they may beware" shows the hearer must accept. (v1/322–324)
  • Objection → Reply: "all the groups together reach tawātur." Reply: a plural attributed to a group matches one to one, and "return to them" fits only one who went out and came back. (v1/323–324) "The warner must warn, but hearers need not accept until it spreads, like one witness." Reply: a lone witness need not testify at all; the warner must, so his word binds, as the Prophet's did. (v1/324)
• Sunna and practice: the Prophet was sent to all people but reached most by envoys and letters (to the kings; governors to Yemen and Bahrain); secluded women learned from their husbands. Objection: the Prophet's envoys were specially protected, a mark of honour for him. Reply: such a special trait would have been transmitted as an honour to them and their descendants, as were Ḥanẓala's washing by the angels and Jaʿfar's two wings. (v1/325) "Let the one present convey to the one absent"; "may God brighten a man who heard from us…": whoever hears is his deputy in conveying. He ate what Salmān and Barīra brought, relying on single reports for what is lawful. Objection: revelation told him the food was lawful. Reply: «مَا كَانَ ينْتَظر الْوَحْي عِنْد كل أَكلَة» (v1/326) — "he did not wait for revelation at every meal": he took a bite of the roasted sheep and asked about it only when it would not go down, and he took a bite of the poisoned sheep. (v1/324–326)
• Testimony in God's rights (ḥadd for drinking, theft, zinā) binds the qāḍī despite remaining doubt. Dealings also carry God's rights (the water is pure; this is a gift to you). (v1/326)
• Knowledge by the apparent: «فَإِنَّهُ عندنَا عمل بِعلم هُوَ ثَابت من حَيْثُ الظَّاهِر وَلكنه غير مَقْطُوع بِهِ» (v1/326) — "in our view it is action on knowledge established on the apparent, though not certain": "we testify only to what we know" (Q 12:81); "if you know them to be believers" (Q 60:10). Ignorance is acting on a sinner's report (Q 49:6). (v1/326–327)
• A fortiori: qiyās is acted on though its doubt is greater, and those who reject single reports fall back on qiyās. The need of later generations is like the need of those around the Prophet, who himself judged by testimony and oaths ("I am only a human… I judge by what I hear") though revelation was open to him. (v1/327–328) ʿĪsā b. Abān gathered the reports on this; Sarakhsī avoids them, since opponents would call it arguing single reports by single reports. A layman must follow a muftī's fatwā though it may err. The claim that this ranks a fallible reporter above the Prophet is "a clear error": miracles were needed to establish certain knowledge of his prophethood, which no single reporter's word gives. (v1/328)

Back to the list

B9a.03 · خبر الواحد لا يوجب العلم · Those who say single reports yield knowledge confuse reassurance with certainty; creedal reports bind the heart, a trial like the trial of action or weightier (v1/329–331)

• Who: some say it yields knowledge: "then inform (aʿlimhum) them" (the ḥadīth of Muʿādh); action needs knowledge (Q 17:36); "lest you harm people in ignorance" (Q 49:6) implies knowledge from the upright; and creedal matters (the grave's torment, Munkar and Nakīr, seeing God) rest on single reports. (v1/329)
• Reply: «وَلَكنَّا نقُول هَذَا الْقَائِل كَأَنَّهُ خَفِي عَلَيْهِ الْفرق بَين سُكُون النَّفس وطمأنينة الْقلب وَبَين علم الْيَقِين» — "We say: it is as if the difference between calm and reassurance of the heart, and certain knowledge, were hidden from this speaker." The possibility of lying remains; reassurance is knowledge in its apparent sense, and that is what "inform them" means. (v1/329) The grave-torment reports, some mashhūr and some single, bind the heart to believe, and binding the heart is like binding action, or weightier; it is not the same as knowing (Q 27:14; 2:146). (v1/329–330)
• Who: al-Naẓẓām: with supporting signs (the washing of the dead, an old woman crying out) a single report yields immediate knowledge. Sarakhsī: «وَهَذَا قَول بَاطِل» (v1/330) — "this is a false view": immediate knowledge does not vary between people; it would abolish evidence, oaths and liʿān; and even knowledge of prophethood is reasoned. A second report ("the owner is hiding from the ruler") makes one doubt again. Objection: "the deniers of the Prophet denied after knowing, so a single report gave them immediate knowledge." Reply: those were obstinate people who knew him from their scripture (Q 2:146), not all unbelievers. (v1/330–331)

Back to the list

B9a.04 · اشتراط العدد · Number is not required in reports; the Companions' requests for a second witness were caution; ⚑ whether ʿUmar would have accepted Abū Mūsā alone (v1/331–333)

• Who: some require the number of witnesses: Abū Bakr asked al-Mughīra for a second witness on the grandmother's share (Muḥammad b. Maslama confirmed); ʿUmar asked Abū Mūsā on asking permission (Abū Saʿīd confirmed); ʿUmar on Fāṭima bt. Qays ("we do not leave our Lord's Book and our Prophet's Sunna for a woman's word"); ʿAlī on Abū Sinān ("a Bedouin who urinates on his heels"). Some went to four. (v1/331)
• Reply: "you are the best community… enjoining the right" (Q 3:110) describes every believer, so each one's word in religion binds. Their own reports would have to reach them two-from-two to count. (v1/331–332)
  • Abū Bakr doubted something, or the event was public. ʿUmar sought confirmation because the matter was of general need, and would have accepted Abū Mūsā anyway. ⚑ Some later shaykhs say he would not, since narrators were many then. «وَلَكِن الْأَصَح هُوَ الأول وَعَلِيهِ نَص مُحَمَّد» — "but the sounder is the first, and Muḥammad stated it" in al-Istiḥsān: ʿUmar accepted al-Ḍaḥḥāk on a widow's share of blood-money, and ʿAbd al-Raḥmān b. ʿAwf on the plague and on jizya from Zoroastrians, without a second witness. (v1/332)
  • Fāṭima's report was rejected for contradicting the Kitāb (Q 65:6) and Sunna; ʿAlī had his own practice of not accepting Bedouins and of making narrators swear (except Abū Bakr), while Ibn Masʿūd accepted Abū Sinān gladly. (v1/332)
• The number in testimony is known by text, with no reason we can grasp, since two upright men do not add knowledge. Reports differ: no testimony wording, no court, men and women equal, and one ruling for all subjects, while testimony varies (women's testimony; Khuzayma counted as two). (v1/332–333)

Back to the list

B9a.05 · أقسام ما يكون خبر الواحد فيه حجة · Four domains: religious rulings (⚑ ḥudūd), binding claims between people, non-binding dealings, and the partly binding (v1/333–338)

• 1. Rulings of religion that admit abrogation: in what is not averted by doubt (acts of worship), an upright single report binds action with no number and no fixed wording, «لِأَن الْمُعْتَبر فِيهِ رُجْحَان جَانب الصدْق لَا انْتِفَاء احْتِمَال الْكَذِب» — "because what counts here is that truth outweighs, not that the possibility of lying is removed." ʿAlī's making narrators swear was caution, not a condition. (v1/333)
  • ⚑ In what is averted by doubt (ḥudūd): Abū Yūsuf in al-Amālī, chosen by al-Jaṣṣāṣ: a single report is a proof, as evidence and dalāla are used there despite doubt. al-Karkhī: it is not, since its link to the Prophet holds a doubt, like qiyās; testimony in ḥudūd rests on a specific text (Q 4:15). (v1/333–334)
• 2. People's rights with pure binding force: only with number, the wording of testimony, and the witness's capacity, since these settle real disputes between claim and denial, where forgery is common. (v1/334–335) The Ramaḍān crescent under cloud is domain 1 (God's right: fasting); the ʿĪd crescent is domain 2 (people's benefit). Reports of prohibition by nursing in a marriage or ownership are domain 2, since they end ownership; purity of water and lawfulness of food are domain 1. ⚑ Endorsement of witnesses (tazkiya): Abū Ḥanīfa and Abū Yūsuf, domain 1 (no number); Muḥammad, like domain 2 (number). (v1/335)
• 3. Dealings without binding force (agency, muḍāraba, permission to a slave to trade, buying from agents and owners): any discerning reporter counts, upright or not, child or adult, unbeliever or Muslim, if the hearer believes him. The Prophet accepted gifts from anyone and bought from unbelievers; markets never demanded uprightness. Two reasons: real necessity (one rarely has an upright messenger), and these reports bind no one. (v1/335–336) al-Istiḥsān cases: "this was mine, usurped, and I took it back" signals dispute (not relied on); "he repented and returned it" signals peace (relied on). Likewise a later bar to a wife (nursing) is accepted from a single reporter, but a bar present at the contract is not; a wife may rely on a report that her absent husband divorced her, not that the contract was void. (v1/336)
• 4. Partly binding (restricting a trading slave, dismissing an agent; also a virgin told of her marriage who stays silent, the pre-emptor told of a sale, the master told his slave committed a crime): ⚑ Abū Ḥanīfa requires one of the two conditions of testimony, number or uprightness, since the report binds in one way and a sinner's report cannot bind (Q 49:6), unlike the owner's messenger. Abū Yūsuf and Muḥammad: any discerning reporter, as with permission. Two sinners: some say it counts (one condition met), some say not. The text's wording "two men or one upright man" may mean "two upright men or one"; al-Istiḥsān prefers the upright reporter over the non-upright, and two upright over one, on the purity of water. (v1/337–338)
• Who: ⚑ someone who became Muslim in enemy territory and was unaware of his duties: our shaykhs say it is disputed: on Abū Ḥanīfa's view a sinner's report does not oblige him (the Ṣāḥibān: it does) to make them up. Sarakhsī: «وَالأَصَح عِنْدِي أَنه يعْتَبر الْخَبَر هُنَا فِي إِيجَاب الْقَضَاء عِنْدهم جَمِيعًا» — "the sounder, to me, is that the report counts here for requiring make-up, according to all of them", since the reporter conveys for the Prophet ("let the present convey"), like the owner's messenger. (v1/338)

Back to the list

B9b.01 · أقسام الرواة · Narrators are known or unknown; the known are faqīh or merely reliable (v1/338)

• Position: «الروَاة قِسْمَانِ مَعْرُوف ومجهول» — "Narrators are of two kinds: known and unknown." (v1/338)
  • «فالمعروف نَوْعَانِ من كَانَ مَعْرُوفا بالفقه والرأي فِي الِاجْتِهَاد وَمن كَانَ مَعْرُوفا بِالْعَدَالَةِ وَحسن الضَّبْط وَالْحِفْظ وَلكنه قَلِيل الْفِقْه» — "The known are two types: one known for fiqh and ijtihād-level judgement, and one known for probity, sound retention and memory but of little fiqh." (v1/338)
  • Examples Sarakhsī gives of the first type: the Rightly-Guided Caliphs, the ʿAbādila, Zayd b. Thābit, Muʿādh, Abū Mūsā, ʿĀʾisha. Of the second: Abū Hurayra and Anas. (v1/338–339)
• [Editorial] Debate note: "Of little fiqh" is a comparative category, set against the jurist-Companions of the first type. Sarakhsī defends Abū Hurayra's probity and memory explicitly (B9b.03); do not present the category as a charge of ignorance.

Back to the list

B9b.02 · خبر الفقيه مقدم على القياس · A faqīh Companion's report beats qiyās, agreeing or not (v1/338–339)

• Position: The report binds action; its knowledge-status is probable, not certain. «للْعلم الَّذِي هُوَ غَالب الرَّأْي ويبتنى عَلَيْهِ وجوب الْعَمَل» — "[It yields] knowledge in the sense of preponderant opinion, on which the obligation to act is built." (v1/339)
  • «وَإِن كَانَ مُخَالفا للْقِيَاس يتْرك الْقيَاس وَيعْمل بالْخبر» — "If it contradicts qiyās, qiyās is abandoned and the report is acted on." (v1/339)
• Who: Opponent: Mālik, as Sarakhsī reports him. «وَكَانَ مَالك بن أنس يَقُول يقدم الْقيَاس على خبر الْوَاحِد فِي الْعَمَل بِهِ» — "Mālik held that qiyās is put before the single report in practice." (v1/339)
  • Mālik's reason, as reported, runs through ijmāʿ: qiyās is a proof by the consensus of the early authorities (al-salaf) among the Companions and by the proof of the Kitāb and Sunna, and «وَالْإِجْمَاع أقوى من خبر الْوَاحِد فَكَذَلِك مَا يكون ثَابتا بِالْإِجْمَاع» — "consensus is stronger than the single report, and so is what is established by consensus." (v1/339)
• Proofs
  1. The early generations routinely left qiyās for a report, so much so that such rulings were called «معدولا بِهِ عَن الْقيَاس» — "departed from qiyās." (v1/339) Cases cited: ʿUmar and Ḥamal b. Mālik on the ghurra for the foetus; Ibn ʿUmar leaving mukhābara on Rāfiʿ b. Khadīj's report. (v1/339)
  2. The doubt in a report lies in its transmission, an accident; the doubt in qiyās lies in its root (is this the operative attribute?). «وَمَا يكون الشُّبْهَة فِي أَصله دون مَا تكون الشُّبْهَة فِي طَرِيقه بعد التيقن بِأَصْلِهِ» — "What is doubtful at its root ranks below what is doubtful only in its route, once its root is certain." (v1/339)
  3. The report states; the attribute (waṣf) is silent. Hearing beats opinion, and the strong is not left for the weak. (v1/339)
• [Editorial] Debate note: Keep the epistemic status exact: Sarakhsī says the report yields ghālib al-raʾy and obliges action. He does not claim it yields certainty (see B9a on the ʿilm debate).

Back to the list

B9b.03 · رواية المعروف غير الفقيه إذا خالفت القياس · A reliable non-faqīh's report yields to sound qiyās only if the umma did not accept it and raʾy has no room (v1/339–342)

• Position: «مَا وَافق الْقيَاس من رِوَايَته فَهُوَ مَعْمُول بِهِ وَمَا خَالف الْقيَاس فَإِن تَلَقَّتْهُ الْأمة بِالْقبُولِ فَهُوَ مَعْمُول بِهِ وَإِلَّا فَالْقِيَاس الصَّحِيح شرعا مقدم على رِوَايَته فِيمَا ينسد بَاب الرَّأْي فِيهِ» — "What in his narration agrees with qiyās is acted on. What contradicts qiyās is acted on if the umma received it with acceptance; otherwise sound, legally-grounded qiyās is put before his narration in matters where the door of raʾy is closed." (v1/341)
  • Conditions stated in the text before the report is set aside:
    1. It contradicts sound qiyās in every respect. Such a qiyās rests on the Kitāb, the Sunna and ijmāʿ, so the report is then "in meaning" against them: «فَمَا خَالف الْقيَاس الصَّحِيح من كل وَجه فَهُوَ فِي الْمَعْنى مُخَالف للْكتاب وَالسّنة الْمَشْهُورَة وَالْإِجْمَاع» — "What contradicts sound qiyās in every respect is, in meaning, contrary to the Kitāb, the well-known Sunna and ijmāʿ." (v1/341)
    2. The umma has not received it with acceptance (talaqqī bi'l-qabūl). (v1/341)
    3. "The door of raʾy is closed." Sarakhsī's own gloss ties this to the qiyās condition: «إِذا انسد بَاب الرَّأْي فِيمَا رُوِيَ وتحققت الضَّرُورَة بِكَوْنِهِ مُخَالفا للْقِيَاس الصَّحِيح» — "when the door of raʾy is closed regarding what was narrated, and necessity is established by its contradicting sound qiyās." (v1/341)
       • [Editorial] The text leaves open whether condition 3 is a separate test (no reasoned way to accommodate the report) or a restatement of condition 1. Do not settle it silently in debate; quote the gloss.
  • The reason is paraphrase, not probity: Companions commonly narrated by meaning, and a narrator conveys only as much as he understood of speech that was "jawāmiʿ al-kalim". Such a shortfall is not hard to detect when set against the Prophet's meaning. (v1/341)
  • Explicit disclaimer: «وَلَعَلَّ ظَانّا يظنّ أَن فِي مقالتنا ازدراء بِهِ ومعاذ الله من ذَلِك» — "Someone may think our position belittles him; God forbid." He is "foremost in probity, memory and precision." (v1/341)
  • Contrast: this shortfall is not supposed in a faqīh narrator. «ثمَّ هَذَا النَّوْع من الْقُصُور لَا يتَوَهَّم فِي الرَّاوِي إِذا كَانَ فَقِيها» — "This kind of shortfall is not supposed in a narrator who is a faqīh." That is why the jurist-Companions narrated little (Ibn Masʿūd trembling as he said "samiʿtu"). (v1/342)
• Who: No Ḥanafī is named as the originator, and no Ḥanafī dissent is recorded here; Sarakhsī speaks in the first person plural (qulnā, “we say”). The Companions and Successors he cites as precedent: Ibn ʿAbbās, ʿĀʾisha, ʿUmar, Ibrāhīm al-Nakhaʿī. (v1/340–341)
• Proofs: The Companions and Successors themselves weighed some of Abū Hurayra's reports against qiyās (v1/340–341):
  1. Ibn ʿAbbās on wuḍūʾ from what fire has touched ("would you make wuḍūʾ from hot water?"). Abū Hurayra's reply: «يَا ابْن أخي إِذا أَتَاك الحَدِيث فَلَا تضرب لَهُ الْأَمْثَال» — "Nephew, when a ḥadīth reaches you, do not strike analogies against it." Ibn ʿAbbās also objected to wuḍūʾ from carrying a bier. (v1/340)
  2. On "the child of zinā is the worst of the three", ʿĀʾisha cited a verse: «كَيفَ يَصح هَذَا وَقد قَالَ الله تَعَالَى» ﴿Q 6:164﴾ — "How can this be, when God says: no bearer of burdens bears another's burden?" (Q 6:164) (v1/340). Sarakhsī comments on her proof: «وَهَذَا عَام دخله خُصُوص» — "and this is a general text into which specification has entered." (v1/341)
  3. The qiyās-type objection to the same report comes next, apparently from al-Nakhaʿī (the text says only "and he said"): «لَو كَانَ ولد الزِّنَا شَرّ الثَّلَاثَة لما انْتظر بِأُمِّهِ أَن تضع» — "If the child of zinā were the worst of the three, [her punishment] would not have been deferred until she gave birth." Sarakhsī comments: «وَهَذَا نوع قِيَاس» — "and this is a kind of qiyās." Also al-Nakhaʿī: «كَانُوا يَأْخُذُونَ من حَدِيث أبي هُرَيْرَة وَيدعونَ» — "They used to take some of Abū Hurayra's ḥadīth and leave some." (v1/341)
  4. Further attestations of scrutiny, without a new argument: ʿĀʾisha's remark on the volume of his narration, and ʿUmar's warning «لتكفن عَن هَذَا أَو لألحقنك بجبال دوس» — "Stop this, or I will send you back to the mountains of Daws." (v1/341)
• Objection → Reply: "Ibn ʿAbbās rejected it by another text (the Prophet ate a shoulder and prayed without wuḍūʾ), not by qiyās." Reply: had he had a text he would not have argued by analogy, or he would have sought the chronology, or restricted meat by it. He argued by qiyās, so he held the report open to scrutiny where it contradicted qiyās. (v1/340)
• Furūʿ test
  • The muṣarrāt: «وَبَيَان هَذَا فِي حَدِيث الْمُصراة فَإِن الْأَمر برد صَاع من تمر مَكَان اللَّبن قل أَو كثر مُخَالف للْقِيَاس الصَّحِيح من كل وَجه» — "the command to return a ṣāʿ of dates in place of the milk, whether little or much, contradicts sound qiyās in every respect", because liability for wrongs is by like or by value, fixed by the Kitāb, Sunna and ijmāʿ. (v1/341)
  • Salama b. al-Muḥabbiq's report on a man who has intercourse with his wife's slave-girl: «فَإِن طاوعته فَهِيَ لَهُ وَعَلِيهِ مثلهَا وَإِن استكرهها فَهِيَ حرَّة وَعَلِيهِ مثلهَا» — "if she complied, she is his and he owes [his wife] one like her; if he forced her, she is free and he owes one like her." Sarakhsī: «فَإِن الْقيَاس الصَّحِيح يرد هَذَا الحَدِيث» — "sound qiyās rejects this ḥadīth." (v1/342)
• Limit on the rule: the senior Ḥanafīs still rely on such narrators. Abū Ḥanīfa took Anas's view on the length of menstruation, though Abū Hurayra ranks above Anas. «مَا تركُوا الْعَمَل بروايتهم إِلَّا عِنْد الضَّرُورَة لانسداد بَاب الرَّأْي» — "They only left acting on their narration out of necessity, because the door of raʾy was closed." (v1/342)
• [Editorial] Debate note: State this position with its conditions and the disclaimer, or it becomes a straw man ("Ḥanafīs reject Abū Hurayra"). Sarakhsī here names no originator and no Ḥanafī dissenter; any attribution to a named early Ḥanafī, or report of internal dissent, needs a source outside this book.

Back to the list

B9b.04 · رواية المجهول · The "unknown" Companion-narrator: five cases, decided by how the early jurists treated him (v1/342–345)

• Definition: «من لم يشْتَهر بطول الصُّحْبَة مَعَ رَسُول الله إِنَّمَا عرف بِمَا رُوِيَ من حَدِيث أَو حديثين» — "one not famed for long companionship with the Messenger, known only through one or two reports he narrated." Examples: Wābiṣa b. Maʿbad, Salama b. al-Muḥabbiq, Maʿqil b. Sinān. (v1/342)
• Position: «وَرِوَايَة هَذَا النَّوْع على خَمْسَة أوجه» — "This type's narration falls into five cases." (v1/342)
  1. The jurists accepted and narrated from him → he ranks with the well-known. (v1/342–343)
  2. They stayed silent after it spread → the same, since silence when there is need implies approval. (v1/343)
  3. They differed over it → accepted “in our view” (ʿindanā). Case: Maʿqil b. Sinān on Barwaʿ bint Wāshiq's mahr al-mithl; Ibn Masʿūd accepted it, ʿAlī rejected it, and the jurists of the second generation (al-qarn al-thānī: ʿAlqama, Masrūq, al-Ḥasan, Nāfiʿ b. Jubayr) accepted it. «فَصَارَ معدلا بِقبُول الْفُقَهَاء رِوَايَته» — "He became vouched-for through the jurists' acceptance." (v1/343)
     • Sarakhsī's tentative explanation of the split: «وَكَأن عليا إِنَّمَا لم يقبل رِوَايَته لِأَنَّهُ كَانَ مُخَالفا للْقِيَاس عِنْده» — "It is as if ʿAlī did not accept his narration only because it contradicted qiyās in his view", while Ibn Masʿūd accepted it as agreeing with his qiyās. «فَتبين بِهَذَا أَن رِوَايَة مثل هَذَا فِيمَا يُوَافق الْقيَاس يكون مَقْبُولًا» — "This shows that such a narrator's report is accepted where it agrees with qiyās." (v1/343) This is the bridge to case 5.
  4. They unanimously rejected it → not acted on. «فاتفاقهم على الرَّد دَلِيل على أَنهم كذبوه فِي هَذِه الرِّوَايَة» «وَعَلمُوا أَن ذَلِك وهم مِنْهُ» — "Their agreed rejection shows that they declared him wrong [lit. 'belied him'] in this narration, and knew it was a mistake on his part." (v1/343) The analogy: «وَلَو قَالَ الرَّاوِي أوهمت لم يعْمل بروايته» — "if the narrator himself said 'I erred', his narration would not be acted on", and evidence of it from those above him is stronger. (v1/343)
     • Case: Fāṭima bint Qays, with ʿUmar's words on "the Book of our Lord and the Sunna of our Prophet". (v1/343)
     • Grouped with it: «وَمن هَذَا النَّحْو حَدِيث سهل بن أبي حثْمَة» on qasāma, Busra's report on touching the penis, and Abū Hurayra's "whoever wakes junub has no fast". (v1/344)
  5. It neither spread nor was challenged → «فَإِن الْعَمَل بِهِ لَا يجب وَلَكِن يجوز الْعَمَل بِهِ إِذا وَافق الْقيَاس» — "Acting on it is not obligatory, but it is permitted if it agrees with qiyās." (v1/344)
     • ⚑ Parallel, and a time-bound Ḥanafī split: «وَلِهَذَا جوز أَبُو حنيفَة الْقَضَاء بِشَهَادَة المستور وَلم يُوجب على القَاضِي الْقَضَاء» (v1/344) — "For this reason Abū Ḥanīfa permitted judging on the unexamined (mastūr) witness's testimony, but did not oblige the qāḍī to judge by it", because he lived in the third generation, when truthfulness prevailed. The Ṣāḥibān did not permit it: «وَلِهَذَا لم يجوز أَبُو يُوسُف وَمُحَمّد الْقَضَاء بِشَهَادَة المستور قبل ظُهُور عَدَالَته» — "Abū Yūsuf and Muḥammad did not allow judgement on the mastūr's testimony before his probity is shown." (v1/345)
     • Sarakhsī applies the same logic to case 5: «فَأَما فِي زَمَاننَا رِوَايَة مثل هَذَا لَا يكون مَقْبُولًا» — "In our time the narration of such a one is not accepted" until upright narrators accept it. (v1/344)
• ʿĪsā b. Abān's reading of ʿUmar's words, as Sarakhsī reports it: «مُرَاده من الْكتاب وَالسّنة الْقيَاس الصَّحِيح» — "By 'the Book and the Sunna' he meant sound qiyās" (a qiyās al-shabah: maintenance measured on lodging). (v1/344)
• Objection → Reply: "ʿUmar said 'we do not know whether she spoke truly', not 'she erred'." Reply: accepting a report rests on truthfulness visibly outweighing falsehood; he said it had not, and raʾy pointed the other way, so it is the same as rejecting it. Like a qāḍī telling a fāsiq witness "bring another witness". (v1/344)
• Summary in his words: «فَصَارَ الْحَاصِل أَن الحكم فِي رِوَايَة الْمَشْهُور الَّذِي لم يعرف بالفقه وجوب الْعَمَل وَحمل رِوَايَته على الصدْق إِلَّا أَن يمْنَع مِنْهُ مَانع وَهُوَ أَن يكون مُخَالفا للْقِيَاس» — "The upshot: the well-known narrator not known for fiqh must be acted on and taken as truthful, unless a bar prevents it, namely contradicting qiyās." The majhūl is not proof for action unless supported by the early generations' acceptance. (v1/345)
• [Editorial] Debate note: In this passage, "majhūl" is a Companion with a thin record, defined by the text above. Check which sense an interlocutor means before comparing it with other uses of the term.

Back to the list

B9c.01 · شرائط الراوي · Four conditions: reason, accuracy, uprightness and Islam; an unbeliever's report is refused for an added motive to lie, not for unbelief itself (v1/345–346)

• Position: «اعْلَم بِأَن هَذِه الشَّرَائِط أَرْبَعَة الْعقل والضبط وَالْعَدَالَة وَالْإِسْلَام» — "Know that these conditions are four: reason, accuracy, uprightness and Islam." (v1/345)
  • Reason: speech counts only when it comes from discernment, not rote or raving (a bird's recited letters are a tune, not speech). Accuracy: truth depends on good retention from hearing to relating. Uprightness: one who shuns what he believes forbidden is likely to shun lying; the sinner's conduct and belief pull against each other, so one must wait. (v1/345–346)
  • Islam: unbelievers' hostility to the religion drives them to insert what is not in it ("they spare no effort to corrupt you", Q 3:118), as they hid the Prophet's description. So their testimony against Muslims is refused, but among themselves it is accepted. «وَبِهَذَا تبين أَن رد خَبره لَيْسَ لعين الْكفْر بل لِمَعْنى زَائِد يُمكن تُهْمَة الْكَذِب فِي خَبره» — "This shows that his report is refused not for unbelief itself, but for an added factor that makes lying suspect", like a father's testimony for his son. (v1/346)

Back to the list

B9c.02 · العقل · Reason is a light in the breast; the law fixes full reason at puberty without defect; a child may hear and later relate (v1/346–348)

• Reason is a light by which the heart sees proofs, as a lamp lets the eye see; it grows gradually, so the law takes puberty with no defect as its sign, for ease. A child's partial reason counts only where it purely benefits him. (v1/346–347)
• So a discerning child may hear (and bear testimony) and relate after puberty, as some Companions did; he cannot bind others while a child. The imbecile who understands is like the child. (v1/347–348)

Back to the list

B9c.03 · الضبط · Accuracy is outer (the wording and its sense in language) and inner (its legal meaning, i.e. fiqh); hence the faqīh's report prevails; Abū Ḥanīfa narrated little for this reason (v1/348–350)

• Definition: «فَأَما الضَّبْط فَهُوَ عبارَة عَن الْأَخْذ بِالْجَزْمِ» — "Accuracy means taking firmly": hearing fully, understanding the meaning intended, memorising it and holding to it until relating it. So Abū Ḥanīfa did not let a witness testify from his handwriting on a deed without remembering the event. (v1/348)
• Two kinds: outer (the wording and its linguistic meaning) and inner (its meaning for legal rulings, which is fiqh). So the report of one with grave inattention is refused, and the early generations did not set a non-faqīh's report against a faqīh's. When Jābir b. Zayd was told that Yazīd b. al-Aṣamm said the Prophet married Maymūna while not in iḥrām, he answered that Ibn ʿAbbās (who said "in iḥrām") knew better, and asked how Yazīd, "one who urinates on his heels", could be set against Ibn ʿAbbās. The reason: relating by meaning was common, and the faqīh does not fall short in it. «وَلِهَذَا قُلْنَا إِن الْمُحَافظَة على اللَّفْظ فِي زَمَاننَا أولى من الرِّوَايَة بِالْمَعْنَى» (v1/349) — "Hence we say that keeping to the wording is better in our time than relating by meaning." (v1/348–349)
• Objection → Reply: the Qurʾān is validly transmitted by those who do not understand it. Reply: its transmission began with the best of people and continues with great effort, God promised to protect it (Q 15:9), and rulings attach to its wording (recitation, prayer, inimitability). In reports only the meaning matters, so full accuracy needs understanding; hence Abū Ḥanīfa and Muḥammad do not allow testimony to a sealed letter whose content the witness does not know. (v1/349–350)
• The salaf narrated little: Abū Bakr ("refer people to the Book of God"), ʿUmar ("narrate little from the Messenger"), Zayd b. Arqam ("we have grown old and forgotten"), Ibn ʿAbbās. For this reason Abū Ḥanīfa narrated little, so that some attackers said he did not know ḥadīth: «وَلم يكن على مَا ظن بل كَانَ أعلم أهل عصره بِالْحَدِيثِ وَلَكِن لمراعاة شَرط كَمَال الضَّبْط قلت رِوَايَته» — "It was not as he thought: he was the most knowledgeable of his age in ḥadīth, but his narration was little out of regard for the condition of complete accuracy." One who joined a gathering late may miss the start, so he should relate only what he retained with full effort. (v1/350)
• [Editorial] Debate note: this is Sarakhsī's answer to the charge that Abū Ḥanīfa was weak in ḥadīth. His ground is a stricter accuracy standard (outer plus inner ḍabṭ), not a count of narrations. Use the passage as Sarakhsī's claim, not as a historical finding.

Back to the list

B9c.04 · العدالة · Uprightness is shunning what one believes forbidden: no major sins and no persistence in minor ones; the heretic who shuns sins is accepted; the unknown of the first three generations is upright, his report proof only as settled in B9b.04 (v1/350–352)

• Outer uprightness comes with religion and reason; inner uprightness is known from dealings. Since desire precedes reason and stays with it, «فَعرفنَا أَن الْعدْل مُطلقًا من يتَرَجَّح أَمر دينه على هَوَاهُ» (v1/351) — "we know that the fully upright is one whose religion outweighs his desire", and who abstains from what he believes forbidden. So a major sin removes uprightness in testimony, and persisting in a minor one does too. Strictly any sin would, but that would close the door, while avoiding persistence and major sins is possible. (v1/350–351)
• The heretic (ṣāḥib al-hawā) who shuns what he believes forbidden is accepted as a witness: he erred from too much zeal for proof, which does not bring suspicion of lying. Slavery, femaleness and blindness do not affect uprightness, though they bar or reduce testimony. The sinner and the unexamined (mastūr) are not fully upright for testimony: a qāḍī may not rule on a sinner's testimony, though if he does it takes effect, but he need not rule on a mastūr's before his state is known. (v1/351–352)
• Who: al-Shāfiʿī: if the sinner's and mastūr's reports are not proof, the unknown's is even less so. «وَقُلْنَا نَحن الْمَجْهُول من الْقُرُون الثَّلَاثَة عدل بتعديل صَاحب الشَّرْع إِيَّاه مَا لم يتَبَيَّن مِنْهُ مَا يزِيل عَدَالَته» — "We say: the unknown narrator of the first three generations is upright, by the Lawgiver's endorsement, unless something appears that removes his uprightness", and so his report is proof "in the way we settled", i.e. only with support such as the early authorities' acceptance. (v1/352) → B9b.04

Back to the list

B9c.05 · الإسلام · Islam is outer (birth and upbringing) and inner (assent); a general description of the faith suffices, against detailed interrogation (v1/352–353)

• Outer Islam comes by birth among Muslims; inner by assent to God with His attributes and names, His angels, books, messengers, the resurrection, decree, and His rulings. al-Jāmiʿ al-kabīr: a girl who reaches puberty and cannot describe Islam when asked is separated from her husband (treated as apostasy). (v1/352)
• Who: ⚑ some of our shaykhs require knowing the reality of what one describes (a Christian "believes in ʿĪsā" as God's son). Sarakhsī: «وَلَكنَّا نقُول فِي الْمصير إِلَى هَذَا الِاسْتِقْصَاء حرج بَين» (v1/352) — "But we say: going to this length is a clear hardship." A general description suffices: the Prophet asked the Bedouin who saw the crescent only the two testimonies; Jibrīl's questions were answered in general terms; "test them" (Q 60:10). Ask in question form ("do you believe this and this?"); regular congregational prayer also counts. (v1/352–353)

Back to the list

B9c.06 · العبد والمرأة والأعمى والمحدود في القذف · Slaves, women and the blind narrate like others; ⚑ the one punished for qadhf, after repentance: ẓāhir al-madhhab vs al-Ḥasan's report (v1/353–355)

• The four conditions do not vary with slavery, sex or sight, so their reports are equal. Testimony differs because it imposes one's word on another (wilāya). The reporter binds himself first, and the hearer binds himself by his own belief, like the qāḍī bound by his office, not by the witness. A slave's testimony on the Ramaḍān crescent is accepted for this reason. (v1/353–354)
• The Prophet accepted invitations from slaves, Salmān's charity and gift while a slave, and Barīra's reports; the Companions relied on the Prophet's wives ("take two-thirds of your religion from ʿĀʾisha"). Blindness does not touch uprightness (Shuʿayb, Yaʿqūb; Ibn Umm Maktūm; Ibn ʿAbbās, Ibn ʿUmar, Jābir), and no one asked whether they related before or after losing sight. (v1/354)
• Who: ⚑ the one punished for qadhf, after repenting. On the ẓāhir al-madhhab he narrates like others (Abū Bakra's reports are accepted, with no one asking when he related them), since refusing his testimony completes his ḥadd and narration is not testimony. On al-Ḥasan's report from Abū Ḥanīfa he is not accepted, since the text declares him a liar (Q 24:13). (v1/354–355)

Back to the list

B9c.07 · النقل بالمعنى · Keeping the wording is better; relating by meaning is allowed after good accuracy, by type of text; the Prophet's "comprehensive words" must be kept verbatim, on Sarakhsī's view (v1/355–357)

• Who: some ḥadīth scholars (from Ibn Sīrīn): the wording must be kept; some theorists: a Companion's paraphrase is not proof ("an abandoned view"). «وَقَالَ جُمْهُور الْعلمَاء مُرَاعَاة اللَّفْظ فِي النَّقْل أولى وَيجوز النَّقْل بِالْمَعْنَى بعد حسن الضَّبْط» — "The majority of scholars say: keeping the wording is better, and relating by meaning is allowed after good accuracy" (al-Ḥasan, al-Shaʿbī, al-Nakhaʿī). (v1/355)
• Proofs: the Companions' "the Messenger commanded us…", Ibn Masʿūd's "or something like it", Anas's "or as the Messenger said". The ḥadīth's wording is not inimitable; the aim is its meaning. Even the Qurʾān has the concession of the seven modes. (v1/355–356)
• By type: the firm (muḥkam): anyone who knows the language may paraphrase. The apparent (ẓāhir: general, literal): only one who also knows fiqh, lest he widen or narrow it. The problematic and the shared: never, since their meaning comes by interpretation. The summary and the ambiguous: impossible to paraphrase. (v1/356–357)
• Who: ⚑ "comprehensive words" (jawāmiʿ al-kalim: "revenue follows liability"; "the animal's damage is unaccountable"): some of our shaykhs allow paraphrase on the condition set for the apparent. Sarakhsī: «وَالأَصَح عِنْدِي أَنه لَا يجوز ذَلِك» — "the sounder, to me, is that it is not allowed": the Prophet was singled out for this concision, and "relate it as he heard it" refers to this kind. (v1/357)

Back to the list

B9c.08 · الضبط بالكتابة · Writing as a reminder is fine; ⚑ relying on handwriting without remembering: Abū Ḥanīfa never; Abū Yūsuf in the register and ḥadīth, and in a deed only when it is safe in the witness's hands; Muḥammad in all three; Sarakhsī adopts the concession (v1/357–359)

• Writing as a reminder: one who reads and remembers may relate, whoever's hand it is in ("bind knowledge by writing"; forgetting is human, except as God exempted the Prophet, Q 87:6). (v1/357–358)
• Who: ⚑ relying on handwriting without remembering, in ḥadīth, in a qāḍī's register, and in a witness's deed. Abū Ḥanīfa took the strict view: «لَا يجوز لَهُ أَن يعْتَمد الْكتاب مَا لم يتَذَكَّر» (v1/358) — "he may not rely on the writing unless he remembers", since handwriting resembles handwriting, and a mirror that shows nothing counts for nothing. Abū Yūsuf (via Bishr b. al-Walīd): allowed in the register and in ḥadīth, not in the deed. His basis is safety from tampering: the register sits sealed in the qāḍī's own case, and is not relied on once out of his hands, since forgery is usual in lawsuits; a ḥadīth book may be relied on in his own hands or those of any trustworthy keeper, since forgery is not usual there; the deed sits with the opponent, «حَتَّى إِذا كَانَ فِي يَد الشَّاهِد كَانَ الْجَواب فِيهِ مثل الْجَواب فِي السّجل» (v1/359) — "so if it is in the witness's hands, the answer is the same as for the register." Muḥammad (via Ibn Rustam): allowed in all three, once he knows it is his hand beyond doubt. Sarakhsī adopts the concession ("what we have gone to is a concession, for ease on people"), not Abū Ḥanīfa's strict view. (v1/358–359)
• Finding a known trustworthy man's writing: one may say "I found in so-and-so's hand", and it counts only with some other support; this is for reports only, not testimony or judging (Q 43:86; "if you see it like the sun, testify"). (v1/359)

Back to the list

B9d.S1 · الانقطاع صورة: مرسل الصحابي · Break in form: the Companions' mursal is proof by agreement (v1/359)

• Position: a break (inqiṭāʿ) is of two kinds, in form and in meaning (the latter: B9d.01–08, contrary proof in .02–.05 and narrator deficiency in .06–.08). In form, it is the mursal. «وَلَا خلاف بَين الْعلمَاء فِي مَرَاسِيل الصَّحَابَة أَنَّهَا حجَّة» — "There is no disagreement among scholars that the Companions' mursal reports are proof", since they heard from the Prophet or from each other, and they were truthful (al-Barāʾ: "not all we tell you we heard from the Messenger… but we do not lie"). (v1/359)

Back to the list

B9d.S2 · مراسيل القرون الثلاثة · The mursal of the second and third generations is proof, against al-Shāfiʿī (on the preferred view, unless the narrator is known to relate from the non-upright, B9d.S3); ʿĪsā b. Abān: even stronger than a musnad (v1/360–363)

• Position: «فَأَما مَرَاسِيل الْقرن الثَّانِي وَالثَّالِث حجَّة فِي قَول عُلَمَائِنَا» — "The mursal reports of the second and third generations are proof on the view of our scholars." (v1/360)
• Who: al-Shāfiʿī: not proof unless supported by a verse, a well-known sunna, the early generations' practice, or a connected route; hence he accepts Saʿīd b. al-Musayyab's, "because I traced them and found them connected". His case: proof depends on the narrator's qualities, which are unknown if he is unnamed; a narrator may relate from someone he himself thinks weak (al-Shaʿbī: "al-Ḥārith told me, and by God he was a liar"); testimony on testimony must name the originals; and people's effort to keep chains would otherwise be pointless. (v1/360)
• Proofs: the same proofs as for single reports; and the Companions and their successors did it openly. Abū Hurayra on waking junub ("al-Faḍl told me"); Ibn ʿAbbās heard only some dozen ḥadīths directly; al-Nuʿmān b. Bashīr one; al-Ḥasan and Ibn al-Musayyab often. al-Ḥasan: "when four Companions agreed on a ḥadīth for me, I sent it without a chain." Ibn Sīrīn: "we did not give chains until the fitna." Ibrāhīm: "when I say 'ʿAbd Allāh said', it is from more than one." Hence ʿĪsā b. Abān: «الْمُرْسل أقوى من الْمسند» (v1/361) — "the mursal is stronger than the musnad", since one who heard by many routes drops the chain. (v1/360–361) Objection: then abrogate by it. Reply: its extra strength comes by a kind of ijtihād, like qiyās. (v1/361)
• The trilemma: the great ones sent reports without chains either from narrators they thought not upright (unthinkable), or from upright ones while thinking a mursal no proof (hiding proof), or believing the mursal proof like a musnad. «فَتعين الثَّالِث وَهُوَ أَنهم اعتقدوا أَن الْمُرْسل حجَّة كالمسند وَكفى باتفاقهم حجَّة» (v1/362) — "So the third is settled: they held the mursal to be proof like the musnad, and their agreement is proof enough." (v1/361–362)
  • al-Shāfiʿī's "they sent them so that the musnad would be sought" fails: either they had no chain (inventing), or they had one and dropped it only to tire us. A muftī's "the Messenger ruled thus" binds without a chain, and "so-and-so from so-and-so" is accepted without "he told me". A trustworthy man attesting directly to the Prophet is more telling than attesting to a narrator. Testimony on testimony differs: the originals may owe compensation on retraction, and the branch stands in for the original. Seeking chains is like seeking many routes, which does not disprove single reports. (v1/362–363)

Back to the list

B9d.S3 · مراسيل من بعد القرون الثلاثة · ⚑ After the three generations: al-Karkhī accepts all; ʿĪsā b. Abān accepts the famous carriers of knowledge; al-Rāzī's view, the soundest: only from one known to relate solely from the upright (v1/363–364)

• Who: al-Karkhī: no difference between ages; whoever is accepted with a chain is accepted without one. ʿĪsā b. Abān: whoever is famous as one people take knowledge from (Muḥammad b. al-Ḥasan and the like) is accepted with or without a chain; one known only as a narrator: his musnad is proof, his mursal is suspended «إِلَى أَن يعرض على من اشْتهر بِحمْل الْعلم عَنهُ» (v1/363) — "until it is put before one famous for carrying knowledge from him."
• Position: «وَأَصَح الْأَقَاوِيل فِي هَذَا مَا قَالَه أَبُو بكر الرَّازِيّ» — "The soundest view here is what Abū Bakr al-Rāzī said": the mursal of the first three generations is proof «مَا لم يعرف مِنْهُ الرِّوَايَة مُطلقًا عَمَّن لَيْسَ بِعدْل ثِقَة» (v1/363) — "so long as he is not known to relate freely from one who is not upright and reliable"; after them, only from «من اشْتهر بِأَنَّهُ لَا يروي إِلَّا عَمَّن هُوَ عدل ثِقَة» (v1/363) — "one famous for relating only from the upright and reliable." The Prophet attested to the truthfulness of the three generations and foretold that lying would spread after them. ʿUrwa told ʿUmar b. ʿAbd al-ʿAzīz, "the upright, approved man told me", and was accepted. (v1/363)
• Broken by one route, connected by another: some ḥadīth scholars drop it, as if taking the break as a sign of jarḥ; «وَأَكْثَرهم على أَن هَذَا يكون حجَّة» — "most of them hold it is proof": the broken route is silent about the narrator, the connected one names him, and the silent does not oppose the speaking. (v1/364)

Back to the list

B9d.01 · الانقطاع معنى · Break in meaning: a report is cut off by a contrary proof or by a defective narrator (v1/364, 369–370)

• Position: Inqiṭāʿ maʿnawī «يَنْقَسِم قسمَيْنِ إِمَّا أَن يكون ذَلِك الْمَعْنى بِدَلِيل معَارض أَو نُقْصَان فِي حَال الرَّاوِي» — "divides in two: the break comes either from a contrary proof or from a deficiency in the narrator's state." (v1/364) This pilot covers the first; the narrator-deficiency half (v1/370–374) is outside it.
  • The contrary-proof kind has four forms: «إِمَّا أَن يكون مُخَالفا لكتاب الله تَعَالَى أَو لسنة مَشْهُورَة» — "either it contradicts the Book of God or a well-known Sunna" — or it is a lone report on a matter of general need, or the early authorities ignored it while disputing the very question. (v1/364)
• Who: Contrast with al-Shāfiʿī, as Sarakhsī frames it: «وَالشَّافِعِيّ أعرض عَن طلب الِانْقِطَاع معنى واشتغل بِبِنَاء الحكم على ظَاهر الِانْقِطَاع فِي الْمُرْسل» — "al-Shāfiʿī turned away from seeking the break in meaning, and busied himself with building rulings on the outward break in the mursal." (v1/370)
  • «وعلماؤنا يبنون الْفِقْه على الْمعَانِي المؤثرة» — "Our scholars build fiqh on effective meanings." (v1/370)
• Coin simile: filters 1–2 expose a counterfeit by comparison, «بِمَنْزِلَة نقد الْبَلَد إِذا قوبل بِنَقْد أَجود مِنْهُ تظهر الزيافة فِيهِ» — "like local coin set beside finer coin, which shows its baseness." Filters 3–4 expose it by extra suspicion of a break, like coin found to carry more alloy than usual. (v1/369)
• [Editorial] Debate note: This is the Ḥanafī answer to "you accept mursal but reject ṣaḥīḥ reports": the two cases use different kinds of break (outward vs in meaning). Know both halves (B9d mursal and these four filters) before debating either.

Back to the list

B9d.02 · مخالفة الكتاب · Filter 1: a single report contradicting the Kitāb is not acted on (v1/364–366)

• Position: «فَإِنَّهُ لَا يكون مَقْبُولًا وَلَا حجَّة للْعَمَل بِهِ عَاما كَانَت الْآيَة أَو خَاصّا نصا أَو ظَاهرا عندنَا» — "It is neither accepted nor a proof for action, whether the verse is general or specific, explicit or apparent, in our view." (v1/364)
  • This rests on the view that the ʿāmm is definitive, so a single report cannot specify it at the outset, nor turn a ẓāhir to a figurative sense. «وَالْحمل على نوع من الْمجَاز لَا يجوز بِخَبَر الْوَاحِد عندنَا خلافًا للشَّافِعِيّ» — "Taking [the verse] in a figurative sense is not allowed by a single report in our view, contrary to al-Shāfiʿī." (v1/364) → see B4 (ʿāmm qaṭʿī).
• Who: Opponent: al-Shāfiʿī. (v1/364)
• Proofs
  1. "Every condition not in the Book of God is void", read as "contrary to the Book"; it cannot mean "not found in it", since that ḥadīth itself is not in it. (v1/364)
  2. The report of presenting ḥadīth to the Book: «فاعرضوه على كتاب الله تَعَالَى فَمَا وَافقه فاقبلوه» — "Present it to the Book of God: what agrees with it, accept." (v1/365)
  3. «وَلِأَن الْكتاب مُتَيَقن بِهِ وَفِي اتِّصَال الْخَبَر الْوَاحِد برَسُول الله شُبْهَة» — "The Book is certain, while the single report's link to the Messenger carries doubt." Weighing starts at the wording (matn) before the meaning, and the Qurʾān's wording is mass-transmitted. (v1/365)
• Furūʿ test
  • Wuḍūʾ from touching the penis: «وَلِهَذَا لم يقبل عُلَمَاؤُنَا خبر الْوضُوء من مس الذّكر لِأَنَّهُ مُخَالف للْكتاب» — "For this reason our scholars did not accept the report of wuḍūʾ from touching the penis, since it contradicts the Book." Q 9:108 praises those who "love to purify themselves" (istinjāʾ with water, which involves touching), and «لِأَن الْفِعْل الَّذِي هُوَ حدث لَا يكون تطهرا» — "an act that is ḥadath cannot be purification." (v1/365)
  • Fāṭima bint Qays (no maintenance for the irrevocably divorced woman) against Q 65:6. Sarakhsī reads the verse as covering maintenance: «وَلَا خلاف أَن المُرَاد وأنفقوا عَلَيْهِنَّ من وجدكم فَالْمُرَاد الْحَائِل» — "there is no dispute that the meaning is 'and spend on them from your means', so the non-pregnant woman is meant", since pregnancy has its own clause. (v1/365)
  • Judgement on one witness plus the plaintiff's oath against Q 2:282. The verse's list is the full explanation of what is required, so the extra is an addition: «وَالزِّيَادَة على النَّص كالنسخ عندنَا» — "Addition to the text is like abrogation, in our view." (v1/365–366) [print p. 366; turath misplaces the top of p. 366 onto p. 365, verified against the printed page]
    • Further points: "that is nearer to preventing doubt" (Q 2:282) is read as fixing the minimum: «نَص على أَن أدنى مَا تَنْتفِي بِهِ الرِّيبَة شَهَادَة شَاهِدين» — "a text that the least by which doubt is removed is the testimony of two witnesses", and nothing is below that minimum (v1/366). The verse moves to two women rather than to an oath, and Q 5:106 moves to non-Muslim witnesses rather than to an oath; the same verse also mentions the witnesses' oath when suspicion arises (not proof today; the printings add "because of abrogation" in brackets, which the Asfār edition records only as a variant), so if the plaintiff's oath could remove suspicion or complete the proof, mentioning it would have been more fitting. (v1/366)
• [Editorial] Debate note: Filter 1 only works together with B4 (ʿāmm qaṭʿī) and B11 (ziyāda = naskh). An opponent who concedes neither will not feel its force; debate those premises first. Expect proof 2 to be contested on its transmission; check its takhrīj before relying on it.

Back to the list

B9d.03 · مخالفة السنة المشهورة · Filter 2: a gharīb report contradicting the well-known Sunna is set aside (v1/366–368)

• Position: «وَكَذَلِكَ الْغَرِيب من أَخْبَار الْآحَاد إِذا خَالف السّنة الْمَشْهُورَة فَهُوَ مُنْقَطع فِي حكم الْعَمَل بِهِ» — "Likewise, an isolated single report that contradicts the well-known Sunna is cut off as regards acting on it." (v1/366) The mutawātir, mustafīḍ and agreed-upon Sunna stands with the Kitāb in certainty. The mashhūr is stronger than the gharīb, which is why abrogation is allowed by the mashhūr and not by the gharīb. (v1/366)
• Furūʿ test
  • One witness plus oath, against «الْبَيِّنَة على الْمُدَّعِي وَالْيَمِين على من أنكر» ("proof is on the claimant and the oath on the denier"), in two ways: the oath belongs to the denier, and oath and proof are not combined. (v1/367)
  • ⚑ Fresh dates for dried (ruṭab bi'l-tamr): «وَلِهَذَا الأَصْل لم يعْمل أَبُو حنيفَة بِخَبَر سعد بن أبي وَقاص» — "On this principle Abū Ḥanīfa did not act on Saʿd b. Abī Waqqāṣ's report" ("does it shrink when it dries?"). It contradicts "dates for dates, like for like" in two ways (v1/367):
    1. That Sunna requires equality by measure without condition, so «فالتقييد بِاشْتِرَاط الْمُمَاثلَة فِي أعدل الْأَحْوَال وَهُوَ بعد الجفوف يكون زِيَادَة» — "restricting it by requiring equality in the most balanced state, i.e. after drying, is an addition." (v1/367)
    2. That Sunna makes the forbidden excess the one shown by measure, so «فَجعل فضل يظْهر عِنْد فَوَات وصف مَرْغُوب فِيهِ رَبًّا حَرَامًا يكون مُخَالفا لذَلِك الحكم» — "making an excess that appears on the loss of a desired quality into forbidden ribā contradicts that ruling." (v1/367)
    3. The Ṣāḥibān disagree on the application, not the principle: «إِلَّا أَن أَبَا يُوسُف ومحمدا قَالَا السّنة الْمَشْهُورَة لَا تتَنَاوَل الرطب» — "But Abū Yūsuf and Muḥammad said the well-known Sunna does not cover fresh dates", as the oath cases show, so Saʿd's report governs. Abū Ḥanīfa's reply: "tamr" names the fruit from formation to ripeness, and oaths follow custom. (v1/367)
• The payoff, in his words: «فَإِن أصل الْبدع والأهواء إِنَّمَا ظهر من قبل ترك عرض أَخْبَار الْآحَاد على الْكتاب وَالسّنة الْمَشْهُورَة» — "The root of innovations and sectarian whims appeared from abandoning the presentation of single reports to the Book and the well-known Sunna." Some made the doubtful the base and interpreted the certain by it. (v1/367)
  • The same charge against the report-deniers: «بِمَنْزِلَة من أنكر خبر الْوَاحِد فَإِنَّهُ لما لم يجوز الْعَمَل بِهِ احْتَاجَ إِلَى الْقيَاس» (v1/367) — "just like one who denies the single report: since he does not allow acting on it, he needs qiyās", or istiṣḥāb, which is no proof at all. «وكل وَاحِد مِنْهُمَا زيف مَرْدُود» — "Each of the two is counterfeit and rejected." (v1/368)
  • The right path: «وَإِنَّمَا سَوَاء السَّبِيل مَا ذهب إِلَيْهِ عُلَمَاؤُنَا من إِنْزَال كل حجَّة منزلتها» — "The straight path is what our scholars took: putting each proof in its rank." (v1/368)
  • The positive rule (do not omit): «وَمَا لم يَجدوا فِي الْكتاب وَلَا فِي السّنة الْمَشْهُورَة لَهُ ذكرا قبلوه أَيْضا وأوجبوا الْعَمَل بِهِ» — "What they found no mention of in the Book or the well-known Sunna, they also accepted and made action on it obligatory." Only then, «وَمَا لم يجدوه فِي شَيْء من الْأَخْبَار وصاروا حِينَئِذٍ إِلَى الْقيَاس» — "what they found in none of the reports, they then resorted to qiyās for." (v1/368)
• [Editorial] Debate note: This passage is Sarakhsī's own answer to "Ḥanafīs put qiyās over ḥadīth": the order is Kitāb and mashhūr, then single reports that do not contradict them, then qiyās. The ruṭab case shows a Ḥanafī split sitting on the application (does the mashhūr text cover the case?) while the principle is shared.

Back to the list

B9d.04 · عموم البلوى · Filter 3: a lone report on a matter of general need is suspect; the act's legal grade (wujūb) may still rest on a single report (v1/368–369)

• Position: «وَأما الْقسم الثَّالِث وَهُوَ الْغَرِيب فِيمَا يعم بِهِ الْبلوى وَيحْتَاج الْخَاص وَالْعَام إِلَى مَعْرفَته للْعَمَل بِهِ فَإِنَّهُ زيف» — "The third: an isolated report on a matter of general need, which elite and commoner alike must know to act on, is counterfeit." (v1/368)
  • Reason: the Lawgiver was commanded to explain what people need, and they to transmit it. So if it was not widely transmitted, «عرفنَا أَنه سَهْو أَو مَنْسُوخ» — "we know it is a slip or abrogated." (v1/368)
  • Parallels: one townsman's sighting of the Ramaḍān crescent in a clear sky; a guardian claiming huge spending on an orphan in a short time. (v1/368)
• Furūʿ test: «وعَلى هَذَا الأَصْل لم نعمل بِحَدِيث الْوضُوء من مس الذّكر لِأَن بسرة تفردت بروايته» (v1/368) — "On this principle we did not act on the ḥadīth of wuḍūʾ from touching the penis, because Busra alone narrated it." Also wuḍūʾ from fire-cooked food and from carrying a bier (v1/368–369), and: «وعَلى هَذَا لم يعْمل عُلَمَاؤُنَا بِخَبَر الْجَهْر بِالتَّسْمِيَةِ وَخبر رفع الْيَدَيْنِ عِنْد الرُّكُوع وَعند رفع الرَّأْس من الرُّكُوع» — "On this, our scholars did not act on the report of saying the basmala aloud, or of raising the hands at rukūʿ and on rising from rukūʿ", because «لم يشْتَهر النَّقْل فِيهَا» — "its transmission did not become widespread", despite general need. (v1/369)
• Objection → Reply: "You accepted single reports making witr obligatory, and rinsing mouth and nose in ghusl, though these are general-need matters." Reply: «لِأَنَّهُ قد اشْتهر أَن النَّبِي فعله وَأمر بِفِعْلِهِ فَأَما الْوُجُوب حكم آخر سوى الْفِعْل» — "because it is widely known that the Prophet did it and commanded it; obligation is a further ruling beyond the act", which a few specialists could carry. The act is proven by wide transmission; only its obligatory status rests on the single report. (v1/369)
• [Editorial] Debate note: The defence turns on separating the act from its legal grade. Questions to prepare: does the distinction apply the same way to the cases Sarakhsī rejects under this filter? Answer from the text's own criterion ("its transmission did not become widespread"), not from outside claims.

Back to the list

B9d.05 · إعراض الصحابة · Filter 4: a report the Companions ignored while disputing the very question is suspect (v1/369–370)

• Position: «وَأما الْقسم الرَّابِع وَهُوَ مَا لم تجر المحاجة بِهِ بَين الصَّحَابَة مَعَ ظُهُور الِاخْتِلَاف بَينهم فِي الحكم فَإِنَّهُ زيف» — "The fourth: a report the Companions did not argue with, though they openly disagreed on the ruling, is counterfeit." (v1/369)
  • Reason: they were the roots of transmission and would not argue by raʾy while holding a report, so their silence shows it is a later slip, abrogated, or to be interpreted. (v1/369)
• Furūʿ test
  • «الطَّلَاق بِالرِّجَالِ وَالْعدة بِالنسَاء» ("divorce goes by the men, the ʿidda by the women"): the senior Companions disputed this and never cited it, so it is not established, or it means only that pronouncing divorce belongs to men. (v1/369)
  • «ابْتَغوا فِي أَمْوَال الْيَتَامَى خيرا كَيْلا تأكلها الصَّدَقَة» ("seek profit with orphans' wealth so zakāt does not consume it"): the Companions disputed zakāt on a child's wealth without citing it. (v1/369)
• [Editorial] Debate note: Filter 4 is an argument from silence. Its strength depends on showing that the dispute was real, that the report would have settled it, and that nobody cited it. Test those three facts before relying on it.

Back to the list

B9d.06 · المستور والفاسق · The unexamined narrator's report is not relied on "in our time"; the sinner's report is not proof in religion, though accepted in dealings of necessity (v1/370–371)

• Mastūr: ⚑ Muḥammad in al-Istiḥsān: like the sinner's report. al-Ḥasan from Abū Ḥanīfa: like the upright's ("Muslims are upright toward one another"), hence Abū Ḥanīfa allowed judging on a mastūr's testimony in what stands despite doubt, if the opponent does not object. «وَلَكِن مَا ذكره فِي الِاسْتِحْسَان أصح فِي زَمَاننَا فَإِن الْفسق غَالب فِي أهل هَذَا الزَّمَان» — "but what he said in al-Istiḥsān is sounder in our time, since sin prevails among the people of this age." Narration binds, so it needs uprightness shown by inquiry ("do not narrate from one whose testimony you do not know"). (v1/370) → B9b.04
• Fāsiq: al-Istiḥsān: on the purity of water or lawfulness of food, the hearer follows his best judgement of the sinner's report; ⚑ some of our shaykhs extend this to his narration. Sarakhsī: «وَالأَصَح عِنْدِي أَن خَبره لَا يكون حجَّة لِأَنَّهُ غير مَقْبُول الشَّهَادَة» (v1/370) — "the sounder, to me, is that his report is not proof, since his testimony is not accepted." Water and food are a particular necessity; for ḥadīth there are many upright narrators. (v1/370–371)
• Dealings: his report is accepted there of necessity, since it binds no one; lawfulness has a binding side, so there it needs the hearer's best judgement too. (v1/371)
• Objection → Reply: "the verify verse (Q 49:6) came down on al-Walīd b. ʿUqba's report, which was a dealing." Reply: it was a disowned report that a tribe had apostatised and meant to kill him, which would bind jihad against them. (v1/371)

Back to the list

B9d.07 · الكافر والصبي والمعتوه · The unbeliever's report counts not at all in religion; the child and imbecile are like the unbeliever, not like the sinner (v1/371–373)

• «فَأَما الْكَافِر فَإِنَّهُ لَا تعتمد رِوَايَته فِي بَاب الْأَخْبَار أصلا» (v1/371) — "As for the unbeliever, his narration is not relied on in reports at all", nor on water: if the hearer thinks he is truthful that it is impure, it is better to pour it out and then do tayammum, but tayammum without pouring it out does not validate the prayer. The sinner's report of impurity is different: tayammum without pouring is valid. (v1/371–372)
• Who: al-Istiḥsān adds "and likewise the child and the imbecile, if they understand what they say". ⚑ Some of our shaykhs: this joins them to the sinner. Sarakhsī: «وَالأَصَح أَن المُرَاد عطفه على الْكَافِر فَإِن الصَّبِي لَيْسَ من أهل الشَّهَادَة أصلا» — "the sounder is that it joins him to the unbeliever, since the child is not a witness at all": he binds others without binding himself. (v1/372)
• Who: some accept a child's narration, like a slave's, citing Ibn ʿUmar telling the people of Qubāʾ of the qibla change while under age. Reply: Anas brought the news (both may have come), or Ibn ʿUmar was already adult: he was turned away from fighting for weak build, and at fourteen one may be adult. (v1/372–373)
• The inattentive and the careless: if alertness dominates, they are like others; if inattention dominates, they are like the imbecile. (v1/373)

Back to the list

B9d.08 · صاحب الهوى · The heretic's testimony is accepted but his narration in religion is not, since partisanship may drive him to corrupt the way of truth (v1/373–374)

• His testimony is accepted, except the Khaṭṭābiyya (a Rāfiḍī group who testify for whoever swears to them) and those who hold inspiration (ilhām) to be a proof yielding knowledge. (v1/373–374)
• But «فَلهَذَا لَا تعتمد رِوَايَته وَلَا تجْعَل حجَّة فِي بَاب الدّين» — "hence his narration is not relied on and not made proof in religion": in disputes between people he has no reason to lie against someone of true belief, but in reports of religion partisanship may lead him to corrupt the path to truth. (v1/374) → B8.09

Back to the list

B9e.01 · أقسام الأخبار · Four kinds of report: certainly true, certainly false, equally balanced, and weighted; the upright narrator's report is the weighted-true kind (v1/374–375)

• Position: «هَذِه الْأَقْسَام أَرْبَعَة خبر يُحِيط الْعلم بصدقه وَخبر يُحِيط الْعلم بكذبه وَخبر يحتملهما على السوَاء وَخبر يتَرَجَّح فِيهِ أحد الْجَانِبَيْنِ» — "These kinds are four: a report known to be true, a report known to be false, one bearing both equally, and one in which one side outweighs." (v1/374)
  1. «أَخْبَار الرُّسُل المسموعة مِنْهُم» (v1/374) — "the Messengers' reports as heard from them", since they are protected from lying and their missions are proven by miracles: believe them and obey (Q 59:7). A prophetic report that reaches us through a narrator falls under kind 4.
  2. Pharaoh's claim to lordship, idols as gods, the claims of Zoroaster, Mani and Musaylima: believe them false and resist them as needed.
  3. The sinner's report in religion: wait until one side outweighs (Q 49:6).
  4. The sinner's testimony once a qāḍī rejects it, or the report of one punished for qadhf «عِنْد إِقَامَة الْحَد عَلَيْهِ» (v1/375) — "at the carrying out of the ḥadd on him": the lie side now outweighs, and it may not be acted on. (Whether he narrates after repenting is the separate question of B9c.06.) And the upright narrator's report in religion, where the truth side outweighs: this is the kind meant here. (v1/374–375)

Back to the list

B9e.02 · السماع · Hearing: the teacher reading or the student reading back are true hearing; writing and messenger are near it; permission (ijāza) is valid only when the content is known to the recipient (v1/375–379)

• Strict ways: (1) the teacher reads and you hear; (2) you read to him and he confirms. The ḥadīth scholars rank the first higher. Abū Ḥanīfa, reported: reading to the teacher is stronger when the teacher relates from a book (from memory, his own reading is stronger), since the Prophet's reading was protected from error and later teachers read from books, where the student reads with more care. Omissions by the teacher's inattention are easier to bear than a misreading. (v1/375–376) (3) A letter and (4) a messenger, as the Prophet conveyed by letters and envoys. Say "ḥaddathanī" for (1)–(2) and "akhbaranī" for (3)–(4); "God told us (akhbaranā)" is allowed, "God narrated to me" or "God spoke to me" is not said (speech was for Moses alone). al-Ziyādāt: an oath not to "speak of" a secret is not broken by a letter; an oath not to "tell" it is. (v1/376–377)
• Concessions: permission (ijāza) and handing over (munāwala), valid if the content is known and understood by the recipient and the teacher is accurate; he says "ajāza lī" or "akhbaranī", not "ḥaddathanī". (v1/377)
• Who: ⚑ permission when the recipient does not know the content. Some of our shaykhs: invalid for Abū Ḥanīfa and Muḥammad, valid for Abū Yūsuf, on the model of their split over the qāḍī's letter. Sarakhsī: «وَالأَصَح عِنْدِي أَن هَذِه الْإِجَازَة لَا تصح فِي قَوْلهم جَمِيعًا» (v1/377) — "the sounder, to me, is that this permission is invalid on all their views": Abū Yūsuf's allowance for letters was for secrecy, which ḥadīth books do not need. Making children who cannot yet discern or understand "hear" is a blessing people approved of, not a way to establish transmission (unlike valid reception in childhood by one who understood, B9c.02). Likewise one present at the session but busy reading another book, writing, chatting or dozing has not heard soundly enough to narrate; unavoidable lapses are pardoned out of necessity, but «فَأَما عِنْد الْقَصْد فَهُوَ غير مَعْذُور» (v1/378) — "when it is deliberate, he is not excused." A blanket permission for "all my hearings" is invalid by agreement; some later scholars allowed it, but «وَلَكِن فِي هَذِه الرُّخْصَة سد بَاب الْجهد فِي الدّين وَفتح بَاب الكسل» (v1/378) — "this concession closes the door of effort in religion and opens the door of laziness." (v1/377–378)
• Well-known books: one who understands them may say "so-and-so said" or "the school of so-and-so is", without "ḥaddathanī", since they are like mashhūr reports. Some ignorant ḥadīth scholars attacked Muḥammad's books because he learned much by study-circle, not by hearing; the Muwaṭṭaʾ is used the same way. (v1/378–379)

Back to the list

B9e.03 · الحفظ والأداء والتدليس · Memory and delivery have strict and concessionary forms; dropping a link is not blameworthy tadlīs unless done to fake a higher chain (v1/379–380)

• Memory: the strict form is keeping it from hearing to delivery, which was Abū Ḥanīfa's way in reports and testimony, "hence his narration was little"; the concession is relying on writing. Delivery: the strict form is the wording, the concession the meaning. (v1/379)
• Tadlīs: saying "so-and-so said" of one met but not heard from (al-Aʿmash and al-Thawrī did it; Shuʿba abhorred it). «وَالصَّحِيح القَوْل الأول» (v1/379) — "the sound view is the first": the Companions did it, and no Companion may be called a mudallis. Blameworthy tadlīs is dropping a narrator to fake a higher chain; dropping to shorten, or to assert firmly that the Prophet said it, is fine. One known to drop only reliable narrators may be narrated from. (v1/379–380)

Back to the list

B9e.04 · قول الصحابي أمرنا والسنة كذا · A Companion's "we were commanded" or "the sunna is" does not by itself mean the Prophet, against al-Shāfiʿī's old view (v1/380–381)

• Position: «فَالْمَذْهَب عندنَا أَنه لَا يفهم من هَذَا الْمُطلق الْإِخْبَار بِأَمْر رَسُول الله» — "Our position: this unqualified wording is not understood as reporting the Messenger's command" or sunna. (v1/380)
• Who: al-Shāfiʿī in his old view: it refers to the Prophet; so he took Saʿīd b. al-Musayyab's "the sunna" on separating a spouse who cannot provide maintenance, and on women's blood-money matching men's up to a third. In his new view: not without clarification (Mālik's "the sunna in our town" meant a local official's practice). On the blood-money case we did not follow Saʿīd, knowing he meant Zayd's sunna, and preferred ʿAlī and ʿAbd Allāh by sound qiyās. (v1/380)
• Proofs: command and prohibition come from others too ("those in authority", Q 4:59), and "the sunna of the rightly guided caliphs"; the Companions name the Prophet when they mean him (ʿUmar: "you were guided to your Prophet's sunna"; ʿUqba; Ṣafwān). (v1/380–381)

Back to the list

B9f.01 · إنكار الراوي · ⚑ When the original narrator denies or forgets a report: al-Shāfiʿī acts on it; "our scholars" did not; within the school, Abū Ḥanīfa and Abū Yūsuf against Muḥammad (v2/3–5)

• Four ways the narrator undermines a report: he denies it; he acts or rules against it; he fixes one of its possible meanings; or he leaves it altogether. (v2/3)
• Who: the ḥadīth scholars differ on denial. Suhayl forgot the one-witness-and-oath ḥadīth that Rabīʿa related from him ("Rabīʿa told me from me"): «وَقد عمل الشَّافِعِي بِالْحَدِيثِ مَعَ إِنْكَار الرَّاوِي وَلم يعْمل بِهِ عُلَمَاؤُنَا» (v2/3) — "al-Shāfiʿī acted on the ḥadīth despite the narrator's denial; our scholars did not." Likewise al-Zuhrī did not recognise "any woman married without her guardian's permission". ⚑ Muḥammad (with al-Shāfiʿī) acted on it; Abū Ḥanīfa and Abū Yūsuf did not; and "they said" the question should follow the school's split over a qāḍī who does not remember his own judgment (Abū Yūsuf rejects the proof of it, Muḥammad accepts it), and of the exchange between Abū Yūsuf and Muḥammad over three reports from Abū Ḥanīfa in al-Jāmiʿ al-ṣaghīr. (v2/3–4)
• Some of our shaykhs held that denial should void the report only on Zufar's view, by analogy with the husband who reports his wife's ʿidda ended while she denies it; «وَالْأول أصح» — "the first is sounder": the husband is believed as trustee of what is between him and his Lord. (v2/4)
• Proofs on both sides: acceptance: Dhū al-Yadayn (the Prophet accepted Abū Bakr and ʿUmar though he did not recall); ʿUmar accepted Anas on al-Hurmuzān's safety; forgetting is common. (v2/4) Rejection: ʿUmar did not accept ʿAmmār's tayammum report, which he did not recall; the original narrator's denial is his own proof, and a branch narrator may mistake whom he heard from. The Prophet recalled when told; ʿUmar may have recalled or taken caution. (v2/4–5)

Back to the list

B9f.02 · مخالفة الراوي لما روى · If the narrator rules or acts against his report after relating it, the report is no longer proof, being read as his knowledge of abrogation; his interpreting it does not bind others (v2/5–7)

• If his contrary act or ruling is before the report, or undated, the report stands (he changed his view on hearing it). If it is after: «فَإِن الحَدِيث يخرج بِهِ من أَن يكون حجَّة لِأَن فتواه بِخِلَاف الحَدِيث أَو عمله من أبين الدَّلَائِل على الِانْقِطَاع» (v2/6) — "the ḥadīth ceases to be proof, since his ruling or act against it is among the clearest signs of a break." The best reading is that he knew it abrogated. (v2/5–6)
• Furūʿ: Abū Hurayra's "wash seven times" and his fatwā of three: read as his knowing it abrogated, «أَو علم بِدلَالَة الْحَال أَن مُرَاد رَسُول الله النّدب فِيمَا وَرَاء الثَّلَاثَة» (v2/6) — "or knowing from the circumstances that the Messenger meant washing beyond three as recommended"; ʿUmar's "two mutʿas… I forbid them"; ʿĀʾisha's "without her guardian" and her marrying off her niece; Ibn ʿUmar's raising the hands at rukūʿ and Mujāhid's "I accompanied Ibn ʿUmar for years and he raised them only at the opening takbīr". (v2/6)
• Interpretation: «فَإِن ذَلِك لَا يمْنَع كَون الحَدِيث مَعْمُولا بِهِ على ظَاهره» (v2/6) — "that does not stop the ḥadīth being acted on in its apparent sense": his reading is his own ijtihād. Ibn ʿUmar read "the two parties have the option until they separate" as bodily separation; we do not follow it. al-Shāfiʿī likewise rejects Ibn ʿAbbās's exempting the apostate woman from "whoever changes his religion, kill him". (v2/6–7)
• Leaving it altogether is like acting against it (Ibn ʿUmar on raising the hands). (v2/7)
• [Editorial] Debate note: the raising-the-hands argument here rests on the Mujāhid report and on reading a narrator's later practice as knowledge of abrogation. Expect the opponent to contest the Mujāhid report; the argument stands or falls with it.

Back to the list

B9f.03 · الصحابة والأئمة يخالفون الحديث · A leading Companion's contrary practice voids a report he could not have missed; not one he may have missed; ḥadīth critics' vague criticism is not jarḥ (v2/7–11)

• Companions (ʿĪsā b. Abān): if a leading Companion who could not have missed it acts against it, it is read as abrogated or not binding: the rightly guided caliphs did not combine flogging with stoning, or flogging with exile (ʿUmar: "by God I will never exile anyone"; ʿAlī: "exile is enough of a trial"); ʿUmar left the Sawād to its people despite Khaybar's division. Ibn Masʿūd's taṭbīq was reading the knees-grip as a concession, not missing it. If he may have missed it, the report stands: Ibn ʿUmar on the menstruating woman's farewell ṭawāf; Abū Mūsā on laughter in prayer; Ibn ʿUmar's "no one performs ḥajj for another". (v2/7–8)
• Ḥadīth critics: «فَأَما الطعْن الْمُبْهم فَهُوَ عِنْد الْفُقَهَاء لَا يكون جرحا» — "Vague criticism is, for the jurists, not jarḥ", since uprightness holds by outward religion, especially in the first three generations; even a defendant's vague criticism of a witness is not jarḥ. (v2/9)
• Specific criticisms that are not jarḥ: that Abū Ḥanīfa planted his son to get hold of his teacher Ḥammād's books and narrated from them (if true, a sign of care, checking memory against the books); "tadlīs" by "ḥaddathanī fulān ʿan fulān"; concealing a narrator's name (al-Thawrī's "Abū Saʿīd"; Muḥammad's "the trustworthy told us"); Ibn al-Mubārak's refusal to narrate to Muḥammad: «فَإِن هَذَا إِن صَحَّ لم يصلح أَن يكون طَعنا» (v2/10) — "even if this were true, it would not be fit as criticism", since jurists stand as models and ascetics in seclusion, and each station suits different manners; and besides it is unsound, since Ibn al-Mubārak named Muḥammad as the reviver of his time; horse-racing (training for jihād); much joking, which is permitted so long as he says nothing untrue, on condition he is not a reckless, rash man habitually out to undo a proof and confuse; youth when hearing, on condition of accuracy when receiving as a child and when narrating after puberty (Ibn ʿAbbās, Ibn ʿUmar; hence the half-ṣāʿ ḥadīth of ʿAbd Allāh b. Thaʿlaba preferred for its better text); not being a habitual narrator; many furūʿ; and irsāl. (v2/9–11)
• Partisan criticism: specific jarḥ from one known for partisanship does not count: heretics against Ahl al-Sunna, «وَطعن بعض من ينتحل مَذْهَب الشَّافِعِي فِي بعض الْمُتَقَدِّمين من كبار أَصْحَابنَا فَإِنَّهُ لَا يُوجب الْجرْح لعلمنا أَنه كَانَ عَن تعصب وعداوة» — "and the criticism by some who claim al-Shāfiʿī's school against some of the great early members of our school does not establish jarḥ, since we know it came from partisanship and enmity." The valid grounds number about forty, in the books of jarḥ. (v2/11)
• [Editorial] Debate note: this node is Sarakhsī's defence of Abū Ḥanīfa and Muḥammad against the ḥadīth critics. Quote it as his position. He points to the jarḥ books for the valid grounds of jarḥ (about forty); the critics' charges he answers are the ones listed in the node.

Back to the list

B9g.01 · حقيقة التعارض · Revealed proofs do not truly conflict; conflict is in our knowledge (unknown dates); it needs two equal proofs, one time and one subject, each able to abrogate the other (v2/12–13)

• Position: «اعْلَم بِأَن الْحجَج الشَّرْعِيَّة من الْكتاب وَالسّنة لَا يَقع بَينهمَا التَّعَارُض والتناقض وضعا» — "The legal proofs of the Kitāb and Sunna do not conflict or contradict in themselves", since that would mark incapacity; «وَإِنَّمَا يَقع التَّعَارُض لجهلنا بالتاريخ» — "conflict arises only from our ignorance of the dates". (v2/12)
• Definition, pillar, condition: conflict is mutual blocking; its pillar is two equal proofs requiring opposites (lawful/forbidden); its condition is one time and one subject (marriage makes the wife lawful and her mother forbidden, without conflict), and each must be able to abrogate the other if dated. (v2/12–13)
• So conflict occurs between two verses, two readings, two sunnas, and a verse and a mashhūr sunna. «وَلَا يَقع التَّعَارُض بَين القياسين» — "there is no conflict between two qiyās", nor between Companions' opinions, since opinion does not abrogate opinion. (v2/13)

Back to the list

B9g.02 · حكم التعارض · Conflicting verses of unknown date send you to the Sunna; undated conflicting sunnas to Companion opinion or qiyās; between two qiyās the mujtahid chooses only when neither can be preferred, one being right and one wrong (v2/13–17)

• Two verses: seek the dates from the occasion of revelation; the later abrogates; «فَإِن لم يعلم ذَلِك فَحِينَئِذٍ يجب الْمصير إِلَى السّنة» (v2/13) — "if that is not known, then one must go to the Sunna." Two sunnas: go to what follows (a Companion's view or sound qiyās), as if the two did not exist; two equal proofs of marriage cancel each other. (v2/13)
• Two qiyās: prefer the one with a strength the other lacks, which works like knowing the date between texts; «وَإِن لم يُوجد ذَلِك فَإِن الْمُجْتَهد يعْمل بِأَيِّهِمَا شَاءَ» (v2/14) — "if that is not found, the mujtahid acts on whichever he wishes", not because both are right: «فَالْحق أَحدهمَا وَالْآخر خطأ على مَا هُوَ الْمَذْهَب عندنَا فِي الْمُجْتَهد أَنه يُصِيب تَارَة ويخطىء أُخْرَى» (v2/14) — "one is right and the other wrong, on our doctrine that the mujtahid is sometimes right and sometimes wrong", but he is excused, since his effort is right even if he misses (the believer's insight; searching for the qibla). Choice is only there because nothing lies after qiyās. And conflict between texts comes from our ignorance of their dates, and ignorance cannot ground a legal choice. (v2/14–15)
• Furūʿ: two vessels, one impure: search for drinking, but do tayammum for wuḍūʾ (a substitute exists); mixed slaughtered and dead carcasses only in necessity; never for mixed free and slave women. After acting on one qiyās, he is bound unless a stronger proof comes, like praying in one of two garments (unlike the qibla, which can move). A qāḍī's judgment in one case is fixed; in later cases he follows his settled view. One who divorced a specific wife (or freed a specific slave) and forgot which has no right of choice, since his ignorance gives none; one who divorced "one of them" unspecified does. (v2/14–17)
• Where proofs cancel and one falls back on the prior state: a donkey's leftover water (pure in origin, so use it and add tayammum); the hermaphrodite; the missing person (alive for his own estate, dead for inheriting). (v2/17)

Back to the list

B9g.03 · المخلص من التعارض · Resolve apparent conflict in order: by the proofs themselves, by the rulings, by circumstances, by an explicit date, then by an inferred date (v2/18–21)

• Position: «يطْلب هَذَا المخلص أَولا من نفس الْحجَّة فَإِن لم يُوجد فَمن الحكم فَإِن لم يُوجد فباعتبار الْحَال فَإِن لم يُوجد فبمعرفة التَّارِيخ نصا فَإِن لم يُوجد فبدلالة التَّارِيخ» — "The way out is sought first from the proof itself; failing that, from the ruling; then from the circumstances; then from an explicit date; then from an indication of date." (v2/18)
  1. The proof: firm over summary; Kitāb or mashhūr over a single report; a general open to specification is specified by the other (the theft verse and the mustaʾmin, Q 9:6; "pray it when you remember" and the three forbidden times). General practice by one of two texts shows it is later, or the other is interpreted or a narrator's slip. Consensus can also create a conflict (the aunt and maternal aunt "get nothing" vs "the maternal uncle inherits"; our scholars prefer the affirming text, al-Shāfiʿī the original state of no inheritance). (v2/18–19)
  2. The ruling: "what you knotted" (Q 5:89) is expiation in this world, "what your hearts earned" (Q 2:225) punishment in the next, so the ghamūs oath falls under the second only. (v2/19)
  3. Circumstances: "until they are pure" read yaṭhurna (after ten days) and yaṭṭahharna (bathing, under ten); "and your feet" read in the accusative (washing, feet bare) and genitive (wiping, feet in khuffs). (v2/19–20)
  4. Explicit date: Ibn Masʿūd on the pregnant widow ("the shorter Sūrat al-Nisāʾ came after the longer"). (v2/20)
  5. Inferred date: prohibiting reports come after permitting ones (lizard, hyena), because permission was the state before the prohibition. Some of our shaykhs ground this in permissibility being the default in things (as Muḥammad indicates in Kitāb al-Ikrāh); «وعَلى أقوى الطَّرِيقَيْنِ» (v2/20) — "and on the stronger of the two routes", permissibility was the visible practice before the Prophet's mission, in the interval between prophets, and lasted until our law proved the prohibition. So one abrogation is supposed instead of two; the prohibition also adds reward and sanction, which is the cautious course. (v2/20–21)

Back to the list

B9g.04 · المثبت والنافي · ⚑ Affirming vs denying reports: al-Karkhī prefers the affirming, ʿĪsā b. Abān sees a conflict; Sarakhsī: a denial based on evidence conflicts, a denial from mere absence of knowledge does not (v2/21–24)

• Who: ⚑ al-Karkhī: the affirming report is preferred, as testimony is accepted on affirmation, not negation. ʿĪsā b. Abān: they conflict, since what proves the one narrator's truth proves the other's. The early imams' practice varies: Maymūna (we took "in iḥrām"), Barīra's husband (we took "free"), Zaynab returned to Abū al-ʿĀṣ (we took "by a new marriage"), water purity reports (conflict), tazkiya (jarḥ preferred). (v2/21–22)
• Sarakhsī's reconciliation: «وَذَلِكَ الْوَجْه أَن خبر النَّفْي إِمَّا أَن يكون لدَلِيل يُوجب الْعلم بِهِ أَو لعدم الدَّلِيل الْمُثبت أَو يكون مشتبها» (v2/22) — "the way to reconcile them: a denial is either based on evidence that yields knowledge of it, or on the absence of affirming evidence, or unclear." The first conflicts with the affirmation (as testimony that the husband said no exception is accepted, al-Siyar al-kabīr); the second does not (it rests on the prior state, like the hearer's own ignorance); the unclear is questioned. (v2/22–23)
  • Hence: jarḥ beats tazkiya; water reports conflict; Maymūna conflicts (the muḥrim looks different), then Ibn ʿAbbās's precision decides; Zaynab, Barīra, Anas on qirān over Jābir on ifrād, and Ibn ʿUmar that the Prophet prayed in the Kaʿba over Bilāl's denial are all "affirmation over mere absence". (v2/23–24)
• [Editorial] Debate note: two live disputes (the Prophet's ḥajj type, prayer inside the Kaʿba) are settled here by this rule; Maymūna is not (both reports rest on evidence, so they conflict, and Ibn ʿAbbās's precision decides). When an opponent cites the denying report, the question to put is whether its narrator denied from evidence or only did not witness the act.

Back to the list

B9g.05 · الترجيح بالعدد والزيادة · No preference by number of narrators (Abū Ḥanīfa and Abū Yūsuf, against Muḥammad); an addition from the same narrator is taken; two narrators give two reports; preference is by proof, not by the rank of men (v2/24–26)

• Who: some theorists prefer the report with more narrators, or with free narrators (from al-Istiḥsān). Sarakhsī: «وَالَّذِي يَصح عِنْدِي أَن هَذَا النَّوْع من التَّرْجِيح قَول مُحَمَّد خَاصَّة» (v2/24) — "what is sound to me is that this kind of preference is Muḥammad's view alone" (al-Siyar al-kabīr: he took what two of the three regional schools of sīra agreed on); ⚑ Abū Ḥanīfa and Abū Yūsuf refused it, «وَالصَّحِيح مَا قَالَا فَإِن كَثْرَة الْعدَد لَا يكون دَلِيل قُوَّة الْحجَّة» (v2/24) — "and the sound view is theirs: a greater number is no sign of a stronger proof" (Q 12:21 and others). The Companions did not prefer by number, and testimony rules are not transferred to reports. The al-Istiḥsān rulings prefer two, or free men, only in people's rights; in rulings of the law one and two are equal. (v2/24–25)
• An addition: «فمذهبنا فِيهِ أَنه إِذا كَانَ الرَّاوِي وَاحِدًا يُؤْخَذ بالمثبت للزِّيَادَة» (v2/25) — "our doctrine: if the narrator is one, the version with the addition is taken", and its omission is put down to weaker retention ("if the parties to a sale differ while the goods remain, they swear and return"). ⚑ Muḥammad and al-Shāfiʿī: act on both. We say: one origin is certain, two reports only possible, and the addition must have a point. With two narrators, they are two reports, both acted on (the ban on selling food before taking possession, and the general ban). (v2/25–26)
• Who: ḥadīth scholars rank narrators in tiers for additions, for raised vs stopped, and for connected vs mursal. «وَلَكِن الْفُقَهَاء لم يَأْخُذُوا بِهَذَا القَوْل لِأَن التَّرْجِيح عِنْد أهل الْفِقْه يكون بِالْحجَّةِ لَا بأعيان الرِّجَال» — "but the jurists did not take this view, since for jurists preference is by the proof, not by the persons of men." (v2/26)

Back to the list

Previous · All chapters · Next