Previous · All chapters · Next

B9 · باب الكلام في قبول أخبار الآحاد والعمل بها · Single reports: when one narrator's report binds, which narrators count, and how reports are tested (v1/321–v2/26)

1. In short
2. Why it matters
3. B9a.01 · حجية خبر الواحد · A single report from an upright narrator binds action in religion, but does not give certain knowledge
4. B9a.02 · أدلة حجيته · Why a single report binds: whoever must convey must be believed, and the Prophet reached people through single envoys
5. B9a.03 · خبر الواحد لا يوجب العلم · A single report gives reassurance, not certainty; reports on matters of belief bind the heart without being certain
6. B9a.04 · اشتراط العدد · A report needs no second narrator; the Companions' requests for a second witness were caution
7. B9a.05 · أقسام ما يكون خبر الواحد فيه حجة · Where a single report binds: religious rulings, disputes between people, dealings that bind no one, and cases in between
8. B9b.01 · أقسام الرواة · Narrators are known or unknown; a known narrator is either a jurist, or reliable but of little fiqh
9. B9b.02 · خبر الفقيه مقدم على القياس · A jurist Companion's report beats qiyās, whether it agrees with it or not
10. B9b.03 · رواية المعروف غير الفقيه إذا خالفت القياس · A reliable Companion who was not a jurist: his report yields to sound qiyās only if it contradicts it in every respect, the community did not accept it, and the door of reasoned opinion (raʾy) is closed
11. B9b.04 · رواية المجهول · A Companion known only from one or two reports: his report depends on how the early jurists received it
12. B9c.01 · شرائط الراوي · A narrator needs reason, accuracy, uprightness and Islam
13. B9c.03 · الضبط · Accuracy includes understanding the legal meaning; keeping the wording is better; relying on handwriting alone is disputed
14. B9c.06 · العبد والمرأة والأعمى والمحدود في القذف · Slaves, women and the blind narrate like anyone else; ⚑ one punished for slander may narrate after repenting, on the school's main view
15. B9d.S1 · الانقطاع صورة: مرسل الصحابي · A Companion's report with a missing link (mursal) is proof, by agreement
16. B9d.S2 · مراسيل القرون الثلاثة · A report with a missing link from the second and third generations is proof, against al-Shāfiʿī, on the soundest view unless its narrator is known to relate from unreliable people
17. B9d.S3 · مراسيل من بعد القرون الثلاثة · ⚑ After the first three generations, on the soundest view, a report with a missing link counts only from one known to relate solely from upright, reliable people
18. B9d.01 · الانقطاع معنى · A report can also be broken in meaning: by a stronger contrary proof, or by a flaw in its narrator
19. B9d.02 · مخالفة الكتاب · Test 1: a single report that contradicts the Qurʾān is not acted on
20. B9d.03 · مخالفة السنة المشهورة · Test 2: an isolated report that contradicts a well-known Sunna is set aside
21. B9d.04 · عموم البلوى · Test 3: a lone report on something everyone needed to know is suspect, though the grade of an act may still rest on a single report
22. B9d.05 · إعراض الصحابة · Test 4: a report the Companions never cited while openly disputing that very question is suspect
23. B9d.06 · المستور والفاسق · A narrator whose state is unexamined is not relied on "in our time"; a sinner's report is not proof in religion
24. B9d.07 · الكافر والصبي والمعتوه · An unbeliever's report counts not at all in religion; on the sounder view, a child's or an imbecile's is treated like the unbeliever's
25. B9d.08 · صاحب الهوى · A heretic's testimony is accepted, but his narration in religion is not
26. B9e.01 · أقسام الأخبار · Reports are of four kinds; an upright narrator's report is the kind where truth outweighs
27. B9e.02 · السماع · How a report is received: hearing the teacher, reading back to him, a letter or a messenger; permission only when the recipient knows the content
28. B9e.03 · الحفظ والأداء والتدليس · Memory and delivery have strict and lenient forms; dropping a link is blameworthy only when done to fake a higher chain
29. B9e.04 · قول الصحابي أمرنا والسنة كذا · A Companion's "we were commanded" or "the sunna is" does not by itself refer to the Prophet
30. B9f.01 · إنكار الراوي · ⚑ If the original narrator denies or forgets a report: al-Shāfiʿī acts on it; the Ḥanafīs did not, though Muḥammad did
31. B9f.02 · مخالفة الراوي لما روى · If a narrator rules against his own report after relating it, the report is no longer proof; his interpretation of it binds no one
32. B9f.03 · الصحابة والأئمة يخالفون الحديث · A leading Companion acting against a report he could not have missed voids it; vague criticism by ḥadīth critics does not count
33. B9g.01 · حقيقة التعارض · Revealed proofs never truly conflict; a conflict exists only in our knowledge, for want of dates
34. B9g.02 · حكم التعارض · When proofs conflict and their dates are unknown: from two verses go to the Sunna; from two sunnas to a Companion's view or qiyās; between two qiyās choose, only when neither can be preferred
35. B9g.03 · المخلص من التعارض · Resolve an apparent conflict in a fixed order: the proofs, the rulings, the circumstances, an explicit date, then an inferred date
36. B9g.04 · المثبت والنافي · ⚑ An affirming report against a denying one: a denial based on evidence conflicts with it; a denial from mere absence of knowledge gives way
37. B9g.05 · الترجيح بالعدد والزيادة · More narrators do not make a report stronger; an added detail from the same narrator is taken; jurists prefer by proof, not by the rank of men

In short

• This chapter asks when a report that is not mass-transmitted binds us [Overview]. From an upright narrator it binds action, but not certainty [B9a.01]. A jurist Companion's report beats qiyās [B9b.02]. Four tests can still set a report aside [B9d.01]. Conflicting proofs are resolved in a fixed order [B9g.03].

Back to the list

Why it matters

• These tests are why the Ḥanafīs did not act on the report that touching the penis breaks wuḍūʾ, or on raising the hands at rukūʿ [B9d.02, B9d.04].

Back to the list

B9a.01 · حجية خبر الواحد · A single report from an upright narrator binds action in religion, but does not give certain knowledge

• In plain words: Does a ḥadīth reported by one narrator, and not mass-transmitted, bind us? Sarakhsī says a single report from an upright narrator is a proof for acting on in religion, the view of the jurists of the major cities. But it does not establish certain knowledge. Two groups disagree from opposite sides: some deny it any force in religion, and some ḥadīth scholars say it gives certain knowledge (B9a.03).
• Who disagrees: Some, whose view Sarakhsī says does not count, deny it any force in religion. God forbids following what one does not know (Q 17:36), they say, and a report that may be false is not knowledge. Sarakhsī answers that acting on it is acting on knowledge in its apparent sense, though not certain (B9a.02).
• Deeper: v1 B9a.01

Back to the list

B9a.02 · أدلة حجيته · Why a single report binds: whoever must convey must be believed, and the Prophet reached people through single envoys

• In plain words: Why should one person's report bind anyone? Sarakhsī argues from the Qurʾān: God binds each one who knows to convey and not conceal (Q 2:159, 3:187), which would be pointless unless the hearer must accept. "A group from every party" who go out to learn and warn their people (Q 9:122) can be one man, on the sounder reading. The Prophet reached most people through single envoys and letters. And a qāḍī judges on witnesses though some doubt remains.
• Example: The Prophet ate what Salmān and Barīra brought him, relying on their word that it was lawful; he did not wait for revelation at every meal. So a single report can make food lawful to eat.
• Who disagrees: The deniers say accepting a fallible reporter without a miracle ranks him above the Prophet. Sarakhsī calls this a clear error: miracles were needed for certain knowledge of prophethood, which no single report claims to give.
• In his words: "In our view it is action on knowledge established on the apparent, though not certain": «فَإِنَّهُ عندنَا عمل بِعلم هُوَ ثَابت من حَيْثُ الظَّاهِر وَلكنه غير مَقْطُوع بِهِ» (v1/326)
• Deeper: v1 B9a.02

Back to the list

B9a.03 · خبر الواحد لا يوجب العلم · A single report gives reassurance, not certainty; reports on matters of belief bind the heart without being certain

• In plain words: Does a single report give certain knowledge? Some say yes, pointing to "inform them" in Muʿādh's ḥadīth, and to beliefs such as the torment of the grave that rest on such reports. Sarakhsī says they confuse the heart's calm and reassurance with certain knowledge; the possibility of lying remains. Reports on belief bind the heart to believe, which is like binding action or weightier, but that is not knowing for certain.
• Who disagrees: al-Naẓẓām said a single report with supporting signs, such as seeing the dead being washed, gives immediate knowledge that needs no reasoning. Sarakhsī calls this false: immediate knowledge does not differ between people, and it would make evidence and oaths pointless.
• In his words: "It is as if the difference between calm and reassurance of the heart, and certain knowledge, were hidden from this speaker": «وَلَكنَّا نقُول هَذَا الْقَائِل كَأَنَّهُ خَفِي عَلَيْهِ الْفرق بَين سُكُون النَّفس وطمأنينة الْقلب وَبَين علم الْيَقِين» (v1/329)
• Deeper: v1 B9a.03

Back to the list

B9a.04 · اشتراط العدد · A report needs no second narrator; the Companions' requests for a second witness were caution

• In plain words: Must a report, like testimony, have two narrators? Some said yes, pointing to Abū Bakr asking al-Mughīra for a second witness on the grandmother's share, and ʿUmar asking Abū Mūsā. Sarakhsī says no. Every believer is described as enjoining the right (Q 3:110), so each one's word in religion binds. Those requests were only caution, not a condition: Abū Bakr had a particular doubt, and ʿUmar's matter touched everyone. The number two in testimony is known only by text, and a report is not testimony.
• Example: ʿUmar accepted al-Ḍaḥḥāk's report on a widow's share of blood-money, and ʿAbd al-Raḥmān b. ʿAwf's on the plague, without a second witness; one narrator's report counts, so the widow inherited from her husband's blood-money.
• Who disagrees: ⚑ Some later shaykhs say ʿUmar would not have accepted Abū Mūsā alone, since narrators were many then. Sarakhsī calls the opposite view sounder, and says Muḥammad stated it.
• Deeper: v1 B9a.04

Back to the list

B9a.05 · أقسام ما يكون خبر الواحد فيه حجة · Where a single report binds: religious rulings, disputes between people, dealings that bind no one, and cases in between

• In plain words: Does one narrator's word count everywhere? Sarakhsī gives four domains. First, rulings of religion that can be abrogated: an upright narrator's report binds, with no number and no set form of words, because what counts is that truth outweighs. Second, binding claims between people: only with the rules of testimony, meaning number, its wording and the witness's capacity. Third, dealings that bind no one, like agency: any discerning reporter counts if the hearer believes him, upright or not. Fourth, cases that are partly binding, like telling an agent he is dismissed: ⚑ Abū Ḥanīfa requires number or uprightness; the Ṣāḥibān accept any discerning reporter.
• Example: The Ramaḍān crescent seen under cloud belongs to the first domain, so one upright report counts. The ʿĪd crescent concerns people's benefit, so it needs the rules of testimony.
• Who disagrees: ⚑ In punishments that doubt averts (ḥudūd), Abū Yūsuf, whose view al-Jaṣṣāṣ chose, accepts a single report, since testimony too is acted on there despite doubt. al-Karkhī does not, since its link to the Prophet holds a doubt, like qiyās. Sarakhsī gives both reasons and picks neither.
• Deeper: v1 B9a.05

Back to the list

B9b.01 · أقسام الرواة · Narrators are known or unknown; a known narrator is either a jurist, or reliable but of little fiqh

• In plain words: Does it matter who narrates a report? Sarakhsī sorts narrators into known and unknown. A known narrator is either known for fiqh and reasoned judgment, or known for uprightness, sound memory and accuracy but of little fiqh. His examples of the first are the Rightly Guided Caliphs, Zayd b. Thābit, Muʿādh, Abū Mūsā and ʿĀʾisha; of the second, Abū Hurayra and Anas. "Of little fiqh" compares them with the jurist Companions, and Sarakhsī insists it is no slight on them (B9b.03).
• Deeper: v1 B9b.01

Back to the list

B9b.02 · خبر الفقيه مقدم على القياس · A jurist Companion's report beats qiyās, whether it agrees with it or not

• In plain words: If a jurist Companion reports something that goes against qiyās, which wins? Sarakhsī says the report: it binds action and yields a strong likelihood, and qiyās is left for it. The doubt in a report lies only in its transmission, while the doubt in qiyās lies at its root: is this feature really the cause? The report states the ruling outright; the feature qiyās relies on says nothing. Hearing beats opinion.
• Example: ʿUmar left qiyās for Ḥamal b. Mālik's report on the payment (ghurra) owed for a foetus, so the payment follows the report, not qiyās.
• Who disagrees: Mālik, as Sarakhsī reports him, puts qiyās first, since qiyās rests on the Companions' consensus, and consensus is stronger than a single report. Sarakhsī answers that the early generations left qiyās for reports so often that such rulings were called "departed from qiyās".
• In his words: "What is doubtful at its root ranks below what is doubtful only in its route, once its root is certain": «وَمَا يكون الشُّبْهَة فِي أَصله دون مَا تكون الشُّبْهَة فِي طَرِيقه بعد التيقن بِأَصْلِهِ» (v1/339)
• Deeper: v1 B9b.02

Back to the list

B9b.03 · رواية المعروف غير الفقيه إذا خالفت القياس · A reliable Companion who was not a jurist: his report yields to sound qiyās only if it contradicts it in every respect, the community did not accept it, and the door of reasoned opinion (raʾy) is closed

• In plain words: What if a reliable non-jurist Companion, like Abū Hurayra, reports something against qiyās? Sarakhsī says: where it agrees with qiyās, or the community received it with acceptance, act on it. Sound qiyās is put first only when three things hold. The report contradicts it in every respect; the community did not accept it; and, in his own phrase, the door of reasoned opinion (raʾy) is closed. The reason is not his honesty but paraphrase: a non-jurist narrating by meaning may convey less than the Prophet's full meaning. He adds, "God forbid" that this belittles him.
• Example: The report on the tied-udder animal (muṣarrāt) orders a ṣāʿ of dates for its milk, whatever the amount. That contradicts sound qiyās in every respect, since damages are made good by a like or by value. So the buyer who returns the animal does not pay a ṣāʿ of dates.
• Who disagrees: An objector says Ibn ʿAbbās rejected Abū Hurayra's report by another text, not by qiyās. Sarakhsī answers that he argued by qiyās ("would you make wuḍūʾ from hot water?"), treating such a report as open to scrutiny where it contradicts qiyās.
• Deeper: v1 B9b.03

Back to the list

B9b.04 · رواية المجهول · A Companion known only from one or two reports: his report depends on how the early jurists received it

• In plain words: What about a Companion known only through one or two reports? Sarakhsī gives five cases. If the jurists accepted it, or stayed silent after it spread, he counts as well known. If they differed over it, the Ḥanafīs accept it. If they all rejected it, it is not acted on. If it neither spread nor was challenged, acting on it is not required, but allowed where it agrees with qiyās. In Sarakhsī's time such a report is not accepted until upright narrators accept it.
• Example: Maʿqil b. Sinān, one such Companion, reported that Barwaʿ bint Wāshiq, whose husband died without naming a dower, gets the fair dower. Ibn Masʿūd accepted it, ʿAlī rejected it, and the jurists of the next generation accepted it, so the report counts.
• Who disagrees: ⚑ Take the parallel case of a witness whose state is unexamined (mastūr). Abū Ḥanīfa allowed a qāḍī to judge on his testimony, in matters that stand despite doubt and when the other party does not object, without obliging him to. He lived when truthfulness prevailed. Abū Yūsuf and Muḥammad did not allow it before his uprightness is shown.
• Deeper: v1 B9b.04

Back to the list

B9c.01 · شرائط الراوي · A narrator needs reason, accuracy, uprightness and Islam

• In plain words: What must a narrator be? Sarakhsī names four conditions: reason, accuracy (B9c.03), uprightness and Islam. For reason, the law takes puberty without defect as its sign. Uprightness belongs to one whose religion outweighs his desire, who avoids major sins and does not persist in minor ones. For Islam, a general statement of belief is enough; detailed questioning is not required. An unbeliever's report is refused not for unbelief itself, but because hostility to the religion makes lying suspect.
• Example: A discerning child who heard a ḥadīth may relate it after puberty, and it counts, as with some Companions. While still a child he cannot bind others by it.
• Who disagrees: al-Shāfiʿī says that if the sinner's and the unexamined narrator's reports are not proof, an unknown narrator's is even less so. Sarakhsī answers that the unknown narrator of the first three generations is upright by the Prophet's own praise of those generations, unless something removes it. Even so, his report is proof only with support, such as the early jurists' acceptance (B9b.04). ⚑ Some Ḥanafī shaykhs require a detailed grasp of the faith; Sarakhsī calls that a clear hardship.
• Deeper: v1 B9c.01; B9c.02; B9c.04; B9c.05

Back to the list

B9c.03 · الضبط · Accuracy includes understanding the legal meaning; keeping the wording is better; relying on handwriting alone is disputed

• In plain words: What does a narrator's accuracy (ḍabṭ) require? Hearing fully, understanding the intended meaning, memorising it, and holding to it until relating it. It has an outer side, the wording, and an inner one, its legal meaning (fiqh). Most scholars, he reports, prefer keeping the wording but allow relating by meaning after good accuracy. Sarakhsī calls it sounder that paraphrase is never allowed for the Prophet's concise "comprehensive words", short sayings that carry a whole rule; ⚑ some Ḥanafī shaykhs allow it even there. This strict standard, he says, is why Abū Ḥanīfa narrated little.
• Example: On Abū Ḥanīfa's strict view, a witness may not testify from his handwriting on a deed unless he remembers the event. Sarakhsī follows the concession below.
• Who disagrees: ⚑ On relying on one's handwriting without remembering, Abū Ḥanīfa never allows it. Abū Yūsuf allows it for a qāḍī's sealed register and a ḥadīth book kept safe, and for a deed only while in the witness's own hands. Muḥammad allows it in all three once he is sure it is his hand. Sarakhsī goes with the concession, not with Abū Ḥanīfa's strictness, for ease on people.
• Deeper: v1 B9c.03; B9c.07; B9c.08

Back to the list

B9c.06 · العبد والمرأة والأعمى والمحدود في القذف · Slaves, women and the blind narrate like anyone else; ⚑ one punished for slander may narrate after repenting, on the school's main view

• In plain words: Do slavery, being a woman or blindness affect a narrator? Sarakhsī says no: the four conditions do not change with them, though they can bar or reduce testimony. Testimony imposes one's word on another. A report binds the reporter first, and the hearer binds himself by believing it.
• Example: The Companions relied on the reports of the Prophet's wives, and no one asked whether blind Companions related before or after losing their sight; their reports count.
• Who disagrees: ⚑ Take one punished for slander (qadhf) who has repented. On the school's main transmitted view (ẓāhir al-madhhab) he narrates like others, since the ban on his testimony is part of his punishment, and narration is not testimony. On al-Ḥasan's report from Abū Ḥanīfa he is not accepted, since the text calls him a liar (Q 24:13).
• Deeper: v1 B9c.06

Back to the list

B9d.S1 · الانقطاع صورة: مرسل الصحابي · A Companion's report with a missing link (mursal) is proof, by agreement

• In plain words: What if a Companion reports from the Prophet without saying whom he heard it from? Sarakhsī says a report can be broken in two ways: in form, when a link is missing (a mursal report), or in meaning (B9d.01). For the Companions there is no disagreement: their mursal reports are proof, since they heard from the Prophet or from each other, and they were truthful.
• In his words: "There is no disagreement among scholars that the Companions' mursal reports are proof": «وَلَا خلاف بَين الْعلمَاء فِي مَرَاسِيل الصَّحَابَة أَنَّهَا حجَّة» (v1/359)
• Deeper: v1 B9d.S1

Back to the list

B9d.S2 · مراسيل القرون الثلاثة · A report with a missing link from the second and third generations is proof, against al-Shāfiʿī, on the soundest view unless its narrator is known to relate from unreliable people

• In plain words: What about a Follower, or a man of the next generation, who reports from the Prophet with a link missing? The Ḥanafīs say it is proof. On the view Sarakhsī calls soundest, that holds unless the narrator is known to relate freely from people who are not upright (B9d.S3). The early authorities did it openly: al-Ḥasan said that when four Companions agreed on a ḥadīth, he sent it without a chain, naming no one between himself and the Prophet. So they held such reports to be proof, and their agreement is proof enough.
• Who disagrees: al-Shāfiʿī: not proof unless supported, for example by a connected chain, since proof depends on the narrator's qualities, unknown when he is unnamed. Sarakhsī answers that a trustworthy man who attests directly to the Prophet says more than one who names his source. ʿĪsā b. Abān even held the mursal stronger than a report with a full chain.
• Deeper: v1 B9d.S2

Back to the list

B9d.S3 · مراسيل من بعد القرون الثلاثة · ⚑ After the first three generations, on the soundest view, a report with a missing link counts only from one known to relate solely from upright, reliable people

• In plain words: What about a report with a missing link from someone who lived after the first three generations? ⚑ Sarakhsī gives three Ḥanafī views. al-Karkhī sees no difference between ages. ʿĪsā b. Abān accepts it from one famous as someone people take knowledge from, like Muḥammad b. al-Ḥasan; from one known only as a narrator, it waits until put before such a scholar. al-Rāzī's (al-Jaṣṣāṣ's) view, the soundest: the first three generations' mursal is proof unless the narrator is known to relate freely from unreliable people. After them, it counts only from one known to relate solely from upright, reliable people. The Prophet attested to the first three generations and foretold that lying would spread after them.
• In his words: "The soundest view here is what Abū Bakr al-Rāzī said": «وَأَصَح الْأَقَاوِيل فِي هَذَا مَا قَالَه أَبُو بكر الرَّازِيّ» (v1/363)
• Deeper: v1 B9d.S3

Back to the list

B9d.01 · الانقطاع معنى · A report can also be broken in meaning: by a stronger contrary proof, or by a flaw in its narrator

• In plain words: Can a report with a full chain still be set aside? Yes, Sarakhsī says, if it is broken in meaning. That happens in two ways: a contrary proof shows it cannot stand (four tests, B9d.02 to B9d.05), or its narrator falls short (B9d.06 to B9d.08). The first two tests work like setting local coins beside finer coins, which shows up the base ones.
• Who disagrees: Sarakhsī says al-Shāfiʿī looked only at the outward break of a missing link, while "our scholars build fiqh on effective meanings".
• Deeper: v1 B9d.01

Back to the list

B9d.02 · مخالفة الكتاب · Test 1: a single report that contradicts the Qurʾān is not acted on

• In plain words: What if a single report contradicts a verse? Sarakhsī says it is neither accepted nor acted on, whether the verse is general or specific, explicit or apparent. The Qurʾān is certain, while a single report's link to the Prophet carries doubt. This rests on two rules set out elsewhere. A general word is certain (B4.03). And a report that adds to what a verse requires counts as abrogating it; so the report of judging on one witness and an oath fails against Q 2:282 (B11.08).
• Example: The report that touching the penis breaks wuḍūʾ is not accepted. Q 9:108 praises those who love to purify themselves, which covers washing that involves touching, and an act that breaks wuḍūʾ cannot be purification; so the wuḍūʾ stands.
• Who disagrees: al-Shāfiʿī allows a single report to turn a verse to a figurative sense. Sarakhsī answers that the certain is not set aside for what carries doubt.
• Deeper: v1 B9d.02; B4.03; B11.08

Back to the list

B9d.03 · مخالفة السنة المشهورة · Test 2: an isolated report that contradicts a well-known Sunna is set aside

• In plain words: What if an isolated single report (gharīb) contradicts a well-known Sunna? Sarakhsī says it is cut off as regards acting on it, since the well-known report is stronger. He calls the Ḥanafī way the straight path: putting each proof in its rank. Reports not found in the Qurʾān or the well-known Sunna are still accepted and acted on. Only where there is no report at all do the Ḥanafīs turn to qiyās.
• Example: A claimant with one witness and his own oath does not win. That would contradict "proof is on the claimant and the oath on the one who denies".
• Who disagrees: ⚑ Saʿd's report forbids selling fresh dates for dried ones. Abū Ḥanīfa did not act on it, as against "dates for dates, like for like", so for him the sale is valid if equal in measure. Abū Yūsuf and Muḥammad share the principle but say that Sunna does not cover fresh dates, so Saʿd's report governs and the sale is not valid.
• Deeper: v1 B9d.03

Back to the list

B9d.04 · عموم البلوى · Test 3: a lone report on something everyone needed to know is suspect, though the grade of an act may still rest on a single report

• In plain words: What if only one person reports something that everyone needed to know? Sarakhsī calls it counterfeit. The Prophet was commanded to explain what people need, and they to pass it on; so if it did not spread, it is a slip or abrogated. But once an act itself is widely known, its grade can rest on a single report.
• Example: The Ḥanafīs did not act on the reports of saying the basmala aloud or raising the hands at rukūʿ, since they did not spread despite general need. Witr is different: that the Prophet did it and commanded it is widely known, and only its being wājib rests on single reports.
• Deeper: v1 B9d.04

Back to the list

B9d.05 · إعراض الصحابة · Test 4: a report the Companions never cited while openly disputing that very question is suspect

• In plain words: What if the Companions disputed a question and none of them cited a report that would settle it? Sarakhsī calls the report counterfeit. They were the roots of transmission and would not argue by opinion while holding a report, so their silence shows it is a slip, abrogated, or meant otherwise.
• Example: "Seek profit with orphans' wealth so zakāt does not consume it." The Companions disputed zakāt on a child's wealth without ever citing it. So it does not count as proof on that question.
• Deeper: v1 B9d.05

Back to the list

B9d.06 · المستور والفاسق · A narrator whose state is unexamined is not relied on "in our time"; a sinner's report is not proof in religion

• In plain words: What about a narrator whose uprightness has not been examined (mastūr), or a known sinner (fāsiq)? ⚑ Muḥammad treated the unexamined narrator like the sinner; al-Ḥasan reports Abū Ḥanīfa treating him as upright. Sarakhsī calls Muḥammad's view sounder in his own time, since sin prevails. Sarakhsī's sounder view: a sinner's report is not proof in religion, since his testimony is not accepted. In dealings that bind no one, like agency, it is accepted out of necessity.
• Example: A sinner says the water is impure. The hearer follows his own best judgment of the report; if he believes it, tayammum is valid even without pouring the water out.
• Who disagrees: ⚑ Some Ḥanafī shaykhs extend that water rule to a sinner's narration of ḥadīth. Sarakhsī answers that water is a particular necessity, while ḥadīth has many upright narrators.
• Deeper: v1 B9d.06

Back to the list

B9d.07 · الكافر والصبي والمعتوه · An unbeliever's report counts not at all in religion; on the sounder view, a child's or an imbecile's is treated like the unbeliever's

• In plain words: Can an unbeliever, a child or someone mentally deficient (maʿtūh) narrate? Sarakhsī says an unbeliever's narration is not relied on in religion at all. A child and an imbecile, even if they understand what they say, are joined to the unbeliever, not the sinner, since a child is not a witness at all.
• Example: An unbeliever says the water is impure and the hearer believes him. It is better to pour it out and then do tayammum; tayammum without pouring it out does not make the prayer valid.
• Who disagrees: ⚑ Some Ḥanafī shaykhs join the child to the sinner; Sarakhsī calls it sounder to join him to the unbeliever. Others accept a child's narration, citing Ibn ʿUmar, who told the people of Qubāʾ of the qibla change while young. Sarakhsī answers that Anas brought the news, or Ibn ʿUmar was already adult.
• Deeper: v1 B9d.07

Back to the list

B9d.08 · صاحب الهوى · A heretic's testimony is accepted, but his narration in religion is not

• In plain words: What about someone with a heretical belief (ṣāḥib al-hawā) who otherwise avoids what he thinks forbidden? Sarakhsī says his testimony is accepted, but his narration is not made proof in religion. In disputes between people he has no motive to lie against someone of true belief. In reports of religion, partisanship may lead him to corrupt the way to truth.
• Example: His testimony in a dispute counts. The exception is the Khaṭṭābiyya, a group who testify for any claimant who swears to them he is in the right, holding that a Muslim does not swear falsely.
• Deeper: v1 B9d.08; B8.09

Back to the list

B9e.01 · أقسام الأخبار · Reports are of four kinds; an upright narrator's report is the kind where truth outweighs

• In plain words: How sure can we be of any report? Sarakhsī names four kinds. Some are certainly true: the Messengers' reports as heard from them. Some are certainly false, like Pharaoh's claim to lordship. Some are evenly balanced, like a sinner's report in religion, where we wait. In the fourth, one side outweighs: a sinner's testimony once a qāḍī rejects it leans to lying, and an upright narrator's report in religion leans to truth. A prophetic report that reaches us through a narrator falls under this fourth kind: it leans to truth, which is enough to bind action.
• Example: The report of one punished for slander, at the time the punishment is carried out on him, leans to falsehood and may not be acted on.
• Deeper: v1 B9e.01

Back to the list

B9e.02 · السماع · How a report is received: hearing the teacher, reading back to him, a letter or a messenger; permission only when the recipient knows the content

• In plain words: How must a narrator have received a report? Sarakhsī's strict ways: the teacher reads and you hear, or you read and he confirms; a letter and a messenger come next. The concessions are permission to narrate (ijāza) and handing over a written copy (munāwala). These are valid only if the recipient knows and understands the content and the teacher is accurate. A blanket permission for "all my hearings" is invalid by agreement.
• Example: Children too young to understand who are made to "hear" gain a blessing, not a report they may later relate. One who spent the session chatting, or let himself doze, has not heard soundly and may not narrate it; only unavoidable lapses are excused.
• Who disagrees: ⚑ Some Ḥanafī shaykhs say permission without knowing the content is invalid for Abū Ḥanīfa and Muḥammad but valid for Abū Yūsuf. Sarakhsī's sounder view: invalid for all three, since Abū Yūsuf allowed it for a qāḍī's letter to another qāḍī, whose contents must stay secret, and ḥadīth books need no secrecy.
• Deeper: v1 B9e.02

Back to the list

B9e.03 · الحفظ والأداء والتدليس · Memory and delivery have strict and lenient forms; dropping a link is blameworthy only when done to fake a higher chain

• In plain words: Must a narrator keep every report in memory, and repeat its exact words? The strict way is to keep it in memory from hearing to delivery, which was Abū Ḥanīfa's way; the concession is relying on writing. Delivering the wording is strict; delivering the meaning is a concession. Dropping a narrator from the chain (tadlīs) is blameworthy only when done to fake a higher chain. Dropping one to shorten it, or to assert firmly that the Prophet said it, is fine.
• Example: One known to drop only reliable narrators may be narrated from; his reports count.
• Who disagrees: Shuʿba condemned saying "so-and-so said" of someone met but not heard from. Sarakhsī calls the first view, that it is allowed, the sound one, since the Companions did it and no Companion may be called a deceiver.
• Deeper: v1 B9e.03

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 refer to the Prophet

• In plain words: When a Companion says "we were commanded" or "the sunna is", does he mean the Prophet? The Ḥanafī position: this unqualified wording is not taken as reporting the Prophet's command. Commands also come from those in authority (Q 4:59), and there is "the sunna of the rightly guided caliphs". When the Companions meant the Prophet, they named him.
• Example: Saʿīd b. al-Musayyab said "the sunna" is that women's blood-money matches men's up to a third. The Ḥanafīs knew he meant Zayd's sunna and preferred ʿAlī's and ʿAbd Allāh's view by sound qiyās, so his words did not bind them there.
• Who disagrees: al-Shāfiʿī, in his old view, took such wording as the Prophet's; in his new view, not without clarification.
• Deeper: v1 B9e.04

Back to the list

B9f.01 · إنكار الراوي · ⚑ If the original narrator denies or forgets a report: al-Shāfiʿī acts on it; the Ḥanafīs did not, though Muḥammad did

• In plain words: What if the person a report is traced to later says he does not remember it? Sarakhsī notes four ways a narrator can undermine his own report: denying it, acting against it, fixing one of its meanings, or ignoring it. On denial or forgetting, al-Shāfiʿī acted on the report, and "our scholars" did not. ⚑ Within the school, Muḥammad acted on it; Abū Ḥanīfa and Abū Yūsuf did not.
• Example: Suhayl forgot the ḥadīth of one witness and an oath that Rabīʿa related from him, and later said "Rabīʿa told me from me". al-Shāfiʿī acted on it; the Ḥanafīs did not, so one witness with an oath does not win a case.
• Who disagrees: Those who accept it point to Dhū al-Yadayn: the Prophet accepted Abū Bakr's and ʿUmar's word though he did not recall. Those who reject it point to ʿUmar refusing ʿAmmār's tayammum report, which he did not recall, and answer that the Prophet recalled once told.
• Deeper: v1 B9f.01

Back to the list

B9f.02 · مخالفة الراوي لما روى · If a narrator rules against his own report after relating it, the report is no longer proof; his interpretation of it binds no one

• In plain words: What if a narrator later rules or acts against the report he related? If his contrary act came before the report, or its date is unknown, the report stands: he may have changed his view on hearing it. If it came after, the report is no longer proof; the best reading is that he knew it was abrogated. But if he only chose one of its possible meanings, that is his own reasoning, and the report is still acted on in its plain sense.
• Example: Abū Hurayra related "wash the vessel licked by a dog seven times" but gave a fatwā of three. The Ḥanafīs read this as his knowing it was abrogated, or that washing beyond three was only recommended; so three washes purify it.
• Deeper: v1 B9f.02

Back to the list

B9f.03 · الصحابة والأئمة يخالفون الحديث · A leading Companion acting against a report he could not have missed voids it; vague criticism by ḥadīth critics does not count

• In plain words: What if a leading Companion acts against a report? Sarakhsī, following ʿĪsā b. Abān, says: if he could not have missed it, the report is read as abrogated or not binding; if he may have missed it, the report stands. Criticism of a narrator (jarḥ) by ḥadīth critics must be specific: vague criticism does not count for the jurists. A narrator counts as upright on the strength of his outward practice of religion. Specific criticism from someone known for partisanship does not count either.
• Example: The rightly guided caliphs did not combine flogging with exile, and ʿUmar swore never to exile anyone. So the exile clause is read as abrogated, and the punishment is flogging alone.
• Who disagrees: Some ḥadīth critics attacked Abū Ḥanīfa and Muḥammad, for instance that Ibn al-Mubārak refused to narrate to Muḥammad. Sarakhsī answers that this, even if true, would not be fit as criticism, since a jurist's manners are a public model and an ascetic's are those of seclusion. He dismisses the criticism of great early Ḥanafīs by some who claim al-Shāfiʿī's school as partisanship.
• Deeper: v1 B9f.03

Back to the list

B9g.01 · حقيقة التعارض · Revealed proofs never truly conflict; a conflict exists only in our knowledge, for want of dates

• In plain words: Can two proofs from the Qurʾān and Sunna truly contradict each other? Sarakhsī says no, since that would mark incapacity in the Lawgiver. A conflict (taʿāruḍ) arises only from our ignorance of which came later. It needs two proofs of equal strength requiring opposite rulings, at one time and on one subject, each able to abrogate the other if dated. So two qiyās never conflict in this sense, nor do Companions' opinions, since opinion does not abrogate opinion.
• Example: Marriage makes the wife lawful and her mother forbidden, with no conflict, since the two concern different persons.
• In his words: "Conflict arises only from our ignorance of the dates": «وَإِنَّمَا يَقع التَّعَارُض لجهلنا بالتاريخ» (v2/12)
• Deeper: v1 B9g.01

Back to the list

B9g.02 · حكم التعارض · When proofs conflict and their dates are unknown: from two verses go to the Sunna; from two sunnas to a Companion's view or qiyās; between two qiyās choose, only when neither can be preferred

• In plain words: What do you do when two proofs conflict? For two verses, first look for their dates in the occasions of revelation; the later abrogates. Only if that is unknown do you go to the Sunna. For two sunnas of unknown date, go to what comes next, a Companion's view or sound qiyās, as if the two did not exist. Between two qiyās, prefer one with a strength the other lacks. Only if there is none does the scholar reasoning (mujtahid) act on either; even then one is right and the other wrong, though he is excused. Once he has acted on one, he is bound by it unless a stronger proof appears.
• Example: Two vessels, one of them impure. For drinking he searches and drinks from the one he judges pure. For wuḍūʾ he must do tayammum, since a substitute exists.
• Deeper: v1 B9g.02

Back to the list

B9g.03 · المخلص من التعارض · Resolve an apparent conflict in a fixed order: the proofs, the rulings, the circumstances, an explicit date, then an inferred date

• In plain words: How is an apparent conflict resolved? Sarakhsī gives an order. First the proofs themselves: the Qurʾān or a well-known report over a single report, or one text narrows the other. Then the rulings: each may concern a different thing. Then the circumstances: each may fit a different case. Then an explicit date. Last, an inferred date: a prohibiting report is taken as later than a permitting one, since permission was the earlier state. Some Ḥanafī shaykhs ground this in permission as the default in things; Sarakhsī calls it the stronger route that permission was the visible practice before the Prophet's mission.
• Example: "And your feet" is read two ways in the wuḍūʾ verse: washing when the feet are bare, wiping when they are in khuffs; so both readings stand.
• Deeper: v1 B9g.03

Back to the list

B9g.04 · المثبت والنافي · ⚑ An affirming report against a denying one: a denial based on evidence conflicts with it; a denial from mere absence of knowledge gives way

• In plain words: If one narrator affirms that the Prophet did something and another denies it, which wins? ⚑ al-Karkhī prefers the affirming report, as testimony is accepted on affirmation. ʿĪsā b. Abān sees a conflict, since whatever proves one narrator truthful proves the other too. Sarakhsī reconciles them. A denial based on evidence that yields knowledge conflicts with the affirmation. A denial resting only on not having seen gives way to it, and an unclear one is looked into.
• Example: When one man criticises a witness and another praises him, the criticism wins (jarḥ over tazkiya). The praise rests only on not knowing of a flaw; the critic affirms a flaw he found. So here the praise is the denial and the criticism the affirmation.
• Deeper: v1 B9g.04

Back to the list

B9g.05 · الترجيح بالعدد والزيادة · More narrators do not make a report stronger; an added detail from the same narrator is taken; jurists prefer by proof, not by the rank of men

• In plain words: Does a report with more narrators beat one with fewer? Some theorists say yes. Sarakhsī says that within the school this is Muḥammad's view alone; ⚑ Abū Ḥanīfa and Abū Yūsuf refuse it, and he calls their view sound. If one narrator gives a report both with and without an added detail, the version with the addition is taken. For jurists, preference is by the proof, not by the standing of the narrators.
• Who disagrees: ⚑ On the added detail, Muḥammad and al-Shāfiʿī act on both versions. Sarakhsī answers that there is certainly one report behind both versions, and the added detail must be there for a reason.
• In his words: "The sound view is theirs: a greater number is no sign of a stronger proof": «وَالصَّحِيح مَا قَالَا فَإِن كَثْرَة الْعدَد لَا يكون دَلِيل قُوَّة الْحجَّة» (v2/24)
• Deeper: v1 B9g.05

Back to the list

Previous · All chapters · Next