Previous · All chapters · Next

B8 · باب بيان الحجة الشرعية وأحكامها · The binding proofs: the Qurʾān, mass-transmitted and well-known reports, and consensus (v1/277–321)

1. In short
2. Why it matters
3. B8.01 · الحجة والبينة والبرهان والآية والدليل والشاهد · The words for proof: some proofs give certainty, some only bind action
4. B8.02 · أصول الحجج · Three roots, the Qurʾān, the Sunna and consensus, with qiyās drawn from them; the certain proofs go back to hearing from the Prophet
5. B8.03 · الكتاب · The Qurʾān is known only by mass transmission, not by its inimitability
6. B8.04 · المتواتر · A mass-transmitted report gives certain knowledge, and that knowledge comes at once, not by reasoning
7. B8.05 · المشهور · ⚑ A well-known report gives reassurance, not certainty, on the sound view, yet it can add to the Qurʾān
8. B8.06 · حجية الإجماع · The community's consensus gives certain knowledge, as an honour to it for the religion, not because many people cannot err
9. B8.07 · سبب الإجماع · A consensus may rest on revelation, on an inference from a text, or on opinion; a single report or qiyās can ground it; and it can reach us by single reports
10. B8.08 · ركن الإجماع · ⚑ Consensus forms when all agree, or when one view spreads and the rest stay silent after it reaches them and time to reflect has passed
11. B8.09 · أهلية الإجماع · Every scholar able to do ijtihād counts, if free of heresy and open sin; in rulings that need expertise, only such experts
12. B8.10 · شرط الإجماع · A consensus need not wait for its generation to die out; ⚑ a lone dissenter blocks it only if the others accepted his reasoning
13. B8.11 · حكم الإجماع · Consensus has grades: the Companions' explicit consensus is certain and its denier an unbeliever; lower grades bind without that

In short

• Which proofs bind, and how surely [Overview]? The roots are the Qurʾān, Sunna and consensus; qiyās is drawn from them [B8.02]. A mass-transmitted report gives certainty [B8.04]; on the sounder view a well-known report gives reassurance, yet can add to the Qurʾān [B8.05]. Consensus binds as an honour to the community [B8.06].

Back to the list

Why it matters

• Only mass transmission makes words part of the Qurʾān, so a prayer recited with words that only Ibn Masʿūd read is invalid [B8.03]. And once the Companions settle on a set of views, no one after them may invent another [B8.11].

Back to the list

B8.01 · الحجة والبينة والبرهان والآية والدليل والشاهد · The words for proof: some proofs give certainty, some only bind action

• In plain words: What do the Arabic words for "proof" tell us? Sarakhsī explains six. Ḥujja (proof) comes from "he overcame", or from "returning", since in practice one must turn to it. Bayyina (clear evidence) makes the binding force plain to the heart, whether it gives certain knowledge or less. Burhān is used like ḥujja. Āya (sign), when the law uses it without qualification, means what gives certainty, like the prophets' miracles. Dalīl is a proof that brings out what was hidden, and shāhid (witness) is like it, certain or not, as with witnesses in court.
• Who disagrees: An objector says some denied the prophets after seeing miracles, so miracles cannot give certain knowledge. Sarakhsī answers that miracles give knowledge through fair reflection, not by force; the deniers turned away from reflecting, or denied out of obstinacy.
• Deeper: v1 B8.01

Back to the list

B8.02 · أصول الحجج · Three roots, the Qurʾān, the Sunna and consensus, with qiyās drawn from them; the certain proofs go back to hearing from the Prophet

• In plain words: What are the sources of the law, and how sure is each? Sarakhsī says the roots are three: the Qurʾān, the Sunna and consensus (ijmāʿ). Qiyās, the fourth, is the meaning drawn out of these three. Proofs are of two kinds. Some give certain knowledge. Others only require action, without certainty. The certain ones are four: the Qurʾān, the Sunna heard directly from the Prophet by those present, the Sunna mass-transmitted from him, and consensus. Their root is hearing from the Prophet, who is protected from falsehood. Consensus binds because he said God would not unite his community on error.
• In his words: "Know that the roots of the legal proofs are three: the Kitāb, the Sunna and ijmāʿ; the fourth root, qiyās, is the meaning drawn out of these three": «ثمَّ اعْلَم بِأَن الْأُصُول فِي الْحجَج الشَّرْعِيَّة ثَلَاثَة الْكتاب وَالسّنة وَالْإِجْمَاع وَالْأَصْل الرَّابِع وَهُوَ الْقيَاس هُوَ الْمَعْنى المستنبط من هَذِه الْأُصُول الثَّلَاثَة» (v1/279)
• Deeper: v1 B8.02

Back to the list

B8.03 · الكتاب · The Qurʾān is known only by mass transmission, not by its inimitability

• In plain words: How do we know which words are Qurʾān? Sarakhsī says only by mass transmission: the Qurʾān is what is written in the muṣḥafs and has reached us by mass transmission in the seven well-known modes of reading. Inimitability is no proof: less than a verse is not inimitable, yet is Qurʾān. The basmala, written in every muṣḥaf, is Qurʾān. al-Rāzī reports the school's sound view: it is a verse revealed on its own to separate the sūras, not part of any sūra.
• Example: Words that only Ibn Masʿūd read are not Qurʾān, so a prayer recited with them is invalid. His "three consecutive days" for the oath expiation is still acted on, as a report from the Prophet.
• Who disagrees: Two disputes turn on inimitability. ⚑ For the least recitation in prayer, Abū Yūsuf and Muḥammad require three short verses or one long one. Abū Ḥanīfa accepts one short verse, though stopping there is disliked. ⚑ Abū Ḥanīfa allowed reciting in Persian in prayer, though it is disliked, since the meaning too is inimitable; the Ṣāḥibān allow it only for one who cannot recite Arabic.
• Deeper: v1 B8.03

Back to the list

B8.04 · المتواتر · A mass-transmitted report gives certain knowledge, and that knowledge comes at once, not by reasoning

• In plain words: Can a report ever give certain knowledge? Sarakhsī, with most jurists, says a mass-transmitted report (mutawātir) does. At every stage it is passed on by a group so large and widespread that their agreeing on a lie cannot be imagined. Examples are the number of rakʿas and the amounts of zakat. People with different aims could agree only on what they all heard. The Ḥanafī scholars add that this knowledge is immediate, like what we see, not reached by reasoning.
• Example: A qāḍī learns by mass report that Zayd's property belongs to ʿAmr. Sarakhsī gives two answers: he may rule on that knowledge; or he still waits for witnesses, since judging has conditions beyond knowledge.
• Who disagrees: Some say mass reports give only reassurance, citing the Jews' and Christians' report that ʿĪsā was killed, a mass report that is false. Sarakhsī answers that that report goes back to a few people in one house, so it was never mass-transmitted. al-Shāfiʿī's followers say the knowledge is certain but reasoned; he answers that a child knows his parents by report without reasoning.
• Deeper: v1 B8.04

Back to the list

B8.05 · المشهور · ⚑ A well-known report gives reassurance, not certainty, on the sound view, yet it can add to the Qurʾān

• In plain words: What about a report that began with one or a few narrators, then was received by the scholars with acceptance and practice? This is the well-known report (mashhūr), like the reports on wiping over khuffs or the ban on mutʿa. ⚑ al-Rāzī counts it as a kind of mass-transmitted report that gives certainty, since scholars of every kind could only have agreed to accept it if its narrators were truthful. ʿĪsā b. Abān says it is not mass-transmitted, and Sarakhsī calls his view the sound one. It rests at its root on one or a few narrators. So a doubt of falsehood remains, and it gives reassurance, and one who denies it is not called an unbeliever. Even so, it can add a rule to a Qurʾān verse, which for the Ḥanafīs counts as abrogation (B11.08), because the scholars' acceptance of it is itself a binding proof.
• Example: Wiping over khuffs is known by such a report. One who denies it is wrong, and one fears for him in sin, but he is not an unbeliever.
• Deeper: v1 B8.05; B9d.02

Back to the list

B8.06 · حجية الإجماع · The community's consensus gives certain knowledge, as an honour to it for the religion, not because many people cannot err

• In plain words: Why does the community's consensus bind with certainty? Sarakhsī, with the jurists and most theologians, says it is an honour God gave this community for the religion. It is not because large numbers cannot err: Jews, Christians and Zoroastrians outnumber us and agreed on error. The Qurʾān calls this community "witnesses over the people" (Q 2:143), beside the Messenger's witness. The Prophet said his community would not unite on error. And since he is the last prophet, his law can last to the Hour only if the community is protected from agreeing on error. How certain a particular consensus is depends on its kind and on how it reached us (B8.07, B8.11).
• Who disagrees: al-Naẓẓām and some Imāmīs say consensus is only fallible people added together. Sarakhsī answers that joining together brings something new, as people together carry a beam none can carry alone. Here the new thing is the proof that united them, not their number.
• In his words: "Whoever denies that consensus is a proof yielding knowledge has destroyed the root of the religion": «وَمن أنكر كَون الْإِجْمَاع حجَّة مُوجبَة للْعلم فقد أبطل أصل الدّين» (v1/296)
• Deeper: v1 B8.06

Back to the list

B8.07 · سبب الإجماع · A consensus may rest on revelation, on an inference from a text, or on opinion; a single report or qiyās can ground it; and it can reach us by single reports

• In plain words: Must a consensus rest on a certain text? Sarakhsī says no. It may rest on revelation, on an inference from a text, or on opinion. It can also reach us by mass transmission, by fame, or by single reports. A consensus reported by single reports binds action only, as the Prophet's own word does when it reaches us that way.
• Example: In the wine punishment, the whipping came from the Sunna. The number eighty came from what was done at the Prophet's time (forty men, each striking with both sandals). Only the change from sandals and palm stalks to lashes was by opinion, taking lashes from the punishment for slander. So eighty lashes became the agreed punishment.
• Who disagrees: Ibn Jarīr said a consensus that gives certainty cannot come from a single report or qiyās, since what rests on an uncertain proof cannot itself be certain. Sarakhsī calls this a clear error: consensus is a proof in itself, not by virtue of its basis.
• Deeper: v1 B8.07

Back to the list

B8.08 · ركن الإجماع · ⚑ Consensus forms when all agree, or when one view spreads and the rest stay silent after it reaches them and time to reflect has passed

• In plain words: Must every scholar speak for a consensus to form? The strict form is when all agree in word or deed. The concession, which the Ḥanafīs accept, is when a view spreads from some scholars and the rest stay silent, neither dissenting nor refuting it. Silence counts only after the fatwā was put to them, or became widely known, and a period for reflection has passed. After that, none of the silent scholars may retract by a new opinion.
• Example: Sharecropping (muzāraʿa) spread after Abū Ḥanīfa while his students no longer censured those who practised it. That was no consensus, since their dissent had been argued openly.
• Who disagrees: Some scholars say silence never makes a consensus, since silence may be awe or caution, and al-Shāfiʿī is reported to accept it only when most spoke. ⚑ Among the Ḥanafīs, al-Karkhī said a mujtahid may not censure another in matters of ijtihād, so silence shows no agreement. Sarakhsī answers that requiring all to speak means consensus would never form, and a mujtahid who disagrees may not stay silent.
• Deeper: v1 B8.08

Back to the list

B8.09 · أهلية الإجماع · Every scholar able to do ijtihād counts, if free of heresy and open sin; in rulings that need expertise, only such experts

• In plain words: Whose agreement makes a consensus? Sarakhsī says every scholar in each age able to do ijtihād (independent legal reasoning) who is not tied to heresy and not openly sinful. Expertise matters in rulings known only to scholars; there a theologian without uṣūl al-fiqh does not count. The number need not reach mass transmission; a group suffices. Consensus is not limited to the Companions, the people of Medina, or the Prophet's family.
• Example: A ḥadīth scholar with no insight into legal reasoning is not counted in a consensus on such rulings, being like a layman there.
• Who disagrees: ⚑ Some Ḥanafī shaykhs exclude a known heretic only on the issue he errs in. The Iraqi Ḥanafī shaykhs exclude a sinner even of sound belief, since a sinner is not fit to testify. Sarakhsī's sounder view: one who displays his heresy or sins openly does not count at all. One who does not display them still counts, even if they are known, since he keeps the honour of the religion. But such a heretic does not count on the issue he errs in.
• Deeper: v1 B8.09

Back to the list

B8.10 · شرط الإجماع · A consensus need not wait for its generation to die out; ⚑ a lone dissenter blocks it only if the others accepted his reasoning

• In plain words: Must a consensus wait for its scholars to die before it binds? The Ḥanafīs say no: it is like a text, and once it forms no one may dissent by opinion. And does one dissenter block it? ⚑ al-Karkhī, with al-Shāfiʿī, said one or two dissenters block it. Sarakhsī calls sounder the view al-Rāzī indicated: if the others accepted the dissenter's reasoning as legitimate, there is no consensus without him; otherwise it stands.
• Example: Ibn ʿAbbās once allowed unequal exchange of the same kind (ribā al-faḍl), and the others rejected his view, so the consensus stood without him. So a qāḍī's ruling allowing one dirham for two is void, as Muḥammad says.
• Who disagrees: al-Shāfiʿī, among others, makes the death of the generation a condition, since a scholar may change his mind. ʿAlī and ʿUmar had agreed that an umm walad (a slave woman who bore her master's child) is not sold; then ʿAlī said "I thought to keep them slaves". Sarakhsī answers that ʿAlī meant only that the master's death does not free her until the heir frees her, not that she may be sold, so the agreement stood.
• Deeper: v1 B8.10

Back to the list

B8.11 · حكم الإجماع · Consensus has grades: the Companions' explicit consensus is certain and its denier an unbeliever; lower grades bind without that

• In plain words: Are all consensuses equally strong? Sarakhsī gives four grades. What the Companions agreed on is certain, like the Qurʾān and Sunna, so its denier is an unbeliever. When they differed on a set of views, no new view may be invented, but its denier is not an unbeliever. A later age's consensus is like a well-known report: its denier is not an unbeliever, yet it can abrogate an earlier ruling. ⚑ When one age disputed a question and a later age agreed, some deny it is consensus. The Ḥanafīs say it is, but like a single report it binds action, not knowledge.
• Example: The Companions differed over selling an umm walad; those after them agreed she may not be sold. ⚑ Sarakhsī's teacher said this is consensus only on Muḥammad's view, since only Muḥammad voids a qāḍī's ruling allowing the sale. Sarakhsī's more cogent view: it is consensus for all of them. Abū Ḥanīfa lets such a ruling stand only because this kind of consensus is itself disputed, so the question stays one of ijtihād, where a qāḍī's ruling is not undone.
• Deeper: v1 B8.11

Back to the list

Previous · All chapters · Next