Schacht, isnads and the dating of hadith: 8 answers

The objection begins with a fact: false hadith circulated. Political partisans and pious storytellers alike put sayings in the Prophet’s mouth. Nobody disputes this, and Muslim scholars were the first to say it. Ibn Sirin, who died in 110 AH (about 728 CE), is reported to have said that after the first civil strife scholars began to ask transmitters to name their sources.^[1] Abdullah ibn al-Mubarak is reported to have said that the isnad is part of the religion, and that without it anyone would say whatever he liked.^[2]

A whole literature grew out of that concern: collections of fabricated reports such as Ibn al-Jawzi’s Kitab al-Mawdu’at, and manuals that name the forgers and the motives behind them.^[3] So the first point of the critique is accepted. The question left open is how much of the material survives scrutiny, and whether the scholars’ own method for sorting it was sound. The next four entries take those questions in turn.

Ismail al Faruqi states the point plainly: Muslims know that reports of the Prophet’s practice were tampered with during the first generations of transmission, and so they qualify the authentic reports as ṣaḥīḥah (“verified”).^[4] He notes that the criticism examined the content as well as the chain: its language, its coherence with the Quran, and its agreement with known history.^[4]

The claim of a century or more of purely oral transmission is the one most easily tested. Reports record written material from the start. The Prophet ordered that a sermon after the conquest of Mecca be written down for a man of Yemen named Abu Shah.^[5] Ali is reported to have kept a sheet of rulings, and Abu Hurayra said that Abdullah ibn Amr wrote down more hadith than he did.^[6] The text attributed to Abu Hurayra’s student Hammam ibn Munabbih survives, and its contents appear in Ahmad ibn Hanbal’s Musnad in the same wording.^[7]

There is a real concession to make. Other reports record a prohibition on writing anything but the Quran, and early scholars disagreed about it.^[8] Much teaching was oral, and memory carried a great load. But Umar ibn Abd al-Aziz is reported to have ordered the gathering of hadith around the year 100 AH,^[9] and Malik’s Muwatta was composed in Medina in the second century AH. The great collections of the third century came from many cities, among them Bukhara, Nishapur and Sijistan, and they do not point to one Baghdad project under political pressure.

The objection holds that chains are late inventions, attached to sayings to make them look old. If so, the science built around them would be an exercise in decoration. In practice the scholars of jarh wa ta’dil (criticism and praise of transmitters) compiled biographical dictionaries that record the birth, death, teachers, travels and reputation of thousands of narrators.^[10] A narrator had to be upright and precise, and each link in a chain had to be able to have met the one before. A chain with a gap, or with a transmitter known to be careless, was graded weak.

They also tested the text itself (matn). A report that contradicts the Quran, a firmly established practice or an established fact was suspect however clean its chain looked.^[11] Modern research has weighed in on this. Harald Motzki studied early collections such as the Musannaf of Abd al-Razzaq and argued that much of the legal material can be traced to the first century AH.^[12] Other Western scholars remain more cautious, and Schacht’s own case rested mainly on legal traditions. The honest summary is that the dating of individual reports is debated, and that the blanket claim of late invention has lost ground.

Schacht’s working rule was that the best way to prove that a tradition did not exist at a certain time is to show that it was not used as a legal argument in a discussion that would have required it.^[13] Muhammad Mustafa al-Azami, in his book-length reply, points out that the rule collides with Schacht’s own thesis. Schacht held that the ancient schools of law resisted the hadith of the Prophet, so scholars who resisted them would hardly have quoted them.^[14]

The rule also fails a direct test. Azami compares Malik’s Muwatta’ with the recension of his younger contemporary al-Shaybani. Malik’s chapter on the times of prayer holds thirty reports, and al-Shaybani keeps only three. Several omitted reports go back to the Prophet, and in some chapters whole sections are missing.^[15] Nobody concludes that al-Shaybani invented a doctrine of ignorance. Elsewhere he cites a report he left out of his recension, so he knew it. Abu Yusuf writes that there are many reports on a matter and that he would have quoted very many if he had not feared making the book too long.^[16]

Azami adds a question of scale. Schacht asks the reader to believe that scholars across the Muslim world, with the transport and communications of the second century AH, agreed to forge matching material in widely separated cities.^[16] Silence can sometimes carry weight, when a text could not have avoided a report. The burden of showing that a particular silence is such a case rests on the one who makes the claim, and Azami’s survey of Schacht’s examples finds that burden unmet.^[14]

The strongest form of the argument says that chains improved over time, so a complete chain going to the Prophet is a sign of later invention. This has real support in one observation: a number of chains converge on a single transmitter, and some scholars, following Juynboll, read that point as the origin of the report, or its source of fabrication.^[17]

The difficulty is the logic. If a chain is incomplete it is suspect, and if it is complete it is suspect, so no chain could ever count as evidence. A claim that nothing could confirm is a stipulation, not a finding. The Muslim scholars also never treated a clean chain as enough. They looked for the same report arriving by separate routes from different cities and different teachers, and they graded reports with a gap as weak by that very fact.^[18] A forger would have to plant a matching text in several unconnected lines, and that is harder than planting one. Convergence on one transmitter may equally show a real collector who taught widely, which is how Motzki read it. For related discussion, see 5 Western scholars of Islam and where Muslims differ with them.

Schacht described a common link in several chains as a strong sign that a tradition began with that transmitter, and he built the claim on the pattern of branching chains.^[19] Azami notes that Schacht offers a single example, a report in al-Shafi’i’s Ikhtilaf al-hadith with three chains that meet at Amr ibn Abi Amr. In Schacht’s diagram the teacher Muttalib appears twice, which makes the report look as if it came to Amr by three routes. When the diagram is redrawn, and al-Shafi’i’s own comparison of the three students is read, only one channel remains: Jabir to Muttalib to Amr. One student had named the wrong authority, and al-Shafi’i says so.^[20]

The classical critics already knew the pattern. Al-Dhahabi explains that every Companion and every Successor held some knowledge that others did not, so a report with one narrator is not suspect for that reason alone. A trustworthy and accurate narrator who is alone in a report gives a sound but unusual (gharib) report. A narrator of lower grade who is alone gives a rejected (munkar) report. A narrator whose reports are mostly unconfirmed is abandoned.^[21] Azami adds an analogy: on Schacht’s logic, a journalist who collects news from many sources and publishes it would be accused of inventing it, since thousands of readers can name only that journalist as the source.^[21]

Azami also finds an inconsistency in how Schacht treats evidence. Schacht takes one unsourced notice in Ibn Sa’d, about the sale of the right of loyalty (wala’) over a freed slave by a man named Abu Ma’shar, as proof of general custom. The same author gives eight reports, with their sources, on the case of Barira, and Schacht dismisses all eight.^[22] A common link is a question to ask of a chain. Treated as a verdict, it needs more than one disputed diagram.

Schacht wrote that the most perfect and complete chains are the latest, and that early authors such as Malik and al-Shafi’i show chains still being improved.^[23] Azami answers that this reading mistakes a habit of abbreviation for a lack of information. The early lawyers state openly that they chose not to quote every authority available to them. Abu Yusuf says brevity stopped him, and al-Shafi’i says in places that he has heard an unbroken chain but cannot recall it at the time of writing.^[23]

The pattern is checkable. Authors sometimes gave no chain, or only the nearest authority, or a single route out of several, or the words “from a reliable man” for a person they elsewhere name. Azami’s test case is a report that Malik gives with a gap between the Successor and the Prophet. Al-Shafi’i has it with two complete chains through Jabir. Al-Tahawi notes that some pupils of Malik relate it with a full chain through Abu Hurayra, while the most reliable pupils relate it with the gap. Al-Shaybani, after quoting it, says there are many other reports on the subject from the Prophet.^[24] The fuller chains were therefore known in the lifetime of the authors who left them out.

Azami adds a point about method. Western studies of chains have leaned on biography and on law books that quote hadith for an argument, instead of the collections of hadith themselves. He compares this to learning physics from science fiction: some of it is useful and the picture stays incomplete.^[23]

The number 600,000 usually comes from a report that al-Bukhari selected his book from about that many hadith.^[25] It never meant 600,000 separate sayings that appeared from nowhere. Scholars counted each route as a separate hadith, so one saying with ten chains counted as ten. The Sahih itself holds a few thousand reports once the repeats are removed, and the exact count depends on the edition and the method of counting.^[26] Reports of the starting figure differ as well. Ismail al Faruqi and Lois Lamya al Faruqi, in The Cultural Atlas of Islam, give it as 300,000 reports from a thousand authorities, from which al-Bukhari accepted 7,275; removing the repeats leaves 2,602.^[27]

Nor is the figure of 99.5 percent spurious a measured result. It follows from counting chains as sayings and from the premise in the previous entry. The scholars did reject most of what circulated, and that rejection shows the method working. Muslims also grade hadith unequally: a report ranked sound has a different standing from a weak one, and the Quran stands apart because it was transmitted by a different route and does not depend on any single chain. For how chains are used in a comparable case, see 5 things a Catholic Bible says about 2 Peter.