How Reliable Are the Hadith?, A Short Look at the Honest Answer

A Question Worth Asking Carefully

Most Muslims grow up with two competing impressions of the hadith. On one side: the canonical collections, Sahih al-Bukhari, Sahih Muslim, the four Sunan, are substantially reliable, authenticated by the rigorous isnad (chain of transmission) methodology of the classical muhaddithun. On the other side: there is awareness, often unspoken, that fabrications existed, that the collections were compiled long after Muhammad’s death, and that disagreements among the canonical collections exist.

The question this document asks is narrower than either “the hadith are reliable” or “the hadith are unreliable.” It asks: how should an honest reader evaluate the canonical hadith corpus, and what tools does the historical-critical tradition provide for distinguishing more authentic material from less?

The argument here is not that the hadith should be wholesale rejected. That position, sometimes associated with the early-twentieth-century Western scholar Joseph Schacht, has been substantially moderated by subsequent academic scholarship, and it cannot be sustained in its strong form. The basic structure of Islamic ritual practice (the Five Pillars in essential form) is not what critical evaluation targets.

But neither can the orthodox claim of wholesale reliability be sustained against the evidence the Islamic tradition itself preserves. There is a principled middle position: the corpus contains both authentic and fabricated material, and a critical method can distinguish them with reasonable (though not certain) confidence. The key tool is one that historical-critical scholars use across all religious traditions: the criterion of embarrassment.

You can read this in twenty minutes. The argument operates from inside the Islamic tradition’s own admitted facts.

1. The 200-Year Gap

The six canonical Sunni collections (Sihah Sitta) were compiled approximately 200 years after Muhammad’s death:

Sahih al-Bukhari, d. 256 AH / 870 CE
Sahih Muslim, d. 261 AH / 875 CE
Sunan Abu Dawud, d. 275 AH / 889 CE
Jami al-Tirmidhi, d. 279 AH / 892 CE
Sunan al-Nasa’i, d. 303 AH / 915 CE
Sunan Ibn Majah, d. 273 AH / 887 CE

Earlier semi-systematic collections existed (Malik’s Muwatta in 795 CE, Ahmad ibn Hanbal’s Musnad in 855 CE), and there is physical-manuscript evidence for some earlier written hadith documentation (the Sahifa of Hammam ibn Munabbih, dated by Muhammad Hamidullah to roughly 670-680 CE, though this dating is contested by Juynboll and others).

The 200-year gap does not, by itself, establish that the canonical collections are wholesale fabrication. The transmission was not purely oral, the Sahifa tradition and the broader picture (Gregor Schoeler’s The Oral and the Written in Early Islam) show a hybrid system with written notes supporting oral transmission. The gap does, however, establish that the canonical collections cannot be treated as substantially reliable on the strength of the orthodox isnad methodology alone.

The early ban on writing hadith

The Islamic tradition itself records that the early community deliberately discouraged the writing of hadith. Reports include:

Sunan al-Darimi 449 (a report attributed to the Prophet declining to permit writing of hadith, grading contested by classical critics).
Reports of Umar ibn al-Khattab’s deliberation against systematic compilation, preserved in al-Khatib al-Baghdadi’s Taqyid al-‘Ilm.

Al-Baghdadi’s work presents both the prohibitionist and the permissionist traditions. The early Islamic community treated the question as contested. The orthodox tradition that came to depend on written hadith collections preserved reports of an early prohibition on the very practice that produced the corpus. Whether by constraint (the prohibitionist reports were too well-attested to suppress) or by integrity (the classical compilers preserved them as a matter of professional honesty), the preservation indicates the prohibitionist tradition was historical, not invented later.

This is the criterion of embarrassment applied for the first time in the document.

2. The Filtration Rate

Al-Bukhari is reported by al-Dhahabi in Siyar A’lam al-Nubala’ to have examined approximately 600,000 hadith and accepted only about 7,275 distinct hadith for inclusion in his Sahih. The orthodox interpretation notes that 600,000 includes many duplicate transmission events counted separately, complicating simple arithmetic, but even granting this, the filtration process necessarily eliminated a large fraction of hadith in circulation.

A rejection ratio this high requires explanation. The orthodox account is that al-Bukhari’s filter rejected fabrications while accepting authentic material. But the rigor of a filter establishes its selectivity, not its accuracy. A highly selective filter operating on isnad criteria, chain of transmission rather than content, can let plausible fabrications through (with credible-looking chains) while rejecting clumsy ones. The classical methodology was sophisticated, but it could not detect a well-constructed fabrication with a fabricated but credible chain.

This does not establish that any specific hadith is fabricated. It establishes that the orthodox apologetic claim, “if it is in Bukhari, it is reliable because the methodology was rigorous”, overclaims what selectivity can demonstrate.

3. The Common-Link Phenomenon

The scholar G. H. A. Juynboll documented a structural feature of the hadith corpus that is significant for evaluating reliability. When parallel isnads for the same hadith are diagrammed, they typically converge on a single transmitter. Above that point, multiple chains exist; below that point, the chain becomes a single strand back toward the Prophet. The “common link” is the point at which the transmission narrows.

Juynboll’s interpretation, that the common link is often the originator of the tradition, is contested by Harald Motzki and others who argue for a more moderate reading. But the structural fact is not in serious dispute: the isnad network frequently shows narrow transmission at early points, with multiplication of chains only at later points. This is not the signature of authentic mass transmission from the Prophet outward. It is consistent with originator-and-spread patterns where one transmitter passes a tradition to multiple students.

Motzki’s refined methodology, isnad-cum-matn analysis (combining chain analysis with content analysis), can sometimes push the origination of specific traditions back to the early first century AH, within the Companion generation. But this works for individual traditions, not for the corpus as a whole. The corpus contains both early-attested traditions and traditions whose common-link patterns suggest later generation.

4. The Internal Contradictions

The canonical hadith corpus contains substantial internal contradictions, documented by Ignaz Goldziher in the late nineteenth century and engaged by every serious modern scholar since. Hadith on virtually every major theological and political question of the late Umayyad and early Abbasid periods exist in conflicting forms, predestination versus free will, the createdness of the Quran, the status of Ali, legal rulings on disputed fiqh questions, biographical details about the Prophet.

The classical Islamic tradition was aware of this. It developed sophisticated methodology, mukhtalif al-hadith (“apparently contradictory hadith”), for handling contradictions. The methodology offers four resolutions: abrogation (one tradition supersedes the other), harmonization (the two address different cases), preference (choose the more strongly attested), and suspension of judgment.

What this framework accomplishes: it shows the muhaddithun were aware of contradictions and developed principled tools for them. Treating internal contradictions as automatic evidence of wholesale fabrication is methodologically naive. Many apparent contradictions are resolvable by the classical framework.

What this framework does not accomplish: it assumes the corpus is substantially authentic and provides tools for reconciling apparent inconsistencies within an assumed-authentic corpus. It does not address the prior question, whether the corpus accurately preserves what Muhammad said and did. The pattern of contradictions across theological-political disputes (Goldziher’s documented pattern) is more economically explained by partisan generation during the early Islamic disputes than by Prophetic transmission of contradictory material requiring centuries of harmonization.

5. The Criterion of Embarrassment

The principled middle position rests on a standard tool of historical-critical scholarship: the criterion of embarrassment.

The principle: when a religious tradition preserves material that embarrasses its own theological position, the preservation indicates either that the material was too well-attested to suppress, or that the methodology of preservation operated with sufficient integrity to preserve uncomfortable material despite the pressure of orthodoxy. Either mechanism legitimates the inference: the preserved material is more likely authentic than other material in the corpus.

This is not a Christian apologetic tool. It is a standard tool deployed in New Testament historical-Jesus scholarship by John P. Meier (A Marginal Jew, 1991), Bart Ehrman (Did Jesus Exist?, 2012), and Dale Allison (Constructing Jesus, 2010). It is applicable to any religious tradition’s source material, including the hadith corpus.

The orthodox Muslim response, articulated by Jonathan Brown in Misquoting Muhammad (Oneworld, 2014), is that the muhaddithun preserved embarrassing material because their methodology was honest, they explicitly warned that “lying in the name of the Prophet is among the most serious of sins” and rejected vast quantities of material on that basis. They preserved embarrassing material as a matter of professional integrity.

Brown’s response is correct, and it is fully compatible with the criterion of embarrassment. The criterion does not require that the muhaddithun were trying to suppress and failed. It requires only that material preserved despite discomfort is more likely authentic than material that smoothly serves doctrinal interests. The integrity of the muhaddithun is the mechanism by which the embarrassing material was preserved; it is not a refutation of the inference about which preserved material is most likely authentic.

6. The Criterion Applied, Examples

The criterion identifies a set of canonical hadith and tafsir attributions that carry significant historical weight precisely because the orthodox tradition would have preferred to suppress them, or that the muhaddithun’s methodology preserved as a matter of integrity despite their awkwardness.

Uthman’s standardization of the Quran (Bukhari 4986-4987)

The narration of Uthman’s standardization of the Quranic text, with the destruction of variant codices, is graded sahih in the most authoritative Sunni collection. Multiple companion attestations (Zayd ibn Thabit, with branching transmission through Anas ibn Malik, Kharija ibn Zayd, and others) provide independent confirmation.

This narration embarrasses the orthodox doctrine of perfect Quranic preservation: it admits variant codices existed, that they were destructible enough to require destruction, that the standardization was an editorial act under political authority. The fact that this narration is preserved at the highest grade of authenticity, by multiple independent transmission lines, is exactly the criterion-of-embarrassment pattern. It is unlikely to have been fabricated by partisans of perfect-preservation doctrine. It is more likely to be historical.

The classical commentary tradition (Ibn Hajar al-Asqalani’s Fath al-Bari) reads this hadith as testifying to Uthman’s wisdom in preventing community fracture. Even granting that interpretive frame, the historical facts the hadith records remain. The interpretive frame is downstream of the facts.

Umar’s testimony on the missing stoning verse (Bukhari 6829, Muslim 4194)

Umar ibn al-Khattab, second caliph, one of the ten promised paradise, publicly testified that a Quranic verse mandating stoning for adultery had been revealed, was recited and memorized by the early community, and was no longer in the written Quran. Umar’s status makes the testimony unimpeachable on orthodox grounds. The narration’s preservation in both Sahih al-Bukhari and Sahih Muslim indicates double-canonical attestation.

This narration embarrasses the orthodox doctrine of complete Quranic preservation. The orthodox response is the doctrine of naskh al-tilawa (abrogation of recitation while ruling persists), but this is a later doctrinal construct that resolves the embarrassment by relabeling it. The fact that the hadith is preserved at the highest possible grade, in both Sahihayn, by Umar himself, points to the criterion-of-embarrassment pattern: the muhaddithun preserved testimony to textual loss because they could not suppress it without compromising the integrity of their methodology.

The early prohibition on writing hadith

Already discussed in Section 1. The orthodox tradition that came to depend on written hadith collections preserved reports that the early community discouraged the practice. This is preserved-despite-embarrassment material.

The variant codices (Ibn Abi Dawud’s Kitab al-Masahif)

Ibn Abi Dawud (d. 928 CE) catalogued the personal codices of major Companions in the period before Uthman’s standardization, documenting substantial textual variations. Ibn Mas’ud’s codex omitted three surahs; Ubayy ibn Ka’b’s codex contained two surahs not present in the Uthmanic text. The cataloguing of these variants by a senior figure of the hadith-collection generation is preserved-despite-embarrassment.

The classical tafsir on Q 4:157, the variant substitutes

Al-Tabari, Ibn Kathir, al-Qurtubi, and al-Razi on Q 4:157 preserve multiple substitute candidates for the crucifixion, Judas Iscariot, Simon of Cyrene, a volunteer disciple, unnamed others. The inconsistency of substitute candidates embarrasses the orthodox claim that the substitution doctrine is a clear traditional reading. The classical tafsir tradition preserved the disagreement.

The classical tafsir on Q 3:55, the death-reading attributed to Ibn Abbas

Al-Tabari in Jami al-bayan preserves multiple transmitted readings of mutawaffika in Q 3:55. Among them is the reading attributed to Ibn Abbas (the prophet’s cousin and the most authoritative early exegete) that Allah caused Jesus to die before raising him. This reading is preserved alongside the orthodox-preferred no-death reading.

A methodological caveat: tafsir isnads tolerate more interpretive layering than hadith isnads. The Ibn Abbas attribution in classical tafsir is more contested than analogous attributions in Sahih hadith. This is best understood as supplementary evidence within a broader argument about Q 3:55, not as load-bearing on the same level as Bukhari 4986-4987. The death-reading is in the classical tradition; the attribution to Ibn Abbas is preserved; al-Tabari thought the reading was worth preserving as part of the diversity of transmitted opinion on the verse.

What these applications establish

In each case, the orthodox tradition preserves material that embarrasses developed orthodox doctrine. The classical methodology either could not suppress the material (because it was too well-attested) or chose to preserve it as a matter of methodological integrity (Brown’s account). Either way, the inference holds: these specific traditions are more likely to reflect early authentic material than the surrounding apologetic apparatus.

7. The Criterion Applied, Skepticism Where Appropriate

The mirror application: hadith that defend orthodox positions, particularly when those positions were contested in the formative period of hadith compilation, should be treated with appropriate skepticism.

Theological-dispute hadith taking sides in the late Umayyad and early Abbasid disputes (predestination, the createdness of the Quran, Ali’s status). Goldziher documented the pattern: each faction in these disputes possessed hadith favoring its position. The pattern is more economically explained by partisan generation than by Prophetic transmission of contradictory material.
Detailed legal hadith providing specific rulings on disputed fiqh questions, particularly when rulings align with specific madhhabs against rivals.
Hadith with common-link patterns at known Umayyad-era jurists, where the matn supports positions characteristic of that jurist’s school.
Specific Christological hadith generated in response to Christian-Muslim debates of the formative period.

The skepticism does not require claims of certainty about any specific hadith’s inauthenticity. The criterion is probabilistic. What the critical method requires is that the burden of proof shifts: hadith defending orthodoxy require positive evidence of their early origin, rather than being accepted on the strength of the isnad-based filter alone.

8. What This Means for Islamic Practice

The critical method, in its bounded application, does not undermine the core of Islamic ritual practice. The classical orthodox source theory distinguishes mutawatir hadith (mass parallel transmission attaining certainty) from khabar al-wahid (single-chain or few-chain transmission, attaining only probability). The essential elements of Islamic practice, the obligation of salah, the direction of qibla, the obligation of zakat, the basic structure of hajj and Ramadan, are established by tawatur transmission and by ‘amal mutawatir (continuous mass communal practice). These are largely impervious to isnad-based criticism.

The Maliki tradition (Imam Malik, d. 795 CE) developed an alternative source-theory: ‘amal ahl al-Madina, the living practice of the people of Medina, is the most authoritative form of sunnah transmission, more reliable than any single-chain written hadith. This is structurally similar to the critical method’s distinction between mass-attested practice and single-chain transmission. The Maliki approach shows the orthodox Islamic tradition has not unanimously rested on the canonical hadith corpus alone.

What the critical method targets:

Specific detailed elaboration of practice where the details depend on khabar al-wahid.
Legal rulings derived from single-chain traditions.
Biographical material attested only in single-chain or family-chain transmission.
Theological positions defended by hadith with common-link patterns suggesting late generation.
The contextualizing apparatus (asbab al-nuzul, exegetical chains) used to read Quranic verses in particular orthodox ways.

A Muslim reader can maintain the basic structure of Islamic practice, the tawatur-transmitted Five Pillars, while accepting the critical method’s evaluation of more specific material. The two are not in conflict.

9. What This Means for the Bigger Picture

The reframing of the hadith question from wholesale rejection to principled critical evaluation has direct implications.

First, specific hadith that meet the criterion of embarrassment, Bukhari 4986-4987 on the Uthmanic standardization, Umar’s testimony on the stoning verse, the variant codices in Ibn Abi Dawud, the multiple substitute candidates in classical tafsir on Q 4:157, the death-reading of Q 3:55 attributed to Ibn Abbas, can be used as historical evidence by anyone, including someone critical of broader orthodox claims. The orthodox interlocutor’s first move (“you cannot use Bukhari while questioning hadith”) is answered by the principled distinction: the embarrassing-to-orthodoxy material is exactly what the criterion identifies as most likely authentic.

Second, the orthodox apologetic apparatus that depends on khabar al-wahid traditions to defend specific theological positions cannot rest on the unmoderated reliability of the canonical corpus. Where these positions are contested, the burden of proof shifts to the orthodox apologist to establish that the supporting hadith reflect early authentic material rather than later partisan generation.

Third, the critical method is honest about its boundaries. The criterion of embarrassment does not establish certainty about any specific hadith’s authenticity. It is a probabilistic framework triangulated with multiple-attestation, common-link analysis, coherence, and historical plausibility. Specific authentications and rejections remain matters of scholarly judgment.

What the method offers is a principled way to engage the hadith corpus that does not require either the wholesale acceptance the orthodox tradition demands or the wholesale rejection some Western critiques attempted. The middle position is academically defensible, methodologically transparent, and consistent with the use of specific hadith as historical evidence where they embarrass orthodoxy.

10. The Choice

The situation, stated plainly:

The canonical Sunni hadith collections were compiled approximately 200 years after Muhammad’s death, from a transmission environment that included both written notes and oral transmission, after an early period of explicit prohibition on writing hadith.
The most rigorous classical compiler (al-Bukhari) rejected the vast majority of hadith he examined. The rigor of the filter establishes selectivity, not accuracy.
The corpus shows the common-link transmission pattern documented by Juynboll: narrow transmission at early points, with multiplication at later points.
The corpus contains substantial internal contradictions on theological-political questions of the formative period. The classical mukhtalif al-hadith framework can harmonize many but does not address the prior question of authenticity.
The canonical collections preserve material that embarrasses developed orthodox positions, including admissions of textual loss, variant codices, alternative substitution candidates, and a death-reading of Q 3:55. This material is preserved at the highest grades of authenticity by multiple independent transmission lines.

The honest reader, looking at this evidence, does not need to choose between wholesale acceptance and wholesale rejection. A principled middle position is available:

Use hadith that meet the criterion of embarrassment as historical evidence.
Treat hadith defending contested orthodox positions with appropriate skepticism.
Acknowledge that the basic structure of Islamic ritual practice rests on mass transmission and continuous communal practice that is not what the critical method targets.
Recognize that the orthodox apologetic apparatus often deploys hadith that do not meet the criteria of either embarrassment or robust common-link multiple-attestation, and that the burden of proof for these traditions is real.

This conclusion does not require leaving Islam, and it does not require accepting Christianity. It requires a methodologically transparent engagement with the hadith corpus that the Islamic tradition itself has the resources to support, through the Maliki ‘amal tradition, through the tawatur / khabar al-wahid distinction, through mukhtalif al-hadith, and through the integrity of the classical muhaddithun themselves.

The Muslim reader who has been told that the canonical hadith are uniformly reliable, or that they are uniformly unreliable, is invited to examine the evidence and apply the criterion of embarrassment in their own reading. The corpus itself is public. The classical methodology is documented. The honest answer is more interesting than either extreme.

A Note on Where to Go Next

If you want to engage this question seriously, start with Jonathan Brown, Hadith: Muhammad’s Legacy in the Medieval and Modern World (Oneworld, 2009) and Misquoting Muhammad (Oneworld, 2014). Brown is a contemporary Muslim scholar (Georgetown University) who defends a moderated version of traditional reliability while engaging the academic critique honestly. He acknowledges fabrications, the limits of isnad methodology, and the historical complexities, while defending the substantial reliability of the corpus when subjected to critical scrutiny.

Then read Harald Motzki, Analysing Muslim Traditions (Brill, 2010), for the strongest academic case that some traditions can be dated to the early first century AH using isnad-cum-matn methodology.

Then read Ignaz Goldziher, Muslim Studies Volume 2, for the original documentation of the partisan-generation pattern in the corpus.

The honest answer lies between these voices. You can engage it for yourself.

This is a condensed version of the longer argument in hadith-reliability-collapse.md, which engages the strongest orthodox and academic responses in fuller scholarly detail, Wael Hallaq’s critique of Schacht’s e silentio argument, Motzki’s isnad-cum-matn methodology, the Hammam Sahifa dating debate, Jonathan Brown’s integrity-versus-impotence response, the i’jaz-based versus hadith-based preservation distinction, the classical mukhtalif al-hadith framework, and the methodological selective-application concern. For the everyday reader, the points above are sufficient. The longer document is available for anyone who wants the full scholarly engagement.