How the Quran You Hold Was Assembled, A Short Look at the History

A Question Most Muslims Are Never Encouraged to Ask

Most Muslims grow up hearing that the Quran in their hands is identical, letter for letter, to what Muhammad recited. The text, we are told, has been preserved perfectly by Allah, transmitted by mass memorization across thousands of independent witnesses, free from the corruption that affected earlier scriptures. Q 15:9 is cited: “Indeed, it is We who sent down the Quran and indeed, We will be its guardian.”

It is worth asking whether the history Islamic tradition itself records matches this claim.

This short document does one thing. It walks through what Sahih al-Bukhari, the most authoritative book of hadith in Sunni Islam, records about how the Quran was assembled, alongside what other major Islamic sources (Sahih Muslim, Ibn Abi Dawud’s Kitab al-Masahif, the classical uloom al-Quran tradition) add to the picture. Every fact below is from inside the Islamic tradition. The argument is not Christian. The sources are not hostile. They are the sources orthodox Muslim scholars themselves treat as authoritative.

The conclusion the evidence points to is that the Quran was not preserved as orthodox doctrine claims. It was assembled, twice, under political authority, from a fragmentary base, with competing copies burned. You can read this in twenty minutes and judge for yourself.

1. The First Compilation, After Yamama (~633 CE)

A year after Muhammad’s death, many of the huffaz (memorizers) were killed at the Battle of Yamama against Musaylima the false prophet. Umar ibn al-Khattab came to the first caliph Abu Bakr with a warning. Sahih al-Bukhari records the conversation:

Umar has come to me and said, “The casualties were heavy among the qurra [Quran-reciters] on the day of Yamama, and I am afraid that more casualties may occur among the qurra on other battlefields, by which a large part of the Quran may be lost. Therefore I suggest you order that the Quran be collected.”

, Sahih al-Bukhari 4986 (Volume 6, Book 61, Hadith 509). Grade: sahih.

Abu Bakr ordered Zayd ibn Thabit to collect the Quran. Zayd’s own account of what he did follows:

So I started locating the Quranic material and collecting it from parchments, scapula, leaf-stalks of date palms, and from the memories of men (who knew it by heart). I found with Khuzayma two verses of Surah al-Tawbah which I had not found with anybody else.

, Sahih al-Bukhari 4986.

Three things are admitted here by orthodox tradition’s own strongest source.

First, the Quran was not, in 633 CE, preserved as a single coherent written document. It was scattered across personal notes on bones, leaves, palm fragments, and the memories of individual companions. The compilation was a reconstruction.

Second, the deaths of memorizers at one battle were enough to make the second caliph fear “a large part of the Quran may be lost.” If the text had been mass-preserved across thousands of independent witnesses, as the later preservation doctrine claims, the loss of one battle’s qurra would not have caused this panic.

Third, two verses of Surah al-Tawbah (Q 9:128-129) were found with one man only, Khuzayma ibn Thabit. Zayd accepted them on a single witness. In Islamic terminology this is ahad transmission, single-chain, not mutawatir (mass-attested). The orthodox doctrine that the Quran is preserved by tawatur cannot survive the admission, in Bukhari, that two verses of the present text entered the compilation on the memory of one man.

Abu Bakr’s collected text was held by his successor Umar, then passed to Umar’s daughter Hafsa, one of Muhammad’s widows. This is the manuscript that Uthman would use twenty years later.

2. The Second Compilation, Uthman’s Standardization (~650 CE)

Twenty years later, under the third caliph Uthman ibn Affan, regional differences in how Muslims were reciting the Quran had become severe. Iraqi and Syrian troops fighting together in Armenia and Azerbaijan were arguing over whose recitation was correct. The Companion Hudhayfah ibn al-Yaman returned to Medina alarmed:

O Chief of the Believers! Save this nation before they differ about the Book (Quran) as Jews and Christians did before.

, Sahih al-Bukhari 4987.

Uthman responded by sending for Hafsa’s manuscript and appointing a committee of four, Zayd ibn Thabit, Abdullah ibn al-Zubayr, Sa’id ibn al-As, and Abdul-Rahman ibn Harith ibn Hisham, to produce a single standard text. Uthman’s instructions:

In case you disagree with Zayd ibn Thabit on any point in the Quran, then write it in the dialect of Quraysh, as the Quran was revealed in their tongue.

, Sahih al-Bukhari 4987.

When the committee finished, Uthman sent one copy to each Muslim province. Then comes the instruction that the orthodox tradition has spent fourteen centuries trying to absorb:

Uthman… ordered that all the other Quranic materials, whether written in fragmentary manuscripts or whole copies, be burnt.

, Sahih al-Bukhari 4987.

Read this carefully.

The variation among regional recitations was severe enough that a senior Companion feared communal fracture. Trivial differences in pronunciation do not produce this fear.

The standardization was an editorial act under political authority. A committee of four men, chosen by the caliph, three of them Qurayshi (Uthman’s own tribe), worked from one manuscript supplemented by witnesses available in Medina. The tie-breaking rule favored the Qurayshi dialect, Uthman’s tribal dialect.

All other Quranic materials were ordered burned. Not archived. Not preserved for cross-reference. Burned.

The Quran you hold is descended from this single standardized copy. The other materials, including the personal codices of major Companions, were destroyed.

3. The Codices That Were Burned

The standardization had targets. Several senior Companions had compiled their own masahif (codices), and what they contained is partially recoverable from Islamic literature even after their destruction. The most thorough early treatment is Ibn Abi Dawud’s Kitab al-Masahif (d. 928 CE).

Abdullah ibn Mas’ud

Ibn Mas’ud was not a peripheral figure. Muhammad himself named him as one of four authoritative reciters of the Quran:

Take (learn) the Quran from four: Abdullah ibn Mas’ud, Salim, Mu’adh, and Ubayy ibn Ka’b.

, Sahih al-Bukhari 3758. Grade: sahih.

His codex differed from the Uthmanic text in well-documented ways. It omitted Surah 1 (Al-Fatihah) and Surahs 113 and 114 (the Mu’awwidhatayn, the protection surahs). The surah order differed. Numerous textual variants existed in the body.

Ibn Mas’ud refused to surrender his codex for destruction. When he heard that Zayd ibn Thabit had been put in charge of the standardization, he is reported to have said publicly: “By Allah, I accepted Islam while he [Zayd] was still in the loins of an unbelieving man” (Ibn Abi Dawud, Kitab al-Masahif).

Ubayy ibn Ka’b

Ubayy, another of the four authoritative reciters Muhammad himself named, possessed a codex that contained two additional surahs not present in the Uthmanic text, Surat al-Khal’ and Surat al-Hafd. Their content is preserved in Islamic literature as liturgical prayers.

What this means

The Companions Muhammad himself singled out as the most authoritative reciters of the Quran disagreed about what was in it. Ibn Mas’ud’s codex was missing three surahs that today are in every Quran. Ubayy’s codex contained two surahs that today are in no Quran. These were not opponents of Islam or marginal figures, they were the men Muhammad pointed to as authoritative.

The orthodox response is that Ibn Mas’ud “regarded” Al-Fatihah and the Mu’awwidhatayn as prayers (du’a) rather than as surahs for the purpose of his personal codex. Even granting this, the response requires admitting that Muhammad’s own designated authoritative reciter did not know what belonged in the Quran. And the response cannot apply to Ubayy: Ubayy did not omit; he included. Either Ubayy added non-scripture, or the Uthmanic recension omitted scripture from him. Either way, the orthodox doctrine that the Quran was perfectly clear to the early community is conceded.

4. A Verse That Is Missing, Umar’s Testimony

The most direct admission of textual loss in Islamic tradition comes from Umar ibn al-Khattab himself, the second caliph, one of the ten promised paradise. Speaking publicly on the minbar in Medina:

Verily, Allah sent Muhammad with the truth and revealed the Book to him, and the verse of stoning was among what was revealed to him. We recited it, retained it in our memory and understood it. Allah’s Messenger awarded the punishment of stoning to death, and after him we also awarded the punishment of stoning. I am afraid that with the passage of time, the people may say: We do not find the verse of stoning in Allah’s Book, and thus go astray by abandoning an obligation which Allah has revealed.

, Sahih Muslim 4194; Sahih al-Bukhari 6829. Both graded sahih.

Umar testified that a verse was revealed by Allah, recited by the early community, memorized, applied as law, and is no longer in the Quran. He said this publicly, and the muhaddithun preserved it in the two most authoritative collections of Sunni hadith.

The orthodox response is the doctrine of naskh al-tilawa, “abrogation of recitation”: Allah deliberately removed the text while preserving the ruling. The doctrine is articulated in classical uloom al-Quran (al-Suyuti, al-Itqan) and is the standard apologetic response.

But notice what this response does. It does not save preservation; it redefines it. “Preservation” comes to mean “preservation of whatever remains after Allah’s deletions.” That is not the plain meaning of the word. Q 15:9 promises that Allah is the Quran’s guardian; if guardianship includes deleting verses without leaving any trace in the text, the promise has been emptied of content.

Look also at Umar’s own framing. He is not explaining a theological feature (“the naskh al-tilawa of the stoning verse, brothers, is a manifestation of divine wisdom”). He is afraid. He is worried people will say “we do not find the verse of stoning in Allah’s Book.” His anxiety is the anxiety of a man who knows the text is incomplete, not the calm exposition of a theologian who knows the text was divinely edited. The doctrine of naskh al-tilawa reads later theology back into Umar’s anxiety, and his anxiety, the source the doctrine is built on, contradicts the reading.

5. The Seven Ahruf, A Defense That Doesn’t Defend

The classical orthodox defense against the variant-codex problem is the doctrine of ahruf. Several sahih hadith record:

This Quran has been revealed to be recited in seven different ways (ahruf), so recite of it whichever is easier for you.

, Sahih al-Bukhari 4992; Sahih Muslim 818.

The orthodox argument: the variants among the Companion codices are not evidence of textual instability, they are evidence of divinely revealed multiplicity. The Quran was always multiple.

This defense breaks down on close examination.

First, no one knows what the seven ahruf are. Classical Muslim scholars offered at least six major interpretations: seven Arabic dialects (al-Tabari); seven categories of variation (Abu Ubayd); seven theological themes (Ibn Qutaybah); seven aspects of meaning (al-Razi); seven modes of pronunciation; seven historical layers. If the doctrine were a clear positive teaching, its meaning would not be contested fourteen centuries later.

Second, the doctrine contradicts Uthman’s action. The Prophet’s instruction was “recite whichever is easier for you”, a direct authorization of multiplicity. Uthman did the exact opposite: he sent one version to each province and burned all others. If the seven ahruf were divinely revealed, Uthman destroyed six-sevenths of revelation. The orthodox tradition cannot say both that the ahruf are revelation and that Uthman appropriately destroyed them.

Third, the doctrine doesn’t reach the variants we actually have. Ibn Mas’ud’s codex omitted three surahs. Ubayy’s codex contained two extra surahs. These are not differences in pronunciation or dialect within a shared text, they are differences in what counts as scripture. The ahruf defense, even at its strongest, has nothing to say about them.

Fourth, the hadith makes sense only in a community where multiplicity was a real feature. A revealed text uniformly transmitted does not need an instruction to “recite whichever is easier.” The hadith’s existence is itself evidence of an original plurality the tradition has spent centuries explaining.

6. The Manuscript That Was Erased, Sanaa

In 1972, workmen repairing the Great Mosque of Sanaa in Yemen discovered a cache of ancient Quranic manuscripts in the roof. One of them, designated DAM 01-27.1, was a palimpsest: a parchment whose original text had been scraped off and overwritten. Using ultraviolet imaging, scholars have recovered the lower (erased) text.

The most detailed study is Behnam Sadeghi and Mohsen Goudarzi, “Ṣan’ā’ 1 and the Origins of the Qur’ān,” Der Islam 87 (2012), pages 1-129. Their findings:

The lower text dates close to Muhammad’s lifetime. Carbon dating places the parchment with 95% probability between 578 and 669 CE.
The lower text contains hundreds of variants from the Uthmanic standard, different word orders, missing phrases, additional phrases, a different surah arrangement.
The variants do not match any of the canonical qira’at. The text represents an independent tradition.
The lower text was scraped off and overwritten with text matching the Uthmanic standard.

A physical, ancient, non-Uthmanic Quranic manuscript existed close to Muhammad’s lifetime. Someone scraped it off the parchment and wrote the Uthmanic version on top of it.

Behnam Sadeghi, who published the study, has been careful to resist sensationalized readings, and he should be. The Sanaa palimpsest does not “prove the Quran is a forgery.” What it does prove is that a non-Uthmanic textual tradition existed close to the Prophet’s lifetime and was suppressed by overwriting. That is enough. The orthodox claim of perfect preservation requires that no such alternative tradition existed; the Sanaa text shows one did.

7. The Qira’at, Standardization That Didn’t Standardize

Even after Uthman’s burning, the text did not become uniform. The Quran today exists in multiple canonical qira’at, readings, that differ from one another. Nearly three centuries after Uthman, Ibn Mujahid (d. 936 CE) canonized seven of them as authoritative. Later scholars expanded the canon to ten and then fourteen. The two most widely used today are:

Hafs ‘an ‘Asim, dominant in most of the Muslim world.
Warsh ‘an Nafi’, dominant in North and West Africa.

These differ in vocalization, in some consonantal letters, and occasionally in meaning. Examples:

Q 2:184, Hafs reads “ransom of feeding a poor person” (singular); Warsh reads “ransom of feeding poor people” (plural). The legal application differs.
Q 3:146, Hafs reads qatala (“fought alongside”); other readings have qutila (“was killed”). The theological implication of prophets being killed in battle versus accompanying combatants is non-trivial.
Q 19:19, variant readings produce either “I shall give you a holy boy” or “He shall give you a holy boy” (addressed to Mary).

These are canonical readings, all considered valid, all recited in salah by different Muslim communities today. They have semantic consequences.

The orthodox defense is that the qira’at represent “managed diversity” preserved through disciplined chains of transmission (isnad). The transmission discipline is real. But the discipline does not change what is being transmitted. Multiple words are canonically accepted for the same verse-position in the same book that is supposed to be one fixed text. The discipline of transmitting variation is not the same as the absence of variation.

8. The “Bible Is Corrupted” Claim, In Light of This

Orthodox Muslim apologetics frequently teaches that the Christian Bible has been corrupted (tahrif) while the Quran has been preserved. Compare the two textual histories honestly.

The New Testament:

Earliest fragments c. 125-200 CE (P52 of John).
Earliest near-complete codices c. 325-360 CE (Sinaiticus, Vaticanus), predate Uthman’s standardization by nearly three centuries.
About 5,800 Greek manuscripts, plus 10,000 Latin and 9,300 in Syriac, Coptic, Armenian, Georgian, Ethiopic, Gothic.
Independent transmission streams across politically and theologically hostile communities, Latin West, Greek East, Syriac East, Coptic, Armenian, Ethiopian.
Variants are preserved in the manuscripts. They are openly catalogued in any standard textual-criticism reference.

The Quran:

Earliest authoritative compilation by a committee under political authority.
Variant codices burned.
Sanaa palimpsest scraped and overwritten.
Surviving variant tradition (the qira’at) limited to a fraction of the original multiplicity.
Hadith preserved record of textual loss (Umar’s stoning verse).

The structural comparison is straightforward: the New Testament preserved its variants in independent multi-language transmission. The Quran suppressed its variants through centralized political action. A tradition that preserved its variants is more transparent about its textual history than a tradition that destroyed them.

This does not prove Christianity. It establishes only that the tahrif charge against the Bible inverts the actual textual situation. If either scripture has been corrupted in the sense the tahrif doctrine asserts, it is not the Bible.

9. What the Orthodox Position Has to Hold Together

When the evidence above is gathered, the orthodox apologist has to hold the following positions simultaneously:

The Quran was perfectly preserved, and Uthman’s standardization with manuscript burning was necessary, but a perfectly preserved text does not produce community-fracturing variants twenty years after the Prophet.
The seven ahruf are divinely revealed, and Uthman appropriately destroyed six of them, but God’s revelation is not, on orthodox premises, something a caliph appropriately destroys.
The Quran has not been changed, and the stoning verse was divinely deleted from the text, but “preservation” has been redefined to mean “preservation of whatever remains after divine editing,” which is not the plain meaning.
The Quran is preserved by tawatur, and two verses entered the compilation on the testimony of one man (Khuzayma), but tawatur and ahad are technical categories; they cannot be interchanged.
Ibn Mas’ud was Muhammad’s chosen reciter, and he was wrong about which surahs belong in the Quran, but a tradition cannot say its foundational reciters were mistaken about scripture.
Umar testified that a verse is missing from the written Quran, and Umar must be a reliable witness, Sunnis cannot call Umar a fabricator. The only consistent position is that Umar was right.

Each of these tensions has been engaged by serious classical and modern scholars (al-Suyuti, Ibn al-Jazari, al-Razi, Mustafa al-Azami). They have not failed to notice the difficulties. What they have done is develop interpretive frameworks, ahruf, naskh al-tilawa, tawatur, ‘ard al-akhir (the Final Review of Gabriel in Muhammad’s last Ramadan), that distribute the tensions across multiple doctrinal layers rather than resolving them.

A reader who follows the frameworks all the way to their internal limits finds that the strong popular preservation claim, “the Quran in my hand is identical to what Muhammad recited”, is sustained only by readers who do not press the frameworks past their breaking point.

10. The Choice

The situation, stated plainly:

The Quran was assembled twice, both times under caliphal authority, from a fragmentary base.
Two verses of the present Quran entered the compilation on the memory of one man.
The second caliph testified that at least one verse was revealed and is no longer in the text.
The Companion Muhammad himself named as an authoritative reciter possessed a codex missing three surahs.
Another such Companion possessed a codex with two extra surahs.
All competing codices were burned by order of the third caliph.
A physical, ancient, non-Uthmanic Quranic manuscript still exists in scraped-off form (Sanaa).
Multiple semantically-distinct canonical readings of the Quran exist today.

The orthodox doctrine that the Quran is the perfectly preserved word of Allah cannot be held alongside the historical record that the Islamic tradition itself preserves in its most authoritative sources. The doctrine is sustained by interpretive frameworks (ahruf, naskh al-tilawa, tawatur) that, examined honestly, absorb the difficulties rather than resolve them.

This does not require anyone to convert to anything. It only requires looking at what Bukhari and Muslim actually say, and what Islamic tradition itself preserves about the compilation. The sources are not Christian. The argument is not hostile. The conclusion is built entirely from inside the tradition.

If the Quran has not been preserved as the orthodox doctrine claims, then the Quran’s own commands to consult the Christian and Jewish scriptures (Q 5:46-47, Q 5:68, Q 10:94) take on a different weight. Those verses, which Uthman’s recension preserved, direct Muslims to engage seriously with prior revelation. The orthodox response, that tahrif makes those verses inapplicable today, depends on a corruption charge that the actual textual evidence inverts.

You can decide for yourself what to do with this. The sources are public; the manuscripts can be examined; the hadith can be read in their full Arabic. What is asked here is only that the question be asked honestly.

This is a condensed version of the longer argument in uthmanic-corruption.md, which engages the strongest orthodox responses in fuller scholarly detail, the ‘ard al-akhir (Final Review) defense, al-Azami’s specific critique of Arthur Jeffery, the oral tawatur / written shahid distinction on the Khuzayma problem, the qira’at isnad-quality defense, and the orthodox “managed diversity” response to the New Testament comparison. For the everyday reader, the points above are sufficient. The longer document is available for anyone, Muslim or otherwise, who wants the full scholarly engagement.