The Variant Readings (Qira’at): How Many Qurans Are There?

Walk into a mosque in Cairo or Istanbul and pick up the Quran on the shelf, and you will almost certainly be holding a text printed in the reading of Hafs ‘an ‘Asim. Walk into a mosque in Rabat, Nouakchott, or much of West Africa, and the Quran in your hands may instead be printed in the reading of Warsh ‘an Nafi‘. Open them side by side and you will find them overwhelmingly the same book—the same surahs, the same order, the same overarching message. But you will also find, scattered across the pages, hundreds of small differences: a lengthened vowel here, a different verb form there, occasionally a different word altogether, and in a handful of places a difference that changes what a verse means or even what it commands a believer to do.¹

Both of these texts are printed and certified as authentic by Muslim authorities, including Saudi Arabia’s King Fahd Complex for the Printing of the Holy Quran.² Neither is a forgery or a sectarian deviation. They are two of the canonical qira’at—the readings of the Quran—and they raise a question that deserves to be asked carefully and answered honestly: in what sense is there one Quran?

Stating the Question Carefully

It is worth being precise from the outset, because this topic is easy to overstate in both directions. This essay is not about whether the Quran has been wildly corrupted or rewritten beyond recognition—it has not been. The Arabic consonantal text (the rasm) is remarkably stable, and the readings differ within a narrow band. Nor is this the broader question of the Quran’s overall preservation history, the Uthmanic standardisation, or the companion codices; that is treated in a separate deep dive.

The narrow question here is the qira’at themselves. The popular apologetic claim—heard often in dawah (Islamic outreach)—is that the Quran has been transmitted letter-for-letter, dot-for-dot, identical across the centuries and across the world, and that this perfect preservation is itself a proof of its divine origin. The aim of this essay is to test that specific claim charitably against the Islamic tradition’s own data. The honest finding, stated plainly, is this: one perfectly identical text is a more complicated claim than the popular version allows. There is one Quran in the sense that matters most to a Muslim—one message, one stable skeleton—but there is genuine, divinely sanctioned variety in how it is read, and the system that defines which readings count was the work of named human scholars over several centuries. That is the whole of the claim, and it is enough to be worth examining.

The Foundation: The Hadith of the Seven Ahruf

Any fair account must begin where the Islamic tradition itself begins: not with later scholars but with Muhammad. The most authoritative hadith collections record that the Quran was revealed in multiple forms. In Sahih al-Bukhari, ‘Umar ibn al-Khattab reports that he heard Hisham ibn Hakim reciting Surah al-Furqan in a manner different from his own, seized him, and dragged him to the Prophet. After both men recited, Muhammad responded: “It was revealed thus,” and then, “This Quran has been revealed to be recited in seven ahruf, so recite of it whichever is easy for you.”³

The word ahruf (singular harf) is notoriously difficult. It can mean “letters,” “modes,” “dialects,” or “ways”—and the Islamic tradition never settled on a single meaning. The tradition itself preserves an extraordinary range of views: the hadith scholar Ibn Hibban (d. 354 AH) listed some thirty-five different opinions on what the seven ahruf actually were, and later scholars round the figure to around forty.⁴ This is not a Christian polemic; it is the tradition’s own record of its disagreement. What the hadith establishes, on the tradition’s own terms, is that variation in recitation was permitted by the Prophet from the beginning—and that even the earliest community did not agree on how to classify it.

From Ahruf to Qira’at: How the System Was Built

The qira’at are not simply the seven ahruf preserved intact. They are a later, scholarly system, and tracing how it was assembled is the historical heart of the matter.

When the third caliph Uthman standardised the written text around 650 AD, the manuscripts he produced recorded only the consonantal skeleton. Early Arabic script had no vowel signs and few of the dots (diacritics) that distinguish otherwise identical consonants. A single written word could therefore be vocalised in several legitimate ways. For roughly the next three centuries, reciters across the Muslim world transmitted a wide variety of readings, in a period scholars call ikhtiyar—“choosing.” Master reciters in Medina, Mecca, Kufa, Basra, and Damascus each had their own systems.

The decisive narrowing came in Baghdad around 936 AD, when the scholar Abu Bakr Ibn Mujahid (d. 324 AH) wrote his Kitab al-Sab‘a fi al-Qira’at, “The Book of the Seven Readings.” Ibn Mujahid selected seven reciters as canonical and treated readings outside this set as shadhdh (anomalous, non-canonical).⁵ The Harvard-trained scholar Shady Nasser, in the standard academic study of this process, documents that Ibn Mujahid’s choice of exactly seven was controversial in its own day: some contemporaries objected that it caused confusion by inviting people to conflate his seven readings with the seven ahruf of the hadith—which were not the same thing.⁶ Three further readings (those of Abu Ja‘far, Ya‘qub, and Khalaf) were later raised to canonical status, chiefly through the work of Ibn al-Jazari, to give the ten readings recognised today; a wider set of fourteen is sometimes discussed.⁷

Each of the ten readings is transmitted through two principal rawis (transmitters), giving the familiar pairs: ‘Asim of Kufa is transmitted by Hafs and by Shu‘ba; Nafi‘ of Medina by Warsh and by Qalun. So the “readings” a Muslim encounters in print—Hafs ‘an ‘Asim, Warsh ‘an Nafi‘—are reader-plus-transmitter combinations.⁸ The point to grasp is simply that the authoritative list of how the Quran may be recited was assembled, debated, and revised by named scholars over centuries; it was not delivered complete and fixed at the start.

What the Differences Actually Look Like

Honesty cuts both ways here, and the first thing to concede is that the great majority of differences between the readings are minor. They concern vowel length, points of pronunciation, the assimilation of letters, and rules of elongation—features that disappear entirely in translation. A reader of the King Fahd Hafs text and a reader of the King Fahd Warsh text are reciting the same surah, and a listener who did not know the technical system might hear only a difference of “accent.”

But not all the differences are merely phonetic. Some reach the consonants, the words, and the meaning.

Owner or King? (Surah 1:4)

In the opening chapter, al-Fatiha—the most recited passage in all of Islam—verse 4 is read by ‘Asim and al-Kisa’i as mâliki yawmi al-din, with a long vowel, “Owner of the Day of Judgment,” and by other canonical readers as maliki, with a short vowel, “King of the Day of Judgment.”⁹ The difference is one letter (an alif) and a clear shift in sense—owner versus king. Muslim scholars reasonably treat these as complementary: God is both. But it is a difference of meaning, not merely sound, and it sits in the chapter every Muslim recites in every unit of every daily prayer.

Fought or Were Killed? (Surah 3:146)

Surah 3:146 describes prophets and their followers in battle. In the Hafs text the relevant word is written with an alif, qatala (“[many a prophet] fought”); in the Warsh tradition it is read qutila (“[many a prophet] was killed”).¹⁰ This is a genuine difference of meaning—fighting alongside the prophet versus being slain as a martyr—yet both are read from a single consonantal skeleton, since every canonical reading conforms to the one ‘Uthmanic rasm. That is significant in the other direction: it shows the variation is not always a trivial matter of accent, but can change what the verse says even where the written skeleton is identical.

Wash or Wipe? A Difference of Law (Surah 5:6)

The most consequential example concerns ritual law. Surah 5:6 lays out the rules of ablution (wudu) before prayer, including how to treat the feet. The key word, arjulakum (“your feet”), is read in two canonical ways differing by a single vowel: in the accusative case it parallels “wash your faces and hands,” yielding “wash your feet”; in the genitive case it parallels “wipe your heads,” yielding “wipe your feet.”¹¹ This is not a nuance of beauty. It is two different ritual obligations, and the divergence tracks a real difference in practice between the Sunni and Shia traditions of how ablution is performed. A second example, Surah 2:222, turns on whether a husband may resume relations with his wife once menstrual bleeding has ceased or only after the full ritual bath—again a legal fork that the schools of law debated, resting on a difference of verb form.¹²

The Mainstream Muslim Explanation, Stated at Full Strength

Before assessing any of this, the traditional Muslim account deserves to be heard in its strongest form—not a caricature, but the case as a thoughtful Muslim scholar would actually press it.

First, the variety is divinely authorised, not a human accident. The seven-ahruf hadith shows the Prophet himself sanctioning multiple recitations, and the Quran (39:23) calls itself “a Book... mathani”—understood by some as confirming its multiple consistent modes. The qira’at are therefore revelation, not corruption: God gave His word in a way that flexed to the dialects and tongues of the early Arabs, “so recite whichever is easy for you.” Variation here is a mercy, an ease, woven into the revelation by design.¹³

Second, the readings are transmitted by tawatur—mass transmission. A report is mutawatir when it is conveyed by so many independent chains at every link that collusion or error becomes inconceivable. The traditional claim is that each canonical reading was recited by vast numbers of Muslims in an unbroken chain back to the Prophet, so that the readings carry the same epistemic certainty as the consonantal text itself. On this view the qira’at are not the residue of scribal confusion but a living oral tradition of unmatched rigour, sustained by a community that memorised the entire text generation after generation.

Third, and crucially, the differences complement rather than contradict. Where one reading says “Owner” and another “King,” both are true and the pair enriches the meaning. The qira’at, on this account, are a feature of the Quran’s inimitability (i‘jaz): a single text bearing multiple layers of legitimate sense, all flowing from one source. Far from undermining preservation, the readings are the preservation—evidence of a text guarded so carefully that even its authorised variants were tracked and named.

This is a serious position, held by serious scholars, and it deserves a fair and complete hearing before any reply — which is exactly what the next section gives it.

Weighing the Explanation

Several parts of this account can simply be conceded, and a Christian should concede them gladly. The variety is rooted in the tradition’s earliest layer—the ahruf hadith is in Bukhari, not a later invention. Most differences are minor. Many readings genuinely are complementary in the way described. And the Arabic consonantal text is, by the standards of ancient texts, impressively stable. None of that is in dispute. The question is whether the popular conclusion—one perfectly identical, letter-for-letter text, miraculously preserved—follows. Three points suggest it needs qualifying.

The system is humanly constructed, not self-evidently given. If the qira’at were simply the seven ahruf handed down whole, one would expect the tradition to know clearly what the ahruf were and how the readings map onto them. Instead, as Ibn al-Jazari himself records, the tradition preserves dozens of incompatible theories of the ahruf, and the relationship between the ahruf and the qira’at remains unsettled.⁴ The number seven that Ibn Mujahid fixed upon in the tenth century was, even contemporaries complained, partly a numerical echo rather than a derivation—and the canon then grew to ten and was discussed as fourteen. A revelation whose authorised readings had to be selected, debated, and expanded by named scholars across three centuries is not the same thing as a text that arrived complete and self-authenticating.

“Complementary” is the right word for synonyms, not for a legal fork. The complementarity defence works well for “Owner” and “King”: both are true of God, and the pair enriches. It works far less well for Surah 5:6. “Wash your feet” and “wipe your feet” are not two facets of one truth; they are two different commands, and a worshipper cannot do both at once in the way the readings specify. Here two canonical readings yield two incompatible ritual obligations from what is supposed to be a single perfectly transmitted divine text. One may hold that God authorised both practices—but that is a different and larger claim than “one identical text,” and it is the claim the popular dawah argument quietly drops.

The tawatur claim is exactly where modern scholarship presses. This is the most technical point and must be stated carefully. The traditional guarantee of perfect oral preservation is that every canonical reading is mutawatir. Yet Shady Nasser, examining the actual chains of transmission, argues that the system of readings “failed to meet the conditions of tawatur” as the legal theorists themselves defined them: at decisive junctures the chains narrow to one or two named transmitters—technically ahad (singular), not mass-transmitted—creating what Nasser calls a paradox between the certain transmission of the written mushaf and the far less certain transmission of its oral recitation.¹⁴ If that analysis holds, then the readings are not guaranteed by the kind of mass transmission the popular claim invokes, and the question of which readings are canonical becomes, at least in part, a matter of later scholarly judgment—which is precisely what the canonisation history shows.

The Honest State of the Scholarship

Fairness requires giving the strongest reply the last word before the conclusion, and naming where the dispute genuinely stands. The tawatur question is contested, not settled. Traditional Muslim scholars have sophisticated rejoinders to Nasser: they distinguish tawatur in the early generations (where the readers were many) from the later named “bottlenecks,” which they treat as a documentary convenience rather than the real channel of transmission; and they argue that the conditions for tawatur should be assessed at the level of the reading-community, not the isnad as written. A careful critic should label this a live scholarly debate, not a knockdown. Nasser is making a serious, well-documented academic case; he is not the unanimous verdict of the field, and many Muslim specialists in ‘ilm al-qira’at disagree with him.

It is also worth naming the academic on the other side of the meaning question. Adrian Brockett, in his St Andrews study comparing the Hafs and Warsh transmissions in detail, concluded that while the differences are real and worth cataloguing, their effect on the overall meaning of the text is limited—the two transmissions, he argued, are close enough that neither tells a substantially different story.¹⁵ Brockett is a sympathetic witness for the modesty of the claim being made here: the readings differ, sometimes in meaning and law, but they do not produce two different Qurans in the sense of two different messages. That is exactly the line this essay walks. What the evidence unsettles is the perfect-identity showpiece, not the substantial integrity of the book.

Conclusion: What This Does and Does Not Establish

So how many Qurans are there? The honest answer is layered. In the sense the popular argument intends—one letter-for-letter, dot-for-dot identical text—the picture is more complicated than the slogan: there are multiple canonical readings, certified by Muslim authorities, that differ not only in pronunciation but at times in word, meaning, and even legal obligation, and the system defining them was built by named scholars over three centuries. In the sense that matters most to a believing Muslim—one message, one stable consonantal text, one book—there is indeed one Quran, and it has been transmitted with real care.

What the qira’at do not establish should be stated as plainly as what they do. They do not show that the Quran has been corrupted into chaos, or rewritten, or that its message is unrecoverable—to claim any of that would be to overstate the evidence badly. What they show is narrower and quieter — but on the popular boast it is a plain no: the specific claim of flawless, identical, miraculous preservation, often offered as the showpiece proof of the Quran’s divine origin, does not hold. It claims more than the Islamic tradition’s own data will bear. The variant readings are evidence of a careful and reverent transmission—but a human, historical one, with the marks of human scholarship on it.

None of this is an occasion for triumph, and it is not meant to embarrass anyone. The point of understanding the qira’at is not to win an argument but to move the conversation to the ground that actually matters. Notice what even a flawlessly copied text would and would not settle: careful transmission can show that a book has been faithfully preserved, but it cannot by itself show that the book is true. So the preservation boast was always the wrong foundation to rest a claim of divine origin on. The real question — worth asking gently, accurately, and with respect (1 Peter 3:15) — was never the manner of the copying, but who God has shown Himself to be, and whether He has entered history to make Himself known. That is where this conversation begins.