Developmental canonization of the Qur'an — a walkthrough

Developmental canonization of the Qur’an One diagram, read from about 600 to about 1100, with the writing on one side and the recitation on the other.

The diagram tracks two things that are usually collapsed into one: the Qur’an as a written text and the Qur’an as a recited text. They run in parallel columns, they feed each other, and they are canonized separately. Scroll to walk down the timeline; the diagram lights up the era you are reading.

This is a scholarly reconstruction in the historical-critical tradition, built backwards from surviving manuscripts and from the later literature on variant readings. It is an argument, not a record, and its two question marks are placed deliberately. The traditional Islamic account of the same period differs on several points noted at the end.

Start at about 600
Dialect Written tradition Timeline Oral tradition Hijazi Arabic Classical Arabic ? ~600 ~630–650 ~750 ~900 ~1100 Qur’anic collection — about 600 Qur’anic collection An unknown number of recitation traditions Uthmanic archetype — about 630–650 Uthmanic archetype Uthmanic recitation — about 630–650 Uthmanic recitation supplemental surahs? — source unspecified supplemental surahs? Fully vocalized Qur’ans — about 750 Fully vocalized Qur’ans Developing recitations — about 750 Developing recitations Semi-canonical Qur’ans — about 900 Semi-canonical Qur’ans Canonical Qur’ans — about 900 Canonical Qur’ans Canonical recitations — about 900 Canonical recitations

written recited uncertain

1 of 6about 600

Swipe the diagram sideways to see both columns.

About 600

Two traditions, and only one of them can be traced

In the diagramA single box at the head of the written column. Opposite it, five strands fanning out of nothing, marked with a question mark, converging as they fall into one arrow.

The reconstruction opens with a written body of Qur’anic material already in existence. What it cannot open with is a single oral tradition. The fan stands for an unknown number of recitation practices, and the question mark is load-bearing: how many there were, what they contained, and how they related to the written collection cannot be recovered from the surviving evidence.

Everything below this row is inference worked backwards from manuscripts and from the later literature on variant readings. This row is where the inference runs out. The strands fade at the top for the same reason the question mark is there.

About 630 to 650

The written stream narrows to a single ancestor

In the diagramThe trunk arrives at one box, the Uthmanic archetype. A double-headed arrow ties it to the timeline; a second arrow carries across to a recitation of the same name. The dialect axis on the far left now reads Hijazi Arabic.

An archetype is a common ancestor reconstructed from its descendants. The claim of this row is that the surviving Qur’anic manuscripts converge on one written exemplar produced in the middle of the seventh century, and that a recitation practice sat alongside it.

The pairing is not decorative. The script of this period records consonants and long vowels — the rasm — and omits short vowels and most of the dots that distinguish otherwise identical letters. A reader supplies the rest from memory. So the page constrains the recitation and the recitation disambiguates the page, which is why that arrow points both ways.

The language of the archetype is Hijazi, the Arabic of the region, and not the standardized Classical Arabic that grammarians codified later. Hold on to that; it matters two rows down.

An open question

A side branch the diagram declines to assert

In the diagramA small box off to the left of the trunk, entered by a faded, broken arrow from an unspecified source and leaving into the vocalized copies. Its label ends in a question mark.

This is the second admission of uncertainty, and it works differently from the first. It allows for material entering the written tradition from outside the archetype — surahs, or parts of them, carried by some other route and later absorbed into copies that otherwise descend from the archetype.

The arrow into the box fades out at the top because the source is not named. The question mark asks whether the box belongs on the diagram at all. Marking a possibility without asserting it is a real move in a stemma, and it is worth noticing that the diagram makes it twice.

About 750

Vowels reach the page, in a different language

In the diagramThe trunk reaches fully vocalized Qur’ans. A second, heavier arrow leaves the archetype and sweeps clear across to developing recitations on the right. The dialect axis switches to Classical Arabic.

Later hands add short vowels, the dots that separate ambiguous consonants, and other reading aids. The page now carries much of what a reciter previously had to supply, and the written tradition becomes far more self-sufficient than it was.

Two consequences follow. The additions are made in Classical Arabic, so a vocalized copy reads a Hijazi consonantal text through a later grammatical norm — the skeleton and the pointing come from different stages of the language. Meanwhile the recitations are multiplying rather than settling.

That long diagonal is the strongest structural claim on the page. The developing recitations descend from the archetype directly, not by way of the vocalized copies. Text and recitation are cousins here, not parent and child, and neither one is simply a transcription of the other.

About 900

Canonization is an act of exclusion

In the diagramThree boxes where there were two. The recitations become canonical; the written column splits, one branch canonical, one semi-canonical.

A limited set of recitations is recognized as authoritative and the remainder are demoted. The readings literature of this period does exactly that, naming a small number of readers and their transmitters and treating the rest as something less.

Written copies are then sorted by whether they agree with the recognized readings. Copies preserving other, often older readings do not vanish, which is why the left-hand branch keeps running instead of stopping at this row. Nothing is added to the text at this point. A boundary is drawn around part of what was already there.

After about 1100

Three streams leave the frame

In the diagramA break in the timeline, a final marker, and three arrows running off the bottom edge.

The break compresses time; the arrows mean the streams continue past the edge of the reconstruction. There are three of them: a canonical written text, a canonical recitation tradition, and the semi-canonical copies that fell outside the boundary.

The first two are the ones that reach the printed Qur’an and the recited Qur’an of today. The third is why the diagram can be drawn at all — those variant copies are the evidence the whole reconstruction is built from.

What this diagram is, and what it argues with

Read as a whole, the reconstruction makes one main claim and one secondary one. The main claim is that a single written archetype stands behind the surviving manuscript tradition, and that canonization of the text happened centuries after it, by selection rather than composition. The secondary claim is the shape of the right-hand column: recitation is treated as a tradition with its own history, running alongside the written text and canonized separately.

The traditional Islamic account covers the same period differently. There the collection is made by the Prophet’s companions and standardized under the caliph Uthman, and the canonical readings are understood as transmitted from the Prophet himself rather than selected from a widening field in the ninth or tenth century. The disagreement is about transmission and about what the later readings literature was doing, not about the date of the consonantal text, which both accounts place in the seventh century.

The labels, dates, arrows and both question marks here are the diagram’s own. The explanations of what they imply are context added around them.

Walk the timeline again