Thread by @PhDniX

In my recent BSOAS article, I show that all manuscripts must derive from a single archetype, they share a single "fingerpint" but, as with all manuscripts, they're not quite identical. A little thread on what the differences look like.

The picture below is a transcription of the above text, with yellow being places where the text differs from the standard text most used today. Every single other letter is identical to what we have.

This is the Birmingham fragment, which is carbon dated to the 7th century

So what are the differences like? Two are rather banal.

fa-qāla "so he said" (quran.com/20/10/) is spelled defectively. This is common in early manuscripts, and creates some ambiguity, the reading could also be fa-qul "so say!", but that makes little sense here.

The other case is bi-lisāni-ka 'in your tongue' (quran.com/19/97), with once again an unwritten long ā where modern editions do write it. Vacillation between written ʾalif for /ā/ is very common, but has little effect on the archetypical "fingerprint" or meaning.

There's what looks like a verse marker after the basmalah. Modern print Qurans don't use this, and instead print the basmalah in a separate line. This is more of a difference in layout than it is a difference in text.

However, after the basmalah come the "mysterious letters" ṭāhā is missing a verse break. The verse break of the standard text is typical for the Kufan mushafs and readings (ʿĀṣim was a Kufan reader). Warsh ʿan Nāfiʿ (Medinan print Kurans) lack this verse break even today.

Keeping track of verse breaks in fact is one way we can identify a Quranic fragment's regional identity, and generally agrees with the regional rasm variants as well.

The Birmingham fragment therefore seems to stem from the non-Kufan manuscript tradition.

At quran.com/20/6/ the text of the Birmingham fragment has only two dots on top of aṯ-ṯarā "the soil"; seemingly writing at-tarā. There is no plausible reading that fits here, so this is presumably simply a forgotten dot by the scribe.

The final variant prompted Yasin Dutton to write his "When Did the Shawādhdh Become Shādhdh?" article. While the current text actually conforms to the standard text, clearly an ʾalif has been removed. quran.com/20/12/

doi-org/10.1163/1878464X-00801001/

So what's going on here? It's difficult to understand it as a scribal error, why would anyone place an ʾalif here? The word involved ṭuwan is somewhat obscure placename. And this spelling keeps on showing up in other manuscripts (sometimes corrected too).

That spelling is so widespread in early manuscripts that it almost certainly has to be the original spelling that got corrected in later manuscripts. Collections of non-canonical readings transmit a memory of this spelling too.
The reasons for this change, however, remain obscure

The Birmingham fragment is by no means unique in being this similar to the standard text, this is instead quite similar. But there are other variants that don't affect the rasm, but do affect the language and wording of the Quran: consonantal dotting and vowels.

Al-Kisāʾī and Ḥamzah read Q19:91, 92 ولدا as wuldan instead of waladan "son". No effect on the meaning.

Šuʿbah, Ḥamzah, al-Kisāʾī read the mysterious letters Q20:1 ṭēhē; Warš and ʾAbū ʿAmr ṭāhē; The rest ṭāhā.

Ḥamzah reads Q20:10 as fa-qāla li-ʾahli-hu mkuṯū, the rest reads fa-qāla li-ʾahli-hi mkuṯū. "he said to his family: "sit down"

No effect on the meaning, but having -hu for 'his' after the genitive rather than -hi is irregular in Ḥamzah's linguistics system. Usually has -hi.

Q20:12 Ibn Kaṯīr and ʾAbū ʿAmr read fa-lammā ʾatā-hā nūdiya yā mūsā ʾannī ʾana rabbu-ka "and when he came to it, it was proclaimed, O Moses, that I am your lord".

The rest reads ʾinnī which changes the meaning somewhat: "... he was called: "O Moses, I indeed am your lord"

In the same verse the al-Kisāʾī, Ḥamzah, ʿĀṣim and Ibn ʿĀmir read ṭuwan. The rest reads ṭuwā.

Ḥamzah reads Q19:13 as wa-ʾannā ḫtarnā-ka : "and that we have chosen you", the rest reads wa-ʾana ḫtartu-ka "and I have chosen you".

While other variants were ambiguous in this manuscript, the dotting of the manuscript actually allows us to identify that the majority wa-ʾana ḫtartu-ka is the reading intended by the scribe.

I how this overview gives a bit of a sense what kinds of differences and correspondences we find in early Quranic manuscripts! There are sections where the readings have more significant differences in meaning, and also some more rasm variants, but this page is quite typical.