When is it likely that the Qur’an was composed, and how many strata of composition does it consist of?
The tradition is that the Qur’an was revealed to the Prophet Muhammad over a period of twenty-two years, between 610 and 632. Muslims believe that Muhammad was not literate, and so he could not himself have produced the Qur’an. He memorized it, and a few of his followers memorized some parts of it or wrote them down. So when he died in 632 there was no codex as such. Different disciples had variants written down, creating a need around 650 to produce a canonical version, because ‘Uthman—the Caliph at the time—feared these differences might cause splits among Muslims. So he put together a committee to produce a standard text for the faithful. That’s the traditional account of the origin of the written text, which was in circulation by the middle of the seventh century ce. We have scarcely any documentary evidence from the seventh century itself, only oral traditions whose authenticity is hard to verify, with many conflicting narratives, which makes the task of the scholar very difficult. A very few exceptions exist, such as the inscription inside the Dome of the Rock in Jerusalem (built in 692) which documents a few verses from the Qur’an (e.g. 4.171–172).
As a historian, when you start examining the Qur’an, you realize that this is a very difficult text. It is not like the Hebrew Bible, and can’t be compared to the Gospels. It doesn’t tell us the story of a person or a people. The Hebrew Bible is the story of the Israelites; the Gospels record the story of Jesus’s ministry. The Qur’an is not a story of the Arabs or of Muhammad’s ministry. Its unique character as a narrative poses particular problems for the historian. If you only read the Qur’an, you would not know much about Mecca, Muhammad, and Arabia. Muslims have always read the Qur’an by using the books on the Life of Muhammad to help explain what the Qur’an intends. There is nothing like it, which actually is something that the Qur’an acknowledges. Aside from some verses that address legal matters, the Qur’an has a tendency to be brief and alludes to stories, events and issues with the assumption that its direct audience already knows them.
Some modern historians have postulated a human agent to have produced it, but was there one author or several authors? For a long time scholars in the field—the most influential was John Wansbrough—believed that the Qur’an was finalized at the end of the eighth century, or the beginning of the ninth. Since we have some early inscriptions and a recent discovery of partial manuscripts of the Qur’an that can be dated to the late seventh or early eighth century, that view is now discredited. The material and the type of script used tell us when these were produced. For example, the Muslim world was by and large using paper by the middle of the eighth century, or the early ninth. Leather, papyri and parchment were abandoned because they were so much less practical. Script is also important. In the seventh century the principal script was Kufic, believed to have been invented in Iraq. Later, as Muslims started to develop new styles of writing, this was completely abandoned. So, by and large, anything written in Kufic script must come from the seventh or eighth century, especially if it is on papyri or parchment.
Where was this new manuscript evidence found?
When the Great Mosque in Sana‘a was being renovated in the early 1970s, a secret attic was discovered above a false ceiling, containing a mass of old manuscripts. The Middle Eastern tradition (which applies to Christians and Jews as well) is that if a manuscript has the name of God or the name of the Prophet on it, you can’t simply destroy it. The best thing you can do is put it away, or bury it, as with the Dead Sea Scrolls or the Nag Hamadeh texts. You do so not to hide them for hiding’s sake, but to keep them from getting corrupted and thus insulting God. That was the case in San‘a. A German scholar was allowed to study the finds, but she has published very little on them for fear of the political consequences of doing so; it seems the Yemeni government threatened Germany with repercussions if anything embarrassing appeared. But from a few of what are believed to be very early parchments in the cache, using Kufic script, we know that they date to the late seventh or early eighth century, and we can already see one significant difference with the canonical version of the Qur’an. The traditional story tells us there were no serious variations between the different versions assembled by Caliph ‘Uthman around the year 650, though we know that down to the eighth century more popular versions of the Qur’an, without major discrepancies from the canonical text, were retained in certain regions—Iraq or Syria—out of local pride. The Yemeni manuscript, however, contains a very serious divergence. In the canonical Qur’an, there is a verse with the imperative form ‘say’ [qul]—God instructing Muhammad—whereas in the San‘a text, the same verse reads ‘he said’ [qala]. That suggests some early Muslims may have perceived the Qur’an as the word of the Prophet, and it was only some time later that his reported speech became a divine command. There is also some serious variation with respect to the size of some chapters.
One other early variation is attested. In the Dome of the Rock, three sets of verses from different parts of the Qur’an referring to Jesus can be seen. In the original of one of these Jesus says: ‘Peace on me the day I was born, the day I die, and the day I am raised from death.’ The inscription in the Dome of the Rock, however, reads: ‘Peace on Jesus the day he was born, the day he died, and the day he was raised from death.’ Here the switch from the first to the third person isn’t a major change, but could be a kind of variation from the official Qur’an.