Writings: Islam

How were Islam's texts handed down?

Muslim tradition holds that God revealed the Quran over about 23 years, and David Thomas reports that tradition dates the revelations from 610 to 632 CE.¹²³ Abdullah Saeed writes that the standard traditional Muslim view, Sunni and Shia alike, holds that God alone composed it and Muhammad passed it on without addition or alteration.⁴

Yousef Wahb of the Yaqeen Institute for Islamic Research, whose paper names no branch, writes that recitation from memory was the primary method of preserving the text, with scribes writing alongside.¹

In al-Bukhari's Sunni Hadith collection (a hadith is a report of what Muhammad said or did), the caliph (leader of the Muslim community) Abu Bakr has the Quran gathered. Uthman later has copies sent out.⁵⁶⁷³ Abu al-Qasim al-Khoei, a Shia scholar of Najaf, rejects the reports that Abu Bakr or other caliphs collected the Quran, as contradicting one another. He holds that it existed as a complete written collection in Muhammad's lifetime, and that even if Abu Bakr collected it the collector only recorded what memory had preserved. He also holds that Uthman united the Muslims on one authoritative reading.²⁸

Mohammad Hashim Kamali writes that Muhammad is reported to have discouraged writing hadith down at first, lest they be confused with the Quran, and later to have allowed it. Companions kept small written collections, he adds, and scholars compiled large ones in Islam's second and third centuries. Joseph Schacht and others doubted that most hadith were authentic, he reports, while Muslim scholars hold they were documented much earlier than Schacht suggested.⁷

Stephen Shoemaker, whose book argues one thesis, holds that the Quran was recited from memory for decades after Muhammad's death. It was composed and recomposed in that transmission, he argues, largely outside the Hijaz (the region that includes Mecca and Medina). He adds that much of its content was presumably inspired by Muhammad's teachings but also heavily influenced by the religious milieu of the late ancient Near East.⁹ Nicolai Sinai writes in a 2020 review that the Quran, as the scripture of Islam, is as much oral and aural as written. His 2014 two-part article argues one position (below).¹⁰

How much survives, and how old are the oldest copies?

No source used here counts all Quran manuscripts. The sources place seven entries in Europe (two, the Heidelberg papyri and roll, only as older accounts do), one mainly in Yemen, and one in Jerusalem. Qaf 47 is mostly in Cairo, with seven leaves in Berlin.^{11121314151617181920}

The list is Quran-heavy because the Quran is, for Muslims, the revealed word of God and the Hadith the most authoritative source next to it.⁴⁷ No early manuscript of a major Hadith collection, Sunni or Shia, or of another Shia text was found in the sources used here. Kamali reports that al-Kafi is rated first of the four most authoritative Twelver Shia collections.⁷

• The text's date. Thomas reports Muslim tradition as dating the final written text, under Uthman, to about 650 CE, and Saeed to his caliphate, 644 to 656 CE.³⁴ Sinai holds 650 CE or earlier (apart from some spelling features) as the default date of the standard consonantal text (the letters without vowel signs). That holds unless passages of a distinct style and vocabulary are compellingly placed in the late seventh century. Shoemaker holds that the canonical text took shape under the caliph Abd al-Malik (ruled 685 to 705 CE). Sinai's and Shoemaker's dates do not overlap.²¹⁹
• The copies' dates, by radiocarbon. Four entries give radiocarbon ranges, which date the parchment, not the writing, at a stated probability.²² Birmingham is 568 to 645 CE and Tubingen 649 to 675, each at 95.4 percent, and Qaf 47 is 606 to 652, from two of its seven Berlin leaves.¹⁴²² Sana'a's 578 to 669 (95 percent) is for one detached leaf with no recorded origin. Sinai and Shoemaker think it likely to have come from the manuscript, though Shoemaker writes that this is not certain. Shoemaker reports that laboratories disagreed on samples from another leaf, Lyon giving 388 to 535 CE, Zurich 565 to 660, and Oxford 595 to 658.⁹¹⁰
• The copies' dates, without radiocarbon. Francois Deroche dates the Codex Parisino-petropolitanus (a codex is a bound book; Parisino for short) to 670 to 700 CE by handwriting and spelling, as Eleonore Cellard summarizes it. The online catalog Corpus Coranicum estimates 640 to 680 in its header.²³¹⁶²⁴ The Hamburg papyrus's editors date it by script to the late seventh or early eighth century, computed here as 667 to 719 CE.¹⁷²⁵ Or. 2165 has three datings. Corpus Coranicum's is 640 to 680, which points to archaic spelling. Yasin Dutton's is about 671 to 719, computed here from Marijn van Putten's table. The British Library's is the eighth century, which gives no method. Dutton's overlaps both of the others, which do not overlap each other.²⁶²⁷¹⁸
• Which is oldest. No source used here settles it. Corpus Coranicum's overview calls the Berlin leaves the oldest Quran fragment, and Cellard's online thread asks whether Parisino is the oldest preserved Quran.¹³²³
• The one inscription entry. The Dome of the Rock in Jerusalem quotes the Quran and gives 72 AH (the Islamic calendar), 691 or 692 CE, in its dedication.²⁸²⁹³⁰ Estelle Whelan, whose article argues one thesis, reads its wording as a sign that the Quran was already well known. Patricia Crone and Michael Cook, whose 1977 book argues one account, say it shows Quranic material but not the form in which it circulated.²⁸³¹
• The gap, computed here. On the tradition's dates, Birmingham, Qaf 47, and the 578 to 669 leaf overlap the years of revelation (610 to 632 CE), and Tubingen begins 17 years later. Qaf 47, Tubingen, and that leaf contain 650; Birmingham ends five years before it. On Shoemaker's 685 to 705 CE, the four ranges end 10 to 40 years before 685. Michael Marx and Tobias Jocham hold that a wait of decades between making parchment and writing on it is unlikely for economic reasons. Shoemaker holds that unwritten parchment could wait decades or longer.²²⁹ Parisino (on Deroche's dating), Hamburg, and the Dome begin 17 to 42 years after 650, within Shoemaker's years. Corpus Coranicum's 640 to 680 for Parisino contains 650.

How should the entries be read?

Equal slots do not mean equal evidence. The tradition's side rests on reports by two scholars, Saeed and Thomas, and three voices from within, covering Sunni and Shia sources but not every community. The scholars' side rests mainly on Shoemaker and Sinai. Shoemaker argues for dating the canonical text to Abd al-Malik's reign. Sinai argues for keeping the traditional date as the default, in a two-part 2014 article (abstracts only).³²²¹ No source counts scholars on either side.

The David papyrus bears 229 AH (844 CE in its editor's conversion). In Khoury's 2014 account the life of Muhammad bears no date, though other writers give 228 or 229 AH. The Ibn Lahi'a roll bears no date; a 1995 university report places it at the end of the eighth century. The Turkish Islam Encyclopedia, a Sunni reference work, gives no year for composing either text. Khoury holds that Wahb's family, with Wahb writing part, wrote his material down at the start of the second half of the eighth century. By a comparison computed here, that is after Wahb's death (728 to 734 CE).^193334353637

See The Quran and the Hadith and how ancient texts are dated.