2 Al-Bukhārī, al-Jāmiʿ al-ṣaḥīḥ, ed. Muḥibb al-Dīn al-Khaṭīb and Muḥammad Fu’ād ʿAbd al-Bāqī, 4 vols (Cairo: al-Maṭbaʿa al-salafiyya, ah 1400), vol. 3, 337–8, no. 4986–7 (66.3).

3 The Islamic tradition is contradictory on the question whether Zayd or somebody else was the first to have collected the Quran (see Mingana, Alphonse, “The transmission of the Kur’ān”, Muslim World 7, 1917, 223–32CrossRefGoogle Scholar, at 224–5).

4 The campaign mentioned in al-Zuhrī’s account is probably to be identified with a campaign that al-Ṭabarī reports for ah 30 in Annales, ed. M. J. de Goeje et al. (Leiden: Brill, 1879–1901), series 1, vol. 5, 2856 – thus Nöldeke, Theodor, Geschichte des Qorāns, revised by Schwally, Friedrich, Bergsträsser, Gotthelf and Pretzl, Otto, 3 vols (Leipzig: Dieterich'sche Verlagsbuchhandlung, 1909–38Google Scholar, henceforth GdQ), vol. 2, 49.

5 According to al-Yaʿqūbī (d. early tenth century) ʿUthmān ordered the people to recite ʿalā nuskhatin wāḥidatin (al-Yaʿqūbī, , Historiae, ed. Houtsma, M. Th., vol. 2, Leiden: Brill, 1883, 197Google Scholar).

6 Mingana, , “Transmission of the Kur’ān according to Christian writers”, Muslim World 7, 1917, 402–14CrossRefGoogle Scholar, at 414, citing Casanova, Paul, Mohammed et la fin du monde: Étude critique sur l'Islam primitif (Paris: P. Gauthier, 1911), 141–2Google Scholar.

7 GdQ, vol. 2, 1–121.

8 Crone, Patricia and Cook, Michael, Hagarism: The Making of the Islamic World (Cambridge: Cambridge University Press, 1977), 17–18Google Scholar.

9 Wansbrough, John, Quranic Studies: Sources and Methods of Scriptural Interpretation (Oxford: Oxford University Press, 1977), 49Google Scholar.

10 E.g. Donner, Fred, Narratives of Islamic Origins: The Beginnings of Islamic Historical Writing (Princeton: Darwin Press, 1998), 35–63Google Scholar and Crone, Patricia, “Two legal problems bearing on the early history of the Qur'ān”, Jerusalem Studies in Arabic and Islam 18, 1994, 1–37Google Scholar, here 16–18.

11 Crone, “Two legal problems”.

12 Robinson, Chase, ʿAbd al-Malik (Oxford: Oneworld, 2005), 100–4Google Scholar.

13 de Prémare, Alfred-Louis, Les fondations de l'islam: Entre écriture et histoire (Paris: Éditions du Seuil, 2002), 278–323Google Scholar; de Prémare, , Aux origines du Coran: questions d'hier, approches d'aujourd'hui (Paris: Téraèdre, 2004)Google Scholar; de Prémare, , “ʿAbd al-Malik b. Marwān et le processus de constitution du Coran”, in Ohlig, Karl-Heinz and Puin, Gerd-R. (eds), Die dunklen Anfänge: Neue Forschungen zur Entstehung und frühen Geschichte des Islam (Berlin: Verlag Hans Schiler, 2005), 179–210Google Scholar.

14 Powers, David S., Muḥammad Is Not the Father of Any of Your Men: The Making of the Last Prophet (Philadelphia: University of Pennsylvania Press, 2009)Google Scholar.

15 Shoemaker, Stephen J., The Death of a Prophet: The End of Muhammad's Life and the Beginnings of Islam (Philadelphia: University of Pennsylvania Press, 2012), 136–58Google Scholar.

16 Motzki, Harald, “The collection of the Qur'ān: a reconsideration of Western views in light of recent methodological developments”, Der Islam 78, 2001, 1–34CrossRefGoogle Scholar.

17 Shoemaker, Death, 148.

18 Sadeghi, Behnam and Bergmann, Uwe, “The codex of a companion of the Prophet and the Qur'ān of the Prophet”, Arabica 57, 2010, 343–436CrossRefGoogle Scholar, at 344.

19 Sadeghi, Behnam and Goudarzi, Mohsen, “Ṣanʿā' 1 and the origins of the Qur’ān”, Der Islam 87, 2012, 1–129CrossRefGoogle Scholar. Sadeghi has informed me that the Grand Mosque of Ṣanʿā’ houses forty more folios of the palimpsest.

20 Ibid., 23.

21 Sadeghi and Bergmann, “Codex”, 348 and 353–4. According to Déroche, François (Qur'ans of the Umayyads: A First Overview, Leiden: Brill, 2014: 13)Google Scholar, a carbon dating of two more samples of the Ṣanʿā' palimpsest has been commissioned by Christian Robin, yielding the date ranges 543–643 ce and, bizarrely, 433–599 ce. Since Déroche does not supply further details, it seems preferable for the time being to rely on Sadeghi and Bergmann's results, although further testing is probably called for.

22 The parchment of another early Quranic folio has been dated, on a 95.2 per cent probability, to 609–94; see Dutton, Yasin, “An Umayyad fragment of the Qur'an and its dating”, Journal of Qur'anic Studies 9, 2007, 57–87CrossRefGoogle Scholar, at 63–4. For a discussion of the limits of carbon dating see Déroche, Qur'ans of the Umayyads, 11–14, noting, inter alia, that C14 dating of the famous “Qur’ān of the Nurse” which, according to its colophon was completed in 1020, has yielded a date range between 871 and 986 ce, with a probability of 95 per cent. See also the previous note.

23 Sadeghi and Bergmann, “Codex”, 383–4. Sadeghi's attempt to show that the standard rasm preserves an older prototype of the Quran more faithfully than C-1 will be discussed in the second part of this article.

24 See, for example, Robinson, ʿAbd al-Malik, 104.

25 On the date see Johns, Jeremy, “Archaeology and the history of early Islam: the first seventy years”, Journal of the Economic and Social History of the Orient 46, 2003, 411–36CrossRefGoogle Scholar, at 424–6.

26 The inscriptions are transcribed in Kessler, Christel, “ʿAbd al-Malik's inscription in the Dome of the Rock: a reconsideration”, Journal of the Royal Asiatic Society, 1970, 2–64Google Scholar; for a translation of the inscriptions and the plaques see Whelan, Estelle, “Forgotten witness: evidence for the early codification of the Qur’ān”, Journal of the American Oriental Society 118, 1998, 1–14CrossRefGoogle Scholar.

27 The phrase is lahu l-mulku wa-lahu l-ḥamdu [from Q 64:1; Q 57:2 begins with the similar phrase lahu mulku l-samawāti wa-l-arḍi] yuḥyī wa-yumītu [from Q 57:2] wa-huwa ʿalā kulli shay'in qadīrun [concludes both Q 57:2 and Q 64:1] (Kessler, “ʿAbd al-Malik's inscription”, 4 and 9).

28 Hoyland, Robert, “The content and context of early Arabic inscriptions”, Jerusalem Studies in Arabic and Islam 21, 1997, 77–101Google Scholar, at 87–8.

29 Robinson, ʿAbd al-Malik, 103. Robinson also draws attention to similar divergences in early literary texts, such as Ḥasan al-Baṣrī's letter to ʿAbd al-Malik (cf. Cook, Michael, The Koran: A Very Short Introduction, Oxford: Oxford University Press, 2000, 120–22CrossRefGoogle Scholar).

30 Shoemaker, Death, 148.

31 On the use of Q 112 on Marwanid coinage see Heidemann, Stefan, “The evolving representation of the early Islamic empire and its religion on coin imagery”, in Neuwirth, Angelika, Sinai, Nicolai and Marx, Michael (eds), The Qur'ān in Context: Literary and Historical Investigations into the Qur'ānic Milieu (Leiden: Brill, 2010), 149–95Google Scholar, at 184–6.

32 De Prémare, “Processus de constitution”, 183.

33 Whelan, “Forgotten witness”, 6.

34 Shoemaker, Death, 321, n. 132.

35 Ibn Dāwūd, Abī, Kitāb al-maṣāḥif, ed. in Jeffery, Arthur, Materials for the History of the Text of the Qur'ān: The Old Codices (Leiden: Brill, 1937)Google Scholar, 117 (Arabic text).

36 Fondations, 293–4.

37 See Small, Keith, Textual Criticism and Qur’ān Manuscripts (Lanham, MD: Lexington Books, 2011), 36–44Google Scholar; Déroche, François, La transmission écrite du Coran dans les débuts de l'islam: Le codex Parisino-petropolitanus (Leiden: Brill, 2009), 51–75CrossRefGoogle Scholar.

38 See GdQ, vol. 3, 256. For a different interpretation of the tradition, which presupposes the reading alifay ḥarfin, “the two alifs of a word”, see Hamdan, Omar, Studien zur Kanonisierung des Korantextes: Al-Ḥasan al-Baṣrīs Beiträge zur Geschichte des Korans (Wiesbaden: Harrassowitz, 2006), 135–7Google Scholar. According to Hamdan, what ʿUbaydallāh did was to emend li-llāhi in Q 23:87 and 23:89 to allāhu by inserting two alifs. This interpretation has the merit of allowing one to see how Yazīd al-Fārisī was able to explain ʿUbaydallāh's measure by saying that the latter had been born in the Basran quarter of Kallā’: the reading allāh instead of li-llāhi seems to have been a specifically Basran variant that was reportedly contained in the codex of al-Ḥasan al-Baṣrī (Hamdan, Studien, 136) and in the codex that ʿUthmān had dispatched to Basra (Cook, Michael, “The stemma of the regional codices of the Koran”, Graeco-Arabica 9–10, 2004, 89–104Google Scholar, at 94). Hamdan also quotes a tradition transmitted by al-Dānī which states that ʿUbaydallāh “added two alifs” to Q 23:87–89.

39 See Hamdan, Studien, summarized in Hamdan, “The second Maṣāḥif project: a step towards the canonization of the Qur'anic text”, in Neuwirth et al. (eds), The Qur’ān in Context, 795–835.

40 Ibn Abī Dāwūd, Kitāb al-maṣāḥif, 119–20, and al-Zarkashī, , al-Burhān fī ʿulūm al-Qur’ān, ed. Muḥammad Abū al-Faḍl Ibrāhīm, 4 vols (Cairo: Dār Iḥyā’ al-Kutub al-ʿArabiyya, 1957–8)Google Scholar, vol. 1, 249. In many Biblical manuscripts, similar word counts – called the “final Masorah” – appear at the end of individual books (Tov, Emanuel, Textual Criticism of the Hebrew Bible, 3rd edition, Minneapolis: Fortress Press, 2012, 67Google Scholar).

41 Ibn Khallikān, Wafayāt al-aʿyān wa-anbā’ abnā’ al-zamān, ed. ʿAbbās, Iḥsān (Beirut: Dār Ṣādir, 1972 (according to the last volume))Google Scholar, 8 vols, vol. 2, 32. Hamdan cites a very similar tradition from Ḥamza al-Iṣfahānī (see the Arabic quotation in Studien, 146, n. 84). On Naṣr ibn ʿĀṣim see Sezgin, Fuat, Geschichte des arabischen Schrifttums, 11 vols, 1967–2000, vol. 9, 32–3Google Scholar.

42 Ibn ʿAṭiyya, al-Muḥarrar al-wajīz fī tafsīr al-kitāb al-ʿazīz, ed. Ṣādiq al-Mallāḥ, Aḥmad, 2 vols (Cairo: al-Majlis al-aʿlā li-l-shu’ūn al-islāmiyya, 1974), vol. 1, 66–7Google Scholar. On Yaḥyā ibn Yaʿmar (also a student of Abū al-Aswad al-Du'alī) see Sezgin, Geschichte, vol. 9, 33–4. While Hamdan has pioneeringly worked through an enormous number of Arabic sources, he proceeds on the basis of the questionable assumption that all reports relating to al-Ḥajjāj's interest in the Quranic text or to his interaction with Quran scholars are to be interpreted on the model of a unified editorial project involving the appointment of a “project committee”, the successive implementation of various “project goals”, and finally the publication of the results. This highly orderly framework seems to be inspired by Ibn ʿAṭiyya (on the basis of whom Hamdan, Studien, 140–1, dates al-Ḥajjāj's measures to 703–04), but Hamdan does not address the possibility that the latter's tidy narrative could be a retrospective attempt at imposing some kind of overarching order on the material about al-Ḥajjāj. For instance, apart from Ibn ʿAṭiyya, reports describing how al-Ḥajjāj initiated a counting of the text's consonants and its division into sections do not mention the insertion of diacritics, nor that these measures took place at Wāsiṭ.

43 See al-Mizzī, Tahdhīb al-kamāl fī asmā’ al-rijāl, ed. Maʿrūf, ʿAwwād, 35 vols (Beirut: Mu'assasat al-risāla, 1983–1992), vol. 22, 437–41Google Scholar.

44 Ibn Abī Dāwūd, Kitāb al-maṣāḥif, 49–50 and 117–8; Hamdan, Studien, 166–70. Two examples are lam yatasanna > lam yatasannah (both words can be synonyms: Lane, Edward William, Arabic–English Lexicon, London: Williams and Norgate, 1863, 1149bGoogle Scholar) at Q 2:259 and sharīʿatan > shirʿatan at Q 5:48.

45 Mingana, “Transmission”, 409, and Casanova, Mohammed, 119. For the Arabic text see George Tartar, “Dialogue Islamo-Chrétien sous le calife al-Ma'mun [sic.] (813–834): Les Épîtres d'al-Hâshimî et d'al-Kindî”, Thèse pour le Doctorat de 3^e cycle, Strasbourg 1977, 117–8.

46 Jeffery, Arthur, “Ghevond's text of the correspondence between ʿUmar II and Leo III”, Harvard Theological Review 37, 1944, 269–332CrossRefGoogle Scholar, at 298. See Hoyland, Robert, Seeing Islam as Others Saw It: A Survey and Evaluation of Christian, Jewish and Zoroastrian Writings on Early Islam (Princeton: Darwin Press, 1997), 490–501Google Scholar.

47 Ibn Shabba, Ta'rīkh al-Madīna al-munawwara, ed. Fahīm Muḥammad Shaltūt, 4 vols (Mecca: n.p., 1979)Google ScholarPubMed, vol 1, 7–8; al-Samhūdī, Wafā’ al-wafā bi-akhbār dār al-muṣtafā, ed. Qāsim al-Sāmarrā'ī (London: Mu'assasat al-furqān li-l-turāth al-islāmī), 5 vols, vol. 2, 457. See also Hamdan, Studien, 171, n. 198 and n. 200.

48 On Ibn Zabāla's lost Akhbār al-Madīna see now Munt, Harry, “Writing the history of an Arabian holy city: Ibn Zabāla and the first local history of Medina”, Arabica 59, 2012, 1–34CrossRefGoogle Scholar.

49 “Al-Ḥajjāj ibn Yūsuf sent codices to the metropolises (ummahāt al-qurā), and he sent a big one of these codices to Medina. He was the first who sent codices to the towns (wa-huwa awwalu man arsala bi-l-maṣāḥifi ilā l-qurā) ...”. De Prémare contends that this statement contradicts the traditional narrative about ʿUthmān dispatching copies of his recension to the provincial capitals, for he summarizes al-Samhūdī's report as describing “le premier envoi d'un muṣḥaf officiel dans les capitales, alors que cette primeur est habituellement attribuée à ʿUṯmān” (“Processus de constitution”, 200; similarly Fondations, 296). Yet what generates the purported contradiction is only the fact that de Prémare here equates qurā with “capitales” (the passage is translated correctly, with “capitales” for ummahāt al-qurā and “villes” for qurā, in “Processus de constitution”, 199, and Fondations, 461). However one judges the historicity of ʿUthmān's measures, there is surely no inconsistency between the proposition that ʿUthmān sent Quranic codices to the amṣār and the proposition that al-Ḥajjāj was the first to distribute codices to the qurā.

50 Ibn ʿAbd al-Ḥakam, Futūḥ Miṣr wa-akhbāruhā, ed. Torrey, Charles C. (New Haven: Yale University Press, 1922), 117–8Google Scholar, quoted after Mathieu Tillier, review of Déroche, , Transmission, Journal of Qur'anic Studies 13, 2011, 109–15Google Scholar; Vollers, Karl (ed.), Description de l'Egypte par Ibn Doukmak (Cairo: Imprimerie nationale, 1893)Google Scholar, vol. 1, 72; Mingana, “Transmission”, 231; GdQ, vol. 3, 104, n. 1; Hamdan, Studien, 172, with n. 201; de Prémare, “Processus de constitution”, 198–9.

51 Al-Samhūdī, Wafā', vol. 2, 456–7 (cf. Hamdan, Studien, 172). This information is quoted on the authority of Mālik, who then expresses his disapproval of the innovation. This in turn is followed by a statement defending the reading from codices in mosques, and another tradition, cited from Ibn Shabba, which claims that the practice of having the Quran read from a codex in the mosque every morning was already established by ʿUthmān. The most straightforward reconstruction of the material would seem to be that al-Ḥajjāj was indeed responsible for instituting the practice; that his innovation then became a point of dispute, generating both supporting and disapproving comments; and that defenders of the practice finally took recourse to circulating a legitimizing tradition invoking an earlier precedent by ʿUthmān.

52 Hamdan, Studien, 170–1. Hamdan places this report under the heading “Spreading the new copies of the Quran produced during the Maṣāḥif Project”, but this link is not evident from the quotation itself. Beck, Edmund, “Der ʿuṯmānische Kodex in der Koranlesung des zweiten Jahrhunderts”, Orientalia nova series 14, 1945, 355–73Google Scholar, suggests that al-Ḥajjāj only attempted to eliminate codices used for public recitation and teaching.

53 Al-Farrā', Maʿānī al-Qur'ān, vol. 3, ed. ʿAbd al-Fattāḥ Ismāʿīl Shalabī and ʿAlī al-Najdī Nāṣif (Cairo: al-Hay'a al-miṣriyya al-ʿāmma li-l-kitāb), 1955–72, 68 (ad Q 48:26). I owe this reference to Beck, “Der ʿuṯmānische Kodex”, 355, n. 4.

54 De Prémare, “Processus de constitution”, 189–206.

55 Mingana, “Transmission”, 230; de Prémare, Fondations, 297. De Prémare's interpretation of the utterance is endorsed in Kohlberg, Etan and Amir-Moezzi, Mohammad Ali, Revelation and Falsification: The Kitāb al-qirā’āt of Aḥmad b. Muḥammad al-Sayyārī (Leiden: Brill, 2009)Google Scholar, 20.

56 Muslim ibn al-Ḥajjāj, Ṣaḥīḥ, ed. Muḥammad Fu’ād ʿAbd al-Bāqī, 5 vols (Beirut: Dār Iḥyā’ al-Kutub al-ʿArabiyya, 1991)Google Scholar, vol. 2, 942 (15:50).

57 De Prémare, “Processus de constitution”, 200–01. De Prémare discusses two further statements ascribed to Abd al-Malik and al-Ḥajjāj (ibid., 194–7 and 204–05), but his construal does not appear at all compelling to me.

58 See Motzki, “Collection”.

59 Thus Hoyland, Seeing Islam, 501.

60 This is how the tradition is understood by al-Thaʿālibī, who substitutes khatamtu for jamaʿtu (Laṭā’if al-maʿārif, ed. Muḥammad Ibrāhīm Salīm, Cairo: Dār al-Ṭalā'iʿ, 1992, 110Google Scholar). Note that jamaʿa can undoubtedly have the meaning “to collect in one's heart” = “to learn by heart”, as illustrated by al-Bukhārī, Ṣaḥīḥ, vol. 3, 348, no. 5036 (66:25), citing Ibn ʿAbbās as saying, jamaʿtu l-muḥkama fī ʿahdi rasūli llāhi, and Nuʿaym al-Iṣfahānī, Abū, Ḥilyat al-awliyā' wa-ṭabaqāt al-aṣfiyā', 10 vols (Cairo: Maktabat al-Khānjī / Maṭbaʿat al-Saʿāda, 1932–38)Google Scholar, vol. 1, 285, overlapping with Ibn Mājah, Sunan, ed. Muḥammad Nāṣir al-Dīn al-Albānī (al-Riyāḍ: Maktabat al-maʿārif li-l-nashr wa-l-tawzīʿ, n.d.), 239 (5:178).

61 Unsurprisingly, this is what al-Nawawī assumes the command must mean. He cites a deliberation by al-Qāḍī ʿIyāḍ as to whether the command refers to the canonical order of the sūras or, which is deemed to be the more obvious meaning, to the order of verses within a given sūra (quoted in Muslim, Ṣaḥīḥ, vol. 2, 942, n. 1).

62 Al-Ḥajjāj is said to have threatened to behead anyone reciting Ibn Masʿūd's recension and to “remove it from the codex, if need be even [by scraping it off] with the rib of a pig” (Ibn ʿAsākir, al-Tārīkh al-kabīr, vol. 4, ed. ʿAbd al-Qādir Badrān, Damascus: Maṭbaʿat Rawḍat al-Shām, 1332, 69; for further invectives see de Prémare, “Processus de constitution”, 202–3; Sadeghi and Goudarzi, “Ṣanʿā' 1”, 28–9, n. 62). On ʿUbaydallāh ibn Ziyād's provocative recitation of Q 113 and 114 (missing in Ibn Masʿūd's recension) see Hamdan, Studien, 137–8.

63 Cf. Sadeghi and Bergmann, “Codex”, 365, n. 36.

64 Hamdan (Studien, 146–8) accepts that al-Ḥajjāj initiated the use of diacritics in Quran manuscripts (presumably on the basis of Ibn Khallikān and Ḥamza al-Iṣfahānī), but rejects Ibn ʿAṭiyya's claim that al-Ḥajjāj also introduced vowel signs. This view is confirmed by the fact that Yaḥyā ibn Yaʿmar and Naṣr ibn ʿĀṣim, the two Basran Quran readers who are portrayed as working for al-Ḥajjāj by Ibn ʿAṭiyya and Ibn Khallikān (see notes 41 and 42 above), both figure as “the first person to have dotted codices” in traditions cited by al-Dānī, al-Muḥkam fī naqṭ al-maṣāḥif, ed. ʿIzzat Ḥasan, Damascus: Maṭbūʿāt Mudīriyyat Iḥyā’al-Turāth al-Qadīm, 1960, 5–6 (main text). Three caveats are in order here: (i) Manuscripts and papyri show that it would be anachronistic to conceive of al-Ḥajjāj's codices as fully dotted (see Small, Textual Criticism, 16–30 and Kaplony, Andreas, “What are those few dots for? Thoughts on the orthography of the Qurra Papyri (709–710), the Khurasan Parchments (755–777) and the inscription of the Jerusalem Dome of the Rock (692)”, Arabica 55, 2008, 91–112CrossRefGoogle Scholar). (ii) The extent to which Quranic manuscripts employed diacritical marks continued to vary considerably during the following centuries (cf. Small, Textual Criticism, 22–3, on BNF Arabe 333c). (iii) Diacritics as such are older; see Adolf Grohmann, Arabische Paläographie. II. Teil: Das Schriftwesen. Die Lapidarschrift, Vienna: Hermann Böhlaus Nachf., 1971, 41; ʿAli ibn Ibrahim Ghabban and Hoyland, Robert, “The inscription of Zuhayr, the oldest Islamic inscription (24 ah/ad 644–645), the rise of the Arabic script and the nature of the early Islamic state”, Arabian Archaeology and Epigraphy 19, 2008, 209–36Google Scholar).

65 See Sadeghi and Bergmann, “Codex”, 365, n. 36. The skeletal modification that is most likely to stem from a conscious decision to correct the text is the alleged substitution of li-llāhi at Q 23:87.89 by allāh: since the two verses quote the answer to a preceding question formed with man, the variant allāh certainly makes for a smoother text. Yet already the ʿUthmānic codex sent to Basra reportedly had allāh instead of li-llāhi, and the alteration li-llāhi > allāh is also ascribed to ʿUbaydallāh ibn Ziyād (see above, n. 38). Al-Ḥajjāj's text may therefore simply have followed an existing Basran reading.

66 On the deposition of master copies as a form of publication see Gregor Schoeler, “Writing and publishing: on the use and function of writing in the first centuries of Islam”, Arabica 44, 1997, 423–35.

67 As late as 323/935, the Quran reader Ibn Shannabūdh was tried for reciting variants deviating from the standard rasm (see Melchert, Christopher, “Ibn Mujāhid and the establishment of seven Qur'anic readings”, Studia Islamica 91, 2000, 5–22CrossRefGoogle Scholar). Note, however, that al-Ḥajjāj seems to have targeted not just the recitation of non-ʿUthmānic variants, but proper non-ʿUthmānic codices.

68 See above, n. 65. I owe this point to a comment by Behnam Sadeghi.

69 A French translation of the passage is contained in Tillier's review of Déroche (see above, n. 50), which was kindly brought to my attention by Marie Legendre.

70 On the text and the question of its date see Hoyland, Seeing Islam, 459–65.

71 See Hoyland, Seeing Islam, 194–200.

72 Mingana, “Transmission”, 406.

73 Motzki, “Compilation”, 14.

74 Griffith, Sydney H., “Disputing with Islam in Syriac: the case of the monk of Bêt Ḥālê and a Muslim emir”, Hugoye 3.1, 2000, 29–54Google Scholar, citing 34.

75 De Prémare, “Processus de constitution”, 184–9.

76 Crone and Cook, Hagarism, 17 with n. 14.

77 Hoyland, Seeing Islam, 465–72.

78 Quoted (slightly modified) according to Griffith, “Disputing with Islam”, 47–8. The end of the sentence is garbled; “G-y-g-y” and “T-w-r-h” may refer to the Gospel (Arabic injīl) and the Torah (Hoyland, Seeing Islam, 471–2, and Griffith, “Disputing with Islam”, 47).

79 Hoyland, Seeing Islam, 471.

80 The authenticity of the chapter on Islam has been challenged, but see the discussion in Sahas, Daniel J., John of Damascus on Islam: The “Heresy of the Ishmaelites” (Leiden: Brill, 1972, 60–66)Google Scholar. On the date of John of Damascus's death see ibid., 47–8, and Hoyland, Seeing Islam, 482–3.

81 De Prémare, “Processus de constitution”, 186; Sahas, John of Damascus, 89–93.

82 Sahas, John of Damascus, 91.

83 De Prémare connects his hypothesis to Muqātil ibn Sulayman's commentary on Q 26:155–8, which he understands to preserve “les traces d'un texte antérieur aux différents passages coranique actuels sur la chamelle de Ṯamūd” (“Processus de constitution”, 188). He very much works with a Wansbroughian analysis of the Tafsīr Muqātil here, which views the occasionally seamless interposition of brief expansions and additions between scriptural segments as documenting a stage when proto-Quranic material was still closely linked with proto-exegetical material (see Wansbrough, Quranic Studies, 119–48). For a different description of the literary makeup of the Tafsīr Muqātil see Sinai, Nicolai, Fortschreibung und Auslegung: Studien zur frühen Koraninterpretation (Wiesbaden: Harrassowitz, 2009)Google Scholar.

84 De Prémare, “Processus de constitution”, 186.

85 One might rejoin that the ultimate triumph of the Marwanid Quran came only after a protracted struggle spanning several decades (i.e. after the 730s), but this would aggravate the challenge of explaining why in the end all Muslim groups unanimously adopted it, without leaving behind any literary trace of the entire process (see the second part of this article).

86 Robinson, ʿAbd al-Malik, 100–4; cf. Kohlberg and Amir-Moezzi, Revelation and Falsification, 20–23.

87 Robinson, ʿAbd al-Malik, 102.

88 Schacht, Joseph, The Origins of Muhammadan Jurisprudence (Oxford: Oxford University Press, corrected edition, 1953)Google Scholar, 224.

89 Motzki, Harald, The Origins of Islamic Jurisprudence: Meccan Fiqh before the Classical Schools, translated by Katz, Marion H. (Leiden: Brill, 2002), 108–17Google Scholar.

90 Crone, “Two legal problems”, 37.

91 Ibid., 19.

92 But see the critical assessment in Barnes, Jonathan, “Roman Aristotle”, in Philosophia Togata II, ed. Barnes, Jonathan and Griffin, Miriam, Oxford: Oxford University Press, 1997, 1–69Google Scholar.

93 This is pointed out by Sadeghi and Goudarzi (“Ṣanʿā’ 1”, 3, n. 3), who remark that Crone argues less for “a late date for the attainment of textual stability” than for “the late canonization of a largely stable text”.

94 Crone, “Two legal problems”, 14. The following two paragraphs are based on Nicolai Sinai, Fortschreibung und Auslegung, 39–58 and 261–7.

95 As Christopher Melchert aptly puts it, “the Qur'an was not primarily a collection of propositions to be looked up but a liturgy to be recited” (“Ibn Mujāhid and the establishment of seven Qur'anic readings”, Studia Islamica 91, 2000, 5–22Google Scholar, citing 16). William Graham has found that in prophetic traditions the term qur’ān occurs mainly in the context of prayer and other devotional practices (“The earliest meaning of ‘Qur’ān’”, Die Welt des Islams 23/24, 1984, 361–77Google Scholar). This is not necessarily to deny that there may have been a limited use of Quranic material in early Islamic theology and law, as reflected, for example, in the so-called Epistle of al-Ḥasan al-Baṣrī: but the Epistle – the early dating of which is criticized in Mourad, Suleiman Ali, Early Islam between Myth and History: Al-Ḥasan al-Baṣrī (d. 110h/728ce) and the Formation of His Legacy in Classical Islamic Scholarship (Leiden: Brill, 2006), 161–239Google Scholar – at most shows that given a controversial theological or legal issue, early Muslims did indeed equip themselves with suitable scriptural ammunition against their opponents, not that they would necessarily have subjected the entire corpus to a sustained analysis.

96 See Graham, William, Beyond the Written Word: Oral Aspects of Scripture in the History of Religion (Cambridge: Cambridge University Press), 110–5Google Scholar.

97 Bulliet, Richard W., Islam: The View from the Edge (New York: Columbia University Press, 1994)Google Scholar, 29. Cf. also the anecdotes indicating very limited scriptural knowledge on the part of some early Muslims gathered in GdQ, vol. 2, 7–8.

98 This description is inspired by John Burton, according to whom Quranic pronouncements entered Islamic legal discourse – i.e. took on the status of a normative source – only after they had already attracted a substantial amount of narrative amplification (for an illustration of this view see his “Law and exegesis: the penalty for adultery in Islam”, in Hawting, Gerald R. and Shareef, Abdul-Kader A. (eds), Approaches to the Qur'ān (London: Routledge, 1993), 269–84Google Scholar).

99 Crone, “Two legal problems”, 20. On Burton's distinction between the Quran as a document and as a source see his The Collection of the Qur'ān (Cambridge: Cambridge University Press, 1977)Google Scholar, 111 and 187.