𝐅𝐨𝐫𝐠𝐨𝐭𝐭𝐞𝐧 𝐖𝐢𝐭𝐧𝐞𝐬𝐬: 𝐄𝐯𝐢𝐝𝐞𝐧𝐜𝐞 𝐅𝐨𝐫 𝐓𝐡𝐞 𝐄𝐚𝐫𝐥𝐲 𝐂𝐨𝐝𝐢𝐟𝐢𝐜𝐚𝐭𝐢𝐨𝐧 𝐎𝐟 𝐓𝐡𝐞 𝐐𝐮𝐫’𝐚𝐧
Mohamad Mostafa Nassar
(We do not necessarily agree with everything written in this article – Islam Compass
Three kinds of historical evidence are examined here that have not previously being seriously considered in relation to the question of codification. The Umayyad inscriptions from the Dome of the Rock have generally being ignored or dismissed because of apparent departures from the “canonical” text, as represented by the Cairo edition; here they are analyzed for the evidence they nonetheless provide for the state of the Qur’anic text toward the end of the first hijri century.
Equally informative are al-Walid’s inscriptions at the Great Mosque of al-Madinah, datable about twenty years later; they were described by eyewitnesses in the first half of the tenth century, when they were still partly visible. Finally, from scattered indications it is suggested that there was group of professional Qur’an copyists at al-Madinah at the end of the first and the beginning of the second century.
In the last two decades a controversy has arisen over the period in which the text of Muslim scripture became codified. The traditional Islamic view can be summarized as follows. Both Abu Bakr (A.H. 11-13 / A.D. 632-34) and ‘Umar (13-23 / 634-44) made efforts to gather together the scraps of revelation that had been written down by the faithful during the lifetime of the Prophet, on bones, on palm leaves, on potsherds, and on whatever other materials were at hand, as well as being preserved in “the breasts of men.”
But it was the third caliph, ‘Uthman (23-35 / 644-61), who first charged a small group of men at al-Madinah with codifying and standardizing the text. Alarmed by reported divergences in the recitation of the revelation,
he commissioned one of the Prophet’s former secretaries, Zayd b. Thabit, and several prominent members of Quraysh – ‘Abd Allah b. al-Zubayr, Sa‘id b. al-‘As, and ‘Abd al-Rahman b. al-Harith are those most often mentioned – to produce a standard copy of the text, based on the compilation in the keeping of Hafsah, daughter of ‘Umar. If there was disagreement over language among members of the commission, it was to be resolved in accordance with the dialect spoken by Quraysh.
Once the standard text had been established, several copies were made and sent to major cities in the Islamic domain, specifically Damascus, al-Basrah, al-Kufah, and perhaps others. Although there are variations in detail, for example, in the list of names of those who served on ‘Uthman’s commission and in the list of cities to which copies were sent, this basic outline is not in dispute within the Muslim world.
Oral recitation nevertheless remained the preferred mode of transmission, and, as time passed, variant versions of the text proliferated – the kind of organic change that is endemic to an oral tradition. In addition, because of the nature of the early Arabic script, in which short vowels were not indicated and consonants of similar form were only sometimes distinguished by pointing, writing, too, was subject to misunderstanding, copyist’s error,
and change over time. In the early tenth century, at Baghdad, Abu Bakr Ibn Mujahid (d. 324 / 936) succeeded in reducing the number of acceptable readings to the seven that were predominant in the main Muslim centers of the time: al-Madinah, Makkah, Damascus, al-Basrah, and al-Kufah. Some Qur’an readers who persisted in deviating from these seven readings were subjected to draconian punishments. Nevertheless, with the passage of time, additional variant readings were readmitted, first “the three after the seven,” then “the four after the ten.”
The modern Cairo edition, prepared at al-Azhar in the 1920s, is based on one of the seven readings permitted by Ibn Mujahid, that of Abu Bakr ‘Âsim (d. ca. 127 / 745) as transmitted by Hafs b. Sulayman (d. 180 / 796).
Early efforts by Muslim scholars to establish the sequence of the revelation, particularly the verses revealed at Makkah and those revealed at al-Madinah, were emulated by European scholars, who focused on similar problems, though often adopting somewhat different criteria for determining solutions. Nevertheless, already in the early twentieth century Alphonse Mingana seriously challenged the entire historical framework outlined here.
Mingana, whose approach was patently tendentious, argued that the Qur’an had not been codified in book form until several decades later than was generally accepted, in the reign of the fifth Umayyad caliph, ‘Abd al-Malik b. Marwan (65-86 / 685-705). In the 1970s John Wansbrough went much farther, concluding, on the basis of textual and linguistic analysis, that there is no evidence for a “canonical” version of the Qur’anic text before the very end of the eighth century at the earliest.
Wansbrough argued that the nature of the text itself presupposes “an organic development from originally independent traditions during a long period of transmission . . . juxtaposition of independent pericopes to some extent unified by means of a limited number of rhetorical conventions.”
In support of his conclusion he noted that Muslim traditions about early revelation, indeed about the life of the Prophet and early Muslim history as a whole, are known only from later Islamic literature; Qur’anic exegesis, for example, first evolved in the late eighth and ninth centuries.
Nor can most early Muslim traditions be confirmed in contemporary non-Muslim sources. Wansbrough’s entire analysis was based on the assumption that the “canonization” of the Qur’anic text and its role in the development of the Muslim community followed a trajectory similar to that of Hebrew scripture.
For example, in connection with “exegetical” (Wansbrough’s characterization of much of the content of the Sirah of Ibn Ishaq, ca. 85-150 / 704-67, edited by Ibn Hisham, d. 218 / 833) reports of material that also appears in the “canon,” he declared:
“For Hebrew scripture the priority in time of such reports over the actual reproduction in literary form of prophetical utterances has been established. To postulate a similar, if not identical, process for Muslim scripture seems to me not unjustified, though in this particular instance complicated by the redaction history of the Sira itself.”
He also cited “the likelihood of a Rabbinic model for the account of an authoritative text produced in committee, namely the Jamnia tradition on the canonization of Hebrew scripture.” The vastly different historical contexts in which these supposedly parallel processes took place were not explicitly recognized or taken into account in Wansbrough’s literary analysis. In fact the results of this analysis were frequently cited as grounds for rejecting the supposed historical evidence presented in such texts as the Sirah.
By means of this reasoning Wansbrough arrived at the conclusion that “concern with the text of scripture did not precede by much the appearance of the masoretic [exegetical] literature as it has in fact been preserved”:
that is, in his view the Qur’anic text assumed its canonical form more or less simultaneously with the appearance of commentaries on it (tafsir). He took as confirmation of this view Joseph Schacht’s conclusion that the Qur’anic text did not serve as a basis for Muslim law before the ninth century.
Particularly crucial to Wansbrough’s argument is the term “canonical,” for which he assumes a high standard of precision.
It is clear that even in the Muslim tradition the fact was acknowledged that readings of the Qur’an continually diverged from a supposed original; it is clear also that steps had repeatedly to be taken to impose or protect a unitary text of revelation – in the time of ‘Uthman, again in the time of Ibn Mujahid, and even as recently as the 1920s, when scholars at al-Azhar prepared the currently most widely used edition.
This edition is nonetheless not treated as uniquely “canonical” in parts of India and North Africa, where versions that differ in titles of the surahs, divisions between ayat, and occasionally vocalizations are in use; furthermore, it is clear from surviving manuscripts that such variants have persisted through the history of Islam. Wansbrough’s difficulty appears to be that these divergences are not substantive but rather involve details that he perceives as formalistic, perhaps even trivial.
Yet there is abundant evidence from the relatively well-documented period of the ninth and tenth centuries that such divergences were not perceived as trivial within Islam itself.
Perhaps the most valuable results of Wansbrough’s study for the historian are his analyses of aspects of the text that, though already familiar, had not previously been so carefully delineated or explored in all their implications.
One of these aspects is the polemical character of much of the Qur’an, which, as Wansbrough convincingly demonstrates, was focused on Jewish scripture and tradition, implying an important Jewish opposition as one of the motivations behind the “canonization” of Islamic scripture.
A second is the nature of the text itself, a series of “independent pericopes” placed side by side but expressed in a unified language and style.
The essential challenge to historians of the early Islamic period is to reconcile these undeniably useful observations with historical evidence that Wansbrough has not admitted into his analysis. Because of the relentless opacity of his own writing style it is tempting to ignore this challenge, but the implications of his argument are too far-reaching to permit such self-indulgence.
It is important to recognize that his analysis was guided predominantly by generalizations drawn from the history of the biblical text, which were then applied to Muslim scripture. Most formidable is the conclusion, not stated explicitly but inescapable from Wansbrough’s analysis, that the entire Muslim tradition about the early history of the text of the Qur’an is a pious forgery, a forgery so immediately effective and so all-pervasive in its acceptance that no trace of independent contemporary evidence has survived to betray it.
An important related issue involves the dating of early manuscripts of the Qur’an. If Wansbrough is correct that approximately a century and a half elapsed before Muslim scripture was established in “canonical” form, then none of the surviving manuscripts can be attributed to the Umayyad or even the very early ‘Abbasid period; particularly, one controversial manuscript discovered in San‘a’ in the 1970s, no. 20-33.1, for which a date around the turn of the eighth century has been proposed, would have to have been copied at a much later period.
The purpose of the present study is to call attention to some types of evidence that Wansbrough did not take into account and that seem to contradict the historical conclusions that he has drawn from his essentially ahistorical analysis.
Primary documents for the condition of the Qur’anic text in the first century of Islam are ‘Abd al-Malik’s two long inscriptions in blue-and-gold glass mosaic, which encircle respectively the inner and outer faces of the octagonal arcade at the Dome of the Rock in Jerusalem.
They are still preserved in their entirety, except for the substitution of the name of the ‘Abbasid al-Ma’mun (198-218 / 813-33) for that of ‘Abd al-Malik; al-Ma’mun did not, however, change the foundation date included by his predecessor, 72 / 691-92, which thus ensures that the inscriptions were actually executed in the reign of ‘Abd al-Malik.
The main inscription consists of brief invocations combined with a series of passages taken from what are now various parts of the Qur’an, all concerned with a single theme – challenging Christian dogma in the main Christian pilgrimage city.
The text was originally read as a single inscription by Max van Berchem, who began with the outer face of the arcade and thus located the foundation notice in the middle, supposedly followed by the Qur’anic verses on the inner face of the same arcade; this sequence has been accepted without question by most subsequent scholars.
Van Berchem’s arrangement contradicts the normal sequence of Islamic foundation inscriptions, in which the foundation notice occurs at the end. In fact, this mosaic text should be recognized as comprising two distinct inscriptions. As Christel Kessler has transcribed them, it is clear that the band on the inner face of the arcade contains the main message.
The outer inscription is experienced first by those entering the building, who may read only the proximate segment, but the main text, on the inner face of the arcade, was meant to be read in its entirety by those who were returning as they had entered, which involved circumambulation of the middle ambulatory.
It begins on the south side of the octagon with part of the shahadah, the declaration of faith, in the same form in which it appears on the reform coinage of ‘Abd al-Malik introduced five years later, and is followed by a series of excerpts from different parts of the Qur’an as it is now constituted:
“In the name of God, the Merciful the Compassionate. There is no god but God. He is One. He has no associate” [beginning of the shahadah]. “Unto Him belongeth sovereignity and unto Him belongeth praise. He quickeneth and He giveth death; and He is Able to do all things” [a conflation of 64:1 and 57:2]. “Muhammad is the servant of God and His messenger” [variant completion of the shahadah], “L//o! God and His angels shower blessings on the Prophet.
O ye who believe! Ask blessings on him and salute him with a worthy salutation” [33:56 complete]. “The blessing of God be on him and peace be on him, and may God have mercy” [blessing, not in the Qur’anic text]. “O, People of the Book! Do not exaggerate in your // religion (dini//kum) nor utter aught concerning God save the truth.
The Messiah, Jesus son of Mary, was only a messenger of God, and His Word which He conveyed unto Mary, and a spirit from Him. So believe in God and His messengers, and say not ‘Three’ – Ce//ase! (it is) better for you! – God is only One God. Far be it removed from His transcendent majesty that He should have a son. His is all that is in the heavens and all that is in the earth.
And God is sufficient as Defender. The Messiah will never scorn to b//e a servant unto God, nor will the favoured angels. Whoso scorneth His service and is proud, all such will He assemble unto Him” [4:171-72 complete]. “Oh God, bless Your messenger and Your servant Je//sus son of Mary” (interjection introducing the following passage). “Peace be on him the day he was born, and the day he dies, and the day he shall be raised alive!” [19:33 complete, with change from first to third person].
“Such was Jesus, son of Mary, (this is) a statement of the truth concerning which they doubt. It befitteth not (the Majesty of) God that He should take unto Himself a son. Glory be to Him! Wh//en He decreeth a thing, He saith unto it only: Be! and it is” [19:34-35 complete]. Lo! God is my Lord and your Lord. So serve Him.
That is the right path” [19:36 complete, except for initial “and”]. “God (Himself) is witness that there is no God save H//im. And the angels and the men of learning (too are witness). Maintaining His creation in justice, there is no God save Him, the Almighty, the Wise. Lo! religion with God (is)
The Surrender (to His will and guidance). Those who (formerly) received the Book differed only after knowledge came unto them, through transgression among themselves. Whoso disbelieveth the revelations of God (will find that) lo! God is swift at reckoning” [3:18-19 complete].
The outer inscription also begins on the south side:
“In the name of God, the Merciful the Compassionate. There is no god but God. He is One. He has no associate” [beginning of the shahadah]. “Say: He is God, the One! God, the eternally Besought of all! He begetteth not nor was begotten. And there is none comparable unto Him” [112 complete except for the introductory basmalah]. “Muhammad is the Messenger of God” [completion of the shahadah], “the blessing of God be on him” [blessing].//
“In the name of God, the Merciful the Compassionate. There is no god but God. He is One. He has no associate. Muhammad is the Messenger of God” [shahadah, complete]. “Lo! God and His angels shower blessings on the Pro//phet. O ye who believe! Ask blessings on him and salute him with a worthy salutation” [33:56 complete].
“In the name of God, the Merciful the Compassionate. There is no god but God. He is One” [beginning of the shahadah]. “Pra//ise be to God, Who hath not taken unto Himself a son, and Who hath no partner in the Sovereignty, nor hath He any protecting friend through dependence.
And magnify Him with all magnificence” [17:111 complete except for the initial “And say”.]. “Muhammad is the Messenger of G//od” [completion of the shahadah], “the blessing of God be on him and the angels and His prophets, and peace be on him, and may God have mercy” [blessing].
“In the name of God, the Merciful the Compassionate. There is no god but God. He is One. He has no associate” [beginning of the shahadah]. “Unto // Him belongeth sovereignty and unto Him belongeth praise. He quickeneth and He giveth death; and He is Able to do all things” [conflation of 64:1 and 57:2].
“Muhammad is the Messenger of God” [completion of the shahadah], “the blessing of God be on him. May He accept his intercession on the Day of Judgment on behalf of his people” [blessing and prayer].//
“In the name of God, the Merciful the Compassionate. There is no god but God. He is One. He has no associate. Muhammad is the Messenger of God” [the shahadah complete], “the blessing of God be on him” [blessing].
“The servant of God ‘A//bd [Allah the Imam al-Ma’mun, Commander] of the Faithful, built this dome in the year two and seventy. May God accept from him and be content with him. Amen, Lord of the worlds, praise be to God” [foundation notice].
With minor variations, these Qur’anic passages reflect the text as known from the standard Cairo edition, and it is possibly the existence of these inscriptions that led Mingana to propose that the original codification of the Qur’an had taken place during the caliphate, not of ‘Uthman, but of ‘Abd al-Malik.
It is, in fact, puzzling that, although the inscriptions in the Dome of the Rock have been known to scholars for more than a century and have repeatedly been the subject of interpretation, little attention has been paid to the elements from which they were composed. On the inner face of the octagon the declaration of faith is followed by conflated verses describing the powers of God.
Next the Prophet is introduced, with a blessing that, though not directly quoted from the Qur’an, was clearly already in use in 72 / 694. Then comes an exhortation to Christians that Jesus was also a prophet and mortal, followed by the claim that God is sufficient unto Himself. Finally comes a command to bend to His will and the threat of reckoning for those who do not.
The inscription on the outer face consists, as Kessler has pointed out, of six sections set apart by ornaments, the last being the actual foundation notice. Each of the other five sections begins with the basmalah. In each of the first four it is followed by the Umayyad shahadah and a Qur’anic verse arrayed in such a way as to form a self-contained and coherent statement, followed by a blessing on the Prophet. The fifth section is the complete shahadah alone. Each of these sections is thus a miniature composition encapsulating the major themes of the inscription on the inner face.
Within this context it is clear that the minor textual variations noted have been introduced to fit the sense. Such alteration of the standard Qur’anic text in order to express a particular theme seems always to have been acceptable in Islamic inscriptions, however rigidly the actual recitation of the Qur’an may have been regulated; even inscriptions of much later dates, when there is no question that a “canonical” text of the Qur’an had been established, embody such variations.
It is difficult to believe that the selection and coherent arrangement of passages in the time of ‘Abd al-Malik would not have influenced the “canonical” arrangement of the text had codification taken place in his reign or later. It seems particularly unlikely that the combination of phrases from 64:1 and 57:2, repeated twice, could originally have been a unitary statement that was then “deconstructed” and incorporated into different parts of the Qur’an.
Nevertheless, the types of minor variation mentioned, juxtaposition of disparate passages, conflation, shift of person, and occasional omission of brief phrases, led Patricia Crone and Michael Cook to question the value of the mosaic inscriptions at the Dome of the Rock as evidence for the “literary form” of the text as a whole at that early date.
Their skepticism appears to have been engendered rather by two contemporary inscriptions on hammered copper plaques installed on the exterior faces of the lintels over the inner doors in the eastern and northern entrances respectively:
“There is extensive agreement with our text in [the mosaic inscriptions] . . .; on the other hand, there is extensive deviance from our text in [the copper plaques]. . . .” Closer scrutiny of the two copper plaques suggests that the question is not one of “extensive deviance”; rather, one inscription is not primarily Qur’anic in character, and the other is a combination of Qur’anic fragments and paraphrases that makes sense only as a manipulation of a recognized standard text.
The copper plaques include, respectively, seven and four lines of the Umayyad originals; in each instance the remainder of the text, no doubt including an original foundation inscription in the name of ‘Abd al-Malik, was replaced by an attached sheet of copper inscribed in the name of al-Ma’mun – substitutions comparable to that at the end of the outer mosaic inscription.
In the first instance, the plaque over the eastern entrance, the remaining lines (indicated below by paragraph breaks) of the original inscription contain the following text:
“In the name of God the Merciful, the Compassionate” [basmalah], “praise be to God than Whom there is no god but He” [tahmid], “the Living, the Eternal, the Originator of the heavens and the earth and the Light of the heavens and the earth and the Pillar of the heavens and the earth, the One, the eternally Besought of all” [a series of epithets] – “He begotteth not nor was begotten and there is none comparable unto Him” [112:3-4], “Owner of Sovereignty! Thou givest sovereignty unto whom Thou wilt, and Thou
withdrawest sovereignty from whom Thou wilt” [3:26]; “all sovereignty belongs to You and is from You, and its fate is (determined) by You, Lord of glory
the Merciful, the Compassionate” [words of praise]. “He hath prescribed for Himself mercy” [6:12], “and His mercy embraceth all things” [7:156, with shift from first to third person], “may He be glorified and exalted” [words of praise]. “As for what the polytheists associate (with You), we ask You, oh God by
Your mercy and by Your beautiful names and by Your noble face and Your awesome power and Your perfect word, on which are based the heavens and the earth and
through which we are preserved by Your mercy from Satan and are saved from Your punishment (on) the Day of Judgment and by Your abundant favor and by Your great grace and forbearance and omnipotence
and forgiveness and liberality, that You bless Muhammad Your servant, Your prophet, and that You accept his intercession for his people, the blessing of God be upon him and peace be upon him and the mercy of God and” [prayer] . . . .
The northern portal inscription begins in a fashion identical to that on the eastern portal but incorporates more passages from the Qur’anic text:
“In the name of God the Merciful, the Compassionate” [basmalah], “praise be to God than Whom there is no god but He” [tahmid], “the Living, the Eternal”; “He has no associate, the One, the eternally Besought of all” [epithets] – “He
begetteth not nor was begotten, and there is none comparable unto Him” [112:3-4, as in the eastern portal inscription] – “Muhammad is the servant of God” [introductory statement] “and His messenger, whom He sent with the guidance and the religion of truth, that He may make it conqueror of all religion,
however much idolators may be averse” [61:9, with an adjustment at the beginning to introduce Muhammad]; “we believe in God and that which was revealed unto Muhammad and that which the Prophets received from their Lord. We make no distinction between any of them, and unto Him
we have surrendered” [2:136 or 3:84, with change of person and omission of the central section, where Ibrahim, Isma‘il, Ishaq, Ya‘qub, the “tribes,” Musa, and ‘Isa are mentioned individually], “the blessing of God be upon Muhammad, His servant and His prophet, and peace be upon him and the mercy of God and His blessing and His forgiveness and His acceptance . . .” [blessing].
The copper inscriptions do not appear to represent “deviations” from the current standard text; rather, they belong to a tradition of using Qur’anic and other familiar phrases, paraphrases, and allusions in persuasive messages, in fact sermons, whether actual khutbahs or not. Of a number of such texts two examples cited by al-Tabari should suffice to demonstrate the point.
In a sermon supposedly delivered to the people of Khunasirah in northern Syria in 101 / 719-20, ‘Umar b. ‘Abd al-‘Aziz included the phrase “nor will you be left aimless,” a clear reference to Qur’an 75:36: “Thinketh man that he will be left aimless?” A more extended example, involving some of the same passages used at the Dome of the Rock, is the first part of a sermon delivered by Da’ud b.‘Isa, governor of Makkah, in 196 / 811-12.
“Praise be to God, Owner of Sovereignty unto whom He wills and withdraws sovereignty from whom He wills, who exalts whom He wills and abases whom He wills. In His hand is the good; He is Able to do all things” [3:26, with change from direct address to God to the descriptive third-person singular].
“I bear witness that there is no God save Him . . . there is no God save Him, the Almighty, the Wise” [3:18, with shift from the third-person plural to the first-person singular and concomitant omission of references to angels and men of learning as beating witness].
“And I bear witness that Muhammad is His servant and His messenger, whom He sent to bring the religion, through whom He sealed the prophets” [further declaration of faith] “and whom He made a mercy for the peoples” [21:107, with shift from first-person plural to third-person singular].
A narrow focus on the Qur’anic text and continued efforts to establish and preserve a standard version without deviation have persisted throughout the history of Islam, but side by side with that concern there has been a tradition of drawing upon and modifying that text for a variety of rhetorical purposes. Such creative use of familiar scriptural associations was hardly unique to Islam, and indeed it would be more surprising if no such tradition had developed.
The tradition was, however, dependent upon recognition of the text by the listeners, or readers – a strong indication that the Qur’an was already the common property of the community in the last decade of the seventh century. The inscriptions at the Dome of the Rock should not be viewed as evidence of a precise adherence to or deviation from the “literary form” of the Qur’anic text;
rather they are little sermons or parts of a single sermon addressed to an audience that could be expected to understand the allusions and abbreviated references by which ‘Abd al-Malik’s particular message was conveyed.
They thus appear at the beginning of a long tradition of creative use of the Qur’anic text for polemical purposes. The brief Qur’anic passages on coins issued from the time of ‘Abd al-Malik’s reform in 77 / 697 to the end of the dynasty in 132 / 750 are additional examples of such use;
these passages include, in addition to the shahadah, verses 112:1-3 (or 4) complete (except for the initial basmalah and the introductory word “say”) and part of 9:33, with slight variations in the reading of the latter, so that it makes sense by itself:
“He sent him with the guidance and the Religion of Truth, that He may cause it to prevail over all religion. . . .” In parallel to the contemporary inscriptions at the Dome of the Rock these extracts are clearly intended to declare the primacy of the new religion of Islam over Christianity, in particular.
More instructive in relation to the literary form of the Qur’anic text is the inscription on the qiblah wall of the Mosque of the Prophet at al-Madinah, long since lost but observed and described by Abu ‘Ali Ibn Rustah during the pilgrimage of 290 / 903.
According to him, this inscription, which extended from the Bab Marwan (Bab al-Salam) in the western wall around the southwestern corner and across the qiblah wall, then around the southeastern corner to the Bab ‘Ali Bab Jibril), began with Umm al-Qur’an, that is, surah 1, complete, then continued with “wa-al-Shams wa- duhaha” through “Qul: A‘udhu birabb al-nas” to the end, thus the complete text of surahs 91-114.
Ibn Rustah’s report was confirmed by the eyewitness account of an anonymous Spanish traveler, who visited the Haramayn between 307 / 920 and 317 / 929 and reported that the inscriptions consisted of “the short chapters” of the Qur’an. According to this traveler, the inscription was written in five lines of gold on a blue ground contained within a marble panel; it was thus probably executed in gold-and-blue glass mosaic, as at the Dome of the Rock.
This conjecture is confirmed by a report given by al-Tabari: “[I]t was as if I had entered the mosque of the Prophet of God and I raised my head and looked at the writing in mosaic that was in the mosque and there was what the Commander of the Faithful al-Walid b. ‘Abd al-Malik had ordered.” Another parallel to the Dome of the Rock was the inscription’s characters, described as squat and thick, in a stroke the width of a finger.
The inscription belonged to the reconstruction of the mosque sponsored by ‘Abd al-Malik’s son al-Walid I (86-96 / 705-15) and carried out between 88 / 706 and 91 / 710 by his governor in the city, ‘Umar b. ‘Abd al-‘Aziz. Because of this early date it is particularly significant, for three reasons.
First, it suggests that the sequence of the Qur’anic text from surahs 91 to 114 had already been established by 91 / 710. Second, the clustering of the short surahs in this sequence probably means that the arrangement of the entire Qur’an generally in the order of the length of the surahs had already been adopted.
Finally, surahs 1 and 113-14, which the compiler of one pre-‘Uthmanic codex, ‘Abd Allah b. Mas‘ud (d. 32/653), had supposedly refused to accept as part of the revelation, had already been incorporated into the text. ‘Umar b. ‘Abd al-‘Aziz, the one Umayyad whose piety was respected even by the ‘Abbasid enemies of his family, is unlikely to have admitted anything but the officially recognized version of the Qur’anic text; indeed, the inclusion of these passages at the Prophet’s own mosque may have constituted official recognition.
The choice of this extended passage for the qiblah wall is difficult to explain in terms of a single coherent message. It appears from a study of reports by Ibn Rustah and other observers that there had been an inscription of al-Walid on the southern facade of the courtyard, which had been destroyed by Kharijites in 130/747, during the reign of Marwan II (127-32/744-50).
It would have been appropriate in the Prophet’s own mosque to adorn the entire courtyard, as well as the surrounding arcades and those of the sanctuary, with the complete text of the revelation, which the faithful could theoretically follow in sequence as they progressed through the building, finishing with the text on the qiblah wall, and several sources seem to support that conclusion.
The fifteenth-century Egyptian historian Nur al-Din ‘Ali b. Ahmad al-Samhudi cited the early-ninth-century informants Muhammad b. ‘Umar al-Waqidi (d. 207 / 823) and Ibn Zabalah to the effect that there were inscriptions inside and outside and on the doors of the mosque.
It might also be possible to interpret Ibn Rustah’s report, “‘Umar b. ‘Abd al-‘Aziz is the one who ordered to be written the inscription that is in the mosque and the one who ordered to be written the inscription that is on the qiblah wall of the mosque of the Messenger of God, the blessing of God and peace be on him,” as evidence that there were inscriptions throughout the building.
The expression of political claims through Qur’anic quotations and allusions suggests wide familiarity with these verses and their implications in the early Islamic community, between 72 / 691-92 and 132 / 750. In fact, although Wansbrough has noted, in his argument for a late compilation of the Qur’an, that the text was not used as a basis for legal decisions before the ninth century,
there is abundant evidence from the Umayyad period that it was already sufficiently familiar to the community at large to provide easily recognizable claims to political legitimation and for religious propaganda.
There is additional, more oblique evidence bearing on the issue of the Qur’anic text. The aforementioned inscription in the mosque at al-Madinah provides a starting point. Ibn al-Nadim reported in the late tenth century (before 380/990) that one Khalid b. Abi al-Hayyaj, sahib ‘Ali, had been responsible for executing it.
Khalid was in all probability a younger brother of Hayyaj b. Abi Hayyaj (sic), named in another source as one of those who witnessed the testament of ‘Ali b. Abi Talib in 39/660. Khalid also made copies of the Qur’anic text and other manuscripts for al-Walid and ‘Umar b. ‘Abd al-‘Aziz. It was Sa‘d, a scribe in the employ of al-Walid, who initially recruited him;
in fact, Abu ‘Abd Allah Muhammad Ibn al-Najjar (d. 643/1245) credited the inscription at the Great Mosque of al-Madinah to Sa‘d himself, whom he identified as a mawla of Huwaytib b. ‘Abd al-‘Uzza, a member of Quraysh and a Companion, who died at al-Madinah in 54 / 674, in the caliphate of Mu‘awiya b. Abi Sufyan (41-60/661-80).
Sa‘d is also mentioned in the dictionary of nisbahs compiled by Abu Sa‘id ‘Abd al-Karim b. Abi Bakr al-Sam‘ani (d. after 562 / 1167), where he is identified as sahib al-masahif, from which the nisbah of his own mawla, Ziyad al-masahifi, was taken.
The authority cited was Ibn Abi Hatim (240-327 / 854-938), who in turn cited his father (d. 277 / 890). Ziyad was supposed to have transmitted ahadith to Bukayr b. Mismar al-Zuhri (d. 153/770) in al-Madinah, which is consistent with the chronological position of Sa‘d.
Although none of these reports can be traced back earlier than the mid-ninth century, it is nonetheless possible to pursue the matter farther. To begin with, Huwaytib was a member of the clan of ‘Amir b. Lu’ayy and converted to Islam only after the battle of Hunayn. He was said to be one of sixteen Quraysh who knew how to write in the time of the Prophet.)
He was allied by marriage to a number of important early Muslim figures, and his family connections can be traced through several branches over many generations; despite certain legendary aspects of his biography, it is thus certain that he was a historical personage.
Several anecdotes suggest that Huwaytib was known for his avarice; the most important of them for present purposes is the story that at some indeterminate date he sold his house in Makkah to Mu‘awiya for the enormous sum of 40,000 dirhams and moved to al-Madinah, where he settled “on the Balat near the ashab al-masahif,” a group with which his mawla Sa‘d was linked.
From this report it seems that already in the seventh century there may have been a specific area of al-Madinah where manuscripts of the Qur’an were copied and sold. A large fragment of an early history of the city, by Abu Zayd ‘Umar Ibn Shabbah al-Numayri (173-262 / 789-875), a descendent of a prominent Madinan family, has been preserved.
Although his descriptions of the topography of al-Madinah are not always perfectly lucid, they are invaluable for their detail; of particular concern here is his mapping of the area surrounding the Balat al-A‘zam, the paved street extending west from the Prophet’s mosque to al-Musalla. Among the residences facing onto the north side of the Balat al-A‘zam was one near its western end belonging to Huwaytib.
 Ibn Shahbah did not explicitly mention the ashab al-masahif near whom Huwaytib was reported to have settled, though he did use the term ashab for various occupational groups. Particularly intriguing are the ashab al-rabba‘, whom he located at the eastern end of the Balat al-A‘zam, near the northwestern corner of the Great Mosque. It is not clear what they did;
in fact, they were already problematic in the fifteenth century, when al-Samhudi, who was drawing upon Ibn Shabbah’s text, speculated that they might have been those who made and sold copies of the Qur’an, which were sometimes known as ruba‘.
Even if he was correct, however, it is clear that the ninth-century ashab al-rabba‘ of Ibn Shabbah were not the ashab al-masahif mentioned by al-Tabari, for they were not located near Huwaytib’s house on the Balat.
There is growing evidence that al-Madinah functioned as an Islamic intellectual center in the Umayyad period, before the rise of the cities of Iraq. For example, M. S. Belguedj and Rafael Talmon have presented evidence for the existence of a distinct “school” of grammarians at al-Madinah in the first half of the eighth century, anticipating the emergence of the better-known schools of al-Basrah and al-Kufah;
 Talmon also claims that a number of men in this group earned their livings by copying the Qur’an, but he has documented only one example, Abu Hazim (or Abu Da’ud) ‘Abd al-Rahman b. Hurmuz b. Kaysan al-A‘raj, classified as one of the tabi‘un of al-Madinah, who died at Alexandria in 117/735 or 119/737. He was a mawla of the Banu Hashim and used to copy masahif.
It has been demonstrated that at least three individuals in al-Madinah copied the Qur’an professionally in the last quarter of the first hijri century and the beginning of the second. It thus seems not at all impossible that there was already a concentration of such an occupational group in the city.
In those early years there must have been sufficient demand for the newly codified scriptures, both for public use in mosques and schools and for private study by wealthy or pious patrons, to ensure employment for such a group.
The details cited here are scattered almost at random through texts of different character and period, and the references are too peripheral to the main accounts and the individuals too insignificant to have been part of a conscious,
however pious, forgery of early Islamic history concocted at the end of the eighth century. All point to the active production of copies of the Qur’an from the late seventh century, coinciding with and confirming the inscriptional evidence of the established text itself.
In fact, from the time of Mu‘awiyah through the reign of al-Walid the Umayyad caliphs were actively engaged in codifying every aspect of Muslim religious practice. Mu‘awiyah turned Muhammad’s minbar into a symbol of authority and ordered the construction of maqsurahs in the major congregational mosques. ‘Abd al-Malik made sophisticated use of Qur’anic quotations, on coinage and public monuments, to announce the new Islamic world order.
Al-Walid gave monumental form to the Muslim house of worship and the service conducted in it. It seems beyond the bounds of credibility that such efforts would have preceded interest in codifying the text itself.
The different types of evidence cited here all thus lead to the conclusion that the Muslim tradition is reliable, at least in broad outline, in attributing the first codification of the Qur’anic text to ‘Uthman and his appointed commission. The Qur’an was available to his successors as an instrument to help weld the diverse peoples of the rapidly expanding empire into a relatively unified polity.
It is also possible to speculate that the inscriptions at the Dome of the Rock, so distinct in paleographic style from earlier examples of Arabic writing in any medium, owed something to this background as well. As al-Walid called upon a Qur’an copyist to design his inscriptions at the Great Mosque in al-Madinah, it seems that fifteen or twenty years earlier ‘Abd al-Malik would have had to turn to a similar source.
The only pool of such experienced writers that has left a trace, however faint, in the historical sources, is the ashab al-masahif at al-Madinah. As professional copyists of the Qur’anic text, these men must very early have developed a standard script with its own conventions –
for example, horizontal extensions, hollow rounded letters, the use of strokes for diacriticals on certain letters, and the marking of text divisions with simple ornaments. Where else could ‘Abd al-Malik have found an artist capable of laying out his beautiful inscriptions at the Dome of the Rock?
With the expansion of the empire, the professional copying of the Qur’an also spread from al-Madinah to other cities. In the late Umayyad period, Malik b. Dinar (d. probably before 131 / 748), a mawla of the Banu Najiyyah b. Samah b. Lu’ayy, was said to have supported himself by making copies at al-Basrah. Al-Asbagh b. Zayd al-Warraq al-Juhani (d. 159 / 776), a mawla of Juhaynah, was a bookseller who copied the Qur’an text at Wasit.
Ibn al-Nadim distinguished copyists of masahif from those who copied the Qur’an in scripts like muhaqqaq and mashq. From the former group, Khushnam al-Basri and al-Mahdi al-Kufi copied the Qur’an during the reign of the ‘Abbasid caliph Harun al-Rashid (170-93 / 786-809). Khushnam was reported to have written his alifs one cubit high with a single stroke; although this report is obviously an exaggeration,
it does imply that he impressed by means of a monumental style. Beginning in the reign of al-Mahdi, all the mushaf copyists in Ibn al-Nadim’s list were from al-Kufah, and the implication is that they wrote in a style distinct from cursive. They were Abu Jari (or Hadi or Juday), who was active in the time of al-Mu‘tasim (218-27 / 833-42), Ibn Umm Shayban, al-Mashur, Abu Khamirah, Ibn Khamirah (or Humayrah or Ibn Humayrah), and Abu’l-Faraj, the last “in our own time.”
Of these names the most famous is Ibn Umm Shayban, which belonged to the Malikite qadi of Baghdad, Abu’l-Hasan Muhammad b. Salih al-Hashimi, who died in 369 / 979. He was a descendant of ‘Isa b. Musa, designated by the first ‘Abbasid caliph, al-Saffah (132-36 / 749-54), as heir to al-Mansur (136-58/754-75) but forced by the latter to renounce his succession to the throne and exiled to al-Kufah. Abu’l-Hasan’s family was thus ultimately descended from the Companion of the Prophet ‘Abd al-Muttalib.
In no other extant report is it mentioned that Ibn Umm Shayban copied the Qur’an (though he is said to have recited it in the version of Abu Bakr b. Mujahid, the great reformer of the text), and his social status sets him apart from the earlier known copyists, most of whom appear to have been mawali.
If he was the man whom Ibn al-Nadim had in mind, rather than some other member of the same family, it is possible that he did such work early in his career. Abu’l-Faraj ‘Ubayd Allah b. ‘Umar al-Masahifi died in 401 / 1011, about twenty years after Ibn al-Nadim himself, and thus was probably working just at the time that the latter was compiling his book.
It has not yet been possible to identify the other named mushaf copyists, but it should be noted that the readings of their names are ambiguous. Nor can any of the copyists mentioned or their contemporaries be connected with surviving manuscript fragments. At present there is no convincing evidence for the survival of any Qur’an datable earlier than the ninth century.
All that can be stated with any certainty is that the earliest manuscripts that do survive, though the names of the men who copied them are totally unknown, represent part of a long, evolving tradition rooted in al-Madinah in the seventh century.
 The classic Western study of the history of the text as preserved in Muslim tradition is T. Nöldeke, Geschichte des Qorans, 2nd ed., ed. F. Schwally, vol. 2 (Leipzig, 1919). For useful brief summaries, with references, see W. M. Watt, Bell’s Introduction to the Qur’an (Edinburgh, 1970); A. T. Welch and J. D. Pearson, “al-Kur’an,” EI2, 5: 400-432.
 Nöldeke, 13.
 See Welch and Pearson, 416-19, especially p. 411, referring to Gustav Flugel’s edition of the Qur’anic text.
 A. Mingana, “The Transmission of the Kur’an,” Journal of the Manchester Egyptian & Oriental Society (1915-16): 25-47.
 His bias is apparent in statements like the following: “In considering the question of the transmission of the Kur’an according to Christian writers, the reader will feel that he is more in the domain of historical facts than in that of the precarious Hadith . . .” (Mingana, 34).
 J. Wansbrough, Qur’anic Studies: Sources and Methods of Scriptural Interpretation (Oxford, 1977); cf. J. Chabbi, “Histoire et tradition sacree: La biographie impossible de Mahomet,” Arabica 43.1 (1996): 190-94. In a recent article Y. D. Nevo (“Towards a Prehistory of Islam,” Jerusalem Studies in Arabic and Islam 17 : 108-41) has attempted to confirm the interval suggested by Wansbrough by tracing the gradual evolution of rock-cut inscriptions in the Negev from “basic” (pre-Islamic) to “Muhammadan” to “Muslim” religious texts.
Aside from the fact that these terms are not clearly defined, Professor Donner has noted (personal communication) that Nevo’s argument can be taken equally well to support the traditional view that early Islam and the Qur’anic text evolved primarily in al-Madinah and other urban centers, to which the Negev was entirely peripheral. The author is grateful to Professor Donner for calling this article to her attention.
 Wansbrough, 47; cf. pp. 12, 18-20, 44-45, 49.
 For recent efforts to identify fragments of original texts preserved by later writers, see S. Leder, “The Literary use of the Khabar: A Basic Form of Historical Writing,” in The Byzantine and Early Islamic Near East, I: Problems in the Literary Source Material, ed. A. Cameron and L. I. Conrad (Princeton, 1992), 227-315; and W. al-Qadi, “Early Islamic State Letters: The Question of Authenticity,” in ibid., 215-75.
 Wansbrough, 42, 45.
 Wansbrough, 45.
 Wansbrough, 44.
 See, e.g., Welch and Pearson, 409-11; A. Jeffery and I. Mendelsohn, “The Orthography of the Samarqand Qur’an Codex,” Journal of the American Oriental Society 62 (1942): 175-95; A. Brockett, “Aspects of the Physical Transmission of the Qur’an in the 19th-Century Sudan: Script, Decoration, Binding and Paper,” Manuscripts of the Middle East 2 (1987): 45, 52, nn. 2-3.
 See, e.g., Wansbrough, 45.
 H. C. von Bothmer, “Architekturbilder im Koran: Eine Prachthandschrift der Umayyadenzeit aus dem Yemen,” Pantheon 45 (1987): 4-20.
 For a full exploration of the polemical function of this building, expressed not only through the inscriptions but also through the choice of site and the architectural form, see O. Grabar, “The Umayyad Dome of the Rock in Jerusalem,” Ars Orientalis 3 (1959): 33-62; cf. H. Busse, “Monotheismus und islamische Christologie in der Bauinschrift des Felsendoms in Jerusalem,” Theologische Quartalschrifi 161 (1981): 168-78.
More recently N. Rabat, “The Meaning of the Umayyad Dome of the Rock,” Muqarnas 6 (1989): 12-26, has provided some refinements and modifications to Grabar’s interpretation. M. Rosen-Ayalon has cited references in the inscriptions to angels and to the cycle of Jesus’ birth, death, and resurrection out of context in support of her interpretation of the building as a representation of paradise
(The Early Islamic Monuments of al-Haram al-Sharif: An Iconographic Study [Jerusalem, 1989], 67-68). These references are, however, merely details in clearly antitrinitarian messages that would be unlikely to put the reader in mind of paradise.
 M. van Berchem, “Jérusalem “Haram”, Materiaux pour un Corpus Inscriptionum Arabicarum, pt. 2, Memoires de l’Institut franqais d’archeologie orientale, 154.1-2 (Cairo, 1925-27): 22946, no. 215. Cf. Busse, “Die arabischen Inschriften in und am Felsendom in Jerusalem,” Das Heilige Land 109 (1977): 12-14, cf. 22-23; Grabar, The Shape of the Holy: Early Islamic Jerusalem (Princeton, 1996), 58-61.
 C. Kessler, “‘Abd al-Malik’s Inscription in the Dome of the Rock: A Reconsideration” Journal of the Royal Asiatic Society (1970): 2-64. The sequence of Qur’anic excerpts is garbled in E. C. Dodd and S. Khairallah, The Image of the Word: A Study of Qur’anic Verses in Islamic Architecture (Beirut, 1981), 1: 21-24. For the sequence in which the inscriptions are intended to be read, see S. Blair, “What Is the Date of the Dome of the Rock?” in Bayt al-Maqdis: ‘Abd al-Malik’s Jerusalem, ed. J. Raby and J. Johns (Oxford, 1992), 1: 86-87.
 The basic text presented here is that given by Kessler. The recent publication for the first time of a complete and clearly readable set of photographs (though misidentified and presented in incorrect order) has, however, necessitated a few corrections and alterations in her version; for the photographs, see S. Nuseibeh and O. Grabar, The Dome of the Rock (New York, 1996), 82-105. The translations of the Qur’anic passages are those of M. M. Pickthall, with substitution of “God” for “Allah” and “Book” for “Scripture.”
 This rendering seems more appropriate than “slave,” given by Pickthall.
 Brackets enclose the substitution by al-Ma’mun.
 Kessler, 11.
 Pace Busse, “Inschriften,” 10. One example is an inscribed stone block dated 10 Jumada II 550 / 11 August 1155, set into the north wall of the Great Mosque in the town of Cizre (Jazirat ibn ‘Umar), on the Tigris in southeastern Turkey; for an illustration, see E. Whelan, “The Public Figure: Political Iconography in Medieval Mesopotamia” (Ph.D. diss, Institute of Fine Arts, New York University, 1979), fig. 407. Of the eight lines inscribed on it the first is unreadable. The second introduces the main theme of the inscription, the Day of Judgment.
Lines 6-8 include an invocation of blessing for the anonymous donor and the date. Lines 3-5 contain the following fragments from the Qur’an, run together to express a single, coherent message: “On the Day when every soul will find itself confronted with all that it hath done of good . . .” [the introduction to 3:30] “On the Day when We say unto hell: Art thou filled? and it saith: Can there be more to come?” [50:30 complete]
“On the day when the wrong-doer gnaweth his hands . . .” [introduction to 25:27] “the Day of the approaching (doom), when the hearts will be choking the throats . . .” [excerpts from 40:18]. Professor Annemarie Schimmel very kindly helped in deciphering this inscription. See also A. Welch, “Qur’an and Tomb: The Religious Epigraphs of Two Early Sultanate Tombs in Delhi,” in Indian Epigraphy:
Its Bearing on the History of Art, ed. F. M. Asher and G. S. Gai (New Delhi, 1985), 257-67. Professor Bellamy very kindly supplied the reference to Jalal al-Din ‘Abd al-Rahman al-Suyuti, al-Itqan fi ‘ulum al-Qur’an, 2nd ed., ed. M. A. Ibrahim (n.p. [Cairo?], 1363 / 1984), 1:378-80, a fifteenth-century work in which recitation of the Qur’an out of order and in mixed selection is generally condemned.
 P. Crone and M. Cook, Hagarism: The Making of the Islamic World (Cambridge, 1977), 18; 167, n. 18.
 Crone and Cook, 167, n. 18.
 For a complete publication and analysis of these plaques and their inscriptions, see van Berchem, 247-53, nos. 216-17.
 The end of the tahmid and these two epithets have been interpreted by van Berchem and subsequent scholars as a quotation from 2:255 or the identical passage in 3:11; nevertheless, though most of the “beautiful names” of God can be found in the Qur’an, it seems a mistake to attempt to identify every use of such an epithet as a Qur’anic quotation.
The epithets in this inscription, including the subsequent series of three in construct with “the heavens and the earth,” can, like the tahmid, be considered to have had an independent existence and need not be regarded as quotations wherever they occur.
 This extract follows two epithets, “the One” and “the eternally Besought of all,” that also occur, in declarative sentences, in 112:1-2 but need not be considered a “deviation” from the Qur’anic text; it is clear, however, that their inclusion in the series of “beautiful names” was meant to introduce the Qur’anic passage.
 Up to this point the inscription exactly duplicates that on the eastern portal.
 Perhaps from 6:163, though again it seems unnecessary to seek a Qur’anic origin for such a standard phrase.
 In these passages of the Qur’an the words are those of Muhammad, expressed in the first-person plural; in this inscription ‘Abd al-Malik speaks for the community of believers, and Muhammad is thus referred to in the third-person singular.
 For a similar idea, developed in a different direction, see H. Edwards, “Text, Context, Architext: The Qur’an as Architectural Inscription,” in Brocade of the Pen: The Art of Islamic Writing, ed. C. G. Fisher (East Lansing, Mich., 1991), 67-68, 69.
 Abu Ja‘far Muhammad al-Tabari, Ta’rikh al-rusul wa’l-muluk, Leiden ed., 2: 1368; Cairo ed., 6: 570.
 Identified by D. S. Powers, in The Empire in Transition: The Caliphates of Sulayman, ‘Umar, and Yazid, A.D. 715-724 / A.H. 97-105, The History of al-Tabari: An Annotated Translation, vol. 24 (Albany, 1989), 98, n. 347.
 Al-Tabari, Leiden ed., 3: 861-62; Cairo ed., 8: 439.
 These passages have been identified by M. Fishbein, The War between Brothers: The Caliphate of Muhammad al-Amin, A.D. 809-813 / A.H. 193-198, The History of al-Tabari: An Annotated Translation, vol. 31 (Albany, 1992), 126, nn. 477-79.
 For parallel evidence of adaptation of familiar Qur’anic passages in early Arabic literature, see W. al-Qadi, “The Limitations of Qur’anic Usage in Early Arabic Poetry: The Example of a Kharijite Poem,” in Festschrift Ewald Wagner zum 65. Geburtstag, vol. 2: Studien zur arabischen Dichtung, ed. W. Heinrichs and G. Schoeler (Beirut, 1994), 162-81 (p. 179: “. . . early Arabic poetry, like its counterpart Arabic prose . . . tends to reformulate Qur’anic materials more than to quote them literally”); idem,
“The Impact of the Qur’an on the Arabic Epistolography of ‘Abd al-Hamid,” in Approaches to the Qur’an, ed. G. R. Hawting and A. Shareef (London, 1993), 205-313 (p. 307: “. . . no one could be a master at drawing from the Qur’an in the manner that ‘Abd al-Hamid is without having full control . . . of the text of the Qur’an . . . he could appeal to what is familiar to his audience”). Professor van Ess kindly supplied references to al-Qadi’s work.
 The essence of this reform was the adoption of purely epigraphic coinage without imagery.
 Ibn Rustah, Kitab al-a‘laq al-nafisah, ed. M. J. de Goeje (Leiden, 1892; repr. Leiden, 1967), 70; cf. J. Sauvaget, La Mosquée omeyyade de Médine (Paris, 1947), 79.
 For a reconstructed plan of the mosque, see Sauvaget, 91.
 Cf. Ibn al-Nadim, Kitab al-Fihrist, ed. R. Tajaddud (Tehran, 1871), 6; Abu ‘Abd Allah Muhammad Ibn al-Najjar (578-643 / 1183-1245), Kitab al-durrah al-thaminah fi akhbar al-Madinah, Bibliotheque nationale, Paris, MS Ar. 1630, fol. 32a. Ibn al-Najjar apparently wrote his book during a stay in al-Madinah, probably relying on manuscripts in local collections and his own observations of the mosque (C. E. Farah, “Ibn al-Najjar: A Neglected Arabic Historian,” JAOS 84 : 222, 223, 226-27). His sources included Ibn Zabalah (d. 199/814), Abu’l-Qasim al-Muzaffari, and al-Ajzi. The identities of the latter two are uncertain (Sauvaget, 26).
 Cited without attribution by Abu Ahmad Ibn ‘Abd Rabbihi (246-328/860-940), Kitab al-‘iqd al-farid, ed. A. Amin, I. al-Abyari, and ‘A. Harun (Cairo, 1368/1949), 6: 261; cf. M. Shaft, “A Description of the Two Sanctuaries of Islam by Ibn ‘Abd Rabbihi (940),” in A Volume of Oriental Studies Presented to Edward G. Browne on His 60th Birthday, ed. T. W. Arnold and R. A. Nicholson (Cambridge, 1922), 420-21.
There is no evidence that Ibn ‘Abd Rabbihi ever left Spain; cf. W. Werkmeister, Quellenuntersuchungen zum Kitab al- Iqd al-farid des Andalusiers Ibn ‘Abdrabbih: Ein Beitrag zur arabischen Literaturgeschichte (Berlin, 1983), 22-23. It is clear from the details of this report and references to the Great Mosque at Cordoba that the informant who traveled to al-Madinah was also familiar with monuments in Spain. The observer described the Black Stone of the Ka‘bah, which the Qarmatians removed in 317/929, providing a terminus ante quem for the visit (Shaft, 422).
 Al-Tabari, Leiden ed., 3: 535, Cairo ed., 8: 178; tr. H. Kennedy, Al-Mansur and al-Mahdi: A.D. 763-786 / A.H. 146-169, The History of al-Tabari: An Annotated Translation, vol. 29 (Albany, 1990): 254. The report was attributed to a descendant of ‘Ali’s brother Ja‘far in the line of al-Mahdi.
 The earliest source for this story appears to have been Abu Muhammad al-Fadl b. Shadhan (d. 260 / 874), but even by his time the actual facts about Ibn Mas‘ud’s version had become blurred; see A. Jeffery, Materials for the History of the Text of the Qur’an: The Old Codices (Leiden, 1937), 21.
 Ibn Rustah, 70. For a reconstruction of the inscriptions reported by Ibn Rustah, see Sauvaget, 78-80; on the Kharijite rebellion, see C. Pellat, “al-Mukhtar b. ‘Awf al-Azdi,” EI2, 7: 524-25. Ibn Rustah’s date of 128 / 745 for the restoration of the mosaics by Ibn ‘Atiyyah appears to be incorrect, however. According to al-Tabari, the Kharijites entered al-Madinah in 130 / 747; ‘Abd al-Malik b. Muhammad b. ‘Atiyyah retook the city a short time later and appointed his nephew al-Walid b. ‘Urwah b. Muhammad b. ‘Atiyyah as deputy governor over the city (ed. Leiden, 2: 2008, 2014; 3: 11; ed. Cairo, 7: 394, 399, 410-11).
 Al-Samhudi, Wafa’ al-wafa bi akhbar Dar al-Mustafa, ed. M. M. ‘Abd al-Hamid (Cairo, 1374 / 1955), 1:371.
 From an extensive analysis of exegesis on surat Quraysh, Patricia Crone has concluded that “the exegetes had no better knowledge of what this sura meant than we have today. . . . What they are offering is . . . so many guesses based on the verses themselves. The original meaning of these verses was unknown to them or else there had been a gradual drift away from it.
In any case, it was lost to the tradition. . . .” Meccan Trade and the Rise of Islam (Oxford, 1987), 210. Her observations suggest a substantial interval between establishment of the Qur’anic text and the development of exegetical tradition at the end of the eighth century. They thus contradict Wansbrough’s notion that codification of the text and the introduction of exegesis occurred at approximately the same time.
 Ibn al-Nadim, 9; cf. Sauvaget, 79-80, where this man is referred to as Khalid b. Abi al-Sayyaj without further explanation.
 Abu Zayd ‘Umar Ibn Shabbah al-Nuwayri, Ta’rikh al-Madinah al-munawwarah (Akhbar al-Madinah al-nabawiyyah), ed. F. M. Shaltult (Beirut, 1410 / 1990), 1: 225-28. Ibn Shabbah’s source for the “testament of ‘Ali,” which he reproduced, was Abu Ghassan Muhammad b. Yahya, who claimed to have the document in his possession, having received it from his father, a scribe, who had, in turn, received it from al-Hasan b. Zayd (d. 167 / 783), a great-grandson of ‘Ali; according to another source cited by Ibn Shabbah, it was Abu Hayyaj himself who witnessed the testament.
The document cannot be assumed to have been genuine, but internal evidence suggests that, if it was a forgery, it was a forgery of the Umayyad period or the first twenty years of the ‘Abbasid period. For example, the testator called himself only ‘Abd Allah ‘Ali Amir al-Mu’minin.
The ‘Abbasids adopted regnal names, though there is at least one instance in which al-Mansur called himself ‘Abd Allah ‘Abd Allah Amir al-Mu’minin; see al-Tabari, Leiden ed., 3: 208; Cairo ed., 7: 566. (Thanks are owing to Dr. Bates for this reference and his views on this point.) Ibn Shabbah himself complained of errors in the language and spoke of having copied the “letter forms” exactly as he saw them, implying that the document already seemed archaic in the early ninth century.
For a summary of Abu Ghassan’s background and career, see Ibn Hajar al-‘Asqalani, Tahdhib al-tahdhib (Hyderabad, 1326; repr. Beirut, 1968), 9: 517-18, no. 846; cf. T. Nagel, “Ein fruher Bericht über den Aufstand von Muhammad b. ‘Abdallah im Jahre 145 H,” Der Islam 46 (1970): 236-38.
 Ibn al-Nadim, 9. N. Abbott (The Rise of the North Arabic Script and Its Kur’anic Development, with a Full Description of the Kur’an Manuscripts in the Oriental Institute [Chicago, 1939], 54, n. 83) had some reservations about Khalid, noting that he “must have been a very young companion of ‘Ali and an elderly scribe of al-Walid” and calling attention to the omission of his name from the list of al-Walid’s scribes assembled by W. Bjorkman (Beiträge zur Geschichte der Staatskanzlei im islamischen Ägypten [Hamburg, 1928], 57-58). Bjorkman listed only al-Walid’s state secretaries in Damascus, however, whereas Khalid was working in al-Madinah.
Furthermore, it should be noted that sahib means not only “companion” but also “disciple, follower” and that the context is Ibn al-Nadim’s discussion of books collected by a Shi‘ite bibliophile of his own time, implying that he identified Khalid as a Shi‘ite. It seems from Ibn Shabbah’s report about ‘Ali’s will that Khalid’s family was close to ‘Ali.
 Ibn al-Najjar, fol. 32a.
 Al-Sam‘ani, Kitab al-ansab, facs. ed., ed. D. S. Margouliouth (London, 1912), fol. 531b, s.v. al-masahifi. On Bukayr, see Ibn Hajar, 1: 495, no. 914; idem, Kitab lisan al-mizan (Hyderabad, 1330 / 1912), 2: 62, no. 236. Y. Eche, Les Bibliotheques arabes publiques et semi-publiques en Mésopotamie, en Syrie et en Égypte au Moyen Age (Damascus, 1967), 18, has interpreted the term sahib al-masahif as “librarian” and has identified Sa‘d as al-Walid’s librarian in Damascus.
It is clear from the context of all these reports, however, that Sa‘d lived in al-Madinah and that he was not a librarian but one who copied masahif; cf. especially al-Sam‘ani, fol. 120a, s.v. al-jami‘i: “Perhaps it is the nisbah related to the collection, that is, the mushaf. The most famous [person] with this nisbah is Abu Habib Muhammad b. Ahmad b. Musa al-Jami‘i al-Masahifi, who used to copy the jami‘.”
 Ibn Shabbah, cited in Ibn ‘Abd Rabbihi (Cairo, 1363 / 1944), 4: 157-58. The specific source may have been the lost Kitab Makkah.
 He was one of two full brothers, the other being Abu Ruhm, who was married to Barrah bt. ‘Abd al-Muttalib, a paternal aunt of the Prophet; another of Abu Ruhm’s wives was Maymunah bt. al-Harith, who married the Prophet after Abu Ruhm’s death. Huwaytib’s sister was married to Sufyaan (or Aswad) b. ‘Abd al-Asad.
There were also two half-brothers, Makhramah and Abu Sabrah (perhaps, rather, a nephew). A descendant of Makhramah in the sixth generation, Sa‘d or Sa‘id, served as chief qadi of al-Madinah in the reign of al-Mahdi; his son ‘Abd al-Jabbar subsequently served as governor and then as qadi of al-Madinah in the time of al-Ma’mun. Abu Sabrah served briefly as governor of al-Basrah in 17 / 638-39 and was commander-in-chief of the army that invaded Khuzistan in that year.
His son Muhammad was chief qadi of al-Madinah, as was Muhammad’s grandson Muhammad b. ‘Abd Allah. The latter’s half-brother Abu Bakr b. ‘Abd Allah supported the revolt of Muhammad b. ‘Abd Allah b. al-Hasan, “the Pure Soul,” in 145/762. He was imprisoned in al-Madinah by ‘Îsa b. Musa but released on the orders of al-Mansur.
Huwaytib himself was married to Âminah (or Aminah or Umaymah) bt. Abi Sufyan b. Harb, daughter of the supreme commander of the Meccan forces against the Prophet. Âminah was thus a half-sister of the caliphs Mu‘awiyah and Yazid b. Abi Sufyan (60-64 / 680-83). She bore Huwaytib a son, Abu Sufyan, but was subsequently divorced. Abu Sufyan’s grandson Abu Bakr b. ‘Abd al-Rahman was chief qadi of al-Madinah in the time of the caliph Hisham (105-25 / 724-43).
Abu Bakr’s grandson Muhammad b. ‘Abd al-Rahman was killed at Nahr Abi Futrus in Palestine in 132/748-49. Finally, this Muhammad’s own grandson Muhammad b. ‘Abd al-Karim transmitted ahadith at Harran in northern Syria.
Mus‘ab b. ‘Abd Allah al-Zubayri, Nasab Quraysh wa’l ‘asabah, ed. E. Levi-Provençal (Cairo, 1953), 426-30; al-Tabari, Leiden ed., 1: 1184, 1773, 2498, 2549-50, 2552-53, 2556-57, 2564-67; 3: 2326-29, 2453-53; Cairo ed. 2: 331; 3: 166; 4: 50, 81-84, 86, 91-93; 11: 517-19, 611; Ahmad b. Yahya al-Baladhuri, Ansab al-ashraf (ed. M. Hamid Allah, Beirut, 1400 / 1979), 1: 220, 228, 292, 312, 349, 350, 352, 362, 363,404, 441, 444-46; (ed. I. ‘Abbas, Wiesbaden, 1979), 4.1:6 and n. 2; ‘Abd al-Rahman b. Safwan Ibn Sa‘d, al-Tabaqat al-kubra (ed. E. Sachau, Leiden, 1904), 3: 293-94; 8: 174, 192-93; Abu Muhammad ‘Ali Ibn Hazm, Jamharat ansab al-‘Arab (Beirut, 1403 / 1983), 1: 169. For Huwaytib’s other offspring, see al-Zubayri, 430; Ibn Sa‘d, 5: 128-29, 335-36. Dr. Ella Landau-Tasseron kindly provided additional references to Huwaytib.
 For example, he was one of those said to have lived sixty years in the Jahiliyyah and sixty years under Islam. Al-Tabari, Leiden ed., 3: 2326-29; Cairo ed., 11: 517-19; cf. Ibn Sa‘d, 5: 335.
 Al-Tabari, Leiden ed., 3: 2329; Cairo ed., 11: 518-19. This story, which was also reported by al-Ya‘qubi (d. 292 / 905), can be traced to al-Waqidi; cf. W. G. Millward, “The Adaptation of Men to Their Time: An Historical Essay by al-Ya‘qubi,” JAOS 84 (1964): 330, 336, where, according to the translation, Huwaytib bought, rather than sold, the house. Cf. Ibn Hazm, 1: 168-69.
The detail about the ashab al-masahif does not seem to have been preserved by al-Ya‘qubi, however; al-Samhudi 2: 746, cited it from the Tabaqat of Ibn Sa‘d. There were apparently at least four recensions of Ibn Sa‘d’s text, the latest of which, that by Ibn Hayyawayh (d. 381/991), was used by Sachau for his edition; that by al-Harith b. Abi Usama (d. 282 / 895) was used by al-Tabari (J. W. Fück, “Ibn Sa‘d,” [EI.sup.2], 3: 922). As the reference to the ashab al-masahif is not given in Sachau’s edition, al-Samhudi must have been quoting it from one of the other recensions.
 Ibn Shabbah, 4 vols. (Beirut, 1410 / 1990). It seems to have been composed in al-Basrah, but there is little doubt that the author was intimately familiar with al-Madinah. Large segments of the text were reproduced by al-Samhudi, but it is only recently that the original has been published, apparently from a copy in the hand of the fourteenth-century author Ibn Hajar al-‘Asqalani (Ibn Shahbah, I: nun-alif). Certain details, including similar variations and errors in spelling, suggest that al-Samhudi worked from this same manuscript.
 It was located between the compounds of al-Rabi‘, mawla of the Commander of the Faithful (al-Mansur, 136-58 / 759-75) on the west (an anachronism of a kind not uncommon in Ibn Shabbah’s text) and of ‘Âmir b. Abi Waqqas on the east. It was separated from the former by a lane that led to the house of Âminah, daughter of Abu Sarh; the context suggests that Âminah’s house may have stood to the north of Huwaytib’s compound.
Across the Balat to the south was the quarter of the Banu Zurayq, a tribal group originally from Yemen; Huwaytib’s compound apparently faced Dar Hafsah, owned successively in his lifetime by ‘Uthman b. Abi al-‘Âs, Mu‘awiyah, and the latter’s mawla Hafsah, and the compound of Abu Hurayrah (Yaqut b. ‘Abd Allah al-Hamawi, Mu‘jam al-buldan, ed. F. Wustenfeld as Jacut’s Geographisches Wörterbuch , 1: 245-46, 25152; Ibn Shabbah, 1: 240-41,252, 255-56).
Huwaytib owned two other houses in al-Madinah, one of them in the quarter of the Banu Zurayq well away from the Balat, the other, known as Dar Subh, situated between the house of al-Muttalib and the square before the Majlis al-Hukm (Ibn Shahbah, 1: 252-53). The precise location of this third house has not yet been established, but it seems not to have been on the Balat.
 Ibn Shabbah, 1: 231.
 Al-Samhudi, 2: 745-46.
 M. S. Belguedj, “La démarche des premiers grammaires arabes dans le domaine de la syntaxe,” Arabica 20 (1973): 168-85; R. Talmon, “An Eighth-Century Grammatical School in Medina: The Collection and Evaluation of the Available Material,” Bulletin of the School of Oriental and African Studies 48 (1985): 225, 228.
 Muhammad Ibn Hibban al-Busti [first half of the tenth century], Kitab mashahir ‘ulama’ al-amsar, ed. M. Fleischhammer (Wiesbaden, 1959), 77, no. 559; al-Sam‘ani, fol. 44b, s.v. al-a‘raj. Ibn Hibban’s source was again the ninth-century biographer Abu Hatim. Talmon erroneously cites Belguedj, 172-73, as the source for his larger conclusion that several grammarians concerned themselves with “Qur’anic scripts.”
 According to two reports from Malik b. Anas, on the authority of Zayd b. Aslam (d. 136 / 753), ‘Amr b. Rafi‘ and Abu Yunus copied the mushaf for the Prophet’s wives Hafsah bt. ‘Umar b. al-Khattab (d. 45/665) and ‘Â’ishah bt. Abi Bakr al-Siddiq (d. 58 / 678), respectively; Muwattah al-Imam Malik (Cairo, 1386 / 1967), 2: 344, nos. 999-1000. Ibn Sa‘d reported that ‘Amr was the son of a mawla of ‘Umar b. al-Khattab and that Hafsah herself was the source of the story about the mushaf (5: 220); Abu Yunus was ‘Â’ishah’s own mawla, but Ibn Sa‘d does not mention his having copied a mushaf for her (5:218).
 E. Whelan, “The Origins of the Mihab Mujawwaf: A Reinterpretation,” International Journal of Middle East Studies 18.2 (1986): 205-24.
 See, e.g., Abbott, pls. II-V.
 For more elaborated versions of these conventions, see Whelan, “Writing the Word of God: Some Early Qur’an Copyists and Their Milieux,” Part I, Ars Orientalis 20 (1990): 113-47.
 Cf. R. Blachere, Introduction au Coran, 2nd ed. (Paris, 1959), 88. Recently Rabbat, “The Dome of the Rock Revisited: Some Remarks on al-Wasiti’s Account,” Muqarnas 10 (1993): 70-71, has suggested that the designer of the inscriptions may have been one of two men charged with supervising work on the Dome of the Rock, according to an eleventh-century report by Ahmad al-Wasiti in Fada’il al-bayt al-muqaddas (Jerusalem, 1979): 80-81. He was Raja’ b. Haywah, a prominent figure in the employ of several Umayyad caliphs, who was at the beginning of his career in the reign of ‘Abd al-Malik. Although Raja’ may have functioned as a secretary under the caliphs Sulayman (96-99 / 715-17) and ‘Umar II (99-101 / 717-20; according to C. E. Bosworth, “Raja’ ibn Haywa al-Kindi and the Umayyad Caliphs,” Islamic Quarterly 16 : 43 and n. 5, the sources vary), there is no evidence that he was ever a copyist, adhering to a specific set of stylizations of the sort visible at the Dome of the Rock, or that a group of such copyists flourished in Palestine in the time of ‘Abd al-Malik.
 Abu Nu‘aym Ahmad b. ‘Abd Allah al-Isbahani, Hilyat al-awliya’ wa’l-tabaqat al-asfiya’ (Cairo, 1351/1932), 1: 357-89; Ibn Sa‘d, 7.2:11.
 Al-Sam‘ani, fol. 579r, s.v. al-warraq; cf. Ibn Sa‘d, 7.2: 61.
 Ibn al-Nadim, 9-10.
 Ibn al-Nadim, 9-10; for the various cubit measures in use in early Islam, all of them rather large for the present context, see W. Hinz, Islamische Masse und Gewichte umgerechnet ins metrische System (Leiden, 1970), 55-62. The smallest was 49.875 cm.
 Al-Khatib al-Baghdadi, Ta’rikh Baghdad (Beirut, n.d.), 5: 363-64, no. 2889, cf. 362, no. 2888; Ibn Hazm, 1: 32; Abu ‘Umar Muhammad al-Kindi, Kitab al-wulah wa’l-qudah, ed. R. Guest as The Governors and Judges of Egypt (Leiden and London, 1912), 573.
 Al-Khatib, 10:380 no. 5548.