Version classiqueVersion mobile
OpenEdition Books

The Voice and the Quill. The Paths of Canonization of the Quʾrān

François Déroche

The Voice and the Quill. The Paths of Canonization of the Quʾrān

Inaugural Lecture delivered on Thursday 2 April 2015

François Déroche
Traduction de Liz Libbrecht

Texte intégral

1Mr Administrator,
Dear Colleagues and Friends,
Ladies and Gentlemen,

2The greatness of institutions such as the Collège de France stems from their ability to anticipate and innovate. Some events and the emotions they arouse can for a moment muddle our perception of the facts. History has it recorded that the Collège de France Faculty made the decision to elect me to the Chair of the History of the Qurʾān nearly a year ago, far from the noise and fury. For the first time in the life of this institution, the Qurʾān will feature in its programme, thus joining the Bible that Thomas Römer’s Chair introduced eight years ago. Allow me, dear colleagues, to salute your audacity and intellectual courage, you who have chosen to put at the heart of the Collège de France’s teaching two books that each represent a fundamental spiritual reference for millions of people, and who, in my case, agreed to interrupt a long tradition of titles relating to the Arabic language.

3The interruption, though, is only apparent, for the Qurʾān has never been far from the teaching of Arabic in general and from the Collège de France’s teaching in particular. Even if, from a strictly chronological point of view, the Qurʾān is not the first Arabic text, its place in history and literature confers on it absolute primacy. And the Arabic scholars who preceded me here rarely failed to refer to it, if only for the obvious linguistic implications of the language in which it was written and that served as a base on which to establish what we call Classical or Literary Arabic. It has indeed been discreetly present within our walls for nearly five centuries. We could of course question whether or not Guillaume Postel was the first to teach Arabic at the newly founded Collège royal. As we know, he was entrusted with the teaching of Oriental languages other than Hebrew, from 1538 to 1543. I like to see him as the first in a long lineage of professors of Arabic, and to recognize as a decision of major symbolic import his choice to publish, in his Grammatica Arabica, the first Surah of the Qurʾān, the “Introduction”, the one that literally “opens” the sacred book of Islam – even though the characters used by his printer admittedly bear little relation to the Arabic alphabet and differ significantly from the calligraphy to which I have become accustomed through my regular work with manuscripts.

4Thanks to this work, I have been frequenting for a long time, might I say, many of those who in a sense preceded me (and you will see why I am choosing my words with care). Until the nineteenth century it was not unusual for Arabic scholars to have a manuscript copy of the Qurʾān, and some of those owned by Collège royal professors are now in the Library that was first Royal, then for some time Imperial, and is now National. Several of these masters served moreover in the Manuscripts Department. The copies that were thus preserved illustrate in their own way those times when the Alqur’ān, as it was called, could not lie in any other form on an Orientalist’s desk. They served as my guides up to the eighteenth century, when Western typography facilitated access to texts at a turning point in the nature of Islamic studies.

5Like his predecessor Arnoult de l’Isle at the Collège royal, Étienne Hubert was a doctor and Arabic scholar in the days when Avicenne or Rhazès were sometimes still read in the original. Hubert, an Arabic teacher from 1600 to 1614, had an unusual Qurʾān, to say the least, even though its appearance was hardly promising (Fig. 1). The manuscript had been copied by a certain François de Boulogne, a “Turk that [he] had taken out of the King’s guard”, as a note on one of the folios indicates. The most prominent figures at the Collège royal in the seventeenth century were then the Levantines, Gabriel Sionita, Abraham Ecchelensis, and Pierre Dipy, who were more interested in grammar than in the Qurʾān. The latter two nevertheless consistently fulfilled their functions as librarians. We owe them several hand-written notes preserved in copies of the Qurʾān, which entered the royal collections in the seventeenth century via the Cardinal Mazarin collection, or other routes.

Figure 1

Figure 1

Copy of excerpts from the Qurʾān, Paris (?), late sixteenth/early seventeenth century.

Bibliothèque nationale de France, Arabic manuscript 460, f. 1 v°-2 r°. © BNF.

6The situation changed in the following century. The name of Antoine Galland, professor at the Collège royal from 1709 to 1715, has become permanently associated with the translation of Thousand and One Nights. He was also an active and sagacious manuscript collector. From his stay in Constantinople he brought back two manuscripts of the Qurʾān now kept at the Bibliothèque nationale de France: a moving copy of excerpts from the sacred texts, along with a Spanish translation of it, both copied in 1569 in Thessaloniki by a Morisco who had fled Philippe II’s reign in Spain and found refuge in Ottoman territory, and an unabridged version of the Qurʾān. Galland probably used the latter for his French translation of the Qurʾān which remained unpublished, unlike the famous Tales.

7Étienne Fourmont, who succeeded Galland at the Collège royal, held the Chair of Arabic until 1745. This polyglot scholar had a copy of the Qurʾān in his library; more precisely he had one volume from a series of thirty. Fourmont is of particular interest to me because of a piece that is also at the Bibliothèque nationale (Fig. 2). When Louis de Bréhan, Count of Plélo, was French ambassador in Copenhagen, he sent him facsimiles of the folios of two Qurʾānic manuscripts brought from Egypt in the seventeenth century by a Danish trader. Fourmont, who was also Keeper of Manuscripts in the King’s library, had difficulty finding a satisfactory explanation for his correspondent as to what it was about. With his confused answer – that the Arabic scholar Johann Georg Christian Adler was later ferociously to deride – a new chapter in Qurʾānic studies was timidly opened, which was directly linked to what concerns us today.

Figure 2

Figure 2

Facsimile of a folio of the Qurʾān conserved in Copenhagen, from around 1730.

Bibliothèque nationale de France, Arabic manuscript 580 a, f. 3 v°. © BNF.

8The late eighteenth and early nineteenth century saw a change in the approach to the Qurʾān – and a break in the main thread that I have been following, as the Qurʾānic manuscript on Arabic scholars’ desks was replaced by Abraham Hinckelmann’s edition, before that of Gustave Flügel. The teaching of Arabic at the Collège navigated with serene continuity the period of upheaval triggered by the end of the Ancien Régime. The Caussin de Percevals, both father and son, seamlessly covered almost a century of the history of the Chair of Arabic at the Collège de France, from 1783 to 1871. The one of interest to me, the son, Armand Pierre, was even born at the Collège, if his biographies are to be believed. While he held his Chair, he told his audiences: “This is where I was born!” His interest in the pre-Islamic period and the beginnings of Islam are apparent in the book that established his reputation, Essai sur l’histoire des Arabes, in which Muhammad’s biography had a prominent place. He concluded his Essai with a mention of the Prophet, his apostolate and the first caliphs. Accounts of Muslim tradition featured conspicuously in it, especially everything regarding the Qurʾān.

9His successors were to show less direct interest in Islam’s holy book. Paul Casanova, however, Professor at the Collège de France from 1909 to 1926, was one exception. He published Mahomet et la fin du monde (1911), which departed in its approach and offered an original reading of Muhammad’s message. Casanova’s theses triggered strong reactions at the time but paved the way for new research.

10Talking about the Qurʾān in this place is therefore not new. The work of these professors of Arabic who preceded me at the Collège de France inspires my respect, and makes me aware of my own shortcomings. As I present myself before you as a historian, my dear colleagues, I nevertheless feel confident in my ability to offer a renewed approach to the founding text of Islam. In preparing my project and then submitting it to you, I was fortunate to find comfort in the friendly words and kindly wisdom of Nicolas Grimal, who defended my candidature to your Faculty, and to whom I am deeply grateful.

11I am equally grateful to Thomas Römer for his support from the outset. It was all the more valuable insofar as our respective work will inevitably lead us in coming years to prolong an ongoing dialogue started when, after his election to the Collège de France, he was appointed to head the CNRS team to which I belong. In this team I am also fortunate to have as colleagues Jean-Marie Durand and Dominique Charpin, whose advice has been so valuable.

12Writing the history of Islam in its infancy, and even more so that of the Qurʾān, raises the question of Arabic sources on this period. Historians have long wondered about the value of accounts by subsequent authors to the period. The Qurʾān is no exception. The Muslim tradition has preserved various accounts of the writing of the text, mostly compiled in the ninth century, but it gradually retained only one: that in which the caliph Uthmān appears as a protagonist. Western historiography has taken up the main lines of this description, which we find in Barthélémy d’Herbelot’s seventeenth-century Bibliothèque orientale. Former debates on the reliability of these sources erupted anew in the 1970s when the “hyper-critical” school radically challenged the use of information passed down by traditional literature.

13Against this extreme position, it was argued that a number of material relics allowed scholars to define the historicity of Islam from a relatively early time. What were these relics? A few buildings, some spectacular, such as the Dôme du Rocher; objects found at excavation sites; coins and inscriptions; and papyri: actually very little. Everything indicates that this relative dearth led specialists of that period to follow the slightest lead so as to obtain the most abundant documentation possible. Yet a body of documents that was accessible without having to undertake long and hazardous journeys, but simply by visiting reading rooms of the great European libraries, had been waiting since the nineteenth century for scholars to examine them more closely. I am referring to the ancient manuscripts of the Qurʾān, which could reveal an essential aspect of ancient Islamic civilization and enable us to study the history of its founding text on a new basis.

14It is an understatement to say that these ancient copies carried little weight in Western historiography. They had such a narrow readership that, in a book published in 1977, John Wansbrough defended the hypothesis that the Qurʾān had been written in the eighth, even in the early-ninth century and not, as Muslim tradition maintained, in the mid-seventh century! We owe to Islam’s great respect for the language and writing that bore the Revelation our immense fortune today, to have numerous fragments of these first Qurʾāns, copied between the seventh and tenth centuries. We are only starting finally to take this valuable testimony into consideration. For many years, men’s piousness led to the accumulation of these membra disjecta in unused rooms of ancient mosques, like a sleeping beauty in the shadows of forgotten reserves (Fig. 3). The discovery by scholars that they can serve as a documentary source coincided with Bonaparte’s Egyptian expedition, but locally their memory was not entirely erased. Crossing paths with Biblical scholars seeking ancient manuscripts from the Old and New Testaments, some nineteenth-century pioneers intuited that they were important pieces for illuminating the history of the early transmission of the Qurʾānic text. Without always having identified their aim – palaeography or history of art, textual critique or bibliophily –, Jean-Joseph Marcel, Jean-Louis Asselin de Cherville, Ulrich Seetzen, Johann Gottfried Wetzstein, Edward Palmer, and Charles Tyrwhitt Drake collected parchment folios in Fustat and in Damascus, which are now kept in Saint-Petersburg, Paris, Gotha, Berlin, Tübingen, and Cambridge. In the intellectual world, Ernest Renan defended the project of a critical history of the Qurʾān that would incorporate this information. On this point, his appeal failed to have the impact he had hoped for. Although in Egypt and the Ottoman Empire efforts to safeguard these manuscripts made it possible to conserve a variable portion of fragments in Cairo and in Istanbul, and a new find in Kairouan substantially increased the amount of available data, their research came to nothing. The reason lies not in the state of these manuscripts, transformed into puzzles whose dispersed pieces were waiting to be gathered together again, but in a problem of method and in chance circumstances. Various causes – the Battle of Solferino, an accident in the mountains or a huge lie – delayed the recognition of these collections’ value. Having started to study their testimony at a time when many people thought I was wasting my time, I now measure the change of attitude that has taken place. I am almost worried to see the media’s agitation that sometimes surrounds the carbon 14 dating of one fragment or another.

Figure 3

Figure 3

Qurʾānic fragments in the collection of the Great Mosque of Damascus.

Photo: François Déroche.

15It is true that, as regards method, the difficulty of dating manuscripts undoubtedly impeded research. Yet we cannot underestimate the impact that the Muslim tradition’s apparent depreciation of written transmission has had on scholars of Islam. Whereas in (written) French a subtle change of the capital letter into a small letter enables us to distinguish the Qurʾānic message from the material supports conveying it, the Arabic language clearly distinguishes between the former, Qur’an, and the latter that are denoted by a specific term, most often muṣḥaf (in the singular). We see here the origin of the relative indifference that I pointed out, and the value attributed to oral transmission. In practice, the situations are more nuanced; but we clearly recognize this attitude in the glaring absence of ancient and modern copies of the Qurʾān from the list of sources used by the al-Azhar scholars when they worked on the edition which was gradually imposed both on Muslim communities and on scientists studying the Qurʾān from 1923. In Western Qurʾānic studies, this perception has been reinforced by the results of ad hoc surveys, such as, for example the mid-nineteenth-century ones of Theodor Nöldeke, which seemed to confirm that little could be expected from written transmission. In fact, until recently, copies of the sacred Islamic text interested only art historians – and even so they often only ever examined the first illuminated pages (Fig. 4).

Figure 4

Figure 4

Full-page embellishment of a Qurʾānic manuscript in thirty volumes, Near East, latter half of the ninth century. Private collection.

Photo: François Déroche.

16When the workers restoring the roof of the Great Mosque of Sana’a discovered a fourth cache of discarded Qurʾānic manuscripts in 1973, they had no idea that they were going to help to change completely the approach to the Qurʾān. Among the thousands of fragments found in the space between the ceiling of the prayer room and the roof, there were forty sheets of a palimpsest on which the oldest layer of writing was a text differing from the vulgate. This discovery, which was widely covered by the media, played a major part – for better or for worse – in raising awareness on the importance of an approach that takes written transmission into account.

17The existence of testimony to another Qurʾān will not enable us to unravel the entanglement of traditions concerning written records. It does however confirm what the writings suggest: towards the mid-seventh century, several versions were circulating simultaneously, competing with that of the caliph Uthmān. Some were maintained at least until the early tenth century, when what we could call the “second canonization” of the Uthmānnian text took place under the authority of Ibn Mujahid, with the support of the Abbasidi caliphate. Our good fortune lies in the survival of many direct testimonies of the period between these two points in time. On these bases, we can follow the history of the Qurʾān-book precisely, by comparing the data of the tradition with the manuscripts themselves. It started when the herald’s silence finally enabled the book to exist, but its advent had been prepared right from the Meccan period of Mohammed’s preaching by the revelation of this celestial kitāb, meaning “Writing” or “Book”. As we will see, Arabic plays on the different meanings of this word. The announcement was probably coupled with partial collections compiled by members of the community, perhaps even officially because the Muslim tradition has preserved lists – admittedly not very reliable ones – of the “scribes of the Revelation” who recorded the prophetic words. But a book, a codex, was impossible for as long as these collections kept arriving. To take only one example from the data of the tradition: the last verses announced by Muhammad were inserted into the ninth Surah of the Qurʾān. The text’s evolution during the Prophet’s teachings was clearly incompatible with the closed form of the codex.

18The Qur’ān/muṣḥaf dichotomy exhausted the research and prevented the complexity of a situation that I qualify as “the muṣḥaf paradox” from being fully embraced. What place should the young community give to a book for which there were no Sunna in the prophetic tradition? What form should it take? These questions appeared in one of the accounts: “How can we do what the Prophet did not do?” Did Islam actually need a book? The affirmative answer provided by the facts seems to support Richard Bell’s analysis of the chronology of revelations. Bell identified different phases, the most recent of which was marked by the announcement of the kitāb – and, in parallel, by growing precision in the way in which the Qurʾānic text referred to the Old and New Testaments: Tawra and Injil.

19Taking the sequence of “people of the Book” of whom the Muhammadian teaching offered an example, the young community of believers was ready to become, in turn, a depository of the kitāb, “Book” or “Writing”, announced to it. This word, which the Qurʾān used most often to denote itself, recurred in numerous verses. At the end of a very detailed analysis of the various Qurʾānic occurrences, Daniel Madigan suggested that we understand it as “Revealed Writing” and not as “Book”. Yet it seems to me that we should not stop at the text itself, and that we need to consider the way in which the young community understood this term. The fact that it was chosen, on the one hand, in the form used to denote the followers of the other revealed religions – Jewish and Christian – which did have actual books, and, on the other, to refer to the Qurʾānic message, is by no means insignificant in a context in which the book was the accepted authority. In the final analysis, the very real question that the written form raised was therefore not so much that of the legitimacy of the operation, as that of the techniques that it required, and then of the use to which the book would be put. The proliferation of codices during the years following Muhammad’s death was in itself to reveal the fledgling Muslim community’s expectation.

20Conceptually, Islam had entered into the world of the book even before owning one. It just had to invent a tradition. The Qurʾān, as Angelika Neuwirth so rightly pointed out, is a text from Late Antiquity. The artefact that was to convey the prophetic message was also a product of that period. Its origins coincide with the time in which Islam, through its conquests, appeared in a world where those that the Qurʾān called the “people of the Book” prevailed from the Mediterranean shores through to Central Asia. Arabic imposed itself, consecrated by the new religion and supported, from the late seventh century, by a caliph’s decision to change radically the Near East’s cultural landscape. The administration started to use the Arabic language and alphabet, and the coins struck by ‘Abd al-Malik and his successors became Muslim, with text from the hundred-and-twelfth Surah on the reverse. However, as regards the book itself, the traditions of the “people of the Book” lasted. Without their own book tradition, the scribes to whom the writing of the Qurʾān was entrusted drew on the heritage of Late Antiquity. They adopted the techniques used to produce a codex, and the artists who created the first illuminated copies of the Qurʾān at the turn of the seventh and eighth centuries copied its decorative vocabulary. Calligraphers and illuminators subsequently severed these ties with the past and took advantage of the relative but surprising latitude they enjoyed to give the Qurʾān-book its singularity.

21The attention that was paid to it by the first generations of Muslims and that subsequently resulted in amendments says enough about the importance of the Qurʾān-book within the communities of the early centuries. It invites us to rediscover the materiality of these ancient copies and to re-evaluate the role of written transmission. I have pointed out the significant place of oral transmission in the Muslim scholastic tradition. The recited Qurʾān was valued highly. In view of this, the emphasis that traditional accounts of the written recording of Muhammad’s teaching placed on the book’s reliability is striking. It was echoed several centuries later in the attention that the Andalusian al-Dānī paid to ancient copies, which he tirelessly inspected in his quest for knowledge on Qurʾānic writings. It was also apparent in the selection of the consonantal text, the rasm, as one of the three criteria on which Ibn Mujahid based the acceptability of a way of reciting the Qurʾān. It is furthermore clearly apparent in the Fihrist or catalogue of works composed by al-Nadīm in the tenth century, the first chapter of which is devoted first to writing, then to the Qurʾān, and Qurʾānic science.

  • 1 Manuscrit arabe 399, Bibliothèque nationale de France.

22The improbable and Pirennian meeting between Abbasside Harūn al-Rachīd and Charlemagne through the intermediary of a Qurʾān1, which was authentic but posterior to these two legendary figures, raises questions (Fig. 5). What perception did the two contemporaries have of their respective sacred Book? What was its place in their world? To be sure, the two religions did integrate it into their practices in distinctly different ways and under considerably different conditions. I was nevertheless surprised to find that the Bibliothèque nationale de France has a far larger number of Abbasside Qurʾāns than it has of Carolingian copies of the Bible or the New Testament. The figures are by no means insignificant: 250 codicological units of the Qurʾān, as opposed to 70 of the Bible. And the latter figure reflects the result of a process of age-old collection within a vast space, whereas the Abbassid Qurʾāns in the Bibliothèque nationale all come from the same place, the Amr mosque in Fustat. And if, to compare two institutions, we turn to the Carolingian foundation of Saint-Gall, the imbalance is even greater: there are 39 biblical manuscripts in the abbey’s library, only one of which has both the Old and the New Testaments.

Figure 5

Figure 5

Copy of the Qurʾān reportedly given as a gift by Harūn al-Rachīd to Charlemagne, Near East, latter half of the ninth century.

Bibliothèque nationale de France, Arabic manuscript 399, f. 269 v°-270 r°. Source: Gallica/BNF.

23One may rightly object that the Qurʾān is slightly shorter than the New Testament and much shorter than the Bible. Indeed, does the term not refer to a “library”? Yet in the case of many Abbassid Qurʾāns it is certainly a matter of libraries. A large number of them are indeed series of thirty volumes conserved in mosques, in specific chests. Judging by the quantity of parchment used, some of them were extremely expensive books. What can we say about a mid-eighth-century in plano that required a thousand sheep skins? Calligraphy, which was still in its infancy, contributed to this outrageous project for which letters were enlarged to expand the text to the extraordinary dimensions of the volume – and I am deliberately using this word. More interesting still is the fact that at least three manuscripts of this kind survived, at least partially. The undertaking had thus been carefully thought out to produce unusual codices.

24What do these figures mean? I am talking of seemingly useless volumes since they played no role in Friday prayers, for example. So what purpose and who did they serve? Were the believers of Fustat more likely to read the Qurʾān than the monks of Saint-Gall were to read the Bible? Were there ever two-hundred-and-fifty readers or more in the ‘Amr mosque at the same time? Or did this proliferation have a symbolic value at a time when Muslims were still a minority? Reading these manuscripts was an effort, judging by the dichotomy between regular practice, illustrated by numerous papyrus records, and copies of the Qurʾān, in scripts of astounding beauty sometimes albeit difficult to decipher. The divorce between written forms in daily life and Qurʾānic calligraphy was characteristic of this period, which constituted a remarkable episode in the history of hierography, the signification of which warrants closer inspection. We need to scrutinize the question of literacy in Muslim societies of the time, not only in Egypt. Fustat was no exception; what was true for it applied equally to Damascus, Sana’a and Kairouan, where the same bibliothecae coranicae existed. In my lectures I will endeavour to provide answers to these questions raised by the study of the material aspect of early Qurʾānic manuscripts. I will for example explore the book economy in the Medieval Islamic world, along with its aesthetics, as well as literacy in Muslim societies, the symbolism of books, and so on.

25On this basis, I intend to analyse in detail the history of the text that started to be constituted. What does the proliferation of copies correspond to? The first and third caliphs, Abu Bakr and Uthmān, feature in the two most known accounts of the writing down of the Qur’an. The former, we are told, wanted to put a stop to forgetting, while the latter wanted to end heterodoxy. Both saw salvation in writing.

26In an amusing anecdote, Ibn al-Zayyāt, which means literally, “the son of the oil merchant”, and who is the founder of one of the “reading” schools of Kūfa, is said to have read zayt (oil) instead of rayb (doubt) at the beginning of the second Surah, and therefore to have understood the verse to say: “this Writing – where there is no oil, etc.” The joke is indeed a bit weak, but it has the merit of reminding us of the condition of the written text at the time, that is, in the first half of the eighth century. The confusion implies that the points differentiating the letters with the same form – or homographs – did not appear in the copy.

27During the latter half of the seventh century, the first conserved copies of the Qurʾān were effectively written using a highly defective medium (Fig. 6). It was defective not because it lacked the means to distinguish between the homographs, but because the transcribers very rarely used all its resources. In many situations they chose to dot with precision perfectly identifiable letters, and to leave in the dark those forms that might be of importance to the actual meaning of the Qur’an. These dots could, for example, enable one to distinguish between the second and third person of conjugated verbs, but the transcribers decided to leave that up to the reader, who might be as ill-informed as Ibn al-Zayyāt was! There was not even a common standard in spelling, and transcribers working on the same manuscript could have each adopted a particular position in this respect.

Figure 6

Figure 6

Copy of the Qurʾān, Near East, last third of the seventh century.

Bibliothèque nationale de France, Arabic manuscript 328, f. 29 r°. © BNF.

28The notation of vowels remained imprecise for a long time. Long vowels were progressively integrated into the consonant framework, from around 675, while the presence of short ones was indicated later. There is no doubt that the development of Arabic grammar during that period helped and even guided this progress, which spread across three centuries and facilitated Ibn Mujāhid’s task of establishing the standard text in the early tenth century. Looking at our corpus, it first presents successive layers of elements that increasingly limit the spaces in which variations could be made. First, among these elements, some constitute variants that the traditional literature has not recorded, and that enable us to analyse precisely the formation of the canon. Second, the dialectic between the written and the oral Qurʾān is evident in the work, partly as the result of technical reasons: the former oriented certain readings, while the latter helped to specify the method. Thus, the consecration of writing happened at a late stage and – along with variations that were either validated or excluded by Muslim tradition, but in all cases were registered – preserved other cases which however probably remained unknown to ninth- or tenth-century specialists.

29It thus took three centuries for the Qurʾān to begin to meet the objectives that tradition attributes to Uthmān; three centuries to get it to the state in which we know it; three centuries that will be at the heart of the teaching I intend to offer. The manuscripts that were conserved represent an essential counterpoint to the Arabic sources for our understanding of the way in which the Qurʾānic text was transmitted and the route that brought it to its current state. They enable us, in particular, to observe on the spot the activity of the “readers” or qurrā’ that I referred to through the figure of Ibn al-Zayyāt. Their unruly corporation, which specialized in the recitation of the Qurʾān, not only made up for the shortcomings of the first copies but also seems to have put new variations into circulation. On Ibn Mujāhid’s decision, some of these subsequently became canonical and divine, while others were excluded.

30The effort that led to the development thirty years ago of a typology for methodically studying ancient Qurʾāns coincided, by chance, with the discovery of the Sana’a palimpsest. The spectacular nature of the discovery gave this documentation from Islam’s early centuries an exposure that it had never had, at the very time when the Qurʾān-book was starting to be studied far more rigorously and where the two aspects, Qur’ān and muṣḥaf, could be reconciled within the same historical approach.

31This approach was careful not to turn its back on the data preserved by tradition, on which the corpus of ancient manuscripts shed new light. We can now see far more clearly what the medieval scholars’ intentions were, when they collected information that they endeavoured to streamline in treatises which eventually served as a basis for the publication of the Qurʾān compiled in the early twentieth century by al-Azhar scholars. Some of them, and I am thinking particularly of the Andalousian Abū Amr al-Dānī, were surprisingly modern in the interest they showed in ancient manuscripts as a source of information. Confronted however with the difficulty of conceptualizing the complexity of documentation that was literally proliferating and was often even contradictory, and of collecting material scattered over a vast geographical area, these scholars were led to proceed by selection. Part of the material that they chose not to keep survived in the manuscripts – and it is now up to us to seek that information and to subject it to textual critique.

32Let us now move to the last theme that I intend to explore in my teaching at the Collège de France. After Muhammad’s death and the writing of his word, a Lachmannian debate before its time divided the Muslim community. The question was: what was the authentic version of the Qurʾān, that which corresponded fully to the celestial archetype mentioned in the 43rd Surah, for example: “And indeed it is, in the Mother of the Book with Us, exalted and full of wisdom”? The followers of Ibn Mas‘ūd, a companion of Muhammad, maintained that the text he had transmitted was the faithful copy and that Uthmān’s codex did not have the same value. History decided otherwise and today’s believers consider the Uthmānian version to be the exact reflection of the “well-reserved tablet”, to use the Qurʾānic phrase. Shiite communities also challenged the validity of Uthmān’s codex, claiming in particular that he had deleted passages affirming the legitimacy of Alī and his disciples.

33In these conditions, can the philologist hope to obtain an Ur-Text? The Qurʾān, which stresses several times the existence of a reference text, indicates too that forgetting and losses have potentially affected the state of the text. Its history, as Muslim tradition reports it, gives us a glimpse of the difficulties awaiting the editor. One of the main reproaches addressed to the caliph Uthmān by those who were to assassinate him was the alleged destruction of manuscripts anterior to the version that he had commissioned: “The Qurʾān was multiple and you made it one!” they claimed. Were these other Qurʾāns different from the vulgate? We can assume so, but we have no concrete idea. If it was indeed so, the editor would irretrievably have lost valuable information.

34The Sana’a palimpsest strikingly confirmed the existence of competing versions and the possible losses that the accounts of destruction suggested (Fig. 7). It also shed new light on the proto-history of the text. It has conserved textual sequences that make it possible to assess fairly precisely the gap between its version and the vulgate, and thereby what a multiple, polyphonic Qurʾān might have been before the undivided triumph of the latter. In light of these particularities, we can conclude that the information transmitted by medieval Muslim scholars concerning the non-Uthmānian versions had been screened for the sake of orthodoxy, and reduced to a few insignificant variants.

Figure 7

Figure 7

Reconstruction of the bottom layer of a sheet of the palimpsest inv. 01-27.1

(Yemen?, latter half of the seventh century) by Hadiya Gurtmann.

35Reading the sources, one actually reaches the conclusion that the relationship that Muslims in the latter half of the seventh century maintained with the Qurʾānic text was far removed from the profound reverence that subsequently developed towards it. A truly surprising account tells of Muhammad’s favourite wife, Aisha, who declared: “The Qurʾān contains mistakes [against the language], but the scribes will correct them.” Was this a justification a posteriori for amendments to the text? Once again, the editor was concerned by the extent of the corrections and the means of identifying them. The analysis of the oldest testimonies had already evidenced the importance of grammatical corrections. We know, for example, that towards 700 the confusion between “He said” and “Say!” disappeared as the notation of the two verbal forms improved. Through constantly finer knowledge of this corpus, the transmission of the text appears with greater precision and partially challenges the traditional schema that the canonization of the text was achieved at an early date. Conversely, the first transcribers sometimes took a long time to make the texts they were transcribing agree with the prevalent form (it would be premature to talk of a “canon”). For some time, they safeguarded details that could correspond to the initial stage of the writing, or even to older notes.

36The development of a scientific edition of the Qurʾānic texts promises to be an arduous task. The existence of another recension, albeit fragmented, modifies substantially the nature of the editing work. Given the complexity of the task which I have endeavoured to outline for you, there is no doubt that the undertaking can find its place within what is now referred to as the digital humanities.

37The complementarity of the two branches of teaching that I intend to develop in coming years is fully apparent: the muṣḥaf offers special access to the Qur’ān and to its history. The systematic exploration of the considerable corpus of ancient manuscripts and the publication of these Materials for the History of the Text of the Qur’an – to cite the title of a book by Arthur Jeffery, a great specialist of the Qurʾānic text – will make available all the basic data for studying the way in which the text took on its currently known form, and for identifying the different phases of the process and perhaps also the existence of regional developments. Meeting the latter objective is trickier: even though, in the traditional presentation, canonical readings initially had a regional base, in principle at least, the origin of manuscripts still escapes us and their own orientations are not always aligned with the situation that Ibn Mujāhid describes. But I do not doubt that ever-finer knowledge of the early-centuries written testimonies will enable us to provide an answer to this crucial question.

38I therefore intend primarily to work on elucidating the history of the Qurʾān-book during the first four centuries of Islam, and to propose a reflection on the place and the role of the muṣḥaf during this period. The study of its materiality will enable us to restore entire segments of the history of Islam under the Umayyads and the Abbassids, be it in the fields of economics, intellectual movements or art. Resituated in their context, the Qurʾānic manuscripts of the early centuries allow us to glimpse the history of the evolution and transmission of the text. This is the route that historical inquiry will take in the next few years. Do not think that this research, of which I have attempted to show various facets, is of interest only to philologists or, worse still, Orientalists. It is of interest to a far broader public, way beyond the West. The points of view may not coincide, but the interactions and debates that it generates open reassuring perspectives, despite the – understandable – suspicion that is sometimes expressed. The ambitious objective of a scientific edition of the Qurʾānic text is now closer to being achieved, owing to this preparatory work. I can already see myself dreaming of taking up again the thorny issue of the chronology of revelations, once we finally have a well-edited text.

39After a long break of close to two decades, Arabic is back at the Collège de France. And what a comeback! The central text of Islam, the Qurʾān, will be at the heart of the teaching taking place at a time of profound renewal. The Chair “History of the Qurʾān. Text and transmission” is part of a long history of Arabic studies at the Collège de France. It is ushering in at the front door the Qurʾānic text at a time where the studies devoted to it are experiencing a prodigious renewal in all fields and where new avenues are being opened to research. This intellectual profusion requires that we proceed with utmost rigour at a time when some are interpreting the Qurʾān in a spirit of radical exclusivism and total ignorance of a complex past. I believe that the historical approach that I have chosen, based on facts, respectful of a past that can seem surprising because it has been completely forgotten, implicitly guarantees responsible development. It opens fascinating perspectives that will not however make me entirely forget – and I must apologize for this, my dear colleagues – the field that I have been clearing for many long years, in parallel with my research on the Qurʾān: I mean the history of the book in the Muslim world.

40I will say no more, not because of a lack of subject matter, but because it is time for silence. Does Oriental wisdom not recommend us to manage it adroitly? The shadow of Scheherazade, introduced by my mention of Galland and Harūn al-Rachīd, urges me to follow her example. I will tell you the rest another time.


1 Manuscrit arabe 399, Bibliothèque nationale de France.

Table des illustrations

Titre Figure 1
Légende Copy of excerpts from the Qurʾān, Paris (?), late sixteenth/early seventeenth century.
Crédits Bibliothèque nationale de France, Arabic manuscript 460, f. 1 v°-2 r°. © BNF.
Fichier image/jpeg, 4,1M
Titre Figure 2
Légende Facsimile of a folio of the Qurʾān conserved in Copenhagen, from around 1730.
Crédits Bibliothèque nationale de France, Arabic manuscript 580 a, f. 3 v°. © BNF.
Fichier image/jpeg, 952k
Titre Figure 3
Légende Qurʾānic fragments in the collection of the Great Mosque of Damascus.
Crédits Photo: François Déroche.
Fichier image/jpeg, 296k
Titre Figure 4
Légende Full-page embellishment of a Qurʾānic manuscript in thirty volumes, Near East, latter half of the ninth century. Private collection.
Crédits Photo: François Déroche.
Fichier image/jpeg, 2,3M
Titre Figure 5
Légende Copy of the Qurʾān reportedly given as a gift by Harūn al-Rachīd to Charlemagne, Near East, latter half of the ninth century.
Crédits Bibliothèque nationale de France, Arabic manuscript 399, f. 269 v°-270 r°. Source: Gallica/BNF.
Fichier image/jpeg, 1,5M
Titre Figure 6
Légende Copy of the Qurʾān, Near East, last third of the seventh century.
Crédits Bibliothèque nationale de France, Arabic manuscript 328, f. 29 r°. © BNF.
Fichier image/jpeg, 4,5M
Titre Figure 7
Légende Reconstruction of the bottom layer of a sheet of the palimpsest inv. 01-27.1
Crédits (Yemen?, latter half of the seventh century) by Hadiya Gurtmann.
Fichier image/jpeg, 4,7M

© Collège de France, 2020

Conditions d’utilisation :