Wherever we find a religion that has a Scripture, that fact presents scholarship with the problem of the textual history of that Scripture. There are no exceptions to this among the historic religions. In the case of Buddhism, for example, we have the problem of the Pali Canon, the Sanskrit Canon, the Tibetan Canon, and the Chinese Canon. In the case of Zoroastrianism there is the liveliest dispute among Iranian scholars at this very moment as to the Avestan text, and, as is well known, the text of the Pahlavi books is an exceedingly complicated problem. Each generation of students for the last hundred years has found itself faced with new problems concerning the text of the Old Testament, and our own memories are still fresh with the excitement caused by the discovery of the Chester Beatty Papyri and the Ryland's Gospel fragment, both of which raised lively discussions on matters related to the textual history of the New Testament. Whether we face the text of the Book of the Dead, coming from the ancient Egyptian religion, or the text of the Qur'an coming from the youngest of the great historic religions, we have the problem of the history of the text.

In the case of none of the historic religions do we have the autographs of the original Scriptures. What we have in our hands are the documents that have come down in the various communities, and which have been more or less tampered with in transmission. This tampering does not mean tampering with evil intent; it may, indeed, have been with very good intent, but nevertheless it was tampering. The Avesta, for example, was written out in Sassanian times in a new alphabet based on the characters of Sassanian Pahlavi, and we have no knowledge whatever of what the original Avestan script was like. Similarly the Hebrew Scriptures as we know them are in the "square script", but this was not the script used when their originals were written. Moreover, the "pointing" that is in the text of all our copies is a relatively recent addition to the text, and at least three varieties of this "pointing" are known. When we come to the Qur'an, we find that our early MSS are invariably without points or vowel signs, and are in a Kufic script very different from the script used in our modern copies. This modernizing of the script and the orthography, and the supplying the text with points and vowel signs were, it is true, well-intentioned, but they did involve a tampering with the text. That precisely is our problem. We have a received text, a textus receptus which is to be found in all the ordinary copies in popular use. It is not, however, a facsimile of the earliest Qur'an, but a text which is the result of various processes of alteration as it passed down from generation to generation in transmission within the community. What do we know of the history of this process of textual transmission?

There is, of course, an orthodox theory as to this textual transmission. The Parsis of India have an orthodox theory as to the transmission of the text of the Avesta, and in Rabbinic literature we have an orthodox theory as to the transmission of the text of the Old Testament; and though scholarship cannot accept these orthodox theories, they have the interest of being the traditionally accepted account of the textual transmission within these communities. The orthodox Muslim theory can be stated quite simply. Before the creation of the world Allah created the Tablet and the Pen, and at His command the Pen wrote on the Tablet all that was to be. As each successive Prophet appeared the angel Gabriel revealed to him from the Tablet the message therefrom that he was to deliver. When the Prophet Muhammad came, and it was time for his ministry to commence, the angel Gabriel came to him also, and from time to time over some twenty years revealed to him those passages from the Tablet that he was to proclaim as the Word of Allah. Each year Gabriel used to collate with the Prophet the passages revealed that year to make sure that they had been recorded correctly. The last year of the Prophet's life they so collated the material twice. As the Prophet from time to time proclaimed his messages thus received from Gabriel to the people. He had his amanuenses copy them down, so that when he died all the material that had been given him as revelation was already written down and carefully collated, so that it was an exact transcript of what was written on the Preserved Tablet in the Heavens. In the Caliphate of Abu Bakr this material was put into Codex form as a first Recension, which served is the official text for his Caliphate and that of 'Umar his successor. During the Caliphate of 'Uthman, however, it was found that this material was being recited by different groups of Muslims in different dialectal forms, so 'Uthman sent to Hafsah, the daughter of 'Umar and widow of the Prophet, and had her bring out the copy that had been in her possession since her father's death. Then he appointed a Committee of men of the Quraish, and had them write out a new recension in the pure Quraish dialect. When this was done he had four copies of it made and sent one to Kufa, one to Basrah, one to Damascus, and one to Mecca, and ordered all other copies in existence to be destroyed. All our modern copies are the direct descendants of this standard official text of 'Uthman. Indeed, the Egyptian standard text of 1342 A.H. expressly says -

This is not, however, the history of the text as modern scholarship reads it.

To begin with it is quite certain that when the Prophet died there was no collected, collated, arranged body of material of his revelations. What we have is what could be gathered together somewhat later by the leaders of the community when they began to feel the need of a collection of the Prophet's proclamations and by that time much of it was lost, and other portions could only be recorded in fragmentary form. There is a quite definite and early Tradition found in several sources which says, "The Prophet of Allah was taken before any collection of the Qur'an had been made". Muslim orthodoxy holds that the Prophet himself could neither read nor write. But in our generation both Professor Torrey of Yale and Dr. Richard Bell of Edinburgh, working independently of each other, have concluded that the internal evidence in the Qur'an itself points to the fact that he could write, and that for some time before his death he been busy preparing material for a Kitab, which he would leave to his people as their Scripture, to be to them what the Torah was to the Jews or the Injil to the Christians. There is, indeed, an uncanonical tradition current among the Shi'a, that the Prophet had made a collection of passages of his revelations written on leaves and silk and parchments, and just before his death told his son-in-law Ali where this material was kept hidden behind his couch, and bade him take it and publish it in Codex form. It is not impossible that there was such a beginning at a collection of revelation material by the Prophet himself, and it is also possible that Dr. Bell may be right in thinking that some at least of this material can he detected in our present Qur'an. Nevertheless there was certainly no Qur'an existing as a collected, arranged, edited book, when the Prophet died.

At first the leaders of the community, who had charge of the community affairs after the Prophet had gone, do not seem to have felt the need for any collection of the revelations. It was only after the community began to settle down to the new situation in which it found itself, that the need for a record of these revelations began to make itself felt. While the Prophet was alive, the fountain of revelation, so to speak, was still open. New injunctions might at any time come to abrogate earlier injunctions which were no longer adequate for the developing life of the community; or fresh revelations might be forthcoming to meet new situations that were arising. The rapidly developing community life in Madina had meant that the Muslim community was continually being faced with unexpected community problems, and they had grown accustomed to coming to the Prophet for instruction, and for the solution of their problems. The customary form for these instructions to take was that of revelations. With the death of the Prophet, however, the source of revelations automatically ceased to flow, and his immediate successors had perforce to direct community affairs in accordance with what was known of revelations that had been given.

But what revelation material was available to these early successors of the Prophet? There were some passages, particularly passages of a legalistic character, that the Prophet had himself ordered to be written down, and which were still in the possession of the community. There were also some passages of a liturgical nature which were used in the daily prayer services, and which, whether written down or not, would have been memorized by a goodly number of members of the community. There may have been passages in written form among the Prophet's own possessions. There certainly were many passages of revelation which individual members of the community had written down, not because the Prophet had ordered them to do so, but because they themselves were interested in having them so in written form. Then there was the memory of the community. That tradition is probably sound which says that the revelations proclaimed by the Prophet were with few exceptions relatively short, and there would have been many members of the community who could remember numbers of revelations given forth on various occasions. When the early leaders of the community needed to know if there were any injunctions extant regarding one matter or another, it was to these sources of information that they turned.

Perhaps even in the Prophet's own lifetime there were certain members of the community who took an interest in "collecting" the pronouncements of their Prophet. In this there is nothing unusual. It was precisely this that in the earliest Christian community provided those collections of "Sayings of Jesus", that we find among the basic material of the Gospels. Certainly after the Prophet's death we find certain members of the community interested in increasing their collections of the pronouncements of the Prophet, and these presently came to be known as the Qurra - the Reciters who became a kind of depository of revelations to whom the civic leaders could turn for information, when such was needed, as to whether there was any revelation which might decide how they should deal with such and such a situation. Some of these Qurra' might have chosen to memorize as much as they could discover of the various revelations, while others chose rather to commit their collections to written form. There has been a suggestion that the Prophet himself had begun to organize a body of Qurra who were to be the guardians of revelation. But the evidence adduced for this is extremely tenuous and the early history of the Qurra' is still veiled in the greatest obscurity.

Here, however, we have our first stage in the history of the text of the Qur'an. There could not be a definitive text while the Prophet was still alive, and abrogation of earlier material or accessions of fresh material were always possible. With his death, however, that situation ended, and we have what was preserved of the revelation material, partly in written form, partly in oral form, in the hands of the community, and tending to become the special care of a small body of specialists. Tradition says that it was the sudden danger of the loss of these specialists that led to the next stage in the history of the text. We read that at the battle of Yemamah in the year 12 A.H. so many of the Qurra' were among the slain that 'Umar suddenly awoke to the fact that a few more battles like that of Yemamah might mean that a great portion of the revelation material would be irretrievably lost, and so he came to the Caliph Ahu Bakr and urged on him the necessity of getting this material that was in the possession of the Qurra' assembled and written down in some fixed form, ere it was too late. As it is we find numerous references in tradition to verses, which were "lost on the Day of Yemamah". Abu Bakr, the story continues, demurred, asking who was he that he should do a thing which the Prophet himself had not done, and about which he had left no commandment. 'Umar, however, convinced him, and he summoned Zaid b. Thabit, who had been an amanuensis of the Prophet, and bade him assemble from the community all that any of them had of the revelations of the Prophet, and write them out in goodly form. Zaid, it is said, also demurred, asking what business they had undertaking to do a thing which the Prophet had not seen fit to do, and about which he had left no commandment. 'Umar, however, convinced him also of the urgency and necessity of the task, and Zaid, so the tradition records, set about assembling the material from leaves, from white stones, from the shoulder-blades of camels, and from the breasts of men. In other words, he assembled the material available, both oral and written, in an attempt at a first definitive text of the revelations.

The text thus obtained Tradition regards as officially promulgated by Abu Bakr, and so the first Recension of the text of the Qur'an. Modern criticism is willing to accept the fact that Abu Bakr had a collection of revelation material made for him, and maybe, committed the making of it to Zaid b. Thabit. It is not willing to accept, however, the claim that this was an official recension of the text. All we can admit is that it was a private collection made for the first Caliph Abu Bakr. Some scholars deny this, and maintain that Zaid's work was done for the third Caliph, Uthman, but as 'Uthman was persona non grata to the Traditionists, they invented a first recension by Abu Bakr so 'Uthman might not have the honour of having made the first Recension. Someone, however, must have made the collection that Hafsah the daughter of 'Umar, later produced to form part of the material used in 'Uthman's recension, so that we must think of some private collection made either by Abu Bakr or 'Umar, and it may well have been by the first Caliph, but it was a private not an official undertaking.

As a matter of fact, there were others besides Zaid b. Thabit who had busied themselves with this task of assembling in Codex form a complete collection of what still survived of the revelation material which now makes up the Qur'an. Tradition knows the names of several of these, e.g. Salim b. Mu'qib, who was killed at the battle of Yemamah, and who, tradition says was the first to make such an attempt at setting all his material down in Codex form; 'Ali b. Abi Talib, who is said to have endeavoured to arrange the revelations in their chronological order; Anas b. Malik, whose Codex may have been based on that of his uncle Abu Zaid, who was well known as one of the early collectors of revelation; Abu Musa al-Ash'ari whose Codex was a large one, and was familiarly given the name of Lubab al-Qulub; and various others, including the two famous Codices of Ubai b. Ka'b and of 'Abdallah b. Mas'ud from both of which a great body of variant readings has survived. It is frequently asserted that the verb jama'a "to collect", as used in this connection, means only hafaza "to memorize". The verb, it is true, can have this meaning, but since 'Ali is said to have packed up what he "collected" on his camel and brought it along; since what Abu Musa had "collected" was something that had a nickname; and since the friends of Ibn Mas'ud at Kufa supported him in his refusal to give up what he had "collected" to be burned , it is quite clear that we are dealing with collections that were in written form. In the case of the Codices of 'Ali, Ubai and Ibn Mas'ud, indeed, we find tradition which professes to give the order in which the revelation material was arranged in their Codices an order which differed considerably from that found in our present Qur'an.

The most important fact that Tradition has preserved in connection with these early Codices, however, is the fact that certain of them came to attain the position of metropolitan Codices. Thus we read that the people of Kufa came to regard the Codex of Ibn Mas'ud as in a sense their Recension of the Qur'an, the people of Basra the Codex of Abu Musa, the people of Damascus the Codex of one Miqdad b. al-Aswad, and the Syrians other than the folk of Damascus, the Codex of Ubai. This is exactly what might have been expected, and has a close parallel in the case of the New Testament, where the texts that go under the name of the Alexandrian text, the Neutral text, the Western text, the Caesarean text, were recensions of the text, differing slightly from one another, and favouring certain groups of variant readings, which had grown up and come into use in certain important centres of Church life. As Kufa, Basra, Damascus, and Homs began to develop into important centres of the Islamic community, it was quite natural that they, as well as Mecca and Madina, would want their own collection or revelation material, and the Tradition reflects the fact that different recensions of the material came into use at these different centres. Such recensions, while embracing in general the same body of material, always differ from one another in the inclusion or exclusion of certain material, and in their choice among a multitude of variant readings, and this holds of these early metropolitan Codices of Islam. Thus we know that the Codex of Ibn Mas'ud omitted Suras I, CXIII and CXIV, and that both the Codices of Ubai and Abu Musa included two short Suras, which are not in our present text, while a considerable body of variant readings from these Codices is to be gathered from the grammatical, lexical, exegetical and masoretic literature of latter generations which still remembered and discussed them. There were once, indeed, a number of special works, under the name of Kitab al-Masahif, which specially discussed this stage of the Old Codices, and it was a fortunate accident which enabled the present writer to discover and publish the text of the sole surviving example of these, the Codex Book of Ibn Abi Dawud.

It was the existence of these variants in the texts used in different centres that led to the next stage in the history of the text. The story in which the memory of this is enshrined is that Hudhaifah b. al-Yaman on being sent to the armies that were fighting in Azarbaijan was horrified to find the Kufans and Syrians disputing over the correct reading in passages that they were using in their devotional services, and in some cases even denying that what the others were using was really part of the Qur'an. In his distress, he returned to the Caliph 'Uthman at Madina and said "Overtake this people before they differ over the Qur'an the way the Jews and Christians differ over their Scripture." 'Uthman was persuaded and sending for Zaid b. Thabit, laid on him the task of making this official recension. Tradition says that he did four things in this connection. First he made an announcement in the mosque calling on all who had any revelation material to bring it to Zaid b. Thabit. Second, he sent to Hafsah to get the material that had come down to her from her father 'Umar. Hafsah produced this from under her bed, and it was found that the worms had eaten it in places, but apparently its material was used for 'Uthman's recension and then returned to Hafsah, for at her funeral the Governor Marwan, who had tried in vain to get it from her during her lifetime, demanded it of her brother and destroyed it, fearing, he said, that if it got abroad, the readings that 'Uthman desired to repress would recommence. Third, he appointed a Committee to work with Zaid b. Thabit, to scrutinize all the material sent in, to accept only that for which two witnesses could he found, and to see that what was written was written but in the genuine Quraish dialect. Fourth, when the work was completed he had copies made and sent to the great metropolitan centres, with orders that all other Codices or portions of revelation material in circulation be destroyed. Some traditionists tell us that this was known as "the year of the destruction of the Codices", and for long afterwards we hear echoes of the bitter hostility of the Qurra' to 'Uthman because of his work in thus canonizing the Madinan text tradition and prohibiting the use of any other.

'Uthman's official Recension gained rapid and almost universal acceptance. Only in Kufa do we hear of any considerable support for one of the earlier texts, for there the text of Ibn Mas'ud continued for some time to dispute the authority of the new canonical text, but even Kufa had finally to come into line with the rest of Islam, and accept the Madinan text. It is always arguable that as Madina was the Prophet's own city, and was the home of the majority of "Old Muslims" who had been closest to the Prophet, the Madinan text tradition had all the chances of being the best available type of text. It is worth emphasizing, however, that at the time it was only one of several types of text tradition in existence, and 'Uthman's work in recording it in a definite and final form, closed a stage in the history of the text. Up till that time had been the period of the Old Codices, but from then on we trace the history of one Codex only, that which represents the official Recension of 'Uthman. Attempts have been made to avoid this conclusion by claiming that all that 'Uthman did was to remove dialectal peculiarities that had crept into the pronunciation of the Qur'an as it was recited, and have a standardized type of text written out in the pure dialect of the Quraish. This matter of Quraish dialect is indeed mentioned in the traditions referring to this Recension, but to pretend that it was merely a matter of dialectal variations is to run counter to the whole purport of the accounts. The vast majority of dialectal variations would not have been represented in the written form at all, and so would not have necessitated a new text. The stories of Zaid and his colleagues working on the text make it perfectly clear that they were regarded as recording a text de novo, for we read that at times when there was only one witness available for a certain passage they would wait till another witness who knew that passage had come back from the wars, or wherever he had been, and recite it to them; and there were discussions among them as to where certain passages belonged in the collection. Finally, the mass of variant readings that has survived to us from the Codices of Ubai and Ibn Mas'ud, shows that they were real textual variants and not mere dialectal peculiarities.

The text that 'Uthman canonized, however, was a bare consonantal text with marks to show verse endings, but with no points to distinguish consonants, no marks of vowels, and no orthographic signs of any kind. Unfortunately we do not know the precise kind of script in which it was written. The earliest fragments of Qur'anic MSS which survive to us are all written in a kind of script that grew up in the city of Kufa as a special script for the writing of Qur'ans and which we call the Kufic script. None of these fragments however, can be dated earlier than the second century of the Hijrah, and it is indeed, doubtful if any are really older than the third century. One often reads of there being still in existence Qur'ans written by the hands of 'Uthman or of 'Ali; or of 'Ali's sons al-Hasan and al-Husain, but such attributions are merely the fruit of pious imagination. The late Professor Bergsträsser collected some twenty references to claims made by different centres of Islam to possess the famous Codex of 'Uthman himself which he was reading when assassinated and whose pages were discoloured by his blood.

Faced with a bare consonantal text the reader obviously had to interpret it. He had to decide whether a certain sign was a shin or a sin, a sad or a dad, a fa or a qaf, and so on; and when he had settled that he had further to decide whether to read a verbal form as an active or a passive whether to treat a certain word as a verb or a noun, since it might be either and so on. In the first generation this problem would not have been so serious for the Qurra', for memory of what the text should be would in many passages decide the matter of how it was to be pointed and vowelled, and where the pauses that governed the meaning should be. Theoretically one could suppose that this oral tradition as to how the text should be read could be transmitted carefully from generation to generation, as was the case with the old poetry. But actually the enormous body of variant readings that has been recorded proves that there was no consistent tradition on this matter transmitted. From the date of the publication of 'Uthman's official text till the year 322 A.H. we are in the period of ikhtiyar or "free choice", and it is very curious that though the 'Uthmanic text is now taken as the basis, many famous savants, even to the end of this period, were accustomed to state their preference in certain passages for readings from one or other of the old non-'Uthmanic Codices. As we might have expected we find that the iktiyar of certain famous teachers tended to be perpetuated by their students and win acceptance in more or less extensive circles, so that before long we begin to hear of students studying the ikhtiyar of So-and-so as to huruf and the riwaya of So-and-so as to qira'a i.e. their schemes first of pointing and then of vowelling the unpointed unvowelled text. This again tended to crystallize at the great centres where students congregated, so that soon we begin to hear of the tradition of the Kufans, or the traditious of the Basrans, or Syrians, etc. as to the correct way of pointing and vowelling the text, this meaning the tradition that had come to dominate in their Schools in which Qur'anic learning was cultivated. At a fairly early date we hear of three principles emerging and being laid down to guide the ikhtiyar, viz. mushaf 'arabiyya, and isnad. That is, the reading proposed must be one that will check with the consonantal text, will be in consonance with the laws of Arabic grammar, and be a reading that has come down from some reputable authority. There was, of course, dispute about these rules. Some claimed that so long as a reading was good Arabic and made good sense it did not matter whether it came from the 'Uthmanic Codex or one of the other Old Codices, since they also came from the time of the Prophet. Some were contemptuous of isnad, but that a reading must be sound Arabic diction was naturally accepted by all.

The next stage was to indicate these readings in the text itself. One did not have to mark them in the text, of course, for once they had been memorized properly the Reader could take up a copy of the consonantal text and read according to what he had memorized. Memory, however, is a very treacherous thing and very soon the custom was introduced, based apparently on a practice in vogue among the Christians using Syriac Scriptures, of marking the readings by a system of dots, black and coloured. Tradition makes it clear that there was very considerable opposition to the introduction of these points into the Codices, this being regarded as "innovation" and so smacking of heresy. There is no unanimity as to who first introduced the systems of points, the favourite names in connection therewith being those of Yahya b. Ya'mar and Nasr b. 'Asim. There was at first a period of fluidity, and we actually have some fragments of Codices in which by the use of vari-coloured dots different possibilities of pointing one and the same word are indicated, while the great majority of words have no points marked at all. This suggests that at first only those words would be pointed where there was some uncertainty as to what the correct reading should be. That the practice of pointing came to be generally accepted and consistently carried through the whole of a Codex is said to be due to the activity of the famous official al-Hajjaj b. Yusif, who was perhaps the most remarkable figure in Islam during the Caliphate of 'Abd al-Malik. When we come to examine the accounts of the activity of al-Hajjaj in this matter, however, we discover to our surprise that the evidence points strongly to the fact that his work was not confined to fixing more precisely the text of the Qur'an by a set of points showing how it was to be read, but he seems to have made an entirely new Recension of the Qur'an, having copies of his new text sent to the great metropolitan centres, and ordering the destruction of earlier copies in existence there, much as 'Uthman had done earlier. Moreover this new text promulgated by al-Hajjaj seems to have undergone more or less extensive alterations. The Christian writer al-Kindi in his polemical work known as the Apology of al-Kindi makes a controversial point out of the alterations he claimed that al-Hajjaj, as everyone knew, had made in the text of the Qur'an, but this was regarded hy scholars as just a polemical exaggeration such as one might expect in a controversial writing. However, in the Kitab al-Masahif of Ibn Abi Dawud, already mentioned we have recorded in a special chapter a list of readings in our Qur'an text which are due to alterations made by al-Hajjaj. If this is so, our textus receptus is not based on the Recension of 'Uthman, but on that of al-Hajjaj b. Yusif.

The limitation of ikhtiyar came in the year 322 A.H. when the Wazirs Ibn Muqlah and Ibn Isa, guided by the great savant Ibn Mujahad, settled on seven systems of reading the text, and decreed that these alone were canonical, permissible ways of pointing and vowelling the text. Their decision did not go unchallenged, but the severe punishment meted out to two famous scholars, Ibn Miqsam and Ibn Shanabudh, who persisted in their right to ikhiyar, and to read, if they saw fit readings from the Old Codices, soon convinced the Readers that the period of ikhtiyar was over, and they were faced with a limitation which marked a new stage in the history of the text.

The Seven Systems chosen by Ibn Mujahid were those of Nafi of the Madinan School, Ibn Kathir of the Meccan school, Ibn 'Amir of the Syrian School, Abu 'Amr of the Basran school, and 'Asim, Hamza, and Al-Kisa'i of the Kufan school. His choice was not unchallenged. Some seriously objected to the fact that there were three among the seven from the Kufan School, and desired one of them to be replaced by a reader from another School, some favouring Abu Ja'far of the Madinan School, and others Ya'qub of the Basran School. In particular the position or al-Kisa'i in the group was challenged, and the candidature or Khalaf of the Kufan School was for long vigorously pressed. Ibn Mujahid's choice, however held, and the systems of his seven are still the canonical Seven, though in many instances the masoretic works continue, as e.g. in the famous work an-Nashr of Ibn al-Jazari, to record the variants of the Ten, i.e. the Seven with the three whose candidature was pressed. Some masoretic works, indeed, preserved fourteen systems, including besides the Ten the readings of four other Readers, Ibn Muhaisin of Mecca, al-Hasan of Basra, al-Yazidi of Basra, and al-A'mash of Kufa, whose systems had had some backing as more worthy to be included in the Seven and made canonical than some of those chosen by Ibn Mujahid, but which had failed to find any very general acceptance. The famous work al-Ithaf of al-Banna', for example, records the readings of all fourteen. Here and there yet other claimants were supported, but for reasons which are not at all clear, Ibn Mujahid was able to gain official support for his seven, and within half a century they had gained very wide acceptance.

We do not have the systems of any of these seven in the form given it by its founder. These seven systems were transmitted in the Schools, and very shortly after their acceptance as canonical we find a great many riwayas in existence as to how each of them read. In the case of one or two of them the riwayas were very considerable in number. By the time that ad-Dani, who died in 444 A.H. came to write his Taisir, two riwayas from each of the seven had been chosen as canonical, and as alone having official sanction. As to how these were chosen we have no information whatever, and at present cannot even venture a guess. All we know is that the process of fixing the text ne varietur had gone this further stage, and as such had been recorded by ad-Dani. For Nafi' were chosen the riwayas of Qalun and Warsh; for Ibn Kathir, the riwayas of Qunbul and al-Bazzi, for Ibn 'Amir the riwayas of Ibn Dhakwan and Hisham; for Abu 'Amr the riwayas of ad-Duri and as-Susi, for Hamza the riwayas of Khalaf and Khallad; for 'Asim the riwayas of Hafs and Abu Bakr; for al-Kisa'i the riwayas of ad-Duri and Abu 'l-Harith. Any reading from any of these riwayas is canonical. No official decision that we know of was taken to establish these particular riwayas as alone permissible, and so the use of the word "canonical" is not quite accurate, but these riwayas did come to take a position of unique authority for which we have no more appropriate word than canonical. As such, one or other of them would be followed when scribes were writing new Codices and indicating therein the pointing and the vowelling.

These systems for marking the readings, however, were not the only signs now added to the text. Signs for verse endings appear in the very earliest fragments of Codices in our possession, though there was by no means universal agreement as to where these pausal marks fell, so that this now comes to be a matter to be settled in the Schools, and the masoretes record tables of Kufan verse endings, or Basran or Syrian or Madman verse endings, as the case may be, and at times signs to indicate where there was variant tradition as to the place where the ending should fall, were inserted in the text. The Suras had been marked off also from very early times, but without rubrics. Now begins the custom of setting at the head of each Sura its name. Different names were used in different localities and even to the present day there is no complete agreement as to the names that appear at the head of certain Suras in Qur'ans lithographed in different centuries. But besides Suras and verses other divisions of the text began to be marked. Some scribes placed a mark at the end of each group of ten or five verses; some divided the text into sevenths and marked the end of each in the text; some used special marks for the beginnings and endings of halves, fourths, eighths, etc. A more popular practice was to divide the text into thirty parts so that one part might be read each day for the month, and these divisions with the quarters and halves thereof, were carefully marked, and this division into ajza and ahzab became so popular that to this day old-style Muslim savants quote the Qur'an by juz' and hizb rather than by Sura and verse. Of more practical importance was the introduction of pausal signs, which, like the similar set of pausal signs in the Hebrew Bible, are a guide to the sense and serve precisely the same function as our punctuation marks. The earliest set of such pausal signs to come into use was a very simple one to indicate "no pause", "optional pause", and "necessary pause", but in the Schools these developed into more elaborate systems by increase in the types of optional pause recognized, though there was also considerable difference among the Schools as to precisely where some of the "no pause and "must pause" signs should be placed. The addition of these signs to the text, of course, represents a further step in the history of the process of fixing a type of received text, though it is not yet possible to write the whole story of the way in which these various systems of pause developed.

One other step in this process. which was a step of great importance, but whose origin is also at present veiled in obscurity, was the selection of a standard exemplar. Students of the Old Testament will remember that the consonantal text in our Hebrew Bibles is derived from MSS which represent one standard exemplar, which once chosen had to be followed by the scribes and reproduced with minute faithfulness, so that even its misspellings and peculiarities of orthography had to be reproduced in all copies. The same is true of the Qur'an. By the time ad-Dani was writing his Muqni' and Taisir this standard exemplar had been chosen, for in his Muqni', which is a book of instructions for the scribes who are copying exemplars of the Qur'an, he gives them in detail the rules they must observe in the practice of their profession, and so carefully lists all the peculiar spellings and oddities of orthography which they must be careful to reproduce in their copies, even though they may know that these are mistakes. Thus in XIX, I must he written with final and not the normal ; in XVIII, 36 has to be written with a long at the end instead of the normal ; in XX,95 must be written instead of ; in XVIII,47 must be written instead of the correct form ; in XXXVII, 130 has to be written instead of and so on. Who the scribe of this standard exemplar was, and how it was chosen, are questions we cannot as yet answer. The orthodox theory is that these peculiarities were already present in the Imam, the standard consonantal text prepared by the orders of 'Uthman, but as these peculiarities do not always appear in the early fragments of Kufic Codices it is doubtful if this is so.

All the Seven systems mentioned above are equally canonical, and the Qur'an may be recited according to either of the riwayas chosen for each. No written text, however, can express in the text all the variants of all the seven. There are a few known examples of fragments of Kufic Codices where, by the use of differently coloured dots, variant readings in the case of individual words are recorded. There are also some MSS known with marginal annotations which give a selection of the variant readings from among the Seven, or even from among the Ten. The usual, and in fact the only feasible practice, however, is to have the text written according to one of the systems from among the Seven. No systematic survey of all Qur'anic MSS with a view to determining their type of text, has ever been made, but so far as they have been examined the result is to show that only three of the possible fourteen riwayas are known to have had any considerable vogue in the writing of Codices. In the Sudan up to a generation ago, there were apparently texts written according to the Basran system of ad-Duri. In North Africa, from Tripoli along to Morocco, the common form of text found in the MSS, and in many lithographed editions, is the Madinan tradition of Warsh. Everywhere else in the Muslim world the only type of text found in use is that of the Kufan Hafs, the rawi of Asun. This Hafs text has in recent years entirely superseded that of ad-Duri in the Sudan, and is rapidly superseding that of Warsh in North Africa. Thus we reach the end of the history of the text of the Qur'an with the practical dominance of the text tradition of Hafs as the textus receptus for all Islam. It was in view of this that the Egyptian standard edition of 1344 A.H. made an attempt to clear the text from modernizations of orthography and an overloading of masoretic markings, and restore as far as possible a pure type of Hafs text. Owing to the use by its editors of relatively later authorities instead of going back to the earliest sources of our information as to this Hafs text, they have not quite succeeded in producing a pure type of Hafs text, but it is better than anything else available, and very much superior to the Flügel text, which has been used almost exclusively by European savants since it first appeared in 1834.

The next stage will be that of a critical text. The ideal would be to print on one page a bare consonantal text in the Kufic script, based on the oldest MSS available to us, with a critically edited Hafs text facing it on the opposite pages and with a complete collection of all known variant readings given at the foot of the page. The present writer was collaborating with the late Professor Bergsträsser on such a project, and a beginning had been made on both the connected problems. The writer has gone through all the printed literature and a good deal of MSS material to collect all the variant readings. Bergsträsser established at Munich a Qur'anic Archive in which he commenced to gather photographs of all early Qur'anic MSS, and of all masoretic material connected therewith. After his untimely death this Archive was continued and developed by his successor Otto Pretzl, but Pretzl was killed outside Sebastopol during this late War, and the whole of the Archive at Munich was destroyed by bomb action and by fire, so that the whole of that gigantic task has to be started over again from the beginning. It is thus extremely doubtful if our generation will see the completion of a really critical edition of the text of the Qur'an.