By studying problems connected with the Masoretic text of the Hebrew Bible, I came across an Arabic manuscript belonging to the collection of rare Arabic texts of Mr. A. Chester Beatty, of which I was preparing a catalogue at that time. The manuscript deals with the reading of the Koran and seemed to me to be of a certain importance as it offered some striking parallels to the development of the Hebrew Masora. So I published the text with the translation in my Schweich Lectures on the Cairo Geniza and discussed the problems connected with the text. I Later on I came across a second manuscript confirming in several directions the results I had found. I have dealt with this text in an article, "The Qur'an and the `Arabiya," which I wrote as a contribution for the Goldziher memorial volume (pp. 163-82), which as I hear will be published in the near future. But since writing this article, new problems have arisen, especially in connection with the correspondence with Professor Johannes Fuck of Halle University, a great authority on Islamic tradition, and I believe it advisable to discuss here the whole problem again.
At the International Congress of Orientalists held in 1905 in Algiers, Carl Vollers, an expert not only on classical Arabic but also on the Arabic spoken in Egypt, where he had been for many years the director of the Khedivial Library, tried to show that the Koran was read by the Prophet in a language in which many of the rules of classical Arabic were not observed; that this language, therefore, differed clearly from the classical Arabic in which the Arabs are accustomed to read the Koran. The classical language was developed in the century after the Prophet on the basis of Bedouin poetry by early grammarians and readers of the Koran, and to this language the consonantal text of the Koran was adapted.
The reading of this lecture in an Islamic country gave rise to a real revolt among the Muslim members of the Congress. Vollers then developed his theory in his book Volkssprache and Schriftsprache im alten Arabien.2 He attempted here to characterize the language spoken in Mecca at the time of the Prophet on the basis of variant readings in the Koran that he collected from various sources. He presumed these to be unofficial readings, rejected in general by the authorities, and he attributed their great importance to the fact that their mode of transmission gives every guarantee that they belong to the time of the Prophet, his contemporaries, or the succeeding generations. Thus he arrives at the conclusion that these readings belong to the language actually spoken in the seventh century in Mecca. He describes this language in his book. The last paragraph is devoted to the Icrab, the vocalic endings of Arabic words according to the rules of the classical language. He is convinced that, besides this official language, a more simple language existed in Mecca at that time that was more in accordance with Arabic dialects and with other Semitic languages. A chief characteristic of that language was that the I'rab was not generally observed in it. It might have been used by certain tribes of the Bedouins, but even there it was only the privilege of the higher language, in the main restricted to metrical poetry.
From this statement Vollers draws the conclusion that the language of the Koran had been transformed on the whole in accordance with the poetry of the Nejd Bedouins. This transformed language had been victorious and had destroyed or driven aside the local and individual shape of the original language. For more than twelve hundred years this language had been regarded as the original, genuine, and undisputed language of the Holy Book.
It is well known that serious objections were raised to the whole theory of Vollers, especially by an authority like Theodor Noldeke. He pointed out several misunderstandings and shortcomings in Vollers's deductions. The material with which he argues is really very insufficient. Noldeke's chief objection against the theory was: if the Prophet and his faithful companions had read the Koran without hrab, the tradition of it would not have been lost without a trace.
Traditions that the Koran was read sometimes without I'rab have not been generally known so far. Vollers refers to two traditions reported by as-Suyuti (d. 911 A.H.), but he correctly remarks that, for so late an authority as as-Suyuti, it was unthinkable that the Koran should ever have been read without I`rab. As-Suyuti tries to understand these traditions as referring to stylistic and rhetorical elegancies.
When Gotthelf Bergstrasser was writing the new edition of the third part of Noldeke's Geschichte des Qorans,3 he passed over Vollers's arguments in silence. For him the chief source of information on the reading of the Koran is the introduction to the book on the Ten Readers composed by Ibn al-Jazari (d. 833 A.H.) under the title An-nashr fi'l-kira'at al-'ashr. This introduction is of value for him especially from the historical point of view, because it is rich in quotations from earlier books on this matter. Nobody will deny the great value of that book. But I think that Bergstrasser did not reckon with the possibility that in the historical retrospect of such a leading handbook only such problems were reported and discussed as were not in conflict with the generally accepted views of later times. Bergstrasser is not aware that the sources used by him do not help us to understand the real conditions in the beginnings of Islam. He discussed the two conditions that had been stipulated by Ibn Miksam (d. 354 A.H.)-correctness of the language and conformity with the text of `Othman-and declares that the first condition was of no great importance. He tries to explain it by referring to the fact that a great number of clients (mawali) who entered Islam often caused incorrectness in reading the Koran, but they were of no importance, as too many genuine Arabs participated in these readings who had a strong feeling for the correct language.
This is hardly right for the beginning of Islam. As far as we know, one of the reproaches against Ibn Miksam was that points and vowels were added by him to the `Othmanic text on the only condition that the readings were justified by the rules of Arabic grammar. According to Ibn Mujahid, his opponent, every reader had to conform to one of the Seven Readers. When in 322 A.H. Ibn Miksam's practice was condemned by a court of lawyers and readers, before which he had to appear, summoned by the Sultan, he had to recant and his books were burned, and we know of them only from occasional quotations. If these books had been preserved, we might find in them several things that could help us to understand the conditions of earlier times. The same may be said about many books now lost to us.
Even Otto Pretzl did not see clearly the problems which exist here. In his learned article "Die Wissenschaft der Koranlesung,"4 he describes a number of texts dealing with the reading of the Koran. Many of these books were hitherto unknown. Pretzl refers to the fact that books on the reading of the Koran are generally of much later origin than books dealing with more profane matters, such as, for instance, grammar, and he suggests that books devoted to religious problems were submitted to a sharper scrutiny. The fact that on the whole only late commentaries or comparatively recent books were available to us was the reason that no special attention was devoted to these studies in the West. He hopes that on the basis of the material made available by him, new results for the history of the language and the religion may be obtained by these studies.
When he himself, however, would review these problems, he confined himself to reproducing the ideas of ad-Dani, one of the chief authorities for the later orthodox practice. But in the books composed by that authority we rarely find any new light shed on the early period of Islam. The reason why earlier books have disappeared may be that the problems discussed in them were uncongenial to later times. The Muslims themselves hesitated to propagate books of that kind.
Nevertheless, the tradition that the Koran was read sometime without I'rab was not altogether lost without trace, as Noldeke supposed. I have already referred to the two texts that I have found among the manuscripts of Mr. Chester Beatty. With the text of al-Farra', published in my Schweich Lectures on the Cairo Geniza, I will deal later on. First I have something to say on the other text. It is the book At-tamhidft i ma'rifat at-tajwid, composed by al-Malikl (d. 438/1046), where these problems are discussed in full. The first part of this book contains sayings exhorting the people to read the Koran in a dignified and beautiful manner. It is of a more general character, interesting too, but not of special importance. The second part contains ten chapters of which the long chapter 6 deals with the problems discussed here. It has the heading "On Urging the Readers to Make Effort of Seeking the I'rab" and contains more than 120 exhortations admonishing people to use the I'rab in reading the Koran. The first thirty-one traditions are attributed to the Prophet, the next thirty-six (nos. 32-67) to his companions, the rest (nos. 68-122) to the followers and to the followers of followers. The most important of these traditions are those attributed to the Prophet and his companions. I have published a translation of these sixtyseven exhortations in my contribution to the Goldziher memorial volume. The following specimens may give an impression of the kind of exhortations contained in these traditions.
(1) The Prophet said: Whoever recites the Koran and reads with I'rab the whole of it, shall have for every letter forty recompenses. Whoever reads with I'rab a part of it and uses lahn in a part, shall have for every letter twenty recompenses. Whoever does not read anything with Drab, shall have for every letter ten recompenses.
(2) Whoever learns the Koran and reads it with I`rab, shall have the reward of the veracious (siddik), the martyr (shahid).
(3) Whoever recites the Koran and does not read it with Drab, for him God has appointed an angel who writes it down for him (correctly) as it was revealed, and he has for every letter ten recompenses. When he reads a part of it with Drab and a part without Drab, for him God has appointed two angels who write for every letter twenty recompenses. And when he reads the whole with Drab, God has appointed for him four angels who write down for him for every letter seventy recompenses.
(6) Whoever recites the Koran according to any method, for him God has written ten recompenses and canceled ten evil acts and elevated ten degrees. Whoever reads a part with I'rab and a part with lahn, for him twenty recompenses are written, twenty evil acts are canceled, and twenty degrees are elevated. Whoever recites the whole of it with I'rab, for him forty recompenses are written, forty evil acts are canceled, and forty degrees are elevated.
(7) Whoever recites the Koran and reads it with Drab, he has a favor with God which is granted to him, if he desires he hastens it for him in this world, if he desires, he reserves it for him in the other world.
(8) Whoever learns the Koran with I'rab, he is like the warrior on the path of Allah (mujahid fi sabili-llah).
(30) Behold, some of the poetry is wisdom, and if anything from the Koran is dubious to you, look for it in the poetry, for it is the Diwan of the Arabs.
(32) Abu Bekr: I would prefer to read with Drab a verse from the Koran than memorize a verse.
(46) `Abdallah b. Masud: Read with I'rab the Koran, for it is Arabic. There will come a people who would like to correct it, but they are not the best of you.
(67) 'Aisha from her father: Learn the poetry, for it brings the I'rdb onto your tongues.
These few specimens-ten of more than 120-may be considered sufficient. The rich material preserved by al-Malik-1 shows clearly that correctness of language in reading the Koran was of real importance. Offenses against the l'rab, the vocalic endings of grammatical forms, were much more frequent than scholars like Bergstrasser and Pretzl were willing to admit. Professor Johannes Fuck, with whom I had a long correspondence on this problem, pointed out that a great number of these traditions collected by al-Maliki could have been known before. Of the sixty-seven exhortations translated by me, he has found nearly forty in the great Thesaurus of traditions collected by al-Muttaki al-Hindi (d. 975/1567) in his Kanz al `ummal, published in eight volumes in Hyderabad in 1312-14 /1894-96, half a century ago. It is true that in this vast collection these traditions are distributed among different chapters and are not easily found. Al-Maliki brought them together in a very convenient and impressive way on about fifty pages of his book.
Have we any possibility of dating the traditions collected by al- Mdliki? From the fact that they were attributed to the Prophet and his companions we cannot necessarily conclude that they really go back to these authorities. Such texts are, however, of great value for studying the ideas of the people of that time. As al-Maliki composed his book at a comparatively late time (400/1010), it is very likely that not all these traditions are from the same time and origin. Fuck pointed out that none of these traditions are mentioned by al-Bukhari or Muslim or the other canonical collectors of Hadith, and he concludes that these authorities may have had the impression that these traditions did not belong to the first century A.H. That is really not very likely, as, for instance, the grammatical terms and other items occurring in some of these traditions reflect problems in which people were interested from the second century onward. On the other hand, traditions of that kind cannot be later than the second century, as al-Farra', who was living in the second half of the century, knows of them.
The traditions collected by al-Maliki show, therefore, that, let us say, in the second century the classical Arabic language was not generally used for reading Koran-that it certainly did not dominate in such a way as was supposed by Bergstrasser, for whom, as we have seen, correctness of language, one of the conditions stipulated by Ibn Miksam, seemed to be of no importance. What were the conditions in the time before?
There are two alternatives. First, the Prophet and his companions read the Koran according to the rules of classical Arabic. In this case we have to suppose that the language spoken by the Kuraish in Mecca, to whom the Prophet belonged, was in the main identical with it. The incorrectness in reading the Koran that existed in the second century was caused by the Mawali, the non-Arabic clients who had entered Islam in great numbers. That was the supposition of Bergstasser, who, however, did not realize that the incorrectness in reading the Koran had spread in such a way that in about 400 an author like al-Maliki was able to collect more than 120 exhortations to read the Koran correctly. The other alternative is that, in the language spoken by the Kuraish in Mecca, the rules of classical Arabic were not observed; that the Prophet and his companions did not use classical Arabic in reading the Koran, the language now connected with the Holy Book. Then we have to suppose that this model language was worked out on the basis of Bedouin poetry by the early readers and grammarians. To this model language the text of the Koran was adapted by a system of vowels and signs added to the consonantal text. This adaptation was finished in the second century. By that time people had to be admonished to use the I'rab and to observe all the rules of classical Arabic in reading the Koran that had not been observed before. That is in the main the theory of Karl Vollers.
I think that the text of al-Farra' published and discussed by me in the Schweich Lectures on the Cairo Geniza brings a clear solution of this problem. Al-Farra' says:
We have seen that the readers who know the book and the practice and are authorities on correct speech are agreed that it came down in the most correct forms of speech. This was opposed by some of those who investigated the poetry and "the days of the Arabs." They said: Those who claimed the excellence of the Koran have merely done so in accordance with what God made obligatory for honoring the Koran but when we look for correctness of speech, we find it among the Bedouins.
But in this they disagreed. The people of Kufa said: Correctness is to be found among the Asad, because of their vicinity to them. The people of Basra said: Correctness is to be found among the upper Tamim and the lower Kais from 'Ukl and 'Ukail. The people of Medina said: Correctness is to be found among the Ghatafan, because they are their neighbours. The people of Mecca said: Correctness is to be found among Kinana b. Sa`d b. Bekr and Thakif.
We wished to refer them through tradition, analogy and example to the superiority of the speech of the Kuraish over all other languages. So we said: Do not the Kuraish surpass the people in the beauty of their statures, in the sagacity of their minds, in the fullness of their bodies? They answered: We know this as well as anyone. But sagacity and beauty came to them merely because the Arabs were accustomed to come to the sanctuary for Hajj and `Umra, both their women and their men. The women made the circuit round the House unveiled and performed the ceremonies with uncovered faces. So they selected them by sight and thought after dignity and beauty. By this they gained superiority besides those qualities by which they were particularly distinguished. We said: In the same way they were accustomed to hear from the tribes of the Arabs their dialects; so they could choose from every dialect that which was the best in it. So their speech became elegant and nothing of the more vulgar forms of speech was mixed up with it....
Al-Farra' refers then to the fact that in the language of the Kuraish, which he takes as identical with the official language of the Koran, certain irregularities are not to be found that occur in the language spoken by the Bedouins, and he continues: "Correctness came to them from their selection of pronunciation just as they selected their wives. And by this we refuted their arguments and reverted to the arguments of those who know the Koran better than they."
Al-Farra' then quotes a number of traditions dealing with the correct reading of the Koran and exhorting people to use the I'rdb in reading it, quite similar to those collected by al-Maliki. Some of al-Malik-i's traditions are quoted by al-Farra' also.
Al-Farra' admits that the statement that the Koran came down in the most correct form of speech was a dogma that had validity only for a convinced Muslim. There was no doubt that correctness of language was to be found among the Bedouins. From the words of al-Farra' we learn that in the chief centers of early Islam, the early readers of the Koran went out to the Bedouins in their neighborhood in order to study their poetry. So far this poetry had been orally transmitted only. The early readers needed this poetry as a model for the language in which the Koran had to be read. The practical need of readers and grammarians seems to have given the impetus for collecting and recording pre-Islamic poetry at the beginning of Islam. On the basis of this material a model language was established. One of the characteristics of this model language was that the rules of I'rdb were carefully observed. When the Muslims were admonished again and again to use the I'rab in reading the Koran, when they were told that reading the Koran with Drab was much more valuable than reading it without Drab, we have to draw the conclusion that the people admonished in this way had not been accustomed to read the Drab in their language. "Learn the poetry, for it brings the Drab onto your tongues," `Aisha is reported to have said.
Al-Farra' was in a difficult position. As a grammarian he could not deny that correct Arabic was to be found in Bedouin poetry. As a good Muslim he had to believe that the word of God had been revealed to the Prophet in the most correct language. As a theologian he was not allowed to admit any alterations in the language of the Holy Book. So he had to find a compromise. He found it by declaring that the influence of Bedouin language on the language in Mecca had taken place long before the time of the Prophet. From the different tribes of the Arabs who had come on pilgrimage to Mecca, the Kuraish had been able to hear all kinds of Arabic. So they had been able to select from the different forms of Arabic the best of each, just as they had selected their wives. In this manner their language had become superior to all the languages spoken by Arabs, superior also to the language spoken by the Bedouins, since certain inaccuracies that are to be found in the language spoken by the Bedouins had not been taken over by the Kuraish and were not to be found in the language spoken by the people of Mecca. In such a way the model Arabic that was used for reading the Koran from the end of the first century onward was identified by al-Farrel' with the language spoken by the Kuraish in the time of the Prophet.
But the real conditions can easily be recognized. Al-Farra' would certainly not have reported that the early readers went out to the Bedouins in order to study their language, when they had not really done so. Why did they go? They went out in order to study correct Arabic. This correct Arabic was found among the Bedouins. If the people of Mecca were accustomed to hearing correct Arabic spoken at home by the Kuraish, there was no need for them to go to the Bedouins. What could they have learned from the Bedouins under these circumstances?
The traditions preserved in these two manuscripts are really of importance. They not only show something of the real conditions of reading the Koran in the early times of Islam, they also give us an impression to what degree texts on reading the Koran have been corrected, revised, and purged. If neither Noldeke nor Pretzl nor Bergstrasser had been able to discover hints of these problems in all the texts available to them, it does not prove that these problems did not exist. We can only conclude, on the evidence of these two manuscripts, that the texts available to them had been brought into accordance with the later adopted practice.
1. Paul E. Kahle, "The Cairo Geniza," in The Schweich Lectures of the British Academy, 1941 (London, 1947).
2. Carl Vollers, Volkssprache and Schriftsprache im alten Arabien (Strassburg, 1906).
3. Theodor Noldeke, Geschichte des Qorans, 2d ed., Vol. 3: Die Geschichte des Qoranstexts (Leipzig, 1938).
4. Islamica 6 (1934): 1-47, 230-46, 290-331.