The plan of Bergstrasser, Jeffery, and later Pretzl to prepare a critical edition of the Qur'an was not realized, and the collection of variants derived from real old codices failed to survive the bombs of World War 11.1 Many more old manuscripts are accessible now, which would justify a new approach, but no such undertaking is in sight. It is true, unfortunately, that the (scriptural) variants are hardly helpful for a better understanding of much of the text which is still far from being as mubin ("clear") as the Qur'an claims to be! Thus, even if a complete collection of variants could be achieved, it will probably not lead to a breakthrough in Qur'anic studies. Certainly, though, it will help to reveal the stages of Qur'anic (and Arabic) orthography.
An exciting "excavation" of old Qur'anic fragments took place in the Yemeni capital of Sawa, from 1980 onward.2 The fragments were discovered in 1972 in the loft of the Great Mosque. Subsequently the (then) General Authortiy for Antiquities and Libraries took care of them in the Dar al-Makhtatat. Meanwhile, the many thousand pieces of parchment have been cleaned and identified according to Mushaf, Sarah, and Ayah; at this stage a complete microfilm documentation is needed in order to make the fragments available for study and for the preparation of a catalog. Unfortunately, the priorities of neither the German sponsor of the restoration project (Ministry of Foreign Affairs) nor of the Yemeni antiq uities administration seem to favor the idea. Hopefully better times will come. Since no complete microfilm documentation is available,3 the details presented below are left without exact reference to the manuscripts from which they are taken.
Among the fragments of roughly nine hundred different parchment Mushafs, about 10 percent are written in a peculiar "pre-Kufic" variety of script, Hijazi or Ma'il. It is this group of manuscripts that was examined in a preliminary way in order to prepare a questionnaire for a more comprehensive investigation. Examples of these observations are recorded here according to the type of deviation from the Rasm of the Egyptian standard edition of the Qur'an. My observations do not claim to be either new or unexpected,4 except for the last paragraph, which discusses the different arrangements of the Surahs.
(1) Defective writing of the Alif constitutes the most common "deviation" from the Rasm of the printed standard edition. The scriptural appearance of the following examples presupposes an established oral tradition of correct reading, much more than the familiar Rasm-which has the plene Alif-does:
The standard Rasm (~6( is easily recognized as 'aba!ukum, whereas ~y( in the Hijazi manuscripts, again, requires the oral tradition for the same pronunciation! In cases like v" bi-l-haqqi or ulyJ~ ka-l jawabi the Alif of the article is written defectively, but there can be no doubt about the correct reading. Evidently, no orthographic convention was connected with the Alif al-wigayah, either (-)(j instead of i'L ra'aw). On the other hand, does L,. ka-lladhi imply that the second letter Ya' should not be pronounced at all?
(2) If it is true that the defective writing of the Alif is more archaic than the plene version, then the same is true for those cases where the Alif is written in lieu of (Semitic) Hamzah: u~- for shay'in (as if sha'yin was intended), .j" for swat, and even L..JI for as-sayyi'a.
(3) Most of the canonical "readings" (Qira'at) of the Qur'anic text do not presuppose a different Rasm; but although the proportion of the cases that deviate from the standard Rasm is relatively low, it is amazing how many of these cases of deviations-in absolute figures-are reported! We can now easily check any variation of the Rasm with the accumulated tradition of Muslim scholarship on the Qira'at, thanks to the eight-volume dictionary Mu jam al-gira'at al-qur'aniyyah.5 By doing this we discovered that `our' manuscripts contain many more Qira'at than are recorded by the old authorities. These examples may suffice: In XIX.62 the original y la tasmac was later corrected to la tasma`una (instead of the usual la yasma`una). Instead of qul ja'a l-hagqu in XXXIV.49 we find l L,.,P gala ja'a l-hagqu. The systems of the seven, ten, or fourteen Qira'at are, consequently, younger than the variants observed in Sanla'.
4. The same is true for the variants in counting the verses. Even in the most archaic manuscripts, the end of a meaningful portion of the text is marked by dots, strokes, Alifs, or similar signs. Many of the separators in the Yemeni manuscripts are placed in positions, however, that are not counted as the "end of a verse" according to the "Kufan" counting. The Islamic tradition is aware of different regional counting systems, of which Spitaler has compiled a condensed, easy-to-handle survey;6 all together, twenty-one systems are distinguishable according to his sources. The Sanlani early manuscripts in question seem to favour the "Kufan" counting, but in a substantial number of manuscripts we find no thorough correspondence with any of the other traditional systems. If we compare, for example, the verse separators/verse counting realized in one archaic manuscript (no. 00-25.1) with the traditional systems of verse counting, the ratio of identical ( + ) or diverging (-) countings can be summed up as follows:
Basra 10 + /5 - , Kufa 4 + /11 - , Makka 11 + /4 - , Madina 11 + /4 - .
It is noteworthy that in some of the Hijazi manuscripts the Basmalas at the outset of the Surahs are always marked by a verse separator. Would these manuscripts reflect the opinion that the Basmalas are primordial parts of the Qur'anic text?
In general the number of separators seems to exceed the number of verses counted, which is clear from contradictory use of separators and markers for groups of five or ten verses. Separators are observed even at places where the Egyptian standard edition has the recitation mark () (al-waslu awla, "enjambement is preferable")!
(5) Two early Qur'an authorities are reported to have kept their "private" Qur'an manuscripts, which they refused to destroy or harmonize with the official version promoted by the caliph 'Uthman: Ibn Masud and Ubayy b. Kalb. Lists of the different arrangements of the Surahs in their respective Mushafs have been preserved,7 but until now no such differing arrangement has been traced in a manuscript. The implications of the "validity of these reports" are far-reaching and apt to shed some light on the question of what the Qur'an looked like at the time of, say, the "Righteous Caliphs." "But if most of the suras were written down and put into approximately their final form during Muhammad's lifetime, then there would be no strong reason for rejecting the validity of these reports [i.e., on different arrangements) outright," A. T. Welch8 connects the two issues of the arrangement and the time of the Qur'an's composition. Now, since we do have examples of different arrangements in San`a'-are we allowed to invert Welch's argument, concluding from their existence that most of the Surahs were not written down and put into approximately mately their final form during Muhammad's lifetime?-The San`ani specimens are, however, not only proofs for their existence, but allow for the hypothesis that even more arrangements were in use which differed from the official sequence as well as from those reported to go back to the two authorities Ibn Masud and Ubayy:
In one case, the end of Sarah XXVI is followed by the beginning of Sarah XXXVII (on the same page, of course!), which corresponds exactly with the leap reported about Ibn Mascud's arrangement-while Ubayy's Mushaf is said to have lept from Sarah XXVII to XXXVII. Two other leaps observed, namely, from Sarah XIX to XXII and XXXVI to XXXVIII are close to the Ubayy list (who has the sequences 11-19-26-22 and 27-37-38-36), while the leap LXVII to LXXI is somewhat closer to Ibn Mascud's codex (49-67-64-63-62-61-72-71-58), again. Finally, there are the leaps LXXII to LI and LXVII to LXXXIII. which are not even remotely reflected in one of the lists. The last three "leaps" are, of course, not of the same importance as the preceding ones, as they are situated in the higher numbers of Surahs where the placement is rather arbitrary and not as easily determined as with the, say, first fifty Surahs.
NOTES
1. Information about the ambitious project can be gathered from scattered sources, like Arthur Jeffery, Materials for the History of the Text of the Qur'an (Leiden: Brill, 1937), pp. vii, 3-4 (esp. n. 6); Otto Pretzl in Geschichte des Qorans. Dritter Teil: Die Geschichte des Korantexts, ed. Theodor Noldeke, G. Bergstrasser, and O. Pretzl (Leipzig, 1938; reprint, Hildesheim: Olms, 1981), pp. 249-251, 274; Anton Spitaler, "Otto Pretzl, 20. April 1893-28. Oktober 1941. Ein Nachruf," ZDMG 96 (1942): 161-170; A. Fischer, "Grammatisch schwierige Schwur- and Beschworungsformeln des klassischen Arabisch," Der Islam 28 (1948): 5-6 n 4; Arthur Jeffery, The Qur'an as Scripture (New York, 1952), p. 103.
More recently, Angelika Neuwirth in her GAP article "Koran" (p. 112, see In. 3) has given the impression that it was the photographs taken in order to build up the "Koran-Archiv" in Munich which were destroyed at the end of World War II. This impression is false, and thus it is an amazing fact that evidently no attempt has been made since to study the photographs!
2. Under the supervision of Albrecht Noth, Hamburg; the present writer was in charge of the scholarly as well as practical organization of the project from 1981 until 1985, when he was succeeded by his collegue H.-C. v. Bothmer for another two years.
3. Meanwhile, microfilms have been made for the Dar al-Makhtutat in SanIa', and one copy is with my colleague, Dr. H.-C. Graf v. Bothmer, Saar- briicken.
4. Cf. especially Noldeke, Bergstrasser, and Pretzl, Geschichte des Qorans, Werner Diem, "Untersuchungen zur fruhen Geschichte der arabischen Orthographie," Orientalia. Roma. 48 (1979): 207-257; 49 (1980): 67-106; 50 (1981): 332-383; 52 (1983): 357-404.
For a detailed bibliography see Angelika Neuwirth, "Koran," in Grundrif3 der Arabischen Philologie. Band II: Literaturwissenschaft, hrsg. von Helmut Gatje (Wiesbaden: Reichert, 1987), pp. 98-135, passim.
5. 'Abd al-<="" p="">
6. Anton Spitaler, Die Verszahlung des Koran. Munchen 1935 (Sitzungs berichte der Bayer. Akad. d. Wissenschaften. Philos.-histor. Abt., Jg. 1935, Heft 11).
7. Jeffery, Materials, pp. 20-24; Hans Bauer, "Ober die Anordnung der Suren and uber die geheimnisvollen Buchstaben im Qoran," ZDMG 75 (1921): 1-20; Muhammad b. Ishaq an-Nadim, The Fihrist of al-Nadrm. A Tenth-Century Survey of Muslim Culture, trans. Bayard Dodge (New York, London: Columbia University Press, 1970), vol. 1, pp. 53-57, 58-61.
8. "al-Kur'an, " in E12 V 407 b.