page xvii

Leaves from two ancient MSS. of the Coran.

From the time that I obtained possession of this palimpsest, I was perfectly aware that at least four of its quires contained an Arabic under script. Several times I tried to identify this without success. I had a natural reluctance to take an old manuscript to pieces by cutting out the cord which held its several quires together; but without doing so I could not even see the inner margins, and there alone were lines of the ancient Arabic script to be found, perfectly free from the upper writing. Add to this the fact that the script was in Kufic, without diacritical points, and that I was trying to find a Christian text, and the reasons of my want of success are at once apparent.

I was just about to place these leaves in hands more skillful than my own; and for this purpose, on June 21st of this year [1901] I was taking a second


page xviii

quire to pieces, and painting up its margins with the reagent, when two lines of writing appeared which had been hitherto invisible. This encouraged me to make another attempt; and on comparing it with the facsimiles in the Semitic Series of the Paleographical Society's publications, I found that the script was exactly similar to that in Plate LIX. assigned to the 8th century. As that represented a portion of the Coran (Sura 44) it seemed possible that mine might do so likewise, and a few minutes later I had identified the first line on f. 150b with part of v. 57 in Sura 44 and on June 27th I found that the on line II of f. 20b is from Sura 16 v. 37

It was then evident that I had got seven leaves of a very ancient manuscript of the Coran, belonging to the first part of the 8th century, or perhaps even to the latter half of the 7th; also 15 1/2 leaves (forming 31 leaves of the palimpsest) from a manuscript only a little later. Coran I., as I have named it, is a flowing Kufic script, without the slightest sign of a diacritical point or of a vowel point. Each of the seven leaves has unfortunately been clipped on the one side in order to reduce its size to that of the Transitus manuscript, so that a whole word is missing at the end of some of its lines. In December I had the pleasure of placing f. 150 side by side with O.R. 2165, the MS. figured in Plate LIX. The resemblance in the handwriting, the size of the page, and the general appearance was so great that we at first suspected my leaves to be a portion of the same MS. But a closer inspection revealed the fact that there is a difference in the length of the final ya. Mr Ellis thinks, however, that they were both produced at the same time and place, if not by the same hand. Similar portions of Suras 24, 28, 29, 40 exist in both MSS.

Coran II. is quite legible (on the margins) without the reagent, as its script has hardly faded. The great difficulty in reading it was the closely written Arabic which lies across the top of it. It has no vowel points; and the only diacritical points which I have detected (partly by the reagent) are:

In f. 13b the ta in is written .

In the last line of f.15b there are either two dots or a horizontal stroke above the waw of , which have no connection with the line above it.


page xix

In f. 16a. there is a dot over the nun

In f. 17b. there is a dot under the ba of bis and two dots over the ta

f. 55b. In there is a dot on and two dots on

f. 57b. The of has two dots over it: thus

The following word is not for it has a long letter like or just before the and the is in a hole.

f. 58a. There is a dot on the and two on or in

f. 95a. In the nun and the ta have dots, the latter being written . Also the ta in of f. 102b.

f. 98a. The first ba in and that in have each a dot. Also in f. 97a. the ba's in and

f. 101a. The ta in is written

It has been suggested to me that other diacritical points may have once existed, in red or green paint, and that these may have been erased when the pages were palimpsested. But in the case of Coran II. this is an impossibility. The vellum is there so thin that every stroke or the stylus has left an indelible mark.

I have contented myself with printing the first and last lines of every page or half page; although in the case of Coran II. these are by no means all that can be deciphered. The whole text might have been edited with a liberal use of the reagent, but this would have been too great an infliction on the very fine vellum of Coran II.; and would have produced only a transient effect on the more solid one of Coran I. It would also have been a severe trial to my eyes and I cannot see that it would have served any useful purpose. I have added extra lines only in those cases where the terminal ones on a page were imperfect; and in those which show the end of one Sura and the beginning of another. Nothing occurs betwixt these except the words .

At the beginning of Sura 45, as it may be seen in Plate IV., we perceive, after these introductory words, a row of six small circles, each of which shows the remains of some red ornament filling up its interior, as it came up under the reagent. It has been impossible to reproduce this in the facsimile, but this is the less to be regretted as we cannot form the slightest idea as to what its pattern was.

I have printed these texts in the usual Coran script with full points.


page xx

Some of my readers might have preferred an unpointed text, but this would have given no true idea of the script. To do it justice would have required the employment of purely Kufic type; which would have appealed to the understanding of only a few. And I hope that these few may be satisfied with the Plates which have been executed by Messrs Annan and Son of Glasgow, from the photographs of Mr Edwin Wilson, Cambridge.

Whatever opinion may be formed as to the exact age of these fragments, there can be no doubt that they carry us back to a very early period in the history of Islam.

Mohammed died in A.D. 632, and we are told that as he received the revelations they wrote them down on any scrap of material that was available, on bones, white stones, ribs of palm leaves, or in the heart of men. Abu Bekr was the first who caused these scattered fragments to be collected, after the battle of Jemama in the 12th year of the Hegira A.D. 632, by Zaid Ben Thabit. Disputes arose as to the genuineness of some portions, and Khaliph Othman in 24-35 of the Hegira = A.D. 644-655 caused four copies of a normal edition, to be made out of the several copies and their variants. The dialect of the Koraites, which Mohammed spoke, was its basis and all MSS. which differed from it were destroyed. But the uncertainty of the writing, and the lack of diacritical and vowel points, caused fresh disputes. So it was decided, not without opposition, to add the long vowels, coloured, so as not to spoil the original form. The short vowels were next added, coloured, then other orthographical signs like the hamza invented by Khalid ben Ahmed 75 A.H. = A.D. 795.

These statements, which I have taken from the valuable book of Dr J. H. Möller, have an important bearing on the date of my palimpsest leaves. They narrow down the period at which these could have been written to a period between A.D. 655, when the chapters or the Coran fell into their present sequence, and that remarkable year (whichever it was) of the 8th century when diacritical points came into use. The seven lines of Coran I. (as we have already seen), show absolutely no trace of these. If we had not a printed text before us whilst deciphering it, we could not tell whether the sign is a ta, a tha a ba, a nun, or a ya; and of its other signs, there is hardly one which may not have two values. The absence of an occasional alif (though required by grammar) shows that it was written when the controversy


page xxi

about the use of long vowels was not quite closed and was designed rather as an aid to memory than as a "first reading book."

Several other questions arise. Were these two MSS. to which our leaves belong, amongst those whose destruction was ordered by Khaliph Othman, and do they owe their preservation to the cupidity of some faithful Moslem, who saw that they had a little value as writing material? I have been told by Dr. Hirschfeld that this supposition is negatived, in the case of Coran II., by the existence of the letters at the beginning of Sura 14. Or are they parts of an authorized copy, which fell into the hands of Christian soldiers after the capture of some town, and were by them handed over to the monks, who were almost the only "clergy" of the period? We suspect that the cases are very rare indeed where a Christian writing exists on the top of a Mohammedan one.


Source: Lewis, Agnes Smith, Apocrypha syriaca; the Protevangelium Jacobi and Transitus Mariae, with texts from the Septuagint, the Corân, the Peshitta, and from a Syriac hymm in a Syro-Arabic palimpsest of the fifth and other centuries; ed. and tr. by Agnes Smith Lewis. With an appendix of Palestinian Syriac texts from the Taylor-Schechter collection, London, C.J. Clay and sons, 1902.

The Text of the Qur'an
Answering Islam Home Page