The Absence of Diacritics in the Koran:
Rough Drafts of Preachers?

It was long thought that Arabic writings, until the 8th century, lacked diacritical dotting. This clearly is an error, since without these marks above or below certain letters, the ambiguities would have cancelled the possibility of conveying any meaning in writing. The problem is that ancient copies of the Koran are all devoid of diacritics; since no reason could be given (out of ignorance of the true origins of Islam) and since Arab identity was equated with Islamic identity, it was believed that early Arabic writing per se used no diacritics.

    Diacritical marks (or points, or signs) help distinguish between many consonants – by comparison, it is rarely the case in Hebrew or Aramean. When those marks are absent, the meaning is unclear : « t » for instance can be read as « b », « y », « th /t » or « n ». Vowels appeared at a later date, but their absence does not hinder comprehension: at the worst they can sometimes cause some mild ambiguity (for instance when the context does indicates whether the verb is in the active or passive voice).

The hypnotic power of (legendary) presuppositions

These presuppositions about diacritics went along with the Islamic narrative of the « revelation ». It was assumed that the declamations of the « prophet » Muhammad had be jotted on flimsy mediums like camels’ shoulder blades, stones, etc., never on normal mediums like parchment or papyrus. Given this unusual background, the hypothesis was that diacritics had been invented to remedy the declining quality of oral transmitters of the Koran, which in turn explained the existence of variants (Islamic tradition speaks of « seven readings » without much more precision).

Of course, we are looking at a string of myths. The material origin of the pieces of the Koranic text is much more reminiscent of a compilation of writings from different periods (on papyrus or parchment) than of a continuous proclamation heard by scribes and set in writing (as the Islamic legendology would have it); there never existed reciters of the Koran before the first compilation of it (and as we know, memory is a very reliable channel in oral cultures); finally, before we venture to explain the apparition of diacritical signs, we should ask ourselves why they were missing in the first place, and whether the task of adding diacritics wasn’t delayed because the literal content was simply unclear (at least for the proto-muslims) [1].

This would go to explain why a text so poorly written technically was copied verbatim for three centuries. For, as François Déroche, a specialist of Arab manuscripts at the National Library of France, remarked, many copyists persisted in not placing diacritical signs, or for the others did so in an unregulated manner:
«In fact, the comparison between different fragments, even different hands – when several copyist joined efforts to transcribe the Koran – highlights the extremely individual character of [diacritical] punctuation; each one placed signs wherever they saw fit” (p.23
[2]).

We must then go against the grain of conventional Islamic thinking, whose main goal was to captivate both the simple and the learned by masking the veritable origin of a movement that would be called « Islâm » only after 700.

At any rate, the proto-Muslims faced not memorization-related, but text-related problems; the absence of diacritics but also of familiarity with the leaflets from which they wanted to extract a Koran throws some light on the obscurites that riddle the Koranic text [3]. Part of these come without a doubt to wrong attributions of diacritical signs, something that has been the object of studies by Christoph Luxenberg and Munther Younès[4]. The rest are due to the successive manipulations the text underwent throughout its first compilations and also posteriorly (those manipulations were generally followed by the massive destruction of « non-compliant Korans »).

True and false examples of pre-Islamic writing with diacritics

At this point, let us mention that the Internet furnishes several examples of pre-Koranic Arabic writings clearing attesting the use of diacritical points. Not all are trustworhty, like this one found recently and dated from 644, which mentions Zubayr and ‘Umar, and happens to be a fake intended to give credence to the idea of the region of Mecca being the craddle of Islam – particularly in order to have people believe that the type of writing of the most ancient copies of the Koran has its origin in that region (some have dubbed it hijâzi i.e. from Hijâz): on the contrary, studies show that Koranic writing, like the Koranic language, existed at that time only in Arabia Petraea, that is Syria – and along these lines, we recommend that you read this article by two eminent Islamologists respectively from Canada and Netherland, Robert Kerr and Tomas Milo.

By contrast, the following document posted here on the Internet (Perf 558) attests the authentic use of diacritics; it can be seen in Vienna, at the National Museum of Austria. The manuscript is a bilingual Greek Arabic manuscript dating back to the year 642. This manuscript shows diacritical signs on the letters. It is one of the oldest Arabic manuscripts (the other one is the P. Berol. 15002 – as far as early engravings, there are the bilingual ones found on the lintels of the churches of Zabad and Harran in Syria). What is more, this document confirms the title of muhajirun (those who completed the Hegira – transposed in Greek as magaritaï ) by which the proto-Muslims were designated (until about 720).

The Koranic leaflets, rough drafts of different preaching notes

The question of the use of diacritics before the composition of the Koran being now settled, the next step is to establish why the leaflets gathered to form an « Arabic Koran » did not have any.
We are faced with an alternative:
either the Koranic leaflets came from « God », in which case we must assume that they were set to writing in a terribly negligent way since the scribes apparenty forgot to add diacritics;
or these leaflets emanate from one or several authors, who simply transcribed what they intended to say, in the form of rought drafts of preaching notes or memory aids destined to be reworked and therefore crudely written. The question is then: who was this author, or who were they?

This question must be approched as a whole. The author of surah 39 (s.39:27) complains that the Arabs do not try harder to memorize; and he is not talking about his own sermons, but about the lectionary (this is what qur’ân means) translated into Arabic, as we can easily grasp when comparing all the genuine or added occurrences of the term. This translated lectionary (qor’ôno in Syriac) must necessarily have been the same as the one used by the preachers who were trying to convert their Arab neighbors. A systematic investigation (as this video montage – in french) quickly reveals that those preachers were the « nasârâ » of the original Koranic text, who are mentionned five times legitimately (the ten others mentions are later additions, probably dating to the first manipulations of the text).

In fact, there can be only one explanation for the defective state of the ancient copies of the Koran: those were copies of rough drafts of preaching notes left behind by Nazarene propagandists wishing to win the Arabs to their Cause.

This perspective also provides the most plausible explanation for the altered name of Jesus in the Koran, not spelled as one could expect in Arabic (Yasû‘) but rather as Christian Arabs must have pronounced it in their popular language (‘Ysā). Every project to preach them implied an adaptation in their language; but, since the native tongue of the Nazarene propagandists wasn’t Arabic but Aramaic,  they needed to prepare drafts before preaching: those constitute the different leaflets from which the Koran of the Caliphs was assembled (not the Koran referenced inside the Koran itself, which is nothing else than the translated lectionary).

There is beside a complete coherence with what we know of the activity of ‛Uthman, who wanted to have a text to oppose the Jews and the Christians, and had to choose and « collect » it from various sources. It would be absurb to imagine that there were scattered or disseminated notes dictated by someone who passed for a Messenger of God, as this person’s words would have been carefully preserved, in writing (that isn’t the case). It is therefore easy to reconstruct what happened: as ‛Uthman was pressed for time, a book was hastily assembled, and in fact several competing versions emerged at once, as Islamic tradition testifies; the only selection criteria for those leaflets written in Arabic and left behind by the Nazarenes was that the content might support, to some extent, the new orientation of the message: to magnify God’s election of the Arab nation and to form a corpus of texts able to challenge the book of the Jews and the Christians.

Finally, the Shiite tradition attests very clearly that the so-called « original » Koran was three time more voluminous than the versions eventually imposed by the Caliphs (Sunnis) of Damascus [5]. This affirmation could only make sense if the Koran as we know it already existed, but it did not. It does point to a veritable mass of rough drafts and notes left by Nazarene propagandists in Medina and Syria, after the Arab followers had turned against them (about 640): it is easy to understand that the strenuous efforts to compose the actual Koran implied the availability of large pool of documents already written, from which to pick.

As one can expect, the successive proto-Korans resulting from those hesitant compositions only partially met the expectations of ‘Uthman and, later, of the Caliphs of Damascus; moreover, their utilization raised contradictions from Jewish and Christian critics (who as a consequence were forbidden to read the Koran!). Obviously, to reinterpret a text in a new way never goes without creating some difficulties. Hence the need to operate further manipulations, and at times the pure elimination of outdated versions of the « Koran » all the way to the 8th century – something that Islamic traditions document. The political authority was responsible for those manipulations, under pain of death for those who oppose them. This explains why the copies of the successive Korans long remained without diacritical signs : their placement was retarded not just by the uncertain meaning of the texts, but by the question of the meaning to give them: it had to obscure the origin of the first leaflets and at the same time to sacralize the Arab rulers installed in Damascus. Faced with such challenges, the copyists could only grope around and use caution – often by not adding any diacritic signs. 

The rough aspect of primitive Koranic leaflets is therefore one of the many elements which, in all logic, allow us to penetrate the origin of the Koran in the proto-Islamic context, involving a short-lived Arab branch of the old Nazarene movement, before it won its autonomy (by turning against its Judeo-Nazarene instigators after 639 – marking the end of the proto-Islamic period and the beginning of the primo-Islamic one).

____________
Home Page 


[1] According to Christoph Luxenberg (Relikte syro-aramäischer Buchstaben in frühen Koran- Kodizes, in Der frühe Islam. Eine historisch-kritische Rekonstruction anhand zeitgenössischher Quellen, Berlin, 2007, p.377-414), an added difficulty of interpretation (for Arabs) was the existence of leaflets written in Garšuni (or Karshuni). i.e. in Arabic with the Syro-Aramean alphabet (known as the Syriac alphabet). It is possible that the majority of the primitive leaflets of the Qur’an were in that alphabet. It is a fact that most ancient manuscripts in Arabic – particularly all the books of the Bible, for example the Arabic ms L – are written in this way, and the Arabic alphabet began to be predominant only during the reign of the Caliph ‛Abd al-Malik (end of 7th century). Certain words that are incomprehensible in today’s Koranic text could be explained by an erroneous reading of the Garšuni (especially the confusion between two letters that are similar in Syriac but not in Arabic, whose correction sends the reader back to a word from one of those languages). For instance, in Aramean script, is easily confused with l.

     Another example (p.394-395): without diacritics (and vowels), the three consonants  may be read three ways: • as in Yuhannan (the first name John in Aramean and in Christian Arabic – without any diacritics, the final letter y bears a resemblance to the final letter n ; and the double consonnant [n] is itself a later diacritic sign); • as in Yahya (the name of John according to the Koranic text); • and tahana, hardening, which makes no sense in the context. The “Quranists” (those who “accomodated” the Koranic text) wished to prove that the Arab Christians and their gospels were mistaken with the name “John” (and its Semitic meaning: “God gave grace”); therefore, they invented the vowels imposing the reading Yahya (that has a meaning in Arabic: “he is living”). This conclusion not only is perfectly credible, it is also the only conceivable one.

     In the first Korans, the graphemes in use led to confusions between a final n and a final y ; interestingly, Luxenberg pointed out that this final y – at least what appears to be a final y – strongly looks like an Aramean n occurring at the end of a word (for example in the BNF 328a conserved in Paris).

[2] DEROCHE François, Beauté et efficacité : l’écriture arabe au service de la révélation [« r » with a lower case»] in KROPP Manfred ed., Results of contemporary research on the Qur’ân. The question of a historical-critical text of the Qur’ân, Orient-Institut Beirut/ Würzburg, Ergon Verlag, 2007. These are the partial acts of a Conference held at the University of Mayence on September 8-13, 2002.

[3] As far as the content, the difficulties are even more numerous. For example, can “God” swear by something less than Himself, “by Mount Tûr” or “by Mount Sinai”, instead of swearing “by Myself” as He does in the Bible?

[4] He established that verse 3 at beginning of Surah 100 was an addition. Cf. YOUNES Munther, Charging Steeds or Maidens Doing Good Deeds? A Re-Interpretation of Qur’ân 100 (al-‘âdiyât), in Arabica 55, 2008, p.362-386.

[5] AMIR-MOEZZI Mohammad Ali (dir.), Dictionnaire du Coran, Paris, Robert Laffont, 2007, p.36.