"So well has it (the Qur'an) been preserved both in memory and in writing, that the Arabic text we have today is identical to the text as it was revealed to the Prophet. Not even a single letter has yielded to corruption during the passage of the centuries. And so it will remain forever, by the consent of Allah." (The Holy Qur'an, English Translation of the Meanings and Commentary, King Fahd Holy Qur'an Printing Complex, Preface, p. v)
On the other hand compare with the above the comments of Abul Ala Mawdudi in his "Introduction" to Yusuf Ali's popular translation of the Qur'an (The Holy Qur'an, Translation and Commentary, American Trust Publications, 1977, pp. xxxiv-xxxvii) in which he acknowledges and even indicates some variant readings of the Qur'an, and then relates how scholars were appointed during the reign of Abdul Malik (65-85 A.H.) "to assign new symbols for vowel points and dots to distinguish between the similar letters...."
Or even more let us consider a portion of Muhammad Hamidullah's introduction to his French translation of the Qur'an -- and he is hardly a radical textual critic of the Qur'an! -- in which he writes:
"... Finally, a third source of variants comes from the Arabic writing of the first times before diacritical marks came into general use: it is then sometimes possible to read a word as an active or passive verb, as masculine or feminine, and the context sometimes allows several possibilities. For example yas'al (God) will ask, can be read: yus'al (it) will be asked - tus'al (she) will be asked. A small number of cases have been found, but in none of these cases does the meaning of the verse change, and one wonders if the discovery of such variants does not sometimes come from the ingenuity of exegetes. Even Bukhari gives some examples of this: instead of of the vulgate text (see Qur'an 2/259. Bukhari 65, sura 2, ch./44); or instead of of the vulgate text 7/57 (see Bukhari 65, sura 7, ch. 1).
But there are cases, indeed very rare, which cannot be explained either by dialectal variability, or by intercalation of a gloss, or by an error in the deciphering to the text without diacritical marks made by a reader who later became a great teacher. Thus Bukhari (65, sura 92, stories 1 and 2) mention that in the Quran 92/3, great Companions like Abu'd-Darda' and Ibn Mas'ud insisted on reciting instead of of the vulgate text, and affirmed that it was the Prophet himself who had taught them thus. One cannot say that is is a revision of the style. One cannot say that God revises His style, nor that Gabriel, this "faithful Spirit", could make errors, even if he were to correct them later. One could not think that the human nature of the Prophet has some role to play in this. Was he absent-minded, did he forget? We can think of the Hadith mentioned by Bukhari (52/11/1 and 80/19/5), Muslim (6/224 no. 788), Ibn Hanbal (6/138) where the Prophet says: "God have mercy on this man who by his nightly recitation reminded me of such a verse which I had forgotten (or dropped) from such a sura." Or is it because when the divine things are revealed to him - not in writing, as with Moses' tablets, but - orally, sometimes some small shade of meaning escapes him? (Then during the yearly collections ('arda) of the month of Ramadan, when Gabriel is present, and the Prophet is momentarily transported again in a heavenly setting, he understands a more correct reading, and he "corrects" himself.) Let us remember that Abu'd-Darda' and Ibn Mas'ud are Muslims since the beginning of Islam, and sura 92 is chronologically No. 9. As for the annual collections by the Prophet, they seem to have begun in Medina only after the institution of the fast of Ramadan in 2 A.H. Thence the few divergences without importance, for example, between Abu'd-Darda' and Ibn Mas'ud on the one hand and Zaid ibn Thabit on the other, the veterans being unwilling to yield to a young man, even though he is the scribe of the Prophet, concerning the writing down of revelations of the Qur'an. There may be other and better explanations for this problem. I remain extremely hesitant." (From "The Problem of Variants" in LE SAINT CORAN, Traduction et commentaire de Muhammad Hamidullah, avec la collaboration de M. Léturmy, nouvelle édition 1989 corrigée et augmentée, pp. xxix-xxx)
For more information on the subject of variant readings, including even suras omitted form or added to the Quran of Uthman, see Arthur Jeffery, Materials for the History of the Text of the Qur'an, which focuses on Ibn Abi Dawud's Kitab al-Masahif and a collection of variant readings, for the Qur'an collections of Ibn Mas'ud, Ubai and others which existed prior to the textus receptus of Uthman (Brill, Leiden, 1937).
Overview: Textual Variants of the Qur'an
Answering Islam Home Page