Textual Criticism of the Qur’an (Prof. Putten)

Article

1. The Uthmanic rasm
2. Islamic narrative sources tell us that during the lifetime of the prophet Muhammad (ca. 570–632 CE), no standard written version of what was considered to be his divine revelation was created. It was only about two decades after his death that a standard text was commissioned by the third Islamic caliph ʿUthmān (r. 644–656 CE). He is believed to have ordered Zayd b. Thābit (d. 660 CE) to lead a commission to create several master copies and distribute them across the nascent Islamic empire (Nöldeke, et al., The History of the Qurʾān, 251ff). While there are different reports as to the number of master copies, the lowest number, four, is the most likely to be correct, as we will see below. Three of the four copies were dispatched to Syria (either Damascus or Homs) (Sidky, “Regionality of the Quranic Codices,” 171–174), Basra (Iraq), and Kufa (Iraq), and one stayed at Medina (Arabia), then the seat of caliphal power. Sidky argues that Homs was more likely than Damascus, although Damascus ends up becoming more important and as a result more representative of the text type.

Many letters are not distinguished clearly without these dots, e.g., bāʾ, tāʾ and thāʾ are all spelled with the same letter shape ٮ, and can only be distinguished through a system of dots: ب for bāʾ, ت for tāʾ and ث for thāʾ. This bare consonantal skeleton is typically known as the rasm, “tracing.” Indeed, our earliest manuscripts lack ways to express vowel signs completely, and have very sparse consonantal dotting – more or less in line with what the tradition reports.

1. Manuscript Evidence for Early Canonization
2. The earliest manuscripts of the Quran are invariably undated, but with the advent of radiocarbon dating it has become possible to challenge this late canonization on the basis of material evidence. We now have a good number of manuscripts that have been confidently radiocarbon dated to the seventh century CE. Most famously the early dating of the fragment colloquially known as the Birmingham fragment (Cadbury Research Library, Birmingham, Mingana Arabic 1572a) made quite a splash in the media when its radiocarbon dating was published to be between 568 CE and 645 CE – the earliest end of that range would mean the text could potentially predate the prophet Muhammad’s career (cf. Fedeli, “Early Qurʾānic Manuscripts”). For a discussion on the dating of mss through radiocarbon dating, see also Marx and Jocham, “Radiocarbon (14C) Dating.”

Of course, early fragments of the Quran are not necessarily evidence for an early canonization of the text. However, as interest in these early fragments grew, it was noticed quite quickly that the text that we find in these fragments is word-for-word identical to the text as we have it today. This already speaks in favor of an early canonization of the text (Sinai, “When Did the Consonantal Skeleton of the Quran Reach Closure? Part I”; idem, “When Did the Consonantal Skeleton of the Quran Reach Closure? Part II), but this becomes even more obvious when we examine the orthography. In the early Arabic orthography of the Quran, there are several words that can be written in a variety of ways, for example niʿmat aḷḷāh, “the grace of God” can be spelled variably with a final tāʾ or a final hāʾ of the first word. This consistent correspondence across manuscripts does not just occur in this single location, but we find a consistent correspondence for all twenty-three cases of this phrase. Moreover, many other phrases could be examined and show the same pattern (Van Putten, “The Grace of God”). The only way to account for such a precise reproduction of these orthographic idiosyncrasies is if the manuscripts were copied from a written exemplar with the utmost precision from the very start of the tradition. And therefore, all manuscripts that follow the Uthmanic text type (which are all manuscripts except one known so far), must descend from a single written archetype.

1. palaeographical and art-historical arguments brought forth by Déroche, Qurʾans of the Umayyads, which clearly draws a notable distinction between pre-Umayyad and Umayyad codices.
2. Lack of a “Masorah”
3. Despite the highly precise copying tradition of the quranic text in the first centuries of Islam, there is no evidence of specific tools to ensure this high precision in copying, such as the “Masorah” that we see in the text of the Hebrew bible (Tov, Textual Criticism, 72–76). Quranic manuscripts are famously devoid of marginal notes of any kind, and thus also of notes of the type that may give indications of the copying practices. The earliest manuscripts written in the so-called Ḥijāzī script, in fact, lack margins altogether, quite consistently using the full width and height of the parchment for writing the Quran (Déroche, Qurʾans of the Umayyads, 66). In the later Kufic script margins are very large, but still these margins are almost exclusively used for ornamentation of sūrah headers and, on occasion, to mark subdivisions in the text. Otherwise, margins are kept empty, leading to a striking and visually attractive layout but also a rather ostentatious waste of parchment. In later centuries, scholars started to produce companion works to the Quran that detail the precise orthographic idiosyncrasies of the Uthmanic text. Such works are known as rasm literature, and several important early ones are Ibn Abī Dāwūd’s (d. 929) Kitāb al-Maṣāḥif and al-Dānī’s (d. 1052–1053) al-Muqniʿ fī Rasm al-Maṣāḥif. These works relate many subtle details of the Uthmanic text type, but were authored after the Uthmanic rasm had already undergone some evolution.

Stemmatics of the Uthmanic Text Type

The Islamic tradition tells us that, when the Quran was first canonized, at least four master copies were made and distributed across the Islamic empire. The medieval rasm literature records differences between these regional codices, primarily reporting about the Syrian, Medinan, Basran, and Kufan exemplars. Occasionally they report differences for a Meccan exemplar as well. The differences that are reported tend to be tiny differences in the consonantal skeleton that mostly appear to be the result of minor scribal errors. As Cook argues, the fact that these reports form an uncontaminated stemma is a clear indication that we are dealing with a genuine transmission from an archetype.

Four possible stemmata based on regional variants from a single archetype of the Quran (S= Syria, M = Medina, B = Basra, and K = Kufa)

The Multiform Oral Traditions

While the Uthmanic text type is strongly controlled in its written transmission, this does not mean that there is no variation present in its transmission. As we mentioned, the early quranic manuscripts were written in a highly defective script, which did not completely unambiguously reflect the text. A number of different prominent reciters, primarily in the 8th CE, started to transmit their own reading tradition (qirāʾah pl. qirāʾāt). By the tenth century, the scholar Ibn Mujāhid (d. 936) described seven of the reading traditions of these reciters, which, after his time, came to be thought of as canonical (Nasser, The Transmission of the Variant Readings, and Nasser, The Second Canonization of the Qurʾān). Today two transmitters per eponymous reader are considered canonical (Nasser, “The Two-Rāwī Canon”). Athough these seven came to be universally accepted as canonical, scholars continued to collect and describe the reading traditions of other reciters. Several centuries later, three more readings came to be accepted as canonical after their description by Ibn al-Jazarī (d. 751/1350), who likewise described those three with two transmissions for each one (Ibn al-Jazarī, Nashr).

1. Among the specific variants, their function can vary quite wildly. Some of these can affect the meaning of the verse, while in other cases we seem to simply be dealing with different word formations with essentially the same meaning, or representing dialectal variations of one another. The list below gives an overview of some of the typical variants one encounters: – Dialectal variants of the same word: al-quds, al-qudus, “holiness” (Q 2:87). – Different noun formation, same meaning: khaṭaʾan, khiṭāʾan, khiṭʾan, “a sin” (Q 17:31) – Different nouns with different meanings: mālik, “possessor,” malik, “king’” (Q 1:4). – Singular or plural noun: wa-kitābihī, “and his scripture,” wa-kutubihī, “and his scriptures” (Q 66:12). – Active or passive verb: nazzala, “he has sent down,” nuzzila, “it has been sent down” (Q 4:140). – Different verb stem: qatalū, “they killed,” qattalū, “they massacred” (Q 6:140)
2. The transmission of the quranic reading traditions continues to be a living practice among Muslims today. Only a few readings have reached broad popularity, most notably that of Ḥafṣ transmitting from ʿĀ ṣim, which is by far the most widespread reading tradition in the Muslim world today. In North Africa, however, the tradition of Warsh transmitting from Nāfiʿ is dominant, while Qālūn, likewise from the transmission of Nāfiʿ, is adhered to as well, especially in Tunisia and Libya. In Sudan al-Dūrī, transmitting from Abū ʿAmr remains in popular use. The other readers however, continue to be learned and recited by experts in a living tradition. While the teaching of these reading traditions these days relies on written works, the oral transmission is still considered an essential part of its acquisition. People who get certified are expected to have taken an in-person recitation exam, and the readings are traced back through chains of transmission, usually to the great writers of the canonizing works like Ibn al-Jazarī and Ibn Mujāhid, then all the way back to the eponymous readers whose reading traditions they transmit. The canonical eponymous readers, in turn, trace back their reading through the teachers they received instruction from, all the way back to the prophet Muhammad. Today, these readings are, in principle, verbatim reproductions of how the canonical readers recited (although, of course, some disagreement has arisen over time between transmitters and sub-transmissions). However, it does not seem to be the case that verbatim reproduction from what one’s teacher said was necessarily always the goal. This much is clear from Khalaf as canonical reader, whose reading is distinct from that of Ḥamzah, for whom he is also one of the two canonical transmitters.
3. Today, disagreements between any two transmitters from a single reader are often piously explained by suggesting that the teacher simply taught both options, but there is very little to suggest that this is what typically happened and is the process by which every single reading variant should be explained.