Abstract
Until the discovery of the Dead Sea Scrolls, the dominant paradigm of the textual criticism of the Hebrew Bible was the quest for the original form of each of its books—the source texts (Urtext) from which all subsequent editions were copied. Since the discovery of the Dead Sea Scrolls, a number of scholars have proposed significant revisions to this paradigm. These proposals are presented by means of an analogy with David Living-stone's expeditions to find the source of the River Nile, and then evaluated by means of a comparison of the history of the text of the Hebrew Bible with the history of the text of the Qur'ān. According to Ibn Abī Dāwūd (d.928), the widespread oral memorization and recitation of the Qur'ān in the first Muslim generation led to the emergence of a multiplicity of textual and oral versions of its ‘original’ suwar. This, in turn, led to a series of (ultimately successful) attempts to standardize the text of the Qur'ān through the repression of all readings that differed from the one ‘official’ text. Applying a ‘Livingstonian’ text-critical model to the Qur'ān suggests that ongoing research into the earliest forms of the Qur'ān could be revolutionized if it sought to recover the early plurality that was consequent from its popularity. Applying an Ibn Abī Dāwūd text-critical model to ongoing research into the Talmudic and Masoretic periods of the Hebrew Bible suggests that it could be revolutionized if it sought to recover the history of the standardization of the variant texts. Under these paradigms, the purpose of textual criticism must be transformed from the pursuit of an imagined ideal text to become an enquiry into the nature of the texts that have been declared canonical.
The Hebrew Bible and the Dead Sea Scrolls
The Parable of the Expedition
The textual criticism of the Hebrew Bible can be likened to a Victorian expedition that set out to explore three great rivers: the Masor (the largest of the three and well-known to the main local tribe); the Septor (less broad, but better understood by European explorers); and the Samar (the smallest of the three and quite difficult to navigate without the help of a local tribe who know it well). The leaders of this expedition believed these rivers to be the distributaries of one mighty source. They divided the expedition into three parties, one for each of the three rivers, and set out expecting to meet up with each other at their common source. The party assigned to chart the mighty Masor had to paddle against a strong flow. After many days of tireless paddling, however, they found that the flow of the river began to diminish. Subsequently, hundreds of streams began to appear alongside it. As the expedition pressed on it became increasingly difficult to distinguish between the mighty Masor, its tributaries and the marshlands around these tributaries. Confused, the Masor party returned to base camp to await news from the other parties. At base camp, the Samar and Septor parties were waiting for them. Their smaller rivers had presented less of a challenge, but their paddling success had been tempered by their discovery of an even greater confusion of tributary and marshland. Discouraged, the expedition returned home to report its findings to the Royal Society—and request more funds for further research. The Royal Society, however, was perplexed. The exploration of these three rivers had raised more questions than they had answered. If further expeditions were to be launched, which of the hundreds of tributaries and marshlands now charted should they choose to explore next? Which one would turn out to be the most important source? Why did all these tributaries produce only three major rivers, not one (or two or five or three hundred)? What could all this profusion of marshland and tributary mean?
The Parable Explained
Before the discovery of the Dead Sea Scrolls in 1947, the textual criticism of the Hebrew Bible was based upon three significant ‘recensions’ (text types) held by three socio-religious communities: the Masoretic text, the authoritative text of rabbinical Judaism; the Septuagint, a Greek translation of the Hebrew Bible used by the Jewish diaspora in the Second Temple period (from the third century BCE onwards) which became the Old Testament of early Christianity; and the Samaritan Pentateuch, the stylized ‘paleo-Hebrew’ text of a Samaritan community that had been independent from other forms of Judaism from the period of the Hasmonean dynasty (second century BCE). Each one was presumed to have descended from, and so witness to, the Urtext of the Hebrew Bible.
The discovery of the Dead Sea Scrolls began in 1947 when Muhammed ed-Dīb, a young Bedouin shepherd, threw a stone into a small opening in a cliff face of the Wādī Qumrān. The story of the next ten years of exploration and discovery is one of intrigue, community relations and patience, as well as scholarship. 1 As the dust settled on the frantic search for caves, Frank Cross and William Foxwell Albright were the first to observe that some of the Hebrew Dead Sea Scrolls represented the Septuagint translation of the Hebrew Bible better than they represented its later canonical form (the Masoretic Text). This suggested that the Septuagint was at various points not a free translation of a forerunner of the Masoretic Text (a proto-Masoretic Text), but a translation of a different Hebrew Vorlage. Indeed, the subsequent research of Cross proved that this was the case for the book of Samuel. 2 Cross and Albright presented a similar argument in regard to the Samaritan Pentateuch. In terms of our parable, these early conclusions launched the expedition to chart the waters of Masor, Septor and the Samar.
John M. Allegro, The Dead Sea scrolls (Harmondsworth: Penguin Books, 1956); Frank Moore Cross, The Ancient Library of Qumran and Modern Biblical Studies (The Haskell Lectures, 1956-57; London: Gerald Duckworth, 1958); Roland de Vaux, Archaeology and the Dead Sea Scrolls (The Schweich Lectures of the British Academy; London: Oxford University Press, 1973).
Frank Moore Cross, Qumran Cave 4, 1–2 Samuel (DJD; Oxford: Clarendon Press, 2005).
Further research, however, showed that the situation was more complex. Many texts not only deviated from the later Masoretic Text, they also deviated from all possible Vorlagen of the Septuagint and the Samaritan Pentateuch. Shemaryahu Talmon was the first clear proponent of the significance of these findings. 3 In his opinion, any of the texts of the multiple variant versions of the books of the Hebrew Bible could have been adopted as the canonical text of the Hebrew Bible. It was purely an accident of history that only three sects survived from the many active in first-century Judaism: rabbinic Judaism, the Samaritans and the followers of Jesus the Nazarene. As a result of this accident the proto-Masoretic Text became the scripture of rabbinic Judaism, the Septuagint survived as the scripture of the Christian sect and the Samaritan Pentateuch survived as the scripture of the Samaritan sect. All others were lost because, and only because, no other sect had adopted them and survived. In terms of our parable, Talmon has observed the number and significance of the tributaries and explained how only three rivers could have been produced by a myriad of tributaries.
See Shemaryahu Talmon, ‘The Old Testament Text’, ‘Aspects of the Textual Transmission of the Bible in the Light of Qumran Manuscripts'’, and ‘The Textual Study of the Bible—A New Outlook’, in Frank Moore Cross and Shemaryahu Talmon (eds.), Qumran and the History of the Biblical Text (Cambridge, MA: Harvard University Press, 1976), pp. 1-41, 226-63, 321-400, respectively.
As chief editor of the biblical scrolls of Qumrān, Eugene Ulrich's first task was to define which texts were ‘biblical’ (and so in his domain) and which were not. The significance of this issue is evident in regard to the ‘Great Psalms Scroll’ (also known as 11Q5 and 11QPsa). This scroll probably began with Psalm 101 and included all (or almost all) of the subsequent Psalms of the masoretic Psalter, 4 but presented them in quite a different order to that of the masoretic Psalter and included extra material alien to the masoretic Psalter. 5 Furthermore, fragments of two other scrolls (4Q87 and 11Q6) suggest that this was not a one off ‘alternative hymnbook of Qumrān’. It seemed that these two scrolls, most likely, included all the psalms of the Psalter and, most probably, followed the order of the masoretic Psalter in regards to the first hundred psalms and the order of 11Q5 in regard to the last fifty. According to Ulrich, different books of the Bible were at different stages of becoming scripture when calamity hit Qumrān. At one extreme, the books of Moses are widely quoted as scripture in the corpus of Qumrān. At the other extreme, the Temple Scroll presents itself as a new writing (second century BCE) that should be accepted as scripture. The content of the last book of the Psalter lies between these two extremes along with the book of Enoch, Ben Sira, Jubilees and others. In this context, Ulrich argues that the term ‘biblical’ imposes anachronistic values upon the scrolls of Qumrān unless it is given a more flexible and wider meaning. In terms of our parable, Ulrich has begun to describe the marshlands around which the various tributaries appeared to have developed. The explanation of the significance of these marshlands and their relationship to the three rivers has been no easy matter. It has taken Ulrich most of his working life. 6
So Peter W. Flint, The Dead Sea Psalms Scrolls and the Book of Psalms (Studies on the Texts of the Desert of Judah; Leiden: Brill, 1997), p. 39.
The order appears to have been Pss. 101, 102,103, 118, 104, 147, 105, 146, 148, 120–132, 119, 135, 136, 145, 154, ‘A plea’, 139, 137, 138, Ben Sira 51, Apostrophe of Zion, 93, 141, 133, 144, 142, 143, 149, 150, ‘Hymn to the Creator’, 2 Sam. 23.1-7, David the Psalmist, 140, 134 and 151: James A. Sanders, The Psalms Scroll of Qumran Cave 11—11QPs (DJD; Oxford: Clarendon Press, 1965).
Eugene Ulrich, The Qumran Text of Samuel and Josephus (Missoula: Scholars Press, 1978); The Dead Sea Scrolls and the Origins of the Bible (Leiden: Brill, 1999); The Dead Sea Scrolls and the Developmental Composition of the Bible (VTSup, 169; Leiden: Brill, 2015).
As the editor-in-chief for the Dead Sea Scrolls Publication Project. Emanuel Tov and his book Textual Criticism of the Hebrew Bible provide a fitting conclusion to the explanation of our parable. 7 The first English edition of this book was published in 1992 and republished with minor amendments as a second edition in 2001. The third edition (published in 2012), however, witnesses to the chief editor's struggles with the developing evidence (summarized above). It includes extensive revisions and additions. In these works Tov reports that ‘fragments of 210–212 biblical scrolls from Qumrān (representing 224–226 biblical books) were found in eleven Qumrān caves’. 8 Out of these, 121 are ‘sufficiently extensive for analysis’. In the opinion of Tov: 9
Emanuel Tov, Textual Criticism of the Hebrew Bible (Minneapolis: Fortress Press, 1992; 2nd edn, 2001; 3rd edn, 2012). A Hebrew version of this work was published in 1989.
Tov, Textual Criticism of the Hebrew Bible (2012), p. 95.
Tov, Textual Criticism of the Hebrew Bible (2012), p. 108.
55 (45%) are closest to the consonantal framework of the Masoretic Text.
55 (45%) are non-aligned.
6 (5%) are closest to a presumed Hebrew source of the Septuagint.
5 (5%) are closest to the Samaritan Pentateuch.
Tov (1992) concludes that:
The textual reality of the Qumran texts does not attest to three groups of textual witnesses, but rather to a textual multiplicity, relating to all of Palestine to such an extent that one can almost speak in terms of an unlimited number of texts. 10
Tov, Textual Criticism of the Hebrew Bible (1992), p. 161.
Although the 1992 edition of Textual Criticism of the Hebrew Bible provided a clear statement of the situation and the need for a radical re-evaluation, its text also defended the paradigm of ‘an original text’ against a paradigm of ‘multiple pristine texts’ since:
It is ‘the simplest’.
‘The known textual evidence points in the direction of one Original text.’
‘The canonical concept that has been accepted in Judaism’ requires it.
‘The alternative assumption of different pristine versions of the biblical books does not appear to be proven by the facts or logic.’ 11
Tov, Textual Criticism of the Hebrew Bible (1992), p. 172.
Tov (1992) summarized the logic of this position as follows:
At the end of the process of the composition of a biblical book stood one textual entity (a single copy or tradition) which was considered finished at the literary level, even if only by a limited group of people, and which at the same time stood at the beginning of a process of copying and textual transmission. During the textual transmission many changes occurred, making it now impossible for us to reconstruct the original form of the text. These difficulties, however, do not refute the correctness of the assumption. 12
Tov, Textual Criticism of the Hebrew Bible (1992), p. 177.
As can be seen from the above quotations, Tov (1992) presents a passionate argument for the re-assessment of the paradigms of the textual criticism of the Hebrew Bible alongside an equally passionate argument for the preservation of the dominant paradigm! The inherent tension is more evident in the 2012 edition of this work. A unique feature of this edition is the presentation of its text. Whenever the text follows the 1992 edition, it is reproduced in standard font, but whenever the text represents new material, it is found in a smaller font. Tov (1992) argued that the textual criticism of the Hebrew Bible must be founded on either a model of ‘multiple pristine texts’ or a model of one ‘original text’—and favoured the latter. Tov (2012) revised this argument to involve a choice between a model of ‘multiple pristine texts’, and a model of ‘an original or a series of determinative (original) texts’. 13 Although once again the latter option is preferred, it has been modified. One must turn to other works of Tov to discover what he might mean by a model of ‘a series of determinative (original) texts’ and how this might be different to a model of ‘multiple pristine texts’.
Tov, Textual Criticism of the Hebrew Bible (2012), p. 163.
In a baraitā to the Palestinian Talmud there is a report of a tradition concerning three temple scrolls (sěfārim) whose names were Sefer Mě'onā, Sefer Za'aṭuṭê and Sefer Hi'. 14 According to the tradition, one scroll was named Sefer Mě'onā because it contained the unique reading ma'on (‘dwelling’) whereas the other two read mě'onā (‘dwelling’) at Deut. 33.27; another was called Sefer Za'aṭuṭê because it alone used za'aṭuṭê (‘youngsters, kids’) instead of na'arê (‘youths’) at Exod. 24.5; the third was called Sefer Hi' since it used the masculine pronoun hu' in nine places in which the other two scrolls used the anomalous feminine pronoun hi' with masculine meaning. Since it is practically impossible for hand-copied scrolls to be error-free, the ‘Temple Standard Text’ was defined as the majority text derived from three chosen, very similar, scrolls (Sefer Mê'onā, Sefer Za'aṭuṭê and Sefer Hi'). This report is repeated with minor variations in three later works, 15 and can be understood to be consistent with a related tradition in the Babylonian Talmud. 16
For a fuller explanation of the place of this tradition in the thought of Tov, see Armin Lange, ‘They Confirmed the Reading (y. Ta'an. 4.68a): The Textual Standardization of Jewish Scriptures in the Second Temple Period’, in Armin Lange, Matthias Weigold, and József Zsengellér (eds.), From Qumran to Aleppo: A Discussion with Emanuel Tov about the Textual History of Jewish Scriptures in Honor of his 65th Birthday (Göttingen: Vandenhoeck & Ruprecht, 2009), pp. 29–80.
That is, Sifre to Deuteronomy (Sifre Devarim ii. 356), one of the two recensions of Avoth de Rabbi Nathan (B, chapter 46) and the late tractate Soferim (vi.4).
b. Ketuvim 106a: Rabba b. Bar Hana said in the name of R. Johanan, ‘Correctors of scrolls in Jerusalem received their fees from the Temple funds’.
Tov has argued that three features of the Dead Sea Scrolls support this tradition. 17 First, the 55 texts that are consistent with the Masoretic Text indicate that a proto-Masoretic text was in circulation by the second century BCE. Second, the proto-Masoretic recension is the most popular singular recension-type at Qumrān. Third, the proto-Masoretic is the only recension evidenced to have been used by other Jewish religious groups outside of Qumrān in the first and second centuries CE. 18 If the model of (three) determinative texts is accepted, all other textual traditions and textual witnesses form the background against which these (three) determinative texts were chosen. It is impossible to know whether any of the (three) determinative texts were more ‘original’ than other text traditions. It is only possible to know that a choice was made. As a result, the six volumes of the Hebrew University Bible present the ‘Masoretic Text’ 19 without any consonantal alterations. 20 All textual variants are listed in the footnotes—as an essential description of the context in which the canonical text was chosen. These textual variants can be used to evaluate the quality of the choice. They cannot change that which has been chosen.
Emanuel Tov, ‘Some Thoughts about the Diffusion of Biblical Manuscripts in Antiquity’, in S. Metso et al. (eds.), The Dead Sea Scrolls: Transmission of Traditions and Production of Texts (Leiden: Brill, 2010), pp. 151-72; idem, ‘Understanding the Text of the Bible 65 Years after the Discovery of the Dead Sea Scrolls’, Open Theology 1, no. 1 (2014), pp. 89–96.
All 25 scroll fragments found at sites other than Qumrān (that is, Masada, WāCdī Murraba'at, Wādī Sdeir, Naḥal Ḥever, Naḥal Tse'elim and Naḥal Arugot) are consistent with the Masoretic text tradition.
As defined by the Aleppo Codex, or where illegible or lost, Codex L (that is, St Petersburg, Firkovitch collection, 1/19a).
Moshe H. Goshen-Gottstein and Galen Marquis, Sefer Yesha'yahu (3 vols.; Hebrew University Bible; Jerusalem: Magnes Press, 1975); Emanuel Tov et al., Sefer Yirmeyahu (The Hebrew University Bible; Jerusalem: Magnes Press, 1997); Moshe H. Goshen-Gottstein, Shemaryahu Talmon, and Galen Marquis, Sefer Yehezkel (The Hebrew University Bible; Jerusalem: Magnes Press, 2004).
In terms of our parable, Tov and the Hebrew University Bible treat the Samar, Septor and all their tributaries as part of the view that travellers on the Masor should enjoy. This view does not include any of the marshlands. In other words, this solution ignores all the writings found among the Dead Sea Scrolls that appear to be treated as scripture (or, indeed, claim themselves to be scripture) that would not be included in the subsequent canon of the Hebrew Bible. In the first century, the Samaritan scriptures limited the canon to the five books of the Pentateuch, Josephus understood the Jewish scriptures to be limited to twenty-two books, 21 different editions of the Septuagint present the Jewish scriptures as extending beyond these twenty-two. 22 The solution of Tov and the Hebrew University Bible is, in effect, to ignore all issues of canon and composition and to focus on the transmission of the text that would become the canonical scripture of rabbinic Judaism. In a recent work, Ulrich explores these issues and their importance for our understanding of the nature of revelation and scripture. 23 The present study focuses on the parameters set by Tov since these parameters facilitate a comparison of the formation and transmission of the text of the Hebrew Bible with the formation and transmission of the Qur'ān.
In his Against Apion, Josephus' total of 22 appears to count Samuel, Kings, Chronicles, Judges–Ruth, Ezra–Nehemiah, Jeremiah–Lamentations and the twelve ‘Minor Prophets’ as one ‘book’ each. Later rabbinic tradition followed this count, but presented Judges, Ruth, Jeremiah and Lamentations as four books not two (so making 24). Modern Christian editions of this text present Samuel, Kings, Chronicles and Ezra–Nehemiah in two parts each and the twelve ‘Minor Prophets’ separately so giving 39 ‘books’. See Roger T. Beckwith, The Old Testament Canon of the New Testament Church, and Its Background in Early Judaism (London: SPCK, 1985), pp. 235–73.
Codex Sinaiticus, for example, adds 2 Esdras, Tobit, Judith, 1 Maccabees, 4 Maccabees, Wisdom and Ben Sira.
For example, John Barton, The Spirit and the Letter: Studies in the Biblical Canon (London: SPCK, 1997); The Old Testament: Canon, Literature and Theology. Collected Works of John Barton (Aldershot: Ashgate, 2007).
The Temple Standard Text and the Standard Qur'ānic Text of ‘Uthmān
According to Muslim traditions, the scribe Zaid bin Thābit was commissioned first by Abu Bakr, the first caliph (r. 632–634), then by the second caliph, ‘Umar (r. 634–644), to gather together an authoritative text of the Qur'ān from the memories and haphazard recorded writings of the first generation of Muslims, that is, from ‘parchments, scapula, leafstalks of date palms and from the memories of men (who knew it by heart)’. 24 The task proved to be much more difficult than had been estimated since, when scrutinized, more variant versions of each sūrah of the Qur'ān were found to be in circulation amongst this first generation than had been supposed. 25 The chief problem that Zaid b. Thābit faced was a corollary of the significance of the Qur'ān in the rapidly growing Muslim community. On the one hand, it was because the Qur'ān was memorized by so many individuals, families and communities that it was of paramount importance to the unity of Muslims and that there needed to be one standard (canonical) text. On the other hand, it was because the Qur'ān was memorized by so many individuals, families and communities that there was so much diversity and the task of uniting all Muslims under one standard text was so challenging.
Ṣaḥīḥ al-Bukhārī. Muhammad ibn Ismā'īl al-Bukhārī and Muhammad Muhsin Khan, Ṣaḥīḥ al-Bukhārī: The Translation of the Meanings of Sahih al-Bukhari. Arabic–English (Riyadh: Darussalam, 1997), IV, pp. 156–57. The quotation is from al-Bukhārī's most expansive account of the collation of the Qur'ān (volume VI, the Book of Commentary, hadith 4679), but there are many other briefer accounts (for example, hadith 4784, 4984, 4986, 4987, 4988, 4989, 4993, 5003, 5004). The most extensive early account of this first maṢāḥif project under Zaid b. Thābit is found in al-Tabarī: Abū Ja'far Muḥammad B. Jarīr Al-Ṭabarī et al., The Commentary on the Qur'an (Oxford: Oxford University Press, 1987), pp. 24–27. See also Frederick Leehmuis, ‘Codices of the Qur'ān’, in Jane Dammen McAuliffe (ed.), Encyclopaedia of the Qur'ān (Leiden: Brill, 2005), I, pp. 347–51.
The different Muslim traditions of this story, their likely historicity and the most significant alternative proposed historical reconstructions of the editing and collection of the Qur'ān are surveyed in John Burton, The Collection of the Qur'an (Cambridge: Cambridge University Press, 1977), and idem, ‘Collection of the Qur'ān’, in McAuliffe (ed.), Encyclopaedia of the Qur'ān, I, pp. 351-61; John E. Wansbrough, Quranic Studies: Sources and Methods of Scriptural Interpretation (Oxford: Oxford University Press, 1977); Harald Motzki, ‘Alternative Accounts of the Qur'ān's Formation’, in Jane Dammen McAuliffe (ed.), The Cambridge Companion to the Qur'an (Cambridge: Cambridge University Press, 2006), pp. 59–76.
Students of Muslim history regularly express frustration at the lack of sources from the first two Muslim centuries. 26 According to our earliest account (that of Ibn Abī Dāwūd, d. 928) in the time of the third caliph (‘Uthmān b. Affān, r. 644–656), each major Muslim city (Madīna, Mecca, Baṣra, Ḥomṣ, Baghdad, Kūfa and Damascus) had its own recension of the Qur'ān and some had several. For example, when the Yemenite Abū Mūsā was appointed the governor of Baṣra, he brought his own large codex of the Qur'ān for the people of Baṣra to use as their own. This codex, however, differed from that used by this city, the codex of Ibn Mas'ūd, so the criers began to call for all those who desired to recite according to Abū Mūsā to gather to one city gate and those according to Ibn Mas'üd to gather to another. An appeal was, therefore, made to the caliph (‘Uthmān b. Affān) to settle the matter. ‘Uthmān's solution was to create multiple copies of one of the codices of Madīna and order the destruction of all other codices—including both of those held at Baṣra. When ‘Uthmān's delegation arrived to destroy both the codex of Ibn Mas'ūd and that of Abū Mūsā, Abū Mūsā is reported to have agreed on the condition that, ‘Whatever you find in my Codex additional to this one, do not remove it [from my Codex], but whatever you find missing write it in [your Codex]’. 27 It is reported that the codex of Abū Mūsā contained two short suwar and a verse about the greed of men that were absent from the codex of ‘Uthmān. They were not included.
As most eloquently expressed by Tom Holland, In the Shadow of the Sword: The Battle for Global Empire and the End of the Ancient World (London: Abacus, 2013).
Arthur Jeffery and ‘Abd Allāh ibn Sulaymān Sijistānī, Materials for the History of the Text of the Qur'ān: The Old Codices which Present a Type of Text Anterior to that of the Canonical Text of ‘Uthmān (Leiden: Brill, 1937), p. 210.
According to Ibn Abī Dāwūd, the originator of one of the codices of Baṣra was a close companion of the Prophet called ‘Abdullāh b. Mas'ūd who lived in Kūfa. When the delegation came to confiscate his codex, Ibn Mas'ūd refused on the grounds that the Prophet himself had appointed him to guard the authentic Qur'ān while Zaid bin Thābit had only been appointed by the caliphs to do this. The outcome of this stand-off and the ultimate fate of his codex is unclear, but the number of text variants ascribed to it in the commentaries suggests it was, at one time in Muslim history, a significant recension. The codex of Ibn Mas'ūd is reported to have followed a different order and to have lacked the first sürah (Q1, al-Fātiḥa) and the final two suwar (Q113 and Q114) and the verse Q94.6. Ibn Mas'ūd considered the first sūrah to be the daily prayer of all Muslims to God and not, therefore, to be a part of the suwar that God had revealed to the Prophet.
Finally, according to Ibn Abī Dāwūd, a number of codices existed in Madīna alongside the one chosen by ‘Uthmān. These included a codex of Ibn ‘Abbās (the famous exegete), another held by ‘Alī b. Abī Ṭālib (the fourth caliph and first imam of Shī'a Islam) and a third held by Ḥafsa the daughter of ‘Umar. While Ibn ‘Abbās and ‘Alī b. Abī Ṭālib surrendered their codices to the flames for the greater good of Islam, Ḥafsa is said to have kept hers (which included many of the first written copies of different suwar collated in the time of Abū Bakr) hidden under her bed. After the death of ‘Uthmān, when Marwān became Governor of Madīna,
he sent to Ḥafsa demanding her Codex that he might destroy it, but she refused to give it up. When she died Marwān assisted at the funeral and at its conclusion sent, and with much insistence, demanded the Codex from ‘Abdallah b. ‘Umar, Ḥafsa's brother. ‘Abdallah finally sent it him and he had it destroyed, fearing, he said, that if it got abroad the variety of readings that ‘Uthmān desired to repress would recommence. 28
Jeffery and Sijistānī, Materials for the History of the Text of the Qur'ān, pp. 212–13.
This Muslim tradition is analogous in function to the Talmudic tradition of the three scrolls. In both cases, the existence of variant texts is recognized, the details of these variants are dismissed and one authoritative version is legitimized. Muslim memory, however, has retained much more spice and far more intrigue in its traditions. It should also be noted that even this level of spice and intrigue represents a sanitized tradition. As Omar Handon has argued, it was probably not until after the extensive reforms of al-Ḥajjāj (d. 713) that the text of the Qur'ān could be said to be anywhere near uniform across the Muslim Empire. 29 While a similar process of resistance and reform must have been involved in the transition between the textual multiplicity in the time of Qumrān (first century CE) and the textual uniformity of the Tiberian Masoretes (tenth century CE), Jewish tradition is silent on this matter. The Muslim tradition, therefore, provides a model of the spice and intrigue that could have been involved.
Omar Handon, ‘The Second masahif Project: A Step Towards the Canonization of the Qur'anic Text’, in Angelika Neuwirth, Nicolai Sinai, and Michael Marx (eds.), The Qur'an in Context: Historical and Literary Investigations into the Qur'anic Milieu (Leiden: Brill, 2010), pp. 795–836.
A key feature of the Muslim tradition is the active and positive role that is ascribed to oral memorization, not only in the formation of the text of the Qur'ān, but also in the preservation and transmission of these texts. In regard to the formation of the text, oral memorization caused the ‘original’ Qur'ān (that is, the suwar recited by the companions of the Prophet) to be united in content, but variant in detail. Widespread oral memorization also meant that it took two generations before the authorities could expunge the memory of the codex of Ibn Mas'ūd from the people of Küfa and Baṣra and even then imperfectly. The record of their variants remained in the Muslim memory until the time of Ibn Abī Dāwūd, over two hundred years later. 30 However, widespread oral memorization also meant that, once the ‘Uthmānic recension was established, the text could be preserved largely unchanged. 31 The question arises as to the extent to which the known features of the history of the text of the Qur'ān can be used as a model to understand the history of the text of the Hebrew Bible. While many features of the oral-textual world of the Qur'ān relate to those of the Hebrew Bible, there are also significant differences. To answer this question, therefore, it is necessary to consider some of the similarities, and some of the differences in their oral-textual worlds.
Jeffery lists over one thousand verses in which the reading of Ibn Mas'ūd is still remembered in subsequent commentaries of the Qur'ān as having differed: Jeffery and Sijistānī, Materials for the History of the Text of the Qur'ān, pp. 25–113.
This, at least, is the conclusion of two recent exhaustive studies: Keith E. Small, Textual Criticism and Qur'an Manuscripts (Lanham: Lexington Books, 2011); Shady Hekmat Nasser, The Transmission of the Variant Readings of the Qur'ān: The Problem of tawātur and the Emergence of shawādhdh (Leiden: Brill, 2013).
Fifty years ago, when the discoveries of the Dead Sea Scrolls were first coming to light, a number of works emphasized the radical difference between societies founded on oral norms and those founded on the norms of the written text. 32 Since this time, however, it has become clear that written and oral technologies are better understood as symbiotic. 33 In particular, it has been shown that societies do not transition between oral and written ways of thinking overnight and the extensive period of transition does not involve anything like a Kuhnian revolution. 34 Instead, the introduction of writing and written texts, in the first instance, tends to enhance oral memorization while oral memorization tends to enhance the development of complexity in writing. Whether this is the case for every society in every epoch of its development need not detain us here—since it is most evidently the case in regard to the pre-modern world of the Hebrew Bible and in regard to the pre-modern world of the Qur'ān.
So Albert Bates Lord, The Singer of Tales (Cambridge, MA: Harvard University Press, 1960); Claude Levi-Strauss, The Savage Mind (Chicago: University of Chicago Press, 1966); Marshall McLuhan and Quentin Fiore, The Medium is the Massage (New York: Bantam Books, 1967); Jack Goody (ed.), Literacy in Traditional Societies (Cambridge: Cambridge University Press, 1968); Jack Goody, The Domestication of the Savage Mind (Cambridge: Cambridge University Press, 1977); Walter J. Ong, Orality and Literacy: The Technologizing of the Word (London: Methuen, 1982).
Note that Goody and Lord published revisions of their early thought: John Miles Foley, Oral Tradition in Literature: Interpretation in Context (Columbia: University of Mssouri Press, 1986); Jack Goody, The Interface between the Written and the Oral (Cambridge: Cambridge University Press, 1987); Mary J. Carruthers, The Book of Memory: A Study of Memory in Medieval Culture (Cambridge: Cambridge University Press, 1990); Albert Bates Lord and Mary Louise Lord, The Singer Resumes the Tale (Ithaca, NY: Cornell University Press, 1995); Jack Goody, The Power of the Written Tradition (Washington: Smithsonian Institution Press, 2000).
Thomas S. Kuhn, The Structure of Scientific Revolutions (Chicago: University of Chicago Press, 1962).
The roots of the Hebrew Bible reach back to the development of the ‘alphabet’ in which a limited number of phonetic symbols represent a much larger number of sounds. 35 Key to this development was the isolation of consonantal sounds from vowels sounds to eliminate the need for a host of different symbols for each consonant-vowel combination. Paleo-Hebrew used twenty-two signs to provide a consonantal skeleton. Readers familiar with Hebrew are able to determine the appropriate vowel that goes with each consonant (if there is one) when reciting the text. After the destruction of the first Jerusalem temple (586 BCE), Hebrew was less widely used and scribes began to use Aramaic lettering, not paleo-Hebrew. This allowed vowel letters to be added to aid the reader-reciter.
This summary follows Peter T. Daniels, ‘Alphabet, Origins of’, in Geoffrey Khan (ed.), Encyclopedia of Hebrew Language and Linguistics (Leiden: Brill, 2013), I, pp. 87–95.
The oral-textual world of first-century Palestine was influenced by the exhortations of the Hebrew Bible to memorize the text, to discuss it in family gatherings, to meditate upon it night and day as well as to read in congregation. 36 In the later memory of the Mishnā, the Palestinian Talmud and the Babylonian Talmud, the oral memorization and performance of scripture was also considered to be an integral part of Jewish religion and culture. 37 The question does arise, however, as to what extent these texts represent an ideal, to what extent they were only the practice of a minority religious (or scribal) class and to what extent they might represent the common experience of the scriptures. 38 The more popular the scripture, the greater the diversity generated.
So Deut. 6.6-7; 11.19-21; 31.10-13; Ezra 8.1-9; Pss. 1 and 119. See also the use of these texts in, for example, Birger Gerhardsson, Memory and Manuscript: Oral Tradition and Written Transmission in Rabbinic Judaism and Early Christianity (trans. Eric J. Sharpe; Uppsala: University of Uppsala Press, 1961); David McLain Carr, Writing on the Tablet of the Heart: Origins of Scripture and Literature (Oxford: Oxford University Press, 2005); Gary D. Martin, Multiple Originals: New Approaches to Hebrew Bible Textual Criticism (Atlanta: Society of Biblical Literature, 2010).
For example, in the Mishnā, Avoth 3.2-3 advocates the discussion of scripture in table fellowship and Avoth 5.21-22 advocates a life of scripture memorization and meditation. In the Talmud, at b. ‘Eruvin 54b, the teaching of scripture by word of mouth and its memorization through mnemonics is encouraged; and at BT Ketuvoth 77b the example of Rabbi Joshua ben Levi is recommended (he requested his death be delayed by thirty days in order to revise his scripture knowledge).
For a variety of answers to this question, see Hartmut Stegemann, The Library of Qumran, on the Essenes, Qumran, John the Baptist, and Jesus (Leiden: Brill, 1998); Alan R. Millard, Reading and Writing in the Time of Jesus (Sheffield: Sheffield Academic Press, 2000); Philip S. Alexander, ‘Literacy among Jews in Second Temple Palestine: Reflections on the Evidence from Qumran’, in M.F.J. Baasten and W. Th. van Peursen (eds.), Hamlet on a Hill: Semitic and Greek Studies Presented to Professor T. Muraoka on the Occasion of his Sixty-Fifth Birthday (Leuven: Peeters, 2003), pp. 3–24.
We can be reasonably certain that in the earlier periods of the Hebrew Bible, only trained scribes and the educated elite of society would have been able to read the Hebrew Scriptures. 39 It is hard to gauge the situation described by the Gospels. On the one hand, the carpenter-rabbi Jesus is portrayed as being sufficiently familiar with the Hebrew Scriptures to be able to quote accurately from them whenever the circumstance required it. 40 On the other, the Gospel of Matthew presents those taught by Jesus as uneducated folk who were very much dependent on the teaching of the scribes and the Pharisees. 41 What does seem to be certain is that this uneducated crowd, the scribes and Pharisees, and Jesus, all believe that not one iota (the letter yodh?) or one keraia (pen-stroke or part of a letter?) of these scriptures has been (or, indeed, will be) misplaced. 42 This belief in a letter-perfect Torah contrasts with the evidence of Qumrān, in which we find that more than half of the scrolls do not accord with the proto-Masoretic recension and none of those that do are in any way near letter-perfect.
As evident in Neh. 8.1-18 and argued by Carr, Writing on the Tablet of the Heart.
So Mt. 4.4-10; 21.12-17; Mk 7.7; Luke 2.41-50; 10.25-37; and John 10.22-39.
In Mt. 5.21, 27, 31, 33, 38 and 43, Jesus introduces various quotations from the Mosaic Law with the words ‘You have heard that it was said', as if his hearers’ knowledge of these quotations is through a religious scribal elite rather than their own reading of these texts. See also Mt. 17.10; 23.13, 15, 23, 25, 27 and 29.
Mt. 5.17–20.
In sum, five features describe the oral-textual world of first century CE Palestine. First, written copies of Hebrew Scriptures consisted of consonants, so only people familiar with the Hebrew language were able to read them. Second, only a segment of society was trained in the writing and memorization of the Hebrew Scriptures. Third, the majority of the population was reliant upon these specialists. Fourth, it was widely understood that the Torah had one uniform letter-perfect standard. Fifth, contrary to this popular conception, variant texts were being used as scripture throughout Palestine in the first century CE.
According to Muslim tradition, the Qur'ān was revealed as an oral recitation to an ‘unlettered’ prophet (al-nabī al-umiya) 43 who dreamed that the Arabs might have their own Arabic book-bound revelation (kitāb) like that of the Jews and Christians, the people of book-revelation (kitāb). 44 According to Alan Jones, 45 Arabic scripts were in use in the seventh century CE, but these were so primitive as to require some of the letter symbols to represent multiple consonant letters. Early copies of the Qur'ān were, therefore, written in the sparse Kūfic and early Ḥijāzī scripts. These texts could only have functioned as memory aids to guide the reader who already had been taught (and memorized) the words of the Qur'ān orally. Only gradually in the eighth, ninth and tenth centuries CE were diacritical marks added to differentiate between different consonants, indicate vowels and mark pauses in the recitation. 46 The emergence of these marks may well have been influenced by the more complete and complex systems of diacritics that were being used on the text of the Hebrew Bible (or indeed vice versa). By the time of the emergence of these more functional texts, however, numerous traditions of recitation had already developed in the central cities of the Muslim Empire. In the tenth century, Ibn Mujāhid (d. 935) brought some order to the situation by declaring that only the seven most significant eighth-century traditions were legitimate for use in Muslim worship (plus at least three more associated to these seven). 47 This was consistent with a ḥadith in Bukhārī's ninth-century collection in which the Prophet declared, ‘Gabriel read the Qur'ān to me in one way, and I continued asking him to read it in different ways till he read it in seven different ways’. 48 In the context of the fundamentally oral identity of the Qur'ān this admission of plural tradition is remarkable. While by the time of the caliphate of ‘Abd al-Malik b. Marwān (r. 685–705) one basic consonantal text of the Qur'ān (the ‘Codex of ‘Uthmān’) had probably been established across the burgeoning Muslim empire, the variant vocalizations of this basic consonantal text would never be successfully suppressed. Instead, it would take another three hundred years to limit the oral variations to seven main recensions and this plurality was enshrined in Muslim tradition as an intended feature of the original revelation, that is, the revealed recitation (the Qur'ān). 49
The Qur'ān is emphatic that the Prophet Muhammad was unlettered (ummiyya: Surah Al-A'rāf 7.157-58) and that he was sent to teach the truth of the Qur'ān to an unlettered people (Surah al-Jumu'ah 62.2).
So Kenneth Cragg, The Event of the Qur'ān: Islam in its Scripture (London: Allen & Unwin, 1971); Daniel A. Madigan, The Qur'ân's Self-image: Writing and Authority in Islam's Scripture (Princeton: Princeton University Press, 2001).
Alan Jones, ‘Orality and Writing in Arabia’, in McAuliffe (ed.), Encyclopaedia of the Qur'an, III, pp. 587–93.
Efrim Rezvan, ‘Orthography’, inMcAuliffe (ed.), Encyclopaedia of the Qur'an, III, pp. 604–608.
Anna M. Gade, ‘Recitation of the Qur'ān’, in McAuliffe (ed.), Encyclopaedia of the Qur'an, IV, pp. 367–85.
al-Bukhārī and Khan, Ṣaḥīḥ al-Bukhārī, VI, pp. 427-29. Hadith 4991 is quoted; see also hadith 4992 and the nine different hadith quoted in support of this tradition by Al-Ṭabarī et al., The Commentary on the Qur'an, pp. 16–21.
In the modern period communication technology has led to the development of an international standard form of recitation of the Qur'ān (the recitation). The science of the study of the variant vocalization of the Qur'an (qirā'āt) has therefore become the preserve of specialists; see Gade, ‘Recitation of the Qur'ān’.
Whereas the Hebrew Bible was by the first century CE an established written scripture, the Qur'ān in first century AH was primarily an oral phenomenon. Only as the first generation of memorizers (ḥāfiẓ) were dying out did the Qur'an begin to become a written scripture—and then only partially so. Nevertheless, despite these differences, a first-century AH readers' experience of the Qur'ān as a written text did parallel that of a first-century CE reader of the Hebrew Bible in a number of ways:
The text of the Qur'ān was a skeletal consonantal text such that only people familiar with this system of writing and the language of its contents were able to read it.
Only a small segment of society was trained in scribal matters.
The majority of the population was reliant upon these specialists.
The specialists declared the scriptures to be letter-perfect and the population (and many of the specialists themselves) believed this to be true.
In reality, variant texts of the Qur'ān were widely used in the first century AH.
The Torah of first-century CE Palestine and the Qur'ān of first century AH Arabia are texts that must be understood in their own unique oral-textual context. Once their contexts are considered, it is possible to accept that a strong belief in a uniform standard scripture could have been accompanied by a remarkably wide variation in the manuscript traditions. In the case of the Qur'ān, Muslim tradition remembers that Muhammad himself recited each sūrah in seven different ways. Since the Qur'ān is essentially an oral phenomenon, this is a remarkable tradition. It teaches that there never was one ‘original’ recitation for each sürah. Even if these multiple recitations could have been captured by unambiguous consonantal texts, 50 Muslim tradition further teaches that the chosen text of the Qur'ān was an accident of history: one basic consonantal text chosen among many possible basic consonantal texts and one chosen almost at random purely for practical reasons. In sum, there is much in the history of the text of the Qur'ān to encourage the recognition and celebration of plurality in the Hebrew Bible.
On this see the extended discussion of the relationship of the seven recitation traditions to the codex of ‘Uthmān in the Tafsir of al-Ṭabarī: Al-Ṭabarī et al., The Commentary on the Qur'an, pp. 23–31.
Plurality features throughout the Hebrew Bible both in its content and in its vocalization. In regard to content, the Hebrew Bible is replete with parallel texts. 51 For example, Isa. 36.1–38.22 and 2 Kgs 18.13–20.19 both record the siege of Jerusalem in the time of King Hezekiah. In this example, each verse of Isaiah 36–38 corresponds to a verse in 2 Kings 18–20, but almost every corresponding verse has a minor difference—such as a variant spelling, a different preposition, an extra conjunction or a different word. In the parallel Isaiah 36–38//2 Kings 18–20, these differences probably occurred when these texts were copied either from each other or from a common source text. In many other pairs of parallel texts the verses do not correspond so closely. In some of these cases, it appears that the two parallel texts have used different source texts. Ulrich has shown this to be the case with many of the texts in Chronicles that correspond to texts in Samuel. 52 The classification and explanation of the many different kinds of parallel texts in the Hebrew Bible need not detain us here. The point here is that parallel texts are an expression of plurality—and the Hebrew Bible is replete with parallel texts. Furthermore, Jewish tradition and Jewish scribes have worked hard to preserve this plurality. Chapter 8 of the Talmudic treatise Soferim, for example, list the differences between the parallel texts Isaiah 36–38//2 Kings 18–20 and Psalm 18//2 Samuel 22 and forbids any scribe from harmonizing them. Quite why it is important to preserve the minor differences between these texts is never explained. It appears that plurality in the Hebrew Bible is worth preserving for no other reason than because it was perceived to be a part of the heritage of the Hebrew Bible. As a result, contra the practice of many translators and commentators, the canon of the Hebrew Bible does not offer one harmonized version of Psalm 18 and 2 Samuel 22 as scripture. Instead, the Hebrew Bible offers two texts, both of which are scripture and both of which are enriched by their comparison (not harmonization) with the other.
Abba Bendavid lists over five hundred; see Abba Bendavid, Parallels in the Bible (Jerusalem: Carta, 4th edn, 2010). The convention is to refer to variant repetitions in texts with different locations as ‘parallel texts’. Variation repetition within one textual unit is referred to as ‘parallelism’.
Ulrich, The Qumran text of Samuel and Josephus.
At other places in the Hebrew Bible, it does appear that two or three different traditions have been combined to produce a composite text. These composite texts, however, celebrate the plurality of their sources by retaining many of the distinctive features of their sources. This preservation of plurality enabled Julius Wellhausen to propose a theory of the historical evolution of the Pentateuch (Hextateuch). 53 Whether or not his history of the development of the Pentateuch is sustainable need not detain us here. The point here is that the compilers of the Pentateuch preserved these differences of style (such as the appellation of the name of God). It is as if the composers of the text are proud to have preserved elements of the different traditions in the composite text. The fact that these sources repeat the names of the people in quite different stories, 54 or repeat the stories in quite different contexts, 55 does not appear to be a problem for them.
Julius Wellhausen, Die Composition des Hexateuchs und der historischen Bücher des Alten Testaments (Berlin: Georg Reimer, 2nd printing with supplements, 1889); Prolegomena to the History of Ancient Israel (Cleveland: World Publishing, 1957).
For example, the name details of Gen. 4.17-24 are repeated in Gen. 5.9-31.
For example, the deception, ‘My wife is my sister’ is repeated in Gen. 12.10-20, 20.1-18 and 26.7-11.
In regard to the vocalization of the consonantal text, the Tiberian Masoretic Reading Tradition (TMRT) frequently recognizes a plurality of tradition. For example, in Exod. 8.12-14 the consonantal word k-n-m can be vocalized in two different ways: either as kinnim (‘gnats’) or as kinnām (‘lice’). 56 Against narrative sense, the TMRT alternates between both options throughout these three verses. A similar mind-set appears to have motivated the Masoretes to present paragraph divisions in the middle of twenty-five verses (pisqā bê'emṣa 'pasuq). 57 In these cases, the TMRT notes that other reading traditions have a major break in a position where it does not even have a verse break. A further unique example of plurality in the Hebrew Bible concerns the differences between its reading tradition and its consonantal text. In codices written after the ninth century, over one thousand qěrē–kětib (reading–written) notes direct the reader to read the text in a different way to that indicated by the consonants. 58 According to Robert Gordis, 59 in 20 percent of cases an otherwise unintelligible text is rendered intelligible, but in 70 percent of cases the reading tradition presents an equally intelligible alternative to an intelligible consonantal text. Furthermore in 10 percent of cases Gordis finds the reading tradition (qěrē) to be less intelligible than the written consonants (kětib) it supposedly ‘corrects’. Gordis concluded, therefore, that the prime purpose of the qěrē–kětib apparatus was to celebrate plurality in the reading tradition, not to provide corrections to the consonantal text. The reader's mouth and ears recite according to one meaning of the text and the reader's eyes read according to another meaning. In this way, plurality is presented and preserved and becomes part of the experienced of every ba'al qěriā.
In English translations, 8.16-18
The verses with this anomalous presentation are:
As found in the Aleppo Codex, Codex L (Firkovitch 1/19a, St Petersburg) and in most printed versions of the Hebrew Bible.
Robert Gordis, The Biblical Text in the Making: A Study of the Kethib–Qere (Philadelphia: The Jewish Publication Society, 1937).
David Livingstone was a man of intelligence, determination and faith. This remarkable combination of gifts enabled him to ‘discover’ the Victoria Falls, Lake Ngami, Lake Malawi and Lake Bangweulu for European civilization. This same remarkable combination of gifts, however, also made it difficult for him to re-evaluate the paradigms of his adventure, even when his own discoveries demanded it. 60
This understanding of Livingstone follows Tim Jeal, Livingstone (London: Heinemann, 1973).
Emanuel Tov is a man of intelligence, determination and faith. This remarkable combination of gifts enabled him to lead the Dead Sea Scroll research project. The three editions of his magnum opus not only witness to his encyclopaedic knowledge of the Dead Sea Scrolls and the textual criticism of the Hebrew Bible, they also witness to the turmoil of an educated man of faith grappling with this evidence and re-formulating his opinion. In the re-formulated opinion of Tov, research into the Dead Sea Scrolls informs the contexts from which the canonical text of the Hebrew Bible was forged. It cannot lead us to recreate a more ‘original’ form of the Hebrew Bible.
Zaid b. Thābit was most probably a man of intelligence, determination and faith. Muslim tradition has preserved little, if any, 61 information about his character, but much about his work. Muslim tradition does testify to the speed with which plurality became an integral part of the earliest versions of both the oral and the written Qur'ān. In this Muslim memory, the reason for the destruction of all the codices that differed from the ‘Uthmānic recension was not their unreliability, but the need for the community to be unified under one of the many written texts. If circumstances did permit the exploration of any of these codices, they would teach us much about the context in which the ‘Uthmānic codex was chosen. They would not lead us to one more ‘original’ version of the Qur'ān because its earliest form was ‘a series of determinative texts’… and memories, not one.
Despite frequent mention of his deeds, his biography is notable for its absence in the exhaustive account of Ibn Sa'd: Muḥammad Ibn Sa'd and Aisha Abdurrahman Bewley, Kitab at-tabaqat al-kabir: The Companions of Badr (Volume 3) (London: Ta-Ha, 2013).