Loose Canons: Reflections on the Formation of the Hebrew Bible
Philip R Davies
University of Sheffield

1. Prolegomena

Study of the rabbinic-Masoretic canon confronts a number of conceptual problems which, despite a number of recent studies,1 persist here and there. Among the most important are the following:

  1. The persistence of the concept and evaluation of “canon” according to the Christian example of a closed and authoritative list. The role of canons within the system of rabbinic Judaism, as well as in other ancient and modern societies has not been fully evaluated.
  2. The inherited view of pre-rabbinic Judaism as having been, throughout its development, a unitary phenomenon, with a single line of evolution, and thus, implicitly, with a single canonizing process, leading to a single canon. This is often compounded by the teleological fallacy: that within the process of formation of a canon lie the seeds of the final canon itself, so that histories of canonizing begin with the final shape and work backwards, rather than starting from the beginnings and going forwards—as if the final shape of the canon were the outcome of an inevitable growth rather than being the result of discrete historical decisions.
  3. The assumption the scriptural canon provides clear and reliable evidence of its own history. One obvious example of this is to divide the history of canonization into the rabbinic-Masoretic divisions of “torah”, “prophets” and “writings”, without considering that different groupings may have been in force at earlier stages; more generally, there persists a tendency to accept canonical stories such as Ezra as being suitable evidence for the canonization of torah.
  4. Within biblical scholarship, it is rarely asked whether or not “canon” is a good thing; where the matter does get an airing, the answer is a ringing affirmation. Yet an ongoing controversy about whether or not canons do or should exist is raging in the field of English literature.2 Some critics are saying that the notion of “canon” is no more than an attempt by educational (we could read “ecclesiastical”) fascists to administer control of one’s culture (“religion”) and one’s society, to preserve the values of a powerful few against the interests of the less powerful many. Others counter that excellence cannot be relativized and must be recognized, and that canons do and should exist because they testify to the self-authorizing nature of excellence. But values do not lie in texts. Texts can only refract the values of writers and readers. Canons do not impose themselves.

Writing a history of canonizing, especially of Jewish canonizing, does seem to entail some kind of sacrilege wherever the canon is treated (as in the case of Childs’s work) as a religious icon. Canon is a Janus-like phenomenon, facing backwards through the process of canonizing that brought it into being, but also forwards in exerting a “canonical” influence on subsequent study of it. Biblical scholarship, after all, is largely conducted within or among communities for whom that canon is in some way definitional. Any history of canonizing, then, that we construct, though supposedly dealing with the backward-looking face, is being composed under the forward-looking gaze of a final, definitive and authoritative canon, which has helped to shape not just Christian communities, and not just the discourse of biblical scholarship (note the term carefully!) but also Western culture.

A “canonical” culture such as Protestant Christianity may thus harbor an aversion (whether or not conscious) to a process that necessarily treats the creation of its Bible as the result of a series of human decisions. For the discipline of history is about human decisions, about change, about design but also about accident. History shows how things might have been different but were not.

More threateningly, the postmodern view of history sees all histories as narratives rather than as objective representations of a “real” past. Canons are grand narratives, especially when they purvey canonized history. But a canon which is substantially historiographical has in fact worked its effect in Western culture largely as an account of a real past (and still does, even in critical scholarship). To critique this past by means of archaeology or literary criticism is standard in scholarly practice, but does not pose such a great threat to the status of the canon as does an approach that makes the canon itself a product of history rather than the reverse.

The effects of canon and of a secularizing critical history may also conflict in another way. Canons represent eternal, transhistorical values. Whether freezing forever a glorious culture or encoding the eternally valid words of a transcendent deity, they seek to defy or overcome the processes of history, in which cultures age and decay and in which languages shift the meanings of words. Between the perception of a permanent ideal reality and that of a constantly moving flux (between, we could say, Plato and Heraclitus) we are all caught, affirming at some time both one and the other, seeing eternal values behind the transience of our own historical experience. Even historians need to invoke some kind of universal and eternal principles such as providence, human nature or laws of social behaviour. But historians at their best infer these values from the study of individual, the particular and the discrete.

Whether or not any of the above considerations are deemed valid, it remains a puzzle that no attempt to understand or explain why Jewish canonizing took place or why Judaism ended up as a religion with a canon. It has generally been taken for granted that this should be so. The most basic historical questions of how and why have scarcely been addressed. To be fair, Haran’s recent book (cited in n. 1) does make such an effort. But in my opinion it is far too complacent in its use of canonical data as evidence.

2. What are canons?

Rather than begin by asking about the character of Jewish canons, we should take note that canons are an expression of several human cultures and available for comparative analysis. To understand Jewish canonizing entails understanding canonizing generally as human social activity.

The obvious starting point is the term “canon” itself, though the history of the term must not be confused with the history of the phenomenon. The term itself is Greek, and denoted a physical ruler (such as a carpenter would use for measuring) and an abstract standard (as we might nowadays say “yardstick”). Thus, kanon was accordingly used by Greeks to refer to the rules by which poetry or music could be composed, or geometrical shapes measured: so, for example, in the middle of the fifth century bce, the title “canon” was given to two textbooks dealing with sculpture. The apparently curious connection thus established between artistic creation and “canon” remained in Greek culture into the Hellenistic period, and the notion of a perfect work of art as representing the ideal, to be studied and copied, is fundamental to the Greek concept of canon. For the perfect work is itself a “canon” because it both enshrines and demonstrates the “rules” or the “art” in question (the eternal, as against the ephemeral). And since for Socrates and his circle, goodness was the supreme art (and knowledge the way to it), the good man can, as Aristotle says, be a “canon” of human nobility.3

Individual works or collections of works could be created in the Greek and Hellenistic world specifically as canons, and such works could cover a range of topics, whether art, medicine, technology or philosophy. These canons were neither exclusive or closed. The notion of a determinate set of books as forming a canon, which is fundamental to the Christian concept, may have originated in the treatment in Hellenistic schools and libraries of Homer, Herodotus and Thucydides as perfect exemplars of their art, from which it follows that the determination of genuine, as opposed to false, works assigned to these authors was an important issue, as was also the establishment of a reliable text. It was, indeed, probably under the influence of the scholars of the Alexandrian library, who used the word “canon” of collections of ancient authors, that the Christians notion of a canon derives.

Our modern use of the word “canon” has moved some way beyond its classical origins. Yet, if we want to approach Jewish canonizing from a historical perspective, we must ask ourselves what “canon” might mean in terms of Jewish writings, and return to the definitions that governed the earlier age. Indeed, we must go even further than the classical origins of the word “canon”. For, even though the word (or its equivalent) may not have existed, a process of canonizing is also clearly at work in both Mesopotamian and Egyptian cultures, and it is important to place Jewish canonizing historically in the wider context of the great literary cultures to which the classical world was also indebted. Millennia before the Greeks learned to write, the civilizations of the Tigris-Euphrates and the Nile valleys had produced highly complex bureaucratic systems in which the art of writing was indispensable: this in turn necessitated a society of scribes, and over time this society defined and replicated itself through a body of literature that served as a kind of genetic blueprint of its own values and world-view, its theoretical and practical philosophy. By means of its own educational system and the constant copying and refining of this corpus, the Mesopotamian and Egyptian civilizations produced, alongside the much more numerous but transient administrative literature which paradoxically has survived where so much creative literature has been lost, works that we would call canonical, even in the Greek sense.

3. The canonizing process

Canons are the products of canonizing processes. Canonizing generates all kinds of canons which, as the process continues, can assume different shapes over time or between different groups. The production of a single closed list of authoritative writings is not the inevitable end product of this ongoing process. Canonizing, however, is an inevitable by-product of a consciously literary culture. Before the age of mass production of books (i.e. printing), the accumulation of a literary corpus involves many stages: composition, copying, editing; but also classifying, collecting and archiving, since the growth of a corpus depends on its physical preservation. Copying and archiving are the very stuff of canonizing. A work does not become canonized by being included in a formal list: that is a final flourish. A work become canonized (a) by being preserved by copying until its status as a classic is ensured; and (b) by being classified as belonging to a collection of some kind. Scrolls can be canons in their own right, but multiple scrolls need to be archived: that means labelling and storing in a sort of order. This in turn entails collecting. The results are various canons, groups of classic texts or classic collections on scrolls or tablets.

A written document in the ancient world generally had a very short life. Unless written on stone or maintained by unusual climatic conditions, in order to be preserved it had to be copied, as it did, of course, in order to be distributed. The development of writing seems to have occurred for economic reasons, in order to enable the collection of taxes or mark property or verify transactions. It is a function of the development of an economic system. The earliest texts are administrative, and these were copied for an archive, where they were carefully preserved. They were of value for consultation and needed to be classified for retrieval. Such texts are called “documentary”. Literary texts, on the other hand, are not preserved as records to be consulted but as cultural artifacts, whose contents contain the stories, philosophy, laws or prayers that furnish the social reality. As with the society itself, these texts are living and organic: they may be copied with reasonable fidelity, but may also be altered by editing, supplementing, or combining with other texts. We have plenty of evidence to show that copying was rarely exact and often highly creative.

The fate of documentary texts and literary texts is basically different, even though they were sometimes archived together. Literary texts are transmitted—at least, those that survive. Transmission is a selective process: some literary works are copied; some not—at least in principle, for those not copied are lost and we cannot be certain. Some texts are rewritten and copied more than others. As this process takes place, some texts become more familiar, more ancient and more respected. Such works become quoted, and influence other works. In other words, some texts become what we would call “classic”. “Classic” works constitute a canon, even when that canon is not formally listed. The listing of such works can come about in a number of ways: there is cataloguing, necessary both for administrative archiving and for the maintenance of libraries of literary works. There is also curricular listing, in which certain texts form the basic of an educational syllabus. And there is also scholarly listing, in which genuine works of a certain type or author are distinguished from those judged to be inauthentic or inferior. These processes of discrimination and of formal selection constitute the core of the “canonizing process”, and what they produce are canons.

Thus, canonizing comprises a sequence of stages from the creation of texts, through transmission and discrimination to formal lists. Though one stage tends to lead naturally to another, so that we can speak of a sequence of processes, even when the production of final canonical lists does not result, we can, and should, speak of a canonizing process. The notion of a canon can be present without any definitive list (as it does in our own days). Indeed, the actual drawing up of formal canons is only a final stage in the entire process. For even before such lists are created, canons are created on shelves and in boxes, where literature of a certain kind or a certain value is grouped together physically. The tendency to issue canonical lists is in fact typical of a “post-classical” age anxious to control and define exactly the values of a culture that in reality is gone. In such circumstances we see a conscious effort to define and promote the study of a body of work. Such canons can have a long afterlife. The English public school educational system until not so very recently continued to treat the study of “the Classics “(Greek and Latin) in very much this way.

The “literate society” in which canons develop will, of course, differ over time and place, and even in societies of universal literacy, it is not the society as a whole that determines the canons of the future. Literary canons in the ancient world emerged in a specific (sub)culture, that of the “scribal” class. “Scribe” and “scribal” are not always ideal terms for this society, but they have the advantage of underlining the connection between economic activity and literary culture. Religious canons were also not the product of the body of adherents of a religion, but of those (rabbis, bishops) who identified themselves as the leaders and definers of the value of their “society” (their “religion”). Canonizing is elitist in conception and authoritarian in implementation. Canonizing may commence by trying (not even explicitly) to create a culture; but it typically ends by dictating a culture through the medium of a fixed list of what is and what is not canonical. It is thus an entirely open question whether or not fixed, closed and authoritative canons are a good thing at all. Perhaps it depends on how they are used. But typically they are imposed.

Finally, canons are not unilayered or undifferentiated. Some works can be more firmly “canonical” than others. In Greek literature Homer first, Herodotus and Thucydides next, and so on. Let us take the example of the rabbinic-Masoretic scriptural canon: here there are three division, one at least of which was at one time a canon in itself (torah). The torah has also enjoys a higher degree of authority than the rest of the canon, and sometimes the entire Jewish scriptures are called “torah”. In the Christian scriptures, the New Testament is a similar sense more “canonical” than the Old, providing the key to its “correct” understanding, abrogating many of its statements.4 Thus, the notion that a canon differentiates strongly between authorized and unauthorized, authoritative and non-authoritative, inside and outside, is only true up to a point. Neither within nor outside canonical boundaries is there equality. Canons do not grow up with rigid boundaries, and the creation of such boundaries by fixed canonical lists does not eliminate grey areas either within (“law” versus “gospel”, “torah” versus “prophets”) or on the edges of (the “apocrypha”) the canon.

4. Mechanisms of Jewish Canonizing

Historically we start an investigation of Jewish canonizing with the agrarian societies of the Iron Age in which literacy was a monopoly or near-monopoly of the class of scribe-administrators. Scribes were in large measure insulated from the majority of the population; physically (they lived in cities), economically (they were supported by the tax payer) and culturally. At a royal court, perhaps even a provincial governor’s court, traditional story-telling (the cabaret of the ancient world) may furnish a bridge between the two parts of the society; meeting-places such as the city gate, or the market also afford social contact and cultural exchange, and in general we must not rule out all meaningful contact between popular culture and the world of the scribe. The emergence of a significant artisan and merchant class during the Second Temple period afforded the opportunity for social class mobility (in both directions) and a medium for the negotiation of cultural values between peasants, “middle classes” and the governing classes. But these contacts played little part in the forming of the scribal identity. They may, however have modified it to a small degree.

The scribal duties, as has been seen, traditionally embraced a range of activities, amounting to a good deal of ideological control: archiving (possession and control of the present), historiography (possession and control of the past, didactic writing (maintenance of social values among the Èlite), predictive writing (possession and control of the future).5 The traditional ethos of the scribal class itself generated works of instruction, speculation on the meaning of life, social ethics, cosmology and manticism.

Hence, in Judah as elsewhere in the ancient Near East the scribes can be identified as “intellectuals” or as “sages”6 or as “the wise,” and especially responsible for “wisdom literature.” A succinct profile of the Judean scribe has been drawn by M. Weinfeld:

persons who had at their command a vast reservoir of literary material, who had developed and were capable of developing a literary technique of their own, those experienced in literary composition, and skilled with the pen and the book: these authors must consequently have been the soferim-hakamim..7

But Weinfeld is talking about the Deuteronomic school, which he places in seventh century Judah. Weinfeld assumes that at this time the Judean scribal class had reached the point of sophistication achieved in Mesopotamia and Egypt, and developed its own distinctive tradition. Such an accomplishment implies the existence of scribal schools, or at least an extensive educational system, in which not just the writing of Hebrew, but the reading of other languages, mastery of diplomatic forms, principles of archiving and so on would be passed on. At some point the number of scribes and variety of their functions makes the provision of a rationalized educational system inevitable. Mesopotamian and Egyptian civilizations possessed scribal schools. But whether in monarchic Judah such schools existed, or, if they existed, reached more than a rudimentary state, is in my judgment unlikely. What evidence do we actually have of literacy, of administrative complexity, and of scribal education in monarchic Israel and Judah?

There is sharp disagreement on this question. We can start by noting that scribal schools existed in the Hellenistic age, and E. W. Heaton’s recent discussion of Jewish schools, in which he comes to the conclusion that the canon is the product of the scribal school system, takes as its starting ben Sira and Qoheleth.8 He notes that ben Sira invites his readers to attend his school (bet midrash, 51:23), possibly even without payment (51:25). The range of topics in his book, however, makes it clear that he is not now training scribes, but offering an education to any who would acquire the Judean form of worldly wisdom, including the national literature, practical etiquette, sound ethics, piety, and so on. As Heaton says, the conservative scribal values “came to colour the whole ethos of educated society”, the “mobile middle class”.9

Certainly in the second century literacy was more widespread in Judah because of the advent of Greek culture, the growth of international trade, the emergence of a “middle class” and the growth of the administrative class (especially under the Ptolemies). But where education extends beyond the scribal school, it is still likely to be the scribes who educate. And so while it is pretty obvious that ben Sira is acquainted with Greek culture, his curriculum owes a great deal to the traditional scribal school and he himself had almost certainly enjoyed a career as a professional administrator. Heaton infers a tradition of scribal school education in Judah. So far, so good. But he then moves to a description of scribal education in Egyptian and Mesopotamian cultures, and posits the existence of libraries, including a “Temple Seminary”10 preserving the literary tradition of monarchic Judah beyond the demise of the state. This step entails quite a leap in time, and a considerable leap in social context. The appeal to Egypt and Mesopotamia is rather too superficial. First of all, the most obvious and direct parallel to ben Sira’s schooling is Hellenistic, and before automatically turning elsewhere this needs to be considered. Second, the scribal school system of Egypt and Mesopotamia at the time of ben Sira ought to be considered. In fact, it seems that in these cultures the old canonized texts were still being taught and copied, though the administration of Seleucid Babylon and Ptolemaic Egypt was being conducted in Greek (and Aramaic in Babylon). But both these civilizations had accumulated a vast canon of very ancient works. At the time of the Judean monarchy, both had been more advanced and more populous states and in existence over a very much longer timespan than Judah. It is entirely misleading to jump from second century bce Seleucid Judah over half a millennium backwards in time via “Egypt and Babylon”! But Heaton is hardly alone in this.

Evidence for a substantial indigenous scribal culture in monarchic Judah is slight. One of the indices of scribal activity is the complexity of state administration. David Jamieson-Drake11 has offered an anthropological approach to scribes and schools, based precisely on such considerations: a wide range of non-literary data and some sociological modelling: population size and concentration: luxury goods and monumental architecture, in an attempt to discover at what point in its development a state needs, and can sustain, an administrative class. He concludes that from the eight century on Judah became a fully developed monarchic state, but that literacy did not spread very far: all the writing is associated with government and thus with a specialized administrative class. We do have scribes, in several places. But on Jamieson-Drake’s analysis, no likelihood of literacy much beyond this not very large class.

What literature did the monarchic scribal class produce? There is certainly evidence of administrative texts from Israel and Judah.12 We have some ostraca from Samaria, totalling some 66 sherds, probably dating from 8th century bce and recording deliveries of wine and oil. Although they were not found in situ, but had been used as a foundation layer for subsequent building, they probably represent originally part of an accounting system, and presumably an archive. But the area the transactions cover is rather small (a radius of a few miles), which does not suggest a large archive or administrative staff. There are precious few royal inscriptions from the area, and none from Judah, unless we count the Siloam inscription (if genuinely from the time of Hezekiah;13 even so, it is not a royal inscription). The newly-discovered Tel Dan inscription(s)14 are not Israelite, nor is the Mesha inscription. We can assume that Israelite and Judean monarchs had scribes who could erect such inscriptions, but the lack of these remains an embarrassment to any theory of a large scribal-administrative class. Official correspondence has been preserved: the Arad ostraca (if Arad ever belonged to Judah) and the Lachish letters are probably written down by officials (both cities were sufficiently large to contain royal scribes). But again the evidence is not extensive enough to support the notion of a scribal class in the slightest degree comparable to those of Mesopotamia or Egypt, nor to cities of the size of Ebla, Ugarit or even Bronze Age Hazor (indeed, there were no cities of anything like this size in monarchic Israel or Judah).

In the Second Temple period, however, it seems that the Persians employed the Temple personnel to collect imperial taxes and deliver them to the imperial representative; as a result a more substantial scribal activity, combining imperial and cultic business, grew up. 15 The lists of officials the Chronicler assigns to David may be fictitious but it suggests that the Chronicler regarded an extensive scribal-administrative class as plausible. The view of Judah in the Persian period as a cultural backwater and as economically poor perhaps needs to be balanced against such a growth in the scribal-administrative activity of the temple-city. Literary activity in this case is not necessarily to be linked to the size of the state, since the state no longer exists as a political entity; instead Jerusalem is a provincial city.

Obviously in the Ptolemaic and Seleucid periods Judah’s wealth increased considerably, and the later we move in date the easier it is to conclude that the temple could sustain a number of scribal schools with a vigorous literary activity. Basically, the later in time, the better evidence we have for scribal activity. Those scholars who have objected that canonization could not have begin in the Second Temple period ignore the fact that even on their own views the canonizing of texts continued and increased during this period, so that the objection aims equally at their own view. Since canonizing in any case has to be attributed to the Persian-Hellenistic periods, I suggest that the onus of proof lies on those who want to argue that it could also have begun earlier. I do not think this possibility can be denied, but both the process of canonizing and its extent need to be demonstrated and not assumed.

As I have mentioned, the entrenched educational system of the scribal school broadened in the Hellenistic period (perhaps earlier), spreading its values to non-scribal classes; literacy spread, and the scribes themselves found a wider circle for their services, and, concomitantly, expanded their own intellectual interests to accommodate those of their widened intellectual circle. According to 2 Macc. 4:9–14 (cf. 1 Macc 1:14), a gymnasium and an ephebeion were introduced into Jerusalem, in 175 bce. No doubt they were already present in the many Greek-style cities already established (and many still to be built) especially on the Palestinian coast and in Transjordan, but including Samaria and Bethshean.16 According to 2 Maccabees, these were eagerly frequented by the priests especially. If the Hasmoneans officially disapproved of these institutions, they were either unable or unwilling (or both) to halt the spread of Greek education. But they in a position to foster the Hebrew language, create a Hebrew library, and, perhaps, encourage the development of a Jewish version of the Greek style of education. Between the scribal school and the later rabbinic school, whose aim was religious: to turn out good Jews, lie important developments of which we have too little evidence. It seems likely, however, that a specific emphasis on teaching Judaism(in its various forms) emerged, while some of the basic elements of Greek education (music, gymnastics) were discouraged. Given the indispensability of the Greek language, and the presence of so many Greek-speaking Jews both resident in and visiting Jerusalem, it is impossible to imagine that education for the priestly, administrative and ruling classes in Judah did not include many Greek elements.

The distinction between a professional education and a non-professional education entails a distinction between kinds of writing too, which is visible in the canonized literature. We can identity (or hope to identify) literary activity undertaken by the scribes in furtherance of their professional interests: writings that display the scribal ethos itself: historiographic, didactic, liturgical, and legal. Such writings, since they belong in spirit as well as in letter to the scribal class, lend themselves naturally to being canonized by copying, studying and teaching in the schools. Given the likelihood of specialization among the scribes, where different branches dealt with the temple cult, the temple liturgy, fiscal administration, diplomatic correspondence with Persian officials, and perhaps much else, we may be able to identify particular schools as the main agents of canonizing.

But not all the canonized books or stories come from a scribal milieu. Many stories—Joseph, Jonah, Ruth, Esther, Daniel—deal with questions of ethnicity, sometimes to the suppression of piety. They do, of course, diverge: for Jonah and Ruth, non-Jews are not to be shunned; for Esther and Daniel Jewish identity is something to be preserved from threatening foreigners, even though foreign rulers are not necessarily bad. Issues of gender, which have already been noted, may be related: the question of identity, which was identified as a matter of national importance, of class importance, in the torah and prophetic books, becomes a more personal matter. Alongside the personalization comes a personalization of piety too: what does mean for an individual to be a Judean, a Jew? We ought not to consider this purely a diaspora matter, for diaspora Jews did not write in Hebrew: it is a matter of ethnicity within Judah itself.

These stories ascribe little importance to the temple or cult. The visions of Daniel contrast sharply here with the stories. Jonah mocks it in his psalm: it has nothing to do with Esther or Ruth, and certainly Solomon’s antics in the Song are unconnected with his temple building. There are, then, a number of writings, many featured in this chapter, that betray an interest in individual identity. The factors promoting this are several. First, the reading classes for which the stories are told are concerned with their own individual careers: their fortunes depend less on co-operation with others. But in the wider cultural world they inhabit, their own social identity is important. It is, after all, a label they have to wear. Jonah, Esther, Ruth and Daniel all deal with the image of a Jew (or “Hebrew”) among non-Jews. In this they point not only a diaspora world but also to a Judah that is becoming much more cosmopolitan. Their travels, too, force them to face the question of their ethnic identity. Jonah, asked who he is: he answers “I am a Hebrew and worship Yahweh the god of heaven”. Precisely what that meant was what Judean schools would try to teach.

In what circumstances do such writings move on the path towards canonization? How is a hitherto scribal canon opened up to such works? Is it simply that they are widely read? There are two possibilities: one is that these works were used very widely in the school curriculum. Indeed, Jonah, Ruth and Esther are still used as college texts to teach classical Hebrew, because they are short and grammatically simple. Another factor may be a concern deliberately to loosen the control of one class on the canon and to sanction a wider range of literature held in the temple libraries.

5. Evidence of canonizing processes within canonized texts

We can find traces of the canonical process within the canonized texts of scripture. Several collections within the Psalms canon are headed “of David”. Whether or not these originally implied Davidic authorship, there are some Psalms whose headings explicitly make such a claim; while other writings from Qumran and the New Testament appear to assign the whole canon of psalms to David. We may argue as to whether this extension is logically speaking a canonical or postcanonical development. We can also see in Psalms evidence of smaller collections: psalms of David, psalms of ascent, psalms of the Korahites, etc. These form sequences that betray their collection as canons. The present Psalms collection itself is composed of five books, and the evidence from the Cave 11 Psalms scrolls may be interpreted to mean that all but the last book had been fixed into a canonical shape by the end of the 1st century bce. Similarly with instructional literature: Proverbs as a whole is assigned to Solomon, yet within the canon of instructional sayings are some collections assigned to others. The book of Ecclesiastes plays with Solomonic authorship—how seriously we cannot know—and may reflect the existence of a canon of “Solomonic wisdom” (though the instruction of ben Sira, however, makes no such claim, and yet there is evidence that it was included within some canons of Jewish instructional literature).

Psalms and Proverbs thus appear to be composed of once separate collections brought together in a single scroll. The process of writing them on one scroll had implications for archiving. The contents were given, in all probability, a single name. They would be copied, sooner or later, as a single composition. Hence the Psalms scroll is Davidic and the Proverbs scroll Solomonic. A similar process might be suggested for the scrolls of Isaiah and Jeremiah, both far too long for any serious refashioning as a single coherent “book”. Despite recent attempts to argue otherwise, it still looks likely that what we now call chapters 40–66 of Isaiah were at first written on the same scrolls as what we call chapters 1–39 (and archived as “Isaiah”?). The scroll represents the canon of Isaiah: copied and altered, supplemented and cross-referenced as a single entity, it becomes a “book” of Isaiah. A canon of Daniel stories may also be mooted: in this case supplemented by a series of later additions, this scrolls became the “book of Daniel” (in different Aramaic/Hebrew and Greek forms). Much the same process, I suggest, accounts for the canon of Enoch, now known as the book 1 Enoch, but once composed of four or five separate compositions.

Jeremiah’s scroll did not develop by juxtaposition, however, but by inflation. Had Lamentations been written on the same scroll, that scroll would have represented the Jeremiah canon and these poems would have become part of the “book” of Jeremiah. This did not happen. However, there is a prophetic canon represented by a scroll of twelve different composition. But because of the desire to distinguish the individuality of these prophets, they did not merge into a single “book”. A single scroll, but not a single composition.

At the other end of the scale, we also have multi-scroll canons such as the Mosaic books, while scrolls that contain accounts of a period of history will, under the guidance of a process seeking to create a single comprehensive history, become moulded into a sequential narrative. Once a single more or less coherent narrative is achieved, it can become canonical.

The point I am making here is that canonizing is a process that involves all the stages from composition, editing, archiving (combining on a scroll) and collecting scrolls into larger units. There is no single canonical mechanism, nor “trajectory”. There are “canonical processes”. However, and they operate within the formation of scrolls as well as in the grouping of scrolls.

6. Reflections

The impossibility of dealing with canonizing in the shadow of later lists can be illustrated by the following scenario. If we were to find in some church’s library in, say the second century ce, some codices of the Mosaic canon alongside a codex of some letters of Paul (let us say excluding Colossians and Ephesians), a scroll of Enoch and a codex of the letters of Ignatius, how would be decided which of these were canonical? We would have before us (a) a clearly recognized Mosaic canon (b) a collection of works that would be canonized in the Western “New Testament” but does not match the final list, (c) a work that was canonized but not in the Western church, and (d) a collection that was not later canonized. An illustration such as this shows not only how difficult it is to decide what “canonical” might mean at any given time or place, and indeed how inappropriate it is to allow the category “canonical” to get out of hand. “Canonical” does not imply only a fixed status in a list but can reflect a number of degrees of “canonization” prior to that. Even where it does make someone’s list, it may fall out of another’s.

Canonizing begins and continues as an open-ended process. To canonize a work is not an entirely conscious process at all stages and does not entail that other works have to be barred from being canonized, or definitely excluded from such a status. Only when definitive canonical lists emerge does the canonizing process stop. While canonizing does entail listing, organizing and labelling, a single definitive list is not, indeed, the purpose of the canonizing process, any more than death is the purpose of life: just its end.17


[1] e.g., J. N. Lightstone, “The Formation of the Biblical Canon in Judaism of Late Antiquity: Prolegomena to a General Reassessment,” Studies in Religion 8 (1979), pp. 135–42; David M. Carr, “Canonization in the Context of Community,” in R. D. Weis and D. M. Carr (eds.), A Gift of God in Due Season (JSOTS 225). Sheffield: Sheffield Academic Press, 1996:22–64; M. Haran, Ha-asuppah ha-miqra’it: t ahalike haggibush ‘ad sof yeme bet sheni veshinnue hatsurah ‘ad mots’e yeme habbenayyim {The scriptural collection: processes of crystallization up to the end of the Second Temple and changes in form up to the close of the mediaeval period,} Jerusalem: Bialik and Magnes Press, 1996.

[2] For a defence of such a canon (and a definitive list), see Harold Bloom, The Western Canon: The Books and School of the Age. New York & London, Harcourt Brace, 1994.

[3] For Greek canonizing, I have consulted H. I. Marrou, A History of Education in Antiquity Tr. G. Lamb. London: Sheed and Ward, 1956 and R. Pfeiffer, History of Classical Scholarship. Oxford: Clarendon Press, 1968.

[4] In Islam, while the Quran is a single-work canon, certain verses also “abrogate” others.

[5] Oppenheim (Ancient Mesopotamia: 230) classifies the purposes of Mesopotamian writings as follows: administrative recording; codification of laws; formation of a sacred tradition; for annals; for scholarly purposes.

[6] E. g. Gammie and Perdue (eds), The Sage in Israel and in the Ancient Near East., Winona Lake: Eisenbrauns, 1990

[7] M. Weinfeld, Deuteronomy and the Deuteronomistic School, Oxford: Clarendon Press, 1972: 177–78.

[8] E. W. Heaton, The School Tradition of the Old Testament, Oxford: OUP, 1994.

[9] The School Tradition: 13, 15.

[10] The School Tradition: 185.

[11] D. W. Jamieson-Drake, Scribes and School in Monarchic Judah, Sheffield: JSOT Press, 1991.

[12] For the corpus (excluding Tel Dan) see G. I. Davies (ed.), Ancient Hebrew Inscription Cambrdge: CUP, 1991.

[13] J. W. Rogerson and P. R. Davies, “Was the Siloam Tunnel Built by Hezekiah?,” Biblical Archaeologist 59 (1996): 138–49.

[14] A. Biran, and J. Naveh, “An Aramaic Stele Fragment from Tel Dan,” IEJ 43 (1993): 81–98.

[15] J. Schaper, “The Jerusalem Temple as an Instrument of the Achaemenid Fiscal Administration.” VT 45 (1995): 528–39.

[16] For a list and discusssion of these cities, see E. Sch¸rer, (rev. G. Vermes, F. Millar and M. Goodman). The Jewish People in the Age of Jesus Christ II, Edinburgh: T. & T. Clark, 1979, pp. 85–183.

[17] The argument advanced in this article is expanded in my forthcoming book in the series Library of Ancient Israel. [Editor’s note: Philip Davies, Scribes and Schools: the canonization of the Hebrew Scriptures. 1st ed. Louisville, Ky: Westminster John Knox Press, 1998.]