John Barton’s Oracles of God: Perceptions of Ancient Prophecy in Israel after the Exile has recently been reissued after 21 years (Oxford: University Press, 2007). Though called a “New Edition” on the title page, it is new only by including a “Preface to the Second Edition” by Barton, in which he says that this is a reprint (except for errors) of the 1986 edition.
Last year also saw a special article of The Journal of Hebrew Scriptures dedicated to the reissue of the book, with contributions by Ehud Ben Zvi (as the editor), Philip R. Davies, James Kugel, Hindy Najman, and Barton himself, whose response to the other scholars here comprises the aforementioned “Preface to the Second Edition” in the book, with only minor changes.
Barton’s book is clearly important, though as he says in his new preface, it has not impacted scholarship on the reception of the prophets as much as he had hoped. It has had more influence in discussions of the canon, for which it is often cited as an early articulation of the view that the formula “The Law and the Prophets”, which appears often in literature of NT times, references the entirety of Scripture, and not just the first two sections of the Hebrew Bible, which is now divided into Law, Prophets, and Writings. In other words, the “Prophets” in this formulation constituted not just the second section of the Hebrew Bible (as it is now divided), but all non-Pentateuchal literature, including all the literature now found in the Writings (e.g. Psalms, Chronicles, etc.). Though the traditional theory of OT canon formation connects the three sections of the Hebrew Bible to three periods of canonization, with the Law canonized first, then the Prophets, then the Writings, Barton says this is anachronistic. The canon was bipartite, not tripartite, in NT times, as witnessed by the formula, “The Law and the Prophets”.
If this is so, how did it come about that the Hebrew Bible is now tripartite, as it has been at least since Talmudic times (see Baba Bathra 14b–15a)? Barton’s answer is that the creation of the third section, the Writings, is connected to the development of the liturgy (see Oracles, pp. 75–82). In synagogues today, there is regularly a public reading of the Torah, followed by a reading from the Prophets, i.e., the second division of the modern Hebrew Bible. These readings from the Prophets are called haftaroth. Barton’s suggestion for the development of the tripartite Bible is that all those books that were not included as haftaroth were relegated to the newly formed third section of the Bible, the Writings. This leaves only what we now think of as “The Prophets” (in the Hebrew Bible) in this section which formerly encompassed all non-Torah literature. Barton is able to cite (p. 78) as precedent for this view Sid Leiman’s The Canonization of Hebrew Scripture (1976; p. 168 n. 287).
But Leiman and Barton give different answers to the question: “On what basis were certain books chosen for haftaroth readings?” Leiman says that those books that tell of Jewish national history were included in the weekly readings, an explanation Barton dismisses rather easily (p. 79). Barton’s own solution is to suppose that the haftaroth readings were taken from books that were more commonly available in local synagogues, and those books that hadn’t quite established themselves as sufficiently important to possess were left aside. He writes on p. 79:
By New Testament times the scrolls of the ‘Deuteronomistic History’ and of the three great prophets and the Twelve were, we may suppose, widely known, and all synagogues would aspire to possess copies. Later books, such as Chronicles or Daniel, were becoming known but were not yet common property.
He gives a similar explanation as to why early Christians quoted so often from particular books, such as the Psalms and Isaiah. From p. 148:
The horribly simple explanation that their preference had something to do with the distribution of scrolls of these two books—worse still, that these were the only two non-Torah scrolls that happened to be in the book-cupboard of the synagogue at Nazareth or Capernaum—cannot be discounted.
I cannot now evaluate Barton’s position in full, and it may well be that the distribution of particular scrolls had something to do with the development of the haftaroth readings and with Christian preference for Isaiah and the Psalms, but I would be hesitant to say that this was a significant factor for at least two reasons.
First of all, as the second quotation from Barton above makes clear, the Psalms were widely known and used in early Judaism. Why would they not, then, be included in the haftaroth readings, if the decisive factor for inclusion in these readings was the availability of the scroll? The inadequacy of Barton’s explanation with regard to the Psalms is the more apparent since the psalms themselves are so clearly “liturgical”, and they were regarded in some ways as prophetic (see Barton, p. 40), meaning that they could legitimately stand among the Prophets, at least as legitimately as the Book of Judges.
Secondly, the evidence available for book distribution in the first century indicates that none, or very few, of the documents now in the Writings were unavailable to large numbers of Jews. David Goodblatt has recently examined the distribution of biblical scrolls in first century Judah, and has arrived at results that are startling, even to him. (See his Elements of Ancient Jewish Nationalism , chapter 2: “Constructing Jewish Nationalism: The Role of Scripture,” pp. 28–48.)
Goodblatt’s evidence includes the scrolls found in the Judaean Desert (mostly in the caves around Qumran) analyzed according to the survival ratio of ancient texts. A survival ratio allows us to estimate the number of texts that would have existed in a particular time and place based on the number of texts that we now possess from that time and place. Goodblatt says that a survival ratio for first century Judah of 1:5000 (one extant text representing 5000 that did not survive) would be rather conservative. Given that we have about 900 scrolls from the Qumran library, we can estimate that tens of thousands of biblical scrolls circulated in first century Judah. “Even taking into account the fact that these copies span three centuries of production, these are still astronomical numbers” (Goodblatt, p. 45).
These results make Goodblatt somewhat uncomfortable, but not enough to dismiss them. “These extraordinary results suggest that the survival ratio we are using must be off kilter. But unless we are willing to assume that the Qumran collection constitutes a large percentage of all the scrolls in circulation in first-century Judah, as the Golb thesis might allow, then these findings suggest the existence of thousands of scrolls in the country” (p. 45).
This indicates that Barton’s proposal for the reason that certain documents were chosen for the haftaroth is inaccurate. Since every book of the current Hebrew Bible was found at Qumran (except for Esther and Nehemiah), we can estimate that thousands of copies of these books would have existed in first century Judah. To say that certain books were relegated to the Writings because they were not widely available, as Barton does, fails to take account of these data.
This is the more true if one accepts Goodblatt’s explanation for the large number of texts:
And such a large number in a small, predominantly nonliterate population would make widespread public recitation much more likely. (pp. 45–46)
The archaeological evidence from late first century (Qumran and Masada) and early second century (caves with refugees from the Bar Kokhba revolt) Judah thus suggests that biblical scrolls were fairly plentiful and widely diffused. Why were so many texts needed in an overwhelmingly nonliterate society? The most probable explanation is that many of these manuscripts, like many or most ancient books, were performance texts. (p. 47)
Goodblatt proposes that the large number of scrolls is indicative of their use in the public reading of Judaean synagogues. This is obviously the case not only for those books now included in the Prophets, but also for those now included in the Writings. If these documents really were so widely available as Goodblatt’s evidence suggests, and used in the way Goodblatt suggests, then Barton’s idea that the haftaroth were chosen based on availability is shown to be false.
This does not entail that Barton is wrong in thinking that the tripartite canon arose in connection with the liturgy, which I am inclined to accept.[*]