Septuagint and Canon (3): Joosten

This is a third post in my occasional series on issues relating to the canon of Scripture and the Septuagint. See previously here. (Wow, has it really been a year since the last time I did this?)

Some time ago I mentioned the Alexandrian Canon hypothesis, noting that it is making a partial comeback in some circles, and I pointed to Jan Joosten's recent article as a part of this comeback.

Joosten signals his main point a few pages into the article: "To my mind, the hypothesis of an 'Alexandrian canon' ... has been abandoned over-hastily" (p. 690). (Joosten says that Alexandria does not matter to the theory, but Egypt is important.) He says this despite having just admitted that "the hypothesis of an Alexandrian canon lacks proof" in the way of external evidence: Philo does not cite the deuterocanonical books, other Egyptian Jews don't, and the NT does not either. This post will survey Joosten's argument.

Joosten begins by mentioning the 'extra' books in the LXX codices, describing "an irreducible core" consisting of Sirach, Wisdom, Tobit, and Judith. In a note he mentions also the books of the Maccabees as sometimes included, and that Baruch and the Epistle of Jeremiah would form part of the core, though they are not counted separately but as a part of Jeremiah. As for 1 Esdras, it was perhaps considered the Greek equivalent to Ezra-Nehemiah rather than an additional book.

I'm not sure what to make of the statement that "The outlines of a distinct Septuagint canon are recognized also in Patristic and synodic lists of the early Christian centuries" (p. 688). Joosten does not cite anything, except for a discussion by Bogaert (here) and another by Sweet. I guess Joosten is thinking of the canon lists that contain 'extra' books, but those lists are not in Greek, at least not early on, only in Latin, as a perusal of Swete 203–14 makes clear. Bogaert's article referenced by Joosten concentrates on the Latin evidence. To be clear, the early (first four centuries) Greek lists do contain some of these books, like Baruch and 1 Esdras, but Joosten has already told us (quite rightly) that these are special cases. The Greek lists do not contain the "irreducible core" of Tobit, Judith, Wisdom, and Sirach. I find a similar problem in Joosten's comment, at note 5, that in a different article (this one) "Bogaert has convincingly argued that the distinction of two categories within the larger canon--books that are also in the Hebrew canon and books that are not--is secondary and reflects an attempt to reconcile the two canons." Bogaert made this argument primarily in respect to the Latin evidence, not the Greek, though, to be sure, Bogaert does think that the evidence indicates that Christianity in general received from Jews a wide range of religious literature (I don't think he cites Sundberg's famous phrase in this regard) and only subsequently attempted to define the borders in relation to the rabbinic definition of the canon.

Joosten again: "The more extensive list is hardly of Christian origin" (p. 689). The note says that "this statement is true only in regard to the historical origins of the Septuagint canon, not to its explicit definition as a rule of the faith" (n. 6)--meaning, I think, that all of the books were written and valued by Jews, but they never actually developed a list of books (= explicit definition as a rule of faith?); this was done only by Christians. But the note continues: "But the collection as such can hardly have been drawn up first in the Christian church." I find this phrasing somewhat confusing, because if Jews "drew up" a collection, that seems pretty close to defining a list, which I think is what Joosten has already denied.

Joosten sees the problem thus: these books were Jewish writings, not quoted in the NT, but accepted within the wider canon of Christianity. Why? "It is hard to see why Christian groups would have selected them and added them to an existing canon" (p. 689). Joosten says most scholars believe "Christianity inherited the Septuagint canon from some form of Judaism." The common view (argued by Swete) had been the Alexandrian canon hypothesis, but this idea was attacked by Sundberg, which led scholars to turn away from it. (We saw in that previous post that the Alexandrian canon hypothesis was not widely accepted until the very end of the nineteenth century, only about 7 decades before Sundberg's attack.)

Here Joosten turns to his positive evidence for the Alexandrian canon hypothesis. "The hypothesis is favored by internal data ... The Septuagint canon is a Greek canon that could not possibly have existed in a Semitic version. It possesses a degree of coherence that characterizes it as a corpus" (p. 691).

What does Joosten mean by canon? "A canon implies a limited list of writings, but also a definite status, or function, within a given community. ... the list is usually clearly circumscribed ... In the present paper, the term 'canon' will be used in reference to the collection as such, while wider religious or theological implications will play a subordinate role only" (p. 691). I think what he's saying is that he can't (at least, not here) say exactly what status this collection held for Egyptian Jews, but his argument does depend on the idea that it was a clearly circumscribed collection of books, not a wide religious literature without definite bounds, as Sundberg argued. Joosten also insists that "a canon involves not only the number of books belonging to it, but also other features such as the sequence of the books, the inner organization of the corpus, and the names of the single writings" (pp. 691–92).

Joosten makes his case under several headings:

• "The Septuagint canon and the Greek language." He says "the question of language is in fact crucial" (p. 692). The different types of Greek literature in this corpus include both translations and original Greek compositions, and so this corpus could not have existed in a Semitic language. 
• "The Septuagint canon as a coherent corpus." Here Joosten explores "the way linguistic and intertextual links tie together its various parts" (p. 693). Examples: translations of isolated words--not the obvious translations for the Hebrew words--used throughout the Greek Pentateuch. Joosten admits exceptions, with some translation equivalents varying among the books. The Pentateuchal vocabulary then influences other translations. Moreover, the books translated later attest "the creation of intertextual references to the Pentateuch (occasionally also to other Septuagint books), sometimes against the Hebrew text" (p. 694; citing his earlier article and this book by Myrto Theocharous). Joosten's example is Sir 7:31, the Greek translation of which (against the Hebrew Ben Sira) creates a link with the Pentateuch (he points for more detail to this article, available here). The 'extra' books of the LXX also adopt the LXX Pentateuch vocabulary, but more than this, "the Prayer of Azariah in Dan 3 is made up of a dense web of allusions and references to earlier texts: Exodus, Deuteronomy, 2 Samuel, Ezekiel, Micah, Psalms" (p. 695, pointing to an article of his). Since the allusions rely on the Greek text of these previously translated books, Joosten argues that "the prayer was composed from the start in 'Septuagint Greek'" (more here). The books written originally in Greek exhibit the same reliance on the LXX; examples: Wisdom, Judith. Thus the conclusion: 

...the entire Septuagint canon, including the extra books, stems from a milieu where biblical books were studied intensely in their Greek translation. ... The literary coherence makes it unlikely that the Septuagint canon came about by random selection. It seems, rather, that the books making up the canon of the Greek Bible always belonged together. ... a process of conscious amplification. Perhaps it is even possible to speak of a form of canonical awareness: the post-Pentateuchal translators, supplementers and authors may have used the vocabulary and style of the earlier books in order to lend their writings an aura of "scripturality". (p. 695, italics original).

• "The Egyptian background of the Septuagint." The Pentateuch was probably translated in Egypt (see Joosten's article here), and the other books probably were too (see Tov's article), except for probably those books whose original translation is close to Theodotion or Aquila (Song of Songs, Ecclesiastes, Lamentations, Ruth, Second Esdras [= Ezra-Nehemeiah]). Some of the extra books can also be associated with Egypt: Sirach, Wisdom, Judith. 
• "A wider perspective." Although it is true that all Jews, no matter the geography, were hellenized to a greater or lesser extent, Qumran has also shown us the great diversity within Judaism at the time. The point is: Egyptian Jews could have had a Bible distinct from Palestinian Jews. 

He seems to think that this expanded canon has "by-passed" Philo, and it was transferred to Christians after the NT, which "reveals no trace of acquaintance with" it, so probably in "the early second century, when Egyptian Judaism was wiped out by the Romans and its intellectual heritage appropriated by Christian groups" (p. 698). The LXX canon has a distinct sequence and inner organization, as well as unique names for the individual books. Joosten says that future research is needed on exactly which books belong to it. The LXX canon might have lacked some books received in the Tanak, perhaps explaining the absence of an OG translation for Song of Songs, Ecclesiastes, Lamentations, and Ruth.

It seems to me that the major points here are *literary influence and *entrance into the Christian biblical codices. Joseph and Aseneth evinces all the features of Joosten's first three points above--composition in Greek, in Egypt, attesting interextual linkage with LXX ... but "it never became part of the Septuagint canon" (p. 699), meaning it does not appear in Vaticanus and Sinaiticus. The fact of literary influence in the direction of the Greek Pentateuch on later Greek Jewish writings I take as proven. But it seems that Joosten is using this fact to try to explain the existence of  Jewish Greek canon that I think remains unproven. He seems to think its existence can be assumed because of the "irreducible core" of "extra" books in the Christian codices, which makes the patristic evidence rather important in this theory. And it is the patristic evidence that I think is susceptible of different, and better, explanations. As one point: it is curious that patristic authors seem universally to think that Jews do not accept these "extra" books as Scripture. At least, we have all sorts of patristic comments to this effect, and none that I remember claiming that Jews do accept these books.