Tuesday, January 28, 2014

A Temple Collection? (part 3)

I'm finally back at this subject, which has, I admit, ended up being more complicated than I anticipated (what else is new?). These posts are my attempt to get a handle on the evidence and arguments for a temple collection of scripture, especially as that relates to the development of the biblical canon. The first post surveyed some of the ancient testimonia for such a collection, and the second post presented the argument put forward by Emanuel Tov in regard to the textual data arising from the Judean Desert scrolls.

I did promise some evaluation of Tov's arguments for our present purposes (biblical canon), so here goes.

First of all, Tov's views are shared by other scholars. For example, David Carr:
The proto-Masoretic texts that start to appear in the late Hasmonean period probably linked back to reference exemplars stored there [= the Jerusalem temple]. (Formation, p. 163)
Armin Lange also believes that a standard text was created in the second half of the first century BCE by priests in the Jerusalem temple ("They Confirmed the Reading," p. 79). Lange (Flores Florentino): accepts much of Tov's presentation, but adds: (1) diversity of biblical quotations (see now his book and the review of same at JHS); (2) Greek recensions; (3) Greek textual scholarship (but precisely this third argument was refuted in 1978 by Bertil Albrektson, pp. 48-51, citing Lieberman before him). Lange thinks Jewish standardization began in Egypt (evidence: Alexandrian pagan scholarship; LXX mss earliest examples of Jewish standardization; Aristeas) and thence influenced Palestine after 63 BCE. Originally it was not aimed at MT exclusively--"The proto-Lucianic recension as well as the supralinear corrections of 5QDeut demonstrate that originally other texttypes were also favored as standard texts" (p. 126)--but eventually the MT was housed in the temple, perhaps by the late 1st cent. BCE.

But Tov's hypothesis is speculative, as he admits. This is how I summarized his position in my previous post. 
If the non-Qumran Judean Desert texts reflect exclusively MT and if those who used these scrolls can be reasonably considered to have recognized the authority of the Jerusalem temple and its leaders--and to have obtained their biblical texts in reliance on this Jerusalem leadership--and if rabbinic writings testify to the existence within the temple of a scroll of the Torah and possibly other biblical books used for corrections, then perhaps the manuscript evidence confirms these rabbinic statements and we should connect the non-Qumran textual uniformity to the temple.
Tov here says "This is a mere hypothesis" (p. 9) and he goes on in the same passage to admit that the rabbinic evidence really concerns only a Torah scroll in the temple, "but it stands to reason that other Scripture books were also found in the temple." There's the rub for our question (= canon): exactly which scripture books does it stand to reason were to be found in the temple?

So, it's a 'mere hypothesis'. On the other hand: 
The textual unity described above has to start somewhere and the assumption of master copies is therefore necessary. (same article, p. 8; a nearly identical sentence at TCHB p. 31)
So, the assumption of master copies is an assumption, but a necessary one. The location of these master copies in the temple is more of a guess, but a guess based on good sense and on the ancient testimonia that we surveyed in the first post.

That doesn't mean everyone has to like the guess. Eugene Ulrich, for instance:
Despite suggestions to the contrary, the future still awaits demonstration that the texts preserved in the medieval MT transmit the texts guarded by the priests in the Jerusalem temple as opposed to other popular or "vulgar" texts that were less well preserved by less well qualified people. Nor has a line of succession--from temple priests to Pharisees to rabbis--been convincingly shown. (here, p. 155)
In the footnote on the same page, he writes:
Moreover, to my knowledge, no one has demonstrated how we could know either the textual nature of the priests' manuscripts in the Jerusalem temple, or how the Pharisees/rabbis, usually considered a lay group, would have received them in contrast to the (probably priestly) LXX translators and the Qumran leaders who were presumably very strict priests. (p. 155 n. 18)
What does that leave us with?

First, Ulrich seems to accept the idea that the temple in Jerusalem housed some scrolls of scriptural books. The ancient statements affirming this point--and the comparative evidence from other cultures--put the matter beyond dispute to my mind. That is to say, I think it would be very hard to argue that the temple contained no scriptural scrolls, and I'm not sure why someone would want to deny it. The exact content of this temple library can certainly be disputed: besides the Pentateuch, which other scriptures were included? The entire Tanak? More than the Tanak? And which textual form were these scrolls? Just one form or multiple?

Second, Tov seems to have established that there were master copies of some scriptural books that led to the production of scrolls textually equivalent to the later MT. As he says, the idea of master copies is an assumption but a necessary one. While Ulrich does in another essay question to what extent the Masada scrolls actually reflect later MT, he must also admit in regard to his principal example (MasGen, a fragment with 8 complete and 3 broken words, containing 3 variants vis-a-vis the MT): "such small variants are to be expected even within the Masoretic group" (p. 456). Lange also says that MasGen cannot be classified, and he says the same about seven Bar Kokhba era mss (Handbuch, p. 24). Nevertheless, Lange basically agrees that all of these non-Qumran texts are protomasoretic and that they thus stand apart from the texts at Qumran (ibid.).

Granted the existence of master copies (though I'm not positive that Ulrich would grant this point), where would these master copies be housed? The Rabbis possessed MT-like scrolls, and their predecessors are usually thought to have been the Pharisees. Ulrich seems to accept this in the quotation above, so let's go with it. I guess it's possible that the Pharisees were the ones at Masada leaving behind the biblical scrolls, and at Wadi Murabba'at and elsewhere (especially if a wing of Pharisees really were into violent revolution, as Wright has argued here, pp. 185-86, and here, pp. 190-95).

But what about all those 'second-circle' texts (as Tov calls them) at Qumran? These 'second-circle' texts are copies of the proto-MT (I know Ulrich rolls his eyes when he reads that--because I'm sure he reads my blog!--sorry, it's a useful though anachronistic term) that are not quite as precise representations of the MT as are the non-Qumran Judean Desert scrolls, but they seem to have been copied from one of these latter 'inner-circle' texts. That is, they were not copied from the master scrolls themselves but from copies of the master scrolls. While the actual scrolls are earlier than the 'inner-circle' scrolls, they are typologically later. (See the summary of Tov's argument in my previous post.)

Does not the presence of the 'second-circle' texts at Qumran indicate that the master scrolls, or at least the text represented by them, must have been important, or, let's say, available, to groups beyond the Pharisees? What's more, I think the numbers in which they are represented at Qumran would indicate not just their availability but truly their importance. While Qumran undoubtedly exhibits textual pluriformity, it is equally undoubted that the textual form that became the MT was important there.

I don't know that the master scrolls were housed in the temple, but that is a pretty good guess. I'm basing this on the previous two points: (1) the 'necessary assumption' of master scrolls that provided the model for (a) the non-Qumran Judean Desert scrolls, (b) the biblical text adopted by the Rabbis and apparently earlier used by the Pharisees, and (c) many of the Qumran scriptural scrolls at a secondary level (second-circle texts); (2) the presence in the temple of some scriptural scrolls, at least the Pentateuch if not more. Moreover, the Rabbis thought that the temple Torah scroll was used for correcting other scrolls that were more-or-less proto-MT (since that is the only text form the Rabbis seem to attest), so we apparently have some ancient evidence (granted, it's later rabbinic testimony and so open to doubt) connecting the temple Torah scroll to the proto-MT. But, given the presence of proto-MT texts at Qumran and the apparent animosity of the Qumran group to the Pharisees, they probably weren't adopting this text from each other. It seems necessary to imagine that the text was disseminated from another authoritative source.

In light of these 'second-circle' texts at Qumran, I would also want to call into question Ulrich's statement quoted above that the Pharisees somehow ended up with these standard scrolls and the group at Qumran did not. Or, again, from Ulrich:
If any group had temple texts that they preserved and copied, the Qumran group would seem to be the most likely candidate. Their early members are widely believed to have been priests in the temple who separated themselves because they believed the temple had been defiled. (here, pp. 155-56)
It's true that scrolls of the inner-circle were not found at Qumran (except for 4QGen-b; Tov, TCHB, p. 31), but plenty of 'second-circle' texts were. I don't know why the 'inner-circle' is almost completely absent, but the presence of the 'second-circle' is still significant.

[By the way, Armin Lange has his own classification of the Qumran scrolls (Handbuch, 16-17): he counts 20 semimasoretische Handschriften and 7 protomasoretische Handschriften, along with the other text types: pre-SP (2 scrolls), scrolls equally near MT and SP (11), LXX Vorlage (4), independent (47), non-classifiable (83), and some which might not actually attest biblical books (35).]

What biblical books are represented by the 'inner-circle' and 'second-circle'?

(See my previous post, especially the 'update' at the bottom. The following list of biblical books represents the books attested by the non-Qumran Judean Desert scrolls as give in Tov's Revised Lists, pp. 126-29. )

  • Genesis
  • Exodus
  • Leviticus
  • Numbers
  • Deuteronomy
  • Joshua
  • Judges
  • Isaiah
  • Jeremiah
  • Ezekiel
  • Twelve Prophets
  • Psalms
  • Daniel
  • Ezra-Nehemiah
Some of these are of course more impressive specimens than others. Not all of them have enough text to say for sure that they absolutely represent proto-MT and not other textual editions (see the doubts about MasGen expressed by Ulrich and Lange, noted earlier).

(The following list names the biblical books of the Qumran scrolls with exclusive closeness to MT as given by Tov here, pp. 154-57.)

  • Genesis
  • Exodus
  • Joshua
  • Samuel
  • Kings
  • Isaiah
  • Jeremiah
  • Ezekiel
  • Twelve Prophets
  • Psalms
  • Job
  • Proverbs
  • Daniel
  • Ezra
(Once again, I have a bit of a problem with Tov's statistics. Here, p. 22, he mentions "the exclusive closeness of fifty-seven Qumran texts to the medieval texts," but he doesn't list them. In the article cited above, he does list them but they only total 24 texts. He goes on to list other scrolls that are equally close to MT and either SP or LXX, which account for the remaining 33 texts.)

(Another caveat: some of these Qumran scrolls still present problems, even if the extant portions exclusively reflect MT. For example, the Book of Psalms is listed above in the 'second-circle,' but anyone familiar with the attestation of the Psalter at Qumran will know that it's more complicated than that. See Flint.)

If we put these two lists together, we find that all of the Torah is attested in this "standard" MT-like form, all of the Nevi'im, and five of the eleven Ketuvim. The books missing are the Five Megilloth (Ruth, Lamentations, Esther, Qoheleth, Song of Songs) and Chronicles.

If the books represented here--18 of the 24 books of the Hebrew Bible--were copied from master scrolls ('inner-circle') or from copies of the master scrolls ('second-circle'), then we can say that the text form represented by the master scrolls was available and important to the Qumran group and the Pharisees.

It seems to me that the most likely location for such master scrolls to be housed is the Jerusalem temple, especially given the ancient testimonia on precisely that point.

Remaining Problems

The textual diversity presents some difficulties, especially as Ulrich argues it. Not only does he point to Qumran, but he also argues that the SP, the LXX, the NT quotations, and Josephus "all resoundingly confirm this widely accepted state of pluriformity" (p. 155). That means that the evidence from Qumran is "representative of the Jewish scriptures generally in that period." I don't see that this overturns the possibility of a standard text, though it would limit the reception enjoyed by this standard text. Whereas we don't necessarily know what the Qumran group thought about the available textual options--we don't know, in other words, that they valued all these texts the same, as Ulrich wants us to think (p. 155)--the textual diversity exhibited by the LXX vis-a-vis the MT presumably meant that the translator valued his Vorlage over other options (though maybe even here other scenarios could be imagined; and for which books is there a real difference between LXX Vorlage and MT?). But there might still be a standard text promoted by an influential group, even if not all Jewish groups or individuals accepted or had access to this standard text.

Another problem is that MT is textually inferior in some books. This is especially true for Samuel, which is the typical example given in these discussions.
Is it at all plausible to picture the MT of the Books of Samuel as the outcome of a careful comparison of manuscripts and textual traditions, when it is obvious that it is an inferior text, full of errors and lacunae which could easily have been remedied with the aid of contemporary manuscript material? (Albrektson, p. 57)
[...] it is at the same time a good text--as a whole it is clearly superior to other textual traditions like the LXX or the Samaritan--and an uneven text with obvious and in places rather embarrassing defects. (Albrektson, p. 60)
The specific texts for each book in the rabbinic collection as reflected in the MT are, as far as we can tell, not selected or chosen but chance or coincidental. The poor state of the text, for example, of Samuel and Hosea, would seem to preclude conscious textual preference and selection; and the criteria for the choices of the MT versus the Hebrew Vorlage of the LXX could not have been the same for all books (cf., e.g., Jeremiah and Daniel). (Ulrich, p. 156)
I am not sure what to do with this fact. There were scrolls of Samuel circulating in antiquity with a higher quality of text than that of the MT. Why did not one of these higher quality scrolls become the standard text of later generations (i.e., the MT)? It is tempting to say, with other scholars (such as Albrektson and Ulrich) that this simply resulted from chance, that following the desolation of Judah enacted by the Romans, only this defective text of Samuel survived among the Pharisees/Rabbis. Does this undermine the entire idea of a standard text prior to the destruction? I don't think so, because Tov's assumption of master scrolls would still be a necessary assumption. How the text of Samuel works within this assumption requires further thought.


How does this relate to the canon of Hebrew scripture? To quote myself from the first post in this series:
the evidence makes it probable that the temple housed a collection of scripture, but it is hard (impossible) to know the exact contents of this collection, and it seems unlikely that it contained precisely the Jewish canonical books, no more and no less.
A collection of scripture in the Jerusalem temple that may have served the function of housing master scrolls for the dissemination of a particular text form of various scriptural books does give us a possible way of thinking about how the canon (to use again an anachronistic term) might have functioned in antiquity and how the ancient Jews might have conceptualized the canon. Since we are not in a position to know the precise contents of the temple collection, we cannot determine from this line of argument exactly which books counted as canonical or, indeed, whether anyone had given any thought to delimiting the scriptural collection in such a way.


John Meade said...

Thanks, Ed. I agree with the general line of reasoning which you present.

It's interesting that Qoheleth, Lamentations, Song of Songs, and Ruth are all commonly classified as members of the kaige tradition (not rescension). The Greek version of these books follows very closely a Hebrew Vorlage near to MT. I'm unfamiliar with the situation in Esther and Chronicles.

Do you think the evidence of the kaige tradition also supports the idea of master scrolls or a temple text?

Also, I know you have well formed ideas on canon, but I couldn't help but notice that you can't get away from using the word in your last paragraph, even if you say it is anachronistic. If what you have presented over your three posts is true, then how do we describe these master copies of biblical books? They appear to be authoritative and hence generally accepted and recognized as Scripture within broader Judaism. For me these points indicate a canonical consciousness, even if we can't say with certainty which books were in and which were out at this time. Absence of evidence is not evidence of absence. If what we are saying about the temple text is true, then the people would have known what books were in and which were out. Rather than saying canon is an anachronisitc term, why not say at minimum that we can detect a nascent canon from the period at which the evidence appears?

We may disagree on the last point, but I still want to affirm your work in these three posts. Well done!

Ed Gallagher said...


Thanks for your comments and interest in my thoughts on these matters. And thanks for reminding me about how the Kaige works in here. I did intend to highlight that a little more. I briefly mentioned it in the fine print when quickly surveying Lange's views, but Tov also mentions it in some places, e.g., TCHB p. 177: "However, this type of revision only shows that the same influential group that was behind the proto-Masoretic text was also behind the revision of the Greek Bible." So, you should be able to add Qoheleth, Lamentations, Song of Songs, and Ruth to the books for which there seem to be 'master scrolls'. Thanks for mentioning this point.

As for the term 'canon', I (sometimes) try to avoid using it for the pre-rabbinic period because scholars like to argue about its definition and its applicability to that time period. To a certain extent I buy into this idea that there was not (or might not have been) a canon in the strict sense. In other words, before Josephus we don't have the smoking gun that would demonstrate that someone could tell you what was in and what was out. But I think it would be next to impossible to deny a canonical consciousness in the Hasmonean era and even earlier. Even Ulrich allows for that (in his article "Methodological Reflections," linked in the post, p. 149).

I think of it like this: if you asked John Hyrcanus, 'which books are in the Bible?', I don't know that he could name them off and say that's it, no more no less. So Ulrich would say John Hyrcanus didn't have a canon. Fair enough. But if you asked him, 'is Genesis the word of God?', I think he would say, 'certainly.' Same for Exodus, Isaiah, Psalms, etc. If you asked him about Esther, I don't know what he would say, but 2Mac 15:36 makes me think he might have said 'yes.'

It sounds like at the end of your comment you might be trying to say that the temple library was equivalent to the canon of the time. (I might be misinterpreting you.) I'm not confident that we can draw that conclusion. In other words, I'm not sure that the existence of a temple library would have informed people of "what books were in and which were out," if you mean in and out of the canon. Beckwith (and Moses Stuart before him) assumed that no noncanonical book could have been kept in the temple library, but I think that is merely a supposition about what ancient priests "would have allowed"; it is not based on ancient evidence. I believe the ancient evidence would actually suggest that other types of material could also be kept in a temple library. Josephus is pretty explicit about priestly genealogical lists being kept there, and he can't be talking about the biblical genealogies because he says that they come down to the present day. (And I remember you asked me about that passage before. I'm still working on it. :) It seems improbable for a number of reasons that there would have been a copy of 1Enoch in the temple, but what about 1 and 2 Maccabees? Or Tobit? They might not have found a place in the temple library, but I don't think I can rule it out. Nor am I saying that Tobit, if it were in the temple library, must have therefore been considered canonical. I am trying to say that the temple library is not necessarily equivalent to the canon.

Nevertheless, if the reasoning I've presented in these posts (based on the work of Tov, Lange, and others) is cogent, then we can cautiously affirm the existence of a temple library that housed, among other things, master copies of the proto-MT form of many (maybe all) books of the Tanak. I would still suggest that is a gain, even if we can't say that the Hebrew canon equals the temple collection.

Thanks again for your interaction on this issue. It's a blessing to me.

John Meade said...

Thanks for your response. I've enjoyed exploring these matters with you (really after you:)).

As far as Hyrcanus is concerned, we don't know what his list would look like. For all we know, it could be the 24 books of the Hebrew canon that we know obtained. Equally possible, it could contain more books such as Sirach or Maccabees. Regarding Scripture/word of God, do you think Hyrcanus would call Sirach or Maccabees the word of God or Scripture? Are there any sources which would indicate a positive or negative response? Perhaps we simply have no evidence.

I will have to go back and look at Josephus' genealogical lists again. Clearly he is not thinking about these lists in the same way as the 22 books, but perhaps both types of books were kept in the temple, which would rule out a strict equation of temple books with canonical books. But the real question is were the deutero-canonical books housed in the temple? Yadin did find significant fragments of Ben Sira at Masada as well. Thus it is possible to say that this manuscript was brought to Masada by the Pharisees along with the other "inner circle" biblical texts.

A couple of things to investigate. (1) The Rabbis rejected the deutero-canonical books as evidenced as early as Aquila in the first part of the 2nd century. Are the Rabbis the heirs of the Pharisees? You seem to agree with Ulrich and Wright on this point.
(2) Can we guess, however speculatively, what traditions the Pharisees pass on to the Rabbis of the early 2nd century?
(3) Where did the Pharisees get their traditions from and how early?

We may not have any sound answers to these questions, but I wanted to raise them just in case you had answers.

Thanks again, Ed. I will continue to drop in to see what's cooking in that head of yours :). Blessings.

Ed Gallagher said...


Thanks for the comment and questions. Just to clarify, I only pick John Hyrcanus as an example because I wanted the hypothetical questions to be directed to a well-informed Jew of the second century. And I completely agree with you--he might have had a completely closed canon equivalent to the Tanak, but maybe not. We don't have the evidence to say. That's precisely one of the points I critiqued Lim for: it's fine to say that there was no canon in pre-rabbinic Judaism, but then you can't say that there were multiple "collections of scriptures" (= canons?) all disagreeing. It's one or the other: either they had (a) canon(s) or they didn't. I think the evidence points toward, "we can't say for sure, though most Jews definitely accepted as scripture most of the books (if not all) that the later Rabbis deemed canonical."

I assume that Hyrcanus would not have accepted Sirach or Maccabees as scripture, mostly because of their date. There is acknowledgment in 1Mac, for instance, that no prophet was current among the Jews, and that suggests to me that 'scripture' could not be written. A potential problem with this line of thinking would arise from the usual critical dating of Daniel, but (without engaging that issue) perhaps we could say that the setting of the Daniel within the Babylonian period, and its lengthy literary development (even on the usual critical view) allowed it to be accepted as authentic scripture, which it seems to have been even in the second century BCE (judging from citations of Daniel as a prophet in contemporary literature).

As for the deuterocanonical books in the temple, I don't think we can rule it out. It depends on the acquisition policy of the temple library, about which we have almost no information. Many have speculated about it. In my first post in this series, I cited both Beckwith and van der Toorn, who have opposing views. The basic questions: did the temple library house only books that were deemed sacred? Were the deuterocanonical books deemed sacred? For both questions I think we're just taking guesses, though perhaps the acquisition policies of other temple libraries in the ANE can inform our guesses (see van der Toorn). Another question: if the deuterocanonical books were housed in the temple, would there have been 'master scrolls' of those books? I don't know.

I do think the Pharisees were the predecessors of the Rabbis, as the Rabbis themselves seem to represent the matter, and I think this is generally accepted among rabbinics scholars. That does not necessarily mean that the Rabbis inherited a closed biblical canon from the Pharisees (after all, the Rabbis were innovative in their own right), but they may have done. Indeed, Lim and others assume as much.

Josephus is our best evidence for a relatively early Jewish canon (no surprise there); the way he presents it, there are very firm boundaries on the canon, and so it has been for a long time. But how many Jewish groups accepted the canon that Josephus references, and how much earlier than Josephus can we date it? Those are the million dollar questions. I'll continue to work on them.

John Meade said...

Ed, thanks again for your comments. In short, you are right about Josephus and we will need to continue to work on his statements regarding the canon.

It is interesting that Ben Sira was found at Masada. Would we need to affirm about it what we affirm about MasPsa and MasPsb for example? I would hold that the Psalms fragments from Masada are very important and are inner-circle texts. Indeed even the division of poetic sense units of MasPsa agrees with the much later Aleppo codex. It is proto-MT in the truest sense, I think. Therefore, do I need to affirm the same about Ben Sira? That is the sticking point and perhaps we have no more evidence to indicate one way or the other.

So is there no connection between considering a book "Scripture" and thinking it "canonical"? If Hyrcanus (to keep using the same example) could recognize the difference between Scripture and non-Scripture does this not mean he also had an idea of what is in the canon and what was outside of the canon? If it is only the dating of Daniel that is the counter evidence, we will need more than that, since that is simply more speculation and cannot be counted as an Archimedian point by which to gain leverage over the problem.

I realize that this is a different discourse than the Temple collection. But any conclusion of these matters will necessarily be based on a cumulative case.

Thanks again.

Ed Gallagher said...


About Masada: I think what we mean when we talk about the Psalms scrolls being 'inner circle' is that they conform very closely (are identical) to the later MT, not that they are necessarily canonical. Their canonicity is not the issue to which the category 'inner circle' refers. Rather, this term places them in a very carefully transmitted textual tradition. I would be interested if there were such 'inner circle' texts for Ben Sira. If there were, that would mean that some people cared very much about copying Ben Sira carefully; it would not necessarily indicate Ben Sira's canonicity in the minds of those who copied the scrolls. Today (and in antiquity) some scholars cared very much to copy Greek pagan literature accurately, for instance (think: Alexandrian grammarians), without implying any sort of 'divine authority'.

There is a relationship between scripture and canonicity, but the nature of the relationship depends on how one defines a canon. While I have sympathy with the insistence that one may have an 'open canon', I try to word my position in such a way that even someone who thought all canons must be closed by definition would still understand what I was talking about, if not completely agree. In other words, the existence of scripture does not imply canon, if we define canon as having closed borders, a product of a reflective choice (following here Ulrich). In that case, I don't know that Hyrcanus had actually reflected on it. Indeed, we don't have firm everyone that anyone had reflected on it until Josephus. His citing the number 22 is clear evidence that his canon had closed borders.

I'll keep working on Josephus, and maybe sometime we can compare notes. :)