When Geza Vermes first coined the term “Rewritten Bible” a half-century ago, I suspect he did not have any idea of the impact that term would make in Qumran studies. I also suspect that the phenomenon to which he applied the term seemed to him clearly defined and easily recognizable. It certainly has to me for much of the time since I first encountered the book of Jubilees and the Temple Scroll, nearly fifteen years ago in Bernard Levinson’s seminar on Scripture and Interpretation at the University of Minnesota. Rewritten Bible, for me, was simply a biblical text that had been revised according to a later interpreter’s own agenda; as Vermes put it, “In order to anticipate questions, and to solve problems in advance, the midrashist inserts haggadic development into the biblical narrative—an exegetical process which is probably as ancient as scriptural interpretation itself.” Comparing the book of Jubilees to Genesis or Chronicles to Samuel-Kings allows one to grasp immediately what Vermes describes. Granted, my work with the Temple Scroll made clear to me early on that Vermes’ definition, focused as it is on “haggadic development,” must be extended to include law as well as narrative. But the basic image of a Second Temple scribe interpretively reconfiguring the text to reflect a new set of perspectives and priorities remained.
I think many of my Qumran colleagues would agree with my sense that one of the recurring themes of recent scholarship across all sorts of subfields in the study of the Scrolls is the need to apply a “hermeneutic of suspicion” to our own scholarly constructs: in light of the full publication of the Qumran texts, we have come to realize that what we once regarded as intuitively obvious may in fact be more complex than we suspected. (I am thinking of issues like the Essene Hypothesis, the reconstruction of the community’s history, and the distribution of texts among the various caves, among many others.) The concept of “Rewritten Bible,” and with it my image of a lone scribal rewriter, has not been immune from this development. Discussion of the proper use and application of the term has flourished in the past decade in particular; one notable and widely-accepted proposal is that we speak not of “Rewritten Bible” but of “Rewritten Scripture,” in recognition of the absence of a fixed canon until sometime after 70 CE. My own recent work has complicated the idea of scriptural rewriting in other ways. In particular, my ongoing projects on the Temple Scroll and on the 4QPseudo-Ezekiel manuscripts have made clear to me the difficulty of recovering that one individual scribe and the particular scriptural text that scribe is rewriting. In other words, while the scribe/rewriter and the scriptural text remain the two poles around which the entire concept of Rewritten Scripture orbits, the Qumran textual evidence forces us to think about rewriting in a more complex way.
The first issue that complicates our attempts to describe and understand the rewriting of scripture is the question of any given rewriter’s scriptural Vorlage. We can only know the details of how rewriting works if we know what the text looked like before it was rewritten. In earlier periods of Qumran scholarship, not much thought was given to the question of whether this was something we could actually know: it was easy to assume that the Vorlagen of Rewritten Scripture compositions were more or less represented by MT or, at least, by other known versions. Some studies of particular rewritten texts did take into account the possibility that small differences from MT could stem from the Vorlage rather than originating with the rewritten text. But aside from minor variants of a “textual” nature, the form of scripture that the rewriters would have used was generally assumed to be accessible to us.
My recent work has made me increasingly aware of serious methodological problems with this assumption (which I should stress that I have shared to a considerable degree). In the last decade or so it has become clear that scriptural texts continued to be updated, revised, and expanded in the late Second Temple period, and that multiple versions of scriptural texts circulated simultaneously. These revisions could be quite major. We already know that the LXX version of Jeremiah is some 15% shorter than the MT version, and that both were current in the late Second Temple period. It wouldn’t be much of a surprise if we found a Rewritten Scripture composition based on LXX Jeremiah, or on MT Jeremiah. But what if we extrapolated this situation to imagine a case where a rewritten composition is drawing on an unknown version or edition of a given book?
For example, one of the most central features of the Temple Scroll (TS)—the one from which it received its name—is its inclusion of plans for a huge temple complex, equivalent to nothing found in the Hebrew Bible. According to the standard scholarly view, TS combines these temple plans (either of the author’s own devising or drawn from an existing extrabiblical source) with a rewritten version of the legal sections of the Torah. Yet the full publication of the Cave 4 texts brought to light a manuscript, 4Q365, that contains substantial verbatim overlaps with part of TS’s instructions for the temple courts, but otherwise includes texts from across the Pentateuch, with some revisions and additions. 4Q365 was originally classified as a “parabiblical” text, but in light of the full evidence I have recently argued that it is best understood as an expanded edition of the Pentateuch. The overlaps with TS suggest that the combination of a temple plan with pentateuchal law may not be an innovation by the author of TS at all, but rather a feature already present in TS’s scriptural Vorlage. If 4Q365 had never been discovered, we would have had no idea that a copy of the Pentateuch could include plans for a temple, and would likely have continued to locate the origins of the Temple Scroll’s temple either in the author’s own creativity or in some extrapentateuchal source. Though I do not expect universal agreement with my proposal that 4Q365 represents a copy of the Pentateuch, even the possibility that it might be raises the problem of our ability to identify the Vorlage of a given rewritten work.
Here I can mention 4QPseudo-Ezekiel to illustrate the broader implications of this example. Pseudo-Ezekiel presents shorter versions of several episodes known from the book of Ezekiel. Although the manuscripts are very fragmentary, it is clear that in at least one case, Ezekiel’s vision of the dry bones (Ezek 37:1–14), the scriptural text is given a new interpretive spin. Yet this interpretive perspective is introduced through the addition of new material before and after the account of the vision itself. In fact, though the version of Ezek 37:1–14 found in the 4QPseudo-Ezekiel manuscripts is not attested elsewhere, there is nothing in that version that seems directly to support the interpretation provided in the surrounding material. In light of what I’ve said above, it seems like we must raise the question whether we can necessarily attribute the shortening of the vision to rewriting on the part of the author/compiler of Pseudo-Ezekiel, or whether the author might perhaps have found this shorter text in whatever form of Ezekiel he used. There is, of course, no independent evidence that we know of for the existence of such a version. And it may be possible to mount an argument on textual grounds that Pseudo-Ezekiel knew a fuller version of Ezekiel similar to those known to us. What it doesn’t seem responsible any longer to do, however, is simply assume that the author of Pseudo-Ezekiel shortened the text; i.e., ignore the possibility that the shorter form of Ezekiel 37 was already present in the author’s scriptural Vorlage.
If my vision of an individual scribe working with his scriptural text is blurred by our inability to know what exactly that scriptural text looked liked, it is further obscured by the blurriness of the figure of the scribe himself (or herself—I can always hope!). As recent studies of the Damascus Document and the Serekh materials have shown, works later included in the Hebrew Bible were not the only Second Temple texts that underwent multiple revisions and redactions. What happens to our understanding of rewriting when the rewritten text itself may have come to be only in stages? Again, the Temple Scroll provides a particularly knotty example.
Here I am not referring to the still-popular theory that TS was composed by knitting together various preexisting sources, of which I remain skeptical. The manuscript evidence points towards a more complex reality. Alongside the main copy of TS, 11Q19, we have one additional manuscript (11Q20) that, where it is extant, differs little from 11Q19. But a third copy, 4Q524, presents challenges. Although it is very poorly preserved, some of its fragments clearly attest a text that differed very little from 11Q19 (e.g., frs 6–13). Yet one substantial fragment, fr 25, does not overlap at all with 11Q19. Another, fr 14, matches the text of 11Q19 col. 64 for several lines in a section where TS has made substantial changes to the text of Deut 21:21–22. After that, however, 4Q524 differs from 11Q19, jumping to Deut 22:11 while 11Q19 continues with Deut 21:23. In other words, the manuscripts share unique material, but at the same time their demonstrable differences indicate that one cannot simply be a copy of the other. The evidence, slight as it is, thus suggests that the Temple Scroll must have had a substantial literary history of its own.
If TS itself went through multiple stages of redaction and revision, we must adjust the way we think about its rewriting of scripture. (The same point could be made about other rewritten texts. For example, a second manuscript of Pseudo-Ezekiel, 4Q386, has a slightly longer version of the dry bones vision than 4Q385.) Instead of the lone interpreter reconfiguring scripture according to a single set of goals (or even stitching together pre-existing sources!), we must allow for the presence of multiple hands growing the text over time, and multiple agendas superimposed on one another. The problem, of course, is that the evidence is so scanty: if we possessed full manuscripts of both 11Q19 and 4Q524, for instance, it would be much easier to assess their relationship to one another. Arguments could be made on the basis of material unique to one or the other about the particular approach to rewriting evident in each as well as the nature of the rewriting in the material common to both. As it is, we can usually only operate around the margins of these questions. As the figure of our solitary scribe (here, “the author” of the Temple Scroll) is blurred, we sense other shadowy figures alongside—but we don’t have the evidence that would force them into focus.
These examples have forced me to admit that we must largely abandon the idea of a dyadic relationship of rewriter (the new text) and rewritten (the old text), both of which are known to us. I don’t think this means we have to give up on saying anything meaningful about rewriting in general or about the functions and purposes of rewriting in any given text. We may, however, need to be more willing to live with uncertainty regarding the stage or context in which a given reading emerged. Minor variants, even when they do have some interpretive meaning, are the most difficult to place, since they could have been introduced at any stage of transmission of either a rewritten work or its scriptural Vorlage. We need to recognize the pervasiveness of such minor changes and resist the temptation to explain every slight adjustment as an exegetical move on the part of “the” rewriter. At the other end of the spectrum, even very large revisions or additions could have arisen at multiple points or could involve several stages of growth, as the evidence of 4Q365 mentioned above illustrates. The most compelling evidence that would help us identify various parts of a rewritten work as the product of a single scribe would involve repeated occurrence of distinctive terminology, exegetical concerns, or literary features (such as voice or setting). Yet manuscript evidence shows that even these patterns can be extended or mimicked. In the absence of such evidence, of course we can tentatively group such cases as the work of a single scribe, but it would be irresponsible in my view to lose sight of the fact that the reality may be more complex.
All this need for humility and restraint and lack of certainty is definitely frustrating—there are many days when I would be happy to return to my earlier blissful ignorance of the complicated textual histories of the rewritten texts I work on, textual histories that leave traces of their existence but that we can do little more than acknowledge, hardly describe. But, to end on a more optimistic note, the reality ultimately excites more than it frustrates, because a new image of Second Temple scribes comes into focus at the same time as my individual author, and the Vorlage that author used, get fuzzier. In light of the evidence from Qumran, it is becoming clearer and clearer that a scribal approach in which serious intervention into the text was sanctioned, or even expected, must have been pervasive. We see it in text after text, from copies of books that later entered the Bible to Rewritten Scripture texts to sectarian rule documents. It’s not that nobody was copying without making changes—the numerous preserved copies of books very close to their MT form makes that clear—but neither does scribal intervention seem to be an activity restricted to a few. Whether there were particular sociological or institutional factors that may have influenced this activity remains unclear. But the Qumran finds and the light they shed on their Second Temple context, rather than allowing us to assume that texts were copied faithfully unless we have explicit evidence to the contrary, suggest the opposite: that we should assume texts continued to grow and evolve unless we have clear evidence to the contrary. What we have in most cases is likely snapshots of texts still in motion, that likely stayed in motion at least until the Jewish Revolt.
Numerous scholars have called attention in various ways to this fluidity that characterizes many Second Temple textual traditions. Now is the time to begin to integrate this insight and its methodological implications more fully into discussions of the techniques and purposes of scriptural rewriting. Documenting and explaining Rewritten Scripture may be more complex than before—we now see shadows of many more versions of scripture and many more rewriters of it than we once thought. At the same time, all those versions and all those rewriters testify to the significance of the phenomenon, and to the importance of our continued attempts to understand it.
Dr. Molly Zahn is an Assistant Professor in the Religious Studies Department at the University of Kansas. She is the author of Rethinking Rewritten Scripture: Composition and Exegesis in the 4QReworked Pentateuch Manuscripts (Brill).
All content provided on this blog is for informational purposes only. The American Schools of Oriental Research (ASOR) makes no representations as to the accuracy or completeness of any information on this blog or found by following any link on this blog. ASOR will not be liable for any errors or omissions in this information. ASOR will not be liable for any losses, injuries, or damages from the display or use of this information. The opinions expressed by Bloggers and those providing comments are theirs alone, and do not reflect the opinions of ASOR or any employee thereof.
For example, 4Q158 rewrites a pre-Samaritan version of Exodus in part by using precisely the same technique in response to the same type of exegetical issue as in the pre-Samaritan version itself; see Molly M. Zahn, Rethinking Rewritten Scripture: Composition and Exegesis in the 4QReworked Pentateuch Manuscripts (STDJ 95; Brill), 37–40.