Current thought about the Documentary Hypothesis

(From the Introduction to Empirical Models for Biblical Criticism, Jeffrey Tigay, Editor)

Since the nineteenth century, the Documentary Hypothesis has been the best-known, most published, most often criticized, most thoroughly defended, and most widely taught explanation of the development of the first five books of the Bible.  Evidence and arguments in support of it have grown continually more substantial—not just in quantity but in their nature, grounded more in demonstrable, quantifiable data from linguistic, literary, and archaeological products.  Despite this, there have been a plethora of new models and variations proposed in recent years—some with substantial followings—and the number of claims that the Documentary Hypothesis has been overthrown has grown.  Those among our colleagues who assert these claims have never challenged—in fact, have rarely mentioned—the main evidence and arguments for the hypothesis.  They especially do not go near the newest strong evidence, namely: (1) linguistic evidence showing that the Hebrew of the texts corresponds to the stages of development of the Hebrew language in the periods in which the hypothesis says those respective texts were composed; (2) evidence that the main source texts (J, E, P, and D) were continuous, i. e. it is possible to divide the texts and find considerable continuity while keeping the characteristic terms and phrases of each consistent; and (3) as this book shows, evidence that the manner of composition that is pictured in the hypothesis was part of the literary practices of the ancient Near East.

To its earlier critics, the problem with the Documentary Hypothesis was that it was too hypothetical: a clever construction, perhaps, but without enough solid evidence.  Two things have changed that situation over the years: the first is this acquisition of new evidence, and the second is advances in method.  This book, Empirical Models for Biblical Criticism, stands out because it features both.  The authors have brought new texts into the picture, which sheds light on what was or was not possible in the literature of the ancient Near East.  And, at the same time, they have exercised a level of methodological sophistication that was not matched in the early stages of the field.

One of the early attacks on the Documentary Hypothesis was that no other works were ever composed in the manner that the hypothesis attributes to the Pentateuch.  The Torah of Moses, this argument went, is being pictured as a “crazy patchwork,” having a literary history that has no parallel in the ancient Near East or in subsequent ages. Empirical Models for Biblical Criticism responds to this attack.  Jerffrey Tigay and his four colleagues, Mordecai Cogan, Alex Rofé, Emmanuel Tov, and Yair Zakovitch, present what the book’s title promises: empirical models—analogous cases of composition—from the ancient Near East.  They bring parallels to the manner of composition that the Documentary Hypothesis attributes to the Hebrew Bible, including parallels from the Epic of Gilgamesh, Qumran, the Septuagint, and the Samaritan Pentateuch.  In an appendix, Tigay also includes George Foot Moore’s 1890 essay treating Tatian’s Diatesseron—the work that combined the New Testament’s four Gospels—as another, manifest parallel to the combining of the four pentateuchal sources in the Hebrew Bible.

Tigay’s own contributions to this volume, in particular, provide a number of cases of demonstrable parallels in works from the ancient Near East.  Notably, he reviews the stages of composition of the Epic of Gilgamesh. These stages are documented, thanks to the existence of copies from several periods, and so the development of the work can be observed.  Tigay and his fellow contributors also bring examples of conflation of stories in the Qumran scrolls (as in the case of the conflate text of the Decalog in the All Souls Deuteronomy Scroll); in the Septuagint (as in the Bigtan and Teresh episode in the book of Esther); in the Samaritan Pentateuch; in postbiblical literature (as in the Temple Scroll); and even in modern works.  They show where literary conflation in these texts results in inconsistencies and vocabulary variation like those in the Hebrew Bible.

Not only do these ancient works involve doublets and conflation of texts in the manner of the Torah.  They involve the same editorial techniques.  We find, for example, cases of epanalepsis—also known as resumptive repetition orWiederaufnahme—that is, cases in which an editor inserts material into a text and then resumes the interrupted text by restating the last line before the interruption.  This is a well-known phenomenon in the Hebrew Bible as well.  (See, for example, Exod 6:12 and 30, in which the words, “And Moses spoke in front of YHWH, saying, ‘. . . how will Pharaoh listen to me?  And I’m uncircumcised of lips!’” come both before and after a long text that has signs of having been added to the chapter.)  Also we find different names for the same character in different accounts, differences of vocabulary and style in components of a text stemming from different sources, and cases of incomplete revision of reformulated texts. Tigay demonstrates that literary conflation in these texts produces inconsistencies and vocabulary variation like those in the Hebrew Bible.  Emanuel Tov as well reveals cases of homogenizing additions by editors.  Tigay also observes that “a comparison of [four different] stages [in the evolution of the Gilgamesh Epic] reveals a pattern of decreasing degrees of adaptation of earlier sources and versions.”  This, too, parallels the development of the Torah, in which the exilic Deuteronomist (Dtr2) and the final Redactor (R) can now be shown (in my own research and that of others) to practice substantially less adaptation of their sources than did the earlier redactor of the J and E sources (RJE).1     What all of this book’s contributors are doing is uncovering how the editors worked.  We have come a long way in this understanding.  It is now possible to have a sense of the individual editors: their differences, their preferences, their methods.  That is, we have reached a stage of acquaintance with the editors.

These analyses provide the visible, empirically-documented parallels that the challengers of the Documentary Hypothesis asked.  But a point that we should all note here is that, even before these new analyses came to light, this argument was really no argument at all.  A lack of analogies to the Torah’s literary history was never a case against that history’s reality.  As Tigay puts it, “The reluctance of these writers to contemplate the possibility of something unique in Israelite literary history does not commend itself.”  Indeed it does not.  A vast body of evidence led to the identification of distinct source works and layers of editing in the Five Books of Moses.  Numerous lines of evidence converged to point in the same direction.  This convergence of evidence is what made the hypothesis so compelling a century ago and in our own day.  Tigay and Rofé especially emphasize this convergence of many lines of evidence in establishing the Documentary Hypothesis.  This has been the strongest argument for the hypothesis all along and the one least addressed by its opponents. One cannot challenge such a body of evidence with a simple claim that other works did not get written that way.  And, as the researches in this book show, as a matter of fact other works didget written that way.  These researches should bring this argument to an end, and it will not be missed.  But really it never was an argument anyway.

What, after all, was the point of demanding to be shown ancient Near Eastern analogues when there was no ancient Near Eastern prose?!  The early biblical sources, being the first lengthy prose on earth, were without analogues from the get-go.  If we were to acknowledge the soundness of this demand, then should we also say: the biblical text could not be prose because where is there an analogue?!

Indeed, I have occasionally heard people claim that dual variations of stories—known as doublets, such as the two accounts of Moses bringing water from rock—are a common feature of ancient Near Eastern literary works composed by single authors.  But there is no such feature.  Doublets can hardly be a common feature of Near Eastern prose when there is no Near Eastern prose, either in the form of history-writing or long fiction, prior to the biblical source-works that are treated in the Documentary Hypothesis.  It is not even a common feature of ancient Near Eastern poetry.  Where we do find the same sort of characteristics that we see in the biblical text is the Epic of Gilgamesh, and Tigay demonstrates in this volume that the Epic of Gilgamesh is a composite of edited sources very much in the manner of the Hebrew Bible.  This volume thus turns a false claim on its head.  The evidence from the texts of the ancient Near East is a demonstration that composition by combining sources was a known literary activity in that world.

Nonetheless the challenge persists.  The most recent expression of the claim that the Documentary Hypothesis pictures something that was not a literary reality in that world came from an old friend and colleague of mine, Susan Niditch.  Niditch pictures the steps of the process of editing the sources and says, “I suggest that the above imagining comes from our world and not from that of ancient Israel” (Oral World and Written Word: Ancient Israelite Literature, p. 112).

Tigay and his colleagues had shown that what is pictured in the hypothesis was very much a part of the world of ancient Israel.  But Niditch, seemingly unaware that this had already been demonstrated, made this unfortunate criticism in 1996 with no citation of Tigay et al when Empirical Models for Biblical Criticism had been out for over ten years.

Even experienced biblical scholars seem not to understand how texts that are composed by more than one author come to be put together according to the hypothesis.  Niditch, for example, writes:

At the heart of the documentary hypothesis. . . is the cut-and-paste image of an individual pictured like Emperor Claudius of the PBS series, having his various written sources laid out before him as he chooses this verse or that, includes this tale not that, edits, elaborates, all in a library setting.
If the texts are leather, they may be heavy and need to be enrolled.  Finding the proper passage in each scroll is a bit of a chore.  If texts are papyrus, they are read held in the arm, one hand clasping or “supporting” the “bulk” of the scroll, while the other unrolls.  Did the redactor need three colleagues to hold J, E, and P for him?  Did each read the text out loud, and did he ask them to pause until he jotted down his selections, working like a secretary with three tapes dictated by the boss?   (p. 113)

That is a nicely visual and witty formulation, but it is not what is pictured in the hypothesis.  Precisely what makes the hypothesis seem outlandish in this formulation is what Niditch has gotten wrong.  There is no juggling of multiple large, heavy scrolls.  There is no one redactor who works with J, E, and P at the same time.  Only two large texts are combined at each stage of the editing.  J and E are combined not long after 722 BCE.  P is combined with the already-edited JE some three hundred years later.  I have worked with scrolls.  Laying out two scrolls of the size in question and doing the editing process that actually is pictured in the hypothesis is perfectly feasible.  (Anyone who has ever seen a scroll of the entire Torah in a synagogue knows that scrolls containing just JE or just P would be much smaller and quite manageable to a scribe.)  And Tigay et al have shown that it is more than just feasible.  It is something that was done in that literary world.  Niditch says that “the hypothesized scenario behind the documentary hypothesis is flawed.”  With due respect, it is Niditch’s understanding of what the hypothesis pictures that is flawed.

I do not mean to single out Niditch’s challenge as if it were something unique to her.  It is just a very recent example of a long line of this type of criticism of the hypothesis.  When Niditch uses language like “cut-and-paste” and “dicing and splicing source criticism” (p. 114), much like the old phrase “crazy patchwork,” she and the others who have used these phrases reveal that they have not yet come to terms with the way in which literature was in fact produced in antiquity and that they do not appreciate the artistry that it entailed, an artistry that was a collaboration of authors and editors.

One significant thing about the argument that there are no ancient parallels to the Torah as conceived in the Documentary Hypothesis is that it was of a different type from most previous arguments against the hypothesis.  Traditional responses to the  hypothesis had been to argue the items of evidence one at a time.  Each doublet was interpreted as complimentary rather than repetitious.  Each change in the divine names was explained as reflecting different divine aspects.  Each contradictory datum was defended as not contradictory, or ascribed to prophecy, and so on.  The documentary analysis, on the other hand, explained virtually all of the data with a very few, consistent premises.  The item-by-item approach never came to terms with either the convergence of all this evidence or theeconomy of the explanation.  The argument concerning the absence of analogous works at least was an argument against the structure of the hypothesis rather than a one-for-one series of responses that missed the forest for the trees.  That is why the newer, stronger bodies of evidence for the hypothesis are so significant.  They are structural defenses in terms of factors that pervade the text.   The evidence of language, the evidence of continuity of texts, and the evidence regarding the literary practices of the ancient Near East are more persuasive than old claims such as that J and E use two different words for a handmaid.  The force of linguistic evidence is that it is something pervasive and concrete.  The same goes for the evidence in this book.  What these two bodies of evidence offer together is a window into the ancient world that produced the Hebrew Bible.  Works that do not take these two corpora of data into account—either accepting them or responding to them—at this point really cannot be considered to be scholarship of substance.

Among the many failures of method in recent biblical scholarship, one that stands out is the weakness in properly addressing arguments that disagree with one’s own views.  Scholars argue that texts are late without addressing the linguistic evidence that they are early.3   Scholars argue that texts are composite without addressing the evidence that they are part of a single design.  Other scholars argue that particular texts are unities without addressing the evidence that they are composite.   This volume is a lesson to such scholarship.  It faces a classic argument, takes it seriously, and brings the evidence in response to it.

Tigay discusses why it is called the Documentary Hypothesis in his introduction, referring to its hypothetical methodology (p. 2).  Really, it is long past time for us to stop referring to it as a hypothesis.  The state of the evidence is such that it is now—at the very least—a theory, and a well established one at that.  To my mind, in the absence of any proper refutation of its strongest evidence, it is fact.

What is good about all this: Both the defenders and the critics of the Documentary Hypothesis have benefited from the process of arguing their cases.  In the early days the defenders made questionable arguments based on style, foolish arguments about when a certain idea would be developed, and murky pictures of how redaction takes place.  And the critics (and the defenders as well) misunderstood the “names of God” argument.4   And the critics used the argument: if an author would not include a contradiction, why would an editor?5   And they used arguments based on chiastic literary structures that supposedly could only have been fashioned by a single author.6   And they used absurd computer studies.   Truly, both sides lacked a real literary sense.  We needed to be prodded by non-specialists to develop a better sense of artistry and editing.

This book thus has value even beyond its stated purpose.  As I indicated above, its additional value is in terms of method.  Our field developed like a detective story in its early centuries.  We were simply following clues that surfaced, going wherever they led us.  There was little self-conscious addressing of method.  The argument against the Documetary Hypothesis that “there are no ancient Near Eastern literary parallels” had worth because at least it was a methodological criticism.  The authors of Empirical Models for Biblical Criticism, in responding to this argument, came to produce a project in methodology, examining our assumptions, our premises.  Their work is both a vindication of the process of the Documentary Hypothesis and a lesson to scholars who come after them.  The lesson is: you must address the methods by which you arrive at your conclusions.  This book is thus a signpost, a contribution to our field’s evolution.  Our field of research has reached a more self-critical, more sophisticated stage.  To be sure, there are still plenty of scholars who have not yet come to terms with the present state of methodology.  But they pay the price for this as their arguments and conclusions are more easily criticized and belied.

Much of the criticism of the Documentary Hypothesis comes from people who are not actively involved in the research.  So the argument that there are no ancient Near Eastern literary parallels was easy for them.  They could just say there are no parallels—without having to produce anything.  But the authors of this book were willing to do the research, and they showed that there are in fact parallels.  Likewise, critics who say that the “names of God” prove nothing and those who say that the presence of doublets proves nothing are persons who do not know the arguments and have not checked the texts themselves.  If they had, they would know that: (1) the actual evidence concerning the name of God in each of the sources proves a great deal; and (2) the divine name argument was not in fact that different sources use the words El, Elohim, or YHWH characteristically, and the doublet argument was not simply that doublets exist.  Everyone knows that a single author can use different names for the same person in a work.  Everyone knows that a single author can use two similar accounts within a work.  The argument was that these two things converge.  When we divide up the doublets, they fall into identifiable groups in which the right name for God comes in the right place consistently, with hardly an exception in the entire Pentateuch.  (In two thousand occurrences of El, Elohim, and YHWH in the Masoretic Text, there are just three exceptions.)  What’s more, the other categories of evidence line up and converge with these two consistent bodies of evidence as well:  We find the right terms in the right half of the doublets, we find contradictions disappearing as they fall in one group or another, we find the doublets to fall in groups that consistently fit the right stages of the Hebrew language.  These critics simply do not know this evidence.  Likewise they do not take on the linguistic evidence because they simply do not know it.

And then they toss off that phrase “crazy patchwork.”  And they are right.  In the absence of understanding the convergence of the many lines of evidence, and in the absence of awareness of the power of the linguistic evidence, it really does look like a crazy patchwork to them.

Similarly, Emanuel Tov’s chapter shows the actual process, the way this kind of composition works in practice.  For those who say: show us a J text that has been found (or an E or a P), i. e. show us a source, Tov provides it.  The Greek text contains just one of the two sources that were used to form the David-and-Goliath story.  This criticism (“show us a source”) was another of the arguments that did not require the critics of the hypothesis to produce anything.  These critics rather pressure the supporters of the hypothesis to produce something.  Well, the inability to produce such evidence—viz. finding ancient Near Eastern parallels or finding a source text separately—would not have made the Documentary Hypothesis wrong anyway.  But Tigay and Tov have produced it nonetheless.

Those who oppose the hypothesis are left with arguments like: the critics themselves disagree over the identification of various passages as being from one source or another.  They say this as if disagreements among specialists in any field of learning are proof that the entire field is wrong.  The disagreement of scholars over whether a passage is J or E is part of the normal process of refining that we expect in any healthy, honest field.  Pointing to this professional process and claiming that it proves that all of the disagreeing scholars are wrong is, well, absurd.  And it is an argument from weakness, something to be claimed when other arguments have failed.  The only thing weaker is the claim that we still hear from time to time that “no one believes that anymore.”

As the very existence of the Diatesseron shows, the impulse to reconcile four sources is not “crazy” or even unique.  What all of the examples in this book show is that editing was a fundamental part of literature well into antiquity.  To state it simply: from the earliest times of the production of written works until the present day, where there are authors there are editors.  The story of the formation of the Hebrew Bible is the story of a great collaboration of authors and editors (very possibly the greatest in all literature).  Those who claim that this was uncommon, unlikely, uncharacteristic, unparalleled, anachronistic, or a crazy patchwork simply do not know the evidence.  So, please, let us never hear the ridiculous phrase “crazy patchwork” again.  If we need metaphors from the world of sewing, then we would do better to say “brilliant embroidery.”


  • Wednesday, July 6th, 2011,
  • by