Biblical Hebrew Diachrony

This past November’s NAPH annual meeting brought a close to several years of discussions about linguistic dating of biblical texts that will be preserved in a volume of the papers being being edited by Ziony Zevit and Cynthia L. Miller and published by Eisenbrauns. Both of us have papers appearing in that volume that we have posted here (Cook) and here (Holmstedt).

Participating in the meetings has helped clarify the arguments in the debate. Perhaps the foremost of these is the claim that linguistics cannot independently date biblical texts (see Young 2005; 2010; see Cook’s paper for bibliography). We both drew on Hale’s (2007) recent work on historical linguistics to clarify what is really at issue here: the artifact texts and the language represented in those artifacts are distinct and only the latter falls within the purview of the linguistic enterprise (see further in our papers). Thus, it is certainly true that linguistics cannot date the biblical text, because that is a philological task. However, the continued protests misleadingly suggest that linguistics has nothing to contribute to the philological task; this seems to be the point Young (2010) driving at in response to Cook’s paper: “We claim instead that the nature of the biblical texts is such that this chronology, however, is not visible in any way that makes linguistic dating of biblical texts possible.” Given their calls to reevaluate the traditional chronological model of the biblical text from the ground up, claiming that no relevant diachronic data is “visible” in the text is nothing less than prejudging of the data. Regardless of how much the biblical texts may be in disarray due to scribal activities, every author, redactor, and scribe contributes linguistic data that linguistics may legitimately be used to analyze.

Their preclusion of any role for diachronic linguistics in the analysis of biblical texts is evident in the many and varied protests that share the view that the biblical texts are simply not a legitimate source of linguistic data (e.g., there is an “unbridgeable gap” between the artifact text and the “original text”; as a literary text the Bible yields not usable diachronic data because the rate of diffusion differs from that of spoken language, etc.). Such arguments lay bare the real agenda of those that make them: linguistics must refrain from making any diachronic statements about the biblical text. To admit that “there is still a history of the Hebrew language” but then to preclude from the database the major textual artifact of that language is astonishingly illogical.

A second issue also involves prejudgments. Following the work of C.-J. Bailey, A. Kroch, and S. Pintzuk (see esp. Holmstedt’s bibliography), we invoked the “S-curve” of diachronic change. Without getting into particulars (for that, see our papers), diachronic changes in languages show a pattern, whereby the older feature is replaced by the newer feature in a way that reflects the Sigmoid function (hence, the S-curve). Those who are unshakeably opposed to seeing any connection between the linguistic data in biblical texts and the (relative) dating of those texts have served up an odd counter-argument. It goes like this (and we are sincerely attempting to avoid making straw man of the argument): “The results of Holmstedt’s and Cook’s S-curves are so different than what we’d expect, it would require those in biblical studies to engage in wholesale re-evaluation of the dates typically assigned to the book. Ergo, their S-curves cannot be valid.”

Do you see what we see? We have a bit of a problem with this line of argumentation. With our S-curves (which are based on single features and are thus only a starting point for such research), we are proposing that the linguistic data be taken seriously, even if it suggests some radical changes to the dates commonly assigned to books—-or better, since we eschew absolute dating, to the relative order of reflected in the dates commonly assigned to books. We suggest that numerous S-curves could be overlaid, resulting in a relative order of books that reflects a truly methodologically sound approach to the linguistic variation we can all see in the biblical (and extra-biblical) texts. To dismiss our S-curve proposal by judging the results against pre-assigned dates for the books is patently unscientific.



22 Responses to “Biblical Hebrew Diachrony”

  1. Ian Young Says:

    Hi John and Robert,
    I thought we had begun to engage with each others’ ideas, but it appears that we are still talking past each other somewhat.

    My 2010 paper (SBL Atlanta), which you cite, argued that of course linguistics has much to contribute to the study of biblical language, but dating the composition of biblical texts absolutely or relatively is not possible. The point I am arguing is not that linguists do not have a good enough methodology, or that linguists aren’t all brave and handsome and deserve a pay rise. The simple point is that according to the consensus of historians of the text of the Hebrew Bible (textual critics), which I documented extensively in my paper, the task of dating biblical texts as usually understood (e.g. Samuel is pre-exilic; Qoheleth is 4th century BCE) is impossible, whether you are using linguistics or divination via liver omens. This is for the simple reason that biblical texts do not have a single date of composition. They were written and rewritten over time. Each biblical book and passage, to the best of our knowledge, is a composite of the work of a succession of authors over many centuries. Especially in regard to the linguistic features, one of the most fluid elements of the texts, what we have in our current texts is a potential museum of linguistic forms from many different eras. Maybe one day we will have enough extra-biblical, securely datable, sources like inscriptions to give a chronological account of the linguistic forms of our various biblical texts, but that day is not today, in my opinion. So what is the date when a biblical book was written, reflected by its language? This question makes as much sense as asking what single date the Hebrew Bible was written.

    You also chide me for being surprised at some of the results of your S-curves. My comments (in unpublished, draft material that I am still thinking about) regarding your S-curves were more along the lines of how surprised other scholars involved in linguistic dating would be at the chronology implied by your S-curves. For myself, I’d love you to overturn all the accepted dates in biblical studies. I have no investment in any dating of biblical texts; mostly when I dealt with a specific issue I have objected to the certainty of scholars based on linguistic data that certain texts (e.g. Qoheleth; Job) are totally “late” in origin. Beyond that, personally, as I said, I think that dating biblical books in the old sense is asking the wrong questions. In the work you cite, I suggested that our evidence for the composition of the biblical texts wouldn’t lead us to expect a linguistic chronology to emerge from the texts, and that it looked to me like the chaos of your S-curves backed that up.

    As I discussed with John in Atlanta, linguistics needs to take into account the nature of the evidence we have for ancient Hebrew (i.e. textual and literary aspects), or it will keep asking the wrong questions.

    Thanks for the dialogue,
    Ian Young

    • robertholmstedt Says:


      Thanks for responding, thereby making this a dialogue, well, technically a three-way discussion, instead of just John and I talking to each other (is there anyone out there!? We thought this blog would be our path to fame and fortune. Hmm….)

      One thing caught our eye that we want to address first. You mentioned our use of your unpublished draft material; indeed, we both read it and appreciate that you shared it with us. But we want to be clear that we were actually referring to Robert Rezetko’s identical response to our 2009 NAPH paper at the session itself. While your paper mirrors his response on the S-curve issue, we want to be clear that we didn’t actually cite your unpublished work, just as we promised. A small point, but one of importance to us, since we also often share unpublished work.

      Ok, now to the meat of your response.
      1. We most certainly agree about (Hebrew) linguists being handsome and deserving a pay raise!

      2. Here’s a basic problem, though: you keep talking about the consensus of historians of the text (= text critics) regarding the dates the biblical texts. As we’ve pointed out before, though, you are necessarily engaging in reconstruction if you move behind any of the historical documents (B19a, DSS). So why do you privilege the text critics to the point of trumping every other discipline? It is a significant claim and one we don’t find compelling. Our understanding is that cogency can only be achieved when the various disciplines work together and recognize that the only privileged folks are the dirt-diggers** (because they *find* the artifacts) and possibly the chicken-scratch-readers** (because they read the letters on the stone/sherd/vellum/papyrus). So, we disagree that the text critical “consensus” is simply to be followed blindly with regard to the dating of the *reconstructed texts* that are the product of textual criticism. (Bear in mind that the dating scheme that is the target of your criticism was at one time a “consensus” as well!)

      Indeed, as we point out in our articles, reconstructing the texts is actually the domain of philology, of which text criticism is but one component. It seems naive to view this process as one of strict succession: text criticism *first* then other aspects of philology, as you seemed to imply in our (JAC) post-session discussion at SBL. Such discreteness simply does not reflect the reality of philological task. Moreover, I (RDH) am reading a paper at a conference in May in which I point out how text critics have likely made mistakes in a Leviticus passage because they fail to recognize a linguistic issue that lies behind the variation. (I’ll post a summary of my paper soon to this blog.)

      3. You assert that biblical texts do not have a single date of composition. How did you arrive at this conclusion? Did you not use language? The texts are, after all, communicated via human language. And since you had to have used language to make these decisions, how do you know you used it correctly? Perhaps there were diachronic changes — those that accord to known cross-linguistic change types — that led to the coexistence of certain forms that you (or whomever you’re following) have missed and thus misidentified as compositional layers. Heck, perhaps there are changes you’ve missed that point to more or different compositional layers. Even if neither is the case, the point is simple: you cannot dismiss the analytic study of the language of the texts that you’re dating from the dating process itself. As you have often pointed out, we have no external dating for most of the biblical *texts* (≠ composition), which means that all internal clues must be used, including language.

      4. And so, since you by necessity use language in all your reconstruction of the dating of texts, then you are in fact doing what you’re disallowing for the rest of us, because by dating the texts you are actually ordering the linguistic facts to follow your own dating scheme.

      ** Let it be clear that we use these terms affectionately. We both work with dirt-diggers/shovel-bums (a.k.a. archaeologists) and respect their work. Moreover, while neither of us claim to be professional epigraphers, we are sufficiently competent in reading the chicken-scratches ourselves to appreciate the combination of brilliance, insight, and abundant squinting that deciphering newly found texts requires.

  2. » The Role of Linguistics in Dating Biblical Texts Says:

    […] at the Ancient Hebrew Grammar blog there is a fantastic discussion going on between Robert Holmstedt, John Cook, and Ian Young on the […]

  3. Donnerstag Digest (February 24, 2011) « New Testament Interpretation Says:

    […] Robert Holmstedt advocate the relevance of diachronic linguistics for dating biblical texts, and Ian Young responds to the contrary (HT: Charles […]

  4. Robert Rezetko Says:

    Hi John and Robert,

    I wish to add my voice to the discussion between you two, Ian, and others referring to it elsewhere in the blogosphere (Halton, Hobbins, etc.). Mostly I want to clarify several issues and *our real views* which, astonishingly, you continue to misread and misreport (and even mislead others who reproduce the same erroneous thoughts).

    Complete discussions of these and other issues will appear in another book that Ian and I have been working on, presently called Diachrony in Biblical Hebrew: Steps Toward an Integrated Approach. This book is our conversation with the 2009/2010 SBL/NAPH meeting participants on this topic and thus with many of the articles in the forthcoming book edited by Zevit and Miller (which were/are supposed to represent a response to our two volume Linguistic Dating of Biblical Texts = LDBT). However, our new book is more than a response, more than (re-)engaging previous and current scholarship on their own terms. It will take the debate over *diachrony* in BH to the next level–note “diachrony” and not “linguistic dating”–by integrating a range of diachronic approaches and data from an interdisciplinary perspective, including literary criticism, textual criticism, and historical linguistics. The latter, by the way, draws extensively from viewpoints and methods of diachronicians (linguists, philologists, etc.) working outside biblical studies and BH. So, stay tuned. For now, here are some of my thoughts on your blog entry.

    FIRST, DATA. You say that we do not take the/your BH linguistic data “seriously,” that we “prejudg[e]” them, thus we dismiss or downplay their significance. These remarks in particular are in the context of your discussion of S-curves for particular linguistic features that you treat in your papers/articles. To begin, if anything, we wanted LDBT to be as comprehensive as possible in terms of scholarship, bibliography and *data.* Clearly, due to constraints of space, time and energy, we couldn’t do everything, but we did our best. What is my point? While there will always be more data, other kinds of data, new constellations of data, alternative ways of viewing and explaining the same data, etc., we take all the BH linguistic data *very seriously.*

    Furthermore,–and I have documented this elsewhere, mainly in the light of some of Dresher’s remarks–often in LDBT we discuss issues of linguistic *innovation and diffusion*, although we often use other terminology such as distribution, opposition and accumulation, and for the last concept we often use words such as concentration, frequency, rate, ratio, percentage, proportion, etc. In other words, with regard to statistical variations in BH, we neglected in LDBT, mainly because we were reviewing scholarship on its own terms–but not in the forthcoming book–the jargon and graphics of the Constant Rate Effect/Hypothesis, Sigmoid Curve, and so on. However, we examined dozens of so-called early and late linguistic features, superimposing them in the formulation of our hypothesis, and as we did so, we became increasingly aware of the chaotic variation (real situations ≠ traditional assumptions) that we had already started to observe in previous studies and publications. Now, I admit that we did not illustrate this in a very lucid way in LDBT–except tangentially in the table of accumulation–but the data are there in many of the discussions and tables. What is my point? We appreciate the two S-curves in your papers (each based on a single feature, as you say), they are “valid” descriptive representations of data, but, one, as we mentioned in the sessions following your papers, both S-curves exhibit “exceptions” (too quickly dismissed by you) that cast doubt on any plausible (relative) linguistic chronology given familiar periodizations of biblical literature; and, two, based on the results of our own research, some published, some given out in my handout in Atlanta, and some forthcoming, your two S-curves do not coalesce with other evidence to paint a coherent chronology of biblical writings, that is, they are simply illustrative of the same widespread linguistic disorder that we expect given the complex composition-transmission histories of the biblical books (see below). (I will forward my specific criticisms of your treatments of the data when they are published.) In any case, we are not a priori “prejudging” or “dismiss[ing]” your data!

    SECOND, VIEWPOINTS. You say that we “privilege” (also: “strict succession”) textual criticism over “every other discipline” and “other aspects of philology” by which we presume that you mean primarily over linguistic issues and analysis. On the contrary, engaging and working within “various disciplines” and mentally employing “a continuous feedback loop” (Hobbins) led us to formulate the particular thesis argued in LDBT. Indeed, it was the crossing of boundaries, interdisciplinary approach and collaboration, that enabled Ian, Martin and I to work together so enjoyably and productively. Please allow me to explain. You may know this, but perhaps not others. Ian began working on language but then branched out into textual criticism. I began with textual criticism and literary criticism and then became involved in the language debates. Martin is a Hebraist from the beginning. Thus our publications (so far) include not only the articles and books on language that you know, but also a dozen or so articles by Ian on text-critical matters, and several text-language-literary articles and a monograph by me. What is my point? We do not deliberately privilege textual criticism, or literary criticism, or any other criticism or viewpoint, over linguistic criticism. Rather, we sit down and listen attentively to their conversation and then try to explain their relationship with one another. The main objective of LDBT was to deconstruct previous scholarship and results: BH texts (= documents) cannot be dated on a linguistic basis only, as Hurvitz and so many others have asserted. The new book will move the discussion forward by showing how various streams of data and information could be related to one another into a coherent whole.

    I have a few specific comments on your remarks on textual criticism. In your reply to Ian, the second point, you speak twice about the dates of biblical texts that are reached by text critics or by means of textual criticism. First, you say that Ian (we) “keep talking about the consensus of historians of the text (= text critics) regarding the dates [of] the biblical texts,” and second, you say that you “disagree that the text critical “consensus” is simply to be followed blindly with regard to the dating of the *reconstructed texts* that are the product of textual criticism.” Well, I have some news for you. First, we (Ian and Robert) have never spoken a word about the “dates” or “dating” of biblical texts (= compositions) that can be determined on the basis of textual criticism. Second, text critics of the ancient versions of the Tanak do not postulate dates for biblical texts (= compositions) other than the dates of the manuscript artifacts themselves. The current “consensus”–and there is one (Tov, Ulrich, Sanders, Talmon, Trebolle, and many more)–has nothing to do with “dates”, but rather, it relates to the nature of the evidence, textual fluidity and pluriformity in the BCE period, and its implications, including that the Tanak was evidently written, rewritten and edited over the centuries, as was other ANE literature (e.g. Gilgamesh). On the general “consensus,” reread some of the quotations in Ian’s paper (De Troyer, Aejmelaeus, Tov, Ulrich, Hendel, Talmon, again Ulrich) and/or in LDBT, vol. 1, ch. 13.

    I have one more remark on your views of textual criticism. It is prompted both by your general comments on textual criticism and dating and by your specific remark on Leviticus (a paper you also sent us in early 2008). Traditional textual criticism, whether of ancient biblical or medieval English writings, has to do with the editing of texts, the removal of errors, and so on. Thus there are specific instances here and there where linguistic criticism and textual criticism combined may prompt one to call a particular linguistic feature early or late or a particular textual reading primary or secondary. However,–and this is what your remarks on textual criticism overlook or simply give lip service to without seriously engaging–following the discovery of the biblical Dead Sea Scrolls there was a shift (or reentrenchment, of previous textual-exegetical scholarship) in the practice of textual criticism from resolving smaller-scale textual problems to unraveling larger-scale literary formation (again, Tov, Ulrich, Sanders, Talmon, Trebolle, and many more, contra pre-DSS perspectives reflected in works by authors such as Brotzman, Waltke, etc.). I documented this transition and employed the new insights and methods in my monograph on the book of Samuel. Similarly, see the recent essays by textual and literary critics of Samuel in the book edited by Hugo and Schenker, Archaeology of the Books of Samuel: The Entangling of the Textual and Literary History. What is my point? Your remarks on textual criticism mirror old-school thinking and represent an inadequate appreciation of the text-critical discipline as practiced today by many textual and literary critics of the Tanak. Indeed, I will state this more strongly: Your apparent acquiescence to “a single date of composition” and disparagement of “compositional layers” for “biblical texts” (your third point) is pre-critical and undoubtedly unacceptable to most biblical scholars, including textual and literary scholars, who are writing introductions and commentaries on the Tanak. See, for example, the survey of scholarship in LDBT, vol. 2, ch. 1, where we sought to give a solid overview of “consensus” / “mainstream” appraisals of the dates of biblical writings.

    THIRD, DATING. You speak about us in terms of “the texts that [we’re] dating,” “[our] reconstruction of the dating of texts,” and “by dating the texts [we] are actually ordering the linguistic facts to follow [our] own dating scheme.” You are implying that we seek to date, or we actually date, biblical texts (= compositions) to particular periods or points of time on the basis of language and other evidence, especially textual criticism. It seems that others have also understood your remarks in this way, e.g. “linguistics can be used to assign relative dates to the composition of biblical texts (H[olmstedt]&C[ook]–yes, there is a role for linguistics; Young–no, this is the prerogative of text-critics)” (Halton). This is a fundamental misrepresentation of our views and arguments. Ian made this very clear in his initial reply, above: “…This is for the simple reason that biblical texts do not have a single date of composition. They were written and rewritten over time. Each biblical book and passage, to the best of our knowledge, is a composite of the work of a succession of authors over many centuries….So what is the date when a biblical book was written, reflected by its language? This question makes as much sense as asking what single date the Hebrew Bible was written.” So, to clarify, we do not have a dating scheme! We are open to all sorts of preexilic, exilic and postexilic dates for the origins of biblical writings! What is my point? *Our real view* is that individual biblical writings do not have *a* single date of origin but rather take in a multiplicity of dates, a continuum of dates, such that it is meaningless to speak, for example, about a preexilic (only) book of Genesis or Samuel or Pentateuch or Deuteronomistic History. Consequently, perhaps someday we will have the (external) evidence, methods and insights, working with linguistic, textual, literary and other perspectives, to corroborate the relative and/or absolute dates of (initial, transitional, and final forms of) stories, sources and books of the Tanak, but in our opinion scholarship has not yet reached that moment in history. In terms of language, particularly linguistic innovation and diffusion, our findings thus far are that the increasing quantity of well-documented data, when superimposed, do not exhibit patterns of usage in support of a lucid chronology of biblical writings. That said, the S-curves that we (you and I) and perhaps others are assembling may eventually prove convincing of one or another dating scenario, though I admittedly have my doubts.

    FOURTH, AGENDAS. Robert, in your paper/article, you call for an agenda-free assessment of methodological principles and practices following our challenge to the consensus. Good, I agree. To my surprise, though, as the debates over the linguistic dating of biblical texts (= compositions) and diachrony in BH have unraveled, as we have monitored various kinds of responses to LDBT and other publications, as I read and thought about your papers and blog entry, I have increasingly realized the potential weight and role of different sorts of deeply-embedded ideologies. In principle this is nothing new. For example, Sanders has highlighted resistance to the idea and implications of textual fluidity and pluriformity (mentioned above) among particular religious communities and groups (ABD, vol. 1, pp. 847-848), because of pre-conceived ideas about the origin, character, sacredness, etc. of the Bible. So, let’s speak openly. Ian, Martin, and I have gone on record about our personal histories, beliefs, and intentions in writing LDBT. For example, see (“Twenty Questions,” March 2009), and LDBT, vol. 1, p. 4: “…we are uninterested in proving or disproving the antiquity of biblical stories or the historicity of people and events in those accounts. Rather, we are concerned with the role language has played in assigning dates to the extant BH texts [= compositions!] which transmit those stories.” Let me restate this: We deny that we have an axe to grind in terms of the (supposed) early or late dates of particular biblical writings, the (supposed) inspiration and inerrancy of their words or ideas, the (supposed) historicity of people and events mentioned in them, or minimalist, “middle” and maximalist ideologies.

    Finally, “agenda-free” scholarship should also be mindful of name calling and guilt by association. I am thinking in particular of charged words such as minimalist, revisionist, postmodern, etc. and statements like the following ones that explicitly ascribe to us certain views about the (late) provenance and (non-) historicity of biblical writings. “…They can make this claim because at every turn they either dismiss the obvious non-linguistic evidence that is relevant for dating at least some of the biblical books (e.g., many of the prophetic books, or sections thereof) or take a revisionist biblical studies position, THAT THE BOOKS ARE ALL LATE (note that they often couch this position as “mainstream biblical scholarship,” a claim that is both subjective and tendentious)” (Robert H., paper posted on this webpage, pp. 43-44, n. 41; emphasis added). Similarly: “One implication of their conclusion is that although the books from Joshua through II Kings, were composed in the SBH style, they were authored during the Persian period. Since they were written long after the purported events that they claim to describe, there can be little of historical worth in them. So too, all that is narrated in the Torah. Appropriately postmodern, one could say that there is no history in the history because the appearance of history writing conceals essentially fictional narratives that may contain accurate information only here and there” (Zevit, Iggeret newsletter). In short, statements such as these apparently intend to darken or discredit us as scholars, and whether unintentionally (carelessness) or intentionally (misinformation), they misrepresent both our own views and what we have said in publication.

    Thanks for the dialogue,

    Robert Rezetko

  5. February 2011 Biblical Studies Carnival | A Fistful of Farthings Says:

    […] 1:17 and then presented a solution for the textual variant in that verse. John Cook argued for the relevance of diachronic linguistics in biblical Hebrew to which Ian Young responded, as did John […]

  6. Jan Joosten Says:

    Dear Rob and John, I downloaded both your papers and read them (I’d heard John deliver his and read an earlier version of Rob’s before). Although much of what you write is impressive, there is one aspect I don’t like at all and that is the way you almost completely disregard earlier studies on the history of Hebrew in the biblical period. It may be gratifying to you to present yourselves as those with whom the really scientific investigation of diachrony in BH begins, but it is a false picture. Rather than to start with Young and Rezetko’s presentation of the state of the art, I invite you to go back to the writings of Gesenius, S.R. Driver, Kropat, Kutscher and Hurvitz themselves. Working through their studies may seriously retard the moment you can contribute to the discussion, but that is the price to pay for real progress.

    • robertholmstedt Says:

      (RDH: this comment has been edited; I questioned whether my first response to Dr. Joosten reflect the intent of his comment; then as I was editing the comment, my co-blogger approved it while it was still in a slightly muddled state.)


      Thank you immensely for stopping by. I’m not quite sure how to read your comment, though.

      If you are objecting to the fact that our articles acknowledge that Y&R’s LDBT has been the catalyst for more *and more linguistically grounded* research on BH diachrony, I think the very existence of 2 years of NAPH sessions (with a 3rd coming this year) at SBL indicates that this is precisely where LDBT has its greatest service.

      If you are simply troubled by our use of previous scholarship, I can only disagree. Indeed, this is the first time we’ve been accused of this particular sin, i.e., not doing the research on past scholarship and acknowledging it. Frankly, I’m a bit appalled to be the recipient of such a charge.

      In our posts on this topic and in the comments, you’ll find that we *do* appreciate those who worked on the issues in the past — indeed, this is an area where I offered a criticism of Young and Rezetko’s LDBT in my last comments in our second post.

      On our papers in particular, while we list Gesenius, Driver, Kutscher, Hurvitz, etc. (and Kropat simply didn’t make the final cut in mine, although I’ve used him elsewhere) in our bibliographies and cite their works in footnotes, etc., it is true that these two papers are forward looking. Moreover, they are *papers*, not *monographs*, and so including a full history of scholarship is out of the question. And I *really* doubt that what the past masters have said would “retard” what either one of us contribute, since though the data have not changed, our arguments are about methodology as much as our specific conclusions re: the data. Moreover, we both build (in these two papers) on our theses, which you may or may not have read. And, of course, in both of those (as well as other papers, such as my JNES paper on the etymologies of אשׁר and שׁ), we acknowledge and treat those you have mentioned on the specific issues of the relative markers and the verbal system.

      And so, although we cast our papers differently than you may have expected, one cannot seriously infer that we ignored past works in order to present our work or conclusions as somehow more significant. It is the case that linguistics has changed enormously since any in the list you mention (except for Hurvitz, who is still with us, of course). While we do read the older works for insight (and appreciate them), we don’t think that we can simply lift their conclusions and dress them up in new language. Rather, given the state of general linguistic research, the work must be re-done. And that is what we’re doing in both our articles, which we do believe represent “real progress”. So, it seems from your comments we would have to disagree on a basic approach to doing research in this area.

  7. johncookvw Says:


    I chuckle a bit at your post: to accuse us of ignoring past scholarship when our (your’s and my) ongoing debate on the verb boils down in its essentials to my defense of an Ewaldian position (19th century!) versus your “new-fangled” modality approach. Talk of the pot calling the kettle black!

    I will gladly defend my silence with regard to earlier work on diachrony in my paper. It certainly is not due to unfamiliarity, and though this silence is appropriate in an article of this length (as Rob mentioned). The real reason was to side-step Young and Rezetko’s fixation on the previous position. Any alignment or appreciation with those earlier views would (IMO) have needlessly drawn the whole discussion to the debunking of those positions rather than an engagement with a new forward-looking approach. So I am willing to readily acknowledge that YRE have “cleared the deck” of the old approach for the purely pragmatic reason of getting beyond their vitriol to a discussion about how to move forward now. While I agree with Rob that they are too dismissive of these earlier studies, at this point in the debate it is difficult to make much use of them since they are down and still being kicked, so to speak.

  8. Jan Joosten Says:

    John’s reply confirms my initial impression. I don’t think the old lot are “down and being kicked” nor that the deck has been cleared.
    Let me illustrate with a specific question. From Gesenius to Hurvitz there has been an understanding that BH basically divides into two corpora, “classical” and “late”. In your papers, most explicitly perhaps in Rob’s, you reject this dichotomy in favor of a view of more continuous and incremental development. It is true, of course, that languages develop in this way—Gesenius and Hurvitz knew that too. And yet there is a lot to say for the idea of two corpora in the Hebrew Bible.
    My point is not that the earlier guys were right and you’re wrong. My point is you haven’t engaged this issue at all: you just substitute your presuppositions for the older ones. In my view, this is not how science progresses.

    • robertholmstedt Says:

      Yes, I do reject the fundamental dichotomy between SBH and LBH (although they are convenient categories in the classroom–that’s where they should stay). I follow Naude in this regard and he’s quite right. The fact that previous scholarship was deficient in this view of language change (your claim to the contrary notwithstanding; if they understood this, it was incumbent upon then to write it more clearly) is very important.

      What it comes down to is this–do we acknowledge on whose shoulders we stand? Absolutely. But do we intend to spend time sorting out how we might translate past generations’ views on these issues into contemporary frameworks? No, because our work is not a history of the scholarship. Rather, we are more concerned with contributing to a more accurate picture of BH diachrony than wondering what Gesenius or Kropat or Ben-David would have said if they had known about current historical or statistical linguistic research.

      Indeed, I remain somewhat flabbergasted as to why you bothered to bring up this issue.

  9. Jan Joosten Says:

    Rob, the way it looks to me is Y&R say: “Hurvitz etc were completely wrong, there is no history of Hebrew in the biblical period,” while you say: “Hurvitz etc were indeed completely wrong, but there is nevertheless a history of Hebrew in the biblical period.” Athough I like your work more, both of these stances are unacceptable to me. I’ll side with Hurvitz, Talshir, Eskhult, Polak against Jacky Naude any day.

    • robertholmstedt Says:


      I will go back and read precisely how I’ve stated things in my article, because while I do consider Hurvitz’s methodology to be inadequate, as I said in one of my comments to Young and Rezekto in our other post, I also appreciate some of his and our other predecessors’ insight(s).

      But as it stands, I do think Hurvitz’s arguments are untenable precisely because, while you might consider his methodology scientific, it also represents one of the basic problems with much Hebrew research—it’s ad hoc and “home-grown” nature. There is no interaction with linguistics in general or language study in other areas. And when that is done, it shows clearly, in my opinion, that one cannot simply side with Hurvtiz, etc., and consider oneself “scientific”. It is not “scientific” to ignore advances in related fields simply because of one’s respect for the arguments of those on whose shoulders we stand.

  10. Jan Joosten Says:

    Rob, I admire the way you try to integrate general linguistics and Hebrew studies, but I don’t like the pretention that all that was done before is simply inadequate (as opposed to perfectible). That’s all I wanted to say when I posted my first remark.

    • robertholmstedt Says:


      Fair enough. I see now better what motivated your response. But in my defence I don’t think it’s a charitable reading of my work. If I really had such an arrogant stance vis-a-vis older works, I’d ignore them outright. But I don’t and, as my NAPH article concludes, the results of the S-curve analysis of שׁ “supports the traditional explanation” and that the approach I’m advocating should maintain the “philological diligence that has always been a hallmark of our discipline”.

      How is this different in substance from your own approach? For instance, in your 2002 JANES article on the verb, you make this statement in your conclusion: “Could it be that the “the century-old tide of describing the BH verbal system as aspectual” started out from an undue application of Arabic categories to the Hebrew verbal system?” Why do you reserve the right to question Ewald’s methodology (i.e., his so-called misapplication of Arabic categories in his study of the Hebrew verb), but John and I are not allowed to question the methodology of Hurvitz and others before him?

      In all ancient language study, the data *rarely* change, whereas our scientific methodologies do. Therefore, it is going to be the case much more than not that our disagreements with past scholars concern methodology. And as such, “inadequate” is hardly a pejorative or contemptuous term.

      So, I think my definition of “inadequate” is pretty much the same as your “perfectible”. And since in terms of LDBT and Hebrew diachrony we essentially agree with each other on the important issues, I’m don’t see the reason that you have bristled so at our work.

      • robertholmstedt Says:

        By the way, it may contextualize our exchange with Dr. Joosten for readers to know that much of Jan Joosten’s work is required reading for BH linguistics (it’s all good reading, but he also works on Syriac, textual criticism, etc.). And, it is because we respect his work very highly that we are sensitive to the criticism of our BH diachrony work that he has offered in these comments (more so than to the criticisms of Rezetko and Young, with whom we enjoy a relationship established much more on bantering and bluntness).

  11. Bill Schniedewind Says:

    I’m coming late to this party, so to speak, but as I’m going to be speaking at the upcoming NAPH on related issues, I thought I might revive this (or at least add my two cents worth). I’ll confess from the beginning that I have much more sympathy for Joosten’s position, but I’m also in total agreement with the part about paying handsome Hebrew linguists more money.

    I have two reservations with the direction that the field seems to be going, which I hope to address in a forthcoming book, “A Social History of Hebrew: its origins through the Rabbinic Period” (Yale University Press, forthcoming soon?). First, I am a bit skeptical about using the assured results of literary criticism to suggest that biblical texts were being “written and rewritten.” I think we impose our own modern text-centric perspectives on ancient Israel when we reconstruct very complicated literary histories for biblical literature. We may be obsessed with the Bible, but I think the obsession with authoritative literary texts was a relatively late phenomenon in the 2nd Temple period. Moreover, as I argued in How the Bible Became a Book, it’s very difficult to construct as social, political and Hebrew scribal context for this type of work in the Persian and even early Hellenistic context. In sum, I think the “rewriting” argument is overstated. At the same time, I readily grant that there was some recopying and that this recopying sometimes including minor editing and even significant linguistic changes (as we see in IQIsa). This, of course, makes the dating of biblical texts quite complex (and more interesting). For my part, I’m fascinated with the way later 2nd Temple period scribes tried to understand older linguistic forms that we no longer in use. Joosten, for example, has written on pseudo-classicisms. I explored the later scribal transmission of the old asseverative lamed. The scribes of chronicles and 1QIsa “updated” the language of their sources (albeit in different ways). There is real diachonic change in Hebrew that is fascinating and the current direction of this debate seems intent on obscuring this linguistic change, or at least suggesting that it is not diachronic.

    My second reservation is that the touting of “linguistic models.” To suggest that Hurvitz or Polak or Joosten aren’t using linguistic models would certainly not be fair. The problem is really which linguistic model one chooses. In my forthcoming book (hopefully, 2012), I will press not only the use of sociolinguistic (/anthropological linguistic) models, but perhaps even more importantly, I want to press for the use of the Linguistics of Writings Systems (e.g., Florian Coulmas). Most of what I learned about Hebrew historical lingusitics as a student and what I see discussed in the literature is heavily dependent on traditional linguistics, that is, linguistics of vernacular, in spite of the fact that what we possess are text-artifacts. I’d like to see the discussion of Hebrew diachrony show more awareness of the fundamental differences between writing systems and vernacular language. And then, I’d like to see more consideration of sociolinguistics and anthropological linguistics. So much of what we write assumes that if I write the word “colour” is says something about phonology, when in fact it really betrays (or hides) identity.

    • robertholmstedt Says:


      Thanks for your comment. You bring up some important issues. I agree with you on your first point, of course. And I, too, think the changes that were made make the dating task just that much more interesting. And it’s worth noting that Jan and I clarified a few things in a follow-up email exchange, in which it was clear that we are actually much closer on these issues than the comment exchange suggested. (The limits of e-communication are very frustrating.)

      But on your second comment, I can’t say that we’re on the same part of the page (the same page, yes, but not the same part). First, unless you understand that building on a linguistic theory is so open-ended that it can be any approach to the study of language, however undefined, then you have to admit that most of those who have been involved in the dating of BH have not aligned themselves with a recognized theory of language. I encourage them to do so and would heartily welcome that addition to their work, but the simple fact is that most have not situated their work within a modern theoretical approach of some kind. The problem, then, is that it’s hard, if not impossible, to identify the theoretical assumptions of those works and so also the validity of the analytical claims. It’s the same in any scientific endeavour — their is no interpretation of the data without a theory by which the data are processed. So, when I go to, for instance, Hurvitz’s work, in which there are many truly insightful observations, I find it hard to evaluate the claims because there are no points of comparison within the larger linguistic market.

      Second, the move to working with text artifacts has already begun. In the last 10 years there have been a number of generative-oriented work addressing pre-modern languages. The work by Hale, that John and I both cite, or Jan Faarlund** or the collection edited by Katalin Kiss** are just a few examples. I see a new one every time I get a chance to visit the library stacks. These works, in addition to quite a bit of work applying much more rigorous statistical analyses to historical data (e.g., work by Anthony Kroch or Susan Pintzuk), show just how your historical linguistics education (and mine!) puts us quite behind the historical linguistic times. This much and more I realized when I jumped into this discussion. It was humbling and exciting at the same time (exciting because it didn’t take much imagination to see that there is a lot left to be done in historical Hebrew grammar).

      I look forward to your paper and book. However, you might do me the favor of sending me a copy of the paper when you finish it. I won’t be attending SBL this year: it’s time for a break!

      **Faarlund, Jan Terje. 2005. The Syntax of Old Norse: With a Survey of the Inflectional Morphology and a Complete Bibliography. Oxford: Oxford University Press.
      **Kiss, Katalin É., ed. 2005. Universal Grammar in the Reconstruction of Ancient Languages. Studies in Generative Grammar 83. New York: Mouton de Gruyter.

  12. Bill Schniedewind Says:

    Robert, thanks for the references. I shall be very interested in these applications of generative linguistics to “dead” languages. I do know of bit of Chris H. Reintges since he taught Egyptian at UCLA. Of course, just because you can do something, doesn’t mean you should. As you know, I’m not a huge fan of generative linguistics or universal grammar. Indeed, the debate has played out at UCLA between our Linguistics department (Chomskian) and our Linguistic Anthropology group (Duranti, Ochs, and Kroskrity) so I’m well aware of the discussion, though I hadn’t seen these latest studies. At this point, specialists in Sociolinguistics and Generative Linguistics have little interaction with one another. Moreover, in my experience, linguistic specialists tend to roll their eyes when you tell them you’d like to apply linguistic theory to ancient texts. I’d be surprised if most specialists in Generative Linguistics didn’t greet such studies with great skepticism and little enthusiasm.

    Unfortunately, the linguistics field is absolutely enormous with more approaches than is really possible to credibly deal with as an interlocutor. As a result, it seems almost inevitable that we will be talking past each other to a certain extent. It’s not that we don’t read it, but the method one uses partly depends on the questions that one asks and the topics that interest us. There’s no point for us to argue with each other about the “correct” linguistic approach as this debate is taking place (or, actually, mostly not taking place) in the field(s) of linguistics itself. Personally, I find Florian Coulmas’, The Linguistics of Writing Systems (Cambridge, 2002), or Duranti’s, Linguistic Anthropology (Cambridge, 1997), much more useful for what I’m interested in–the people who spoke ancient Hebrew and the role of Hebrew in ancient Israelite culture. But if I were to turn to other linguistic approaches, I am more drawn to the varieties of functional linguistics than generative linguistics.

    I’ll be happy to send along my SBL paper — when I narrow down exactly what I want to say. I think I’ll focus on the “gap” in Hebrew scribal tradition between the Iron Age/Babylonian period (SBH) and the Hellenistic period (LBH) and explore sociolinguistic approaches to the problem. The diachronic differences in Hebrew, however, must play a central role in the sociolinguistic observations that I want to make.

    • robertholmstedt Says:


      In backwards order:

      Thanks. I appreciate getting to read cutting edge stuff before it has to go through the long publication process.

      I agree that we’d get nowhere trying to argue with each other about theory. My point was simpler: it’s often valuable, if not necessary, to know what theoretical underpinnings a particular analysis is built upon in order to evaluate both the conclusions and if/how one can incorporate the study into one’s own linguistic model of, say, Biblical Hebrew. I’ve stopped proselytizing for linguistic theories, but I haven’t stopped arguing for explicit identification of one’s theory and/or assumptions. And if I don’t see such in a given study of some linguistic feature of Hebrew, it’s hard not to set it aside as an argument without a theory.

      Reintges is a good example of using a theory on a dead language. Some generativists may roll their eyes at working on dead languages, but not my colleagues at U of T. In fact, although we all recognize the limits of what I’m trying to do (applying a theory to linguistic data for which I have no speaker input and a relatively limited corpus for example comparison), I’ve had nothing but encouragement.

      It just goes to show that our respective experiences colour (or color?) our assessment of what we do with regard to the broader field of linguistics.

Comments are closed.

%d bloggers like this: