Isn’t It Time to Break Up with Linguistic Dating? Rethinking Hornkohl’s Method—and What Comes Next

The central claim of these scholars is stated clearly:"the linguistic patterns found in the MT (or other sources) are more likely the result of later editorial and scribal activity than of earlier stages of composition." This is certainly a possibility, among many other possibilities (e.g. the infamous monkeys in a room hypothesis) But the standard model is more probable and more consonant with the evidence. Should we abandon the discipline of historical linguistics of biblical Hebrew? Surely not (as I have argued elsewhere). Hornkohl's book is an important contribution to the state of the art, which may be old (begun in the modern period by Gesenius) but not yet superannuated.

Thanks for the comments, Ron.

We completely agree with your “Surely not” to the question, “Should we abandon the discipline of historical linguistics of biblical Hebrew?” However, as we have argued in several places, “linguistic dating,” of which this book is indeed the state of the art, is not the same thing as “historical linguistics” (e.g., Currents, pp. 11-13). Attempting to date works of unknown origin by their language is not a normal procedure in mainstream historical linguistics. In our opinion, linguistic dating of biblical writings does not hold water. Historical linguistics does to the degree that the nature of the sources allows.

One key part of historical linguistics is understanding the nature of the sources before using their evidence. In agreement with the consensus of mainstream scholars, you yourself have written important work on the composite nature of the biblical compositions due to editorial and scribal processes. In line with this, and as mentioned in our article, we have discussed at length (both theoretically and practically in hundreds of pages of cross-textual variable analysis in Masoretic Text parallel texts and MT vs. other text traditions) the high fluidity of non-standard Hebrew linguistic features (e.g., HLBH, pp. 117-210, 413-591, and elsewhere). To our knowledge, nobody has engaged in depth with our data and arguments, so the claim that “the standard model is more probable and more consonant with the evidence” seems to us unfounded.

One question from us that we are curious about. Aaron Hornkohl argues that the language of the whole Pentateuch is no later than about 800BCE. How does that fit with the views in your recent excellent commentary on Genesis 1-11 that the composition (and language) of the sources in this section of the Pentateuch (and presumably elsewhere in the Pentateuch) dates to the 9th-5th centuries BCE, from the monarchic to the early postexilic periods?

Dear Martin,

I'm happy to respond to your question. But first, a clarification. Hornkohl is engaged in linguistic typology, which is an area of historical linguistics. I think "linguistic dating" is a misleading term, implying an analogy with carbon dating or tree-ring dating (dendrochronology). HIstorical linguistics should be used alongside other historical forms of inquiry (as I have argued elsewhere). On this I think we all agree.

Hornkohl does a granular analysis on a number of cases of diachronic change within Classical Biblical Hebrew. But naturally, not all of his analyses are equally compelling. For instance, the relative absence of personal names compounded with Yhwh in the Pentateuch has something to do with its genre, i.e., ancestral stories, compared with texts about later eras in the Former Prophets, which have plenty of names compounded with Yhwh. That is to say, some of stories told in the Pentateuch might be old, and so might the names of some of its protagonists.. But this has little or no bearing on the linguistic profile of the Pentateuch. Hence I don't see the justification for using this case as a criterion for distinguishing two phases of CBH.

The issue of short vs. long forms of the 1st person singular prefix conjugation in weak roots may have to do, as Hornkohl acknowledges, with the relative lack of linguistic updating in one textual family of the Pentateuch (proto-MT), compared to the situation for the Former Prophets. So it's also, in my view, not a strong basis for a distinction between two phases of CBH. But it's an interesting phenomenon, which I'd like to see more about (and related instances).

Having stated some reservations, I think it's exciting to see evidence of diachronic complexity within CBH. This is a rich topic of research, to which Hornkohl has impressively contributed. But I don't see any reason (yet) to revise my description of the linguistic history of Genesis.

Yours,
Ron
.

Dear Ron,

Thanks from all of us for your very interesting response to our question.

Regards,
Martin, Robert, Ian, and Martijn

That the Hebrew language evolved over time seems to be agreed upon. If I understand correctly, the extent to which such texts can be dated is less agreed upon, given that editing may introduce changes. (Though editing, itself, can, to some extent, be analyzed.)

Two proposals or questions:
1) Poetry tends to be less edited than prose.
2) Composition of the Torah did not happen in a short time period, pace, e.g., Russell E. Gmirkin, writing that the five books were "composed in their entirety about 273-272 BCE by Jewish scholars at Alexandria," Berossus and Genesis, Manetho and Exodus (2006) page 1.

Drew Longacre at his OTTC blog raises some valid concerns about our response to Hornkohl’s book.

https://oldtestamenttextualcriticism.blogspot.com/2025/07/diachronic-di…

We value the engagement of such a distinguished textual critic. As we have previously addressed the concerns he raises, we will briefly summarize his objections and point to several earlier discussions where we have dealt with them in more detail.

Longacre acknowledges occasional scribal updates in biblical texts but emphasizes that such changes are typically unsystematic, with the MT tradition in particular remaining largely conservative. We agree. The challenge, however, is that the linguistic dating approach developed by Hurvitz, Hornkohl, and others focuses precisely on relatively minor changes and uncommon forms—features that we have demonstrated in detail tend to fluctuate between CBH and LBH writings, across MT parallel texts, and between the MT and other textual traditions. For further discussion, see our comments on the documentation of variation and change (pp. 37-67), especially the section on common vs. uncommon language (pp. 41-42), in our “Currents” article (above, n. 3), as well as our detailed treatment of cross-textual variation in HLBH (pp. 117-210, 413-591; above, n. 3). (See also our quantitative analysis of distinctive late vocabulary at https://doi.org/10.5508/jhs.2016.v16.1.)

Longacre’s second point is that, while redactional histories of the Hebrew Bible are broadly acknowledged, they often remain uncertain and speculative. As a result, he favors “historical linguistic typology,” which he views as grounded in more objective evidence and thus a more reliable method for dating biblical writings than redaction-critical approaches. It is not uncommon for scholars of biblical Hebrew—though in this case, Longacre speaks as a textual critic—to claim that linguistic criteria are more objective than literary or other criteria for tracing the compositional history of the Hebrew Bible. The core issue, however, is that everything hinges on the nature of the sources involved. We have examined this matter in depth. See, for example, our discussions of source-related challenges within the framework of historical linguistics in “Currents” (pp. 14-37) and HLBH (pp. 21-45). Nonetheless, we would like to expand here on the claim that linguistics provides a “more objective control.”

Linguistics encompasses both highly objective and more interpretive domains of research. Fields such as phonetics are grounded in empirical data and formalized methods; measuring acoustic properties or articulatory movements yields testable, replicable results. By contrast, historical linguistics often relies on fragmentary evidence and reconstructed data, where human judgment and methodological choices play a more significant role. The nature of the sources—often indirect or incomplete—is further complicated by whether they are inscriptional or literary, unedited or editorially-scribally mediated, with each type presenting distinct challenges in reliability and interpretation. Thus, while some linguistic subfields approach the rigor of the natural sciences, others resemble historical or interpretive disciplines, where conclusions are more tentative and shaped by subjective evaluation, theoretical commitments, and evidentiary constraints. This subjectivity is even more pronounced when the sources are editorially-scribally mediated (i.e., texts that have been edited and copied), as opposed to inscriptional sources (i.e., original, unaltered inscriptions, which, sadly, are scarce for ancient Hebrew). Therefore, asserting the objectivity of “historical linguistic typology”—particularly in biblical studies—as a secure and independent control over literary-critical models is problematic. Both approaches are interpretive in nature and demand careful, critical engagement rather than hierarchical privileging. This is the raison d’être for our pursuit of a more integrated methodology for dating biblical writings, a point of agreement also acknowledged in Ron Hendel’s comments earlier.

Hi, Martin. Though:

1) Some poetry stands alone, retained, without an available close parallel to compare. And why focus only on "less common linguistic features"?

2) Given that you agree "in viewing the Pentateuch as having a long compositional history," I nonetheless sense (so far) that you all tend to present the most recent composition possible rather than the most plausible.

A point of sad information. After I posted here, I learned that Russell Gmirkin had died on July 17. His good friend. Gregory Doudna, also informed that he had completed a book manuscript with his interpretation of Kings.

Hi Stephen

1) On common vs. uncommon language features: see our response to Longacre, above; also Currents, pp. 41-42; cf. 29-37; HLBH, pp. 110-115; ch. 5; and really throughout entire book; etc.

2) We are uncertain what you mean. Please elaborate.

We are sad to hear of Russell's passing.

Dear Robert (if I may),
Thanks for your thoughtful interaction with my blog comments. I think we can probably all agree (if we lay down our disciplinary preferences/specialities for a moment) that an integrated approach is necessary for dealing with this complex corpus, and I appreciate your push in this direction. I would also agree that--in principle--no one data set or discipline should a priori be prioritized in a hierarchy above another, and none of our historical disciplines can achieve that highest-level, reproducible objectivity you talked about. At the same time, not all data sets and methods are equally well-suited to answering every question, and an integrated approach needs to delineate how exactly these methods relate to most effectively answer the dating question. No doubt your research is already attempting to provide that answer, but it seems to me like a method that builds on what can be most confidently known and what is most chronologically indicative is eventually going to need to weight the various data sets and methods differently.

I'm happy we can share a degree of common ground. I would say that doubting "any connection" between the MT and original authors is quite a strong claim though, one that seems overstated to me. Not only would that substantively undercut linguistic methods for dating authorial texts, but it could also bring the entire text-critical enterprise (and probably other methods within the historical-critical toolbox) into question. In this regard, I would say I am generally more optimistic than you. Best wishes!

A statement agreeing "about the necessity of an integrated approach" may be undercut by considering "language one of the less useful parts of the dating puzzle." Dating texts, i.e., instantiated language.

Hello everyone

As someone who recently completed a dissertation on Linguistic Dating and the Book of Isaiah, I affirm the call for methodological caution in this article. During the course of my research, I perused several of your documents and derived considerable benefit from your invaluable contributions, which furthered my understanding of the subject matter.

In my study, I examined the extent to which linguistic data in Isaiah can meaningfully inform questions of dating. What emerged is that linguistic features cannot be interpreted in isolation: any diachronic claim must be tested against assumptions about the dating of other biblical books, the nature of editorial processes, and the role of genre, sociolect, or idiolect. Indeed, over two-thirds of the linguistic features proposed for dating the Book of Isaiah have not withstood rigorous examination. Nevertheless, there are some linguistic features that appear to support a pre-exilic or post-exilic classification of text blocks within the Book of Isaiah. These, of course, also vary depending on how other biblical books are dated.

In the fifth chapter of my work, I present a discussion of these linguistic findings in the context of various approaches to the Book of Isaiah. To achieve this, I have orientated myself along three principal lines of enquiry: the "traditional" approach, the redaction-critical/literary-critical approach and the synchronic approach.
It has been demonstrated that divergent assumptions concerning the Book of Isaiah can be accommodated by the linguistic data in different ways. Nevertheless, it is possible to evaluate which assumptions and dating models can be most effectively linked with the linguistic data.

I am in contact with a publisher and hope to be able to publish my work soon.

Greetings
Samuel

Dear Drew and Stephen,

Thank you for your comments. We now recognize that it would have been helpful to phrase our preceding comment about language differently or to clarify more precisely what we intended by it. Specifically, the catalog of forms and usages conventionally employed to sequence biblical writings tends to be less common, more superficial, and less stable in the writings and in their transmission. Our conclusion is not the result of a dogmatic stance on language, nor does it reflect any neglect of linguistic research. Rather, it arises from the detailed analysis presented in our numerous publications, as well as from other features we have studied but not yet published, which consistently raise as many distributional red flags as green ones when used for trying to date the sources in which they appear. We have approached this issue from multiple angles: comparing between CBH and LBH writings, comparing within CBH writings or within LBH writings, comparing across MT parallel texts, and comparing across the MT and other textual traditions. Regarding MT parallel texts, for example, we conclude:

“All of the evidence from parallel texts in the MT indicates that the surviving texts of the biblical books are late, thoroughly revised and reworked versions of earlier texts whose distinctive linguistic features are lost to us. We would like to emphasize the word “distinctive” in this statement. We say again that the basic features shared by almost every BH writing—“early” and “late,” MT and non-MT—are quite stable, which is not unexpected since BH as a whole is a standard literary language, so that the surviving manuscripts give us no reason to doubt that the basic features of BH were characteristic of earlier compositional stages of the biblical books....” (HLBH, p. 167)

“In summary, all the evidence from the parallel texts in the MT itself points in the same direction: Large-scale and basic features of Classical Hebrew only relatively rarely show variation. Less common features of Classical Hebrew are highly fluid, and the current distribution of such forms cannot be relied on as evidence of the language of particular authors at particular times and in particular places. Nevertheless, it is precisely these less common features that have played a big role in historical linguistic and linguistic dating studies of BH writings.” (HLBH, p. 168)

We hope this will help clarify what we meant in context with our “We would just say that, in our view, our research makes dubious any connection between the language of the MT and that of “original authors” in their historical setting. We therefore consider language one of the less useful parts of the dating puzzle.” Our comment, our conclusion, revolves around the issue of the “distinctive” or “less common” language features, rather than the basic or common ones that are pervasive and stable. See especially “Currents,” pp. 29-37, where we look specifically at textual stability on the one hand, and textual instability or fluidity on the other, in relation to “original” biblical language and texts.

Dear Samuel,

I tracked down your dissertation and spent some time reading through it. I’ll need to read it more carefully, especially your analysis of linguistic features in chapter 4, but it’s already clear that your contribution is an important one. In particular, you offer many insightful observations concerning underlying assumptions. I look forward to reading further and continuing the conversation.

I am not a player in this debate, but I thought I would put in a few comments from the bleachers. First, the claim that Hornkohl’s work is “an old paradigm trying to survive in a new era” seems disingenuous. I get the sense from everything I have read – and I don’t think I would be alone in thinking this - that the basic methods and conclusions underlying Hornkohl’s work (and being questioned here) are alive and well in the scholastic world. Second, I do not see any reason to think Hornkohl is discounting data that would reveal his argument to be a house of cards (the implication of saying he “highlights evidence that fits the preconceived model while sidelining anomalies and counterexamples”). If Hornkohl’s arguments really did amount to that, it should have been easy enough for this post to give readers that data instead of a few “illustrative examples” that by no means overthrow Hornkohl’s conclusions. Third, to say that Hornkohl fails to “fully grapple” with the complex compositional and transmissional history of the Hebrew Bible is not a valid criticism. No one could ever “fully grapple” with all of it. Nor should anyone try to. Not all data is either equal or relevant. And the parts that Hornkohl *does* grapple with are sufficiently discussed with caveats and alternate evidence and with reference to multiple textual sources (like the Samaritan Pentateuch or biblical and non-biblical Dead Sea Scrolls). Fourth, this whole post is mostly concerned with disagreeing with some of Hornkohl’s underlying methodological conclusions, but unlike Hornkohl's work which is being criticized in this post, no solid argumentation and analysis is provided in this post to give those disagreements weight. After reading Hornkohl’s work, I am not convinced of his main thesis. But not because of his methodological errors or inability to analyze a wide array of textual data that might or might not support his arguments. The authors of this post would be well-advised that my conclusion in that respect will not be limited to me.

Dear David,

Thank you for taking the time to engage with our article, even from the bleachers, as you put it. Since you’re observing from outside the ongoing debate, it may be helpful to note two things:

First, as with any complex game, much of the nuance and strategy becomes visible only from the field. Many of the “rules,” both methodological and evidentiary, that shape the current conversation are best understood by those who are active participants in the scholarly debate. That context often clarifies why certain approaches or critiques carry weight within the field. This particular debate is especially intricate, as it involves not only linguistic analysis but also textual, literary, and other disciplinary dimensions that interact in subtle and often contested ways.

Second, our article was meant as a brief synopsis rather than a full exposition. For readers looking for more sustained argumentation, including fuller engagement with data and alternative models, we’ve already published these discussions in more detailed venues (as cited, e.g., in HLBH and “Currents”). There, we take up exactly the kinds of issues you raise here with a broader evidentiary base and more extensive argumentation.

Third, you note that you are not a player in this debate, but since you’ve read Hornkohl’s book and presumably have a good command of Hebrew, we’re genuinely curious to hear what specific reasons led you not to be convinced by his main thesis, namely, that the entire Pentateuch and its language date to before 800 BCE. That kind of critical engagement would help advance the conversation constructively.

Again, we appreciate your thoughtful remarks and the opportunity they offer to clarify our intentions.

Editing is possible, but I am old enough to remember the bogus proposal that all Qumran mss had "depositation" in 63 BCE, hence presumed different from small, selectively excluded Masada mss.

Though it is true that there are not many inscriptions to compare "in viewing the Pentateuch as having a long compositional history," I wonder what the authors might write about Pentateuch verses written in Ketef Hinnom silver amulets and "Deir 'Alla" inscriptions.

I’ve been following this admirable site for many years and seen this dispute, with much the same protagonists, pursued more than once - and I’ve asked what method can there be for distinguishing a writer writing in old times, in old time style, from one writing in later times but wishing to use the old time style with which he is familiar - maybe thinking that this is how one should address important matters? I still want to ask that

Dear Martin
This is an important point, because we have argued that there is in fact no method to discover late use of early language in the Bible. The traditional stance, oft repeated in e.g. Avi Hurvitz' writings, is that later authors would invariably betray their lateness by not mastering the classical language, using newer words, expressions, and syntax now and then. This stance is puzzling to us, not only because of text-critical problems. How could we tell if someone did write perfect classical Hebrew after the exile? Besides, there are words, expressions, and syntax considered characteristic of late Hebrew in all texts in the Hebrew Bible, the main difference being one of accumulation. In order for e.g. Hurvitz to declare a text late, there needs to be a number of late features in it.

Your point about addressing important matters in classical language is also correct. Specifically the prophetic genre, which mostly uses classical language. We have post-exilic books such as Third Isaiah, First Zechariah, Haggai, Joel, and Malachi, which in their use of language align with classical biblical Hebrew, without clear late features. Pentateuchal literature is another example. P is seen by most scholars as postexilic, with the notable exception of many language scholars who on the strength of its classical language alone date P early. The book of Ruth is another such example.

Linguistic daters claim that there are no non-biblical examples of postexilic classical biblical Hebrew, or at least no extensive ones, and that it was therefore not feasible to write classical biblical Hebrew in the postexilic period without betraying one’s late origin with slip-ups. But take a look at Ian Young’s work on Pesher Habakkuk, ‘Late Biblical Hebrew And The Qumran Pesher Habakkuk,’ Journal of Hebrew Scriptures 8 (2008) (https://jhsonline.org/index.php/jhs/article/view/6223). Young shows that the pesher is written in classical biblical Hebrew plus Qumran dialectal features. In other words, it is not written in late biblical Hebrew.

It really comes down to dating criteria: Linguistic daters privilege language, others consider a variety of data. We also point out that the linguistic data that is harnessed for dating is already based on preconceived notions of what writings are already early and late, so the entire linguistic dating paradigm is circular.

See also Martin Ehrensvärd, 'Why Biblical Texts Cannot Be Dated Linguistically,' Hebrew Studies 47 (2006), 177-89 (https://www.academia.edu/30537634/Martin_Ehrensva_rd_Why_Biblical_Texts…).

As the Dead Sea Scroll contain all the books of the O.T. except for the book of Esther, do all the various forms of Hebrew you discuss also appear in the DSS?