Skip to: Site Menu | Main content

Unhistorical Hebrew Linguistics: A Cautionary Tale

See A Very Tall “Cautionary Tale”: A Response to Ron Hendel

By Ronald Hendel
University of California, Berkeley
September 2011

While prowling around this website, I came across a surprising article on “Linguistic Dating of Biblical Texts” by Ian Young, Robert Rezetko, and Martin Ehrensvärd.1 (It’s a potted summary of their two-volume work of the same title.) Now Hebraists have long known that Biblical Hebrew changed over time, from the earliest to the latest biblical books. This was established by Gesenius in his 1815 monograph, Geschichte der hebräischen Sprache und Schrift, and many scholars have contributed to this topic since. But Young, Rezetko, and Ehrensvärd say that this scholarship is incorrect. They argue that “the best model for comprehending the evidence is that ‘Early’ BH and ‘Late’ BH, so-called, represent co-existing styles of Hebrew throughout the biblical period.” This is a large claim. For a claim this size, one imagines that they have gathered compelling evidence and can marshal strong arguments. Well, nope. Their arguments have as many holes as Swiss cheese, and their evidence falls apart as soon as one looks closely.

I think this is a shame, because it shows that Hebrew linguistics has come to a sad pass. Most biblical scholars don’t seem to know how to evaluate linguistic and textual data. This is a bad situation, because these are the tools we need in order to construct any kind of sophisticated literary, historical, or cultural reading of biblical texts. These are the building blocks, and without them the whole structure falls down. As my grandmother would say, Oy vey.

Let me dip into some of the problems here. The authors use the example of the lexical contrast of mamlakah and malkut. According to the standard model, mamlakah is the normal word for “kingdom” in Classical Biblical Hebrew (CBH) and was replaced by malkut (from Aramaic) in Late Biblical Hebrew (LBH). This lexical contrast, along with a whole chain of other lexical, morphological, and syntactic contrasts, helps to map the changes from CBH to LBH. The authors contest this model by noting that malkut occurs a few times in books that are not usually considered to be late: “seven of those 91 occurrences in the Hebrew Bible are found in core EBH books like Samuel and Kings.” The rightly note that if the word occurs in early texts, then the word cannot be an indicator of late language. Rather, it must be an option at any time. Therefore LBH features aren’t necessarily late.

But the evidence belies this argument. The instances where malkut occurs in putatively early texts are problematic in one way or another, and as such cannot bear the argument the authors advance. For instance, they say that it occurs in Samuel. Right, it occurs in MT of 1 Sam 20:31. But the MT is notoriously corrupt in Samuel. They don’t tell us (perhaps because they don’t know) that 4QSamb has a clearly superior text in this clause, based on criteria other than this word. 4QSamb reads mamlakah, not malkut. So this evidence for the LBH word in Samuel flies away like the wind.

In the MT of Jeremiah, malkut occurs three times. But two of those instances (10:4 and 49:34) are lacking in the LXX, which represents the earlier edition of the book. This means that malkut was used in these verses only in the later edition, which is precisely what one would expect. The third instance (Jer 52:31) is a doublet of 2 Kgs 25:27. The Kings passage reads malko, not malkut. Since the Jeremiah passage is probably a late revision of the Kings passage, this evidence too flies away like the wind.

The one instance of malkut in Kings (malkuto in 1 Kgs 2:12) is also problematic. It is in a formula that is repeated, with variation, in the last verse of the chapter (2:46). The latter verse uses mamlakah. In this context, malkuto in 1 Kgs 2:12 may well be a secondary reading (or possibly a misvocalization of melukato). The solution is unclear, but it’s anomaly under any construal.

My point here is that linguistic arguments depend on reliable assessments of the evidence. The authors are playing fast and loose with the evidence. Why are they doing this? Perhaps they don’t really understand the evidence, or perhaps they don’t care. Maybe making a large argument—however dubious—is its own reward. In this era of publish-or-perish, maybe publishing a spurious argument is enough.

Here’s another example. They say that Persian loanwords aren’t indicative of texts written after the onset of the Persian period. Their reasoning is as follows: “we argue that it is very likely that the MT intends us to read the Persian words dat in Deuteronomy 33.2, raz (“secret”) in Isaiah 24.16, and peladot (“steel”) in Nahum 2.4.” This is an unobjectionable statement, because the medieval Masoretes did vocalize the words this way. But it does not support their argument. These readings are clearly textual problems. Deut 33:2 can’t possibly mean “fire of the law.” Isa 24:16 doesn’t mean “secrets.” Nahum 2:4 doesn’t mean “steel.” The MT vocalizations are dubious construals of difficult texts. (Deut 33:2 may have suffered a graphic error [resh-dalet], Nahum 2:4 a metathesis [pe-lamed], and the word in Isa 24:16 is from a different root—and in any case is in a late text). In other words, these verses provide no warrant for the authors’ argument. It’s misleading—or simply untruthful—to assert that Persian loanwords occur in these verses. This dog won’t hunt.

There are other ways in which their argument falls flat. We now know that the Septuagint translators and their contemporaries often read CBH words and constructions as if they were LBH.2 This means that erudite scholars had already forgotten details of CBH by the third century B.C.E. So we’re clearly not talking about contemporary literary styles. We’re talking about linguistic history. It’s a complicated linguistic history, to be sure, and we lack comprehensive documentation of all the changes. But changes there were.

Another point – the authors claim that the textual history of the Hebrew Bible makes it impossible to study the history of Biblical Hebrew. They write: “text-critical evidence, therefore, puts a question mark over the whole enterprise of linguistic dating before it has begun.” This is nonsense. My examples above demonstrate that textual criticism is entirely compatible with historical linguistics, and, indeed, that the two pursuits are necessary adjuncts. The authors’ argument evinces a tenuous grasp of the practice and implications of textual criticism. I think that the opposite of their claim is correct. The textual history of the Hebrew Bible corroborates its linguistic history, as indicated above with the two editions of Jeremiah.3

It is obvious that Biblical Hebrew has a history, and if we are patient and respect the data, we can discern the nuances of these changes. Gesenius and his successors (Driver, Kutscher, Hurvitz, Joosten, et al.) weren’t wholly wrong. We can date our texts linguistically, with varying degrees of confidence. But the knowledge and skills necessary to do this are ephemeral. If one generation loses them, they may be forever lost. This is my worry. In the long run, bad scholarship plows over the good, until at the end it’s all wind and fury. Biblical philology is a delicate thing, which we need to tend with loving care, or else it is lost.


2 See Jan Joosten, “Pseudo-Classicisms in Late Biblical Hebrew, in Ben Sira, and in Qumran Hebrew,” in T. Muraoka and J. F. Elwolde, eds., Sirach, Scrolls and Sages (Leiden: Brill, 1999), 146-159; idem, “Biblical Hebrew as Mirrored in the Septuagint: The Question of Influence from Spoken Hebrew,” Textus, 21 (2002), 1-19. See also Joosten’s review (in press) of Young, Rezetko, and Ehrensvärd, available at rc_Joosten.pdf.

3 See further Joosten, “L’excédent massorétique du livre de Jérémie et l’hébreu post-classique,” in J. Joosten and J.-S. Rey, eds., Conservatism and Innovation in the Hebrew Language of the Hellenistic Period (Leiden, Brill, 2008), 93-108.