Overview of Hebrew history
Although not specifically asked for, it is profitable to start with a brief overview of the history of the Hebrew language. For more details on the subject, see A History of the Hebrew Language by A. Sáenz-Badillo (from which the following information comes).
Hebrew is a member of the Semitic language family, a group of about 70 languages historically spoken in Northeast Africa and the Middle East. According to the field of Linguistics, the Semitic languages derive from a single mother language, labelled as Proto-Semitic. Although unattested from archaeology, linguists have largely reconstructed the language and believe it was spoken throughout the region from 3500-3000 BC.
Around 3000 BC, the oldest Semitic language, Akkadian split off in the east. The northwestern (which includes Hebrew) and southern branches split around 2000 BC. The Canaanite languages in the north then diverged from Arabic in the central region. In the far north, the cultural center of Damascus eventually enabled the local dialect, Aramaic, to become the lingua franca of the north. The lack of a major political center to the west meant that none of the other Canaanite dialects achieved dominance. Thus each area had its own language, although they were mostly mutually intelligible. (Biblical Hebrew shares ~80% words in common with nearby languages such as Phoenician and Ugaritic.)
It is hard to say exactly when Hebrew became a distinct language. The oldest accepted inscription dates from 1000 BC, but the archaeological record for the second millennium BC is pretty sparse in the area, so it could have diverged earlier than that. The oldest form of Hebrew is generally labelled proto-Hebrew or Archaic Biblical Hebrew (ABH). Conventional wisdom is that this form is represented in the Bible in only a few poetic passages.
The form of Hebrew used from ~950 BC until the start of the Babylonian exile in 586 BC is known as Classical Biblical Hebrew (CBH), Standard Biblical Hebrew (SBH), or Early Biblical Hebrew (EBH). During this period, the language shows minimal outside influence. For example, there are few attested loanwords. This form of Hebrew is certainly found in the Bible, although there are disagreements about just how much of the Old Testament is written in EBH. (More on this in later sections.)
The period from the Babylonian exile until about Alexander conquered Judah in 332 BC saw the use of what is known as Late Biblical Hebrew (LBH). This period shows influence from Persian and to a lesser degree Aramaic. The majority of the Old Testament is usually assigned to this period. (Again, see below for a critique of this position and further details.)
Hebrew from 332 BC until the destruction of the Temple in 70 AD is labeled Dead-Sea-Scroll Hebrew, so named because it is primarily attested to from the Dead Sea Scrolls. This period saw heavy influence from Aramaic and some influence from Greek.
The final period of Ancient Hebrew is known as Mishnaic Hebrew and lasts until the 2nd or 3rd century AD, when the language ceased to exist as a spoken language. Hebrew remained a scholarly language, however, leading to Medieval Hebrew. Finally, a resurgence in interest in the language in the 19th century led to the creation of Modern Hebrew which again became a living language, culminating in the (re-)creation of the Israeli state in 1948.
It should be noted that the above distinctions and especially the start/end dates are for convenience - people obviously did not just stop speaking a language one day. Language change is a slow process that is only observable in retrospect.
Difficulties of dating the Old Testament
Having now outlined different periods of Hebrew, one might think it would now be easy to get a rough idea when a book was written. One would be wrong, as there are several significant difficulties in dating based on linguistics. Among them, Craig Davis (Dating the Old Testament) lists:
The corpus to compare from is extremely small. In particular, the entirety of the extra-Biblical pre-exhilic text amounts to a couple pages worth. This makes it very difficult to say how much variation is from different authorial styles and genres and how much is from language change.
It is possible that later writers purposely imitated the style of the older, accepted texts.
There is solid evidence that it was acceptable for scribes to update the text, at minimum in regards to spelling and possibly also word choice and grammar, for a period of time.
Our understanding of the language's history has been derived primarily from the estimated (on other grounds) dates of Biblical books, making linguistic arguments "confirming" such dates in danger of exhibiting circular reasoning.
To this Gary Rendsburg ("Late Biblical Hebrew and the Date of 'P'"), based on the work of Joshua Blau, adds:
- The difference between different Biblical texts is unexpectedly slight. That is, the language is largely uniform despite surely spanning at least several hundred years.
Proposed differences
To simplify things a bit, scholars generally divide the Old Testament into just two categories: Early Biblical Hebrew and Late Biblical Hebrew. In 1976, Robert Polzin (Late Biblical Hebrew) attempted to systematically define the differences between EBH and LBH by comparing Chronicles with Kings & Samuel. The approach has some merit: Chronicles being significantly more recent than Kings/Samuel is one of the few relative dates that don't rely on presuppositions and\or the reliability/historicity of internal evidence (Chronicles is certainly dependent on Kings/Samuel). It also has a significant weakness: Chronicles copies wholesale from Kings in places and also says it is utilizing other sources, so it is hard to say how much of the language is contemporary and how much comes from sources.
Anyway, Polzin's work attracted a significant amount of attention, so it is worth looking at his conclusions. He proposes 19 differences in LBH as compared to EBH based on his study.
A. Differences not from Aramaic influence:
- Radically reduced use of 'et with pronominal suffix.
- Increased use of 'et before noun in the nominative case: 'et emphatic.
- Expression of possession by prospective pronominal suffix with a following noun, or lə plus noun, or sel plus noun.
- Collectives are construed as plurals almost without exception.
- A preference for plural forms of words and phrases which the earlier
languages uses in the singular.
- The use of an infinitive absolute in immediate connection with a finite verb of the same stem is almost completely lacking in the Chronicler; the infinitive absolute used as a command is not found at all in Chronicles.
- The Chronicler's use of the infinitive construct with bə, and kə; as Segal (l927) points out, the later books of the OT show less frequent use of the infinitive construct with bə and kə even in the cases when LBH does use the introductory infinitive with bə, and kə, the usage is different from earlier constructions.
- Repetition of a singular word = Latin quivis (to express distributive).
- The Chronicler shows a merging (i.e., a tendency to replace) the third feminine plural suffix with the third masculine plural suffix.
- The first person singular imperfect with -ah (the lengthened imperfect or cohortative) is found but once in the Chronicler's language.
- The use of wayhi greatly recedes in Chronicles and in the younger language.
- In appositional relationship the Chronicler prefers to place the substantive before the numeral and almost always puts it in the plural; this is contrary to the older practice of putting the number first.
- The Chronicler shows an increased use of the infinitive construct with lə.
B. Differences from Aramaic influence:
- In citing material and its weight or measure, the Chronicler often has the order, material weighed or measured + its weight or measure (+ number).
- lə is used often as a mark of the accusative.
- In min 'from', the nun is often not assimilated before a noun without an article.
- The Chronicler uses the emphatic lə before the last element of a list.
- In an attributive usage, rabbim is twice placed before the substantive.
- The use of 'ad lə, [for 'up to', 'until'].
Few, if any, scholars have accepted all the criteria as being valid markers of EBH vs LBH, as opposed to authorial style (or nothing at all). However, many have accepted at least some of them. For example, Rendsburg finds A1, A5, B2, & B3 to be correct; A2, A4, A6, A11, A13, & B5 to be unconvincing and/or likely to be from authorial style; A7 & A8 to be mixed (true for some expressions, invalid for others); and A3, A9, A10, A12, B1, & B4 to be completely invalid. (For a more detailed statistical analysis of Polzin's conclusions, see "Dating Second Zechariah: A
Linguistic Reexamination.")
Subsequent researchers (Avi Hurvitz is a leader in this area) have come up with other markers by broadening the number of texts for comparison in each category. Unfortunately, using more texts to form hypotheses leaves less texts to test the hypotheses generated and creates more danger of conclusions thus being entirely circular.
More recent work
The 2003 volume Biblical Hebrew: Studies in Chronology and Typology (ed. Ian Young) attempted to provide an overview of current scholarship by presenting essays from people with a variety of views on Biblical Hebrew. In the process, the work reignited debate in the field about how valid the existing paradigm really was.
The first half of the the work presents the views of those who support the "standard" chronological framework. These essays primarily focus on loanwords and "Aramaicisms" to support the critical view of most of the Old Testament being dated late. At least one contributor, Frank Polak, does make an argument on grammar, but it isn't the dominant tactic. This suggests to me that efforts to define clear, grammatical difference between EBH and LBH, like those outlined in the previous section, have largely failed to produce persuasive evidence.
The second half of the work presents essays by scholars who challenge the "standard" chronological framework. These authors don't find any substantial difference between Biblical texts (in contrast with, for example, noticeable differences between the Bible and the Dead Sea Scrolls). They variously argue that this means that EBH continued to be used in literary circles in the post-exhilic times or that the language changes of the period were primarily phonetic and thus not reflected in the written text.
Further work by Young, Rezetko, and Ehrensvärd (Linguistic Dating of Biblical Texts) has proposed a new idea: the so-called EBH and LBH variations don't represent different time periods, but rather dialectal variants that co-existed throughout the Biblical period. A chief tenant of their argument is that texts typically labelled as LBH almost universally contain some so-called EBH elements, a fact that supporters of the chronological framework readily admit but attribute to intentional "archaicizing" or "patriarchal coloring".
They challenge whether the accepted LBH features actually mean anything or are just the result of statistical anomalies. Given a small corpus and a predetermination that certain texts are late/early, there are bound to be some features that occur much more often in one set than the other, even if there is no real difference in the language used. They note that conclusions about dating are often circular arguments. For example, Job is supposed to be a late work, so the small amount of late elements in it proves the author was trying to imitate earlier styles but failed to do so completely, while Kings is supposed to be an early work, so the small amount of late elements it in must be from later scribal changes. (Their analysis shows that every book of the Old Testament contains at least some so-called LBH features.)
Young et al. suggest such data is more easily explained by multiple dialects of Hebrew being in active use rather than by authors imitating old grammar. From there, they strongly challenge the idea that the amount of loanwords in a text is a reliable indication of date. For example, they show that many works universally dated late have few or no loanwords. They also challenge how reliably one can really determine loanwords from Persian, noting that certain words in early texts could be considered loanwords based on having the same form, but are rejected as such because borrowing is considered impossible for the time period. (They also say that dating by Aramaicisms is already out-of-style, so spend little time on it.)
Naturally, the work of Young et al. has been challenged by those who hold other views. For example, the 2012 compilation Diachrony in Biblical Hebrew is largely a response to Young's critiques, featuring essay that seek to defend the chronological framework with more vigorous methodology. The dialects hypothesis certainly has not displaced the chronology framework, but the work has had a noticeable effect. Specifically, researchers are generally more cautious about drawing conclusions on linguistic evidence. For example, Robert Holmstedt (who supports the chronological framework) writes:
In a nutshell, while we [Holmstedt and John Cook] agree with the weaker hypothesis that texts cannot be dated absolutely by linguistic means, we disagree with the stronger hypotheses is that no dating at all is possible and argue that the relative dating of features, and the texts in which they cluster, is possible if the analysis is carried out with a sound linguistic and philological methodology. ("Biblical Hebrew Diachrony (continued, again)")
A very recent work, Early Biblical Hebrew, Late Biblical Hebrew, and Linguistic Variability takes yet another approach, seeking to explain the observed differences in style via sociolinguistics. Author Dong-Hyuk Kim concludes that while to texts to exhibit changes that appear to be time-dependent, it is virtually impossible to use those clues to date a given work because it is impossible to say whether a given author was an early or late adopter of a particular feature.
In summary, I think it is fair to say that the question of how much linguistic difference between books of the Old Testament reflects language change is an open question with no strong consensus one way or another.
Conclusion
To directly answer the question of "How much variation is in the Hebrew Language of the Old Testament?" the answer is not much (after controlling for genre differences - naturally poetry uses quite different syntax than prose.) Biblical scholars have proposed dividing the Biblical language into two basic categories - Early (pre-exhilic) Biblical Hebrew and Late (post-exhilic) Biblical Hebrew. A number of markers have been proposed, but there is a wide variation between scholars as to which constitute useful markers and which are false positives. Nonetheless, for a while this distinction was generally accepted as valid.
More recently, the framework has come under attack on the basis of it being largely based on circular reasoning - text dates were largely decided before the markers were identified and those dates were used to search for the linguistic differences. Furthermore, the small corpus makes any conclusions quite tentative. Thus, there is significant doubt about the applicability of using linguistic clues to date the Old Testament.
That is not to say Hebrew did not evolve over the Biblical period - it certainly did. But, the change was probably not very substantial and what change there was is not well reflected in the Bible because of a combination of literary Hebrew not matching spoken Hebrew, authors possibly using archaic forms on purpose, and scribes possibly updating texts over time.
Sources
"Biblical Hebrew Diachrony (continued, again)" (2011) by Robert Holmstedt
Biblical Hebrew: Studies in Chronology and Typology (2003) by Ian Young (ed.)
Dating the Old Testament (2007) by Craig Davis
Early Biblical Hebrew, Late Biblical Hebrew, and Linguistic Variability (2013) by Dong-Hyuk Kim
A History of the Hebrew Language (1993) by A.Sáenz-Badillos
Late Biblical Hebrew (1976) by Robert Polzin
"Late Biblical Hebrew and the Date of 'P'" (1980) by Gary Rendsburg
Linguistic Dating of Biblical Texts (2009) by Ian Young, Robert Rezetko, and Martin Ehrensvaerd