What kind of a linguistic record is the Hebrew Bible? More--was it meant to record a language?
The sociolinguist Sarah Roberts comments, apropos of my previous post, that, in trying to see how complete a picture your written sources give you of a language's lexicon, looking for spoons is not a bad way to go. That is, the method Ullendorff uses (seeing how well those pedestrian, daily-life words are covered; interestingly, this is not at all the same as making a Swadesh list) proves useful, but "Another approach is to calculate the proportion of hapax legomena in the corpus; the higher the proportion, the less representative the corpus usually is. It is also important to pay attention to the kinds of literary genres that comprise your sources..."
Ullendorff is way ahead of me here as well: he cites the great (greatest?) Semitist Noeldeke who had already, in the classic "Semitic Languages" article in the classic 13th edition of the Encyclopedia Brittanica (from 1912, if memory serves, but still unequalled*) noted that the numerous hapaxes are "a sufficient proof that many more words existed than appear in the O.T." Ullendorff goes on to cite a list of these hapaxes which, at 2,440, would constitute about a third of the vocabulary of the Bible! Others have produced somewhat lower numbers but the point is made.
He also notes words that we would have expected to find in Biblical times, for example the Mishnah's massu'ot "fire signals," (as opposed to the Tanakh's semantically diffuse mas'et, which can be anything from "portion of food" to "tax" to "pillar of smoke"; see discussion in comments below) which we then dug up out of the ground, on an ostracon at Lachish (4:10). His conclusion is that Biblical Hebrew is more of a "linguistic fragment," "To be sure, a very important and indeed far-reaching fragment, but scarcely a fully integrated language which in this form...could ever have been spoken and have satisfied the needs of its speakers. The evidence presented by the epigraphical material contemporary with the OT and by the Mishna, its immediate successor, underlines the essentially fragmentary character of the language of the Hebrew Bible. And there is a strong case, in my submission, for looking upon the language of the Mishna as the developed colloquial--otherwise so largely, though by no means wholly, repressed and curbed--of the predominantly formal and elevated diction of the OT."
Ullendorff's article is a shrewd, and remarkably fun piece of work (see the second essay in the volume, "C'est de l'Hébreu pour moi!" a delightful study springing from his discovery that the French expression for the (Shakespearean) "It's Greek to me!" is "It's Hebrew to me!") , but it only scratches the surface. For one thing, he leaves out much of what makes language work: the verbal and deictic systems, the inventory of registers, speech genres, ways of indicating person, status and relationship. In this he is not alone: read some of William Safire's "On Language" columns for a weekly dose of the folk-theory that language is just a bag of words.
In the case of Biblical Hebrew, casting a wider grammatical net may catch only an even greater sense of vertigo, because while translations generally render the Bible into one type of English, Biblical Hebrew itself is not linguistically uniform. Ullendorff could as well have spoken of "shards" as of a "fragment." Reading along one encounters not just different sets of vocabulary and spellings, but even different verbal systems that appear to handle tense, mood and aspect in at least three different ways. Scholars have therefore long argued for at least three types of Hebrew: Archaic (usually said to be exemplified by Exodus 15, Deuteronomy 32, and Judges 5, the "Song of the Sea," the "Song of Moses," and the "Song of Deborah," each marked as poetry (Hebrew uses related terms for these, different derivatives of the root sh-y-r), Standard (usually taken as the bulk of the Torah plus Joshua through II Kings), and Late (Chronicles, Ester, Ezra-Nehemiah).
Yet though composed of shards, Biblical Hebrew is not broken. This is because the language was integrated by a group of Jewish Aramaic (not Hebrew!)-speaking scholars near the coast of Palestine, in Tiberias, who provided it with vocalization and speech melody based on ancient traditions of their own around the 7th and 8th centuries C.E., somewhere in the neighborhood of 1,500 years after the first texts were probably written down in Standard Biblical Hebrew. What is remarkable about this system of vocalization and cantillation is that it gives far more information than necessary for verbal understanding: one of the things that makes the Tiberian vowel system confusing is that, while it almost always gives enough information to tell different words with identical consonants apart, it is obviously not designed to do that; rather, it's designed to record the exact sounds produced by a tradtionally correct liturgical reader in the synagogue. In other words, it's more like Sanskrit, with its elaborate notation of strictly phonetic phenomena, than it is like the more matter-of-fact Greek or Arabic.
This linguistic fact has interesting consequences for popular things like the study of Midrash and the always blossoming fields of Biblical interpretation. This is because the cantillation marks, rarely taught in Biblical Hebrew class, in fact seem to set forth a set of very precise instructions for prosody; that is, they tell you how to intone and express the content of the text. In an environment where the significance of the text was, to put it mildly, disputed, the Tiberians produced a text that not only could only be read one way, but that tried to turn its readers into human tape recorders, playback machines that ventriloquized God's word.
If this is true, could it tell us new things about what the people behind the Masoretic tradition thought Scripture was? One of the great frustrations in reading a wonderful book like Michael Fishbane's Biblical Interpretation in Ancient Israel is that he never renders explicit what he thinks the scribes' precise notion of the text was--what gave them the right to do the things they did to it? What constraints were they under and how did they conceive of what they were doing? His student Bernard Levinson has taken some major, equally wonderful steps towards figuring out a scribal view of the text in Deuteronomy. But what about the people who put the end result together?
That's for a future time. Now, as I promised, we go back to the Iron Age.
*A rare personal note: the memory in question dates years back, to a cherished moment at the Albright Institute in Jerusalem, sleeping on the floor of the library after reading late into the night and all the buses had stopped running.
ARAMAIC ADDENDUM: Ed Cook quite rightly asks where the curious btdwd' kitchen text can be found. My revered teacher Delbert Roy Hillers edited it with Eleonora Cussini as PAT 2743:8, and they cite an original publication in Syria 1926; it was conveniently republished in Rosenthal's Aramaic Handbook, entry 13 under Palmyrene. Curious readers who examine the original will see that something is indeed being cooked up here, but it is not food :-).