tapirrider wrote:It might matter to those with doctorate degrees who have spent their lives in research being challenged by someone lacking the degree or experience in ancient Middle Eastern languages who is proposing just another variation of a long discarded theory.
Whether Stubbs got his PhD is totally irrelevant in a sense, though I suspect if he'd finished he would have been Lyle Campbell's student, and I wonder what Lyle Campbell thinks about this work. His is a review I would take more seriously than a BYU professor and friend of Stubbs. Dirk Elzinga's review in
BYU Studies is more balanced than Robertson's and showed a better grasp of Stubbs's weak points.
Doctor Scratch wrote:Symmachus:
By chance did you have an opportunity
to look at this?
John S. Robertson wrote:As a practitioner of the comparative historical method for 40+ years, I believe I can say what Stubbs’s scholarship does and does not deserve: It does not deserve aprioristic dismissal given the extensive data he presents. It does deserve authoritative consideration because, from my point of view, I cannot find an easy way to challenge the breadth and depth of the data.
I'm not sure what kind of praise that is. Stubbs's problem is not that he lacks a significant data but that he interprets it poorly.
Aside from that vague praise, I found Robertson's review to be lacking in the critical stance you would expect from a reviewer, particularly in a venue of such high caliber like the
Interpreter. It's riddled with misstatements, some of which I assume are inadvertent, but some are significant. An example of the first kind:
Syriac is a Western dialect of Aramaic.
That is wrong, but I assume it is just a slip.
A quick example of a more serious problem is the discussion of Stubbs's correspondence between Nahuatl pronouns and Semitic (page 112, Table 3). Well, Robertson for some reason buys this connection and writes about the Semitic stuff as if those prefixes are pronouns. THEY ARE NOT PRONOUNS. If you put pronouns from any Semitic language next to the Nahuatl set of pronouns, you'll see there is no correspondence between them.
The hypothesis of Stubbs, therefore, must be that Nahuatl's pronouns are borrowed from the verbal prefixes of Semitic verbs. But think about what kind of linguistic phenomena speakers of one language usually borrow from speakers of another. Words get borrowed (Robertson mentions one case), but entire grammatical systems, usually not so much. If Stubbs is right, then some speakers of a Uto-Aztecan language had no way to distinguish between "I," "you," or "she" and the like until they met the Nephites. Pronouns, in fact, are extremely durable and usually archaic in every language, and I could show you how English "I" and Russian "Ya" along with Irish "Me" and Sanskrit "aham" as well as a host of others are all variants of same word that goes back probably to at least the third millenium BCE. Egyptian, to use a more pertinent example, is not obviously related to the Semitic languages, but the pronouns sure are an important clue of their relation. If Uto-Aztecan pronouns were derived from Semitic, you would expect a correspondence with the pronouns, not verbal prefixes.
You tend not to borrow those sorts of words, though. English third person plural pronouns (they, them, from Old Norse) are one notable exception but the reason for that is a socio-linguistic that I can go into if anyone cares to know. But notice that English speakers did not borrow their ENTIRE set of pronouns. Speakers do not borrow entire morphologies, only features and words. Yet according to Stubbs's theory, the entire set of Nahuatl personal pronouns ultimately derives from Semitic verbs. That's typologically improbable.
Did no one ever say "I love you"? Perhaps they had no concept of love until they met the Nephites (although apparently the Nephites didn't love them enough to share their DNA).
Then there's this gem:
Uniting Northwest Semitic and Egyptian with United Airlines sheds light on certain data in United Airlines that would otherwise remain obscure. Among other things, the union reveals two ancient dialects, one the “p-dialect,” which has characteristics of Hebrew/Aramaic and the other the “kw-dialect,” which is Phoenician-like.
It doesn't reveal anything; the two dialects are assumed to have existed in order to account for that fact that Stubbs's sound laws (and consequently his whole argument) don't work without it. That's using a 100% hypothetical assertion as evidence to support another 100% hypothetical argument.
And anyway, it doesn't make any sense at all. He is treating Hebrew and Aramaic as if they are just variants of the same language that are distinct from Phoenician. Actually, Hebrew and Phoenician are so close that one could argue they are dialects of the same language, and they are usually grouped together. Aramaic, by contrast, not only has significant differences in phonology (sounds) that make it a distinct language but even its morphology (its grammar) is different. In other words, it should be "Hebrew/Phoenician as opposed to Aramaic-like." The reason why Stubbs seems to group them this way is that there are a couple of minor phonological differences between early Hebrew and Phoenician that Stubbs exploits in order to derive some sound changes that he then uses to create equations between words in some Uto-Aztecan languages and words in Semitic languages.
A critical reviewer would have spotted this, but Robertson is not a critical reviewer.
I can't believe anyone takes this seriously, not because of his conclusion, which is obviously ridiculous and unsupported by material evidence, but because of his shoddy methodology. Let me just give you a basic rundown of Stubbs's method: he is trying to link up Uto-Aztecan with Afro-Asiatic (Egyptian, plus Hebrew and Aramaic, Arabic, Akkadian, and others). In the latter case, we have a massive corpus of data going back to the very beginning of the third millennium BCE up to the present day.
In the former case, the first written records don't appear until the sixteenth century, mediated initially by Spanish-speaking monks—and that's just one language in the family. Most of the others don't get treatment until the 19th century, and in a very unscientific way until people like Franz Boaz and Edward Sapir come along near the end of that century and the beginning of the next. Consequently, Uto-Aztecan, which Stubbs treats as if it were some kind of unified language, is just a linguistic construct meant to serve as a shorthand to account for the fact that X number of languages have a cluster of features in common. We don't even know if all of those features coexisted at the same point in time, and we have no way of knowing it. James Clackson has a great analogy for this: reconstructed languages are like constellations. Like stars in the sky, they appear to form a synchronic pattern that we can call a "language," but the fact is that the constellation is not real. The constellations are useful for making sense of one's place in a given locale, but they are constructs, not realities.
In sum, there was no attested language called Uto-Aztecan, and all of his data (the words) in that column are often hypothesized constructs, not actually attested words with the meanings he ascribes to them.
So, for example, the equation between the word for "daughter" in Hebrew
batt and Uto-Aztecan
patti looks promising, but the latter word is unattested and is just reconstructed. We know when and where Hebrew was spoken, but we have no idea when and where
patti meant "daughter. Was it in the 12th century AD? BC? If you don't know that, you can't know whether there was even the possibility of influence. You need history to answer that sort of question, and the history is not favorable to the theory that Stubbs presents. What Stubbs and Robertson ignore, in other words, is that "historical linguistics" means that you have to think about history too, not just linguistics, when you make historical linguistic arguments of this sort.
In making his correspondence sets, Stubbs either uses those reconstructed forms like that or just cherry-picks rather indiscriminately from Language Group A (Semitic and Egyptian) to find matches in Language Group A (Uto-Aztecan). His only criteria are 1) his hypotheses that need data for support and 2) semantics, which is notoriously unreliable in forming correspondence sets beyond a few basics (he uses, for example, the Hebrew word for "flesh" as a correspondent for an Uto-Aztecan word for "penis," but while all human penises consist of flesh, not all flesh is a penis. If you're gonna play that game, you could make "vagina" mean "penis").
I deployed the same cherry-picking method above to derive phony sound laws in order to show that English is influenced by Akkadian, which hasn't been spoken in 2,000 years, and Arabic.
Let me know when someone who isn't a personal friend of Stubbs at BYU reviews the work.
"As to any slivers of light or any particles of darkness of the past, we forget about them."
—B. Redd McConkie