Two more things:
1. Stubbs's invented dialects (a dialect where Semitic b becomes p in Uto-Aztecan, and another where it becomes kw) are typologically very odd: there are numerous examples of sound change where labio-velar consonants ("kw" without voicing or "gw" with voicing, or variants with other features like aspiration) develop into labial consonants (kw > p, with no voicing or the voiced consonants gw > b), sometimes with further developments (e.g. Greek). It's not impossible that it could go the other way (p > kw or b > gw), which is the way Stubbs's model has it work in order to make a dialect that explains exceptions to his rules, but I can't find examples (I welcome anyone's pointing them out to me). The reason it is improbable is that a labio-velar, by definition, contains two features that happened together: the point where the consonant is articulated on the soft palate (hence the velar part, the k or g part of the consonant) but with a secondary feature, rounding of the lips (hence the labial part, the w part of the consonant). What happens is that, over time, one of these features gains prominence over the other, so you get a split: either the velar feature is picked up by more speakers and the consonant becomes a velar (k, g), or the labial element gains prominence and the consonant becomes a labial (p, b). As I say, there are numerous examples of this cross-linguistically, but the usual direction from labio-velar to velar (as happened to Hittite, Sanskrit, Latin, Germanic and other branches that descended from Indo-European somewhere) or from labio-velar to labial consonant (as in Greek, Oscan). Labio-velar to velar is less drastic a shift because you don't actually change the articulation, you simply remove the lip rounding when the velar is articulated. Therefore, labio-velar to labial, which is rarer, is usually taken as an innovation that marks a different dialect or language. An example of this happens in Celtic, where Irish has maintained velars, but Welsh (as well as Gaulish) innovated by reanalyzing labio-velar consonants as labials (thus you get the word for "five" in Old Irish cóich
on the one hand vs. Gaulish pinpe
and Welsh pymp
, all descended from the same Proto-Celtic form *kwenkwe
, which Latin readers will recognize as very close to quinque
). The outcome of the labiovelar is used as a shorthand to divide Celtic into to dialectical groupings or areas: Q-Celtic (Irish, Celtiberian, where the labiovelar became velar) and P-Celtic (Welsh, Gaulish, Lepontic, where the labiovelar became a labial).
Stubbs appears to be using Celtic as a model, so that we get from him a Northwest Semitic-Kw dialect which explains the outcome of "kw" in some Uto-Aztecan words where a Semitic or Egyptian word he is using had "b," and a Semitic-P dialect, where the Uto-Aztecan word has "p" supposedly nativized from some Semitic or Egyptian word also containing a "b." In other words, b has two possible outcomes: p or kw.
I have already outlined some inconsistencies in how this model is used, but as while the direction kw > p would not be unusual, b > kw definitely would be. We have to assume the "b" was first devoiced to "p," which is one of his rules that, as I shall explain below, is quite meaningless. But granting it anyway for the moment, the direction p > kw is typologically improbable. We can understand the mechanism by which kw > P: namely, the labial element (the W sound caused by lip rounding) gains prominence, and as a result the velar element drops away. But if you start P, what feature of P would lose prominence such that a speaker would begin to reanalyze it with a velar? Kw can go to p because the labial element is already present in the kw, but there is no velar element present in "p" alone. In any case, Stubbs attempts to remedy this on pages 73-74 in various ways: 1) what his wife tells him about some Argenitine dialects of Spanish, 2) his own perception of Yaqi speakers, 3) another linguist's perception of central Mexican Spanish, unpublished, 4) appealing to the w > gw in medieval French, which is not the same kind of phenomenon in question, 5) two examples from Welsh, whose linguistic shape he doesn't fully understand, and 6) a reference to another work citing examples from Austronesian languages where w > kw.
The examples of French and Welsh, in any case, involve loanwords, so these could be more helpful to his thesis, but they're not. He wrongly calls Welsh a Q-Celtic language (as you see, it is P-Celtic), but the two loanwords from Latin that he cites where Latin v (pronounced as a w) became Welsh gw- at the beginning of a word have to do with consonant mutations in Welsh in light of the fact that such a sound in Welsh does not occur at the beginning of a word, and these words do not result from expected sound changes in Welsh (that is how we can tell they are loanwords!). Something similar occurs in his Medieval French examples: these are words borrowed from Germanic languages representing non-native sounds, and their form cannot be accounted for by the regular phonological development of French (seeing a pattern here? loanwords are nativized, as Robertson keeps insisting, but they are detectable because they are not following the expected patterns of the language).
So, for those who are still awake, here's Stubbs's problem: how to explain the improbable change of a P to a KW. He gives examples of nativized loanwords (with Welsh and Middle French) but not whole shifts that could mark a dialect, and in any case his use of that evidence doesn't show P > KW. They only show W > GW, which is phonetically not that odd, given the point of articulation of semi-consonant like W. His other examples show the same thing. I make no comment about his or his wife's ear for Spanish (maybe useful, but should actually be studied rather than simply asserted), although also no P > KW. He does then go on to do some mental gymnastics to explain it, but with no evidence at all.
So why posit this improbable sound shift? Because it explains a serious problem in his model: Semitic B doesn't always go to P.
2. About that. If you look at the summary of his laws/rules/sound changes in the introduction, you'll notice that B goes to P, D goes to T, and G goes to K. Now, these are three basic sets of a kind of consonant called a stop or a plosive: you can't continuously articulate these they way you can and F or an S or a vowel. In order to pronounce them at all, you have to "stop" their articulation.With B and P this happens at the lips (hence they are labials), with D and T this happens at the teeth or thereabouts (hence they are called dentals), and with G and K at the palatal velum (hence they are called velars...and if they get lip rounding they are labiovelars, as we saw). The only thing distinguishing each member from the other in these sets (B vs. P, D vs. T, and G vs. K) is that the 1st in each involves vibrating the vocal cords ("voicing" the consonants), and the second set doesn't. So B is the voiced version of P, D is the voiced version of T, and G is the voiced version of K. Voicing is the only thing that distinguishes one from the other. According to Stubbs, all of the Semitic and Egyptian consonants that were voiced (B, D, and G) became unvoiced in Uto-Aztecan (they all morphed into P, T, K).
Lots of languages have this voiced vs. voiceless distinction. Most Indo-European languages does; Egyptian did, all Semitic languages like Hebrew and Arabic do. Conveniently for Stubbs's theory, Proto-Uto-Aztecan did not have this distinction. The reason that it is convenient is that for any language does have this distinction, you can say that all of the voiced consonants are going to be unvoiced and you will automatically be correct.
Thus, in addition to the two sound changes that I discovered here
, showing that Latin had some influence on Uto-Aztecan (Latin m- initial remains Uto-Aztecan m- initially, and Latin -rt- becomes Uto-Aztecan -kk- between vowels), I can be absolutely certain that that at least three more can be added:
1) Latin B > Uto-Aztecan P
2) Latin D > Uto-Aztecan T
3) Latin G > Uto-Aztecan K
I really don't have to give you examples because there are no voiced consonants in Uto-Aztecan to be accounted for. But I'm feeling generous. Let us consider a Latin word everyone knows and one that was most relevant to the Nephites, the Latin word for God, Deus
. Now, I know by regular sound change, that the Uto-Aztecans nativized the D to T, just as John Robertson would have expected them to do.
So, now I have teus
But the -s is merely a morphological marker, so I can ignore it and focus on the root of the word.
So now I have teu
But of course the rest of the declension of this word, as well as the history of Latin and comparison with other Indo-European languages, confirms that the original vowel here was not a U but an O (this is appears in all other cases of the noun in the singular except the accusative and its unusual vocative).
So now I have teo
In Classical Nahautl, the best attested of all Uto-Aztecan languages, the suffix -tl
marked a noun, serving a somewhat similar function to what the S was doing at the end of the word in Deus
. I can add that on now, because the word would have been nativized not only to sounds but also to the morphology of Uto-Aztecan languages. Although Nahuatl is far removed from Proto-Uto-Aztecan, it is the earliest attested, and moreover if Stubbs can use medieval Syriac, then surely I can use the earliest attested Uto-Aztecan language. At least I'm trying to be as close to Uto-Aztecan as I can.
So, adding that nominalizing suffix, I get teōtl
. That just happens to be the Classical Nauhautl word for...God.
. This is the expected result of my sound change. Therefore, my hypothesis is confirmed. But again, unlike Stubbs's hypothesis, mine at least has historical evidence that Latin actually did have contact with Uto-Aztecan languages. Moreovoer, as i believe the Nephites certainly had a dialect of Latin that they picked up from the Sabine and Faliscan exiles who had come over with the Jaredites, I see nothing contradicting revealed scripture. Lastly, mine also makes sense in light of scripture. The Book of Mormon never talks about prickly pears and penises or buttocks or other words that would have made it much more valuable as a tool for staying awake in Sacrament meeting, but it does talk an awful lot about God. GIven what we know about Nephites, it is the kind of domain where there would have been a powerful influence. Picking up with Robertson's favorite analogy, the French influence on English, French was most influential in English in the domains of law and administration, but not so much in farming, sailing, sheep-herding, and so on. We should expect certain domains to have a higher proportion of loanwords then others, you see. The Nephites did not come to the Costa Rica to teach the Uto-Aztecans about bears, fingers, hands, penises, chins, insects, and the other domains of ordinary life where borrowing is least likely to occur, which are the areas where Stubbs finds his evidence, quite improbably. No, the Nephites came as exiles from a fallen Judea, and they brought the religion with them. That religion was the core of their culture, so much so that were willing to die and—more significantly—to kill for it even a thousand years after their patriarch Lehi had come over. This wasn't a transitory thing for them. Religion is the domain of life where we should expect loanwords from the Nephite civilization, therefore. That is why my thesis makes more sense. And, so far, I have been able to deduce at least five sound changes from Latin to Uto-Aztecan, and I haven't been doing it for even a single year, let alone thirty, and have appealed to but a single language, rather than half a dozen.
DarthJ had no idea how right he was.
"As to any slivers of light or any particles of darkness of the past, we forget about them."
—B. Redd McConkie