Brian Stubbs’s argument for extensive ancient contact between Semitic and Proto-Uto-Aztecan has received some attention recently in Mormon apologetics, but I don’t think Stubbs’s proposal is going to pan out. First, though, a few important messages.
Stubbs’s book, Exploring the Explanatory Power of Semitic and Egyptian in Uto-Aztecan, is not crackpot pseudo-science. It reflects a lifetime of work by someone with extensive training and experience. He follows established scholarly models and takes into consideration the cautions that go with linguistic reconstruction, and he is familiar with the relevant scholarship. As the author of important works on Uto-Aztecan historical linguistics, he is not an outsider to the primary field of his study.
Also, I fully support the idea of laying out one’s evidence and making a forthright argument for an idea outside the mainstream that may not find acceptance soon, if ever. Stubbs has made his work available online, and he has done quite an admirable job of making his methods and data accessible to reasonably educated readers who want to invest the effort to inspect them.
Stubbs’s hypothesis is that the Uzo-Aztecan family of American Indian languages, which today ranges from Central America to California and the western U.S., arose from an infusion of Semitic and Egyptian speakers into a much larger native community. As evidence, he provides a list of over 1500 cognates along with extensive discussion of phonology and rules of sound change, morphology, and other aspects of Semitic and Uto-Aztecan languages. He finds Semitic or Egyptian roots behind the majority of reconstructed Proto-Uto-Aztecan core vocabulary (7.8, p. 347) and throughout the Uto-Aztecan pronoun system (3, pp. 91-93), and argues that the proposed Semitic input would solve several problems in Uto-Aztecan historical linguistics (6.1-6.7, pp. 306-22).
Uto-Aztecan is not my field, so this is not a review. For positive reviews of Stubbs’s work from scholars trained in Uto-Aztecan and Mayan, see here and here. Instead, consider this post an example of the kinds of judgments people often have to make about academic work outside their narrow expertise.
With all that said, my own experience with historical linguistics tells me that Stubbs’s hypothesis won’t pan out. This is based on three broad areas of concern.
1. Stubbs’s hypothesized derivations of Uto-Aztecan vocabulary from Semitic forms are extensively documented and carefully reasoned. Many of the proposed derivations involve fossilized grammatical forms—the equivalent of linking a word meaning “hear” in one language not with English “hear,” but with hears, hearing, hearer, heard, or overhear. With a Semitic verbal morphology that includes prefixes, suffixes, pronoun object suffixes, and an extremely rich system of derived forms, the possible targets for a hypothetical etymology are very numerous. Stubbs is certainly correct that much remains unknown about the actual spoken language of ancient Israel, but drawing on Aramaic and other forms of Semitic in addition to Hebrew, as well as the hypothesized phonological changes in the derivation from Semitic to Uto-Aztecan, only increase the number of possibilities.
So, for example, Stubbs compares the Arabic root mrr to Uto-Aztecan *miya (Nr. 65, p. 84). The sound changes involved are certainly possible, and Stubbs provides examples of similar changes in the Mayan languages. But with so many possible sources to choose from, it seems that it would be impossible not to find cognates. The long list of cognates is actually a problem. I would actually find a shorter list using stricter criteria much more convincing. In particular, the numerous proposed Egyptian cognates are troubling, as they increase the suspicion that a similar list could be found for almost any language.
2. Historical linguists differ about the status of reconstructed languages. Are they merely schematic representations, a type of intellectual game, or do they represent a historical and linguistic reality? I tend to side with the historians, and I suspect Stubbs would as well. But I don’t see how a real language contact situation could result in what Stubbs proposes, where a creole arose that included a massive amount of Semitic core vocabulary (including words for “daughter,” “eye,” “neck,” and color words), pronouns, and fossilized grammatical forms, but did not include the Semitic grammatical system itself. I don’t see how you get such close contact that pronouns are borrowed, while at the same time grammatical morphemes survive in fossilized forms (7.11, p. 356). It’s as if someone decided to to call a strange yet delicious new food “heygivemebackmynachos” based on what his neighbors said when they pointed at it, while also adopting his neighbors’ words for I, you, and we.
Semantically, the Semitic and Egyptian cognates are all over the map, which is another problem, as it would suggest that the immigrant community was fully bilingual in both Egyptian and a form of Hebrew, instead of reserving one language or the other for specific purposes. If you are part of a larger native population, why adopt foreign words for common body parts, basic verbs, and pronouns? There doesn’t seem to be any logic to why the proposed cognates would be adopted. Here again, a handful of cognates for foreign imports or items of cultural significance would have been far stronger. (See, for example, the case for ancient Polynesian contact with South America, where the strongest evidence currently is not DNA, but the resemblance between words for “sweet potato” in Quechua and Polynesian languages.) The phonological processes Stubbs proposes are all quite plausible, but the resulting semantic patterns of the hypothesized mixed language do not seem convincing at first glance.
3. Stubbs makes a number of comparisons between his proposed scenario of Uto-Aztecan arising from a creolization of Semitic and an indigenous language, and, somewhat closer to my area of expertise, the history of English and other Germanic languages. English, Stubbs notes, provides an example of a language that borrowed massively from other languages. I don’t think the example of English fits well, however, as the numerous loans into English from French, Latin, Greek, and other languages left almost all the core vocabulary of English intact.
An important comparison for Stubbs is Yiddish (pp. 356, 362), which he treats as a parallel case of a Semitic language being infused with a great amount of material from unrelated languages, in this case primarily German (“Yiddish […] results from the original Hebrew-Aramaic idiom being subject to many centuries of mostly German influence…”). But this is not an accurate portrayal of Yiddish. While the history of the language is not entirely settled, no one sees it as developing among speakers of Aramaic. A widely accepted proposal is that Romance-speaking Jews moved to German-speaking areas before 1250 and adopted the local variety of German, so that Yiddish is best seen as a Germanic language with extensive vocabulary drawn from Hebrew and other languages.
I would very much like Stubbs’s hypothesis to be correct. I like my apologetics served rare, and linguistic evidence of ancient contact with Semitic in the Americas would be awesome. Stubbs’s case is worth making and considering, but I think it will not ultimately work out.
But Stubbs is not ridiculous (and, please note, this is not a place to ridicule him or the historicity of the Book of Mormon). Stubbs is a scholar who is asking others to pore over his evidence and come up with something better if they can. Someone should take him up on it.