An American Indian Language Family with Middle Eastern Loanwords:
Responding to A Recent Critique

  • Article Formats:
  • MP3 audio
  • PDF
  • MOBI
  • ePub
  • Kindle store
  • NOOK store
  • Order Print Copy

Abstract: In 2015 Brian Stubbs published a landmark book, demonstrating that Uto-Aztecan, an American Indian language family, contains a vast number of Northwest Semitic and Egyptian loanwords spoken in the first millennium bc. Unlike other similar claims — absurd, eccentric, and without substance — Stubbs’s book is a serious, linguistically based study that deserves serious consideration. In the scholarly world, any claim of Old World influence in the New World languages is met with critical, often hostile skepticism. This essay is written in response to one such criticism.


The most recent issue of the Journal of Book of Mormon Studies included a review of two books by Brian Stubbs.1 It gives me no pleasure writing this response to Chris Rogers’s review of Stubbs. I sat on Chris’s dissertation committee (University of Utah) directed by Lyle Campbell, and I wrote a positive letter of recommendation for his candidacy for the Linguistics Department at BYU. When I heard about his strong objection to Brian Stubbs’s books,2 it piqued my curiosity because I had [Page 2]written a favorable article regarding Stubbs’s 2015 book.3 I understood that Rogers’s displeasure had to do with (a) Stubbs’s shoddy scholarship and (b) the apparent impossibility of Middle Eastern languages effecting changes in Proto-Uto-Aztecan (PUA), a reconstructed American Indian ancestral language.

I immediately emailed Rogers, trying to get some clarification regarding his strong antipathetic attitude toward Stubbs’s works. In his response, Rogers mentioned that it had something to do with “joiners and splitters.”4 I was acquainted with “lumpers and splitters” (explained below), but I could not imagine how the concept applied to Stubbs’s work. Rogers was kind enough to send a preliminary copy of an article, which was since revised, now appearing in the 2019 issue of Journal of Book of Mormon Studies.

When I read the draft, I saw what Rogers’s view of lumpers and splitters meant as applied to Stubbs’s work, and given what he meant, I was deeply disappointed for the reasons outlined below. He sees Stubbs’s work as “so replete with disorganization, numerous assumptions, mistaken definitions or incorrect characterizations of linguistic concepts, inexact methods, pedantry, and apologetic rhetoric that the idea seems dubious, even without careful scrutiny” (260, emphasis added). He went on to say, “the content of the book … suffer[s] from significant analytical and methodological issues” (261).

On reading the final publication, I now understand why Rogers used such derogation in describing Stubbs’s work — he totally misunderstood the very foundation of Stubbs’s work. What Rogers mistakenly concluded led him down a rabbit hole of confusion, misunderstanding, and mischaracterization, which he is now publicly sharing with others. Ironically, Rogers lumps together Stubbs 2015 and Stubbs 2016 in his review, which should have focused solely on Stubbs 2015.

In what follows, I will accomplish three things: I will first provide some background information, then show what Rogers understood Stubbs to have said, and finally give hard evidence of what Stubbs actually said.

[Page 3]Lumpers, Splitters, and the Comparative Historical Method

Prior to examining Rogers’s critique, it is important to understand two fundamental concepts: What the Comparative Historical Method is and what the terms “lumper” and “splitter” mean.

When establishing linguistic relationships — particularly relationships between a language and any derivatives or “daughter languages” — the researcher typically relies on a methodology known as the Comparative Historical Method. According to the renowned Calvert Watkins, the Comparative Historical Method “is one of the most powerful theories about human language that has ever been proposed — and the one most consistently validated and verified over the longest period of time.”5 In applying this method, here are the steps that a successful comparatist follows:

  1. In two or more languages, the researcher scouts out as many words as he can find, which resemble each other both in sound and meaning. Such words are cognates — etymologically ‘born together’ from a given ancestral form.
  2. Based on a series of such cognates, the linguist then reconstructs the unique ancestral forms from which each daughter language descends.
  3. Starting with each reconstructed proto form, the comparatist develops a set of rules for each daughter language.

To be successful, the application of such rules to the unique, reconstructed forms must produce the same forms that are attested in each daughter language.6 It is the rules alone, as applied to the ancestral forms, that validate the reconstructed forms. Without the effective application of the rules, the reconstructed forms remain purely hypothetical. This last step is the only way of telling if the reconstructed forms truly gave rise to the appropriate forms that exist in the daughter languages.

This brings us to the concept of “lumpers” and “splitters.” These terms designate competing approaches to how languages and language families [Page 4]should be viewed, particularly those in the Americas.7 An example of a lumper is found in the work of Joseph Greenberg, who contended that before the Spanish Conquest there were three language groups that populated the Americas: Eskimo-Aleut, Na-Dené, and the remaining thousands of other American Indian languages spoken in North, Central, and South America. These thousands form a Greenbergian macro-group, called “Amerind.”8 His Amerind macro-grouping was not based on the traditional Comparative Historical Method. Greenberg’s “lumper” grouping has since been rejected by the majority of scholars. In fact, Lyle Campbell, Rogers’s dissertation chair, said that Greenberg “should be shouted down.”9 Conversely, “splitters” should presumably not be “shouted down.” They generally seek to apply the Comparative Historical Method to determine how languages evolve and diverge (or split) over time.

Rogers’s Mistaken View of the Core of Stubbs’s Comparisons

Rogers claims that Stubbs’s work catastrophically fails because, like Greenberg’s work, “Stubbs’s proposal falls into the ‘lumper’ camp'” (259). According to Rogers, by wrongly lumping Semitic, Egyptian, and Uto- Aztecan together, saying they came from some sort of proto- language, Stubbs is, like Greenberg, a misguided lumper.

Let me be more specific regarding what Rogers alleges Stubbs said:

The purpose of this piece is to review the long-distance genetic linguistic relationship between languages of the Afro-Asiatic language family and the Uto-Aztecan language family suggested in Stubbs’s Exploring the Explanatory Power of Semitic and Egyptian in Uto-Aztecan and Changes in Languages from Nephi to Now. (258, emphasis added)

Note the emphasized phrase: long-distance genetic linguistic relationship. This phrase has a special meaning when it comes to Rogers’s assessment of Stubbs’s work. Later in his critique, Rogers begins to clarify his use of the phrase:

[Page 5]A proposal for a genetic relationship between two or more languages must be supported by two types of evidence: (1) evidence that the languages discussed are in fact genetically related, and (2) evidence for the reconstruction of the common linguistic ancestor. (261)

Rogers left out an important step of the comparative historical method. Review, again, the three steps utilized under the Comparative Historical Method. Of those three steps, it is the third step — the crucial step — that Rogers left out. Only after this is done can the comparatist speak of a genetic relationship among the derived sister languages.10

The quintessential example of relations due to common ancestry would be Proto-Indo-European (PIE), the reconstructed ancestor of almost all European languages, certain languages typically found in Northern India, as well as the languages of Iran and Afghanistan. Here we consider two branches of PIE, the Italic and the Germanic branch. The Italic branch preserved the speech sounds p, t, and k, which were present probably around 6,000 years ago, in PIE times. On the other hand, the Germanic branch innovated by changing the original p to f (compare Latin pater to English father, both from PIE *pH₂tér), t to th (compare Latin trēs to English three, both from PIE *tréyes), and k to h (compare Latin cornu to English horn, both from PIE *kr̥-no).11 It is the application of many more rules like p > f, t > th, and k > h that help validate the reconstruction of ancient Indo-European spoken all those millennia ago.

However, any person who understands the Comparative Historical Method would never surreptitiously put in people’s minds that Stubbs should have used the method (1) to compare Near Eastern and the New World languages, (2) to postulate an ersatz proto-language, and (3) to devise a set of rules that would predict the cognate forms in each daughter language.

Specifically, the theory that Rogers attributes to Stubbs is that somehow, perhaps 7,000 years ago, in an unknown continent, there existed an ancestor of PUA and Proto-Afro-Asiatic (PAA), namely Proto-PUA-PAA. This ancestral language would have been an even more remote ancestor of what today we know as Uto-Aztecan (New World) and Semitic, Egyptian, Berber and other languages (Old World). According to Rogers, Stubbs 2015 has the duty of giving evidence of genetically related languages, descendants of some remote, ur-language.

[Page 6]This, however, is an unfortunate strawman; it is no wonder that Rogers assesses the content of Stubbs’s book as suffering from “disorganization, numerous assumptions, mistaken definitions or incorrect characterizations of linguistic concepts, inexact methods, pedantry, and apologetic rhetoric” (259). His mischaracterization of Stubbs’s work and the strawman he constructs makes it easy to make Stubbs look like a fool — but as we shall see, Stubbs is no fool.

Rogers further asserts that “one of the main methodological issues of Stubbs’s proposal is the omission of an explanation for why the Uto- Aztecan and Afro-Asiatic languages are being compared in the first place” (261). The only reason I can possibly imagine for Rogers’s statement is owing to his determination that Stubbs is a lumper. According to Rogers, Stubbs omits the reason for comparing Uto- Aztecan and Afro- Asiatic because he should have, but could not, reconstruct a 7,000+ year old ancestral language existing on two separate continents. So, like Greenberg, Stubbs had to “lump” them, without reconstructing an ancestor language, and without providing the rules that would correctly predict the forms of each daughter language. This makes Stubbs, like Greenberg, a lumper.

The Core of Stubbs’s Actual Comparisons

The truth, though, is that Stubbs does make a comparison — but on what grounds? Throughout Stubbs 2015 the comparisons come from borrowing and not from ancestral descent. To be clear, what Stubbs asserts is that at a given time in the past, in the environs of Uto- Aztecan, and in an intimate relationship, borrowing effectively brought Uto- Aztecan and the Near Eastern languages together. Therein lies the grounds of comparison; there was no way that he ever imagined that the relationship was genetic, despite Rogers’s strawman allegations.

To bring Stubbs’s notion of borrowing closer to home, let us consider a hypothetical scenario. Let us suppose that a Martian has accessed thousands of pages of English as well as voluminous pages of French, doing this for the purposes of careful research. His investigation reveals two classes of words. On the one hand he finds words like hand, foot, tooth, child, and sun that he cannot find in French. On the other hand, he finds words like disorganization, numerous, assumption, definition, incorrect, inexact, method, and pedantry12 that he finds in both languages. [Page 7]With no English words in French and nearly half of English words of French origin, our Martian friend rightly concludes that English borrowed from French, which makes our Martian researcher a splitter, not a lumper — a splitter because he separates French loanwords from what was originally Anglo-Saxon English!

And so it is with Stubbs, who has assiduously documented a vast reservoir — literally thousands — of Near Eastern words as well as fossilized grammatical affixes presently found in the sinews of Uto- Aztecan. This makes Stubbs a splitter by the very act of identifying borrowed words, separating Near Eastern loanwords from what was originally Uto-Aztecan. But there are no Uto-Aztecan words in the Near Eastern languages.

The Relationship between Genetic Relatedness and Borrowing

Here I must point out that borrowing does implicate the Comparative Historical Method. First, though, I must note that when one language borrows from another, there is a compulsory transformation of the loanword to accommodate the shared speech habits of the speakers of the receiving language. For example, contrast the donor language Spanish with the receiving language English in these three examples:

  • Spanish: galeón [galeon] > English galleon [gal-ee-uhn, gal-yuhn]13
  • Spanish tomato [tomato] > English tomato [tuh-mey-doh]
  • Spanish bonanza [bonansa] > English bonanza [buh-năn-zuh]

The point is that phonological rules describe the transformation of words of the donor language wherein sounds and meanings are adapted into the ingrained sound patterns of the receiving language. From the examples above, the rule would be, “In English, any vowel of the syllable next to a syllable with a stressed vowel, is reduced to ‘uh‘ as in, for example, buhnăn-zuh.”

Note that borrowing is comparable to the three rules of the Comparative Historical Method:

  1. Finding Similarities
    Genetic Relatedness: Words that are cognates in sister languages are equivalent to the ancestral words from which they descend, having similar sounds and meanings.
    Borrowing: Words in the donor language are [Page 8]equivalent to words in the receiving language, having similar sounds and meanings.
  2. Reconstructing
    Genetic Relatedness: Based on cognates, ancestral forms are postulated from which the words of the daughter languages can be derived.
    Borrowing: The words of the donor language are the words from which the receiving language can be derived.
  3. Derivational Rules
    Genetic Relatedness: If results turn out to be valid, rules applied to the words of the ancestral form must correctly predict the sounds and the meanings of the words attested in the daughter languages.
    Borrowing: If results turn out to be valid, rules applied to the words of the donor language must correctly predict the sounds and the meanings of the borrowed words of the receiving language.

A Word About Rules

Rules take the form “if x, then y.” Rules are indispensable to scientific inquiry because their application predicts repeatedly true results. If I heat water to 212° Fahrenheit, at sea level, then the water will boil, every time. The reason Watkins says that the Comparative Historical Method “is one of the most powerful theories about human language” is because the method takes a form that is largely predictive and replicable. A Spanish word borrowed into English almost always follows the rule outlined in the previous section.

Sadly, Rogers chose to ignore whether Stubbs’s rules are predictive, and this is a glaring omission of his review. Had he turned to page 2 of Stubbs 2015, he would have been able to comment on 31 instances of the rules b > p, d > t, and g > k, where there is an obvious similarity in both sound and meaning between Semitic and PUA words. Had he looked at the next page he would have been able to comment on 18 instances where Ɂ > w/o/u, where the similarities between the donor and borrower words are again transparent. Each of the following pages (4 through 9) is full of rules that show a plethora of predicted relationships of similarity between Near Eastern donor and PUA borrower. These data are included in the 1,528 well-documented instances of the same relationships that can hardly be attributed to a chance happening, to onomatopoeic similarity, to universal traits, or to genetic descent. These are all examples [Page 9]of borrowing — Uto-Aztecan borrowings from languages of the Near East, fully ensconced in the Uto-Aztecan language family.

Rogers pokes at Stubbs by paraphrasing his words: “Yet gullible may better describe those accepting the (assumptions) in the book than those digging in to find the facts” (261). Those same words ironically apply to Rogers’s critique — he obviously does not accept the assumptions of the book he reviews and fails to dig in to find facts that Stubbs provided; he ignores making reference to the rules themselves. In other words, he never interacts with the data. Recall that it is the application of the rules to the donor words which, if successful, validates the proposed relationship between the Near Eastern words found in Uto-Aztecan. More than anything, this omission deprives Rogers’s readers of a genuine understanding of the intellectual merit of Stubbs 2015.

E Pluribus Unum

One consequence of ignoring rules shows up in Table 6 (265). The table wrongly assumes that a multitude of correspondences between sound and meaning magically occur by chance without taking account of rules. Let us look more closely at the role rules play in Stubbs 2015.

A single, proper application of a rule belonging to any of the three languages would have only a 2% or 3% chance of showing a meaningful correspondence of sound and meaning. Another proper application would slightly diminish the possibility of chance. However, what happens if there is a multitude of proper applications of all the rules, each having the same consistent results? In the end, the cumulate effect of direct hits makes the appeal to happenstance so diminished as to rule out even the possibility of chance. Out of many proper applications of the rules, there can be only one, cumulative result for each rule.

Criticism of What Stubbs Did Not Do

Here, let us pick up with Rogers’s discussion on “other explanations of the similarities” (264).

Languages do not have to be genetically related to share similarities (as Stubbs correctly points out). Language similarities can be a consequence of accidents/change, borrowing, onomatopoeia and sound symbolism (or ideophones), universal traits, and genetic inheritance (or a combination of these). (264)

[Page 10]Let us consider each type that Rogers mentions:

  • Accident: Lowland Mayan has ay and Spanish has (h)ay (the initial h was never pronounced). Both mean ‘there is/ are. (H)ay comes from Latin.14 Obviously, Stubbs’s work is not founded on forms that are purely accidental.
  • Sound Symbolism: “moo” English; “moo” Hindi; “møNorwegian; “mu, muh” Uropi.15 Again, Stubbs’s work is not driven by sound symbolism. Rogers’s attempts to use sound symbolism to diminish Stubbs’s work is pointless, to say the least. Rogers says, “at least 100 of the 1,528 suggested similarities in the proposal are likely due to sound symbolism … This leaves conservatively 1,328 [sic] similarities as evidence for the proposal” (264). I do not know how Rogers came up with the round number 100, but I do know that 1,428 data points are left to be accounted for.
  • Universal Traits: All languages have speech sounds that convey meanings. Obviously, this is not the basis of Stubbs’s work.
  • Borrowing vs. Genetic Inheritance: I find it conspicuously damning that Rogers ignores borrowing as one of the explanations for seeing similarities among languages. He goes on to say, “in a proposal of genetic relatedness, these other possibilities [including borrowing] should also be considered, but are not presented in either of Stubbs’s publications” (264). Clearly, borrowing is Stubbs’s only consideration, not the “genetic relatedness” that Rogers misattributes to Stubbs’s thinking. Why should Rogers even think that Stubbs was obligated to consider any of the other reasons for similarities between Semitic and Uto-Aztecan?

Rogers clearly confuses the relationship between borrowing and genetic inheritance. I cannot begin to understand what he means by the following: “Lastly, when similarities due to borrowing are extensive, the result can be a mistaken conclusion of linguistic relatedness” (265). I am guessing that Rogers is using “linguistic relatedness” as equivalent to “genetic relatedness,” as defined above. But who, if anyone, confuses [Page 11]extensive borrowing with “linguistic relatedness?” Is Rogers saying that because similarities due to borrowing are so extensive that Stubbs confused genetic heritage with borrowing? Or was it that Rogers knew all along that Stubbs’s work was based on extensive borrowing, but the “confusion” allowed him to allege Stubbs to be a lumper? Or is this somehow political theater, a kabuki dance?

Rogers goes even deeper into the weeds. His language in the following is so opaque, so difficult to follow, that I include here a bracketed translation from my attempt to get at what he means:

The potential of borrowing resulting from a scenario of contact is not systematically considered as an explanation for the similarities presented in Stubbs’s proposal. [Translation: There might be languages other than Near Eastern languages responsible for the loanwords found in Uto-Aztecan, which Stubbs did not take into account.] That is, the similarities are not put into the context of the other languages spoken all around the Uto-Aztecan languages. [Translation: The loanwords found in Uto-Aztecan might be due to languages “spoken all around Uto-Aztecan.”] Without such a comparison, it is not possible to rule out the scenario that the Uto-Aztecan similarities to Near-Eastern languages are a result of borrowing these features from other languages or from Near-Eastern languages themselves. [Translation: Since there were other languages in the neighborhood of Uto-Aztecan, it is impossible to tell whether the loanwords in Uto-Aztecan came from Near Eastern languages or other languages contiguous with Uto-Aztecan]. (266)

First, it is perplexing that after all that Rogers said about genetic relatedness, out of nowhere he suddenly recognizes that there really are “similarities presented in Stubbs’s proposal” owing to borrowing. However, I am puzzled by the fact that Rogers could even imagine that the loanwords cited by Stubbs as present in Uto-Aztecan could have been borrowed from “other languages spoken all around the Uto-Aztecan languages.” It should go without saying that every sentient human knows that two mutually unintelligible languages are not the same, which is certainly the case regarding Near Eastern languages and languages neighboring Uto-Aztecan. It is unimaginable that Rogers would formally state that the characteristics unique to Near Eastern languages, which are clearly found in Uto-Aztecan, could possibly be the same as the characteristics inherent in the American Indian languages “spoken all around Uto-Aztecan [Page 12]languages.” This makes just as much sense as saying apples are identical to oranges in every possible way — taste, texture, smell, peelability — including all the other features that distinguish apples from oranges.

So, following Rogers, there are two possible explanations for the loanwords found in Uto-Aztecan. They either come from Near Eastern languages or from “the other languages spoken all around the Uto- Aztecan languages.” For Rogers, both explanations are on equal footing. But it is clearly impossible to know what the other languages “spoken all around Uto-Aztecan” were. Rogers even calls Stubbs to task for failing to note that “the potential of borrowing resulting from a scenario of contact [was] not systematically considered as an explanation for the similarities presented in Stubbs’s proposal” (266).

In light of Stubbs’s “failure,” let us do what Rogers omitted doing: choose between Near Eastern languages and his curiously odd alternative. Ockham’s Razor makes the choice easy: “Entia non sunt multiplicanda praeter necessitatem, that is, a hypothesis ought not to introduce complications not requisite to explain the facts.”16 Near Eastern languages are responsible for the loanwords found in Uto-Aztecan because no other language or combination of languages could have features identical to those belonging to Uto-Aztecan. Surely, Rogers’s alternative suggestion that “the other languages spoken all around the Uto-Aztecan languages” is responsible for the similarities between the two languages is a complication “not requisite to explain the facts.”

A Knotty Problem

Of course, Stubbs 2015 raises a knotty problem. It is academic dogma that any prehistoric migration from the Middle East to the Americas never happened, nor could it ever have happened. Any scholar’s work would be anathema if it made such a claim. Some say Stubbs’s work is anathema — but only at the expense of ignoring the breadth and depth of the actual data. There is actually existing evidence that favors such a migration — not an archeological artifact, nor a recorded manuscript — but evidence in the form of factual, predictive, lawful linguistic data found in Stubbs 2015. Such evidence of borrowing exists in abundance, available for proper review and criticism. And certainly, factual linguistic data should carry more weight for professional linguists than for anyone else.

[Page 13]Stubbs’s Bona Fides

Does Stubbs’s professionalism, his modus operandi as a comparatist, measure up to Comparative Historical Method standards? I believe so, and below I explain why Stubbs’s 2015 data meet every expectation of a good comparatist.

Among the many that have contributed to the study of PUA, there are three major players that stand out: Wick Miller, Kenneth Hill, and Brian Stubbs. Miller produced 700 correspondence sets (that is, reconstructions of PUA). Hill added 500 more, for a total of 1,200 correspondence sets. In 2011 Stubbs published a landmark book, Uto- Aztecan: A Comparative Vocabulary,17 which added 1,500 more, more than doubling Miller/ Hill’s 1,200 to 2,700 correspondence sets, without addressing any Near- Eastern contributions. Written for other Uto-Aztecanists, Uto-Aztecan: A Comparative Vocabulary treats Uto- Aztecan reconstructions of vocabulary and grammar of PUA. Stubbs’s careful use of the Comparative Historical Method resulted in Hill saying, “All in all, this is a monumental contribution, raising comparative Uto-Aztecan to a new level.”18 This work has become a standard, which every Uto-Aztecan comparatist must take into account.19

The methodology that all competent comparatists use is the Comparative Historical Method, which is, as already mentioned, “one of the most powerful theories about human language that has ever been proposed.”20 As a student of Watkins, as a comparatist who has studied the Mayan language family from 1967 to 2019, and as one who has 60+ refereed articles in prestigious journals, not to mention books, I feel that I might be somewhat competent to judge Stubbs’s comparative work. My take is that Stubbs’s work of both 2011 and 2015 is spot on, that Stubbs has always been an exacting practitioner of the Comparative Historical Method. In this regard, I would challenge anyone to demonstrate any significant drop off of methodology between Stubbs’s 2011 and 2015 books. Both meet every standard required by the Comparative Historical Method.21 Furthermore, [Page 14]his publications in refereed journals and his presentations at Uto-Aztecan conferences also attest to his professionalism.

In evaluating Uto-Aztecan: A Comparative Vocabulary, Hill goes on to say,

Part III (pp. 47–420) is the core of the work, the comparative vocabulary. Stubbs numbers the sets 1–2703, but in reality there are many more than 2,703 sets because many subsets are given with numbers like 7a, 7b, 7c, for vocabulary that may or may not be groupable into a single more inclusive set. Each set is discussed in some detail and the serious comparativist will delight in the discussions.22

It must be agreed on all hands that Stubbs, or anyone who understands scholarly methods and practices, has the right (if not the responsibility) to object to Rogers’s criticisms. Regarding “splitters” and “lumpers” we can happily say that Stubbs is a splitter because he carefully separates the native aspects of Uto-Aztecan from the borrowed Near Eastern features that are now ensconced, part and parcel, in the fabric of Uto-Aztecan.

Conclusion

As I said initially, I am sorry that this whole thing came up. I am sorry that Rogers abused my words, wrongly giving the impression to the quick or casual reader that I agree with his harsh disavowal of Stubbs 2015:

The result is that when evidence and methods are considered carefully, there is ample reason to “challenge the breadth and depth of the data” … (259, emphasis added)

I was saddened when I saw that Rogers had manipulated the intent of my words. These were the closing words of my favorable review of Stubbs, not conveyed by Rogers:

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 [Page 15]consideration because, from my point of view, cannot find an easy way to challenge the breadth and depth of the data.23

I am sorry that Rogers similarly manipulated Dirk Elzinga’s statement regarding Stubbs 2015. Rogers wrote:

… to remain unconvinced by the “extensive accurate data, to back up his [Stubbs’s] extraordinary claim.” (259, emphasis added).

How can extensive, accurate data constitute a reason for remaining unconvinced? This is what Elzinga actually said:

Stubbs has something the language eccentrics do not have: the training and experience, together with extensive accurate data, to back up his extraordinary claim of significant Old World linguistic influence in Uto-Aztecan, a New World language family.24

I am sorry that Rogers accused Stubbs of postulating a logically impossible theory of some contrived common ancestor of the Uto-Aztecan and the Afro-Asiatic languages, that he was so fixated on “lumpers and splitters” that it blinded him to Stubbs’s obvious claim of borrowing, which functionally makes him a splitter.

I am sorry that Rogers became so overly zealous as to state that Stubbs was someone who made false assumptions, mistook definitions, mischaracterized linguistic concepts, used inexact methods, and was pedantic and apologetic.

I am sorry that Rogers accused Stubbs of wholly inept scholarship, below academic standards, in the face of Uto-Aztecanists’ general acceptance of and acclaim for his past work.

I am sorry that Rogers was blind to Stubbs’s genuine contributions to the field of comparative Uto-Aztecan — his 2011 book (crucial to the field), his publications in scholarly journals, as well as his lifetime of presentations at Uto-Aztecan conferences.

Most of all, I am sorry that his misleading representation of Stubbs’s work may cause LDS, LDS dissenters, and non-LDS scholars not interested in the Book of Mormon to discount out-of-hand the 1,528 data points, the unyielding facts that alone give substance to the predicate of Stubbs’s claim.

1. Chris Rogers, “A Review of the Afro-Asiatic:Uto-Aztecan Proposal,” Journal of Book of Mormon Studies 28 (2019), 258–67. The books reviewed by Rogers were Brian D. Stubbs, Exploring the Explanatory Power of Semitic and Egyptian in Uto-Aztecan (Provo, UT: Grover Publications, 2015), hereafter referred to as “Stubbs 2015,” and Brian D. Stubbs, Changes in Languages from Nephi to Now (Blanding, UT: Four Corners Digital Design, 2016), hereafter referred to as “Stubbs 2016.”
2. This paper treats Rogers’s critique of Stubbs 2015. I do not address Rogers’s critique of Stubbs 2016 because that book treats his perception of his earlier works’ relevance to the Book of Mormon.
3. John S. Robertson, “Exploring Semitic and Egyptian in Uto-Aztecan Languages,” Interpreter: A Journal of Mormon Scripture 25 (2017), 103–16, https://journal.interpreterfoundation.org/exploring-semitic-and-egyptian-in-uto-aztecan-languages/.
4. In his published critique, Rogers uses “lumpers” instead of “joiners.” I will use “lumper” because this is the terminology he uses in his published version.
5. Calvert Watkins, “New Parameters in Historical Linguistics, Philology and Culture History,” Language 65/4 (1989), 783. Unknown to me until recently, Stubbs and I have a connection with the late Calvert Watkins, whose academic genealogy goes back to Ferdinand de Saussure, Antoine Meillet, and Émile Benveniste, all giants in the field of historical linguistics. Watkins was on my dissertation committee at Harvard, for which I am grateful. After retirement, Watkins went to UCLA where Brian Stubbs read one of his papers on comparative Uto-Aztecan. Watkins spoke highly of Stubbs’s lecture to Uto-Aztecan scholar Pamela Munro.
6. If the rules do not properly derive the attested forms, the reconstructed ancestor form or the rules themselves must be revised.
7. There are, in fact, lumpers and splitters in a variety of disciplines, not just linguistics. For more information, see https://en.wikipedia.org/wiki/Lumpers_and_splitters.
8. Joseph H. Greenberg, Language in the Americas (Stanford: Stanford University Press, 1987).
9. Lyle Campbell, Comment on Greenberg, Turner, and Zegura, Current Anthropology 27/5 (December 1986), 488–89.
10. French, Spanish, Portuguese, Dalmatian, etc. come from the common ancestor, Latin (or, to be more precise, “that come from the emergent dialects of Vulgar Latin”).
11. This is the celebrated Grimm’s Law, named after one of the Brothers Grimm.
12. For the record, he also finds affixes that are attached to French words, like dis-, mis-, de-, in-, -ous,tion, -ment, -able, -ain, ism, which are compelling evidence of French loans into English.
13. The “phoneticism” representing English pronunciation was borrowed from https://www.dictionary.com.
14. “From Old Spanish ha i (“it has there”) (compare Catalan hi ha and French il y a), from ha, third-person singular present of haber (“to have”), + i, enclitic form of ahí, from Latin ibī (“there”),” https://en.wiktionary.org/wiki/hay#Etymology.
16. C.P. Peirce, Collected Papers of Charles Sanders Peirce, Vol. IV (The Simplest Mathematics), edited by C. Hartshorne and P. Weiss (Cambridge, MA: Belknap Press, 1933), 4.
17. Brian D. Stubbs, Uto-Aztecan: A Comparative Vocabulary (Flower Mound, TX: Shumway, 2011).
18. Kenneth C. Hill, “Uto-Aztecan: A Comparative Vocabulary by Stubbs,” International Journal of American Linguistics 78/4 (October 2012), 591–92.
19. Hill told Stubbs (personal communication), “I must tell you Brian, whenever I think I have found a new Uto-Aztecan cognate set, I go to your Uto-Aztecan book, and nearly every time it’s already there ” (personal communication from Stubbs).
20. Calvert Watkins, “New Parameters in Historical Linguistics, Philology and Culture History,” Language 65/4 (1989), 783.
21. In some sense, Uto-Aztecan comparatists, all of whom received the 2015 book, are in a bind because, on the one hand, they accept Stubbs’s methods and practices in his 2011 book, dealing only with Uto-Aztecan, but on the other hand, they recognize the same professionalism found in the 2015 book, dealing with Middle Eastern languages and Uto-Aztecan.
22. Kenneth C. Hill, “Uto-Aztecan: A Comparative Vocabulary by Stubbs,” International Journal of American Linguistics 78/4 (October 2012), 591–92, emphasis added.
23. Robertson, “Exploring Semitic and Egyptian,” 114.
24. Dirk Elzinga, Review of “Brian D. Stubbs: Exploring the Explanatory Power of Semitic and Egyptian in Uto-Aztecan,” BYU Studies Quarterly 55/4 (2016): 176.

Posted in Review Essay and tagged , , , , on . Bookmark the permalink.
mm

About John S. Robertson

John S. Robertson is a Professor of Linguistics Emeritus, Brigham Young University. He received a BA in Political Science, an MA in Linguistics, both from BYU, and a PhD from Harvard University, Department of Linguistics. He has over sixty publications which include several books and many articles that treat the reconstruction of proto-Mayan, language change, and the grammar and sound system of the Mayan Hieroglyphs. He is also an inveterate student of the polymath C.S. Peirce, founder of American Pragmatics and Semiotics. His missionary service includes France Paris (1962-1965), England London, Family History (2011-2012), Pathway (2015 to 2017), Cody Historic Mural and Museum (May-September 2017), the latter three with his wife Barbara. Other Church service includes bishop, bishop’s counselor, branch president (Guatemala), high council, Young Mens, and Primary. He is happily married to Barbara Clyde Robertson, with twenty-two grandchildren.

62 thoughts on “An American Indian Language Family with Middle Eastern Loanwords: Responding to A Recent Critique

  1. So, basically, in reconstructing a proto-language, 1-3 apply. In determining borrowing, 1-3 apply in the same way. So how does he tell the difference? Borrowing is limited by phonology only in its possibilities, but that doesn’t predict the actual outcome: the phonology of language A sets the limits of possible realizations for loanwords from language B, but it doesn’t predict them. Without recourse to data external to the phonology of English, how would you predict the realization of the -illa/o in flotilla, guerilla, and armadillo on the one hand against tortilla and quesadilla on the other?

    Usually, violations of expected rules or anomalies indicate borrowing (in this case, the Spanish words do not contain English morphemes or roots, and again the fact that their pronunciation is not consistent despite their spelling calls for an explanation, which you cannot derive from the rules of English phonology).

    His discussion about the method is a bit like talking about the 7-step Scientific Method that you learn about in elementary school: it’s just a memorable shorthand for describing different facets of empirical analysis, but ultimately method alone is not some kind of absolute guarantor of validity whenever you apply it because there are all kinds of externalities that one encounters along the way. Same goes in linguistic reconstruction. An example: his response to Brad Anderson (who is apparently posting my comments there), wherein he imagines himself to have refuted me by pointing out that semantic drifts I am skeptical of are not typologically impossible (i.e. they occur somewhere). In his description of 1-3 that you quoted, he refers to the “similar sounds and meanings.” Well, what the hell is “similar”? “____ and “sit” are pretty similar in sounds, distinguished by one phoneme that is articulate in almost the same manner as the other; furthermore, anyone considering the mechanics of relieving oneself in a modern western lavatory sees the connection to sitting—we even use the word “seat” to refer to part of the equipment! Then there is the comparandum of “stool,” and in Polish, there is a similar semantic drift in with the word sedes.

    Obviously, method alone is not going to untangle this without some thinking about issues that are independent of method.

    His comments about Syriac show he has very little grasp of the history of Aramaic. I wish he had read closely what I wrote: “Never mind how Syriac got to Costa Rica in the 7th century BC when it had yet to develop as a dialect of Aramaic.” Middle Aramaic dialects exhibit a phonological merger not present in earlier dialects, but Stubbs’ reconstructions assume the the post-merger phonology in using Syriac and Bablyonian Aramaic cognates (incidentally, in future editions, Stubbs should remove “the” in his citation of any Late Aramaic evidence, as the alpha suffix no longer denoted definiteness). I can’t prove that the merger didn’t happen in whatever Aramaic dialect the Lehites were supposed to have brought over along with Hebrew, various stages of Egyptian, Akkadian, and Arabic, but nonetheless I don’t see on what basis I should think so, other than that it helps the theory along.

    He also didn’t actually reply to my comment but what he imagined I said. My comment was: “They are in the same semantic field in a vague way, but it’s hard to imagine how they could be borrowed with such drastic shifts of meaning.” If indeed this is a result of borrowing—and to be honest, there are conflicting statements even in the book, which I notice he does not even cite in trying to clarify what its position is—but assuming it is borrowing, I find it rather strange that the drift seems to have happened so close to the point of borrowing, so that a very marginal or questionable meaning of the Semitic is the primary one in a U-A set, which often doesn’t even contain a single instance of the usual meaning of the Semitic set. Take entry 1130, for instance. Here, the Semitic evidence (“Hebrew pԑgԑr ‘corpse’, Aramaic pagr-aa ‘body-the’; Syriac pagr-aa ‘body-the, flesh-the [sic], a carcase’ [sic]”) is quite limited in its semantic field. It’s not hard to see how you get from body to carcass, as they are essentially the same thing (though with a very important difference!). But “skin” seems not to be present in the Semitic evidence, and while the examples he brings in do show that such drift is not typologically impossible (which is not what I claimed), one does wonder why the meaning of “body, carcass” doesn’t show up in the U-A cognate set if this term was borrowed. We get a deer carcass, but that seems actually an instance of semantic drift not from “carcass, body” but from “hide, skin, fur,” as confirmed by the rest of the U-A cognate set.

    pt. 1 of 2

    • p2. 2 of 2

      His reply to “downward” is interesting, though he is responding to a different point than the one I made. All of the I-E material he brings in are verbs, but entry 580 (supporting his claim about Semitic aleph > w/o/o’) gives not only a semantic extension but a grammatical one as well: “but like other denominalizations in the change from Semitic to UA, the adverbial itself became verbalized in UA.” That is an assertion, and accepting it helps his case, but it would help more if just one meaning of the U-A set contained the adverb, otherwise it seems based on nothing more than special pleading.

      There is a pervasive asymmetry throughout the book: on the one hand the Old World set includes cognates exhibiting a small semantic range, but on the other hand, the core meaning of all or most of the U-A cognates is one that is questionably connected or at least on the very edges of the semantic field of the OW correspondence set. Given a situation in which these words were borrowed, this suggests that the OW languages spoken by one group utilized the rarest meanings of these words at the point of contact, or that the semantic drift happened very soon after borrowing. In any case, semantic drift is not actually observable: it is simply assumed. The Albanian preposition për-posh means “under” and is related to the I-E word for foot (possibly from a lengthened e-grade of the root that showed up in the locative case). But here, the drift is observable, because the root meaning of “foot” with that root is pervasive throughout I-E languages. Robertson’s citation of verbs meaning “fall” in Slavic and Indo-Iranian that may be connected to the root for “foot” is another example: we have attested meanings for “foot,” and so we can actually see a drift. That is not so with so much of Stubbs’ U-A evidence; when the Semitic words don’t match up easily with a fairly stable meaning in the U-A set, we are just asked to believe a semantic drift that is not actually observable in the evidence. Why should we?

      Yes, semantic leeway aids reconstruction, but there is a hierarchy of semantic evidence. There is a reason the meaning of an Albanian word carries less weight in reconstructing Proto-Indo-European than a word from Sanskrit. Closely related or overlapping meanings bear more weight than distantly related meanings, and yet very often the overlapping meanings don’t appear at all in the U-A set. We can accept the correspondence sets as cognate only by special pleading—perhaps he can find an instance in Hidatsa of this or that, which shows that it is not impossible, but on what basis should we assume so other than it helps the case? Given the extraordinary claims of Stubbs’ work, “it’s not impossible” is not persuasive because there is no limiting principle.

  2. Continued…

    But these are areal features, not wholesale borrowing of entire phonological systems. 2) is just absurd.

    Stubbs himself has added more particularity, as well as more confusion, in later comments. On Jeff Lindsay’s blog, I find this statement, which seems to endorse 2) on the face of it:

    Quote:
    “all the main UA pronouns are from Semitic or Egytian [sic], as is a relatively high percent of its basic vocabulary: head, eyes, nose, cheek, neck, hair, shoulder, chest, breast, waist, leg, calf, finger/toe, sun, sky, moon, rock, water, several kinds of trees / plants, man, woman, several kinds of animals and insects, etc, etc. Of course, much remains to be figured out of how it all happened, yet it’s beginning to look like, rather than a near east infusion into UA, that other things came into the Near-Eastern base that UA actually is, because both Semitic-kw terms (Mulek) and Semitic-p and Egyptian terms (Nephi) are in all branches of UA, besides the actual Semitic terms for Nephites, both masc plural and feminine pl in some UA languages.”

    Well, let us pretend for now that the confidence with which he makes such assertions is resting on something other than his imagination. Even so, you can see here that he has finally realized a serious problem that I mentioned above in this post and the other above: the core vocabulary. But there are problems with this as well. For one thing, it means that we have no way reconstructing proto-Uto-Aztecan before its affect on the the Nephite mixture of Egyptian and Semitic arrived on the scene—how would we even know that there was a language before that, because all of the rules that Stubbs’ formulates could more easily be explained as just sound change from one stage of the language (or his two dialects) to a later one. There is no reason to posit a proto-Uto-Aztecan before that, and certainly no way to detect anything about it. You see, his claim in this quote is basically one of genetic descent. If UA (as reconstructed by Stubbs) arises out of a “Near-Eastern base” of Egyptian and the two dialects of Semitic that he concocts, which themselves consist of language mixture (presumably a Hebrew base, with a heavy dose somehow of Aramaic from a later period, and some Arabic on the margins), then what language is influencing them? I don’t know what could be linguistically vaguer than “other things came into.” It seems that he must mean a kind of pre-proto-Uto-Aztecan (so, even farther back than proto-Uto-Aztecan, which is that much farther from any actual evidence). But I don’t know. “Things” can mean anything. Someone who has Stubb’s impossible-to-obtain Uto-Aztecan: A Comparative Vocabulary and the time to do so should see how many words remain after removing every instance used in his Book of Mormon work. That would tell you what is left from which to reconstruct Proto-Uto-Aztecan. Would there be anything left?

    As I said early on, genetic descent is the only way that Stubbs’ thesis can make any sense. Scenarios of borrowing or “mixed language” either rely on faulty analogies that simply don’t hold up or require us to posit absurdities. Stubbs, at least in his comments posted on Lindsay’s blog from last year, seems aware enough of these problems to suggest what is in effect genetic descent, even though he seems unwilling to admit it. The most he can claim (and the most he says in the blog comments) is that something else—”things”—effected the Hebreo-Arameo-Egyptian (+ Arabic on occasion) hybrid of the Nephites. If you don’t have record of what that is, why posit it at all? My guess is that Stubbs wants to avoid “genetic descent” for the reasons that Kishkumen eloquently summarizes (“an apologetic theory built on exceptions and the multiplication of ‘hemicycles.’ If the data don’t work, create a new rule, or even a new dialect”) and with the effect that Gadianton discerns with his usual piercing eye (“The advantage of as it stands, with methods employed nobody on their side in the know is willing to clarify, is they can retain the mantra of having something really technical that nobody is willing to investigate, and then work that angle at FAIR conferences for the next ten years”).

    Robertson, at least, should retract this piece. Ostensibly, he is defending Stubbs’ book from a review he thinks is wrong. But I note that in attempting to clarify just what Stubbs’ book says, he does not cite Stubbs’ book under review at all or inform us what it says, so he is at best clarifying what he thinks Stubbs is doing. That he couches this as a correction of Chris Rogers makes it all suspiciously personal. His shoddy reasoning, misleading statements, and espousal of a view that Stubbs’ himself doesn’t seem to hold does no service to the thesis he seeks to defend.

  3. For Collin: As I understand Stubbs’ work, he doesn’t purport to declare where PUA came from. PUA could be as old as the Middle Eastern languages from which PUA and/or its daughter languages eventually borrowed, or older. Stubbs’ proposal doesn’t inject this unnecessary question into his hypothesis. I believe critics are trying to enlarge Stubbs’ claims beyond anything originating from him, in order to make his 2015 work an expanded and more vulnerable target. But his “borrowing” hypothesis only suggests that during a certain time in history, PUA and/or the UA family of languages it spawned was or were infused with many words borrowed from Middle Eastern languages.

    • Scott,

      The enlargement is thanks to the fact that the comparative method was derived to separate the wheat, genetic relation, from the chaff of borrowing. The enlargement is the better of his alternatives. If he can better explain his theoretical underpinning to show how his methods match with borrowing (he says infusion) then the critics are listening.

  4. Based on these very constructive comments, it’s clear that Stubbs has a TREMENDOUS amount of work ahead in order for his research to be accepted by Mormons and non-Mormons alike.

    I look forward to Stubb’s responses.

  5. Professor Robertson,

    In a comment to me on his blog, Hansen said this, “The idea of explanatory power by the way is a very odd idea in this case, since he is claiming contact and not shared inheritance. You don’t expect regular sound laws in borrowed vocabulary. And the phenomena that he claims need explanation in UA, can in fact be explained in a number of other and much more plausible ways, as other UA scholars have already done.”

    This seems very different from what I had understood. Is he correct?

    • Thank you for discovering that quote, Collin, it is exactly the quote that we need to settle the nature of the claim put forth by Stubbs.

      • Do you agree that there are no sound laws governing borrowed vocabulary? Seems like there must be some scientific principle about it. Is merely chance?

        • “Do you agree that there are no sound laws governing borrowed vocabulary? ”

          To my knowledge, there are no *known* sound laws — prior to Stubbs — governing borrowing. Intuitively: borrowing scenarios are too diverse and complex to get predicted by one variable, such as sound change. In contrast, with descent, sound change correlates with one vector, speaking communities diverging geographically.

          “Is merely chance?”

          You mean, as in the case of Canadian Geese migration patterns pacing a stock market index?

          There are two questions here (1) is there a strong correlation? I am not saying ‘yes’ or ‘no’, but correlation has been disputed. One objection is over the real proximity of terms linked together, a friend of mine says:

          “There is a Semitic word for “flea” that he connects with an Uto-Aztecan root for “jackrabbit” (entry 724). At 617 he pairs an Aramaic word for “chin, beard” with a Uto-Aztecan root for “mouth.”

          (2) Granting a strong correlation, however, is there reason to buy into causality from one to the other? That is where it would really help to have a historical record of contact.

          One thought is that if he does have a new law to show how sounds predict borrowing, test the law first in known waters, and show how it works in languages where the legwork has already been done. He’d achieve world acclaim if successful just for that.

          • It does seem that Stubbs would have to do a lot of ground work to show a new principle like that. Like showing that in known cases of borrowing, similar laws apply.
            Let me ask you this: would you agree that if every single work in a given UA language corresponded to a similar sounding word in Aramaic with a similar meaning that that would be evidence of a causal relationship of some kind? If not, why not? If so, then would 99% be enough? 98%? At what point would you say that that was just chance? And in addition to the existence of similarities, what if the similarities all followed a pattern like “b to p” like baraq and pirok for that particular UA language?

          • In languages, there is a wide continuum of possible levels of borrowing, from a single word to a larger mixture, to a creole. Sound change in the mixture and creole are known, and do follow some sound shift principles. See https://www.eching.org/present/ms_dissertation.pdf . Some comments on this blog think there are these “laws” that govern historical linguistics, but there really are guidelines, not “laws” as there are many exceptions and complicating factors when it comes to evaluating historical language evolution. With Stubbs work, one must look at the possibility that there was sound shifting occurring in the original language prior to contact, and also post-mixture. To say that Stubbs need to go out and create a bunch of new historical linguistic laws relating to borrowing before presenting the data is kind of ridiculous tbh. The data presented shows a very large number of correlations with cognates, and just like any historical linguistic analysis some are better than others. Further analysis would certainly be helpful to determine if the group of correlates appear to be from a certain vocabulary group, which might help explain the relationship between the two linguistic groups (ie political dominance, trade, etc.).

          • The Relationship between Genetic Relatedness and Borrowing
            Here I must point out that borrowing *does* implicate the Comparative Historical Method. First, though, I must note that when one language borrows from another, there is a compulsory transformation of the loanword to accommodate the shared speech habits of the speakers of the receiving language. [See the Polynesian/IsiNdebele papers listed above as examples.]

            Note that borrowing is comparable to the three rules of the Comparative Historical Method:

            1. Finding Similarities Genetic Relatedness: Words that are cognates in sister languages are equivalent to the ancestral words from which they descend, having similar sounds and meanings.
            Borrowing: Words in the donor language are equivalent to words in the receiving language, having similar sounds and meanings.

            2. Reconstructing Genetic Relatedness: Based on cognates, ancestral forms are postulated from which the words of the daughter languages can be derived.
            Borrowing: The words of the donor language are the words from which the receiving language can be derived.

            3. Derivational Rules Genetic Relatedness: If results turn out to be valid, rules applied to the words of the ancestral form must correctly predict the sounds and the meanings of the words attested in the daughter languages.
            Borrowing: If results turn out to be valid, rules applied to the words of the donor language must correctly predict the sounds and the meanings of the borrowed words of the receiving language.

            If I am correct in seeing systematic rule governance between genetic descent and borrowing, it is also the case that all borrowing is not random and idiosyncratic.

        • reply to Collin @ 6:32 PM:

          “Let me ask you this: would you agree that if every single work in a given UA language corresponded to a similar sounding word in Aramaic with a similar meaning that that would be evidence of a causal relationship of some kind? ”

          Yes, genetic. Hence the confusion on both sides of the fence as to what Stubbs’ project is actually about. At that point, we would have Proto-UA + Aramaic. The problem still, is that even impressive correlations could be like Canadian Geese and Bible Codes. From a review/summary of Campbell’s text at linguistlist.org:

          ” Campbell argues that many hypotheses proposing distant genetic relationships between languages do not hold up to methodological scrutiny. The Altaic hypothesis, for example, which would group Turkic, Mongolian, and Tungusic, rests on shared features frequently occurring in unrelated languages.”

          Without other kinds of historical evidence (which are absent by default in long-distance comparisons) even tried-and-true comparative methods that show descent can go off the rails with deceiving correlations.

          But Stubbs isn’t showing descent, per the paper you are responding to, nor would that be beneficial, because to support the BoM, you need the languages to be genetically unrelated, but to show influence from contact scenarios like Nephi’s ship.

          It would go something like this: the proto-UA languages would be shown related by the sound innovations Robertson describes here in Grimm’s law etc. Many words in this family would be left without explanation, however. Those words left over, then, are candidates for borrowing, or rather, directly assumed to be borrowed by their exclusion from the sound-change rules. However, once assumed they are borrowed, then how to explain where they are borrowed from?

          Go read Chapter 3 of Campbell’s book for the myriad of messy ways in which languages are known to influence each other in contact scenarios. But suppose, in your scenario, 50% of Proto-UA words were shown related to each other, and 50% were shown by “similar sounding” words to relate to Aramic. You probably wouldn’t be showing borrowing, but accidentally showing genetic relation, and have to divide up the language families differently, or, per the quote above, showing a lucky coincidence that means nothing. Remember, we have no historical record for the UA-Aramaic connection to guide us along. It’s maddening for me to comprehend, making it a worthy challenge, but the very thing that is being touted as strong evidence for borrowing, systematic sound rules, is way too clean and really shows descent. But if he shows only 10% correlation, then absent other kinds of evidence, we shrug it off as coincidence. So how does sound correspondence really work in the world of borrowing?* I have some material Dr. Grover suggested to look into, but after that, I’m stuck.

          *An Interpreter Article with your name on it, Collin? Somebody needs to address it.

          • G.P., thanks for that last comment. Boosted my self esteem when I’m starting to think that this is above my head. I’ve had an interest in linguistics, but no training. Stubbs will have to jump in. But I’m sure he will take time to do serious thinking about this issue before addressing Hansen’s challenge.

        • For Dr. Grover.

          “To say that Stubbs need to go out and create a bunch of new historical linguistic laws relating to borrowing before presenting the data is kind of ridiculous tbh.”

          Guilty as charged. But I didn’t quite say that, I don’t think. To clarify, you said:

          “It is true that they don’t compare languages that don’t have known proximity but it doesn’t mean that they can’t.”

          By presenting the data, if his work is doing something that hasn’t been done before, then he’s already created new method, so it’s a question of explaining it, as there appears to be lots of confusion.

          “In languages, there is a wide continuum of possible levels of borrowing, from a single word to a larger mixture, to a creole. Sound change in the mixture and creole are known, and do follow some sound shift principles. ”

          That’s the only material I’ve seen to date attempting to contextualize his work. Dr. Robertson’s example in this paper settled the question of borrowing, but he didn’t address it in terms of sound correspondence.

          Now, I’ll quit spamming until I’ve had a chance to review your material, and if I can come up with something useful to move the ball forward, I’ll drop by again.

          • Can I offer a speculation? I realize that this may sound ignorant because I’m not a linguist, but I think I have an idea that might avoid both Roger’s and Hansen’s criticisms.
            As I understand it, Rogers criticizes Stubbs because he hasn’t shown a common ancestor to PUA and Semitic, so a relationship can’t be established. Robertson counters that it’s borrowing, not genetic, so they don’t need a common ancestor. Hansen counters that borrowings don’t have established rules, so the cognates are probably coincidence. He thinks that the borrowings would have had to happened at the proto-UA time period for them to be spread about in each UA language. Also, he points out that all other borrowings that we know of have known historical reasons to believe that they were borrowed. Like, (my example) Japanese borrowing from Chinese. They have a long history of dealings with each other. But what about a third way where a language is a hybrid? The example of English has been given. I think it is usually said that English is a germanic language that borrows from French. Couldn’t it be said that English is genetically descended from proto-germanic and Latin? After all, a perfectly good English sentence could be said in both a german way or a french way. Example, “A person exited a vehicle consuming poultry and after inscribing a letter onto a paper with a crayon.” Or in a german way, “The man got out of the car while eating a bird and then writing a writ with a marker.” Could UA be a hybrid; genetically related to Semitic and PUA, independently? Of course this would take a lot of research show, but it seems to me that it avoids both Roger’s and Hansen’s objections. Roger’s because the Semitic words need not be genetically related to PUA and not merely “borrowed” either. French was not merely “borrowed” into English, it was assimilated like Picard into the Borg, if you get the reference. And it avoid Hansen’s because we don’t need to show how the borrowing happened. Only that the strain can be identified like French can be identified in English.
            As a tangent: I am reminded of an experience I had on my mission in Madagascar. The Malagasy people have a perfectly good word for “to jump.” It’s “mitsimbikina.” But they always use “misotey” which they get from the French “Sauter.” They use the French but put on the prefix “mi” which is an indication that it’s a verb. I’m not sure, but I think that they do it because misotey is easier to say than mitsimbikina. A rule for the Malagasy borrowing of French is easy to establish. They merely put a “mi” in front of a french verb and, voila, it’s a Malagasy verb. Doesn’t Stubbs suggest some rules like that? Doesn’t the consistency of these rules lend explanatory power to his thesis? Assuming they are consistent. Anyway, thanks for indulging my speculations.

          • Reply to Collin below:

            Collin,

            I barely saw this. This will be my last response on this thread, mainly because following the format is really difficult once there are a number of comments. I also thank Dr. Robertson for returning to clarify (above) the relation between borrowing and the comparative method. I’m having a hard time synthesizing what he’s saying into what I’ve read, but he’s the expert and I’m not, and so my only recourse would be to ferry back and forth material from my friend who is an accomplished philologist, which means the opportunities for talking past each other soar. Perhaps more material like this from Robertson will help move understanding forward, though.

            “Couldn’t it be said that English is genetically descended from proto-germanic and Latin?”

            My friend had said a few days ago:

            “The infusion of French was an infusion of loanwords, and the grammatical and phonological structures of English were unaffected by this infusion. Ottoman Turkish and Farsi are further examples; both of these manifest a heavy Arabic element in their vocabulary but no phonological effect and no grammatical effect to speak of (the Arabic plurals of Arabic loanwords don’t affect the grammars of these languages, since they can occur only on loanwords and even then don’t occur regularly, which is the same as what English does with certain French loanwords like “attorneys general”).”

            “together with some of his replies in the comments section suggest that he thinks that speakers of proto-UA borrowed a huge part of their core, everyday vocabulary from the Egyptian-Semitic hybrid of the Nephites and adapted it to their phonological system. That at least explains why Stubbs formulates rules, which are discussed as if they were sound changes, but on this interpretation he is actually restating proto-UA phonology as he reconstructs it.”

            And for mixed languages:

            (creole etc. tossed out of court)

            “Does Stubbs’ evidence support a mixed language? No. On the other hand, it is the only scenario that makes any sense. For Nephite speakers of Semito-Egyptian of whatever dialect Stubbs constructs, to import grammar so thoroughly implies a very high degree of bilingualism…. The rules for sound change he describes tell us nothing, even if they were valid. Why? Because if speakers are bilingual enough to code-switch in a mixed-language scenario they are not bound by the phonological constraints of a single language but by both languages, and as result you cannot set predictive rules for how feature of language B will be realized in language A, since speakers have competency in both. I give a contemporary example from a native language: bilingual speakers of English and Lakota, even those whose first-language is Lakota, do not impose Lakota phonology when they code-switch into English.”

            That’s a lot to unpack. If you’re *really* curious, you can do some creative web searches and find more information like this and even participate. I can do what I can to grant safe passage. At any rate, I think you’re on the right track, maybe you can email Stubbs and see what he thinks based on some of this feedback? He seems to prefer private communication to board discussions.

  6. I scrolled through some random pages of Stubb’s material and was able to discover some interesting connections between his reconstructed forms and Latin. The Uto-Aztecan material is entirely from Stubbs; using that material within that twenty minutes, I was able to derive some rules for how Latin was adapted into Uto-Aztecan. I use his material with little modification. Obviously, this is the result of language contact at the time when Uto-Aztecan speakers invaded Italy in Book of Mormon times.

    Latin g- initial remains Uto-Aztecan m- initially

    Latin moveō (mov-) = move

    UACV-1009 *miya ‘go’: M67-197 *miya/*mi; I.Num101 *mi’a ‘go, walk’; KH.NUA; M88-mi6 ‘go’; KH/M06-mi6 *miyaC (AMR): Mn miya ‘go’; NP mia ‘go’; Sh mia ‘go’; Kw miya ‘come, go, walk, pl’; stake president mia ‘travel, journey, vi pl’; CU miyá-y ‘move away from, be far from’; Cm mia/mi’a; TSh mia/mi’a; Gb mya; Sr mi/miaa; Ktn mi; Tb miyat~iimiy ‘go’; Tb(H) miyyat ‘go, take leave’. Add WMU -mi ‘while going/moving, do s.th. while going, v’; Kw mi ‘move while V-ing’; Kw miya ‘go, walk’. [NUA: Num, Tb, Tak]

    Latin mortus (mor-), “dead”

    UACV-655a *mukki (with intervocalic -rt becoming -k- or -kk-; see below) ‘die, be sick, smitten’: Sapir; VVH86 *muuki/*muuku die; M67-126a *muk / *muki; BH.Cup *mukii? ‘a sore’; B.Tep155 *muuki; L.Son155 *muku/*muk-i; M88-mu2; KH.NUA; KH/M06-mu2: Tb muugït~’umuuk ‘die’;
    Tb mugiinat~’umugiin ‘hurt’; Tb muugut ‘spirit of a dead person’; Ls múúki-l ‘sore, boil, knot in wood’;
    Ls múúki- ‘fester, v’; Ls múú- ‘be in eclipse, of sun, moon’; Ca -múk- ‘get sick, weak, die’; Ca múk’ily ‘sore, n’; Ca múki-š ‘sick person, dead person’; Hp mooki ‘die, faint, be numb, suffer from or be afflicted by’;
    Ktn muk ‘be sick, die’; Ktn mukic ‘disease’; Ktn mukim ‘dead people’; Hp mokpï ‘corpse’; TO muuki ‘die, corpse’; Eu mukún ‘morirse [die]’; Wr mugu-ná/mugi-má ‘morir, sg’; Wr muguré ‘corpse’; Tr mukú-mea; My múúke; Yq múúke; Cr mï’ïči ‘dead person, he is dead; etc.’; Cr wamï’ï ‘se murió’; Wc mïïki ‘dead, adj/n’; CN miki ‘die, suffer from’. PUA *u > CN i, CrC ï. Sapir includes SNum terms stake president čaŋwïqqa, čaŋwïkki, čawukki (< *ca-mukki) 'die off, disappear’. It and Tak -k- (vs. -x-) suggest *-kk-, but stake president moġoa does not; thus, Ken Hill rightly separates those.
    UACV-655b *mukki 'sore': Munro.Cup121 *múúki-l 'sore'; M67-128a; KH.NUA: Ls múúki ‘ to fester, v’; Ls múúki-l 'a boil, knot in wood'; Cp múki-ly 'sore'; Cp múkilya’a-š 'sore, pl’; Ca múk’i-ly; Sr mukţ 'a sore, n'; Sr moki’ 'be getting sore, vi’. Cp muhí’i-š ‘suppurating, sore, adj’ a variant with softened medial consonant? Though the semantics vary—e.g., 'spirit' in Numic—this is one of the few etymons found in all eight branches of UA. Note Tb g < *kk rather than Tb h ( Tb g, Wr g, Tak k, not x] [Num, Hp, Tb, Tak, Tep, TrC, CrC, Azt]

    Latin -rt- becomes Uto-Aztecan -kk- between vowels

    Latin artus, a, um, adj. v. arma, prop. fitted; hence (according to Lewis and Short) close, strait, narrow, confined, short, brief; the root is from Proto-Indo-European -*h₂er-. The the zero-grade of this root produces Latin rēte (=net), with initial laryngeal loss inducing vowel lengthening in the following syllable.

    UACV-1519 *ikkaC / *iCkaC ‘carrying net’: BH.Cup *’íkat ‘carrying net’: M88-’i3 ‘net’; Munro.Cup79 ’ííka-t ‘carrying net’; KH/M06-’i3: Cp íkat ‘carrying net’; Ca ’íka-t ‘carrying net’; Ls ’ííka-t ‘carrying net’. Intervocalic -k- in all Cupan languages suggests a geminated *-kk-, and final -t shows in Tak -t vs. -l. [e1i,e2n,e3q,e4t] [NUA: Tak]

    Prove me wrong.

      • I think what Brad did in his satirical post was just to copy and paste from a couple of Stubbs’ entries and then copy a couple of items from Latin, and make up a few random one-time “rules” for his individual “examples” in an attempt to make it seem that Stubbs doing something similar. For example, the UACV-1099 entry is just copied from p. 77 of Stubbs (2015), but perhaps with an autocorrect tool strangely changing “SP” (the abbreviation for Southern Paiute) to “stake president.”

        What follows is addressed to Brad.

        Unfortunately, Brad, the 20-minute exercise seems to miss the strength of Stubbs’ work and his appropriate and detailed data from the comparative method. It’s not about conjuring up a random change to make any two words seem related, but in demonstrating the existence of plausible rules that each apply to multiple examples in a meaningful way with predictive power.

        So for the section you copied from p. 77, Stubbs was illustrating a sound change of “intervocalic -r- became -y-/-i- in non-initial positions” with examples from the proposed Semitic-kw infusion. Changing -r- to -y- or -i- is not a random rule to try to force a Semitic word into Uto-Aztecan, but a very plausible sound change that occurs in other language families. Stubbs gives the helpful example of “Proto-Mayan *r > y in most of Q’anjobalan, Tzeltalan, Cholan, and Yucatecan (Campbell 1977,
        97-100).” He shows several examples of this sound change for the Semitic-kw infusion and notes that many other examples of -r- > -y- “abound throughout” the volume. A related issue is the sound change discussed first on p. 4 of “Clusters with -r- as 2nd consonant show -Cr- > -Cy-, especially -gr-, -qr- > -ky-, or -gra / -qra > Hopi -kya.” These are reasonable sound changes, reflecting the kind of things that can happen naturally in speech, and are supported with multiple examples.

        I don’t think your satire offers anything that can be taken as a serious proposal to evaluate in the first place under the comparative method. While it may be impossible to prove that UA mukki did not have some influence from Latin’s mortus, when it comes to the comparative method, there’s nothing to evaluate. Adding “artus” = narrow, close, confined to compare with *ikkaC for “carrying net” doesn’t help much, IMO.

        Stubbs, on the other hand, gives us far more than what it took to establish newly recognized connections between other languages. There is significant meat that demands consideration, not simple satire that misses the point of what the comparative method is all about.

        • Wait, I do have to give you bonus points for the zero-grade root giving Latin rete for net. Makes it a more interesting satire, but nothing close to the level of plausibility and predictive power that occurs for each of the rule changes from Stubbs. But better than I realized at first.

        • Dr. Lindsay,

          We’ve been having a discussion below about the scope of Stubbs’ project, as there seems to be a great deal of confusion over it. Recall, the comparative method is a tool for reconstructing earlier language based on sound changes, changes that are caused by speakers becoming more remote from each other. Grimm’s law + Verner’s law creates a near deterministic picture of the proto language (indo-european).

          Recall, that the very point of this paper, was to tell Rogers that he misunderstood the scope of Stubbs’ work as an exercise in reconstructing a proto language using deterministic sound change laws.

          Now you are telling Dr. Anderson that Stubbs is doing the very thing that Rogers has been scolded for telling people.

          There are no accepted sound change laws, a “comparative method” to cover borrowing. Borrowing is an exception to rule by definition — it’s what’s left over when genetic drift fails to explain. Certainly, there are sound change rules involved with borrowing, but these don’t stand on their own as a near deterministic system, and so absent further historical data (that we don’t have). Perhaps Stubbs will revolutionize the field here. But he’s not working out borrowing scenarios via the comparative method.

          Surely, Stubbs is well versed in the comparative method as his Uto-Aztecan is his own reconstructed language. The reconstruction, can’t explain everything.

          And so in this exercise, a secondary undertaking, his work, per Dr. Robertson, is to fill the holes left over by solutions such as borrowing. Borrowing scenarios are not normally discovered by appealing to sound change only, even if sound change can be an important factor in understanding a borrowed scenario. Borrowing is not a deterministic process like sound change through drift. One is in danger of following seductive, yet false correlations, no matter how impressive, in a chaotic, under-determined system.

          This doesn’t mean his project can’t be right, or is impossible, but as Dr. Peterson has pointed out, it would be a paradigm shift that would revolutionize all of linguistics if he is right — I’m not just talking about semitic influence, but his means of demonstrating borrowing.

          But then again, perhaps I’ve misunderstood something, and I assume ya’ll won’t be shy to correct me. But I believe action item # 1 is for the core defenders of Stubbs to get on the same page as to what precisely, Stubbs is actually doing.

          • It’s not a perfect analogy, but to try and understand, conceptually, what the forces these various laws at a high level describe, consider a pool table.

            Sound change, as speakers become remote from each other, follow rules somewhat tractable like genetic drift or wind erosion. Perhaps also, to the analogy at hand, we could say that as a cue ball strikes its sisters arranged in a triangle, and they scatter from each other in somewhat an orderly fashion, that this represents sound change, under the primary force of proximity change. Now, suppose the balls bounce off the side of the table and begin to interact with each other. That would represent a more chaotic set of forces influencing language.

        • PART I

          Jeff, would have thought it self-evident that the idea that Latin influenced Uto-Aztecan is not serious, or at least not a serious proposal. The point was to illustrate one fatal problem (in addition to some others I have described on this thread and the other) with Stubbs’ method.

          Second, I think you missed the part of my entry that had Latin rete (= “net”) from that root. I’m puzzled that you think that ” adding artus” = narrow, close, confined to compare with *ikkaC for “carrying net” doesn’t help much, IMO.” I agree, except I don’t see how it’s more outlandish than most of the semantic elasticity that Stubbs relies on in his correspondences.

          For example, there is Aramaic pagrā (=body) and Hopi pïïkya (=skin, fur). There is a Semitic word for “flea” that he connects with an Uto-Aztecan root for “jackrabbit” (entry 724). At 617 he pairs an Aramaic word for “chin, beard” with a Uto-Aztecan root for “mouth.” The Hebrew adverb for “downward” he connects to a verb “to fall” ; the Syriac word for “ditch” he pairs with some UA word for “valley.” Nevermind how Syriac got to Costa Rica in the 7th century BC when it had yet to develop as a dialect of Aramaic—are ditches and valleys really that related? Is a chin a mouth? Is skin the body? Is a flea a jackrabbit (I guess if a horse can be a tapir…)? They are in the same semantic field in a vague way, but it’s hard to imagine how they could be borrowed with such drastic shifts of meaning. It would be one thing if these were abstract terms, but few people, I suspect, would look at a jackrabbit and say, “kind of reminds me of a flea” or the other way around. And there are a lot like these in the book. It’s fine to be persuaded by them, but it’s another thing to pose as if you are hard to persuade in such things.

          Third, we are talking about sound change or borrowing (his word for that is “infusion” here, I suppose)? If it is borrowing, then sound change is an irrelevant category. Since another Stubbs supporter (Robertson) keeps using the example of French and English, I will stick to that: in any of the several waves in which English borrowed significantly from French, it did not induce sound changes in English. Robertson’s repeated point (which I think he uses in misleading way to support Stubbs’ thesis, but let’s humor him) is that English speakers accommodated French words to the English phonological system then current. Pronouncing a borrowed word according to the sound system of your own language is not sound change over a period of time because by definition nothing is changing and it’s not over a period of time. Sound change is when sounds change over a period of time. For example, English used to have a fricative -ch- sound (as you sometimes here in pretentious pronunciation of “BaCH”), but it developed in different ways into other sounds or disappeared but affected neighboring sounds. That is sound change through time. It is traceable. The comparative method, is a technique for tracing sound change, among other things. It’s relevance for borrowing is largely to help determine whether something is borrowed.

          Fourth, this is a bad series of sentences:

          “He shows several examples of this sound change for the Semitic-kw infusion and notes that many other examples of -r- > -y- “abound throughout” the volume. A related issue is the sound change discussed first on p. 4 of “Clusters with -r- as 2nd consonant show -Cr- > -Cy-, especially -gr-, -qr- > -ky-, or -gra / -qra > Hopi -kya.” These are reasonable sound changes, reflecting the kind of things that can happen naturally in speech, and are supported with multiple examples.”

          I will let slide that reasonableness is not a feature in determining sound changes, but assuming we are talking about sound change, then, it doesn’t matter that they “abound.” Nor is does it matter that they are “multiple.” What matters is that exceptions should not exist. Since Robertson mentioned Grimm’s Law, I’ll use that: the change whereby what had been P in proto-Indo-European became F in Germanic languages, T became TH, and K (or C if you like) become that fricative KH (or CH) sound was total. It didn’t just “abound” and that weren’t just “multiple examples.” In the same way, the fricative “ch” sound that used to exist in English doesn’t still exist except perhaps in some wild and barbarous parts of the Hebrides where unfortunately, I hear, some people drink. It’s completely gone otherwise. Exceptions to Grimm’s Law are few but are themselves explainable by a conditioning factor that is itself TOTAL across the language family and had to do with the accent—but this second law (Verner’s Law) wasn’t simply invented to explain the exceptions (which is what Stubbs often appears to do). Doing that might yield consistency, perhaps, but it wouldn’t be based on evidence.

          To Be Continued

          • PART II

            No, Verner’s Law is based on evidence using the comparative method: what Karl Verner figured out was that the apparent exceptions to Grimm’s Law happened in parts of words whose cognates in a related language received an accent in the distant past. He compared languages.

            Now, Stubb’s laws (or is it rules? Is it sound change or a phonological system that he is describing?) are not total. For instance, 527 connects Hebrew bārāq (=”lightning”) with a reconstructed Uto-Aztecan form *pïrok in order to provide evidence for the change of Semitic b > UA p. Ok. But look at what law Lindsay just quoted at my Latin satire: “Stubbs was illustrating a sound change of “intervocalic -r- became -y-/-i- in non-initial positions” with examples from the proposed Semitic-kw infusion. Changing -r- to -y- or -i- is not a random rule to try to force a Semitic word into Uto-Aztecan, but a very plausible sound change that occurs in other language families. Stubbs gives the helpful example of “Proto-Mayan *r > y in most of Q’anjobalan, Tzeltalan, Cholan, and Yucatecan (Campbell 1977,
            97-100).”

            Why didn’t the *r in *pïrok turn into a *y, as his rule (or is it law?) would predict? Ah, see, that’s a problem. How to solve it? How to apply a “Verner’s Law”-type answer to explain the exception? Stubbs’ solution is to invent two dialects, which he calls Semitic-p and Semitic-kw, and put the supposed exception to his rule in a different dialect. What evidence is there for these dialects? None, so it’s not like Verner’s Law, which had external comparative evidence. Stubbs relies on consistency instead, but consistency can be misleading. If you invent a set of rules (or laws), and there are exceptions, by definitions the exceptions are already consistent in one fact: they don’t meet your set. If you propose or invent a second rule to explain the exceptions, it should at least explain all of them (as Verner’s Law does), and it really should have external evidence (again, as Verner’s Law does). But when you have to invent multiples layers of rules that have no evidentiary support, you can’t then go and claim that consistency simply because you’ve solved them all—they have to be solved all in the same way.

            Here’s Stubbs attempt at a rule for this invented dialect to explain the exceptions to the first rule:

            Stubbs:
            “In UA’s Sem-p, Semitic intervocalic -r- usually remains -r- in TaraCahitan (TrC) and Numic and NUA, though often represented as PUA *-t- which is pronounced -r-intervocalically”

            Usually? Often? And whether -t- is pronounced -r- intervocalically is irrelevant if you are trying to explain a phoneme, and in this case he is admitting that the underlying phoneme is a -t- (the pronunciation is an allophone; think of American English butter: the underlying phoneme represented by the -tt- is -t-, but in American English it is pronounced as a -d-. But even Americans pronounce it as -t- when speaking slowly, which tells you that the underlying phoneme in their minds is a -t-).

            He has made a rule about -r- between vowels that doesn’t hold up, so invents a new dialect where the rule doesn’t apply, but even then there are exceptions that aren’t explained. That’s not much of a rule.

            Speaking of twerking Rs, he has the following correspondence (rule? law?):

            Loss of Semitic final -r, without effect on the preceding vowel:

            Then follow a few examples.

            But note entry 1484:

            Stubbs::
            “Syriac dwr ‘to go round’; Syriac duur ‘a circle’; Aramaic(J) ‘to form a circle or enclosure’; Hebrew dwr ‘to stack in a circle’; Arabic dwr ‘turn, revolve, move in a circle, walk or go about, roam, wander about’ UA *tur ‘whirl, roll, twist’: stake president turu’ ‘whirl’; CU turú-kwi ‘roll, roll over, vt’; CU turú-’ni ‘be a whirlwind, dust-devil’; WMU turú-’ni ‘be a whirlwind, dust-devil’; Hopi tori(k-) ‘get twisted’; Hopi tori-k-na ‘twist, vt’.”

            Well, note that -r in the reconstructed UA form looks present to me. So what context is the final -r not lost? Stare at all of those words long enough and you will find a pattern. I can see one. But does that mean that I have discovered an operative rule?

            Fifth, my example was from a single language. If I were to expand this to Indo-European + one more language (as Stubbs take Semitic + Egyptian), I am sure I could easily find many more correspondences, especially if I can be very loose in semantics, if I can invent rules to suit my reconstructions without external confirmation, and if I can ignore the rules from time to time. They just “usually” and “often” have to work.

          • Of course it’s self-evident, Brad! But for the satire to be meaningful, it should do more than find one or two contrived examples. That was my point. But I did give you kudos in my follow-up comment that recognized the merit I missed initially in the find of “rete”. Yes, that made it more interesting.

          • Some of these semantic distinctions are not as unreasonable typologically as they seem to you. There are universal tendencies, even in semantics, both metanymic, and metaphoric. And don’t forget, there are inevitable and legitimate semantic shifts in loanwords well esconced in the language that are 3,500 to 4,000 years old by the time of our modern dictonaries.

            *baq2 ‘mouth’
            Attestations from S. Phil and N. Kal seem to reflect monosylbs. In Tag and northward this root means ‘chin’. S. of that it means ‘mouth’.
            Proto-Austronesian Phonology with Glossary
            Volume 2
            Book by John U. Wolff · 2018

            The IE word for ‘mouth’ persisted in several languages in the literal sense, in some others in secondary uses, but was to a large extent replaced by other words. …There is frequent association between ‘mouth’ and ‘throat’, ‘jaw’, ‘cheek’, ‘chin’ or ‘lip’.
            A Dictionary of Selected Synonyms in the Principal Indo-European Languages Carl Darling Buck

            ChSI., SCr. pasti, padati, etc., general Slavic : Skt. pad- ‘fall, go’, Av. pad-, pað- ‘move downward, plunge down’, prob. the same root as in Skt. pad-, Grk. ‘foot’. Walde-P. ibid.

            We say “fall down” which I take to be a semantic connection. Haven’t heard of falling up.

            I propose that Koromu *mete* ‘body or skin’, *wapi* ‘hands/arms’, and *ehi* ‘legs/feet’ have more than one sense, just as the English words ‘part’ and ‘body’, as in the whole body, the body

            val, a Romansh term can refer to valley, forest or a ditch
            Exploring microtoponyms through linguistic and geographic perspectives

            What in the world did you mean?? “Nevermind how Syriac got to Costa Rica in the 7th century BC when it had yet to develop as a dialect of Aramaic.” It first appeared in the first century CE. But, it’s a virtual dialect of Aramaic, Costa Rica not withstanding. So don’t discount Aramaic and Syriac as belonging to the p-dialect.

          • Oh, I found another body > skin:
            Cf. ‘skin’. Hidatsa suggests that there was a Proto-Siouan *xú•ha ‘body’, reanalyzed to Proto-Mississipi-Valley *xu• + *ha, ‘body’ + ‘covering’, through folk etymology; hence the semantic switch to ‘skin’
            https://csd.clld.org/parameters/1066#5/42.508/-98.187

            as in many Papuan languages, the same form can mean either ‘body’ or ‘skin’.
            Talking about our Bodies and their Parts in Warlpiri: Australian

            Here’s a Native American language use of the same word for “rabbit/flea,” which came from the word “deer.”
            “The word for rabbit or flea synchronically derived from the word for deer.”
            Central Coast Salish Words for Deer

            I’m not sure that some Central Coast Salish speaker looked at a jackrabbit and said, “kind of reminds me of a flea” or the other way around. Maybe a little more to it than that.
            It may be that there are “rabbit fleas” and the Salish/Aztecan figured it out, and metynimically called jackrabbits fleas. Who knows?
            Some Requirements for Mating in the *Rabbit Flea* 🙂

          • Sorry, but the comment about the jackrabbit and flea correlation being far fetched is just proof that you need to read the book better. As the book indicates both the words for jackrabbit and flea derive from the verb form of “to jump.” This is clearly explained in the book. See page 194, Semitic 724. The book can be downloaded for free at http://www.bmslr.org.

      • I think Robertson should give it a try, because it only does take 20 minutes: 1) pick a random page of Stubbs, 2) look for any kind of phonic similarity to any word in a language you know, 3) delete Stubbs’ rule and put your own in. Explain away any oddities by appealing to any known linguistic phenomena or by making other rules, and for any such new rules, simply repeat 1)-3) until you find a match. The overwhelming bulk of Stubb’s book consists of paragraphs in small print that are largely collections of Uto-Aztecan cognates, each headed usually by one word or root from a Semitic language or Egyptian, so you just have to pick something besides Semitic or Egyptian. And there is no limit on the amount of subsequent rules you can create in order to make it consistent.

        For example, there is a lot of “with loss of final -r,” a rule he also includes alongside other developments of -r (it went to y) in other environments. Not unusual for phonemes to have more than one reflex in subsequent stages of language, of course, but it is convenient. The PDF I have doesn’t have printed page numbers, so it is hard to give you reliable references, but check his paragraph 725, where he goes on about the various reflexes of “Semitic R” in several different Uto-Aztecan languages. It developed in so many different ways in UA, apparently, that he will always have an out. Almost all Semitic roots have three consonants, so if, for example, you get to subtract a final -r, then instead of having to find a match for all three columns, you only have to find a match for the first two. It’s not unusual for final -r to be dropped (British English does that, a phenomenon known as “non-rhotacism” as opposed to rhotic dialects like American English, which maintains final -r). It’s just rather easier now to find cognates if you can ignore. If on one side you are picking from Egyptian, Aramaic at any stage, Arabic, Akkadian, and Hebrew or whichever Semitic language, and on the other you have dozens of Uto-Aztecan languages to choose from in finding connections, 20 minutes might in fact be all you need.

        Keep in mind that it’s not as if phonemes (meaningfully contrastive sounds) are infinite. There’s an old debate about how many possible phonemes exist across all languages, but it’s irrelevant really, because there does seem to be an upper limit on how many phonemes a given language can have. English has about 44 or so (depending on dialect, it could be different) and that is on the high end. The highest I’ve seen is in the 70s for Old Irish argued by a great scholar 90 years ago who just didn’t know what a phoneme was, and nobody accepts that today. But Classical Nahuatl has a few more than 20. I don’t know what language among the UA family has the highest number of phonemes, but assuming it’s less than 30, then certainly someone can do the math on how hard it is to find correspondences between roots. Or how easy, rather. The trick then is to make the semantics match up, and nothing in historical-linguistic reconstruction is easier than fudging semantics, because the only real checks are 1) what you can imagine in your head and 2) what criticism you can tolerate from other people’s mouths.

        Just randomly looking at pages:

        He ignores double consonants, as convenient. At section 671, for example, to the Semitic root ђmm (bathe, wash), he connects UA huma (wash). Ok, but you’ve just reduced again one of the columns, making it easier to find your match. And the problem is that in Semitic these are not incidental. For example, if I were to refer to my father-in-law as ђamām rather than ђāma, I’m calling him a bathroom.

        I just scrolled to another page (entry 107) and see Hebrew hu and related Semitic forms (=”he” and with the definite article = “that”). This is supposed to be evidence of influence on the Uto-Aztecan pronoun system, because the UA form he has reconstructed for “that” is “hu.” So what? The Greek article for “the,” which was originally the demonstrative “that” (you can see it in Homer) was pronounced “ho.” These are all third person pronouns, and come to think of it, the English 3rd relative pronoun is pronounced “hu” though obviously spelled “who.” In that entry, he gives us the Southern Paiute form ungwa, presumably part of the evidence for his reconstruction. Now, in Lakota, ung- is part of the pronoun system, though it’s first person plural. Lakota is not Uto-Aztecan (it’s Siouan), but on the other hand until the 19th century Paiutes in Nevada were much closer to Siouan speakers than they were to Hebrews, Englishmen, or Ancient Greeks…or were they? Anyway, how can I be sure the Paiutes weren’t conquered by a roving band of Lakota speakers from whose language they massively borrowed? It’s not as if the Lakota didn’t get around and slaughter speakers of Uto-Aztecan languages.

        • Sorry. Mispelled metonymic. But why are you so angry, Brad. Is it because you think Stubbs is so terribly off base, or some other reason? It would be nice to carry on a civil conversation, don’t you think?

          • The lack of clarity about his thesis remains one of Stubbs’ central problems. The book implies genetic descent, not least of all by formulating the rules for sound change, and says as much in a quote I provided on the other thread. The same book also suggests a “mixed language” scenario, so it is hardly consistent in its argument.

            The concept of a mixed language is itself a little bit controversial, but in any case Stubbs’ discussion of mixed languages in his book is quite inaccurate. He mentions Spanglish, which is closer to what a mixed language is, but while I have indulged Stubbs’ and Robertson’s preferred point of comparison throughout this, the fact is that English is not a mixed language. Stubbs and Robertson treat Middle English (the forerunner of Modern English, obviously) as a mixture of Old English and Norman French, but that isn’t so. Perhaps they could contact Royal Skousen to learn more, but the fact is that quite a lot of the grammatical features of Middle English were already well along in the later period of Old English by the time of the Norman conquest in the late 11th century. Most Middle English texts from the 11th to the early 14th century contain very little French influence. More than Old English, to be sure, but anyone who wishes to slog through Layamon’s Brut or to delight in Ancrene Wisse will not get very far with French: if you don’t know Old English, these Middle English texts will look like they were written by a German imagining what English is. The French element in English consisted of loanwords, with infusion really coming in large doses in the late 13th century and especially the 14th century. A look at history will tell you why: the expansion of the chancery under Edward I on the one hand and the Hundred Years’ War on the other. Both of these developments fueled the spread of French throughout all levels of English society. But I emphasize: that spread didn’t effect the grammar of English and didn’t effect the morphology. The infusion of French was an infusion of loanwords, and the grammatical and phonological structures of English were unaffected by this infusion. Ottoman Turkish and Farsi are further examples; both of these manifest a heavy Arabic element in their vocabulary but no phonological effect and no grammatical effect to speak of (the Arabic plurals of Arabic loanwords don’t affect the grammars of these languages, since they can occur only on loanwords and even then don’t occur regularly, which is the same as what English does with certain French loanwords like “attorneys general”).

            They really need to retire this analogy, because it is not at all what Robertson seems to be proposing, and his use of this analogy adds more confusion about just what it is that Stubbs is proposing. Robertson uses the term “borrowing,” but what was borrowed? Assuming that Robertson has actually thought of the implications of what he has said, I can see two possibilities:

            1) Robertson’s analogy with English and French, bad as it is, together with some of his replies in the comments section suggest that he thinks that speakers of proto-UA borrowed a huge part of their core, everyday vocabulary from the Egyptian-Semitic hybrid of the Nephites and adapted it to their phonological system. That at least explains why Stubbs formulates rules, which are discussed as if they were sound changes, but on this interpretation he is actually restating proto-UA phonology as he reconstructs it. In any case, this is the position that has claimed most of the attention of this and the other thread, which discuss the problems with this view.

            2) Since Stubbs’ examples consist of core vocabulary (where are the borrowings of Nephite religious or military or political terms!?), the other possibility is that the Nephites borrowed the phonology of proto-UA, because it is extremely unlikely that a speakers of a language would borrow most of their language’s core vocabulary and do so without borrowing any of the phonology of the donor language—and in that case, they would essentially be adopting a new language. Therefore, whatever language that group spoke before adopting the new one would be beyond recovery in the absence of written records or independent testimony. In short, since 1) is not really possible Robertson on this view leaves us with a scenario in which the Nephites borrowed the phonology of proto-Uto-Aztecan in pronouncing their own language. That is unlikely in the extreme. Yes, you can find situations where close proximity to language A means that language B will be influenced. The Parisian R, for instance, is an areal phenomenon that affected the northern zone of High German, but the south still pronounces R with a trilled R, and until the mid 20th century, large parts of southern France still pronounced R with a trill, as well. Czech, almost alone among the Slavic languages, has word-initial stress, probably because its proximity to German.

            To be continued….

  7. Not everyone qualified believes that Rogers is seriously mistaken and so let’s not rescind his PhD just yet.

    http://nahuatlstudies.blogspot.com/2019/09/an-evaluation-of-nahuatl-data-in-brian.html

    “Rogers points out some serious flaws in the work: Primary of these is that it does not stick to the established bilateral method of comparing languages with languages and proto-languages with proto-languages, but that it frequently cherrypicks so that a form in any Semitic or Egyptian variety can be compared with a form in any Uto-Aztecan language. As has been demonstrated time and time again this multilateral method hugely increases the risk of mistaking chance resemblances for cognates, and makes it possible to prove virtually anything. As Rogers’ points out since any two languages can be expected to have between 1% and 3% chance cognates, if we add additional languages to the comparison the risk rises incrementally as well. This means that once Stubbs is comparing 30 Uto-Aztecan languages with at least three Semitic and Egyptian varieties (actually more, including at least Egyptian, Coptic, Syriac, Hebrew, Aramaic) the number of expected chance similarities far exceeds the 1528 proposed cognates…These general criticisms seem absolutely valid to me, now on having reviewed Stubbs’ work myself.”

    Hansen recognizes Robertson’s issue, but apparently doesn’t see
    this issue as the only issue, and doesn’t see the alternative of borrowing as a stronger position.

    “Though I can see why Rogers though that Stubbs was arguing for a genetic relation given Stubbs presentation of the evidence, I wish Rogers had realized that Stubbs’ claim was in fact a proposal of language contact. Because it really is a more problematic claim.

    It is problematic because there is no accepted method for demonstrating borrowing or contact induced changes, and consequently no method for falsifying them. Systematic sound laws do not apply in the transfer of elements of one language to another… This is why linguists normally would never even entertain the idea of a scenario of borrowing, unless there is independent evidence suggesting probable historical contact between the two languages.”

    • Brian has directly contacted Hansen with Hansen backing off most of his blog assessments. Brian will providing a response to that blog. I find it strange that Hansen asserts there is no systematic way of looking at borrowing when the literature is replete with all kinds of systematic approaches. It is true that they don’t compare languages that don’t have known proximity but it doesn’t mean that they can’t.

    • I wish Hansen had read my review of Rogers before writing his take on Stubbs 2015. What follows are some reasonable responses, I believe. When I use quotes, I am referring to http://nahuatlstudies.blogspot.com/2019/09/an-evaluation-of-nahuatl-data-in-brian.html, unless otherwise stated.

      •Before proceeding, I need to respond to this:
      “Today such a belief of visits from the Ancient Near East to Mesoamerica is not common among Christian denominations, but it is found today among the Latter Day Saints (also known as Mormons), whose sacred book, the Book of Mormon tells that four Hebrew tribes made it to the Americas. Where their prophets wrote the original Book of Mormon on metal plates in a language named “Reformed Egyptian.” LDS scholars have, over the years, invested much time and energy in trying to find external evidence in support of the account given in the Book of Mormon, both through archaeology and linguistics (Hansen).”

      Hansen continues:
      “Usually, I would follow Jay S. Gould in considering scientific inquiry and religious confession to be non-overlapping magisteria [domains], and that as long as scholars keep their religious beliefs out of their scientific inquiries then they can believe whatever they want. But sometimes this is not so easy, and this blogpost is about one of those times (Hansen).”

      In response, I would say that Stubbs 2015 stands on its own, LDS interests or non LDS interests notwithstanding. In themselves the data do not prove the authenticity of the Book of Mormon. 2015 only purports to give evidence that “at a given time in the past, in the environs of Uto- Aztecan, and in an intimate relationship, borrowing effectively brought Uto- Aztecan and the Near Eastern languages together [Robertson p. 7].” In short, the data must be reviewed in their own terms, not in terms of whether “scientific inquiry and religious confession” have been merged. The question is not about the Book of Mormon, but is about the validity of the data. Of course, my reviews of Stubbs maintain that Stubbs 2015 does yield authentic results, and I think with good reason.

      °I must also respond to this:
      “[T]here is no accepted method for demonstrating borrowing or contact induced changes, and consequently no method for falsifying them. Systematic sound laws do not apply in the transfer of elements of one language to another, a language may borrow many words or few and change them in fairly random ways as they are adapted to the borrowing language’s phonology, and there really is no good way to disprove a claim about a form in one prehistoric language being borrowed from another.”

      Frankly, there is too much evidence against this statement for it to be accurate. In his seminal article on borrowing, my old teacher, Einar Haugen said, “Only a complete analysis of the sound system and the sequences in which sounds appear could give us grounds for predicting which sounds a speaker would be likely to substitute in each given case.” Citing Polivanov (my translation from French): “On hearing an unknown foreign word, we attempt to call up our means of expressing those sounds from our language, in compliance with our sound laws. (Haugen Language Vol. 26, No. 2 (Apr. – Jun., 1950), pp. 210-231.) Consider “Patterns In The Avoidance Of Marked Segmental Configurations In Japanese Loanword Phonology” http://faculty.human.mie-u.ac.jp/~glow_mie/IX_Proceedings_Oral/16Sano.pdf
      Or the wiki, which goes into extraordinary detail regarding the effective rules of borrowing: https://en.wikipedia.org/wiki/Japanese_phonology. If you look for examples of rule-based borrowing on the web, it’s like picking blackberries in the middle of summer. Stubbs claim cannot be “methodologically falsified” on grounds that there are no rules when and entire speaking community receives loanwords from a donor language.

      I can’t help but think of English and French. When English borrowed the high front rounded vowel ü from French, the speech habits in English could not accommodate that sound. So the universe of English speakers did the best they could: They substituted yu, a high front glide (y) and a high back rounded vowel (u) for French ü, as pür becomes English pyur ‘pure.’ And that precipitated palatalization before s and t, as well as subequent and vowel reduction, French créature [kreatür] became English ˈkrētyər, which then became ˈkrēCHər. French sûr [sür] became English syur, which was then palatalized to SHo͝or. And on and on.

      There are many studies that deal with rules of borrowing. Changes are not random, as Hansen claims, but largely rule governed (See me p. 8). Words from donor languages must undergo rule-like transformations as they accommodate the speech habits of the receiving language. And Stubbs’s UA data show this in spades.
      (More below)

      • Hansen following Rogers:
        °“As has been demonstrated time and time again this multilateral method hugely increases the risk of mistaking chance resemblances for cognates, and makes it possible to prove virtually anything. As Rogers’ points out since any two languages can be expected to have between 1% and 3% chance cognates, if we add additional languages to the comparison the risk rises incrementally as well.”

        This statement is based on the mistaken belief that what Stubbs has found amounts to pure happenstance and chance. This is not true. What Stubbs has found is rule governed behavior, which obviates Hansen’s claim that chance “makes it possible to prove virtually anything.” These very rules belie Hansen’s implied claim that everything must be thrown out because there are no rules. [See Robertson, p. 9, E Pluribus Unum].

        If anyone wishes to see the efficacy of rules, download a pdf of Stubbs 15 (http://bmslr.org/exploring-the-explanatory-power-of-semitic-and-egyptian-in-uto-aztecan/ ) and try this experiment: Do a search for the Semitic voiced pharyngeal fricative [ʕ], and then look at the corresponding UA word and see if rounding [o, u, w] is not involved. Do a search for Hebrew “hard” (dageshed) [b] (word initial [b] or internal [bb]), and see if UA does not displace these sounds with [kw], but [p] after a vowel. Elsewhere non-dageshed [b] > [p], and [d] > [t], and [g] > [k]. Do a search for an initial [r-] and you will see that [t-] displaces [r-] in UA. These rules and others are particular for Hebrew. Stubbs has also found rules particular to Aramaic. These and all other rules that produce expected results cannot be a function of chance!

        °As for Hansen’s analysis of the data, he gives 14 examples out of 500+ instances of comparing Proto-Uto-Aztecan (PUA) to a Near Eastern loanword — and I have seen some good evidence to the contrary from Stubbs, who is in friendly communication with Hansen. And these discussions are grounded (finally) on DATA, unlike other criticisms I have seen (LDS scholars cannot be taken seriously). But even IF 14 examples happen to show reasonable grounds for their exclusion, this is thin gruel in the face of the breadth and depth of the data.

        If there is no general means of dismissing Stubbs 2015 (e.g.,happenstance, no possibility of Middle Easterners to the New World), there is no chance of anyone being able to dismiss the entirety of the rule-based governance found in Stubbs 2015. Stubbs’s superior scholarship of 2011 gives the lie to that.

      • Dr. Robertson,

        Talking passed each other doesn’t seem to be a communication problem that linguists have any special talent for avoiding. 😉 As I sample the commentary on Stubbs’ work, I have difficulty believing anyone is on the same page, including those who are on the same side of the table. It’s humorous to a degree, and so my mood hasn’t destabilized over it, but I do think a great place to start is getting on the same page, most importantly, concerning the “laws” that govern “borrowing”. Dr. Grover says, “the literature is replete with all kinds of systematic approaches.” Could we get an example that illustrates what is meant here?

        Now what *you* mean, is very clear, in this example, “On hearing an unknown foreign word, we attempt to call up our means of expressing those sounds from our language, in compliance with our sound laws.”

        However, I do not believe this has anything to do with what Hansen meant. I would like to clarify what I believe Hansen meant, with two examples.

        First() Your example brings to mind instruction from Lyle Campbell:

        “…older loans reflect sound substitutions before intimate contact brought new sounds and patterns into the borrowing language, while more recent borrowings may exhibit the newer segments or patters acquired after more intensive contact.”

        – HL, p. 68

        I think an implication of Hansen’s observation is that if all you have is two dictionaries, and no historical information detailing interaction, and you see a possible sound substitution, you presume the case where contact is early without justification. There are no *systematic sound laws* to establish, in this example, option (a) or (b). But certainly, there are “rules for borrowing, which include sound laws”.

        Second() Directly to Hansen’s point, consider this statement from Campbell:

        “For the Neogrammarians, sound change was considered regular, borrowings needed to be identified, and analogy was, in effect, everything else that was left over. That is, almost everything that was not sound change or borrowing was analogy.”

        p. 103

        “Sound changes resulting in multiple outcomes in a particular phonetic context compromised the cornerstone of the neogrammarian framework, the regularity principle, by forcing Neogrammarians to posit borrowing and analogy for unexpected developments”

        Janice M. Aski HL 1999 p. 31

        Campbell again – his criteria number one for identifying loanwords (borrowed):

        “Words which violate the typical phonological patterns (canonical forms, morpheme structure, syllable structure, phonotactics) of a language are likely to be loans.”

        p. 68

        My apologies for redundancy, but I think of the lurkers, and multiple contexts help reinforce an idea.

        Hansen actually said, “Systematic sound laws do not apply in the transfer of elements of one language to the other”

        To draw an analogy to medicine, borrowing (and analogy) is an example of a diagnosis of exclusion. Chronic fatigue syndrome presents a regular set of symptoms, but these symptoms correlate with other illnesses discoverable directly by testing. Once these tests come back negative, CFS is on the table. Just as CFS includes real symptoms, real sound laws are at work in borrowing, however, they do not exhaustively establish borrowing.

        More later, as I consider the rest of what you’ve written, but let’s see if we can get aligned on the easy stuff, for now.

        • Chapter 13 of Lyle Campbell’s Historical Linguistics describes certain criteria to identify borrowings and differentiate them. Historical linguistic analysis of creole and pidgin also have utilized systematic approaches to the borrowing integration of borrowing. Brian Stubbs analysis is not showing a creole per se, but many elements of the approach are similar.

        • Response to Gordon P. Richards
          Before responding directly to your kind and interesting comments, I can’t help but cite what I take to be an observation foundational to all scientific inquiry, including the well-established procedures of the comparative method.

          The polymath C.S. Peirce said,
          “nothing can justify a theory except its explaining observed facts,…which is the sole legitimate function of a theory.” “It is a poor kind of theory which … merely supposes the facts to be inexplicable. It is one of the peculiarities of nominalism that it is continually supposing things to be absolutely inexplicable. That blocks the road of inquiry “(CP 1.170). Incidentally, Peirce’s observation reminds me of Chomsky’s LAD, language acquisition device, which was often referred to as a “black box,” as I recall. I have never been fond of mechanistic theories of language which ignore the very purpose of language — to communicate interpretable information.

          I cite Peirce because I have to be on the lookout for “theories” regarding Stubbs 2015 which do indeed “block the road to inquiry” — block the consideration of the actual data found in 2015. The purpose of such such suppositions consist in avoiding the actual data. They are numerous, but a few might be mentioned (in no particular order):
          •not looking at the data because Stubbs is a poor scholar for spurious reasons (read Rogers);
          •not looking at the data because saying my review of Stubbs’s work is to be avoided because I am LDS (see the trolling OK’s claim in Mormanity);
          •not looking at the data because making the mistake of mixing “scientific inquiry and religious confession;
          •not looking at the data because Old world languages cannot be found in the languages native to the New World.
          •not looking at the data by cherry picking and ignoring the overall data.
          •…

          Sorry for the topheavy preface; it does have bearing on what follows.

          But now on to your comments. Please understand that I am not trying to put words in your mouth; I am only putting words in my mouth for the purpose of understanding what you are saying.
          Your citations from Campbell and Aski regarding the Neogrammarians assumes possible confusions between the regularities that result from common descent against irregularities caused by borrowing, analogies as well as whatever Aski means by “This model … can account for regular, neogrammarian-type changes, lexically diffused changes (borrowing), and *all the exceptions which cannot be characterized as a product of either of these models.* These comments are given in the context of descent from a proto-language.

          Stubbs is *not* claiming that the sound changes he is referring to are a product of common descent (Rogers’s misunderstanding), but is a product of borrowing: x → y: language x’s term is received by language y via particular rule. So does borrowing apply here? I think so because borrowing is borrowing — but common descent has no place in Stubbs 15. I have a hard time imagining analogy having any part of the actual process of borrowing, but even if we allow for it, the likely effect would be miniscule compared to the process of rule-based borrowing.

          As regards your take on Hansen’s comment, it would be well to compare the two.

          Hansen’s comment:
          It is problematic because there is no accepted method for demonstrating borrowing or contact induced changes, and consequently no method for falsifying them. Systematic sound laws do not apply in the transfer of elements of one language to another, a language may borrow many words or few and change them in fairly random ways as they are adapted to the borrowing language’s phonology, and there really is no good way to disprove a claim about a form in one prehistoric language being borrowed from another. This is why linguists normally would never even entertain the idea of a scenario of borrowing, unless there is independent evidence suggesting probable historical contact between the two languages. In this case, there is exactly zero independent evidence of contact between Ancient Semites or Egyptians and Uto-Aztecans…except for the Book of Mormon.

          Your comment:
          I think the implication of Hansen’s observation is that if all you have is two dictionaries, and no historical information detailing interaction, and you see a possible sound substitution, you presume the case where contact is early without justification. There are no *systematic sound laws* to establish, in this example, option (a) or (b). But certainly, there are “rules for borrowing, which include sound laws”.

          Correct me if my understanding of your claim does not square with what you are thinking:

          • Because you cite Campbell, “…older loans reflect sound substitutions before intimate contact brought new sounds and patterns into the borrowing language, while more recent borrowings may exhibit the newer segments or patterns acquired after more intensive contact,” you are thinking that
            because there is “no historical information detailing interaction,” and
            because initial loanwords are more apt to show more of the native sound patterns, and
            because later borrowing is more proximate to sound patterns of the donor language, therefore
            I have no way of being sure that Stubbs’s proffered loan words are of the earlier type of borrowing
            because they do not closely match the sound patterns of the donor language.

            However, then you say, “[b]ut certainly, there are “rules for borrowing, which include sound laws,” meaning (I can only guess) that if the “rules for borrowing” stand up under scrutiny, we can never be sure the borrowing is from the inital stage because there’s no independent attested historical record.

            And therefore, by the unstated implication, your takeaway of Hansen’s statement, gives license to Hansen’s assertion that Stubbs 2015 is a “a claim of ancient contact and language mixture, [and therefore,] Stubbs is in fact making a claim that cannot be method[ologi]cally falsified.”

            *If* I have even approximately understood your argument, I have to think (forgive me) that your hypothesis is an artifice, however explicit or indefinite, that evades even looking at the content of Stubbs 2015. I have to ask, (1) if what Stubbs has captured is an earlier state of borrowing or a later state, (2) if it really matters. It really doesn’t matter. I return to Peirce: [N]othing can justify a theory except its explaining *observed facts,*…which is the sole legitimate function of a theory.” The facts to be explained are in the actual *data* of Stubbs 2015. Those should be the observed facts. “It is a poor kind of theory which … merely supposes the facts to be inexplicable,” or to devise a theory that evades the task of looking at the facts. Such behaviour blocks the road to inquiry.”

            Finally, Magnus Hansen really did say, “Systematic sound laws do not apply in the transfer of elements of one language to another,” and for me, it seems difficult to read into it your takeaway on what Hansen really meant.

            Anyway, I hope that we are not “talking past each other, [which] doesn’t seem to be a communication problem that linguists have any special talent for avoiding. 😉 But if we are talking past each other, I hope we can do it in a spirit of *trying* to understand each other.

          • Dr. Robertson,

            This is a reply to both of your comments.

            I think we are talking passed each other to a degree, but I take full responsibility and will try to correct that. I do think we are on the same page — to a degree — also.

            I fully accept your position that Stubbs is not speaking of “descent” but borrowing.

            I also fully admit I am a skeptic. However, I do understand your concern of (1) “artificial” arguments construed to dismiss research on theoretical grounds, in order to avoid looking at the evidence. However, it is also true that (2) it is difficult to evaluate data without first having the proper theoretical context.

            I’ll grant that (1) is at play in these discussions, but (2) is also a problem. Your colleague Dr. Lindsay is defending Stubbs under the same (alleged) confusion as Rogers — that Stubbs work (under review) is an exercise in the comparative method. A friend of mine, a philologist, also has stated “it sure looks like” Stubbs is showing descent.

            My understanding is that the comparative method is performed exclusively by analyzing sound change.

            Stubbs appears to primarily, at least, analyze sound changes. Or do I misunderstand that?

            Now, a square is a rectangle, but a rectangle isn’t a square, I understand that. However, given borrowing is not normally associated with the (near?) exclusive application of sound change rules, it’s easy to (1) understand why folks are getting confused over what he’s doing, it “looks like” descent (2) are skeptical not just because he’s making a radical claim involving religion, or a radical claim involving Israelite contact, but he’s using a radical method to demonstrate his claims.

            Or am I wrong about the novelty of his method? Your colleague, Dr. Grover, has suggested I might find a way forward to understanding Stubbs in chapter 13 of Campbell’s book. I hope to have time to return to that chapter this evening. By the way, do you agree that chapter 13 will help me?

            For some, perhaps finding a theoretical objection to show borrowing by the methods he employs is a way to end the discussion before looking at the evidence, but others may be struggling because they have no theoretical framework for analyzing cases of borrowing in the manner of Stubbs.

            The way I see it then, the supporters of Stubbs need to greatly clarify either one of two things (1) If Stubbs is using tried-and-true methods to show borrowing, then you need an entire Interpreter article introducing that material — for instance, if Chapter 13 of Campbell’s book is it, or if there is another good text out there that explains it, introduce his work by a thorough exposition of that material (2) If Stubbs is using a groundbreaking set of tools for identifying borrowing, then be very clear that indeed, his approach is novel, so that the audience is less likely to confuse his procedures with the nearest thing associated with those procedures.

            I think you “sort of” made an attempt at this, but perhaps what’s due, is a paper taking his best example, spelling it out in plain terms, and contextualizing it as best as possible within known historical linguistic methods, and/or where the updates are to contemporary methods.

  8. Dr. Robinson,

    Thank you for the review. I found it very interesting and informative.

    When I first heard of Dr. Stubbs’ work a few years ago, I also found the possible Uto-Aztecan/Semitic links very intriguing. I am not an academic, (certainly not a linguist), but I noticed something that to my very untrained eye seemed like another possible connection.

    Could the Egyptian crocodile god “Sobek” be related to the Mayan/Popol Vuh crocodile demon “Zipacna”?

    I apologize for wasting your time, especially if this is wrong thinking, or if someone else noticed it, but Dr. Daniel Peterson in one of his excellent talks made mention of the Egyptian deity “Sobek”. In the Book of Abraham, in Facsimile 1, there is a crocodile below the Lion Couch, given the designation Figure 9.

    https://www.churchofjesuschrist.org/study/scriptures/pgp/abr/fac-1?lang=eng

    Joseph Smith interpreted the crocodile as “The idolatrous god of Pharaoh”, which is actually a pretty good description of the Egyptian deity “Sobek” and his role in Egyptian theology.

    While learning about the Popol Vuh, I heard about the crocodile god/demon “Zipacna”. To my untrained ear, the consonants of the two names sound pretty close to each other. Has someone else pointed this out, or is this just a bad analysis on my part?

  9. Rogers’ claim that hundreds, nay thousands, of false cognate words, with both similar pronunciation AND meanings, can be expected as a matter of random chance between two wholly unrelated languages, is an extraordinary claim requiring extraordinary evidence to be taken seriously. Before the realization that the languages of Europe and Iran and northern India had a common ancestry, some observers might have thought those similarities were random in nature, but it was later acknowledged that they were evidence of a relationship that had not been recognized before. To validate a formula for deriving the number of cognates that would appear between two utterly unrelated languages, one needs to do detailed and thorough comparative studies of pairs of languages that definitely do not have common ancestry or extensive borrowing from one to the other.

    Good candidates for a test of the multiple random cognates hypotgesis would be English and Japanese. While English has a broad family tree, Japanese, apart from specific borrowings from Chinese circa 600 AD, and from English and German starting in the Meiji Era and accelerating during the Occupation, has no apparent cognates with Korean, let alone other East Asian languages.

    The two languages I speak are English and Japanese. When the modern loan words are subtracted from Japanese, there are only a handful of words with both similar sounds and meanings in the two languages. The structure of words in Japanese is very distinct from English, with each word constructed of syllables combining a consonant followed by a vowel. It is not obvious that any word in English of more than one syllable could have a cognate in Japanese. A random arrangement of letters would not be likely to even constitute a Japanese word, and if it did, it would almost by definition be disqualified from sounding like an English word. And in my own experience, there are not a hundred native cognates, let alone a thousand. Establishing over a thousand cognates would be such a gross departure from random chance that it is clear evidence for some historical relationship.

  10. Roger’s critique was anticipated given the antipathy held by academia, of any hint of Old World influence in Pre-Colombian civilization. Stubb’s work is a direct threat to the secular worldview and therefore to Linguistics or Genetics
    or whatever other profession or worldview that rejects any suggestion that the Book of Mormon could be true. I therefore understand Roger’s reaction. HIs years of study, his profession, his “worldview” is threatened by Stubb’s work.Therefore, it must be attacked! It cannot be allowed to stand. I feel for Roger’s, I understand his reaction. Thank you for your excellent review! A most needed response and understandably painful undertaking.

  11. This is a very cordial response to what I can only characterize as a jealous academic hatchet job. Probably the best response has been provided by Brian Stubbs, which is found on the website http://www.bmslr.org (near the bottom). The original book be Brian is also available there for free download.

  12. I strongly urge those who have read Professor Robertson’s review of what seems like a very deeply flawed treatment by Chris Rogers of the important academic work by Stubbs that appeared in the recent issue of Journal of Book of Mormon Studies to also read carefully the review of Stubbs work by Professor Robertson, which appeared in Interpreter, as well as the review by Dirk Elzinga that was published in the BYU Studies in 2016.

    I also wonder if the editor (or one of those involved in editing) the Journal of Book of Mormon Studies sought, urged or encouraged Chris Rogers to write that review of the work of Stubbs in what was published by the University of Illinois for the Maxwell Institute. If that publication is still under the editorial control of the Maxwell Institute, they should be held fully responsible for publishing the clearly deeply flawed essay by Rogers.

    By encouraging or allowing this to happen, it seems that those currently in charge of the Maxwell Institute, as well as the editor(s) of a journal over which they have supervision and control, have not taken seriously the scolding and pleading by Elder Holland last November about the direction being taken by the institution that carries the name of Elder Neal Maxwell.

    I think that it is urgent for Chris Rogers to clarify how he came to have his name on the review to which Professor Robertson has responded. However, I feel sorry for Chris Rogers, and especially if he was urged to write that review.

    I would also like to see some official clarification on whether the Maxwell Institute has fully severed its relationship with the Mormon Studies Review, rather than merely ceasing to publish it, while still retaining editorial control over the Journal of Book of Mormon Studies. Perhaps they have fully yielded ownership and hence editorial control of both these publications to the whims of those at the University of Illinois who are now in the business of pumping out academic journals. If they have ceased having control over both, then they cannot be held responsible for publishing the review by Chris Rogers. But they can be held responsible for not just ceasing to publish those two undertakings, both of which were begun by those who sought as well as they could to make them a means for both advancing and defending the faith of the Saints.

  13. Nice work, John. I enjoyed your restraint and objectivity in raking your former student over the coals. But, really? You gave that guy a PhD?

    • Hi Robert,
      I should probably clarify. Chris Rogers was never a student of mine. I taught at BYU, and was asked to sit on his dissertation committee at the University of Utah. The chair was Lyle Campbell whom I genuinely admire. (If you looked at his Wicki you would see why.) Rogers’s dissertation was quality, and he certainly deserved a Ph.D. As I said earlier, his subsequent scholarship is commendable.

      However, I think his review in the JBMS is an anomaly. It seemed to me that it was overly harsh and mean spirited. But that is hardly the reason I wrote my reply. I wrote it because Rogers’ review deprives readers of a genuine understanding of the intellectual merit of Stubbs 2015. It is a landmark book. It deserves better. You can see why I said I found no joy in writing my response.

  14. Excellent, carefully-reasoned, informative article. I hope everyone who’s read Rogers’ critique will care enough about the subject to read this response and the one previously issued by Stubbs himself. But I fear that won’t happen. Those who search for arguments to discredit the Book of Mormon are usually content to find some scholar who aids their cause, but not interested enough to follow up their search by reading objective responses to, and analysis of, the arguments they’ve accepted, even if it comes from those like Stubbs and Robertson, who know more about the subject than the author of the critique. It’s also disappointing that Rogers’ critique appeared in the Journal of Book of Mormon Studies, of all places, and this isn’t the first time I’ve felt this same particular frustation. Meticulous editing of an article’s reasoning and tone should precede publication, especially when the article criticizes work that has been well received in academic circles by LDS and non-LDS scholars alike.

  15. Thank you for this thorough and thoughtful response. It helps clarify just how much merit there is to Brian Stubbs’ surprising work. I can see that it pained you greatly to have to write this, but I am keenly grateful for your scholarship and for Brian Stubbs’ as well, of course.

    I can understand a bright young scholar misunderstanding something as unexpected as Stubbs’ discovery and looking at it in a completely wrong way, but for that work to make it into the Maxwell Institute’s annual journal is perhaps the most painful part of the story. Is anything that knocks apologetics considered to need no scrutiny?

    Has nobody there listened to what Stubbs has been saying and writing on this topic? But even someone not familiar with Stubbs’ work ought to have picked up on many other issues, including the bad math. Stubbs has also posted a response on the Mormanity blog where additional issues are pointed out.

    Thank you again for this!

    • I should add that I hope Rogers will respond to this and Stubbs’ rebuttal and recognize at least reconsider what Stubbs has provided and acknowledge some of the errors he has made. I also hope that this misstep on Rogers’ part will not stop us from respecting his other work and look forward to seeing what he does in ongoing research. Hoping he can correct things where needed and move on successfully. I wish him well.

      • I agree with this wholeheartedly. Chris is a fine researcher and scholar, as shown in his otherwise serious publications and papers. Anyone who reads this should know that.

        It mystifying to me, however, that he sought to publish it in the first place. Sadly, it was unnecessary. Its content is a glaring injustice to an enormous work of a lifetime; to Stubb’s exacting life’s work, and to Stubbs himself.

        Stubbs’ 2015 is out there ready for inspection. If there are effective, well researched reasons to dismiss the whole of, so be it. But dismissal without justification is both unwarranted and without merit.

  16. Does the Journal of Book of Mormon Studies exercise editorial review and evaluation? Or do the editors just print whatever submissions that come along? The errors of Rogers appear not to be matters of debate or differing viewpoints but fundamental errors of understanding and analysis. If I were the editor or publisher, I’d be embarrassed.

    • My company is the publisher of Brian Stubbs 2015 book which was the primary source of Rogers review. The book went through a peer review process with over a dozen linguistic scholars consulted for review and input. Any responses were reviewed and where appropriate the book draft was updated.
      I was absolutely aghast at reading the review from the Journal, as it was academically incompetent (for example making statements that Brian ignored Nahuatl when in fact the data uses over 800 examples of Nahuatl). I contacted Joseph Spencer, the editor of the Journal as to their level of review of book reviews. Here is his emailed response: Hi Jerry,

      As is standard in the academy, book reviews are editor-reviewed rather than fully peer-reviewed. And as with most journals, the JBMS has a book review editor (Janiece Johnson) who (1) receives and makes decisions about occasional (rather infrequent) unsolicited reviews and (2) solicits reviews and makes decisions about them when they come in. Decisions are made in consultation with myself.

      Hoping this helps,

      Joe

      So this was the process.

  17. Thank you for the clarifications, Professor Robertson. This helps explain Rogers’s missteps in his evaluation of Stubbs’s work. It should also help vindicate Stubbs’s exhaustive study while demonstrating the effect of sloppy, “lumping” assumptions by Rogers. Bravo, Mr. Stubbs.

Add Comment

Your email address will not be published. Required fields are marked *

 characters available

All comments are moderated to ensure respectful discourse. It is assumed that it is possible to disagree agreeably and intelligently and comments that intend to increase overall understanding are particularly encouraged.

*