asked what the most impressive evidence is for Book of Mormon
authenticity, serious students of the Book of Mormon often point to one
of a small handful of items: the finding of candidates for Bountiful,
Nahom, and the River Laman in the Arabian Peninsula (particularly Warren
Aston’s Lehi and Sariah in Arabia and In the Footsteps of Lehi); the existence of chiasmus and Hebraisms, particularly
Hebraic wordplays; the diverse
and consistent testimony of the witnesses of the gold plates (see Richard
Lloyd Anderson’s works); and the strength of numerous cultural and
geographical correspondences between Mesoamerica and the Book of Mormon
(e.g., John Sorenson’s Ancient American Setting for the Book of Mormon and Mormon’s Codex, Brant Gardner’s Traditions of Our Fathers, etc.). Of these, I think the Arabian evidence has the most
easily appreciated “wow” factor. The evidence related to Nahom, including archaeological confirmation that that tribal name was in the right spot in Lehi’s day, and its precisely plausible location relative to the leading and amazing candidate for Bountiful. It takes serious effort and a great
deal of advanced scholarship to minimize the growing body of evidence
from Arabia — and so far those failed efforts have only helped to
highlight how improbable it was that Joseph could have fabricated the
details of Lehi’s trail. So Nahom and the Arabian evidence are often considered at the top of the list for impressive Book of Mormon evidence.
While the attacks of critics
have failed to diminish the luster of the Arabian evidence, a new work
from an LDS scholar may actually achieve that unintended effect — not by
attacking past scholarship, but by uncovering what may be an even more
exciting line of evidence for the Book of Mormon which one day may displace
Arabia as the “go-to” topic for Book of Mormon defenders. LDS linguist Brian Stubbs, through his decades of exploration of the Uto-Aztecan language family (spanning southwestern Mexico to the Western United States with languages like Nahuatl, Shoshone,
Navajo, and Hopi), has uncovered a very big surprise. It was not something he was expecting to find. He resisted it for years until the data became overwhelming and demanded some kind of treatment. But his research points to strong influence from Semitic languages in Uto-Aztecan that are consistent with two major infusions from the Near East into the ancient New World. In fact, there is evidence for an infusion bringing Egyptian language and one flavor of Hebrew, with another infusion bringing a different flavor of Hebrew (different set of behaviors in how sounds like “b” in Hebrew shifted in UA).
however, is that his evidence is far more technical than, say, showing
photographs of the proposed Bountiful site at Khar Kharfot in Oman and
listing how perfectly the leading candidate accords with Nephi’s text.
The strong and compelling evidence of ancient Semitic elements in
Uto-Aztecan from a skilled linguist, thoroughly aware of what it takes
to establish relationships between languages, demands a good deal from
reader to appreciate the linguistic data that now exists, and may take
decades before it’s explanatory power is widely recognized in the Church
and among other hesitant scholars. But what has been achieved already
is so remarkable and so interesting that it may well deserve to be the next big
thing in LDS apologetics.
Let me jump to the big
picture and put it in context: Stubbs has documented 1500
correspondences between Uto-Aztecan and ancient Semitic languages,
particularly Hebrew, Aramaic, and Egyptian. The Semitic influence
identified shows patterns consistent with two different infusions, an
infusion of one type of Hebrew/Aramaic along with Egyptian, as if from
the entry of Lehi and his group, and another infusion of a different
dialect of Hebrew that evolved in slightly different ways, as if it were
the Hebrew/Phoenician from Mulekites.
The level of Semitic presence in
Uto-Aztecan turns out to be much greater than the level of Hebrew in the
Yiddish language, which is well known to have developed from Hebrew
speakers coming to Europe. Similar changes and adaptation of local
languages may have happened with both the Nephites and Mulekites, but
the result has left us with much more easily identified remnants of
Semitic influence than we find in Yiddish This is a level far beyond
mere chance and highly contrived pattern seeking.
many New World languages, 100 to 200 such cognates are what was required
to show a legitimate connection and establish a language family.
Cognates and parallels in words and grammar happen by chance all the
time in languages. But when they are due to chance, you’ll find a few
handfuls, as we sometimes do between Chinese and English. Some, like
“mama” for mother may point to common ancient roots shared by many
languages, while others are just random and don’t fit any kind of
meaningful pattern. In related languages like German and English,
however, numerous cognates can be found and they often reflect sound
changes that follow some common patterns, like the hard “H” sound of
German’s buch becoming the k in book and in many other cognates (e.g., kuchen and cook, suchen and seek). Offering far more than just 100 or 200 cognrates, Stubbs so far has found and
published over 1500 cognates and identified many intriguing patterns
that point to a strong relationship between these languages. His work,
inherently highly controversial since it clearly supports Book of Mormon
claims, has been sent to his fellow Uto-Aztecan specialists, with no
public but several private comments so far, and eventually will be ready
for a fair peer review process, but this takes time and faces some
practical and political considerations.
Stubbs’ work is
in two volumes, one intended for LDS readers and one intended for
linguists. The lighter work for LDS audiences is Brian D. Stubbs, Changes in Languages from Nephi to Now
(Blanding, UT: Four Corners Digital Design, 2016). This 210-page book
includes useful background material on the evolution of languages and
the relationships that link languages, as well as some background on the
Book of Mormon. The meat of the book are the large sections exploring
patterns of relationships with many specific examples creating
impressive cases for relationships between Uto-Aztecan and Semitic
languages, including Hebrew and Egyptian.
larger, more technical volume is Exploring the Explanatory Power of
Semitic and Egyptian in Uto-Aztecan (Provo, UT: Grover Publications,
2015). This book has 436 large pages and plenty of small print with
extensive technical detail, offering 1500 detailed examples of
In subsequent posts, I’ll discuss some specific examples and the interesting trends he uncovers. To me, there are some really amazing finds that go far beyond what one might expect from chance. Some of these were shared in Brian Stubbs’ 2016 FairMormon Conference presentation, but there is much more to discuss and ponder.
26 thoughts on “Bigger Than Nahom? The Surprising Link Between Semitic Languages and the Uto-Aztecan Language Family”
One point to clarify, Navajo is not part of the Uto-Aztecan family.
Hi everyone! Just thought I'd drop in. A few comments:
(1) A minor point — there's some text missing at the end of your second paragraph.
(2) Normally, when a book like Exploring the Explanatory Power of Semitic and Egyptian in Uto-Aztecan is published, the publisher sends review copies out to the relevant academic journals. In this case I would expect Grove Press to send copies to, at the very least, the Linguistic Society of America and the American Anthropological Association.
Any idea if this has been done? If not, why not?
The peer-review problem is not going to go away. If people like Brian Stubbs and Stanford Carmack (and yes, you, Jeff) want to be taken seriously outside the immediate confines of the LDS community, they must subject their work to the scrutiny of the larger professional community. All this talk about how this will take place "eventually," but not, alas, quite yet, because of some vague "practical and political considerations," is just blowing smoke.
It is long past time to stop making excuses. Why hide one's light under a bushel, if it is truly light?
(3) I'm not an expert in this stuff by any means, Jeff, but the methodology raises some questions for me. Ditto for the way you're discussing Stubbs's work in your post. Uto-Aztecan is not a language; it's a language family. It's a grouping of many languages, and of course it's misleading to compare individual languages to language families. But that's what you do when you write things like this:
"The level of Semitic presence in Uto-Aztecan turns out to be much greater than the level of Hebrew in the Yiddish language."
Hebrew and Yiddish are individual languages; Semitic and Uto-Aztecan are not languages but groups of languages.
Common sense tells us that we'll find more cognates between language groups than between individual languages.
Suppose we go looking for names that appear in the telephone directories of both Appleton, WI and Jackson, MO. Suppose we next look for names appearing in all Wisconsin directories (Appleton, Madison, Eau Claire, etc.) vs. all Missouri directories (Jackson, St. Louis, Springfield…). Obviously, we'll find more "cognates" the second time around.
Would it make any sense to write a sentence like this? —
"The level of Wisconsonian presence among Missourian names turns out to be much greater than the level of Appletonese among Jacksonite names."
Well, of course it turns out that way. And the fact that it does is completely meaningless.
(4) Again, I'm not an expert in this stuff, but I would think that, in the absence of vast wars of conquest and/or modern trade and communication networks, languages track pretty closely with genetics. If Mohammed and his successors had not conquered so many other peoples, Arabic would pretty much be spoken only by Arabs, just as today the Hopi language is spoken almost exclusively by Hopis.
The question for Stubbs is this: how did so much of the language of these ancient Jews survive and spread so widely, while their DNA disappeared so completely? Obviously one can propose historical models that will account for this, but such models will have to be consistent with what we already know about the region's history. Good luck with that.
— Orbiting Kolob
Dear Orbiting Kolob,
Establishing relationships between languages can't fairly be equated with comparing names in a phone book.
In short, words aren’t necessarily related just because they sound or look similar. Words that are related should be both similar and different in consistent ways. Jeff gave some examples, but perhaps a more explicit one would help:
English: Icelandic of identical meaning.
whale → hvalur
when → hvenær
what → hvað
… And dozens of similar examples.
Virtually every wh word in English corresponds to a hv word in Icelandic with an identical or close meaning. Icelandic and English share many other cognate patterns. For example, most English words with sh correspond to Icelandic words with sk, with the same or related meaning. These patterns of difference and similarity are not only numerous but systematic. And this is sufficient to prove conclusively that the two languages descend, at least in part, from a common ancestor.
That is what Stubbs is attempting to show – not the hit and miss of random similarities, but a systematic relationship.
Furthermore, if there is a relationship between Semitic and Uto-Aztecan languages (and I think his evidence is fascinating) he is under no obligation whatsoever to prove how it may have come about. That is precisely NOT the point. Such a relationship would stand on its own and could theoretically be accounted for in a great variety of ways that do not include reference to the Book of Mormon.
But genetics are certainly no issue here. Perhaps I can be forgiven for again invoking Iceland, where meticulous genealogies have helped establish the dramatic contraction of genetic signatures that naturally occurs over even a century or two. The fact is that we propagate actual genetic material from only a percent or two of our ancestors that lived a mere hundred years ago. Most people's contribution to the gene pool dies completely out in not much more time. So after more than two millennia? The idea of genetics as an argument against the Book of Mormon is simply without scientific merit.
Thanks for the serious response to my comments, VW. My phone book analogy was used only to demonstrate a point about comparing language families to individual languages. And, with all due respect, I note that you have not responded to that point.
Any thoughts on why scholars like Stubbs and Carmack avoid peer review of this fascinating work? If, as you say, it can "stand on its own and could theoretically be accounted for in a great variety of ways that do not include reference to the Book of Mormon," why not take it mainstream? What's the holdup?
He isn't avoiding peer review. He has sent it to all the Uto-Aztecan scholars as the article notes (and as he told my husband and I during a conversation at the FAIR Mormon confernce.) The issue is that not all of them choose to respond for various reasons.
Interesting article. I look forward to hearing more about this. FYI, though, the description of the relationship between Hebrew and Yiddish is incorrect.
I think that if one does not know the Uto-Aztecan language that was spoken 2500 years ago nor the precise mixutres of Semitic languages that could have influenced the Uto-Aztecan language at the time, I think that it is an assumption to use the language family in this case rather than the precise language.
Crediting those who quickly rebutted “evidence”, of possessing not just “advanced scholarship”, but a “great deal of it”, and labeling their efforts as “serious”, seriously minimizes the words “scholarship” and “effort”. Further diminishing the luster of English words and mathematics is the post’s standard for “archaeological confirmation” and “improbable”. Even more convoluted, is describing a rebuttal that is un-rejoined as a “fail effort” of the rebuttal, not the original argument.
The most impressive evidence against the Book of Mormon is the post's concession that only “a small handful of items” are the “most impressive evidence” for it, Nahom “considered at the top of the list”.
Good catch on Navajo, Anon. I crossed that out. Stubbs' service on a Navajo-speaking mission is what got him interested in linguistics and Native American language, but I erred in listing that as Uto-Aztecan. It's a Southern Athabaskan language in the Na-Dené family.
Many of the comparisons in Stubbs work involve the words proposed in Uto-Aztecan, which can be related to many but not all of the words in individual families. Finding strong patterns between the proposed Proto Uto-Aztecan mother langauge (like Indo-European as a proposed common ancestor for Romance languages, Hindi, etc.) and Hebrew is meaningful when there are significant numbers of relationships and when they seem to follow similar patterns. (It is also interesting to see many of the connections reflected in a healthy mix of the modern UA languages of the UA language family.) But in some cases, the only comparisons are directly to modern languages and not a proposed Proto UA word.
Yes, working with a language family versus single language does open up the opportunities for more chance parallels, but not on the order achieved here. If chance gives you a tiny handful of random cognates for a language, using 20 related languages in the same language family will give you more chances for random hits, with, maybe, 5-10 times as many random hits. So 5-10 tiny handfuls. Not on the order of 10 times the quantity of cognates normally required to establish connections between languages. We'll get into some of the details in following posts.
IANAL (I Am Not A Linguist), but it seems to me that someone hoping to use shared cognates to establish historical relationships between languages would want to do something like this:
(1) Create a List of Basic Terms (call it the LBT) for which every language would presumably have a word (e.g., man, woman, water, eat, walk as opposed to coconut, porpoise, ice).
(2) Take two languages known to be historically unrelated (say, Maori and English?) and count up the number of shared cognates from the LBT. Since the languages are known to be unrelated, any such parallels will presumably be the product of chance.
(3) Repeat (2) for a bunch of other pairs of unrelated languages.
(4) Use the resulting data to create a baseline percentage, a figure indicating approximately how many (pseudo-)cognates are likely to crop up, simply due to chance, in any comparison using the LBT.
(5) Finally, use the same List of Basic Terms to compare Uto-Aztecan and Hebrew to determine whether the percentage of "hits" significantly exceeds the baseline percentage.
Of course, there are complications involved with each of these steps. Instead of looking for exact matches, one would want to account for known patterns in the ways that phonemes change over time, etc., etc., etc. But the basic idea should be clear. The basic idea is to avoid the methodological pitfalls of simply scouring entire vocabularies for shared elements and then saying "Wow! Look at how many parallels I found!"
To see what I mean, assume for a moment that the baseline described above is, I dunno, 5%. That is, on average, and simply as a result of chance, two unrelated languages will have shared cognates for about 5% of the words on the List of Basic Terms. If we then use the LBT to compare two languages whose relationship is unknown, and the percentage of shared cognates from the list is, say, 20%, then we might well have something. If it's only 6%, then probably not.
What we can't do is simply take two languages, make comparisons across their entire vocabularies, and say "Look at the number of hits!" And maybe I'm missing something, but this appears to be what Stubbs is doing, and if so, he's got a severe methodological problem. You can't just draw up a big impressive list of "hits"; at some point, you've also got to compare percentages against a baseline.
The general methodological idea here seems so basic to me that surely Stubbs must have done something like it. In fact, the general idea seems so basic that I can't imagine I could be the first to think of it. And sure enough, I'm not. After a bit of googling I found that there's something called a Swadesh List that can be used more or less as I've indicated above. It's "a classic compilation of basic concepts for the purposes of historical-comparative linguistics," as Wikipedia puts it.
So, has Stubbs used a methodology like I've just described? I did a quick search and didn't find any evidence of it, but as I said, I might have missed something.
I believe that the intent of Stubbs' article was to establish the rules governing the differences and then listing the cognates. The article is not about establishing what constitutes a baseline nor trying to guess which words would qualify to becoming a cognate, your List of Basic Terms, as languages influence other languages in different ways such as Old Norse influencing the pronouns in English and French influencing legal, military, political along with meat terms (pork and not pig, beef and not cow) in English. Likewise, Arabic has influenced Urdu as far as religious terms are used and these same religious terms are not used in the language that is nearly identical to Urdu – Hindi or the influence of Arabic in Swahili. Most likely, basic terminology already exists and the cultural influence fills some sort of void whether it is government, trade, science, religion, etc.
From what I can tell, the Swadesh List is used to determine the relatedness of languages within a language family and not the amount of influence one language had over another language.
On pages 90-91 of Changes in Languages from Nephi to Now, Stubbs shows results for a modified 100-item Swadesh list used in comparing languages in the UA family, but I don't recall seeing a Swadesh list used to compare Semitic-UA relationships. I can recall a number of the Swadesh words being among the cognates he discusses, but don't know what the scores will be. Fair question. Let me see if I can find an answer. From what I can see, though, the Swadesh tool looks like a pretty blurry lens for evaluating unexpected or secondary relationships between languages (e.g., the kind that occur when a base language is influence by an infusion by or interaction with another group). For example, we know there's a lot of Greek in English, but will analysis of the Swadesh words reveal a strong connection? The Swadesh list in English seems heavy on Anglo-Saxon, Germanic, Latin, and maybe French but light on Greek. Just a gut feel from scanning the list, no analysis. But if it gives a negligible score for Greek, does that mean that Greek words have not strongly influenced English?
Stubbs uses a chart showing the correlations between UA languages of 100 terms from a modified Swadesh list. Quite a few are in the 20s and 30s, while some are in the teens, and some are in the 70s and 80s.
Get to a common baseline like if Lehi and nephi were sitting here what language would they be speaking the day they landed and jump to years later when all those non Jews were around what language would they all be speaking? Was there a rabbi school for kids? Was some bi or tri lingual?
Stubbs is digging in a deep dark well and I am guessing there will be no fruit to support the BOM language issues. Btw how do you write adieu in Hebrew Aramaic or uztechan?
The "adieu" criticism is, in my opinion, a silly one. The BoM is a TRANSLATION! Just because "adieu" shows up doesn't mean there is some silly business going on. Sure…adieu wouldn't be found in the Hebrew language. But guess what….neither would "good-bye!"
Adieu is a common enough that most English speakers know what it means. It has almost become English. Thus, if there was a certain something, an emotion or a feeling, expressed by the original author of that passage in the original language which would be best captured in the word "adieu," I think a translator would be within his/her rights to use the word "adieu" in the translation.
I don't see a problem here. There are plenty of problems with the BoM. "Adieu" simply isn't one of them.
I do not know that this would be bigger than Nahom (although many critics do not accept that Nahom has any real significance). If Stubbs' work becomes fairly well accepted, there are alternative theories to account for the Semitic linguistic influence and I an sure that some will be proffered.
No one has proffered a plausible secular explanation for Nahom as of yet. Just a lucky shot in the dark.
Mormanity and Daniel Peterson proffered secular explanations for the rough translation of NHM in southern Arabia (no one agrees Nahom has been discovered). Mormanity proffers Walt Whitman’s Leaves Grass and the Moroni-Comoros Islands examples regarding similar statistical voodoo.
Daniel Peterson has this to say, "It's just…you know, if you take a long enough list of place names, you'll find parallels, especially if you're "loosy-goosy" about it. You'll find parallels with just about anything. This is easily done."
@Mormonography. I think you missed the point on "Leaves of Grass" published in 1855.
Don't know what you are talking about "rough" translation.
We aren't working with a loose long list of place names with Nahom.
While it is not proven that the Nehem/Nehm area and the Nihm altars there are a lot of LDS scholars and lay people alike that feel it is strong evidence, in for Nephi's narrative, since the way that he described the journey is the only way that anyone would encounter that name.
By the way, do you expect to win the lottery?
@Glenn. I think you missed the point on "statistical voodoo".
Don't know what you are talking about "Just a lucky shot in the dark”.
“We” aren't working with a strong evidence for a narrative of supernatural events.
While it is not proven that there is any match, there are a lot of scholars and lay people alike (Mormanity and Peterson included) that know many such parallels should occur by random chance and the way Nephi's narrative describes the New World lacks parallels anyone should encounter.
By the way, do you expect no one to win the lottery?
yep, no one should win the lottery that is exactly what the comment means
Yeah, now I am beginning to understand the psychology:
The critics win a large “lottery” so often they must be cheating. After nearly two centuries and spending a large amount on tickets, the apologist only won a ten-dollar prize. Clearly, no one ever wins an "honest" lottery.
The numerical aspect of Stubbs’ work is most clearly manifest in the large numbers of cognates, far greater than are normally required to establish that languages are part of a common language family. Over 10% of the vocabulary of the UA family has apparent Near Eastern influence. Of the 2700 cognates documented among UA languages, about 30% of them are part of the 1500 cognates with Near Eastern languages that Stubbs has found in this work. That's quantitative significance.
Swadesh lists are interesting tools for rough analysis, but this is a very fuzzy instrument (intended for other purposes) that need not give high correlations even when a strong relationship exists between languages. Morris Swadesh developed a list of words while studying a Native American language in Montana. Based on intuition, he proposed a group of words that was later winnowed down to 100 words to be used for dating changes between related languages (glottochronology). These are words that should exist across many different cultures, words like “eat”, “dog,” “liver,” “yellow” and “not.” There are many variations of his original lists, some as small as 35 words and others over 200 words. There is no agreement on which words or how many should be on the list.
As I understand it, this tool is intended to trace changes between languages known to be related. In Changes in Languages from Nephi to Now, Stubbs does provide a table comparing various UA languages using a Swadesh list (pp. 90-91). But there’s no reason to suppose that an infusion of a new language into an older language would tend to displace the common words on a Swadesh list. Why, for example, would the Norman invasion result in the English using “pas” instead of “not” or adopting the French words for, say, hand, knee, or house? Swadesh’s final list of 100 words strikes me as heavily Anglo-Saxon with little hint of the strong influence of French, Latin, and Greek in English. So is that list useful for testing the presence of French, Latin and Greek in English? It’s not the right tool for that, obviously.
The way an infusion alters a language depends very much on the details: is the new language spoken by elites or servants, conquerors or the conquered, wealthy or poor, merchants or cobblers? Does it provide a vocabulary for war, religion, medicine, diet, etc.? This depends on the unique historical context. It may be possible for one language group to make a significant infusion into another but leave any of the many variant Swadesh lists largely untouched.
I don’t think it’s the right tool for evaluating contact and influence with an outside language group, or for establishing whether two remote languages should be considered part of the same language group. For that, you look at the cognates and their abundance. That’s what Stubbs offers, in spades.
If you rearrange the letters in William Shakespeare, you can spell “here was I like a psalm”. In the King James Bible in Psalm 46, the 46 word is “Shake” and the 46 word from the bottom is “Spear”. William Shakespeare was 46 years old when the King James bible was completed. Of course, like Nahom, this is not just a coincidence.