Popular Book of Mormon Criticism Debunked

[tl;dr—Book of Mormon critics can be guilty of confirmation bias, too.]

This must be my week for Book of Mormon analysis, a last hurrah of ex-Mormon geekery before the Fall semester starts next week.

Vernal Holley in his book Book of Mormon Authorship compiled a list of Book of Mormon place names that resemble place names in the vicinity of Palmyra, New York, the site of the purported translation and the first publication of the Book of Mormon. (pp. 6263) I’ve seen his list cited several times by critics as evidence for modern human authorship and against the historicity of the Book of Mormon. [1]

Something about this list always bothered me. I recently sat down and figured out why: this criticism is deeply flawed.

Compiling such a list is an exercise in confirmation bias, the human tendency to look for evidence that confirms our preconceptions while failing to seek evidence that disconfirms them (or ignoring such evidence). The list compiler scans a gazetteer while noting any place names that are similar enough, ignoring the many place names that don’t match.

Further, this process doesn’t qualify as a scientific test which must include a way to know if the test fails. If a test includes no failure condition, then it can only succeed. After such a test, the tester is no more informed than before. Holley made no consideration of what a failed test would look like.

Even more, there is no objective test to determine which names match. This is left to the judgment and biases of the compiler. All-in-all, Holley’s list has the same evidentiary standard as any run-of-the-mill conspiracy theory.

There are at least 91 unique Book of Mormon place names, and Holley found a match for 29. [2] This seems to be a high proportion and a good result, but to be rigorous, we must test this impression to see if this is meaningful. We must find out how many matches we can expect by random chance.

As a test, I gathered a list of 6,836 place names in the state of Virgina to compare against Book of Mormon names. [3] I chose Virginia because it was reasonably distant from New York and therefore shouldn’t have figured in the writing of the Book of Mormon if Holley’s hypothesis is true, but the people who created these place names shared a common language and heritage with the people who named the cities on Holley’s list.

I found the following 30 similarities or exact matches which are no more far fetched than Holley’s (compare his list if you doubt).

This demonstrates pretty well that Holley’s result is not significant. A list with similar quantity and quality can be replicated with another area of the United States. The similarities that Holley found could be explained by random chance and confirmation bias.

There are two considerations that could make Holley’s list more significant. His list had matches between cities above the Great Lakes and cities in what the Book of Mormon calls the “land northward” which was separated from the “land southward” by a “narrow neck” of land, similar to the narrow strip of land between Lake Erie and Lake Ontario. (Alma 63:5) This makes these matches more significant.

Also, if Holley drew from a much smaller list of place names in the states and provinces around Palmyra, it would be more difficult to find a match. This would make the matches more significant. The area was much less settled at the time of the publication of the Book of Mormon, so we should expect fewer place names. On the other hand, Holley’s list draws from a much larger area that covers several states and provinces rather than a single state. Without the list of names that Holley considered, it is difficult to make a comparison.

Nevertheless, this criticism is severely blunted by its deep methodological problems and the demonstration that a similar list can be drawn from another region. It may be true that a human author drew inspiration for Book of Mormon place names from the surrounding area. Holley’s list does little to nothing to reliably demonstrate whether or not this is true. No self respecting ex-Mormon should give this criticism credence.

Update: To further demonstrate the effect of confirmation bias, I compared the Book of Mormon names to a list of 3,224 Chinese place names that came from Russian maps and were later transcribed into the Latin alphabet. [4] Given the shorter list, it would be reasonable to expect about half as many matches.

Despite using a much smaller corpus in an entirely unrelated language, I was able to find 24 similarities, including one exact match. Though these seem reasonable to me, some of the most tenuous matches could be discarded and still arrive at a list roughly half as long as the one obtained from comparing to Virginian place names.