Thursday, August 13, 2015

Mosiah Priority or Changing Authors?

Brent Metcalf made the observation that the usage of wherefore and therefore in the Book of Mormon changes dramatically between Mormon and Ether. I've borrowed an illustrative figure from this blog post:

What made Metcalf's observation interesting is that he found a similar crossover in the Doctrine and Covenants sections 4-19 (and 74 which has zero occurrences of either word), the sections dictated the same year as the Book of Mormon translation. His conclusion was that the crossover is thus evidence of a change in word preference by Joseph Smith at some time in the middle of translation, and thus evidence that Joseph started translation at Mosiah, went to the end, and then finished with Nephi's small plates.

I currently believe this was the order of translation, but I have doubts about the therefore/wherefore crossover being evidence of translation order. The primary argument in favor of it as evidence for translation order is that the words have essentially the same meaning, and are unlikely to have two exactly corresponding words in the original Book of Mormon language. This kind of observation could bring into question the results of stylometry, since stylometry depends on function words including therefore and wherefore. If Joseph's word preference over time can explain the style changes, then evidence of multiple authorship is weakened. This is why Metcalf's observation about the trend in the Doctrine and Covenants is important. That is one potential way to identify independently changes in Joseph's word preferences.

Since I don't have time this moment to go into great detail, I'm going to sketch why I think the crossover is more likely a result of style changes of the original authors than of preference in Joseph's word choice with time. In brief, Joseph used therefore about 211 times in the Doctrine and Covenants, and wherefore about 181. Both were used throughout the ~14 years that Joseph's dictations span. The distribution of both is roughly evenly distributed over the 14 years. However, there are exceptions. There are six sections where therefore appears more than 10 times. That's 35.5% of the appearances in 6 sections, and one of those happens to be in early 1829. Wherefore appears over 10 times in only 2 sections, comprising 14.4% of its occurrences. One of these happens to be from later in 1829. Thus these two outliers define the correlation that Metcalf observed with the Book of Mormon. Maybe a better statistician can show that these outliers make his correlation statistically insignificant. I can't without a lot of work, but I would lay heavy odds on my statistical intuition here. This graph shows the percentage of wherefores or therefores that appear one time in a section, two times in a section, three times in a section, etc.:

One thing this graph shows is that both words were used semi-frequently throughout the Doctrine and Covenants (each in at least 60 sections), and both words were used with rather high frequency in small numbers of sections. To me this raises the question of whether the words are used differently with different topics or genres? It would require a different kind of analysis to make this check, but I don't think we are seeing translation chronology and Joseph's word preference in the wherefore/therefore crossover. I think it more likely we are seeing an actual change in word use of the original authors. Perhaps we are seeing a major language shift in the original Book of Mormon. Maybe the authors of the small plates wrote in a different language or different dialect from Mormon's abridgement. It seems likely to me, since there was a several hundred year separation between them. A word from one language that felt like therefore could easily have been replaced by one from a different dialect that felt like wherefore, I would think. It would be interesting to see if any other homonyms have similar crossovers from the small plates to the abridgement. Brant Gardner's book suggested whoso and whosoever from another author's work, but those cross over between Helaman and 3 Nephi, and don't have enough appearances in sections 4-19 of the Doctrine and Covenants to covenants to conclude anything.

Take home message from taking a closer look at this claim? Correlation is not causation, and that's even more true with weak correlations. If you are going to make statistical claims, it's good to actually do the statistics--or have someone else do it.

No comments:

Post a Comment