Update: Check Jeff Lindsay's response here, which notes some benefits and drawbacks to the new wordprint analysis.
Three scholars from Stanford recently completed another "wordprint analysis" of the Book of Mormon, the newest in a series of studies completed by various researchers over the past few decades. Matthew L. Jockers (Department of English), Daniela M. Witten (Department of Statistics) and Craig S. Criddle (Department of Civil and Environmental Engineering) published "Reassessing authorship of the Book of Mormon using delta and nearest shrunken centroid classification" in December 2008. The study can be found in the Oxford Journal of Literary and Linguistic Computing.
I note out the outset I am not a statistics or linguistic analysis expert. I have read the paper and reached a few preliminary conclusions, though others are currently dissecting the paper and checking the intricate details. I expect a full response by people who know much more than me in the very near future. Be that as it may, I offer my initial thoughts here.
First, it should be stated that other wordprint analysis studies have been said to demonstrate the Book of Mormon had more than one author, and that statistically, the odds of Joseph Smith, Sidney Rigdon or Solomon Spalding (sometimes Spaulding) were found to be slim to none.1 Those who performed these earlier studies included believing Mormons and others unaffiliated with the Church. In the case of the current study, Craig Criddle has participated on the website "Recovery From Mormonism" (RFM) detailing his loss of faith and his leaving the Church, and written on his disbelief in the Book of Mormon in other venues. Thus, like the former studies supposedly used to bolster the claims of Joseph Smith, the new study (potentially used to contradict his claims) is not untainted by ideology or presuppositions.
Second, I see a pretty big flaw in their method. It seems they decided to use the original text of the Book of Mormon as published in 1830 by E. B. Grandin rather than the original manuscript or printers manuscript. Thus additions made by Oliver Cowdery in editing/copying the manuscript for printing, as well as any changes made by the typesetters, are unaccounted for. More importantly, they decided to analyze the current chapter structure, rather than the original:
We opted to use the chapter structure currently recognized by modern Mormon Church editors to create our text samples. This results in a total of 239 text segments for testing. This approach yields texts that are generally of adequate size (verses are too small and books too large), recognizes natural breaks in the narrative, facilitates cross-referencing to online resources, and avoids the chance that we have imposed our own bias.2The current chapter/verse structure is not original to the Book of Mormon. It was not created until 1879 when Orson Pratt published a new edition of the book under approval of the First Presidency. Brant Gardner's analysis of the Book of Mormon shows that Pratt follows a pretty consistent pattern in dividing chapters but actually did quite a bit of damage to the original structural patterns of the book. There are notable differences in chapter division style between the small plates of Nephi and the rest of the book, for example. He also divided chapters and verses to align with the structure of the King James Version of the Bible when a correlation was apparent. Like Pratt, Mormon also followed a consistent pattern in deciding when to begin new chapters and as Royal Skousen's work has shown, the original chapter divisions were a part of the original dictation manuscript.3 I believe this should have been taken into account. Were it shown that chapters overlap where different authors are said to have written the same chapter (now divided into separate chapters) the underlying assumption of the results is undermined. At any rate, I would have hoped that the group would stick to the original composition of the Book of Mormon as it was written.
Further, and I believe most damaging, the authors do not include Joseph Smith in the study. (Or did they include him, but not include the results in the published study?) Either way, that is a massively gaping hole. The study apparently shows that Rigdon and Spaulding are more likely than Pratt or Longfellow (a control subject) to have written the Book of Mormon. This says little about whether they did, only that it matches some better than others, and they did not include Joseph Smith himself. They also did not consider the nature of translation which is a fundamental aspect of approaching authorship of the Book of Mormon in my opinion. If it really is a translation we would expect it to reflect the knowledge or language of the translator. All the more reason Joseph Smith ought to be included in the results.
Finally, the study is premised on the shaky "Spaulding theory." I believe the historical foundation, as a result, is extremely tenuous.4 It also goes without saying that, despite Rigdon's falling out with the Church after Joseph Smith's death he never made a claim to have been involved with the Book of Mormon before his conversion.
Finally, as Ben McGuire pointed out to me, They do not claim to be able to assign a probability to each putative author. The authors of the study claim to be able to assign a relative probability to each putative author - that is, Solomon Spalding was 95% more likely to be the author of passage X than was Sidney Rigdon, and 99% more likely than Longfellow. In other words, it is not really proving any of the men wrote the book, only that some on their list are more likely to have done so, based on other writings, than the others. And again, Joseph is not in the mix. At best the study is a theory, though vastly far from hard proof. I remain unconvinced, just as before, in regards to wordprint analysis and the Book of Mormon.5 Now I wait to see what the experts have to say.
For a brief overview, check the article on FAIRwiki, "Book of Mormon wordprint studies." See especially the cited sources at the end of the page.
Criddle, et. al., p. 6.
While I do not follow all of Skousen's end conclusions, I believe his analysis of the Book of Mormon text itself is crucial and trustworthy. See Skousen, Translating the Book of Mormon, 85-86; Skousen, How Joseph Smith Translated the Book of Mormon, 27-28. His overall study is The Original Manuscript of the Book of Mormon: Typographical Facsimile of the Extant Text, edited by Royal Skousen (Provo, Utah: FARMS, 2001) and The Printer's Manuscript of the Book of Mormon: Typographical Facsimile of the Entire Text in Two Parts, edited by Royal Skousen (Provo, Utah: FARMS, 2001). Gardner's Second Witness series makes great use of Skousen's analysis.
See Wade Englund's response to the Spaulding theory here and Matthew Roper, "The Mythical 'Manuscript Found,'" FARMS Review (City Unknown: FARMS, 2005), 7-140.
In regards to wordprint analysis and the Book of Mormon, a good primer is John B. Archer, John L. Hilton, and G. Bruce Schaalje, "Comparative Power of Three Author-Attribution Techniques for Differentiating Authors," Journal of Book of Mormon Studies, 6:1.