The Book of Mormon text used in my apparatus was 269,551 words long with a unique vocabulary of 5,638 words (compared with the text of 271,240 words used in the Johnson study). The [Late] War was 56,632 words long (compare this to 55,378 words in the blog study – a difference most likely due to the inclusion of the appendix material) with a unique vocabulary of 5749 words. Significantly, the Book of Mormon text, while being nearly five times longer, has a vocabulary of similar size. And the shared vocabulary amounts to roughly forty percent of the respective vocabularies (specifically, they have a shared vocabulary of 2,281 words). My experience is that the vocabulary size of The Late War is consistent with books of similar length, while the Book of Mormon has an unusually small vocabulary. When we calculate the number of unique four-word locutions for each text, we can see the difference in repetition. The Book of Mormon contains 202,830 unique four-word locutions compared with The Late War containing 51,221. Why is this interesting to us? If we follow the weighted matches used by the blog, there are 549 shared four word locutions common to both texts. This means that of all the possible phrases found in The Late War, only 1.07% of them make it into the Book of Mormon. And within the Book of Mormon, of the potential 200,000+ unique phrases, only 0.27% could be derived from The Late War. This is not a high number. This ratio drops substantially when we back out the 75 parallels taken from the copyright application (with 474 parallels it becomes 0.93% and 0.23% respectively).

This sort of ratio (the size of the footprint relative to the size of the text) doesn’t come out in the calculations used.

. . .

Without considering the size of the texts, any sense of relative proportion is lost. Harold Love pointed out that we are likely to find some degree of coincidental overlap between any two texts of sufficient size. This is a relatively small footprint (textually) – finding only 474 parallels in more than 200,000 opportunities. It is much smaller than the connection between Pride and Prejudice and The Officer’s Daughter.

. . .

The final issue is over the challenge of context. When we take texts and reduce them to these strings, we eliminate context. We rip out punctuation. Our four-word phrases cross natural textual lines. Without a more nuanced parsing, this is the only possible outcome. But it doesn’t help us understand the relationship between texts. Because I quoted it verbatim earlier, the copyright statement makes a terrific example. Here are a few lines from it:

In conformity to the act of Congress of the United States, entitled, “An act for the encouragement of learning, by securing the copies of Maps, Charts, and Books, to the authors and proprietors of such copies, during the times therein mentioned;” and also the act entitled, “An act supplementary to an act, entitled, ‘An act for the encouragement of learning

The blog identified a couple of noteworthy parallels in this short text:

entitled-an-act-for : time-therein-mentioned-and : therein-mentioned-and-also : act-entitled-an-act : entitled-an-act-supplementary : act-entitled-an-act : entitled-an-act-for …

Each of these four-word phrases crosses a textual boundary. They move from the immediate statement to a quotation of another text. This movement is lost. These phrases cross sentences and paragraphs. They string words together that don’t belong together except in the sense of an n-gram – a computational model based on removing the markers of these divisions from the text. The relationships that can sometimes be seen in these parallels don’t exist for us as readers (or as writers). These aren’t phrases that occur for us (or in our environment) because they don’t actually exist as phrases (as locutionary acts). These are naturally rarer – because they are created entirely by coincidental circumstance and not by design of any author. And, in using them in a way that weights rarity more heavily, we tend to emphasize a feature of the language that doesn’t exist except in the computational representation of word strings that no longer correlate to real writing or to real speech. These fragments, strung together, cannot provide us indicators to the language usage in comparison because they don’t represent language usage at all.

. . .

It isn’t a particularly difficult feat to reconstruct the Book of Mormon using phrases found from many different sources. In the 1960s, Julia Kristeva coined the term intertextuality to describe this feature of all texts. They were, as she described them, a ‘mosaic of quotations’ all coming from other sources. Some of this is certainly due to textual influence and reliance. There is no doubt that the Book of Mormon owes a great deal of its contents to the King James text. But, as Harold Love points out, given a large enough body of literature, you can also find these phrases caused by coincidence. In the long run we note that there are some real similarities that can be found in the texts of these two books. But, most of these similarities are not discovered by creating a list of these four-word phrases – because these phrases are not themselves meaningful. Does this process attempt to reduce the significance of the Book of Mormon to a few hundred four-word phrases, stripped of punctuation and context? That seems to be the outcome. Hunt wanted to create a text that read like scripture as a marketing tool. In this way we get a lot of biblical sounding text. The Book of Mormon, on the other hand, doesn’t just use biblical language, it engages biblical issues – it asks questions about morality, about agency, about creation. It ponders the meaning of writing and reading. It describes religious experience.