Roger writes:
So after reading the above about 4 times, I think I get the basic jist of it. Assuming I do understand the above, my first observation would be: 40% seems rather low, assuming Spalding wrote the Book of Mormon. However, that, of course, is not what S/R postulates at all. Instead, S/R postulates that a Spalding ms was used by someone else, (for sure Smith/Cowdery and probably also Rigdon) in their production of the Book of Mormon. So… given that factor, 40% seems to fit rather nicely. Do we all agree with that so far?
This is more of a sticky kind of issue. The 40% is really more of what we would expect between two texts of about the same size as the Spalding Manuscript - instead of one that was as long as the Book of Mormon. So, from Book of Mormon to Spalding, its quite low. In the other direction its about average (what we would expect). Now this is unique words. Longer texts will almost always have a higher overlap with smaller texts. And authors will have an impact (after all, compare anything to James Joyce, and your percentages will be rediculously high on one side and horribly small on the other).
This percentage isn't really a big deal in terms of total verbiage. The first 50 words or so will make up fully half the text. The rest scale down rapidly. So even a 60% gap in verbiage results in a very small portion of the over all text. I could calculate it exactly, but its not really going to provide us with highly useful information. What I wanted to demonstrate with these figures is that they are low compared to texts where we know we have plagiarism in one direction.
The other thing is that given the nature of these word lists, you won't find adding Rigdon to Spalding for a vocabulary control to add a lot. In other words, most of the verbal overlap between Spalding and the Book of Mormon will also be the same verbal overlap between Rigdon and the Book of Mormon.
Finally, you have to remember that the word studies (like Criddle's) only use the most common vocabulary words. In general, these words appeared in all authors in the study (although I would have preferred a different criteria that would have modified perhaps a dozen or so of the words used). They look at the frequencies of these most common words - the ones that make up collectively such a huge percentage of the texts. So, the rare and unusual words won't have any impact for the most part on these studies.
Now, if you want to provide me with a clean copy of some Rigdon material, I can provide actual numbers to match with this if you want.
This is where it gets a little fuzzy—for me at least. It seems like there are overlapping factors here that could only be “graphed” by complicated algorithms that go way over my head. For instance, Ben states that “Warren is about as long as the Book of Mormon. Ramsay is about two to three times the length of Spalding (but nowhere near the length of either Warren or the Book of Mormon).” –this, I would think, would greatly impact results when we are talking about percentages that are derived from comparisons of large numbers of shared locutions. If text A is short by comparison to the text we are comparing it with, but text B is not so short by comparison to the text we are comparing it with, it seems that you will almost surely get different results, even if all four were written by the same author. No?
You would, but not necessarily in the direction you might think. Suppose, for example, that to make the test more even, I truncated the Ramsay text to be the same length (word count wise) as the Spalding manuscript. The Spalding manuscript that I used had a word count of 36,930. I eyeballed it, and chopped the Ramsay text at 37,044 - so close enough. The resulting truncated Ramsay text had a total vocabulary of 4,388 (down from 7,351). The new percentages were 86.2% and 29.6% respectively. The longer the text, the more "unique" words we expect to see. (I note that this isn't the case with the Book of Mormon which has a tiny vocabulary - in Warren, for example, there is 1 unique word for every 22.6 total words. In the Book of Mormon that drops to 1 unique word for every 48.3 words). So while you are right, that the results change, when I reduce the size of the Ramsay text by simply truncating it to a pre-determined size, the results become far more skewed instead of far less skewed.
Beyond that, from the outset, this seems to be two very different scenarios. On the one hand you have Warren making use of Ramsey and on the other you have Spalding writing one ms and then allegedly writing another, which is then greatly embellished and altered by someone else in order to produce something different from the original--hopefully different enough to avoid discussions like we are having. Somehow, these differences have to be taken into account and that, I suspect, is what Dale attempts to do by identifying the most “Spalding-like” sections of the Book of Mormon and generally not considering the Book of Mormon as a homogenous text. I am not sure that we should expect similar results going the other direction even if the S/R theory is true… ? But Perhaps I am just not getting the logic.
Perhaps a better gauge would be my comparison between the Book of Mormon and the English translation of Jules Verne's book. I will line it up with the overal similar figures from the Spalding comparison. Since Verne's book is longer than the Spalding text, I will truncate to about the first 37,000 words.
Basic stats:
Length - Book of Mormon: 269,072 Spalding: 36,930 Verne: 37,096
Vocabulary - Book of Mormon: 5,575 Spalding: 5,251 Verne: 5,137
Vocabulary Overlap with the Book of Mormon - Spalding: 2227 (39.9%/42.4%) Verne: 2016 (36.1%/39.2%)
The difference is quite small - and more than easily explained by the fact that the time and location gap between the Book of Mormon's writing (translation or whatever) and Spalding is much smaller than the gap between the Book of Mormon and Verne's book. But in the end, my real point is that these kinds of numbers don't indicate borrowing on any substantial level. While with texts that we do know borrowing occurred (the Warren/Ramsay connection) we do see a significant gap. To demonstrate this, I am going to provide the results of the same comparison using three texts - Book of Mormon X Ramsay and Warren X Ramsay.
Length - Book of Mormon: 269,072 Ramsay: 37,044 (truncated version) Warren: 288,977
Vocabulary - Book of Mormon: 5,575 Ramsay: 4,388 Warren: 12,762
Vocabulary Overlap with Ramsay - Book of Mormon: 1,850 (33.1%/42.1%) Warren: 3,784 (29.6%/86.2%)
Do you see the difference here? Comparing the Book of Mormon with the truncated Ramsay yields some numbers that are reasonably close to those generated by comparing the Book of Mormon to both Verne's book and Spalding's manuscript. But, comparing Ramsay with Warren yields some rather different figures.
Now there are, I am the first to admit, a lot of ways to interpret this data. One of the biggest features (that I don't think you mentioned) is that the small vocabulary of the Book of Mormon is a surprise given the length of the text. And this will tend to skew the numbers in my opinion. But, what I do see is that the numbers between Spalding and the Book of Mormon are rather typical and not particularly extraordinary. And I don't mind running other texts if you had some you wanted me to look at in particular.
Again, isn't that to be expected due to the size of Warren in comparison to Spalding's RS? In other words, if Spalding had continued writing his RS for another 200 pages, wouldn't we expect the 3-word locutions to get closer to that of Warren & Ramsay (assuming S/R is true, that is)? I may be wrong, but it seems to me that this can and should be accounted for by the size difference of the documents.
Running this on the truncated Ramsay does reduce this number. In fact, as expected, it drops it down by a similar ratio as the text was cut, to 6,173. Still a number far, far higher than the similar comparison between the Book of Mormon and Spalding. In fact, the truncated Verne yields 1715 common 3 word phrases with the Book of Mormon (about 5.0%/1.2% compared to the Spalding text's similar 6.4%/1.5%). At 6,173, the percentages in the Ramsay/Warren case are 18.8%/2.7%. This is a huge gap. By the way, unlike with the vocabuary, most 3-word strings in a text are unique. This means that while chopping a text up will result in fewer shared strings, it also means that the number of these strings when viewed as a percentage will not shift much. In this case, using just the first 70,000 words of Ramsay provides a significant increase in that proportion. So at least in this regard, the numbers don't change all that much when the text length is equalized. The difference is quite amazing.
--Just assuming for the sake of discussion that that observation is correct, I think I can further qualify it by again pointing out that even if Spalding were to have written another 200 pages, we still should not expect the numbers to approach those of Warren/Ramsay since Warren/Ramsay is direct, whereas Spalding-Spalding-Rigdon-Smith/Cowdery is obviously not.
Of course, so now the question is, are these numbers significant even considering that aspect of it? Do you know of some indirect borrowing I can test this against? The only thing I can look it is the number I took from Verne's book, written in 1872 (published as a serial in French) and almost immediately translated into English. You would think that this would have little connection to the Book of Mormon. It doesn't contain a lot of the warfare (by intention) but some travel (by intention). And this gives us results that compare quite well with Spalding's manuscript - which arguably has much closer themes, is more contemporary in terms of language, and so on. If you had to request a study to see how normative the Spalding figures are, how many books (and of what sort) should I use in the study, from what time periods, and so on. If it would be convincing to you (and I am not sure anything will be), what would it take? I am fairly confident (from working with texts at this level for a while now) that my kinds of statistics will show that the Spalding-BoM comparison of this nature is not unusual and doesn't even indicate indirect borrowing.
Now don't take this to mean a blanket rejection of the notion - what I am trying to convey isn't that, merely that these kinds of statistics are not any kind of adequate indicator of this kind of relationship that you are proposing.
This is long, so more in the next reply.