Book of Mormon (er, -- of Solomon)

[_Danna]

What if a new word-print study demonstrated that all three sections of "Spaldingish" text very likely came from the SAME writer?
We would not even have to mention Spalding's name. -- If Zeniff, Alma/Mormon, and Ether/Moroni in the Book of Mormon are all from the SAME author, then something is very wrong with the Nephite Record.

Great idea!

If chunks of the book could be tested against other chunks to determine same or different authors, that would be interesting. A start point would be the authorship claimed in the book. If the book is what it claims, each separate author will have a separate voice with traces of either Moroni or Mormon as appropriate. JSjr's voice should be uniform across the lot.

For JSjr alone, we should expect no difference between any claimed authors - this was not supported by Jockers et al., but the counter hypothesis would be strengthened by testing the book against itself.

I don't think that Smith alone is actually at 180 degrees to the S/R theory though. First of all, Jockers et al. (and the new data presented at the exmo conference) indicates both JSjr and Oliver are involved as well. Next, a lot of the data supporting the Smith theory was initially analysed in light of an ancient vs 1800s hypothesis - and does not differentiate between JSjr alone vs JSjr and others.
Was anybody expecting to see Oliver, PPP, and probably JSjr turn up as potential authors?

It would be great to see a synthesis of evidence including the work Dan Vogel and others have done as well, taking into account the liklihood of multiple modern authors.

I think it is great that your textual analysis - looking at word pattern and contextual words ties in so nicely with the statistical analysis - for the Jockers study it would not have mattered if the words in each chapter were randomly arranged. Just brilliant.

[_Uncle Dale]

What if a new word-print study demonstrated that all three sections of "Spaldingish" text very likely came from the SAME writer?
We would not even have to mention Spalding's name. -- If Zeniff, Alma/Mormon, and Ether/Moroni in the Book of Mormon are all from the SAME author, then something is very wrong with the Nephite Record.

Great idea!

If chunks of the book could be tested against other chunks to determine same or different authors, that would be interesting. A start point would be the authorship claimed in the book. If the book is what it claims, each separate author will have a separate voice with traces of either Moroni or Mormon as appropriate. JSjr's voice should be uniform across the lot.

Of course the Mormon apologists will come up with some wild hare explanation --
that's their job. They'll probably say that since God inspired the book, that
overwhelming inspiration made even two totally different authors write in the
same manner, etc. etc.

Also, from Mosiah up until, Ether, a huge block of text was edited by Mormon. So any
similarities in language discovered throughout that lengthy text could all be attributed
to Mormon's homogeneous effects.

Except for lengthy quotations from Isaiah, Malachi, etc., the book theoretically could
have one single word-print, matching well with that of Joseph Smith. Such is the view
that Dan Vogel and the other Smith-alone authorship proponents would have us believe.

But there is no single "voice uniform across the lot."

For JSjr alone, we should expect no difference between any claimed authors - this was not supported by Jockers et al., but the counter hypothesis would be strengthened by testing the book against itself.

My guess is that Vogel and similar writers would oppose such a study, as being not only
useless, but confusing. They would argue that "word-printing" is junk science.

I don't think that Smith alone is actually at 180 degrees to the S/R theory though. First of all, Jockers et al. (and the new data presented at the exmo conference) indicates both JSjr and Oliver are involved as well. Next, a lot of the data supporting the Smith theory was initially analysed in light of an ancient vs 1800s hypothesis - and does not differentiate between JSjr alone vs JSjr and others.
Was anybody expecting to see Oliver, PPP, and probably JSjr turn up as potential authors?

I'm still unconvinced as to the Pratt and Smith "voices" in the Book of Mormon text. Yes, they may be
present -- but I'll need to see some compelling literary explanations, as to what they
supposedly contributed, and why. I have the sneaking suspicion that some of Jockers'
Pratt authorship attributions are mis-readings of complex blocks of text having multiple
authorship, heavy editing, etc.

It would be great to see a synthesis of evidence including the work Dan Vogel and others have done as well, taking into account the liklihood of multiple modern authors.

That is for the next generation of Mormon origins scholars to synthesize -- we stand too
close in time and space to these recent textual discoveries. We do not have sufficient
contextual perspective to present a useful overview.

I think it is great that your textual analysis - looking at word pattern and contextual words ties in so nicely with the statistical analysis - for the Jockers study it would not have mattered if the words in each chapter were randomly arranged. Just brilliant.

What I've done contributes a second, different sort of view of the texts. It is a beginning,
but we really need two or three other, different kinds of studies carried out before the
average student of early Mormonism and early Mormon texts will take notice.

There are some other ways to measure the language and plot distribution patterns.
The grammar of the book could be quantified and reported, as could some other ways
of examining the uniformity/diversity of the English in the texts. Simply counting the
number of nouns on each page -- or the number of subject/verb reversals -- might
provide us with some additional "literary texture" for the book's narrative.

But once that is accomplished -- and people actually do take a second look at the idea
of multiple 19th century authorship -- then, it will be important for somebody to try
and explain WHY that evidence for multiple authorship reads as it does -- and what
it tells us about the earliest Mormons and their secretive activities.

UD

[_Benjamin McGuire]

Ok,

For this quick study, I am using Spalding’s Oberlin Manuscript (following Dale’s designation). Since his methodology seems to be to consider the local text (the text being examined) in page size chunks of material, but to use the source document as a single coherent piece in terms of vocabulary and word strings, I outline my process as follows:

The base local text for the Spalding Manuscript was taken from Dale’s site under the title of corrected e-text (It wasn’t my intention, of course, to get into issues over the textual history of that manuscript). I then separated this into the “pages” of material following Dale’s transcript of the original source. Punctuation and numbers were stripped out of all files.

Unlike Dale, I do have the capacity to determine relatively quickly, all locutions (Dale’s “word strings”) of three words in length. While this is not going to be used directly in the tabulations following Dale’s model, I am going to be indicating some of these statistics as we go - for 2, 3 and 4 word locutions. These will be exact locutions, not making exceptions for plural forms, or other kinds of similarities. While any discussion of “tabulations” is based to some degree on rather subjective qualifications, I am going to try to stick as best I can to Dale’s critiera where I explicitly adopt it, and will not add any of my own criteria.

My Source text for this project is Mercy Otis Warren's History of the Rise, Progress, and Termination of the American Revolution, 1805 - its publication date, provides an adequate window for an alleged influence on Spalding’s writings.

Given the length of the text, my simple normalization process to remove punctuation and other oddities in the text will have an impact, however, since we are comparing it as a source text, errors in this normalization will have less an impact on the kind of statistics Dale generated – generally causing errors on the conservative side which should not influence the final tabulations in any significant way.

I want to point out one thing though, and that is this – the single largest criteria for your study is that of overlapping vocabulary. The impact of your words strings is seems minimal. After all, on page one of the Spalding manuscript, with 180 words, all but 5 were accounted for in Warren’s text – a whopping 97.2 percent. In terms of shared vocabulary, this beats all but 3 of the top 25 pages you list between the Book of Mormon and the Spalding Manuscript.

For the next stage, I broke down the various texts into 2, 3, and 4 word locutions. Then built frequency tables and compared them. Then I produced ratios as you did for all of the common 4 word locutions. I then looked through the 3 word locutions for acceptable groupings - and I chose a few. I did not spend the time to look at gaps or similar words (which is to say, I did not combine similar word forms or plurals with singulars) – I only used exact comparisons. This means that these figures are conservative in another way. In other words, from your list of 7 criteria, this preliminary look only uses criteria 1and 3, but not 2, 4, 5, 6, or 7.

The nice thing about my software is that while the source text took some time to analyze (280,000 words of text), this analysis only had to occur once. Manipulation of an individual page with less than 200 words is very quick, and since I was generating lists of locutions and their frequencies, I merely had to compare these very small lists with the master list from Warren.

My preliminary analysis here covers just the first 15 pages of Spalding’s manuscript. Some statistical notes follow (greater detail has been spent on page 1 than on the other 14 pages).

The following is a chart of just those 15 pages (out of the 171 pages on Dale’s site). I decided to simply tackle the first set, although I could, I think, have tackled all 171 pages, and simply culled the best of the best with some spectacular results. Here is a chart from highest to lowest of those 15 pages. It uses the same calculations that Dale used to produce his chart at the beginning of this thread. Note that instead of my producing a list of which chapters in the Book of Mormon are most “Spaldingish”, I have produced a chart indicating how the pages in Spalding’s manuscript are ranked as most “Warrenish”.

Column 1: Page number, Column 2: Word Count (modified), Column 3: Common Vocabulary on the page, Column 4: Tabulated Word Strings, Column 5: Ratio of Tabulated Words Strings to modified Word Count, Column 6: Total Value (produced by adding columns 3 and 5.

3 120 0.958 10 0.083 1.042
1 180 0.972 9 0.050 1.022
6 180 0.983 7 0.039 1.022
5 175 0.954 8 0.046 1.000
7 161 0.963 5 0.031 0.994
2 189 0.947 8 0.042 0.989
13 151 0.947 6 0.040 0.987
8 177 0.932 9 0.051 0.983
4 186 0.957 2 0.011 0.968
14 143 0.916 6 0.042 0.958
9 140 0.943 1 0.007 0.950
10 152 0.908 6 0.039 0.947
11 170 0.924 2 0.012 0.935
15 169 0.858 0 0.000 0.858
12 135 0.800 2 0.015 0.815


The averages are: 162 words per page, 93.1% common vocabulary, 5.4 tabulated word strings, a tabulated word string ration of 3.4% and a total average score of .965

As a consideration here, if this were close to a typical cross section of the text, I would expect to see the top 25 pages closer to a 97-98% common vocabulary with an average total score in the 1.025-1.035 range.

As you can see though, in the first 15 pages, the results I think justify my position. 3 of these 15 (20%) would fall in the 5 of Dale’s list, with my number one exceeding all of Dale’s best examples. My averages – using just the ratio of tabulated word strings and common vocabulary (in my case 3.4% and 93.1%) compare quite well to the three sets of calculations which Dale produced for his texts – Alma with 94.6% and 3.5%, Mosiah with 93.3% and 2.1%, and Ether with 93.5% and 2.3%.

And finally, I want to point out that I used 2 of Dale’s criteria for tabulating word strings and not all 7. This would suggest that my figures here are quite conservative.

Page 1, the following words found in Spalding are not found in Warren: cave, coneaught, conical, lever, mound, stones, walking. Only the word stones occurs more than once (it occurs twice). However, since the word “stone” does occur in Warren, I am not sure how this fits in with your vocabulary model. I have opted to move these two instances to the matching vocabulary section to sit alongside the two instances of “stone” which does match a word in Warren. Coneaught would seem to fit the Proper Noun rule and likewise has not been counted. This reduces the total word count on page 1 to 180, with 175 words of overlapping vocabulary (97.2%).

In terms of locutions (word strings), the following exist:

2 word locutions (i.e. your “word string”): 108
3 word locutions: 33
4 word locutions: 8

The 4 word locutions were:

“and that it was” (occurs twice)
“by the ravages of”
“from the fort and”
“the assistance of a”
“this was at a”
“to me to be”
“with the assistance of”

I included one 3 word locutions as having sufficient information to be included:

“appeared to me”

Bringing the total significant word strings to 9. This ratio (9:180) was then added in per your formula, bringing the total for page one to: .972 + .05 = 1.022.

I note that this would take the 4th spot in your list.

Page 2:

With 189 total words, there were no proper nouns to eliminate, and there were 10 words (only 8 unique) that occurred in Spalding but not in Warren: cave, cavity, diameter, earthen, parchment, perpendicular, tight, tore: 94.7%

2 word locutions: 79
3 word locutions: 21
4 word locutions: 6

3 word locutions used: “for that purpose”, “gained the ascendancy”,
4 word locutions: “for that purpose and”, “gained the ascendancy and”, “the bottom of the”. “the design of this”, “the form of a”, “to the bottom of”

I realize that both of the 3 word locutions are contained in 4 word locutions. The total difference in counting 8 as opposed to 6 in this case ends up being approximately 1 percent (.01). Neither a .979 or a .989 would make your top 25 (although it would be darn close). But, in any case, these numbers are significantly higher than your average for the Book of Ether comparison you did, and even higher than the averages in Alma (even dropping the two 3 word locutions).

Page 3:

The word count was dropped from 120 to 119 to exclude the word “Latin”. Apart from this, there were only 5 words in Spalding not in Warren: lightly, skilled, thou (X2), treadest: 95.8%

2 word locutions: 71
3 word locutions: 28
4 word locutions: 9

Included 3 word locutions: “variety of subjects”

Total: 1.042 (beats all of your top 25)

Page 4:

186 words, 8 words not common: 95.7%

2 word locutions: 102
3 word locutions: 31
4 word locutions: 2

Total: .968

Page 5:

175 words, 8 words not in common: 95.4%

2 word locutions: 89
3 word locutions: 28
4 word locutions: 6

Included 3 word locutions: “arrival in america”, “a candid mind”

Total: 1.000

Page 6:

180 words, 3 not in common (excluded Constantine): 98.3%

2 word locutions: 112
3 word locutions: 32
4 word locutions: 7

Total: 1.022

Page 7:

161 words, 6 not in common: 96.3%

2 word locutions: 79
3 word locutions: 19
4 word locutions: 3

Included 3 word locutions: “preparation was made”, “spread her sable”

Total: .994

Page 8:

177 words, 12 not in common: 93.2%

2 word locutions: 95
3 word locutions: 29
4 word locutions: 9

Total: .983

Page 9:

140 words, 8 not in common: 94.3%

2 word locutions: 71
3 word locutions: 15
4 word locutions: 1

Total: .950

Page 10:

152 words, 14 not in common: 90.8%

2 word locutions: 79
3 word locutions: 23
4 word locutions: 5

Included the 3 word locution: surprise and astonishment

Total: .947

Page 11:

170 words, 13 not in common: 92.4%

2 word locutions: 84
3 word locutions: 19
4 word locutions: 2

Total: .935

Page 12:

135 words, 27 words not in common: 80%

2 word locutions: 58
3 word locutions: 19
4 word locutions: 2

Total: .815

Page 13:

151 words, 8 words not in common: 94.7%

2 word locutions: 86
3 word locutions: 28
4 word locutions: 6

Total: .987

Page 14:

143 words, 12 not in common: 91.6%

2 word locutions: 66
3 word locutions: 17
4 word locutions: 5

Included “necessary to establish”.

Total: .958

Page 15:

169 words (excluded Jesus), 25 words not included: 85.8%

2 word locutions: 73
3 word locutions: 14
4 word locutions: 0

Total: .858

[_Uncle Dale]

....
The following is a chart of just those 15 pages (out of the 171 pages on Dale’s site). I decided to simply tackle the first set, although I could, I think, have tackled all 171 pages, and simply culled the best of the best with some spectacular results. Here is a chart from highest to lowest of those 15 pages. It uses the same calculations that Dale used to produce his chart at the beginning of this thread. Note that instead of my producing a list of which chapters in the Book of Mormon are most “Spaldingish”, I have produced a chart indicating how the pages in Spalding’s manuscript are ranked as most “Warrenish”.

Column 1: Page number, Column 2: Word Count (modified), Column 3: Common Vocabulary on the page, Column 4: Tabulated Word Strings, Column 5: Ratio of Tabulated Words Strings to modified Word Count, Column 6: Total Value (produced by adding columns 3 and 5.

3 120 0.958 10 0.083 1.042
1 180 0.972 9 0.050 1.022
6 180 0.983 7 0.039 1.022
5 175 0.954 8 0.046 1.000
7 161 0.963 5 0.031 0.994
2 189 0.947 8 0.042 0.989
13 151 0.947 6 0.040 0.987
8 177 0.932 9 0.051 0.983
4 186 0.957 2 0.011 0.968
14 143 0.916 6 0.042 0.958
9 140 0.943 1 0.007 0.950
10 152 0.908 6 0.039 0.947
11 170 0.924 2 0.012 0.935
15 169 0.858 0 0.000 0.858
12 135 0.800 2 0.015 0.815


The averages are: 162 words per page, 93.1% common vocabulary, 5.4 tabulated word strings, a tabulated word string ration of 3.4% and a total average score of .965

As a consideration here, if this were close to a typical cross section of the text, I would expect to see the top 25 pages closer to a 97-98% common vocabulary with an average total score in the 1.025-1.035 range.

Very interesting -- In fact, very, very interesting.

This is the first approximation we have of anything like "control data,"
by which to try an assess just how unique Spalding's shared language
with other texts (namely the Book of Mormon) might be.

I appreciate your taking the time to conduct the experiment.

My suspicions are that the word-string correspondence will increase
significantly, if you continue your testing through the end of Spalding's
"Roman story." Very likely the military terminology found in the last
dozen or so pages of Spalding will be replete with parallels with
similar narrative in Warren.

The drop-off into the 80s% for shared vocabulary surprises me.
I wonder just how "low" the "lowest" Spalding page might be?

Obviously exploration, military and religious narratives of the early
19th century will share many two-word strings, quite a few
three-word strings, and an occasional four-word string. I would
suppose that narratives written in English might tend to share
a bit more phraseology than those written in another language
and then translated into English.

Since Josephus works fall into the latter category, they might be
a little less like Spalding than is Warren.

Then again, Spalding and Josephus would have been thinking
about fighting and social interactions before the inventions of
gunpowder, modern communications, etc. -- so they might have
written about more technologically similar incidents than Warren,
and thus may share some "antique" phraseology.

I wish we could do something similar for a dozen or so books
from that era -- and illustrate the results by mapping the
numbers across charts for all of the books' pages.

Ryan Larsen is going to want to consult your experiment,
before he carries out something similar.

Does your software run on a windows-platform PC?

UD

[_Uncle Dale]

...even higher than the averages in Alma...

There is a reason that I broke Alma into two sections --
and that is, the last 1/3 of Alma is where nearly all the
"hot spot" clusters of language shared with Spalding occurs.

When we add my two sections back together, and look at
the Book of Alma as a whole, the parallels will Spalding
will be considerably less than in the second section, if
viewed by itself.

In other words, while measuring a "full context" can be
important in work like this, it will often be the sub-contexts
which are most interesting.

It is entirely possible that many early 19th century texts will
provide higher scores for shared vocabulary and shared
phraseology with Alma (as a whole) than does Spalding.

But, of course, it is the latter part of Alma that most interests me;
and that is the span of text I'd like to see compared with other
narratives than Spalding's.

UD

[_Uncle Dale]

...
I wish we could do something similar for a dozen or so books
from that era -- and illustrate the results by mapping the
numbers across charts for all of the books' pages.

Ryan Larsen is going to want to consult your experiment,
before he carries out something similar.

Does your software run on a windows-platform PC?
...

I worked up a rough chart of the 15 Spalding MS pages you tabulated.
The top set of numbers are the 1910 LDS edition's page numbers. The
bottom set of numbers are the actual Oberlin MS page numbers.



Dale

[_Uncle Dale]

...
a rough chart of the 15 Spalding MS pages
...

I could, of course, go ahead and draw up the entire chart,
covering the full set of Oberlin MS pages (less the Hurlbut affidavit),
if that would be of any use.

With your automated computer program, could you not develop
the data for matching the 1830 Book of Mormon text directly with the
Spalding text, in a matter of a few hours? We could then fill
in the Spalding chart, to see if it had any noticeable distribution
patterns of shared language with the Book of Mormon (or with Warren, etc.)

???

Dale

[_Roger]

The numbers and ratios portion of this conversation goes over my head, but--and this may be a stupid question--what exactly does Ben think his numbers demonstrate? If his conclusion is that one can run similar tests on contemporary texts--with similar subjects but otherwise unrelated--and come up with similar numbers, then Warren may not be the best example because it is theorized that Spalding plagiarized--or at least made use of--Warren in production of his own RS & MF. So I'm not sure how much value Ben's current set of numbers is--at least to his position. Rather, unless I am missing something, they seem to support Tom Donofrio's suggestion of Spalding's dependence on Warren.

Also, while I think there is certainly value in all this hyper-analysis & comparisons, I think it is also important to consider the picture from a broader perspective... at least to the layman, these numbers seem to be somewhat analogous to examining tree bark in the proverbial "seeing the forest vs the trees" notion. On the "forest" level, I think Holley's observations have relevance here as well in terms of tying it all together. For example, Holley states:

The description of the battle at the hill Riplah between the Nephites and the Lamanites in the Book of Mormon is very similar to that of a battle between the Sciotans and the Kentucks in Manuscript Story.

"By using a little stratagem," Spaulding's Sciotans "marched down the river to a certain place where the army of the enemy must pass... At this place, the hill... came within less than a mile of the river." The Sciotans divided their forces and ambushed the enemy as they crossed the canal. Having compassion for the trapped enemy, Lobaska "conjured the Sciotans not to shed one drop of blood" (Manuscript Story pp. 37-39).

The Book of Mormon's Moroni [made the same defense] "by stratagem." After discovering which course the enemy would take, he divided his army, concealed them by a hill, then ambushed the Lamanites as they were crossing the river. After subduing the Lamanites, Moroni said: "We do not desire to be men of blood... we do not desire to slay you" (Alma 43:30-36; 44:1).

In both stories, the losing army is more than double the number of the winning army, but the warriors are so struck with "terror" by the ambush that they "throw down" their arms and surrender. Then, in both stories, the losing commander has a personal conversation with the opposing commander, asking him to "spare their lives," after which a treaty of peace is made and the losing warriors return to their own country (Manuscript Story pp. 38-39; cf. Alma 43:51, 44:15, 19). Warriors [among] the opposing nations in both stories were related -- being offspring of the same family -- and called each other "brother."

http://www.solomonspalding.com/docs2/vernP1.htm#pg05

The similarity of concept here is obvious and I think such similarities are rather difficult for computers, charts, ratios, etc, to quantify. The only relevant "clue" for a computer analysis that I can see is found in Holley's observation that:

"by stratagem."

in the Book of Mormon and...

"By using a little stratagem"

in Spalding...

...are obviously similar. But I wonder if the broader contextual similarities involved in this type of parallel are picked up by either Dale's or Ben's methods? And if so, what weight are they given? Just another statistic within a larger set of numbers and ratios? I think it is important to consider the broader concept--the forest if you will--in addition to close examination of a lot of individual trees.

Also, the above is merely one example among many broad parallels that Holley discovered between RS and the Book of Mormon.

Holley's observations lend support to Dale's shared phraseology observations which lend support to Jocker's word prints, all of which harmonize with a lot of 19th century testimony.

If the Spalding-Rigdon theory was built on nothing more than false memories and a satanically inspired motivation to bring Smith & his church down, then one would not expect to see word print studies and shared-vocabulary tests supporting it by mere coincidence.

[_Uncle Dale]

...what exactly does Ben think his numbers demonstrate?
...

I'll let him answer that one. I'm more interested in his ability to generate
the numbers. If I had such a software resource it would be wonderful!!

Let's just hope that Ms. Warren didn't write the Oberlin manuscript and
leave her "word-prints" all over its text. If so, then I'll have to correlate
Ben's shared language data with somebody else's word-print data, all
over again, from scratch...

That would be a major headache.

UD

[_Benjamin McGuire]

Sorry, Roger, I am simply not interested in having this discussion with you again.

Ben