Two voices in the Book of Mormon

[Bruce Schaalje]

If you assume a naturalistic view of the origins of the Book of Mormon, as we did, you are left with three theories for the Book of Mormon:

Theory 1. Smith composed it by himself.

Theory 2. Smith had help.

Theory 3. Someone else wrote it. Smith was just a front man.

Craig,

I think you know my view of the origin or the Book of Mormon. But I suppose that you

[MCB]

I also advocate adding Lucy Mack Smith, and others. The initial study simply included the most likely candidates, and controls.

Be patient. I would not be surprised if additional research includes others. You can't expect immediate perfection.

[consiglieri]

Although perhaps out of place in this somewhat technical discussion of word print studies, it was mentioned before that King Benjamin is the only voice in the Book of Mormon that uses the term Lord God Omnipotent; a fact I think was pointed out first by John Welch.

On a similar issue, Anti-Nephi-Lehi also has a unique title that he, and only he, uses for God, as described in the first part of a piece I wrote a few years back. For what it's worth . . .

The

[Uncle Dale]

...

You can't expect immediate perfection.

I'm reminded of the alchemists' search for the philosopher's stone

or for the elixir of life. "What was the proper formula?" they asked.

At least modern computers can provide new possibilities for searching

out literary needles in literary haystacks.

If 1st Nephi 1:12 was not actually written by Nephi, then

what is the proper formula for authorship?

Take four parts Solomon Spalding and add two parts Sidney Rigdon --

Shake well and heat to the temperature of Lucy Smith divided by

W. W. Phelps. Distill off the Oliver Cowdery and filter through

Parley Pratt. Serve with garnishes of Joseph Smith, Jr.

I really, really hope it is NOT that complex.

UD

[MCB]

At least modern computers can provide new possibilities for searching

out literary needles in literary haystacks.

I really, really hope it is NOT that complex.

UD

I hope so, too. However, we need to look at all the possibilities, in order to ease criticism of the area of study. "Testing the null hypothesis," if you will. Listening to the arguments of critics is critical in objective research.

[Benjamin McGuire]

If you read back in this thread, you'll see that I recommended that Bruce transfer his x,y measurements to a bar graph incorporating each BoM chapter in order -- with the x component depicted as color and the y component depicted as bar height. Once that graphic is available, we will have at our disposal yet another rough view of the "texture" patterns discernible across the Book of Mormon.

Either I am not following what it is exactly you want to see represented in this way, or (if I do), I don't think it will be all that meaningful.

But, you'll probably understand why I cannot accept that same isolated chart as constituting preemptive proof that the Book of Mormon could not contain textual contributions from Spalding, Rigdon, Smith, etc.

But I think that this would not be unreasonable. If we accept on some level that this kind of study is capable of telling us which of a set of authors wrote a specific text, then it also ought to be able to tell us which authors did not write a specific text. It seems quite capable of doing this thing exactly if we look at how it is used with the Federalist Papers. I think in this case, we run into this problem with individuals who are hoping that the question of stylometry can answer the question of authorship here. If it shows authorship, there is this clamor of certainty. But, when it seems that this study has in fact produced evidence against the very notions that it set out to prove, there is this suggestion that perhaps we just need to massage the data a bit more.

Pardon me for being skeptical - but, I am still left wondering about this one issue. If you believe that this kind of analysis can eventually show a connection between the unknown author of the Book of Mormon and either Rigdon or Spaulding, at what point would you be willing to admit that this kind of analysis can show that it is quite unlikely that either Spaulding or Rigdon contributed? Or is this an open ended question in one direction only?

Ben McGuire

[Chris Smith]

What about the Whitmers, the Knights, other Cowderys,W W Phelps, Martin and Emer Harris, Ethan Smith, Grandin, E T A Hoffman, and Mordecai Noah (see here)?

Not to mention Alvin, Hyrum, Joseph Sr., Lucy...

[Uncle Dale]

Either I am not following what it is exactly you want to see represented in this way, or (if I do),

I don't think it will be all that meaningful.

OK, then let me ask you a relevant question:

As you consult Bruce's chart, what does it say

about the measured comparison of Alma 51, 52 and 53?

In a bar chart of the Book of Mormon chapter, lined up

in order, would we see those three chapters depicted

at about the same bar height, displaying roughly the

same color?

Or, would those three chapters, as characterized by

Bruce's measurements, show radically different heights

and colors? Would the entire Book of Mormon, so graphed,

exhibit no coherent distribution patterns?

Bruce has already told us that the Isaiah chapters of

the BoM cluster over to the left, with the 19th century

authors. So, even if all the rest of the graphed BoM

chapters exhibited only randomness, we should expect

to see a set of Isaiah chapters within the BoM with

similar coloration and similar bar height. Correct?

But I think that this would not be unreasonable. If we accept on some level that

this kind of study is capable of telling us which of a set of authors wrote a

specific text, then it also ought to be able to tell us which authors did not

write a specific text. It seems quite capable of doing this thing exactly if we

look at how it is used with the Federalist Papers.

Which of the Federalist Papers had joint-authorship?

Which of the Federalist Papers had editorial redactions?

I think in this case, we run into this problem with individuals who are hoping

that the question of stylometry can answer the question of authorship here.

I think such results can provide some of the literary "layers"

in my proposed set of quantified "overlays" in graphing the BoM.

If it shows authorship, there is this clamor of certainty. But, when it

seems that this study has in fact produced evidence against the very

notions that it set out to prove, there is this suggestion that perhaps

we just need to massage the data a bit more.

I think that would be dangerous -- unless there was some specific,

agreed upon reason for so doing. Say, perhaps removing Isaiah-Malachi

from the data set, too see what new results would thus be provided.

Pardon me for being skeptical - but, I am still left wondering about this

one issue. If you believe that this kind of analysis can eventually show

a connection between the unknown author of the Book of Mormon and either

Rigdon or Spaulding, at what point would you be willing to admit that this

kind of analysis can show that it is quite unlikely that either Spaulding

or Rigdon contributed? Or is this an open ended question in one direction only?

...

I do not know how to define any such "point," but my view of things

would probably be impacted by the publication of a dozen papers on

this subject in the non-LDS professional literature, which began to

consolidate a consensus conclusion that Spalding and Rigdon are very

unlikely candidates for having authored any page in the book.

As a Reorganized Saint I am free to follow the evidence, to see

where it leads and to see what can be predicted in the way of

future discoveries, if my conclusions are correct. As a Reorganized

Saint I do not need to consider possible religious exclusion if I

voice the conclusion that "the Book of Mormon is NOT true;" or,

even the conclusion that "the Book of Mormon IS true" -- (although

the latter expression would not get me far within CoC these days).

I am not interested in pressing the S-R authorship explanation to

the detriment of truth --- I am interested in seeking out the truth,

even if it proves historical Nephites.

In the meanwhile, why not let somebody translate Bruce's pca chart

into a bar graph of the entire Book of Mormon? What would it hurt?

UD

.

[MCB]

LOL

I am not interested in pressing the S-R authorship explanation to

the detriment of truth --- I am interested in seeking out the truth,

even if it proves historical Nephites.

Got it!! We have always been RLDS-friendly, as you already know.

[Glenn101]

I'm reminded of the alchemists' search for the philosopher's stone

or for the elixir of life. "What was the proper formula?" they asked.

At least modern computers can provide new possibilities for searching

out literary needles in literary haystacks.

If 1st Nephi 1:12 was not actually written by Nephi, then

what is the proper formula for authorship?

Take four parts Solomon Spalding and add two parts Sidney Rigdon --

Shake well and heat to the temperature of Lucy Smith divided by

W. W. Phelps. Distill off the Oliver Cowdery and filter through

Parley Pratt. Serve with garnishes of Joseph Smith, Jr.

I really, really hope it is NOT that complex.

UD

Modern computers are no better than the algorithms they are programmed with and the data those programs operate on. I am not saying "GIGO", or even lies, damned lies, and statistics. But some are superior to others. The simple solution is one that accepts the Book of Mormon on face value, something that the dyed in the wool critics are opposed to doing. I do not think that the Jockers study can really answer the authorship question on the Book of Mormon in its present guise because it cannot positively eliminate possible authors, (if I understand it correctly), but can only give relative probabilities of the selected set of possible authors.

Also, as you have aleady noted, the parts that should theoretically match Spaulding the closest do not. Kevin Christensen in another post presented a laundry list of problems that any authorship study needs to overcome for any nineteenth century to be deemed credible. The more nineteenth century authors posited as contributors to the Book of Mormon text only magnifies the complexity and the difficulty in producing any coherent means of ascertaining who did what.

Glenn

[Benjamin McGuire]

With regards to the bar chart, Dale, I think that what we would have is more of a three dimensional object (you attempt to deal with the third dimension by using a color shade). I would rather just see the plot with chapters identified - that would probably be sufficient for what you want to see, don't you think? Obviously, that would allow us to look in this sense at the clustering of the different BoM chapters.

Bruce has already told us that the Isaiah chapters of the BoM cluster over to the left, with the 19th century authors. So, even if all the rest of the graphed BoM chapters exhibited only randomness, we should expect to see a set of Isaiah chapters within the BoM with similar coloration and similar bar height. Correct?

But I already see this in the existing chart. We actually have plotted points for the Isaiah/Malachi author. These line up rather well with some of the Book of Mormon chapters. The conclusion is (without identifying the chapters in the plot) that they represent the overlapping material. Of course our seeing this is because we have the other material in the plot.

Which of the Federalist Papers had joint-authorship?

Which of the Federalist Papers had editorial redactions?

And the answer, of course, is that none of them really did. But that is not the point I am making. The point I am making is that the Jockers study is suggesting that it ought to be possible using their method to say something about authorship. Part of that something that ought to be sayable then is that the author of the material is not in the list of candidates. This is clear even if the real author is not present. That is to say, if our best candidate for authorship for a piece of text ends up being Spaulding with a very high likelihood, then those authors with very small likelihoods are really unlikely to be an appropriate candidate for authorship. At some point, we can extend this idea to suggest that none of the candidates are likely (even if there is a "most likely" candidate). The Jockers study did not provide this kind of mechanism - but that is part of the core of this discussion here. You want to talk about redactions, about editing, but on some level, this only becomes an issue if we are at a point where it might matter. As Bruce has pointed out, even with all of that, the Spaulding material is statistically so distant from the Book of Mormon that such distinctions seem unnecessary. At this point you still have to come to some kind of conclusion. It seems to me that either A) the method the Jockers study uses isn't able to tell us anything about the text of the Book of Mormon given the fact that it seems to reveal pretty clearly an issue of multiple authorship, and with multiple authorship, the issue becomes too ambiguous to really make a good deal of sense of it in this fashion. Or, B) the Jockers study is capable of telling us something specific about the authorship of a particular chapter - with reference to specific authors - and it can tell us more about who didn't write it than perhaps who did. Yes there are probably other options, but I suspect that they will largely be variations on a theme.

I think such results can provide some of the literary "layers" in my proposed set of quantified "overlays" in graphing the BoM.

Yes, but I think that your sense of "overlays" is overly ambiguous. In a sense, it seems to me that I have yet to see you provide any system of modeling that argues even a little bit conclusively for a Spaulding authorship. Why should I think that "layering" numerous studies that don't indicate Spaulding as a contributor will somehow magically coalesce into this patter which now argues for a Spaulding contribution?

At any rate, I think that you are right in asking for a distribution which identifies chapters. Since the chart Bruce provided shows a number of potential clusters (and thus potential authors), it would be useful to see how the chapters themselves cluster and so whether we obtain a graph of the sort you ask for, or simply a repeat of the same chart Bruce has already provided with chapter information indicated, it would be about the same in my estimation, but it would be very helpful in looking at what information this data can provide us. If the clusters reveal largely randomness, or if they reveal useful contextual clustering, it would be interesting then to start talking about the text in those terms.

Ben McGuire

[Chris Smith]

Craig's comment that the high false positive rate for Isaiah/Malachi may be due to the BoM's imitation of the KJV is telling, to my mind. It suggests that the method is sensitive to context and genre, and that it is possible to "trick" the method by imitating someone's style. I suspect that if all the individual biblical authors were included as controls (rather than just the Isaiah/Malachi composite), then none of the Book of Mormon would remain attributed to 19th century persons.

[Mola Ram Suda Ram]

Craig's comment that the high false positive rate for Isaiah/Malachi may be due to the BoM's imitation of the KJV is telling, to my mind. It suggests that the method is sensitive to context and genre, and that it is possible to "trick" the method by imitating someone's style. I suspect that if all the individual biblical authors were included as controls (rather than just the Isaiah/Malachi composite), then none of the Book of Mormon would remain attributed to 19th century persons.

That tricky JS. How did he know to write this book in such a manner as to fool the system.

Chris, I love you man.

[Uncle Dale]

With regards to the bar chart, Dale, I think that what we would have is more of a three dimensional object (you attempt to deal with the third dimension by using a color shade). I would rather just see the plot with chapters identified - that would probably be sufficient for what you want to see, don't you think?

How do I overlay THAT with a bar chart measuring the correctness of KJV

English subject/object agreement throughout the BoM? Or with a bar chart

measuring average numbers of words separating occurrences of gerunds and

infinitives? Or with a bar chart measuring percentage of overlap with

the known literary vocabulary of Joseph Smith, Jr.?

Unless these sorts of data sets can be displayed upon a common ground,

I cannot search out patterns of inter-relationships between them. Even

if we only make use of pca charts, we must find a way to lable the

data points and to correlate them with locations in the BoM -- so that

other pca charts can be used as the "overlays." I do not know how to

do that -- but I'm guessing some interesting connections will show up.

What if, for example, the vast majority of faulty examples of KJV English

correlate strongly with the distribution of Smith's vocabulary in the BoM?

Perhaps you will argue that data charted by pca does not lend itself

well to such cross-examination -- but I still want to see for myself.

If the Mormons are uninterested in doing that work, then I'll bug

Jockers to do it for us.

Obviously, that would allow us to look in this sense at the clustering of

the different BoM chapters. But I already see this in the existing chart.

I'm not so much interested the fact of clusters, as I am in the inner

dynamics of those clusters. Does Moroni's editing of Ether "taper off"

in certain spots, or is it abrupt? How different are Moroni and Ether,

when compared to runs of consistency in sections of Alma? etc. etc.

We actually have plotted points for the Isaiah/Malachi author. These line

up rather well with some of the Book of Mormon chapters. The conclusion is

(without identifying the chapters in the plot) that they represent the

overlapping material. Of course our seeing this is because we have the

other material in the plot.

That's nice -- would somebody please post an enlargement of that part of

the pca chart? My eyesight isn't what it used to be, you know.

And the answer, of course, is that none of them really did. But that is not

the point I am making. The point I am making is that the Jockers study is

suggesting that it ought to be possible using their method to say something

about authorship. Part of that something that ought to be sayable then is

that the author of the material is not in the list of candidates. This is

clear even if the real author is not present. That is to say, if our best

candidate for authorship for a piece of text ends up being Spaulding with a

very high likelihood, then those authors with very small likelihoods are

really unlikely to be an appropriate candidate for authorship.

Yes, that is what he is interested in.

I'm interested in the structure of the Book of Mormon itself.

Jockers shines a light into a dark room looking for Spalding.

I'm happy to see whatever else that light happens to illumine.

At some point, we can extend this idea to suggest that none of the candidates

are likely (even if there is a "most likely" candidate). The Jockers study

did not provide this kind of mechanism - but that is part of the core of this

discussion here.

You want to talk about redactions, about editing, but on some level, this only

becomes an issue if we are at a point where it might matter. As Bruce has

pointed out, even with all of that, the Spaulding material is statistically

so distant from the Book of Mormon that such distinctions seem unnecessary.

Allow me to go back to my imagery of the Great Salt Lake basin, from 1976.

One of the variables measured was surface temperature -- mapped out for a

large area in and around the water. A person could have examined that info,

trying to discern water depth, and come away with the notion that the only

really deep spot in the lake was where the Bear River emptied its cool flow

into the water. I could say that there are spots twenty feet deep in the lake,

and somebody else could tell me that there is no indication of any such thing

in the temperature distribution map. I would need to pull out the sonar map,

in order to provide a different view of the same lake -- one showing depth.

I am not convinced that the one, unlabeled pca chart we have looked at

conclusively demonstrates that Solomon Spalding could not have written

half of a certain chapter, with Oliver Cowdery adding the remainder. It

would be foolish of me to cease my examination, just because of what I've

heard argued in this mb thread.

At this point you still have to come to some kind of conclusion. It seems

to me that either A) the method the Jockers study uses isn't able to tell

us anything about the text of the Book of Mormon given the fact that it

seems to reveal pretty clearly an issue of multiple authorship, and with

multiple authorship, the issue becomes too ambiguous to really make a

good deal of sense of it in this fashion. Or, B) the Jockers study is

capable of telling us something specific about the authorship of a

particular chapter - with reference to specific authors - and it can tell

us more about who didn't write it than perhaps who did. Yes there are

probably other options, but I suspect that they will largely be variations

on a theme.

What I see is that Jockers has provided a map of the distribution of

"something" across the text of the BoM. The lack of randomness in that

map indicates to me that some underlying factors are creating the

variations in distributions. I say we should apply some other methods

to examining the same text, to see what other sorts of variations pop

up in the results.

What if, for example, methods x, y, and z all result in showing a major

discontinuity at the Nephi/Mormon juncture and at the Isaiah junctures?

Would you tend to dismiss such a discovery as totally uninteresting?

Yes, but I think that your sense of "overlays" is overly ambiguous.

In a sense, it seems to me that I have yet to see you provide any system

of modeling that argues even a little bit conclusively for a Spaulding

authorship. Why should I think that "layering" numerous studies that

don't indicate Spaulding as a contributor will somehow magically coalesce

into this patter which now argues for a Spaulding contribution?

You always seem to think of me as a Spalding enthusiast first and as

a Latter Day Saint in some distant last place --- My interests are

in determining the truth. I was taught that in Junior Sunday School,

before I ever heard of any Book of Mormon controversies. That remains

my beginning point.

At any rate, I think that you are right in asking for a distribution which

identifies chapters. Since the chart Bruce provided shows a number of

potential clusters (and thus potential authors), it would be useful to see

how the chapters themselves cluster and so whether we obtain a graph of

the sort you ask for, or simply a repeat of the same chart Bruce has

already provided with chapter information indicated, it would be about the

same in my estimation, but it would be very helpful in looking at what

information this data can provide us. If the clusters reveal largely randomness,

or if they reveal useful contextual clustering, it would be interesting

then to start talking about the text in those terms.

...

It will be very interesting if a consolidation/amalgamation of textual study

"overlays" agree that the Book has a single author, whose "voice" corresponds

closely with that of Joseph Smith Jr. (or with Smith's uncanny ability to

consistently mimic complex sets of frequently used, non-contextual words,

for a set of different authors).

I'll guess the former possibility at about .000001% and the latter possibility

at about .1% (just to hold the interest here of Don and Chris).

UD

.

[Nemesis]

We have gotten requests to move this to the pundits section and give access to those that are in this discussion. However I do not want to unintentionally exclude someone so I will keep it here. If we get derailments or posters that do not behave we will consider moving it then. Please report problems if they come up so we can remove the offending party.

Nemesis

[Vance]

. . . , then none of the Book of Mormon would remain attributed to 19th century persons.

An interesting statement.

[Chris Smith]

How did he know to write this book in such a manner as to fool the system.

Imitating the style of the KJV isn't that hard to do. The Brennan paper I cited earlier in this thread showed that even a person who has no awareness of statistical authorship attribution methods can "trick" those methods with a fairly high rate of success by imitating someone else's style.

[Chris Smith]

An interesting statement.

Don't get too excited. Since the authors of the KJV aren't really candidates for authorship of most of the Book of Mormon, for the method to attribute the Book to those authors would only indicate the bankruptcy of the method.

[Mola Ram Suda Ram]

Imitating the style of the KJV isn't that hard to do. The Brennan paper I cited earlier in this thread showed that even a person who has no awareness of statistical authorship attribution methods can "trick" those methods with a fairly high rate of success by imitating someone else's style.

I hope that you do realise that you are making the BoM story (how it came to be) alot more complicated. Because as I have followed this thread it appears that they took out the Isaiah/Malachi chapters when they did their studies. We are trying to incorprate one text to fit into another text all with a flowing continual complicated story.

Were is Bill Hamblin? He is alsways so good at being critical of the critics theories.

At the risk of derail though I think I will remain out of this thread. I would hate to complicate such a good thread.

[Uncertain]

Hi All,

Well with the departure of Craig there are not to many defenders of the Jockers paper left in the thread so it may have just about run its course. But I can always pick a fight with someone . I am not fully convinced the principle component analysis completely rules out Rigdon authorship of at least some of the BOM. Don't misunderstand me I am not claiming Jockers et al were right I am just not sure they are wrong on the basis of the PCA plot. I think they are wrong for other reasons . My concerns with the PCA plot are two fold.

1. As I already explained in an earlier post looking at the pca plot the Spaulding cluster does not look all that different than say the Pratt cluster and yet Spaulding is selected far more often as a potential BOM author than Pratt. For that matter the Cowdery cluster to me looks closer to the BOM cloud than the Spaulding cluster and again Spaulding is selected far more often than Cowdery. This suggests at least to me the principle component analysis is not capturing some information used by NSC. Which in turn suggests using the pca plot to discredit the NSC results may be a bit premature.

2. How exactly the pca analysis statistically rules out Rigdon/Spaulding authorship. Using the eyeball test certainly the BOM cluster(s) look far away from the Rigdon/Spaulding cluster but this is somewhat subjective. It would be nice to quantify what is meant by "distance" and how this metric can be used in a rigorous fashion to rule out Rigdon/Spaulding authorship. Ben in an earlier post suggested we look at the furthest datapoint from the center of the Rigdon cluster call this distance d and see how many BOM chapters fall within the circle centered at the middle of the Rigdon cluster with a radius based on some multiple of the distance d. Well just looking it seems to me if I go say twice d at least some BOM chapters fall within this radius are all of them Isaiah-Malachi chapters? Since the chapters are not labeled it is difficult to tell.

All the Best,

Uncertain

[MCB]

Thats what I told them already. If that is what they want to do--- I can think of more productive things to do, but that is just me.

[Chris Smith]

I hope that you do realise that you are making the BoM story (how it came to be) alot more complicated.

How so?

[Chris Smith]

I am not fully convinced the principle component analysis completely rules out Rigdon authorship of at least some of the BOM.

I agree. The problem for me is that Bruce's PCA plot relies on the same data Jockers, et. al. used, which is biased by factors such as genre, time, and context. So texts of a different genre might appear very distant from each other even if they had the same author. Now, the fact that the Rigdon sample did include some revelations suggests that genre-bias will affect his plot the least as compared to the BoM. But the revelations were written much later in his life, so who knows how his style might have changed in the interim?

[Doctor Steuss]

Uncertain and/or Chris,

I apologize for my ignorance, but

[Uncertain]

I agree. The problem for me is that Bruce's PCA plot relies on the same data Jockers, et. al. used, which is biased by factors such as genre, time, and context. So texts of a different genre might appear very distant from each other even if they had the same author. Now, the fact that the Rigdon sample did include some revelations suggests that genre-bias will affect his plot the least as compared to the BoM. But the revelations were written much later in his life, so who knows how his style might have changed in the interim?

This is a good point.