Intro/FAQ ⎜ Episode 1 ⎜ Episode 2 ⎜ Episode 3 ⎜ Episode 4 ⎜ Episode 5 ⎜ Episode 6 ⎜ Episode 7 ⎜ Episode 8 ⎜ Episode 9 ⎜ Episode 10 ⎜ Episode 11 ⎜ Episode 12 ⎜ Episode 13 ⎜ Episode 14 ⎜ Episode 15 ⎜ Episode 16 ⎜ Episode 17 ⎜ Episode 18 ⎜ Episode 19 ⎜ Episode 20 ⎜ Episode 21 ⎜ Episode 22 ⎜ Episode 23
[Editor’s Note: This is the third in a series of 23 essays summarizing and evaluating Book of Mormon-related evidence from a Bayesian statistical perspective. See the FAQ at the end of the introductory episode for details on methodology.]
The TLDR
It seems unlikely that Joseph Smith could write the Book of Mormon through the dictation process described by witnesses, or that he could have written it in some other way without leaving a trail of evidence.
Many critics seem convinced that Joseph could dictate the Book of Mormon without much trouble, and those who recognize the difficulty of the dictation process rely on him authoring the manuscript in secret—without leaving any evidence of that process. It’s difficult to know how unlikely those scenarios are, but estimates that favor the critics still suggest that it wasn’t likely. I estimate the probability of him dictating a coherent narrative, without in-line revision or prompts at the start of a writing session, to be at most p = .000283. Furthermore, the probability of him being able to conduct basic preparations without being caught should be at most p = .0297. I also estimate the probability of Joseph performing the startling feats of memory associated with the dictation at p = .000283. Even taking into account the unlikely theory of “automatic writing,” the available dictation evidence weighs heavily in favor of authenticity.
Evidence Score = 5 (increases the probability of authenticity by five orders of magnitude)
The Narrative
When last we left you, our eminent skeptic, you were watching the final embers die in the fire of your early nineteenth-century hearth, wisps of smoke fizzling up and out the chimney into the New England air. You still grasp the spine of a thick book in your hand. You know that you should probably find your bed at this late hour, but a dozen questions continue to spin in your mind, and you point them toward the young man on the other side of the room.
“This volume does seem impressive, to be sure. But how can you know that its anything but a man’s work? If this Smith wrote down the words of his ‘translation,’ what would make them God’s words and not his?”
“Well, that’s the curious thing,” the young man says. “He didn’t write the words down, as you might expect—he spoke the words to a scribe as they appeared to him, as if it were in vision.”
You chuckle softly. “Yet more visions. He saw the whole book in his mind’s eye, I suppose, transported by the spirit into the high heavens and reading from a heavenly book?”
The man’s head gently shakes. “Stranger still than that, if you’ll believe it. Joseph himself only ever said that the translation was done by the power of God, but if you speak with those who were with him when the translation was being done—his scribes, his wife, those boarding him—they all tell a similar story. With the plates Joseph also was provided a set of interpreters, clear stones that he could look through, set in a breastplate—when he did so he could see words written, in the King’s English, about 20 at a time, which he would then dictate to a scribe. After the scribe had recorded the words, the words would disappear from the interpreters and a new set would appear.
“I hear tell that he later found another stone in the river that would accomplish the same purpose, but the pattern was the same. He would stick his head in a hat to block out the light and dictate the words that appeared before him. He would go on this way for hour after hour, dictating the whole of the manuscript, his own fingers never touching pen and ink, no notes in front of him, and no revisions. He would dictate entire chapters of biblical verse almost word for word, with no bible to refer to. When he would begin again each morning he would continue to dictate exactly where he left off without having to be told where he was. It was truly extraordinary.”
Your chuckle returns and quickly grows into laughter. “You’re right to call it strange. You should have your head examined to say such silliness with a straight face.”
He turns his eyes to yours. “You can mock if you like, but it’s not that simple. It’s not just his friends saying this silliness, it’s some who despised and hated him, and those who later turned on him. I’ve seen the manuscript that his scribes wrote as the words fell from his lips—if you examine it closely, you can see where the scribes would have to pause as Joseph received a new set of words, and if his scribes ever made errors, they’re errors of mishearing words rather than mis-seeing them. There’s nothing to contradict Joseph’s account, and much to support it.
You see that the young man’s face is filled with sincerity, and you pause a moment before continuing. “If what you’re saying is true, it does seem unlikely that someone could dictate an entire book like this one," you say, hefting the book in your hand, “or, at least, that there would be no evidence that he did otherwise."
The Introduction
It’s very common for critics of the church to poke fun at the dictation process. Refrains of “rock in a hat!” can be heard echoing through the entirety of the ex-mormon world. But these same critics don’t often appreciate that the accounts of Joseph and his scribes are very consistent with the available primary evidence. There’s a general inability to explain this evidence outside of appeals to Joseph’s alleged genius, to large-scale conspiracies, or to the near-supernatural. We’ll be covering the “large-scale conspiracy” angle when we talk about the Book of Mormon witnesses a couple episodes down the road, but until then the dictation process provides plenty for Bayes to chew on.
This analysis will have to be a little looser than any of the others so far—there’s very little solid research on the mechanics of large-scale dictation projects. Nevertheless, we’re going to try to think the problem through as clearly as we can.
The Analysis
This analysis is going to be a little different from the others—we’re going to think through three different hypotheses instead of the usual two—but I’ll try my best to keep it simple.
The Evidence
When it comes to the dictation of the Book of Mormon, there are a few kinds of evidence that can inform our opinion. The first are the actual witnesses of the dictation process—those who served as his scribes, and those who were present in the households where he conducted the work of translation. These are the people we’re familiar with, such as Oliver, Emma, Martin Harris, but also many we don’t hear about, such as Reuben Hale, Michael Bartlett Morse (Emma’s brother-in-law, who was no friend of Joseph’s or the church), and various members of the Whitmer and Knight families. All of them saw the translation firsthand, and though they differ in some of the details, the core of the dictation process remains consistent in each of their accounts. Royal Skousen aptly summarizes the distinguishing features of that process, but we’re only going to concern ourselves with a few points:
- The vast majority of the Book of Mormon was dictated openly, with Joseph in full view of his scribes and others nearby.
- Joseph dictated for extended periods of time, including long stretches of nearly word-for-word biblical quotations, without notes or reference material. (And, no, he wouldn’t have been able to hide them in the hat!)
- When starting a dictation session, Joseph would often continue without being reminded of what had been dictated at the end of the previous session.
- After dictating a section of text, Joseph would have the scribe read back what was written to check its accuracy. Scribes occasionally missed, added, or misheard words, but there’s no indication that Joseph redacted or revised the text as it was being dictated.
Another piece of evidence is the original manuscript of the Book of Mormon, which corroborates these key details whenever possible. Aural errors (rather than visual) can be found throughout, clearly indicating that the text was dictated rather than copied from another manuscript, and the pages themselves are remarkably clean—Oliver or the other scribes didn’t have to go back and revise key portions of the text or alter word-choices as you might expect from someone drafting the text as they went along.
We also can keep in mind the kind of book that Joseph is alleged to have produced. The Book of Mormon is not just any old narrative. It has a number of features that would have greatly increased the technical difficulty of any proposed dictation. These features include, but are not limited to, the following:
- Extended word-for-word quotations from the King James Version of the bible. (The nature of which are interesting enough to deserve their own episode, later on.)
- Comprehensive internal allusions and quotations.
- A complex internal geography with a scale to rival Tolkien’s Middle Earth. (We’ll also have a separate episode dedicated to the nature and consistency of the book’s geography.)
All told these details and features certainly do not match what we’d expect from a dictation. So how might we have ended up with the evidence as we now have it?
The Hypotheses
Absent an airtight conspiracy, there seem to be a limited number of ways that the evidence we have could have been produced:
Joseph’s account of dictation of the Book of Mormon is authentic. In short, this hypothesis posits that Joseph really did use a set of stones as a medium to see and then dictate the words of the Book of Mormon as translated from the plates.
Joseph’s dictation is a product of “automatic writing.” We also consider another, less-mainstream explanation, that Joseph tapped into the near-superhuman powers of the unconscious mind to dictate a draft in an incredible stream-of-consciousness manner.
Joseph fraudulently dictated the Book of Mormon via a conscious writing process. Even assuming that Joseph was a literary prodigy, this hypothesis would require Joseph to spend the time necessary to plot, outline, and otherwise write the Book of Mormon, and to learn it well enough to be able to dictate a surprisingly clean manuscript in a single draft absent of in-line revisions (by Joseph). There are a number of ways that this sort of process could have worked, and we’ll be considering those in turn as we go along.
Prior Probabilities
With three hypotheses to work with, our estimates of prior probabilities will work a little differently than in previous episodes. We’ll continue to assume that Joseph was a fraud, which means that we’ll estimate the probability of the less conventional explanations first (authentic dictation and automatic writing) and then whatever probability remains will belong to Joseph producing the dictation via a conscious writing process.
PH—Prior Probability of an Authentic Dictation—We get this probability from the result of our previous analysis. Our skeptic did adjust the probability of an authentic Book of Mormon downward on the basis of the consistency of the First Vision, but not by much. Our current estimate, based on what we’ve considered up to this point, is p = 8.57 x 10-40. Here’s a figure showing us where we’ve been thus far:
PA1—Prior Probability of Automatic Writing—Given the unusual nature and extreme rarity of instances of automatic writing, our prior probability that it would serve as a valid explanation of a given piece of writing should be quite low. It’s somewhat more believable on its face than angels and seer stones, but it should still have a considerable handicap relative to human fraud (which is very, very common). The fairest shake I can give this probability is to add up the number of books attributed to automatic writing, and then divide it by the number of books that have been published in modern history. There are only about 20 good examples of supposed automatic writing available, while, as of 2010, Google estimates that 129,864,880 books have been published, making our initial estimate p = .000000154, or just over 1 in 10 million.
PA2—Prior Probability of a Fraudulent Dictation—Since the only other live option is that Joseph somehow produced the book consciously via dictation, we subtract the sum of PH and PA1 from 1, which would mean that our prior probability of fraud would be p = .999999846.
Consequent Probabilities
CH—Consequent Probability of an Authentic Dictation—The evidence we’ve considered above fits a faithful assumption quite well, mostly in that once the interpreters or the seer stone is in play, there’s nothing that needs explaining. There’s no need to explain the absence of anyone noting Joseph writing the manuscript or using notes, and no need to explain the apparent feats of memory that would otherwise characterize the dictation process.
Still, it’s worth considering those who have attempted to replicate Joseph’s dictation process. There have been a few who have tried to do what Joseph claims to have done—reading sets of 20-30 words from a page while someone else writes down those words, having that person read the words back, making corrections before continuing on. Their experience reproduces what we see from the Book of Mormon dictation in all material regards—they get the same approximate dictation rate and the same types of scribal errors. In short, under an assumption of authenticity, nothing in Joseph’s story is unexpected (except, of course, glowing rocks in hats, but that’s already accounted for in our vanishingly small prior probability). There’s no reason that I can see not to assign a consequent probability of p = 1.
CA1—Consequent Probability of Automatic Writing—We next turn to the other potentially strange explanation. Does Joseph’s dictation process line up with the sort of material produced in alleged automatic writing? There have been a couple articles that have summarized the links between Joseph and automatic writing, and documented some of the key problems with the automatic writing hypothesis.
I won’t pretend to have much unique to add here, other than to document my own experiences with the tales of automatic writing. I took a hard look at the stories of several automatic writers, such as Pearl Curran’s “The Sorry Tale”, Catherine-Elise Muller and the material documented in “From India to the Planet Mars”, Geraldine Cummins and “The Scripts of Cleophas”, and Helen Schucman and “A Course of Miracles.”
The stories themselves are deeply interesting even if you don’t buy the premise of supernatural communication. It’s pretty easy to understand how contemporary witnesses could have come to believe their stories, and why they’re such a tempting morsel for Joseph’s critics. There are still plenty of paranormal researchers who continue to hold that at least some of these cases are authentic, though there are plenty that plainly aren’t. (See the Appendix for my brief notes on these works.)
Yet the tie between automatic writing and Joseph’s dictation is far from a slam dunk. We simply don’t know enough about the phenomenon of automatic writing to say for sure whether it explains the evidence we have. No modern scientific researchers have ever documented such capabilities. And the accounts we do have too closely intermix more convincing cases with obvious frauds. How are we to distinguish the incredible but authentic powers of the human unconscious from clever forgeries?
From what I can tell, though, there’s very little about works of automatic writing that suggest they’re anything but ordinary books, written by the people who put the words to paper. The material may have come from their unconscious minds, but they certainly didn’t produce published books at any prodigious rate or with any startling content. Of the eight works I looked at, all of them took years to produce—none of them can come close to matching the 65-day dictation speed of the Book of Mormon. They average a pace of 206.89 pages/year (SD = 127.62), which is only a fraction of Joseph’s rate of 3076 (assuming 531 pages over 65 days). Even if you count the time it took to typeset and publish (which would’ve been much slower on the nineteenth-century frontier than for any of the automatic writers), we still have a rate of 542.9 pages/year for the Book of Mormon, 2.63 standard deviations off the automatic writer’s average. Translating that into a z-score, we can estimate the probability that the Book of Mormon was written using the same process as the examples of automatic writing, putting our best guess for a consequent probability of this particular hypothesis at p = .0085 (assuming a two-tailed test).
In short, Joseph’s rate of production is much faster than most works of automatic writing.
CA2—Consequent Probability of a Fraudulent Dictation—So how would Joseph have been able to produce the evidence we have, assuming that there were no interpreters, no gold plates, and no quasi-mystical powers of automatic writing? Well, he would’ve had to have written it the old-fashioned way. Either 1) he would have had to write it on the spot as he went along, or 2) he would’ve had to have written it entirely in secret, memorizing it well enough to present it confidently during dictation sessions. Both of those options have their difficulties lining up with the data, and we’ll cover them both in turn.
Writing during dictation. First, dictating material on the spot is hard—it’s at least as hard as writing with paper, and probably harder. What makes it hard is not being able to refer back to the text in progress—it’s harder to keep a train of thought going when you can’t see what the whole train looks like. The only thing that makes it faster is that you don’t have to take the time to physically write or type the words. Professional writers who use dictation still have to make a ton of revisions as they go—a lot of the engineering challenge involved with dictation software is finding an easy way for writers to go back and make changes.
Thankfully, the digital nature of modern dictation means that we have some decent data on how much people have to go back and revise what they write when they dictate. Researchers can tell the difference between changes that people make due to the software making an error (repairs) and changes to content or wording (revisions). On average, people make around 3.37 total repairs and revisions per 150 dictated words (with a standard deviation of .5).
If Joseph was dictating the Book of Mormon out of his own head, he would’ve had the same sorts of alterations to the manuscript. There would’ve been times that his scribes misheard him (repairs) and times when he would’ve wanted to correct his own wording (revisions). The original manuscript shows plenty of evidence of the former, but none of the latter, which is consistent with what the scribes report. We don’t see the types of revisions that we would see if he was writing it on the spot.
But let’s pretend that both the scribes and the manuscript scholars are incorrect, and that the revisions we do see in the original manuscript are all from a normal dictation process. Is the manuscript cleaner than we’d expect? Well, based on my own count using the portion of the original manuscript that’s available here, there were about 1.68 changes made per 150 words, well short of the average we see from modern dictation studies. In fact, it’s 3.38 standard deviations lower than we’d expect, allowing us to estimate the probability of observing so few changes at p = .000725. In other words, it’s not particularly likely.
In recognition of the difficulty of this kind of writing process, some have recently argued that Joseph could have essentially improvised the Book of Mormon through a process of extemporization—memorizing the key points, and then expanding and elaborating those points through oration. This is the same kind of process that many use in public speaking, and it was very common among nineteenth-century preachers, and even for Joseph himself in speeches such as the King Follet discourse. These speakers could talk for hours without relying on notes, a fact which seems to apply well to Joseph’s dictation process.
The problem here, recognized even by those making this proposal, is that it doesn’t explain the complex features of the Book of Mormon or the long quotations of other Book of Mormon passages and the KJV. Joseph would still have needed access to a Bible, previous portions of the manuscript, and likely to a copious set of notes for this explanation to hold any water. We’ll be sticking with our estimate of p = .000725 for this portion of the hypothesis.
Writing in secret. How about the other scenario where Joseph is writing the Book of Mormon beforehand and memorizing it well enough to perform the dictation? Now, just stay with me here for a bit because this might seem a little off the wall. We have lots of cases where people stand in front of us and present well-crafted, well-structured material, without any evidence that they had spent much time writing it ahead of time. We call those people comedians. Think about the last time you saw someone like Jerry Seinfeld on stage. He’s not sitting there with a bunch of cue-cards or reading from a teleprompter—he’s up there with nothing but a mic and his own brain—the material he presents just seems to sprout from his head fully-formed. If he claimed that he’d just written that material on the spot, you’d be tempted to believe him, and you certainly wouldn’t have any evidence to prove him otherwise.
But Jerry would never make that claim. Why? Because for every minute he spends onstage he spends an ungodly amount of time writing his material. He writes every joke down—on paper—and has to polish and refine it over and over until it shines. If he tried to pretend it was all ad-libbed, he would have to try to find an alibi to cover for a substantial portion of his life that would otherwise be unaccounted for, and people would notice. He lives with people. He has friends. People would wonder where he was going all the time. They would walk in on him and ask what he was doing. They would stumble across his meticulous notes. Even if these people were his friends, they’d be likely to call him on his lies. And the longer he went on pretending, the riskier it would be.
The same would hold true for Joseph. Yes, it’s conceivable that he could write and memorize the Book of Mormon, but it would take him a while, and the odds that he’d produce some incriminating evidence along the way would get rather high. Joseph didn’t have a ton of time alone, or a private office where he could stash any alleged copious notes—he spent most of his time in the fields or in other laborious endeavors working with family and friends, and otherwise he was in crowded houses with one or two rooms. It’s also probably not responsible to assume that Joseph had much leisure time in which to get that kind of work done. People generally worked more hours than we do today, as working hours markedly declined in the period following the industrial revolution, and farm laborers like Joseph generally put in more hours of work than other professions. For Joseph to have spent that much time in undetected writing would have been a feat in and of itself.
Just how long would it have taken him to write and memorize the Book of Mormon? I was able to find this reddit thread where comics share how long it takes them to come up with 5 minutes of material and get it ready to perform. Most of the responses say that it takes them around 3 hours just to get something basic ready (followed by months of practicing and performing to hone it further), though one person said they could do it in just an hour. Let’s assume that Joseph was on the low end of the spectrum here, and that it would require an hour to produce. Given average speaking rates, that would give us a writing pace of 750 words/hour, meaning that he would have to find at least 356 hours of writing/memorization time to produce the 267,000 words of the Book of Mormon.
Now remember, when professional comics perform this sort of work, we’re not talking about them idly thinking up jokes as they’re going about their day. This kind of writing and memorization work requires just as much concentration as any other kind of writing. I can’t quite swallow the idea that Joseph could have given us the Book of Mormon we have by daydreaming it up during his work in the fields, in part because of my own experience with writing. Before writing the couple amateur novels I have under my belt, I spent many hours daydreaming about the plot and the characters and the setting. None of this daydreaming made it any easier to put the words themselves on paper—each took 300+ hours for me just to get a first draft, and they weren’t nearly as long or as complex as the Book of Mormon.
The only option then, in my mind, is that Joseph would have had to find a way to spend at least 10 working weeks of his life laboriously writing and rehearsing the book without anyone realizing what he was doing, and each hour spent in that kind of work would expose him to the risk of discovery. We can assume that such risk would apply independently to each hour of work, an assumption which allows us to model how the overall risk of discovery would change as he logged more hours of writing time. Even if he only had a 1% chance of getting caught every hour, that would mean that he would have a probability of .99356 of not getting caught, which would be p = .0297.
[OH NO! HE USED RED TEXT! WHAT DOES THAT MEAN? It means that I’m making a guess here, and that I have no particular data to back that guess up. It’s almost impossible to know for sure how likely Joseph would’ve been to get caught, and unless you’d like to observe someone while they live in close quarters and try to write a secret book, we’re not going to get any extra insight on that anytime soon. I’m trying hard to bias my guess against Joseph and an authentic Book of Mormon, but a guess is what we’ll have to live with.]
Now, that’s not all that bad as far as things go. But it’s important to note how precarious that probability is. If Joseph could only write even a bit slower than that—say, at 500 words/hour—then that probability takes a nosedive to p = .0046. And if the risk of getting caught increases to 5%/hour instead of 1%, then we’re really in trouble, putting us at p = .0000000117. We’re going to give the critics the benefit of the doubt here, but it wouldn’t be unreasonable to see it as a near impossibility.
Dictating without prompts. So, of the two options, it’s probably easier to believe that Joseph wrote and memorized the Book of Mormon outside of dictation sessions rather than dictated it as he went (which also seems to be the consensus of critics). And if all we had was an otherwise unremarkable 267,000-word book it wouldn’t be all that hard to believe that he could pull it off. But we have more than that—which will be amply demonstrated in future episodes—and in this case, we have some other evidence to account for as well. In particular, we have the reports from scribes that Joseph didn’t need to be prompted before starting a new dictation session. That seems, at the very least, a remarkable feat of memory.
Now, despite the massive literature on human memory, we don’t really have a close experimental analogue to help guide us here. But we can get pretty close. Cognitive psychologists have tried to get people to do some pretty crazy memory tasks, and we have a decent sense of how well humans do on them. The most applicable one that I was able to find is called the reading span task. Researchers had people read a set of sentences, and then immediately recall the last word in each sentence, with the number of sentences increasing as the task went on. This might sound easy, but it’s not—the average score was 4.63 out of a possible 10 (with a standard deviation of 1.48).
Now hopefully we can all agree that it would be easier to believe that someone got a 10/10 on that task (though no one in the sample did), than to believe that Joseph could start dictating without a prompt (at least 126 times straight over 63 working days, assuming he only took one break per day), so the probability of him being able to do should be at least as low as someone getting a perfect score, if not a great deal lower, and that probability would be p = .000283.
“But ho!” remark the critics. “What if Joseph had a photographic memory? Wouldn’t that let him accomplish such memory-related feats with ease?” To that I reply: what evidence do we have that he had a photographic memory, or that a photographic memory would allow him to do this kind of thing? It’s true that some people have very impressive, savant-level memories, being able to recount thousands of decimals of pi without having to spend time memorizing them, or to tell you what day of the week you were born on without checking a calendar. But these savant-level abilities are very rare, with one estimate at 1 in a million, and another at just 50 people in the entire current human population. And the reality of photographic memory doesn’t live up to the myth perpetuated in movies and TV shows. Photographic memory (or eidetic memory) is extremely rare in adults and is still quite prone to error. We’re actually doing the critics a favor by assuming Joseph fell on the spectrum of normal human memory, since otherwise we’d have to radically adjust our prior probability estimate.
One usual evidence provided for Joseph’s impressive memory is in reference to Joseph’s dictation of D&C 132, where Joseph, after being asked by Hyrum to confirm the revelation through the Urim and Thummim, claimed that “he knew the revelation perfectly from beginning to end," and elsewhere claimed that he could rewrite it at any time if necessary. But claims don’t necessarily align with reality, and those claims don’t seem to square with other known facts, such as Joseph forgetting that Jerusalem had walls. We have no evidence and little reason independent of the Book of Mormon itself to assume Joseph had an incredible memory, meaning our estimate of p = .000283 remains reasonable, if not generous.
Summary. So, all told, we can now calculate our consequent probability. What’s the probability that we would both have no evidence of working from notes or from a Bible and evidence that Joseph didn’t require prompting at the start of dictation sessions? We’ll use the probability that he would have been able to spend 356 hours writing and memorizing the material without getting caught (p = .0297) and the probability of performing a startling feat of memory (p = .000283), which we’ll assume are independent. Multiplying them together gives us a probability of p = .0000084. For those who like visual aids, here’s one that summarizes how we got to this point.
And with that, I believe we’re ready to tie this all together.
Posterior Probability
The Bayesian formula we’re going to use has to be modified slightly to account for our additional hypothesis, but it’s still pretty straightforward.
PH = Prior Probability of the Hypothesis (our guess at how likely an authentic dictation would be based on what we know so far, which is p = 8.37 x 10-40)
CH = Consequent Probability of the Hypothesis (our estimate at how well the data we have would align with what Joseph and others claimed about the dictation process, or p = 1)
PA1 = Prior Probability of Automatic Writing (our guess at how likely it is that a given book is a product of automatic writing, or p = .000000154)
CA1 = Consequent Probability of Automatic Writing (our estimate of the probability of an automatic writer producing published material as quickly as the Book of Mormon, or p = .0085)
PA2 = Prior Probability of a Fraudulent Dictation (our assumption that a fraudulent dictation is overwhelmingly likely, p = .999999846)
CA2 = Consequent Probability of a Fraudulent Dictation (our estimate of how likely a fraudulent dictation would be to produce the evidence we have, or p = .0000084)
PostProb = Posterior Probability (our new estimated probability of an authentic dictation)
| PH = 8.37 x 10-40 | |
| PostProb = | PH * CH |
| (PH * CH) + (PA1 * CA1) + (PA2 * CA2) | |
| PostProb = | 8.37 x 10-40 * 1 |
| (8.37 x 10-40 * 1) + (.000000154 * .0085) + (.999999846 * .0000084) | |
| PostProb = | 9.96 x 10-35 |
Lmag = Likelihood Magnitude (an estimate of the number of orders of magnitude that the probability will shift, due to the evidence)
Note: In this case, since we have more than one consequent probability for the alternative hypothesis, we’ll do a weighted sum of the two CAs, weighted by their respective PA’s contribution to the overall PA. Given that the alternative for automatic writing has such a small prior, that means we’ll essentially be using the consequent for a fraudulent dictation.
Lmag = log10(CH/CA)
Lmag = log10(1/.0000084)
Lmag = log10(119047)
Lmag = 5
Conclusion
This is the sort of evidence that should make our skeptic pay a bit more attention, increasing the probability of an authentic Book of Mormon over 100,000 times. Some of our estimates are admittedly a little light on data, but that’s why I’ve tried hard to stack the deck against Joseph and his story. More in-depth analyses on any one of the issues I’ve raised here could easily bring the dictation into smoking-gun territory, and they’re unlikely to make the case any better for the critics. Joseph’s dictation narrative is not only incredible, attempts to otherwise account for it are quite unlikely. That makes it notable evidence in Joseph’s favor no matter how much some would prefer to stuff it away in the nearest hat.
Skeptic’s Corner
When I say above that my estimates are a little light on data, I wasn’t kidding. I would much prefer to base these estimates on solid observational studies, especially ones that specifically replicate the experience of resuming dictation unaided, without receiving a prompt. To be fair, that could just as easily make this evidence stronger rather than weaker, so critics should be a bit cautious there. A stronger line of argument would be to go after the historical evidence itself—the only source we have for that particular detail is Emma, and it’s possible that she was exaggerating or that he did indeed need prompting at times that she wasn’t present. A good review of how Joseph’s memory played out in the historical record would be really useful.
In terms of the risk of Joseph being discovered while writing or rehearsing the text of the Book of Mormon, we have very little to go on. I picked a 1% chance of discovery because that seems like a relatively small risk, but it’s not unfathomable that the risk is even smaller, and that Joseph was just really good at not being caught. Judging by how easily he was caught in other aspects of his life, though, I’m not sure that idea holds water.
And lastly there’s the idea that Joseph really did make use of a Bible and extensive notes while writing the Book of Mormon, and that the dictation witnesses were lying or just simply weren’t present when he did so. The breadth of biblical quotation in the text makes the latter idea unlikely—neutral observers like Reuben Hale wouldn’t have had to watch very much of the dictation before Joseph would’ve been forced, sooner or later, to refer to a Bible or to a previous spot in the manuscript. And the former, as we’ll see in a couple episodes, would make matters worse for the critics. Bringing people in on a conspiracy is a really bad idea, and having all the dictation witnesses successfully keep Joseph’s secret would probably be an even less likely prospect than Joseph just memorizing the entire book. If you’re intent on discounting the dictation-related evidence, it’s better to assume the latter, as unlikely as it is.
Next Time, in Episode 4:
In the next episode, we’ll look at the available DNA evidence surrounding the Book of Mormon, examining the likelihood of failing to observe Near-Eastern DNA in the Americas given what’s described in the Book of Mormon.
Questions, ideas, and wooden horses full of sweaty Greeks may be gifted to BayesianBoM@gmail.com or submitted as comments below.
Appendix
Examined works of automatic writing
| Author | Work | DoB | Writing Sample | Dictation Time | Notes |
|---|---|---|---|---|---|
| Pearl Lenore Curran | The Sorry Tale | 1883 | “And lo, there rode, upon a camel’s pack, a one upon the in-road unto her, and this shewed within the white light. And the sun came up and lo, reds crept and golds glinted, and upon the young sun’s redded ball the camel man shewed black and the camel sunk and rose upon his loose legs. And the one cried out: “E-e-e-o-e! E-e-e-o-e!” and the beast stealth-slipped on.” |
1500 words/hour 658 pages in 3 years |
The writing is mostly nonsense–feels much more like “stream-of-consciousness” than does the Book of Mormon. It’s also way too fast. As the purported rate we’d have had the book of Mormon in days, not months (though The Sorry Tale itself seems to have taken three years to produce). “Either our concept of what we call the subconscious must be radically altered, so as to include potencies of which we hitherto have had no knowledge, or else some cause operating through but not originating in the subconsciousness of Mrs. Curran must be acknowledged.” – Walter Franklin Pierce – The Case of Patience Worth |
| Catherine-Elise Muller | From India to the Planet Mars | 1861 | “Helene told me one Sunday that she had been possessed several times during the day by the hallucinatory image of a straw hat.” (Okay, this one’s not a writing sample, but it’s still hilarious) |
“Flournoy was able to show her Martian language was an artful fabrication. Although it sounded decidedly foreign, frequency analysis of words and letters and an examination of the syntax convinced Flournoy that the language had all the basic structural characteristics of French.” The writing is mostly gibberish with some random words thrown in. |
|
| Geraldine Dorothy Cummins | Swan on a Black Sea | 1890 | “Titus called upon the Gentiles to confess the evil of their lives. One by one they passed before the high seat, halting to make their lewd confession. Then, bowing their heads and declaring contriteness, they went forward, making way for one another.” | 220 pages over 3 years | Geraldine had a long history of writing plays and fiction prior to becoming a medium. The nature of the writing doesn’t appear at all unusual. |
| Helen Schucman | A Course in Miracles | 1909 | “Forgive us our illusions, Father, and help us to accept our true relationship with You, in which there are no illusions, and where none can ever enter. Our holiness is Yours. What can there be in us that needs forgiveness when Yours is perfect? The sleep of forgetfulness is only the unwillingness to remember Your forgiveness and Your Love. Let us not wander into temptation, for the temptation of the Son of God is not Your Will. And let us receive only what You have given, and accept but this into the minds which You created and which You love. Amen.” | 1312 pages over 7 years | Helen was a clinical psychologist. The book underwent extensive revision and edits prior to publication. The content itself is rather bland religious platitudes. |
| Jane Roberts | Seth Speaks | 1929 | “There is no such thing, in your terms, as nonliving matter. There is simply a point that you recognize as having the characteristics that you have ascribed to life, or living conditions—a point that meets the requirements that you have arbitrarily set” | 480 pages over 2 years | Jane was a poet prior to encountering Seth. She started with a Ouija board and then moved to dictation. She remained a life-long atheist and was extremely skeptical of her own story. The content is extremely interesting, and not altogether incompatible with LDS thought (though it’s a bit rambly and repetitive). The cadence and fluency seems similar at first glance to the introductory and commentative material provided by Robert Butts, her husband. |
| John Ballou Newbrough | Oahspe: A New Bible | 1828 | “After the creation of man, the Creator, Jehovih, said unto him: That thou shalt know thou art the work of My hand, I have given thee capacity for knowledge, power, and dominion. This was the first era.” | 900 pages over 2 years | Newbrough manually wrote the material with a typewriter and/or pen and paper. Lots of strange historical and scientific oddities contained within. |
| Anonymous | The Urantia Book | 1924 |
Starting from any one element, after noting some one property, such a quality will exchange for six consecutive elements, but on reaching the eighth, it tends to reappear, that is, the eighth chemically active element resembles the first, the ninth the second, and so on. Such a fact of the physical world unmistakably points to the sevenfold constitution of ancestral energy and is indicative of the fundamental reality of the sevenfold diversity of the creations of time and space. |
2097 pages over 10 years | “[26] Martin Gardner states that an explanation concerning the origin of the book more plausible than celestial beings is that the Contact Commission, particularly William Sadler, was responsible. Gardner’s conclusion is that a man named Wilfred Kellogg was the sleeping subject and authored the work from his subconscious mind, with William Sadler subsequently editing and authoring parts.[27] Brad Gooch believes Sadler wrote the book, possibly with help from others on the Contact Commission.[28] A statistical analysis using the Mosteller and Wallace methods of stylometry indicates at least nine authors were involved, and by comparatively analyzing the book against The Mind at Mischief, does not indicate authorship or extensive editing by Sadler, without ruling out the possibility of limited edits.”
The book contains a great many scientific claims that reflect scientific understanding at the time (with a couple exceptions that aren’t particularly profound). Evidence of plagiarism from 15 other books. |
| Levi Dowling | The Aquarian Gospel of Jesus Christ | 1844 | He said, the Lord had given me this wealth; I am his steward by his grace, and if I ive not to his children when in need, then he will make this wealth a curse. | 274 pages over 4 years | Rather clearly the product of Dowling. |
In your conclusion, you say that “Joseph’s dictation narrative is incredible,” but you clearly do not believe that it is impossible, which raises the notion of Sherlock Holmes that “When you have eliminated the impossible, whatever remains, however improbable, must be the truth.”
That suggests a need to take seriously the notion that Joseph had access to an already extant manuscript, but (due to its Early Modern English nature) it could not have been authored by him. This would also take account of the strong opinion of a number of scholars from B. H. Roberts to Dan Ludlow that the blocks of KJV quotations therein can only be accounted for by someone referring directly to a copy of the KJ Bible.
This further demands that we take seriously the testimony of several key witnesses that the seerstone somehow presented short English phrases to Joseph’s eyes in bright Roman letters, and he simply read them to his scribes. No one would object to that scenario nowadays since all Joseph would have needed today is a smartphone.
While it is true that such technology did not exist yet in Joseph Smith’s time, God would presumably not be prevented from using such technology then or at any other time in history. Even atheist astrophysicists admit that it is very likely that very advanced civilizations have existed in our galaxy for a long time, and they would seem godlike to us humans on Earth. What would Sherlock say? In Joseph’s time, of course, the word “translation” could also mean “transmit,” and perhaps that is really what he did.
Still, this is not a very promising section of your evidentiary inquiry. As with your last episode, this one does not deal directly with the book itself. Moreover, as with trying to estimate the likelihood of the Resurrection of Jesus, or the likelihood of the Exodus of Moses, hard forensic evidence seems to be rather scant.
Part 2 of 2
When looking at the length of the Book of Mormon, you noted that it was longer than most (but not all) 19th century literature. Not fitting this pattern of naturally written books counted as significant evidence against it being a natural book.
Applying the same reasoning here requires asking at least two questions: First, what percentage of authentic *translations* are created by reading words off of a seer stone in a hat? Second, what percentage of authentic *revelations* are received by reading words off of a seer stone in a hat?
Your estimated of CH (of how well the data we have would align with the BoM being an authentic translation or revelation) is 1.00. This implies you think 100% of real revelation and real translation comes about by the process of a prophet dictating words he is reading off of something he sees while his head is in a hat.
I don’t think this is true. For example, the Book of Abraham wasn’t translated by reading words off of a seer stone in a hat. Nor the dead sea scrolls nor any other real document was ever translated this way. None of the revelations in the D&C were read off of a stone in a hat. Nor anything in the Bible. The Family Proclamation didn’t come about by a prophet reading what he saw in a hat.
The translation of the Book of Mormon, with 6 years of preparation for a 65-day dictation period that constituted reading words the prophet saw with his head in his hat is unique among revelations and translations. Consistently using your reasoning process dictates that this constitutes strong evidence against its authenticity. It doesn’t fit the established pattern.
In contrast, lengthy timeframes of preparation and props such as magic stones and magic hats are essential tools of the trade for magic. In aggregate, the details of how the BoM was dictated counts as evidence against the BoM. I’ll conservatively assign a likelihood ratio of 1-in-10 to this, lowering my probability of authenticity from 1-in-400 to 1-in-4,000.
I’m glad you went down this route Billy, and hopefully we can talk this idea through a little bit together.
“The dictation process didn’t start until about 6 years after the plates were first discovered.”
I’m not sure how the gap is relevant, other than it gave Joseph and potential conspirators time to plan. My analysis assumes that they would’ve had essentially as much time as they needed, but that time needs to be paid for with risk of discovery. You’re also apparently forgetting that the hat was a rather late innovation, so I’m not sure how the six years factors in there.
Speaking of risk of discovery, I’d be interested in your thoughts on what I’ve actually done here–it feels more like you’re talking past my analysis rather than engaging with it. Do you agree that we should prefer theories that would’ve taken less time to put together over ones that would’ve taken more time? If so, how would you propose to handicap ideas that would’ve required more man-hours? Can you think of a better way to do it than the method I’ve put forward?
“Under this third hypothesis, the Book of Mormon was already written behind the scenes, possibly with the help of somebody like Dartmouth-educated Hyrum or Lucy Mack.”
If so, then you’re just playing a shell game. You’ve pushed the problem from Joseph onto someone else without considering the potential problems with doing so. Not only does the length problem still apply here, but there would be numerous other difficulties, such as expanding the corpus of potential co-conspirators (which we’ll get to in a couple weeks). And, of course, they’d have to find a way to secretly author the manuscript the same as Joseph would’ve, with a similar effort required and a similar risk of discovery.
“I’ll call the third naturalistic hypothesis the “magic trick” hypothesis.”
I can understand why this hypothesis would appeal to you, but all it does for me is bring up a host of questions. How likely would it have been for Joseph to have the sort of slight-of-hand skills that such a trick would’ve required? What did the slight-of-hand actually look like? How many pages would he have needed to have in the hat at any given time, how small would the text had to have been, and how would he have constantly manipulated those pages without being detected? Were such trick-hats common in the early 19th century, and how much effort would it have taken to procure or develop one? And what about the biggest problem, which is that Martin and other scribes apparently had full access to the hat and stone (as indicated by his rock-swapping maneuver)? All of these issues would need to be addressed and assigned probability values before I could take this hypothesis seriously. At the very least, I’d need estimates for: 1) the probability of Joseph having access to slight-of-hand techniques, and 2) the risk of discovery under those conditions during the estimated 520 hours of dictation.
The real magic trick, in my mind, is that someone got you to accept this hypothesis with complete credulity.
“I’ll conservatively assign a likelihood ratio of 1-in-10 to this”
I’d already built in the strangeness of rocks, angels, and seer stones into my 10^-41 prior, so it’s perfectly reasonable for you to do the same here. At some point, though, I’d like to see you statistically swallow some of the hard pills that the BofM has to offer. Identifying a seemingly plausible explanation (and then waving your hands with a flourish) isn’t enough. Adopting something like the magic trick hypothesis with any degree of integrity requires that you critically examine that hypothesis in an a fortiori manner and, if you can’t concretely instantiate it in the math, adopt appropriate penalties for doing so.
Otherwise I’ll have to assume that you’re simply committed to ignoring any of the sincere questions and issues brought forward by faithful scholars. You’re obviously smart enough that you could rise above that ideological commitment if you wanted to. Whether you’ll choose to do so remains to be seen.
KR: “I’m not sure how the gap is relevant, other than it gave Joseph and potential conspirators time to plan.”
I would think dictating a book in 65 days is a lot easier if you’ve had 6 years to prepare.
KR: “ You’re also apparently forgetting that the hat was a rather late innovation, so I’m not sure how the six years factors in there.”
Before the hat innovation, the translation process was separating Joseph Smith from his scribe with a blanket draped across the room and again, that was still being done after years of preparation. That was awfully suspicious, don’t you agree? I think they switched over to the stone in the hat as soon as they devised it. Translating that way was more impressive to his audience than translating behind a curtain. That makes a heck of a lot more sense than saying God was upset with him and as punishment made him translate with his trusty old seer stone rather than with the Urim and Thummim.
KR: “Speaking of risk of discovery, I’d be interested in your thoughts on what I’ve actually done here–it feels more like you’re talking past my analysis rather than engaging with it.“
On this one, there isn’t a lot to engage with. In principle, I agree that it is unlikely that a random New England farmer could pull this off as a trick. Sure. But it’s also unlikely that a random New England farmer could convince his neighbors to pay him to dig for buried treasure because he could see it underground with his seer stone. Joseph Smith was not an ordinary farmer. And in general, not knowing how a magician does a trick doesn’t count as evidence that he really used magic. And regardless of how unlikely all that is, it needs to be kept in context of the unlikeliness of authenticity. You do that in a valid manner by estimating CH, not by padding PH.
KR: “At the very least, I’d need estimates for: 1) the probability of Joseph having access to slight-of-hand techniques, and 2) the risk of discovery under those conditions during the estimated 520 hours of dictation.”
You would also need an estimate for the covariance between those two things. Let’s say the probability is 1/1000 that he had the slight-of-hand skills to do this. *If* he did have the skills to do it, the risk of discovery would be very low, so the probability of having the skills to do this and *not* getting caught wouldn’t be much lower than 1/1000.
KR: “I’d already built in the strangeness of rocks, angels, and seer stones into my 10^-41 prior.”
My point isn’t that those things are “strange.” My point is that they are *unlikely*. The relative unlikeliness of Joseph being able to dictate this in 65 days in the manner that he did under the natural hypothesis needs to be kept in context of the unlikeliness of seeing the same evidence under the supernatural hypothesis. If you set CH = 1.00 you aren’t doing this, regardless of what you set PH equal to.
“I would think dictating a book in 65 days is a lot easier if you’ve had 6 years to prepare.”
You’d have more hours to work with, but those hours would still need to happen, and each bit of additional effort risks giving the game away.
Reducing the number of years could potentially reduce the number of hours available–something similar will apply when I talk about ship construction in a few weeks–but with my estimate of the number of hours required it doesn’t make much of a difference whether he had 6 years or just one.
“That was awfully suspicious, don’t you agree?”
Maybe, but also an unnecessary bit of rigamarole relative to just saying that he was receiving the words through vision (as you point out, there’s nothing that would’ve stopped him from using the same method as the BofA). And the curtain makes quite a bit of sense from the faithful hypothesis, as a sacred artifact for which there would be a limited number of witnesses. So it’s not entirely clear which way that cuts or by how much.
“In principle, I agree that it is unlikely that a random New England farmer could pull this off as a trick.”
Then perhaps you should consider accounting for that in your analysis.
“But it’s also unlikely that a random New England farmer could convince his neighbors to pay him to dig for buried treasure because he could see it underground with his seer stone. Joseph Smith was not an ordinary farmer. ”
Now this would be a separate bit of evidence worth considering, and I’ll admit that it didn’t make my list. But I’m not sure there’s much meat behind the idea that treasure-digging was an unexpected practice in an early 19th century context.
Also, it does nothing to take away from the question of the dictation evidence, as framed here. Handy distraction attempt there Billy! Squirrel!
“You would also need an estimate for the covariance between those two things.”
True, though it’s possible to handle that using a conditional probability. You calculate the probability of Joseph having slight-of-hand skills (which I’d reckon would be quite a bit lower than 1 in 1000), and then calculate the probability of him succeeding over the course of 520 hours on the condition that he has those skills. It would take the greatest magician in the world (and probably then some) to guarantee 100% success over that timeframe and under those general circumstances (i.e., where Martin can freely peek in over his shoulder at any given moment, and also has almost unmitigated access to the hat during breaks, etc.).
“If you set CH = 1.00 you aren’t doing this, regardless of what you set PH equal to.”
Feel free to change my prior to 10^-40 in your head then, and change the CH in your head canon to .1. Since the result would be identical, I’m not going to sweat it much.
KR: “Then perhaps you should consider accounting for [it being unlikely that a random New England farmer could pull this off as a trick] in your analysis.”
I did. That is why my likelihood ratio is 10 rather than 10^20.
I’m being generous here. If we assumed document X were an authentic translation of an existent ancient manuscript, what are the probabilities we would have this translated document because the translator dictated words he read off of a seer stone in a hat in 65 days? It should be CH = 0.00. No revelation is received this way. No translations are done this way. When estimating this, there is nothing other than 0 to put in the numerator. But I’m being generous and setting this to only being slightly less likely than the plausible-yet-unlikely mundane explanation.
“It should be CH = 0.00.”
And there’s that hidden impossible prior (expressed at a CH, in this case) that I knew was lurking about. Didn’t take long for it to show itself.
“No revelation is received this way.”
Unless you count the Israelite Urim and Thummim. Or the Mesoamerican zaztun.
The idea of receiving an entire book that way would be unique, for sure. But if you wanted to do that sort of thing, using seer stones as a medium to assist in its delivery would probably be as good a method as any.
There is no evidence of any curtain or blanket placed between Joseph and his scribe at any time. As Elizabeth Ann Whitmer stated: “Joseph never had a curtain drawn between him and his scribe while he was translating.” Her brother David Whitmer said that “a blanket was used only to prevent visitors from seeing the work – the entire household (Oliver, Emma, et al.) viewed the work of translation firsthand.” Joseph’s wife Emma said: “In writing for your father I frequently wrote day after day, often sitting at the table close by him, he sitting with his face buried in his hat, with the stone in it, and dictating hour after hour with nothing between us.”
> There is no evidence of any curtain or blanket placed between Joseph and his scribe at any time.
Except Martin Harris’s interview with John A. Clark in 1840 (see Welch 2005 chapter “The Miraculous Translation of the Book of Mormon”).
It is probably this account (and some other?) that caused Royal Skousen to write in the 1997 article Rasumussen linked:
> Martin Harris, in a couple of early statements, said that a blanket or curtain separated Joseph from him at the time he (Martin) obtained a sample transcript and translation to take to Professor Anthon in New York City.
There are other accounts that include the blanket (antagonistic sources, though, but still constitute some level of “evidence” from a historical perspective).
Part 1 of 2
This episode is about the evidence of how the BoM was dictated. There are three salient points:
1- Once the dictation started, it went at a fast clip with no revisions. The way the scribe read back material indicates Joseph was comparing what the scribe said to what he saw in his hat.
2- The dictation process didn’t start until about 6 years after the plates were first discovered.
3- The dictation process entailed Joseph Smith putting a rock in his hat, sticking his head into his hat, and then reading what he saw there.
In addition to the “automatic writing” hypothesis and the “Joseph Smith conscious writing process” hypothesis you described, I’d like to introduce a third naturalistic explanation. This third one happens to be the only hypothesis, either naturalistic or supernaturalistic, that explains all three salient features.
I’ll call the third naturalistic hypothesis the “magic trick” hypothesis. Under this third hypothesis, the Book of Mormon was already written behind the scenes, possibly with the help of somebody like Dartmouth-educated Hyrum or Lucy Mack.
But how did the dictation happen? A trained magician looked at the story of how the BoM was dictated and thinks the answer is clear. The hat Joseph used was actually translucent—plenty of ambient light crept into the hat. Therefore, if a manuscript containing a page of the BoM was in the hat with the rock, Joseph easily could have read the manuscript in the hat. The rock was merely a magic prop used to distract his audience (see Radio Free Mormon: 161: Magic and the Book of Mormon for more details).
There are several appealing elements to this. It explains the 6-year gap between when the BoM was discovered and when the dictation process began. It explains the rock. It explains the hat. And it fits neatly with witness accounts that indicate Joseph was reading something.
Understanding the plausibility of the magic trick hypothesis makes it clear the odds of this happening in a naturalistic way are much greater than the 1-in-100,000 you ascribed.
Continued in Part 2…