The Comprehensive Bible of Book of Mormon Evidence

Chapter 1: How to Read Evidence Honestly

The Tools Before the Test

I. Introduction

Picture a Quora answer being typed at midnight. The question reads, in some form or another: Has any archaeology ever confirmed anything in the Book of Mormon? The respondent, a Latter-day Saint apologist with strong feelings, opens with the strongest hit he knows: a name found on a recently excavated altar in Yemen that matches a place-name in 1 Nephi. He follows with another item, this one about cement structures in Mesoamerica. And another, this one about steel swords in Iron Age Jerusalem. By the time he hits the word count limit, the answer is a cascade of vindications, every one of which he believes is decisive. The reader who clicks the question's other answers will read, in nearly identical form, an opposing cascade: no horses, no chariots, no wheat, no smelted iron, no DNA, no inscriptions. Both writers leave their respective audiences satisfied. Neither writer has changed a single mind on the other side.

I have been on the apologetic side of that exchange more often than I can count. Over five years of answering questions on Quora about Latter-day Saint belief, by my last count somewhere over a hundred answers viewed more than forty-two thousand times. I have learned something the cascades obscure: the problem with the public Book of Mormon debate is not, in the main, that one side has facts and the other does not. The problem is that both sides are bad at using the facts they have. The good evidence and the bad evidence get bundled together; the strong claims get diluted by the weak ones; the diagnostic items get lost in a pile of items that prove nothing on either side.

This chapter is, in a sense, the precondition for the rest of the book. Before we examine any specific piece of evidence about the Book of Mormon, before we even open the Arabian travelogue or assess the population genetics of the Americas, we have to agree on how to weigh evidence in the first place. The chapters that follow advance specific arguments, some strong and some weak, and rate every one of them on an explicit scale. The scale only works if the underlying methodology is sound. So we begin with the methodology.

I will introduce five tools. None of them is original to me. None of them is even original to the Book of Mormon debate; they are standard moves in any field where evidence is in dispute, from textual criticism to forensic science to historical Jesus studies. What is unusual is the consistency with which they have been ignored, on both sides, in the popular Book of Mormon literature. Apply them rigorously and the apologetic case looks both stronger and weaker than the cascade approach suggests, stronger because the genuine items become visible once the noise is removed; weaker because the noise is, in fact, substantial.

The five tools are: the leads-vs-follows test, Bayesian framing without the math, the loan-shift principle and its limits, the proper use and the proper limits of stylometric analysis, and the difference between single-point evidence and conjunction evidence. To these I will add, at the close of the chapter, a sixth discipline that runs through every chapter of this book: the strength rating system, by which every evidentiary item discussed is tagged with a transparent indicator of how much weight I judge it to bear.

II. The Leads-vs-Follows Test

The single most useful question one can ask of any piece of supposed evidence for the Book of Mormon's authenticity is this: did the text say something specific, and did the world later confirm what the text said? Or, conversely, did the world discover something, and did interpreters then reach back to the text and find language that could be made to fit?

A. When the text leads

The clearest example of the text leading is Nahom. The Book of Mormon, dictated and published in 1830, says that Lehi's party stopped at a place called Nahom, where Ishmael was buried (1 Nephi 16:34); that from Nahom they turned nearly eastward (1 Nephi 17:1); and that they eventually arrived at a fertile coastal land they called Bountiful, with timber, fruit, honey, and the materials needed to build a ship. In 1830 these were claims, not confirmations. The place-name Nahom was unknown to the English-speaking world's standard biblical and geographic references (Aston & Aston, 1994). The eastward turn was not on any map a frontier farmer in upstate New York could plausibly have consulted. The Dhofar coast of Oman, fertile, anomalously green, with the timber, water, and cliff geography Nephi describes, was not part of any popular geographical literature of the period.

Then the world caught up. In 1988, German archaeologists excavating the Bar'an temple at Marib in Yemen recovered three inscribed limestone altars donated by men identifying themselves as members of the Nihm tribe, dated on epigraphic and archaeological grounds to the seventh or eighth century BC, contemporary with Lehi (Aston & Aston, 1994; Aston, 2015). The site sits along an ancient incense trade route that does, in fact, turn eastward across the Empty Quarter precisely where the Book of Mormon's narrative places the turn. The Dhofar coast of Oman, fully described by Western explorers only in the late nineteenth and twentieth centuries, was systematically surveyed by Latter-day Saint researchers and others in subsequent decades; the candidate site Khor Kharfot satisfies the text's specific topographical and botanical claims.

This is what the text leads looks like. A specific claim, name, direction, terminus, was made before the world could check it. The world has now checked it, and the check has gone in the text's direction. The match is not airtight, and we will see in Chapter 2 exactly where its discounts lie. But it is the cleanest example we have of a Book of Mormon detail that was specific and confirmed by independent evidence the author could not plausibly have manufactured.

B. When the text follows

The opposite case is the popular argument that Abraham 3, Joseph Smith's revelation describing the celestial sphere and the star Kolob, anticipates Einstein's theory of special relativity. The argument runs, roughly: the text says one revolution was a day unto the Lord ... after the time of Kolob, which Kolob is after the reckoning of the Lord's time (Abraham 3:4), and modern physics has demonstrated that time runs differently in different reference frames. Therefore Joseph Smith, in 1842, anticipated Einstein.

The strongest form of this argument fails. Special relativity is a specific physical theory: velocity-dependent time dilation governed by the invariance of the speed of light and the Lorentz transformations. Nothing in Abraham 3 mentions velocity, light-speed, or reference frames; the text describes a rotational period (revolutions) used as a calendar (reckoning), not a relativistic effect. To extract relativity from Abraham 3, one must discard the words the text actually uses and substitute concepts the text does not use. And the specific 1000:1 ratio is biblical, not scientific: 2 Peter 3:8 reads one day is with the Lord as a thousand years, echoing Psalm 90:4. The ratio appears in Abraham 3 because it appears in Scripture, not because it was revealed from physics not yet discovered. (Appendix C revisits this item at length. The strongest form of the case turns not on physical mechanism but on the conceptual claim that the rate of time flow is locality-dependent. That conceptual claim is striking against the Newtonian-absolute-time framework of 1830 and survives on its own terms as a borderline [C] item. The dismissal here is of the strong physical-prediction-of-relativity reading, not of the conceptual claim Appendix C treats more fully.)

But a more modest version of the same argument does not fail, and the test that rewards specificity should reward it. Abraham 3:9 reads: thus there shall be the reckoning of the time of one planet above another, until thou come nigh unto Kolob. That sentence makes a claim that 1842 Newtonian physics explicitly rejected. Newton's Principia (1687) had defined time as absolute, true, and mathematical ... flowing equably without relation to anything external; Kant treated time as a universal a priori category. The dominant physical assumption of Joseph Smith's day was that there is one universal time, the same everywhere. Abraham 3:9 says no: time is reckoned differently in different cosmic locations.

That is not special relativity. It is also not nothing. It is the conceptual break, the rejection of universal simultaneity, that relativity, both special and general, would later vindicate. The text does not give the mechanism (gravity curving spacetime, velocity-dependent dilation); it gives the conceptual claim, and the conceptual claim contradicts the physics of 1842. By the leads-vs-follows test, the strong claim (Kolob predicts Einstein) is retrofit; the modest claim (Abraham 3:9 breaks with Newtonian universal time) is partial lead.

A second illustration of the same principle. Moses 1:33 and Abraham 3:3 speak of worlds without number, intelligences without number, and stars beyond counting. The strongest form of the apologetic argument here, that the Book of Abraham specifically predicts the Hubble-era discovery of two trillion galaxies in the observable universe, is retrofit: nineteenth-century cosmic-pluralism literature (Bruno, Fontenelle, Herschel, and Thomas Dick whose works are documented to have circulated in early Latter-day Saint contexts) already posited very large numbers of stars and worlds, and biblical idiom (Hebrews 11:12, Genesis 15:5) already provided innumerable as standard scriptural language for the heavenly host.

But here too a more modest claim survives. Every one of those 1830s sources operated with a finite number. Herschel posited approximately 600 million stars in the Milky Way; Dick calculated specific trillions of inhabitants; even Bruno's philosophical infinity was not taken by his contemporaries as empirical measurement. The actual post-Hubble universe contains on the order of 10 to the 23rd stars across approximately two trillion galaxies, none of which was known to mainstream science until 1923 at the earliest and not in remotely modern scale until the Hubble Deep Field images of the 1990s. The text's non-quantified language, without number, innumerable, turns out to match the empirical reality more closely than any 1835 quantified estimate did, because every quantified estimate underestimated by factors of billions. That is not a full prediction of Hubble cosmology, but it is more than retrofit, because the text declined to specify a number in an era when all serious sources specified numbers, and the declined specification turned out to be the right move.

Retrofits, properly defined, are unfalsifiable, any later discovery in any field can be mapped onto vague enough language by a motivated interpreter, and they therefore carry essentially no evidential weight. Partial leads, by contrast, are bounded but real: they should be rated honestly at [B] or [C], not promoted to [A] and not dismissed at [D]. We will see in Chapter 13 how this graded treatment looks in practice.

C. The principle stated

So the leads-vs-follows test, formally: a putative confirmation counts as positive evidence in proportion to (i) how specific the text's prior claim was on its plain reading, (ii) how independent the later confirming evidence was of the apologist's interpretive effort, and (iii) how unavailable the confirming information was to the author at the time of writing. A claim that scores high on all three dimensions is a real evidentiary asset. A claim that scores low on any of them is either neutral or, in the case of motivated retrofit, actively damaging to the apologist's credibility.

Throughout this book, I will apply this test before tagging an item as strong. The reader will see it work in real time. Nahom passes cleanly (Chapter 2). The metal-plates concept passes (Chapter 4). Several onomastic discoveries pass (Chapter 3). Other items produce a more interesting result: the strongest forms of the Kolob and worlds-without-number arguments fail the test as full predictions of relativity and modern cosmology, but modest readings of both pass partial versions of the test. Abraham 3:9 breaking with Newtonian universal time, and the text's non-quantified innumerability matching the post-Hubble universe in a way no 1835 quantified estimate did, both clear that lower bar. I will sort them carefully in Chapter 13. The same test, applied consistently, produces both a strong apologetic case and a credible one.

III. Bayesian Framing Without the Math

The second tool is a way of thinking about evidence that I will not formalize in equations but will rely on throughout the book. The thinking is Bayesian, which means, in plain English, that evidence does not simply prove or disprove a claim. Evidence shifts a probability. The amount of the shift depends on the evidence itself and on the starting position the evidence is being measured against.

A. Why evidence is never proof

Consider a fact about which all sides agree: the Book of Mormon mentions horses. Both the believer and the skeptic acknowledge that the noun horse appears in the text, in contexts that seem to describe practical animal use. The dispute is about what this fact means. To the skeptic, it is a fatal anachronism: horses were absent from the pre-Columbian Americas in the relevant period, therefore the text cannot be ancient. To the believer adopting the loan-shift defense, the fact is consistent with a translator using a familiar word for an unfamiliar Mesoamerican animal. The same fact updates two different priors in two different directions. Neither side is being dishonest. They are simply running the same evidence through different starting positions.

This is what evidence does. It does not produce verdicts; it adjusts confidences. A piece of evidence that seems decisive to one reader may seem trivial to another, not because either reader is irrational but because their starting positions assign different weights to the same observation. Recognizing this is not relativism. It is realism about how human reasoning under uncertainty actually works.

B. Priors are not biases. They are starting points

A common move in popular apologetic literature is to accuse the skeptic of bias for refusing to accept the text. A similarly common move on the skeptical side is to accuse the believer of bias for failing to abandon the text. Both moves misread what is actually happening. Bias, properly speaking, is reasoning improperly from evidence to conclusion. A prior is a starting confidence about a claim before evidence is consulted. The two are not the same.

It is not biased for a skeptic to start with the prior that miraculous translations of ancient golden plates do not occur. That is the default assumption of post-Enlightenment empirical inquiry, applied without prejudice to every claim involving the supernatural, to the resurrection of Christ, to the visions of Muhammad, to the Buddha's enlightenment, equally. Nor is it biased for a believer to start with the prior that personal religious witness provides reliable information about reality. That is the default assumption of essentially every functioning religious community in human history. These are different starting points, both internally consistent, and the same evidence will travel different distances from each.

What is biased, what is a methodological failure, is the move of treating any starting point as if it required no defense. The skeptic who treats methodological naturalism as obviously true, in a tradition that takes that question to be exactly the one in dispute, has helped himself to an assumption. The believer who treats personal witness as obviously decisive, in a conversation with someone who does not share that witness, has done the same. Honest engagement requires both sides to name their priors and to follow the evidence with those priors visible.

C. The Deutero-Isaiah example, revisited

The Prolog walked the reader through a single fact. The Book of Mormon's quotation of chapters of Isaiah that critical biblical scholarship dates to the Babylonian exile, after Lehi's departure from Jerusalem, and showed how three different readers receive it in three different ways. Let me state the lesson more formally here.

The Deutero-Isaiah objection is a hard disconfirmer of Book of Mormon historicity only if one accepts the prior assumption that genuine long-range predictive prophecy does not occur. The critical late-dating of Isaiah 40-55 rests substantially on that assumption: the exilic standpoint of those chapters, the announced return from captivity, the naming of Cyrus in chapter 45. These can mean either that the chapters were composed during the exile, or that an eighth-century Isaiah was given to foresee them. The first reading is the consensus of critical biblical scholarship, and is reached by applying methodological naturalism. The second reading is the traditional position, and is reached by allowing for predictive prophecy. Neither reading is internally incoherent. They are different conclusions from different priors operating on the same textual facts.

So the Deutero-Isaiah anachronism is fatal to Book of Mormon historicity for a reader with naturalistic priors. It is inert for a reader who accepts prophecy. And here is the important methodological move: it can do no independent work in deciding between those two prior sets, because it presupposes one of them. A disconfirmer that is hard only on the prior it presupposes is not a brute fact about the world; it is a reflection of the starting position chosen.

Consider what the prophecy-accepting prior actually requires the reader to accept, because it is not an abstract metaphysical commitment but a claim about empirically observable history. On Christmas Day 1832, twenty-eight years before the firing on Fort Sumter, Joseph Smith dictated a revelation later canonized as Doctrine and Covenants Section 87. It begins: Verily, thus saith the Lord concerning the wars that will shortly come to pass, beginning at the rebellion of South Carolina, which will eventually terminate in the death and misery of many souls. The revelation continues to describe the Southern States dividing against the Northern States, calling upon Great Britain, and slaves being marshalled and disciplined for war. In April 1861 the Civil War began at Fort Sumter in South Carolina, the first state to secede; the Confederacy pursued diplomatic recognition from Britain throughout the war; the conflict was, of course, fundamentally about slavery; and the casualties exceeded six hundred thousand.

The prophecy has discounts I will not hide. South Carolina was the epicenter of the Nullification Crisis in November 1832, weeks before Smith's revelation; tariff-related tension between Charleston and Washington was the political news of late 1832, and a general prediction that trouble would start in South Carolina was not historically arbitrary. The immediate Nullification Crisis was resolved peaceably by the Compromise Tariff in early 1833, within months of the prophecy, and the war that did come in 1861 was animated by different issues than the ones at stake in 1832. Critics have also pointed out that war was not, in the event, poured out upon all nations as the revelation states (D&C 87:2). The prophecy is not a clean unambiguous hit, and I would not present it as one.

But what matters for the present argument is more modest. The reader being asked to evaluate Deutero-Isaiah's prophetic credibility, to consider whether long-range predictive prophecy is the kind of thing that occurs in the world, has, internal to the same prophetic tradition, a concrete, documented, datable instance to weigh. The Civil War prophecy is, at minimum, a real partial hit at a twenty-eight-year lag in a specifically named location, and the text was demonstrably in circulation among Latter-day Saints decades before the war began. Whether it is decisive or merely suggestive, the abstract objection that long-range predictive prophecy is absolutely impossible is not consistent with the believer's own evidentiary record. The Deutero-Isaiah debate, in other words, is not adjudicated by treating one side's prior as obviously true and the other's as obviously false. It is adjudicated by weighing each side's prior alongside its evidentiary track record, and on that test, both sides have empirical work to do.

D. Prior-dependent vs. prior-independent disconfirmers

Not all disconfirmers are like Deutero-Isaiah. Some are hard on all prior sets. The clearest example, which is examined in detail in Chapter 13, is the appearance of italicized words from the 1769 King James Bible in the Book of Mormon's Isaiah passages. KJV translators set certain words in italics because those words were added to the English to smooth the translation. They correspond to nothing in the Hebrew. No mechanism of ancient transmission, oral or written, can deposit a seventeenth-century English translator's editorial insertions into a record that left Jerusalem in 600 BC. This problem is hard regardless of whether one accepts prophecy. It does not presuppose any particular metaphysics.