dabble deepDon't know, don't care, doing it anyway.
report
synced October 20, 2026

"And it came to pass": the Book of Mormon against Hebrew wayhi, English and Egyptian

Every count, test and method behind the "it came to pass" post, as generated by the bom-stylometry code.

Generated by the project's code and copied from e3b5789; only its links to code and data were pointed at the repo. CC BY 4.0.

Generated by uv run bomstylo report wayhi from the result files in results/wayhi/. The prose is in reports/templates/wayhi.md; every table and every number in it comes from a result file, so the report changes only when a result does. Licensed CC BY 4.0 (code: MIT).

Contents. 1. Summary · 2. Questions · 3. Sources · 4. Methods · 5. Results · 6. Corrections · 7. Limitations · 8. Reproducing

1. Summary

The Book of Mormon uses “it came to pass” far more often than the King James Bible. That is often offered as evidence of a Hebrew original, because the phrase translates the Hebrew narrative form wayhi. This report counts the Hebrew, not the KJV’s English. It compares the two texts genre for genre, and it splits each use by what follows the formula:

  • setting-first: a time or circumstance comes first (“and it came to pass, when they had eaten, that…”). This is how Hebrew wayhi works as a narrative marker.
  • straight to the event: the next thing that happened comes right after (“and it came to pass that they did march”). Hebrew wayhi doesn’t do this.
  • ordinary verb (Hebrew only): wayhi with its own subject (“and there was light”). The KJV renders it “it came to pass” in only 8 of 317 verses (2.5%), so it’s excluded.

Labels: registered results come from tests whose method and decision rule were written down and hashed before the measurement (preregistrations/). Everything else is exploratory. Each analysis’s headline, as recorded in Requiem:

  • Book of Mormon uses by bucket and genre step, current and 1830 texts (bom-buckets, exploratory).
    • Narrative frame, current text: 7.76 past-tense uses per 1,000 words (336 setting-first, 788 straight to the event, 144,801 words).
    • Narrative frame, 1830 text: 8.01 per 1,000 words.
    • Straight-to-the-event share, narrative frame, current text: 70% (67%–73%), 788 of 1,124.
    • Uses removed after 1830: 42 (12 past setting-first, 30 past straight-to-event, 0 future); current-text uses not found in the 1830 OCR by the aligner: 51 (an earlier hand check found these present but garbled or misaligned; its per-case labels weren’t kept).
  • Genre ladder: Hebrew setting-first wayhi vs the Book of Mormon (genre-ladder, exploratory).
    • Whole Old Testament vs whole Book of Mormon: 7.3x the Hebrew setting-first wayhi rate (7.9x the KJV’s English).
    • 17 Hebrew narrative units vs Book of Mormon narrative frame: 2.7x (2.3-3.1) current text, 2.7x (2.4-3.2) 1830 text, per 1,000 KJV-scale words.
    • Setting-first uses alone: Book of Mormon 2.32 vs Hebrew 2.92 per 1,000 = 0.80x (0.67-0.94). Hebrew has 0 straight-to-the-event wayhi; the Book of Mormon 788.
    • Densest Hebrew unit (setting-first): Elijah cycle at 5.81 per 1,000, against the Book of Mormon frame’s 7.76.
  • Egyptian narrative openings, and the Hebrew + Egyptian additive model (egyptian, exploratory).
    • Egyptian connected narratives matched in genre and date run 0–22 old-particle openings per 1,000 estimated translation words by text (Piye c. 727 BC: 8.26); they bracket the Book of Mormon’s 8.01 (1830 frame), so the rate test is inconclusive, not negative.
    • Hebrew settings + Demotic ḫpr: 5.7–6.2 per 1,000 against the 1830 frame’s 8.01; the Demotic texts fail on genre, so this does not decide the question.
    • pRylands 9 narration (c. 513 BC): 12 ‘it came to pass’ openings in 5,602 German words (2.14), 1 setting-first, 11 straight to the event.
  • English rule vs hand labels, 120 Book of Mormon cases (english-rule-audit, exploratory). English rule vs hand labels on 120 Book of Mormon cases: revised rule agrees on 119 (95%-100%), registered rule on 119. Disagreements (revised): Alma 49:28.
  • Event-first share in English comparison texts, the JST, Moses and Abraham (event-shares, exploratory).
    • Pastiche event-first shares (revised rule): Snowden 43% of 113, First Book of Napoleon 77% of 13, The Late War 29% of 77, Book of Nullification 38% of 16; rates 0.65–2.10 per 1,000 words, at most a third of the Book of Mormon’s.
    • KJV Old Testament: 4% event-first (14/388); New Testament 12%.
    • JST additions, revised rule, boundary-corrected: 38/47 event-first (81%); formulas kept from the KJV 5%.
    • Abraham (claims an Egyptian source) 15/16 event-first at 2.91 per 1,000; Moses 34/43 at 3.42. Fisher p = 0.26: no difference.
  • KJV parallels in the Book of Mormon, and the step at which each leaves the comparison (exclusions, exploratory). 39 KJV-parallel passages found; 3 keep verses in the narrative frame: 1 Nephi 22:20 (Acts 3, 1 verse, 73 words); 3 Nephi 10:4–10:6 (Matthew 23, 3 verses, 146 words); Mormon 9:22–9:24 (Mark 16, 3 verses, 110 words).
  • ‘It came to pass’ per verb and per source word (per-verb, exploratory).
    • Per 100 verbs, Book of Mormon ÷ Hebrew = 2.55× (2.2–2.9) current, 2.63× 1830 (2.7× per KJV word).
    • Validation: KJV has 1.026 verbs per Hebrew verb in the same passages, but 1.83 words per Hebrew word.
    • Additive model per 100 verbs: 3.60–4.16 against the 1830 book’s 5.60.
  • Greek New Testament ἐγένετο by bucket (greek-egeneto, exploratory).
    • Luke + Acts narrative frame: 2.16 per 1,000 KJV words, 10% straight to the event; the 1830 Book of Mormon is 3.7× (2.8–5.0) that rate.
    • Greek NT ἐγένετο: 56 setting-first, 8 straight to the event, 3 subordinate, 135 ordinary verb; English rule agrees on 52/52 rendered settings.
  • Forms of ‘and + hayah’ in the Hebrew Bible, and wayhi by bucket (hebrew-forms, exploratory).
    • Hebrew Bible: 784 wayhi by morphology (421 setting-first, 0 straight to the event, 363 ordinary verb); 357 wehayah.
    • Unpointed ויהי + והיה = 1,224, the only counting rule near 1,204; it pools past wayhi with future wehayah and their look-alikes.
  • KJV rendering of each Hebrew bucket, and calibration of the English rule (kjv-rendering, exploratory).
    • The KJV renders 84% of setting-first wayhi as ‘came to pass’ (334 of 396 single-form verses); ordinary-verb wayhi 2.5%; future setting-first wehayah 49%.
    • KJV past ‘it came to pass’ with no wayhi in the verse: 31 of 388 (8%).
    • English rule (revised) agrees with the Hebrew on 331 of 334 rendered past settings (registered rule 302 of 334); future settings 77 of 85.
  • Pre-registered: what the narrative particle introduces (PREREG 2) (prereg-event-share, registered). PREREG 2 (registered): Egyptian particles 1.00 event-first, Book of Mormon frame 0.71; gap 0.29 > 0.15. Verdict: MISS: the book does not match the Egyptian proportion
  • Pre-registered: is ‘it came to pass’ Smith’s all-purpose ‘then’? (PREREG 30, JST) (prereg-jst-then, registered).
    • PREREG 30 (registered): 52 added formulas in JST OT narrative, 39 event-first, S = 0.75 (0.61–0.86). Verdict: R1, the habit transfers.
    • Without the five boundary spill-overs (Amendment 2): 47 added, 38 event-first, S = 0.81 (0.67–0.91); verdict unchanged. Shows the idiom was Smith’s by 1830; whether he had it before the Book of Mormon is untested.
  • Published figures vs measured counts (published-claims, exploratory). Published vs measured: wayhi 1,204 vs 784; KJV Old Testament ‘came to pass’ 727 vs 391; Book of Mormon 1,404 vs 1,353.

In the narrative frame (the book’s own narration, without quotations and embedded speeches), the Book of Mormon uses the formula 7.76 times per 1,000 words (8.01 in the 1830 text): 336 setting-first and 788 straight to the event.

Most of the familiar gap is genre. Whole text against the whole Old Testament, the book uses the formula 7.3× as often as Hebrew uses setting-first wayhi. Narrative against matched narrative (17 Hebrew units), it is 2.7× (2.3–3.1) per 1,000 words and 2.55× (2.2–2.9) per 100 verbs.

The Hebrew-style use is at the Hebrew rate. Counted alone, the book’s setting-first uses run at 0.80× (0.67–0.94) the Hebrew rate in the same kind of text: a little below it, not above. All of the remaining difference is the straight-to-the-event use. Share of the book’s formulas that go straight to the event: 70% (67%–73%), 788/1,124.

Hebrew has no straight-to-the-event wayhi. None of the Hebrew Bible’s 421 narrative-marker wayhi goes straight to the event. The other sources measured on that use:

  • Egyptian narrative particles (ꜥḥꜥ.n, wn.ꞽn): every one leads into an event (438 of 438).
  • Demotic ḫpr “it came to pass”, Lehi-era sentences: 7 of 15 straight to the event; pRylands 9 (c. 513 BC): 11 of 12.
  • Joseph Smith’s own additions to KJV narrative in the JST: 81% (67%–91%), 38/47 straight to the event.
  • English imitations of biblical style written before the book: The Late War 29% (19%–40%), Snowden 43% (34%–53%), The First Book of Napoleon 77% (46%–95%) on only 13 uses. The KJV Old Testament: 4% (2%–6%).

The book’s share lies between a Hebrew-style use (none straight to the event) and an Egyptian-style use (all of them). It also lies between the KJV’s settings and Smith’s own later habit. The share alone can’t separate those readings. PREREG 2 tested whether the book matches the Egyptian particles alone, within 0.15: MISS: the book does not match the Egyptian proportion. It did not test a mixed Hebrew and Egyptian use. PREREG 30 tested whether Smith’s later habit is enough to explain the share: R1 HABIT TRANSFERS: H_then is a sufficient explanation of the event-first share. Whether he had the habit before dictating the book is untested.

Egyptian rates bracket the book. The Egyptian texts that match the book in genre and date (connected narratives composed before about 500 BC) run 0–22 openings per 1,000 estimated translation words, by text, around the book’s 8.01. So the Egyptian rate test is inconclusive: it neither supports nor rules out an Egyptian contribution. The Lehi-era Demotic texts are lower, but they fail on genre.

2. Questions and why each was asked

analysis label question why it was asked
bom-buckets exploratory How often does the Book of Mormon use the formula, setting-first vs straight to the event, at each genre step, in the current and 1830 texts? The Book of Mormon side of every comparison, cut the same way as the Hebrew side.
genre-ladder exploratory How does the Book of Mormon’s rate compare with Hebrew wayhi at each matched genre step, and which use of the formula carries the difference? Replaces the whole-Bible English-to-English ratio with a like-for-like comparison, showing the effect of each control.
egyptian exploratory Do Egyptian narratives open clauses with ‘then / it came to pass’ at a rate and mix that, alone or added to Hebrew, matches the Book of Mormon? Tests the ‘reformed Egyptian’ alternative on the texts that match in genre and date.
english-rule-audit exploratory How often does the English rule agree with a hand reading of Book of Mormon instances? Measures the rule’s error on the text it is used to classify, not only on the KJV.
event-shares exploratory Do English texts in biblical style reach the Book of Mormon’s event-first share and rate? English in biblical style is the competitor to an Egyptian source for the event-first share; it gets the same measurement.
exclusions exploratory Which Book of Mormon passages run parallel to the KJV, and does any stay in the narrative frame? Checks the exclusion annotation against the texts, so no KJV text is counted as the book’s narration.
per-verb exploratory Do the Hebrew ratio and the additive model survive when the unit is verbs or source words rather than English words? English word counts don’t translate between languages; verbs roughly do.
greek-egeneto exploratory How often do the Gospels and Acts use the formula, by bucket, against the Book of Mormon? Luke-Acts is the nearest non-Hebrew model of a heavy ‘it came to pass’ style, and the hand labels test the English rule on a second language.
hebrew-forms exploratory How many wayhi, wehayah and look-alike forms are there, and how do wayhi and wehayah split into setting-first, straight-to-event and ordinary-verb uses? Separates forms that are easy to pool by accident, which is where published counts diverge.
kjv-rendering exploratory How consistently did the KJV render each kind of wayhi/wehayah as ‘came to pass’, and how often does the English rule recover the Hebrew bucket? Tests the ‘translators varied it’ premise, and measures the English rule’s error rate on text whose bucket is known.
prereg-event-share registered Does the Book of Mormon’s event/time split of ‘it came to pass’ match the Egyptian narrative particles (within 0.15), and do English controls fail to? The registered test of the ‘Egyptian particle’ reading of the event-first share.
prereg-jst-then registered When Joseph Smith added ‘it came to pass’ to the KJV in the JST, was it event-first at the Book of Mormon’s share? Tests the main competitor to a non-English source: Smith’s own English, in his later scripture.
published-claims exploratory Do the published counts (Parry 1992/2020) reproduce from the texts? Shows where the familiar numbers come from before the comparison replaces them.

3. Sources

No source text is in this repository. Each is downloaded from a pinned address, checked by sha256, and prepared by a named step. The same entries in corpora/manifest.yaml generate this list.

bom-current: The Book of Mormon (Project Gutenberg eBook #17, 13th edition, mormon13.txt).

  • Edition: Text of the 1981 LDS edition (e.g. 2 Nephi 30:6 “pure and delightsome”), as released by Project Gutenberg on 8 March 1992 and last updated 22 May 2005.
  • Cite as: The Book of Mormon. Project Gutenberg eBook #17, file mormon13.txt (13th edition, last updated 22 May 2005). https://www.gutenberg.org/files/17/old/mormon13.txt
  • Licence: Project Gutenberg licence (US public domain etext)
  • Download: https://www.gutenberg.org/files/17/old/mormon13.txt (sha256 192d7b51bcaaac46…)
  • Preparation: step bom_current in src/bomstylo/prepare.py; prepared files pinned in corpora/checksums/bom-current.sha256.
  • Notes: The current PG #17 files and the 2024–25 Wayback snapshots have a different layout and don’t reproduce this text; use the old/ file.
  • Read by: bom-buckets, exclusions, per-verb, prereg-event-share, published-claims

kjv: King James Version (Project Gutenberg eBook #10).

bom-1830: The Book of Mormon, first edition (Palmyra, 1830), archive.org OCR.

  • Edition: archive.org item bookofmormonacco1830smit (University of Pittsburgh Library System scan), uncorrected OCR (_djvu.txt).
  • Cite as: Smith, Joseph, Jr. The Book of Mormon. Palmyra: E. B. Grandin, 1830. Scan and OCR, Internet Archive, https://archive.org/details/bookofmormonacco1830smit
  • Licence: Public domain by date (no licence stated by archive.org)
  • Download: https://archive.org/download/bookofmormonacco1830smit/bookofmormonacco1830smit_djvu.txt (sha256 06c116b20e7d2d52…)
  • Preparation: step copy in src/bomstylo/prepare.py; prepared files pinned in corpora/checksums/bom-1830.sha256.
  • Notes: archive.org has no versioned URLs and re-derives OCR occasionally; its own sha1 for this file is f57b45ffb072ac896cb3199e53254a31ee705f70. No Wayback snapshot exists.
  • Read by: bom-buckets

wlc: Westminster Leningrad Codex with OpenScriptures morphology (morphhb).

  • Edition: openscriptures/morphhb, commit 3d15126fb1ef74867fc1434be1942e837932691f (wlc/ last changed in 8de8960, 2022).
  • Cite as: Open Scriptures Hebrew Bible (morphhb), Westminster Leningrad Codex with lemma and morphology tagging. https://github.com/openscriptures/morphhb
  • Licence: OSHB tagging CC BY 4.0 (“Original work of the Open Scriptures Hebrew Bible available at https://github.com/openscriptures/morphhb”); WLC text public domain
  • Files: 39 from https://raw.githubusercontent.com/openscriptures/morphhb/3d15126fb1ef74867fc1434be1942e837932691f/wlc/{file}
  • Preparation: step copy in src/bomstylo/prepare.py; prepared files pinned in corpora/checksums/wlc.sha256.
  • Read by: genre-ladder, per-verb, hebrew-forms, kjv-rendering, prereg-event-share

web: World English Bible, read-aloud edition (eBible.org).

  • Edition: eng-web_readaloud.zip as generated 19 August 2026; the chapter files of Genesis–Esther, Jonah, the Gospels and Acts (557 files).
  • Cite as: World English Bible (read-aloud edition). eBible.org. https://ebible.org/Scriptures/eng-web_readaloud.zip
  • Licence: Public domain (“World English Bible” is a trademark of eBible.org)
  • Download: https://ebible.org/Scriptures/eng-web_readaloud.zip
  • Preparation: step web in src/bomstylo/prepare.py; prepared files pinned in corpora/checksums/web.sha256.
  • Notes: The zip is unversioned and regenerated by eBible; no copy of the 19 Aug 2026 build is archived. The 18 Oct 2026 zip (sha256 1b5e9d5d…6bc0) differs from it in one used file, John 3:2, restored by patch (2 words in John’s narration; no formula). A later zip may differ more: verify will say which files.
  • Read by: genre-ladder, per-verb, greek-egeneto

sblgnt: SBL Greek New Testament with MorphGNT parsing.

  • Edition: morphgnt/sblgnt, commit aaed91e57c8e4a8dc9a2383e129ca5e75fe6393d (content as of b4d1e66, 2023). Not tag 6.12, which differs in 9 books.
  • Cite as: Holmes, Michael W., ed. The Greek New Testament: SBL Edition. Society of Biblical Literature and Logos Bible Software, 2010; MorphGNT parsing by James K. Tauber, https://github.com/morphgnt/sblgnt
  • Licence: SBLGNT text under the SBLGNT EULA (http://sblgnt.com/license/); morphology CC BY-SA 3.0
  • Files: 27 from https://raw.githubusercontent.com/morphgnt/sblgnt/aaed91e57c8e4a8dc9a2383e129ca5e75fe6393d/{file}
  • Preparation: step copy in src/bomstylo/prepare.py; prepared files pinned in corpora/checksums/sblgnt.sha256.
  • Read by: greek-egeneto

tla: Thesaurus Linguae Aegyptiae open sentence subsets (Hugging Face).

aed-tei: AED-TEI named Egyptian texts (Ancient Egyptian Dictionary, TEI export).

  • Edition: simondschweitzer/aed-tei, commit 462c722e0323e05641aea2eee8cdf1e27303d939 (files as of 0817b82, 2019); the 21 texts used.
  • Cite as: Schweitzer, Simon D. aed-tei: Ancient Egyptian texts of the Thesaurus Linguae Aegyptiae in TEI XML. https://github.com/simondschweitzer/aed-tei
  • Licence: CC BY-SA 4.0
  • Files: 21 from https://raw.githubusercontent.com/simondschweitzer/aed-tei/462c722e0323e05641aea2eee8cdf1e27303d939/files/{file}
  • Preparation: step copy in src/bomstylo/prepare.py; prepared files pinned in corpora/checksums/aed-tei.sha256.
  • Read by: egyptian, prereg-event-share

jst-1867: The Holy Scriptures, Inspired Version (1867), archive.org OCR.

  • Edition: archive.org item holyscriptures00smit (BYU Harold B. Lee Library scan), uncorrected OCR (_djvu.txt).
  • Cite as: The Holy Scriptures, Translated and Corrected by the Spirit of Revelation, by Joseph Smith, Jr., the Seer. Plano, Ill.: Reorganized Church of Jesus Christ of Latter Day Saints, 1867. Internet Archive, https://archive.org/details/holyscriptures00smit
  • Licence: Public domain by date (no licence stated by archive.org)
  • Download: https://archive.org/download/holyscriptures00smit/holyscriptures00smit_djvu.txt (sha256 02f3b8c19f76e6f6…)
  • Preparation: step copy in src/bomstylo/prepare.py; prepared files pinned in corpora/checksums/jst-1867.sha256.
  • Notes: bomstylo.jst segments this file by line number (GENESIS_START, NT_FROM), so a re-OCR by archive.org breaks the registered test. archive.org sha1: e12af71a26ce18695aa7fe9920aca0fa4bf2cf38.
  • Read by: event-shares, prereg-jst-then

late-war: Gilbert J. Hunt, The Late War (1819), archive.org OCR.

  • Edition: archive.org item latewarbetweenun00inhunt (University of Pittsburgh scan), uncorrected OCR.
  • Cite as: Hunt, Gilbert J. The Late War, between the United States and Great Britain, from June 1812, to February 1815. Written in the Ancient Historical Style. New York, 1819. Internet Archive, https://archive.org/details/latewarbetweenun00inhunt
  • Licence: Public domain by date (no licence stated by archive.org)
  • Download: https://archive.org/download/latewarbetweenun00inhunt/latewarbetweenun00inhunt_djvu.txt (sha256 589afbc822cfc513…)
  • Preparation: step copy in src/bomstylo/prepare.py; prepared files pinned in corpora/checksums/late-war.sha256.
  • Notes: archive.org sha1 5027b8d63d9ec29bda6da9b76ac347db93880646. Cleaned at read time (bomstylo.controls).
  • Read by: event-shares, prereg-event-share

period-pastiches: Snowden (1793), The First Book of Napoleon (1809), The Book of Nullification (1830).

moses-abraham: Book of Moses and Book of Abraham (Pearl of Great Price), bcbooks/scriptures-json.

4. Methods

4.1 Buckets and rules

Every use is put in a bucket by a rule that looks at the words right after the formula.

  • Hebrew rule (src/bomstylo/hebrew.py, classify). It reads the lemma and morphology of the next word or two:

    • setting-first if a setting word follows (“after”, “when”, “as”, “end”, “time”, “if”, “until”, “before”), or a time noun after a preposition or numeral (“on the day”, “at the end of days”), or a circumstantial clause (subject, then a participle, then a main verb later in the verse);
    • straight to the event if a finite verb follows directly (it never does with wayhi; the bucket exists so the absence is measured, not assumed);
    • ordinary verb otherwise.

    Forms: wayhi is morphology HC/Vqw3ms of lemma 1961 (hayah); its future counterpart wehayah (“and it shall come to pass”) is HC/Vqq3ms.

  • English rule (src/bomstylo/english.py). It finds “it came to pass” and “it shall / should / will come to pass”, and reads the next 40 characters. Two versions are kept and both are reported:

    • registered: PREREG 2’s rule, unchanged. A short list of time words after the formula.
    • revised: the registered list plus settings that the calibrations below showed it missing (“the same day”, “in process of time”, “one day”…), relative conditions (“whosoever…”), and a setting placed just before the formula in the same clause (“And when the feast was ended, it came to pass…”). Every addition is a grammatical setting found by checking the rule against the KJV, where the Hebrew (or Greek) bucket is known. None was tuned on the Book of Mormon.

  • Calibration. The English rule can only be trusted as far as it agrees with a reading whose answer is known:

English rule KJV past settings (Hebrew known) KJV future settings Book of Mormon hand audit (120)
revised 331/334 (99%) 77/85 (91%) 119/120 (95%–100%)
registered 302/334 (90%) 71/85 (84%) 119/120 (95%–100%)

On the Greek: the revised rule agrees on 52 of 52 setting-first ἐγένετο the KJV renders with the phrase.

4.2 Genre steps

The often-quoted comparison sets the whole Book of Mormon against the whole KJV Old Testament. That mixes in poetry, law and prophecy, where the construction barely occurs. Here both sides are cut the same way, one step at a time, and every step is reported:

step Old Testament side Book of Mormon side
1 whole Old Testament whole text
2 17 narrative books (Genesis–Esther) minus the KJV block quotations
3 story books: minus Leviticus, Numbers, Deuteronomy (mostly law and address) minus the KJV block quotations
4 story books, speech removed narrative frame: minus embedded speeches, letters, prophecies and records
5 17 named narrative units, speech removed narrative frame
  • Hebrew speech is removed per instance. The KJV marks no quotations, so each Hebrew instance is placed at the same relative position in the World English Bible verse and checked against the WEB’s quotation marks. Narrative-frame word counts use each verse’s share of words outside quotes.
  • The 17 units are connected narrative blocks as conventionally delimited, chapter-aligned. Whole books are too coarse: Genesis has genealogies, Exodus turns into law at chapter 20, Kings is padded with regnal formulae. Daniel is left out on purpose: its court tales are close analogues, but Daniel 2–7 is Aramaic, which has no wayhi, so it would read as a false zero.
unit chapters frame words (KJV scale) setting-first wayhi ordinary-verb wayhi setting-first per 1,000
Abraham cycle Genesis 12–25 5,541 15 10 2.71
Jacob cycle Genesis 26–36 5,314 13 8 2.45
Joseph novella Genesis 37–50 5,991 20 10 3.34
Exodus narrative Exodus 1–19 6,589 17 13 2.58
Conquest Joshua 1–12 6,314 23 4 3.64
Judges tales Judges 3–16 7,243 22 10 3.04
Ruth Ruth 1–4 1,090 4 1 3.67
Samuel and Saul 1 Samuel 1–15 6,043 17 10 2.81
Rise of David 1 Samuel 16–31 6,878 19 9 2.76
Court narrative 2 Samuel 9–20 5,829 16 4 2.75
Solomon 1 Kings 1–11 7,574 11 15 1.45
Elijah cycle 1 Kings 17–22 3,616 21 8 5.81
Elisha cycle 2 Kings 2–13 7,158 29 7 4.05
Hezekiah narrative 2 Kings 18–20 1,534 5 1 3.26
Esther Esther 1–10 4,196 5 1 1.19
Nehemiah memoir Nehemiah 1–7 3,765 11 1 2.92
Jonah Jonah 1–4 673 1 4 1.49

4.3 What the Book of Mormon side leaves out, and why

The comparison is about the book’s own narration, so text the book takes from the KJV must not be counted as its own. The layers are data in annotations/embedded.tsv, read by src/bomstylo/spans.py, not logic inside an analysis.

  • KJV block quotations leave from step 2. These are the long Isaiah and Malachi chapters, which follow the KJV nearly word for word:
KJV block quotation (layer quotation) out of the comparison
1 Nephi 20:1–1 Nephi 21:26 step 2 onward
2 Nephi 7:1–2 Nephi 8:25 step 2 onward
2 Nephi 12:1–2 Nephi 24:32 step 2 onward
Mosiah 14:1–Mosiah 14:12 step 2 onward
3 Nephi 22:1–3 Nephi 22:17 step 2 onward
3 Nephi 24:1–3 Nephi 25:6 step 2 onward

The annotation has 73 spans: 6 KJV block quotations and 67 embedded speeches, letters, prophecies and records (1 allegory, 46 discourse, 9 epistle, 3 liturgy, 1 narration, 7 prophecy).

  • Embedded speech leaves at step 4: sermons, letters, prophecies and embedded records, as the book itself frames them. This also removes the KJV parallels inside speech, such as the Sermon at the Temple (3 Nephi 12–14, Matthew 5–7), Isaiah 29 in 2 Nephi 27, Isaiah 52 and Micah in 3 Nephi 20–21, the Decalogue in Mosiah 12–13 and 1 Corinthians 13 in Moroni 7.
  • Zeniff’s record (Mosiah 9–10, the first-person part of the record the book embeds; chapters 11–22 switch to third person and stay in the frame) is counted as narrative frame. It is a first-person account of events, like Nephi’s own, which is also frame, and PREREG 2 registered it as frame. The earlier published figures counted it as speech; the frame figures with it counted that way are reported in bom-buckets.
  • Checked against the texts. The exclusions analysis finds every passage that shares long word runs with the KJV, and says at which step it leaves. Passages that stay in the frame are flagged in bold:
Book of Mormon KJV (best match) verses leaves the comparison at note
1 Nephi 20:1–21:26 Isaiah 48–49 47 step 2 (without KJV quotations)
1 Nephi 22:20 Acts 3 1 stays in the narrative frame all verse in frame, 73 words
2 Nephi 6:6–6:7 Isaiah 49 2 step 4 (narrative frame)
2 Nephi 6:16–8:25 Isaiah 49–52 35 step 2 (without KJV quotations)
2 Nephi 9:50 Isaiah 55 1 step 4 (narrative frame)
2 Nephi 12:1–24:32 Isaiah 2–14 261 step 2 (without KJV quotations)
2 Nephi 26:4 Malachi 4 1 step 4 (narrative frame)
2 Nephi 27:2–27:5 Isaiah 29 3 step 4 (narrative frame)
2 Nephi 27:25–27:35 Isaiah 29 11 step 4 (narrative frame)
2 Nephi 30:9–30:15 Isaiah 11 6 step 4 (narrative frame)
Mosiah 12:21–12:24 Isaiah 52 4 step 4 (narrative frame)
Mosiah 12:34–12:36 Exodus 20 2 step 4 (narrative frame)
Mosiah 13:13 Deuteronomy 5 1 step 4 (narrative frame)
Mosiah 13:14–13:20 Exodus 20 5 step 4 (narrative frame)
Mosiah 14:1–14:12 Isaiah 53 12 step 2 (without KJV quotations)
Mosiah 15:16–15:18 Isaiah 52 3 step 4 (narrative frame)
Mosiah 15:29–15:31 Isaiah 52 3 step 4 (narrative frame)
Alma 42:2–42:3 Genesis 3 2 step 4 (narrative frame)
Helaman 10:7–10:9 Matthew 18–21 2 step 4 (narrative frame)
3 Nephi 10:4–10:6 Matthew 23 3 stays in the narrative frame all verses in frame, 146 words
3 Nephi 11:25 Matthew 28 1 step 4 (narrative frame)
3 Nephi 12:5–12:45 Matthew 5 32 step 4 (narrative frame)
3 Nephi 13:1–15:2 Matthew 6–13 56 step 4 (narrative frame)
3 Nephi 15:17 John 10 1 step 4 (narrative frame)
3 Nephi 15:21 John 10 1 step 4 (narrative frame)
3 Nephi 16:15 Matthew 5 1 step 4 (narrative frame)
3 Nephi 16:18–16:20 Isaiah 52 3 step 4 (narrative frame)
3 Nephi 20:16–20:19 Micah 4–5 3 step 4 (narrative frame)
3 Nephi 20:23–20:27 Acts 3 5 step 4 (narrative frame)
3 Nephi 20:32–20:45 Isaiah 52 12 step 4 (narrative frame)
3 Nephi 21:8 Isaiah 52 1 step 4 (narrative frame)
3 Nephi 21:12–21:21 Micah 5 6 step 4 (narrative frame)
3 Nephi 22:1–22:17 Isaiah 54 17 step 2 (without KJV quotations)
3 Nephi 24:1–25:6 Malachi 3–4 24 step 2 (without KJV quotations)
3 Nephi 27:33 Matthew 7 1 step 4 (narrative frame)
Mormon 9:22–9:24 Mark 16 3 stays in the narrative frame all verses in frame, 110 words
Ether 4:18 Mark 16 1 step 4 (narrative frame)
Moroni 7:45 1 Corinthians 13 1 step 4 (narrative frame)
Moroni 7:48 1 John 3 1 step 4 (narrative frame)

Each passage is placed by the KJV book with the most shared runs. 1 Nephi 22:20 quotes the prophet-like-Moses saying, which the KJV has in Deuteronomy 18 and, nearest in wording, Acts 3.

4.4 Units and statistics

  • Per 1,000 words. Book of Mormon rates use its own English words. Hebrew rates use the KJV’s English words for the same verses (narrative frame: the WEB share outside quotes), so both sides are on an English scale. Demotic rates use words of the TLA’s German translation; Egyptian token counts are scaled by the translation-words-per-token ratio measured in each TLA subset. These are not the same unit across languages. per-verb repeats the comparisons per 100 verbs, counted in each language’s own text, and checks that verbs (unlike words) survive translation.
  • Intervals are exact and 95%: Poisson (Garwood) for rates, Clopper–Pearson for shares, and the conditional binomial interval for rate ratios (src/bomstylo/stats.py).
  • Comparability of Egyptian texts. An Egyptian rate counts for or against the book only against a text that matches it in genre (connected narrative) and date (composed about 600–500 BC or earlier; older is fine, because formal Egyptian kept the older literary register and Saite scribes revived it). Every Egyptian row carries date, genre and matched flags. When no matched text decides the question, the result is called inconclusive.

5. Results

5.1 Published figures

Question. Do the published counts (Parry 1992/2020) reproduce from the texts?
Why. Shows where the familiar numbers come from before the comparison replaces them.
Label. exploratory.
Code. analyses/wayhi/published_claims.py; result results/wayhi/published-claims.json, fingerprint 9af674e0790f; Requiem record results/wayhi/published-claims/headline.

published figure published measured note
wayhi in the Hebrew Bible 1,204 784 unpointed ויהי + והיה = 1,224
‘came to pass’ in the KJV Old Testament 727 391 535 with ‘come to pass’; 622 in the whole KJV, any tense
‘it came to pass’ in the Book of Mormon 1,404 1,353 current text
in the original text 1,494 — manuscript-level count; outside this check (see bom-buckets for the 1830 edition)

The phrase in passages the same literature names:

passage words ‘it came to pass’ per 1,000
Psalm of Nephi (2 Ne 4:16-35) 684 0 0.00
King Benjamin (Mosiah 2-5) 5,573 6 1.08
Abinadi (Mosiah 12-16) 4,368 8 1.83
Alma 5 2,787 1 0.36
Alma 32-34 4,268 3 0.70
Jesus (3 Nephi 11-18) 8,052 30 3.73
Moroni 7-10 5,154 0 0.00
Alma war chapters (43-62) 29,423 231 7.85
1 Nephi (whole book) 25,319 200 7.90
Ether (whole book) 16,661 164 9.84

5.2 Forms of wayhi in the Hebrew Bible

Question. How many wayhi, wehayah and look-alike forms are there, and how do wayhi and wehayah split into setting-first, straight-to-event and ordinary-verb uses?
Why. Separates forms that are easy to pool by accident, which is where published counts diverge.
Label. exploratory.
Code. analyses/wayhi/hebrew_forms.py; result results/wayhi/hebrew-forms.json, fingerprint 7ace071f7a44; Requiem record results/wayhi/hebrew-forms/headline.

The forms are easy to pool by accident: unpointed ויהי is also wihi (“and let it be”), and unpointed והיה is also plain “and” + perfect. Pooling the past with the future (wayhi + wehayah = 1,141) compares two different constructions with one English phrase.

counting rule count
wayhi only (morphology) 784
wayhi + wihi (unpointed ויהי) 816
wayhi + wehayah (morphology) 1,141
unpointed ויהי + unpointed והיה 1,224
all 3rd-person wayyiqtol (wayhi, wayyihyu, wattehi) 975
every hayah form with ‘and’ 1,775
form (morphology) meaning count
HC/Vqw3ms wayhi (and it was / and it came to pass) 784
HC/Vqq3ms wehayah (and it shall be / shall come to pass) 357
HC/Vqw3mp wayyihyu (and they were) 105
HC/Vqq3cp wehayu (and they shall be) 98
HC/Vqq3fs wehayetah (and she/it shall be) 87
HC/Vqw3fs wattehi (and she/it [fem.] was) 86
HC/Vqj3ms wihi (and let it be; jussive) 27
HC/Vqw1cs wa’ehi (and I was) 23
HC/Vqw3fp wattihyenah (and they [fem.] were) 10
HC/Vqi3ms weyihyeh (and it will be; plain imperfect) 5
HC/Vqw1cp wanniheyeh (and we were) 4
HC/Vqw2ms wattehi (and you were) 1
29 other forms plain (non-consecutive) perfects, participles, imperatives, infinitives and other persons; counted only by the every-form rule 188

By bucket (Hebrew rule), past = wayhi, future = wehayah:

setting-first straight to event ordinary verb
past 421 0 363
future 184 6 167

5.3 How the KJV rendered each kind

Question. How consistently did the KJV render each kind of wayhi/wehayah as ‘came to pass’, and how often does the English rule recover the Hebrew bucket?
Why. Tests the ‘translators varied it’ premise, and measures the English rule’s error rate on text whose bucket is known.
Label. exploratory.
Code. analyses/wayhi/kjv_rendering.py; result results/wayhi/kjv-rendering.json, fingerprint f5452bc09697; Requiem record results/wayhi/kjv-rendering/headline.

Hebrew form and bucket single-form verses KJV says ‘came / shall come to pass’ share
wayhi, setting-first 396 334 84.3%
wayhi, ordinary verb 317 8 2.5%
wehayah, setting-first 172 85 49.4%
wehayah, ordinary verb 148 7 4.7%
wehayah, straight to event 5 2 40.0%

KJV past ‘it came to pass’ with no wayhi in the verse: 31 of 388 (8%). Verses whose Hebrew and English numbering differ: 12 (left out).

5.4 The English rule on the Book of Mormon itself

Question. How often does the English rule agree with a hand reading of Book of Mormon instances?
Why. Measures the rule’s error on the text it is used to classify, not only on the KJV.
Label. exploratory.
Code. analyses/wayhi/english_rule_audit.py; result results/wayhi/english-rule-audit.json, fingerprint 229ef35b2450; Requiem record results/wayhi/english-rule-audit/headline.

rule agrees share (95% CI) disagreements
revised 119/120 99% (95%–100%) Alma 49:28
registered 119/120 99% (95%–100%) Alma 49:28

Hand labels: 44 setting-first, 76 straight to the event.

5.5 The Book of Mormon by bucket and genre level

Question. How often does the Book of Mormon use the formula, setting-first vs straight to the event, at each genre step, in the current and 1830 texts?
Why. The Book of Mormon side of every comparison, cut the same way as the Hebrew side.
Label. exploratory.
Code. analyses/wayhi/bom_buckets.py; result results/wayhi/bom-buckets.json, fingerprint e1f0dbbc663e; Requiem record results/wayhi/bom-buckets/headline.

Current text, revised rule:

genre level words past, setting-first past, straight to event future (all) past per 1,000
whole text 268,275 412 940 61 5.04
without KJV quotations 256,149 410 938 46 5.26
narrative frame 144,801 336 788 10 7.76
speech (not KJV) 111,348 74 150 36 2.01

1830 text (current + uses removed after 1830), revised rule:

genre level words past, setting-first past, straight to event future (all) past per 1,000
whole text 268,443 424 970 61 5.19
without KJV quotations 256,317 422 968 46 5.42
narrative frame 144,949 348 813 10 8.01
speech (not KJV) 111,368 74 155 36 2.06
straight-to-event share value
current 70% (67%–73%), 788/1,124
1830 70% (67%–73%), 813/1,161
current, registered rule 71% (68%–74%), 800/1,124
speech, future 58% (41%–74%), 21/36

By narrator (narrative frame, current text):

narrator past uses straight to event (revised) (registered)
Nephi1 151 75% (67%–82%), 113/151 75% (68%–82%), 114/151
Jacob 14 71% (42%–92%), 10/14 71% (42%–92%), 10/14
Enos 6 67% (22%–96%), 4/6 67% (22%–96%), 4/6
Jarom 4 100% (40%–100%), 4/4 100% (40%–100%), 4/4
Omni 2 100% (16%–100%), 2/2 100% (16%–100%), 2/2
Amaron 2 100% (16%–100%), 2/2 100% (16%–100%), 2/2
Abinadom 1 100% (2%–100%), 1/1 100% (2%–100%), 1/1
Amaleki 8 75% (35%–97%), 6/8 75% (35%–97%), 6/8
Mormon 780 66% (62%–69%), 513/780 67% (64%–70%), 524/780
Moroni 156 85% (79%–90%), 133/156 85% (79%–90%), 133/156

1830 alignment: of 1,451 past uses in the 1830 OCR, 1,288 are in the current text, 42 were removed after 1830 (12 setting-first, 30 straight to the event), and 121 could not be aligned because of OCR damage (treated as present). Reverse direction: 51 current-text uses were not found in the 1830 OCR by the aligner (89 more unalignable).

Sensitivity, Zeniff’s first-person record (Mosiah 9–10) counted as speech: 142,980 words, 335 setting-first, 775 straight to the event (7.76 per 1,000).

5.6 The genre ladder

Question. How does the Book of Mormon’s rate compare with Hebrew wayhi at each matched genre step, and which use of the formula carries the difference?
Why. Replaces the whole-Bible English-to-English ratio with a like-for-like comparison, showing the effect of each control.
Label. exploratory.
Code. analyses/wayhi/genre_ladder.py; result results/wayhi/genre-ladder.json, fingerprint aa8a7d75317f; Requiem record results/wayhi/genre-ladder/headline.

Rates per 1,000 words (KJV-scale for Hebrew). The ratio compares the Book of Mormon’s past formula with Hebrew setting-first wayhi.

Hebrew side Book of Mormon side KJV English per 1,000 Hebrew setting-first per 1,000 BoM per 1,000 BoM 1830 ratio (current) ratio (1830)
whole OT whole text 0.64 0.69 5.04 5.19 7.3× (6.5–8.2) 7.5× (6.8–8.4)
17 narrative books without KJV quotations 0.94 1.02 5.26 5.42 5.2× (4.6–5.8) 5.3× (4.7–6.0)
story books (no Lev/Num/Deut) without KJV quotations 1.18 1.28 5.26 5.42 4.1× (3.7–4.7) 4.3× (3.8–4.8)
story books, narrative frame narrative frame 1.86 2.02 7.76 8.01 3.8× (3.4–4.4) 4.0× (3.5–4.5)
17 narrative units, narrative frame narrative frame 2.71 2.92 7.76 8.01 2.7× (2.3–3.1) 2.7× (2.4–3.2)

Narrative frame by bucket (17 Hebrew units vs the Book of Mormon frame), per 1,000 words:

bucket Hebrew Book of Mormon BoM 1830 BoM ÷ Hebrew
past, setting-first 249 = 2.92 (2.57–3.30) 336 = 2.32 348 = 2.40 0.80× (0.67–0.94)
past, straight to event 0 = 0.00 (0.00–0.04) 788 = 5.44 813 = 5.61 —
past, ordinary verb 116 = 1.36 (1.12–1.63) 0 = 0.00 0 = 0.00 0.00× (0.00–0.02)
future, setting-first 44 = 0.41 (0.30–0.55) 15 = 0.13 15 = 0.13 0.33× (0.17–0.61)
future, straight to event 0 = 0.00 (0.00–0.03) 21 = 0.19 21 = 0.19 —
future, ordinary verb 37 = 0.34 (0.24–0.47) 0 = 0.00 0 = 0.00 0.00× (0.00–0.10)

Hebrew instances whose verse the WEB numbers differently (counted as frame): 11. KJV Old Testament English, straight-to-the-event share: 4% (2%–6%), 14/388.

5.7 Per verb and per source word

Question. Do the Hebrew ratio and the additive model survive when the unit is verbs or source words rather than English words?
Why. English word counts don’t translate between languages; verbs roughly do.
Label. exploratory.
Code. analyses/wayhi/per_verb.py; result results/wayhi/per-verb.json, fingerprint c2060ffad677; Requiem record results/wayhi/per-verb/headline.

per 100 verbs, past, narrative frame setting-first straight to event total
Hebrew 2.13 (1.87–2.41) 0.00 (0.00–0.03) 2.13 (1.87–2.41)
Book of Mormon (current) 1.62 (1.45–1.81) 3.81 (3.55–4.08) 5.43 (5.12–5.76)
Book of Mormon (1830) 1.68 (1.51–1.87) 3.92 (3.66–4.20) 5.60 (5.28–5.93)
demotic loose, all dates 0.78 (0.45–1.24) 0.96 (0.59–1.46) 1.73 (1.23–2.38)
Hebrew + demotic loose, all dates 2.91 (2.58–3.37) 0.96 (0.59–1.46) 3.86 (3.36–4.51)
demotic loose, 700–300 BC 0.89 (0.43–1.64) 0.63 (0.25–1.29) 1.52 (0.88–2.43)
Hebrew + demotic loose, 700–300 BC 3.02 (2.56–3.77) 0.63 (0.25–1.29) 3.65 (3.02–4.56)
demotic strict, all dates 0.84 (0.47–1.39) 1.18 (0.73–1.81) 2.03 (1.42–2.80)
Hebrew + demotic strict, all dates 2.97 (2.60–3.52) 1.18 (0.73–1.81) 4.16 (3.55–4.93)
demotic strict, 700–300 BC 0.91 (0.39–1.79) 0.79 (0.32–1.64) 1.70 (0.95–2.81)
Hebrew + demotic strict, 700–300 BC 3.04 (2.52–3.92) 0.79 (0.32–1.64) 3.83 (3.08–4.94)
pRylands 9 0.12 (0.00–0.68) 1.35 (0.67–2.41) 1.47 (0.76–2.57)
Hebrew + pRylands 9 2.25 (2.13–2.81) 1.35 (0.67–2.41) 3.60 (2.89–4.70)

Check: in the same 17 units the KJV has 1.026 verbs per Hebrew verb but 1.83 words per Hebrew word. Words per verb: KJV (17 units) 7.12; Book of Mormon 6.99; German, demotic loose, all dates 6.09; German, demotic loose, 700–300 BC 5.55; German, demotic strict, all dates 6.21; German, demotic strict, 700–300 BC 5.58; German, pRylands 9 6.87.

Per 1,000 words, by which words are counted:

rate per 1,000
Hebrew per 1,000 Hebrew words 5.35
Hebrew per 1,000 KJV words 2.92
BoM 1830 per 1,000 words 8.01
demotic loose, all dates per 1,000 Demotic words 3.10
demotic loose, all dates per 1,000 German words 2.84
demotic loose, 700–300 BC per 1,000 Demotic words 2.91
demotic loose, 700–300 BC per 1,000 German words 2.74
demotic strict, all dates per 1,000 Demotic words 3.59
demotic strict, all dates per 1,000 German words 3.26
demotic strict, 700–300 BC per 1,000 Demotic words 3.24
demotic strict, 700–300 BC per 1,000 German words 3.05
pRylands 9 per 1,000 Demotic words 2.41
pRylands 9 per 1,000 German words 2.14

5.8 Greek ἐγένετο

Question. How often do the Gospels and Acts use the formula, by bucket, against the Book of Mormon?
Why. Luke-Acts is the nearest non-Hebrew model of a heavy ‘it came to pass’ style, and the hand labels test the English rule on a second language.
Label. exploratory.
Code. analyses/wayhi/greek_egeneto.py; result results/wayhi/greek-egeneto.json, fingerprint 93f067ccfd0e; Requiem record results/wayhi/greek-egeneto/headline.

book setting-first straight to event subordinate ordinary verb
Mt 6 0 0 7
Mk 2 2 0 15
Lk 38 1 0 30
Jn 0 0 0 17
Ac 10 5 3 36
Ro 0 0 0 2
1Co 0 0 0 1
2Co 0 0 0 1
1Th 0 0 0 1
2Ti 0 0 0 2
Heb 0 0 0 3
Re 0 0 0 20

KJV rendering, verses with one ἐγένετο: setting 52/56; event 8/8; sub 1/3; verb 1/124.

narrative frame setting + event per 1,000 KJV words per 100 Greek verbs straight to event BoM ÷ this, per word per verb
Gospels + Acts 51 + 7 1.28 0.71 12% (5%–23%) 6.3× (4.8–8.3) 7.9× (6.1–10.5)
Luke + Acts 44 + 5 2.16 1.24 10% (3%–22%) 3.7× (2.8–5.0) 4.5× (3.4–6.2)
Matthew + Mark + John 7 + 2 0.40 0.21 22% (3%–60%) 20.2× (10.6–44.4) 26.4× (13.9–57.9)
Book of Mormon 1830 frame 8.01 5.60 70%

5.9 Straight-to-the-event share in English texts

Question. Do English texts in biblical style reach the Book of Mormon’s event-first share and rate?
Why. English in biblical style is the competitor to an Egyptian source for the event-first share; it gets the same measurement.
Label. exploratory.
Code. analyses/wayhi/event_shares.py; result results/wayhi/event-shares.json, fingerprint 12915f9cc5a3; Requiem record results/wayhi/event-shares/headline.

text words past uses per 1,000 straight to event (revised rule) (registered rule)
KJV Old Testament 610,746 388 0.64 14/388 = 4% (2%–6%) 12% (9%–16%)
KJV New Testament 180,667 65 0.36 8/65 = 12% (5%–23%) 22% (12%–33%)
The Late War (1819) 47,500 77 1.62 22/77 = 29% (19%–40%) 35% (25%–47%)
Snowden, American Revolution (1793) 53,916 113 2.10 49/113 = 43% (34%–53%) 48% (38%–57%)
First Book of Napoleon (1809) 20,003 13 0.65 10/13 = 77% (46%–95%) 85% (55%–98%)
Book of Nullification (1830) 14,888 16 1.07 6/16 = 38% (15%–65%) 75% (48%–93%)
Moses 1 1,472 10 6.79 7/10 = 70% (35%–93%) 70% (35%–93%)
Moses 2–4 2,979 1 0.34 1/1 = 100% (2%–100%) 100% (2%–100%)
Moses 5–8 8,108 32 3.95 26/32 = 81% (64%–93%) 84% (67%–95%)
Moses (all) 12,559 43 3.42 34/43 = 79% (64%–90%) 81% (67%–92%)
Abraham 5,507 16 2.91 15/16 = 94% (70%–100%) 94% (70%–100%)
Book of Mormon, narrative frame (bom-buckets) 1,124 70% (67%–73%), 788/1,124 71% (68%–74%), 800/1,124

JST, revised rule:

panel added kept from the KJV added (boundary-corrected)
primary (OT narrative) 75% (61%–86%), 39/52 5% (3%–8%), 15/317 81% (67%–91%), 38/47
secondary (NT narrative) 20% (1%–72%), 1/5 16% (7%–27%), 9/58 20% (1%–72%), 1/5

Abraham vs Moses (revised rule), Fisher’s exact test: p = 0.26.

5.10 Egyptian

Question. Do Egyptian narratives open clauses with ‘then / it came to pass’ at a rate and mix that, alone or added to Hebrew, matches the Book of Mormon?
Why. Tests the ‘reformed Egyptian’ alternative on the texts that match in genre and date.
Label. exploratory.
Code. analyses/wayhi/egyptian.py; result results/wayhi/egyptian.json, fingerprint ce5522c23c94; Requiem record results/wayhi/egyptian/headline.

The old narrative particles (ꜥḥꜥ.n, wn.ꞽn) are the unambiguous clause-opening “then” of Middle and Late Egyptian narrative. Counting them is a floor: a text that chains events another way reads as zero (Wenamun uses the Late Egyptian continuative instead). Demotic marks “it came to pass” with ḫpr. Both lead into events, not settings.

text stage genre composed tokens particles per 1,000 est. translation words date ok genre ok matched
Sinuhe (pBerlin 3022, B) Middle Egyptian literary tale c. 1900 BC 3,194 7 1.28 yes yes yes
Shipwrecked Sailor Middle Egyptian literary tale c. 1900 BC 1,300 27 12.17 yes yes yes
Tales of pWestcar Middle Egyptian literary tales c. 1600 BC 3,391 126 21.77 yes yes yes
Eloquent Peasant (B1) Middle Egyptian literary tale with speeches c. 1850 BC 3,161 11 2.04 yes yes yes
Herdsman’s Tale Middle Egyptian literary tale c. 1800 BC 176 0 0.00 yes yes under 500 tokens
Carnarvon Tablet I (Kamose) Middle Egyptian royal campaign narrative c. 1550 BC 440 1 1.33 yes yes under 500 tokens
Kamose Stela I Middle Egyptian royal campaign narrative c. 1550 BC 171 0 0.00 yes yes under 500 tokens
Kamose Stela II Middle Egyptian royal campaign narrative c. 1550 BC 837 1 0.70 yes yes yes
Tale of Two Brothers Late Egyptian literary tale c. 1200 BC 4,337 70 11.02 yes yes yes
Contendings of Horus and Seth Late Egyptian literary tale c. 1150 BC 4,776 157 22.44 yes yes yes
The Doomed Prince Late Egyptian literary tale c. 1300 BC 1,630 34 14.24 yes yes yes
The Taking of Joppa Late Egyptian literary tale c. 1250 BC 569 1 1.20 yes yes yes
Truth and Falsehood Late Egyptian literary tale c. 1250 BC 1,215 30 16.85 yes yes yes
Astarte and the Sea Late Egyptian literary tale c. 1400 BC 761 9 8.07 yes yes yes
Report of Wenamun Late Egyptian first-person travel report c. 1050 BC 3,044 0 0.00 yes yes yes
Bentresh Stela Late Egyptian temple legend in royal style c. 400 BC (set in Ramesside times) 972 23 16.15 no yes no
Piye Victory Stela Late Period royal campaign narrative c. 727 BC 4,379 53 8.26 yes yes yes

Pooled by stage (texts of 500 tokens or more):

group texts tokens particles per 1,000 tokens per 1,000 est. translation words
Middle Egyptian narrative 5 11,883 172 14.47 8.48
Late Egyptian narrative 8 17,304 324 18.72 12.78
Late Period narrative 1 4,379 53 12.10 8.26
all named narrative texts 14 33,566 549 16.36 10.55

Translation words per Egyptian token, measured in each TLA subset: Earlier Egyptian 1.71; Late Egyptian 1.47; Demotic 1.16.

25th–26th-dynasty hieroglyphic royal inscriptions (the Saite check):

inscription tokens old particles ḫpr openings per 1,000 (×1.47 scale) note
Piye Victory Stela (c. 727 BC) 4,379 53 2 8.54
Taharqa flood stela, Cairo JE 48400 (c. 680 BC) 381 0 1 1.79 under 500 tokens
Taharqa stela Cairo JE 36861, lower text 377 0 0 0.00 under 500 tokens
Banishment stela of Aspelta(?) (c. 600 BC) 216 0 0 0.00 under 500 tokens
Stela of Aryamani 435 0 0 0.00 under 500 tokens

Demotic ḫpr openings, TLA narrative sentences (selected by rule; genre not matched):

selection sentences German words ḫpr openings per 1,000 setting-first straight to event date ok
demotic loose, all dates 1,567 13,365 38 2.84 (2.01–3.90) 17 21 no
demotic loose, 700–300 BC 792 6,213 17 2.74 (1.59–4.38) 10 7 yes
demotic strict, all dates 1,306 11,030 36 3.26 (2.29–4.52) 15 21 no
demotic strict, 700–300 BC 641 4,916 15 3.05 (1.71–5.03) 8 7 yes

pRylands 9 (c. 513 BC; a petition; genre not matched):

selection sentences German words openings per 1,000 setting-first straight to event
all sentences 1,012 9,421 15 1.59 (0.89–2.63) 1 14
unquoted (narrative frame) 627 5,602 12 2.14 (1.11–3.74) 1 11

Additive model, per 1,000 words (Hebrew setting-first only; its straight-to-event bucket is empty):

model setting-first straight to event total Egyptian side matched
Hebrew + demotic loose, all dates 4.19 1.57 (0.97–2.40) 5.76 (4.93–6.82) no
Hebrew + demotic loose, 700–300 BC 4.53 1.13 (0.45–2.32) 5.65 (4.51–7.30) no
Hebrew + demotic strict, all dates 4.28 1.90 (1.18–2.91) 6.18 (5.20–7.44) no
Hebrew + demotic strict, 700–300 BC 4.54 1.42 (0.57–2.93) 5.97 (4.63–7.95) no
Hebrew + pRylands 9 3.10 1.96 (0.98–3.51) 5.06 (4.02–6.66) no
Hebrew + Piye (old particles, all straight-to-event) 2.92 8.26 11.18 yes
Book of Mormon, 1830 frame 2.40 5.61 8.01

Matched narratives, by text: 0.00–22.44 per 1,000 estimated translation words.

5.11 KJV parallels left in the frame

Question. Which Book of Mormon passages run parallel to the KJV, and does any stay in the narrative frame?
Why. Checks the exclusion annotation against the texts, so no KJV text is counted as the book’s narration.
Label. exploratory.
Code. analyses/wayhi/exclusions.py; result results/wayhi/exclusions.json, fingerprint 457f8ba2ad54; Requiem record results/wayhi/exclusions/headline.

39 passages found (table in Methods). Passages that keep verses in the narrative frame:

passage KJV frame verses words
1 Nephi 22:20 Acts 3 1 Nephi 22:20 73
3 Nephi 10:4–10:6 Matthew 23 3 Nephi 10:4, 3 Nephi 10:5, 3 Nephi 10:6 146
Mormon 9:22–9:24 Mark 16 Mormon 9:22, Mormon 9:23, Mormon 9:24 110

5.12 Pre-registered: PREREG 2, what the narrative particle introduces

Question. Does the Book of Mormon’s event/time split of ‘it came to pass’ match the Egyptian narrative particles (within 0.15), and do English controls fail to?
Why. The registered test of the ‘Egyptian particle’ reading of the event-first share.
Label. registered, as fixed by PREREGISTRATION_2.md.
Code. analyses/wayhi/prereg_event_share.py; result results/wayhi/prereg-event-share.json, fingerprint 4d36f05adcbe; Requiem record results/wayhi/prereg-event-share/headline.

text (registered rules) event time event share
Egyptian particles, pooled (E) 438 0 1.00
Hebrew wayhi, 17 units (H) 0 239 0.00
Book of Mormon frame (B) 799 324 0.71
KJV Old Testament 48 340 0.12
KJV New Testament 14 51 0.22
The Late War (1819) 27 50 0.35
rule result
R0 boundary E-H
R1 E >= 0.50 True
R2 B-E
R3 controls KJV Old Testament: separate; The Late War (1819): separate

Registered verdict: MISS: the book does not match the Egyptian proportion

The revised-rule figures beside it: the Book of Mormon frame is 70% (67%–73%), 788/1,124 straight to the event, the KJV Old Testament 4% (2%–6%). They don’t change the registered verdict.

5.13 Pre-registered: PREREG 30, Joseph Smith’s own “then”

Question. When Joseph Smith added ‘it came to pass’ to the KJV in the JST, was it event-first at the Book of Mormon’s share?
Why. Tests the main competitor to a non-English source: Smith’s own English, in his later scripture.
Label. registered, as fixed by PREREGISTRATION_30.md.
Code. analyses/wayhi/prereg_jst_then.py; result results/wayhi/prereg-jst-then.json, fingerprint ddedb53b5e99; Requiem record results/wayhi/prereg-jst-then/headline.

book JST words KJV words token match KJV formulas removed by Smith
Genesis 50,632 38,515 0.93 63 2
Exodus 32,164 32,768 0.94 26 1
Numbers 33,185 32,942 0.96 10 1
Joshua 19,068 18,866 0.96 28 4
Judges 20,472 19,006 0.97 26 2
Ruth 1,216 2,583 0.46 3 2
1 Samuel 25,000 25,145 0.97 33 2
2 Samuel 20,491 20,717 0.96 26 2
1 Kings 24,495 24,586 0.96 40 1
2 Kings 23,369 23,589 0.97 36 0
1 Chronicles 20,047 20,414 0.94 9 1
2 Chronicles 25,693 26,122 0.95 19 0
Ezra 7,915 7,453 0.96 0 0
Nehemiah 11,803 10,497 0.97 13 0
Esther 3,981 5,714 0.68 4 2
Matthew 27,785 23,726 0.92 6 0
Mark 14,909 15,186 0.79 4 2
Luke 28,266 25,986 0.91 40 4
John 18,816 19,127 0.92 0 0
Acts 24,245 24,282 0.97 14 0
panel formulas n event-first share (registered rule)
primary (OT narrative) added 52 75% (61%–86%), 39/52
primary (OT narrative) retained 317 13% (10%–17%), 42/317
secondary (NT narrative) added 5 40% (5%–85%), 2/5
secondary (NT narrative) retained 58 26% (15%–39%), 15/58

Registered verdict: R1 HABIT TRANSFERS: H_then is a sufficient explanation of the event-first share. Without the five boundary spill-overs (Amendment 2): 81% (67%–91%), 38/47; verdict R1. Headnote-limit robustness: 2 lines → 53 added, S = 0.755 (R1); 3 lines → 53 added, S = 0.755 (R1); 6 lines → 52 added, S = 0.750 (R1); 8 lines → 52 added, S = 0.750 (R1).

Reading, from the pre-registration’s Amendment 3: R1 shows that the straight-to-the-event idiom was Joseph Smith’s by 1830, in new narrative with no source text behind it. Every Smith narrative was written after the Book of Mormon was dictated, so whether he had the idiom before it is untested. No pre-1828 Smith narrative survives. Under the revised rule, formulas Smith kept from the KJV are 5% (3%–8%) straight to the event.

6. Corrections to earlier figures

Earlier versions of these figures circulated in notes and drafts. Where a figure changed, both are given, and the reason. Corrections carry the same weight as the results they replace.

earlier figure now reason
Hebrew narrative-marker rate 3.04 per 1,000; Book of Mormon 2.5× (2.6× for 1830) setting-first only, 2.92; 2.7× (2.3–3.1) (2.7× (2.4–3.2) for 1830) the earlier Hebrew rule missed settings after l- prepositions and bare time nouns
Book of Mormon event-first share 71% 70% (67%–73%), 788/1,124 revised English rule (setting before the formula; more time phrases)
KJV Old Testament event-first share 12% 4% (2%–6%) mostly rule misses (“the same day”, a setting placed first)
KJV New Testament 22%, then 20% 12% (5%–23%) revised rule; then a bug fix: the revised pattern was sliced one character short and lost “the day”
The Late War 35%; Snowden 48%; Napoleon 85%; Nullification 75% 29% (19%–40%); 43% (34%–53%); 77% (46%–95%); 38% (15%–65%) revised rule. Nullification puts the setting before the formula
JST formulas kept from the KJV 13% event-first 5% (3%–8%) revised rule. JST additions unchanged
Hebrew + Demotic additive model “fits” the Book of Mormon (7.7–8.1 per 1,000) 5.7–6.2 against 8.01 the first model added the Hebrew ordinary-verb uses as if they were formula uses
Egyptian texts before 600 BC set aside as “too old” matched texts run 0–22 per 1,000; the rate test is inconclusive the cut-off wasn’t principled: formal Egyptian kept the older register, and Saite scribes revived it
wehayah 353; wayhi + wehayah 1,137 357; 1,141 the earlier count used a regular expression on the text; this one uses the morphology tag
Isaiah quotation span 2 Nephi 12:1–24:23 2 Nephi 12:1–24:32; frame 144,801 words, 7.76 per 1,000 the span stopped 9 verses short of the end of Isaiah 14, leaving 236 words of KJV Isaiah in the frame (no formulas)
Moses 80% event-first at 3.75 per 1,000; Abraham 15/16 at 3.15; p = 0.42 Moses 34/43 at 3.42 per 1,000; Abraham 15/16 at 2.91 per 1,000; p = 0.26 recounted from the pinned Pearl of Great Price text; the earlier source and word counter weren’t recorded. No difference either way

The registered verdicts of PREREG 2 and PREREG 30 are unchanged by any of these.

7. Limitations

  • The rules are word-after rules, not full readings. Their agreement with the Hebrew, the Greek and a hand reading is in section 4.1: 91%–100% of checked cases, lowest on future settings. The residual error is a few percent each way.
  • Speech removal is asymmetric. The Old Testament side removes every quotation marked in the WEB. The Book of Mormon side removes sermons, letters and records, but not short dialogue inside narration. The book’s frame therefore keeps some speech, which makes its frame rate, and the ratio, slightly conservative.
  • Versification. The Hebrew text is numbered as in Hebrew Bibles, the KJV and WEB as in English ones. Hebrew instances in verses numbered differently can’t be placed in or out of speech; they are counted as frame (genre-ladder). Single-form verses that don’t match a KJV verse are left out of kjv-rendering.
  • The 1830 text is uncorrected OCR. Uses removed after 1830 are restored from an alignment. 51 current-text uses weren’t found in the 1830 OCR by the aligner. An earlier hand check of these candidates found them present but garbled or misaligned (with 4 not found), but the per-case labels weren’t kept, so the claim that no use was added after 1830 rests on an unrecorded check. Uses lost between the manuscripts and the 1830 printing are outside this study; Royal Skousen’s critical text would settle them.
  • Three short KJV parallels stay in the frame (section 4.3), together a few hundred words with one future-tense formula. They are reported, not removed, because they are short quotations inside narration, and the frame’s rule removes framed speeches, not quotations.
  • Egyptian rates depend on scale factors. Egyptian token counts are put on an English-word scale with a ratio measured in other texts of the same stage. Demotic rates are per German word. Neither scale has an interval. The per-verb comparison avoids this, but the TLA’s VERB tag and the Hebrew verb tag haven’t been validated against each other.
  • The Demotic selections are by rule. TLA sentences have no text names, so narrative sentences are chosen from their German translations. The sentences are mostly from documents, later copies and a petition. That is why they fail on genre.
  • The Greek labels are one reader’s. 70 ἐγένετο were read in context and labelled by hand; the labels and their notes are in annotations/egeneto_labels.tsv.
  • The Greek is the SBLGNT, not the KJV’s Textus Receptus. Where the two Greek texts differ, an ἐγένετο may be missing from one side; those differences were not counted.
  • The WEB text drifts. The pinned WEB build isn’t archived; the current zip differs in one used verse (John 3:2), restored by a recorded patch. See the sources section.

8. How to reproduce

git clone https://github.com/nicklecoder/bom-stylometry.git && cd bom-stylometry
uv sync --extra verbs            # Python 3.12; spaCy model for per-verb and greek-egeneto
uv run bomstylo fetch            # download, check by sha256, prepare
uv run bomstylo verify           # prepared files match their pins
uv run bomstylo run              # every analysis, in dependency order (about 6 minutes)
uv run bomstylo status           # all should read "current"
uv run bomstylo report wayhi     # rebuilds this file

Each result file records the hashes of everything it was computed from. If a rerun changes a headline number outside its previous interval, uv run bomstylo record says so in the Requiem record. The analyses are listed, with their inputs, in ANALYSES.md.