The computational studies published on this site were run with the explicit goal of being checkable. Everything needed to verify or challenge them is provided here: the analysis code, the corpora as analysed, per-document provenance records for every text (title, date, source URL, authorship confidence, and scribal status), all numeric outputs including the full phrase lists, and every figure. Nothing was selected by hand: every number and figure in the documents regenerates by running the scripts.
Complete replication package
All analysis scripts, source and cleaned corpora with provenance manifests, results files, phrase lists, and figures — everything needed to re-run the studies from scratch. Python 3.13 with numpy and matplotlib; run order and source credits in the included README.
Download replication-package.zip (10.3 MB)The documents
Each page on this site is also available as a Word document:
- The Case Against the Truth-Claims of the LDS Church (.docx)
- A Stylometric Study of the Book of Mormon (.docx)
- Stylometric Analysis of the Book of Mormon — Full Manuscript (.docx)
- Four Computational Studies on the Origin of the Book of Mormon (.docx)
- The Book of Mormon and The Late War (.docx)
- The Book of Mormon and View of the Hebrews (.docx)
Numeric results (browsable)
The raw output files behind every reported number — validation grids, per-narrator recalls, the permutation test, attribution distances and votes, the open-set percentages, the full italics inventory, shared-phrase lists, and the New-Testament phrase census:
- attribution.txt
- distinctive_words.txt
- internal_voices.txt
- permutation_cosine.txt
- study1_exclusive_LateWar.txt
- study1_exclusive_Napoleon.txt
- study1_shared_phrases.txt
- study1b_pairwise.txt
- study1c_sizematched.txt
- study2_italics_detail.txt
- study2_kjv_italics.txt
- study3_nt_census.txt
- study3_tierA_5grams.txt
- study3_tierB_5grams.txt
- study4_register.txt
- validation.txt
How to reproduce
1. Download the replication package and unzip it.
2. Install Python 3.13 with
pip install numpy matplotlib.
3. Run prepare_bom.py,
prepare_spalding_kjv.py, and
prepare_controls.py (corpus preparation), then any
analysis script. Outputs land in results/ and match the
published documents.
Corpus sources
All corpus texts are 19th-century (or earlier) public-domain works. Transcription sources, credited per document in the manifests: the Joseph Smith Papers (verbatim transcripts; holograph/dictated documents only), Project Gutenberg, Internet Archive, Wikisource, centerplace.org, sidneyrigdon.com, eBible.org, and bcbooks/ scriptures-json.
All figures
fig1_kappa.png
fig2_pca.png
fig3_candidates.png
fig4_openset.png
fig5_shared_phrases.png
fig6_register.png