Known limitations
- AI review: The release was reviewed by AI systems operating under a strict consolidated protocol. Agreement between AI systems is corroborative convergence, not human reliability evidence.
- Textual apparatus access: Full licensed NA28, UBS5, or UBS6 apparatus adjudication was not completed for every textual issue. Main-text observations are not equivalent to manuscript-witness verification.
- Blank textual placeholders: Forty-three textual-control rows contain no supplied reference or issue summary and remain transparent not-applicable placeholders.
- Provisional mappings: Forty-five occurrence-to-central mapping controls were confirmed provisionally because the reviewers did not have the exact Pilot 27 source wording mounted in their environments.
- Segmentation dependence: Counts depend on declared proposition-segmentation rules and controlled vocabularies. They should not be treated as inspired numerical rankings.
- Multi-label categories: Soteriological relations and inclusive topics may produce totals exceeding 179.
- Declared theological framework: The dataset is conservative evangelical, free-choice/non-Calvinist, conditional-security, moderately dispensational, and cautiously continuationist. It does not claim theological neutrality.
- Source coverage: Original-language, ancient-source, and modern-scholarly coverage is substantial but not exhaustive.
- Licence: The release does not add a separate open-data licence. Website use and sharing terms remain controlling.
How to use the dataset responsibly
- Read Matthew 5–7 before using the analytical classifications.
- Check propositions against their literary units and source occurrences.
- Distinguish analytical counts from theological priority.
- Do not represent provisional textual or mapping controls as settled manuscript facts.
- Do not represent the release as human peer reviewed.
- Report possible errors with the exact page, identifier, issue type, proposed correction, and supporting evidence.
What the release does establish
Within its declared framework and segmentation rules, the release provides a stable, internally validated public structure: 179 reviewed central propositions, 213 preserved historical records, 220 explicit links, thirty literary units, and documented command-family, topic, normativity, evidence, soteriological, textual-control, and mapping profiles.
Licence and sharing
The website release page and documentation follow the site's established non-commercial attribution and no-derivatives use-and-sharing terms. The tabular dataset does not grant an additional open-data or open-source licence. Do not assume permission for commercial republication, data resale, derivative redistribution, model training, or altered republication.