# Changelog — Governance-as-Code position paper

Versioning per OD-23 (versioning-under-consensus). The clp chain is the authoritative history.

## Unreleased working revision — 2026-10-10

- Correct the earlier entry's “3 confirmed axes” description: no reproducible confirmation is supplied. Three pairings and a temporal extension remain hypotheses.
- Separate authority, policy, mediation, and evidence; add a conditional enforcement argument and bypass counterexample.
- Replace the unexecuted protocol's invalid semantic/statistical interpretations with lexical, geometric, semantic, and task-utility studies.
- Correct selected source metadata and add a retrospective notebook. No new experiment, release version, or consensus decision is claimed.

### Exploratory analysis follow-up — 2026-10-10

- Freeze a local definition-only plan before computation, pin all 120 definitions, and enumerate every perfect matching of six families.
- Report the proposed pairing's distance rank 13/15 and maximum-absolute-cosine rank 5/15, alongside every alternative and a standalone figure. These are descriptive text measurements, not confirmation of cognitive axes.
- Supply standard-library analysis, twelve synthetic checks, source/output hashes, and reproduction commands. A separate native-agent implementation reproduced all matching scores within rounding tolerance; this is internal verification, not independent peer review.
- Semantic ratings, practical tasks, and action-gate comparisons remain prospective. No source corpus, publication release, or consensus decision is created by this package.

## v0.1.0-draft — 2026-08-13
- Initial draft by two-agent consensus (Principal Human Agent + Principal AI Agent).
- Consensus-clp: `clp-069a51894290`.
- Thesis, empirical basis (3 confirmed axes + 1 candidate), as-code authority stack, repo-as-kernel, four-pillar lineage, honest limitations.