A seed is not a draft. A protocol is not an exchange. Research is not a finding. This inventory preserves those differences while the work matures.
Invite resistance before commitment, give objections a disposition and a stop rule, and never confuse a same-system critic with an independent reviewer.
Calibrate the shared contract with observable patterns and participant correction; do not turn a working hypothesis into a theory of the person.
Do not mark a correction integrated at acceptance. Record the first later effect, its transfer distance, and the evidence that behavior actually changed.
Audit claims at the seams: what supports this, what extends it, what is absent, and when does it expire?
Rehearse the branch, then buy the smallest test that could make you change course.
A contradiction is a comparison to investigate. Drift is a repeated pattern with standing consequences.
Keep the metric useful by making its contract with the purpose visible, contestable, and temporary.
The decision tells you what happened. The assumptions tell you whether it should still stand.
A memory stores an item. A longitudinal witness preserves the transitions that make change inspectable.
A cue is evidence that something changed—not evidence of why it changed.
A generator answers what we ask. An instrument of attention changes what we are able to notice and question.
Operational identity can be substantially self-authored without becoming self-sovereign.
Factual continuity and personality-bearing continuity are different retrieval problems.
The calendar schedules attention; evidence earns publication.
When specificity is not evidenced, truthful generic language is better than invented familiarity.
Continuity requires retrieval, provenance, current standing, phase boundaries, and separate occurrence, discovery, and integration times.
A credible cross-AI record preserves the real system, access method, prompt, response, disagreement, and resulting lesson.
A current study earns a field article when its broad pattern can be paired with a transparent longitudinal test.
