Skip to content

Test Protocol · 27 September 2026

Test protocol

The next phase is empirical operationalisation. Until the items below exist, the programme is testable in principle, not experimentally demonstrated.

The S3 files in the current freeze record the carrier path and falsifiers. They do not yet constitute a preregistered pilot. The on-site reading remains at Carrier protocol. The canonical files are the Full Method and the Illustrated Summary, and the summary does not replace the method.

A future pilot must specify

  • the observable dependent variables constituting Y;
  • the construction of baseline Y0;
  • the candidate active history H_a;
  • archive-present and archive-stripped conditions;
  • stranger-instance or equivalent controls;
  • blinded raters;
  • inter-rater agreement;
  • counterbalanced conditions;
  • sample size;
  • preregistered thresholds;
  • the method used to estimate or reject k_r.

The ELIZA strip test belongs inside that design: if the claimed relation survives the removal of dated scaffolding, the control has won that trial.

Do not invent additional protocol doctrine here. Write the pilot when it can fail.

The Probe, filed 30 September 2026, is the instrument for that pilot. It does not replace this list, and it does not replace Document 02.