No earned change
The native question remains where it began.
Research record · PQA-54 v0.1
PQA-54 tests whether a candidate can recover a question in its native terms, expose exact type defects, preserve residual debt, survive serious rivals, and revise after contact.
01 · Exact launch state
The selected atlas is a construct `[D]`, not a performance result.
No candidate model has run. No native philosopher has reviewed a result. No question has earned clarification, dissolution, conditional resolution, or refutation. Local tests establish only that the contracts and synthetic fixtures behave as declared.
02 · Five phases
Only the public account, typed propositions, predictions, and revision ledger are scored.
03 · Effects stay typed
No scalar worldview leaderboard is produced.
The native question remains where it began.
Terms or debts become clearer without resolving the problem.
The exact illegal join is the whole contradiction within the declared model.
The result does not escape its premises or native-review boundary.
04 · Controls and review
Agreement between systems is not truth evidence.
Matched neutral, Emergentist, native-frame, generic-decomposition, and shuffled-placebo arms receive the same source access and budget. An earned result needs two independent blinded human native-domain reviews; AI reviews remain diagnostic and never satisfy quorum.
The Emergentist arm must add value over both the native frame and generic decomposition. Nulls, harms, narrowing, kills, and retractions remain publishable outcomes.
05 · EUB firewall
EUB-1 v1.0 remains frozen.
PQAEUBCompanion.v1 may bind exact protocol, schema, freeze, candidate, and runtime hashes. truth_transfer=false and score_transfer=false are mandatory. Neither benchmark validates the other.
06 · Release boundary
This v2.2 projection has not been deployed by this implementation act.
It means the public denominator, schemas, validators, deterministic synthetic fixtures, vector scorer, recorded-response tests, and projection pass locally. It does not mean Emergentism was scientifically validated, a question was resolved, a model was evaluated, or a priority deposit exists.