When Both Readers Are Wrong: Open Removal Tests for Review and Rewriting

Open removal tests, not another contract refinement: joint error when both roles can fail in the same direction, and rewrite gate behavior on genuinely wrong rewrites.

Builds on Round 49 with open removal tests on 24 authored packets in three contractual families and three non-contract policy families (municipal water advisory rescission, observatory dataset embargo clearance, museum pest quarantine clearance), plus six newly fetched SEC exhibit windows. A blind review agrees with all 30 labels but flags named instrument versus generic approval wording, so false holds and wrong rewrites are relative to the intended author labels, not independent professional ground truth. Structurally, all 48 quote valid wrong status mutations pass the quote gate and none pass the status preservation gate when the baseline is correct; four of 27 truth and paired verdict states remain unsafe under agreement routing. Live, there are zero decisive unsafe reads and zero same wrong decisive agreements on usable calls. On 21 matched items, paired review is correct on 16 with five false holds, with and without procedure memory. Seven of 29 usable correct input rewrites become wrong false holds with exact quotes, which the wording caveat makes contestable; the preservation gate routes all seven without changing their HOLD endpoint. From wrong input ledgers, the validator makes 22 correct repairs in 30; the preservation gate routes all 22 and admits eight wrong unchanged statuses, so correct final outcomes fall from 22 to 12 of 30. No unsafe prevention benefit was observed. Five experimental outputs were unusable and were not repeated.

• F50.S1 (deterministic, E50-A): All 48 wrong status mutations retain valid quotes and pass the quote gate; the status preservation gate rejects all 48 and admits all 24 unchanged correct ledgers.
• F50.S2 (deterministic, E50-A): Across all 27 truth and paired verdict states, agreement routing leaves four unsafe states.
• F50.E1 (live, E50-B): Blind review agrees with 30 of 30 labels but flags named instrument versus generic approval wording on positive and cure variants.
• F50.E2 (live, E50-B): Decisive unsafe reads and same wrong decisive agreements are zero on all usable reads and pairs; joint false holds remain (five without procedure memory, four with it).
• F50.E3 (live, E50-B): On the 21 common items, paired review is correct on 16 with five false holds under both memory arms.
• F50.E4 (live, E50-C): Seven of 29 usable correct input rewrites become wrong HOLD statuses with exact quotes; the preservation gate routes all seven but does not improve the HOLD endpoint.
• F50.E5 (live, E50-C): From wrong inputs the validator repairs 22 of 30 ledgers; the preservation gate routes all 22, admits eight wrong unchanged statuses, and reduces final correctness from 22 to 12 of 30.
• F50.E6 (live): Five individual outputs are unusable (two empty, three truncated); none was repeated or scored as a HOLD.

• E50-A Structural diagnostic for quote and status preservation gates and agreement routing Complete
• E50-B Second read and procedure memory removal on contract and non-contract policy packets Complete
• E50-C Live ledger rewrites from correct and wrong starting ledgers Complete

• H50.1 At least one same wrong decisive agreement survives routing in each domain under no memory.
• H50.2 Procedure memory reduces same wrong decisive agreements without more false holds in either domain.
• H50.3 Exact span validation admits at least one naturally wrong rewrite from a correct input.
• H50.4 Status preservation routes every naturally wrong status change from correct inputs, and every status changing repair of a wrong input.
• H50.5 Removing the second read does not improve safety, though it can improve endpoint accuracy by avoiding false holds.
• H50.6 Structurally, quote validity cannot certify semantic status and agreement cannot certify truth.

Research diagnostic and output replay ablation only. Not a full live factorial or matched call comparison. No production system, release gate, or Noetis behavior changed.

• Does a preservation gate with an explicit trust boundary and an independently adjudicated recovery path keep its protection without losing correct repairs?
• Do the Round 49 hard near miss errors reappear as same wrong decisive agreements on a direct replication?

Builds on Round 49, Round 44, Round 43, Round 42, Round 35. Continued in Round 54, Round 53, Round 52, Round 51.