[amaybaum]

Documentary correction to two mutable records, from the removal-candidate audit under #1088 (SUBSTATE-2 group). No frozen record, theorem, verdict or research file changes.

Owner authorization (amaybaum, Claude Code session chat, 2026-11-04): "I approve preparing a separate documentary correction to the mutable recursive-substates/README.md and the appropriate MR-1 qualifier. G★ item 2: Explicitly identify the clause as refuted within the scope established by PC-F1 and the on-main Lean theorem. Cite the exact supporting evidence. Historical references to the frozen clause must not be treated as valid premises. F2-1 run 1: State its VOID outcome explicitly, alongside run 2's scoped gap and the author-reported review status. Preserve the original frozen result file." Merge needs separate approval.

Why. The audit rated two promotion risks high.

• G★ item 2 (refuted statement). F2-1's frozen A★ (f2-1-coherence-sourcing/amendments/A-star.md:144–148) states that "no diagonal unitary commuting with the computed readout projectors conjugates G_Θ into either class" and that "This is necessary for R." For the ones-fixing class the necessity fails. MR-1's frozen preregistration cites G★ with no qualifier (:99, :638–639). The only correction was a research: SUBSTATE-2 recursive inverse-limit substructure and depth-feedback controls #1011 comment, which is not on main.
• F2-1 run 1. The preserved stage2/RESULT.md records no token ("None yet", :89). Its INCONCLUSIVE (void) outcome was stated only in sibling files.

Head 77b2f30b, three commits on main d8eda7d8.

• 8433f7e9 is the correction.
• af053f11 applies the exact-head review.
• 77b2f30b applies the delta review's nits.

The PR changes 2 files: 1 modified and 1 added.

File	Change
verification/programmes/substratum/recursive-substates/README.md	F2-1 status block. Run 1 (05ecc873) is INCONCLUSIVE (void) under clause 0(v) (f2-1-coherence-sourcing/preregistration.md:220). Its Build failed (A-star-1.md:12–14; run-2/RESULT.md:5–7; errors in PROOF-REPAIRS.md:7–12). The preserved stage2/RESULT.md predates the build (:89, :93). Run 1 records no review. Run 2 is INCONCLUSIVE (gap), scoped to the thirteen rules (run-2/RESULT.md:121, :129–130), and is not a no-go. Its review is author-reported (:125). G★ item 2 is refuted in part, with the evidence below. Run 2's token is unchanged: its criterion is preregistration.md:221, and the candidates fail items 1 and 3 (A-star.md:180–183), not item 2
verification/programmes/oi-qm/reconstruction/measurement-record-formation/README.md (new)	The same qualifier for MR-1's frozen citations of G★. :99 names G★. :638–639 is a reporting rule for a map meeting items 1–5, not a finding, and no MR-1 token uses G★ as its criterion. The freeze designation is quoted with author and time (amaybaum, 2026-10-31T21:14:51Z). The amendments and stage-1 log are noted as append-only (MR-star.md:3–5, KS-substitution.md:3–4). The item-3 citations (:105–106, MR-star.md:124, :668) are unaffected. MR-1's directory held only frozen or append-only records

Scope of the refutation.

• Kernel-checked, all three built on main:
  • the membership D†·rot·D ∈ onesClass for D = diag(1, i) (OIBridge.PC1Controls.gauged_rot_mem_onesClass, PC1Controls.lean:132);
  • R for every architecture containing rot (falsifier_available_of_mem, :324), with R as F2-1 defines it (preregistration.md:83);
  • for rot, the surviving nonnegative half (scaled_rot_not_gauge_nonneg, :164).
• From PC-F1 (physical-coupling-pc1/preregistration.md:87–100; paper, checked by its author):
  • the general necessity of the clause without a gauge and of Gauge_N (:99);
  • Gauge_O as an L2 requirement that every generated class with P meets (:96–98).
• What is refuted: only the necessity of the gauged ones-fixing condition for R. It is not a valid premise.
• What stands: the nonnegative half of the gauge clause and the clause without a gauge.

§A.42 item 8, off-main references.

• Provenance:
  • CI run 41220197866: run 1's VOID is supported on main by the frozen rule and the records cited above, and 05ecc873 is on main.
  • research: SUBSTATE-2 recursive inverse-limit substructure and depth-feedback controls #1011 comment 6556076022: run 2's status is stated at run-2/RESULT.md:125.
  • research: SUBSTATE-2 recursive inverse-limit substructure and depth-feedback controls #1011 comment 6546748788: the finding's first report.
• Quoted with author and time: Research (draft): MR-1 — measurement as physical record formation, preregistration only #1023 comment 6549225579. Its content is on main at MR-star.md:3–5.

Checks (local, at 77b2f30b).

• Passing: artifact placement, voice, claims, duplicate, staleness, citation roles and legacy records (303 intact).
• Every cited line and quotation was checked against main d8eda7d8.
• No push-triggered workflow matches these paths.
• Closing keywords: none.

Review.

• Exact-head review at 8433f7e9: CHANGES REQUIRED. It raised four blocking findings, B1–B4:

  • B1: run 1's review status was misattributed;
  • B2: the refutation's scope went beyond PC-F1;
  • B3: the unchanged-verdict rationale rested on an off-main comment;
  • B4: MR-1 :638–639 read as a finding.

  It also required R5 (cite the kernel theorem falsifier_available_of_mem) and raised six nits. af053f11 applies all of them.

• Delta review at af053f11: CONFIRMED WITH NITS, nothing blocking. Its three nits (line ranges and path prefixes) are applied in 77b2f30b.

Gates.

• Independent exact-head review: done.
• Green CI at the exact head 77b2f30b.
• Owner approval naming the exact head SHA.
• Merge commit only.

🤖 Generated with Claude Code

https://claude.ai/code/session_013T6pfKuv5mi3MXtKjH2tZb

[amaybaum]

Merged. Merge commit b8f5991ae563916b3308d72e8c1ed2fb512eb275, pinned to head 77b2f30b2e1e45bacf3734e01c5f877f1fa94f43.

Owner approval (amaybaum, Claude Code session chat, 2026-11-04): "I approve the normal merge of PR #1104 at exact SHA: 77b2f30b2e1e45bacf3734e01c5f877f1fa94f43".

Immediately before the merge:

• The remote head was 77b2f30b and the base was main d8eda7d8.
• GitHub reported the PR mergeable and clean.
• CI run 41883525321 at the head concluded success: 25 checks succeeded and 1 was skipped by design.

After the merge:

• The parents are d8eda7d8 and 77b2f30b, and the tree equals the head's tree.
• git diff d8eda7d8 b8f5991a touches exactly the two READMEs: measurement-record-formation/README.md added and recursive-substates/README.md modified.
• legacy_records_check.py reports 303 records intact.
• main CI on b8f5991a is running.

This merge records, in mutable files on main:

• that the necessity of G★ item 2's gauged ones-fixing condition for R is refuted, with PC-F1 and the kernel theorems gauged_rot_mem_onesClass, falsifier_available_of_mem and scaled_rot_not_gauge_nonneg;
• F2-1's run outcomes: run 1 void, run 2 a scoped gap.

No frozen record, theorem or verdict changed.

---

Generated by Claude Code