[amaybaum]

This PR is correction unit S3 of the Main proof-independence work on #1029. It implements Option B, which the owner selected (2026-11-02) after the independent proposal review on #1029 (comment 6583722893, ACCEPT-WITH-CHANGES, recommending B). Merge is not authorized.

Change

The PR deletes one sentence: the last sentence of Main:652 (ii).

The common local quantum composite/instrument structure is the content of the inert-spectator and iterated-composition conditions of the operational-completion characterization (Remark on the operational-extension boundary; [GR §3.3]).

The attribution is refuted, not merely unsupported. The kernel theories diagTheory and everywhereAvailable satisfy both named conditions in the same FiniteOperationalTheory setting and under the same condition predicates, and neither is exactly quantum:

• diagTheory: control_independent / cell_C, and diag_not_exactAll;
• everywhereAvailable: validity_independent, with validity_of_exactComposite.

The claim is therefore withdrawn without a recovery prerequisite. A version of the claim with added hypotheses would be a different proposition and is not substituted here.

The stronger valid result is preserved at full strength. The five-condition characterization stays in Main as headline (4) at Main:82, anchored in §3.4. Each direction is kernel-checked:

• physical_of_exactAll (exact ⇒ five conditions);
• exactAll_of_physical_general (five conditions ⇒ exact).

The kernel citations sit at GR:218, which Main's sites point to through [GR §3.3]. Bringing them into Main is the separate C1/R1 unit.

It holds within the assumed complex-matrix kinematics, the carrier–ancilla composites and the setting's rules (a)–(f). GR §3.3 states those rules; Main itself does not. Which of the conditions OI's native substratum physically supplies remains open.

Unchanged

• No LEDGER, census-registry, grade or SC-42 change.
• No other site: the review found no other place asserting the two-condition attribution.
• No P1, C1, Correction planning: scientific overclaims and status corrections (from the 2026-11-01 disposition) #1028, Main:356 or S2 site is touched.

Checks

• Rebuilt with sh ./build.sh Main: 99 pages (unchanged), no dropped glyphs, no U+FFFD, and the stamp matches. The PDF word change (−25 net) tracks the 26 deleted words.
• Local checks: lean_manuscript_census, coverage_check, voice_check, claims_check, duplicate_check, mirror_check, citation_check, dependency_label_check --strict and staleness_check all pass. The R7-CTN and R7-PTR guards are true.
• SC-42 and proof census: SC-42 --verbose output is byte-identical to the checked-in result, and proof_census --json is identical to base.

One commit on 06df90c9, 3 files. #1036 and #1037 stay paused, and the five standing holds are unchanged.

---

Generated by Claude Code

[amaybaum]

Exact-head review record: f5ee87c6

Verdict: ACCEPT. No blocking findings. The review was independent. The head is f5ee87c6afce74b3c79b18bc4768844ad7378556, single parent 06df90c9. main has since moved to 1cfc20f6 (the #1047 merge, AGENTS.md only). git merge-tree --write-tree origin/main f5ee87c6 is clean.

Fidelity.

• Applying edits_altB.py to 06df90c9 reproduces the head's Main.md byte for byte.
• Only Main.md, Main.tex and Main.pdf change.
• The deletion is exactly the 26-word last sentence of Main:652 (ii). The other 917 lines are identical.

Science.

• No recovery step needed. The withdrawn sentence is a refuted claim, so the theorem-recovery rule requires no recovery step.
• Five-condition characterization intact. Headline (4) at Main:82 and its anchor at Main:564 still state it at full strength.
• Item (ii) reads correctly after the deletion.
• No other site makes the two-condition attribution.
• Clarification. Main cites neither physical_of_exactAll nor exactAll_of_physical_general. Both directions are cited at GR:218, which Main's sites reach through [GR §3.3]. That is the existing gap the C1/R1 unit covers; this PR neither creates it nor widens it.

Generated files.

• The stamp matches, and pandoc regeneration of Main.tex is byte-identical.
• 99 pages; median characters per page 2707 (base 2712); no U+FFFD.
• The word diff shows only the deletion plus reflow.

Checks at the head.

• The full battery passes. The R7-CTN and R7-PTR guards are true.
• SC-42 output is byte-equal to the checked-in result in both normal and -O runs, and the self-test controls are byte-equal too.
• proof_census --json is byte-identical to base.
• No probe pin falls on the deleted text.

Commit message and PR body.

• Both are accurate, including the refutation witnesses and the statement that a version with added hypotheses is a different proposition.
• The PR body has been clarified on three points:
  • validity_of_exactComposite is now cited alongside validity_independent;
  • the kernel citations sit at GR:218 rather than in Main;
  • the rules (a)–(f) are stated in GR §3.3, not in Main.
• The commit message's "composites and rules" describes the theorem. Rewriting it is not recommended.

CI at f5ee87c6. Run 41494227213 is green: 25 jobs passed, and 1 (A42 exclusion (${{ matrix.part }})) was skipped by its condition.

No merge is authorized. Merging needs separate exact-SHA approval from the owner.

---

Generated by Claude Code

[amaybaum]

Exact-head owner approval and verified S3 normal merge — 2026-11-02

APPROVED AND MERGED S3 Option B at #1048 exact reviewed head f5ee87c6afce74b3c79b18bc4768844ad7378556.

• Independent exact-head review: ACCEPT, no blockers. Source change is solely deletion of the refuted 26-word final sentence of papers/Main.md:652(ii); 917 other source lines unchanged.
• Exact-head CI 41494227213 completed successfully: 25 jobs passed and one conditionally skipped.
• Prior to merge confirmed current main 1cfc20f6bbf17bc4b45457b9fed59a7464096b3e, exact PR head unchanged and cleanly mergeable; marked draft ready then performed normal merge with expected-head SHA guard.
• GitHub returned merged=true, merge SHA 7b389414a71e08486738ff7069930c9e09acff98. Re-fetched main at exactly 7b389414a71e08486738ff7069930c9e09acff98, with parents 1cfc20f6bbf17bc4b45457b9fed59a7464096b3e and f5ee87c6afce74b3c79b18bc4768844ad7378556. The only changed files are papers/Main.md, and generated Main.tex/Main.pdf.
• Post-merge main CI 41497171388 was queued, not yet successful when checked.

Science: diagTheory and everywhereAvailable refute the two-condition sufficiency attribution. No theorem-recovery step is required for a genuinely refuted claim. The five-condition equivalence remains unchanged in Main's headline and §3.4 anchor under its named assumed kinematics, and its physical native sourcing remains open. This merge does not certify Main's independence from GR: the existing C1/R1 requirement to provide independently supported five-condition theorem statements and kernel proofs remains unresolved. S3 has no new SC-42/LEDGER/census changes.

H1 retrospective note on #1029 was already complete. H2 scoped source/refutation writer can continue; its clause-by-clause proof comparison precedes any manuscript commit. P1 follows H2; C1/C2/L1 retain separate ownership. #1036/#1037 and five holds unchanged.

[amaybaum]

Post-merge verification: 7b389414

• Merge commit: 7b389414a71e08486738ff7069930c9e09acff98, with parents 1cfc20f6 (previous main, the AGENTS.md: theorem recovery before concession #1047 merge) and f5ee87c6 (the reviewed head).
• Tree: e49454cd1d15f195fef248dbf6bb79a0eb2c1baf, which equals git merge-tree --write-tree 1cfc20f6 f5ee87c6. Only Main.md, Main.tex and Main.pdf differ from the previous main.
• main CI on 7b389414: verify, push event, run 41497171388, is green. 25 jobs passed and 1 (A42 exclusion (${{ matrix.part }})) was skipped by its condition.

S3 is complete. The refuted two-condition attribution is withdrawn from Main, and the five-condition characterization is unchanged. Bringing its kernel citations into Main remains with C1/R1.

---

Generated by Claude Code