[amaybaum]

Draft foundational release candidate — attribution correction complete; new-head CI and owner approval required.

Current head: 7ac361440e879f449aa108e127347544f932c83c. Tree: b763c37272bf58d8e15f2159d008b6bcdd0f5b48. Sole parent: previously reviewed 20f818f072036ea99fc4588fd2e3116dec898393. Targets main; draft and unmerged.

Attribution-only follow-up

This seven-file increment changes one manuscript, its canonical TeX/PDF, and four audit/manifest records. It adds explicit Specht–Wiegmann attribution beside Proposition 24.2 and the Schützenberger/Carlyle–Paz/Fliess lineage beside Proposition 24.3, retaining the latter's existing Berstel–Reutenauer and Thon–Jaeger citations. Six primary/historical bibliography entries are added.

Both proposition statements and complete proof blocks are byte-identical to 20f818f. All 134 canonical fingerprints/grades, Lean source, coverage ledger, registry and workflow are unchanged. The spanning premise remains in the kernel-certified statement; the general no-spanning classical criterion is attributed, not presented as a new OI or Lean result.

Substratum rebuilt through the canonical script: 54 pages, zero missing glyphs, matching source stamp, no out-of-page word boxes, and visual inspection of pages 28, 29, 53 and 54. Coverage, Lean/manuscript census and whitespace checks pass. The current 88-input manifest and attribution-check.json retain the source/proof/publication identities.

Current verification and review scope

Fresh normal verify 40767146156 is pending on this head. Current-head compiler, release-gate and complete job conclusions must pass before approval is requested.

The preceding source at 20f818f passed 40756692508: 25 successful jobs and one skip, including Lean kernel, Mathlib Build/Release gate, core B and KTOKEN. That run does not cover the attribution increment.

The retained independent scientific and provenance/publication acceptances are scoped and incremental, not a claim that a fresh reviewer exhaustively re-reviewed the entire expanded PR. They cover their exact recorded snapshots and protected scientific inputs. This citation repair preserves the accepted mathematics and does not broaden those verdicts or imply new physical correctness, a complete gauge classification, or a new compiler certificate. The primary references were checked directly for the added attribution. Earlier negative reviews, acceptance reports and receipts remain intact.

Overlapping PR disposition after an approved foundational landing

No PR is closed or merged by this list; preserve branches and evidence until their surviving dependencies are accounted for.

PR	Relationship to #982	Disposition
#965	Restricted first-order chirality/onsite converse is incorporated and formalized in the recovered SM Theorem 3 chain.	Superseded as a separate landing for that result; retain its original proof/control provenance.
#966	Unapplied seven-file patch-generator proposal overlaps source already corrected and rebuilt here.	Superseded as the proposed source-patch route; do not apply it onto reconciled main. Preserve its record.
#980	SC-44 partial correction is incorporated and superseded by the integrated corrected source.	Keep as provenance; no separate automatic merge.
#981	SC-45 counterexample, controls and required fixed-interface correction are incorporated/reconciled.	Keep the separate research/provenance record; no duplicate merge.
#992	Narrow regress/independent-ancilla correction is incorporated as source.	Preserve its separate checkpoint and review provenance; no duplicate merge.
#969	Alternative second-order/normalized-branch Theorem 3 route is materially different from the recovered first-order statement.	Not fully superseded. Reconcile its independent result and downstream scope on their merits before closure; no route choice or blanket closure is implied by #982.

#990 → #993, #991, #989 and R23 #984 remain separate workstreams, not superseded by this attribution edit. #918 remains held for its dependency and SC-39 reconciliation sequence.

Preserved foundational scope

#992's correction, SC-44/45 provenance, the exact cubic decomposition, conditional stabilizer, correlated quotient, faithful charged-family image, complete fixed-family anomaly classification, predictive reconstruction and record-fibre results remain. Written results and kernel certificates retain their respective statuses; Pair C's written infinite-ancilla argument remains separate from its finite-carrier Lean certificate.

Evidence: verification/audits/manuscript/foundational-release-20261028/. Current publication counts are SM 147, Substratum 54, Explainer 68, Methodology 47, Juno 22 and book 548. Only Substratum changed in this increment.

Keep this head unchanged for CI. If it passes, present this exact SHA and the stated review scope for explicit owner approval. No merge, special workflow dispatch, native F/E designation, PR closure or branch deletion is authorized or performed.

Historical checkpoint and scoped evidence at 20ca05f

Draft research and manuscript reconciliation — do not merge.

Current checkpoint: 20ca05fe9e7e460de29e08f7a21029042fa45b12. This PR remains stacked against
corpus/sc44-lemma241-core (#980); SC-45 #981 remains a separate correction unit.
No merge, native-round F/E designation or special workflow launch is authorized.

Results and scientific scope

• Preserve Lean-certified Proposition 24.2's block-word rigidity with its full
  word-trace and block-spanning premises. Theorem 24's restricted coherent-channel
  invariances, broader population relabellings and open all-partitions completeness
  problem remain distinct.
• Retain Theorem 23's conditional outputs for the selected cubic branch. The
  companion-family inverse calculation identifies parameters within that family;
  it does not select the actual physical substratum.
• Add the written finite-dimensional predictive-realization and passive quotient
  results (Propositions 24.3–24.4) under their reachability, observability, support
  and state-bound hypotheses. They do not establish a complete physical gauge group.
• Add Proposition 24.5's known-model record-fibre identification criterion. The
  reversible swap and cyclic-register controls separate predictive memory from
  residual hidden-state uncertainty. Preparation, retained records and permitted
  operations are explicit hypotheses. P24.5 is GAP, without a Lean certificate.
• Preserve the restricted SM3 chirality formalization and the conditional gauge,
  faithful-image and hypercharge recovery results, including the exceptional
  anomaly-free branch. Their physical selection assumptions remain explicit.

The hidden-state correction preserves all 133 prior canonical statement bodies
and fingerprints, with P24.5 bringing the census to 134. The source corrections
propagate through the papers, book chapters and FULL mirror. Physical branch
selection, the coherent measurement/Bell mechanism, the observable-to-gauge bridge
and the matter/gravity coupling remain research obligations.

Publication and review evidence

Markdown, canonical TeX and rebuilt PDFs are committed. The latest scientific
increment rebuilt Main, Computation, Explainer, Structure, Substratum and the book
with zero missing glyphs; earlier recovery builds are retained. Source stamps,
PDF hashes and page statistics are recorded in
verification/audits/manuscript/hidden-state-identifiability-20261028/.

Independent candidate-v2 review accepted the scoped P24.5 correction and its
propagation, with 49 pinned scientific-source/control hashes. Those files remain
unchanged by both verification-only increments. That acceptance is separate from a full
repository release, broader physical correctness or a Lean proof of P24.5.
Independent review accepted the census repair at f420020 (tree
34c9e4da4945deb58e62d79ebd1a74e42da0b80a), retaining all 34 original patterns
and rejecting four controlled mutations. Its report SHA256 is
7f01ac9bc36e4dddd895935d23445fc669173b57bd9b4b311c00781b4fd2ca2a.

Independent review also accepts the latest publication-guard repair at tree
eab5fff9a0e9470d4b4ef3393f7254ed2b863964. All 62 pinned inputs match the
published tree, all 253 existing function ASTs are unchanged, and the isolated
A6/Audit B checks and eight added deletion controls pass. Review SHA256:
0571840e43a452f263412fcbd34eab60c01294013fd0621dd81f9c72ba1d0a24.
These are scoped content reviews, separate from full release acceptance.

Concrete forward source reconciliation

The Introduction reconciliation identified a residual book Chapter 3 summary that still claims a proof terminating the deeper-substratum regress. #992 provides a narrow correction and canonical book rebuild on a separate branch; it is not integrated into this held head. This concrete source-consistency finding must be reconciled before full release acceptance. The 25-job CI result below remains valid for the unchanged 20ca05f tree; green CI alone does not close this scientific-source finding.

Current verification and remaining gates

Normal PR CI run 40703739036
completed successfully: all 25 executed jobs passed for current head 20ca05fe9e7e460de29e08f7a21029042fa45b12. Ordinary automatic PR checks are
not native F/E attestations.

On preceding head f420020, Lean kernel, Mathlib, core A and KTOKEN passed,
and the audit census passed. Core B then reached two stale publication guards.
The latest increment reconciles Audit B with SM's complete selected-family
qualification and five A6 positive text contracts with the recovered
stabilizer/quotient/charged-action distinctions. It retargets the dead A6
state-dependent mutation while preserving the positive text and its rejection.

Bounded isolated checks pass both guard sections, six Audit B deletion mutations,
all eight A6 baseline functions and 22 existing conjunctions, plus two additional
qualification-deletion controls. All other A6 predicates and frozen control-plane,
legacy-record, census-prefix and chronology pins remain required. This is mutable
downstream content-guard maintenance, not a re-execution or amendment of the old
native round. Neither guard-only increment changes manuscripts, theorems, Lean,
TeX or PDFs; the 49 pinned scientific-source hashes remain unchanged.

Normal PR CI is green on this integrated source. Independent integrated release
acceptance remains required. The shallow local checkout leaves three historical V3
vectors explicitly undecidable; no local full-history or full release attestation
is claimed. Owner merge approval is still required, and no merge is proposed here.

[amaybaum]

Theorem 23 scientific review finding / writer correction (exact draft head 09239e06e764327ec774e73dd557a7c1b2b578bd). The earlier reconciliation preview still assumed, without proof, that E1–E7 and A1–A6 uniquely select the local branch. In fact, Stage 2(b) needs the nearest-neighbor second-order class, Stage 2(c) uses H-link and m=1, and later steps use the selected cubic/octrahedral representation plus additional [SM] assumptions. The Stage-2 sentence claiming every step (a)–(f) was an iff was not supported. The writer has removed that inference and narrowed Theorem 23 to conditional local-branch outputs and Lemma 23.0 to shared outputs on the same selected branch—not equality of full substrata modulo gauge. Direct opening, theoretical-status, interpretation, synthesis and conclusion claims have been adjusted. The proof-coverage ledger keeps T-23 and L-23.0 at GAP; Theorem 24 at GAP; only Proposition 24.2 is K3 with its two Lean directions.

Review record: verification/audits/manuscript/intro-corpus-audit-20261026/tracks/SC44-SC45-reconciliation/README.md (section F-23). Current CI verify run 40645479455 was queued at the last check. Earlier run 40643741304 failed downloading Lean 4.33.0 with SSL/TLS errors before the Mathlib checking (missing downstream reports, not scientific negatives). The separate SC-44 Lean kernel installation previously failed downloading 4.35.0. Canonical Substratum .tex/.pdf rebuild, complete exact-head checks, independent scientific acceptance and external-site propagation remain blockers. This is a draft source correction, not merge authorization.

[amaybaum]

Foundational merge record — #982

Owner authorization: "Proceed as suggested" after the exact-head merge recommendation.

Merged: #982 using GitHub's normal merge method, with expected PR head 7ac361440e879f449aa108e127347544f932c83c. GitHub returned merged: true, merge commit dffe1945989f60632c75b2b3e2e96afad2af8509; the PR is now closed/merged.

Pre-merge gate: main was 0d917665d32287e5d0ce649012175f2c7bf75b09; #982's head, scientific/provenance acceptance and the narrow attribution-only follow-up were reviewed. Exact-head verify run 40767146156 completed successfully (25 jobs passed, 1 optional skip), including Lean kernel and Mathlib bridge. The PR was marked ready before merging.

Post-merge identity: comparison 7ac36144...dffe1945 shows the merge commit is directly ahead by one commit and has zero changed files; main points to the new merge commit (comparison dffe1945...main is identical). The connected commit-read action omits parent/tree fields, so the ordered parent objects were not directly re-read from Git metadata; normal merge method and the pre-merge main/head identities were independently checked.

Post-merge CI: no successful main-push run is claimed by this record. The available commit-run query covers PR-triggered runs only and returned no post-merge run; the combined status endpoint returned no statuses. Main's exact merge-commit CI therefore remains to be confirmed from Actions.

Scope: #992 was incorporated as source; SC-44/45 and #992 source/evidence PRs remain separate provenance, not duplicate merges. All scoped proof-grade qualifications remain. #969 is not automatically superseded. #918 remains on hold for correction of its SC-39 Bell/card statement and dependency reconciliation.

Next: reconcile recovery #990 onto current main, verify and independently review its resulting landing candidate; then #993, followed by other scoped recovery work. No further PR has been merged or automatically closed.