Supporting correction report · Website reading copy · Released 12 November 2026
CM Model Correctness Audit
This supporting resource records which conclusions require which algebra, tensor product, transformations and readout rule. It belongs beneath Correspondence Matrices / CM-LM. It does not replace the foundations paper or certify all later programme results.
Read the dates correctly. The original report is dated 18 September 2026 and keeps its original byline: “Prepared for Brian / Quantum CM project.” It predates the current September 27 foundations revision. The current foundations manuscript already disclaims physical measurement and speedup; this package is not a claim that all historical mistakes persist there.
Original 28-page PDF · Exact source · Fresh bounded checks and test receipt · Historical source and evidence · Correction ledger · Provenance
Four contracts to keep separate
- Native CMs/LMs: truth tables, formula-valued operators and logical pairing.
- Rotation operators: the selected commuting algebra generated by a four-cycle.
- Shared module: a shared coefficient register, tensoring over the rotation algebra and restricted linear operations.
- Literal independent registers: Boolean vector-space tensor factors with explicitly declared restricted or enlarged invertible gates and modal readout.
Corrections and current relevance
AU-01 · Native CM versus selected rotation operator
The 2×2 truth-pattern identity remains idempotent; its distinct selected 4×4 operator can be square-zero. Pointwise logic, matrix composition and cyclic operator composition have different types.
Original report: Sections 3–4; corrected nilpotent. Already compatible with the current typed CM/LM framing; retain as a misuse-prevention example.
AU-02 · Logical pairing versus modal measurement
Logical truth extraction does not supply a possible/impossible effect rule, state-update rule or probability law. Those require additional contracts.
Original report: Section 3; LM extraction. Current public CM/LM manuscript and project explicitly disclaim physical measurement. Supporting clarification, not a new erratum against that edition.
AU-03 · Shared module versus independent registers
Their tensor products and dimensions differ. Nonzero shared-module preparations can tensor to zero. The coefficient embedding need not preserve separability.
Original report: Sections 5–6. Applies to the historical modal branch and future BR/PC/P01 releases; not a refutation of the Boolean foundations calculus.
AU-04 · Resource and analyzer accounting
Literal one-copy universal eight-dimensional transfer requires rank-eight resources and a complete 64-outcome analyzer under the declared enlarged gate set. A logical-only four-label readout is insufficient.
Original report: Sections 9–10. Retain with historical protocol claims and future specialist releases; no physical quantum advantage follows.
AU-05 · Finite-copy activation
The shared rotation-linear no-activation result cannot be exported to literal tensoring with unrestricted local maps and coordinate filters. A two-copy rank-six witness is supplied. Heralded possibility has no numerical success probability here.
Original report: Section 11; finite-copy activation. Model-specific correction; do not make it an unconditional statement about quantum teleportation.
AU-06 · Restricted classifications
Fifteen Smith classes and the 192-basis contextuality results describe declared restricted families. Enlarging local transformations changes the invariant problem. Zero is not an allowed modal preparation.
Original report: Sections 5–8 and class tables. Owner papers P01/P02/PC carry broader generalizations. This audit retains bounded evidence and correction provenance.
AU-07 · Orthogonal corrections
The stated even-dimensional rank-one complete Bell scheme cannot have all receiver corrections transpose-orthogonal. The enlarged P,T gate set does not preserve a nonzero common four-dimensional bilinear form.
Original report: Section 12. The architecture restrictions matter; no universal no-go for every ancilla-assisted instrument or symplectic model is asserted.
AU-08 · Copying stored coefficients versus modal cloning
Stored classical bit strings can be copied. The modal cloning restriction concerns unknown tensor-state vectors and a fixed linear evolution class.
Original report: Section 13. Use as a contract distinction, with established modal-theory attribution; not a new general cloning theorem.
AU-09 · Support versus probability
Possibilistic no-signalling does not guarantee strictly positive probability for every allowed event. Six allowed events in the supplied Bell table are forced to zero by probabilistic no-signalling constraints. Weak resolution remains possible.
Original report: Section 14. Established modal antecedents; retain as the exact finite regression example, not a new nonlocality discovery.
AU-10 · Evidence and claims
Inherited tests, full driver enumerations, finite instances, general proofs and new package checks have different scopes. Timing receipts establish neither speedup nor physical realization.
Original report: Sections 2, 16–17 and Appendix A. Applies across the programme. Historical manuscript file-library citations do not mean the audit reviewed today’s full CM/LM edition.
Evidence status
The report records 59 historical tests, a 31-test independent suite and large finite enumerations. These are inherited claims unless a fresh result is explicitly recorded. This preparation checks the source/PDF identity, package completeness, selected mathematical witnesses and the independent regression suite; it is not a replay of every driver or formal proof certification.
The historical root manifest has eight CRLF/LF differences and one unresolved timing-file mismatch. All three nested historical manifests match. The new packaging manifest binds the current bytes; it does not repair or overwrite the old record. Timing logs are not performance benchmarks. Some original reconstruction paths refer to the historical workspace.
Antecedents and limits
Finite-field modal states, Bell/no-cloning phenomena and related information-processing constructions have established antecedents: Schumacher–Westmoreland, Modal quantum theory; Almost quantum theory discusses weak probabilistic resolution; Diamond–Schumacher treats cloning, deleting and hiding. This package claims correction provenance and exact model-specific examples, not priority for those general ideas.
For B-Theory and The MIND, its value is guarding transfers from representation to observation and execution. It supplies no physical quantum device, empirical cognitive mechanism, consciousness criterion, compression guarantee or computational speedup.
The HTML is a reading guide and correction ledger, not a full mathematical transcription. Use the original PDF and source for proofs, equations and tables. Extracted text is a limited prose aid. No new license or external peer review is inferred.