Skip to content

Prove an unbounded field-to-dyadic correction rank gap (no kappa) - #7675

Open
aliromman wants to merge 2 commits into
CrocSwap:mainfrom
aliromman:codex/consecutive-result-020
Open

aliromman wants to merge 2 commits into
CrocSwap:mainfrom
aliromman:codex/consecutive-result-020

Conversation

@aliromman

@aliromman aliromman commented Nov 18, 2026 •

Copy link
Copy Markdown
Contributor

Result and scope

The field-minimum correction rank in PR #7654 need not be attainable over Z[1/2]. This package classifies the minimum rank(P-I) for every normalized two-role dyadic row/column contract. It also gives an unbounded gap: m prescribed pairs need rank m over Q and 2m over Z[1/2], even with arbitrary coupling to extra coordinates.

At m=1 the contract is P*e1=4e1+e2 and e2^T*P=e1^T. The unique rank-one completion has integer entries but determinant 3, so its inverse is not dyadic. A three-ADD word attains rank two with determinant one. The larger family uses 3m ADDs. These are algebraic operation counts, not physical tape prices or instruction optima.

Read the proof, exact receipt, verifier and scope record. The two-role criterion tests whether the required determinant is a dyadic unit. The larger-family lower bound reduces modulo 3 and bounds the fixed space; it does not assume independent pairs in the competing completion.

No new kappa, supplier admission or selected-assumption discharge is claimed. All five assumptions and selected PR #7000 remain unchanged. This is a scoped proof with finite exact corroboration, completing the advance claim.

Evidence and reproduction

Normal and optimized CPython 3.12.16 checks passed on macOS arm64. The verifier examines all 6561 matrices in the stated half-integer grid, including 2480 dyadic invertible matrices, and all 308 normalized integer pairs in [-5,5]^4. It constructs attaining completions for all 928 distinct contracts; 336 have a strict field/dyadic gap. A separate field formula agrees. Literal words and inverses pass on 16 family shapes. All 11330 admissible reduced completions in four complete modular domains satisfy the lower bound, including coupled free blocks. Six controls reject invalid coefficients, determinants or endpoints.

Metadata, committed bytes, executable modes, source ancestry and the official seven-file export passed. The optimized read-only export replay passed without Git or a usable PATH; all eight payload/manifest hashes and modes stayed unchanged. The proofs and executable have no external run-time input.

From the repository root, with CPython 3.12:

python3.12 -I -W error -B research/dyadic-correction-rank-gap/verify.py
python3.12 -I -W error -B -O research/dyadic-correction-rank-gap/verify.py
python3.12 -B scripts/research_package.py --package research/dyadic-correction-rank-gap
python3.12 -B scripts/check_research_commit.py --package research/dyadic-correction-rank-gap --check-source-history --check-modes

Expected status markers:

PASS_DYADIC_CORRECTION_RANK_GAP
PASS_REGISTERED_RESEARCH_PACKAGE
PASS_RESEARCH_COMMIT_BYTES

Hosted CI did not certify this package. The plan job failed in the shared test_current_ci_section_matches_selected_record: the base guide names frontier6030 while the selected record requires make frontier7000-verify. The log reports 760 tests, one failure and 24 skipped; package replay was skipped. This package changes no shared CI file. Local replay remains the scientific evidence. Predecessor scientific replays, the selected cold replay, Lean and Docker were not run. Independent human and formal-kernel review are unclaimed.

Credit and remaining obligations

Prepared by @aliromman with substantial OpenAI Codex assistance. Nick Petalas (@petalas) supplied the field formula and dyadic question in PR #7654, exact proof, and the known completion parameterization in PR #6016, exact proof. Ali Romman's PR #5611, exact proof, supplies the helper interface. @Dugongue posed GQ08 in PR #5803. No predecessor code was copied. Classical unit algebra and modular rank reduction are credited as standard tools.

The next arithmetic task is a criterion for general paired contracts. Individually saturated prescribed and fixed modules can have a sum with odd index, as the explicit example shows. An actual helper cut, operation optimality, range/precision bounds, physical delivery, native charts, restoration and the complete paid/target invoice remain open. Correction rank alone is not a completed supplier price. Base: canonical main e54523517d0f03e58352e647c403a5af3e525a57, priority revision 8.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant