Skip to content

Exclude eight-row fifteenth-root partial matrices and the full q+3 family - #7762

Open
aliromman wants to merge 1 commit into
CrocSwap:mainfrom
aliromman:codex/consecutive-result-075
Open

aliromman wants to merge 1 commit into
CrocSwap:mainfrom
aliromman:codex/consecutive-result-075

Conversation

@aliromman

@aliromman aliromman commented Nov 19, 2026 •

Copy link
Copy Markdown
Contributor

Scoped negative result: there is no (q+3)-by-(q+2) partial Butson matrix over roots of order 3q^b, for any prime q>=5 and b>=1. In particular, the eight-row, fifteenth-root input left open in PR #7744 does not exist. No new kappa follows.

Read the proof, complete eight-row certificates, exact checker, source pins and credits.

Projection to third roots gives a Gram matrix whose off-diagonal entries have magnitude q. A nontrivial triangle phase gives a negative minor, so all phases must agree. The resulting columns are directed edges; every compatible family with at least three edges is a star. This forces the missing column to retain its original alphabet. A full completion then has the form Q+(omega-1)I, up to conjugating omega, with Q Hermitian and diagonal one. Orthogonality forces Q²−3Q−qI=0. Its two irrational eigenvalues cannot have integer multiplicities with trace q+3. Both the completion and nonexistence steps are proved; this is not a failed-search claim.

The checker exhausts all 319,770 eight-term root multisets, all 6,561 ternary words and all 1,540 pairs of the 56 possible edge words. The 15 vanishing multisets match exactly one three-cycle and one five-cycle. A maximal-clique search and a separate fixed-size search agree on all 112 possible six-edge configurations. All missing-column signs agree. Checks over the 24 primes from 5 through 101 cover 648 phase assignments and 2,408 row residues. Exact matrix arithmetic and controls reject changed conjugation, diagonal shift, linear coefficient, phase and sign. The pinned source verifier runs unchanged.

Python 3.12, standard library only:

python3.12 -B scripts/research_package.py --package research/mixed-cycle-column-obstruction

Both registered commands print:

PASS_MIXED_CYCLE_COLUMN_OBSTRUCTION
319770 root multisets; 15 vanishing sums; 6561 ternary words; 1540 edge pairs
44 maximal cliques; all 112 six-edge configurations; exact sign and trace controls
No (q+3)-by-(q+2) input over roots of order 3q^b, q prime >=5: written proof

Normal and optimized registered replay, syntax/indentation, committed bytes/source history/modes, export and sealed read-only Git-free serial replay pass. 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.

Nick Petalas (@petalas) supplied the two-prime proof, reduction and open question and unchanged checker. His completion proof from PR #5865 is a dependency. The retained notices credit David Ganz’s atlas and pilot applications in PR #2849 and PR #2855. Lam and Leung supply the classical cycle decomposition.

The arbitrary-prime and arbitrary-power result is the written proof. No valid mixed-case matrix is constructed. Other row counts, larger powers of three, more prime factors, noisy recovery, physical implementation and a complete caller bill remain open. The atlas/pilot suites, selected cold replay, Lean and Docker were not run. All five framework assumptions and the selected frontier remain unchanged. No global literature priority or independent human/formal-kernel acceptance is claimed.

Base: canonical main e54523517d0f03e58352e647c403a5af3e525a57, priority revision 8. Six source files are retained at the exact head above; no unmerged ancestry is asserted. Prepared by Ali Romman (@aliromman) with substantial OpenAI Codex assistance in research, implementation, verification and writing. Apache-2.0. Canonical claim.

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