Ferrers Constructions: Corrected Tableau Maps and a KR1 Boundary at Weight 41
Research update: 26 October 2026. The revised Ferrers-diagram preprint corrects two mathematical errors and records a failure of the proposed general KR1 construction. This work does not provide a complete bijective proof of the Kanade–Russell I₁ identity. The erratum explains precisely which claims have changed.
Christopher Yamba’s suggestion to explore an additional Ferrers coordinate motivated useful local tableau moves and partition-coordinate experiments. Revision 2 preserves those results while separating their different levels of support.
- Corrected tableau routing: the map selects the first candidate that is both a valid tableau and preserves graph-inversion weight. The earlier displayed inequalities were insufficient. Complete routing and selected degree-by-degree cancellations are checked exhaustively over all 42 Hessenberg vectors at n=5 and all 132 at n=6.
- Corrected determinant recovery: under the stated row-to-column convention, the decreasing shifted-key order is the determinant permutation itself, σ=τ. The revised written proof establishes both inverse identities; tests recover the columns, permutation and signs through n=6.
- Finite KR1 success and a boundary: full roundtrips pass for 5,461 states of weights 0–40. At weight 41, two distinct clusters map to the same columns, (19,12,6,3,1). Thus the all-weight extension of the current rules is false. The construction also uses a finite apex table.
- A useful restricted extension: the a=3,b=0 sector passes 4,491 states through weight 60. This extends the computational evidence for that sector; it does not prove validity on its infinite domain.
- Formal assurance: 19 indexed supplementary Lean statements cover specified local moves under their hypotheses and arithmetic identities. They do not establish the complete routing algorithms, the general beta-rearrangement theorem or a KR1 bijection. GitHub CI rebuilt the unchanged Lean Std certificate, reporting 73 theorem results across the full certificate; the verification package includes its source-bound record.
The research has improved in precision and tested coverage. A complete KR1 construction remains a goal. The immediate task is to repair the +6,+3,+1 orbit near (19,12,6), preserving disjointness from the lower-sector image and proving a unique inverse. Adding missing apex assignments alone cannot repair the demonstrated collision.
Read the erratum · Revised article and abstract · PDF · LaTeX · Verification package · Manifest and hashes · Research PR #15
Author: Christopher Yamba. Original collaborator: Antigravity (Google DeepMind). The original recorded timestamp, 2026-10-24T22:42:00Z, is retained as provenance; this corrected update is dated 26 October 2026. No claim of a new complete KR1 proof or verified novelty is made. File hashes and timestamps identify artifacts and their history; they do not establish mathematical validity or priority over prior work.