Alphabeta Math
LemmaStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedjudge pass (gpt-6.1-sol)
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Bender--Knuth involutions permute the weights of semistandard tableaux

Statement

Fix r≥1 and a skew shape ν/λ (for a partition shape take λ=∅), let k∈{1,…,r−1}, and let T be a semistandard skew tableau of shape ν/λ with entries in {1,…,r} (Skew diagrams and semistandard skew tableaux, Semistandard tableaux and Kostka numbers). Call an entry k or k+1 of T free if there is no k+1 respectively no k in the same column. Then:

(i) the free positions in each row occupy consecutive cells of that row; (ii) replacing in every row the free k's and free k+1's by their complementary counts (if the row has ai free k's and bi free k+1's, then after the replacement it has bi free k's and ai free k+1's in the same free cells, the remaining entries unchanged, the free cells filled from left to right by the bi copies of k followed by the ai copies of k+1) produces again a semistandard skew tableau σk(T) of the same shape, with entries in {1,…,r}; (iii) σk is an involution of the set of semistandard skew tableaux of shape ν/λ with the same set of free positions, and wt⁡(σk(T))=skwt⁡(T), where sk is the transposition of k and k+1 acting on weights (Semistandard tableaux and Kostka numbers, with weights read as vectors in Zr).

Consequently, for every weight α∈Z≥0r the number of semistandard skew tableaux of shape ν/λ and weight α equals the number of weight skα, and the generating function ∑Txwt⁡(T), over semistandard skew tableaux of shape ν/λ with entries in {1,…,r}, is symmetric in x1,…,xr (Partitions, English diagrams, and conjugation).

Facts & Assumptions

Given: r≥1, a skew shape ν/λ with λ,ν partitions and [λ]⊆[ν], an index k∈{1,…,r−1}, and a semistandard skew tableau T of shape ν/λ with entries in {1,…,r}.

[F1]

A semistandard skew tableau fills the cells of the skew diagram ν/λ with positive integers, weakly increasing from left to right in each row and strictly increasing from top to bottom in each column; its weight is wt⁡(T)=(a1,…,ar) with ai the number of entries equal to i, and its monomial is xwt⁡(T) (Semistandard tableaux and Kostka numbers, Skew diagrams and semistandard skew tableaux).

[F2]

The Young diagram [ν] consists of the cells (i,j) with 1≤i, 1≤j≤νi (English coordinates), and its columns are the j's with j≤νi, so a cell (i,c) belongs to ν/λ exactly when λi<c≤νi (Partitions, English diagrams, and conjugation, Skew diagrams and semistandard skew tableaux).

Proof

Given: r, ν/λ, k, T as above.

1.1F2givenalgebra

In a skew diagram, row i consists of the interval λi<c≤νi, and column c consists of the interval λ′c<i≤ν′c. Two rectangle-completion properties will be used. If (i,c),(i,c′),(i′′,c) are cells with i<i′′ and c′<c, then (i′′,c′) is a cell: λi′′≤λi<c′<c≤νi′′. If (i,c),(i,c′),(i′′,c) are cells with i′′<i and c<c′, then (i′′,c′) is a cell: λi′′<c<c′≤νi≤νi′′. These use the weakly decreasing row lengths of both partitions.

1.2F1givenalgebra

No column contains two free cells: a column contains at most one k and at most one k+1 by strict increase, a column containing a free k contains no k+1 at all (freely), and a column containing a free k+1 contains no k at all (freely); so a column cannot contain both a free k and a free k+1, and cannot contain two entries equal to the same letter. Consequently, in the modification of (ii) each column changes in at most one cell.

2.1F1F2givenstep 1.1algebra

Structure of the free cells of a row. Let (i,c) be a cell of T with entry k that is not free. Then some cell (i′′,c) of the same column has entry k+1; by strict increase of the column i′′>i. If (i,c′) is a cell of the same row with c′<c and entry k, then (i′′,c′) is a cell by step 1.1, and inside row i′′ one has T(i′′,c′)≤T(i′′,c)=k+1, while strictly increasing column c′ gives T(i′′,c′)>T(i,c′)=k; hence T(i′′,c′)=k+1 and (i,c′) is not free. So the non-free k's of a row form an initial segment of its block of k's, read from the left. Symmetrically, if the entry k+1 at (i,c) is not free, there is a cell (i′′,c) with i′′<i and entry k; for a cell (i,c′) with c′>c and entry k+1, step 1.1 makes (i′′,c′) a cell, and T(i′′,c′)≥T(i′′,c)=k by weak increase in row i′′ and T(i′′,c′)<T(i,c′)=k+1 by strict increase in column c′, so T(i′′,c′)=k and (i,c′) is not free. Hence the non-free k+1's of a row form a final segment of its block of k+1's. Since the entries of a row are weakly increasing, the cells carrying k precede those carrying k+1, and combining the two statements the unblocked cells carrying k (a final segment of the k-block) and those carrying k+1 (an initial segment of the k+1-block) form one consecutive block of cells of the row, proving (i).

3.1F1step 2.1step 1.2algebra

The modification produces a semistandard tableau. Rows: by step 2.1 the free cells of a row form consecutive cells, the entry immediately left of the block, if present, is a non-free k or a smaller letter, the entries of the block after the modification lie in {k,k+1} and are filled weakly increasingly, and the entry immediately right of the block, if present, is a non-free k+1 or a larger letter; so rows stay weakly increasing. Columns: by step 1.2 only one entry of a column can change. If a free k at (i,c) is replaced by k+1, then column c contains no k+1, so every entry above (i,c) is <k<k+1 and every entry below is >k+1, and strict increase persists; if a free k+1 is replaced by k, column c contains no entry k, so every entry above is <k and every entry below is >k+1>k, and strict increase persists. The entries stay in {1,…,r} because 1≤k<k+1≤r. This proves (ii).

4.1F1step 2.1step 1.2step 3.1algebra

Involution and weight. The modification is reversible: it is performed on the free cells, and by step 1.2 the free cells of σk(T) are the same cells (a cell that was free remains the only cell of its column with an entry in {k,k+1}, hence remains free), while every other cell is unchanged, so no free cell is created or destroyed. Hence σk is an involution with the same free cells. For the weight, let Ak and Ak+1 be the numbers of entries equal to k and to k+1 in T, and let a=∑iai, b=∑ibi be the total numbers of free k's and free k+1's. The non-free k's are in bijection with the non-free k+1's: send a non-free k to the unique k+1 below it in its column (existence is the definition of non-free, uniqueness is strict increase in the column, and the image is non-free because its column contains that k); the inverse sends a non-free k+1 to the unique k above it. Hence Ak−a=Ak+1−b. The modification deletes the a free k's and b free k+1's and inserts b copies of k and a copies of k+1 in their place, so the number of k's in σk(T) is Ak−a+b=Ak+1=(skwt⁡T)k and the number of k+1's is Ak+1−b+a=Ak=(skwt⁡T)k+1, all other letter counts being unchanged. Thus wt⁡(σk(T))=skwt⁡(T).

5.1F1step 3.1step 4.1algebra∎

Consequences. By step 3.1 and 4.1, σk is a weight-sk-equivariant involutive bijection of the set of semistandard skew tableaux of shape ν/λ with entries in {1,…,r}; hence it restricts to a bijection between the tableaux of weight α and those of weight skα, so those two sets have the same cardinality. Consequently the generating function G(x1,…,xr)=∑Txwt⁡(T) satisfies G=∑Txskwt⁡(T), which is G with xk and xk+1 interchanged; thus G is invariant under each adjacent transposition of the variables. Every permutation of {1,…,r} is a product of adjacent transpositions (bubble-sort any ordering), so G is invariant under all permutations of the variables, that is, symmetric.

Depends on

Used by

Dependency tree · two levels

5 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources