Schur polynomials are symmetric #
A Schur polynomial is the generating function of the semistandard tableaux of its shape, one
monomial per tableau, the exponent of xᵢ being how often the letter i occurs. Nothing in that
description is symmetric in the letters: the tableaux are ordered objects, and the alphabet
{0, …, N - 1} enters TauCeti.diagramSchurPoly through the row-weak and column-strict conditions.
Symmetry is a theorem, and this file proves it.
The mechanism is the Bender-Knuth involution of
TauCeti.Combinatorics.Young.BenderKnuth: for each letter v it is a bijection of the tableaux of
a fixed shape which exchanges how often v and v + 1 occur and fixes every other multiplicity
(SemistandardYoungTableau.content_benderKnuth). Reindexing the defining sum along it therefore
exchanges the variables x_v and x_{v+1}, which is
TauCeti.rename_swap_diagramSchurPoly. Adjacent transpositions generate the symmetric group
(Equiv.Perm.mclosure_swap_castSucc_succ), so this one exchange gives the whole symmetry,
TauCeti.isSymmetric_diagramSchurPoly.
The involution moves only the letters v and v + 1, so it preserves the alphabet only when both
of them belong to it; that is why the exchanged variables are given as a pair a, b : Fin N with
(b : ℕ) = a + 1 rather than as an arbitrary pair of letters. Exchanging two letters that are not
adjacent is not a Bender-Knuth move, and is obtained here only through the group generated by the
adjacent ones.
Symmetry is what makes TauCeti.schurPoly independent of the ordering of its alphabet, so it also
retro-justifies the definition of TauCeti.schurPoly on an arbitrary finite alphabet by an
arbitrary choice of ordering, and it places the Schur polynomials in
MvPolynomial.symmetricSubalgebra, where the elementary and complete homogeneous symmetric
polynomials they generalize already live.
Main results #
TauCeti.rename_swap_diagramSchurPoly: a Schur polynomial is unchanged by exchanging two adjacent variables.TauCeti.isSymmetric_diagramSchurPolyandTauCeti.schurPoly_isSymmetric: Schur polynomials are symmetric.TauCeti.schurPoly_mem_symmetricSubalgebra: the same, read inMvPolynomial.symmetricSubalgebra.
References #
- R. P. Stanley, Enumerative Combinatorics, Volume 2, §7.10, Theorem 7.10.2: the skew Schur
function is symmetric, proved by the same reduction to interchanging
xᵢandx_{i+1}. The involution is that of E. A. Bender and D. E. Knuth, Enumeration of plane partitions, J. Combinatorial Theory 13 (1972), 40--54. - W. Fulton, Young Tableaux, Section 2.2.
- Schur--Weyl roadmap, Layer 7.
The Bender-Knuth involution on bounded tableaux, exchanging two adjacent letters a and
b of the alphabet. Adjacency is what keeps the alphabet: the involution writes no letter other
than the ones already present and the two it exchanges.
Equations
- T.benderKnuth hab = ⟨(↑T).benderKnuth ↑a, ⋯⟩
Instances For
The Bender-Knuth involution exchanges two adjacent letters, read on the weight of a bounded tableau.
The weight of the Bender-Knuth image of a tableau, as the transported weight.
A Schur polynomial is unchanged by exchanging two adjacent variables: the Bender-Knuth involution at the corresponding letter reindexes the sum of monomials defining it.
Schur polynomials are symmetric. Adjacent transpositions generate the symmetric group, so
the single exchange TauCeti.rename_swap_diagramSchurPoly supplied by the Bender-Knuth involution
gives the whole symmetry.
Schur polynomials are symmetric, on an arbitrary finite alphabet. This is what makes
TauCeti.schurPoly independent of the ordering of the alphabet chosen to define it.
Schur polynomials are symmetric, read as membership in
MvPolynomial.symmetricSubalgebra, where the elementary and complete homogeneous symmetric
polynomials they generalize already live.