English

Skew Howe duality and limit shapes of Young diagrams

Representation Theory 2023-09-25 v3 Combinatorics Probability

Abstract

We consider the skew Howe duality for the action of certain dual pairs of Lie groups (G1,G2)(G_1, G_2) on the exterior algebra (CnCk)\bigwedge(\mathbb{C}^{n} \otimes \mathbb{C}^{k}) as a probability measure on Young diagrams by the decomposition into the sum of irreducible representations. We prove a combinatorial version of this skew Howe for the pairs (GLn,GLk)(\mathrm{GL}_{n}, \mathrm{GL}_{k}), (SO2n+1,Pin2k)(\mathrm{SO}_{2n+1}, \mathrm{Pin}_{2k}), (Sp2n,Sp2k)(\mathrm{Sp}_{2n}, \mathrm{Sp}_{2k}), and (Or2n,SOk)(\mathrm{Or}_{2n}, \mathrm{SO}_{k}) using crystal bases, which allows us to interpret the skew Howe duality as a natural consequence of lattice paths on lozenge tilings of certain partial hexagonal domains. The G1G_1-representation multiplicity is given as a determinant formula using the Lindstr\"om-Gessel-Viennot lemma and as a product formula using Dodgson condensation. These admit natural qq-analogs that we show equals the qq-dimension of a G2G_2-representation (up to an overall factor of qq), giving a refined version of the combinatorial skew Howe duality. Using these product formulas (at q=1q =1), we take the infinite rank limit and prove the diagrams converge uniformly to the limit shape.

Cite

@article{arxiv.2111.12426,
  title  = {Skew Howe duality and limit shapes of Young diagrams},
  author = {Anton Nazarov and Olga Postnova and Travis Scrimshaw},
  journal= {arXiv preprint arXiv:2111.12426},
  year   = {2023}
}

Comments

57 pages, 15 figures, 2 tables; v3 fixed typos, added comparison to Biane's result, updated references, fixed typos; v2 fixed typos in Theorem 4.10, 4.14, shorter proof of Theorem 4.6 (thanks to C. Krattenthaler), proved of Conjecture 4.17 in v1