English

Cubulation of Bruhat graphs

Representation Theory 2026-05-14 v2 Combinatorics Group Theory

Abstract

For (W,S)(W,S) an arbitrary Coxeter system and any yWy \in W, we investigate the condition that the Bruhat graph for the interval [1,y][1,y] can be cubulated, meaning roughly that this graph can be spanned by a product of subintervals of Z\mathbb{Z}. Results of Carrell-Peterson and Elias-Williamson imply that if [1,y][1,y] can be cubulated, then the Kazhdan-Lusztig polynomial Px,y=1P_{x,y} = 1 for all xyx \leq y. We consider the converse to this result. For (W,S)(W,S) finite and w0w_0 the longest element in WW, so that Px,w0=1P_{x,w_0} = 1 for all xWx \in W, we use normal form forests to construct cubulations of [1,w0][1,w_0] in types AA and B/CB/C. However, in some exceptional types, we determine elements yWy \in W such that P1,y=1P_{1,y} = 1 but [1,y][1,y] cannot be cubulated. We then prove that if there are infinitely many yWy \in W such that [1,y][1,y] can be cubulated, then (W,S)(W,S) must be of type A~n\tilde{A}_n for some n1n \geq 1. Finally, for (W,S)(W,S) of type A~2\tilde{A}_2, we exhibit a cubulation of [1,y][1,y] for each of the infinitely many yWy \in W such that Px,y=1P_{x,y} = 1 for all xyx \leq y.

Cite

@article{arxiv.2504.03046,
  title  = {Cubulation of Bruhat graphs},
  author = {Alex Bishop and Elizabeth Milićević and Anne Thomas},
  journal= {arXiv preprint arXiv:2504.03046},
  year   = {2026}
}

Comments

35 pages; 10 figures, most in color. Version 2: new title, reorganization and shortening of some content, added discussion of equivalence between Lehmer codes and cubulations, added discussion of our computations

R2 v1 2026-06-28T22:46:01.538Z