Let $X$ be the poset of positive roots of a finite root system of Dynkin type $Q$.

Question 1: In Dynkin type $A_n$, is it true that the poset of order ideals of $X$ is isomorphic to the poset of [2,1,3]-avoiding permutations under strong Bruhat order?

I tested this experimentally in Sage, for example:

W=[p for p in S if p.avoids(  [2,1,3])]
B = Poset([W, lambda x,y: x.bruhat_lequal(y)])

Question 2: Is there a generalisation to the other Dynkin types so that the poset of order ideals of $X$ is isomorphic to the poset of $\{g_i \}$-avoiding permutations under the strong Bruhat order of the Weyl group for some group elements $g_i$?


1 Answer 1


Regarding Q1: I believe this is proved in "A Distributive Lattice Structure Connecting Dyck Paths, Noncrossing Partitions and 312-avoiding Permutations" by Barcucci et al. https://doi.org/10.1007/s11083-005-9021-x. (Note that 213-avoiding and 312-avoiding permutations are in bijection via inversion, a poset automorphism of the Bruhat order.)

Regarding Q2: I doubt this coincidence will extend in a straightforward way to other types. One reason is because in other types the number of order ideals of the root poset (the "W-Catalan number") is not the same as the number of fully commutative elements of the Weyl group. In Type A, fully commutative elements are 321-avoiding: not exactly the same as 213-avoiding, but equinumerous with them.

EDIT: By the way, even to define what pattern avoidance means in other types is a nontrivial task, but there is such a notion, due to Billey and Postnikov https://doi.org/10.1016/j.aam.2004.08.003.

  • 2
    $\begingroup$ Another related paper towards Q1 is Noncrossing partitions and Bruhat order by Thomas Gobet, Nathan Williams. $\endgroup$ Jan 25 at 14:49
  • 4
    $\begingroup$ At some point, I (and probably also others) have tried to identify subsets (such as noncrossing partitions or sortable elements) in other Weyl groups W counted by the W-Catalan numbers that yield, under (strong) Bruhat order, the distributive lattice of the root poset. I do not remember any details, but certainly did not succeed in finding such a subset of group elements. $\endgroup$ Jan 25 at 14:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.