Small roots, low elements, and the weak order in Coxeter groups

TL;DR

This paper introduces low elements in Coxeter groups, showing they form a Garside shadow, which helps prove all finitely generated Artin-Tits groups have a finite Garside family, with results on bipodality of small roots.

Contribution

It defines low elements from small roots in Coxeter groups and proves they form a Garside shadow, advancing understanding of Artin-Tits groups and root system structures.

Findings

Low elements form a Garside shadow in Coxeter groups.

The set of small roots is bipodal.

Proved bipodality for affine Coxeter groups and certain labeled groups.

Abstract

In this article we provide a new finite class of elements in any Coxeter system (W,S) called low elements. They are defined from Brink and Howlett's small roots, which are strongly linked to the automatic structure of (W,S). Our first main result is to show that they form a Garside shadow in (W,S), i.e., they contain S and are closed under join (for the right weak order) and by taking suffixes. These low elements are the key to prove that all finitely generated Artin-Tits groups have a finite Garside family. This result was announced in a note with P. Dehornoy (P. Dehornoy, M. Dyer, and C. Hohlweg. Garside families in Artin-Tits monoids and low elements in Coxeter groups. Comptes Rendus Mathematique, 353:403-408., 2015.) in which the present article was referred to under the following working title: Monotonicity of dominance-depth on root systems and applications. The proof is based on a fundamental property enjoyed by small roots and which is our second main result; the set of small root is bipodal. For a natural number n, we define similarly n-low elements from n-small roots and conjecture that the set of n-small roots is bipodal, implying the set of n-low elements is a Garside shadow; we prove this conjecture for affine Coxeter groups and Coxeter groups whose graph is labelled by 3 and infinity. To prove the latter, we extend the root poset on positive roots to a weak order on the root system and define a Bruhat order on the root system, and study the paths in those orders in order to establish a criterion to prove bipodality involving only finite dihedral reflection subgroups.