arXiv reader

Rigidity of Right-Angled Coxeter Groups

David G. Radcliffe

arXiv
math/9901049v4 · PDF · 13 May 2026
Categories
math.GR
Licence
CC BY 4.0 The site shows the text.

Abstract

If S and S’ are two finite sets of Coxeter generators for a right-angled Coxeter group W, then the Coxeter systems (W,S) and (W,S’) are equivalent.

Items 8

Theorem

Theorem 1

X1JN

Let W be a right-angled Coxeter group. If S and S' are finite Coxeter generating sets for W, then the Coxeter systems (W,S) and (W,S') are equivalent.

Proof (Proof of Theorem 1) of Theorem 1

For each s\in S, define the membership pattern of s to be

\mathcal A(s)=\{\,A\in\mathcal N^{\max}(W,S): s\in A\,\}.

By inclusion–exclusion, the number of elements s\in S whose membership pattern is exactly \mathcal U is

\sum_{\mathcal V\subseteq \mathcal N^{\max}(W,S)\setminus \mathcal U} (-1)^{|\mathcal V|} \left| \bigcap_{A\in \mathcal U\cup \mathcal V} A \right|.

The corresponding number for S', with membership pattern \mathcal U^*=\{A^*:A\in\mathcal U\}, is

\sum_{\mathcal V\subseteq \mathcal N^{\max}(W,S)\setminus \mathcal U} (-1)^{|\mathcal V|} \left| \bigcap_{A\in \mathcal U\cup \mathcal V} A^* \right|.

These two sums are equal term by term by Theorem 4. Hence the two systems have the same number of generators of each membership pattern.

It follows that there is a bijection \phi\colon S\to S' such that, for every A\in\mathcal N^{\max}(W,S) and every s\in S,

s\in A \quad\Longleftrightarrow\quad \phi(s)\in A^*.

Now let s,t\in S with s\ne t. Since (W,S) is right-angled, m(s,t)=2 exactly when \{s,t\} is contained in some maximal simplex A\in\mathcal N^{\max}(W,S). By the defining property of \phi, this occurs exactly when \{\phi(s),\phi(t)\} is contained in the corresponding maximal simplex A^*\in\mathcal N^{\max}(W,S'), which is exactly the condition that m(\phi(s),\phi(t))=2. Otherwise both exponents are \infty.

Thus

m(s,t)=m(\phi(s),\phi(t))

for all s,t\in S. Therefore \phi respects the Coxeter matrices and extends to an automorphism of W carrying S onto S'. Hence (W,S) and (W,S') are equivalent.

Lemma

Lemma 2

M6XC

Every finite subgroup of W is contained in a finite parabolic subgroup. Consequently, the maximal finite subgroups of W are exactly the groups

wW_Aw^{-1},\qquad w\in W,

where A\in\mathcal N^{\max}(W,S).

Proof of Lemma 2

Let G be a finite subgroup of W. By the fixed-point theorem quoted above, G fixes a point of X. The action of W on \Sigma(W,S) is without inversions: if an element stabilizes a simplex, then it fixes each vertex of that simplex. Hence G fixes a vertex of \Sigma(W,S).

The vertices of \Sigma(W,S) are precisely the cosets wW_A with A\in\mathcal N(W,S)\cup\{\varnothing\}. If G fixes the vertex wW_A, then G is contained in its stabilizer, namely wW_Aw^{-1}. This is a finite parabolic subgroup. Maximal finite subgroups are therefore the maximal finite parabolic subgroups, which are exactly the conjugates wW_Aw^{-1} with A\in\mathcal N^{\max}(W,S).

Theorem

Theorem 3

X58K

For every A\in\mathcal N^{\max}(W,S), there exists a unique A^*\in\mathcal N^{\max}(W,S') such that

q(W_A)=q(W_{A^*}).

Moreover, the map A\mapsto A^* is a bijection from \mathcal N^{\max}(W,S) to \mathcal N^{\max}(W,S').

Proof of Theorem 3

Let A\in\mathcal N^{\max}(W,S). Then W_A is a maximal finite subgroup of W. Applying Lemma 2 to the Coxeter system (W,S'), there are w\in W and A^*\in\mathcal N^{\max}(W,S') such that

W_A=wW_{A^*}w^{-1}.

Since W_{\mathrm{ab}} is abelian, conjugate subgroups have the same image under q. Thus q(W_A)=q(W_{A^*}).

The set A^* is unique because q is injective on the union V_{S'} of the finite standard parabolic subgroups for (W,S'). The same argument with S and S' interchanged gives an inverse correspondence, so A\mapsto A^* is a bijection.

Theorem

Theorem 4

90NZ

If A_1,\ldots,A_r\in\mathcal N^{\max}(W,S), then

\left|\bigcap_{i=1}^r A_i\right|= \left|\bigcap_{i=1}^r A_i^*\right|.
Proof of Theorem 4

Using the standard parabolic intersection formula and the injectivity of q on V_S, we have

\left|W_{\cap_i A_i}\right| =\left|\bigcap_{i=1}^r W_{A_i}\right| =\left|\bigcap_{i=1}^r q(W_{A_i})\right|.

By Theorem 3, q(W_{A_i})=q(W_{A_i^*}) for every i, and therefore

\left|\bigcap_{i=1}^r q(W_{A_i})\right| =\left|\bigcap_{i=1}^r q(W_{A_i^*})\right| =\left|\bigcap_{i=1}^r W_{A_i^*}\right| =\left|W_{\cap_i A_i^*}\right|.

Since all relevant standard parabolic subgroups are elementary abelian 2-groups, |W_B|=2^{|B|} for every simplex B of either nerve. The result follows.

Source and licence

The text on this page is from arXiv:math/9901049v4 by David G. Radcliffe, under the licence CC BY 4.0. The corpus made these changes: it read the items out of the LaTeX source, it expanded the macros of the paper, it put each proof after its statement, and it converted the text to HTML. The formulas are the TeX of the paper, and KaTeX renders them. A formula or a reference can look different from the PDF. The page is published under CC BY 4.0.

The record of each item is in analysis/objects/9901/math-9901049.jsonl.