• Boundaries
• Section3

3 Boundaries of Coxeter groups

Let

dvipng error! exitcode was 2 (signal 0), transscript follows:

be a finitely-generated Coxeter group with Coxeter presentation

dvipng error! exitcode was 2 (signal 0), transscript follows:

. This presentation determines a Davis-Vinberg complex

dvipng error! exitcode was 2 (signal 0), transscript follows:

(see MR2360474), whose dimension equals rank of the maximal finite special subgroup of

dvipng error! exitcode was 2 (signal 0), transscript follows:

with respect to the above presentation. The complex

dvipng error! exitcode was 2 (signal 0), transscript follows:

admits a natural piecewise-Euclidean

dvipng error! exitcode was 2 (signal 0), transscript follows:

metric. The group

dvipng error! exitcode was 2 (signal 0), transscript follows:

acts on

dvipng error! exitcode was 2 (signal 0), transscript follows:

properly discontinuously and cocompactly. Hence,

dvipng error! exitcode was 2 (signal 0), transscript follows:

has visual boundary

dvipng error! exitcode was 2 (signal 0), transscript follows:

, which we can regard as a boundary of

dvipng error! exitcode was 2 (signal 0), transscript follows:

. Topology of

dvipng error! exitcode was 2 (signal 0), transscript follows:

was studied in MR1684267, MR1804695. For instance, MR1684267 constructs examples of hyperbolic Coxeter groups whose boundaries are both orientable and non-orientable Pontryagin surfaces and 2-dimensional Menger compacta. Recall that a Pontryagin surface is obtained as follows. Let

dvipng error! exitcode was 2 (signal 0), transscript follows:

be a connected, compact (without boundary) triangulated surface. Define

dvipng error! exitcode was 2 (signal 0), transscript follows:

by replacing each closed 2-simplex

dvipng error! exitcode was 2 (signal 0), transscript follows:

in

dvipng error! exitcode was 2 (signal 0), transscript follows:

with a copy

dvipng error! exitcode was 2 (signal 0), transscript follows:

of the closure of

dvipng error! exitcode was 2 (signal 0), transscript follows:

. We get the map

dvipng error! exitcode was 2 (signal 0), transscript follows:

by sending each

dvipng error! exitcode was 2 (signal 0), transscript follows:

to

dvipng error! exitcode was 2 (signal 0), transscript follows:

. Set

dvipng error! exitcode was 2 (signal 0), transscript follows:

. Then the corresponding Pontryagin surface

dvipng error! exitcode was 2 (signal 0), transscript follows:

based on

dvipng error! exitcode was 2 (signal 0), transscript follows:

is inverse limit of the sequence

dvipng error! exitcode was 2 (signal 0), transscript follows:

It turns out that

dvipng error! exitcode was 2 (signal 0), transscript follows:

can have only three distinct topological types:

1. If

   dvipng error! exitcode was 2 (signal 0), transscript follows:
   
   

   , then

   dvipng error! exitcode was 2 (signal 0), transscript follows:
   
   

   .

2. If

   dvipng error! exitcode was 2 (signal 0), transscript follows:
   
   

   is oriented but has genus

   dvipng error! exitcode was 2 (signal 0), transscript follows:
   
   

   , then

   dvipng error! exitcode was 2 (signal 0), transscript follows:
   
   

   is oriented (i.e.

   dvipng error! exitcode was 2 (signal 0), transscript follows:
   
   

   ) but not homeomorphic to

   dvipng error! exitcode was 2 (signal 0), transscript follows:
   
   

   .

3. If

   dvipng error! exitcode was 2 (signal 0), transscript follows:
   
   

   is not oriented then

   dvipng error! exitcode was 2 (signal 0), transscript follows:
   
   

   is unoriented. In this case, the rational homological dimension of

   dvipng error! exitcode was 2 (signal 0), transscript follows:
   
   

   equals 1.

Problem 3.1 (Comments) [Alexander Dranishnikov] Is it true that isomorphic Coxeter groups have homeomorphic boundaries?

Remark It appears that the answer is positive provided that all labels are powers of

dvipng error! exitcode was 2 (signal 0), transscript follows:

. REFERENCE?

Problem 3.2 (Comments) [Alexander Dranishnikov] Does there exist a Coxeter group

dvipng error! exitcode was 2 (signal 0), transscript follows:

with

dvipng error! exitcode was 2 (signal 0), transscript follows:

-dimensional boundary

dvipng error! exitcode was 2 (signal 0), transscript follows:

, so that the rational homological dimension of

dvipng error! exitcode was 2 (signal 0), transscript follows:

equals

dvipng error! exitcode was 2 (signal 0), transscript follows:

?

Problem 3.3 (Comments) [Alexander Dranishnikov] Under which conditions on the Coxeter diagram of

dvipng error! exitcode was 2 (signal 0), transscript follows:

, the boundary of a Coxeter group is

dvipng error! exitcode was 2 (signal 0), transscript follows:

-connected and locally

dvipng error! exitcode was 2 (signal 0), transscript follows:

-connected?

Partial results in this direction are obtained in MR1422863. The main motivation for this problem comes from the problem of realizing Menger spaces as boundaries of Coxeter groups.

Problem 3.4 (Comments) [Misha Kapovich] Can exotic homology manifolds as in MR1394965 appear as ideal boundaries of Coxeter groups?

Next Section