Alcoves

Alcoves

To study how the groups and permute the hyperplanes in , we examine how they permute the collection of connected components of . Each element of is called an alcove. Fix a set of simple roots in . We assume moreover that is irreducible. It is convenient to single out one particular alcove and define it to be the fundamental alcove: Equivalently for irreducible, we have the following definition: with the highest root.

It is clear (since elements of act as homeomorphisms) that does permute .

What do alcoves look like? Notice first that is open in . Given , for each root there is some such that lies between and , so we can find an open neighborhood of meeting no -hyperplane. Intersecting these neighborhoods for all roots yields an open neighborhood of in . Since is open, its connected components are also open.

Any element outside is separated from it by one of the hyperplanes or ; so is a connected component of . In general, an alcove is defined by a set of inequalities (some of which may be redundant) of the form: , .

It is obvious that is included in this set. On the other hand, if satisfies the indicated inequalities, note that also for all positive . Since is a sum of simple roots, , so and thus .

This description shows that is simply an intersection of open half-spaces. Moreover, it is a euclidean simplex (whereas if has a number of irreducible components, will be a product of simplexes). (Question: What are the vertices of ?)

Walls of Alcoves

Define the walls of to be the hyperplanes , and , and define to be the corresponding set of reflections: The walls of can then be defined to be the images of these hyperplanes under for any . As soon as we show that acts transitively on , we will have well-defined walls for each alcove.

Proposition

The group permutes the collection of all alcoves transitively, and is generated by the set of reflections with respect to the walls of the alcove .