Topics in Representation Theory: SU (n), Weyl Chambers and the Diagram of a Group

Another Example: G = SU (n)

Last time we began analyzing how the maximal torus T of G acts on the adjoint representation, dening the roots of G as the non-trivial irreducibles that occur in this representation. The analysis proceeds by complexifying the Lie algebra, after which the adjoint actions of all H tC are diagonalizable, with eigenvalues the complex roots and eigenvectors the root spaces. In general we have the commutation relations [H, Ei ] = i (H )Ei Here H tC , the complex root is a complex linear map i : tC C and Ei gC is one of the root spaces. In the case G = SU (2), gC = sl(2, C), we can take elements of the Cartan subalgebra tC to be 2 by 2 diagonal matrices of the form H = 3 = 0 0

and the two root spaces consist of matrices with entries proportional to X+ = with root (H ) = 2 and X = 1 (1 i2 ) = 2 0 1 0 0 1 (1 + i2 ) = 2 0 0 1 0

with root . The theory of roots and root spaces is one part of representation theory where the SU (2) example is too trivial to be useful since here the rank is one, and there is only one root (and its conjugate). We need to begin looking at higher rank examples, and will begin with the case of G = SU (n), gC = sl(n, C). Note that while some formulas are simpler for the case of G = U (n), gC = gl(n, C), this is not a semi-simple group and one needs to analyze its normal U (1) subgroups

separately. For this case T can be chosen to be the diagonal matrices D with entries Djj = eij , eij = 1
j =1,n

and the corresponding Cartan subalgebra consists of diagonal matrices H with entries
n

(H )jj = j ,
j =1

j = 0

Here the subscript now refers to the entire set of n numbers i . The root spaces are labelled by pairs j, k, j = k of integers from 1 to n and correspond to matrices proportional to the matrix Ejk all of whose entries are 0 except for the j, k th which is 1. Since [H , Ejk ] = (j k )Ejk the complex roots are given by jk (H ) = j k The Weyl group in this case is the group S n , acting on T by permuting the Djj . Note that roots come in pairs of opposite sign. Making a choice of one of each pair is a choice of which roots are positive. One could for example choose in this example positive roots to be the jk , j < k but note that this choice is not invariant under the Weyl group. The Weyl group acts on the set of roots, but does not respect the decomposition into positive and negative roots. A specic example to continually keep in mind, one where we can easily draw things, is the case of G = SU (3). Here the rank is two, and G/T is a complex manifold of six real dimensions or three complex dimensions. There are a total of six roots 12 (H ) = 1 2 23 (H ) = 2 3 13 (H ) = 1 3 and their negatives. The three positive roots are not independent, they satisfy the relation 13 = 12 + 23 We will draw these as vectors in a plane that can be identied with it tC Welll see a little later that the Killing form gives a natural inner product on this space, but for now well avoid using this, just dealing with the parts of the theory of roots that dont actually need the inner product. 2

Weyl Chambers and the Diagram of a Group

Recall that we dened the Weyl group W (G, T ) to be N (T )/T , this is a group of automorphisms of T that come from conjugation by an element of G. We saw in our example that the Weyl group S n of SU (n) permutes its roots. In general the Weyl group acts on the set of roots since a root is a homomorphism : T GL(g) and an element w W (G, T ) corresponds to a homomorphism w : t T xtx1 T for some x G. Composing these two maps gives a new root w(). Elements t T for which the group N (t) = {g G : gtg 1 = t} has dimension greater than dim T are said to be singular, those with dimension equal to dim T are said to be regular. In our example of G = SU (n), the regular elements are those where all the diagonal elements in T are distinct, the singular ones are those where two or more diagonal elements are identical. To understand the structure of the set of singular points in T , for each root dene Denition 1. For each root , there is a codimension one subgroup of T U = ker( ) = ker( ) The point t U is a singular points of T , since the root space g generates a one-parameter subgroup of G, not in T , of elements that commute with t, so the the dimension of N (T ) is greater than the dimension of T . Note that the intersection of all the U consists of elements that commute with all of G, the center Z (G) of G. The Stiefel diagram of a group is the set consisting of all exp1 U and will consist of innite sets of parallel hyperplanes in t. If we pass to the Lie algebra (now will be thought of as elements of t ) and just look at those that go through zero we can dene Denition 2 (Diagram of G). The innitesimal diagram or diagram of G consists of the hyperplanes Lie(U ) = ker t i.e. those elements v t such that (v ) = 0.

12

23
13

13

23

12

12

13

23

Diagram for SU (3)

The hyperplanes of the diagram divide t into nitely many convex regions Denition 3 (Weyl Chambers). Given a choice of positive roots, a Weyl chamber is a set of the form {v t :
i i (v )

> 0}
i

for all positive roots i and some choice of signs

= 1.

C2 U

C1 U
23

13

C3

C6

C4

C5

Weyl Chambers for SU (3)

The roots are not independent, we can express some in terms of others. A set of roots that form a basis are called simple roots Denition 4 (Simple Roots). A subset S of the set R of roots is called a system of simple roots if the roots in S are linearly independent and every root R can be written as = m
S

with m integers all 0 or 0. In the case of SU (3), a set of simple roots will have two elements. One possible choice of the set S of simple roots is S = {12 , 23 } A choice of a set S of simple roots is equivalent to a choice of positive roots. Positive roots are those where all m are positive, negative roots ones where all m are negative. Corresponding to each choice of S there is a distinguished Weyl chamber Denition 5. The fundamental Weyl chamber corresponding to a set S of simple roots is K (S ) = {v t : i (v ) > 0} for all i in S .
U
12

Fundamental Weyl Chamber

23

23

12

Simple Roots and Fundamental Weyl Chamber for G = SU (3) We will not go into much more detail about the general theory of the Weyl group and how it acts on the diagram of G, referring to extensive discussions in the texts [1],[2],[3]. but there are two basic facts that will be of importance. 1. The Weyl group acts simply transitively on the set possible choices of S . 2. The Weyl group is generated by reections in the hyperplanes corresponding to simple roots in S .

References
[1] Adams, J. Frank, Lectures on Lie Groups, University of Chicago Press, 1969. [2] Br ocker, T. and tom Dieck, T.,Representations of Compact Lie Groups, Springer-Verlag, 1985. [3] Simon, B., Representations of Finite and Compact Groups,American Mathematical Society, 1996.