# Springer corresponding for nullcones other than the standard nilpotent cone

I understand the ordinary Springer correspondence gives a bijection between orbits in the nilpotent cone for the adjoint representation and irreducible representations of the Weyl group, through action of the Weyl group on the top intersection cohomology. (I'm still learning about this, I know a very vague picture). My question is: does the same method give a correspondence (not injective nor surjective) between orbits in the nullcone of any representation (not the adjoint representation), and representations (not necessarily irreducible) of the Weyl group? Or is the construction so specific that it breaks down entirely when we replace the nilpotent cone by more generally, a nullcone, and we cannot get any Weyl group representations out of it?

-

Let me make another guess as to what you are seeing:

The Lie/algebraic group acts on the coordinate ring for the closure of the nilpotent orbit (since it acts on the nilpotent orbit). Take the weight 0 weight space of this coordinate ring; the Weyl group (realized as N(T)/T) acts on it since T acts trivially on this weight space.

In the case of SL_n, this gives the total cohomology ring of the Springer fiber associated to the conjugate partition (if I remember correctly, and I might not, this result is due to DeConcini and Procesi back in the early 80s). I seem to recall there is a way to recover the grading on the cohomology ring, which allows you to recover the top cohomology (and hence the original Springer correspondence), but cannot remember what it is.

This isomorphism between the weight 0 subring and the cohomology ring is specific to SL_n.

What I just wrote does make sense for any nullcone, though a priori there is no guarantee that you get a finite dimensional representation of the Weyl group.

-
The reference I failed to put in above is: Deconcini and Procesi, Symmetric functions, conjugacy classes, and the flag variety, Inventiones 64 (1981), 203--219. –  Alexander Woo Dec 3 '09 at 20:34

Certainly, there's no reason to believe there will be. The Springer correspondence depends very strongly on the existence and structure on the Springer resolution.

-
In the case of my nullcone, I have an analogue of the Springer resolution for it. What properties of the Springer resolution do you need to construct the Springer correspondence? –  Vinoth Dec 2 '09 at 3:21
What does this resolution look like? Is it an actual resolution of singularities? Is it semi-small? As for the Springer correspondence, it's not easy to say exactly what you need. The fact that components of the Steinberg variety are the conormals to the W-orbits is very important. Another way of explaining it is that the Springer resolution is related by Fourier transform to the Grothendieck simultaneous resolution, which is small and generically a Galois $W-$cover. –  Ben Webster Dec 2 '09 at 3:45
Yes it's a resolution of singularities, I'm now trying to work out if it's semi-small or not (but I am hoping it is). Some of what you said went over my head, but I would like to understand - what paper do you recommend I read? Currently I was thinking of going for Springer's original paper, "A construction of representations of Weyl groups" or something like that. –  Vinoth Dec 2 '09 at 9:35