2
$\begingroup$

What is the standard resolution of singularities, for the nilpotent cone (of the adjoint representation) for the symplectic group? I know how to do this for the general linear group, but am having trouble finding a good reference for the symplectic group. I understand it uses Richardson orbits. (I do know what the closure ordering should be though).

$\endgroup$

2 Answers 2

2
$\begingroup$

Ben gave the general answer above. If you care specifically about the symplectic group and are interested in a "flag-like" description of its flag variety, then one exists. It is given by all half-flags of isotropic subspaces (this is just like for $SL_n$, the symplectic group acts transitively and the stabilizer of the standard half-flag will be the standard Borel). With this description, it's just as straightforward computing Springer fibers and the like as it is for the $SL_n$ case, which you're presumably familiar with.

A reference for these flag-like descriptions can be found in the section of Fulton and Harris on "Homogeneous Spaces" (there's a similar description for the special orthogonal groups).

$\endgroup$
3
$\begingroup$

It's the Springer resolution; this works for all semi-simple groups. The resolution is the moment map from the cotangent bundle of $G/B$ to $\mathfrak{g}^*$. Look at section 6 of Ginzburg's notes.

Not to steal Mike's thunder, but there's an even more specific description of this resolution: given a nilpotent element $X$ of the symplectic Lie algebra, the fiber in this resolution over it is the space of complete flags $V_1\subset V_2\subset \cdots \subset \mathbb{C}^{2n}$ such that $V_i$ and $V_{2n-i}$ are symplectic orthogonal (this immediately implies that all the spaces in this flag are isotropic or coisotropic) which are preserved by $X$ ($XV_i\subset V_{i-1}$).

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.