MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $G$ be a complex connected reductive Lie group, $\theta$ an involution, and $K = G^\theta$ the fixed-point subgroup. Then $K$ has finitely many orbits on $G/B$, one of which is open and (quite often) several of which are closed.

We can pick a $\theta$-stable Borel $B$ and maximal torus $T$, and using Springer's map $$ \varphi : K\backslash G \to G, \qquad Kg \mapsto \theta(g^{-1}) g $$ we get a map of the same name $$ \varphi: K\backslash G/B \to B\backslash G/B \qquad \cong W := N(T)/T. $$ This map isn't usually injective. For example, if $G = SL(2)$ and $K = SO(2)$, then the LHS has three elements (North and South poles and the rest) while the RHS has only two.

If we fix a closed orbit $v \in K \backslash G/B$, and restrict $\varphi$ to the order ideal {$u \in K \backslash G/B : u \geq v $}, does it become injective?

Motivation: the poset $K \backslash G/B$ is somewhat mysterious, whereas the poset of "twisted involutions" {$w \in W : (\theta\cdot w)^2 = 1$} is at least less so. I need to sum over chains in this order ideal and would rather do it inside $W$.

share|cite|improve this question
up vote 2 down vote accepted

Malheureusement, this is not true, not even for the weak order. This can be seen for example when $G = GL(4)$ and $K = GL(2) \times GL(2)$. Then $K \backslash G / B$ is parameterized by involutions with signs attached to fixed points and the map $\varphi$ simply forgets the markings on the fixed points.

For example, the closed orbit associated to $(1^+)(2^-)(3^-)(4^+)$ lies below both $(14)(2^-)(3^+)$ and $(14)(2^+)(3^-)$. Figures with weak order for a plethora of examples of symmetric subgroups appear in a preprint by Ben Wyser.

share|cite|improve this answer
Helas! $\ \ \ \ $ – Allen Knutson Mar 22 '12 at 1:17

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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