4
$\begingroup$

Let $G$ be a connected complex reductive group with a maximal compact subgroup $K$. Let $\lambda$ be a dominant weight in the interior of the positive Weyl chamber. Let $V_\lambda$ denote the irreducible representation with highest weight $\lambda$ and fix a highest weight vector $v_\lambda$. Then $gB \mapsto [g. v_\lambda]$ gives a $G$-equivariant embedding of the flag variety $G/B$ into the projective space $\mathbb{P}(V_\lambda)$.

Fix a $K$-invariant Hermitian product on $V_\lambda$. This defines a $K$-invariant Fubini-Study form on $\mathbb{P}(V_\lambda)$ and hence on the image of $G/B$ in there.

My question is whether this symplectic form on $G/B$ coincides (up to a scalar) with the Kostant-Kirillov form on the coadjoint orbit of $\lambda^*$ (dual weight)? (via the moment map $\mu: \mathbb{P}(V_\lambda) \to Lie(K)^*$) restricted to the image of $G/B$.)

$\endgroup$

1 Answer 1

2
$\begingroup$

The answer is yes, due to formal properties of moment maps (that depend on almost no details of your situation). Namely, suppose $K$ acts transitively on any symplectic manifold $(X,\omega_X)$, with equivariant moment map $\mu:X\to\mathfrak k^*$. Then $\mu(X)$ is a coadjoint orbit $\mathcal O$ and $\omega_X=\mu^*\omega_{\mathcal O}$.

Proof: Write $Z(x)$ for the infinitesimal action of $Z\in\mathfrak k$ on $x\in X$. The moment map definition $i_Z\omega_X=-d\langle\mu(\cdot),Z\rangle$ says $$ \omega_X(Z(x),\delta x)=-\langle D\mu(x)(\delta x),Z\rangle $$ for all tangent vectors $\delta x\in T_xX$. By transitivity, any such writes $Z'(x)$ for some $Z'\in\mathfrak k$, and equivariance then gives $D\mu(x)(Z'(x))=Z'(\mu(x))=\langle\mu(x),[\cdot,Z']\rangle$. So the above equality becomes $$ \begin{align} \omega_X(Z(x),Z'(x))&=\langle\mu(x),[Z',Z]\rangle\\ &=\omega_{\mathcal O}(Z(\mu(x)), Z'(\mu(x))), \end{align} $$ where the last equality is the definition of the Kirillov-Kostant-Souriau 2-form $\omega_{\mathcal O}$.

$\endgroup$
0

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.