Let $G$ be a semi-simple, compact Lie Group. Consider its complexification $G_{\mathbb{C}}$. Does there exist a Kähler structure on $G_{\mathbb{C}}$ which is $G$-invariant (maybe in a neighbourhood of $G$ in $G_{\mathbb{C}}$)?


  • $\begingroup$ If $G$ be compact then $T^*G\cong G\times \mathfrak g^*\cong G^{\mathbb C}$, since $T^*G$ is K\"ahler, then it induces a Kähler structure on $G^{\mathbb C}$ which is $G$-invariant $\endgroup$ – user21574 Jan 19 '17 at 5:18

Yes, such a Kähler form always exists: Embed $G$ as a matrix group in $\mathrm{SU}(n)$ for some $n$ and then let $G_\mathbb{C}\subset \mathrm{SL}(n,\mathbb{C})\subset M_n(\mathbb{C})$ be the complexification. Choose a Kähler form on this latter vector space, pull it back to $G_\mathbb{C}$ and then, using the compactness of $G$, average its pullbacks under left and right multiplications. This will yield a Kähler form on $G_\mathbb{C}$ that is $G$-invariant.

  • $\begingroup$ Is there any reference where I can see how the proof is done? $\endgroup$ – hapchiu Dec 2 '12 at 8:27
  • $\begingroup$ what do you mean by average its pullbacks under left and right multiplications if $G$ is compact??? $\endgroup$ – hapchiu Dec 2 '12 at 8:59
  • $\begingroup$ @hapchiu: It's a standard technique. If you have a compact Lie group $H$ acting on the left as biholomorphic transformations on a Kähler manifold $M$, then, letting $dh$ be Haar measure on $H$ and $\Omega$ be a Kähler form on $M$, one can construct an $H$-invariant Kähler form on $M$ by averaging: $$\bar\Omega = \int_H L_h^*(\Omega)\ dh\ ,$$ where $L_h:M\to M$ is left action by $h\in H$. In your case, $M=G_\mathbb{C}$ and $H = G\times G$, with $L_{(g_1,g_2)}(g) = g_1g{g_2}^{-1}$. $\endgroup$ – Robert Bryant Dec 2 '12 at 13:21
  • $\begingroup$ You can see arxiv.org/abs/1307.0454 $\endgroup$ – user21574 Nov 5 '13 at 23:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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