Let $G$ be an even-dimensional compact Lie group with Lie algebra $\mathfrak{g}$ and let $T \subset G$ be a maximal torus with Lie algebra $\mathfrak{t}$.

We can construct a left-invariant complex structure over $G$ in the following way. Let $\Sigma$ be the set of roots associated to the Cartan algebra $\mathbb C \mathfrak{t}$ (the complexification of $\mathfrak{t}$) and $\mathfrak g_\alpha$ the eigenspaces associated to the root $\alpha \in \Sigma$. We denote by $\Sigma_+$ a choice of positive roots of $\Sigma$.

Now, let $\mathfrak{m} \subset \mathbb C \mathfrak{t}$ be any subalgebra such that $\mathfrak{m} \oplus \overline{\mathfrak{m}} = \mathbb C \mathfrak{t}$, we then define $\mathfrak h = \mathfrak{m} \oplus \bigoplus_{\alpha \in \Sigma_+} \mathfrak g_\alpha$. The Lie algebra $\mathfrak h$ defines a complex structure on $G$ such that every left-translation of $G$ is holomorphic. And since $T$ is abelian, the structure $\mathfrak m$ is a bi-invariant complex structure of $T$ and thus $T$ is a complex Lie group.

In [1], it is said that $T \to G \to G / T$ is ``obviously'' a holomorphic principal bundle. I tried to verify this claim, I can prove that $G$ can be regarded as a smooth principal bundle with structure $T$, but I could not prove that this bundles is holomorphically locally trivializable.

How can I prove that such bundle is holomorphically locally trivializable? Are there some references I could follow?

I do not know if you could call this bundle a holomorphic principal bundle without this property, the the authors of [1] need to apply the Borel spectral sequence, and in the construction of such spectral sequence, it is necessary to use a local trivialization compatible with the complex structure of the fiber and of the base space.

[1]: Vanishing theorems on Hermitian manifolds - Bogdan Alexandrov and Stefan Ivanov (https://www.sciencedirect.com/science/article/pii/S0926224501000444)


1 Answer 1


Proposition 4.26 in Félix, Oprea, Tanré, Algebraic Models in Geometry applied to $T \to G \to G/T$ shows what you want since $G/T$ is Kähler manifold. (See also Example 4.32.)

  • $\begingroup$ Thank you very much for your answer. Unfortunately, the mentioned proposition does not answer the question. The proposition just explain how to construct a complex structure on G, which the authors claim that could be chosen to be either left or right-invariant. Still, it is not clear why this bundle would by holomorphically locally trivializable. Also, it is not clear that (in the case of G having a left-invariant structure) why the right action by T would be holomorphic (also necessary for it to be a holomorphic bundle). $\endgroup$ Mar 4, 2020 at 16:34

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.