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

For a holomorphic vector bundle $E$ over a complex manifold $M$, we denote its space of smooth sections by $\Gamma^{\infty}(E)$, and its space of holomorphic sections by $\Gamma^{hol}(E)$. Now I've been looking at the line bundles $L_k$ over the complex projective spaces ${\bf C} P^N$, and I have managed to show that $\Gamma^{\infty}(L_k)$ is generated as a $C^{\infty}({\bf C} P^N)$-module by $\Gamma^{hol}(E)$, which is to say that every element $\Gamma^{\infty}(E)$ is a sum of elements of the form $ef$, where $e \in \Gamma^{hol}(E)$, and $f \in C^{\infty}({\bf C} P^N)$.

I am guessing that this result is extremely well known, and an example of a well understood general phenomenon. So I would like to ask if there is a characterization of the manifolds for which this result holds, for both the case of line bundles alone, and holomorphic vector bundles of general dimension?

share|cite|improve this question
up vote 5 down vote accepted

Swan has proved that taking global section gives an anti-equivalence between finitely generate projective $\Gamma^{\infty}(M)$-modules and $C^{\infty}$ vector bundles on $M$; this correspondence is functorial in $M$. Hence a set of section of a $C^{\infty}$ vector bundle $E$ on $M$ generated $\Gamma^{\infty}(E)$ if and only if it generates each fiber of $E$. So if $E$ is holomorphic, the holomorphic sections generate if and only if $E$ is globally generated. In particular, this is always true if $M$ is Stein. In the case of $L_k$ on $\mathbb{CP}^N$, this is true if and only if $k ≥ 0$.

share|cite|improve this answer
Excuse my ignorance, but what is a globally generated vector bundle? – Jean Delinez Feb 10 '13 at 15:56
It is a vector bundles whose global sections generate each fiber. – Angelo Feb 10 '13 at 19:47
Sorry, but I would like to get this totally clear for my mind. What you are saying is that there exists for $v \in E_p$, for any $p \in M$, an element $s \in \Gamma^{\infty}(E)$ of the smooth sections (or global sections as you call them), such that $v = s(p)$. Is this correct? – Jean Delinez Feb 11 '13 at 16:06
.... but I can't see why every vector is not globally generated. Surely, for $v \in E_p$ as above, we have a local section for which $s(p) = v$, why can one not just use a partition of unity argument to extend $s$ to a global section, and hence conclude that your vector bundle is globally generated? I can't see the problem in my logic here. – Jean Delinez Feb 11 '13 at 16:13
A holomorphic vector bundle is globally generated when for every point $p$ of $M$ and every $v \in E_p$ there exists a holomorphic global section $s$ with $s(p) = v$. There are lots of holomorphic vector bundles which have no non-zero sections, for example, the line bundles on $\mathbb{CP}^N$ with negative degree, so they can't be globally generated. – Angelo Feb 11 '13 at 17:08

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.