2
$\begingroup$

Suppose X,Y are two complex manifolds and E,F are vector bundles over them respective, what can I say about their global sections? Does this formula $\Gamma (X\times Y,p_1^{*}E\otimes p_2^{*}F)=\Gamma (X,E)\otimes\Gamma (Y,F)$ hold?

If it holds,then we will get every holomorphic function of two complex variables will be the form $f_1(z_1)g_1(z_2)+\cdots+f_n(z_1)g_n(z_2)$. It seems plausibile.

$\endgroup$
2
  • $\begingroup$ Please make the title of your question an english sentence, and also check out mathoverflow.net/howtoask $\endgroup$ – David Roberts Dec 15 '11 at 6:37
  • 1
    $\begingroup$ This is completely false when neither of them is compact. In the case when one of them is compact, you can use the projection formula. $\endgroup$ – Angelo Dec 16 '11 at 9:11
1
$\begingroup$

Perhaps the easiest counterexample comes from letting $X$ and $Y$ be countable infinite disjoint unions of points, and setting the vector bundles to be one-dimensional. Then you just do a dimension count.

A more sophisticated counterexample is the function $e^{xy}$ on $\mathbb{C}^2$. It is globally holomorphic, but it is not a finite sum of products of single-variable entire functions. You can check this by considering the restriction to $\mathbb{N} \times \mathbb{N}$, and writing the values as entries in a big matrix. The matrix is a union of nested van der Monde matrices of nonzero determinant, so it has infinite rank. Therefore, it can't be the sum of finitely many rank 1 matrices of the form $f_i^T g_i$

When one of the manifolds is compact, the isomorphism is a corollary of cohomology and base change. You can find a proof in pretty much any book on several complex variables. See, e.g., Grauert, Remmert Coherent analytic sheaves.

$\endgroup$
0

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.