0
$\begingroup$

Let $X,Y,Z$ be compact complex manifolds or smooth complex projective varieties, do we have the following commutative diagram similar to flat/proper/smooth base change for quasi-coherent sheaves? $$\require{AMScd} \begin{CD} H^\bullet(X\times Y,\mathbb{Z}) @>u^*>> H^\bullet(X\times Y\times Z,\mathbb{Z}) \\ @Vq_*VV @VVp_*V\\ H^\bullet(X ,\mathbb{Z}) @>v^*>> H^\bullet(X\times Z,\mathbb{Z}) \end{CD}$$ Here the maps are pullback/pushforward through either canonical projection or closed embedding.

$\endgroup$
2
  • 1
    $\begingroup$ Does this wiki page along with references there answer your question? $\endgroup$
    – Z. M
    Commented May 21, 2022 at 19:41
  • $\begingroup$ This follows from base change in the category of sheaves (of vector spaces) over topological spaces. A good reference might be Iversen's sheaf theory book. $\endgroup$
    – Pulcinella
    Commented May 23, 2022 at 21:06

1 Answer 1

3
$\begingroup$

This claim is weaker than base change.

By duality between homology and cohomology, it suffices to consider the dual diagram in homology, and check this commutes.

The dual to integration (which I guess is the pushforward map here) is taking the inverse image of homology cycles, and the dual to pullback is taking the image of homology cycles.

Then the claim is that if we take a cycle on $X \times Z$, the inverse image in $X \times Y$ of its image in $X$ is equal to the image in $X\times Y$ of its inverse image in $X\times Y \times Z$. But this is obvious.

$\endgroup$
1
  • $\begingroup$ Maybe "integration" is a bit of a misnomer here; that would be the de Rham version (which doesn't make sense for integral coefficients). $\endgroup$ Commented May 21, 2022 at 23:55

You must log in to answer this question.