9
$\begingroup$

Whilst reading Hartshorne's appendix C I came across the comparison theorem for etale cohomology and singular cohomology:

Let $X$ be a smooth projective variety over a number field $K$ and $\ell$ a prime number. Fix an embedding of $K$ into $\mathbb C$. Then there is a "natural" isomorphism of $\mathbb C$-vector space $$\mathrm H^n(X_{\bar K,et},\mathbb Q_\ell)\otimes \mathbb C \cong \mathrm H^n(X(\mathbb C),\mathbb C).$$

Do we have the slightly stronger "natural" isomorphism of $\mathbb Q_\ell$-vector spaces

$$\mathrm H^n(X_{\bar K,et},\mathbb Q_\ell) \cong \mathrm H^n(X(\mathbb C),\mathbb Q)\otimes \mathbb Q_\ell?$$

I would appreciate references to the literature.

$\endgroup$

1 Answer 1

9
$\begingroup$

Yes. In fact what Artin proves in SGA4 exp XI thm 4.4 is that étale cohomology and singular cohomology agree for smooth schemes over $\mathbb{C}$ with finite coefficients. The statement you want will follow from this by taking inverse limits to get to $\mathbb{Z}_\ell$ and then extending scalars to $\mathbb{Q}_\ell$. If you don't feel like looking at SGA, you can find treatments of this in the books by Freitag-Kiehl, Milne,…

Added (in response to comment). The isomorphism $H_{et}^*(X_{\bar K}, \mathbb{Q}_\ell)\cong H_{et}^*(X_{\mathbb{C}}, \mathbb{Q}_\ell)$ follows from the smooth base change theorem (cf. Milne, Etale cohomology, p 231 cor 4.3).

$\endgroup$
1
  • $\begingroup$ This only gives $\mathrm{H}^n(X_{\mathbb C,et},\mathbb Q_\ell) \cong \mathrm{H}^n(X(\mathbb C), \mathbb Q)\otimes \mathbb Q_\ell$. What I need now is to know whether $H^n(X_{\mathbb C,et},\mathbb Q_\ell) = H^n(X_{\bar K,et},\mathbb Q_\ell)$. I guess this follows from the equivalence of categories "etale covers of $X_{\bar K}$" and "etale covers of $X_{\mathbb C}$". Correct? $\endgroup$
    – user12
    Commented Aug 2, 2014 at 17:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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