18
$\begingroup$

Let $X$ be a variety over a $p$-adic field $K$.

Is there a simple or intuitive explanation of why the $G_K$ representation $H^i(X_{ét},\mathbb{Q}_p)$ is Hodge-Tate? More precisely, why do the powers of the cyclotomic character appear and not other characters?

$\endgroup$
  • $\begingroup$ For clarity, does "variety" assume smooth and proper as well? $\endgroup$ – Matt Sep 8 '13 at 20:21
  • 5
    $\begingroup$ Just to be precise, what is true is that, after tensoring over $\mathbb{C}_p$, the subspaces on which the (semi-linear!) Galois action is via powers of the cyclotomic character generate the whole space. As for whether there is an 'intuitive' explanation for this phenomenon, I don't know! $\endgroup$ – Keerthi Madapusi Pera Sep 9 '13 at 3:07
  • 5
    $\begingroup$ As Keerthi pointed out, it's after you tensor by Cp that the powers of the cyclo char appear. But a statement like that is true for any Galois repn V by Sen's theory. What is specific to the étale cohomology is that it's integer powers of the cyclo char that appear. This is where you use the geometric input. As to why this is so... $\endgroup$ – Laurent Berger Sep 10 '13 at 8:02
2
$\begingroup$

You need to assume that your variety is proper and smooth.

The properness implies that the vector space $H^{1}_{et}(X_{\overline{K}}, \mathbb{Q}_p)$ is of finite dimension. And the smoothness implies the following decomposition which is proved by Falting: $(\mathbb{C}_p \otimes H^{i}(X_{\overline{K}}, \mathbb{Q}_p)=\oplus_{j \in \mathbb{Z}} \mathbb{C}_p(-j) \otimes H^{i-j}(X,\Omega^{j}_{X})$

$\endgroup$
  • 1
    $\begingroup$ In fact neither smoothness or properness is necessary for the cohomology to be Hodge-Tate, though the isomorphism of your last line does require both smoothness and properness. $\endgroup$ – Daniel Litt Mar 28 '18 at 18:38

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.