5
$\begingroup$

Let $K/\mathbb Q_p$ be a local field, $X/K$ a proper scheme with semi-stable reduction.

Question: What is the possible range of Hodge-Tate weights of the etale cohomology $H^i(X_{\overline K}, \mathbb Q_p)$ as a $G_K$ representation?

The only "understanding" that I have is that when $i=1$, things reduce to abelian schemes. And the Hodge-Tate weights are all in {0,1}. Is that right?

What will happen with $i \geq 2$? Will the range be between 0 and $i$? It is true for abelian schemes right? Since $H^i$ in this case is the $i$-th wedge product of $H^1$.

$\endgroup$

2 Answers 2

10
$\begingroup$

The Hodge-Tate weights are just the numbers where the filtration in the de Rham cohomology jumps, so as you suspected will be in the range $[0,i]$. To prove this I guess one needs to observe, amongst other things, that the associated graded ring of $B_{dR}$ is $B_{HT}$, and that there's a comparison theorem relating de Rham and etale cohomology, with the field of periods being $B_{dR}$. It's all downhill from there.

$\endgroup$
5
$\begingroup$

If $X$ is an abelian variety, by a result of Tate there is an isomorphism:

$(\mathbb{C}_p(j) \otimes H^{i}(X_{\overline{K}}, \overline{\mathbb{Q}}_p)^{G_K}=H^{i-j}(X,\Omega^{j}_{X})$ when $j\leq i$ or $j\leq 0$ and zero in the other case.

Generally, Falting shows the following isomorphism for a proper smooth scheme $X$. $(\mathbb{C}_p \otimes H^{i}(X_{\overline{K}}, \overline{\mathbb{Q}}_p)=\oplus_{j \in \mathbb{Z}} \mathbb{C}_p(-j) \otimes H^{i-j}(X,\Omega^{j}_{X})$.

$\endgroup$

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.