2
$\begingroup$

Let $H$ be a finite dimensional $\mathbb{Q}$-vector space. I've heard that to give an effectif Hodge structure of weight 1 on $H$ is the same as to give a complex structure on $H_\mathbb{R}$. Why is this true?

If I have a complex structure, that is an endomorphism $I \in \mathrm{End}(H_\mathbb{R})$ such that $I^2=-Id$, then I can define $H^{1, 0}$ (resp. $H^{0, 1}$) as the eigenspaces of the eigenvalues $i$ and $-i$.

How to prove the converse. Here is my idea:

consider the inclusion $j: H_\mathbb{R} \hookrightarrow H_\mathbb{C}=H^{1, 0} \oplus H^{0, 1}$. For any $u \in H_\mathbb{R}$, $j(u)$ decomposes as

$j(u)=v+w, \quad v \in H^{1, 0}, w\in H^{0, 1}$

Define $I(u)=i(v-w)$ for the corresponding $v$ and $w$. One has to verify that $I(u) \in H_\mathbb{R}$. But:

$\overline{I(u)}=-i(\bar{v}-\bar{w})=-i(w-v)=i(v-w)=I(u)$

where I use the Hodge symmetry to say that $\bar{v}=w$ and $\bar{w}=v$.

Is that correct? Is it the standard argument? Is there a more conceptual way of understanding things?

Thanks

$\endgroup$
5
  • 1
    $\begingroup$ I suppose you mean weight $1$ Hodge structure? $\endgroup$
    – dhagbert
    Apr 11, 2013 at 21:54
  • $\begingroup$ You can use the equivalence between Hodge structures and representations of the group scheme $\mathbf{S}$. In weight 1, this information can be encoded by the eigenvalues of your operator $I$, but in higher weights, you lose some information. $\endgroup$
    – S. Carnahan
    Apr 12, 2013 at 2:54
  • $\begingroup$ @dhagbert: Yes I mean weight 1, I edited my question. $\endgroup$
    – SH1
    Apr 12, 2013 at 7:25
  • $\begingroup$ @Carnahan: Could you develop a bit more your answer? Anyway, what you are saying should be the generalization to higher weight of the equivalence "Effective HS of weight 1 --- complex structures" $\endgroup$
    – SH1
    Apr 12, 2013 at 7:27
  • $\begingroup$ Help please!!!! $\endgroup$
    – SH1
    Apr 16, 2013 at 16:41

0

Your Answer

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

Browse other questions tagged or ask your own question.