6
$\begingroup$

Let $V_1$, $V_2$ be two polarised simple $Q$-Hodge structures which are non-isomorphic. I am assuming that the Mumford-Tate groups of $V_1$ and $V_2$ are semi-simple adjoint.

Is it true in this case that $MT(V_1 \times V_2) = MT(V_1) \times MT(V_2)$?

(I can easily see that this is not true when $MT(V_1)$ and $MT(V_2)$ are tori !)

$\endgroup$
7
  • $\begingroup$ Hmm, I have the feeling that this is not true; but I am far from confident, and not an expert. I think you might have to look at the Hodge group $\textrm{Hdg}(V)$, which is the Mumford–Tate group without the homotheties. Now you have ‘too many scalar multiplications’ on the right hands side. But maybe I am mistaken (this is just a quick guess). $\endgroup$
    – jmc
    Commented Nov 13, 2013 at 8:58
  • $\begingroup$ Oh, and you can download to pdfs here math.ru.nl/~bmoonen/research.html#lecturenotes that deal with similar questions, and also give a precise definition of Hodge groups. $\endgroup$
    – jmc
    Commented Nov 13, 2013 at 9:00
  • $\begingroup$ Yes, but in my case the two M-T groups are adjoint; is it not the case then that the M-T groups and Hodge groups are the same? $\endgroup$
    – user42721
    Commented Nov 13, 2013 at 9:42
  • $\begingroup$ I do not know the answer to that question. I am not familiar with adjoint MT-groups. $\endgroup$
    – jmc
    Commented Nov 13, 2013 at 9:51
  • 1
    $\begingroup$ MT groups of abelian varieties may well never be semisimple, but there are plenty of Hodge structures in general whose Mumford-Tate groups are adjoint (I think it forces them to be of weight 0). Given a Q-Hodge structure V, one can make a tensor construction that will kill the centre of the Mumford-Tate group. Am I wrong? $\endgroup$
    – user42721
    Commented Nov 13, 2013 at 12:38

1 Answer 1

6
$\begingroup$

The answer is no. Namely, let $E$ be an elliptic curve without complex multiplication and $V=H^1(E,Q)$ its first rational cohomology group, which is a 2-dim'l rational Hodge structure of weight 1. Its Hodge (resp. Mumford-Tate) group is $SL(V) \cong SL(2)$ (resp. $GL(V)$). Let $m$ be a positive integer and $V_m=Sym^{2m}(V)$ be the $2m$-th symmetric power of $V$, which is the (absolutely) irreducible rational Hodge structure of weight $2m$ with Hodge group $PSL(V)\cong PSL_2$. (The irreducibility follows from the representation theory of $SL_2$.) Now its twist $\tilde{V}_m=V_m(-m)$ is the irreducible rational Hodge structure of weight 0 with ``the same" (adjoint) Hodge/Mumford-Tate group $PSL(V)$; the Hodge structures $\tilde{V}_m$ are not isomorphic for different $m$, because they have different dimension, namely, $1+2m$. Now take two distinct positive integers $m$ and $n$. Then $MT(\tilde{V}_m \times \tilde{V_n})$ is still $PSL(V)$, which is strictly less than $PSL(V)\times PSL(V)$. This provides a desired counterexample.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .