Let $E_1$ and $E_2$ be elliptic curves over a field $k$, and let $l$ be a prime coprime to the characteristic of $k$ (if $char(k) \ne 0$). Let $\varphi$ denote the canonical map

$Hom(E_1,E_2)\otimes_{\mathbb{Z}} \mathbb{Z}_l \rightarrow Hom_{G_k}(T_lE_1,T_lE_2)$ `.

For any field $k$, $\varphi$ is easily shown to be injective, and (a case of) the Tate conjecture says that if $k$ is a finite field or a number field then $\varphi$ is surjective (though the proof is hard). What can we say if instead $k$ is a local field?

More precisely, my questions are:

1)Is there a reason/counterexample explaining why $\varphi$ will not be surjective for $k$ a local field?

2)If the answer to (1) is `yes', is there a weaker statement along the lines of the above which is (or is expected to be) true for $k$ a local field? (Sorry this is rather vague).

Thanks, David

  • 4
    $\begingroup$ See mathoverflow.net/questions/53014/… $\endgroup$ Aug 20, 2011 at 12:25
  • $\begingroup$ Thanks Felipe Voloch, the answers there are exactly what I was looking for. Sorry for duplicating your question, I didn't think to search for the name `isogeny theorem'. $\endgroup$ Aug 20, 2011 at 13:27

1 Answer 1


Hi David,

A nice question. The map $\varphi$ can be very far from surjective! One way to see this is as follows. Let us work over $Q_p$, and suppose first that $l\neq p$. Working with elliptic curves with good reduction, the corresponding Galois representation is determined by $a_p$ (as this determines the characteristic polynomial of Frobenius).

Now suppose instead that $l=p$, and consider elliptic curves with good ordinary reduction. Then the Galois representation is again determined (up to a finite number of possibilities) by $a_p$. More generally, the classification of crystalline Galois representations by weakly admissible modules shows that the local Galois representation is always determined up to a finite number of possibilities by $a_p$.

Now thinking about cardinality shows that there must be many pairs of elliptic curves over $Q_p$ which are not isogenous, but have isomorphic $l$-adic Galois representations.

  • $\begingroup$ Thanks very much JT, that seems to do it (combined with Felipe's link for part (2)). $\endgroup$ Aug 20, 2011 at 13:26

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.