Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Let $A/k$ be an abelian variety with real multiplication by some ring of integers $\mathcal O \subset F$. Let $n$ be an integer prime to the characteristic of $k$.

We have the standart $e_n$ pairing $A[n] \times A^\vee[n] \to \mu_n$. It satisfies $e_n(a \cdot x, y) = e_n(x, a \cdot y)$ for any $a \in \mathcal O$.

It is desirable to "extend" this pairing to an $\mathcal O$-linear pairing with respect to the $\mathcal O$-action on $A[n] \times A^\vee[n]$.

In Rapoprt's "Compactifications de l'espace de modules de Hilbert-Blumenthal" 1.21 we find such an extension:

$e_{n_{\mathcal O}} \colon A[n] \times A^\vee[n] \to (\mathcal D^{-1}/n\mathcal D^{-1})(1)$

where $\mathcal D$ is the different of $F$ and with $(1)$ we denote the Tate-twist. Concerning the definition of the pairing it is only said that $Tr(e_{n_{\mathcal O}}) = e_n$.

Can someone make this definition more explicit?

share|improve this question
Isn't it just nothing to do with abelian varieties and just a case of unravelling some algebra and using that $Hom_{\mathbf{Z}/n)(\mathcal{O}/n,\mathbf{Z}/n)=\mathcal{D}^{-1}/n$? No time to do the exercise but hopefully shouldn't be hard. Think of $e$ as a map from $A^\vee[n]$ to $Hom_{\mathbf{Z}/n}(A[n],\mathbf{Z}/n)(1)$. –  Kevin Buzzard Jun 3 '11 at 18:05

1 Answer 1

up vote 5 down vote accepted

Yes, it's just a linear algebra stuff. Probably, it's more convenient to look at the Weil pairing between the $\ell$-adic Tate modules of $A$ and its dual, taking into account that these modules are free $O\otimes Z_{\ell}$-modules of the same rank.

From the linear algebra point of view the situation is as follows.

Let $e: M \times N \to P$ be a perfect pairing of free $Z_{\ell}$-modules $M$ and $N$ of the same rank that takes values in a free $Z_{\ell}$-module $P$ of rank $1$. Assume also that $M$ and $N$ are free $O\otimes Z_{\ell}$-modules of the same rank. Then we get a natural $Z_{\ell}$-linear map

$M \times N \to Hom_{Z_{\ell}}(O\otimes Z_{\ell}, P),$ $(m,n) \mapsto [a \mapsto e(ax,y)]$ for all $a \in O$. Taking into account that $$Hom_{Z_{\ell}}(O\otimes Z_{\ell}, Z_{\ell})=D^{-1}\otimes Z_{\ell}$$ (via the trace map), we get the natural pairing

$$M \times N \to P \otimes_{Z_{\ell}}[D^{-1}\otimes Z_{\ell}].$$ Here $M$ and $N$ are the Tate modules of $A$ and its dual while $P$ is $Z_{\ell}(1)$ and $e$ is the Weil pairing.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.