MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $E$ be a CM elliptic curve defined over a quadratic imaginary field $K$ with maximal order, that is, $\mathrm{End}_K(E)\cong \mathcal{O}_K$. Suppose the class number of $K$ is equal to $1$. Let $\mathfrak{p}$ be a prime ideal of $\mathcal{O}_K$ which satisfying $E(K)/\mathfrak{p}E(K)=0$. In this case, is it possible to derive finiteness of $E(K)$ using Mordell-Weil?

share|cite|improve this question
"Using Mordell-Weil" ? If you know that $E(K)$ is a finitely generated $\mathcal{O}_K$-module, then it is quite clear that $E(K)$ won't contain a copy of $\mathcal{O}_K$ using the decomposition theorem for finitely generated $\mathcal{O}_K$-modules because $\mathcal{O}_K/\mathfrak{p}\neq 0$ . – Chris Wuthrich Dec 21 '13 at 19:42
@ChrisWuthrich I am sorry, could you explain more details? In general, for a fixed ring $R$, a finitely generated $R$-module contains a copy of $R$, doesn't it? – Kevin.lijh Dec 22 '13 at 1:40
No. $\mathcal{O}_K/\mathfrak p$ is a counter-example for $R=\mathcal{O}_K$ and $\mathbb{Z}/2\mathbb{Z}$ a counter-example for $R=\mathbb{Z}$. – Chris Wuthrich Dec 22 '13 at 15:09
up vote 4 down vote accepted

Let $Q\in E(K)$. Since $E(K)/\mathfrak{p}E(K)=0$, we know that $Q\in\mathfrak{p}E(K)$. You've assumed that $\mathfrak{p}$ is principal, say $\mathfrak{p}=(\pi)$, so $Q=u_1\pi Q_1$ for some unit $u_1\in\mathcal{O}_K^*$ and some point $Q_1\in E(K)$. Repeating $n$ times, we get $Q=u_n\pi^nQ_n$. (For simplicity, I'll assume that $p=N\mathfrak{p}$ is odd.) Then the easiest way to finish the proof at this point is probably to use canonical heights to conclude that $$ \hat h(Q)=\hat h(u_n\pi^nQ_n)=p^n\hat h(Q_n). $$ Next we observe that $$ h(Q_n) = \hat h(Q_n) + O(1) = p^{-n}\hat h(Q) + O(1), $$ so $(Q_n)_{n\ge1}$ is a set of bounded height in $E(K)$, hence is finite. Hence we can find $n>m$ with $Q_n=Q_m$, and hence $$ Q = u_n\pi^nQ_n = u_n\pi^nQ_m = u_nu_m^{-1}\pi^{n-m}u_m\pi^mQ_m = u_nu_m^{-1}\pi^{n-m}Q. $$ Therefore $$ (u_nu_m^{-1}\pi^{n-m}-1)Q = 0, $$ which proves that every point in $E(K)$ is a torsion point.

(There may well be easier ways to do this, but this is the first thing that occurred to me.)

share|cite|improve this answer
Maybe, I cannot get the result just from the statement of Mordell-Weil. – Kevin.lijh Dec 21 '13 at 3:27
The proof of MW has two parts, first show $E(K)/mE(K)$ is finite, then use heights to deduce finite generation of $E(K)$. So you just want to use the fact that $E(K)$ is finitely generated, without using heights. Doesn't the fact that $Q$ is infinitely $\pi$-divisible contradict the finite generation of $E(K)$ (assuming that $Q$ is non-torsion, of course)? – Joe Silverman Dec 21 '13 at 16:45

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.