3
$\begingroup$

Let $E_i\colon y^2=4x^3+A_ix+B_i$, for $i=1,2$ be two elliptic curves where $A_i,B_i \in \mathbb C$ are algebraic over $\mathbb Q$. For $i=1,2$ let $\Lambda_i\subseteq \mathbb C$ be the unique lattice that parametrizes $E_i$, namely such that $\displaystyle g_2(\Lambda_i)=60\sum_{0\neq \omega\in \Lambda_i}\frac{1}{\omega^4}=A_i$ and $\displaystyle g_3(\Lambda_i)=140\sum_{0\neq \omega\in \Lambda_i}\frac{1}{\omega^6}=B_i$.

The question goes as follows: suppose that $\psi\colon E_1\to E_2$ is an isogeny. Then I can write down rational maps for $\psi$ with coefficients in $\overline{\mathbb Q}$. Now $\psi$ corresponds to a unique complex number $\alpha$ such that $\alpha\Lambda_1\subseteq \Lambda_2$. Does $\alpha$ need to be algebraic over $\mathbb Q$?

$\endgroup$
2
$\begingroup$

I think it is. The point is that you can identify $\Lambda _i$ with $H_1(E_i,\mathbb{Z})$, embedded into $\Bbb{C}$ by $\gamma \mapsto \int_\gamma \eta _i$, with $\eta _i=\frac{dx}{y} $ (there may be some normalization factor, but it should be in $\mathbb{Q}$ so that doesn't matter). Now multiplication by $\alpha $ on $\mathbb{C}$ induces $\psi _*:H_1(E_1,\mathbb{Z})\rightarrow H_1(E_2,\mathbb{Z})$; since $\int_{\psi _*\gamma }\eta _2=\int_\gamma \psi ^*\eta _2$, this implies $\psi ^*\eta _2=\alpha \eta _1$. But both forms are in the $\bar{\mathbb{Q}}$-vector space $H^0(E_1,\Omega ^1_{E_1})$, so $\alpha \in \bar{\mathbb{Q}}$.

$\endgroup$

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.