4

Notation: For a finite abelian extension $F / K$, let $\mathfrak{f}_{F / K} \subset \mathcal{O}_K$ denotes its conductor such that $F$ is contained in the ray class field $K(\mathfrak{f})$. In particular, the set of primes dividing $\mathfrak{f}_{F / K}$ consists precisely of those prime ideals which ramify in $F / K$. Whenever $\chi: \mathbf{A}_L^{\times} \rightarrow \mathbf{C}^{\times}$is a Hecke character of some number field $L$, we let $\mathfrak{f}_\chi$ denote its conductor, i.e. the smallest ideal such that $\chi$ is trivial on $U_f\left(\mathfrak{f}_\chi\right)$.

Let $K$ be an imaginary quadratic field and $F / K$ a finite abelian extension. Let $E / F$ be an elliptic curve with complex multiplication by $\mathcal{O}_K$. Let $F\left(E_{\text {tors }}\right)$ denote the field extension of $F$ generated by all the torsion points of $E$. Let $\psi: \mathbf{A}_F^{\times} \rightarrow \mathbf{C}^{\times}$be the Hecke character associated to $E / F$. Let $N_{F / K}: \mathbf{A}_F^{\times} \rightarrow \mathbf{A}_K^{\times}$denote the idele norm. Assume that there exists a Hecke character $\varphi: \mathbf{A}_K^{\times} \rightarrow \mathbf{C}^{\times}$such that

$$\psi=\varphi \circ N_{F / K}$$

Show that $F\left(E_{\text {tors }}\right)$ is an abelian extension over $K$.

Hint: It is equivalent to show that the $G_K$-module $\operatorname{Ind}_{G_F}^{G_K}\left(E_{\mathrm{tors}}\right)$ is abelian. Note that the Artin map $$\left[-, K^{\mathrm{ab}} / K\right]: \mathbf{A}_K^{\times} \rightarrow \operatorname{Gal}\left(K^{\mathrm{ab}} / K\right)$$

factors through the finite ideles. We may use that the kernel of $\left[-, K^{\mathrm{ab}} / K\right]$ : $\mathbf{A}_{K, f}^{\times} \rightarrow \operatorname{Gal}\left(K^{\mathrm{ab}} / K\right)$ is the topological closure of $K^{\times}$in $\mathbf{A}_{K, f}^{\times}$.

I would really appreciate any help with this exercise with which I am stuck.

Mario
  • 742
  • Isn’t this “trivial”? By CM theory, $E$ has to be defined over the Hilbert class field $H$ of $K$, and $H(E_{tors})$ is well-known to be abelian over $K$. Then $F/K$ is contained in the compositum or $F$ and $H(E_{tors})$ – two abelian extensions of $K$ – so we’re done? I guess I missed something, but what…? – Aphelli Jan 09 '24 at 22:08

0 Answers0