Table of Contents
Fetching ...

Entanglement of elliptic curves upon base extension

Tori Day, Rylan Gajek-Leonard

TL;DR

The paper investigates how the $(p,q)$-entanglement type of an elliptic curve $E$ over a number field $K$ changes under base extension, focusing on the finite set of entanglement types $\operatorname{Ent}_{p,q}(E/K)$. It proves that for every prime $\ell$ dividing $d(E,p,q)=\gcd(|\operatorname{im}\rho_{E,p}|,|\operatorname{im}\rho_{E,q}|)$ there exists a base field $L/K$ with $T_{p,q}(E/L)\cong \mathbb{Z}/\ell\mathbb{Z}$, i.e., cyclic entanglement of order $\ell$ can be forced. The authors develop a group-theoretic framework, relate entanglement to subfield structure via $K_{p,q}(E)=K(E[p])\cap K(E[q])$, and introduce entangleable fields to study how entanglement can be glued or shrunk under base change, with explicit results for $(2,q)$-entanglements over $\mathbb{Q}$. They also provide a detailed classification of $(2,q)$-entanglement types, including constructions ensuring the presence of $S_3$-type entanglement and identifying exceptional primes and small primes where the behavior deviates. Together, these results give constructive tools to control and predict how torsion-field intersections influence the full mod $pq$ Galois image under base extension, with concrete implications for the arithmetic of elliptic curves.

Abstract

Fix distinct primes $p$ and $q$ and let $E$ be an elliptic curve defined over a number field $K$. The $(p,q)$-entanglement type of $E$ over $K$ is the isomorphism class of the group $\operatorname{Gal}(K(E[p])\cap K(E[q])/K)$. The size of this group measures the extent to which the image of the mod $pq$ Galois representation attached to $E$ fails to be a direct product of the mod $p$ and mod $q$ images. In this article, we study how the $(p,q)$-entanglement group varies over different base fields. We prove that for each prime $\ell$ dividing the greatest common divisor of the size of the mod $p$ and $q$ images, there are infinitely many fields $L/K$ such that the entanglement over $L$ is cyclic of order $\ell$. We also classify all possible $(2,q)$-entanglement types that can occur as the base field $L$ varies.

Entanglement of elliptic curves upon base extension

TL;DR

The paper investigates how the -entanglement type of an elliptic curve over a number field changes under base extension, focusing on the finite set of entanglement types . It proves that for every prime dividing there exists a base field with , i.e., cyclic entanglement of order can be forced. The authors develop a group-theoretic framework, relate entanglement to subfield structure via , and introduce entangleable fields to study how entanglement can be glued or shrunk under base change, with explicit results for -entanglements over . They also provide a detailed classification of -entanglement types, including constructions ensuring the presence of -type entanglement and identifying exceptional primes and small primes where the behavior deviates. Together, these results give constructive tools to control and predict how torsion-field intersections influence the full mod Galois image under base extension, with concrete implications for the arithmetic of elliptic curves.

Abstract

Fix distinct primes and and let be an elliptic curve defined over a number field . The -entanglement type of over is the isomorphism class of the group . The size of this group measures the extent to which the image of the mod Galois representation attached to fails to be a direct product of the mod and mod images. In this article, we study how the -entanglement group varies over different base fields. We prove that for each prime dividing the greatest common divisor of the size of the mod and images, there are infinitely many fields such that the entanglement over is cyclic of order . We also classify all possible -entanglement types that can occur as the base field varies.
Paper Structure (12 sections, 26 theorems, 60 equations, 1 figure)

This paper contains 12 sections, 26 theorems, 60 equations, 1 figure.

Key Result

Theorem 1.2

Let $\ell$ be a prime dividing $d(E,p,q)$. Then there exists a field $L\subseteq K(E[pq])$ such that

Figures (1)

  • Figure 1:

Theorems & Definitions (63)

  • Theorem 1.2
  • Corollary 1.3
  • Example 1.4
  • Remark 1.5
  • Corollary 1.6
  • Definition 2.1
  • Proposition 2.2
  • proof
  • Proposition 2.3
  • proof
  • ...and 53 more