Torsion of Abelian varieties over large algebraic fields (Q2567158)

Let \(A\) be an Abelian variety over a number field \(K\) and let \(\widetilde K\) denote the algebraic closure of \(K\). The authors' main result is that \(K\) has a finite Galois extension \(L\) such that for almost all (all but a set of \(\sigma\) of Haar measure \(0\)) \(\sigma\in\text{Gal}(L)\), there are infinitely many prime numbers \(l\) with \(A_l(\widetilde K(\sigma))\neq 0\). Suppose \(K\) is an infinite finitely generated field over its prime field. Let \(K_s\) denote the separable closure of \(K\) in \(\widetilde K\) and let \(\text{Gal}(K)\) denote the absolute group of \(K\). Then the latter group is profinite and is therefore equipped with a unique Haar measure \(\mu K\) satisfying \(\mu_K(\text{Gal}(K))= 1\). For all \(\sigma= (\sigma_1,\dots, \sigma_e)\in \text{Gal}(K)^e\), let \(K_s(\sigma)\) be the fixed field of \(\sigma_1,\dots,\sigma_e\) in \(K_s\) and let \(\widetilde K(\sigma)\) denote the maximal purely inseparable field extension of \(K_s(\sigma)\). Properties of \(K_s(\sigma)\) and \(\widetilde K(\sigma)\) that hold for almost all \(\sigma\i\text{Gal}(K)^e\) reflect fundamental theorems of arithmetic geometry like the Hilbert irreducibility theorem and Mordell-Weil theorem which hold over finite extensions of \(K\). The authors then provide a summary of these properties. They consider the torsion part of \(A\) over the fields \(K_s(\sigma)\). They state a result for elliptic curves. They then discuss and state a conjecture concerning arbitrary Abelian varieties: Let \(A\) be an Abelian variety over \(K\). Then the following properties hold for almost all \(a\in\text{Gal}(K)^e\): (a) If \(e= 1\), then there are infinitely many prime numbers \(l\) with \(A_l(\widetilde K(\sigma))\neq 0\). (b) If \(e\geq 2\), then there are only finitely many \(l\) with \(A_l(\widetilde K(\sigma))\neq 0\). (c) If \(e\geq 1\), then for all \(l\) the set \(\bigcup^\infty_{i=1} A_l(\widetilde K(\sigma))\) is finite. By taking \(L= K\) in the authors' main result, they are able to prove part (a) of their conjecture in some special cases. The paper contains many interesting lemmas leading to the proof of their main result.