Abelian varieties isogenous to no Jacobian (Q2296300)

This paper is motivated by a question of \textit{C.-L. Chai} and \textit{F. Oort} [Ann. Math. (2) 176, No. 1, 589--635 (2012; Zbl 1263.14032)] who ask for the existence of an abelian variety defined over the field \(\overline{\mathbb{Q}}\) of all algebraic numbers that is not isogenous to the Jacobian of any (stable) curve. Classical results show that a positive answer is only possible for abelian varieties of dimension at least \(4\). Positive answers are finally given independently by \textit{C.-L. Chai} and \textit{F. Oort} [Ann. Math. (2) 176, No. 1, 589--635 (2012; Zbl 1263.14032)] and \textit{J. Tsimerman} [Ann. Math. (2) 176, No. 1, 637--650 (2012; Zbl 1250.14032)] who construct special abelian varieties with complex multiplication CM. These varieties are quite rare, a refined version of the first question is given by Chai and Oort as follows. They start with an algebraic (special) hypersurface \(\mathcal{H}\) over \(\overline{\mathbb{Q}}\) in the Siegel moduli space \(\mathcal{A}_g\) and ask for something explicit in \(\mathcal{A}_g\) without CM, or with CM but not Weyl CM, that is not isogenous to anything over \(\overline{\mathbb{Q}}\) in \(\mathcal{H}\). In the present paper the authors prove the existence of such abelian varieties without CM and being Hodge generic. In a first result (Theorem 1.1) they show that for an algebraic hypersurface \(\mathcal{H}\) in \(\mathcal{A}_g\) \((g\geq 2)\) there exists a Hodge generic abelian variety \(A\) in \(\mathcal{A}_g\) that is not isogenous to any \(B\) in \(\mathcal{H}\), while it is defined over an extension of \(\mathbb{Q}\) of degree at most \(2^{16g^4}\). This implies (Corollary 1.2) that there is a principally polarized and Hodge generic abelian variety of dimension \(g\geq 4\) that is not isogenous to any Jacobian and defined over a number field of degree at most \(2^{16g^4}\). They not only answer Chai's and Oort's question, but they also give a bound for the degree of the field of definition. Their next result (Theorem 1.3) focuses on the set of these special abelian varieties. They consider a finite covering \(\tilde{\mathcal{A}}\) of \(\mathcal{A}_g\) together with a finite map \(\Psi\) from \(\tilde{\mathcal{A}}\) to affine \(\mathbb{A}^G\) defined over \(\overline{\mathbb{Q}}\) and a hypersurface \(\mathcal{H}\) in \(\mathcal{A}_g\). Depending on a chosen integer \(N\) and a few mild conditions, they show that for finite \(\mathbf{n}\in\mathbb{Z}^G\) the projection of any element of \(\Psi^{-1}(\mathbf{n})\) to \(\mathcal{A}_g\) is either not defined over a number field of bounded degree or isogenous to an element of \(\mathcal{H}\). As a consequence (Corollary 1.4) they show for \(g\geq 2\) that there is a set of principally polarized \(g\)-dimensional abelian varieties, dense in the euclidean topology, with each defined over a number field of degree at most \(2^{16g^4}\) and not isogenous to any of the others or to any Jacobian. In contrast to the case of CM, this result tells us that in some sense the majority of abelian varieties is not isogenous to any Jacobian. For \(g=2,3,4,5\), \(\mathcal{A}_g\) defined over \(\mathbb{Q}\) and under the assumption that there is a dominant rational map \(\Xi\) from \(\mathbb{A}^G\) to \(\mathcal{A}_g\), they give in Theorem 1.5 a similar statement of Theorem 1.3. They prove under a few assumptions that for any integer \(N\) there are only finitely many \(\mathbf{n}\in\mathbb{Z}^G\) such that \(\Xi(\mathbf{n})\) is either not defined over \(\mathbb{Q}\) or isogenous to an element in the hypersurface \(\mathcal{H}\). In their Corollary 1.6 they conclude under the assumption that \(\mathcal{A}_4\) is unirational over \(\mathbb{Q}\) that there exists a principally polarized abelian fourfold, defined over \(\mathbb{Q}\) and Hodge generic, that is not isogenous to any Jacobian. The last result in Theorem 1.7 considers elliptic curves and says that for any integer \(N\) and a real algebraic curve \(\mathcal{C}\) in \(\mathcal{A}_1\) there are only finitely many elliptic curves that are either not Hodge generic or isogenous to an elliptic curve connected with an element in \(\mathcal{C}\). The general strategy of the proofs is to construct many Hodge generic abelian varieties such that there is no isogeny up to a large degree \(M\) to any element in \(\mathcal{H}\). The property Hodge generic is provided by using Serre's version of the Hilbert Irreducibility Theorem. Continuing by contradiction, they take an abelian variety isogenous to an element in \(\mathcal{H}\). As a consequence the degree of the isogeny is larger than \(M\) and they further generate an upper bound for the degree of the isogeny depending on the degree of the field of definition of the variety.