Torsion points on families of products of elliptic curves (Q2450100)

Motivated by work of the second author on unlikely intersections [Some problems of unlikely intersections in arithmetic and geometry. With appendixes by David Masser. Annals of Mathematics Studies 181. Princeton, NJ: Princeton University Press (2012; Zbl 1246.14003), (Chapter III)], the authors proposed the following Conjecture. Let \(\mathscr{S}\) be a semiabelian scheme over a variety defined over \(\mathbb{C}\), and denote by \(\mathscr{S}^{[c]}\) the union of its semiabelian subgroup schemes of codimension at least \(c\). Let \(\mathscr{V}\) be an irreducible closed subvariety of \(\mathscr{S}\). Then \(\mathscr{V}\cap \mathscr{S}^{[1+\dim(\mathscr{V})]}\) is contained in a finite union of semiabelian subgroup schemes of \(\mathscr{S}\) of positive codimension. This conjecture is related to a conjecture of \textit{R. Pink} [``A common generalization of the conjectures of André-Oort, Manin-Mumford and Mordell-Lang'', Preprint] and to the conjectures of \textit{B. Zilber} [J. Lond. Math. Soc., II. Ser. 65, No. 1, 27--44 (2002; Zbl 1030.11073)]. It is a relative form of the Manin-Mumford conjecture. Recent work of the authors [Am. J. Math. 132, No. 6, 1677--1691 (2010; Zbl 1225.11078)] implies the conjecture when \(\mathscr{S}\) is isogenous to the fibred product of two isogenous elliptic schemes and \(\mathscr{V}\) is a curve. In the paper under review the authors prove the conjecture in the case where \(\mathscr{S}\) is a fibred product of two arbitrary elliptic schemes and \(\mathscr{V}\) is a curve. This gives the following impressive theorem towards the above conjecture in the case where \(\mathscr{S}\) is proper and of relative dimension two. Theorem. Let \(\mathscr{A}\) be a non-simple abelian surface scheme over a variety defined over \(\mathbb{C}\), and let \(\mathscr{V}\) be an irreducible closed curve in \(\mathscr{A}\). Then \(\mathscr{V}\cap \mathscr{A}^{[2]}\) is contained in a finite union of abelian subgroup schemes of \(\mathscr{A}\) of positive codimension. Note that a recent preprint of the authors [``Torsion points on families of simple abelian surfaces'', Preprint] settles the conjecture also in the case where \(\mathscr{S}\) is a \textit{simple} abelian surface scheme and \(\mathscr{V}\) is a curve. It turns out, however, that some care is necessary in connection with the above conjecture when \(\mathscr{S}\) is not proper over the base (even if the relative dimension is still two): \textit{D. Bertrand} [``Special points and Poincaré biextensions'', Preprint] discovered a counterexample related to so-called Ribet sections in a situation where \(\mathscr{S}\) is an extension of an elliptic scheme by \(\mathbb{G}_m\). On the other, hand recent work of \textit{D. Bertrand}, \textit{A. Pillay} and the authors [``Relative Manin-Mumford for semi-abelian surfaces'', Preprint, \url{arXiv:1307.1008}] shows that Ribet sections are essentially the only obstruction to the conjecture if \(\mathscr{S}\) is of relative dimension two over a one-dimensional base.