Uniformity of stably integral points on principally polarized Abelian varieties of dimension \(\leq 2\) (Q5937684): Difference between revisions

The starting point of this well-written paper is the following principle regarding uniformity of rational points: \[ \begin{matrix} \text{Lang's conjecture in}\\ \text{arbitrary dimension}\end{matrix} \implies \begin{matrix} \text{Uniform version of Lang's conjecture}\\ \text{in a fixed dimension. } \qquad\qquad\qquad\end{matrix} \] Recall that Lang's conjecture asserts that for a projective variety of general type \(X\), defined over a number field \(K\), the \(K\)-rational points \(X(K)\) of \(X\) are contained in a proper Zariski closed subvariety of \(X\). The first instance of the validity of the above principle was shown in the seminal paper of \textit{L. Caporaso, J. D. Harris} and \textit{B. C. Mazur} [J. Am. Math. Soc. 10, 1-35 (1997; Zbl 0872.14017)]; precisely they showed that if Lang's conjecture holds in arbitrary dimensions, then a uniform version of it holds in dimension one. This was generalized to arbitrary dimension by \textit{B. Hassett} [Duke Math. J. 85, 95-107 (1996; Zbl 0874.14030)], \textit{D. Abramovich} [Invent. Math. 127, 307-317 (1997; Zbl 0898.11020)] and \textit{Abramovich} and \textit{J. F. Voloch} [New York J. Math. 2, 20-34 (1996; Zbl 0887.14010)]. A variety (over a field of charateristic zero) is of logarithmic general type if there exists a desingularization \(\widetilde{X} \rightarrow X\) and a projective embedding \(\widetilde{X} \subset Y\) where \(D=Y\setminus \widetilde{X}\) is a divisor of normal crossing, such that the invertible sheaf \(\omega_{Y}(D)\) is big. Now for varieties of logarithmic general type (over a number field) there is a generalization of Lang's conjecture (Lang-Vojta conjecture) that goes as follow: Let \(S\) be a finite set of places of \(K\) and let \({\mathcal O}_{K, S}\) denote the ring of \(S\)-integers. Fix a model \({\mathcal X}\) of \(X\) over \({\mathcal O}_{K, S}\). Then the set of \(S\)-integral point on \({\mathcal X}\) is not Zariski dense in \({\mathcal X}\). The principle inspiring the work of this paper is the natural generalization to the logarithmic case of the above uniformity principle, namely: \[ \begin{matrix} \text{The Lang-Vojta conjecture in}\\ \text{arbitrary dimension}\qquad\qquad\end{matrix} \implies \begin{matrix} \text{Uniform version of Lang-Vojta} \\ \text{conjecture in a fixed dimension.}\end{matrix} \] The first observation of the authors is that one has to restrict to stably integral points. They actually provide an example in dimension one in which without the restriction to stably integral points no uniformity can be proven. The main result of the paper is the following theorem: Assume that the conjecture of Lang and Vojta holds true. Then there exists a number \(N(K,S)\) such that for any principally polarized Abelian surface with a theta divisor \((A,\Theta)\), the number of stably \(S\)-integral points of \(A\setminus \Theta\) is uniformly bounded by \(N(K,S)\). A similar result for elliptic curves was proven by the first author in [Invent. Math. 127, 307-317 (1997; Zbl 0898.11020)]. As the authors point out, most of their arguments work in arbitrary dimension and they expect the result to hold true for Abelian varieties of any dimension.