Good reduction of algebraic groups and flag varieties (Q2257988): Difference between revisions

The finiteness result considered here stems from the classical theorem of Hermite and Minkowski in number theory: given a finite set \(S\) of places of \(\mathbb{Q}\) and \(n \geq 1\), there exists only finitely many extensions with degree less than \(n\) and unramified outside \(S\). The analogues for curves and abelian varieties are proposed and proved by Shafarevich and Faltings, in which ``unramified'' is replaced by ``good reduction'' outside a finite set \(S\) of places of a number field \(K\). Remarkably, there exists also an automorphic analogue: see [\textit{S. W. Shin} and \textit{N. Templier}, Compos. Math. 150, No. 12, 2003--2053 (2014; Zbl 1308.11053)]. In this paper, the authors consider a variant of Faltings' theorem for connected reductive groups, or for some homogeneous spaces thereof. The first result asserts the following: Theorem 1.1. Let \(B \subset \mathrm{Spec}(\mathcal{O}_K)\) be an open dense subscheme. Then the set of isomorphism classes of \((G,E)\), where \(G\) is an \(n\)-dimensional connected reductive \(B\)-group scheme and \(E\) is a torsor under \(G\), is finite. From this, the authors deduce finiteness results for (i) the case \(E=G\), i.e. \(n\)-dimensional \(K\)-groups with good reduction outside \(S\), (ii) \(n\)-dimensional flag varieties with good reduction outside \(S\). This is actually one step in the authors' investigation of Shafarevich-type conjectures for Fano varieties. A counterexample for unipotent groups is also given, showing that the reductivity assumption is essential.