Schottky groups acting on homogeneous rational manifolds (Q2317008): Difference between revisions

The object of this paper is to construct new families of Schottky group actions leading to new examples of non-Kähler compact complex manifolds having free fundamental group. The classical concept of a Schottky group action on \({\mathbb P}^1\) was extended by \textit{M. V. Nori} [Contemp. Math. 58, 195--197 (1986; Zbl 0597.32026)] to actions on \({\mathbb P}^{2n+1}\). It was subsequently shown by \textit{F. Lárusson} [Ann. Inst. Fourier 48, No. 1, 223--246 (1998; Zbl 0912.32020)] that, under a technical assumption, the quotient manifold \(Q_\Gamma\) has Kodaira dimension \(-\infty\), is rationally connected and is not Moishezon. Moreover, in the case of \({\mathbb P}^3\), \(Q_\Gamma\) has algebraic dimension \(0\). So far, the only known examples of Schottky groups are discrete subgroups of the automorphism group of \({\mathbb P}^{2n+1}\). In this paper, the authors extend the definition of Schottky groups to allow actions on connected compact complex manifolds \(X\) and particularly on homogeneous rational manifolds. They begin by defining a concept of Schottky pair and more particularly a movable Schottky pair in \(X\); these exist in certain cases and can give rise to Schottky group actions (see Section 2 and Proposition 3.2 for details). The main result (Theorem 4.2) is as follows. Let \(G\) be a connected semisimple complex Lie group, let \(Q\) be a parabolic subgroup of \(G\) and let \(G_0\) be a non-compact real form of \(G\) whose minimal orbit is a hypersurface in \(X:=G/Q\). The Schottky pairs giving rise to Schottky group actions are precisely those on \({\mathbb P}^{2n+1}\), the quadrics \(Q_{4n+2}\) and the isotropic Grassmannians \(\operatorname{IGr}_n({\mathbb C}^{2n+1})\). The authors prove that the quotient manifold \(Q_\Gamma\) is non-Kähler, rationally connected and has Kodaira dimension \(-\infty\) (Proposition 6.1). They give a criterion for the algebraic dimension to be \(0\) (Theorem 6.2) and give examples where the algebraic dimension is strictly positive. They also determine the Picard group pf \(Q_\Gamma\) (Theorem 6.9) and establish the dimension and smoothness of its Kuranishi space of versal deformations (Theorem 6.13). Some of these results are new even in the case of \({\mathbb P}^{2n+1}\). There is an appendix classifying all triples \((G,G_0,Q)\) such that the minimal \(G_0\)-orbit in \(G/Q\) is a hypersurface.