Towards a Mori Theory on Compact Kähler Threefolds, I (Q4358519)

The fundamental results of S. Mori about the existence and classification extremal contractions of projective manifolds \(Y\) whose canonical bundles are not numerically effective has become a cornerstone for the birational classification of projective manifolds, e.g. the minimal model program. With the paper under review the authors start a program to establish analogous methods for compact Kähler \(n\)-folds \(X\) which can be approximated by projective manifolds in the sense that there is a smooth family \(\pi: X\to \{z\in\mathbb{C}^m \mid |z|<1\}\) of compact Kähler \(n\)-folds with \(\pi^{-1} (0)=X\) and \(\pi^{-1} (\alpha_\nu)= : X_\nu\) projective for some sequence \(0\neq \alpha_\nu \to 0\). The idea behind this setting is the following: If \(K_X\) is not nef, then \(K_{X_\nu}\) should not be nef for \(\nu\gg 0\), and suitable contractions of the \(X_\nu\) should lead to a ``limit contraction'' of \(X\). For projective manifolds the notion ``nef'' for a line bundle can be defined by estimates for the curvature of a suitable metric on the bundle. In this sense the notion ``nef'' is to be understood here also. The authors restrict their considerations to the case \(n=3\). The main result of the paper is the construction of special contractions of compact Kähler threefolds \(X\) from a given compact non-splitting family of rational curves on \(X\). These constructions are analogous to those for projective threefolds.