Bordism-finiteness and semi-simple group actions (Q1597477)

A direct consequence of the Atiyah-Bott-Segal-Singer fixed-point theorem in equivariant index theory is the fact that all the Pontryagin numbers of an oriented closed manifold with smooth fixed-point-free \(S^1\)-action vanish. In this paper the author gives bordism-finiteness results for smooth \(S^3\)-manifolds. The class of oriented manifolds which admit an \(S^1\)-action with isolated fixed points such that the action extends to an \(S^3\)-action with fixed point is considered. The author exhibits various subclasses, characterized by an upper bound for the Euler characteristic and properties of the first Pontryagin class \(p_1\), for example \(p_1=0\), which contain only finitely many oriented bordism types in any given dimension. Also, the auhor shows finiteness results for homotopy complex projective and complete intersections with \(S^3\)-action satisfying the above property, which is denoted as: (*) \(S^3\) acts with fixed point and has isolated \(S^2\)-fixed points. At first, the author states the main results: Theorem 1.1. Let \(C\) and \(m\) be natural numbers. The class of connected \(m\)-dimensional oriented manifolds with vanishing first Pontryagin class and Euler characteristic \(\leq C\) which admit an \(S^3\)-action satisfying (*) contains only finitely many oriented bordism types. Theorem 1.2. In a fixed dimension \(2n\) there are only finitely many homotopy complex projective spaces with an \(S^3\)-action satisfying (*). Section 2 deals with tangential weights and weights of equivariant complex line bundles for manifolds which admit an \(S^3\)-action with fixed point. The author gives conditions in terms of the first Pontryagin class which guarantee that these weights are determined ``up to finite ambiguity''. In Section 3 the author extends Theorem 1.1 to the case when a negative multiple of the first Pontryagin class is a sum of squares. In Section 4 the author shows, by examples, that bordism-finiteness fails if one relaxes the conditions on the \(S^3\)-action or the condition on the first Pontryagin class. Finally, in Section 5 the author proves the result on homotopy complex projective spaces (see Theorem 1.2 above) and he uses similar methods to show that in a given dimension there are only finitely many complete intersections which admit an \(S^3\)-action satisfying (*).