Homology of finite K-area (Q379238): Difference between revisions

Gromov introduced the notion of K-area (which can be infinite), with the motivation that its infinity is an obstruction to positive scalar curvature of closed spin manifolds. This concept systemizes the use of ``almost flat twist bundles'' to obstruct the existence of metrics of positive scalar curvature. The paper at hand strives to make an even more systematic study of the homological meaning of finite and infinite K-area. Indeed, Listing defines a subgroup of the ordinary homology groups of ``homology classes with small K-area'' essentially as those homology classes which pair trivially with all K-theory classes which have almost flat representatives. The main focus of the paper is the systematic study of this subgroup of homology and of its calculation in specific examples. In particular, functoriality, behavior under coverings and with respect to product classes is discussed. One should be aware that, despite the fact that Listing calls these subgroups K-area homology, they do not obey (as is shown in the article) the decisive axioms of homology, namely excision or Mayer-Vietoris.