This paper continues my application of theories of concepts developed in cognitive psychology to clarify issues in Kuhn's mature account of scientific change. I argue that incommensurability is typically neither global nor total, and that the corresponding form of scientific change occurs incrementally. Incommensurability can now be seen as a local phenomenon restricted to particular points in a conceptual framework represented by a set of nodes. The unaffected parts in the framework constitute the basis for continued communication between the communities supporting alternative structures. The importance of a node is a measure of the severity of incommensurability introduced by replacing it. Such replacements occur incrementally so that changes like that from the conceptual structure of Aristotelian celestial physics to the conceptual structure of Newtonian celestial physics occur in small stages over time, and for each change it is in principle possible to identify the arguments and evidence that led historical actors to make the revisions. Thus the process of scientific change is a rational one, even when its beginning and end points are incommensurable conceptual structures. It is also apparent, from a detailed examination of the conceptual structure of astronomy at the time of Copernicus, that the kind of conceptual difficulty identified as incommensurability may occur within a single scientific tradition as well as between two rival traditions.