Abstract

Since the publication of Thomas Kuhn's The Structure of Scientific Revolutions, many philosophers and psychologists have supposed that the key concepts of a linguistic community are transmitted to initiates by means of exemplars, particular items which the community accepts as representative exemplifications of the concept. Although this view has several advantages over the traditional view that concepts are learned from definitions or analyses, it is incomplete in the absence of a theory of how the transition is made from viewing group-licensed exemplars to possessing a concept which is shared with one's linguistic community. Further, although there is general agreement that the concepts of a given society may change with time, there is not agreement on how this conceptual evolution occurs and whether or not it is a rational process. The dissertation contributes to a resolution of these issues by arguing for a particular descriptive theory of the dynamic processes of conceptual activity. ;According to the theory proposed, concepts are capacities for acquiring states which permit the person who is in those states to treat certain things as distinct from others in his ken. An idealized model of the items in a given exemplar set bridges the gap between examining an exemplar set and possessing the associated capacity. The model completes an explanation of certain properties of the exemplar set in terms of certain other properties and a subject's background beliefs. Such models differ from actual exemplars and, thus, stand in need of justification. ;Case studies of the development of the concept of a feedback control system in engineering and the concept of liability for negligence to noncontracting parties in law reveal some types of arguments which are used to justify conceptual models. Interestingly, the same types of arguments which justify acceptance of a given model are evidently also employed in choosing among competing models and in justifying the extension of a concept to nonexemplars. The battery of arguments, plus an understanding of their function, provides the beginnings of a picture of the transition from exemplars to concepts for societies, as well as for individuals; however, the actual or possible particulars from which a society develops its concepts are not given as exemplars of a concept. Individuals in the society forge new concepts out of items which seem to have valuable or noteworthy properties. The attempt to fulfill this objective is concept "optimization." ;The justifying arguments used in model construction, acceptance, and application all serve as optimizing arguments. In additon, there is a second class of optimizing arguments: those which utilize a concept's record of success or failure in maximizing access to valued properties to justify alterations in the background beliefs and exemplar set which support the associated model. ;The notion of concept optimization provides a picture of conceptual evolution as a rational process. If the account is accurate, the appropriate stance for the learner of complex concepts is optimization. The function of the definitions and analyses which play the central role in traditional theories of conceptual activity is to accelerate the optimizing process for the individual concept learner by directing his attention and suggesting fruitful models