Abstract

This dissertation is primarily an analysis of the nature and functions of theoretical models in modern science. In attempting an analysis of a methodological concept of science, however, certain issues of meta-philosophy of science are encountered. Thus, the first chapter is devoted to the formulation of principles for analyzing methodological concepts. The resultant set of criteria for formulating and evaluating characterizations of methodological concepts reflects what is desirable in each of two extreme themes in philosophy of science, prescriptivism and descriptivism. Prescriptivism is the view that there are objective criteria for "perfect" or "full-fledged" theories, models, explanations, etc., by reference to which we can evaluate the actions of scientists and guide future scientific practice. It is a normative position in that it purports to show how scientists ought to proceed. Examples of prescriptivist philosophy of science may be found in the writings of Carnap, Hempel, Nagel, Suppes and Braithwaite. Descriptivism is the view that philosophers of science should merely describe how actual scientists practice. Typical descriptivist method is to engage in historical analysis and present case histories. Descriptivist methodology is exemplified in the writings of Kuhn, Feyerabend, Toulmin, Harre and Hanson. It is argued that neither method, by itself, nor a disharmonious combination of the two, is adequate to obtain a generally applicable and accurate account of a methodological concept. ;This study is a rational reconstruction of 'theoretical model' which attempts to clarify and improve this methodological concept with respect to the framework of actual scientific practices and goals. In concluding that a theoretical model is a set of assumptions about the thing modelled, which invariably states an analogy between the thing modelled and some other, better understood, analogue object or system, several other possibilities were considered and rejected. For instance, it is argued that a theoretical model is not an actual object or system, itself, although the assumptions of a theoretical model make reference to an analogue object or system. In addition, it is argued that a theoretical model is not a deductively connected set of statements which is an interpretation for a calculus and is isomorphic to the calculus of the theory it is a model for. This position is the semantical theory of models which is endorsed by Brodbeck, Nagel, Braithwaite and Suppes. ;Several theoretical models, most notably the black body radiator of astrophysics and the general problem solver of psychology, are examined to illustrate the explanatory value of models. The explanatory value of theoretical models is seen to derive primarily from: the degree of analogy between the thing modelled and the analogue object or system, and the degree of understanding of the nature and behavior of the analogue object or system. ;Theoretical models are intimately connected with the growth of theoretical understanding and with scientific progress, in general. On my view, there is a strong correlation between the sorts of functions that a theoretical model is employed to fulfill and the stage of theory formation and level of understanding in the domain under investigation. However, nearly all theoretical models serve some sort of heuristic and meaning-constituting functions. Even though the fate of most theoretical models is eventual modification or abandonment, vestiges of earlier models can usually be found in subsequent models and theories. Theoretical models have a certain versatility and flexibility which enable them to serve as a repository for the principles and concepts which have proven fruitful in previous models and theories, and which continue to be useful in achieving and applying theoretical understanding