Abstract

Duhem's thesis holds that there is no logical way to decide the issues between competing theories so that crucial experiments cannot exit. This dissertation shows, contrary to Duhem's thesis, that it is sometimes possible for scientists to perform crucial experiments and that those experiments follow the logic of a crucial experiment. To solve the problem raised by Duhem's thesis, first, this dissertation suggests a robust interpretation that Duhem's thesis consists of four sub-thesis. According to this interpretation, Duhem's thesis is different from Quine's thesis and the underdetermination thesis. The second part deals with nine versions of Duhem's problem found in typical theories such that those cannot be solved within scopes of their theories. With these backgrounds the third part of the dissertation suggests a solution to Duhem problem set up by the robust interpretation. First, two components of the incommensurability thesis, the theory-ladenness of observation and meaning variance, are critically examined by suggesting counter-examples and a distinction between a guiding theory and a target theory and by showing meaning invariance using the dynamic type hierarchy theory. Secondly, a solution to Duhem's problem is suggested by presenting the logic of a crucial experiment, by criticizing Duhem's thesis in terms of the logic of a crucial experiment, and by giving case studies of four historical episodes.