Abstract

The thesis of the incommensurability of competing scientific theories, as it has been formulated and defended by Thomas Kuhn and Paul Feyerabend, implies the following: there will be competing scientific theories between which there will exist no neutral perspective from which to evaluate the merits of those theories with regards to truth. If this thesis were true, it would threaten what I take to be basic epistemic values, the values of seeking truth and avoiding error. To pass from one theory to another incommensurable with it, one must pass to a theory which, from one's original perspective, appears absurd, trivially false, or even meaningless. Doing so obviously flouts error avoidance. ;Unlike many who have opposed incommensurability, I argue that the incommensurabilist's semantic vision of alternative and widely divergent conceptual schemes is a coherent and plausible one, but I also contend that such a vision cannot possibly characterize rival scientific theories. This last fact removes the threat to epistemology since semantic incommensurability will only arise in situations in which comparison between theories with respect to truth is not at issue. My arguments for these claims are based on a positive account of scientific theories and their interpretation which makes clear how we can discover conceptual overlaps and ultimately find a neutral point of view from which non-question-begging adjudication between rival theories can take place. ;Turning from the general, philosophical arguments which I present in support of my thesis, I examine particular epsiodes from the history of science to illustrate how the view I defend captures the actual evolution of scientific theories. Drawing upon disputes which were central to the Copernican Revolution as well as a more contemporary debate between particular interpretations of Quantum Mechanics, I demonstrate that, although these cases may have required scientists to specify types of disagreements that they were not accustomed to specifying, the rival theories in these instances are nonetheless not incommensurable in the epistemologically threatening way described above