Abstract

The Duhem/Quine thesis, as it is customarily understood, asserts that empirical evidence cannot falsify a scientific theory. This dissertation has three general objectives. First, I make the customary version of the thesis percise. Second, I argue that, as I have formulated it, it is true, but not for the reasons often cited. I further argue that once the genuine reasons for its truth are understood, the limits of the thesis are apparent. Third, I explore the relationship between my precise version of the thesis and the epistemological problems that arise when one tries to justify the claim that, on the basis of empirical evidence, a scientific theory is false. In particular, I try to clarify the dispute on this point between Karl Popper and Imre Lakatos. ;In the first chapter, I formulate a number of versions of the thesis and explore its relationship to crucial experiments. I conclude that the Duhem/Quine thesis can be reduced to the assertion that it always requires more than one logically independent hypothesis whenever an experimental result is deduced. ;The second chapter is an extended investigation of why this might be the case. I consider at length Duhem's claim that epistemological problems inherent in justifying the accuracy of a measurement entail that all statements expressing the result of a measurement are themselves unverifiable hypotheses. I also consider problems generated by hypotheses containing theoretical terms other than those referring to quantities. I conclude that although strictly speaking the Duhem/Quine thesis is true, preserving a hypothesis in the face of experimental results that differ widely from those predicted would imply near absurd things about the process of measurement and the link between theoretical terms and experience. ;In the final chapter, I use these results to show that Popper's "methodological rules" are insufficient to cope with criticisms based on the Duhem/Quine thesis. But I also show that Lakatos' use of an analogue of the thesis to criticize Popper is misdirected, and that it is sometimes possible to justify the claim that a theory is false even though one lacks logically conclusive refuting evidence