Crucial Experiments

Abstract

Crucial experiments have been largely neglected by philosophers of science because most of them have accepted the Duhem conclusion that they are impossible. This skeptical conclusion rests on the assumption that the purpose of a crucial experiment is to confirm (or even verify) a hypothesis or theory and decisively refute all its possible rivals. In this view, crucial experiments must be decisive; otherwise, they are not possible. I argue that we must distinguish between crucial and decisive experiments and that crucial experiments do not need to be decisive. A crucial experiment is one that tests two or more rival hypotheses or theories and, if successful, confirms only one of them and refutes the others. It does not have to be decisive in any sense of the term, nor does it have to refute all the possible rivals of a given hypothesis. I then distinguish two types of crucial experiments and analyze their logical structure within the standard framework of the hypothetico-deductive account of confirmation. Finally, I conclude with some remarks on the scope and limitations of crucial experiments.