Abstract

Crucial experiments have been largely neglected by philosophers of science. The main reason for this predicament is that Duhem’s criticism of that kind of experiment has been accepted as sound and definitive. In this article, I start by revisiting the main argument against the possibility of crucial experiments, which is based on epistemological holism. I contend that the argument rests on the confusion between crucial and decisive experiments. When crucial experiments are deprived of their supposed decisive character, the argument loses its bite. Epistemological holism applies to any experiment, whether crucial or not, but it does not imply that experiments are not possible or that they do not have any epistemological import. This variety of holism simply shows that any evidence has to be interpreted and assessed within a theoretical context that includes many auxiliary hypotheses and presupposed theories, which are regarded as accepted background knowledge. This knowledge is not put to the test in a given experiment, but it is rather employed in describing the experimental result and interpreting its theoretical consequences. The meaning of any crucial experiment has then to be extracted from the theoretical context in which the experimental result is interpreted. When the background of accepted knowledge undergoes a drastic change, a crucial experiment may be reinterpreted in such a way that confirms or refutes hypotheses or theories not available at the moment in which it was performed. I will illustrate this kind of reinterpretation with the historical cases of Fizeau’s 1851 experiment, the Michelson and Morley 1887 experiment, and Eddington’s 1919 experiment. I will conclude by vindicating crucial experiments.