Circularities in Experimental Test

Abstract

The experimental test of scientific laws has two essential characteristics: (a) the holism of predictious generation (the predictions are deduced from sets of laws and not from individual laws) and (b) the theoretical weight of measurements (the measurements are based on laws). The former aspect brings about "Duhem's problem": if a prediction is refuted, it is necessary to decide which of its different premises must be modified. Point (b) above implies circularities in experimental test that can be extreme: the same laws may be part of the deduction of a prediction and, simultaneously, of the fundaments of the prediction that allows to verify it. The analysis of one of such cases, the Venturi effect, reveals that circularity does not protect the theories of experimentation and, furthermore, that its methodological effects reduce to reinforce holism, thus increasing "Duhem's problem".