The domain of hypotheses and the realism of assumptions

The principle that theories should be tested by the accuracy of their predictions but not by the realism of their assumptions needs to be qualified. As a practical matter we often need to evaluate the applicability of theories to cases for which they have not been tested by their predictions. Here we rely on the fact that theories are applicable only within a specific domain. In determining whether a specific case, for which no direct tests are available, is within the theory's domain, we look primarily at whether the assumptions of the theory are as applicable to it as they are to the cases for which the theory has been successfully tested.