Idealization and empirical testing

It has been argued that the presence of idealizations in physical theories implies that scientific realism is false. Furthermore, it has also been argued that because physical theories incorporate idealizations scientific theories must be accepted a priori, because the objects, processes, etc. quantified over in such expressions cannot be observed in the actual world. I argue against the former view by showing that scientific realism is compatible with the fact that all theoretical claims depend on idealizing assumptions, at least when such claims are properly regimented as a special kind of counterfactual conditional. To this end a logic incorporating such conditionals, VCP, is constructed and given a semantic basis. Given VCP, the latter problem is considered in light of extant theories of confirmation, and all are found to be incapable of dealing with the acceptance of theoretical claims that depend on idealizing assumptions. In addressing this problem it is also shown that a new problem arises for subjective Bayesians concerning prior probabilities of theoretical claims that depend on idealizing assumptions. Insofar as I claim that all theoretical claims depend on idealizing assumptions, this problem for Bayesian confirmation theory appears to be quite serious. I conclude that only inference to the best explanation, when understood in terms of VCP, can provide a rational basis for the empirical acceptance of such theoretical claims