Abstract

According to scientific realists, the predictive success of mature theories provides a strong epistemic basis for thinking that such theories are approximately true. However, we know that many theories once regarded as well-confirmed and predictively successful were eventually replaced with successor theories, and some claim this undermines the epistemic confidence we should have in the approximate truth of current science. Selective scientific realists in turn argue that if one can show that the predictive success of some rejected theory T is a function of theoretical claims consistent with current science, then T’s failure doesn’t undermine the claim that current successful theories are approximately true. As such, Selective Scientific Realism (SSR) can be tested through historical examples. Showing that the predictive success of a failed theory is the result of theoretical features later rejected provides a counterexample to SSR. Conversely, SSR is supported if its explanation of the predictive success of failed theories--namely, that the factors which lead to predictive success in the failed theory survive in the successor theory--is able to handle a wide array of historical cases.