Document Type


Publication Date



Being able to make correct novel predictions has long been trumpeted as one of the most important virtues a scientific theory could have. Imre Lakatos (1970, pp. 91-195) has even gone so far as to claim that a research program can supersede its rival only by making some correct novel predictions. Others have based an entire defense of realism on the claim that realism is the only adequate explanation of the success of scientific theories, particularly their success in making correct novel predictions.1 For still others, the ability of the theories of science to make correct novel predictions (and hence be tested) is what separates science from other explanatory enterprises, such as history, economics, natural theology, and psychoanalysis.


Originally published as:

Collins, Robin. "Against the Epistemic Value of Prediction Over Accommodation." Noûs 28, no. 2 (1994): 210-24. doi:10.2307/2216049.