Paul Draper argues that all else held equal, "naturalism is much more probable than theism," and therefore "theism is very probably false"; moreover, naturalism is simpler and smaller in scope than theism, and has much greater predictive power than theism with respect to evolutionary facts about suffering. In this response, Alvin Plantinga disputes that theism has larger scope than naturalism, and argues that what is really at issue for epistemic probability is not simplicity as Draper understands it (as "uniformity"), but "epistemic naturalness"—and that theism is more epistemically natural than naturalism. Moreover, if we treat theism as a hypothesis (rather than as a fact), theism might be subject to prima facie defeat by facts about suffering and misery, but nevertheless explain or predict a whole range of other data better than naturalism, such as our possession of reliable cognitive faculties, the existence of objective morality, the fine-tuning of the universe, the existence of abstract objects, and so on. But if some theists know that theism is true (in virtue of religious experiences. say), then their theism is not subject to defeat by facts about suffering even disregarding these explanatory advantages.
Paul Draper's critique of the evolutionary argument against naturalism (EAAN) alleges that it does not follow from the probability thesis that evolutionary naturalists cannot rationally believe in unplanned evolution, but that this conclusion does follow from the amended but demonstrably false thesis that the probability of reliable cognitive faculties given unplanned evolution is low. Though Plantinga disputes that the probability thesis does not entail this conclusion, he does not take up that argument here. Rather, he aims to show (contra Draper) that the probability of reliable cognitive faculties given unplanned evolution is indeed low, given that evolution selects only for "indicators," not full-fledged beliefs. If evolutionary naturalism is true, then both true and false belief content will yield equally adaptive behavior, and thus natural selection will not select for true belief content over false belief content; but then naturalism is indeed self-defeating in the sense that naturalists cannot trust the cognitive faculties that lead them to believe that naturalism is true.
In this chapter, Alvin Plantinga argues that naturalism entails that our beliefs cannot affect our behavior, but natural selection only selects for beneficial behaviors. Consequently, natural selection cannot select for beneficial beliefs on naturalism, and thus the probability that human beings have evolved reliable cognitive faculties is low or inscrutable if evolution has occurred without supernatural guidance. Evolutionary naturalism, then, is "self-defeating" in the sense that if it were true, we would have no good grounds to believe that our cognitive faculties are reliable, and thus no good grounds to believe that evolutionary naturalism is true. Moreover, because our cognitive faculties are reliable, evolution actually provides evidence that naturalism is false—and thus there is a "religion/science conflict" between the quasi-religion of evolutionary naturalism and the science of evolution.
