Abstract
In this paper I evaluate Herbert Simon 's important computational approach to scientific discovery, which can be characterized as a contribution to both the "cognitive science of science" and to naturalized philosophy of science. First, I tackle the empirical adequacy of Simon 's account of discovery, arguing that his claims about the discovery process lack evidence and, even if substantiated, they disregard the important social dimension of scientific discovery. Second, I discuss the normative dimension of Simon 's account, here I argue that Simon 's project is best understood as a contribution to "android epistemology." I conclude with some comments on the direction a naturalized yet normative philosophy of science might take