Abstract
Appealing to science is a popular suggestion for separating projectible predicates. According to this suggestion, we can expect science, eventually, to separate such predicatesfor us, rendering it unnecessary to make further attempts to explicate the criteria for pro-jectibility. In this essay, I address three theoretical challenges to this suggestion. The firststems from the inductive character of science, which casts doubts on its efficacy in separat-ing projectible predicates, since induction itself requires this separation. The second is theinferential externalism implied by this suggestion, whereas the problems of induction havebite only if inferential internalism is presumed. The third challenge appears when a strongrelationship between projectibility and kindhood, on the one hand, and kindhood and sim-ilarity, on the other, is posited, such that a more projectible predicate is believed to be apredicate which tracks better similarities in nature. Now the question appears: whether andhow science enables us to track better similarities? I distinguish two conceptions of similarity,one intuitive, the other theoretical, and I argue that the theoretical one is to be preferred,showing that how scientific practice involves shifting from an intuitive idea of similarity to atheoretical one. Through answering these three challenges, I attempt to support the appealto natural science as far as Goodman’s problem of induction is concerned.