Abstract

I carefully analyze the two major arguments for scientific realism, what I call the Explanatory Argument for Realism and what I call the Epistemic Arbitrariness Argument . According to advocates of EAR, we should believe in realism because it is the only adequate explanation of the success of science or scientific methodology. On the other hand, according to advocates of EAA, the distinction between observable entities and unobservable entities is epistemically arbitrary, and thus, contrary to what nonrealists claim, inference to the existence of unobservable entities is no more suspect than inference to the existence of observable entities. ;The importance of EAR and EAA for the issue of scientific realism can hardly be overemphasized. For instance, Richard Boyd, one of the most distinguished proponents of scientific realism, has claimed that EAR is "probably the argument that reconstructs the reason why most scientific realists are realists"; and, according to Arthur Fine, something like EAA is what really motivates scientific realists, in their heart of hearts, to be realists. ;I argue that both EAR and EAA fail. However, I also argue that nonrealists such as Bas van Fraassen have failed to provide good reasons to adopt their own position. Thus, I conclude, we are left in a sort of stalemate between realism and nonrealism. After drawing this conclusion, I argue for a new form of nonrealism--what I call categorical empiricism. According to categorical empiricism, we can only be rationally obligated to believe in theoretical entities that fall into categories that are grounded in our everyday experience. I argue that this form of nonrealism has the advantage of casting doubt on the existence of unobservable entities such as quarks and curved space-time without at the same time casting doubt on the existence of unobservable entities such as dinosaurs