PAUL TELLER WHITHER CONSTRUCTIVE EMPIRICISM? 1 INTRODUCTION In this paper I will set out my understanding of Bas van Fraassen’s constructive empiricism, some of the difficulties which I believe beset the current version, and, very briefly, some valuable lessons I believe are nonetheless to be learned by considering this view. We’ll need to begin with a review of how van Fraassen conceives of this kind of discussion. I. WHAT IS CONSTRUCTIVE EMPIRICISM? The Terms of Engagement: Van Fraassen’s Voluntarism Van Fraassen rejects realism, for the moment briefly described as the view that science aims to give us a literally true story of what the world is like, in favor of constructive empiricism (CE), summarized as the view that the objective of science, so far as belief is concerned, is limited to providing theories which are empirically adequate. I and many others (see, e.g., Giere, IS 85; Musgrave, IS 205–206; and Leeds, 1994, 199ff.) 2 have read van Fraassen as arguing for CE, and, consequently, against realism. In turn we assumed that to argue for any such position was to attempt to offer rationally compelling reasons for adopting the view in question. But van Fraassen takes the constraints of rationality to be much weaker than do most of the rest of us. He refers to a distinction between two conceptions of rationality which . . . is analogous to that between . . . the Prussian and English concept of law. In the former, everything is forbidden which is not explicitly permitted, and in the later everything permitted that is not explicitly forbidden. When Russell is still preoccupied with reasons and justification, he heeds the call of what we might Philosophical Studies 106: 123–150, 2001. © 2001 Kluwer Academic Publishers. Printed in the Netherlands.