An Empiricist’s Guide to Objective Modality
Modality is a sticking point for empiricists. Some of them (e.g., Ladyman 2004) recognize that science comes with a heavy dose of modal commitment—a whole slew of beliefs not only about how things are, but how they might have been, could have been, or would have been had things been otherwise—and accept modality on the strength of their commitment to science. But others shun it on the grounds that talk of non-actual possibilities is epistemologically and metaphysically suspect.1 If we take the semantics of modal belief at face value, moreover, it seems that we are committed to the existence of non-actual possible worlds, and it has never been clear what these are, or how we could know about them. So understanding modality is a matter of some urgency for philosophers of science. On the one hand, there is a near-universal recognition that modality is central to science. Science cares not only about the pattern of actual events but also what they reveal about the modal substructure behind the phenomena. On the other hand, there is among many a desire to reject metaphysical commitment to non-actual, possible worlds. To satisfy that desire while preserving scientific practice, one needs an account of modality that makes modality immanent in the actual world, i.e., one that does not take the semantics of modal belief at face value. 
Early attempts to remove metaphysical commitment took the form of attempts to reduce modal concepts to non-modal ones. Views of this form hold that modal concepts do not add anything to a description of the pattern of actual events. Such views run into difficulties, however getting the contents of modal belief correct. In this paper, I defend an empiricist account of modality that keeps a substantive account of modal commitment, but throws out the metaphysics. I suggest that if we pair a deflationary attitude toward representation with a substantive account of how scientific models are constructed and put to use, the result is an account that deflates the metaphysics of modal commitment without deflating the content of modal claims.
In section 2, I review the recalcitrant stumbling block for reduction. In section 3, I look at the function of models and the practical and epistemic role that modal structures play. In section 4, I introduce a generic account of modal structures as partially prepared solutions to frequently encountered problems (PPS’s to FEP’s). In section 5, I compare my account of laws and chances with the account that comes out of David Lewis’s Best Systems Analysis. In section 6, I suggest that the account avoids the pitfalls of both reification and reduction.
-  Van Fraassen takes rejection of modality to be one of the defining features of empiricism: “To be an empiricist isto withhold belief in anything that goes beyond the actual, observable phenomena and to recognize no objectivemodality in nature” (1980, 202; emphasis mine).
-  I use “structure” in a generic way to refer to any property or relation. For examples of discussion in this mold, seePrice (2011) and Kment (2014). I would not follow Kment in calling it an external standpoint, because it will itselfuse modal concepts. I explicitly deny the possibility of stepping outside the practice, as though we could describeor conceptualize or come to understand anything without using modal notions. The idea is rather to take a side-onview of the practice in which these concepts arise that is internal to a fully articulated scientific picture of the world.Doing this kind of genealogy for everyday modal belief is more difficult, since the epistemology of everyday belief isless explicit and systematic, and since the logic of common sense is less regimented than that of scientific discourse.