Each modern physical theory posits a collection of mathematical models that represent a range of physical phenomena or states of affairs within that theory’s scope. Usually, each model of a theory represents some way or ways for that phenomena or state of affairs to be—its representational capacities. This observation alone raises several general issues for modality in physical theory. One quite general question that arises in the interpretation of a physical theory is to what modal facts one commits when one commits to such a theory. Most discussions of the modal commitments of a physical theory assume a binary classification of putative models—possible or not—and that explanations involving that theory use only a single model of the theory representing the phenomena explained. The widespread use of variational principles across physics seems to challenge both of these convictions.