Abstract:

According to the new orthodoxy, a scientific theory is a class of models (i.e. the "semantic view of theories" championed by Suppe, van Fraassen, Ladyman, et al.). I show that this account misidentifies relations between theories -- such as when two theories are equivalent, or when one theory is embeddable in another. I then outline a more adequate view, and discuss open technical problems that such a view suggests.