Invariance and Scientific Representation

This paper argues for an invariantist view of scientific representation. First, it introduces the deflationary view that sees models in science as no different in essence from symbolic vehicles, which are derivative and adopted pragmatically, as a matter of convention. After analyzing this deflationary view and pointing out its shortcomings, it argues that representations play at least two radically different roles: one purely symbolic and therefore conventional, and the other epistemic. Finally, it argues that although it is correct to say that all particular external vehicles are introduced via some conventions, they get their jobs done because they are invariant with respect to particular conventions.