Structural realism and theory classification

Abstract Ontic structural realism constitutes a promising take on scientific realism, one that avoids the well‐known issues that realist stances have with underdetermination and theory change. In its most radical versions, ontic structural realism proposes a type of eliminativism about theoretical entities, ascribing ontological commitment only to the structures, and not to the objects appearing in our theories. More moderate versions of ontic structural realism have also been proposed, allowing for ‘thin’ objects in the ontology. This work connects these takes on structural realism with the independent notion that there are two categorically different kinds of scientific theories, namely those that deal with interactions and those that deal with structural constraints – what are known as interaction and framework theories, respectively. By taking this classification seriously, one can arrive at a selective version of ontic structural realism which is better adapted to our scientific knowledge, in which eliminativist ontic structural realism constitutes a natural fit to framework theories, whereas the moderate version is connected to interaction theories.