Abstract

The ontologies of scientific theories include a variety of objects: point-mass particles, rigid rods, frictionless planes, flat and curved spacetimes, perfectly spherical planets, continuous fluids, ideal gases, nonidentical but indistinguishable electrons, atoms, quarks and gluons, strong and weak nuclear forces, ideally rational agents, and so on. But the scientific community currently regards only some of these objects as real. According to Paul Teller, a group sometimes can be justified in regarding competing ontologies as real and the ontologies we are justified in regarding as real are inexact, because the theories that give those ontologies characterize what things are like rather than what they are. In this paper, I argue that Teller's view is incomplete and suggest that one way to remove this incompleteness is to adopt a criterion for when we are justified in regarding a theory's ontology as real that is based upon a theory's comparative degree of confirmation. I argue that this criterion is prima-facie plausible and that Teller's view is false if this criterion is correct.