Abstract

It is widely believed that science is in the business of finding out what the world is really like. The philosophical version of this belief is scientific realism -- a doctrine about science that tells us that we ought to believe that the best theories in science are true, and that the world is occupied with the objects that those theories contain. If scientific realism was not correct, the argument goes, the incredible success of science would be a miracle. The best explanation for the success of science however, is not that it is a miracle, but that scientific theories are true. This argument is an instance of inference to the best explanation, or IBE. Skeptics have questioned why scientific success must imply its truth given that there are so many abandoned, false scientific theories in the history of science that were nevertheless successful. One of the controversies in the debate between scientific realists and anti-realists surrounds the legitimacy of reasoning in accordance with IBE. Realists need IBE to be a justified and reliable guide to truth. In this compilation thesis, I address various questions related to IBE and scientific realism. Paper 1 argues that scientific realism without IBE loses too much of its epistemic optimism, and that it in some contexts even becomes more pessimistic than the most prominent rival philosophical doctrine about science -- constructive empiricism. To avoid deflating realism, I argue, a defense of IBE is necessary. Paper 2 addresses whether methodological similarities between science and metaphysics force scientific realists to also be realists with respect to metaphysics. If IBE is legitimate, it should not only be valid in science, but also in metaphysics, effectively inflating the ontology that scientific realists are rationally bound to accept. I argue against this conclusion. Paper 3 offers a proof of concept regarding a novel way to justify inferences to unobservable objects. Paper 4 establishes a novel critique of non-probabilistic versions of IBE in scientific realism.