Structural realists claim that structure is preserved across instances of radical theory change, and that this preservation provides an argument in favor of realism about structure. In this paper, I use the shift from Newtonian gravity to Einstein’s general relativity as a case study for structural preservation, and I demonstrate that two prominent views of structural preservation fail to provide a solid basis for realism about structure. The case study demonstrates that (i) structural realists must be epistemically precise about the (...) concrete structure that is being preserved, and (ii) they must provide a metaphysical account of how structure is preserved through re-interpretation in light of a new theory. Regarding (i), I describe a means of epistemic access to the unobservable that I call “thick detection” of structure, which isolates the structure that will be preserved. Regarding (ii), I argue that thickly detectable structure is preserved across theory change through a process of extracting the old structure from the new structure, much like what has been done with geometrized versions of Newtonian gravity. With these two responses in hand, the structural realist can adequately account for the preservation of structure and can provide a strong argument in favor of structural realism. (shrink)

‘Shallow’ and ‘deep’ versions of scientific realism may be distinguished as follows: the shallow realist is satisfied with belief in the existence of the posits of our best scientific theories; by contrast, deep realists claim that realism can be legitimate only if such entities are described in metaphysical terms. We argue that this methodological discussion can be fruitfully applied in Everettian quantum mechanics, specifically on the debate concerning the existence of worlds and the recent dispute between Everettian actualism and quantum (...) modal realism. After presenting what is involved in such dispute, we point to a dilemma for realists: either we don’t have the available metaphysical tools to answer the deep realist’s demands, and realism is not justified in this case, or such demands of metaphysical dressing are not mandatory for scientific realism, and deep versions of realism are not really required. (shrink)