In this talk I focus on two ways that explain salient differences in epistemic accessibility: Bas C. van Fraassen’s ‘observable/unobservable’ distinction and Azzouni’s ‘thin/thick access’ distinction. For both, exercising an ability entails satisfaction of an epistemic burden. I argue Azzouni’s thick/thin access distinction provides an alternative approach wherein detection and credibly inferred manipulation are also earned abilities, so although these abilities do not satisfy the epistemic burden for tables and chairs, they do satisfy epistemic burdens for scientific posits. This provides an avenue for articulating a potential truth tracking ability beyond observability or direct manipulatability. Therefore, the anti/realist and realist agrees that we have abilities that track truth about hands, tables and chairs; however, we have different abilities in regards to scientific posits. Specifically, the antirealist doubts that our abilities track the truth of scientific posits while the realist affirms that this ability can, in certain contexts, track truth. This approach helps clarify how scientific posits, although they are not observable, are open to satisfying epistemic burdens. This approach reframes the realism/anti-realism debate by speaking of satisfaction of epistemic burdens in a way that does not beg the question over whether belief in scientific posits is ever justified. I conclude with listing the positive and negative aspects of both approaches, whether they account for paradigmatic examples of each kind or how they classify edge cases, and the preservation of intuitions over members of kinds.
A revised version of this talk was given at the Advanced Research Seminar at King’s College London on Oct 21, 2015.