Reliabilism and the Generality Problem

In the post-Gettier era, process reliabilism is one of the most influential epistemic theories on offer. According to process reliabilism, a belief counts as knowledge only if that belief was arrived at by use of a reliable belief-forming process. This reliable process condition is the central constituent in both reliabilist accounts of warrant and justification. Over the course of the past three decades, the generality problem has been seen as one of the most pressing objections to process reliabilism. Roughly, the generality problem begins with the observation that we currently understand very little about the relevant belief forming process types that are measured for reliability for any given belief forming process token. According to the generality problem objection, there is some normative burden for the reliabilist to supply some account of type relevance. This dissertation is an extended investigation into the generality problem—what it is, and what the reliabilist can say in response.

For all of the attention given to the generality problem, there’s been no attempt to systematically formulate how the objection to reliabilism is supposed to go. In chapter one, I present what I take to be the four most reasonable approaches to formulating the generality problem objection, and find all of them subject to compelling objections. In chapter two, I argue that if the reliabilist has some special burden to produce a theory of type relevance, then so do theorists of almost every other competitor epistemic theory. Hence, the generality problem fails to constitute some unique objection to reliabilism.

In chapter three, I criticize recent attempts to ground type relevance in subjective factors like a subject’s higher order beliefs, or a subject’s practical interests. In chapter four, I argue that relevance theories that merely take into account causal features of the tokens aren’t extensionally correct. Rather, extensionally correct relevance theories must make relevant types a function of the token’s modal properties. Finally, in chapter five, I present relevance principles for arithmetical intuition belief formation while highlighting two neglected aspects for crafting a correct relevance theory.