Before we can evaluate specific accounts of self-knowledge, we need to understand the claim of “specialness” made on its behalf. This chapter takes up two questions: What is (purportedly) special about self-knowledge? And what kinds of self-knowledge are thought to be special in this way? Section 3.2 addresses disparate assessments of what is special about selfknowledge. These include epistemic assessments, according to which first-person access to one’s own states is epistemically privileged; and nonepistemic assessments, according to which self-attributions qualify as special on nonepistemic grounds. (Those who construe the basic asymmetry between self-knowledge and other sorts of knowledge as nonepistemic may allow that there is also an epistemic difference between these-unless, that is, they are skeptical about self-knowledge.) Section 3.3 focuses on the idea that self-knowledge is epistemically

special in that we have privileged access to some of our own mental states. It presents a range of empirical and philosophical considerations that exclude a broad variety of mental states from the domain of privileged first-person access. Section 3.4 surveys the types of mental states that have not been excluded, and uses the results of this discussion to define the parameters of the succeeding chapters.