This thesis has two aims. One is to motivate the claim that challenging what I call a “sameness of evidence thesis” is a particularly promising approach to external world scepticism. The other is to sharpen an underexplored issue that arises when challenging the sameness of evidence thesis. The second aim is the primary aim of the thesis. Pursuing the first aim, I start by examining a predominant formulation of external world scepticism known as the “closure argument” for knowledge. I examine three main strategies for responding to external world scepticism and highlight their major challenges (DeRose 1995; Dretske 1979; Nozick 1981; Sosa 1999). The goal is not to demonstrate that these challenges cannot be met, but rather to highlight a deeper issue that arises when responding to the closure problem for knowledge. In particular, I take the discussion to motivate looking at what I will call “scepticism about evidential justification” (Feldman 2000; Kornblith 2001; Pritchard forthcoming). The general argument in favour of a shift to scepticism about evidential justification is based on considerations about what an adequate response to external world scepticism should hope to achieve. I argue that one condition of adequacy is being able to account for radical forms of scepticism that challenge not only that our beliefs enjoy the epistemic status of knowledge (however that status is conceived) but also that our ordinary empirical beliefs are justified, or that we are reasonable in holding them. There are different varieties of scepticism about evidential justification. I focus in some detail on the anti-sceptical strategies of Pryor (2000; 2004) and Wright (2004) as examples of strategies that engage with scepticism about evidential justification. But I argue that one form of evidential scepticism known as the “underdetermination argument”—which Pryor and Wright do not directly engage with—is of particular importance. The main assumption in the underdetermination argument I focus on is about the nature of evidence. More specifically, the underdetermination argument presupposes that one’s evidence is the same in so-called “bad” and “good” cases in which an agent forms an empirical belief. This is the “sameness of evidence thesis.” Pursuing the main aim of the thesis, I introduce two forms an anti-sceptical strategy that involves challenging the sameness of evidence thesis. The two forms I consider differ in their commitments concerning a condition of accessibility on our evidence. Pritchard (2006; 2007; 2012; forthcoming) maintains that one’s evidence is “reflectively accessible.” Williamson (2000; 2009) rejects this claim. The central issue I aim to sharpen is that while accepting the condition of accessibility leads to serious challenges in rejecting the sameness of evidence thesis, rejecting it leads to counterintuitive consequences if we grant that there is a normative principle that requires us to proportion our beliefs to the evidence. A central part of the thesis involves examining these counterintuitive consequences and showing what accounting for them requires. This is an underexplored project in the context of external world scepticism. I look at three different approaches to spelling out the counterintuitive consequences. My preferred account turns on a distinction between three different kinds of responsibility (Shoemaker 2011). I claim that there is a notion of responsibility – “attributability” – that is centrally connected to normative judgments. I argue for a “condition of accessibility” on attributability. Taken together, these two claims comprise an account of what is problematic about rejecting an access condition on our evidence. I then claim that there are two ways forward. One is to accept the condition of accessibility on our evidence that my account implies; the other is to challenge my claims about the connection between attributability and normative judgments, or the accessibility condition on attributability, or both. Although I claim that the prospects look better for taking the second option when it comes to rejecting the sameness of evidence thesis, drawing on recent work from Gibbons (2006; 2013) and Daniel Greco (2013), I argue that the first option is still a live possibility. The main aim in this part of the thesis is not to decide what the best way of rejecting the sameness of evidence thesis is, but rather to examine the challenges that arise when we reject it in one way or another. The question of what sort of access we have to our normative requirements is the focus of an increasingly sophisticated discussion in contemporary epistemology. An important upshot of this thesis is that it brings the problem of external world scepticism directly within the scope of that debate.

In this thesis, I explore the issue of epistemic responsibility. I start by examining an argument against the use of critical thinking made by Michael Huemer. Huemer argues that critical thinking is not epistemically responsible, because it is not as truth conducive as credulity. Huemer instead argues that credulity should be the default approach taken by non-experts. After dissecting this argument, I go on to examine one of the critics to Huemer’s argument: David Kary. Kary argues that critical thinking and credulity are not mutually exclusive and can therefore be combined in an epistemically responsible way. Kary further argues that one must consider the social components of epistemic responsibility, and that when one does so, it is evident that truth conduciveness is not the only component of epistemic responsibility. I extend Huemer and Kary’s discussion by arguing that epistemic responsibility is even more complex. Epistemic superiority, equality and inferiority are fluid positions that change depending on context, and this must be considered when evaluating epistemic responsibility. The consequence of this approach is that a combination of critical thinking and credulity is the most responsible alternative. I end by arguing for the intellectual virtues and benefits of embracing this argument.