1Davidson's "A Coherence Theory of Truth and Knowledge," Harman's "Thought" and "Change in View," and BonJour's "The Structure of Empirical Knowledge," are representative examples of coherentist approaches.
2See Cohen's "How to be a Fallibilist," DeRose's "Solving the Skeptical Problem," and Lewis' "Elusive Knowledge" for examples of differing contextualist theories. Further, Pritchard's "Two Forms of Epistemological Contextualism," and Pryor's "Highlights of Recent Epistemology" survey the landscape of contextualism.
3For a discussion of what counts as a legitimate objection under contextualism, see the discussion on the Default and Challenge approach to justification in the section responding to DeRose and Black in the third chapter, page 92 and 93.
4It is worth noting that economic factors do not necessarily pertain to monetary costs and benefits.
5Perhaps it is the case that DeRose intends for 'sense experience' and 'perception' to mean different things, but without an explicit account of what that difference is it is hard to say anything about it. DeRose seems to equate sensory experiences with perceptions, by saying that we do not have any sensory experiences, but rather, all perceptual beliefs are caused directly by neural events; that is to say, our sensory experience beliefs are not produced by outside stimulus, but only by neural events; it does not seem to make sense to say our sensory beliefs are not caused by stimulus, but rather perceptual beliefs are caused by neural events, unless perceptual beliefs just are sensory experience beliefs. 'Sensory experiences' and 'perception' seem to be equated. Yet he also says that we take ourselves to know what we believe through perception. If perception is just sense experience, then we have a straightforward contradiction; we do not believe that we have sensory experiences, yet we take ourselves to know through sensory experiences. If DeRose intends by 'perception' something other than what he means by 'sense experience' it would seem that it would have to mean something like "all experiences that are not of the sensory sort;" if the premises are to not contradict one another, 'perception' would have to be the negation of whatever 'sense experience' is. It is of course very difficult to say what sort of experiences these would be; would they be perceptions of memory, ideas, thoughts, or the like? This seems strange. Perhaps DeRose means by "perception" direct knowing of the world, and means by "sense experience" knowing the world by experiences taken as intermediaries. It is hard to say exactly what knowing the world directly entails. Again, without an explicit account of what the difference between perceptions and sensory experiences is, it is very difficult to analyse. Even so, if "perception" is to mean direct apprehension of the world, then it seems a brain-in-a-vat argument would not be effective; for if we take ourselves to know the world directly, then a brain-in-a-vat argument, which relies on us having sensory experiences of the world that are intermediary between us and our knowledge of the world, would not be persuade the knowers that DeRose asks us to imagine. The knowers DeRose asks us to imagine would not be convinced by brain-in-a-vat arguments, because they would take themselves to know the world directly, and to know, full stop, that they are not brains in vats.
6Descartes does make use of arguments from error in his formulation of scepticism. We might make the argument that, if we have made errors in the past, then we could be making errors all the time and not know it. But this seems to go beyond the argument from error, because it adds a premise that says we do not ever know whether or not we are making an error. The possibility of error does not rule out the possibility of coming to know that we were in error, and does not rule out the possibility of coming to know aright, unless we add a premise that says that we can never know whether or not we are in error, such as a premise that says we do not know whether or not we are dreaming, or a premise that says all cases of error look like cases of being correct. Further Descartes, in his skeptical argument, does progress from cases of error to cases of dreaming or insanity, so that it seems as if he builds upon familiar error situations to make his skeptical conclusion. But this does not mean that such a progression is necessary for skeptical argumentation; for it seems perfectly plausible that we may go straight to the possibility of dreaming, without making mention of common error. The distinction between common sources of error and genuinely skeptical premises, such as dream-possibilities, seems to be a legitimate one, even if Descartes did not seem to make it.
7The third portion of the Trilemma is infinite regress, but as we have seen, we quickly run out of creative answers when responding to the sceptic's challenges. It seems reasonable to suppose that once the challenge has been moved to an epistemological context, we run out of novel support, and are forced into assumption or circularity.
8Of course we are given reason to believe this in a particular context. The topic of why we should or should not accept this context over skeptical contexts is discussed in the next sections, in response to McGinn and Skorupski.
9Skorupski asks us to suppose that there is a function that determines the degree of justification of a given proposition for all contexts. It could be the case, given Williams' contextualist position on content, that context and content of the given proposition are variables that determine the degree of justification of the proposition; even if Williams is correct about propositions having no content outside of any context, it may still be the case that propositions do have content in a context, and that this content at least partially determines the proposition's epistemic status. The sceptic's point would be that no proposition would ever be justified given this function. It is doubtful that Williams would allow that such a function could ever exist under contextualism (Williams, Problems of Knowledge 166). For if we allow that content is an independent variable in the function that determines degree of justification, then we have just allowed that content is an objective, independent determinant in degree of justification for a proposition; and this is of course foundationalism. If we allow that content and context determine degree of justification, it is hard to see how this is not contextualism; if degree of justification depends on context, so that content does not always confer the same degree of justification on a proposition, then it seems we are just saying that context determines the degree to which content determines justification, which is to say that context determines justification, which is contextualism. The only way to truly allow that content independently determines degree of justification is to adopt a foundationalist epistemology; we cannot have it both ways, so that content and context contribute to justification while still being contextualists about justification. But for the sake of argument it is granted.
10Foundationalists would of course disagree; they may hold that there are objective facts about what norms are correct. We look at a such theories in the second last section of this chapter, which is focused on Graham.
11Graham abbreviates each principle, so that a priori insight is AP, Memory is MEM, and so on. These methods of belief forming are fairly well-known, but for an account of each, see Graham, pages 3 and 4.
12This second premise is required because the possibility of error alone is just the argument from error, which does not rule out the possibility of overcoming error, and so is not radically sceptical.