Cartesian Skepticism

Cartesian skepticism is the problem of explaining how knowledge of (or justified belief about) the external world is possible given the challenge that we cannot know (or justifiably believe) the denials of skeptical hypotheses. The problem has its source in Rene Descartes’ Meditations on First Philosophy, and in particular, the First Meditation. In general, a Cartesian skeptical hypothesis is a possibility which is incompatible with the truth of what we believe about the world (or incompatible with what we believe being knowledge), but which is indistinguishable from what we take to be our normal circumstances, where our beliefs are not systematically false (or systematically unknowable). For example, in the First Meditation, Descartes considers the hypothesis that there is a powerful evil demon who renders his beliefs about the world false, while making it seem to him just as if they are true. The challenge Descartes raises is: how can we know that the evil demon hypothesis is false, if such a scenario is indistinguishable from what we take to be our actual scenario? Skepticism about the external world, then, is the thesis that knowledge of (or justified belief) about the external world is impossible. And a proponent of this form of skepticism is a Cartesian skeptic if they appeal to skeptical hypotheses in order to show that we cannot know (or justifiably believe) anything about the external world. The Cartesian skeptical argument is often presented as follows: (1) if we know that a proposition about the external world P is true, then we know the denial of the skeptical hypothesis SH. But (2) we do not (or cannot) know the denial of SH. Therefore, (3) we do not (or cannot) know that P. We can organize the literature on Cartesian skepticism around two central areas of focus: (a) the nature and structure of Cartesian skeptical arguments, and (b) responses to the Cartesian skeptical argument. Epistemologists are divided on the nature of the Cartesian skeptical argument. Proponents of the closure-principle formulation of the Cartesian skeptical argument maintain that the argument depends on a suitable closure principle for knowledge or justification. Proponents of the underdetermination-principle formulation maintain that the argument depends on a suitable underdetermination principle. Responses to the Cartesian skeptical argument can be divided into those which maintain that premise (1) is defective and those which maintain that premise (2) is defective. Moorean's reject premise (2), and argue that since we can know that P (e.g., I have two hands), and we know that P implies that SH is false, we are in a position to know that SH is false. Externalist Moorean’s reject premise (2), and argue that we can know that P if our believing that P is the product of a reliable belief-forming process (reliabilism), or if our belief that P is safe (safety theories). We can thus make a knowledge-preserving deduction from P to the denial of SH. Knowledge-first varieties reject premise (2), and argue that our evidence for P in our actual case is different from our evidence in the case in which SH is true, because our evidence is what we know, and in the actual case, we know that P, whereas, in SH, we do not know that P. Internalist Moorean's reject premise (2), and argue that our perceptual experiences provide us with knowledge (or justification) to believe P, and so deduce ~SH. Dogmatist varieties argue that these perceptual experiences alone provide us with prima facie justification to believe P, even though the strength of this justification is the same in the actual case as it is in the case in which SH is true. Epistemological Disjunctivist’s argue that the strength of the epistemic justification is different in the actual case than the case in which SH is true, and that the strength of our justification in our actual case is factive and accessible to us. A priorist’s reject premise (2), and argue that we can know that SH is false, even though this is not grounded in empirical evidence or reasons. Entitlement varieties ague that we have a default, non-evidential entitlement to reject SH. A priori argument varieties argue that there is an a priori argument for the conclusion that SH is false (e.g., from considerations about meaning or content). Explanationist’s reject premise (2), and argue that SH fails to explain our experiential evidence and beliefs better than what we take to be our actual scenario. Tracking theorist’s reject premise (1), and argue that our belief that P can track its truth even though our belief that SH is false fails to track its truth. Relevant alternatives theorist’s argue that SH is not a relevant epistemic alternative to P, so that even though our evidence is not sufficient for rejecting SH, it is sufficient for P. Contextualist’s argue that the truth-conditions of our knowledge-ascriptions are sensitive to context, allowing that, in ordinary contexts, ascriptions of “S knows that P” can be true even though, in skeptical contexts, ascriptions of “S knows that P” are false.

For the original presentation of Cartesian skepticism and the Cartesian skeptical argument, see Descartes 1996. For work on the nature of the Cartesian skeptical argument, see Unger 1975 , Nozick 1981, Stroud 1984, Williams 1991, and Pryor 2000. For work on closure-based and underdetermination-based formulations of the argument, see Yalçin 1992, Brueckner 1994, Cohen 1998, Vogel 2004, and Pritchard 2005. A classic response to Cartesian skepticism is Moore 1959. For Moorean responses from epistemic externalism, see Hill 1996, Sosa 1999, Greco 2007, and Pritchard 2005. For knowledge-first variants, see Williamson 2000. For dogmatist responses, see Pryor 2000, and Huemer 2000. For epistemological disjunctivist responses, see McDowell 2008 and Pritchard 2012. For explanationist responses, see Vogel 2005 and Vogel 1990. For entitlement responses, see Wright 2004. For a priori argument responses, see Kant 1998, Putnam 1981, and Davidson 1989. For truth-tracking responses, see Nozick 1981, and Zalabardo 2012. For relevant alternatives responses, see Dretske 1970 and Stine 1976. For contextualist responses, see Cohen 2000, Lewis 1996, and DeRose 1995.