Blake Roeber, Northern Illinois University
Contextualists maintain that the truth-values of knowledge attributions vary from conversational context to conversational context. In one conversational context, a knowledge attribution may express a true proposition; in another conversational context, the same attribution may express a false proposition. 1 Almost invariably, contextualists defend their position as necessary for preserving our intuitions in the face of the so-called “skeptical paradox.” 2 In addition to DeRose 1995, see Cohen 1988 and Lewis 1996.[/note] Contextualists often proceed, not by appealing to linguistic data, but by arguing that contextualism uniquely preserves our commonsense belief that we know a lot andour philosophical belief that skeptical hypotheses destroy knowledge andour unqualified commitment to closure. 3As its defenders make clear, contextualism is a linguistic thesis. Rather than a theory about knowledge, it is a theory about the word “knows.” 4 In this paper, I attempt to undermine the case for contextualism by showing how an epistemic thesis—that is, a theory of knowledge—might preserve our intuitions where contextualism cannot. In what follows, I outline the contextualist solution to a familiar skeptical paradox. After highlighting the intuitions at stake, I present several compelling objections leveled against contextualism by Kent Bach. I conclude by pointing the way to a theory of knowledge that might preserve our intuitions where contextualism fails. My contention in this paper is neither that this theory of knowledge is true nor that contextualism is false. Here, I argue simply that contextualism cannot preserve the intuitions it claims to preserve and that a properly Chisholmed theory of knowledge might succeed where contextualism fails. 5
- The Skeptical Paradox
Where “N” is a non-controversial proposition about the world (e.g., that Jones has hands) and “H” is a well-tailored skeptical hypothesis with respect to N, (e.g., that Jones is a bodiless BIV), 6 the following sentences confront us with a paradox:
- (1) Jones knows N.
- (2) If Jones knows N, and Jones knows “N” entails “∼H,” then Jones knows
- (3) Jones does not know ∼H.
These sentences confront us with a paradox for the following reasons. Normal people seem to know a lot, there are few clearer examples of the things normal people seem to know than that they have hands, and Jones is a normal person. So (1) seems true. At the same time, Jones doesn’t seem to know ~H. After all, how could she? She can’t point to a single experience she wouldn’t be having ifH were true. So, nothing in her experience is evidence for ∼H. So (3) seems true. But (2) is true if Jones knows that “N” entails “∼H” and knowledge is closed under known entailment. We’re assuming the former and it is exceedingly difficult to deny the latter, so (2) seems true, too. Yet, assuming the second conjunct in (2)’s antecedent, (1) and (2) entail ∼(3), and (2) and (3) entail ∼(1). So at least one of (1), (2), and (3) has got to go. But denying any of them requires rejecting a strong intuition.
II. The Contextualist Solution
Contextualists embrace (2) across contexts. The conjunction of (1) and (3) in a single context is abominable, however. 7 Contextualists resolve the tension between (1) and (3) via the claim that knowledge attributions express different propositions in different contexts. According to contextualists, (1) expresses a true proposition in everyday, practical contexts, but it expresses a false proposition in philosophical contexts, where skeptical hypotheses are salient. It follows that our intuition favoring (3) is preserved in philosophical contexts, while our intuition favoring (1) is preserved in non-philosophical contexts. So, contextualism preserves our unqualified intuitions about closure and our philosophical intuitions about skeptical hypotheses and our commonsense intuition that Jones knows she has hands. Or so the claim goes.
III. Contextualism and “Changing the Subject”
According to Kent Bach, contextualism fails to address “full-blooded” skepticism. 8 Skeptics draw our attention to H, says Bach, in order to demonstrate that, even according to normal, everyday standards, Jones’s evidence for N isn’t sufficient for Jones to know N. Skeptics don’t argue for the conclusion that there are conversational contexts in which (1) expresses a false proposition. Rather, skeptics argue that Jones’s having sufficient evidence to rule out every ~N alternative is a necessary condition for Jones to know that N. And since the evidence Jones possesses does not vary from conversational context to conversational context, skeptics conclude that, as a matter of context-independent fact, Jones does not know that N. Of course, that Jones doesn’t know that N entails that (1) is false. But the skeptic’s point is fundamentally epistemic, not linguistic. As Bach puts it, “skeptics are concerned with knowledge, not [the word] “knowledge.” 9
The same can be said about “full-blooded” commonsense. Mooreans are also “concerned with knowledge, not [the word] ‘knowledge.’ 10 When Mooreans argue from (1) and (2) to the negation of (3), they aren’t arguing for the conclusion that there are conversational contexts in which (3) expresses a false proposition. Mooreans draw our attention to the irresistibility of N to demonstrate that, in conjunction with the closure principle, Jones has overwhelming evidence for ~H. Since the evidence Jones has does not vary from context to context, Mooreans conclude that there is a context-independent fact of the matter whether Jones knows that she is not a BIV, and they argue that, as a matter of context-independent fact, not only does Jones know that she has hands, she knows that she isn’t a BIV. That Jones knows she isn’t a BIV entails that (3) is false, but, again, the Moorean’s point is epistemic, not linguistic. Like the skeptic, the Moorean is also concerned with knowledge, not the word “knows.”
The explicitly epistemological concerns of skeptics and Mooreans highlight another counter- intuitive feature of contextualism. Suppose Smith, a commonsense philosopher, is sitting at the bar. At t1, Smith tells the bartender, “Jones knows she has hands.” A nearby skeptic hears Smith’s claim and, at t2, confronts Smith with an argument involving H. According to contextualists, as soon as our skeptic mentions H, the context changes and the standards shift from practical to philosophical. At tt, (1) was true, and now, at t2, (3) is true. The conjunction of (2) and (3) entails ~(1), however. So, in spite of the fact that (1) was true at t1, (1) is false at t2. The skeptic concludes, “It’s not the case that Jones knows she has hands,” and what the skeptic says is true.
According to contextualists, “Jones knows she has hands” expresses different propositions at t1 andt2. But if “Jones knows she has hands” expressed different propositions at t1and t2, it follows that the skeptic’s conclusion at t2 didn’t contradict Smith’s claim at t1. The skeptic changed the subject on Smith, says the contextualist. Smith and the skeptic took their claims to be contradictory, but they only did so because neither of them realized that the skeptic changed the subject. 11
Contextualism’s Failure to Preserve Our Intuitions
Section III highlights four ways contextualism fails to preserve our intuitions. First, we have the intuition that, given Jones’s lack of evidence with respect to ~H, as a matter of context- independent fact, Jones doesn’t know N. So, there is no context in which “Jones knows N” is true. Call this intuition “the skeptical intuition.” Because contextualism neither denies that Jones knows N nor denies that there are contexts in which “Jones knows N” is true, it is unclear how contextualism preserves the skeptical intuition. Second, we have the intuition that, given the irresistibility of N, as a matter of context-independent fact, not only does Jones know N, Jones knows ~H. That is, we have the intuition that N is irresistible in every context, and that there is therefore no context in which “Jones knows ~H” is false. Call this intuition “the Moorean intuition.” Because contextualism neither maintains that Jones knows ~H nor denies the existence of contexts in which “Jones knows ~H” is false, it is not clear how contextualism preserves the Moorean intuition. Third, we have the intuition that, when Mooreans assert (1) and skeptics counter by arguing from (2) and (3) to the negation of (1), skeptics actually succeed in contradicting Mooreans. That is, we have the intuition that skeptics—meaning to disagree with Mooreans—succeed in disagreeing with Mooreans. Call this intuition “the disagreement intuition.” Because contextualism claims that “Jones knows she has hands” expressed a different proposition at t1 than it expressed at t2, contextualism amounts to an all-out denial of the disagreement intuition. Fourth, we have the intuition that skeptics are making a monumentally important claim; if they are right, then Jones ought to give up most of her beliefs. Reflecting the fact that people do epistemology at least in part because they worry that skeptics might be right, call this intuition “the epistemological intuition.” Because contextualism denies that the truth-value of (3) in philosophical contexts has any bearing on the truth-value of (1) in everyday contexts, contextualism denies that skepticism is relevant outside contexts in which skeptical hypotheses are raised. But this denial trivializes the epistemological intuition. As DeRose concedes, contextualism “might lead us to believe that traditional inquiry into skepticism has not been concerned with a very important question.” 12
Philosophers have debated skepticism for thousands of years. 13 Without the skeptical intuition at work, there would never have evolved any alternative to the Moorean position, and without the Moorean intuition at work, there would never have evolved any alternative to the skeptical position. Moreover, without the disagreement and epistemological intuitions at work, skeptics and Mooreans would never have bothered to debate one another. The skeptical, Moorean, disagreement and epistemological intuitions are widespread, deeply held, and ignored by contextualism. Whatever contextualists claim about their ability to resolve skeptical paradoxes, contextualism mishandles these strong intuitions.
How to Preserve Our Intuitions about Knowledge and “Knows”
The closure principle is so intuitive that (2) is nearly indubitable. Indeed, many philosophers think closure-failure amounts to a reductio ad absurdum on any theory that results in it. As a first step toward preserving our intuitions, then, we must choose a theory that preserves closure.
DeRose, Cohen, Sosa, and others acknowledge an asymmetry between the Moorean intuition and the skeptical intuition, and this asymmetry is the second thing we must account for. 14 In everyday contexts—when we’re eating breakfast, using the fax machine, etc.—the possibility that Jones is a bodiless BIV rarely enters our minds. When it does, we naturally dismiss it as absurd. 15 Without stopping to reflect on the matter—without putting our practical concerns aside and working ourselves into a philosophical frame of mind—(3) makes little purchase on our intuitions. Consequently, in everyday contexts, we rarely have any intuition favoring (3). Even in philosophical contexts, however—even contemplating the possibility that our senses are deceiving us—we find ourselves strongly inclined toward (1). As Dretske and Nozick make clear, 16 this inclination is so strong that we sometimes believe (1) even while admitting that we don’t know ~H. The intuition in favor of (1) transcends context. As a second step toward preserving our intuitions, then, we must pick a theory of knowledge that avoids skepticism in every context—a theory according to which Jones knowsN and does so across contexts.
Alvin Plantinga’s virtue theory is as easy to wield as any. 17As we shall see below, it preserves closure and doesn’t result in skepticism. (As an added bonus, it delivers intuitive results in Gettier situations. 18) According to Plantinga, S knows that p if and only if S believes p, p is true, and p was formed in S “by cognitive faculties functioning properly (subject to no dysfunction) in a cognitive environment that is appropriate for S’s kind of cognitive faculties, according to a design plan that is successfully aimed at truth.” 19
Any viable theory of knowledge will serve the purposes of this paper, 20so long as it preserves closure and allows ordinary people to know ordinary things. For the sake of argument, then, let’s assume Plantinga’s account is correct and that S knows p if and only if the following conditions obtain (readers not fond of Plantinga’s account are welcome to substitute their own):
BELIEF: S believes p.
TRUTH: p is true.
VIRTUE: S’s belief that p is the product of an intellectual virtue—a cognitive faculty
that is (a) properly functioning, (b) successfully truth-aimed, and (c) functioning in the maxi- and mini-environments for which it was designed.
We can call this theory “V”—for “virtue theory.”
Assuming that Jones is not a bodiless BIV, and that she is just the sort of person you and I
take ourselves to be, the belief and truth condition non-controversially obtain with respect to Jones and N. But what does it mean to say that the virtue condition obtains? The function of a cognitive faculty is the production of belief. So a successfully truth-aimed cognitive faculty functions properly just in case it produces mostly true belief. 21The cognitive faculty in question is Jones’s vision. 22 Because we are assuming Jones is a normal person, we can safely assume that this faculty is successfully truth-aimed, 23and we can safely assume that it is functioning properly. So, Jones’s belief meets conditions (a) and (b). 24 Jones isn’t a BIV, she’s in normal light on the planet earth, etc., so Jones’s faculty is functioning in the maxi-environment for which it was designed. Jones’s eyes are opened, nothing is obstructing her view, and she’s not in any kind of Gettier situation, so Jones’s faculty is also functioning in the mini-environment for which it was designed. 25 So, Jones’s belief meets condition (c), and it follows that, according to V, Jones knows that she has hands.
How does V handle closure, then? It succeeds in preserving it, even in the most difficult cases. Suppose Jones is driving through Fake Barn County, in which there are hundreds of fake redbarns and one real blue barn. 26 Not knowing she’s in Fake Barn County, Jones looks at the blue barn and forms the belief that there’s a blue barn in front of her. She then deduces that there’s a barn in front of her. Jones is in a Gettier situation with respect to her belief that there’s a barn there, so she doesn’t know there’s a barn there. But if V produces the result that she does know there’s a blue barn there, we get closure failure. 27 Fortunately, it doesn’t. The cognitive faculty responsible for Jones’s belief that there’s a blue barn in front of her isn’t functioning in the mini-environment for which it was designed, so condition (c) fails to obtain. Jones doesn’t know there’s a blue barn nearby, and closure is preserved.
So far, so good, with respect to our intuitions about (1) and (2). (If the reader thinks otherwise, she is more than welcome to replace V with any account she prefers.) We still have a problem, however. (1) and (2) entail ~(3) and, as we have seen, we have a strong intuition favoring (3). While V falls short here, we can buttress it with the work of Ernest Sosa.
Referring to the epistemological intuitions that give raise to skepticism, Sosa tells us “[s]uch intuitions reflect a long tradition and still demand their due.” 28 Giving these intuitions their due leads Sosa to distinguish between levels of knowledge. In addition to knowledge simpliciter, Sosa introduces reflective knowledge. 29 V is a variety of externalism. As such, it is subject to the standard internalist criticisms. 30 Adding reflective knowledge to V fills this gap. Following Sosa, we can say that S has reflective knowledge that p if and only if S knows that p (i.e., meets the belief, truth and virtue conditions, as specified above) and the following condition also obtains:
PERSPECTIVE: S knows that (q) the virtue condition obtains with respect to S’s belief that p. 31
Call the conjunction of V and reflective knowledge “VP”—for “virtue perspectivism.” 32According to VP, reflective knowledge is epistemically superior to knowledge simpliciter. This superiority bears on our intuitions about (3). According to (3), Jones does not know ~H. VP commits us to the conclusion that (3) is false, in every context, so VP can be accurately characterized as Moorean. But Jones is a normal person (we are assuming). She has never verified that the virtue condition obtains with respect to ~H. So, Jones lacks reflective knowledge of ~H. According to VP, Jones’s belief fails to manifest an important epistemic desideratum. And VP lets us emphasize this failure as much as we please. VP allows children and animals to have knowledge, 33 so, if we choose, we can emphasize what Jones lacksby deriding her knowledge as juvenile or animal. If we choose, we can write entire books praising reflective knowledge and blaming Jones for lacking it.
Moreover, if our skeptical sensibilities are still not satisfied, we can note that reflective knowledge doesn’t give us what we’re really after—certainty. Reflective knowledge can’t produce certainty because every instance of reflective knowledge rests ultimately on an item of unreflective knowledge. S has reflective knowledge that p only if S knows (reflectively or unreflectively) that q. If S knows unreflectively that q, then S’s reflective knowledge of p depends on unreflective knowledge of q. But suppose S has reflective knowledge of q, too. In this case, S has knowledge of some proposition (r) about the virtue condition with respect to q, and either r is known reflectively or unreflectively. If the latter, then S’s reflective knowledge of p depends on S’s unreflective knowledge of r. If the former, then S knows some proposition (s) about the virtue condition with respect to r, and either s is known reflectively or unreflectively . . . and so on ad infinitum. But since an infinite chain of known propositions is impossible, there is some proposition in S’s noetic structure with respect to p that S knows unreflectively, which means that S’s reflective knowledge rests ultimately on unreflective knowledge. So, even reflective knowledge that p falls short of certaintythat p.
For the sake of preserving our intuitions, then, let’s say the highest form of knowledge is fully reflective knowledge, which S has of p if and only if she has reflective knowledge of p and the following condition obtains:
CERTAINTY: S has reflective knowledge that (q*) the virtue condition obtains with respect to S’s belief that p, and S’s noetic structure with respect to q* contains no item of unreflective knowledge.
On the one hand, VP claims that fully reflective knowledge is the highest, most desirable kind of knowledge. On the other hand, VP admits that fully reflective knowledge isn’t attainable, since it requires a God’s-eye perspective. So, VP affords us resources for emphasizing the limits of human knowledge, and here we have a further means of preserving our intuitions about (3). We can say that Jones lacks reflective knowledge of ~H, and we can chide her by calling her knowledge of ~H mere animal knowledge. But nobody has fullyreflective knowledge of ~H. So, whenever we’re in the mood to view the glass as half-empty rather than half-full, we can underscore the appeal of fully reflective knowledge, then point out that we can’t have it. While VP endorses the Moorean response, it preserves many of our intuitions favoring (3) by giving us resources for emphasizing, as much as we please, the following: first, that it would be epistemically better to believe reflectively that p than to believe unreflectively that p; 34 second, that we can’t have what we’ve been after all along: unshakable Cartesian certainty.
So, VP preserves our intuitions about (1) and (2), and it goes a long way toward preserving our intuitions about (3). Moreover, because VP is a traditional theory of knowledge, nothing in VP commits us to a contextualist account of knowledge sentences. 35 (Of course, the linguistic data might turn out to support a contextualist account—one according to which “knows” sometimes expresses only the satisfaction of the belief, truth and virtue conditions and, in other contexts, expresses the satisfaction of the perspective condition as well—but that would be another matter.) It follows that, in addition to preserving (most of) our intuitions about (1), (2) and (3), VP does relatively little violence to the skeptical, Moorean, disagreement and epistemological intuitions mentioned in Section IV.
In this paper, I have tried to undermine the case for contextualism by showing that contextualism is not necessary for preserving our intuitions. I have argued for this conclusion, first, by showing that contextualism does violence to our intuitions about knowledge and knowledge sentences, and, second, by showing how a properly Chisholmed virtue perspectivism might succeed where contextualism fails. If my argument has been successful, contextualists must abandon their favorite argument thus far—that contextualism delivers the goods with respect to our intuitions.
Bach, K. (2005). “The Emperor’s New ‘Knows’.” In Contextualism in Philosophy: Knowledge, Meaning, and Truth. Eds. Gerhard Preyer and Georg Peter. New York: Oxford UP, 51-89.
Cohen, S. (1988). “How to be a Fallibilist.” Philosophical Perspectives 2, 91-123.
Florida Philosophical Review Volume IX, Issue 1, Summer 2009 66
DeRose, K. (1995). “Solving the Skeptical Problem.” The Philosophical Review 104 (1), 1-52. DeRose, K. (2004). “Sosa, Safety, Sensitivity, and Skeptical Hypotheses.” In Ernest Sosa and His
Critics. Ed. John Greco. Oxford: Blackwell, 22-41.
Dretske, F. (1970). “The Case Against Closure.” In Contemporary Debates in Epistemology. Eds. Ernest
Sosa and Mattias Steup. Oxford: Blackwell, 13-26.
Dretske, F. (1971). “Conclusive Reasons.” Australasian Journal of Philosophy 49 (1), 1-22.
Greco, J. (1992). “Virtue Epistemology.” In A Companion to Epistemology. Eds. Ernest Sosa and Jonathan Dancy. Oxford: Blackwell, 520-22.
Greco, J. (2004). “Introduction.” In Ernest Sosa and His Critics. Ed. John Greco. Oxford: Blackwell, xv-xxiv.
Lewis, D. (1996). “Elusive Knowledge.” Australasian Journal of Philosophy 74 (4), 549-67. Nozick, R. (1981). Philosophical Explanations. Cambridge, MA: Harvard UP, 172-217. Plantinga, A. (1993a). Warrant: The Current Debate. New York: Oxford UP.
Plantinga, A. (1993b). Warrant and Proper Function. New York: Oxford UP.
Plantinga, A. (2000). Warranted Christian Belief. New York: Oxford UP.
Sosa, E. (1997). “How to Resolve the Pyrrhonian Problematic: A Lesson from Descartes.”
Philosophical Studies 85, 229-49.
Sosa, E. (1999). “How to Defeat Opposition to Moore.” Philosophical Perspectives 13, 141-53.
Stough, C. (1992). “Pyrrhonism.” In A Companion to Epistemology. Eds. Ernest Sosa and Jonathan Dancy. Oxford: Blackwell, 408.
- For an excellent summary of contextualism, see Bach 2005, pp. 51-67.
- As DeRose puts it, “Contextualist theories of knowledge attributions have almost invariably been developed with an eye toward providing some kind of answer to philosophical skepticism.” See DeRose 1995, p. 4.
- Hence DeRose’s comment (see footnote 2, above) that contextualism is a theory about knowledge attributions, not about knowledge itself.
- As defined in Daniel Dennett’s The Philosophical Lexicon—chisholm, v. To make repeated small alterations in a definition or example.
- See DeRose 1995, p. 1. “BIV” is, of course, short for “brain in a vat.”
- See DeRose 1995, p. 1. “BIV” is, of course, short for “brain in a vat.”
- DeRose 1995, pp. 27-29.
- See Bach 2005, p. 68.
- See Bach 2005, p. 69.
- Bach 2005, p. 69.
- DeRose 2004, p. 38.
- Since at least the time of Pyrrho, around 390 BC. See Stough 1992, p. 408.
- See DeRose 2004, p. 39, Cohen 1988, pp. 111–115, and Sosa 1999, p. 147. As DeRose puts it, “Since I first encountered [the skeptical paradox], I personally have been fairly strongly inclined to think that I do [italics his] know the various skeptical hypotheses to be false.”
- It strikes Cohen (a contextualist) as absurd even in philosophical contexts. See Cohen 1988, pp. 111- 15.
- See Nozick 1981 and Dretske, 1970 and 1971.
- See Plantinga, 1993a, 1993b, and 2000. Plantinga views his epistemological framework as a variety of naturalized epistemology (see his 1993b, p. 46); nevertheless, John Greco takes Plantinga’s work to be a paradigmatic example of virtue epistemology. See Greco 1992, p. 521.
- See Plantinga 2000, pp. 156-161.
- Plantinga 2000, p. 156. Plantinga defends the philosophical neutrality of the word “design” in Plantinga 1993b, Chapter 11.
- Viable, that is, with respect to the standard epistemological conundrums.
- Some cognitive faculties are apparently aimed at survival rather than truth. See Plantinga’s comments on this at Plantinga 2000, pp. 362-63.
- See Plantinga 1993b, Chapter 5.
- Either by Darwinian natural selection or God’s designs—take your pick.
- See Plantinga 2000, p. 146.
- Plantinga 2000, 156-61.Plantinga distinguishes between mini- and maxi-environments explicitly for the sake of dealing with Gettier scenarios.
- I owe this example to conversation with Mylan Engel.
- That is, the following instantiation of the closure principle would be false: If Jones knows there’s a blue barn nearby, and Jones knows there’s a blue barn nearby only if there’s a barn nearby, then Jones knows there’s a barn nearby.
- Greco 2004, p. xxii.
- Sosa thinks of knowledge in terms of safety rather than proper function. So, in this respect, Sosa’s account differs significantly from VP.
- See, for example, Sosa 1997, p. 231.
- Recall that, according to VP, S knows that q if the belief, truth and virtue condition obtain with respect to S’s belief that q.
- This label comes from Greco 2004.
- Assuming animals have beliefs, of course.
- Those inclined to object that reflective knowledge is beyond our ken should note that this point holds even if we can’t ascend to reflective knowledge. Regardless of whether or not reflective knowledge is possible, it is intuitively obvious that it’s better to believe reflectively that p than to believe unreflectively that p.
- Recall that, according to VP, Jones does know she’s not a bodiless BIV and (3) is therefore false. When the skeptic says, “It’s not the case that Jones knows ~H,” he might mean that, while Jones satisfies the belief, truth and virtue conditions, she fails to satisfy the perspective condition. In this case, the VP theorist can admit that what the skeptic means is true. Nevertheless, because VP maintains that Sknows p so long as the belief, truth and virtue conditions obtain with respect to S’s belief that p, the VP theorist can claim that, in spite of what the skeptic means, Jones does know ~H and the sentence the skeptic utters is consequently false.