This is not final draft. Do not cite without permission. Comments are welcome.
Virtue Epistemology and Environmental Luck Masashi Kasaki
Abstract: Virtue epistemology has it that knowledge is a kind of success through ability, and explains the value of knowledge in terms of the general value of success through ability. However, Duncan Pritchard, in a series of recent writings, argues that knowledge is not merely a success through ability, and the virtue theoretical explanation of the value of knowledge fails. His argument relies on the claims derived from examples involving what he calls ‘environmental luck.’ First, I propose counterexamples to Pritchard’s claims about environmental luck. Second, I offer a diagnosis of both Pritchard’s and my examples, according to which they involve different means-end structures of performance. Different means-end structures make for different conditions for success to be fully creditable to a subject. Given this, virtue epistemology can deal with all the examples, while maintaining its main tenet. Third, I show that my response to Pritchard’s argument against virtue epistemology is more plausible then the ones offered by John Greco and Ernest Sosa.
1. Introduction
A recent development in virtue epistemology, brought about by Ernest Sosa (2007, 2011) and John Greco (2003, 2007, 2008, 2009, 2010), has it that knowledge is a kind of success through ability, skill or competence (I use these terms more or less interchangeably, and nothing important hinges on my word choice). In more detail, the core thesis of recent virtue epistemology is put as follows:
1
S knows that p if and only if S believes the truth (with respect to p) because S’s belief that p is produced by intellectual ability. (Greco 2010: 71)
Belief amounts to knowledge when apt: that is to say, when its correctness is attributable to a competence exercised in appropriate conditions. (Sosa 2007: 92)
If the core thesis is true, virtue epistemology has theoretical advantages over other accounts of knowledge. Most importantly, it offers the most promising solution to the value problem about knowledge. In general, when S succeeds at achieving the aim of a performance because of S’s abilities, S is credited for the success of the performance. This is because success through ability is generally more valuable than success by other means, e.g., success by luck. The core thesis applies this general idea to the value of knowledge. According to Greco and Sosa, the aim of S’s forming beliefs about the relevant subject matter is to acquire truth. When S succeeds at believing truth because of S’s intellectual abilities and thereby has knowledge, S is credited for acquiring truth. This explains why knowledge is more valuable than mere true belief or true belief by luck.1 Duncan Pritchard, in a series of writings (2008a, b, 2009a, b, 2010), develops a twotiered argument against the core thesis of virtue epistemology. He argues that the core thesis is susceptible to two different kinds of counterexamples: one implies that success through ability is not sufficient for knowledge, and one implies that it is not necessary. These kinds of counterexamples, then, jointly constitute a dilemma for virtue epistemology. On the one hand, in order to evade counterexamples of the first kind, the condition for knowledge figuring in the core thesis must be strengthened. On the other hand, in order to evade counterexamples of 1
On Greco’s definition, the ‘because’ relation in the core thesis holds if S’s ability is the most salient component of the causal explanation of why S believes truly that p. On Sosa’s definition, it holds if S believes truly that p by manifesting S’s competence. In this paper, I commit myself to neither of these definitions, and use ‘because’ in as much a neutral sense as possible; defining this relation in one way or another is not crucial for the purposes of the paper.
2
the second kind, it must be weakened. Pritchard claims that virtue epistemology, insofar as it aims to give a necessary and sufficient condition for knowledge, has no way out of the dilemma.2 The focus of this paper is on the counterexamples of the first kind that are reckoned to constitute a horn of the dilemma. I argue that it is possible to deal with them with the resources available in virtue epistemology. My argument is not inevitably vulnerable to the other horn of the dilemma. For Pritchard’s alleged dilemma is genuine only if virtue epistemology is meant to provide a set of conditions for knowledge that are invariable from case to case, or from context to context. As a matter of fact, neither Sosa nor Greco envisages their conditions for knowledge, as they are cited above, as invariant. For example, the conditions allow such parameters to vary across cases or contexts, as which of S’s abilities are relevant, what the obtaining of the ‘because’ relation between S’s ability and true belief requires, and what degree of success is required, for knowledge.3 Perhaps then, Pritchard’s dilemma is better taken as a challenge to the effect that the conditions must not be ad hoc, merely adjusting the parameters on a case-by-case basis. This, however, is a general constraint on any account of knowledge. There is reason to think, as will be explained in a moment, that counterexamples of the first kind are especially pressing for virtue epistemology, and my concern here is with the specific challenge they propose against the core thesis of virtue epistemology.4
2
Pritchard accredits the counterexamples of the second kind to Lackey (2007). Lackey originally proposes them against Greco (2003) in her (2004), and herself puts forth a dilemma very similar to Pritchard’s in her (2009). 3 The variability of the first parameter is obvious, and commonly endorsed by virtue epistemologists. In addition, Greco (2007, 2009, 2010) endorses the variability of the second parameter, and Sosa (2009, 2011) that of the third parameter. 4 Lackey (2009) argue that virtue epistemologists’ attempts, in particular, Greco’s (2007) and Sosa’s (2009), to deal with counterexamples of the second kind all fail. Lackey’s main argument in part presupposes that most, if not all, forms of reductionism about testimonial knowledge are false, although the core thesis per se does not favour reductionism or anti-reductionism. Counterexamples of the second kind do not raise a special challenge to the core thesis, for two reasons: first, they would lose their critical power to some extent, were reductionism to turn out to be false; second, the core thesis can be combined with a further condition specific to testimonial knowledge.
3
Essential to Pritchard’s counterexamples of the first kind are what he calls ‘environmental luck’. He distinguishes between two kinds of luck involved in Gettier or Gettier-like cases: intervening and environmental luck. Intervening luck is operative in ‘standard’ Gettier-like cases, where circumstances are not propitious for the exercise of the relevant ability, but luck intervenes so as to result in a success. Environmental luck is operative typically in Goldman-Ginet’s Barn Façade case, where, ‘while circumstances were indeed, as it happens, propitious for the exercise of the relevant ability, they could so very easily have not been’ (Pritchard 2010: 38).5 On Pritchard’s diagnosis, epistemic cases of environmental luck are such that (i) S believes the proposition in question truly because of the relevant ability, but (ii) S’s belief does not amount to a case of knowledge. Hence, such cases are allegedly counterexamples to the core thesis of virtue epistemology. With cases of environmental luck at hand, Pritchard does not merely intend to offer counterexamples to the core thesis, but also to deny the account of the value of knowledge that is provided by virtue epistemology. By generalizing from the Barn Façade case and other similar epistemic cases, Pritchard concludes that knowledge is never compatible with environmental luck. On the other hand, by generalizing from non-epistemic cases of environmental luck, he concludes that success through ability in general is always compatible with environmental luck. Thus, this difference in compatibility with environmental luck suggests that more than success through ability is involved in knowledge, and hence virtue epistemology fails to explain the value of knowledge. My argument against Pritchard is structured in the following way. In §2, I propose counterexamples to his general claims: that environmental luck is never compatible with 5
This is the closest to a definition of environmental luck Pritchard offers. It is, however, quite difficult to pin down what it is for environments to propitious. Propitious environments cannot be identified with those in which S succeeds. This implausibly entails that if S succeeds in any environment, it is propitious for the relevant ability. Pritchard seems to hold that environments are propitious only if S succeeds because of the relevant ability. Intervening luck involved in standard Gettier cases prevents S’s success of having true belief from being attributable to the relevant ability, while S exercises the relevant abilities. For this reason, only non-standard Gettier cases, i.e., epistemic cases of environmental luck, are problematic for the core thesis.
4
knowledge, and that environmental luck is always compatible with success through ability. At this stage, my points, just as Pritchard’s, merely rely on the intuitive judgements of the cases I propose. In §3 and §4, I offer and defend a theoretical diagnosis of the cases, which specifies the conditions under which success through ability is undermined by environmental luck, and those under which it is not. My diagnosis has it that in the Barn Façade case, S believes that the structure in front of him is a barn by performing two things: selecting the relevant target of perception and perceptually identifying it as such. Corresponding to these two things, perceptual and selecting abilities are relevant in the Barn Façade case. Pritchard wrongly assumes that the only relevant ability is perceptual ability. On his diagnosis, S believes truly because of S’s perceptual ability in the Barn Façade case, but environmental luck prevents S from having knowledge. From this, he concludes that the core thesis is not sufficient to accommodate the Barn Façade case. I argue that S believes truly in part because of S’s perceptual ability but not because of S’s selecting ability, and hence S is not fully credited with the success of believing truth. Full creditability requires all, but not some, of the relevant abilities to be responsible for success. Greco (2009, 2010) responds to Pritchard’s objection to the core thesis, and Pritchard (2010), in turn, raises problems for his response.6 Sosa (2007, 2011) provides resources to cope with Pritchard’s objection. Each of Greco’s and Sosa’s response has some similarities with mine. In §5, I explain where my response differs from theirs, and argue that my response is more plausible than their responses.
2. Success through Ability and Environmental Luck
6
For the exchange between Greco and Pritchard, see also Greco (2007) and Pritchard (2008a, b).
5
Pritchard compares a non-epistemic case of environmental luck with the Barn Façade case, and claims that success is achieved because of the relevant ability in both cases. He describe each case as follows:
(Archie) Archie (an archer) . . . selects at target at random, skillfully fires at this target, and successfully hits it because of his skill. On the account of achievement on the table, his hitting of the target is a genuine achievement. Suppose, however, that unbeknownst to Archie there is a force-field around each of the other targets such that, had he aimed at one of these, he would have missed. It is thus a matter of luck that he is successful, in the sense that he could very easily have not been successful. (Pritchard 2010: 35)7
(Barney) Barney forms a true belief that there is a barn in front of him by using his cognitive abilities… The twist in the tail, however, is that, unbeknownst to Barney, he is in fact in ‘barn façade county’, where all the other apparent barns are fakes. Intuitively, he does not have knowledge in this case because it is simply a matter of luck that his belief is true. (ibid.: 35–6)
Pritchard holds that both the Archie and the Barney cases are such that the success of the subject’s performance—hitting the target in the Archie case and believing truth in the Barney case—is attributable to the subject’s abilities, and that Barney, however, does not have knowledge. By generalizing from these and other similar cases, he concludes that environmental luck is always compatible with success through ability, but never compatible with knowledge. 7
Pritchard’s description of the Barney case is a bit unfair to virtue epistemology; his judgement that Barney is credited for believing truth is made before the Gettiering twist is introduced. Nevertheless, I accept that his judgement is intuitive and does not change even after the case is Gettierized.
6
It is easy to refute Pritchard’s claim that environmental luck does not undermine success through ability, by way of example. Consider the following case:
(Cutie) Cutie is a bird hunter. She has heard that black-tailed godwits are good targets for hunting, and decides to go to a hunting spot known for them. Now, on her way to a hunting spot for black-tailed godwits, she finds another hunting spot where she sees many godwits. She carefully aims at one of them, and shoots it. As a result, she gets a black-tailed godwit. Unbeknownst to her, however, this is a famous hunting spot for bar-tailed godwits — which are only distinguished from black-tailed godwits by barred tails and lack of white wing bars. The black-tailed godwit she shot just happened to be in a flock of bar-tailed godwits.
Intuitively, the success of Cutie’s performance—hunting the relevant target—is not attributable to her abilities. It owes too much to luck rather than to her abilities. She has picked a black-tailed godwit as a target, just by luck. In addition, it is easy to refute Pritchard’s claim that environmental luck undermines knowledge, again by way of example. Consider the following case:
(Davie) Davie is taking a logic exam at university. Each student in the exam room is assigned a computer, and asked to prove a relatively complex theorem in predicate logic, which is randomly selected by the computer from the big pool of exam problems. He is well prepared for the exam; he has practiced a lot and developed his derivation skills. He has successfully proved the theorem that shows up on the monitor of his computer, and believes that the formula he got on the exam is provable. Unbeknownst to him, however, the exam system has a massive error, and other
7
problems in the pool are impossible to prove (perhaps, the teacher failed to correctly transcribe the other problems from the textbook in inputting data for the system).
Again, intuitively, Davie knows that the formula is provable. The Cutie case is meant to be analogous to the Barney case, and the Davie case is meant to be analogous to the Archie case, in all the relevant respects that Pritchard appeals to in constructing the cases of environmental luck. For this reason, the Cutie and the Davie cases are cases of environmental luck, insofar as the Archie and the Barney cases are so.8 Then, if one shares my intuitions about the Cutie and the Davie case, they undermine both of Pritchard’s claims: the intuition that Cuties has not achieved her aim primarily because of her abilities refutes his general claim that success through ability is always compatible with environmental luck; and the intuition that Davie has achieved his aim primarily because of his abilities refutes his more particular claim that knowledge is never compatible with environmental luck. For the time being, however, I grant that one shares Pritchard’s intuitions about the Archie and the Barney cases. Then, if there is no relevant difference between the Archie and the Cutie cases, no general conclusion about the relation between success and ability is to be derived; similarly, if there is no relevant difference between the Barney and the Davie cases,
8
Some may doubt that the Davie case satisfies the definition of environmental luck on the following grounds: first, the proposition in question is ambiguous between de re and de dicto reading: on the de re reading, Davie believes of the formula he got that it is provable, and on the de dicto reading, he believes that whatever formula he gets on the exam is provable. This ambiguity may render it indeterminate whether the Davie case is a case of environmental luck. Reply: The same ambiguity exists in the Barney case. If ambiguity prevents one from having a clear judgement that the Davie case is a case of environmental luck, the same is true of the Barney case. Moreover, I do not believe that there is ambiguity in the Davie and the Barney cases; it is clearly the de re reading that matters in both cases. Second, on Pritchard’s analysis, cases of epistemic environmental luck are also cases of veritic luck—the genus of knowledge-undermining luck: according to the account developed in Pritchard (2005), roughly, a belief is true by veritic luck only if it is false in some nearby possible world where S has that belief. However, Davie believes in a necessary truth, and his belief cannot be true by veritic luck. Then, the Davie case cannot be a case of environmental luck. Reply: Pritchard’s account of veritic luck should not be presupposed here. Insofar as it is granted that Davie’s belief is true by luck in the relevant sense, it is a counterexample to Pritchard’s account of veritic luck. In addition, the Davie case can be easily modified in such a way that the proposition Davie believes is contingent. The de re reading may seem to be incompatible with this point. This problem is well recognized and discussed by proponents of modal accounts of knowledge. I do not rehearse their discussion here. Note that as just pointed out, if this is a problem for the Davie case, it is so for the Barney case as well.
8
no general conclusion about the relation between knowledge and ability is to be derived. This is a problem not only for Pritchard but also for the virtue theorist. To break this stalemate, obviously, some theoretical work is necessary, to which I turn now.
3. Performance and Its Means-End Structure
In this section, I offer a diagnosis of the cases at hand, and it justifies the intuitive judgements about the cases at hand. On my diagnosis, the Archie and the Davie cases, where the subjects are fully credited for success, involve a different structure of means-end relation than do the Barney and the Cutie cases, where the subjects are not fully credited for success. Our performances may or may not have an aim, whether it is conscious or not. Sosa calls performances with an aim ‘endeavors’. An endeavor being directed at a certain aim is consistent with the possibility that it serves other aims. Thus, Sosa (2011: 14) claims:
[O]ne can aim to flip a switch by operating on it in a certain way with one’s fingers. In doing what one does with one’s fingers, one endeavors to flip the switch. In one’s plan, the switch-flipping endeavor might itself serve a further endeavor: one might be aiming to turn on the light. One might also do something else thereby, such as altering a prowler, even if this last is not assessable for its success. One’s altering of the prowler is not a success if it was not one’s aim.
What Sosa claims of endeavors and their aims here is reminiscent of the point made of action by Anscombe (1957) and Davidson (1971): we often describe action using one or a series of ‘by’-relations: ‘one has φ-ed by ψ-ing,’ ‘one has ψ-ed by θ-ing,’ and so on. For example, Sosa’s example just given contains the following structure of nested means-end
9
relations: (((you have turned on the light) by flipping the switch) by moving a finger in a certain way). On Anscombe-Davidson’s view, attributing a certain means-end structure to action is giving a description of the action. One and the same action is described differently by attributing different (or no) means-end structures; for example, ‘you have turned on the light’, ‘you have turned on the light by flipping the switch’ and ‘you have flipped the switch by moving your finger in a certain way’ do not describe different actions; they are different descriptions of one and the same action. I leave it open whether this contentious ontological thesis is right or not.9 It will do for my purposes here that actions or performances can be described using one or more ‘by’-relations, whether the relata of such relations refer to the same thing or not. Even though a performance can have multiple aims, I’ll follow Sosa in accepting the simplifying assumption that a performance has exactly one aim, and its success is equated with achieving that aim. In other words, I’ll talk as if every performance is an endeavor in Sosa’s sense. Even if it is assumed that every performance has an aim, the means-end structure involved in a given performance may be quite complex. First, as Sosa points out in the passage cited above, it may be true that S achieves the aim of a certain performance by conducting other performances. Second, the very idea of success through ability implies that when the success of a performance is attributable to a certain ability, S achieves its aim not just by conducting the performance but also by exercising the ability. Third, these two points suggests that S may achieve the aim of a performance by exercising multiple abilities. In Sosa’s terminology, an exercise of ability is a performance, but distinguishing between them is useful for clarifying the means-end structure of a performance and its relevance for the creditability of success to a subject. Suppose, for example, that S has succeeded at making good soup by making a good soup base and boiling it with vegetables 9
This problem arises not only for action but also for mental activities. The notorious generality problem is a form of the problem. To settle the problem one way or another is beyond the scope of the paper.
10
properly. S has used her high skill in making the good soup base, but this may not be enough for her to gain a full credit for success. S has failed to heat it properly due to lack of boiling skill. It is merely a matter of luck that she has stopped heating at a perfect timing. Cases of partial credit like this are mundane.10 They suggest the following condition for the full creditability of success to S: when the relevant description of the case at hand is ‘S succeeds at achieving the aim of a performance φ by conducting a set of performances φ1, …, φi,’ S is fully credited for success iff (i) S has a set of abilities ψ1, …, ψi that are responsible for (the referents of) φ1, …, φi respectively, and (ii) S achieves the aim of φ because of ψ1, …, ψi.11 More simply, full creditability of success amounts to attributability of success to all, but not some of, the relevant abilities. The following list describes the relevant means-end structure of the performance at stake in each of the four cases offered in §2:
Cases Archie Barney
Aim hitting the target acquiring truth
Performance shooting at the target believing that the structure is a barn
Cutie
hunting black-tailed godwits acquiring truth
shooting at a blacktailed godwit believing that the formula is provable
Davie
By shooting an arrow selecting a target object of perception perceptually identifying it as a barn selecting a target of hunting shooting it deriving the formula
Barney and Cutie achieve their aims by conducting more than one performance, one of which pertains to selecting the relevant target, i.e., the target that will lead to the
10
Sosa (2007, 2011) introduces the notion of partial credit in response to counterexamples of the second kind: Credits can be attributed partly to S and partly to S’s informant in cases of testimonial knowledge. Sosa’s distinction between full and partial credit is different from mine spelled out here. 11 The left-hand side of the biconditional is open to different construals of the ‘because’ relation as noted in footnote 1. On Greco’s condition, the left-hand side amounts to the condition that ψ1, …, ψi are all salient factors of the causal explanation of why S achieves the aim of φ. On Sosa’s definition of the ‘because’ condition, , the left-hand side amounts to the condition that ((S achieves the aim of φ) (by ((φ1) by manifesting ψ1), …, and by ((φi) by manifesting ψi))).
11
achievement of their aims; whereas Archie and Davie achieve their aims by conducting but one performance, and it does not involve selecting the relevant target. One’s hitting the archery target, unlike hunting, neither involves nor requires selecting the relevant target, since it is simply assigned to one. Similarly, one’s believing that the formula on the exam is provable neither involves nor requires selecting the relevant copy of exam, since it is simply assigned to one. This explains why Barney and Cutie are not fully credited for success, as opposed to Archie and Davie. In order for success to be fully creditable to Cutie, it is required that she both select the relevant target of hunting and shoot it because of the corresponding abilities. By the same token, in order for success to be fully creditable to Barney, it is required that he both select the relevant target of perception and identify it as such because of the corresponding abilities. Luck matters with regard to the ability of selecting the relevant target. Cutie selects a target of hunting, a black-tailed godwit, at random, and it could very easily have been a bird of a different species, a bar-tailed godwit. It is a matter of luck that she has selected a black-tailed godwit as the relevant target of her performance. This is consistent with the fact that Cutie’s success is attributable to her shooting ability. After all, she succeeds at shooting the target luckily selected because of her shooting ability. Again, by the same token, Barney selects a target of perception at random, and it could very easily have been a fake barn. It is a matter of luck that he has selected the real barn as the relevant target of perception.12 This is consistent with the fact that Barney’s success is attributable to his
12
Greco and Sosa both hold that abilities and competences are by their nature reliable. So one way to cash out the non-obtaining of the because relation in the Barney and the Cutie cases is to say that neither Barney nor Cutie has an ability in selecting the relevant targets, relative to the environments they are in. I am open to the possibility of an ability or competence being unreliable (This is suggested by Turri (forthcoming)). Then, another way to cash out the non-obtaining of the because relation is to say that the selecting abilities Barney and Cutie have do not explain why the success is achieved, because too much luck is involved in the success of selection. The difficulty of choosing one of these ways is related to the difficulty of specifying what kind of luck the success of selection is susceptible to. It is not clear whether the pertinent kind of luck is intervening or environmental. If it is intervening, Barney’s and Cutie’s environments are propitious for the exercise of the selecting abilities; if it is environmental, they are not. Or perhaps, it is a third kind of luck. Sosa (2011: 25-9) discusses a similar difficulty.
12
perceptual ability. He succeeds at perceptually identifying the object luckily selected as it is because of his perceptual ability. By generalizing from the Archie and the Barney cases (and other similar cases), Pritchard claims that success through ability is always compatible with environmental luck, and that knowledge is never compatible with environmental luck. His claims rely on the assumption that the two cases are relevantly similar. In particular, Pritchard assumes that the performances in question have a simple, one-performance-one-means structure, and the relevant ability corresponding to the means is but one. Once this is questioned, all Pritchard shows of the Barney case is that Barney’s success is attributable to some but not all of the relevant abilities. Insofar as the core thesis requires full-credit rather than partial credit to have knowledge, it can handle the Barney case.
4. The Barn Façade Case and Selecting Ability
It may seem doubtful that the Barney case relevantly involves two kinds of abilities on the following ground: it is not plausible that ability to select the relevant target is always involved in perception, and the Barney case is not so much different from ordinary perceptual cases, at least when it is described in the way Pritchard does. In this section, I defend my diagnosis of the cases on the two grounds. It must be the case that the Barney case is not the same as ordinary perceptual cases in every respect, insofar as S does not have knowledge in the former while S does in the latter. The two grounds I offer explain what the relevant difference is. First, DeRose (2009, ch. 1, fn. 24: 23) reports that when the case is described in a way similar to Pritchard’s, many epistemologists including himself do not have the strong
13
intuition that Henry does not know.13 DeRose, then, suggests a modification of the Barn Façade case. Since his modification is relevant for my argument below, I cite it at length:
(Modified Barn Façade) Imagine that Henry has encountered many of the fake barns (let’s say 19 of them), and has been fooled by all of them that he has encountered, believing them to be real barns. Now, he’s encountering the only real barn in the region (so he is 1-for-20 in his barn judgements), but it seems the same to him as all the other objects he has taken to be barns: He is not in any way more confident of his identification of the object as a barn in the case as compared with all the other instances. (‘Yet another barn,’ he thinks.) Make sure you’re imagining the case so that the fakes present an extremely convincing real appearance—just like the real thing—and Henry should have the same visual evidence as he would have had if all the barns around were real, and should be just as justified in thinking that what he is presently looking at is a barn as he would have been if all the barn-like things he had encountered had all been real.
With this modification, DeRose is confident that Henry does not know that it is a barn. I agree with DeRose (and presumably, the other epistemologists he queried) that the mere presence of fake barns in the vicinity is not sufficient to pump the intuition that Henry or Barney does not know. Then, what makes a difference in our intuition? A plausible answer is that the modified Barn Façade case puts emphasis on the fact that there is a set of barn-like structures and Henry has no ability to select real ones out of them. There are twenty barn-like structures b1, …, b20, of each of which Henry believes one of propositions p 1, . . ., p20, to the effect that it is a barn. Among these only b20 is a real barn, and thereby only p20 is true (the
13
See also DeRose (2009, ch. 2, fn. 2: 49–50)
14
subscripts indicate the order of Henry’s encounter with barn-like structures). Nevertheless, as DeRose accentuates, Henry is equally confident of, and has equal qualitative evidence for p1, . . ., p20, since all the fakes, b1, . . ., b19, are extremely convincing fakes. To pump the intuition that Henry does not know, I submit, it need not be the case that the real barn in the Barn Façade County is the very last barn-like structure he encounters therein (the subscript indexed to the real barn need not be 20). What DeRose’s modification makes explicit is, then, that Henry is not able to select the real barn among structures b1, . . ., b20. He selects the relevant target of perception—the real barn—without differentiating it from other structures, and hence his success of selection is not attributable to his selecting ability. Henry’s confrontation with many fakes (or describing the case with an emphasis of this point) makes salient his lack of ability to select the relevant target. Second, given DeRose’s modified description of the Barn Façade or Barney case, it is more similar to lottery cases in Vogel’s (1990) sense than to normal perceptual cases (following Hawthorne (2004), I take it that lottery cases arise for externalist as well as internalist accounts of knowledge ). A lottery case can be constructed in terms of a lottery containing one winning and nineteen losing tickets. This lottery case is similar to the Barn Façade case, and both cases satisfy the four conditions that Vogel requires for lottery cases. First, there is a set of propositions p1, . . ., p20, to the effect that b1, . . ., b20 are barns for the Barn Façade case, and that b1, . . ., b20 are losing tickets for the lottery case. Second, it is not abnormal in the Barn Façade case if pi turns out to be false, since bi could have been a fake barn; by the same token, it is not abnormal in the lottery case that pi turns out to be false, since bi could have been a winning ticket. Third, as exposed above, the evidence S has for each of p1, . . ., p20 is not different in quality: it has the same (or at least very similar) sensory quality in the Barn Façade case, and it is the same statistical evidence in the lottery case.
15
Fourth, it is the case that at least one of p1, . . ., p20 is false.14 The difference between the lottery case and the Barn Façade case is that in the former only one of p1, . . ., p20 is false, whereas in the latter all but one is false. In the lottery case, even though S believes truly some of p1, . . ., pi that it is a losing ticket, S’s beliefs do not amount to knowledge. S bases her beliefs on good statistical evidence, and hence there is a sense in which S believes truly in virtue of her good inferential ability. However, there is little sense in which S believes truly in virtue of her selecting ability. For, if S has such a selecting ability, S can select a winning ticket rather than losing tickets. By the nature of the lottery case, S cannot select a winning ticket, and S does not have the relevant selecting ability. Or at the very least, it is hard to say that S’s selecting ability brings S to have the true beliefs.15 If I am right that the Barn Façade case is similar to the lottery case in the relevant respects, the same kind of explanation of the lack of S’s knowledge is to be applied to the Barn Façade case.
5. Sosa’s and Greco’s Responses
My response to Pritchard’s alleged counterexamples is similar to each of Greco’s and Sosa’s, to some extent. Greco (2009; 2010: 88–9) offers a case similar to the Cutie case to undermine Pritchard’s claim that success through ability is always compatible with environmental luck. Sosa discusses cases analogous to the Barn Façade case in his (2007: 98–112), and points out
14
Vogel’s original fourth condition is that it is known that at least one of p1, . . ., pi is false. He does not specify by whom it needs to be known (this is partly because he is open to the option that it is known only by the attributer of knowledge); S may not know this, but even if not, this does not seem to affect our judgement about knowledge in lottery cases. I here use a rather simplified fourth condition, merely requiring that at least one of p1, . . ., pi is false. Complication arises for the fourth condition, since, as Hawthorne (2004: 15-6) points out, it is not necessary for lottery cases that there is a guaranteed winner of the lottery. I do not here address the question of what essential features lottery cases have. 15 This, I submit, is the most plausible explanation of ignorance in the lottery case given by virtue epistemology. Greco (2003) attempts to handle the lottery case in terms of his account of the context-dependence of causal explanation. I doubt that Greco succeeds at his attempt. The explanation he appeals to seems to be of why the content of S’s belief that a ticket is a loser is true rather than of why S forms the true belief.
16
that they are similar to a case like the Cutie case in his (2011: 7–13). His discussion provides recourses to respond to Pritchard’s examples. In what follows, I briefly describe similarities and differences between mine and each of their responses. Sosa’s Response: he denies that Barney does not know that the structure in front of him is a barn. On his view, Barney has animal (first-order) knowledge that it is a barn, insofar as his believing truth is achieved through his reliable perceptual ability. However, he lacks reflective (second-order) knowledge of the same proposition, i.e., does not know that he believes truly that the structure in front of him is a barn because of his perceptual ability. On Sosa’s view, the lack of reflective knowledge consist in the lack of the ability to respond to internally available signs or evidence that indicates that his perceptual ability is reliable. He calls the ability relevant to reflective knowledge ‘selecting ability.’ Thus, this response has a similarity to mine, in that it postulates two kinds of abilities in the Barney case: perceptual and selecting abilities. My response, however, does not assume any distinction in kind of knowledge, and hence the selecting ability I posit may not be related to the possession of evidence with regard to one’s reliability. For Sosa, the selecting ability pertaining to reflective knowledge is primarily applied to his perceptual abilities rather than perceptual objects; it is the ability to control the application of perceptual abilities to objects by using internal evidence for the likelihood and risk of the application. Thus, the selecting ability is inherently meta-level. This sits well with attributing the following nested means-end structure to Barney’s performance of believing: (((he believes that the structure in front of him is a barn) by perceptually identifying it as a barn) by selecting the relevant object of perception). On the other hand, my diagnosis of the Barney case does not require this particular meansend structure, though it does not exclude it. At minimum, my diagnosis applies if the meansend structure true of the Barney case to be that ((he believes that the structure in front of him is a barn) ((by perceptually identifying it as a barn) and (by selecting the relevant object of
17
perception))). The two performances that are responsible for Barney’s believing are not nested, i.e., one does not control the other, but are merely exercised at the same time; nor does one require having internally available evidence as to how well the other fares in achieving its aim. For these reasons, the selecting ability I posit is not necessarily meta-level. The requirement I impose on the Barney case being weaker than that Sosa does makes my diagnosis less contentious.16 Greco’s Response: he relativizes abilities to environments. On his view, Barney has the ability to tell barns from non-barns in normal environments, but he does not have the same ability when he is in the Barn Façade County. Pritchard (2010) replies that it is counterintuitive that Barney lacks perceptual or barn-discriminating ability in the Barn Façade County, while he has it outside of the County.17 My response circumvents this problem of Greco’s response.18 My response is rather modest in comparison to Sosa’s and Greco’s. They are committed to denying one of the two plausible ideas concerning the Barney case: (a) he lacks first-order knowledge that the structure in question is a barn, and (b) he has the perceptual ability to identify barns or to tell barns from non-barns. Sosa endorses (b) but not (a), whereas Greco endorses (a) but not (b). It is desirable for virtue epistemology to have a response to
16
For problems with Sosa’s diagnosis of the Barn Façade case, see Kornblith (2009). Pritchard puts forth the second reply to Greco that even if abilities are relativized to environments individuated in fine-grained manners, it is not denied that environmental luck is compatible with success through ability, and this point still makes virtue epistemology toothless. I agree with Pritchard here. Greco (2010) proposes a case involving a hypothetical sport in which the success condition of one’s performance includes discrimination of the target. The case is structurally similar to the Cutie case, and in part undermines Pritchard’s second reply. 18 Greco espouses contextualism of some sort. He argues that the truth of the knowledge attributions varies with the attributer’s context of practical reasoning. On his contextualism, different contexts assign different environments to abilities. I am sympathetic to the contextualist treatment of the Barn Façade case, and what has been said in the paper is consistent with contextualism of a sort. My favored view is that context determines which attribution of a means-end structure of the performance in question is relevant. Kvanvig’s (2009) objection to Greco points towards the implausibility of the contextualist solution to the Gettier problem. The Barn Façade case, unlike other Gettier cases, seems to be amenable to a contextualist treatment. Gendler and Hawthorne (2005) enumerate cases similar to the Barn Façade case to show how shifty our intuitions about these cases are, and at the end suggest that some contextual shift may be relevant to the shiftiness of our intuitions. Contextualism may explain why, as DeRose reports, we have different intuitions when the case is described differently. 17
18
Pritchard’s objection to the core thesis, without denying either of (a) and (b). I have offered such a response.
Acknowledgement: The research on this paper is supported by the Government of Canada Post-Doctoral Research Fellowship. Early drafts of the paper are read at the 2011 meeting of the Canadian Society for Epistemology and the 2012 annual meeting of the Canadian Philosophical Association. I thank audiences for discussion. My deepest appreciation goes to Carrie Ichikawa Jenkins for helpful comments and suggestions on drafts.
REFERENCES
Anscombe, G. E. M. (1957/2000). Intention, Cambridge, MA: Harvard University Press. Davidson, Donald. (1971). Agency, in Agent, Action, and Reason, ed. R. Binkley, R. Bronaugh and A. Marras, Toronto: University of Toronto Press: 3–37; reprinted in his Essays on Actions and Event, Oxford: Oxford University Press, 2001: 43–62. DeRose, Keith. (2009). The Case for Contextualism: Knowledge, Skepticism, and Context, Vol. 1, Oxford: Oxford University Press. Gendler, Tamar Szabó and John Hawthorne. (2005). The Real Guide to Fake Barns: A Catalogue of Gifts for Your Epistemic Enemies, Philosophical Studies 124(3): 331–52 Goldman, Alvin. (1976). Discrimination and Perceptual Knowledge, Journal of Philosophy 73(20): 771–91. Greco, John. (2003). Knowledge as Credit for True Belief, in Intellectual Virtue: Perspectives from Ethics and Epistemology, ed. M. DePaul and L. Zagzebski, Oxford: Oxford University Press: 111-34.
19
Greco, John. (2007). The Nature of Ability and the Purpose of Knowledge, Philosophical Issues 17(1): 57–69. Greco, John. (2008). What’s Wrong with Contextualism?, Philosophical Quarterly 58(232): 416–36. Greco, John. (2009). Knowledge and Success from Ability, Philosophical Studies 142(1): 17– 26. Greco, John. (2010). Achieving Knowledge: A Virtue-Theoretic Account of Epistemic Normativity, New York: Cambridge University Press. Hawthorne, John. (2004). Knowledge and Lotteries, Oxford: Oxford University Press. Kornblith, Hilary. (2009). Sosa in Perspective, Philosophical Studies 144(1):127–36. Kvanvig, Jonathan. (2009). Responses to Critics, in Epistemic Value, ed. A. Haddock, A. Millar and D. Pritchard, Oxford: Oxford University Press: 339–51. Lackey, Jennifer. (2004). Review of Intellectual Virtue: Perspectives from Ethics and Epistemology, Notre Dame Philosophical Review, eds. G. Gutting and A. F. Gutting, URL = Lackey, Jennifer. (2007). Why We Don’t Deserve Credit for Everything We Know, Synthese 158(3): 345–61. Lackey, Jennifer. (2009). Knowledge and Credit, Philosophical Studies 142(1): 27–42. Pritchard, Duncan. (2005). Epistemic Luck, Oxford: Oxford University Press. Pritchard, Duncan. (2008a). Greco on Knowledge: Virtues, Contexts, Achievements, Philosophical Quarterly 58(232): 437–47. Pritchard, Duncan. (2008b). Virtue Epistemology and Epistemic Luck, Revisited, Metaphilosophy 39(1): 66–88. Pritchard, Duncan. (2009a). Apt Performance and Epistemic Value, Philosophical Studies 143(3): 407–16.
20
Pritchard, Duncan. (2009b). Knowledge and Virtue: Response to Kelp, International Journal of Philosophical Studies 17(4): 589–96. Pritchard, Duncan. (2010) (with Alan Millar and Adrian Haddock). The Nature and Value of Knowledge: Three Investigations, Oxford: Oxford University Press. Sosa, Ernest (2007). A Virtue Epistemology: Apt Belief and Reflective Knowledge, Vol. 1, Oxford: Oxford University Press. Sosa, Ernest (2011). Knowing Full Well, Princeton, NJ: Princeton University Press. Turri, John. (forthcoming). Review of John Greco, Achieving Knowledge, Mind. Vogel, Jonathan. (1990). Are There Counterexamples to the Closure Principle?, in Doubting: Contemporary Perspectives on Skepticism, ed. M. Roth and G. Ross, Dordrecht: Kluwer Academic Publishers: 13-27.
21
