JOHN GRECO

WORRIES ABOUT PRITCHARD’S SAFETY ABSTRACT. I take issue with two claims that Duncan Pritchard makes in his recent book, Epistemic Luck. The first concerns his safety-based response to the lottery problem; the second his account of the relationship between safety and intellectual virtue.

0. INTRODUCTION Duncan Pritchard’s (2005) Epistemic Luck is full of important insights and makes a number of contributions that advance the state of the literature. There is so much in the book that is valuable, and so much I agree with, that it seems inappropriate to focus only on points of contention. But that is the nature of the business, and so I move straight away to two objections. The first concerns Pritchard’s claim that his safety condition solves the lottery problem. The second concerns a claim that Pritchard makes about the relation between safety and intellectual virtue.

1. SAFETY AND THE LOTTERY PROBLEM The lottery problem trades on two stable intuitions. On the one hand, we think that inductive evidence can ground knowledge. On the other hand, we think that a ticket holder does not know that she will lose the lottery, even if the odds of her winning are very small. But how could both intuitions be right? How can it be both (i) that we can have knowledge on inductive grounds, but (ii) that we don’t know we will lose the lottery, even though our inductive grounds for believing this are excellent?1 To sharpen the problem, consider two cases of inductive reasoning. The Lottery Case. S buys a ticket for a lottery in which the chances of winning are ten million to one. A few minutes later, reasoning on the basis of past experience and relevant background knowledge, S forms the true belief that she will lose the lottery. Of course her grounds for so believing are merely inductive: it is possible that she buys the winning ticket, although this is extremely unlikely.

1

The Garbage Chute Case. On the way to the elevator S drops a trash bag down the garbage chute of her apartment building. A few minutes later, reasoning on the basis of past experience and relevant background knowledge, S forms the true belief that the bag is in the basement garbage room. Of course her grounds for so believing are merely inductive: it is possible that the trash bag somehow gets hung up in the chute, although this is extremely unlikely.2

Pritchard thinks that we can solve the lottery problem by making “safety” a necessary condition for knowledge. More specifically, he defines safety as follows (p. 163): (P-SAFETY)

For all agents, A, if an agent knows a contingent proposition p, then, in nearly all (if not all) nearby possible worlds in which she forms her belief about p in the same way as she forms her belief in the actual world, that agent only believes that p when p is true.

The application to the Lottery Case is straightforward: Even when the odds of S’s winning the lottery are very small, and even when S does in fact lose the lottery, there are some close worlds where S wins. And therefore S’s belief that she will lose the lottery does not satisfy Pritchard’s safety condition. Hence S does not know she will lose the lottery. But what about the Garbage Chute Case? Pritchard thinks that he can handle the case by disambiguating with respect to an important detail. Specifically, we can ask whether the garbage bag gets hung up in close worlds. If the answer is no, then there is a relevant difference with the Lottery Case: S’s belief now satisfies the safety condition and we can say that S knows. If the answer is yes, however, then according to Pritchard we lose the intuition that S knows in the case. Either way, Pritchard argues, his safety condition gives us the right result. It seems to me, however, that Pritchard has not appreciated the full force of the lottery problem. Consider the following cases, all of which are supposed to be structurally analogous to the Lottery Case. a.

Vogel’s Hole-In-One Case. “Sixty golfers are entered in the Wealth and Privilege Invitational Tournament. The course has a short but difficult hole, known as the 'Heartbreaker'. Before the round begins, you think to yourself that, surely not all sixty players will get a hole-in-one on the 'Heartbreaker'.” (Vogel 1999, p. 165)

b. The Typing Monkey Case. A monkey sits down at my computer and starts banging away at the keyboard. I believe that he won’t type out a perfect copy of War and Peace.

2

c. Vogel’s Rookie Cop Case. “Suppose two policemen confront a mugger, who is standing some distance away with a drawn gun. One of the officers, a rookie, attempts to disarm the mugger by shooting a bullet down the barrel of the mugger’s gun. . . . Imagine that the rookie’s veteran partner knows what the rookie is trying to do. The veteran sees him fire, but is screened from seeing the result. Aware that his partner is trying something that is all but impossible, the veteran thinks (correctly as it turns out) [that the] rookie missed.” (Vogel 1987, p. 212) d. Hawthorne’s Quantum Mechanics Case. “Suppose that there is a desk in front of me. Quantum mechanics tells us that there is a wave function that describes the space of nomically possible developments of the system that is that desk. On those interpretations of quantum mechanics according to which the wave function gives probability of location, there is some non-zero probability that, within a short while, the particles belonging to the surface of the desk remain more or less unmoved but the material inside the desk unfolds in a bizarre enough way that the system no longer counts as a desk. Owing to its intact surface, the system would be reckoned a desk by normal observers.” (Hawthorne 2004, pp. 4-5)

Each of these cases (along with the Garbage Chute Case) is supposed to be structurally similar to the Lottery Case: there is a very close world where a highly improbable possibility is actual. More exactly: there are many p-worlds and only a small number of not-p worlds, but some not-p worlds are very close to the actual world. Nevertheless, it is absurd to think that there is no knowledge in the cases. Of course we know that ten players in a row will not get a hole in one! Of course I know that the object across the room is a desk! But then the problem remains: How is it that we know in cases a through d but not in the Lottery Case? Let’s go at it from one more angle. We can define a weak safety condition and a strong safety condition as follows: Strong Safety: In close worlds, always if S believes p then p is true. Alternatively, in close worlds never does S believe p and p is false. Weak Safety: In close worlds, usually if S believes p then p is true. Alternatively, in close worlds, almost never does S believe p and p is false.

Cases a through d satisfy the weak safety condition, but so does the Lottery Case. The Lottery Case fails to satisfy strong safety, but so do cases a through d. Now consider Pritchard’s formulation of safety. It is ambiguous between strong and weak safety, because it uses the locution “nearly all (if not all).” But either way, it is hard to see how

3

Pritchard can treat cases a through d and the Lottery Case differently—either they all satisfy Safety III or none do. The lottery problem remains. 2. SAFETY AND VIRTUE A major thesis of Epistemic Luck is that we don’t have to refer to a reliable ability or power in an account of knowledge—all the relevant work can be done by a safety condition (see especially chapter seven). Contra Pritchard, however, there is a kind of example that shows that knowledge requires both a safety condition and a second condition specifying that the seat of safety is a broader ability or power. Consider a colour-blind perceiver and pick a belief where the truth in question has good stability across close worlds. For example, suppose S believes that some frog he sees is green. Suppose also that frogs are by nature green, due to some feature of frog DNA. Given that S is blind-blind, S could easily be wrong about some objects in the environment being green. But S could not be easily wrong that the frog is green, since this (we are supposing) is a stable fact about a natural kind. Intuitively, S cannot know by perception alone that the frog is green even though S’s perceptual belief is safe. I submit that what S lacks is a broader perceptual ability (for discriminating green objects from non-green objects) that is the seat of safety in this particular case. Similar considerations apply to Pritchard’s case of the innocent son (pp. 152-4). In Pritchard’s example a mother believes truly that her son did not commit a murder, but her belief is based only on her maternal love for him. Intuitively, she does not know that her son is innocent. But now stipulate that the son has a saintly character, so that the truth that he did not commit the murder is stable across close possible worlds. In that case the mother’s belief will satisfy Pritchard’s safety condition and will therefore come out as knowledge on Pritchard’s account. In both cases, the problem is that the belief in question satisfies the safety condition too easily. More exactly, it satisfies the condition by a kind of default, by virtue of being about a truth 4

that is stable across close worlds. Pritchard might respond by insisting that P-Safety is supposed to be read as strong safety, and that if p is contingent then there must be some close world where not-p is true. But trying to handle the cases that way will run us straight into scepticism about contingent truths. That is, if S’s beliefs in the Frog Case and the Innocent Son Case do not satisfy the safety condition, it is hard to see how any of our beliefs about contingent truths do. There is a solution to the problem, however. Again, we can add that safety must have its seat in S’s cognitive abilities (or virtue), where an ability is understood as a broader disposition to form safe beliefs in a relevant field under relevant conditions. Put another way, knowledge requires agentbased safety, or safety that is grounded in the cognitive abilities of the knower.

NOTES 1 2

See Cohen (1988, 92-3). The example is adapted from Sosa (2000, 13).

REFERENCES Cohen, S.: 1988, ‘How to Be a Fallibilist’, Philosophical Perspectives 2, 91-123. Hawthorne, J.: 2004, Knowledge and Lotteries, Oxford University Press, Oxford. Pritchard, D.: 2005, Epistemic Luck, Oxford University Press, Oxford. Sosa, E.: 2005, ‘Skepticism and Contextualism’, Philosophical Issues 10, 1-18. Vogel, J.: 1987, ‘Tracking, Closure, and Deductive Knowledge’, in S. Luper-Foy (ed.), The Possibility of Knowledge, Rowman and Littlefield, Totowa, New Jersey, pp. 197-215. Vogel, J.: 1999, ‘The New Relevant Alternatives Theory’, Philosophical Perspectives 13, 155-80.

Department of Philosophy Fordham University Bronx, NY 10458 U. S. A. E-mail: [email protected]

5