Certain Doubts

I’m working on a paper about epistemic possibility, and I need some help. Can someone tell me what epistemic possibility is? (This is a trick question.) To show why this is puzzling, here is a ‘problem case’:

S, a person of ordinary mathematical abilities, performs a complex mathematical calculation, which leads to a certain mathematical claim, P. S and I then have the following dialogue:

S: P.
Me: That was a pretty difficult calculation, and you’ve made mistakes before. Could you be wrong about P?
S: Yes. I could be wrong in thinking that P.

S speaks wisely here. But, if P is true, it is necessary in (what is usually called) the strongest sense: logical necessity. So in what sense could S be wrong? Well, it’s epistemically possible that S is wrong: yes, but how is that to be interpreted, such that it yields a kind of possibility broader than logical possibility?

Somewhere between the second and third editions of Theory of Knowledge, Chisholm changed his account of the epistemic goal, from believing the truth and avoiding the false to believing rationally and not believing the irrational. I never could quite see what the value was in the change (except some unremarkable ways to represent the first idea as requiring us to identify justified beliefs with true ones).

I now think I can find a reason. Suppose we recharacterize the second idea of the epistemic goal as believing in accord with what the totality of one’s evidence confirms. We will define the concept of evidence, and confirmation, as truth-related in some sense or other (which I’ll leave unspecified for now), so the truth-related appeal of the first characterization doesn’t disappear entirely.

So here’s the reason.

(more…)

In an earlier post, I quote Williamson’s argument why ‘wrong’ is not a contextual term:

In cases of decision-making, one context is distinguished above all others: that of the agent at the moment of action. The primary question is whether the sentence ‘It would be wrong for me not to resign’ expresses a truth as uttered in the context in which the speaker is Clare and the time is that for resigning if she is to do so at all. Call that context the agent’s context, and the proposition which the relevant first person present tense sentence (such as ‘It would be wrong for me not to resign’) expresses in it the agent’s proposition. If at some point in her agonizing Clare uses the sentence to express a proposition other than the agent’s proposition, which might fail to match the agent’s proposition in truth-value, she is no longer concentrating on the relevant practical problem. Similarly, an external commentator who uses the sentence ‘It would be wrong for Clare not to resign’ to express a proposition other than the agent’s proposition is no longer concentrating on the relevant practical problem for Clare. But if the sentence expresses the agent’s proposition in all the contexts at issue, then it expresses the same proposition in all those contexts, and contextualism fails for this case.

The distinguished context, according to Williamson, is one that defines the relevant practical problem for Clare, and because of the existence of such a distinguished context, contextualism purportedly fails for ‘wrong’. I’ve also noted that this argument can be transposed into the epistemic context by delineating a distinguished epistemic problem in the way that, e.g., Foley does in defining the notion of egocentric rationality. So if Williamson’s argument is sound, and Foley’s position well-taken, we should expect contextualism to fail for the concept of egocentric rationality.

For most versions of contextualism, I think this conclusion is correct, but there is a caveat: Ralph Wedgwood’s contextualism differs from other versions in a way that makes this anti-contextualist conclusion precarious.

(more…)

Some of the recent discussion of the Gettier problem has suggested that we simply abandon the intuitions that Gettier cases are not cases of knowledge. The plausibility of doing so depends on how obvious it is, regarding a particular case, that it is a case of knowledge. On this score, some of the examples in the literature are less plausible than others, and it might be interesting to find out what epistemologists think are the least compelling.

Among the least compelling, as I see it, are some of the social cases, such as Harman’s Jill/assassination case. Harman questions how Jill could know since she lacks some evidence that all those around her have. Maybe that’s right, but it must also be noticed that the evidence all those around her have is concocted and misleading evidence. I don’t find this point compelling against the case, but it makes me suspicious of it.

Lycan has an explanation in Judgement and Justification for why he is suspicious of both this case and the fake barn case, but I don’t remember the explanation and don’t have the book here at home to include the explanation. But it’s worth noting that Bill views both of these cases as not compelling. Are there other cases that epistemologists view suspiciously?

Here’s a problematic line of thought about the main competitors in confirmation theory and Bayesianism in particular, arising from Branden Fitelson’s nice paper here. Consider the following theories of confirmation. First, the law of likelihood preferred by likelihoodists such as Sober:

(LL) Evidence E favors hypothesis H1 over hypothesis H2 if and only if H1 confers greater probability on E than H2 does.

Second, he weak law of likelihood:
(WLL) Evidence E favors hypothesis H1 over hypothesis H2 if Pr(E |H1) > Pr(E |H2) and Pr(E |~H1) ≤ Pr(E |~H2).

Third, an old-fashioned view:
(‡) E favors H1 over H2 if and only if Pr(H1 |E) > Pr(H2 |E).

We can eliminate the third one quickly.

(more…)

I’ve been reading some of Branden Fitelson’s bayesian papers recently, (one in particular on the issues between likelihoodists and bayesians is fantastic), and came across a review by him, Stephens, and Sober of Dembski’s book on design theories (F/S/S, for short). The review is long as far as reviews ordinarily go–15 pages or so–and is an excellent source for the inadequacies of Dembski’s metaphilosophical position on when design explanations should be accepted.

The piece has a really interesting passage, however, one that contains a common mistake that philosophers tend to make. The mistake is this: Philosopher X says p is true; philosopher Y isn’t convinced that p is true, but has no direct argument to show that p is false; so, Y attacks some generalization that implies p instead.

My favorite example of this tendency among philosophers occurs in conversations with my former colleague Michael about where to go to lunch. “Where should we eat?” “I don’t know, how about Shakespeare’s?” “Why do you want to go there?” “No special reason, we just haven’t been there in awhile.” “Do you always want to go to lunch at places where you haven’t been in awhile?”

(more…)

Just a quick note on spam here. I’ve put in place a filter that stops spam, and sometimes when you comment (as happened to Keith this morning), your comment gets queued as spam that I have to approve before it appears. I won’t say how this works but if you try to post a comment and nothing appears immediately, that’s likely the reason. To check, email me since I approve as soon as I see them (occasionally I have to teach, so don’t approve immediately–and sometimes I sleep, too…).

In an earlier post, I quoted the following passage in which Williamson defends anti-contextualism about the concept of what is right or wrong:

In cases of decision-making, one context is distinguished above all others: that of the agent at the moment of action. The primary question is whether the sentence ‘It would be wrong for me not to resign’ expresses a truth as uttered in the context in which the speaker is Clare and the time is that for resigning if she is to do so at all. Call that context the agent’s context, and the proposition which the relevant first person present tense sentence (such as ‘It would be wrong for me not to resign’) expresses in it the agent’s proposition. If at some point in her agonizing Clare uses the sentence to express a proposition other than the agent’s proposition, which might fail to match the agent’s proposition in truth-value, she is no longer concentrating on the relevant practical problem. Similarly, an external commentator who uses the sentence ‘It would be wrong for Clare not to resign’ to express a proposition other than the agent’s proposition is no longer concentrating on the relevant practical problem for Clare. But if the sentence expresses the agent’s proposition in all the contexts at issue, then it expresses the same proposition in all those contexts, and contextualism fails for this case.

This passage employs the concept of what is distinguished, which context robs others of the autonomy needed for contextualism to be true about right and wrong. Williamson characterizes this distinguished context in various ways here: it is the “primary question”, it is the context that delineates the “relevant practical problem”. It is easy to gloss these three different characterizations in value terms: what really matters, given the particular situation Claire finds herself in, is which agent proposition is true; nothing else could be important enough to override the importance of this question for Claire, and anyone else should recognize this fact. So contextualism about the terms of this problem–e.g., the concepts of right and wrong–is false.

Notice what happens in epistemology if we accept such an argument. It doesn’t follow that contextualism about knowledge is false, but something very interesting does follow. Underneath every version of contextualism should be something akin to the question Claire faces. There will be some distinguished epistemic problem which defines the primary epistemic question for the person involved, and this issue will make the agent’s context distinguished in such a way that the agent’s context robs other contexts of the autonomy needed for contextualism to be true about this particular epistemic concept. Call this concept the noncontextualist concept (I’m being sloppy metaphysically here, since the issue isn’t really one about concepts at all; but it’s easiest to talk this way, so I’ll leave it for now).

(more…)

Here’s a rather lighter issue than the recent posts, suitable for (some of our) weekend activities, an interesting epistemic issue in ordinary life: cases in which one disagrees with sports officials. I’ll use a baseball example, one where the umpire calls a player out at home, and you’re livid since the guy was obviously safe.

Here’s the facts relevant to the epistemic assessment of your opinion. First, you’re not entitled to think that you’re in a better position to observe the play than the umpire–there’s a reason they put them on the field, close to the action. Second, you’re probably not entitled to think that your eyesight is better than theirs. There will be special cases where you are entitled to this opinion, but not in the most common case. Third, you’re clearly not entitled to the opinion that you are better trained to observe such situations accurately, since the officials have had years of training before being hired for their jobs. Last, you also know that umpires make mistakes, and you know that you make mistakes.

Now, suppose you know all of these things to be true. You still hold that the guy was obviously safe. And I suppose you think your belief is rational, too. How can that be?

Here, in “Knowledge, Context and the Agent’s Point of View,” Timothy Williamson defends a view surprisingly akin to Foley’s view. Williamson defends the view as a response to contextualism, and begins by using the notion of an action being wrong, arguing that the agent’s context trumps the autonomy of any other context (when autonomy of other contexts is lost, the standards of the trumping context govern ascriptions in the non-autonomous context). He uses the following example:

(more…)