Ernest Sosa

Ernest Sosa (Rutgers University--New Brunswick) is presenting "Epistemic Normativity." The two invited commentators for his paper are Ram Neta (University of North Carolina--Chapel Hill), and Duncan Pritchard (University of Stirling).

Download ernest_sosa.pdf

Download neta_on_sosa.pdf

Download pritchard_on_sosa.pdf

Comments

[Clayton Littlejohn]

I confess that my own intuitions about this may well be idiosyncratic, but I'm not entirely convinced that aptness can solve the value problem.

I've asked Prof. Sosa about this before, but I thought I'd toss it out there for discussion to see what others might think.

Suppose a friend and I are teammates shooting arrows in an archery tournament. Suppose we're equally skilled archers, we've invested the same effort into training, etc... I shoot my arrow and the arrow strikes the center of the target as it often does. My friend shoots his arrow and it hits the center of the target as it often does. Let's suppose that the target is close enough and we're skilled enough that someone could truthfully say that one of us intentionally hit the bull's eye. An official notifies us that one of our arrows was deflected ever so slightly off course by an unexpected gust of wind and set back on its path by a second unexpected gust of wind. Nobody yet knows which of us intentionally hit the center and which of us merely hit the center as intended. Thus far, we know that we both displayed the same skill, we both achieved what we intended, but one of us was lucky to have done so, and one of us intentionally did so.

Upon learning all this, it makes no sense to me to think 'I hope I'm the one whose shot wasn't influenced by the wind'. I don't think it would be better or worse from my perspective whether it was my arrow whose path was twice influenced by the wind or whether it was my friend's. In effect, that is to say that I can't see why either agent looking back on the situation would care if theirs was the shot that was apt. But this is odd. If there were a sense in which one of these shots is better than the other, shouldn't it matter to the agents whose shots those were?

I suppose someone could say that they'd prefer to be in the situation of the archer whose arrows are not affected by the wind at all, but the natural way to understand this thought is to say that the reason for this preference is simply that the odds of success are higher when you don't have to rely on these sorts of variables to achieve your end. But this is to explain the preference for the one position over the other in terms of an instrumental value, and I thought that one of the ground rules of this debate is that the value in question is supposed to be something that isn't purely instrumental.

[Simon Rippon]

This is a tremendously interesting paper and related discussion.

In reply to Pritchard's objection of the barn façade case (pp.4-5): This seems to me to be a case where the belief isn't apt because the agent lacks the relevant ability (which includes the ability to discriminate between barns and barn façades). Why, then, is the successful shot in the corresponding archery case so clearly apt? Because what we count as great skill in archery does not include the ability to defeat forcefields! An archer may qualify as greatly skilled even if he were always defeated by forcefields, whereas a believer wouldn't count as greatly epistemically skilled if she were always defeated by the many local barn façades.

Consider another archery case, where the targets are arranged at different distances, and our archer has the peculiarly limited ability to hit a target only from an exact distance of 103 yards. Environmental luck may lead to his choosing the one target that happens to be at 103 yards distance, and to then to hit it through his great, but peculiarly limited, skill. The shot might then count as apt "103-yard-archery", but would not count as apt archery, because our archer would lack a constitutive component of an archer's skill. We might say that apt archery needs to be safe-with-respect-to-the-distance-to-the-target (within reasonable limits), but not safe-with-respect-to-forcefields. In other words, for the apt archery shot, varying the distance to the target a bit would not have made the successful hit unsuccessful, but adding a forcefield around the target might well have.

Apt believing would need to be safe in different respects to apt archery. Our apt believing perhaps needs to be safe-with-respect-to-contextually-relevant-fakes. Arguably, apt believing needs to be safe in every respect, in which case one cannot aptly believe something as a result of environmental luck of the sort Pritchard discusses. But if this is a possibility, then the apt belief account can properly account for knowledge, even if knowledge is understood as "of its nature" safe.

(As I side note, I doubt that knowledge or apt belief needs to be safe in every respect. In particular, it seems implausible that knowledge needs to be safe-with-respect-to-demonic-interference. In the case where I have a true belief as a result of my epistemic skill, I believe that I would have knowledge even if I were extraordinarily lucky that a demon did not intervene on that particular occasion, as he has on every other.)

[Ernest Sosa]

My grateful thanks for the posts on this thread, which give me much to ponder. Special thanks to Ram Neta and Duncan Pritchard for the very interesting challenges contained in their extensive replies.

Ram makes two main points on which I'd like to comment. (He himself replies to his third point.)

First, Ram holds that a "reliabilist account of epistemic justification is not refuted, or in any way disconfirmed, by the value problem. That's because this reliabilist account of epistemic justification predicts nothing at all about what is good, epistemically or otherwise. It speaks only of epistemic justification, not of epistemic goodness."

This will depend, however, on what counts as "discomfirmation." If some reliabilist account of justification cannot well explain how it is that a justified belief that p is as such better than would be an unjustified belief that p, and if another account of justification can give a better explanation, the latter account acquires thereby an advantage over the former. Even if this does not constitute "discomfirmation" in some strict sense, it does speak against that former, reliabilist, account.

Second, Ram doubts that beliefs are epistemically "good." He tells us that predicating "is a good belief" of a belief is perhaps unacceptable, and is in any case "not at all natural." Moreover, it is "not at all clear" how to understand such predications. While granting that it would be natural to say "it's good for you to believe that" or "it's good to believe that," Ram thinks that such sentences as these last two are not about the epistemic properties of a belief, but rather about the "typical or actual practical or moral consequences of holding the belief." He remarks that the value problem "is the problem of understanding why beliefs that are true because they're justified are any better epistemically than beliefs that are merely true." But, for the reasons suggested, he does not "understand" this so-called problem.

The much discussed value problem is, more exactly, the problem of understanding why beliefs that constitute knowledge are as such better than corresponding beliefs that are true but fall short of knowledge. And by extension there is the value problem of why a justified true belief is better than one that is true without being justified. One might wonder, as does Plato, why knowledge should be considered practically or instrumentally better than the corresponding true belief that falls short of being knowledge. Even if we put aside practical advantage, however, it is surely better to know than to believe with truth but without knowing. How do we understand this nonpractical, epistemic added value? Suppose that, given one's evidence and the rest of one's epistemic position, "P is a good thing for one to believe." I myself find it incredible that this sentence could only be about what would be good practically. Even sticking just to epistemic assessment, it would be better, surely, to believe in a rationally justified way than in a way that fell seriously short in that regard, and better also to believe with knowledge than to get it right without knowing. If it were thought more natural, we could switch to talk of someone's belief being "deplorable" or "ill-formed," or one that "falls short" (of being knowledge) instead of its being a "bad" belief. And this would not be just an assessment of the practical value of that belief. Even once we put practical value aside, if a belief were deplorably irrational or ill-formed, or if it fell short of being knowledge, by contrast with another, then the latter would be better at least in that regard, on an epistemic (non-practical) dimension. Take then a belief that is rational and indeed constitutes knowledge, and compare it with someone else's belief in that same proposition, so that it too is a true belief, but suppose this latter belief is deplorably formed epistemically. If we say that the former belief is better than the latter even though both are true, this is enough to pose a value problem for truth monism.

Duncan raises two main objections to my account of knowledge as apt belief. He argues that a belief can be apt without being knowledge, and can be knowledge without being apt. The examples he presents are of sorts that I consider extensively in forthcoming work (in A Virtue Epistemology, OUP 2007), so here let me just sketch the sort of approach I take to each.

First, take a belief that hits the mark essentially through the exercise of what seems a considerable cognitive competence, such as a belief that one sees a barn (while in fake barn country), or that one sees someone one knows (when it could easily have been his twin instead). In these cases does the correctness of the belief manifest a competence to tell or discriminate the true from the false? Arguably, it does not, in which case the belief may be adroit but it is not apt. Let me say a word in defense of this line.

Competences may plausibly be thought not to be wholly intrinsic to the thinker, not always anyhow. On the contrary, some at least might well be thought to encompass his relevant environment. Commonsense is conflicted on how and to what extent dispositions are extrinsic and encompass the environment. We say that in outer space astronauts are weightless, yet we are reluctant to say that going to outer space is a way to lose weight. (No commercial potential here!) Do we lose our normal competence to discriminate when we enter an epistemically problematic environment, such as fake barn country? Perhaps we do, just as we lose our ability to jump so high when we return to Earth from the moon. In that case, we are not clearly believing aptly in the barn or twin cases. (This is in line with Simon Rippon's comments elsewhere on this thread.)

But there is a second way to respond to Duncan's concern, which is compatible with taking the subject to retain his competence even in the bad environment. He still retains his ability to identify a barn by sight, since he retains his 20/20 vision, and has a clear view in good light at a good distance, etc. On this alternative way to view the case, the fact that the subject could easily have gone wrong if he had looked at any nearby barn does not take away his ability to tell that the barn he sees is a barn. The fact that nearby facades are unattached to real barns is like the fact that a gust of wind might easily have come along when the good archer hits the mark through the exercise of his competence when the air is still. Very easily one might have believed the de dicto, Fregean proposition that one now believes, that one faces a real barn, while one faced only a barnless facade, just as when the wind is nearby, very easily one might have shot without success. So, on this alternative view we'd say that the subject in fake barn country does know when he gets it right in the case of a barn-attached facade, just as an archer hits the mark creditably when he hits it with no wind, no matter how easily a gust might have diverted his shot. At least, our subject has knowledge of a sort, of the animal sort, since he has apt belief.

Obviously, we now need to say something about the very powerful and very widely shared intuitions to the contrary. And in fact the subject does lack a second, important sort of knowledge, reflective knowledge. He has apt belief but he lacks apt belief aptly noted. His belief (or assumption or presupposition or implicit taking for granted) that he knows, that he aptly believes, is not itself apt. (This is the solution developed in my forthcoming publication.)

Note that this may well require us to distinguish two sorts of "Gettier" cases, those that preclude even animal knowledge, such as perhaps Gettier's own cases and Lehrer's Nogot/Havit case, and those that preclude only reflective knowledge, such as, arguably, the Ginet/Goldman fake barns case.

About Clayton Littlejohn's lack of preference for apt performance over inapt but accurate performance, I'd suggest the following. If there is variation in the extent to which the success of a performance is due to the competence of the performer, and if accordingly there are degrees to which that success manifests such competence, then aptness is a matter of degree. Take Clayton's case where the gust is very slight and would make the arrow miss but only barely so, so that the second gentle gust easily compensates and puts the arrow back on course. Many of us might still comfortably take a lot of the credit for such success, thinking that the shot was to a high degree apt. But compare the case of an ensuing tornado that picks the arrow up into its whirlwind and eventually by fantastic luck buries its tip in the target. In that case, or at least in many cases intermediate between the gentle diverting wind and the tornado, the "shot" would hit the target. (Even if we say that there is no such thing as the archer's shot that reaches the target in the case of the tornado, there will come a point along the intermediate cases where it is no longer plausible to deny that there is such a shot, one whose success is still very largely due to the luck of the wind and negligibly due to the competence of the archer.) In these cases, few would take much credit for the success, and, correspondingly, many would prefer to have authored the creditably successful shot rather than the lucky success due overwhelmingly to the wind.

Let's turn finally to Duncan's second main problem (which he credits to Jennifer Lackey): that a belief might be knowledge without being apt, without its success in hitting the mark of truth being a manifestation of the subject's competence. The point is basically that when we believe correctly by trusting some testimonial source, our hitting the mark is often largely creditable to the discoverer(s) of the truth and not nearly so much to those later in the testimonial chain, including oneself as final recipient. Duncan concludes that "...it is not the hero alone that is deserving of credit, since it is not because of her cognitive ability that she was cognitively successful. So if we are to allow such cases, it follows that there is sometimes a lot less to knowledge than apt belief." The mistake here is to suppose that apt belief requires that the correctness of the belief be creditable to the believer alone. It does not; apt belief requires only that the correctness of the belief be sufficiently creditable to the believer, not that it be exclusively creditable to her. The believer may be sufficiently creditable for the truth of his belief, even if our knowledge of that truth is not much to his credit. Similarly, the good deeds of someone steeped in a certain culture or religion may be sufficiently creditable to him even if the way of acting (turn the other cheek, or some other) is one that he absorbed from the group or tradition along with much else in a way of life. Of course, these suggestions only point in the direction where I think I see a fuller solution to our interesting problem.

[Ray Cheung]

I find the analogy between delicious coffee and true beliefs interesting. Prof. Sosa mentioned that maybe the fundamental value of delicious coffee depends on other things such as the pleasure it produces. Although this does not seem to be the main concern of the paper, I am just wondering if people would agree that the value of beliefs also relates to pleasure. A case in point: on my way to the airport I wonder whether the plane really departs at 8pm, but not earlier. I check the air ticket, and I feel reassured that it departs at 8pm. The feeling of reassurance upon being more certain about that it departs at 8pm is, arguably, pleasurable. If this is a correct description of the situation, perhaps certainty and pleasure are intimately related.