Riggs on Epistemic Luck

Wayne Riggs has a new (or at least new to me) paper up on his website entitled Why Epistemologists are So Down on Their Luck. Take a look!

UPDATE: The paper is a clear instance of value-driven epistemology, which I like, and it argues for a version of the knowledge-as-credit thesis, considering at length Duncan’s views about epistemic luck. On this topic, I think Jennifer presented a paper at the Pacific meeting claiming counterexamples to the thesis here, but I don’t find it online yet.

FURTHER UPDATE: Wayne notes in the comments that the paper is on Jennifer’s website, which I inexplicably overlooked. The paper is here.


Riggs on Epistemic Luck — 6 Comments

  1. Lackey’s counterexamples involved knowledge gained by testimony. As I recall, the idea was that you don’t get lots of credit just by being a good listener.

  2. I would think that the credit comes not from the listening, but from assessing the reliability of one’s source, even checking on the sources reliability, along with assessing how well the claims made by the source fits with one’s other well corroborated beliefs, and trying to find a reasonable amount of independent corroboration for the claims. The more of this one does, the more credit one gets. I hope no one thinks that simply believing whatever you hear is supposed to bring epistemic credit.

  3. Lackey’s point was that in a lot of cases you don’t really need to do much at all of the “assessing for reliability, checking on the source, etc.” Imagine a case like Grice’s original implicature case, where A asks B whether there is a gas station nearby. (A is broken down at the side of the road, and B is the first person to come by.) If B says yes, A can justifiably believe that without checking on B at all. At least in normal cases that is the intuition being pushed. And it’s hard to motivate that this is a case of credit.

  4. Jon–Thanks for the plug!

    Though I did not have the pleasure of attending the APA at which Lackey presented her paper, I did read a copy of it on-line (it’s on her website). I think the problem she raises for the credit-theory of knowledge is a serious one. Indeed, I think it requires a revision of the view to a certain extent.

    Incidentally, I first encountered this problem when I delivered an ancestor of this paper to the good people at the University of Missouri, including of course, our good host Jon Kvanvig. Someone in the audience there (alas, I cannot remember who) compared the epistemic situations of two people–on the one hand a brilliant and dedicated scientist who spends years verifying result X. She clearly knows that X and deserves a great deal of credit for having acquired such knowledge. But suppose in her excitement she goes to a local pub and tells a friend over drinks that she has conclusively established X. Her friend now believes X on the basis of this testimony, and presumably knows it. The scientist and her friend are on an epistemic par with respect to knowing that X, yet they seem clearly not to be on an epistemic par with respect to the credit they deserve in coming to know that X. This is, I think, a slightly more dramatic illustration of (at least one of) the problem(s) that Lackey is pointing to.

    I haven’t worked out the details, but I think I have the outline of a solution to the problem. I tremble to throw such an unfinished product before the epistemic glitterati (episterati?)who frequent this site, but here’s the general idea:

    Knowing that p is primarily a matter of whether or not your having true belief p is attributable to you. This is sometimes put (by me, for instance) as “S knows that p iff S deserves (full or sufficient) epistemic credit for believing-p-when-p-is-true” or something in that neighborhood. What Lackey’s objection points out, I think, is that we have to disambiguate the use of “credit” in that locution. All I mean by it is that the having of true belief p is properly attributable to you. As Sosa puts it, it is your “deed.” But we also use the term “credit” in these instances to indicate the value of the deed.

    Consider a similar pair of cases to the ones presented above, but in the preferred idiom of all credit-oriented epistemologists–sports. Suppose A (this is a variable that ranges over extremely talented professional athletes who play sports that can be played one-on-one) plays two games of S (you get the idea) on two consecutive days. On the first day, he plays a second-string junior high school player. On the second day he plays a talented rival from his league. He wins both games. Barring extenuating circumstances, it seems that A deserves credit for both wins–at least in the sense that both wins are attributable to him rather than to dumb luck or sabotage or what have you. So A’s two performances are on a par in this sense. Yet they are also importantly different. A deserves more praise for beating his rival than for beating the junior high player. We might say that he deserves more credit for the one win than for the other. But this does not imply that somehow the win against the junior high player was less attributable to him.

    What this shows is that, while the value of X’s occuring accrues to an agent A only when and to the extent that X is attributable to A, the praiseworthiness appropriate to A for X is a function of more than just the degree of attributability of X to A. Thus, the scientist and her friend both have the true belief that X, and in both cases this is attributable to them. So in terms of meeting the conditions of knowledge, they are on a par. But that is compatible with the scientist’s knowing that X being more valuable, even more epistemically valuable, than her friend’s knowing that X. This is what is reflected, I think, in our unwillingness to say that both the scientist and her friend deserve equal credit, despite the fact that they both meet the credit condition for knowing.

    An interesting consequence of this kind of move is that it complicates the whole question of the value of knowledge. Various proponents of the credit-theory of knowledge have argued that it has the benefit of accounting for the value that knowledge has beyond the value of its consitutuent parts (what Jon calls the ‘Meno problem’). But I now think that the whole issue of the value of knowledge is more complex than that problem makes it seem. Sosa long ago (I forget where) said that trivial bits of knowledge (I think his example was about the number of grains of sand in a pile) are not as valuable as more significant instances of knowledge. Maybe that’s right. If the value of knowing is a function of more than just the extent to which the true belief is attributable to the knower, then that opens up a number of possibilities for different sources of value. Maybe some instances of knowing are more valuable because the content itself is more epistmeically valuable. Maybe a state of knowing gains value when epistemic virtues are employed in its acquisition.

    Well, now that I have moved into the realm of bald speculation, I should stop. But I do think that it is attributability that matters to whether or not one has knowledge, and then other factors can come in to account for the epistemic value of that knowing state. It is these other factors that make some knowing states more valuable than others, even when what is known is the same.

  5. >Knowing that p … is that the having of true belief p is properly attributable
    >to you. As Sosa puts it, it is your �deed.�

    OK, I’ll bite. If I believe that p, and p is true, how could the true belief that p fail to be properly attributable to me? (Who else’s deed would it be?) In other words, why isn’t knowledge = true belief on this model?

  6. It seems to me that if credit comes in degrees, and knowledge comes by way of credit, then knowledge should also come in degrees. (Either that, or the credit view will need a credit threshold for knowledge — which sounds problematic to me.) Perhaps this helps with some of the problem cases.

    In the “gas station case” the strength of your knowledge depends (at least) on how reliable your informant appears to be. (If he stammers, “go that-a-way, then left, then right,” and then he staggers back into the bar, I’d say you don’t get much epistemic credit, even if his directions turn out to be correct.) So even in seemingly simple cases epistemic credit depends on your generally reliable tendency to believe someone who looks to be in his right mind, and seems to have no motive to lie or deceive. You get credit for your ability to make such judgements reliably. (This kind of thing may seem easy — but the AI researchers would give heaps of credit to any automated system with this capability.)

    Perhaps part of the problem is that “credit” is being associated too much with “hard work”. Perhaps one need not “work hard at it” in order to deserve epistemic credit. Rather, it suffices to have “epistemic skill” and to employ it adequately to the given task. Getting the skill *may* take hard work; and properly employing it *may* also require work. But it’s not the amount of work that’s important — it’s the skill and its appropriate application to the task at hand. (If I have to work really hard in order to figure out some theorem that one of my brighter colleagues can see through immediately, I surely don’t get more credit for that.)

Leave a Reply

Your email address will not be published. Required fields are marked *