Plantinga on defeaters

I’ve been working through Plantinga’s latest volume, and was struck by a feature of his discussion of defeaters. For Plantinga, by definition a defeater prevents one from having knowledge. That means that one and the same belief can be a defeater for one person and not a defeater for another, since one of the two might have additional information that prevents the belief in question from counting against knowledge. Hence, being a defeater is not really a two-place relation between one belief and another, but rather (at least) a three-place relation between a certain belief (or experience), the belief being evaluated, and a noetic structure.

Even if unproblematic, about which I am still thinking, this result is surprising. I can’t see any good reason to develop a view of defeat in this way rather than the more common way.


Plantinga on defeaters — 5 Comments

  1. Well, it does seem to me that some proposition, say d, could be a “defeater for one person and not for another” although I’m not sure I would put it quite that way. Let’s consider the old Grabit case — the one in which Tom Grabit steals a book, but his twin Buck was in the library on the day Tom stole the book. Now consider two people, Jon and Peter. They both have the same experiences on the day in question–i.e., they both see Tom steal the book and they both come to have the justified, true belief that Tom stole the book. Jon, however, is unaware that Tom has a twin, Buck; Peter does know about the twin and knows that Buck recently broke his arm and that, since it is still in a cast, Buck would not be mistaken for Tom. Seems to me where d is “Tom has an identical twin who was in the library on the day in question”, d is a defeater for Jon but not for Peter.

    More generally, I think there are lots of cases in which propositions that did not play a role in the actual production of a belief (e.g., Buck has a broken arm) can prevent a proposition from being a defeater.

    The effect of a defeater can be overcome by a restorer or what can be called a “defeater- eater”. If Peter has a defeater-eater of d in his true belief set and Jon does not, then Peter’s true justified belief can be knowledge whereas Jon’s can’t be knowledge.

    Now, perhaps you could deem the defeater-eater, say o, a relevant bit of evidence for the proposition and hence you could say that the evidence for “tom stole the book” includes “Buck has a broken arm” — but that seems to blur a distinction we might find useful to maintain, namely, that some propositions that played no role in the justification of a belief can play a role in warding off defeaters. Think of a case in which Jon and Peter both have evidence, say e, which is adequate, ceteris paribus, for coming to know that p. Let us also suppose that Peter believes truly that ~o but Jon either believes that o or withholds that o. Now, suppose further that o entails ~p. What are we to say about the true proposition, (o -> ~p)? Couldn’t it plausibly be held that (o -> ~p) is a defeater for Jon but not for Peter?

    In fact, I think one way of reading Descartes’ First Meditation is that he is seeking to find a defeater for “there are extended objects” to which he does not (yet) have a response. He’ll try one, say “my senses have decieved me in the past” and reject it since he also believes that he can distinguish the times when his senses are reliable from those times when they are not. In other words, he has a defeater-eater, namely, “my senses decieve me when things are not presented clearly by the sense faculty, otherwise not.” He has a defeater eater for “I have always been sleeping”, namely, “dream images are distorted copies of the images I obtain while awake” (so there must be some waking images). Folks who do not have a way of distinguishing occasions when the senses are to be trusted from those when they are not, or folks who don’t have Descartes’ account of dreams (i.e., distorted waking images) might not have the relevant defeater-eaters.

    So, I think it is quite easy to imagine circumstances in which e justifies p for S and that e is defeated for S, but that there is an S’ such that e is not defeated for S’. Just imagine that S’ has a defeater-eater of d, say o, in his/her belief set and S does not have a defeater-eater of d in his/her belief set. Now, as mentioend above, you could rule that S’ has an undefeated justification for p, namely (e & o). And I agree that there is such a justification. But that would seem to obscure an important distinction, namely that the blief that o, Bo, played no role in what brought about the belief that p, Bp, for S’. Bo might play a role in sustaining Bp after S’ considers d, but it didn’t beforehand. And I suppose we want to say that S’ knew that p before considering d.

  2. Very useful, Peter, but in your case, I wonder why you would say, as you do in the first paragraph, that the claim in question is a defeater for Jon but not for Peter. Why don’t you say it’s a defeater in both cases, but in one case, there’s also a defeater-eater present, so that the defeater doesn’t undermine knowledge. The straightforward way to construe what a defeater-eater is just that: there must be something to be eaten!

    I agree completely, though, that the defeater-eater should not be construed as part of the justification of the relevant belief, even though it is part of the explanation of why Peter has knowledge. But that is consistent with claiming that both for both Peter and Jon there is a defeater present. It’s just that in Peter’s case, the defeater is eaten by further information that Peter possesses, whereas that is not so for Jon (I feel abused by this way of constructing the example–Peter ends up so epistemically superior to Jon!).

  3. Okay. Good thought. You could do it that way, but then we’ll have to revise the definition of knowledge as true, justified and undefeated belief, because in the case in question, Peter’s true, justified BUT defeated belief, would still count as knowledge. Perhaps, nothing at all is at stake here if we grant that true, justified but defeated, belief can count as knowledge as long as we then say that knowlege is “ultimately undefeated,” true, justified belief.

    Put another way: The primary target is to give the “right” analysis of knowledge. The notion of “defeat” is a philosopher’s invention. (That’s not a criticism — anymore than it would be a criticism of Newton to point out that “acceleration” became a technical term not equivalent to the ordinary notion of change in velocity.) So, how one characterizes ‘defeat’ should be evaluated by the success of the general account of knowledge in which ‘defeat’ plays a role.

  4. Perhaps this is just verbal, but I’m much more comfortable with saying that a believed proposition, or a sufficiently justified believed proposition or a proposition that is believed to be sufficiently justified and is believed could be a defeater for someone S but not for S’. I also think it’s odd to think of all defeaters as defeaters of knowledge–in the sense that they prevent one from having knowledge. Why not think that a defeater may be a partial defeater? This might work two (three perhaps) ways. Suppose my belief that p and my belief that q (or justified beliefs..whatever) each are accessible via introspection to me and each gives me sufficient justification for my belief that r such that I know that r. Now, suppose I (justifiedly) come to believe that p is false, or that my justification for believing that p is no good–whatever. It doesn’t seem nutty to me to say that you have a defeater for your belief that r, though you still know that r. Or, suppose that in addition to having a good undercutting or rebutting defeater for p, you have a moderately-strong defeater for q. The amount of justification for your belief that q is reduced. But it’s not reduced to the point where it can’t provide sufficient justification for your belief that r such that you know that r. So, here, you might have two defeaters for your belief that r. We also might have a defeater that works directly on r, giving us some reason to think that r is false (and, unless we’re noetically challenged, reason to doubt either the epistemic connection between p and r and/or q and r; or reason to doubt the truth of p or q.) I probably shouldn’t spout because I don’t have Al’s book with me; perhaps he’s already said this sort of thing.


  5. Matt, these are very good ideas. I think you’re right that something like a defeater ought to come in degrees. Perhaps we could call them “diminishers,” and reserve the concept of defeat for a diminisher that is sufficiently strong to undermine knowledge. Even if that issue is purely verbal, which it may be, you’re right that there is a substantive point here that is not sufficiently attended to.

Leave a Reply

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