Another argument for the JTB analysis

Brian Weatherson offers another argument for his view that knowledge is justified true belief, here. The argument presents some interesting ideas, though in a different direction than intended. What’s interesting about it is its reliance on testimony.

So, consider the following possibility. Take an ordinary gettier situation, such as the one involving ten coins in a pocket. You have a justified true belief that someone in the room has ten coins in their pocket, but you don’t know this to be true. I come in and tell you that I’ve conducted a thorough investigation of the matter, and there really is someone in the room with ten coins in their pocket, adding that it’s not the person you think it is. You then believe that there is someone in the room with ten coins in their pocket on the basis of my testimony, and this basis can be sufficient for knowledge (if, e.g., no new gettierizing occurs). Since the basis of your belief has changed, there is no longer any reason for insisting that you don’t know that someone in the room has ten coins in their pocket.

With that case as background, consider Brian’s argument:

1. S is a speaker and H a hearer such that (a) S JTBs p and S doesn’t know that p (b)S and H know each other to be generally reliable informants (c) S tells H that p, (d) although S is uncertain whether H already knows that p (e) On this basis H comes to believe that p (f) Then H tells S that p (g) S comes to regard H’s testimony as her primary reason for believing that p.
2. H knows that p, since belief based on true testimony from a known to be reliable source constitutes knowledge.
3. After receiving H’s testimony, S knows that p, since belief based on true testimony from a known to be reliable source constitutes knowledge.
4. Before receiving H’s testimony, S knows that p, since H’s testimony doesn’t improve S’s epistemic position, and she knows that p after receiving H’s testimony.
5. Since the existence of H satisfying the conditions in 1 is irrelevant to whether S knows that p, all S’s such that they JTB that p know that p.

I’m interested in the case presented in premise 1 more than I am in the argument itself, since the argument is not very convincing. First, premise 2 is problematic, since you can obtain testimony from a source that is known to be reliable and yet form a belief that doesn’t count as knowledge. Second, premise 3 contains a bit of a slip, since the setup of the case only requires that S come to regard H’s testimony as the primary reason to believe p, and from this it doesn’t follow that S bases her belief that p on H’s testimony. So let’s change the setup so that basing is involved. Even given this change, premise 3 suffers from same fault as premise 2: you can obtain testimony from a source that is known to be reliable and yet form a belief that doesn’t count as knowledge.

Where the argument gets interesting is in the claims involved in premise 4. In particular, premise 4 claims two things:
a. After receiving H’s testimony, S knows that p,
b. S’s epistemic position after hearing H’s testimony isn’t improved.

It’s these points I’m interested in, given the setup in premise 1. Suppose we insist on a connection between the state of being justified and the process of justifying a belief, holding that the factors that justify a belief are those that could be cited in the process of defending, or justifying, a belief. H’s testimony is regarded by S as something to cite in defense of the belief, so given the process/state thesis, H’s testimony as the justifier of the belief. So if S knows that p, this knowledge is constituted in part by H’s testimony.

It is uncomfortable, though, to grant that S can come to know in this way. It looks too circular, too easy, too “boostrappy”. If we endorse these criticisms, then we must either say that S knew all along or that S doesn’t know after coming to base belief on H’s testimony.

Both options have problems. The first one, that S knew all along, succumbs because of the failures of the above argument. Regarding the second option, return to the gettier case with which I began above. Even if you’re gettiered, you can remove this impediment by basing the belief on additional evidence such as testimony. To avoid granting that S knows that p, we’ll have to say that H doesn’t know that p on the basis of S’s testimony, I think. But why think that? It is possible to come to know p by testimony from a source that doesn’t know p, so why not here?

But if we grant that H knows that p, I doubt we can deny that S can come to know p by relying on H’s word. That sounds unacceptable, however, so how do we avoid this conclusion?


Comments

Another argument for the JTB analysis — 2 Comments

  1. Is it possible that it is exactly the circularity that you scent in the air, that is causing the trouble here? There are already a number of generally recognized defeater conditions on the transmission of justification through testimony; e.g., you must not have good reasons to think the informant doesn’t know what they’re talking about, or may seriously be attempting to dissemble, etc. Those are anti-unreliability conditions, but perhaps there’s also an anti-circularity defeater condition looks something like this:

    If S is justified in believing that p on the basis of H’s testimony, then H’s belief is not itself justified on the basis of S’s testimony that p.

    To see that perhaps we do operate with something like that condition in our everyday epistemic practices, suppose that you were S, and I was H, and at some point we decide to evaluate the general status of our shared beliefs; perhaps we are about to collaborate on a paper, and want to make sure we’re on the same wavelength. At some point our exchange would look something like this: “Wait a minute — did you just say that you believed p because _I_ said so? Why, I’m only believing p because _you_ said so!” Wouldn’t we thereupon decide that either we were both unjustified, or that our justification had better come from somewhere else (e.g., by your returning to your original basis)?

  2. Jonathan, yes it certainly is the circularity worry that bothers me about the case, but I don’t think we should rule it out in general. At a quite general level, I think some sort of bootstrapping account of the justification of the reliability of sense perception is going to be required.

    About your specific proposal, I think I’d rather change it so that the circularity issue arises when evaluating the defeasibility condition for knowledge rather than the justification condition. The gettier condition is going to have to be an externalist condition, and since the circularity here is clearly of that kind, it is most at home there.

    Even so, I think the general principle is false. Note that in the case in the post, there is a time difference between when the basing of belief on testimony occurs. So suppose I know that p and you don’t. On Tuesday, we talk and I tell you that p is true, and you come to know it on the basis of my testimony. That afternoon, I am knocked unconscious and in helping to revive me, the question of p comes up. I’m confused, uncertain about it, and you assure me that p is true, and I come to know it again, but this time on the basis of your testimony. I don’t think there’s anything epistemically untoward about this sequence of events.

Leave a Reply

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