Monday, January 12, 2009

Bits about bits of evidence

Good or bad objection to Williamson?

Knowledge has an irreducibly normative element. Evidence, however, ought to be construed in non-normative terms. While facts about evidence might ground certain normative facts, those normative facts are not themselves conditions that ground facts about evidence. Thus, evidence does not require knowledge. It might require acquaintance, awareness, or some other such thing that is similar to knowledge. However, if we focus on, say, inferential knowledge it seems that there's more involved to that sort of knowledge than there is to evidence.

In stating this objection, I've been intentionally ambiguous between two ways of talking about evidence. There's the evidence you have and there's the evidence there is. I don't think E = K is intended as an account of the evidence there is. It's instead intended as the evidence you have. (I think, but correct me if I'm wrong.) So understood, it seems subject to the following sort of counterexample:

(D) Suppose I believe p on the basis of the apparent perceptual experience that p but fail to know p simply because I know q: that there are operative devices in my area that tend to create the false appearance that p.

'D' is for defeat. I don't think it's wrong to say that p is part of my evidence or the evidence I have. I can't rightly use that evidence, but I don't feel the intuitive pull towards saying that p is in no way included in my evidence. Don't expect an argument for that, it's just an intuition. What I don't think is crazy is that we distinguish between the evidence a subject possesses and the evidence that the subject can rightly reason from once we take account of the possessed defeaters the subject has. And, if we can make sense of that distinction, it should lend some credence to the idea that the lack of a right to reason from p is no sure indication that p is missing from someone's evidence.

The objection first occurred to me in thinking up possible objections to the following principle linking justified belief and evidence:

JE: If S's belief that p is justified, S's evidence includes the proposition that p.

It seems to put the cart before the horse to move from normative claims about the standing of beliefs to claims that don't appear to be normative about the possession of evidence. There are other objections to JE, besides. That's the next bit about bits of evidence.

Good or bad objection to JE?

Hob and Nob have observed that for 300 days running, at least one egg was waiting for them in the hen house. (Assume they know that there are hundreds of healthy hens in their hen house.) It seems that both Hob and Nob have sufficient justification for the belief that on the 301st day, there will be at least one egg waiting for them in the hen house. Suppose Hob believes that on the 301st day (i.e., the next day for them) there will be at least one egg waiting for them in the hen house. Suppose Nob does not believe this. Before opening the door to the hen house, both Hob and Nob consider the disjunctive proposition that either there will be at least one egg on this day or a witch has cursed their birds. According to JE, Hob can deduce from his evidence that this is so. Nob cannot. That seems weird.

I think there's something like this objection in one of Neta's papers but I have to confess that while this was part of the inspiration for the example, I'm not confident that this was in his paper.

No comments: