Many people liked the quotes board from the APA Pacific, so I should try to provide a similar service for other conferences that I attend. Unfortunately I neither tape papers I’m at, nor take them down in shorthand. So I have to do things from memory. I can only remember two notable quotes from the Rutgers Epistemology Conference. (In neither case do I vouch for the literal accuracy of the quotes, but I will stand by their approximate accuracy.)

Jonathan Vogel: Grue is hard.

Timothy Williamson: We need to recover a philosophical innocence so we can see that the phenomenal conception of evidence is laughable.

Whatever else was in the Williamson quote, the words ‘philosophical innocence’ and ‘laughable’ were memorably used. This is really rather a strong claim. It’s very intuitive to a lot of people, even non-philosophers, that how things seem to us is our evidence about how things are, and so we could have the same evidence as we actually have while being a brain in a vat. In my two anti-indifference papers (here and here) I had to point out that the theorists I was responding to simply assumed that phenomenal states constituted evidence. They thought, not unreasonably I guess, that such a conception of evidence could be simply assumed.

But I don’t want to defend the ‘laughable’ theory here. What I do want to point out are some ways in which Williamson’s response goes well beyond what might be a natural reaction against the phenomenal conception. I was reminded of this a bit by reading Quine’s Roots of Reference, which starts by basically dismissing the phenomenal conception of evidence as being based in a faulty metaphysics of mind, but then quickly insists that evidence is still a ‘local’ matter, being constituted by sensory irritations. For Williamson locality goes with phenomenality. He thinks our evidence is just what we know.

At the conference, John Hawthorne noted that Williamson said some odd things about the following three cases. (I had previously discussed something similar here – I can’t remember whether I got the important example from John or somehow else.)

1. I hallucinate a gas guage showing a full tank.
2. I see a well-functioning gas guage showing a full tank.
3. I see a mis-functioning gas guage showing a full tank, but the tank is in fact empty.

On the phenomenal conception, in all 3 cases we have the same evidence. John noted that even if that isn’t plausible, it is very plausible that we have the same evidence in 2 and 3. At the conference, Williamson replied that sometimes we presumably could use the fact that the tank was full in our evidence. But presumably we could also do that sometimes in case 4 (which John didn’t discuss).

4. I see a mis-functioning gas guage showing a full tank. By coincidence, the tank is in fact full, but the guage would have said this even if it were empty.

Williamson is committed to the claim that we have different evidence in case 2 to case 4. And that’s very hard to motivate.

The idea that evidence is local can be motivated further by a case that Stephen Yablo used at the conference.

Knowing that the mine contains gold, you keep digging until you find it … [Y]ou would not have known the mine contained gold if some misleading testimony given in Carson City — testimony you should have been aware of but weren’t — had not been refuted by court records in Reno — with you again unaware of the fact. The court records in Reno play no role in your continued digging here. But they are a factor in your knowing.

If the Reno judge had been on a bender last night and the court had not opened, the documents refuting the Carson City testimony would not have been entered into the court records. If that had happened, and those documents had not become part of the public record some other way (which we’ll assume they would not have) then you would not have known the mine contained gold. In short, what you know about the mine turns on how much the Reno judge drank last night. But if you’re far enough from Reno, if you are in fact down the mine digging, what evidence you have really does not depend on how much the Reno judge drank. This is pretty bad for the equation of evidence with knowledge I think, especially if we are to follow Yablo and tradition on the influence of social factors on knowledge.

The upshot is that if evidence isn’t constituted by phenomenal states, I think it’s best to go with Quine and say that it’s constituted by sensory irritations. John suggested that we go with something slightly more external, say what is perceived, but either way it’s important to make evidence be closer to home than Williamson does.