Tuesday, June 9, 2009

Evidence and Armchair Access (Revisited)

There's a tension between these two claims:

(Armchair Access): It is sometimes the case that: one’s evidence includes some
proposition E, and one knows from the armchair that one’s evidence includes E.
(Evidential Externalism): It’s possible that: A and B are internal twins and A
and B do not have the same evidence.

Suppose A knows p and a specific version of EE is true, E = K. If p is part of A's evidence, by AA, A is in a position to know from the armchair that p is part of A's evidence. If A knowss EE from the armchair, then A can deduce that p is true given just armchair knowledge. Thus, we have to deny AA, EE, or deny that these claims can be known from the armchair.

I don't know why this just dawned on me, but this just dawned on me. Suppose EE is false. Assume:

(Evidential Internalism): It’s impossible that: A and B are internal twins and A
and B do not have the same evidence.

Suppose someone tells A that she has been 'slow switched'. A knows that if this hypothesis is true, her evidence would be different than it would have been had she not been slow switched. A knows that she cannot tell from the armchair whether this is a hoax or not. So, A doesn't know from the armchair what her evidence is.

Suppose no one tells B that she has been slow switched. B reasons as follows. If I had been slow switched, I wouldn't know from the armchair what my evidence is. But, I do know from the armchair what my evidence is. Thus, I know from the armchair that I haven't been slow switched. But, that's absurd. If AA is false whether EE or EI is true, since EE or EI must be true, AA must be false.

I know there are details to tidy up, but this seems rightish.


Aidan said...

Hey Clayton,

Could you please say a little more about this claim?:

'If p is part of A's evidence, by AA, A is in a position to know from the armchair that p is part of A's evidence.'

As stated AA is a 'sometimes' claim, so it by itself won't entail that A is able to know that p is part of A's evidence from the armchair. Am I missing something, or are you just stipulating that this is one of the cases (the existence of which AA commits us to) in which one can know that some proposition that is part of one's evidence is part of one's evidence?

Clayton said...

Hey Aidan

Good question. The 'sometimes' claim makes AA hard to evaluate, so my initial response is to stipulate. I'd like to argue that if a defender of AA plays the 'sometimes' card, I can then try to show that 'sometimes' will really mean 'rarely if ever' and that should be enough to make people cool to the principle. My strategy for doing this is to run the sort of argument hinted at towards the end of the post.

Pavel Davydov said...


Two points. First, I don't think it'll do to respond to Aidan's complaint by stipulating that AA holds "rarely, if ever". Once we accept IE, it suffices to block the slow-switching argument that AA holds for all first-order evidence, but not for higher-order evidence, such as evidence (partly) about one's evidentiary state.

Second, I think it's not quite correct to say that E=K is a "version" of EE. In absence of supplemental assumptions about the nature of knowledge, E=K is compatible with IE. A disjunctivist, for instance, would be able to deny that E=K is a version of EE, as would an internalist about knowledge. Both could (or should) deny EE, but could in principle accept E=K.

Clayton said...

Hey Pavel,

I think you're right about E = K and EE. I'm sort of following Silins' lead here and ignoring some important assumptions. (A skeptic, for example, who thought that we knew nothing about the external world could easily reconcile E = K with EI.)

I'm not sure I'm reading your other point right because it's late. Are you suggesting that someone who adopts EI can say both that we have armchair access to our first order evidence and the evidence that concerns that first order evidence? That might be, but I'll have to think more about it.