To: Julian P.
From: Geoffrey Klempner
Subject: Scepticism, possible worlds and brains in vats
Date: 14 April 2005 13:03

Dear Julian,

Thank you for your email of 5 April, with your University of London essay on Scepticism.

As I'm sure you already know (!) this is a superb piece of work. And because of the ground it covers, the work will serve you well when it comes to the exam.

I'm not the least convinced by any of the arguments. Or, let me rephrase that. The arguments, for the most part, are fine. The problem is that your case depends ultimately on two key assumptions.

Here are the relevant extract regarding the first assumption:

'Assuming you are not a BIV, all of these BIV PW are 'far away' from the actual world. The possibility that you are not reading this essay because you are a BIV is far more remote than the possibility that you are not reading this essay because you are doing the gardening. But the question of whether your belief that you are reading this essay is knowledge is settled by appeal to the nearest PW in which p becomes false, or continues to be true. So the BIV PW are simply irrelevant to that question. So according to CToK, if you are not a BIV, you know that you are reading this essay.'

The idea of a conditional theory of knowledge is initially plausible because we take one step back from our own situation and consider what one would say about another subject. A form of 'externalism' (call it 'epistemological externalism') is built in from the start. Nozick's conditions give a good guide (with refinements where necessary) to making accurate judgements whether or not a given subject knows that P, as you ably demonstrate.

The problem, the traditional sceptic will say, is that I'm not interested in how I would judge whether or not JP knows whether he is reading this essay. Given that JP is not a BIV, JP assuredly does know. If JP were a BIV he would not. But my question is, how do I know that I am reading this essay? (That was the question you asked, wasn't it?)

It seems reasonable to assume that if I am reading this essay, then possible worlds in which I am not reading this essay because I decided to spend more time adjusting my webcam are much closer than possible worlds in which I am not reading this essay because I am a BIV. What are the grounds for that assumption? According to David Lewis, we judge the nearness of possible worlds by using a notion of 'similarity'. But on what basis are such judgements made? The conditional theory, when used as a response to scepticism, assumes what one is trying to prove.

Let's say, unknown to you or me, a secret government department has been bulk manufacturing brain-vats, and all the disappearances we heard on the news recently were actually kidnappings undertaken by MI6 for the research project. If you knew that, how would you assess the nearness of the BIV world? The point of this example is that judgements of 'nearness', 'similarity', 'probability' are relative to evidence. As a sceptic, my problem is that I don't know whether I am in possession of sufficient evidence to make a reliable judgement.

Now, onto the different question how I know I am not a BIV:

'In the possible worlds in which you are a BIV, you no longer believe that you are not a BIV; you cannot believe that, because you cannot refer to 'BIV'. Instead, you would believe truly that you are a non-BIV*. So there is no PW in which it is false that you are a non-BIV, but you still believe that you are a non-BIV.'

I am not going to pick a quarrel with Putnam's semantic externalism. Once again it is epistemological externalism which is my target.

If anyone (including GK) is not a BIV then he knows that he is not a BIV. If.

If anyone is a BIV, maybe they 'know' something or maybe nothing, who the hell cares?

But how is that supposed to help me in deciding whether I am really reading this essay or not? As the Matrix shows, there are ways in which you can discover you are a BIV, or something similar, but obviously I can't rely on the fact that none of these 'clues' has turned up. (I recall that you suggested one, in a recent exchange.)

Now, you'll probably say that this line of objection confuses knowledge with certainty. We can have knowledge without certainty, and even with a large amount of doubt (as in the nervous schoolboy example - that old chestnut). But that objection itself presupposes epistemological externalism.

What is it to claim knowledge? That is a different question from what is it to attribute knowledge. The conditional account explains how we make judgements which attribute knowledge. But any claim that any person makes at any time, including claims about what this or that person knows, is a claim to know. 'You can take it from ME. I am a reliable authority with regard to this question.'

This is the question which traditional scepticism is fixated upon. Who says I have this authority? Who can help me, if I allow a doubt to creep in? But without authority, how can I ever make any statement to anyone at any time?

The only answer is, 'We just do.' - That would be the beginning of my response. But that's another story.

All the best,

Geoffrey