Wednesday, June 6, 2012

Does 'S knows that P' entail the truth of P?

To: Stuart B.
From: Geoffrey Klempner
Subject: Does 'S knows that P' entail the truth of P?
Date: 25 October 2007 12:06

Dear Stuart,

Thank you for your email of 10 October, with your University of London Epistemology essay in response to the question, 'Is the truth of "P" necessary for S to Know that P?'

I take that this is not a question from past examination papers, but rather prompted by my remark in a recent email that two of my UoL students had argued against the 'Truth Condition'.

This made me smile, because your essay while purporting to defend the 'Truth Condition' in fact does quite the opposite. So now it seems I have three students who are at least prepared to cast doubt on the Truth Condition!

The first point to make is that when we are considering whether there can be an adequate definition of knowledge without TC, we are not simply considering your definition K1, which is the standard tripartite justified true belief definition, minus the Truth Condition. Both students offered a *substitute* for the Truth Condition, which purportedly did the same work, but involved (in their eyes) a less problematic metaphysical commitment. Both were agreed that no-one can 'really know' the truth, so any talk of truth is either empty rhetoric or else requires us to suspend judgement indefinitely.

Would could this third condition be? Well, we are looking for a state of affairs, which can be recognized when it obtains (or recognized at least some of the time -- this is up for debate) in which S believes that P and has good justification for believing that P but, in our view, is wrong about P. That is how we are able to say things like, 'S thinks she knows, but she's wrong'.

(I wonder whether you can see where this is going?)

You make a very good case for the necessity of a third condition. Let's consider what you say:
The truth-status of P, while evidence-transcendent for any one party, is asymptotically approachable by a multitude over time... The truth-condition of the JTB model is intended to capture this asymptotically approachable limit... Alice's claim is not just that she has judged her justification sufficient for a claim to knowledge, but she has also warranted that as the asymptotic limit of evidence is approached by a relevant population over time, she will not be proved wrong.

It could be objected that talk of 'asymptotic limits' is a bit obscure. No-one has ever seen this limit, nor does it play any real role in everyday discourse. What does play a role, is the idea of 'community agreement', albeit extended over time but not necessarily indefinitely.

So how about this:

S knows that P iff (a) S believes that P, (b) S is justified in believing that P, (c) investigation carried out by S's linguistic community converges in agreement on P.

If we are unsure whether investigations do, or will, converge in agreement on P, then we are correspondingly unsure about the truth of the claim that S knows that P. We can consider the hypothesis that everyone whose views are canvassed believes that S knows that P, although ('unknown' to us at the present time) further investigation will converge in agreement on the falsity of P. And so on.

The response to this from a defender of the Truth Condition is that truth is not 'agreement' or an 'asymptotic limit'. Truth is truth. The disquotational or redundancy theory is all that is needed to explain what we mean. If S knows that P, then P. How much evidence we may gather for or against the proposition that S knows that P, at any given time is another matter.

One can raise the question whether we are being 'realist' or 'anti-realist' about truth. But in fact, it is irrelevant (as I think you see) to the logic of the claim that S knows that P.

In terms of the logic of the claim that S knows that P -- the logic of ''P' is True' -- we can state that S believes that P, is justified in believing that P, and moreover all the members of S's linguistic community do, and will forever more believe that 'P' is true. But, in fact, we know something that S's linguistic community do not know: e.g. they are in the Matrix and we are the scientists twiddling the knobs, or whatever.

Or to put the matter another way, you cannot define 'P' is true in terms of justifications, however protracted, or agreement by however many persons you like. (In the Metaphysics program I argue that an anti-realist can and should say this: however, there are anti-realists, like Crispin Wright -- in 'Truth and Objectivity' -- who defend the view that truth is in some sense equivalent to 'satisfaction of criteria' in Wittgenstein's sense. That would force me to radically revise what I say here.)

A defender of the Truth Condition, in other words, is defending the tri-partite definition of knowledge as an account of the *truth conditions* for S knows that P. Of course, it is understood that any statement of truth conditions must relate to things that we can discover through investigation. But as a matter of logic truth is always different from what, at any point in time, we happen to believe.

All the best,