To: Stuart B.
From: Geoffrey Klempner
Subject: Gettier on knowledge as justified true belief
Date: 21 December 2006 13:07

Dear Stuart,

Thank you for your email of 14 December, with your University of London essay in response to the question, ''Knowing that P is at least a matter of having a belief that P which is both true and justified.' Is this an adequate definition of knowledge? If not, how should it be improved?'

My first reaction is that it is a bit of a 'trick' question. The correct response is to say that 'A is at least B, C and D' can never be an adequate definition of A because in order to have a definition you need necessary *and* sufficient conditions. That would be my first sentence.

Now the question splits into two parts. Are truth and justification each necessary for knowledge? (discussed in our previous exchange); and are the conditions put forward jointly sufficient as well as severally necessary?

I was momentarily alarmed when you stated, 'this single sentence does not provide an adequate understanding of just what is meant by either 'true' or 'justified',' because it looked as if you were going to base your essay around this point. If the question had said, 'A causing B is at least a matter of A occurring before B in time', it is obviously irrelevant to object that 'the simple sentence offers no guidance as to which notion of 'time' is being referred to.'

However, as you go on to show with your exhaustive survey of responses to Gettier, we are in fact very much concerned with different notions of 'justification', if not of 'truth'.

In a couple of places you allude to examples without describing them. I haven't actually come across the 'assassination' example (or if I did, I have forgotten) so that makes it difficult for me to follow this point. As I may have said before, in an exam you sometimes have to do this because time is so short, but a general rule is that you should at least attempt to describe the example, however briefly. The same applies to your statement, under the heading 'No-False-Inference', 'But like all other proposed Gettier solutions, counter-examples have been offered for this one as well.'

Your decompositional theory is initially interesting and looks to be your own contribution to the debate. Whenever you do this in an essay, it is worth while stating clearly that this is your idea and not something that you have come across - otherwise the only clue to its originality the examiner has is that the idea is unsound! (If your idea is sound and original then that would be in most cases sufficient for a 'first'.)

However, after the initial response, 'this looks interesting', my next thought (following very close behind) about the compositional theory was that it can't work. You can dispose of the 'lazy' Gettier examples which rely on disjunction, but as you yourself appear to note it doesn't stop there. Each new 'reduced' claim is open to new counter-examples and you are pushed back and back until the only things that you can claim are things which you have infallible knowledge of.

The question your theory raises is this. Forget about the explicit logical structure of language, what are the real 'atoms' of knowledge? For example, if I say that the mug is on the table, would an 'atomic' knowledge claim be the precise position of the mug? (but how can one tell just by looking?) or the position within margins of error allowed for in the absence of a ruler? or...?

The compositional theory is still a very good point to make, even if turns out that the objections are insuperable. Just say, 'One might be tempted to think... but...'. There's always a way to turn things around.

Incidentally, my objection to your theory is an example of a philosophical judgement which one makes without necessarily 'thinking things through'. I just 'know' that the theory 'can't work'. But could I be missing something? Suppose it turns out that I'm right (after exhaustively examining all the possibilities). Was my initial judgement a case of philosophical knowledge? (If I'm wrong, tell me. But briefly please!)

Judgement involves a leap. When we judge correctly, and circumstances are appropriate (plug in your favourite theory) then we have knowledge. However, the very fact that there is a leap there is the thing that causes the problems.

My own feeling is that the crucial step is deciding what use we have for a concept of 'knowledge' (cf my previous email). Of course we can claim that Fred has knowledge but be wrong, not because the thing Fred believes is false but because circumstances are such that it would be wrong to describe this as knowledge. If we can never be sure whether Fred has 'knowledge' (or something else, less than knowledge) then the concept of 'knowledge' has no utility. You might as well get rid of it. So any theory must give us a reasonable shot at being able to tell whether Fred 'knows' what he claims to know or not. This is the admittedly pragmatist perspective, I would argue, from which the various proffered alternatives may be judged.

I sympathise with your objections to 'ridiculous complexities'. But I don't think that the allusion to our practice of 'attributing knowledge to animals and pre-intellectual children' is adequate to justify your claim. For me, it would be an open question whether exercises of this kind are ultimately valuable or not. It is really a matter of how one conceives of the point of the activity of 'philosophy'.

All the best,

Geoffrey