To: David U.
From: Geoffrey Klempner
Subject: Objection to the coherentist account of knowledge
Date: 9 February 2006 11:22

Dear David,

Thank you for your email of 30 January with your essay in response to the University of London Epistemology question, 'My beliefs could form a coherent set even if none of them is true, so the coherent account for knowledge must be wrong.'

This is not a bad essay, although there are a number of points on which you could be clearer.

The first thing I would do in answering this question is distinguish between a coherence theory of truth and a coherence theory of knowledge. You can be a 'coherentist' about knowledge without being a coherentist about truth, i.e. if you believe that beliefs are justified through belonging to a coherent set, while also holding a correspondence view of truth.

It would also be possible to hold the reverse view. You can be a foundationalist about knowledge, believing that there needs to be a special class of privileged propositions whose truth is in some sense self-evident or not requiring justification, while also holding that this 'truth' ultimately consists in coherence.

You can make quite a bit of headway in answering this question by exploring the difference between coherentism about truth and coherentism about knowledge and the way that the objection impacts on these two different positions.

'My beliefs could form a coherent set even if none of them is true' looks like a very serious argument against a coherence theory of truth. If truth just *is* coherence, then it is logically impossible for a coherent set of beliefs to be false. The only line of defence for the coherentist about truth would be to argue that a 'set of beliefs' must include everything about which one might form a belief. A belief is true if and only if it coheres with a set of coherent beliefs which is in some sense 'maximal', including everything about which one might form a belief. It would still be possible for a person to have a coherent set of beliefs which were all false, because that person simply didn't ask questions which they should have asked. If they had asked those questions, they would have discovered that the answer which they found could not be made coherent with their other beliefs.

The objection to a coherence theory of justification, by contrast, is that coherence of beliefs is not necessarily a reliable guide to truth, or adequate grounds for truth. Construed in this way, the objection looks weaker than before because it would also apply to a foundationalist account of justification. You can have excellent justification for a belief which nevertheless turns out to be false. Yet we would not conclude that 'justification' of the appropriate kind is not a reliable guide to, or adequate grounds for truth.

Your example of your car getting bumped is a good one. However, it does not seem right to say that a coherentist 'would pass the paint color because it would raise more questions than anything, and instead take a witness testimony.' What you should have said is that the 'simple explanation' is not always the correct one. Suppose the pain that comes off your neighbour's car does match the bump on your car. What we have is seemingly overwhelming evidence that your neighbour was to blame. It is a much more 'simple' explanation than any alternative that might be put forward. Yet it might still turn out to be the case that your neighbour is innocent.

Foundationalists accept that what seems to us to be the 'best explanation' does not always turn out to be the true one. So why is this point more of a problem for the coherentist? It looks as though the coherentist is committed to the view that the 'best explanation' must ultimately be the true one. But this only holds if our explanation takes everything into account. In the real world, explanations are never like that, because there is always a limited amount of evidence or data available.

One remaining problem to deal with, however, is the issue of 'rationality', where a person reasons in a way which is different from the 'normal' way. It doesn't have to be just one individual. Imagine a society where everyone 'agrees' to methods of reasoning and standards of evidence which seem to us totally absurd and arbitrary. (This problem has actually been explored in relation to anthropology, see P. Winch 'The Idea of a Social Science', Routledge, and Hollis, Ed. 'Rationality' Blackwell.)

In this case it looks as though the coherentist about knowledge is committed to saying that, in this society, it is a 'known fact' that the earth is flat. The response would be to say that in this case the justification condition is met but the truth condition is not met. The earth is not flat, therefore what these people believe can't be knowledge. However, the example still leaves us wondering whether the response given above to the worry about whether coherence is an adequate guide to, or grounds for truth is sufficient.

A general observation. Although you are not required to quote specific books or philosophers, I would like to see some evidence of reading you have done on the topics which you write essays for. For example, you could add a bibliography (obviously there is no time to do this in an exam). As a rough guide, aim to do some serious reading of philosophical texts - say 50 pages - for each essay that you write. I can't tell you what would be best to read. This is work you have to do for yourself. But I will be very interested in what you find in your research.

All the best,

Geoffrey