To: David T.
From: Geoffrey Klempner
Subject: Mind-body problem and the definition of identity
Date: 11 June 2007 12:21

Dear Dave,

Thank you for your email of 3 June, with your second essay for the Philosophy of Mind program, in response to the question, 'What is identity? What is the relevance of a definition of 'identity' to the problem of the relation between mind and body?'

Kant says somewhere that a particular text would have been much shorter had it not been so short. I think this applies to the present essay to some extent.

In your email you say that you are 'ignoring epistemological questions completely' but it can be argued that it is impossible to divorce the concept of a sortal from the question of applicable criteria for identity over time. It is one thing to give the logical properties of a sortal concept in the abstract, but quite a different matter to establish that such a concept could actually be used to identify and individuate objects.

That is not to say that we cannot distinguish logical and epistemological questions with regard to identity. What, conceptually, must be the case - say, with regard to the question whether Lot's wife can be turned into a pillar of salt - is a different question from what, as a contingent matter of fact, I can come to know.

For example, according to the criterion of identity for the concept 'human being', it is not logically possible for me to change into a lizard. If you see me one moment, and a wriggling lizard the next then you know, as a matter of logic, that the lizard isn't me. However, this is a different question from how you know I am a human being. For all you know for certain, I may be an alien from the planet Zog which possesses the power to take on a range of forms. For Zogs, different criteria of identity apply than for humans, but, nevertheless, as a matter of logic, there must be applicable criteria of identity of some kind or other.

There is more on these issues in David Wiggins 'Sameness and Substance' (based on his earlier, 'Identity and Spatio-Temporal Continuity'). I think I have mentioned this text before.

Anything can be a Leibnizian object. Quine talks about objects as the 'logical minima, the result of distinguishing everything that can be distinguished.' For Quine, however, what 'can be distinguished' depends at any given time on science, on our network of beliefs, changing and developing in response to experience. He is talking about concepts as applied, not merely what might be a concept in some possible world.

You give the example of a moon-ear (see the quote from Kolakowski http://www.follydiddledah.com/image_and_quote_4.html !) taking this in the sense analogous to Goodman's Paradox, 'moon before T, my ear after T'. However, from the point of view of constructing Leibnizian objects one could just as easily regard the moon and my ear as physically parts of a single extended 'object'. Hence, when I scratch my ear I have a causal effect on the moon, conceived as part of object of which my ear is also a part.

There is something wrong with this concept, and it is not merely a matter of pragmatic inconvenience. There are no applicable criteria of identity for moon-ears, at least in the spatial sense. In the temporal sense, our criteria of identity for moon are trivially adequate for moon-ears so long as the time is before T, but arguably the diagnosis is the same.

Again, from the perspective of Leibnizian objects, we can happily define three classes of entities, minds, brains and mind-brains but this says nothing about the actual possibilities for identifying or distinguishing minds from minds or minds from brains (or brains from brains).

There were in fact to aspects to the question. You have concentrated on the question of identifying mind and body, but there is also a problem about identifying mind independently of body, hence the argument in the unit that the question of criteria of identity cuts both ways.

The mind-body dualist is saying, in effect, 'IF you give me my concept of mind as an entity with an identity, then there is nothing that you can do to establish an identity between the mind and anything physical.' If we were talking about Leibnizian objects then, of course, you can make mind-brains just as you can make Kolakowskian moon-ears. But if we grant the dualist's assumption (that mind is an entity with an identity) then the objection to unitary mind-brains is essentially the same as the objection to moon-ears.

Raising the issue of functionalism merely obscures the issue. The functionalist (e.g. Dennett) may talk of minds as 'immortal' (because a mind is essentially a Pythagorean-type formula which can be stored transmitted, uploaded and downloaded) but unless we are persuaded in the first place by Descartes arguments for immaterial 'substance', any form in which the formula is either preserved as data or implemented as a program will be physical. (In 'Consciousness Explained' Dennett says the uploading scenario gives far better prospects for immortality than having one's head or body stored in a deep freeze!)

All the best,

Geoffrey