To: Gordon F.
From: Geoffrey Klempner
Subject: Should we get rid of the concept of personal identity?
Date: 26 April 2007 10:57

Dear Gordon,

Thank you for your email of 21 April, with your essay for the Philosophy of Mind program in response to the following two questions:

1. 'Can one give adequate criteria for the identity of a person over time? Illustrate your answer with thought experiments describing 'problem cases' of personal identity.'

2. 'What the thought experiments of fission, fusion etc. show is that 'personal identity' is a philosophically incoherent concept, and ought therefore to be dispensed with.' - Discuss.

Richard Parfit is an Oxford philosopher who argues in his book 'Reasons and Persons' against the 'importance' of personal identity. He sees his project as ultimately arguing for a moral point of view where the best action is one that leads to the maximum 'benefit' for all, where the fact that one of the persons is 'me' is not a relevant consideration.

Leaving aside the claim about ethics, there are two main reasons for resisting Parfit's conclusion. The first is that we find the thought experiments sufficiently gripping to be incapable of feeling indifferent to the question whether 'I' will survive or not in a given scenario. The second is that, contra Parfit, there are very important consequences of identity which would require turning our moral and social institutions upside down if we decided to rid ourselves of the notion of 'same person'.

Parfit can reply that our 'intuitions' are merely symptoms of philosophical illusion, and that if our institutions are based on an illusion then we are better without them.

You give a clear and logical appraisal of different candidates for the 'i-factor' which embodies the necessary and sufficient conditions for personal identity over time. You claim 'agnosticism' on the question whether not there is a 'soul' which would provide the ultimate, non-negotiable (but also, alas, unknowable) i-factor.

However, even if one is agnostic about souls, there remains room for debate over whether we can, or should, attempt to defend the notion of personal identity in the face of the various counter-examples.

For example, suppose that someone successfully constructed a person-duplicating machine. This threatens the logic of identity, because of A=B and A=C then B=C, as a matter of logic. The statement not-(B=C) entails not-(A=A) which is a contradiction.

However, as David Lewis has pointed out in one of his articles, we can save the logic of identity by retrospectively deeming that when a person A enters into the duplicating machine, and B and C emerge, there are and always were two persons, A-B and A-C sharing part of their life history. In the same sense, we can talk of two 'branches' of a tree, which share a common part.

I like this solution. It doesn't allow us to 'say what we like'. B and C are insistent that they were both A, and these claims are fully justified. A did not die. Being duplicated is a way of surviving.

Objections have been raised, however, on the grounds that it reduces persons to 'life histories in the making'. But what's so bad about that? If duplication did become part of everyday life, this would be the most logical response, requiring the least violence to our intuitions and social institutions. (Of course, there would be significant stress on those institutions, but there is no solution that avoids stress.)

Memory is the other big obstacle. As you point out, memories fade in time. Parfit relies on this for his 'Methuselah' thought experiment. Imagine beings like us, except that they live for a hundred thousand years. During this time memories fade and are replaced by other memories. When someone tells me something that GK did 1000 years ago, this means absolutely nothing to me, because there is not the faintest recollection. Yet there was no point where psychological continuity ceased.

Someone who thinks that personal identity 'matters' - in other words, someone who disagrees with Parfit's conclusions - needs to have a principled way to answer this challenge. Because it seems to show that the logic of identity simply breaks down and can't be saved when applied to the identity of persons over time.

My response is that I don't accept the assumption that personal identity must be an all-or-nothing matter. Persons have an identity, which is sufficiently robust to cope with all sorts of counterfactual situations, including the science fiction scenarios like body-duplication, or 'beam me up Scotty'. However, we are still dealing with the empirical world, not the world of numbers. There will always be vague and borderline cases, because things in the real world don't always have sharp outlines.

All the best,

Geoffrey