Monday, September 12, 2011

Looking for loopholes in the argument against free will

To: Marcus S.
From: Geoffrey Klempner
Subject: Looking for loopholes in the argument against free will
Date: 9 November 2004 10:54

Dear Marcus,

Thank you for your email of 28 October, with your second essay for the Philosophy of Mind program, in response to the question, 'Summarize the argument that free will is impossible, whether on the assumption of determinism, or on the assumption of indeterminism. Do you see any loopholes that the defender of free will might exploit?'

In searching for a loophole in the argument against free will, you dismiss the alternative that determinism might be false, agreeing with David Hume that this would make our 'uncaused' actions random, not free. Some way must be found, therefore, to reconcile determinism with free will.

Your solution is basically the one Hume proposed. According to Hume, the ascription of responsibility is compatible with determinism (hence the technical term used for this line of argument, 'compatibilism'). When we say that a human action is 'free' we do not mean that it is uncaused, but rather that it was *not* made under overwhelming duress (gun pointed at your head); that you *were* mentally competent to make the decision; that you *weren't* physically pushed, and so on. When deciding whether an ethical or legal transgression deserves punishment, these are the kinds of considerations that we appeal to.

In your language, certain chains of causes and effects put people in situations where the power of human choice comes into play. Situations call for the 'necessity of choice'. So long as this power is not interfered with in one of the ways mentioned above, the action is free, despite the fact that given the state of the universe at the time of the big bang, and the laws of physics, determinism decrees that no action could have been done other than the action which the agent actually did.

To put the matter more bluntly, human beings are choosing machines.

The problem is that although this line gives us a 'useful' concept of responsibility, one that we can and do apply in daily life with a considerable degree of sophistication, the image is, for many people, repellent. The thought that the way the big bang banged all those billions of years ago determined that I would these words, rather than any number of variations that I might have 'freely' chosen, goes against our basic intuitions about what we are and what we can do.

I don't see any way round this for the compatibilist. Suppose that, unhappy with the tone of my criticisms of your essay, you email back, 'You could have been more polite!' What could those words possibly mean? 'Should' implies 'can', but given the way the big bang banged, and the laws of physics, and the truth of determinism, there was no logical possibility that I might have written different words then the words that I did write.

P.F. Strawson in his British Academy lecture 'Freedom and Resentment' (reprinted in a number of places including, 'Freedom and Resentment and Other Essays) comes closest to giving a plausible 'compatibilist' explanation of what human beings are doing when they criticize one another's actions. But it still does not address the fundamental point that such criticism - given what it implies - appears ultimately irrational, even if it does have a practical *use*.

You might not agree with the statement that 'human beings are choosing machines'. An example of a choosing machine would be the humble chess computer (well, not so humble in the case of Deeper Blue, which beat Kasparov). However, the very same contrast you draw between a human being and a billiard ball can be made between a chess computer and a billiard ball.

Modern chess computers have randomisation devices so that they do not always play the same move in the same position (the device analogous to spinning a coin: the result is still determined, but not one that anyone could predict). The computer considers the available moves, works out the consequences and assigns a numerical value, plus or minus, to each line. It then 'selects' the move with the optimal value.

Of course, even Deeper Blue is just a calculating machine, not an example of genuine artificial intelligence. No chess computer is 'free' because freedom implies conscious, rational choice. However, from the lofty height from which we are considering the question, the difference is unimportant. The way the big bang banged determined the letter I would write to you, just as it determined the moves that Deeper Blue would make in its first game against Kasparov.

We can reassure ourselves that we are not like chess computers, we are not mere calculating machines. But how reassuring is that, ultimately?

However, having said all that, something does niggle me about your idea of the human will being made of 'a different kind of material'. There could be something in this. One thing that this seems to imply is that, contra the supporters of AI, there is not, nor could there ever be a 'program' for the human brain. The human brain is not analogous to a computer but works in a fundamentally different way, one which science not yet fathomed. But if that is true (as it may well be) how does that help with the problem of free will?

All the best,

Geoffrey