The problem of free will is one of the most discussed in philosophy, and although I believe it has something of the character of a mirage - the closer you get to it, the less it seems there is to it - it does not disappear entirely even when it is pinned down. Once you get past the false dichotomies and irrelevant abstractions, there are important ethical ramifications to the nature and extent of our free will.

I should start by saying something about what I mean by 'free will', which is one of these terms - like 'consciousness', or 'beauty' - that people bandy about, and have whole elaborate arguments around, as if everybody knows what everyone else is talking about. Then, when people get explicit about what they actually mean, it often turns out that most of the disagreements are to do with different people having different definitions in mind, if they have definitions in mind at all. The meaning I am most interested in is something like 'the capacity of an agent which means that in some set of circumstances, they could have made different choices'. The big question, then, is what it would have taken for them to choose differently.

The first thing to say about the will is that it's not free in the sense that it costs something to keep ourselves from doing something that some part of us wants to do1. Our stomachs are always telling us that we should be eating more, and many other parts have their own demands. If I want to overrule those wants, I need willpower to do that, and it doesn't come free. Like a muscle, the will gets tired, and a sapped will must be replenished before it can really be relied on. It's not much use saying 'you just need to exercise more willpower' when it's an experimentally verified fact that the stuff runs out, which suggests that there are limits to how much we can rationally blame people for failing to exercise enough of it. I won't delve deeply into the moral implications of this fact here, but it seems likely that they are substantial.

The next sense in which our will is importantly un-free is that both our inclinations and our actions are constrained by what makes up our selves and our environment. This ties in with the debate about determinism - the idea that everything that happens is determined by what has gone before. I would argue that this idea is seductive, but largely meaningless; and to the extent that it is meaningful, it is largely irrelevant. The ways in which our will is significantly constrained by our environment have very little to do with determinism in the usual philosophical sense.

Determinism is seductive because it is obviously true that if our actions are determined by what has gone before, they cannot be said to be entirely our own. It is, however, largely meaningless because of chaos - because there is no way for anyone to make reliable predictions based on known starting conditions, for anything but a very simple physical system. Complexity gives rise to something that is either identical to or indistinguishable from genuine novelty. Maybe an all-knowing deity could predict what you're going to do by completely simulating everything about you and your environment, but so what? Physical considerations mean that such predictions are impossible in principle within our own universe - and even if we were rigorously simulated, maybe the simulations would have free will!

Determinism is largely irrelevant because whatever it is that constitutes ourselves - our minds, brains, bodies, particles, ideas or whatever else - plays a major part in determining what happens. We may or may not be machines, but if we are, we are machines with rich, emergent powers of decision-making. Our decisions are our own, and there is no way, in general, for anyone to predict your decisions or mine. This happens to be true whether or not the present and future are in fact uniquely determined by the past. Standard interpretations of quantum mechanics suggest that the universe is fundamentally random, and subatomic particles routinely 'choose' one outcome or another just by chance, but this fact buys us nothing when it comes to figuring out the extent to which anybody is responsible for their own actions. If your decisions depend on the configuration of subatomic particles in your head, what do I care if those particles move at random or not? There are several levels of description in between the level of subatomic physics and human-level behaviour, and as far as anyone can tell, none of them are affected in the slightest by the question of whether there is yet another, deterministic layer of reality underlying the probabilities expressed by the quantum wave function. Personal responsibility depends on psychology, not physics.

However, the fact remains that we are the product of our environment (and our genes, and the laws of physics, and so on and so forth). We have limited powers of self-transformation - much about me is inescapable, however much I might wish to change it. In principle I may have the liberty to become a breatharian if I want, and just stop eating entirely, but I very much doubt I could manage it in practice. If I wanted to go the whole hog and just stop breathing, too, I would be guaranteed to fail because - to gloss over the neurology of it - my lungs would overrule the part of me that ever thought that was a good idea. The notion that the conscious part of my mind is free to make any choice it wants ignores a mountain of evidence against the unity of the self2.

Our decisions are profoundly constrained by our natural inclinations, our capacities and the ways they relate to the social and material contexts we find ourselves in. This makes it impossible to support the proposition that any of us is completely responsible for our own actions.

Once we acknowledge that the ideas of absolutely free will and absolute personal responsibility are untenable, we are left with some interesting questions. For example, to what extent is personal responsibility just a convenient fiction? How much does it matter, when it comes to things like justice and social order? How much responsibility do we all have for other people's wrongdoings - particularly those of us in positions of power, like parents, teachers, politicians and company bosses? Has the concept of personal responsibility been allowed to act as a smokescreen for a broader concept of social responsibility - or, indeed, vice versa? What are the dangers inherent in acknowledging the limits of either personal or social responsibility, and can they be avoided?

These questions are probably better addressed elsewhere, but I hope I have shown that even after millennia of discussion, the notion of free will is vital and relevant. Science has a lot to say about it, and some of the questions that it throws up are of profound importance - but the genuinely interesting questions are not necessarily the ones that philosophers and others have spent the most time on.

1 Roy Baumeister at Florida State University has done some very interesting work on this phenomenon, which he calls 'ego depletion'. It turns out to be tied up with glucose levels. Here's a brief summary in Wired; here's a book he co-wrote about it; here's a handy Google Scholar search. Note also that stress has a way of shutting down the ability of our 'higher self' to overrule the demands of our hindbrain.

2 One classic experiment on this was conducted by Benjamin Libet et al in the 1980s. Subjects were asked to flip their wrist at a random time, and report the exact time they decided to do so. This was compared with measurements of activity in their brain, which turns out to start about half a second before the participants thought they had made the decision to move. Although it is sometimes taken to refute the idea of free will, it only addresses 'free will' as a capacity of consciousness as such; a naive notion of the conscious mind being solely responsible for all of our decisions.