'Freedom Evolves' can be seen as the final instalment of a life-long philosophical project, the first two parts of which were 'The Intentional Stance' and 'Consciousness Explained'. It gives an account of free will and moral responsibility which is complementary to Dennett's other views on consciousness and personhood. Dennett deserves considerable credit for realising that such an account was needed to round out his theories: but the result does not match the vigour and confidence of 'Consciousness Explained'

As in 'Consciousness Explained', Dennett advertises the controversial nature of his views extensively in advance. He expects hostility from those who fear that a sceptical analysis of freedom will undermine people's belief in the reality of moral considerations: he likens himself to an interfering crow who insists on telling Dumbo he doesn't really need the feather. But in fact, his view of free will is not particularly original or threatening: it comes down to compatibilism, the view that although in the strict physical sense our actions are pre-determined, we can still be free in some other important senses. The important point for a compatibilist case is exactly what senses these might be: but Dennett is not very clear about this.

He goes on to give altruism a similar treatment; it turns out that there is no such thing as altruism, strictly speaking, but certain kinds of selfish (or 'benselfish') behaviour will do instead.

So far, the views presented are convincing, if not particularly new or interesting: but Dennett also suggests that adherence to high ethical standards might pay off for the individual because if others know your behaviour is restricted in this way, the scope for certain beneficial mutual arrangements is enhanced. While there might be something in this, it's odd to find arguments which equate ethics with handcuffs in a book purporting to explain moral freedom. One might have expected the opposite case: that free will pays off because it makes you scarily unpredictable, and hence a person to be treated with circumspection. Dennett concludes by contemplating the possibility that people might be able to opt in or out of moral responsibility: surely, he suggests, given the benefits, they would choose to opt in. This is strange stuff: morality involves duties as well as rights, and you cannot therefore opt out of it: moreover if people could opt out of moral responsibility, wouldn't they have to be locked up like dangerous animals?

Dennett seems conscious that in tackling ethics he has moved into a swamp which is not his normal territory. He imagines Kant suggesting that he, Dennett, has written only about why we do things, not about why we ought to do them, and hence, not about ethics at all. His response is to demand: if the kind of behaviour I've described isn't ethical behaviour, how come it attracts so many well-intentioned friends? If the venerable German were alive and working at Tufts University, I think his answer might be 'Kind of a dumb question for a smart guy like you, ain't that, Dan?'.

Dennett seems to caught a mild case of something we might call 'Penrose disease' - a tendency to extended digressions of limited relevance to the main point of the book (did the world really need another exposition of Conway's game of Life?): but one of these is among the best parts of the book: a splendidly clear and convincing sceptical analysis of the famous experiments carried out by Benjamin Libet.

All in all, two cheers for 'Freedom Evolves', then: not Dennett at his best, but still commendably readable and thought-provoking.