ÖйúAƬ

Daniel Dennett proposes that freedom evolves ("An outrageous optimist finds freedom in the cells", THES , February 28). I agree: the evolution of the nervous system has given us brains that enable us to make choices. But the idea that brains are computers is wrong.

Our free will could not evolve in robots, however adaptable their programming. Robots are deterministic in operation: the next state is the automatic outcome of the current state and inputs. The computer I use to write these words has no choice. I have a measure of freedom as a conscious agent, a self-determining being, whereas it is an automaton. I agree with Dennett that this agency is not a result of a ghost in the machine but neither is it a result of computation. We design evitability out of computational systems because it subverts their correct operation, which relies on strict determinism.

The enabler for free will and, I think, consciousness evolved by the specialisation of cells into neurons. The difference between neurons and computer chips is that neurons are analogue and can support oscillation. We know neural network oscillations are kept under control because uncontrolled feedback is disastrous in analogue circuitry as sophisticated as the brain: it causes seizure. The point is that oscillation defeats determinism: cause and effect merge. Thus we have a mechanism for free will.

You could construct a brain from analogue silicon devices but better to stick with Dennett's digital computers and avoid ethical considerations.

Richard Wilson
Bicester