ÖйúAƬ

D. W. Salt's response (THES, March 17) to my article on consciousness (THES, March 3) misses several of the points I tried to make. I never used the epithet "epiphenomenal" for consciousness. Indeed the thrust of my article, and more general work on neural modelling of the mind, is that consciousness involves the most crucial circuitry in the brain. As such consciousness is no mere epiphenomenon, but has great importance of its own.

The possibility of a distinction between conscious and unconscious processing in a machine was questioned by Salt, but needed more development than I could give in the article. Such a division can begin to be achieved by means of neural models being developed to account for subliminal processing effects in humans. Such models are different from standard computer programs. I agree that there are very serious problems raised by the claim that the human mind can never be modelled by a computer. I tried to answer them, albeit briefly. But to say that an attempt to build a neural model of creativity is simplistic is incorrect. For the nature of human understanding and the creative act is what is at issue, and especially the claim that it cannot be modelled. It is necessary to explore how far the neural substructures of the brain contribute to such activity. My thesis is that they are sufficient to support it, but that is a claim that can be tested.

The difficult questions are not being ignored or trivialised by the neural approach. We are beginning to look at the brain in a global way and should be able to answer these questions in a scientific manner.

John Taylor

Professor of mathematics, King's College, London