Researchers working on consciousness could benefit greatly from incorporating notions of design employed by engineers and software creators into their work, according to Aaron Sloman of Birmingham University's school of computer science.

In a lecture last week at the Royal Society of Arts, Professor Sloman, a philosopher, advocated a systems approach to consciousness. He said that much research wrongly uses such words as "conscious", "aware" and "experience" as if they had clear fixed meanings and there were clear distinctions between things they apply to and things they do not apply to.

This generates the illusion that consciousness is something that is either present or absent in an object and tempts researchers to ask "pseudo- questions" such as which animals have consciousness, how it evolved, could a robot have it and whether consciousness was reducible to physics?

Another tempting mistake is to present consciousness as a matter of degree as in differences between states of consciousness and differences between animals.

ÖйúAƬ

Professor Sloman, giving one of a series of lectures on consciousness at the RSA, said an alternative to these two "inadequate views" of consciousness was the notion of a "cluster" of a vast number of capabilities including perceptual capabilities such as controlling posture and movement and seeing possibilities: different kinds of memory (short, long) and of learning.

He said: "Different subsets of the cluster occur in different organisms, different machines, in different people and even in the same person at different times, for example during infancy, adulthood or after brain injury. The point is that there is no unique subset of capabilties that defines consciousness."

ÖйúAƬ

Professor Sloman does not claim to have a definitive theory of consciousness but believes he has developed an outline for a research programme that would address the problem, partly by transforming it.

"Simply trying to work out what consciousness is from general principles is not, I believe, a good idea. It is much more fruitful to ask what I would need and why to design something that approaches the capabilities of human beings and other kinds of animals," he said.

Answering this requires a multidisciplinary approach involving psychology, brain science and engineers designing complex information processing systems. Research would extend and revise the language and concepts for mental states and processing in ourselves, other animals and in machines - including machines of the future. "This might include a 'periodic table' of mental states and processes," he said.

Details of the forthcoming Tucson 2 Consciousness conference are on The THES Internet Service at Thesis.newsint.co.uk.

ÖйúAƬ