The Muse in the Machine

Emotions, metaphors and analogies lie at the heart of human thought, asserts Yale computer scientist Gerlernter, a founder of parallel processing, in a brilliant and wholly accessible book about the theoretical underpinnings of artificial intelligence. His heretical new model of thought posits a ``cognitive spectrum'' extending from ``high-focus'' activities like reasoning, analysis and abstraction to ``low-focus'' thought whereby emotions make possible unexpected connections, leaps of awareness and creative juxtapositions. Gerlernter ambitiously applies this model to illuminate dreaming, sleep, hypnosis, spirituality and the emergence of the modern mind from an ancient, prelogical mindset that he likens to children's thought processes. He also describes a software program, developed by his research team at Yale, that he sees as a first step toward a working model of a truly ``thinking machine'' embodying the cognitive spectrum. But even so, Gerlernter finally concludes that it's the ``observer illusion'' that distinguishes the ``mystery of consciousness''--and that ``there is no reason to suppose . . . that adding emotion, or performing any other sophisticated programming trick, will ever endow a computer with the illusion of an observer-self.'' (Apr.)