An evolutionary view of intelligence

The traditional view of artificial intelligence accepts the symbol processing hypothesis of Newell and Simon (1976). Within the past decade, Brooks (1991) argued that since naturally intelligent systems are all biological, intelligent systems should be studied in terms of their biological niches. Beer (1990) has extensively studied how intelligent behavior emerges from a hexapod simulated in a lifelike environment. We propose yet a third perspective to study the general topic of intelligence. In our view intelligence may be studied by analyzing how organisms evolved into multi-celled organisms, with specialized cell. One specialized type of cells was the neuron. We believe that the ability to control is an emergent property of a neuron's ability to integrate its inputs. To further study this third perspective, based upon a rhythmic neural firing circuit suggested by Stent and Kristan (1981), we employ the more accurate membrane equations of MacGregor (1993), to enhance the less biologically accurate equations used by Beer in his simulated hexapod.