Lundy Lewis

Book review: Designing Autonomous Agents Theory and Practice from Biology to Engineering and Back Edited by Pattie Maes (MIT Press, 1991)

Over two decades ago, Herbert Simon argued that a physical symbol system (PSS) has the necessary and sufficient means for general intelligent action [12, 13]. A PSS consists of (i) a set of symbolic expressions from old ones and (ii) operational definitions of expressions such that some expressions cause a system to affect external objects and other expressions cause a system to carry out internal processes. The PSS hypothesis holds that both human and computer systems are PSSs and, therefore, that computer science is an empirical inquiry into the nature of intelligence. As evidence, Simon offered that computer systems already exhibits some degree of intelligence. Today, computer systems that can play a good game of chess and perform tasks in well-structured, predictable environments are relatively mature. Importantly thought, the PSS hypothesis implies that a computer system could exhibit intelligent behavior comparable to that of humans in the everyday world. If a computer system could be designed to execute a real world task such as Go to the Market, we would have made a long stride towards explaining intelligence of the sort that we see in humans. The PSS hypothesis has provided a license and a framework for pursuing research in this direction.