Human intelligence is a formidable beast whose extents are poorly understood. Subsumption takes advantage of the way natural organism's are organized, which is rather strikingly similar to the way in which we ourselves are organized. No-one think about breathing, but we carry on doing so, the same with the beating of a heart. How about walking? How about walking and chewing gum at the same time? Concurrency of action and independent processes are the foundations of our day to day lives, and as stated by Morovec, the most difficult to teach a machine to accomplish.
Brooks, R. A., “Intelligence Without Representation”, Artificial Intelligence Journal (47), 1991, pp. 139–159.
Brooks, R.A., “Integrated Systems Based on Behaviors”, SIGART Bulletin (2:4), August 1991, pp. 46–50.
Brooks, R. A., “Elephants Don't Play Chess”, Robotics and Autonomous Systems (6), 1990, pp. 3–15.
Brooks, R. A., “How To Build Complete Creatures Rather Than Isolated Cognitive Simulators”, Architectures for Intelligence, K. VanLehn (ed), Erlbaum, Hillsdale, NJ, Fall 1989, pp. 225–239.
Maybe should get another viewpoint… Marvin Minsky looks like he may have some thoughts. Any good ideas for a strong opposition to the Subsumption methodology?
To read:  Rodney A. Brooks, A robust layered control system for a mobile robot, IEEE J. Robotics and Automation, RA-2, April (1986) 14-23.
 Rodney A. Brooks, A robot that walks: Emergent behavior form a carefully evolved network, Neural Computation, 1(2) (Summer 1989) 253-262.
 Jonathan H. Connell, A colony architecture for an artificial creature, MIT Ph.D. Thesis in Electrical Engineering and Computer Science, MIT AI Lab Tech Report 1151 (June 1989).