ABSTRACT
The first approach to designing artificially intelligent computers assumed that intelligence was the ability to solve problems. Turing, Damasio, Winograd, and Flores all view intelligence as predicated on social activity. According to Haugeland, much of what we consider to be human intelligence is not an internal quality of the mind. Our designs and behaviors arise through and out of interaction with the environment. One might argue that Roomba shows as much intelligence as many animals, in its ability to navigate in a local environment, avoid hazards, and forage for sustenance. For computers, the relational nature of intelligence suggests that the model of an artificial intelligence that holds a separate identity and acts by itself in the world, as a replacement for human intelligence, is the wrong model. One would expect that the movement among computer scientists toward understanding intelligence as both embodied and relational would have dispelled the Cartesian dualism of separation between body and soul once and for all.
