Embedded or embodied? a review of Hubert Dreyfus' What Computers Still Can't Do

A review essay on a book by Hubert L. Dreyfus, What Computers Still Can't Do: A Critique of Artifical Reason (Cambridge, MA: MIT Press, 1992 [(1972) see listing in IRPS No. 86]). It is suggested that despite the validity of Dreyfus's arguments & predictions, three essential faults undermine his theory regarding articial intelligence. His "professional mistake" arises from his selective application of philosophical critique. By refusing to subject science, technology, & mathematics to the same form of analysis practiced on culture & the social world, Dreyfus creates a false division between formal (computer) & informal (societal) knowledge. His "philosophical elision" incorrectly attributes group characteristics to all individual members & thereby leads to the false conclusion that the failure of all computers to achieve human characteristics presupposes the impossibility of any computer achieving human qualities. Lastly, Dreyfus makes the "sociological error" of overlooking the apparent capabilities of computers that are actualized through user interaction. Further, the focus on what computers cannot do leads to an incorrect understanding of what can be done through overlapping systems & user innovation, & it is concluded that research must shift toward the realms of understanding success & the possibility of mechanizing socializability. 7 Tables, 23 References. T. Sevier