โ Scribed by Mark Stefik; Stephen Smoliar
No coin nor oath required. For personal study only.
What Computers Still Can't Do is the latest in a series of books by Herbert Dreyfus i:n which he challenges the foundations and methods of artificial intelligencle. Within the field of artificial intelligence, Dreyfus' work has been controversial and has led to responses both in this column and elsewhere [l, 21. Some people cite his work as influential, arguing that at various times AI has been complacent in its methods and that issues raised by Dreyfus have later become areas of investigation in the field. Others say that the technical and scientific basis for Dreyfus' arguments is too shallow. Practitioners in AI and cognitive psychology use multiple models and understand their limitations. Critics say that Dreyfus speaks to a strawman, inadequately presenting the state of the art and the practitioner's deeper understanding of models used in AI and cognitive psychology.
The relationships between scientists and philosophers and scientists and historians are sometimes tinged with suspicion. We begin with historians. Scientists often view anyone outside of their field as simply not knowing enough to have anything significant to say. They do not like watered down accounts of scientific history or the conclusions drawn from them. For example, in his book Genius: The Life and Science of Richard Feynman, James Gleick observed that Feynman resented the polished myths of most scientific history, which submerge the false steps and halting uncertainties under a surface of orderly intellectual progress. Historians, on the other hand, have professional concerns that scientists interpret the past too much from the vantage point of the present. Scientists reporting on their own field often have an ego investment, and their reports may overemphasize the significance of their own roles. Suspicion is even greater when an author offers an extended critique of current and recent history.
Philosophers also have a difficult relationship to science, although the line is difficult to draw. One view is that philosophers are generalists and that they see scientists as missing the big picture. Philosophers see themselves as providing independent criticisms and suggestions for guiding science. Reacting to this, empirical scientists may criticize philosophers as spending too much time in their armchairs and not enough time in libraries and laboratories. But this characteriza-Elsevier Science B.V.
๐ SIMILAR VOLUMES
Hubert Dreyfus claims that "symbolic AI" is a "degenerating research program", i.e. is not making progress. It's hard to see how he would know, since he makes no claim to have read much of the recent literature. In defending "symbolic AI", I shall concentrate on just one part of symbolic AI-the log
The thing to do with a dead horse is to bury it as expeditiously as possible." Anonymous Seven years ago I reviewed one of Dreyfus' earlier books on computers and cognition [22]. My position at that time was that Dreyfus' critique of AI appeared ill-informed. His arguments for why AI, as a research