A nice piece on embodied knowledge and the failure of mechanism by Herbert Dreyfus.

Achinson

Drefyus says that cognitive science, which looked to be a:

"....promising proposal for understanding human intelligence, while bypassing the body and, indeed, experience altogether, seems to have run its course. In mid-twentieth century, philosophers, linguists, psychologists, and computer scientists joined in proposing a new discipline called Cognitive Science that promised to work out how the logical manipulation of formal, symbolic representations enabled minds and suitably programmed computers to behave intelligently......In the early seventies, however, Minsky’s AI lab ran into an unexpected problem.Computers couldn’t comprehend the simple stories understood by four-year-olds."

it seemed to me that the real problem wasn’t storing and organizing millions of facts; it was knowing which facts were relevant. One version of this relevance problem is called the frame problem. If the computer has a representation of the current state of the world and something changes, how does the computer determine which of the represented facts stay the same, and which representations have to be updated? It seemed to me obvious that any AI program using frames to solve the storyunderstanding problem by organizing millions of facts was going to be caught in a regress, and that therefore the project was hopeless. And, indeed, Minsky has recently acknowledged in Wired Magazine that AI has been brain dead since the early 70s when it encountered the problem of common sense knowledge.

Dreyfus then asks a question:

How, then, do we manage to organize the vast array of facts that supposedly make up commonsense knowledge so that we can retrieve just those facts that are relevant in the current situation? The answer is: “We can’t manage it any more than a computer can, but fortunately wedon’t have to.” Only if we stand back from our engaged situation in the world and represent things from a detached theoretical perspective do we confront the frame problem. That is, if you strip away relevance and start with context-free facts, you can’t get relevance back. Happily, however, we are, as Martin Heidegger and Maurice Merleau-Ponty put it, always already in a world that is organized in terms of our bodies and interests and thus permeated by relevance.