Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!purdue!decwrl!sun!pitstop!sundc!seismo!uunet!mcvax!ukc!strath-cs!glasgow!gilbert From: gilbert@cs.glasgow.ac.uk (Gilbert Cockton) Newsgroups: comp.ai Subject: Re: Question on Chinese Room Argument Message-ID: <2577@crete.cs.glasgow.ac.uk> Date: 13 Mar 89 14:39:16 GMT References: <2233@tank.uchicago.edu> Reply-To: gilbert@cs.glasgow.ac.uk (Gilbert Cockton) Organization: Comp Sci, Glasgow Univ, Scotland Lines: 32 In article harnad@elbereth.rutgers.edu (Stevan Harnad) writes: >The historian J. H. Hexter once wrote: > > in an academic generation a little overaddicted to "politesse," it > may be worth saying that violent destruction is not necessarily > worthless and futile. Even though it leaves doubt about the right > road for London, it helps if someone rips up, however violently, a > `To London' sign on the Dover cliffs pointing south... > Thanks for broadening things out again. For those who missed it, my first trade was history, though Hexter's comments never seemed to apply to British history, where demolition of the shoddy has always been the order of the day. Hexter is a good humanist source for sensible perspectives on matters of mind. In his "History Primer", he discusses three explanations of a "muddy pants" episode. Needless to say, the really dumb one reads like a reasoning trace (an explanation? - hah!) from an expert system. The preferable explanation, which is perfectly adequate for *UNDERSTANDING* is more cogent and less fussy. The common sense context of human understanding rules out a rule-based approach. No-one in AI is up to encoding it, and no-one could maintain it. I'd like to define understanding as the ability to integrate relevant knowledge with any current context. For many tasks, it is impossible to see how a machine could even pass a LTT. AIers cannot program what they do not understand. AIers understand very little. -- Gilbert Cockton, Department of Computing Science, The University, Glasgow gilbert@uk.ac.glasgow.cs !ukc!glasgow!gilbert