Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!bloom-beacon!apple!voder!pyramid!prls!philabs!linus!mbunix!bwk From: bwk@mbunix.mitre.org (Barry W. Kort) Newsgroups: comp.ai Subject: Re: Question on Chinese Room Argument Summary: That reminds me of an interesting anecdote of Feynman. Keywords: Understanding, Comprehension, Learning Message-ID: <45126@linus.UUCP> Date: 20 Feb 89 06:55:57 GMT References: <4298@pt.cs.cmu.edu> <8174@netnews.upenn.edu> <51123@yale-celray.yale.UUCP> Sender: news@linus.UUCP Reply-To: bwk@mbunix.mitre.org (Barry Kort) Organization: IdeaSync, Inc., Chrono, VT Lines: 23 The problem I have with Searle's notion of symbol manipulation is that such a system appears unable to learn anything new. In _Surely You're Joking, Mr. Feynman_, Richard Feynman recounts an attempt to teach physics in Brazil. The students had become very adept at formal symbol manipulation. They could regurgitate the definitions and formulas, but they had no idea that the symbols actually referred to anything in the outside world! It seems to me that "understanding" (or "comprehension", as I prefer to call it) entails the construction of a mental map between symbols and their referents in the world external to our minds. Once we buy into this notion of "understanding", we automatically buy into the notion of "learning" (knowledge acquisition). Searle's Chinese Room could be considered to understand Chinese if it could use the medium of Chinese to acquire knowledge about the world outside the room. Such a system would evolve its "rules" over time. Instead of just translating stories, it would respond with its own anecdotal accounts, maintaining a thematic thread suggested by the preceding stories. --Barry Kort