Path: utzoo!utgpu!utstat!jarvis.csri.toronto.edu!mailrus!cornell!rochester!rutgers!att!whuts!homxb!houdi!marty1 From: marty1@houdi.ATT.COM (M.B.BRILLIANT) Newsgroups: comp.ai Subject: Re: Question on Chinese Room Argument Message-ID: <1540@houdi.ATT.COM> Date: 20 Feb 89 22:54:47 GMT References: Organization: AT&T BL Holmdel NJ USA Lines: 56 From article , by harnad@elbereth.rutgers.edu (Stevan Harnad): > > engelson@cs.yale.edu (Sean Engelson) of Computer Science, > Yale University, New Haven, CT 06520-2158 asks: > > " for those denying (Searle + rules) understanding of Chinese: What is > " your ["effective"] definition of "understanding"... one that does not > " beg the question... by defining understanding to be > " symbol-processing or... that which humans do. > ..... > Understanding is what is "+" of Searle (and you) with respect to > English, and "-" with respect to Searle (and you, and the computer > running the program he's executing) with respect to Chinese. Lacking > any other evidence for "+" on the computer's behalf, that makes the score > on understanding: Searle 1, computer 0. As an educated native speaker of (American) English, I know enough about English to believe that if I did not understand English, I would not be able to persuade an English-speaker that I could speak English, no matter how many rulebooks I had. So I assume, in fairness, that I could not pretend to speak Chinese, with or without a rulebook, if I did not understand Chinese. Did Searle really suppose that he could speak passable Chinese if only he had a rulebook? Could you posit a Chinese with a rulebook who could pretend to speak English, without in fact understanding it? In other words, pose a corresponding "English Room puzzle" and you will see the fallacy. I am persuaded that any native, foreigner, machine, or simulation thereof, that can carry on a respectable conversation with me in English, must understand English. Incidentally, there is a language proficiency examination developed by Educational Testing Service, used in New Jersey (and other places, I suppose) to test the language qualifications of bilingual and ESL teachers, that might be adaptable for use in a Chinese Room trial. An interviewer converses with the subject to elicit speech in the test language on a variety of topics, and tapes the interview. A rater listens to the tape and judges how well the subject has succeeded in expressing ideas in the test language. The highest score, which would be attained by a native speaker with no trace of an accent, is 5. A subject with some accent, but full command of linguistic structure and demonstrable ability to discuss non-trivial topics, would be rated 4. I'd guess a machine or simulation with no semantic proficiency would score below 3 (but I'm not thoroughly familiar with the rating scale). So let me propose as a partial definition of "understanding": that anything that can score a 4 on the language proficiency examination must understand the language. M. B. Brilliant Marty AT&T-BL HO 3D-520 (201) 949-1858 Holmdel, NJ 07733 att!houdi!marty1 Disclaimer: Opinions stated herein are mine unless and until my employer explicitly claims them; then I lose all rights to them.