Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ukma!rutgers!elbereth.rutgers.edu!harnad From: harnad@elbereth.rutgers.edu (Stevan Harnad) Newsgroups: comp.ai Subject: Re: Question on Chinese Room Argument Summary: Objective Test for Subjectivity: The standard LTT vs. the TTT: The TTT is immune to Searle's Argument (but not to the other-minds problem) Keywords: Understanding, Comprehension, Learning Message-ID: Date: 26 Feb 89 20:55:23 GMT References: <4298@pt.cs.cmu.edu> <7220@polya.Stanford.EDU> Organization: Rutgers Univ., New Brunswick, N.J. Lines: 57 geddis@polya.Stanford.EDU (Donald F. Geddis) of Stanford University writes (in a pair of successive postings): " it might be true that a computer system could not converse " intelligently without being embodied in the real world. But the real " question Searle considered was: How do you determine when a system " is intelligent...? The AI answer is "treat it as a black box and observe " its behavior (have conversations, in this case)". Searle (mistakenly) " disputes this view, and wants us to look inside the system for some " "causal powers"... Just because it requires careful probing and the " examiner can be fooled, doesn't mean that "external behavior" is not " the proper criteri[on] for deciding when a system understands. " Just what, exactly, is being proposed as an alternative test? There are two alternative OBJECTIVE tests for having a mind, the (standard) Linguistic Turing Test [LTT] (symbols-in, symbols-out) and my stronger (robotic) Total Turing Test (TTT) (proximal-projections-of- objects-on-sensors-in, effector-action-on-objects-out). The LTT is a subset of the TTT, but one that is, as I have indicated repeatedly, EQUIVOCAL about the issue of "embodiment" and the putative autonomy of symbolic function from many forms of nonsymbolic function that may be needed in order to pass the LTT in the first place. Searle is only addressing the LTT, and my reply to Searle is that the TTT is immune to his arguments against the LTT. Neither the TTT nor the LTT, however, provides a guarantee that the candidate has a mind. There is and can be no objective test for that, only a first-person subjective one: To perform that, you have to BE the candidate. ONLY this subjective test is decisive. There are two senses in which Searle is advocating "looking inside": One is to look at the functions of the brain, because we have pretty good reason to believe that candidates with brains have minds (because, as I would put it, candidates with brains can pass the TTT). The second sense of "inside" is the first-person test for subjectivity, which we can all perform on ourselves. It's THAT "causal power" that he reminds us brains have but symbol-crunchers do not. My reply is that candidates OTHER than the brain that can pass the TTT (if and when we come up with any) are immune to his Chinese Room Argument that they cannot have a mind (though, of course, I repeat, no objective test can demonstrate that anyone, EVEN ourselves, has a mind). Searle's argument against (hypothetical) candidates that pass the LTT only, with symbols only, is decisive, however. I've always thought this reasoning was quite easy to understand, but from the fact that very few people have given me any objective evidence that they've understood it, I've concluded that it must be difficult to understand. Maybe by trying to put it slightly differently each time, tailoring it to the latest misunderstanding, I'll succeed in making it understood eventually... -- Stevan Harnad INTERNET: harnad@confidence.princeton.edu harnad@princeton.edu srh@flash.bellcore.com harnad@elbereth.rutgers.edu harnad@princeton.uucp BITNET: harnad@pucc.bitnet CSNET: harnad%princeton.edu@relay.cs.net (609)-921-7771