kck@g.gp.cs.cmu.edu (Karl Kluge) (02/20/89)
> From: harnad@elbereth.rutgers.edu (Stevan Harnad) > Searle's argument is simple but deep. Its simplicity has > led a lot of people who have not understood the deeper point it is > making into irrelevancies of their own creation. To show it to be > incorrect you must first understand it. What deeper point? It appears to be nothing but a form of vitalism -- brains have these mysterious "causal powers" without which understanding is not possible. It looks to me like the normal sort of confusion one might expect from someone not used to layered sytems in which each layer interprets/runs the layer above it. Further, Searle engages in gratuitous non-sequitors when he says things like "For example, my stomach has a level of description where it does information processing, and it instantiates any number of computer programs (true -- ed.), but I take it we do not want to say that it has any understanding (also true -- ed.). Yet if we accept the systems reply, it is hard to see how we avoid saying that stomach, heart, liver, etc. are all understanding subsystems, since there is no principled way to distinguish the motivation for saying the Chinese subsystem understands from saying that the stomach understands (yes there is -- we have posited that the I/O behavior of the Chinese system passes the Turing Test, we have never posited that wrt the information processing description of the stomach -- ed.)." There's something "deep" here, all right, but it's not the philosophy. > 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. No one cares what Searle does or doesn't understand when he is simulating a physical symbol system capable of passing a written Turing Test in Chinese. Period. If what Searle calls Strong AI is true I still wouldn't expect Searle to understand Chinese in the Chinese room. If Searle-doesn't-understand-Chinese-in-the-Chinese-Room --> ~(Strong AI), then it must be true that (Strong AI) --> ~(Searle-doesn't-understand-Chinese-in-the-Chinese-Room). Unfortunately, that isn't true. Therefore, Searle not understanding Chinese in the Chinese Room is not sufficient to disprove Strong AI. I may be an ideologically blinded AI fanatic (death to the heretical unbelievers!), but I'm still capable of applying the law of the contrapositive. > [This is the negative note on which Searle's Argument ended in 1980; > not to leave it at that, let me add that in "Minds, Machines and > Searle" (1989) I've tried to take it further in a positive direction, > showing that it's only the symbolic approach to modeling the mind > that's vulnerable to Searle's Argument; nonsymbolic and hybrid > symbolic/nonsymbolic models are not. Does Searle agree with you? It would certainly seem that he anticipates this sort of argument in the paper reprinted in "Mind Design" when he discusses "The Brain Simulator Reply". You have to have those mysterious "causal powers" that neurons have. Karl Kluge (kck@g.cs.cmu.edu) --
harnad@elbereth.rutgers.edu (Stevan Harnad) (02/21/89)
kck@g.gp.cs.cmu.edu (Karl Kluge) of Carnegie-Mellon University, CS/RI write:; " What deeper point? [Searle's position] appears to be nothing but a " form of vitalism -- brains have these mysterious "causal powers" " without which understanding is not possible. It looks to me like the " normal sort of confusion one might expect from someone not used to " layered systems in which each layer interprets/runs the layer above it... The deeper point is that Searle's view is NOT vitalism but a reductio of a particular KIND of (putative) model of the mind: The kind advocated by "Strong AI." Searle has said repeatedly that he's not claiming that only brains could have the requisite causal powers, just that brains clearly DO and symbolic models (as shown by his Chinese Room Argument) clearly [sic] DON'T. It would require a grasp of this deeper point to realize why hand-waving about "layered systems" is NOT a satisfactory reply to this; it just misses the point and begs the question yet again. To show that symbolic models have mental powers (e.g., "interpreting") you can't just wave your hand and baptise them with it. [My own approach, in "Minds, Machines and Searle," has been to argue for replacing the mere Teletype version of the Turing Test with the full Robotic version -- the Total Turing Test (TTT), calling for all of our capacities, linguistic and nonlinguistic (e.g., sensorimotor). It turns out that because at least some of the functions of the system that successfully passes the TTT would have to be nonsymbolic, Searle couldn't simulate them, and hence the system would be immune to the Chinese Room Argument. This would not, of course "prove" (or even give empirical support of the usual kind for the hypothesis) that the system actually had mental powers, but it would capture as many of the causal powers of the mind or the brain that we could ever expect to capture empirically. The other deep implication of Searle's Argument is that it points out why a purely symbolic approach to mind-modeling is a nonstarter. It's useful to know that... It suggests you should try something else instead. I in turn describe an alternative hybrid approach to grounding symbolic representations bottom-up in nonsymbolic (analog and categorical) representations.] " No one cares what Searle does or doesn't understand when he is simulating a " physical symbol system capable of passing a written Turing Test in Chinese. " Period. If what Searle calls Strong AI is true I still wouldn't expect " Searle to understand Chinese in the Chinese room. First of all, SOME people apparently do care about this -- care enough to engage in some rather strained arguments to the effect that Searle (or "something/someone") IS understanding in the Chinese Room (see some of the postings that have appeared on this topic lately, and my replies to them). Not to care is either (1) not to care whether AI can capture understanding (which is fine, but then why should mind-modelers be discussing it with such modelers at all, any more than with auto mechanics? and why do such modelers persist in using words like "understanding" to describe their models?) or (2) it is not to care about the inconsistency of claiming that the computer understands but Searle doesn't, even though he is doing exactly the same thing the computer is doing! " Does Searle agree with you [about hybrid symbolic/nonsymbolic models]? " [In his response to] "The Brain Simulator Reply" [he says].. [y]ou have " to have those mysterious "causal powers" that neurons have. I am holding out for a model that captures all of the brain's OBJECTIVE powers, namely, its TTT powers, trusting that the subjective ones will piggy-back on them, or if not, accepting that we can never hope to be the wiser. Searle is holding out for something that captures ALL of the brain's powers, objective and subjective. We both agree on the essential point here, however, which is that symbolic models don't have the requisite powers. -- 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