dmocsny@uceng.UC.EDU (daniel mocsny) (12/13/89)
Perhaps I don't follow Searle's arguments entirely, but to me the "Chinese Room" seems to be equivalent to claiming that humans cannot "understand" mathematics. As I read Searle, I am struck by the similarity between the Chinese Room and how I felt in many of my mathematics classes. The maths instructor stood at the board and made squiggly marks, and then proceded to show the class a set of rules whereby we should transform the squiggly marks into other squiggly marks. When I went home to practice putting squiggly marks into other squiggly marks, at first I was entirely at the mercy of the rulebook. But after some time I memorized the rules, and applying them became considerably less tedious. What is more, at some point (and I'm not at sure just when) the symbols began to take on some kind of *meaning* for me. That is, I didn't just have the syntax, I started to obtain semantics. When that occurred, I felt I was "understanding" mathematics. In fact, this feeling was as solid and reassuring as my sense of "understanding" anything else. Now the problem in light of Searle's argument is obvious. Namely, *where* did this semantics come from? For what is mathematics, but a vast body of purely abstract symbol manipulation? No concept in mathematics has any direct grounding in any observable physical phenomenon. I cannot go into the field and observe the number "3" directly. Nothing I saw in maths class had any power of physical causation. Every symbol the instructor put on the board, (s)he described entirely in terms of other symbols. If I accept Searle's conclusion, that neither Searle nor the Chinese Room "understands" Chinese, then I must also conclude that neither I nor any other entity can "understand" mathematics. What I believe to be my understanding of mathematics must be illusory. By saying this I hope to point out that I think Searle is overlooking something. He gives us the Chinese Room to show us how pure symbol pushing can't produce understanding, but he neglects to mention what happens to *people* who spend sufficient time there. I submit that if Searle spent years in the Chinese Room, dutifully accepting incoming squiggles, looking up the rules to process them, and totting up the proper outgoing symbols, in time he would come to "understand" Chinese EXACTLY as he has already come to "understand" mathematics. That is, as he practiced he would progressively find himself going to the rulebook less and less, yet still producing "correct" answers. In time, he would internalize enough of the rulebook that he could toss it aside altogether. If he was clever enough, then (just like a budding mathematician) he would start making up *new* rules (perhaps by superposing sets of the original rules). He would eventually reach a point where he could be making sense of incoming squiggles that weren't in the original rulebook at all, because he would by then be extracting new information from novel incoming symbols. In other words, syntax *can* (somehow) give rise to semantics. Yes, you may need something "smarter" in the Chinese Room than a combination of present-day digital computer hardware and software, but I think that is immaterial. Searle's thrust is that *no matter who* is in the Chinese Room, as long as he/she/it starts off with no knowledge of Chinese and only a list of rules, he/she/it can never arrive at an "understanding" of what is going on. If that is true, then all mathematicians are imposters. They believe they have semantics, but according to Searle they do not. Dan Mocsny dmocsny@uceng.uc.edu
rapaport@gort.cs.Buffalo.EDU (William J. Rapaport) (12/14/89)
In article <3120@uceng.UC.EDU> dmocsny@uceng.UC.EDU (daniel mocsny) writes: >.. I am struck by the similarity between the Chinese Room and how I felt in .. >mathematics classes. The .. instructor .. made squiggly marks, .. then >proceded to show the class a set of rules .. [to] transform the squiggly marks >into other squiggly marks. .. the symbols began to take on some kind of >*meaning* for me. .. I didn't just have the syntax, I started to obtain >semantics. When that occurred, I felt I was "understanding" mathematics. Precisely this analogy is discussed in: Rapaport, William J. (1986), ``Searle's Experiments with Thought,'' Philosophy of Science 53: 271-279 and in Rapaport, William J. (1988), ``Syntactic Semantics: Foundations of Computational Natural-Language Understanding,'' in J. H. Fetzer (ed.) Aspects of Artificial Intelligence (Dordrecht, Holland: Kluwer Academic Publishers): 81-131
bwk@mbunix.mitre.org (Kort) (12/14/89)
In article <14698@eerie.acsu.Buffalo.EDU> rapaport@gort.cs.Buffalo.EDU.UUCP (William J. Rapaport) reacts to Dan Mocsny's objection to Searle's assertion that symbol manipulation cannot engender understanding: > Precisely this analogy is discussed in: > Rapaport, William J. (1986), ``Searle's Experiments with Thought,'' > Philosophy of Science 53: 271-279 > and in > Rapaport, William J. (1988), ``Syntactic Semantics: Foundations > of Computational Natural-Language Understanding,'' in J. H. Fetzer (ed.) > Aspects of Artificial Intelligence (Dordrecht, Holland: Kluwer Academic > Publishers): 81-131 Bill, could you favor us lazy net.readers with a synopsis? Did you come to the same conclusion as Dan? I for one, am persuaded that one can detect the isomorphism between a formal symbol system and real-world experience. I do it whenever I adopt an apt metaphor. --Barry Kort