Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: Notesfiles; site ea.UUCP Path: utzoo!linus!philabs!cmcl2!floyd!vax135!houxz!houxm!ihnp4!inuxc!pur-ee!uiucdcs!ea!mwm From: mwm@ea.UUCP Newsgroups: net.philosophy Subject: Re: Can computers think? - (nf) Message-ID: <9800002@ea.UUCP> Date: Tue, 29-May-84 12:28:00 EDT Article-I.D.: ea.9800002 Posted: Tue May 29 12:28:00 1984 Date-Received: Sat, 2-Jun-84 10:40:25 EDT References: <1698@stolaf.UUCP> Lines: 79 Nf-ID: #R:stolaf:-169800:ea:9800002:000:4125 Nf-From: ea!mwm May 29 11:28:00 1984 #R:stolaf:-169800:ea:9800002:000:4125 ea!mwm May 29 11:28:00 1984 /***** ea:net.philosophy / stolaf!johnsons / 4:04 pm May 18, 1984 */ >Computers are NOT "effectively Turing machines." Turing proposed his >test to answer the question, Can computers think? As has already been pointed out, a Turing Machine (something Turing invented to solve the problem of whether all mathematical problems could be done by machine) is unrelated to the Turing Test (something Turing thought up while pondering whether machines could think). >In the literature >that I have studied, no psychologist or computer scientist even claims >that there is a machine now that passes Turing's test satisfactorily. The first case of a machine passing the Turing test was over a decade ago. The program in question was "Perry," a paranoid program (the mafia was out to get it - bad gambling debts, if I remember correctly). 60% of the psychologists who "talked" to Perry incorrectly identified it as a paranoid human. Of course, simulating a paranoid is *much* easier than simulating a "normal" human being, as you only have to handle one subject to any depth, and a correct simulation will return to that subject at will. Nobody thinks Perry is intelligent; but this case sums up the state of AI nicely: able to handle small subsets of intelligent behavior reasonably well, but nowhere near capturing all the nuances of intelligent behavior. >I am not a mind/body dualist. However, the people who claim that >computers can think seem off hand themselves to be dualists. They >separate a mental process, thinking, from the physical organism and >give it to a computer, saying that a computer can indeed think. This >seems to me absurd logic. You have two different kinds of dualism in your hands. The mind/body dualists hold that the mind (whatever it is that thinks/is you) is separate from the body. I (as a person who thinks that computers will some day think) hold that an algorithm is seperate from the hardware it runs on. Thus, implementing the algorithm on new hardware is possible. >Thinking cannot be surgically removed from >the biological organism; to do so is to create a monumental blunder, >the very same blunder Descartes made in creating the mind/body dualism >in the first place. I don't understand this statement. What do you mean by "cannot be surgically removed ... "? At one level, this is a tautology; there isn't a "thinking" to be identified and cut out of a biological organism. On another level, it's completely wrong, as AI isn't an attempt to remove thinking from an organism, but to add thinking to an organism. >And he I repeat myself--we we understand the word >"thinking" we understand it in terms of certain criteria, those criteria >being observable behavior; talking, writing, acting logically. The >computer possesses none of the behavior that we normally ascribe to >the word "thinking." Computers *do* talk (I heard one talking to the blind lawyer that owned it via a votrax) and write (they've been generating stories, poetry and music for years - none of it very good, but...). As for acting logically, they are a lot better at that than most people I know (including me). These three acts are only a small part of what we mean when we say "thinking," and aren't even necessary parts (or do blind illiterates not think?). >In this sense, computers do not think as we >think. They do not partake of our language, our understanding of what >it means to think. I obviously don't partake of your understanding of what it means to think. If I didn't speak English, would that mean that I didn't think? Your first sentence, however, hits the nail on the head: "computers do not think as we think." They don't think (period) now, and will probably never think as we think. There are just to many things involved in being human for that to occur; if it does, then what you have isn't a computer, it's a human being. However, there is nothing to prevent computers from thinking in ways other than we do. This won't prevent them from participating in whatever society happens to be around then, just as it hasn't prevented humans from interacting with each other.