Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.2 9/5/84; site yale.ARPA Path: utzoo!linus!philabs!cmcl2!seismo!lll-crg!ucdavis!ucbvax!decvax!yale!andrews From: andrews@yale.ARPA (Thomas O. Andrews) Newsgroups: net.math Subject: Re: Mind as Turing Machine: a proof *and* a disproof! Message-ID: <108@yale.ARPA> Date: Thu, 7-Nov-85 00:06:05 EST Article-I.D.: yale.108 Posted: Thu Nov 7 00:06:05 1985 Date-Received: Sat, 9-Nov-85 06:15:32 EST References: <509@klipper.UUCP> <1096@jhunix.UUCP> <2081@umcp-cs.UUCP> <702@ecsvax.UUCP> Reply-To: andrews@yale-comix.UUCP (Thomas O. Andrews) Distribution: net Organization: Yale University CS Dept., New Haven CT Lines: 64 In article <702@ecsvax.UUCP> hes@ecsvax.UUCP (Henry Schaffer) writes: >Following this path, the rephrased question is "Can the mind be modelled >by a Turing machine?", and it can't be answered by showing speed >differences -- but it could by showing that the mind can "compute" >something that a Turing machine can't. (Vice-versa isn't possible, because >it is evident that the mind can simulate a Turing machine.) >--henry schaffer n c state univ Is it so evident that a mind can emulate a Turing machine? Even a mind with access to paper cannot necessarily emulate a Turing machine. A Turing machine theoretically has perfect, infinite memory. One thing a Turing machine cannot be is interactive in any sane sense of the world. It starts with a fixed set of inputs at time zero. This doesn't allow for a conversation, say. It may be able to respond to a questions, but it cannot ask questions in a simple way, and expect an answer. This would involve changing the values on a tape. What about interactive turing machines? Two turing machines "discussing something." This would involve an entirely new system - one in which each turing machine could write on blank squares of any other turing machine, but could not read the strip of any turing machine but it's own. This would mirror our communications. On the other hand, this pair of turing machines can be computed, and hence can be computed by a single turing machine - in fact whole thousands of "communication turing machines" could be represented in one Turing machine! Scizo-Turing devices! But if we get Billions and Billions of these turing devices together, and have them talk at each other, we could easily get something as complicated as the human brain (er, I think.) For instance, if we have one turing machine that behaves somewhat the way a neuron does, we could create a hoarde of them and set the machine(s) rolling. I'm not making myself clear, am I? I think I'll go to sleep. But first, a brief flame: Someone said something to the effect that "at the moment, Turing machines are not capable of ..." What the hell does 'at the moment' mean when talking about an abstract concept. After define a Turing machine, what it can do and what it can't do is defined, too. "At the moment, we do not know whether a Turing machine can ..." It is mathematically illiterate and bad thinking to interpret our inability as a deficiency in the machine. Since there are things we *know* turing machines can't do, it is best to be careful when talking about "current limits" of our own knowledge. Good night. Oh, and by the way, any more polar bear problems, and I think I'll kill someone. I didn't know there were so many high school students on the net. -- Thomas Andrews "Gosh, I used to know how to do that." Favorite excuse of engineers