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