Newsgroups: sci.nonlinear,comp.ai.philosophy,sci.cognitive
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!uhog.mit.edu!bloom-beacon.mit.edu!news.media.mit.edu!minsky
From: minsky@media.mit.edu (Marvin Minsky)
Subject: Re: Obstructions to Shadowing
Message-ID: <1995Mar31.022212.8953@news.media.mit.edu>
Sender: news@news.media.mit.edu (USENET News System)
Organization: MIT Media Laboratory
References: <3levh8$r7i@portal.gmu.edu> <3lfb9g$9f5@mp.cs.niu.edu>
Date: Fri, 31 Mar 1995 02:22:12 GMT
Lines: 43
Xref: glinda.oz.cs.cmu.edu sci.nonlinear:2783 comp.ai.philosophy:26455 sci.cognitive:7130

In article <3lfb9g$9f5@mp.cs.niu.edu> rickert@cs.niu.edu (Neil Rickert) writes:

>My personal opinion is that it is a mistake to treat the brain as a
>chaotic dynamical system.  It may be, but that is not important.
>Everything is a chaotic dynamical system.  The computer sitting on my
>desk is a dynamical system.  If I were to measure with enough
>precision, I would find plenty of evidence that it is a chaotic
>dynamical system.

[...]

[About a computer] (Two similar computers will show small scale
differences)  It happens that those
>differences make no difference in the output we are looking for, for
>the gates are designed to ignore these small differences.
>
>Any complex system will look chaotic, if its operations are not
>understood.  To say that the brain is a chaotic dynamical system is
>merely to say that we don't yet understand its principles of
>operation.

This is the basic point everyone else seems to miss.  The reason
brains work so well, in my opinion, is that they have evolved to be
substantially independent of microfluctuations, quantum, thermal, or
otherwise.  You can still function after eating a pretzel because of
perhaps a dozen layers of mechanisms of isolation and redundancy of
function that have evolved to protect the higher principles of
operation from chaotic vagaries of the smallest subsystems.  

I don't mean Shannon-type redundancy so much as McCulloch-type -- that
is, "redundancy of potential command", as he put it.  The reason some
of us are good at mathematics is not, as Penrose might have it,
because we can correctly compute the noncomputable, but because we can
try a variety of incomplete and inconsistent methods, guess a
conclusion, and then say "oops" when we find disturbing
counter-evidence.  

I wonder how often Penrose says "oops"?  Well, actually the answer
stands before us, for it would seem that each chapter of "Emperor"
appears to be an "oops" for the deficiencies of its predecessor.

