Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!yale!cs.yale.edu!blenko-tom
From: blenko-tom@CS.YALE.EDU (Tom Blenko)
Newsgroups: comp.ai
Subject: Re: Searle and biology
Message-ID: <25559@cs.yale.edu>
Date: 13 Jul 90 02:09:39 GMT
References: <14265@venera.isi.edu>
Sender: news@cs.yale.edu
Reply-To: blenko-tom@CS.YALE.EDU (Tom Blenko)
Organization: Yale University Computer Science Dept, New Haven CT  06520-2158
Lines: 41

In article <14265@venera.isi.edu> smoliar@venera.isi.edu (Stephen Smoliar) writes:
|Before I left on my vacation, Tom Blenko had raised the issue of whether or not
|Searle was claiming any superiority of biological implementations over those of
|silicon regarding the issue of "understanding."  Unless I am mistaken, he
|introduced this claim when he last spoke at UCLA.  His basic claim was that
|there was something about the biological substrate which enabled the
|implementation of intentionality.  He never said what that "something"
|was, but that did not prevent him from asserting that it was absent in
|any silicon implementation.  Needless to say, I had trouble buying into
|this claim.

I said Searle's claim is that a system composed of a program running on
a silicon processor need not be equivalent to the same program running
on a biological "processor" (for the purposes of talking about
intelligence).

His argument goes like this: even if a program running on a silicon
processor produced the same observable, extensional behavior as the
biological system, there are intensional properties (emotional states
of being happy, angry, alert, confused, and so forth) that he (and, he
asserts, others) take as being essential to an intelligent entity.

Searle's solution to the mind/body problem is that these intensional
properties *are identically* collections of physical states of the
underlying system.  So an intensional property ("being hungry") is
exactly some set of possible states of the system's neurons.  The
biological substrate doesn't "implement" the intensional properties, it
exactly is the intensional properties.

It follows that since the underlying states of the silicon and
biological processors need not be equivalent, the systems constructed
by applying them to the same program are not equivalent.


Searle may have made stronger claims, he's not claiming (by this
argument) that non-biological implementations of intelligence are
unattainable, just that the specification of a program (independent of
its processor) is insufficient.  This is presented as an argument
against "top-down" AI (my terminology) in its various guises.

	Tom