From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!csd.unb.ca!morgan.ucs.mun.ca!nstn.ns.ca!news.cs.indiana.edu!spool.mu.edu!sol.ctr.columbia.edu!bronze!chalmers Thu Apr 16 11:34:26 EDT 1992
Article 5086 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!csd.unb.ca!morgan.ucs.mun.ca!nstn.ns.ca!news.cs.indiana.edu!spool.mu.edu!sol.ctr.columbia.edu!bronze!chalmers
>From: chalmers@bronze.ucs.indiana.edu (David Chalmers)
Newsgroups: comp.ai.philosophy
Subject: Re: A rock implements every FSA
Message-ID: <1992Apr14.064323.16366@bronze.ucs.indiana.edu>
Date: 14 Apr 92 06:43:23 GMT
Article-I.D.: bronze.1992Apr14.064323.16366
References: <1992Apr2.155348.19580@oracorp.com>
Organization: Indiana University
Lines: 47

In article <1992Apr2.155348.19580@oracorp.com> daryl@oracorp.com (Daryl McCullough) writes:

>I am not arguing for Putnam's thesis that everything implements everything,
>but I am arguing for the thesis that functionalism is trivially different
>from behaviorism. The single-state machine can be made functionally equivalent
>to the 5-colour-map-checker by adding a clock (so that no two consecutive
>states are the same).

If you're going to make claims like that, you have to give an argument.
So far, I haven't seen any constructions that work.  Every attempt has
depended on some invalid trick such as using material conditionals
instead of strong conditionals, or not mapping inputs to inputs.
A clock will simply not serve to distinguish states on different
potential pathways (a list of inputs might, as I said before, but
that's a different matter).

>I think it is mistake to treat all transitions inside the brain as
>input transitions, since most transitions are not affected by external
>influences. That is the reason why I separated the input transitions from
>the internal transitions.

The transitions in a FSA aren't just "input transitions"; they're
overall transitions in the state of the system, part of which change
is due to input.  You might want to argue that the part of the
transition that's due to input should be separated from the part
that's independent, but that would best be handled by separating
states themselves into two or more components, hence by moving
from the monadic FSA formalism to a formalism involving FSAs
with combinatorially structured states.

>Anyway, you haven't answered the question about whether a brain
>without connections to ears, eyes, muscles, etc. is functionally
>equivalent to a rock (or a clock). In that case, there are no inputs
>and no outputs, and there are only silent transitions, and it seems like
>to me that any system with only silent transitions is functionally equivalent
>to any other (with the right number of states).

As I've said a number of times, the brain can handle inputs, even if
it might not be receiving actual inputs at a given time (or only null
inputs).  It's sensitive in complex ways to all kinds of stimulations
at its surface, and in virtue of this sensitivity implements a complex
FSA structure that the rock does not.

-- 
Dave Chalmers                            (dave@cogsci.indiana.edu)      
Center for Research on Concepts and Cognition, Indiana University.
"It is not the least charm of a theory that it is refutable."


