From newshub.ccs.yorku.ca!ists!torn.onet.on.ca!utgpu!utcsri!rpi!usc!sdd.hp.com!ux1.cso.uiuc.edu!mp.cs.niu.edu!rickert Tue Jun 23 13:21:07 EDT 1992
Article 6301 of comp.ai.philosophy:
Newsgroups: comp.ai.philosophy
Path: newshub.ccs.yorku.ca!ists!torn.onet.on.ca!utgpu!utcsri!rpi!usc!sdd.hp.com!ux1.cso.uiuc.edu!mp.cs.niu.edu!rickert
>From: rickert@mp.cs.niu.edu (Neil Rickert)
Subject: Re: Transducers
Message-ID: <1992Jun18.020154.18895@mp.cs.niu.edu>
Organization: Northern Illinois University
References: <60837@aurs01.UUCP> <1992Jun17.182829.18441@mp.cs.niu.edu> <60839@aurs01.UUCP>
Date: Thu, 18 Jun 1992 02:01:54 GMT
Lines: 55

In article <60839@aurs01.UUCP> throop@aurs01.UUCP (Wayne Throop) writes:
>> rickert@mp.cs.niu.edu (Neil Rickert)
>> Once you have an analog signal, it is quite hard to apply a Searle -
>> style argument and say that the signal is symbolic and/or syntactic.
>> Harnad interprets this as implying that analog is essential.  I prefer
>> to interpret it as persuasive evidence of the bogosity of Searle's
>> argument.
>
>Hmmmm.  Yes.  Well, here's a thought pump that attempts to show that it
>*would* be possible to apply a Searle-style argument to analog,
>"transducer-only" systems.
>
>Let's take as an example a feedback mechanism to keep a turntable
>rotating at constant speed.

 [example of speed governor deleted for brevity]

>But still.  The analog regulator has no more "semantics",
>"groundedness", "causal powers" or anything else than the digital one.

  As I am sure you are aware, you can't really argue something like this
from one example.  Keep in mind that Searle's argument didn't prove
anything; all it did was make AI sound preposterous.

  Try describing an analog system, where all type of sensory data are
represented.  Now ask if it seems plausible that this could contain
semantic information.  Many people will not find any obvious reason
that it could not contain semantics.  But when you do the same thing
with symbolic and syntactic data, the idea that it carries semantics
sounds preposterous.  The difference is that the terms "symbolic" and
"syntactic" conjure up in our minds only very specific ways of using
computers.  We might think of expert systems, or chess playing programs,
but we don't tend to think of digital TV, or the simulation of tornadoes
on a supercomputer.  Searle's argument is really a very clever deception,
for it tricks you into a very restricted view of what a computer can
do.  [I don't mean to accuse Searle of any intentional deception here;
he probably deceived himself too].

>No matter how hard I try to understand it, this attempt to distinguish
>"nonsymbolic" processes as somehow distinct in and of themselves from
>"symbolic" ones makes no sense to me.  Processes are processes.  The
>symbolic or nonsymbolic distinction is purely in the eye of the
>beholder.

 I agree, and that is why I consider Searle's argument to be bogus.

>           The attempt to label computer processes as always symbolic
>simply begs the question of whether computer processes can themselves
>ever be legitimate beholders.

 Right.  And you are getting closer to the real question.  Is there
necessary information which is not subject to any kind of computation
and transduction?  I tend to doubt it, but would be interested to see
arguments in that direction.



