From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!psych.toronto.edu!michael Tue Mar 24 09:57:29 EST 1992
Article 4611 of comp.ai.philosophy:
Newsgroups: comp.ai.philosophy
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!psych.toronto.edu!michael
>From: michael@psych.toronto.edu (Michael Gemar)
Subject: Re: Definition of understanding
Organization: Department of Psychology, University of Toronto
References: <1992Mar11.122705.22342@neptune.inf.ethz.ch> <1992Mar11.185921.10347@psych.toronto.edu> <1992Mar16.233438.45463@spss.com>
Message-ID: <1992Mar19.203226.18201@psych.toronto.edu>
Date: Thu, 19 Mar 1992 20:32:26 GMT

In article <1992Mar16.233438.45463@spss.com> markrose@spss.com (Mark Rosenfelder) writes:

[discussion of program to calculate potential (or is it electrostatic?) 
 energy deleted.  I argued that just because you have a variable called
 "Potential_Energy" doesn't mean the computer has semantics.]     

>Exactly.  I am afraid that some AI types confuse the names of symbols with
>semantics.  The two expressions you write are strictly equivalent to the
>computer (indeed, many compilers would generate identical object code from
>them).  The first expression is intelligible to a human observer; this 
>must not be confused with understanding on the part of the computer.

I agree 100%.

>I don't think this makes artificial intelligence unattainable.  But 
>if a computer understands a term, it will be because it can relate it
>to an enormous mass of information, experience, and procedures (much as
>happens in a human being), and not because the variables it uses
>have names that resemble English words.

This is where we part company, simply because I don't see how adding more
meaningless symbols is going to help things.  You can put in additional
variables that represent *to you* the ductility of the material that a
capacitor is made of, the weight of the capacitor itself, the level of pain
that one would experience when you were shocked by a capacitor discharging
at the given voltage, etc.  But the same argument as above still applies.  These
are *still* just so many marks to the computer.  They may be marks that depend
on other marks, but there's still no *meaning* for them *in* the computer.
We as observers can *give* them meaning, but that's not the same thing.

- michael



