From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!psych.toronto.edu!michael Tue Mar 24 09:54:39 EST 1992
Article 4376 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: <1992Mar2.190455.17079@mp.cs.niu.edu> <1992Mar2.214012.22715@psych.toronto.edu> <1992Mar10.143639.10130@neptune.inf.ethz.ch>
Message-ID: <1992Mar10.171755.7458@psych.toronto.edu>
Date: Tue, 10 Mar 1992 17:17:55 GMT

In article <1992Mar10.143639.10130@neptune.inf.ethz.ch> santas@inf.ethz.ch (Philip Santas) writes:
>
>In article <1992Mar2.214012.22715@psych.toronto.edu> michael@psych.toronto.edu (Michael Gemar) writes:
>>Let's take an example from physics.  You teach a child how to calculate
>>the potential energy of a pendulum by the formula:
>>
>>PE = 1/2 k x^2
>>
>>where PE = potential energy
>>       k = spring constant
>>       x = displacement from equilibrium
>>
>>You do so as a purely mechanical operation.  If asked, he would answer
>>that he knows nothing about potential energy, or spring constants, or
>>the like.
>>
>>You then say, "Aha, but the *system* that calculates potential energy
>>in a spring does!"  However, someone who knows electrostatics says,
>>"But wait!  That's also the formula for calculating the electrostatic
>>energy *in a capacitor*.  Simply *interpret* k as C (capacitance) and
>>x as V (potential across the capacitor plates)."  *Now* what does the
>>system "understand"?  Only pendulums?  Only capacitors?  Pendulums *and*
>>capacitors?  I'd vote for neither, myself.
>
>If the system has as input only the mathematical formula, it is
>possible that you get a variety of solutions according to the input.
>k can be number of number of cars, and x anything else.

Absolutely.  This is the point of the example.  But what else *are*
the inputs to computers, *except* functions describable in mathematical
terms?

>The human or the computer needs the definitions that you add after the 
>formula. This means that we talk about mechanical and not electrical
>energy.

How do you give definitions that have semantic content to a computer?
How does it know that when you respond to the request "Please input
the capacitance constant => " that you are talking about capacitance,
and not number of cars?  The problem of generating "definitions" with 
semantic content for a computer *is* the central problem under discussion.

>>BTW, this is a good example of how interpretation can play a large role
>>in the attribution of "understanding" to a program.  If an electronics
>>engineer had a program he used to calculate the PE of a capacitor, he or she
>>would probably say that the program *actually did calculate electrostatic
>>energy*.  But it doesn't.  It merely submits the inputs to certain syntactic
>>rules, and provides outputs.  The *exact same program* could be used to
>>calculate the potential energy in a spring system.  The program itself
>>does not *refer* to capacitors - it doesn't "refer" to anything.  It is
>>only our *interpretation* of the inputs and outputs which give meaning.
>
>Your formula itself does not refer to capacitors either. It is the set
>of definitions of the variables that clear things up

Agreed.  But I do not see a way of providing variable definitions that would
"clear things up" from the point of view of the computer.

- michael



