From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!utgpu!pindor Thu Jan  9 10:34:06 EST 1992
Article 2556 of comp.ai.philosophy:
Xref: newshub.ccs.yorku.ca comp.ai.philosophy:2556 sci.philosophy.tech:1761 sci.logic:762
Newsgroups: comp.ai.philosophy,sci.philosophy.tech,sci.logic
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!utgpu!pindor
>From: pindor@gpu.utcs.utoronto.ca (Andrzej Pindor)
Subject: Re: Penrose on Man vs. Machine
Message-ID: <1992Jan8.191127.27952@gpu.utcs.utoronto.ca>
Organization: UTCS Public Access
References: <1992Jan7.031553.24886@oracorp.com> <1992Jan7.105117.7193@husc3.harvard.edu> <1992Jan7.191853.17310@gpu.utcs.utoronto.ca> <1992Jan7.162542.7202@husc3.harvard.edu>
Date: Wed, 8 Jan 1992 19:11:27 GMT

To my request for a definition of 'understanding' Mr. Zeleny writes:
(article <1992Jan7.162542.7202@husc3.harvard.edu> zeleny@zariski.harvard.edu (Mikhail Zeleny))
>
>Understanding is a faculty of grasping with the mind the concept or
>proposition that constitutes the cognitive content of a singular term or a
>declarative sentence, in virtue of expressing which the term or sentence
>can be said to be meaningful, and may denote a certain object or a
>truth-value.  This definition, like all other attemps at real definition,
>is theory-laden, and should be handled with great care.
>
All this mumbo-jumbo does is replacing 'understanding' with 'grasping with
the mind'. In another posting Mr. Zeleny defines 'understanding' using 
a following phrase 'meaningful i.e. capable of being understood' plus more
mumbo-jumbo. No comment. And it is not a matter of 'handling with great care',
but being useful to decide in a given situation whether there is 
'understanding' or not.
Let's take a thought experiment which I have suggested:
>AP:
...
>>if we feed CR with basics of abstract group theory and then proceed to ask it
>>questions which it answers correctly, how will it's 'understanding' differ
>>from that of a human being?
MZ:
>
>I eagerly await the day when some kindly soul would proffer an argument,
>rather than yet another opinion.  Alas, I see no evidence of an argument in
>the above.  As for the last question, I trust that you understand the
>concept of a finite order group.  On the other hand, no Turing machine
>could duplicate that feat, given that the second-order concept of finitude
>is inherently beyond the ken of a recursive mind.  Case closed.
>
This is not a matter of trust! How would you decide if I understand 'the 
concept of a finite order group'? You would ask me a number of questions and
if you thought that my answers were correct you would (trust or not) declare
that I do. Or do you have a better method, using your impressive definition?
If CR gave the same answers, then how do you know that its understanding is 
superior to mine or inferior to yours? All this talk about first-order
concepts, second-order concepts, recursive mind etc. is truly awing but does
nothing to help us decide if there is understanding in the above simple case.
So stop this mumbo-jumbo, please, and get to the point. You have complained
that you do not get arguments, but only opinions. Sorry, you obviously
did not understand (!) my argument. The argument is: There is no way to decide
that you understand abstract group theory better than a properly equipped CR,
other then asking you and CR test questions. Since there are already programs
pretty good a group theory, CR would have to be said to understand the group
theory as well.
By the way, did you notice that your definitions of understanding sound very 
much like produced by a program of 'Eliza' type?

Andrzej Pindor

PS. It is often said that if one truly understands something one has to be
able to explain it others who don't. How about this as a definition?
(in fact I do not understand your definition of 'understanding'. Do you?).



-- 
Andrzej Pindor
University of Toronto
Computing Services
pindor@gpu.utcs.utoronto.ca


