From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!pindor Fri Jan 31 10:26:38 EST 1992
Article 3227 of comp.ai.philosophy:
Newsgroups: comp.ai.philosophy
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!pindor
>From: pindor@gpu.utcs.utoronto.ca (Andrzej Pindor)
Subject: Re: Intelligence Testing
Message-ID: <1992Jan28.220534.1523@gpu.utcs.utoronto.ca>
Organization: UTCS Public Access
References: <11979@optima.cs.arizona.edu>
Date: Tue, 28 Jan 1992 22:05:34 GMT

In article <11979@optima.cs.arizona.edu> gudeman@cs.arizona.edu (David Gudeman) writes:
>In <1992Jan27.201926.20498@gpu.utcs.utoronto.ca> Andrzej Pindor writes:
>]Several people can read the same instrument and decide whether they agree what
>]it says. Then they accept what they agree on. That's how science is done.
>
>Do tell.  Is that how science is done?  A committee sets up an
>experiment, then at various points in the proceeding, each member of
>the committee steps up to the meter and says "Yup, that's the reading
>all right".  Of course this is ridiculous, scientists don't verify
>experiments by verifying individual readings.  That would be
>hopelessly inconvenient and would also lead to methodological errors.
>Scientists verify experiments by repeating the experiment in a
>different place, with different subjects and different instruments.
>
I am puzzled that you take my words so literally. Is it so hard to understand 
that what I wanted to say was that science is done by different people observing
the same phenomena and reaching consensus on what they observe?

>]But no one except you can introspect on your 'understanding' (or whatever). 
>
>Of course not.  Others introspect on their own understanding.  The
>fact that the thinker as well as the subject is different is a
>methodological _advantage_, not a problem.
>
How can this be a 'methodological _advantage_' if essence of science is people
_agreeing_ on what they see or conclude. If everyone talks about their OWN
introspection, which by your definition is incommunicable to others, there is
no place for consensus. Science is supposed to study objective phenomena, 
independent of observer's bias, hence a requirement of consensus.

>]Hence it is not science.
>
>Hence you are inventing arbitrary and unfounded "rules of science" to
>prove your point.
>
Am I? I bet majority of people would agree that science is supposed to be
independent of observer's bias and requires others to agree that they all
observe the same thing. In mathematics it means working with precise defini-
tions, again to make sure that everyone is talking about the same thing.

>]Haven't you found out that your introspection tells
>]you sometimes contradictory things (am I happy or not? am I conscious or am I
>]dreaming?)
>
>Haven't you found out that people have reached contradictory
>consensuses?
>
Sure, but if this was science, then the controversy was resolved by a reference
to an experiment(s), in case of mathematics by producing an example or a proof
everyone (almost) agrees upon. If you are talking about a field where there
is no way, even in principle, to resolve the controversy, then this is not
science. Perhaps religion is a better description. And do not say again that
I introduce my own definitions, since even on this list several people already
expressed a similiar understanding of what is science and what should rather
be called religion.

>] In case of reading an instrument you can compare your experience
>]with that of others, in case of feeling that you understand, you can't. The 
>]first is science, the second is not.
>
>How do you know that comparing your experience with that of others
>leads to more certainty in empirical observations?  The answer is that

By experience! 
I have found by experience that if I rely only on my own judgement I am
more likely to err then when a given judgement is shared also by many other
people. There have of course been cases when individuals were  proven right,
even though no one shared their views at a time when they came up with them.
However there were infinitely more cases to the contrary.

>you determine this by introspection.  Of course you can put off
>introspection by one step by appealing (rather weakly, in my opinion)
>to inductive reasoning, but at the end the reason you finally believe
>anything involves believing because introspection tells you that you
>have enough evidence.  
>
It looks to me here like you are using word introspection in a slightly diffe-
rent sense. If you are talking about post-experimental mental processing, which
can be communicated to others, then I have no objection. It is a crucial part
of science. But originally you were talking about introspection of your
subjective feelings which are incommunicable. I do it too, but I claim that as
long as it is not amenable to examination by others it is not science and 
an opinion whether a machine can be built which is capable of such introspection
is in a category of (religious) beliefs.

>     ....Therefore, _nothing_ you know can be more
>certain than what you know by introspection alone.

Although Mr. Zeleny objected to me calling this opinion solipsism, I believe
that it is not very far from it. If the above is true, why bother to do 
anything else but introspection? Why spend billions on scientific research
instead of sitting in an armchair and introspecting? Have you seriously thought
about logical consequences of this statement? 

>]>No, I have a third choice.  I can simply observe that there is no
>]>known method of determining consciousness _in others_ which is free of
>]>_believing that they are like you_, and reserve my opinion on whether
>]>such a thing might ever be possible.  This, coincidentally, is the
>]>choice I take.
>]>
>]This seems to imply to me that according to you consciousness is possible 
>](by your definition) only in beings that _are like us_, unless proven
>]otherwise.
>
>Can't you see the difference between claiming that a thing is
>impossible and claiming that there is no known way to detect it?
>Especially given that I have already said I do not believe it is
>impossible.
>
I am a bit confused here, may be I misunderstood you. So you think it might
be possible (at least in principle) to detect 'sef-awareness of understanding'?
You've said on another occasion that human understanding is semantic (as opposed
to syntactic understanding in machines, present or imagined). So you think
that there is a physical basis to semantics? Perhaps difference between our
views is not as great as I've imagined.
On the other hand you've also expressed grave doubts as to whether intelligence
could arise by accident, as opposed to design. This would suggest that you
think that humans were designed by an entity with a knowledge superior to
ours, since it knew (knows?) how to construct machines endowed with semantic
understanding (I assume here that you agree that we are machines - see below).
This of course cannot be proven, neither disproven, so I put it into a bin
marked religion.

>]It also seems that you cannot imagine how this could ever be proven and even
>]tend to think it is impossible in principle.
>
>Only because you are reading into my comments what you want to see.  I

Why should I want you to see any particular views on your part? I am not
interested in discussing my views of your opinions but what I understood were
your views. 

>don't accept your blind faith in the behaviorist method of detecting
>mental things, but that does not mean that I deny the possibility of
>any other method of detecting them or the possibility of their
>existence in other things.
>
I do not have blind faith in the behaviorist method of detecting mental things
(here you seem to be reading into my comments what you want to see :-(), but
for now I do not see any other alternative. If you convince me that there is
a better method, I'll only be too happy. I find the behaviorist method far from
satisfactory, but I am resigned to it (for now at least) since not only myself,
but no one else (including most vocal anti-AIers) has come up with anything
better. So I see two alternatives: either use this defective method or shut
up and wait for better times. Note however that no one has even proven that
a better method exist, so may be nothing better is possible.

>]To me it sounds like: 'I am convinced that humans have souls until you prove
>]otherwise. In fact it might be impossible in principle to disprove existence
>]of a soul'
>
>That's because you aren't reading what I'm writing.  I never mentioned
>souls, and if you replace "soul" with "mind", the sentence still does
>not reflect what I said.  Why do you AI'ers all insist on trying to
>tell me what I said and always get it wrong?  Could it be an
>unconscious defense mechanism?
>--
I think I wasn't the only one who got an impression that you consider human
beings as endowed with something special that no machine can have (above
I used the word 'soul', but we are not going to argue about a specific term,
are we?). Now I am a bit confused - so would you agree that we are machines
(i.e. can be constructed, when we finally find out how this machine works)
or not? As I said, I have no opinion one way or another (in lack of evidence
it would be a religious belief), but existing experimental evidence seems to
me consistent with the possibility that we are. Saying that we are not would
mean assuming existence of something for which there is no evidence. My
experience favours using Occam's razor. 
>					David Gudeman
>gudeman@cs.arizona.edu
>noao!arizona!gudeman


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


