Newsgroups: comp.ai.philosophy
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!vixen.cso.uiuc.edu!uwm.edu!news.alpha.net!news.mathworks.com!udel!gatech!rutgers!argos.montclair.edu!hubey
From: hubey@pegasus.montclair.edu (H. M. Hubey)
Subject: Re: Bag the Turing test (was: Penrose and Searle)
Message-ID: <hubey.787006567@pegasus.montclair.edu>
Sender: root@argos.montclair.edu (Operator)
Organization: SCInet @ Montclair State
References: <CzFr3J.990@cogsci.ed.ac.uk> <D00167.91w@spss.com> <3bu0gs$fff@sun4.bham.ac.uk> <hubey.786764192@pegasus.montclair.edu> <3c47mj$fvr@toves.cs.city.ac.uk>
Date: Fri, 9 Dec 1994 20:56:07 GMT
Lines: 108

jampel@cs.city.ac.uk (Michael Jampel) writes:

>H. M. Hubey <hubey@pegasus.montclair.edu> wrote:

>>Now we are in a paradoxical situation: it has to made less
>>"intelligent" to pass the TT.

>This is not a criticism of the above post, but perhaps a co-operation to
>discuss the TT's characteristics....

ok.

>1. There is a difference between having knowledge and being intelligent.

here's where we start running into trouble with what we somehow
feel is intelligence and is defined to be the same for all of us. 
But it is not so. I'm not so sure that there's such a fine line
between knowledge and intelligence. Certainly the IQ tests cannot
make such a clear distinction. 

And finally if you are saying that intelligence is some kind of
an inference engine and not the data on which it operates, it
still doesn't make much difference because then the argument will
simply shift to whether this inferencing engine can be given to
machines. And then remember the HLT vs code problem. There's
a time-space trade-off like the macro vs subroutine dichotomy.
Theoretically we can expand the subroutines everywhere so that the 
expansion need not be done in time, but in space so that they behave
like inline macros.


>2. There is a difference between intelligent behaviour and socially
>acceptable behaviour.

YES. But if humans are the yardstick for intelligence how do we know
that we are not confusing our behaviour with intelligence. Or is it
certain that our socially acceptable behavior is not a result of
our intelligence? There's some variation among societies but the
variation if looked at from a point of view of all possibilities is
not so great. All modern societies resemble one another. Put another
way, men and women are more alike than unlike if looked at within
a space of all possibilities.



>3. The TT is not about intelligence, but about fooling a human into
>thinking that you (the computer) are a human, which may be related to
>intelligence but also has something to do with social behaviour.

If we don't know what intelligence is then how do we know that? If
we can't define it, then maybe we can at least recognize it. Could
you understand things like space, time, temperature from dictionary
descriptions? Can a blind man comprehend colors ? the way we do?

And once again, as pointed out by so many people already, all of
these discussions are really about "what is intelligence". On the
one side we have people who are attmepting to slay dragons with 
other dragons by introducing even less understood concepts to
define intelligence; understanding, awareness, pure consciousness,
self-awareness, creativity...

On the other side we have people who are saying that we should get
started with some kind of a yardstick. Psychologists attempt this
with IQ tests,and AI people have the TT.


>So if a computer is to pass specifically the TT, which includes some
>kind of idea of mimicking human behaviour, even when that behaviour is
>driven by social conditions, then the computer will have a `"Don't
>appear to be a smarty-pants" module'' or a `"Everybody hates a
>smart-arse" module'. 

It's still the same problem. Is becoming socially acceptable
a part of intelligence?  Is not harming others, what is called
enlightened self-interest or just morality (whatever it means).

>say that the humans under test have access to a computer, so that if you
>ask "What is the 123rd decimal place of pi" you can get the answer "My
>computer says it is x" from the human, so the computer under test can
>say something similar.


Let's look at this slightly differently. Let's ask people
what they think is intelligent by asking for examples. Mathematicians?
Yes, most people will say so. just ask the mathematicians :-)..
Chess players? yes. Quick learners? Yes. Those who can size up social
situations very quickly? Yes, people will say they are "bright".

But the anti-AI party keeps upping the ante at each step. As soon
as machines start to outperform humans, that specific mental
task becomes something that no longer measures intelligence.

well, I think the anti-AI crowd owes the world some kind of
an explanation of what kind of behavior intelligence is. They
don't have to provide a definition but maybe they can give
examples of what kinds of actions constitute what we might
consider intelligent. When examined closely they all fade
away as not being particularly good measures.

The best that I can think of at the moment is what might be
termed "creativity" since it seems to be the closest thing
that humans can do that is not or does not seem "mechanical"
or "rote".  

--
						-- Mark---
....we must realize that the infinite in the sense of an infinite totality, 
where we still find it used in deductive methods, is an illusion. Hilbert,1925
