From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!news-server.csri.toronto.edu!rpi!uwm.edu!linac!mp.cs.niu.edu!rickert Tue May 12 15:48:57 EDT 1992
Article 5401 of comp.ai.philosophy:
Newsgroups: comp.ai.philosophy
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!news-server.csri.toronto.edu!rpi!uwm.edu!linac!mp.cs.niu.edu!rickert
>From: rickert@mp.cs.niu.edu (Neil Rickert)
Subject: Re: Systems Reply I (repost perhaps)
Message-ID: <1992May4.181702.13708@mp.cs.niu.edu>
Organization: Northern Illinois University
References: <6637@skye.ed.ac.uk> <1992May1.185606.31991@mp.cs.niu.edu> <6648@skye.ed.ac.uk>
Date: Mon, 4 May 1992 18:17:02 GMT
Lines: 119

 Sigh.

 Why is it that people will not read what I say, but instead insist on
putting words in my mouth, and then going ahead and criticizing their own
invention which they have incorrectly imputed to me.

In article <6648@skye.ed.ac.uk> jeff@aiai.ed.ac.uk (Jeff Dalton) writes:
>In article <1992May1.185606.31991@mp.cs.niu.edu> rickert@mp.cs.niu.edu (Neil Rickert) writes:
>>
>>  Perhaps you should tell us what you think constitutes "thought" or what
>>would be "thought" in a computer.
>
>I'm sorry, but I'm not going to play the definition game.  If you're
>interested what "thought" should mean, you might try looking in the
>philosophical literature.

  There is an example.  I did not ask for a definition.  I know how well
you like to refuse giving definitions.  I didn't ask 'what "thought" should
mean' but only asked for an informal idea of what you would consider to
be "thought" in a computer.  To illustrate what I was asking, I said:

>>                                              To put this in perspective,
>>I am treating a computer chess program as using thought (or the computer
>>equivalent),

  Face it.  Your "I'm not going to play the definition game" is nothing
but a deliberate obfuscation to conceal the fact that you don't have
the faintest idea what you are talking about.

  Prove me wrong by discussing what would constitute computer thought.  Maybe
the discussion will be enlightening.

>             Since many things can be done w/o thought, why can't
>passing the Turing Test be one of them?

  Do we have to make the Turing Test into a religious war.  The Turing Test
is just an ad hoc test, since at the time Turing proposed it (and for
that matter, at the present time) nothing better was available.  Wouldn't
it be better to defer judgement on this until something actually passes
a full Turing Test?

>                       What I had in mind was that a human could think
>A-0 while producing behavior B, or could think A-1 while producing B,
>and so on, where the A-i are thoughts (or multi-thoughts).  That is,
>you can't tell what someone is thinking by how they behave.

  I don't see where computers need be different in this respect.  There
are many different ways to produce the same output in a computer.

>>>(Suppose a computer had been turned off and when booted claimed
>>>to have been thinking all the while.  Would you believe it?
>>>Are you convinced that it's behavior would have to show it
>>>had been turned off?)
>>
>>  The same question could be posed to a human who claims to have
>>spent time in thought.
>
>Only if you can turn them off and reboot them and they nonetheless
>claim to have been awake and thinking the whole while.  Something
>like this happens for dreams, but it now seems reasonably clear
>that dreams take place over time while asleep and are not just
>instantly constructed as fake memories on waking.
>
>In any case, how about answering the questions?  Would you believe it?
>Are you convinced that it's behavior would have to show it had been
>turned off?

  I did answer the question.  However, since you prefer to ignore my answer,
I will comment in more detail.  If a computer were turned off, then on reboot
claimed to have been thinking all the time, including the time it was
turned off, and if there had been no infusion of external data (a disk
transfusion), I would probably treat this as confusion.  No big deal, here,
since people are often confused about past events too.  A computer
implementation of AI which was never confused and never made simple
mistakes would be a quite poor implementation of human level cognition,
and would surely fail the Turing Test.  On the other hand, if the computer
claimed to be thinking the whole time only in the sense that it was
completely unaware of the time gap, I would treat that as quite
unsurprising.

  Would the computer's behavior show that it had been turned off?  Only in
the sensed that it would not be aware of events that occurred while it was
turned off.  This need not be much different from a person who went into
a brief coma, then on recovery was unaware that there had been any
interruption of consciousness.

>>  Behavioral tests cannot prove that thought
>>was absent, because they cannot distinguish between absence of
>>thought and the existence of purely worthless thought.
>
>Do you think they can prove that thought is _present_?  How?
>
>For instance, to use your Chess example, are you sure you can distinguish
>between a player who looks at the board a while and forms a judgement
>w/o thought and one who looks at the board for the same time while
>thinking?

  Of course you can't.  This is why the Turing Test is not allowed to be
a limited test.  Everything must be allowed.  Almost anything can be
faked by a human, or by a machine.  But the more extensive the test, the
lower the probability that it could be all faked.

  When a computer first passes an extensive Turing Test, you can bet that
lots of people will go through its programming with a fine tooth comb to
see whether this was fakery or real intelligence.

>What I am suggesting is that (1) it may turn out that as a matter
>of fact humans produce certain behavior (eg, passing the TT) with
>the aid of though while ai programs produce similar behavior w/o
>the aid of thought; and (2) merely looking at the i/o behavior
>is not enough to tell what's going on inside.

  Then let's stop the silly arguments, and wait till a computer passes
the TT.  Then let's look inside and see whether it was really faking or
not.  My main reason for doubting that it would be faked is that to
successfully fake an extensive TT would require a computer program of
unimagineable combinatorial complexity, and I consider that unlikely in
the extreme.



