Newsgroups: comp.ai.philosophy
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!cs.utexas.edu!uunet!swissbank!root
From: gerryg@il.us.swissbank.com (Gerald Gleason)
Subject: Re: definitions of `Strong AI'
Message-ID: <1994Nov14.214726.19229@il.us.swissbank.com>
Sender: root@il.us.swissbank.com (Operator)
Nntp-Posting-Host: ch1d264nwk
Organization: Swiss Bank Corporation CM&T Division
References: <HPM.94Nov13000910@cart.frc.ri.cmu.edu>
Date: Mon, 14 Nov 1994 21:47:26 GMT
Lines: 87

Hans Moravec writes
> 
> Aaron Sloman opines:
> > I did have some other [positions on AI] that are obviously false, e.g.
> > (5)
> >    There exists an algorithm such that ANY computation expressing
> >    it would be intelligent.
> >
> > But it is not obviously obviously false: you have to think a while
> > to see that it is obviously false
> 
> I don't consider this position absolutely false at all.  It has an
> apparent falsity only relative to our limited interpretive capacity:
> WE couldn't look at such a description and have a conversation with
> it.  But a very clever observer, or one very attuned to the particular
> representation on the paper, could have such a conversation, and
> appreciate any portion of the life of the entity described.  To that
> observer the intelligence would be as real as if it were encoded in a
> robot body.

But would the intellegence be in the algorithm on the piece of paper, or  
in the system, that is, paper plus interpretive system?  As soon as you  
consider the later, and try to formalize your definition of the  
interpretive system, you are in danger of getting sucked into an infinite  
regress of interpretive systems.  I don't see any choices besides  
grounding interpretation in a physically realized system (i.e either  
naturally evolved living organizms or systems invented by living systems),  
or in some "supernatural" entity.

Perhaps it is a human limitation, but I have difficulty attributing  
intellegence to a static artifact, unless you are relying on onmiscience  
for grounding.  Beyond this, you are basically stating the "other minds"  
problem.  In other words, just because we can't interpret any signals from  
a system as intellegent, doesn't mean it isn't, and conversely, just  
because a system exibits complex behaviour, doesn't mean it is  
intellegent.

> Call this the Platonic position on mind: minds (and the experiences
> they imply) are abstract entities that physical representations,
> whether bodies, robots or books, merely encode.

I suggest that this should be turned on its head.  The physically  
discoverable universe is much more of an abstract entity than an  
unmodified Platonic position implies.  What if the many-worlds  
interpretation of QM is true, and the physical world is equivalent to the  
full extension of a symetric energy event to all possible consistent  
world-lines.  Something like Universe == All-that-Is ==  
All-that-is-possible.  This is no longer dualistic, as is the Platonic  
position because it makes abstract principles just as "real" as any  
physically observed phenomenon.

> At any time, our own limitations create a practical interpretive
> horizon, within which we can see real minds, and beyond which we
> can't.  [ ... ]    Without the interface
> machinery, these people are just meaningless electrical noise in some
> chips.  To themselves, of course, they are real either way, because
> self-interpretive machinery, constituting their consciousness, is part
> of their abstract description.

Again, this is "other-minds", but I don't see it as meaningful to propose  
that such an intellegence is "real" or "exists" if the systems is never  
expressed "in the world".  We may not be able to interpret this  
"electrical noise" as intellegent, but there must be some type of signal  
to interpret, and there must be entities that can and do interpret the  
signals.  I'm saying that a "self-interpretive" system is not possible  
unless you consider the "system of everything" (i.e. the Universe).   
Finally, I will just state a complete understanding of this system is an  
unbounded task.  The proof is an exercise for the reader.

> I realize that this position implies that all possible experiences,
> including arbitrarily horrible ones, exist in the infinite realm of
> Platonic minds, which are perfectly real to themselves, whether or not
> particular observers happen to appreciate them.  We can't spare them
> those experiences by choosing to not build, simulate or otherwise
> bring them to our awareness.  Such implement/don't implement decisions
> can affect only our own experiences.  I can live with that.

Although I have suggested that what you are calling "Platonic minds" don't  
exist in the way you have suggested, my speculation on many-worlds  
supports a simmilar observation about the existence of "all possible  
experiences".  As to what it would mean for us that these possible worlds  
exist (or not), that is an open question, and meaning is always  
problematic in any absolute sense.

> 			-- Hans Moravec    CMU Robotics

Gerry Gleason
