From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!wupost!darwin.sura.net!mojo.eng.umd.edu!mimsy!harwood Mon Dec  9 10:47:52 EST 1991
Article 1845 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!wupost!darwin.sura.net!mojo.eng.umd.edu!mimsy!harwood
>From: harwood@umiacs.umd.edu (David Harwood)
Newsgroups: comp.ai.philosophy
Subject: Re: Penrose
Message-ID: <44174@mimsy.umd.edu>
Date: 4 Dec 91 07:36:24 GMT
References: <1991Nov20.002510.5654@husc3.harvard.edu> <1991Nov20.174026.6107@spss.com> <24983.292ae33f@oregon.uoregon.edu>
Sender: news@mimsy.umd.edu
Distribution: usa
Organization: UMIACS, University of Maryland, College Park, MD 20742
Lines: 74

In article <24983.292ae33f@oregon.uoregon.edu> stevew@oregon.uoregon.edu writes:
>I have two things to say about Penrose and "quantum effects." First, I
>think for a scientist to say "It can't be done. EVER." is unscientific.
>Second, I think that to attribute the thought process to something as
>undefined as a quantum phenomena is romantic, and perhaps possible, but it
>also serves his purpose to put AI out of human reach.
\\\\\\\\\\
	As I recall - and I read his book the first week of its publication-
physicist Penrose claims that human neurology and intelligence are not
recursive. His physical arguments are not very persuasive, by themselves,
and neither are his appeals to his own creative mathematical intuition.
	They are a possibility. There does have to be some explanation
of human creative intelligence. Nevertheless, his theories seem to be
essentially physicalist and deterministic. Not the sort of thing that
should trouble an atheistic believer in AI. At worst, you might have
to reverse-engineer a brain - something not very far-fetched in the future
- suitably rebuilding it (only approximately) as some other non-recursive
physical process. You simply could not determine the behavior of either
by a program for a Turing machine.
	Some other comments on quantum mechanics and neurology:
	As I recall, it has been known for 50 years that a single photo-
receptor (a specialized neuron) can reliably, but uncertainly, "detect"
by "firing", the incidence of one quantum of light. And psychophysical
experiments of dark-adapted human vision (with shifted S/N sensitivity)
demonstrate that humans can consciously detect an incident flash of only
10-20 quanta, upon the retinal array, with statistical significance.
(see Cornsweet's text, for references).
	Some people have asked me why should the nervous system be 
so physically sensitive, if this would make for unreliable neurological 
information processing. My reply has been that it may be selected by 
evolution: in the night, any glimmer of light, however faint, might be 
a reflection of a predator. So our dark-adapted sensitivity is at the 
limit of physical information.
	I don't see how it is possible for a Turing machine to 
deterministically compute human or neurological behavior on the basis of 
physically observable information. (This is not to say that such behavior
is proved to be not recursively computable, or determined, with hidden
information; also there might still be probabilistic simulations, will fail
to be deterministically predictive.)
>
>My final comment is to say that the ever-present denial of the possibility
>for AI among many people I have talked to often comes down to a religious
>sentiment that human beings are either created by God, ARE gods, or are
>LIKE gods. To a non-believer like myself, this is ludicrous. I don't mind
>my role as a natural machine, so long as I have free will :-)
>
\\\\\\\\\\
	Well - I am religious, believing in existence of a supernatural
Creator. I do not claim to know how powerful C is. (I might suspect, as
a heretic, that C has some limitations, but I am not prepared to be 
burned at the stake for this suspicion.) Anyway, it does seem to me that
if we admit the possibility of creating artificial intelligence - in our
own image, in some sense - then, logically, we have to admit that we also
may have been created similarly by Someone else. (Nevertheless, I do
not believe in C because of these possibilities.)
	Actually, you do have "free-will" - the question is what does
this mean. Very much of the time people actually do what they want-
they do satisfy their intentions as allowed by circumstances; this is
a pretty remarkable coincidence of our existence, by itself. It could be
otherwise- that we were forever dissatisfied, no matter what we did.
	You have "free-will" in a second sense: it is perhaps theoretically
impossible for anyone or anything in this world to determine what your
behavior will be, to predict what human history will be.
	You have "free-will" in a third sense: a significant part of our
everyday existence is, after all, predictable, so that we can act and think,
to some large extent, to satisfy our intentions, which we could not if
existence was largely random (and unintelligible). We have trouble with
sentient beings and very sensitive physical processes.
	 But as Einstein said, perhaps paraphrasing Schopenhauer, 
"We can do what we want, but we cannot want what we want." (However,
Einstein did not profess to believe in a Creator, although his rhetorical
allusions suggest that his religious intuitions were uncertain.)




