X-Message-Number: 9290
Subject: more careening about the cranium
Date: Sun, 15 Mar 1998 22:56:09 -0500
From: "Perry E. Metzger" <>

> From: Thomas Donaldson <>
> Date: Sat, 14 Mar 1998 01:14:18 -0800 (PST)
> 
> HI everyone!
> 
> To Mr. Metzgar:

"Metzger". The last vowel is an "e", not an "a".

> It is flatly false that neurons are at all simple.

The entire design for a neuron is, by definition, contained in at most
the several billion bits of information located in the DNA in the
nucleus of a human cell. That is less storage than I have attached to
the computer I am typing this on.

Perhaps the term "simple" is throwing you.

I will agree that by the standards of someone dealing with 19th
century technological equipment, a neuron is a complex device. We are
no longer living in the 19th century, however. Humans now routinely
deal with devices composed of billions of individual designed parts -- 
literally billions. A neuron is not nearly as complicated as the
computer on my desk.

It is true enough that to completely understand the computer I am
using requires years of study. It is, none the less, a fairly
straightforward device, composed of well understood
components. By modern standards, these are not grotesquely complicated 
devices. I'm sure they would be thought of as fiendishly complex by
the standards of an earlier age, but as I continue to emphasize, we no 
longer live in that age.

Similarly, neurons are not, by modern standards, complicated
devices. They are far simpler than even the most inexpensive
commercially sold personal computer on the market today. This is not
to say that I believe such computers are particularly suitable for
simulating a neuron, but that the standards of complexity have
shifted.

> I should go on at this point in great detail, but suggest that you
> start your education in this subject

I know far more about neurobiology than you do about computers. The
howlers you and the rest put out on a daily basis, filled with
misunderstandings of theoretical and applied computer science, are
among the funniest things I routinely see here on Cryonet.

Unfortunately, most of the people here feel no compunction about
discussing topics they barely understand with authoritative zest.

> It may also help if you did a little study of the differential
> equations involved in predicting the motion of planets,

Thank you, Mr. Donaldson, but I took physics in engineering school,
long ago.

> an excellent and beautiful example of chaos in the technical
> sense. As for how chaos becomes involved, it becomes involved
> because we could only make an accurate prediction of the motion of
> all the bodies in our Solar System if we had completely accurate,
> infinite-precision arithmetic.

Here we get the fallacy of perfect prediction show up yet again.

To recap for the reader, the argument is this: can a computer simulate 
a brain?

Certainly a computer cannot PERFECTLY simulate a particular analog
system. On the other hand, if I made a molecule for molecule copy of
your brain, and fed it identical inputs, the copy would quickly fall
out of sync with the original, too. Analog systems can no more
"perfectly simulate" identical analog systems than anything else can.

The question is not "can I produce absolutely perfectly identical
output", but instead "can I produce output indistinguishable by any
conceivable test". If the range of outputs of the digital system is
indistinguishable from that of the analog system, the digital system
is functionally simulating it, never mind whether it produces
"identical" results, which even two runs of the same analog equipment
would not produce.

> Furthermore, I strongly suggest you look up the reference which
> Mr. Freeman kindly put back on Cryonet. A counterexample is a
> counterexample; no one claimed that the nonTuring "computer" was
> better, merely that it was not a Turing computer.

The "counterexample" is not a counterexample. Its not even remotely
correct. It assumes that we could construct devices that cannot be
constructed given the physics of the universe we live in -- that is
not a "counterexample", that is mental masturbation. It is like
"disproving" special relativity by saying "assume we have a
faster-than-light ship".

> As for the illusions produced by digital pictures, so far I have not
> seen one which I would mistake for reality.

Boy, you've lived a deprived existence. *I* can't distinguish half the 
stuff they show at SIGGRAPH, and I used to be a professional in the
field.

If you wish, I'll happily bet you a large sum of money that I can
produce twelve photographic images for you, of which six are computer
generated and six are "unretouched", and that you would not be able to
select which are which better than flipping a coin would be able to.

> What I am conscious of, and how my brain runs, are two different things; it
> often happens that consciousness comes AFTER you make a response.

Ah, so you are a believer in the Cartesian Theater. I suppose we'll be 
hearing references to your pineal gland shortly.

I have sad news for you: trying to pick the point at which you become
"conscious" of something is about as silly an exercise as trying to
determine the number of degrees of grace of the Virgin Mary.

> Not long after, but after ... several experiments show this.

Hehehehehehehe!

Okay. If you can pinpoint the moment of consciousness, could you
please explain away experiments like Color Phi perception, in which
subjects happily pinpoint the location of perceptions they never had,
or the "cutaneous rabbit" experiment, in which subjects happily report
after the fact sensations which they have previously contradicted?

To have a moment of "consciousness" would imply a Cartesian
theater. There isn't a Cartesian theater. There is just your brain,
which is made up of nice little neurons made of ordinary everyday
matter.

> The main problem with any digital imitation of the world is chaos.

The problem with this notion (the latest in the series of desperate
attempts to pretend humans couldn't be simulated by computers) is, as
I have previously noted, that "chaos" would doom any analog simulation 
of you as well. In fact, by your argument, since a person molecularly
identical to you won't behave the same way, he couldn't possibly be
conscious.

As I've noted, the question isn't one of "identical simulation" which
is a canard -- the same neuron couldn't put out the same response
every time it was stimulated, either -- the question is merely one of
producing something statistically indistinguishable.

> To apply that to brains, I think it unlikely that a digital picture
> will give such an exact representation of reality that the response
> of my brain, over time, to the digital representation will not
> diverge more and more from its response to a real experience.

This is a meaningless observation.

I mean, your "experience" in another bloody ROOM will be different
from the experience you are having right now. Of course a simulated
world will produce a different experience than the real world -- as
would sitting in an identically furnished room, or even sitting in the 
same room if I could only somehow "run you again" through that
room. So what?

> As for quantum mechanics, it's not at all clear to me that the notion of
> "digital" retains any meaning.

It isn't clear your sentence retains any meaning, either.

Perry

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=9290