Artificial Sight

Matt Jones jonesmat at physiology.wisc.edu
Mon Aug 6 12:47:55 EST 2001


In article <R8Sa7.5507$xj.819641 at typhoon.southeast.rr.com> Glen M.
Sizemore, gmsizemore at triad.rr.com writes:
>Subject: Re: Artificial Sight
>From: Glen M. Sizemore, gmsizemore at triad.rr.com
>Date: Sat, 04 Aug 2001 12:46:41 GMT
>>MJ: Hi Glen,
>
>
>GS: The phrase "semantic
>argument" is usually preceded by "merely a..."
>Obviously, I disagree with this view. And I'm
>not sure that it doesn't characterize your
>position, despite what you have written
>above.
>


Glen,

"merely a semantic argument" is what people say when they really mean:
"It doesn't matter how you define the terms, it all means the same
thing anyway". This is explicitly -not- my position. My position is
that it matters a great deal how you define the terms, because that
affects what you are able to do with the concepts to which they refer.
A good definition allows you to make progress. A loose or ambiguous
definition impedes progress.


Shannon used colloquial meaning to guide his choice of definitions. He
uses logarithms and entropies because using them satisfies the
colloquial intuition that information should be additive, etc. That
is, two books ought to convey about twice the information as one book,
etc. He also attempted to quantify the entropy of English, and a bunch
of other stuff that we would probably associate with trying to get at
the colloquial meaning. Nonetheless, I'd say that his final product
far surpasses the colloquial meaning of information as far as being
able to do anything useful is concerned. Maybe another way to think of
it is that his final theory -crystalizes- the central aspects of the
colloquial meaning, and jettisons all the nonessentials.



>
>GS: Again, this is metaphor. The news server
>doesn't do anything like what a person does when
>he or she is said to interpret. Now, mind you (and
>this is extremely important), I am not offering
>"interpret" as a useful scientific term (I hold that it is
>not).
>


You say that the newsserver doesn't act like a person when he or she
is said to interpret. I say that depends very much on context. A
newsserver isn't very much like a person interpreting a lecture, or
reading a novel. But it's an awful lot like a mailman delivering
letters, or a switchboard operator routing phone calls. The fact that
switchboard operators have been almost universally replaced by telco
switching circuits illustrates the -fact- that these circuits are at
least as good at performing the kind of interpretation that is
necessary for that task as a person is.

What's so special about people anyway? It seems that no analogy or
metaphor would satisfy you unless it includes an actual Homo sapiens.
What's the point of that? Do you think there is nothing to be learned
at all by studying non-human subjects. Is there something magical
about human beings that allows them to "interpret", whereas chimps,
rats, snails, bacteria and electrical circuitry lack this special
quality? What exactly is this quality anyway?

Or is it that it's only -organisms- that have this special quality? 
And why would that be? What is it about organisms that separates them,
in principle, from machines? Some sort of special life-force, no
doubt.


>And there is much about your assertion that is sort
>of funny. Your reference to multiple transformations
>is perhaps more apropos to physical systems in
>which energy of one sort is transferred through a
>system, especially when it is changed in form - as in
>a steam engine for example. 

Yeah. OK.


>Now that raises
>interesting questions! Is the fire sending a message
>to the piston? 

Yes. 


>Are the wheels of the steam
>locomotive interpreting this message? 

Yes.


>What about
>systems in which the energy doesn't change form?
>Am I sending a message to the rock when I lift it
>with a lever? 

Actually, no. You would however be sending a message if you were to
move the rock in such a way as to alter its entropy. Anyway, in this
example, energy is changing form: chemical->mechanical.


>Or maybe,
>and if I were you I would make this argument, your
>kind of "explanation" only comes into play when the
>system cannot be understood either in terms of the
>transference or transformation of energy?
>Yeah.....that's it! You could argue that only when
>these conditions do not hold do you have the kind of
>system shown by the unconditioned reflex. In a way
>that's true......but it still doesn't make your position
>any more cogent.


No. This is exactly the opposite of what I am saying. Information only
makes sense in systems where the transferrence or transmission of
energy is the central aspect of what the system is doing. More
precisely, it's the entropic component of the energy, and changes in
it, that carry information. But I think I see where you're coming from
a little better now. You seem to think that the redistribution of
energy has nothing to do with the transmission of information. This
idea fits in with your attitude that there is something special about
how people interpret things.


GS: Of course I mean something special when I say
>behavior. The "behavior" of a telephone has little to
>do with the behavior of organisms.


Yeah, you keep saying things like that. I mean, trying to state that
this definition isn't the same as that definition, that this system
has nothing to do with that system, etc.  But so far you have not
stooped to actually explain -why- this system isn't the same as that
system. What is it about an -organism- that makes it so different in
principle from the internet, as far as how it handles information is
concerned?

As far as the principles used to process information, I fail to see
any fundamental qualitative difference between a steam engine, a
telephone, the internet, bacteria, a snail, a cat or a person. No
doubt this is because I am so silly, naive and abstruse.


>
>
>MJ: Have a nice day,
>
>Matt
>
>GS: Au revoir,
>
>Glen



Kindest Regards,


Matt





P.S.: This discussion reminds me of a quote from Stanislaw Lem
(paraphrased):

Question: "Do you believe that machines can think?"

Answer: "I believe that -only- machines can think!"




More information about the Neur-sci mailing list