NetNews Usenet Archive 1992 #19/NN_1992_19.iso/spool/talk/origins/10588

Text File | 1992-09-01 | 6.2 KB | 117 lines

Newsgroups: talk.origins
Path: sparky!uunet!cis.ohio-state.edu!zaphod.mps.ohio-state.edu!sdd.hp.com!wupost!uwm.edu!linac!att!cbnewsm!mls
From: [e-mail] (mike.siemon)
Subject: Re: The rise of consciousness
Organization: AT&T
Date: Tue, 1 Sep 1992 20:05:18 GMT
Message-ID: <[e-mail]>
Summary: this still misses Edelman's central point
References: <emc.715304048@keiko> <[e-mail]> <emc.715368211@uhura>
Keywords: brain,evolution,consciousness
Lines: 104



In article <emc.715368211@uhura>, [e-mail] (Eli Chiprout) writes:

> I find this very much more interesting.  The reason a system is
> indeterminate however, is not because it is analog or digital,
> appearances aside.   This can be shown mathematically.
> In fact digital processing carries information much
> more accurately and over longer periods of time than can analog.

That is what I was noting in my aside.  Nor would Edelman object;
his position is much more thoroughgoing than you seem to be getting
from my statements of it (which I would not claim to be particularly
clear -- I just hope to suggest the realm of discourse involved.
I have only just stumbled on this stuff, and have spotty background
in the neural biology Edelman invokes -- I have a fair amount of
homework to do before I could conclude that I either fully understood
or agreed/disagreed with his theories.  I am not offering judgment,
pro or con, expert or otherwise, on his theories.  I *do* again
suggest that his work provides an interesting and provocative take
on the issues of consciousness and its "embodiment."

Information processing becomes a relevant consideration AFTER the
eleaboration of perceptual categories and the rest of the mental
apparatus built on them.  Once that is in place, information models
are both possible and have distinct engineering advantages.  Edel-
man is saying that such a view misses the point of how such systems
come to form in the first place.

Your critique would seem to depend on assumptions that Edelman (and
I, for what little that is worth) would reject -- most basically an
assumption that there *is* in some objective sense "information"
in the real world, as some kind of Platonic thingie that we have
a "picture" of.  Edelman poses his recognition system explicitly as
a rejection of representational and information-theoretic models of
(human) cognition.

> is why we are turning to digital phones, HDTV, cd's, DATs and soon
> digital radio transmission.  That is why DNA is carried as digital
> information.

Fine -- but the manner in which DNA "information" constrains actual
neural development is *highly* indeterminate, dependent on accidents
of place and timing that generate a huge variational population that
is then continually, throughout life modified on an adaptive basis.
This is the substrate of Edleman's theorizing.  All of this, at the
DNA level and the neuronal level, is adaptive -- but the logic in
use is that of Edelman's recognition systems, and depends on PRIOR
variation with selection dependent on an undefined and arbitrary
environmental interaction in time.  It is that, the "blooming,
buzzing confusion" of reality that HAS no inherent catgeorical
structure but from which the brain constructs, by a sort of stochastic
jumble of processes, the very definition of what *is* a signal and
what *is* information in a given context -- THAT is the analogue
nature Edelman is talking about.  How does an engineer construct a
digital process for processing information when he does not know
WHAT counts as a signal or WHAT its bandwidth and other relevant
parameters are?

> But noise introduced into a system cannot be the cause of a sophisticated
> end product such as consciousness.

Edelman would, I suspect, say that you have the whole process backwards --
the noise is initially ALL that there is, and the "sophisticated end
product" is explicable in precisely the same way as the "marvellous design"
of biological species.

For there to *be* a "system" into which "noise" is introduced already
mis-states the nature of neural systems (again, pace Edelman).  We can
make marvelous digital systems for sound recording, on the basis of a
spectral analysis of the kind of sound signals that we, in fact, are
aware of.  But the evolutionary origins of our response to sound and the
organs we construct for that and most particularly the mapping from the
sensory apparatus to cortical functions in the brain are ALL a matter
of adaptation.  The system EMERGES from the "noise" and what ONCE was
noise, is no longer so.  We *learn* the relevant discriminations on
the basis of what our genes have "learned" in their somewhat more meta-
phorical education through natural selection.

I fear that the quotation, with its use of digital vs. analog has led
you on a wild-goose chase -- the main thing he is getting at is that in
design of automata, there is a determined engineering effort to eliminate
noise and to make state transitions as quick and effective as possible
and among as few states as possible.  Human nervous systems by contrast
wind up with highly ambiguous "transitions" (which may never complete)
and adaptive construction of new states and elimination of older ones,
all in response to "whatever-it-is-we-learn-to-respond-to" -- which
changes over time.  Anything at all in the external (and internal)
world could be modulating "signals" and in some more or less extended
spectrum -- that is the kind of "analogue world" Edelman means -- a
world in which we adaptively TRY OUT variant systems built out of an
ineradicable (and vital) population of neuronal variation.

I've already spent more effort on this thread than I intended -- I
just wanted to point people to a good read.  As I say, I found his
arguments, so far as I have read them, cogent -- you may of course
think otherwise.  "Neural Darwinism" is in no position (at this time,
anyway) to become a consensus theory of cognition -- but neither is
anything else, so go thou and read and consider.
--
Michael L. Siemon        The Son of Man has come eating and drinking;
                and you say "Behold, a glutton and a drunkard,
[e-mail]            a friend of tax collectors and sinners."  And
standard disclaimer        yet, Wisdom is justified by all her children.