Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site umcp-cs.UUCP
Path: utzoo!linus!philabs!seismo!rlgvax!cvl!umcp-cs!mark
From: mark@umcp-cs.UUCP
Newsgroups: net.religion,net.misc
Subject: Re: Faith in Evolution.
Message-ID: <567@umcp-cs.UUCP>
Date: Thu, 7-Jul-83 23:06:25 EDT
Article-I.D.: umcp-cs.567
Posted: Thu Jul  7 23:06:25 1983
Date-Received: Fri, 8-Jul-83 05:32:04 EDT
References: <245@cbscd5.UUCP>
Organization: Univ. of Maryland, Computer Science Dept.
Lines: 80


	As far as I can see Holland's book does not even mention
	spontaneous generation.  His book is a mathematical analysis
	of the adaptation process.

Your original message did not mention spontaneous generation.  You
asked for a detailed analysis of the development of living organisms
from an amino acid soup.  (Aside: when I studied at Michigan with
Holland we called this soup Hollandaise sauce.  :-) That is exactly
what holland does.  He shows how the "not in a billion years (or
15) could random processes develop an eye" arguement is wrong by
detailed analyses of how STABLE ADAPATATIONS speed up the adaptations.
Another way to think of this is that all the way to an eye have
been eye-like adaptations which were a little bit adaptive, and so
stayed around.  Once you get a little skew like this in the randomness
(and it doesn't take much) amazing things happen.  His favorite
example was the "2-armed bandit problem":

Imagine you are in Los Vegas playing a two armed bandit.  You
suspect that one arm may give you better odds in the long but
don't know which one it is.  So you must experiment, but at the
same time you must risk your money.  How should you allocate your
coins so that you have the maximum expected return?  Remember,
the good arm could still give you a long run of bad long, so you
can never cease playing both arms every once in a while...
The solution, Holland shows, is to play the arm which has been
better SO FAR exponentially more often than the other arm.
This means evolution is fast.  However, exponentially more often
is not all the time.  Every once in a while, play the other arm.
With this play algorithm, you'll always recover from initial bad
guesses about which arm is the good one.  What does this have
to do with evolution?  The allocation of genes in a gene pool
(population genetics now) corresponds to exactly this exponential
allocation function.  That is, the genes corresponding to
more adaptive features will increase exponentially.

Of course, one doesn't actually need genes.  Holland has
applied this to the evolution of simple strings of integers
(corresponding to RNA or DNA) under random variations.
It works fine there too.  This is what I consider to be
a detailed mechanism of evolution from the Hollandaise sauce
on up.

	He makes applications to Biology (genetics, i.e. living
	systems), but these applications are based on current
	evolutionary assumptions.

The Holland's arguments have nothing to do with current evolutionary
assumptions.  They are based on statistics and logic and little
else.   It just so happens that when he applies them to genetics
he gets out evolution of biological organisms as a theorem...sorry
if this bothers you.

	Please, when you cite a reference, at least give the page
	numbers and a brief description of how the reference applies.
	No one likes being sent on a wild goose chase.

Goose chase not intended.  It seems to me one must understand
Holland's whole approach, that is why I cited no page.  His book
is not easy reading.

	I think you ought to consider the publication as a whole
	before you make such a judgement.

Was not your sample representative?  Why did you post it?

	The article I selected is only a sample to illustrate the
	point that belief in evolution requires a fair amount of
	faith, as does belief in creation.

Going to bed each night hoping the sun will rise requires blind
faith.  When it doesn't, then I'll deal with it.  When I see things
created by a supreme being, then I'll deal with it.  But
right now, I see evolution.
-- 
spoken:	mark weiser
UUCP:	{seismo,allegra,brl-bmd}!umcp-cs!mark
CSNet:	mark@umcp-cs
ARPA:	mark.umcp-cs@UDel-Relay