Newsgroups: comp.ai.alife
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!gatech!bloom-beacon.mit.edu!eru.mt.luth.se!news.luth.se!sunic!news.chalmers.se!news.gu.se!gd-news!d6245
From: sa209@utb.shv.hb.se (Claes Andersson)
Subject: Re: Lamarckian Evolution
Message-ID: <1995Feb9.192347.13781@gdunix.gd.chalmers.se>
Sender: usenet@gdunix.gd.chalmers.se (USENET News System)
Nntp-Posting-Host: d6245.shv.hb.se
Organization: Dept. of economy and computer science.
X-Newsreader: News Xpress Version 1.0 Beta #2.1
References: <3gs6v7$2og@laplace.ee.latrobe.edu.au> <3gu1cb$3l1@morrow.stanford.edu> <1995Feb5.132734.948@gdunix.gd.chalmers.se> <1995Feb6.201628.3154@ictv.com>
Date: Fri, 10 Feb 1995 01:43:06 GMT
Lines: 52

tmh@ictv.com (Todd Hoff) wrote:
>In article <1995Feb5.132734.948@gdunix.gd.chalmers.se>,
>Claes Andersson <sa209@utb.shv.hb.se> wrote:
>> Oh! What a relieve, reinforcements finally arrived! I have tried to explain
>>for several weeks why Lamarckism is impossible and why it's completely
>>unimportant question weather it would be good to as someone put it "encode"
>>accuired traits. It would be like telling your maths teacher that it is more
>>convenient to look up the anwers in the solutions than to solve the tasks.
>>And this without seeing that someone must have solved them once.
>>
>>Claes Andersson. University of Bors. Sweden
>
>How is the pattern/solution lookup different than the fight/flight reaction
>we have that seems built in? We have other externally triggered behaviors that
>are built in as well. Where does one draw that line at what is possible?

 I've tried to tell you.. Evolution is a very special process that allows search thru
great searchspaces and optimisation of unlinear problems. Darwinian evolution
does this because it is the only possible way for it to be done, it is the operator
on the genes. Whatever operators we have for other things is irrelevant. Lamarckian
evolution doesn't qualify. It's like I've said to:

 EXCHANGE KNOWLEDGE OF MATH WITH A SOLUTIONS BOOK

You solve an equation with certain methods. You suggest that it would be faster to
don't care about the methods and go right for the solutions. Very well, here's some
news: Someone used the methods you suggest we bypass to create the solutions
once. The logic thinking and the mathematical methods are a paralell to the Darwinian
evolution and Lamarckism suggests that you should solve the problems without
solving them.

 You can use learning, but it is NOT Lamarckism. Do you really understand how
tremendously complicated the genetic code is? Do you understand that a mechanism
for encoding traits and introducing them in the ontogeny so that the change don't affect
other traits would be numbercrushing of an impressing magnitude. I assure you: there
is no enzymes or something that can figure out how to create a gene for a specific trait.
I think you ought to study some embryology to get a hunch of how genes actually work.

 Just a simpe problem.. a typical Lamarckian inheritance would be that if you stretched
to reach things often, your offspring would get longer legs.
1) After how many stretchings is this function called, is it called all of the time.
2) If the problem of short legs had got evolutionary attention enough to build
    a tremendously complex recognition of the event, wouldn't it have been easier
    just to lenghten the legs at once?
3) What exactly happens that is by any means possible to recognise when you stretch
     for something? Billions of cells are involved and the effect it has an single cells are
     often identical to effects of other actions. Where does this "thing" that recognice
     of stretch sit? Has it got little strings to all cells and tally the number of stretching?

Claes Andersson. University of Bors. Sweden


