From newshub.ccs.yorku.ca!torn!cs.utexas.edu!sun-barr!ames!haven.umd.edu!darwin.sura.net!zaphod.mps.ohio-state.edu!rpi!batcomputer!cornell!uw-beaver!pauld Wed Aug 12 16:52:25 EDT 1992
Article 6563 of comp.ai.philosophy:
Newsgroups: comp.ai.philosophy
Path: newshub.ccs.yorku.ca!torn!cs.utexas.edu!sun-barr!ames!haven.umd.edu!darwin.sura.net!zaphod.mps.ohio-state.edu!rpi!batcomputer!cornell!uw-beaver!pauld
>From: pauld@cs.washington.edu (Paul Barton-Davis)
Subject: Re: Memory and store/retrieve.
Message-ID: <1992Aug4.181806.28275@beaver.cs.washington.edu>
Sender: news@beaver.cs.washington.edu (USENET News System)
Organization: Computer Science & Engineering, U. of Washington, Seattle
References: <1992Aug3.200351.3632@mp.cs.niu.edu> <1992Aug3.220654.20920@beaver.cs.washington.edu> <1992Aug4.165958.17775@mp.cs.niu.edu>
Date: Tue, 4 Aug 92 18:18:06 GMT
Lines: 106

In article <1992Aug4.165958.17775@mp.cs.niu.edu> rickert@mp.cs.niu.edu (Neil Rickert) writes:
>In article <1992Aug3.220654.20920@beaver.cs.washington.edu> pauld@cs.washington.edu (Paul Barton-Davis) writes:
>  I've avoided giving detailed explanation of my assumptions on evolution,
>and have used rather simplistic language in describing them, because I
>sense that most readers of this group are more concerned about the issues
>relating to intelligence.

Well, as someone said in the philosophy session at ALife III recently,
"we'd be in much better shape if we had had 30 years of Artificial
Life and were just getting around to thinking about AI." :-))

>>	   For my part, I actually wrote a small test program to
>>           try this out. Its appalling simple, but it confirmed my
>>           own intuition - mutations that affect reproductive fitness
>>           have to have a *very* dramatic effect in order to affect
>>           their own (rather then a system of which they later become
>>           a part) survival, in the presence of environmental 
>>           noise (weather, predators, disease, etc.)
>
>  I'm not sure what is your definition of "*very* dramatic".  

Neither am I. Yet.
							       Evolution
>takes place over very long periods of time, so that even very small
>statistical advantages can have effects.  I believe there are plenty of
>natural experiments which demonstrate that beneficial traits will tend
>to persist at levels higher than could be explained by their mutation
>rate.

The point of this experiment is to show that in fact, many traits that
persist confer no advantage at all, or at least, no advantage when
they have to compete with environmental "noise".

>  However, all this is beside the point.  You are raising the question
>"if a genetic change is beneficial, what is the probability that this
>change will result in evolution toward a new species?".  

Absolutely not. The question I am asking is "given that a trait
exists, what is the chance that it is actually beneficial ?", which is
the one you suggested next. I claim, very tentatively, that the state
space for biological organisms, along with the actual ecological
terrain in which they interact, are structured so that much of the
diversity we see in living systems is merely diversity for its own
sake, and that the most of the traits reflected by that diversity are
neither beneficial nor harmful.

>  Here, then, are my assumptions:
>
>  It is generally agreed that intelligence has evolved, and that it is
>beneficial.  I admit the possibility that perhaps intelligence is not
>beneficial at all, and it is due to mere homocentric hubris that we
>assume that there is a benefit - but I intend ignoring this extreme
>view.

Why is this extreme ? If you buy into a generally neutral theory of
genetic evolution, one would expect this to be the norm, and for
"genuinely beneficial" developments (such as bipedalism, stereo vision
and so forth) to be the exception.

>  Now if we look at the evolution of intelligence, say from primitive
>fish through primates, we seem to have two possible explanations:
>
>   (a)  There were many small genetic changes, most of which were
>	beneficial to the species.
>
>   (b)  There were many small changes most of which had little
>	beneficial effects, but by random luck just happened to contribute
>	toward development of a highly complex intelligence structure.
>	Finally, at later stages all the parts fell together and the
>	combination suddenly became highly beneficial.
>
>  Personally I find (b) highly implausible, and (a) quite plausible.
>This plausibility argument is all I am depending on.  I don't claim this
>proves anything.  We will prove something only when we either fully
>understand the workings of the brain, or we have truly been able to
>mechanically replicate human-like intelligence, or preferably both.

Or much more likely, when ALife gets to the point where the
phylogenies of its objects are sufficiently rich that evolutionists
get seriously interested. This is happening already, and is why I
mentioned Ray's loop-unrolling optimization as merely one example,
which happens, it seems, to be of type (b). Some references:

Ray, T. S.  1991.  ``Is it alive, or is it GA?''
Proceedings of the 1991 International Conference on Genetic Algorithms,
Eds. R. K. Belew and L. B. Booker, San Mateo, CA: Morgan Kaufmann, 527-534.

Ray, T. S.  1991.  ``An approach to the synthesis of life.''
Artificial Life II, Santa Fe Institute Studies in the Sciences of
Complexity, vol. XI, Eds. C. Langton, C. Taylor, J. D. Farmer, &
S. Rasmussen, Redwood City, CA: Addison-Wesley, 371-408.

Ray, T. S.  1991.  ``Population dynamics of digital organisms.''
Artificial Life II Video Proceedings,  Ed. C. G. Langton,
Redwood City, CA: Addison Wesley.

Ray, T. S.  1991.  ``Evolution and optimization of digital organisms.''
Scientific Excellence in Supercomputing: The IBM 1990 Contest Prize
Papers, Eds. Keith R. Billingsley, Ed Derohanes, Hilton Brown, III.
Athens, GA, 30602, The Baldwin Press, The University of Georgia.
Publication date: December 1991.


-- paul
-- 
"Fun" is a word invented by advertising executives to sell soft drinks.


