Alexei Sharov wrote:
> ***************************************************************
> Now we come to the current point of our discussion. This is
> my reply to Luis:
>
> It looks like our discussion goes around the notion of open-ended
> evolution. I view evolution as a branching process; each branch is a
> new species (or quasispecies, following Eigen and Schuster). The
> evolution is open-ended if the number of branches has no finite
> limit when time -> infinity. My point is that there is no effective
> method to determine if the evolution is open-ended.
I agree that there is no effective means to determine if evolution is
open-ended at this time. However, in the logic of self-reproduction, we
can very well compare systems as to their ability to produce and
re-produce themselves and/or their components. As I said in previous
messages (posted to the list in the last message from Alexei),
template-reproduction and reproduction by self-inspection are much more
restricted than reproduction by coded descriptions. The problem with the
first type of reproduction is that the reproducing objects must be, at
the same time, both messages and products. Therefore life-forms based on
this kind of reproduction are restricted to physical arrangements that
are able to take both of these roles. This is quite distinct from a
description based reproduction where ANY product can be made out of a
description. In this latter case, the description does not have to be a
product, and the product does not have to be a description. This way, we
are able to produce many more products, indeed, we can produce ANY
product that is representable in the description language -- this is
what Pattee means by open-endedness. In addition, the description
language, or syntax, can also be much more efficient because its
components do not need to function as products.
Arantza Etxeberria and Jesus Ibanez have written a very good paper
outlining in detail the distinction between these types of
self-reproduction: "Semiotics of the Artificial: the "Self" or
Self-Reproducing Systems in Cellular Automata". I am not sure if this
paper is available on the web, but it is very worth reading.
> RNA are able both to fold and to copy themselves.
Yes, but that is precisely why an RNA world is not as effective in
evolution as a DNA/RNA -> Protein world. Because the RNA world is
restricted to those RNA molecules that need to be both non-reactive
enough to function as messages, and reactive enough to function as
catalysts, its evolutionary potential is quite limited. In fact, recent
developments in RNA research show that if a RNA world existed, then it
may have required some sort of intermediate coding mechanism between
non-reactive RNA molecules (messages) and reactive molecules (products).
THe process of RNA editing found in most genetic systems today may have
offered that coding relation by altering and integrating several
non-reactive RNA messages into reactive RNA products. If this is true,
then an ancient RNA world could have been endowed with a symbolic code
between messages and products, in which case the two roles would be
clearly distinguished.
Some people think RNA editing processes in current day genetic systems
(DNA -> Protein) is a sort of mithocondrial relic of an ancient RNAA
world. But if this is the case, then the RNA world relied itself on a
code and not on molecules with simultaneous message/product roles.
> Protein is not
> a good example. You argue that the variability is limited to those
> molecules that can fold and unfold. But this is the same
> self-organization constraint that you seem to recognize in the
> next paragraph:
>
> >Sure, but those are the self-organizing (both enabling and restraining)
> >constraints that I acknowledge in reproduction. I have argued for the
> >importance of this embodiment elsewhere.
No. The constraints I acknowledge refer to the bio-chemistry of the
system/environment coupling, and naturally, since natural symbol systems
have to be implemented physically somehow, they will also be subjected
to such constraints. Open-endedness does not mean physically
unconstrained -- everything has to follow the laws of nature. It means
simply that given some physical product of a self-reproducing system
(e.g. proteins), the description-based mechanism of self-reproducing
system can produce any and all (physically) possible configurations of
that product. Again, the products are not restricted to those capable of
re-producing themselves by self-inspection or "templating".
Imagine the difference between building a house by looking at a similar
house. The house needs to be simple enough that all its structure is
available for you to look it an copy it. It also needs to be fairly
sequence independent, because by just looking at it, you don't know
quite where to start copying. Now, on the other hand, if you are given
blueprints to build a house, the only limits you have are those of
physical constraints. Endowed with a description, you know where to
start, and you can use all sorts of scaffolding mechanisms to be thrown
away at the end of the construction (therefore not available for
self-inspection reproduction). In other words, with blue-prints, the sky
is the limit for what you can build -- as manhattan attests.
>
> You view open-ended evolution as the ability to produce ANY protein.
> But some proteins may be deadly harmful and kill the organism.
> This is tha same as molecules that can not fold and unfold.
> These molecules will be just less successful in natural selection
> than those that can fold and unfold. In both casees, natural
> selection reduces variability. No difference!
Yeah, some blue-prints will lead to buildings that collapse, but I still
can build more buildings than by self-inspection. The difference between
slef-inspection and descrription based reproduction is truly huge.
Cheers,
Luis
_____________________________________________
Luis Rocha (Postdoctoral Associate)
Computer Research Group (CIC-3)
Los Alamos National Laboratory, Mail Stop P990
Los Alamos, NM 87545
e-mail: rocha@lanl.gov or rocha@santafe.edu
http://www.c3.lanl.gov/~rocha