Continuation of "group selection" and "emergence" discussion

From: Eric Hawthorne <egh.domain.name.hidden>
Date: Wed, 11 Feb 2004 01:52:52 -0800

Tianran Chen wrote (in private reply to my earlier post, but I thought
this discussion
generally interesting, hope that's ok Tianran):

>i do agree that many very valuable point of view had been
>criticised unfairly due to their 'group selection' nature.
>however, i am quite convinced that there are fundamental
>problems embedded in it.
>
>first of all, i think to look at evolution through a
>individual or group point of view is always dangerous. since
>they are not the fundamental unit of evolution. evolution,
>(in the sense which it is referred to in most biology
>context) DOES NOT manipulate individual NOR species directly,
>instead, it is always the small packets of genetic
>information (such as gene or meme) that is being
>manipulated. so, it is a safe theory about evolution should
>always able to be translated into languages in terms of gene
>or meme, and if a theory cannot, then it is not safe to use.
>and this is exactly the problem of 'group selection'.
>
>
But just because it is the genes are being manipulated DIRECTLY does not
mean that
other factors are not important to understanding what's going on, and
understanding what
factors may be most ESSENTIALLY driving a particular evolutionary
direction.
For example, biology now understands that it is not single genes, but
sets of genes acting together in
regulatory networks, that are the fundamental units of "functionality"
and therefore, of
"adaptive or maladaptive functionality" in organisms.

>
>
>>BECAUSE THE EVOLVABLE "GOAL" IS NOT SIMPLY TO MAXIMIZE THE
>>CHANCE OF SURVIVAL OF AN ORGANISM OF THE NEXT SHORT-TERM ENCOUNTER.
>>THE "GOAL" IS TO MAXIMIZE THE PROBABILITY OF SURVIVAL OF THE SUM TOTAL
>>OF ALL OF THE ORGANISM'S ENCOUNTERS UP TO WHEN THE ORGANISM REPRODUCES.
>>
>>
>
>disagree. a gene's 'goal' is to maximize the availability of
>its own copies in the entire gene pool. so to look at it in
>individual level, it implies that an individual is more
>likely to behave in such a way that it tend to maximize the
>chance of some gene to replicate. and 'some gene' here,
>refered to not only genes in its own body, but also in
>other's body. one thing to notice here is that very often,
>individual try to do so at the cost of its own chance for
>breed. such behavior can be found commonly in social
>animals, symbiosis systems, and etc.
>
>again, here mention about the 'goal' of a gene. but what i
>really mean is that due the the selection pressure, genes
>who had survived selections behave in the way as if they had
>the 'goal' although they are really blind about future. so
>the 'goal' is simply a short hand notation, do not take it
>literally.
>
>
You'll note that I was the one who started the practice of putting
"GOAL" in quotes, indicating
that it is not to be taken literally, but as a stand-in or short-hand
for a complex set of factors
that lead to a tendency of evolution to support one kind of trait over
another.

>valid theories has to be general enough to explain all sorts
>of things in the domain, not just part of them. and now
>better theories does exist (such as self-gene, and memic
>evolution). so 'group selection' SHOULD be marked as
>obselete.
>
>

I disagree with your last clause.
To me, a fan of general theories of emergent complex ordered systems,
of which life evolution is only one example, one of the most fundamental
questions
is what is the best scope-boundary that best defines what is the most
"interesting" or
"systematic" or "robust" system. What I mean by this is that we have a
degree of free reign
about what elements (of the world, universe, what have you) that we
choose to include in
our definition of a system. Or in other words, for ANY particular set of
elements that have something
to do with each other, some crazy guy will have the right to call THAT
collection a perfectly
valid system; his most important system perhaps. So you can imagine
possible "system-scopes"
or "system-boundaries" as being an infinitely variable set of concentric
spheroids overlapping,
Venn-diagram-like, being at wide ranges of spatiotemporal scales, and
including/excluding
different elements.

Faced with such a scenario, one is forced to ask "are there any
universal principles that would
let me decide which "boundary-spheroid" is the most "systematic" or
consideration-worthy system
(at this spatiotemporal scale in this vicinity, anyway)?

Or, both generalizing and specializing a bit; given that, for example
natural systems tend to be
"fractally functional"; that is, comprised of nested layers of
smaller-scale functional systems, we can
ask "how (at what scale boundaries) do we best divide this natural
system up into nested layers (i.e. where
if anywhere are the best-defined layer boundaries (those layer
boundaries that are the best at separating of
distinct functionalities) , and for EACH spatio-temporal-scale layer of
this natural system, what are the
best-defined system-scope spheroids? What are the system-scope spheroids
that perhaps are the best
clusters of negative entropy at that spatiotemporal scale?

Now to me, the evolutionary issue is to figure out how and why
"fractally functional" natural systems evolve
their functionality at all of their spatio-temporal scale layers (and
thus how they maximize, AT EACH
LAYER-SCALE (and fractal-functionality system level) their
pattern robustness and pattern-longevity, compared to competitor patterns.

The direct variations (mutations, recombinations) in genes are an
important part of this picture, to be
sure, but analyzing the whole evolution-trend question in terms of
individual mutations of individual genes
is akin to saying that "it's all the quantum wave equation, really"
about all of the phenomena in physics, large
and small. It's true, but it's also fairly trivially true. Every theory
has a scope of application, and there is room
for other theories at different levels to explain aggregate phenomena
that are not directly or concisely
explainable by the purely reductionist "atomic" theory.

Many of the "aggregate-level" theories will be wrong, and/or hard to
demonstrate, but that does not say
that all of them are wrong. Just that the right ones will be hard to
tease out.

Cheers,
Eric
Received on Wed Feb 11 2004 - 04:55:16 PST