Re: OOL X - The origin of the RNA world.

> > > From: Tim Tyler <tim@xxxxxxxxxxx>
> > > Lovelock made no such mistake. His theis [sic] was that the planet acted
> > > on a global scale as a self-regulating homeostatic system.

> Robert Maas, see http://tinyurl.com/uh3t <rem642b@xxxxxxxxx> wrote or quoted:
> > Regardless of whether he used the word "organism" or merely used a
> > phrase which is its definition, he said the same thing, which was wrong
> > IMO. [...]

> From: Tim Tyler <tim@xxxxxxxxxxx>
> Got a quote from Lovelock making the supposed mistake you mention?

Nope, nothing direct. I trust what you said his thesis was above, was a
basically accurate statement of it. What you said above is basically
the same as I've read from other sources for many years, so that's why
I trust your summary-words so easily.

> The theory does not suggest the system arose through competition with
> others like it. It suggests that feedback mechanisms are generated
> by living organisms that tend to maintain a homeostatic state.

It is true that in just about any complex ecosystem, there are various
kinds of feedback occurring. What I don't believe is that these
feedback mechanisms are guaranteed to be negative (diminishing
deviation rather than amplifying deviation) and sufficiently stable so
as to yield a nice healthy stasis in the environment. For an example of
feedback that is disasterous, connect a microphone into a high-power
amplifier and speaker system and plug your ears as you turn the gain
all the way up causing it to self-destruct by overloading the
amplifiers with horrible oscillations. For another example, place a
child who has never driven any vehicle behind the wheel of a car going
100 MPH and watch that driver over-steer and crash.

In the absense of group selection of organisms (as contrasted with
group selection of the various innerds of each individual organism),
the only thing a living organism does is protect itself and its genes
and copies of its genes by any means necessary including cooperating
with copies of its genes and cooperating with some others who cooperate
with it. But on the grand scale, beyond such limited cooperation, it's
war, not homeostasis. Creatures have been going extinct all the time
because they lose the war. That's not any homeostatic state. Sometimes
the total population is reduced by an order of magnitude, less often
it's reduced by two orders of magnitude. There is no intrinsic limit
how many orders of magnitude such an extinction event can reduce the
total population. There's no "gaia" putting a limit on the number of
orders of magnitude lost during an extinction event. So-far Earth's
ecosystem has been lucky, no extinction totally wiping out all life. If
such an extinction event would have happened, we wouldn't be here to
discuss the issue. There may be other worlds just like Earth where all
life has been completely wiped out.

> Lovelock's most famous example of this happeneing is his "daisy
> world" model. There organisms acting locally and selfishly produce a
> global system that acts against externally imposed temperature
> changes. It's not natural selection that produces the effect he is

If you construct a mathematical model with lots of entitites, with
totally random cause-effect relationships between them, and run a
simulation, in some cases you'll get a stable system and in some cases
you won't get a stable system. The fact that one (1) such stable system
has occurred in Earth's natural ecosystem, among the gazillions of
various local ecosystems overall, is no surprise at all. What would be
a surprise is if *every* local ecosystem self-destructed. Until
somebody provides a convincing argument that the small daisy system is
a good model for the complete Worldwide ecosystem, whereby the
stability achieved by the daisies imples global stability, the example
has no relevance to the "gaia" claim.

Also, do you have any conclusive evidence that the daisy stability will
last an additional one billion years?

> Lovelock explicitly predicts such disasters.

So his theory predicts anything whatsoever, doesn't un-predict
anything? It sounds as worthless as Creationism and Intelligent Design.

> Make his daisy world *too* hot and the homeostatic system fails - the
> system oscillates for a while - and then a new equilibrium
> establishes itself.

What if the "new equilibrium" equals total anihilation of all life?

> The theory goes that the periods of instability are uncommon - and
> that most time is spent in a state of stasis - so the chances of
> observing stasis are high.

It doesn't comfort me at all. All it takes is one period of
super-instability whereby the photosynthetic system totally collapses,
such as some toxin that cleaves all the photosynthesis enzymes, or that
simply blocks photosynthesis, oxygen levels fall to near zero, and all
eukariotic life on Earth goes extinct.

> Gaia theory is popular among environmentalists in part because it
> predicts the opposite of what you say. It doesn't suggest the
> homeostasis resists all kinds of perturbations. It suggests that
> a sufficiently large knock will shift the system into a new
> equilibrium. Environmentalists love to suggest that if we
> keep messing with the environment, it will keep shifting
> gear - and we might not like the new environments it presents
> us with.

Gaia theory isn't necessary for that sort of warning.

W. Ford Doolittle and Richard Dawkins ,,,
... argued that Lovelock's hypothesis that
the Earth's climate was regulated "by and for the biota" was
teleological, implying impossible foresight and planning on the part of
the biota.''

There is no such thing as "the biota", except as a set of various
replicators with complex relationships between them. What's good for
one part of the biota may be bad for another part. There's virtually no
valid sense of "the biota" as a single point of view regarding
advantage/disadvantage.

Accordingly, D&D are even more correct than they state above.

Even in the case of co-evolving genome parts trapped within a
micro-ecosystem, such as a cell, it's not exactly correct to speak of a
benefit for the cell as a whole. Rather there is benefit or
anti-benefit for each individual part of the genome, with extremely
high correlation between each individual such viewpoint, but not 100%
correlation, hence treating the entire genome as if it were a single
point of view is not exactly correct.

The fundamentally correct way of understanding this is the "selfish
gene", with at least three common mechanisms for cooperation as an ESS
(evolutionarily-stable strategy).

> > Also within a single cell, all parts are equally related to each other,
> > so it's not possible to cooperate with kin at the expense of non-kin,
> > except when invading disease is recognized.

> ...or when you have a polyploid genome ;-)

OK, I'll grant that it's theoretically possible for one gene in a cell
to code for some mechanism that kills off other genes without killing
off that particular gene or any copies of it. But any such gene would
very quickly kill it's host cell and thereby kill itself. So I don't
believe any such gene would be around at almost any time you pick at
random.

> I /think/ you are calling this "group selection" because you are
> regarding the genes as a sort of group.

Yes, you understood me correctly. The fundamentally correct
understanding of the genes within a cell is that they are trapped
together co-evolving over a long time period, where there's a lot of
selection pressure to cooperate, so on the whole they've indeed already
evolved to cooperate. But there's also selection pressure to "cheat"
(to work more for yourself and copies thereof than to work for the
common good of your cellmates) if you can get by with it. So there are
a bunch of genes that code for behaviour in such a way as to promote
copies of that same gene at the (slight) expense of copies of other
genes within the same genome which copies don't happen to be in the
same cell as copies of the self gene. (For example, if A represents
this very selfish gene trying to cheat, and B represents an alternative
allelle for A, and X,Y,Z represent other sets of genes not containing
that locus, then A would code for behaviour to help genomes AX AY and
AZ but not BX BY or BZ, so on the average it promotes A at the slight
expense of X,Y,Z, promoting *all* copies of A it encounters, but only
*some* copies of X Y or Z.)

> You may find the term "gene group selection" useful - this sort of
> group selection is pretty different from the conventional sort.

Yes. Convenional "group selection" treats each entire cell or organism
as a single unit, and defines the term only for groups of more than one
cell or organism which are trapped together in a large unit. But I
apply the term at levels within an individual cell too. (This
application is necessary to analyze the very earliest just-barely-life,
per my OOL scenerio, which are just starting to co-evolve after
previously being independent replicators.) Note that group selection
isn't absolute, so nothing larger than a single genetic unit is exactly
correctly understood as a single evolving unit.

> However, I take your point - this sort of selection *does* produce
> cooperation - and is neither kin selection or reciprocal altruism.

Yes, you understand correctly. Note that none of the three methods of
achieving cooperation as an ESS is perfect. Evolution produces better
and better approximations to pure cooperation, but that final state is
never exactly achieved even with perfection selective pressure toward
that pure-cooperation final state. Also there are difficulties that
sometimes deter selective pressure away from that optimum. For example,
individual parts of the co-evolving genome might sometimes leave the
cell then later re-enter, and while they are away they would evolve
away from cooperation. (And of course every time a new bit enters the
genome, it dillutes the nearly-perfect cooperation already achieved by
the old residents. Kin selection likewise can be disrupted by
impersonation, such as Cuckoo laying eggs in other birds nests. And
tit-for-tat game-theoretic strategy is disrupted by difficulty
correctly recognizing specific past opponents and remembering how they
behaved previously. Summary: Each of the three methods achieves, in
many circumstances, a pretty good approximation to full cooperation,
but not exactly 100% cooperative behaviour.

> The hypothetical organism under discussion would *originally* [be] the
> product of natural selection. It would be our final descentant.

Yes, that's why I listed that as one possible exception, the last
member of a species that evolved for a long time but is now on the
verge of extinction. It's no longer competing with alternate
individuals of the same species, but it did in the past, hence it *now*
is highly cooperative internally due to *past* group selection.

In the case of "gaia", there *never* was a time when Earth's ecosystem
replicated and the different mutated copies competed against each
other, so there's no reason to believe Earth's current ecocystem is
more cooperative than any old random ecosystem of similar complexity,
i.e. the current level of global cooperation is a random element from a
sample space, not biassed toward cooperation compared to the full
distribution of possibilities, except insofar as the weak anthropic
principle applies (that portion of sample space where all life has gone
extinct already, is eliminated from the possibilities to consider,
because we are alive to observe this system from the inside). Hmm, new
idea just now: Perhaps the ecosystem on Mars was of similar complexity
to Earth's, but it suffered a 100% extinction instability. Since we're
looking at Mars from the outside, our point of view doesn't depend on
Mars's ecosystem still being alive, our point of view isn't biassed by
the weak anthropic principle, we see Mars as an unbiassed sample
(except insofar as Mars is in a solar system where some *other* planet
still has a living ecosystem, but that's much less of a bias).

> > But short of a supernatural being or a scientist space alien
> > constructing such an organism, or that organism evolving by natural
> > selection, there's no way such an organism could ever come into being.

> The organism's ancestors would have evolved via natural selection.

I was referring specifically to very early life just getting started
from abiogenesis, before any trapped micro-ecosystem could co-evolve to
achieve cooperation within the cell then join into colonies and achieve
cooperation within the colony then evolve more structures to achieve
sufficient brain power to understand its own nature and devise science
and technology. Humans, having already gone through all that evolution
to achieve s&t, can someday in the near future start to deliberately
guide our own evolution. But simple chemical replicators in the time
just after abiogenesis have nowhere near the capabilities to invent
science and "decide to cooperate" as somebody earlier in this thread
seemed to have claimed, which I was arguing against.

.

Relevant Pages

  • Re: OOL X - The origin of the RNA world.
    ... >> which doesn't reproduce, and instead maintains itself. ... > Accordingly it is best not to call it an "organism" in the first place. ... > level of cooperation needed to make it an actual organism. ... > group selection to encourage cooperation. ...
    (sci.bio.evolution)
  • Re: Rapid acceleration in human evolution described
    ... result of reduction of natural selection. ... critical mechanism for this type of cooperation was _cognitive_ mutualised ... cognitive_forms of trading dominate most other species. ... Trade was so stunningly successful that we as a species were able to ...
    (sci.bio.evolution)