Re: ecological complexity

John J. Kineman (jjk@NGDC.NOAA.GOV)
Thu, 3 Sep 1998 11:16:33 -0600


Kineman comments on Vaneechoutte:

At 03:52 PM 9/2/98 +0200, you wrote:
>Alexei Sharov wrote:
>Mario Vaneechoutte wrote:
>
>> >Could you explain your agreement in more detail? To be honest, as a
biologist
> I
>> >agree with Dawkins, Gould, and most other biologists that natural
selection
> is
>> >nothing more than a blind seive, without teleology, and the above
sounds like
>> >blasphemy or scientific nonsense! And what do you mean by free will? It
>> does not
>> >really exist.I would like to understand what good reasons you have to
agree
> on
>> >the opposite of the current paradigm.
>>
>> Mario, you can find the discussion of this in Jesper Hoffmeyer's
>> book "Signs of Meaning in the Universe" 1996.
>
>I have read Emmeche & Hoffmeyer 1991 (which is online: Emmeche, C., & J.
> Hoffmeyer.
>1991. From language to nature. The semiotic metaphor in biology. Semiotica
84:
>1-42.http://alf.nob.dk/~emmeche/cePubl/91a.frolan.htm
>) and find it most interesting. I dislike books, since it always takes months
> (and
>money!) before I can get them. Don't you have an online summary of this. I'd
> love to
>read it.
>
>
>> In short, there is no "blind seive". Any seive is an instrument
>> and somebody makes use of it, otherwise it is not a seive. What
>> we call "natural selection" (a very confusing term!) is a mechanism
>> by which the population probes the environment and selects which
>> organisms should reproduce and which should not.
>
>Well, this puts thing on their heads! The population does not probe anything,
> I'd
>say.When you have variation on a theme, then none, one, some or all of the
>variations will be able to exist in a certain environment (which contains the
> other
>variations as well of course). Whether we call this a seive or not, it is
what
>selection is about in general (natural or not).

The problem with this view is the presumed independence between the
environment and what it selects. We think of them as independent for the
purposes of constructing Darwin's mechanism, but we realize they are not
entirely independent. What is the nature of the dependence and does it
modify the mechanism? If the dependence is complex, as in these
discussions, then one should expect some modification of the evolutionary
result.

> I find it truly confusing and
> even
>erroneous to say that there is a population which probes the environment and
> which
>selects. Selection is something which is the result of having certain
> environmental
>conditions and variations on a theme with possibly different 'viability'
in that
>environment.

This is the traditional (modern-synthesis) view, which I consider
"first-order" process. On top of this, however, is the more subtle
possibility of complex interactions between phenotype (including
populations) and environment. In this complex sense, the organism does
imput to the environment. The question is if these inputs can be ignored
(because we decide they are all mechanically derived and thus themselves
the product of the first-order process) or if they must be considered
because they are in some partly independent of the mechanically derived
ontology, i.e., have any intelligent freedom, however that might be defined
or explained.

>There is no 'which SHOULD reproduce or not'. Those that could reproduce are
> still
>here, that's all.

We always begin be wishing for such simplicity, then, fortunately, gain
greater job security because it doesn't turn out that way. You may come to
appreciate this in old age. If we are to think that organisms are capable
of any innovative inputs to the system at all (the basic question), then,
at the phenotype level, their input is surely associated with goals and
intentions (semantics?) -- that translates to "should."

>
>> Unfortunately,
>> many biologists like Gould can think only of individual adaptations
>> and dismiss ideas of adaptations at higher hierarchical levels
>> (e.g., populations). In order to make use of "natural selection", a
>> population must "remember" which organisms performed best; thus, it
>> needs inheritance which is population's memory. Also, it needs some
>> sort of a search engine which generates variability. This engine
>> includes mutations, recombinations. Epigenetic constraints are used
>> to prevent blind repeating of previous mistakes. All these mechanisms
>> are adaptations at the population level which were selected at a
>> higher level of "natural selection". There is a hierarchy of
>> "natural selections" that correspond to different time scales.
>> Michael Conrad wrote a book "Adaptability" in which he explains this
>> hierarchy of natural selections.
>
>Possible, all kind of books have been written. But this is not the
standard view
> of
>biology and there are good reasons. (See Evolution by M. Ridley as cited
in the
>language article (see below): Group selection can occur, but is limited and
> mostly
>to weak to overwhelm natural selection.

Yes, this is true, but because "group selection" has been considered only
within a mechanical model. The 2nd order processes I suggest are at work
would increase the effect of group selection, as it clearly does in the
human case. The transition that is needed in biology is to break this
mechanical barrier to thought.

>I touched this problem in the article on
>language as cited in the signature below, and give an example of where group
>selection is possible. Gould is probably the last one to blame for being
>adaptationist, since others like Dawkins blame him - among other things -
to be
> too
>group selectionist and to be not enough adaptationist.
>

They are both determinist, and that's the failing.

>>
>>
>> However, it is impossible to consider natural selection at all
>> infinite levels. Thus, for practical purposes, we cut it at some
>> level and call everything that is above "physics". For example, we
>> can consider selection of individuals without thinking of adaptations
>> at the population level. This is like an 8-digit calculator that cuts
>> off extra information. But people like Gould think that the 8-th digit
>> is the last one and there is nothing beyond.
>
>I would be the first to say that natural selection has limited applicability.
> E.g.
>I' d say that the origin of life cannot be explained by NS, while the
standard
> view
>is that of the RNA-world whereby 'self replicating' molecules undergo natural
>selection, because of differential reproduction rate (which by the way is not
> the
>same as differential survival rate: survival is unimportant in biology,
> reproduction
>is). Also selection comes only after variation has been 'created', through
> mutation
>and recombination (through symbiosis, sexual (bilinear) recombination or
> cultural
>(multilinear) recombination). A lot of evolution can be explained through the
>absence of selective constraints, e.g. when a new building plan comes into
being
>like was the case for animal multicellularity: at that moment any
variation goes
>because there is nothing like it and the niche is empty. This empty niche,
> always
>open for more complex organisms may explain why evolution almost inevitably
> leads to
>more complexity: there is always room for more complex organisation.
Evolution
> is
>open ended towards more complexity, (more complex organisms initially have no
>competitors).

Yes, this is Gould's argument as well in "Full House." But I don't buy it
as a "nothing but" argument. First, and here I'm a little unsure of myself,
in a deterministic view are not the number of comlex states still finite?
If not, then are we not subtly introducing freedom at the universal level
while denying its possibility at smaller scales?

Second, natural selection involves more than competition. In fact
competition is today taking a real beating as a selective factor. Why
should more complex organisms be adaptive in the long term even granting
that they have no initial direct competitors? Gould's argument is that it
is merely a statistical liklihood: there are more complex states than
simple ones, hence, since we're just starting out in evolution, we see
"progress" toward this open-endedness. Whenever evolution is "reset" it
starts over with simple forms that can't get simpler, so they get more
complex. Its a "nothing but statistics" arguement which he compares to
baseball scores. But do we observe regression to simpler states from more
complex states as an equal liklihood, as this theory would require
(correcting for the numbers - i.e., looking on a case-by-case basis)? I
think not. There are examples of a quasi staisis - i.e., sharks remaining
essentially unchanges for millions of years, etc. but they didn't evolve
back to single celled organisms because life was easier that way and it
would maximize their numbers. The magnification of complexity seems to
confer selective advantages that are so strong it tolerates fewer numbers
(i.e., lower fitness?). What's driving this?

>
>>
>>
>> The term "freedom" has a lot of connotation. But for scientific
>> purposes we can view freedom as making use of variability. Obviously,
>> it is outside of physics because physics does not study usefulness.
>> But biology studies adaptation which is the same as usefulness. Thus,
>> freedom can be studied within biological sciences. The most simple
>> example of freedom is mutation.
>
>I'd hesitate to call mutation freedom, since it is deterministic, but
still then
>this has a different connotation than free will.
>

I agree. Mutation, if it is associated with "freedom" is associated at the
most fundamental level of freedom, the quantum level (if one is willing to
think about levels). I think this is also freedom, but the most elemental
form of it, hardly something we could find analogous with the "free-will"
concept we are used to in the human case. Still, it is possible this is a
continuum.

>>
>>
>> >Why should we need QM all of a sudden to explain consciousness, while we
> could
>> > do
>> >without to explain the rest of evolution? This is a bit like pretending
>> that, to
>> >explain human existence, you need some kind of deus ex machina solution.
>>
>> Here I agree with you completely! There is no need for QM to explain
>> variability. In most cases we simply don't need to explain variability,
>> we simply postulate it (the same 8-digit calculator!). I don't mean to
>> forbid using QM for explaining evolution or free will. I just say that
>> we can understand a lot about evolution and free will even without QM.
>> May be eventually QM or some other physical theory will be able to add
>> some details, but now we are not talking about details.
>>
>> -Alexei
>> -------------------------------------------------
>> Alexei Sharov Research Scientist
>
>Best regards
>--
>Mario Vaneechoutte
>Department Clinical Chemistry, Microbiology & Immunology
>University Hospital
>De Pintelaan 185
>9000 GENT
>Belgium
>Phone: +32 9 240 36 92
>Fax: +32 9 240 36 59
>E-mail: Mario.Vaneechoutte@rug.ac.be
>
>J. Memetics - Evolutionary Models of Information Transmission:
>http://jom-emit.cfpm.org/
>
>The memetic origin of language: humans as musical primates
>http://jom-emit.cfpm.org/1998/vol2/vaneechoutte_m&skoyles_jr.html
>

----------------------------------------------
Appreciatively,

John J. Kineman
Bear Mountain Institute
1101 Bison Dr.
Boulder, CO 80302
BMI@bayside.net
http://www.bayside.net/NPO/BMI