Re: Levels of selection (Was: Can cognition override natural

in article [EMAIL PROTECTED], Tim Tyler at [EMAIL PROTECTED]
wrote on 11/12/03 8:43 AM:

> Guy Hoelzer <[EMAIL PROTECTED]> wrote or quoted:
> 
>> [I agree] with those who argue that the primary level at which selection
>> occurs is marching its way up the levels of biological
>> organization.  In the beginning it was the gene, then the gene network,
>> then the cell, then the multicellular organism, then groups of
>> multicellular organisms, ...
>> 
>> This view contrasts with the Dawkins view that selection has stayed mainly
>> at the level of the gene, and the rest represents macroscopic constructions
>> of genes that do not take over as levels of selection.  IMHO, the structures
>> of meiosis and mitosis as genomic processes alone provides overwhelming
>> evidence against the Dawkins view, because they generally remove fitness
>> differences among genes at the stage of gene reproduction.
> 
> I'm sure Dawkins would not agree.
> 
> You don't need high level-selection to account for the low frequency
> of segregation distorters that are not suppressed.
> 
> Such genes typically work by destroying half the gametes - or by
> other destructive means.
> 
> I'm sure the species takes a dim view of this - but the other nuclear
> genes are not very keen either - since their fitness is also reduced.
> They make immediate gains if they can supress the segregation distorter.

I appreciate this argument.  It is not new to me, and I understand why this
would seem to be a sufficient explanation when someone is generally
uncomfortable with explanations positing causation from the top-down.
However, I think you will agree that having a gene-level explanation in hand
does not render alternative explanations less valid.  Nevertheless, I agree
with you and Dawkins that meiotic drive alleles proliferate due to selection
at the gene level, and that suppressor alleles at other loci can also
proliferate due to selection at the gene level.  I do not see this as
inconsistent with the multilevel selection model I have been advocating.
The argument that the primary level of selection has been marching up in
scale does not imply that all opportunities for selection at levels other
than the primary one don't occur.  I think the evidence and the logic
clearly indicate otherwise, and meiotic drive is my favorite example of
selection manifesting at the level of the allele.  I am not sure it is
technically right to say this is selection at the level of the gene, because
it is not a matter of which genes will proliferate or be lost from the
population of genes; but this is probably unhelpful technical haggling in
this context.

> That meiosis is fair is interesting - but it hardly seems like evidence
> for higher levels of selection.

I agree that it is not the strongest kind of evidence for selection at
higher levels, but that is not what I said about it.  My point was that it
is evidence of the lack of opportunity for selection at the level of the
gene (or allele).  It is also very consistent with the notion that selection
at higher levels tends to result in the evolution of mechanisms suppressing
opportunity for selection at lower levels.  This is expected because
conflict at lower levels of organization tends to constrain coherent
function at higher levels of organization.

> I suggest that the result of powerful high-level selection would be
> no SD genes at all.

I wouldn't expect this.  It is equivalent to saying that you would expect
the advent of policing in human societies to completely eliminate crime.  It
can have a huge effect reducing crime, but crime will never go away
altogether.  In fact, the evolutionary origin of new mechanisms causing
lower level agents to coordinate their activities to promote function of the
whole (e.g., policing or meiosis/mitosis) often creates new opportunities
for selfish gain by the lower level agents.  It like looking for loopholes
in a new law, or new niches in a recently reorganized community (ecosystem,
if you prefer).  It might be just too expensive to develop and maintain
perfect control from the top-down to make it worthwhile.  Very effective,
but not perfectly effective, mechanisms might be the most viable solutions
from the top-down view, simultaneously leaving some opportunities for
selection at all lower levels and generating new opportunities for selection
at higher levels.  You might ask what is to stop this upward trend in the
primary level of selection, and I would say that it stops when the
metasystem (e.g., the bioshpere as a whole) has maximized its potential for
capturing, processing, and releasing degraded energy back to the external
universe.  In other words, the complexity of the biosphere is constrained by
the rate at which energy is delivered to metasystem from its surroundings.

> The fact that nature is full of SD genes held under house arrest by
> other genes indicates to me that high-level selection is failing to
> do its job properly - and that the selfish genes are largely
> having their wicked way.

Yes they are;  but the effectiveness of higher-level selection in
suppressing this sort of activity should be measured as the difference in
the extent of this activity in the presence vs. absence of higher level
selection.  How much selfish, meiotic drive-like activity would occur in the
absence of higher level selection.  This is a tough question to ask
empirically, but we can ask how much of this activity we would have expected
to see before higher level systems emerged (e.g., cells, multicellular
organisms, ...).  I expect that selfish gene activity would have been the
norm back then, and has become the exception as a result of the emergence of
higher level biological systems.

Cheers,

Guy