perplexed in peoria

I am inserting here a portion of Rice's essay that
John clipped. It is important, though John probably
did not realize this.

[Rice writes:]
I suggest that a useful test of whether selection is
acting at some level is to ask the following question:

In order for selection to act, is it necessary and sufficient
for there to be multiple (>1) units at that level and for
there to be variation between these?

If the answer to this question is "yes", then we say that
selection is acting at the level under consideration.

We can see how this test works by applying it to the case
of kin selection. Consider a population composed of a single
family of full siblings. Now imagine that some of these
offspring carry an allele for altruism and others do not.
In this case there are multiple individuals and variation
in the trait (altruism) that influences fitness. Individuals
are also related to one another and interact. Thus, if kin
selection is a case of selection acting on individuals or
alleles, it should happen here.

Since r = 0.5 for every pair of siblings in this scenario,
[here, John begins his quotation from Rice with the remainder of this sentence]


I, and I think most other defenders of Hamilton in this
group, agree that in Rice's scenario, altruism will
decrease. We agree, John, and we have always agreed.

However, Rice is making two mistakes. The first and
obvious one is that he left the word "heritable" out
as a modifier of "variation" in his proposed criterion.

The second is less obvious, and may be debatable. In
fact, I have just completed a debate with Dr. Hoelzer
on this subject:

Rice, I claim, is wrong to state that r = 0.5 for every
pair of siblings in this scenario. Actually, in this
highly artificial scenario, r = 0. Hamilton's rule, if
properly understood, does NOT predict that altruism will
increase in this scenario.

What Rice has done is similar to what Hoelzer did. He
has constructed a scenario in which IBD is no longer a
good estimator of r.

For details, refer to my debate with Hoelzer on the thread
"Kin Selection Contradiction?"

Hamilton's model is sometimes described as involving gene
level selection, sometimes as involving individual level
selection, and sometimes as involving group level selection.
Rice's conclusion from his toy model seems to be that group
selection is the best characterization of Hamilton. The
odd thing is that Rice is a proponent of group selection,
and of the kin selection model, yet Edser welcomes his
analysis because Edser believes in neither group selection
nor kin selection.

To repeat my own view on this, I think that all three
viewpoints are defensible, but that the individual
selection viewpoint is best. Calling it gene level
selection simply invites the standard objection to gene
level models - namely epistasis. It turns out that
epistasis isn't troublesome to Hamilton, but this takes
some effort to see. Similarly, calling it group level
selection invites various standard objections too. These
can be answered in the Hamilton case, but only with effort.

This is not to say that the individual level viewpoint is
without conceptual problems of its own. Hamilton's model
turns causality on its head in that the success of the
altruistic gene derives not from what it causes in the
donor (the altruism), but rather from what it is merely
correlated with in the donor (his status as a disproportunate
RECIPIENT of altruism). And this is a subtlety that some
people have trouble apprehending, let alone comprehending.

guy hoelzer

in article dj91vg$1tkn$1@darwin.ediacara.org, Perplexed in Peoria at


First, I am flattered by this comparison with Rice. Sean and I were in
graduate school together and I consider him to have an unusually great
intellect.

Second, I want to point out that I made this point explicit in my argument.
It was not a mistake, an oversight, or a deception on my part. I also claim
that the "scenario" I proposed describing the changing relationship between
IBD and r is implicitly true in your (everybody's, really) argument
regarding kin selection. The "scenario" was simply that the frequency of
the allele was changing. I simply addressed the issue of whether the change
in frequency affected the relationship between IBD and r (I assume we are
using McGinn's version of "r" here, in which "r" represents the probability
of sharing an allele). I have yet to see any counter argument, as opposed
to unjustified assertions, that impinges on my claim.


I agree with this point as a practical matter. It is related to the reason
I often claim that empirical studies are rarely able to disentangle kin
selection from reciprocal altruism as explanations for altruism, despite
frequent claims of support for kin selection while ignoring the potential
for reciprocal altruism. However, I don't see where the receiving of
altruism by altruism-donors plays into Hamilton's model.

Guy

an588

Guy Hoelzer (hoelzer@unr.edu) writes:

?????

But isn't that the whole point of Hamilton's rule?

Are you using very different definitions of some
of the terms from the ones I would think you would
be using?
--
Cathy Woodgold
http://www.ncf.ca/~an588/par_home.html
We are all Iraqis now.

perplexed in peoria

Technically, Guy is correct. Receipt of altruism by *donors* is not
absolutely necessary in Hamilton's model.

What IS necessary - the reason the model works - is the receipt of
altruism by individuals who *carry the gene(s)* for altruism.

Of course, it is difficult to imagine any scenarios in which kin
selected altruism is common in which the donors DON'T receive
more altruism than do non-donors.

perplexed in peoria

How did "McGinn's version of r" suddenly show up in this conversation?
McGinn's version is not an 'r' at all!

My understanding (you do realize, don't you, that we are now continuing
a discussion that is 15 months old?) is that the only two versions of
'r' under discussion were the regression (1970) version and the IBD (1964)
version. The regression version is tautologically correct (as proved,
for example, by NAS) and the IBD version is a pretty good approximation.

My claim is that IBD is generally a good approximation at any frequency of
the focal allele. Your claim (as I understand it) is that it is not
a good approximation at some frequencies and your example in support of
this claim was that when there is exactly one copy of the allele in the
population, then none of the relatives of the unique altruist will
bear the gene, hence the allele will not spread under kin selection.

I consider this argument to be irrelevant. You are not providing an
example of the failure of IBD at low allele frequencies. You are instead
giving an example of the failure at low population size. Simply take
whatever frequency you are talking about in your example, and put that
same frequency into a population ten times as large. Now there are
several copies of the allele in the population, and they probably
appear in just a few families - each of which is likely to have a
higher inclusive fitness than does the general population.

Felsenstein, Woodgold, NAS, and I have given many such arguments over the
past few weeks. I gave such an argument 15 months ago. Do you really
mean to say that those were all unjustified assertions?

However, if you are claiming that 'McGinn's version of r' does not
provide a good approximation to Hamilton's r at some frequencies, then
no one is going to disagree. As far as I can tell, McGinn's version
of r is a horrible approximation at ANY frequency.

tim tyler

I would say that Hamilton's rule /usually/ concerned itself with
the receiving of altruism by altruism-recipients.
--
__________
|im |yler http://timtyler.org/ tim@tt1lock.org Remove lock to reply.

guy hoelzer

in article djf5gi$1g0e$1@darwin.ediacara.org, Catherine Woodgold at


I think you are confounding Hamilton's kin selection model with Trivers'
reciprocal altruism model. Let me expound on my position. Hamilton's model
does not require that an altruist receive altruism for the altruism allele
to increase in frequency. The model accounts for all of the necessary costs
and benefits relating to even a single act of altruism. Therefore, it is
not the expectation that altruists will receive altruism from relatives
carrying the allele that makes the model work. This factor, which could
certainly further facilitate an increase in the altruism allele through the
mechanism of reciprocal altruism, is also not accounted for in the precise
prediction of the tipping point predicted by Hamilton's Rule.

This is the point I have trying to make in previous posts. IMHO, kin
selection is very often confounded with the notion of reciprocal altruism.
I think you have done this in concept, while I think that empiricists most
often do this more often out of uncritical cultural tradition. [I just know
that I stepped in it this time. :-(]

Cheers,

Guy

an588

Tim Tyler (tim@tt1lock.org) writes:

LOL!
--
Cathy Woodgold
http://www.ncf.ca/~an588/par_home.html
We are all Iraqis now.

an588

"Perplexed in Peoria" (jimmenegay@sbcglobal.net) writes:

Good point. I was a little off in my description
of what the "whole point" of Hamilton's rule was.
--
Cathy Woodgold
http://www.ncf.ca/~an588/par_home.html
We are all Iraqis now.

an588

"Perplexed in Peoria" (jimmenegay@sbcglobal.net) writes:

Assuming Guy's claim to be as you say, I would describe
the fallacy in it differently. I would say the fallacy
is in assuming a specific scenario in which the allele
experiences what could be called, by the teleologically
inclined, "bad luck". One could contrast it with a
situation in which all the recipients of the altruistic
acts just happened to have copies of the gene.
One could also point out that, even if there is only
one copy of the gene in one generation, in the next
generation there might happen to be two or more copies
and spreading by kin selection could still then occur.
--
Cathy Woodgold
http://www.ncf.ca/~an588/par_home.html
We are all Iraqis now.