Re: Group selected altruism - (was: Hamilton's rule)
- From: "Perplexed in Peoria" <jimmenegay@xxxxxxxxxxxxx>
- Date: Tue, 22 Nov 2005 23:25:47 -0500 (EST)
"Jim McGinn" <jimmcginn@xxxxxxxxx> wrote in message news:dlvnf0$mco$1@xxxxxxxxxxxxxxxxxxxxxx
>
> Perplexed in Peoria wrote:
> > "Catherine Woodgold" <an588@xxxxxxxxxxxxxxxxxxx> wrote in message news:dlotn3$h0s$1@xxxxxxxxxxxxxxxxxxxxxx
> > >
> > > Someone in some post in the last couple of weeks said that
> > > genes for altruism towards distant relatives don't tend to
> > > propagate. I disagree with this.
> > >
> > > Suppose the following scenario: a species is organized
> > > into large groups such that each group almost always breeds within
> > > itself, not interbreeding much with other groups. Suppose
> > > the groups are distinguished by visible phenological differences
> > > but are not restricted to geographical areas. Suppose some
> > > groups have genes that promote altruism towards individuals
> > > within their group. If the benefit exceeds the cost, then
> > > the genes will promote prosperity of the group, which will
> > > tend to take up a larger geographical space. At the same
> > > time, altruism genes within the group will tend to diminish
> > > in frequency; the rates of these two opposing processes
> > > depend on the ratio of cost to benefit. If the benefit is
> > > large enough, the process of altruistic groups growing and
> > > splitting into smaller groups and taking over territory or
> > > habitat from non-altruistic groups will support the altruism
> > > genes faster than those genes diminish in frequency within
> > > each group.
> > >
> >
> > You have given the standard argument for group selection as the
> > mechanism for the spread of altruism. Hence the change in the
> > title of the thread.
> >
> > Yes, by careful 'tuning' of the parameters, you can set up
> > starting conditions such that altruism increases in the population
> > as a whole under the group selection mechanism. However, the
> > same selective dynamic that spreads the altruism also tends to
> > 'detune' the parameters.
>
> How so?
I said how so in the following paragraphs.
> > So, it is very likely that the increase
> > in altruism will be short lived.
> >
> > You pointed out the reason for this yourself: although altruistic
> > groups grow faster than nonaltruistic ones, the frequency of
> > altruism within the altruistic groups must decline. So a very
> > altruistic group grows very fast, and then splits into two
> > moderately altruistic groups which grows moderately fast and
> > then split into four barely altruistic groups which barely grow
> > but eventually split into eight groups with no particular
> > altruistic tendencies.
> >
> > Most group selection scenarios have this same problem. In order
> > for some groups to succeed and others to fail, there have to
> > be differences between groups. But the dynamic of success always
> > seems to have the effect of diminishing the differences between
> > groups.
>
> How so?
One example was given above. Altruism tends to decrease in frequency
in all groups, and it decreases most rapidly in the groups with the
most altruists.
Another example comes from the dynamics of group creation and destruction.
Groups with low frequencies of altruism tend to go extinct. Groups with
high frequency of altruism tend to grow and split. Hence the variation
among groups tends to disappear.
> Group selection rather rapidly 'runs out of gas'.
> > (Incidentally, individual selection avoids this problem because
> > mutation is continually adding 'fuel'.)
> >
> > (Also, you postulated that there are visible differences between
> > groups - presumably you are also postulating that a new visible
> > difference arises every time a group splits! But lets ignore
> > that problem.)
> >
> > Maybe we can split the group asymetrically so that the most altruistic
> > members form one child group and the least altruistic members form
> > the other child group. If you can keep this dynamic going through
> > many group generations, it is possible that at least some later
> > generation groups will be just as rich in altruism as the richest
> > ancestral group. But then what is the reason that causes groups
> > to split in this odd way?
> >
> > If the group splits into kinship clans, that could produce an
> > asymmetry, but then we are back to a kind of kin selection -
> > altruism mostly goes to fairly close relatives.
> >
> > If the groups are small enough, you may get sufficient asymmetry
> > due to sampling error, but now the groups may be too small to
> > do within-group breeding without significant inbreeding.
> >
> > Personally, I think that the only group selection mechanism that
> > can work (for altruism or anything else) is some variety of
> > D.S. Wilson's 'trait group selection'. But a lot of people don't
> > consider this as 'true group selection' since the groups don't
> > live any longer than individuals. In the Wilson model, organisms
> > spend a crucial portion of their life-cycle in groups, but the
> > groups break up before the organisms breed and then reform somewhat
> > randomly in the next generation. Such groups can be small enough
> > so that sampling error at group-formation time provides enough
> > intergroup variance so that you get an ongoing process.
>
> All that is necessary for group selection is that there situational
> factors that prevent or reduce gene flow between groups, as Catherine
> described, and situational factors that cause differential
> survival/reproduction between the groups, as Catherine also described.
That is not even close to "all that is necessary". For one thing, you
have to have competition between groups. If you have a bunch of islands,
with one group per island, then you may have the reduced gene flow, and
even the differential survival/reproduction of groups. But you don't
have group selection unless you have a scheme whereby the different groups
compete. One way of doing this is to have the successful groups colonize
the islands that are left vacant by the extinction of unsuccessful groups.
But you have to do some 'fine tuning' of the parameters if you want
unsuccessful groups to go extinct often enough to make up for the fact
that on every island, the altruistic individuals are less successful than
the non-altruistic ones.
I would claim that the 'reduced gene flow' actually works against you
(even though it is usually assumed as an ingredient in classical group
selection). The usual argument is that the reduced gene flow keeps
the altruistic group from being diluted by non-altruistic migrants. But
this argument looks at only part of the problem. Reduced gene flow also
prevents the successful group from exporting the ingredients of its
success to the population as a whole. A process like the 'trait group'
model seems to work better because everyone is a migrant and new groups
are formed each generation.
.
- References:
- Re: Group selected altruism - (was: Hamilton's rule)
- From: Jim McGinn
- Re: Group selected altruism - (was: Hamilton's rule)
- Prev by Date: Re: The uncorrected simplifications/oversimplifications of
- Next by Date: Re: Life is more stable than inorganic chemicals
- Previous by thread: Re: Group selected altruism - (was: Hamilton's rule)
- Next by thread: Re: Group selected altruism - (was: Hamilton's rule)
- Index(es):
Relevant Pages
|
