Re: Logic of kin selection
From: Perplexed in Peoria (jimmenegay_at_sbcglobal.net)
Date: 02/15/05
- Next message: Tim Tyler: "Re: Haldane's Dilemma"
- Previous message: Perplexed in Peoria: "Re: Haldane's Dilemma"
- In reply to: John Edser: "Re: Logic of kin selection"
- Next in thread: John Edser: "Re: Logic of kin selection"
- Messages sorted by: [ date ] [ thread ]
Date: Tue, 15 Feb 2005 18:23:35 -0500 (EST)
"John Edser" <edser@tpg.com.au> wrote in message news:curull$22nv$1@darwin.ediacara.org...
> The first question that has to settled before
> anything else that anybody calculates can make
> any rational sense is how _exactly_ can altruism
> be defined as scientifically _meaningful_? If Neo
> Darwinists are going to borrow emotionally and politically
> charged terms from general usage (which they should
> not do) then they are going to have to be extremely
> careful that they mean what everybody thinks that they
> mean.
Doing without words like "altruism", "mutualism", and
"investment" strikes me as a good idea. Though, personally,
I have more trouble with "mutualism" and "investment" than
I do with "altruism".
It is difficult pick words that mean what "everybody" thinks
you mean. It is in fact, practically impossible. There
is always a "somebody", claiming to represent "everybody",
who misunderstands. So all that can be done is to make
the definitions precise and then to correct any
misunderstandings that arise.
> Only one possible definition can exist within
> any rational usage of the term "altruism" as it
> applies to critical fitness measures within Neo
> Darwinism:
> ___________________________________________________
> Fitness Altruism:
> Any reduction in the total fitness for a selectee
> via a donation by that selectee.
> ___________________________________________________
That is a perfectly acceptable definition, assuming that the
word "donation" is taken to imply that the total fitness of
the recipient is increased by the "donation".
> Any other definition remains rationally _meaningless_.
> This is because only this definition allows
> anything at all to be meaningfully _non_ altruistic.
> If a total fitness gain can be altruistic then nothing
> at all can be rationally non altruistic so altruism becomes
> entirely non refutable in the Popperian sense of the word.
>
> It is obviously irrational to define altruism
> as a total fitness _increase_ for any actor who
> is supposed to have donated because the actor
> has provided a _net_ benefit for itself.
Do you know of any such irrational person, John? Are you
sure you have understood them?
> Very
> clearly any _gross_ cost cannot be validly used
> as a rational measure as to if that cost was a
> donation or not a donation. Only if the donation
> provides a _net total loss_ can the donation be
> validly altruistic. If it provides a _net total increase_
> for the actor then the donar is making a selfish investment
> and not a donation.
Hmmm. Some people do define "altruism" based on the direct
or "gross" cost. Other people use a "net" cost - subtracting
out any "investment returns" that result from the "donation".
You seem to disapprove of the "gross cost" people. In fact,
you call them "irrational".
OK. The process that Trivers called "reciprocal altruism"
involves an "investment" in the sense that you seem to mean.
So, you would argue (and you have argued) that "reciprocal
altruism" is bad terminology. But what does this have to
do with Hamilton? Hamilton's kin selection model uses
a net cost definition. It assumes, in fact, that there are
no "investment returns". It doesn't apply if there are
"investment returns".
> Here, if the donation increases
> the total fitness of each separate recipient (NOT a
> total of all recipient fitnesses) can the action
> now be defined as mutualistic. An investment by
> the actor could produce a net _reduction_ in every
> recipients total fitness providing a net gain
> for the actor and a net loss for every recipient.
> In this reverse situation, the recipients become
> the altruists and not the donar.
No, that situation is called "selfishness" under the
usual definitions. It would not be correct to call it
"recipient altruism". Not correct at all.
It is important to realize that the definition of "altruism"
requires that the "donor" be active. One cannot "rationally"
say that the victim of a robbery was "altruistic".
> Very clearly,
> just donating something is not necessarily
> to the advantage of the recipient. If you
> donate tons of ice cream to recipients they
> may die of premature heart disease where this
> may lower and not increase their total fitness.
> Any gross measure of any cost cannot
> tell you if the actor was altruistic, the recipients
> were altruistic in accepting the donation or the
> action was mutualistic unless a net measure of that
> cost becomes available. This is only possible when
> gains are totalled in every case. No other solution
> exists.
And here we are getting at the heart of what I think is
your misunderstanding.
To use the notation that NAS used in his proof of Hamilton's
law, the question whether "c" should be measured by
- Regression[w,z]
or by
- PartialRegression[w,z|z']
Or, to put the same question in different words, exactly
what adjustments do we make in converting a gross cost to
a net cost?
It seems that you want to treat any reception of altruism
by a donor as an "investment return". You believe, apparently,
that we distinguish altruism from mutualism by simply checking
whether "donors" have higher total fitnesses than non-donors.
If donors have higher total fitnesses than non-donors (on average)
then you say it is "mutualism" rather than "altruism". So,
apparently, you want to use the total regression formula to
estimate net cost.
Hamilton, on the other hand, would say that you are computing
the net cost incorrectly. You have to take into account that
the donors may actually benefit by receiving more altruism
than non-donors receive. These extra receptions of altruism
should NOT be subtracted out of the gross cost to get the net
cost. They are not "investment returns" in any sense. There
is no causal chain such that one can say that these extra
receptions are caused by the altruistic **action**.
If you want to estimate the cost of the altruism by looking at
the total fitnesses in the population, you have to control for
different reception of altruism between donors and non-donors.
That is, you have to use a partial regression to determine
whether the cost is positive or negative.
And, it turns out that you can get a positive cost using the
partial regression (indicating that this really is altruism),
and at the same time get a negative "cost" using the total
regression (indicating that the trait is favored by Darwinian
natural selection). When does that happen? When rb>c.
[snip remainder]
- Next message: Tim Tyler: "Re: Haldane's Dilemma"
- Previous message: Perplexed in Peoria: "Re: Haldane's Dilemma"
- In reply to: John Edser: "Re: Logic of kin selection"
- Next in thread: John Edser: "Re: Logic of kin selection"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
