"Guy Hoelzer"  wrote in message
news:cc1bkf$2le4$1{at}darwin.ediacara.org...
> in article cbumke$1qm4$1{at}darwin.ediacara.org, Perplexed in Peoria at
> jimmenegay{at}sbcglobal.net wrote on 6/30/04 8:35 AM:
> > "Guy Hoelzer"  wrote in message
> > news:cbs4a3$1276$1{at}darwin.ediacara.org...
> >> After all, if we are both toxic to predators, then I benefit by
> >> being near you and you benefit by being near me.  I think that
> >> reciprocity is fundamentally about positive feedback whether
> >> or not social interaction is involved.
> >
> > I agree that "reciprocity is fundamentally about positive
feedback",
> > but I strongly disagree that "I benefit by being near you and you
> > benefit by being near me" constitutes feedback.  For feedback, you
> > require something like "I benefit *because* you benefit".  Or, to
> > look at it in another way, the distinguishing feature of a feedback
> > *loop* is that it can be broken.  If I cease to provide benefit to
> > you, then thru some chain of causality, this must result in you
> > ceasing to provide benefit to me.
> >
> > You don't have feedback just because you have causal "arrows"
> > pointed in both directions.  Those arrows have to be connected.
>
> Good point.  So it may be more accurate to describe this model, let's call
> it the "proximity=mutual benefit" model, as something more
like quantum
> entanglement than positive feedback.

A very evocative analogy!  I share what it evoked in me below.  But first...

> I agree that this models lacks what we
> might call reciprocity, but don't you think that the symmetry of positive
> effect alone is sufficient to drive the reciprocal altruism model?

Symmetry of positive effect, along with some other conditions, may well
be enough to lead to the increase in altruism in an evolutionary model.
It does so in Hamilton's model.  My main concern here was to defend
the received meaning of "reciprocal".  In both common usage and in
Trivers' model, the word assumes a causal linkage in the feedback.
It also assumes that the feedback is directed to an individual.  A
reciprocator is not "repaying a debt to society".  He is repaying
a "debt" to an individual.  In Hamilton's kin selection model, by
contrast, there is symmetry of positive effect, but not "reciprocity".

"Reciprocity" need not involve intentionality, however.  I think
that the interchange of materials between cytoplasm and mitochondria
is reciprocal, in that the various antiport machinery constrains
the cytoplasm to stoichiometrically supply food to the organelle
to the extent that it extracts energy.  That is perfectly good
reciprocity, by my definition.  In operation, it is not even
teleonomic - merely teleomatic, if that.

> I was actually going to say something about symmetry, rather than "positive
> feedback" originally, but I wasn't sure that I could express myself well
> enough in that framework to find a resonance with this audience.

One of the first things they teach you in an applied statistics text
is that "correlation" is not the same thing as
"causation".  And one
of the canonical examples of how you can have a correlation without
direct causation is when the two things that are correlated are both
results of the same fundamental cause.

However, in spite of this warning, we continue the practice in
mathematical science of reasoning by writing down equations in
which the two sides are joined by the symmetric equal sign.
One cannot tell by inspection which side of the equation is
causal of the other.  And, in fact, in our mathematical reasoning,
the direction of the causality is not important.  Models can
still work, and can be explanatory, even though they may involve
what looks like causal nonsense.

A prime example of this is Hamilton's kin selection.  It is
essential to that model that there be genetic correlation between
the donor and the recipient.  In the case of siblings, or cousins,
neither the donor nor the recipient is the cause of the genetics
of the other.  However, there is a correlation between them,
caused by a shared fundamental cause - namely the genetics of a
shared ancestor.  It was Hamilton's genius to recognize that
correlation is enough, that you don't need causality, or rather
that the math still works if some of the causal arrows must be
traced backward.  Edser doesn't seem to accept this.  Similarly,
in physics Bohr recognized that correlation is enough, that
you don't need conventional causality.  But Einstein, along
with Podolsky and Rosen, couldn't seem to accept this.  (I don't
know whether the stroking implicit in this analogy will soften
or harden John's position.  Time will tell ;-)

All of this tends to weaken my claim that "causality", rather
than "symmetry" is important in the meaning of "feedback" or
of "reciprocity".  I guess that it is not absolutely essential
that an explanatory model have all of the causal arrows
connected and pointing in the same direction.  But still,
I would like to reserve the words "feedback" and "reciprocity"
for those (most common) cases in which the arrows DO all
connect properly.

One final observation, or perhaps speculation:  The reason
why science so rarely needs to consider explanations in
which causal arrows run in the wrong direction is that
such explanations usually don't work.  Something else,
not yet considered here, is needed before backward causality -
the future determining the past - can actually be effective.
It is evocative, for me at least, that physicists sometimes
refer to this "something else" as a "selection rule".
Wilkins is perhaps wise in trying to rein in my use of
teleological language before I say something really dangerous.
;-)
---
þ RIMEGate(tm)/RGXPost V1.14 at BBSWORLD * Info{at}bbsworld.com

---
 * RIMEGate(tm)V10.2áÿ* RelayNet(tm) NNTP Gateway * MoonDog BBS
 * RgateImp.MoonDog.BBS at 7/1/04 10:36:16 PM