[Martin Taylor 2000.06.28 21.44]
Going back to before I went on holiday...
[From Mike Acree (2000.0501.1340 PDT)]
Rick Marken (2000.05.01.1310)--
>I have some questions.
>
>1. What is "cooperation"?
>
>2. How did Darwin (or any evolutionary biologist) know when
>it was occurring and when it was not?Reasonable questions. It was my intention to be granting the most liberal
possible definition of cooperation in my example of the bees and the
flowers. But we should probably ask Martin. I don't necessarily fault him
for having introduced the terms without defining them, but I do think he
owes us an explanation (when he returns) of the very strange idea that
cooperation and competition are themselves in competition.
I thought I had defined them pretty well in the Web site discussed by
Mike: <http://www.mmtaylor.net/PCT/Mutuality/index.html>, but
apparently not. Here goes, again.
In my terminology, "cooperation" occurs when the actions of one
control system eases the control exercised by another. "Competition"
occurs when the actions of one control system makes it more difficult
for another to control. Neither need perceive the existence or the
actions of the other. In the analysis that led to this discussion,
the effects were largely side effects. Neither "cooperation" nor
"competition" need be reciprocal.
"Cooperation" often happens because the actions of one control system
either reduce the influence of a disturbing variable on the
perceptual signal of another, or because the actions of one control
system affect the environmental feedback path through which the other
influences its complex environmental variable.
"Competition" often happens because the environmental variables of
two control systems are not orthogonal, so that the control actions
of one necessarily disturb the perceptions of the other. But it can
happen that it is the side effects of one's control actions that
disturb the perceptions of the other. In that case it is the output
vector of one that is not orthogonal to the perceptual vector of the
other, even though the two perceptual vectors may be orthogonal.
The two effects are labels for the poles of a continuum, in which the
central point is orthogonailty.
"Cooperation" wins out over "competition" in the sense that
reorganization tends toward conditions of better control. Better
control occurs more readily when the actions of one control system
ease control by the other system than when its actions make control
by the other more difficult.
Naturally, in a world of many control systems, a _global_ measure of
competition and cooperation will be a vector measure, and some of the
elements of the vector will be on the cooperative side and some on
the competitive side. The vector will have n*(n-1) elements of the
form Xij,
which is the degree to which the actions of control system i
aids/inhibits the control performance of control system j.
My contention is that in a long-evolved system of interacting control
units, the mean of Xij over all ij tends toward more cooperation,
despite the obvious fact that some competition is inevitable (some
Xij will be strongly competitive). That is the sense in which I claim
"cooperation wins out over competition." See
<http://www.mmtaylor.net/PCT/Mutuality/index.html> for more detail.
Sorry to drag this up again after so long, but Mike did say I owed an
explanation when I returned, in a message I didn't see until after I
returned from a subsequent trip.
Martin