Social Evolution

Harry Erwin erwin at trwacs.fp.trw.com
Mon Mar 29 10:16:14 EST 1993


There is an interesting pattern seen in the behavior of both
bower birds and modern man, but not seen (generally) in other
species: the existence of fads. In a fad pattern, the behavior
involved evolves through learning, but the goal is to "trump"
existing behavior, not optimize it. The result is that the
dynamics tend to "chase their own tail" and can be chaotic.

However, recent studies by Huberman and Glance and some work
done by my son, Jeremy Erwin, back in 87-88 suggest that
we should expect behavior to usually converge to a fixed
point if natural selection dominates, even if behavioral
innovation is intentional rather than random. The one area
where this is not to be expected is in sexual selection
("female choice") where Fisher's runaway process can
occur. This is particularly important early in the process
when the intentional component of innovation is weak. 
(See Hines and Bishop, 1983. "On learning and the 
evolutionarily stable strategy." J. Appl. Probab. 20:689-695.)
Fisher's runaway process is characterized by trait development
to the point that the advantage in sexual behavior is balanced
by the disadvantage in other situations. (See Smith, J. M., 1989.
Evolutionary Genetics, Oxford University Press. 264-269.)

The implication? Sexual selection was probably responsible
for the emergence of innovation, art, and fads in Homo
sapiens behavior 100,000 years ago. The original selective
advantage of art was that it demonstrated the planning and
crafting skills of the individual creating it. Of course, once
Fisher's process takes over, art is produced "for art's sake,"
and (as seen in bower birds) style evolution begins "chasing its
tail," leading to chaotic dynamics. We may well be -->more<--
innovative than optimal because of this. This also seems to
imply that bower birds constitute an "animal model" of human
innovation.

Cheers,
-- 
Harry Erwin
Internet: erwin at trwacs.fp.trw.com
Looking for interesting work...



More information about the Bioforum mailing list