[From Rick Marken (950902.1000)]
Bruce Abbott (950902.1100 EST) --
The physicists with their ferromagnetic model are simply viewing
certain phenomena at a higher level of abstraction (than the
generative model), and finding that certain useful regularities emerge
If this were all they were doing then I think we would have no problems
with the article Tom critiques. But my impression, from reading Tom's
post, is that the physicists think that their model IS a model of fish
behavior. The same kind of misapplication of physical models to living
systems is seen throughout the literature on artificial life, complex
systems and, for that matter, the life sciences in general. I certainly
wouldn't have a problem with S-R, reinforcement and cognitive (planned
output) models of behavior if these models were used simply to describe
interesting regularities in behavior.
Unfortunately, the people who develop S-R, reinforcement and cognitive
models of behavior think (like the physicists) that these models actually
characterize the nature of the systems being modelled. Thus we have
students of living systems who actually think of cause-effect as
an accurate characterization of the processes that generate behavior --
rather than as a convenient fiction.
I read a paper on traffic models in American Scientist some time ago.
I had no probelem with this article because the author made no effort
to suggest that the model was anything other than a useful fiction.
So I have nothing against the kind of descriptive model that you say the
physicists are using to describe the behavior of the fish. If they
are using their maodel simply to predict the best place to throw
down their fishing nets, then I have no problem with the model. If, however,
they are using the model to describe the nature of fish (as I get the
impression they are) then I have a BIG problem with the model.
One thing that seems to distinguish PCTers from conventional behavior theorists
is that we take our theory seriously. That is, we believe that the PCT
model -- to the extent that it is correct-- REALLY characterizes the
nature of living systems; all living systems are controlling their
perceptions all the time. Conventional behavior theorists seem to treat
modelling as a game:if a reinforcement model fits the data that's nice
but it doesn't mean, apparently, that all living systems are controlled
by reinforcement all the time (witness how Skinner, for example, exempted
himself from this control). I think the physicists that Tom describes have
more the conventional behavior theorist rather than the PCT attitude toward
theory. I think they must becuase it doesn't seem to trouble them that,
if taken seriously, their model implies that their own behavior is guided
by external forces. I would think that people who are clever enough to
develop a model like the one Tom describes would want to think that they
had come up with it on their own. But, maybe not.
Bill Powers (950902.0100 MDT)
Bill, your posts on competition have been sensational. It was
competition that was my own undoing in the recent exchanges on
the net. We can't stop being control systems but it really helps
if we understand PCT; it sure helped me.
Best
Rick