[From Gary Cziko 921025.0050 GMT]
I said (921024.0300 GMT):
I am
thinking of (c) showing how PCT provides a unified theory of psychology in
which the various flavors of psychology now existing (e.g., cognitive, S-R,
reinforcement theory, etc.) can be seen as dealing with special, (very)
limited subdomains of behavior and cognition.
Rick Marken (921026.0930) replied:
I do think this is a great idea. But there is still one little, teensie,
weensie problem with this friendly approach; it is hard to keep your
audience from noticing that their "flavor" of psychology is being revealed
as a misinterpretation. The "Blind men" paper doesn't just show that the
cognitive, S-R and reinforcement data can be seen as "special cases" of
control; it also shows that the explanations of these data are wrong.
For example, cognitive models (like the language models we've started
discussing again on the net) are output generation models; the appearance
(that cognitive behavior is generated output) is taken at face value and so
the explanations have been completely off-base. So when you say that PCT
is a way of integrating different approaches to understanding behavior,
you are also saying that the conclusions that were based on that approach
are wrong.
Maybe so, but don't you think it was first necessary to think of behavior
as generated output and then see the problem with it? In this way, the
generated-output view leads eventally to a better view and the
generated-output people can be seen in a friendly way as leading to PCT.
To use one of your favorite examples, do you think it was possible to come
up with a heliocentric view of the solar system before coming up with a
geocentric one? A generated-output view of behavior is probably better
than one which says behavior is caused by angels or the soul interacting
with the body through the pineal gland. But all these theories are just
guesses anyway. It just happens that some of these guesses are easier to
refute (S-R) than others (PCT) and so we think that PCT is closer to the
truth (and probably is).
there is no getting away from the fact that PCT says that
psychologists can take their explanations of behavior and toss them in
the waste basket. In terms of the "Blind men" analogy, it means that
the fellow who has developed all those terrific models of the elephant
that explain its "wallness" can now just throw those models away -- the
elephant's "wallness" is just a side effect.
But if the blind man is never going to move from the side of the elephant,
he might as stay with his "wallness" hypothesis. If I'm never going to
venture beyond the Arctic or Antarctic circles, I might as well stick with
my hypothesis that the sun rises and sets each day. And if a psychologist
is going to stick with hungry rats and pigeons and schedules of
reinforcement, he might as well stay with operant conditioning (as Greg
Williams has said, Skinner was very good at working organisms, but not very
good at understanding what makes them work).
I think you have to take this into consideration, Gary, if you want to
try to make the "Blind men" paper palatable. It's true that the
feedback analysis integrates observations -- and I can see that that
fact can be presented to psychologists in a friendly way. But it also
shows (and there is just no getting around this) that current models
of behavior (that were based on taking these observations at face
value) are just, flat out, downright wrong.
But doesn't the "rightness" or "wrongness" depend on the context in one
which is working? I don't think that NASA worries too much about the fact
that the mass of the space shuttle approaches infinity as its velocity
approaches the speed of light or that the engines increase their mass as
they heat up. Newtonian physics works very well for NASA, thank you. But
you can't stick with Newton if you want to understand physics in less
limited domains, if you want a general theory.
"So what if PCT
let's me see SR,cognitive and reinforcment as special cases of a
bigger picture; I'm only interested in reinforcement so I'll just keep
doing my operant conditioning studies."
And you would have every right to do so. But people interested in a more
general theory for understanding behavior will not have that option. We
can give them another.
I think the paper already gets
this "so what" response; because it is only in a little section at the
end that I gently suggest that ths PCT point of view requires a whole
new approach to explaining SR,cognitive and reinforcement phenomena --
an approach based on recognition that what you are seeing is control of
perception. So, I would really be interested in seeing what you have
in mind to make PCT an attractive option for conventional psychologists.
But it's not really a whole new approach, is it? In some ways, its a
welding together of cognitive and S-R ideas, as your paper shows. It's
nothing but input-output connections through and through. Ah, but when you
connect the output back to the input, include a reference level and
increase the loop gain, something very new emerges--control. Emergence is
trendy today, and we could push that as well. Control emerges from
connecting input to output. Social phenomena emerge by putting individual
control systems in a social setting.
Yes, I think the paper can be made friendly and influential as well. But
it's just a guess. We won't know until we try.
Yours friendly, Gary.
ยทยทยท
------------------------------------------------------------------
Gary A. Cziko Telephone: (217) 333-8527
Educational Psychology FAX: (217) 244-7620
University of Illinois E-mail: g-cziko@uiuc.edu
1310 S. Sixth Street Radio: N9MJZ
210 Education Building
Champaign, Illinois 61820-6990
USA
------------------------------------------------------------------