Lookingy for the heart of reinforcement theory

[From Rick Marken (950706.2110)]

Bill Powers (950706.0905 MDT) --

What we're doing isn't exactly developing an acceptable test to
distinguish reinforcement theory from control theory. It's more along
the line of trying to figure out what reinforcement theory IS. We have
to do that before any test is possible.

I am more of a phenomena freak than a modeller. I like to build models
but I especially like to see what observations would be expected if
the model were correct. That's one reason I like PCT so much; it's
a simple model (so even I can understand it) and it predicts many
fascinating and surprising phenomena. Many of these phenomena are
as fascinating as they are (to me) because they contradict one's
expectations based on other models of behavior. S-R theory leads
one to expect a correlation between cursor and handle movement in a
tracking task -- and there is none. Cognitive theory leads one to
expect that consistent (programmed) means will be used to produce
a consistent end -- in fact, the means vary (as necessary) to compensate
for invisible changes in the connection between means and ends,
so that the end is consistently produced. Reinforcement theory leads
one to expect that random consequences will randomly stregthen and
weaken S-R bonds -- yet the S-R bonds that we see are not only non-random
but they are exactly those that keep the consequences under control.

I find it kind of frightening that in 1995 we would have to be trying to
figure out what reinforcement theory (developed c. 1930) is. It suggests
a rather deep lack of interest in or understanding of modelling in the
EAB community. Either way, it bodes ill for anyone seeking to move EAB
toward PCT. If EABers have no interest in modelling then what good is it
to present them with an accurate representation of their model? The same
holds if they don't understand modelling; what will they learn from your
model of reinforcment theory if they don't know what a model is?

I am working on another demonstration of the relatoinship between
consequences and behavior; the aim of the demonstration is to show
that the same consequence is produced by variable means. It seems to
me that this result would be inconsistent with reinforcement theory;
but maybe not. Since no one seems to know what reinforcement theory
IS, how would anyone know what behavior phenomena to expect if organisms
actually operated according to the theory? SO I've stopped working on this
demo; I'm waiting to find out what "the real reinforcement theory" would
actually lead one to expect to see.

By the time we have finished, his reinforcement model will probably be
totally unacceptable to his colleagues in EAB :slight_smile: He will find himself
ostracized by his former friends, and he will have to join the ranks of
the institutionally homeless like you and Tom Bourbon and all the others
who have contracted this disease.

I sure hope not. But I think you give the wrong impression here. Tom
was never ostracized by friends and I was ostracized only becuase I'm,
well, me;-) And we are institutionally homeless by choice; we just
couldn't, in good conscience, keep teaching the stuff that we had to
teach in order to provide students with the required curriculum. Augsburg
was happy to let me teach a regular course on PCT; but then what would
I have done when I had to teach Intro, Research Methods, Learning, etc? Give
a discalimer at the beginning of each semester saying "You have to learn
the following c**p to pass the GRE; but it is all BS; to see why, take my
PCT class next semester"?

Anyway, good luck on your search for the heart of reinforcement theory.
I can hardly wait to see what you find.

Best

Rick