[From Bill Powers (960604.0800 MDT)]
Bruce Gregory (960603.1500 EDT) --
Those of us teaching the course are committed to help provide the
feedback each student needs, but since we cannot read their minds
(determine their reference levels), they will have to be
responsible for seeing that the gaps between what they perceive and
what they want to perceive are closed.
It is this that gives you a way to read their minds. What you do is make
a proposal (to yourself) about some perception that a student might be
controlling. Of course you have to do this in terms of your own
perceptions of the environment as you imagine the student might
interpret it. Then, when you see disturbances of this perception, you
look for actions that would tend to oppose the disturbance or correct
its effects on that perception. If you have guessed right, every
disturbance tending to change the perception will result in a change of
actions which, if successful, would restore the perception to the
original state, or even prevent it from changing in the first place. Of
course if a perception (as you define it) is disturbed and the student
does nothing about the change, your definition is not sufficiently like
the student's perception and you have to change or refine it. This is
the Test for the Controlled Variable, of course. You can use it
anywhere.
This method can be used to teach, too. If you were in a class where
letter grades are awarded, you might ask the students if they're trying
to get a good grade in the class. Most of them would probably say yes.
So then you say, "All right, I'm going to give everyone in the class an
A, and you won't have to do any homework or study anything or come to
class or take any tests. OK?"
The initial reaction would probably be amused agreement by some of the
students, maybe all of them. But a little discussion would start to
bring out the objections. If everyone gets an A, how can the good
students get appropriate (differential) credit for their intelligence or
hard work? If the students don't have to study, how will they learn the
course material? If they don't take tests or do homework for practice,
how will they know if they've understood the material correctly? If they
all get As, what will happen when they go on to other courses that
assume understanding of the material in this one? Some students who want
As might not be aware of the competitive feelings behind this desire.
Others who want a high grade may never have thought about wanting to
learn what the course teaches just to know it. Others may never have
considered the later disadvantages of being thought to know more than
they actually know. And so on.
Students must then determine for themselves whether they want to
adopt that goal as their own. Further I plan to make it clear to
the students that they bear responsibility for getting the feedback
they need. Those of us teaching the course are committed to help
provide the feedback each student needs, but since we cannot read
their minds (determine their reference levels), they will have to
be responsible for seeing that the gaps between what they perceive
and what they want to perceive are closed.
The students already are responsible for getting the feedback they need;
they can't behave any other way. You don't have to teach them to be
control systems. The only goals of a teacher that a student can adopt
have to do with the teacher-student relationship, a social contract. But
even then the teacher's goals must be different from the student's. The
teacher's goal is for someone else to learn how to do something; that
can't also be the student's goal. The student's goal is to learn
something; that's not the teacher's goal, because the teacher already
knows it. The idea that the students and the teacher are engaged in a
cooperative search for knowledge is phoney unless the course has no
substantive content -- unless there is nothing to teach, and the teacher
has no idea of how the course is going to come out or what will be
learned from it.
The idea of a teacher "providing feedback" confuses feedback (an effect
created by the student on the student) with a disturbance (an effect
created independently by a teacher on a student). For a teacher to be
part of a student's feedback loop, what the teacher does must be caused
by the student. Students can act on teachers to produce effects that the
students want, such as getting information by asking a question, but in
order for this to constitute a control loop the teacher's response must
be dependent on the student's action in a predictable way. For example
if the student's goal is to make the teacher angry, then there must be
things the student can do or say which will predictably result in an
angry response from the teacher. In order to provide the feedback that
the student wants, the teacher must agree to become angry whenever the
proper words or deeds are emitted by the student -- and only then.
I think that what you're after is not to provide the feedback that the
student needs, but to provide the disturbances that the student would
benefit from. The feedback that the student is seeking would consist of
a perception that matches a goal-perception that the student already
has. But if the student's action results in this matching perception,
the student has learned nothing. That perception is already under
control. If anything is to be taught, the feedback must differ from the
expectation in some not-too-large way, so the student can experience the
error, and one hopes see how the error was caused by his or her own
action and find a new action that will correct the error. In other
words, the "feedback" that is "provided" must actually be a carefully-
calculated disturbance, small enough that a little reorganization will
suffice to find a way to correct it. We learn from the unexpected.
I hope that this post is an example of what I'm talking about.
···
-----------------------------------------------------------------------
Jeff Vancouver 960603.1430 EST --
Congratulations on the Psych Bull paper and the Behavioral Science
paper!
I have gotten a few recent books on motivation (Mook, 1996; Petri,
1996) that deal with hunger and other physiological regulation
processes. These books rely heavily on (i.e., teach) the control
process. When I talk to colleagues in these areas they also agree
that the control process is an accepted concept.
I'll have to get hold of these books before I can comment on the control
process that is accepted in this area. Actually, physiological control
systems have been recognized quite explicitly for 50 years or more, in
fields where working models are regularly employed (the British did a
lot of modeling of physiological systems using analog computers,
starting in the 1950s). Back in my innocent youth, I thought that this
sort of work would help to support my efforts to extend control theory
to cover all of behavior rather than just a few internal systems. But it
didn't work out that way: psychologists were quite comfortable with
recognizing control processes, but only as long as they were "merely"
physiological or "neuromotor," meaning confined to the brain-stem
reflexes. That kept control processes separated from real psychology.
Anyway, I would appreciate if anyone knew of a popular (14 cites
does not strike me as popular) reference that condemns control
theory. It helps focus counter-arguments.
I'll leave it to others like Avery Andrews to supply some references (I
don't have a finger on the pulse of the literature). One thing you
should keep in mind that that NO idea is universally "popular" among
behavioral scientists. Each school of psychology judges all ideas by
what is accepted within that school. The behaviorists, for example,
speak of "modern" or "scientific" psychology, but what they mean is only
the latest writings in their own field, which includes only a small
minority of psychologists.
Finally, I am conducting some experimental studies using a PCT
perspective and have developed a model of the process. The model
is programmed in Visual Basic, so posting the code is probably not
useful. I hoped to create an .exe file that I could send to anyone
interested, but I have not gotten it to work.
Excellent. I will certainly want to see the results -- even reading the
code would be helpful. Could you describe how the model is supposed to
work?
That is, positive feedback is telling participants only when the
action is correct and negative is telling them only when an action
is incorrect. This is quite different from the way PCT talks about
pos and neg feedback...
It certainly is. You're not even "talking about" positive and negative
feedback. Please don't take your attempts at reconciliation so far as to
amount to spreading ignorance.
----------------------------------------------------------------------
Best to all,
Bill P.