teaching goals; RT and PCT

[From Bill Powers (960330.0600 MST)]

Chris Cherpas (960328.0909 PT) --

     Regarding the discussion on how to use PCT for designing
     educational programs, I would say it's probably a mistake to think
     that you can "just teach goals" and have students discover how to
     get to them on their own. Centuries of cultural evolution would be
     wasted if each generation of kids had to rediscover everything
     themselves ...

I agree with this; I wasn't taking the Socratic position that all
knowledge lies within us and all we have to do is discover it. That
would truly be starting from scratch. What I meant to say was that
instead of teaching children what to do, we should teach them what to
perceive that will tell them when they have done it (because this is
really what they learn anyway). For example, in order to know that you
have solved C = A - B, you should find a number C such that C + B = A.
In other words, the goal of subtraction is a number which, when added to
the diminuend, yields the subtrahend, if I have those terms right. This
makes subtraction merely a permuation of addition. I believe that
children are now taught subtraction this way, in at least some places.

     There are specific topographies of behavior that are better at
     getting to a given goal than other topographies.

This may well be true. Even when it is true, however, you still have to
teach the behavior. When you come right down to it, you can't "teach a
behavior." All you can do is _require that a behavior be produced_. It
is beyond the capacities of a teacher directly to show a child who can't
perform a behavior how to perform it. The teacher can show what the
behavior should look like, feel like, or sound like once the child has
learned to perform it, but there is no way to help the child discover
the internal acts that must be performed in order to reach that goal.
That process is part of a child's internal capacities; if it's not
there, no learning will happen.

If a child has trouble learning how to produce a given behavior, then
the teacher has to go _down_ a level and teach the child what lower-
order perception has to be controlled in order to reach the higher-order
one. A child who has difficulty writing letters may be holding the
pencil in an awkward way. The teacher may then shift the goal to be
taught from that of forming the letter to that of creating a specific
configuration of pencil and hand. The teacher can't tell the child how
to operate the muscles in order to achieve the new configuration, but
the teacher can help the child to perceive the new way and encourage the
child to adopt this perception as the new goal for holding the pencil,
and at least try it to see if it works better. It takes tact and empathy
to do this, because children (at least one child I knew intimately --
me) are often so happy to have found one method that works that they
strongly resist changing it to a better one. Reorganization is not fun,
even if the results are pleasing.

Or maybe that's just the opinion I formed after years of ridicule,
mockery, criticism, and shaming that some teachers used as a way of
getting me to become semi-skillful at motor acts.

The Skinnerian technique of shaping is probably a good model for
teaching. Since a teacher can't really teach outputs, but can only
require that they be produced, the best way to teach is to require new
outcomes only a little different from those that the child can already
control. If the new perception to be controlled is not too different
from the old one, only a small amount of reorganization is needed at
each step. Of course the steps should not be so microscopic that change
takes forever.

     I don't think grossly stated goals like, "What principle might
     allow one to understand how forces interact with masses and their
     movements?" will lead to f=ma in the amount of time a student has
     available to learn what is possible to to learn in a lifetime.

You're right, of course. But think hierarchically. What is a means of
control at one level is a goal at a lower level. If you started out by
defining the highest-level goal first, you would not have a base of
lower-level perceptions and goals to build on. Before you can teach f =
ma, you have to teach f, m, and a, and then the general concept of
representing relationships as mathematical functions, and only at the
end the particular physical relationship of interest. If the student
hasn't become comfortable with algebraic manipulations, so that f = ma
and m = f/a and a = f/m look like the same relationship, it's futile to
go on to use this formula as a means of expressing Newtonian physics.


      ... there has been a neglect of forms of representations that can
     easily be manipulated and recognized quickly/visually. Perhaps
     this amounts to "going some levels down" -- making higher level
     perceptions representable at lower, more efficient/faster levels.

As you can see, I agree. But don't overlook the hierarchical
relationships: in order to control a perception at one level, you must
be able to _vary_ perceptual goals at a lower level.

     Provide exercises that are closer to the tracking experiments, in
     contrast to the usual question-and-answer formats. Within minutes,
     one might become much more fluent in controlling some educationally
     relevant perceptions than hours of lecture-question-answer

This is a really good idea. Just imagine trying to teach a child how to
tie a knot without ever letting the child experiment with making actual

Have you gone through my Demo1 and Demo2 disks? The Demo1 disk presents
the phenomenology of control without any theory -- it just shows what
happens and what relationships come out of it, with a running commentary
in the form of explanatory screens. Demo2 introduces the theory of
control, based on the experiences of Demo1. An "interactive block
diagram" is used, with the ability of the user to alter parameters and
see the effects at each stage of building the diagram. There's even a
teacher's guide on the disk! I would like to know what you think of
these programs as pedagogical tools.
Rick Marken (960329.1300) --

     I think that most competent therapists will adjust what they do in
     therapy in order to avoid the interpersonal conflicts that would
     result from treating their patients like objects; and patients
     probably adjust to therapists in order to get what they want from
     the therapy. This must be why so many different approaches to
     therapy seem to work; there is no single "correct" technique for
     doing therapy; the "correct" techniques are those that are
     consistent with the PCT nature of people.

That's a good observation. I think it was Kubie who observed that with
long experience, therapists of different schools come to resemble each
other more and more in terms of how they interact with patients, and are
less and less distinguishable on the basis of their theoretical
backgrounds (although they may talk about what they do very

     Once therapists know how people work they will know why certain RT
     techniques work and why some don't; they will also know why certain
     non-RT techniques work and why some don't. They will learn that
     when your goal is to fix something (like a human psychological
     problem) it is just as important to know what's wrong with the
     person (based on PCT) as it is to know possible ways to fix it
     (based on technique).

Right, and if they don't learn how people work from their "home-base"
schools of psychology, they learn it from experiences with patients --
particularly failures.

I wonder, though, about psychiatrists who rely on the Big Book (is that
the DSO?) for diagnosis and treatment. There is a whole school of
psychotherapy that is based on the medical model, where you assess
symptoms "objectively" and apply the "indicated treatment." This mass-
production approach, in which only population statistics is used to test
the efficacy of treatments, seems designed to prevent one from seeing
what is wrong in individual cases, and from judging success and failure
one person at a time. I think this approach is an exception to the
principle you describe -- there seems to be little convergence with the
approaches of non-invasive therapists (on the whole -- there are
individual exceptions, as usual). If you listened to some proponents of
the medical approach, you could get the idea that any personal problem
can be solved by the ingestion or injection of the appropriate mix of
Best to all,

Bill P.