A PCTer among the manual controllers

[Avery Andrews 930923.0943]
(Rick Marken (930922.0830))

Can you describe a concrete case where real problems are caused by some
manual control engineers failing to understand the difference
between the `objective' quantity that the customer of a manual control
system wants to control, and the perception that the operator actually
controls? It's my impression that professionals tend to be pretty
resistant to conceptual points until they see a concrete illustration
of how understanding the point will help them do their jobs better
(make more money and stay out of court).

Avery.Andrews@anu.edu.au

[From Rick Marken (930922.0830)]

Greg Williams (930922) suggests once again that PCTers willfully
reject the PCT-like ideas of others who apply control theory to
various aspects of performance:

Greg cites a number of quotes from Wickens' ENGINEERING PSYCHOLOGY
AND HUMAN PERFORMANCE (HarperCollins, 1992) which are mercilessly
(and I think quite accurately) exposed by Bill Powers
(930922.0710 MDT) as evidencing a rather substantial lack
of understanding of the controlling nature of living systems.

I suspect that Bill's remarks (which I would have made myself
had I not been stuck in the Pacific time zone) will be taken by
Greg as another piece of evidence that PCTers are trying to
drive an unnecessary wedge between themselves and "establishment"
scientists who are applying control theory to behavior. Just
for the heck of it (knowing how tough it is to say anything
that would change Greg's mind on this) here is a little parable
about "A PCTer among the manual controllers".

Some years ago I gave a talk to a group of "engineering
psychologists" on the possible applications of PCT in
their (our?) profession. There was at least one 'manual
control' guru in the audience. The talk was politely received
but the questions from the audience after the talk (including
those from the guru) made it clear that I might as well have
given the talk in Greek. I suggested that one of the main goals
of a "task analysis" is to discover the objective correlates
of the perceptual variables that the "operator" is controlling.
There was no sign that this rather basic point was understood
AT ALL. For a group that was presumably already looking
at people as living control systems, there seemed to be some
fuzziness about fundemental points. When I put up a diagram
with the reference signal (and, therefore, error) inside the
organism there was obvious puzzlement. And when I did a version
of the "mind reading" demo to show that it is difficult to tell
what a person is controlling just by looking at the person's
actions (or the results thereof) -- something that I would have
imagined they should have been well aware of -- nobody "got it"
at all.

So there seemed to be a couple of areas where our friends the
manual controllers were a bit hazy on some basic PCT concepts;
they didn't seem to understand 1) that it is perception (not
output) that is controlled and 2) that perception is controlled
relative to potentially varying reference specifications INSIDE
the controller.

If they DID understand this, then they didn't seem to be doing much
about it. I have never seen a description of how to do "the test
for the controlled variable" in the literature of 'manual control'.
Without this methodology, how can one do anything but get lucky
about the way one draws their diagrams, as I believe Milsum did?
If Milsum actually understood his diagram, he would have seen that
the observer of a living control system cannot simply "see" what
the system is controlling. Milsum would also have understood that
unless one knows what perception the system is controlling, the
observer has no way of knowing why the system is acting as it is.
So Milsum could have developed "the test" (if he understood
his own diagram) and he would have, indeed, turned the study of
manual control into PCT -- and I would hail his work as a
parallel development of PCT.

So far, I have seen no evidence of a parallel development
of PCT; a Wallace to Powers' Darwin. I personally think that
it would be wonderfully exciting to find such a parallel
so I hope you keep looking. By the way, here are MY criteria
for what would count as a parallel development of PCT:

1. Explicit statement that organisms control perceptual
analogs of variables in the "objective" world experienced
by the observer of organisms.

2. Explicit statement that organisms themselves determine
the preferred states of these perceptual variables.

3. Explicit description of a methodology like "the test"
that is designed to determine which variables are being
controlled by an organism.

In a parallel development of PCT, ALL three criteria would
be met.

Happy hunting

Rick