What's in a name? Reply to Lintern

[From Rick Marken (931001.0900)]

Dag Forssell (930930 2240) --

I was told:

"PCT is a novel approach, but one we would prefer not to pursue.
Even in the most positive sense, "control" is a paradigm that
doesn't leverage the "boundaryless" culture we are creating."

I hate to say this but, after reading this, I have become
convinced that if you want to make a living with PCT I think
you should jettison the word "control" completely from all of your
materiels. While I agree with Bill P. that "control" is the right
word for what PCT is about, it takes quite a bit of work to get to
the point where people understand, not only the "good" sense of
control but, also, the deeper implications of the fact that humans
are controllers. It takes quite a bit of education -- and an
audience that is willing to be educated, as you always point
out -- to get people comfortable with the term "control" or even
with the term "perceptual control theory". You can't get people
to control for your product (PCT) if they are busy defending
against it because of what they imagine it is (because of the
name "control"). Even on this net, from control engineers yet, we
get posts about "control is violence". Obviously, "control" is
a lousy name for your product. Why not repackage it?

Academic types like me (who care mainly about a world where the
rubber does NOT meet the road) can go on talking about "control"
all they want; they have nothing to sell; they're just trying to
give PCT away. But if you want to eat (or buy a new Lexus, or
whatever) then I say chuck the word "control" becuase it is
WORDS (images) that sell. My personal suggestion is that you
appropriate a synonym for "control" from the trendy sciences
that have become popular mainly because they have selected cute
names ( eg."artificial intelligence", "expert systems", "chaos",
etc).

Wherever you use the word "control" in your literature, I
suggest that you might substitute the words "non-linear
attractor" or, perhaps, just "attractor". Another possibility
is "affordance" --"affordance theory" sounds very facilitative,
doesn't it? Or "coordinative structure" theory, "coordination"
theory, "goal achievement", or "quality" theory.

I'm serious about this. If you want this to work, eliminate
the word "control" from your documents; we're talking "sales"
here. What you substitute is up to you -- but that's what
makes a good salesman. What's in a name? In sales (which is
based almost entirely on getting people to IMAGINE what the
right stuff) I'd say EVERYTHING.

Gavan Lintern (931001) --

>There is a strong sense in your reponses that the fact that you have a
>working model is central to your argument.

It's just ONE component of the central argument of PCT which is
that, when people control, they are controlling perceptual
variables relative to an internally specified reference. Behavior
(meaning "purposeful behavior" or "control") is the control OF
perception (most conventional psychologists reword that phrase
to themselves so that it is consistent with their preconceptions:
thus I have had reviewers of my papers say "of course, perception
controls behavior". This is flat out wrong. Perception IS controlled
(some of them -- OK Martin?); perception does NOT control, guide,
cause or tell people what to do). We use working models to show
that this phrase (BEHAVIOR IS THE CONTROL _OF_ PERCEPTION) is,
indeed, a correct decription of what is going on in purposeful
behavior.

there are many working models
around, many of which I suspect you would not like.

I like working models becuase I can understand what they do
and whether they actually behave like the system being modelled.
The only working models I might "not like" are those that 1) don't
work like the system modelled (that includes control models as
well as non-control models) 2) that do not include an appropriate
model of the environment in which the system behaves or 3) are
purely stochastic models, where statistical measures of model
behavior are compared to statistical measures of system behavior.
But I'm pretty liberal about 3); stochastic models do have their
place. We've even used some in PCT (for e. coli type behavior).

Thus you presumably
are not claiming that the existence of a model establishes
your case.

Of course not; it is the existence of the phenomenon of control
that establishes our case. The fact that we can model this
phenomenon quite precisely with a simple model is just the
nail in the coffin (for S-R and cognitive type models, anyway).

My earlier comments on the CROWD model as udercutting the primary
position of HPCT came from an earlier discussion with Gary Zciko
in which we seemed to agree that this model worked in a similar
fashion to the termite nest construction model.

It's difficult to see how CROWD can "undercut" HPCT since each
individual in the crowd is modelled as an HPCT -- a hierarchical
control system (just two levels, I think, but, still,
hierarchical).

What significnce is placed on the patterns that emerge.

The OBSERVER places "significance" on them. The "rings" of
actors, for example, look "significant" to an observer who
sees them as a special kind of pattern. But these rings are
not "controlled" results of the actions of ANY of the actors;
in fact, as I believe Bill Powers explained in his post to
you, it is easy to show that these rings are not under control
AT ALL -- not by any individual actor nor by any magic "uber"
controller. The patterns that emerge are "irrelvant side effects"
of individual control actions. Their significance is that they
show how "irrelvant side effects" can look significant to an
observer who doesn't understand the possibility that the
individuals might be controlling their own perceptions. A
social scientist, for example, might be tempted to study
"ring forming" behavior. This would lead to the waste of a
lot of govenment research grant money on the study of an
irrelevant side effect of a real phenomenon -- individual
control of perceptual input. The study of rings (observable
behavior) would be very unlikely to lead to the discovery of
what was really going on -- individual actors controlling
their own perceptions of distance from objects and actors,
their rate of approach to these entities, etc. (as is being
done by the actors in CROWD).

Best

Rick