On The Uses of PCT

CSGnet Archive CSG1993
1

[From Rick Marken (930410.1000)]

Recent posts from Bill Cunningham (930409.1410), Ken Hacker [930409],
and Avery Andrews (930410.1200) as well as the continuous background
beat provided by Greg Williams, have led me to meditate a bit on the
uses of PCT.

Most of what I came up with was negative: what are the non-uses of PCT.
I could think of only one positive use for PCT: PCT is a way to understand
behavioral phenomena that involve control. For some of us, this makes
PCT plenty useful because there are lots of reasons to believe that
most behavioral phenomena involving living systems (and their interaction)
involve control. But I can tell that this particular "use" of PCT seems
pretty ethereal to many of those out there who are doing "real useful
stuff". So here is a list of what seem to me to be the major non-uses
of PCT:

PCT cannot be used to explain why an independent variable (like learning
method) produces an average difference in measures of a dependent
variable (like test scores). This means that PCT is not useful to
behavioral scientists who are trying to do and explain the results of
conventional IV-DV research -- ie, virtually all of them.

PCT is not useful for estimating population parameters from sample stat-
istics. Thus, PCT will not help people who want to predict the results of
the next election or who want to determine how many people will stop
smoking as the result of an anti-smoking ad campaign. People who
want to do these things will, quite appropriately, find PCT to be
irrelevant to achievement of their goals (as Greg Williams keeps
pointing out).

PCT will not be useful if you want to write a paper on the value
of IT for understanding human control behavior. PCT will also be
irrelevant to the discussion of information theory as a model of
living systems as communication channels. PCT does not explain
communication channel behavior -- it explains control behavior.

PCT will not help you teach the introductory course in psychology,
sociology, political science, economics, linguistics, or any other
social or behavioral science. If you taught such courses using PCT
you would quickly start down a path quite different from that taken
by conventional texts. It would not be fair to the student to teach
these topics based on PCT -- they would not be prepared for the next
courses in the sequence.

PCT would not be a good topic for a textbook in any of these fields
-- it will not sell (how did the Robertson/Powers "Intro to Psychology"
do compared to Atkinson, et al?) and the book will not help you get
tenure (Robertson had tenure already/Powers didn't need it).

PCT is useless for promoting your career in any of the above areas.
PCT is not about the things that are considered important or "state
of the art" by those giving out the academic prizes. PCT (unless it
it disguised as something that is relevant to current academic
concerns) is not a useful academic career move.

PCT is useless for people who are interested in people's outputs
rather than their purposes. Thus, most managers, teachers,
psychiatrists and politicians will not find PCT useful because
"useful" to them means "getting people to act in a particular way".

There are probably more ways that PCT is not useful. But I just thought
of another way that PCT is useful:

PCT is useful in the way that art is useful; it's good for
the soul. It helps you understand yourself and others. It helps
you deal with the difficulties and appreciate the beauties of life.
If you appreciate the joy of the pursuit of knowledge then PCT is
a very useful tool in that pursuit. PCT is useful for the soul.
Whether or not it can be useful for more mundane matters -- like
making a living -- well, that depends on how many other people end
up finding it soul-satisfying too. Since most people in my field
(psychology) don't find PCT (and the phenomenon of control)
anything other than threatening, PCT is not useful to me as a means
of making a living as a psychologist. Maybe that's why I get so
passionate about these net discussions? Maybe I feel bad that more
people don't get PCT (and find it "useful") so that PCT would be
real popular and I could go back to teaching -- as a professor of PCT.
Ah well.

Best

Rick

2

From Ken Hacker [930410]

For once I agree with just about every line Rick Marten wrote. No, I am
not drinking... Nor am I recanting (I like playing with religious metaphors
too) anything I have said before. BUT, I think Rick is totally correct in
noting that the domains of social science and PCT are different to the
extent that one cannot explain the concerns of the other. As for behavioral
science activities, most behavioral scientists, but not all, will also being
studying things other than control. Social interaction, communication, and
information have structures and dynamics that are not just about control.
Control theory simply (parsimoniously) and elegantly explains control of
individual perceptions and behaviors. I now why my neighbor, a professor
of electrical engineering specializing in control systems, says that he gave
up years ago trying to apply EE control theory to social systems. He says
the applicability was just too stretched and did not work. I think that Mary
Powers gave good suggestions on how feedback is first of all internal to any
controlling system and that messaging between systems is something I may
call social feedback, but really consists of input and ouput interfaces.
Marten is right in saying that control theory explains control and that is
a major, major accomplishment. Ken

3

[From Dick Robertson] (930410.1725 CDT)
Well I have to disagree with you about one point regarding what pct is not
good for, Rick. After teaching introductory general psychology wholly from
the PCT viewpoint (to modest class sizes of about 10-15) four or five times
now, I haven't heard any complaints from those who "got it" as they went on
to other courses, except that a couple have said they had a ball arguing with
the instructors in some other courses about things that just didn't make sense
in the traditional approach.
Best, Dick

4

From Tom Bourbon [930412.1158 CDT]

While catching up on the mail, I just came across a post from Ken
Hacker (930410) with the assertion that PCT cannot be applied to social interaci
ons -- that those interactions are not about control. Say what? Have you
read any of the things by Clark McPhail, Chuck Tucker, Kent McClelland,
or have you seen or run the CROWD/GATHERINGS program, or have you seen
my work on PCT modeling of interferemce, cooperation, control of others,
helping, or the like? In every one of those examples, people, or PCT
models, or combinations of people and models interact. In every instance,
each controller (whether person or model) controls her-his-its own
perceptions and all of the interactions occur via mutual influences on
variables in the environment. The models both recreate and predict results
of human interactions, with the same .99+ correlations between model and
person that we find in single-person tracking. And models run with
persons, at the same speed at which persons interact.

My work, and that of several of my thesis students, shows cleasrly that
PCT could play a role in social science, but that is not likely. My publication
s on the topic are in a book and a journal issue edited by members of this
net. My rejections, which outnumber the publications, are from "real"
journals on social psychology. I have been told numerous times that,
"we (the real social psychologists) are not interested in this." (sic)

Never, in our modeling, did we see a need to consider measures of
information, or channel capacity, or bandwidth. I will not deny that
those measures might be useful for people who try to ask and answer
questions different from ours, but they are not necessary in the
study of how people, each of whom controls personally specified perceptions,
interact by way of mutual influences on variables in a shared environment.
In that context, the basic ideas of PCT that apply to one person who
is tracking apply to two or more who interact. Those results seem,
to me, to set an agenda for further studies of social phenomena and
of communication. For example, some of my thesis students have made
starts at observing and recording communications (verbal and non) between
people who are learning to master a cooperative control task. It seems
clear to us that the communication that works is about how things will
be perceived when all is right, and not about which actions to use as
a way of making it right.

The people, programs and interactions I mentioned at the top certainly
represent no more than the first steps towardapplying PCT to social
phenomena, but they already show that no one can ever again safely
dismiss PCT in that domain.

Rick's list of places where PCT does not apply was wry. He was listing
both the polaces where it really does not apply, and the places where the
experts only believe it doesn't. The former class is large; the later,
shrinking.

Until later,
  Tom Bourbon

5

From Ken Hacker [930412]

In response to Tom Bourbon --

I have not yet read your experimental reports. I have read two articles
about purposive collective behavior by McPhail & Tucker (1990) and
McPhail, Powers & Tucker (1992). Both articles are instructive and
aided me in recognizing the following:

a) All behavioral and social science theories explain (or describe)
   higher-order reference signals. MT, p. 84.

b) Higher order and lower order control systems depend on each other.
   MT, p.84.

c) Purposive collective action requires the setting of similar reference
   signals. MT, p.86.

d) Planned programs of collective action facilitate complex and sustained
   collective action. MT, p. 89.

e) Humans with similar reference signals are more likely to move in
   similar directions. MT, p. 90.

f) Collective action can happen from not only similar independent
   objectives, but also objectives generated interdependently and
   as adopted from third parties. MPT, p. 1.

These two articles state some new ideas about communication and restate
some old ones. The new ones, in my opinion, are about varied sources
for reference signals (and the computer simulations of collective
behavior). The old ones are about people doing collective action
by having common reference signals. In communication theory, we have
many views of this cast in various teminology such as convergence of
representations, (collective task representations --me), coordinated
management of meaning, etc. etc. The ideas are certainly important,
but it seems to me that PCT should not reiterate what is already being
said with new language and then attack the uses of old language for having
the entire set of concepts wrong!

Tom, I would love to read some reprints of your work, not to criticize
them, but to understand them and respond with a critical (but open) eye.
My address is: Dept. of Communication Studies, New Mexico State University,
Las Cruces, NM 88003. I think that my discipline may be open to more
about your simulations that other social sciences have been. There are
many reasons for this, one of them being that our discipline has had
to differentiate itself from the older social sciences and from electrical
engineering approaches to communication (signal processing) so much that
we are open to new ways of exploring, describing, and explaining human
communication processes. Also, because of technological relationships,
most of us are very open to computer simulations that are grounded in
strong theory and which show new ways of testing hypotheses about social
interactions and relationships.

I would like to propose the following kind of experiment. Let me suggest
a research question involving real people talking, not doing exercises,
but relating verbally and nonverbally to accomplish some task. Let us
then study and describe their behaviors from two points of view: the
variable analytic and the PCT. Let's see what the differences are in all
areas of study design, data gathering, and data interpretation. I cannot
make any strong predictions about this, but I would love to see what happens.
My hunch (a weak prediction) is that the two approaches will show some
overlap but mainly different data about different questions, with both
data sets supporting each other at different levels of analysis.

When I suggest that social interactions are not about control, I mean that
a) Control is endemic in each communicator, but does not explain the
key processes of producing information, creating meanings, and forming
relationships, and b) Control systems are interdependent within each
communicator, but we do not know how control systems relate to each other
in terms of specific qualities of influence on reference signals, etc.

Thank you for the thoughts and many thank yous for not telling me that
because I do not accept, I do not understand. KEN

From Tom Bourbon [930412.1158 CDT]

While catching up on the mail, I just came across a post from Ken
Hacker (930410) with the assertion that PCT cannot be applied to social interaci
ons -- that those interactions are not about control. Say what? Have you
read any of the things by Clark McPhail, Chuck Tucker, Kent McClelland,
or have you seen or run the CROWD/GATHERINGS program, or have you seen
my work on PCT modeling of interferemce, cooperation, control of others,
helping, or the like? In every one of those examples, people, or PCT
models, or combinations of people and models interact. In every instance,
each controller (whether person or model) controls her-his-its own
perceptions and all of the interactions occur via mutual influences on
variables in the environment. The models both recreate and predict results
of human interactions, with the same .99+ correlations between model and
person that we find in single-person tracking. And models run with
persons, at the same speed at which persons interact.

My work, and that of several of my thesis students, shows cleasrly that
PCT could play a role in social science, but that is not likely. My publication
s on the topic are in a book and a journal issue edited by members of this
net. My rejections, which outnumber the publications, are from "real"
journals on social psychology. I have been told numerous times that,
"we (the real social psychologists) are not interested in this." (sic)

Never, in our modeling, did we see a need to consider measures of
information, or channel capacity, or bandwidth. I will not deny that
those measures might be useful for people who try to ask and answer
questions different from ours, but they are not necessary in the
study of how people, each of whom controls personally specified perceptions,
interact by way of mutual influences on variables in a shared environment.
In that context, the basic ideas of PCT that apply to one person who
is tracking apply to two or more who interact. Those results seem,
to me, to set an agenda for further studies of social phenomena and
of communication. For example, some of my thesis students have made
starts at observing and recording communications (verbal and non) between
people who are learning to master a cooperative control task. It seems
clear to us that the communication that works is about how things will
be perceived when all is right, and not about which actions to use as
a way of making it right.

The people, programs and interactions I mentioned at the top certainly
represent no more than the first steps towardapplying PCT to social
phenomena, but they already show that no one can ever again safely
dismiss PCT in that domain.

Rick's list of places where PCT does not apply was wry. He was listing
both the polaces where it really does not apply, and the places where the
experts only believe it doesn't. The former class is large; the later,
shrinking.

Until later,
  Tom Bourbon

···

----------------------------Original message----------------------------