Higher levels; PCT and social phenomena

[From Bill Powers (930510.0830 MDT)]

Bob Clark (930509.2130) --

Putting the concepts of PCT into ordinary language as you suggest
is a fine idea, and I endorse it. There are sticky spots in doing
this, however: those where PCT and common sense part company.
Many people speak of emotion, for example, as if it's something
that the outside world does to them, and with which they must
then try to cope. It's not easy to present a compelling case in
ordinary language for the idea that emotion is part of voluntary
action and is the product of the person's own attempts to seek
goals.

The higher-level definitions of behavior in HPCT aren't
meaningful until you translate the terms into real experiences.
For example, in your interactions with government types, you have
probably seen that many of them state "facts" about human nature
-- what "people" are like, what to expect of them, and so on.
These are system concepts as I think of them. You probably also
hear many people stating generalizations; not specific programs
for actions, but _principles_ of action. In government they are
often called policies, where the program-level stuff consists of
rules, regulations, or laws stated in if-then terms, designed to
suit an overall policy at the principle level such as equal
treatment, fair pay for adequate work, loyalty, and so forth.

The concept of PCT itself is a system concept. It is composed of
principles like control of input and resistance to disturbance,
which describe no particular control system but are meant to
apply to all control systems however they are designed. At the
program level control becomes a mathematical-logical model
containing specific quantitative relationships, no one operation
being a control system or accomplishing control of input in
itself. The "emergence" of control from the quantitative
relationships among parts of control systems is evident only to a
higher level of perception, the principle level at which we
perceive the principles of control. And from these emergent
principles, once we can perceive enough of them, there emerges
the concept -- the yet-higher-level perception -- of an
autonomous self-organizing hierarchy that constitutes human
nature itself: a system concept.

Once you start translating from the too-formal terms of HPCT, you
can begin to see the phenomena to which these terms were intended
to point. Everyone has principles and system concepts. All you
have to do to believe this is to sit in a blue-collar lunchroom
day after day and listen in on the conversations. Listen to the
people talking about union problems, about work rules, about
unfair treatment given to one person or another. Ask their advice
on how to get along in the company and you will be drenched in
principles. Ask how they think the company should be organized,
and you'll get clear statements of system concepts, not to
mention lots of descriptions of errors at that level (all of
which, I must admit, makes me wonder about the relationship of
language to these levels -- how can such things be described?
Some aspects of language must surely operate at the system
concept level, too -- or higher).

Behind the simple terms in my proposed levels there are phenomana
that I think are quite real and observable, at all the levels.

ยทยทยท

---------------------------------------------------------------
Ken Hacker (930509) --

What we DO have to choose is either Marken's hypothesis -- that
social interaction is unimportant to human behavior, or my
hypothesis -- that social interaction is essential to human
behavior.

Well, if Rick wants to play up a stereotype, I suppose he
deserves to be misinterpreted that way. Actually, however, I
doubt strongly whether Rick considers social interaction
unimportant in human behavior. Rick's argument is along a
different dimension from the one in which you're perceiving it.

The basic argument is one that McPhail and Tucker (and I suppose,
McClelland) have noted in sociology. There are many sociologists
who believe that social laws are completely independent of laws
of individual behavior -- that they are superordinate to the
properties of individuals. These sociologists argue vehemently
against "psychologizing" sociology -- just as traditional
economists do against psychologizing economics. While I don't
think you are doing that with communication theory, some of your
arguments could easily be taken that way.

Many arguments on CSGnet, including some arguments with Bruce
Nevin regarding linguistics, are really about this territorial
dispute. Is there a way to study JUST social interactions, in
such a way that the properties of individuals simply don't come
into it? Well, obviously there is, but the real question is
whether this is the best way to go about it in the long run. My
own feeling is that "pure" social studies are impossible, because
they inevitably involve assumptions about human nature that are
grounded in individual characteristics.

I think that Bruce Nevin's argument is that each individual comes
into a world that is already structured socially, and that those
preexisting social conditions determine what is available for the
person to learn. This is incontrovertible. But the next step is
the one to which I object, which is to treat these social
conditions as if they were independent of the individuals in the
society; as if they were causes of behavior. There's some tricky
reasoning in here, because it's true that no individual can have
much effect on the existing social conditions (barring an unusual
access to publicity or a position of unusual power). However, the
underlying question is whether we are to treat social conditions
IN ANY WAY as having a life of their own, or (as I would prefer)
as continually being produced by the members of the society. And
I would dispute any claim that social conditions are
DETERMINISTIC with regard to any individual. I claim that the
society can present conditions and contingencies to the
individual, but can't determine or even much influence how that
individual will find ways of controlling the local environment to
that individual's satisfaction. There are simply too many degrees
of freedom in individual behavior for other people to be able to
constrain an individual to just one way of behaving.

I think that those whose primary interests have to do with social
phenomena tend to give too much weight to social influences on
the individual. Individuals have many interests in addition to
those having to do with other people, perhaps more other
interests than the average social scientist has. A physicist, for
example, probably spends as much time being concerned with non-
living matter as with people. Every person with a non-social
skill such as carpentry or animal husbandry or road-grading or
electronic design devotes large portions of every day interacting
with non-living aspects of nature and solving problems that
require no other person's direct participation. This is not to
say that social factors don't enter -- the road-grader itself is
obviously a product of a social enterprise -- but only that
interactions with other people do not occupy every person's
attention to the same degree. Nor is there much in social
influences that can help an individual deal with the everyday
problems of coping with disturbances in the non-living world like
a new circuit that doesn't do what it's supposed to do.

What I hope for is a way of deriving social laws from the
interactions of individuals modeled as control systems. Modeling
is basically about interactions: what happens when you connect
components with individually-known properties into a system that
makes the inputs and outputs of one component part of the inputs
and outputs of other components. When you think about it, what
else can there be in social interactions? Each person's actions
affect the world in which many other persons live. All the
actions of other persons in the vicinity affect each individual
to some extent. Each person lives in an environment that has been
shaped to suite the purposes of others, that has properties given
to it by others. Yet those properties were invented by someone,
for some individual purpose. The possible modes of direct and
indirect interaction among individuals are multitudinous, yet in
the final analysis they reflect what individuals are doing to
shape their own worlds nearer to their hearts' private desires.

I see no conflict between control theory and social theories of
any kind. PCT offers a bridge between theories of individuals and
theories of society. At present that bridge is only a sketch on
the back of an envelope, but I think it points to a merging of
the sciences of social and individual behavior.

I think that in speaking of "socially-derived and socially-
applied reference variables" you are imagining that society is a
monolith, in which only one choice is open to the individual. In
fact every person grows up faced with a wide choice of reference
levels for the kinds of experiences that occur, including not
only different reference levels but contradictory ones. The most
difficult part of growing up is the realization that the world
pulls in many different directions at the same time, and that
there is no source of guidance for which of all possible lives to
seek, how to act with other people, how to speak, what kinds of
manners to adopt, which demagogue to serve, which attitude toward
sex to adopt, what brand of politics to espouse -- or ignore.
Even if there are choices of guidance -- role models, teachers,
helpers in general -- one is still left with the ultimate
decision of whose advice or help to accept. And there is only one
person who can make that choice, and only one set of criteria in
one place that applies. That is why individuals do not turn out
all alike. Some swear in church and some do not. Some take the
high road, and many the low. Some care enormously about what the
people around them think; others do not care sufficiently to
avoid trouble.

There is no standard person influenced in a standard way by a
standard society. The statistical approach to behavior is
designed to give that impression, but it is false.
------------------------------------------------------------
Best,

Bill P.