[From Bill Powers (920904.0700)]
Greg Williams (920903 - 2) --
The problem with reading all those quotes from Skinner (for which
labor many thanks!) is that one mistakenly tends to put meanings into
his words that make common sense (and in our case that jibe with PCT).
Skinner arranged his choices of words this way deliberately. Common
sense generally sides with control theory. But it is possible to talk
what is heard as good control theory in words that are defined by the
speaker in absolute contradiction with control theory. This makes it
seem that the speaker was ahead of his time, when in fact he was
crusading against the purposive overtones of common-sense descriptions
in part by trying to pre-empt the language.
He said:
I have used technical terms in making a technical point. I have
preferred a technical term elsewhere when it could be used at no
great >cost.... But elsewhere [in this book] I have freely used the
lay >vocabulary while accepting the responsibility of providing a
technical >translation upon demand....
An example of this:
In traditional terms, one person arranges positive or negative
contingencies in order to create interests, provide encouragement,
instill incentives or purposes, or raise consciousness in another
person. In doing so, he brings him under control of various features
of >his environment.
You would think from this that people can have interests, feel
encouragement, seek incentives and have purposes, and be conscious,
wouldn't you? But if you asked Skinner what he meant by each of these
terms, he would describe them "technically" in terms of the
environmental causes of the behaviors from which we infer these inner
states. In other words, he did not believe that these inner states
were causal in any sense. At best they were epiphenomena. In the above
quote, he links his technical terms, positive and negative
contingencies, with common-sense terms, making it seem that the one
flows naturally into the other. This can seduce the reader into
believing that interests and purposes are the sort of thing that
contingencies produce. And the swift passage across "positive or
negative" simply asserts a characteristic of contingencies and hurries
on to the main point, having left these immensely important little
adjectives to be taken for granted.
Also, although Skinner often uses the term "control" in contexts where
we would read it as "negative feedback goal-directed control of
perception," the fact that he always, in the end, referred the control
back to the environment shows that he could not have meant the term in
that way. The environment is simply a collection of stimuli that have
reinforcing properties and which have the capacity to control behavior
-- which we know is not possible for anything but a purposive
environment.
We must also recognize that Skinner consciously bypassed the problem
of variable behavior producing consistent results. He did this by
EXPLICITLY defining behavior -- responses -- as classes of outcomes,
not outputs. This means that when he spoke of external agents or
things controlling _behavior_, he had to mean controlling outcomes,
not actions. We know from control theory that this is precisely what
external agencies can't control, except by winning conflicts. While
Skinner couldn't see the contradiction here, we can.
Here's another problem:
A physical world generates both physical action and the physical
conditions within the body to which a person responds when a verbal
community arranges the necessary contingencies.
While he never says so out loud, he is implying that the physical
world generates _specific_ physical actions and conditions, those
particular effects that we observe as systematic behavior. This
strange directedness of physical effects is handled in the same way he
handles the problem of variable actions producing consistent effects.
He defines a stimulus not as a specific physical situation or energy
input, but as a class of situations defined by their effect on
behavior! So the actual physical situation can vary widely, but if the
same behavior persists, by Skinner's definition the "same stimulus"
exists.
Yet Skinner always spoke as if he were simply applying basic
principles of "science" to an analysis of behavior. He blithely
invoked mysterious nonphysical properties of the environment to make
this view sound plausible. All of the critical questions were begged.
But he was so clever at the use of language that he could hide these
gaps in his argument and make it seem that he was simply applying
logic in an impersonal and rigorous way.
Another thing that Skinner did very well was to slip in non-sequiturs
that easily passed without notice:
In self-management the controlling self is different from the
controlled. But all selves are the products of genetic and
environmental histories.
Wait a minute: even if all selves are products of genetic and
environmental histories (which applies, of course, to everything in
the universe, not just selves), what does this have to do with
controlling selves and controlled selves? Control systems are products
of genetic and/or environmental histories, but they still control.
This is like dismissing behavioral theories by saying that they all
come down to physics and chemistry in the end, so why bother? It makes
a hell of a difference WHAT the physics is and WHAT the chemistry is.
In fact that makes ALL the difference.
If Skinner had not dismissed a controlling self a priori, he might
have come across the concept of hierarchies of organization. But he
did not. Instead he made it seem as if an irrelevant statement was a
refutation of the concept of a controlling self, by prefacing it with
"But". He got away with this a lot.
Skinner was also dishonest:
It would be absurd for the behaviorist to contend that he is in any
way >exempt from his analysis. He cannot step out of the causal stream
and >observe behavior from some special point of vantage...
But Skinner constantly did this. He knew something about the rest of
us that we did not know. There was no question of his view of behavior
being simply a product of random environmental influences; if he had
really believed that, he could not have argued against anyone else's
view because that would be like saying the other person should have
been exposed to different contingencies.
Skinner's view admits of no concept of truth or correctness of an
idea. Yet Skinner always spoke from conviction, and he did not
hesitate to criticize other views. Trying to sound nice, he advocated
manipulating contingencies so that the world would be a better place,
but he never admitted that non-scientists could pick their own
definitions of a better place, and it never occurred to him that his
idea of a better place might seem like Hell to someone else. There was
always a layer of attitudes in Skinner that remained a secret, mainly
from himself. Perhaps it is unfair to call him dishonest. But if he
was not dishonest, he was singularly unaware of his own assumptions.
Much of what Skinner criticized about other approaches to behavior
theory was worth criticizing. Trait psychology and statistical
explanation were correctly assessed by him, in my view. But Skinner
often used the obvious wrongness of other approaches as a way of
making his seem right by contrast. It just doesn't follow that if you
see what's wrong with the other fellow's idea, your ideas are
automatically right. But that's how Skinner used these arguments.
ยทยทยท
----------------------------------------------------------------------
Greg Williams (920903 - 3) --
1. Is controlling a person possible with THREAT of force (which might
be included in what you mean by "direct use of force")?
Yes, I think so. It doesn't even have to be an offered threat; it can
simply be a perceived threat. A protester might see cops in the
vicinity, and think "If I do what I was planning to do, they'll throw
me in jail. So I guess I'll do something a little less spectacular."
So you could say (imprecisely) that the cops, just by being visible,
have controlled the protester's choice of goals (higher-level
behavior). You could say this even if the cops were really on their
way to some other location and knew nothing of this protester. It's
the perceived threat that matters. If the cops threaten to arrest the
protester, but the protester doesn't believe they actually will, the
threat has no effect. Until it's carried out, of course.
2. Suppose I yell "Fire!" when we're both in a theater. You (being
prudent) join the crowd heading for the exit. Were you, BOTH before
and >after I yelled, ACTIVELY controlling for not getting burned (or
some >such)?
What does "actively" controlling mean when there's no disturbance? If
you ask whether my reference level for being burned was zero before
and after you yelled, the answer is yes. If you ask whether the
pathways in the associated control systems existed, again the answer
is yes with some reservations because of the possibility of switching
connections. If you ask whether I was actively countering a perceived
or imagined disturbance before you yelled, the answer is "No."
Did your yelling "FIRE" affect my behavior? Certainly; I took it for
an honest report (this time). I was avoiding getting burned, or so I
thought. Did your yelling "FIRE" control my behavior? Well, to know
that we'd have to have some more details. Suppose I was one of those
who didn't start toward the exit. How would you alter your yelling of
"FIRE" to counteract that error between what I did and what you wanted
me to do? If you did try something else, and it worked, then you would
have been controlling my behavior. But even then, you couldn't control
the _outcome_ I was controlling for by varying my position. That is,
you couldn't make me want or not want to get burned by yelling "FIRE".
I still think you're confusing "control" with "affect" in many of your
examples. You can't tell whether control is present by looking at a
snapshot. You can't distinguish control from mere influence by looking
at apparent cause-effect pairs. You can't identify control by looking
at a single action. Control is a process that takes time and involves
relationships that persist through time.
My problem is with your downplaying the interactivity and,
especially, >significance of manipulation. You say that the manipulee
"ignores" the >manipulator. But this is simply not true in general,
even though it >might be true sometimes. Complex interactions are the
rule, not the >exception, between con men and marks, for example.
There is a complex >"dance" going on between influencer and influencee
so as to maintain >the controlling relationship. In many such
"dances," control can even >go BOTH ways.
While all control of one person by another involves interactions, not
all interactions between people (however complex) involve control. Two
people dancing in the opening of a narrow door, trying to get through
in opposite directions, are not controlling each other; they're simply
trying to get through the door. Their control systems, taken as a
single system, are oscillating. If one of them is deliberately and
systematically stepping in front of the other to prevent passage, the
other will realize that after a time and take action to keep from
being controlled: a vigorous punch in the snoot, for example, would
clear the way if the other guy is littler than you.
Furthermore, your claim that manipulating is "just diddling" seems
purely ideological, in the sense that it ignores the potential
effects >of some examples of manipulation (for ill OR good) on the
manipulee.
I don't ignore that at all. If the effect of manipulating output
begins to matter to the manipulee -- if it disturbs something that IS
under control -- the manipulee will push back, and the situation will
be converted from manipulation into conflict.
It seems to me that you are going to extremes to defend an extra-PCT,
imported notion of extreme autonomy and anti-environmentalism.
Hey, I like spotted owls a whole lot better than I like people who go
around cutting down beautiful old forests.
And anyway, what's "extreme" about autonomy? When you track down the
causes of behavior, you don't end up back in the environment, but at
the highest levels, and if you try to backtrack farther than that, you
end up with the whole species' struggles to stay alive and in control,
over the last 3 billion years. That sounds perfectly reasonable to me.
Conservative, even.
PCT itself says that one's actions (output variables) are conjointly
determined by the person's reference signals AND disturbances. It
does >NOT say that either the reference signals or the disturbances
are to be >construed as in some way more important or significant in
producing the >actions.
Oh, but the reference signals are far more important than all the
disturbances put together, because they define what constitutes a
disturbance. It doesn't work the other way around. Anyway, what
matters to the organism is not its action, but the outcome, and when
there's no immovable obstacle or irresistable force, the reference
signal determines the outcome quite independently of the environment.
The environment has very little to say about the outcomes that matter
to the organism. That's what the actions are for: to make sure it
doesn't.
You are sounding like Skinner turned upsidedown -- he (wrongly, I
think) overemphasized environmental determination, while you are
overemphasizing organismic determination.
Skinner thought that environmental influence was determination; I'm
trying to avoid that mistake. Even the effect of disturbances on
actions is not deterministic; the reference signal can change, too,
which means that in the long run nobody can predict the effect of the
disturbance. Particularly not if there are other disturbances too, as
there usually are. The only deterministic aspect of the organism-
environment interaction is the effect of the reference signal in
determining what the input variable's state will be within narrow
limits, and that holds true only when external interference isn't too
radical.
True, one's INANIMATE environment doesn't control one's variables (of
ANY sort), but one's ANIMATE environment can control some of one's
variables IN WAYS WHICH RESULT IN IMPORTANT CONSEQUENCES TO THE
CONTROLEE.
But so can inanimate variables have important consequences. If you're
caught in an avalanche, that will have important consequences. The
severity of the effect isn't what makes the difference. It's whether
the effect is intended, and whether changes in the effect give rise to
opposing changes in the manipulation that is producing it. Only that
is control. People bump into each other all the time, probably more
often than they bump into the environment. In these encounters there
are many mutual effects. Sometimes there are unforseen effects that
turn out to be extremely inconvenient, like lethal. People adapt their
own modes of control to avoid the worst of these effects -- not
instantly, but over time. Sometimes they don't succeed. Usually they
do. And they make these adaptations to suit their own higher
intentions, not someone else's.
Don't you see that, much (but, I'll grant, not all) of the time,
people >want to control others' ACTIONS, NOT others' DESIRES? That
means that >the slogan "you can't be controlled without overwhelming
physical >force" is, in practice, not such a Big Deal.
I'd say that most of the time, people are controlling for their own
perceptions, and treat the actions of others simply as disturbances to
be counteracted or avoided. I'd say that most people don't even think
about controlling other people's actions -- that is, they don't
systematically vary their own actions in order to keep someone else
producing a given action. The other's action is usually just an
incidental part of what one is doing or wants to happen. When I buy
toothpaste, I hand the clerk my money and expect to get my change, and
the toothpaste, back. I don't care how the clerk does that, or whether
another clerk does it instead. My reference level is for the outcome
that matters to me. It's not for the clerk's behavior.
We are being controlled and are controlling others on a
daily basis without direct physical force -- with, I claim, hugely
significant effects.
We are AFFECTING AND BEING AFFECTED on a daily basis. That's not
control. When General Motors lays off 5000 people, it doesn't do this
in order to put any person out of work. It is laying them off so it
doesn't have to pay them salaries, so it can become leaner and meaner,
with the consequence of course that it is removing buying power from
the hands of its own customers, but businesses in America never look
that far ahead. These are hugely significant effects, but they aren't
control.
There is control involved in such actions, at another level. It's
backed up by a credible threat of overwhelming physical force, called
The Law. A laid-off worker can't show up for work the next day, and at
the end of the week demand a paycheck. There is no way that worker can
continue to get a paycheck from that company -- no way, that is, that
will not end up with physical force being applied to prevent it. The
worker might counter that threat by bringing a gun, but that would be
hopeless, too, in the long run. The force available to The Law is
greater than what any individual can resist or overcome. There are too
many people on the side of the law.
It isn't that I don't believe that control exists in our society. I do
believe it. I believe it is part of the fabric of our society, and all
the others I know of in this last part of the 20th Century. It is true
control. It's not diddling around manipulating people's actions. It
rests solidly on the ability to seize a person and throw him or her in
a cell with bars, and to kill the person in the process if necessary,
in total disregard of that person's attempts to remain free. It's the
kind of control you see on a football field, or in a drive-by
shooting, or in torture of political dissidents, or in shelling a city
or loosing a firestorm of bombs over an army for two weeks.
This is the only kind of social control that I consider serious enough
to worry about. It shows a complete ignorance of human nature. It is
just as if people can't distinguish living systems from rocks.
It never crossed Skinner's mind that his control over his trained
animals depended entirely on his size and strength relative to theirs.
Skinner could reduce a rat's weight to 80 percent of the free-feeding
weight just by seizing the rat, putting it into a cage from which it
could not escape in search of food, and not feeding it. He could force
the rat to press a bar by lifting it up and putting it in another cage
in which the only way to get food was by pressing the bar. The
"control" of the rat's behavior by operant conditioning was an effete
and polite little exercise in comparison with the brute force that
made the experimental procedure possible in the first place. This is
why you hear about operant conditioning being used so often in
prisons, mental hospitals, and "detention centers" in which the
contingencies -- the rules -- can be set up without fear that the
inmates can refuse to live by them. Play the game or it's off to
maximum security conditions, walking or being dragged. This underlying
threat of physical force, made credible by frequent application, makes
a mockery of any attempts to "influence" behavior by more theoretical
means. The only reason any method works in prison is that the inmates
know that if it doesn't work, they will suffer the consequences.
So I am not naive, Greg, about social control. I just want the
conversation to stay focused on the real problem, which comes from not
understanding how people actually work. Our social system needs an
overhaul. We need to stop training people from birth to treat everyone
else as an enemy to be overcome by force. We need to develop a little
respect for the will of others. We need to figure out what people want
and try to give it to them, instead of withholding it to gain an
advantage over them. We need a conception of life together that goes
beyond competition, swaggering, and bullying. The message of control
theory is that yes, you can control other people, but no, you can't
make them hold still for it, and no, you're not going to get what you
really want from life in that way.
I believe passionately in autonomy, not as a recommendation but as a
fact. It is the nature of an organism to be autonomous. And it is the
nature of an organism to control. When the one is invalidated by the
other, we must consider this to be a malfunction, whether it is one
person or a whole society that is destroying its own means of
survival.
---------------------------------------------------------------------
Best,
Bill P.