Scribbling chimps

[From Bill Powers (920923.1500)]

Greg Williams (920923-2) --

In your post to Chuck Tucker you said something that brought this
subject into focus:

Do you mean to say that no one actually will sign his/her name in
rubber-banding if presented with certain kinds of disturbances by
another? This is an instance of what I call "purposive >influencing."

Clearly you can cause another -- even a chimpanzee -- to write any
name in the rubber-band situation by applying a suitable pattern of
disturbances. I had thought at one time how much fun it would be to
stand before a plenary session at the annual meeting of the AAAS,
invite a chimpanzee onto the stage, and say "Will the mystery guest
sign in, please?"

What a furor! Is this chimpanzee actually writing his name? Can a
chimpanzee write? A behaviorist would have a hard time explaining why
this chimpanzee does NOT know how to write his name. But a control
theorist could do so easily. The agent writing the name is not the
chimpanzee, but the human being operating the other rubber band.

Does this represent an "influence on the chimpanzee's behavior?" In
once sense, certainly. But it is an influence on _how the chimpanzee
behaves_? No, it is not. The chimpanzee has acquired no new way of
behaving. All it knew how to do before, and all it knows now, is how
to keep the knot in the rubber bands over a target, as a way of
getting an M&M later. All right, 100 M&Ms; who am I to argue with a
chimpanzee's lawyer?

When I talk about "changing" people's behavior, I automatically think
in terms of doing something so that a person will then know how to do
something new, or do something differently from then on. I don't count
manipulating people's actions by using disturbances as changing their
behavior because as soon as I'm not there to supply the carefully-
chosen and patterned disturbance, that behavior will simply stop. It
never had any reason to exist in the first place except to protect
some controlled variable from being changed. If the behavior were
continued after the disturbance ceased, it would CAUSE an error. There
is absolutely no reason for the "manipulee" to continue that behavior,
and every reason not to do so, when the disturbance is removed.

This is what I mean by "coping." Sure, people adjust their actions
because of effects from other people and from the nonliving
environment. But that just prevents their hierarchies from being
disturbed at the level where the disturbance has its first effect on a
controlled perception. They can interact with others 10,000 times a
day, and they will have learned nothing new. They will still know only
how to control their perceptions in the same ways they knew yesterday.
When all the disturbances go away, the people will be exactly as they
were before. Nothing of any importance has changed.

Why has nothing changed? Because the disturbance-resisting behaviors
have been successful. The whole reason for the existence of these
control actions is to prevent intrinsic error; the current
organization is the latest product of a series of reorganizations
driven by intrinsic error. As long as these control systems continue
to work successfully, protecting all controlled perceptions from
significant deviations from their reference levels, and as long as
maintaining those perceptions at those reference levels is sufficient
to prevent intrinsic error, there will be no further reorganization.
So nothing will change.

This is why I dismiss all the "purposive influences" you bring up as
being trivial. All they illustrate is how control systems with fixed
organizations will act when their actions affect other people and
other people's actions affect them. Out of these fixed organizations
there can arise detailed behavior of immense complexity, in the same
way that all those "people" in the crowd program exhibit complex
patterns of interaction -- while behaving in a simple and unchanged

So it seems clear to me that if people are to be changed in any
meaningful way by interactions with each other, those interactions
have to amount to something more than simply mutually disturbing each
other, or messing about with each other's environments to alter the
"schedule of reinforcements."

Without reorganization, all you will ever see in human interactions is
variations on a theme as people with static organizations play out all
the possible ways of disturbing each other and trying to maintain
control at the same time. Those variations lost their fascination for
me a long time ago, when I began to see them repeating. For most
people, their lives are a replay of their parents' lives, either
symmetrically or antisymmetrically. You hate your parents and want to
be totally different from them? You love your parents and want to be
just like them? What a bore.

Let's talk about reorganization. That's the only area where anything
interesting happens in human behavior.



Bill P.