Side effects, controlling restriction

[From Rick Marken (961003.1330)]

Me:

What do you mean by "genuinely social & cultural dynamics"?

Bill Benzon (961002) --

Hmmm....might not be too different from what you mean by "side effect" in
this context. When a group splits in two for whatever reason, is that a
side-effect? When a group creates a hierarchical social control structure (A
reports to B, who reports to C, who gives orders to B, etc.), is that a side-
effect? Is a conversation a side-effect?

I use the term "side-effect" to refer to observable aspects of control system
behavior that are not under control. For example, movement of the handle in a
tracking task is not under control; it varies as necessary to counter
disturbances to the controlled variable. You might observe that the handle
movements trace out an interesting pattern over time. For example, the
handle movements might trace out the formula E=mc2. This would certainly be
an interesting observation but it is an irrelevant side effect of control.

When a group of people splits in two it may or may not be a side effect. It
is a side-effect if half the group members wants the Pizza at the store on
one side of the mall and the other half want the Chinese at the Panda
Express on the other; the split is a side effect of controlling for food. It
is not a side effect if each member of the group wants the split ("OK, let's
divide into two groups"; in that case, being split is a controlled result for
everyone.

I find it hard to imagine a situation where a hierarchical social control
structure would be a side effect. When such structures exist it seems to me
that they are formed only when each member of the structure agrees to "play
along" -- ie. one member controls for reporting to B based on reports from A,
another controls for reporting to C based on B, etc .

Some aspects of a conversation are side effects and some are controlled. The
loudness of the speakers voices is probably a side effect of their efforts to
control for hearing themselves (over each other when necessary and over the
prevailing background noise). The sequence of talking is probably controlled,
each speaker waiting for a pause from the other before initiating talk.

So my earlier blanket statement that social phenomena are just a side effect
of the controlling done by individuals is obviously wrong. Some social
phenomena are side effects of control and some are intended results of
control. But you can't tell which are side effects and which are intended
results by just looking (as in the case of the people separating into two
groups at the mall). You have to Test (using the ol' Test for Controlled
Variables) to detemine whether a social variable (one whose state is
determined by more than one person) is under control.

Bruce Abbott (961003.1325 EST) --

I am really sorry to hear that you suffer from asthma. I know from my wife's
experience that even a mild case (as she has) can be terribly frightening and
uncomfortable.

The object, of course, was to develop a training program through which
asthmatics could learn to recognize the early changes in lung-state that
precede an asthma attack, so they could take action abort or minimize the
attack before it got worse.

Since the goal is to recognize early changes in lung state, why not have the
training involve tracking continuous changes in restriction? The "correct"/
"incorrect" approach might help a person develop a perceptual function that
allows discrimination of signal from noise when the signal is near threshold.
But it seems that it would really help if a person could monitor
continuous _changes_ in the perceptual variable that is relevant to
determining lung state.

Why not set up the training so the the restriction is varied continuously
(and slowly) over some "ecologically valid" range (a range that includes the
restriction levels encountered "normally" and during much of an asthema
attack)? Let the subject continuously indicate the perceived level of
restriction using handle position or something like that. Then provide a
continuous indication of the difference ("error";-)) between handle position
and actual restriction level. The subject would be asked to keep the measure
of difference at zero. To do this, the subject would still have to learn to
perceive in a way that correlates strongly with the actual level of
restriction in the tube. But I think that what the subject learns to perceive
in order to control for a continuous match of one handle position to tube
restriction would be more useful in the control of asthma than what he learns
in order to control for the binary "correct"/"incorrect".

Best

Rick

[From Bill Powers (961003.1500 MDT)]

Rick Marken (961003.1330) --

Methinks Rick is having a lucid day. Nice post.

For example, the handle movements might trace out the formula E=mc2. This
would certainly be an interesting observation but it is an irrelevant side
effect of control.

For those unacquainted with this phemonenon, I have thought of an
elaboration on my "Professor Challenger" fantasy (of Conan Doyle's "The Lost
World," in which the Professor astounded the Royal Society by bringing a
live dinosaur on stage). In my fantasy I am speaking before a plenary
session of the AAAS annual meeting, on the subject of "Irrelevant
side-effects of control behavior." At a suitable point in the presentation,
I announce that I have a mystery guest who will now come on stage and sign
in on the blackboard. The mystery guest turns out to be an amiable
chimpanzee. In this elaboration, I have a black box mounted near the
blackboard with a long rubber band drooping out of a hole in the side of the
box pointing to the right next to the blackboard. Attached to the the end of
the rubber band is a chunk of chalk.

On my signal, the chimpanzee waddles over to the black box and picks up the
piece of chalk. He places the chalk against the blackboard where I point,
and I press a large red button on the box. The chimpanzee proceeds to write
on the blackboard in shaky but legible script,
"MyNameIsAlbertAndE=MCSquared." Shocked silence, followed by applause that
rises to a hysterical roar. I bow and, with Albert, leave the stage, both of
us smiling cryptically. A year later, the Nobel committee creates the first
prize for behavioral science and awards it to Albert.

Of course inside the black box is a two-dimensional computer-controlled
servomechanism that is moving the other end of the rubber band to write,
upside down and backward, what Albert just wrote on the blackboard. What
Albert was doing was keeping a small mark on the rubber band centered in the
hole in the box. And what Albert wrote on the blackboard was an irrelevant
side-effect of controlling the position of the mark.

···

------------------------------

When a group of people splits in two it may or may not be a side effect. It
is a side-effect if half the group members wants the Pizza at the store on
Express on the other; the split is a side effect of controlling for food. It
is not a side effect if each member of the group wants the split ("OK, let's
divide into two groups"; in that case, being split is a controlled result
for everyone.

Crystal clear. A socially controlled variable must be explicitly controlled
by each person involved. If there is to be "cooperation," an exclusively
social variable, all those involved must be controlling for cooperation.

A crowd of people streaming toward the exit of a burning theater shares a
common goal and a common perception, but there is no social control.
---------------------------------
RE: Bruce's regrettable asthma problem

Why not set up the training so the the restriction is varied continuously
(and slowly) over some "ecologically valid" range (a range that includes the
restriction levels encountered "normally" and during much of an asthema
attack)? Let the subject continuously indicate the perceived level of
restriction using handle position or something like that. Then provide a
continuous indication of the difference ("error";-)) between handle position
and actual restriction level. The subject would be asked to keep the measure
of difference at zero. To do this, the subject would still have to learn to
perceive in a way that correlates strongly with the actual level of
restriction in the tube.

Not quite, I think. The subject could do this task without sensing the
degree of constriction: just keep the visual error at zero.

Another way would be to give the subject a handle that causes a
constriction. The task would be to keep the degree of constriction constant
at several different values (one at a time) while an independent disturbance
is added to the effects of the handle on the actual constriction. This
requires continuous sensing of the effect of the constriction, and
furthermore allows estimating the accuracy of the perception.

Best,

Bill P.

RM:

When a group of people splits in two it may or may not be a side effect. It
is a side-effect if half the group members wants the Pizza at the store on
Express on the other; the split is a side effect of controlling for food. It
is not a side effect if each member of the group wants the split ("OK, let's
divide into two groups"; in that case, being split is a controlled result
for everyone.

BP:

Crystal clear. A socially controlled variable must be explicitly controlled
by each person involved. If there is to be "cooperation," an exclusively
social variable, all those involved must be controlling for cooperation.

And if we look at the right place in the brain we should find the tissue
"concerned" about cooperation. (The late) David Hays & I would guess that
tissue is in the right front cortex & would expect to see high activity
there (in, e.g.a PET scan) in people intensely cooperating.

···

********************************************************
William L. Benzon 518.272.4733
161 2nd Street bbenzon@global2000.net
Troy, NY 12180 Account Suspended
USA
********************************************************
What color would you be if you didn't know what you was?
That's what color I am.
********************************************************

RM:

When a group of people splits in two it may or may not be a side effect. It
is a side-effect if half the group members wants the Pizza at the store on
Express on the other; the split is a side effect of controlling for food. It
is not a side effect if each member of the group wants the split ("OK, let's
divide into two groups"; in that case, being split is a controlled result
for everyone.

BP:

Crystal clear. A socially controlled variable must be explicitly controlled
by each person involved. If there is to be "cooperation," an exclusively
social variable, all those involved must be controlling for cooperation.

And further....As I've said, research indicates that when groups reach a
certain size, they tend to split up. Thus, regarding the size of villages:

     "There is a magic number here, and the number according to
Robert L. Carneiro* is 200. If a village grows beyond that
number, the interpersonal tensions lead to a split. Some fami-
lies move away and start another village. Flannery ( Po12BIBL* )
reports that Tiv compounds usually include about 20 persons, and
rarely reach 80. "

(From David Hays, The Evolution of Technology Through Four Cognitive Ranks,
5.2.1, URL: Account Suspended)

Something about the number of people in relation to the existing social
"technology" leads to interpersonal tensions. Given those tensions, lots of
folks form the purpose of splitting the group. PCT has something to say
about what happens once those purposes are formulated. But it is going to
take another body of thought to account for just why those particular
purposes arise when the group reaches a certain size. The urge to split
may well be a side-effect of individual actions -- when the group gets too
large, it becomes difficult for people to satisfy some important reference
levels, but calling it a side effect isn't very usefull.

If you will, in this particular case, the fact that the cursor traces E =
mc^2 IS what we have to explain. If the explanation turns out to be that
there is a "hidden hand" (to invoke a famous metaphor), then the task of
the sociologist is to find the little man behind the curtain who is the
possessor of that hidden hand. However, as far as I can tell, most
sociologists don't believe in (conspiracies) of little men behind curtains.
So, coming up with the mechanism is a bit harder than simply unmasking the
conspirators.

···

********************************************************
William L. Benzon 518.272.4733
161 2nd Street bbenzon@global2000.net
Troy, NY 12180 Account Suspended
USA
********************************************************
What color would you be if you didn't know what you was?
That's what color I am.
********************************************************