Test and conflict [was: Shared references]

[From Bruce Nevin (2004.07.14 18:10 EDT)]

Bill Powers (2004.07.14.1442 MDT)–

In the test, the experimenter does not try to
alter the controlled variable. So there is actually no conflict at all

If I want to test your control of hand
position, I apply a force to your hand that would be enough to move your
hand measurably if your muscles didn’t change their tension. If you are
controlling hand position, your muscles will produce a force sufficient
to balance out the force I am applying, and your hand position will
remain very nearly constant.

If your hand does not move[…], I do not
increase the force I am applying in order to achieve a goal of seeing it
move. I have successfully applied the disturbance; the applied force is
what I intend it to be. I take the failure to move as evidence that the
hand position is under control, pending the remainder of the Test. There
is no conflict, because I am controlling force (the state of the
disturbing variable) and the hypothesis is that you are controlling
position.

You are controlling force. You are also applying force. How are these two
purposes related?
Let us say for simplicity that you are applying this force with your
hand. While you are applying a force to my hand that would be enough to
move my hand measurably if my muscles didn’t change their tension, my
muscles do in fact change their tension so as to apply an opposing force
to your hand which would be enough to move your hand measurably if your
muscles didn’t sustain sufficient tension to apply the force to my hand
that I am resisting. At that moment, we are in conflict.
The fact that you do not follow through to sustain the conflict is due to
control of force at a higher level. Given that your purpose is to try to
prove that your guess about my control of perceived hand position is
wrong (as you have often said), your immediate purpose in applying that
force must be to move my hand from its position. However, your purpose
next above that is to allow my behavior to show that your guess is
correct, if in fact I am controlling perceived hand
position as you guess. This sets the reference at the lower level where
you are applying the force to my hand, limiting the force that you apply,
and preventing the lower level of control from increasing the force so as
to achieve its purpose of moving my hand despite my resistance. At this
higher level, you want to perceive my resistance to the disturbing force
applied at the lower level, so when you have perceived my resistance
sufficiently this level of control sets the reference for force to zero.

I may be wrong to suppose that these are on two levels. Another
possibility is that they are two purposes at the same level, with a
purpose at a higher level of deciding (observing) which is correct. I
don’t see how it would work as internal conflict, however.

That’s my attempt. How would you analyze what is going on when a Tester
applies a disturbing influence?

    /Bruce

Nevin

···

At 02:58 PM 7/14/2004 -0600, Bill Powers wrote:

[From Bill Powers (2004.07.14.1849 MDT)]

Bruce Nevin (2004.07.14 18:10
EDT)–

Let us say for simplicity that you
are applying this force with your hand. While you are applying a force to
my hand that would be enough to move my hand measurably if my muscles
didn’t change their tension, my muscles do in fact change their tension
so as to apply an opposing force to your hand which would be enough to
move your hand measurably if your muscles didn’t sustain sufficient
tension to apply the force to my hand that I am resisting. At that
moment, we are in conflict.

No. At the meeting I will show you the difference between force control
and position control. There is no conflict. If I am controlling the force
my hand is applying (as I sense it), I can’t also control the position of
my hand. If I am controlling the position of my hand, I can’t also
control the force I apply.

The fact that you do not follow
through to sustain the conflict is due to control of force at a higher
level. Given that your purpose is to try to prove that your guess about
my control of perceived hand position is wrong (as you have often said),
your immediate purpose in applying that force must be to move my hand
from its position.

No. Stop guessing, Bruce. My intention is to apply a known force to your
hand (however it moves or does not move) so I can measure the resulting
deflection, if any. I can apply such a force myself, since I know how to
apply a popsition-independent force, or I can tie a weight to a string
and hang the weight over a pulley, the other end of the string being tied
to your wrist to apply a known position-independent disturbing force. The
result of that measurement, plugged into the control system model, will
give information about control properties related to your hand position.
This is a matter of applying various models, mainly physical models and
control-system models. It has been a long time since I felt the need to
try to guess the outcome of a scientific procedure so I could claim to
have been right before I could possibly have known whether I was right or
wrong. I am willing to put any proposal to a test and accept the results
if they are clear. Why else bother to concoct a theory and a
methodology?

Well, I see that your next paragraph is a direct denial of what I just
said.

However, your purpose next
above that is to allow my behavior to show that your guess is correct, if
in fact I am controlling perceived hand position as you
guess.

The Test is designed to show whether there is control or not,
particularly not, regardless of what the person doing the test has
guessed. There is no point trying to prove anything before doing the
test. The hypothesis being tested is the only guess, and it is neither
wrong nor right until it’s tested. If you care whether it is wrong or
right, instead of caring about testing it properly, you will inevitably
make “mistakes” in favor of or against your guess (depending on
how you want the results to come out). Make enough mistakes of that kind
and you will no longer have a reputation to uphold.

That’s my attempt. How would you
analyze what is going on when a Tester applies a disturbing
influence?

I did that above. The tester forms a guess about what the
controlled variable might be. He applies a known disturbance in
such a way that if there is no control, the effect of the disturbing
variable on the proposed controlled variable can be calculated. If the
observed effect is significantly less than the calculated effect, there
is the possibility that a control system is acting. The characteristics
of that system can be worked out on the basis of a control model proposed
to fit the situation.

The disturbance that is applied is not varied according to the size of
its effect. There is no goal for the amount of effect that is to be
observed; whatever the amount, even if zero, it provides information
about the kind and degree of control as well as the reference level that
the model should be given.

There should be no conflict if you are doing the test right.

Best,

Bill P.

[From RIck Marken (2004.07.14.2020)]

Bill Powers (2004.07.14.1849 MDT)--

There should be no conflict if you are doing the test right.

Bruce N. could demonstrate this to himself by doing the demo of The Test for the Controlled Variable at
http://www.mindreadings.com/ControlDemo/ThreeTrack.html. There is absolutely no conflict between yourself (the subject) and the computer (the experimenter) in this demo. This lack of conflict is demonstrated by the fact that it is easy to move the selected square to any point in the display area. You (the subject) are controlling, as usual, producing intended results (movements of the selected square) in a disturbance prone environment. The fact that the disturbances are produce by an experimenter (the computer in this case) rather than by the wind, say, does not turn the situation into a conflict.

Best regards

Rick

···

---
Richard S. Marken
marken@mindreadings.com
Home 310 474-0313
Cell 310 729-1400

[From Bill Powers (2004.07.15.0415 MDT)]

Rick Marken (2004.07.14.2020) --

You (the subject) are controlling, as usual, producing intended results
(movements of the selected square) in a disturbance prone environment. The
fact that the disturbances are produce by an experimenter (the computer in
this case) rather than by the wind, say, does not turn the situation into
a conflict.

There. You have said the words I couldn't find. It doesn't matter how the
disturbance is applied as long as it's applied. The control system goes
right on resisting disturbances -- successfully -- as usual. There is no
conflict.

Best,

Bill P.