an anti-valentine, output & error

[From Rick Marken (980223.1320)]

Jeff Vancouver (980223.1430 EST) --

If we were doing brain surgery and could identify a nerve (or set of
nerves) that represented the path from the comparator to the output
function of some ECU, we could stimulate that nerve and watch the
output. Thus, error _could_ be an independent variable.

No Jeff. The error variable is still in the loop. The stimulation
from the electrode is an independent variable, but the error variable
is not. Try simulating this on a computer and see what happens.

Closed loop is cause-effect-cause-effect-cause-effect . . .

No Jeff, it's not. Each variable in the loop is both cause and
effect _at the same time_. Try simulating a sequential state
control system on a computer and see what happens.

We've been over these and other basic point 100s of times in the
last 7 years, Jeff. What's the problem?

Jeff Vancouver (980223.1455 EST) --

PCT is about control, not conventional psychology, so I am not
sure what your "according to PCT" means.

It means that the behavior you see organmisms "producing" is a
side-effect of the process of control of perception. If conventional
psychology is not about control then it is not about the behavior
of living organisms.

There is a lot less of this [testing for controlled variables] than
there should be [these is actually _none_ of it -- RM], but there
are other things they do that help us understand control.

Such as?

I respectfully submit that _you and Bill and others_ see more of
an absence then exists.

Why not fill in the gaps for me? Why not tell me what conventional
psychologists have done to help us understand control? It's not
a secret, is it?

Bill Powers (980223.1313 MST) --

Reading the posts for the past few days has left me with an
overwhelming sense of futility. It is futile even to try to
explain why.

I find it invigorating to watch PCT confirmed again and again,
with every post to CSGNet. The controlled variables are as clear
as bells, aren't they?

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

[From Bruce Gregory (980224.1010 EST)]

Rick Marken (980223.1320)]

Bill Powers (980223.1313 MST) --

> Reading the posts for the past few days has left me with an
> overwhelming sense of futility. It is futile even to try to
> explain why.

I find it invigorating to watch PCT confirmed again and again,
with every post to CSGNet. The controlled variables are as clear
as bells, aren't they?

Would you do me the favor of telling me what variable(s) I am
controlling?

Bruce

[From Rick Marken (980224.0800)]

Bruce Gregory (980224.1010 EST) --

Would you do me the favor of telling me what variable(s) I am
controlling?

You are controlling for understanding PCT, and doing a damn
good job of it. You are certainly not one of those who is
controlling for variables that make our efforts to teach PCT feel
futile -- at least you are not making my efforts to teach PCT
feel futile. I think it's pretty clear who those the people are
who are making our efforts to teach PCT seem futile; it's those
who are defending (in one way or another) the lineal causal view
of behavior. I bet you can spot them very easily -- as easily
as you can spot those who are defending the closed loop control
view of behavior.

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

[from Jeff Vancouver 980224.0900 EST]

[From Rick Marken (980223.1320)]

Jeff Vancouver (980223.1430 EST) --

If we were doing brain surgery and could identify a nerve (or set of
nerves) that represented the path from the comparator to the output
function of some ECU, we could stimulate that nerve and watch the
output. Thus, error _could_ be an independent variable.

No Jeff. The error variable is still in the loop. The stimulation
from the electrode is an independent variable, but the error variable
is not. Try simulating this on a computer and see what happens.

You make a distinction between the manipulation and the effect on the
variable of interest. This is legitimate, but could be make when talking
about disturbances affect on the CV as well. Yet, you rarely go to the
trouble. I am not suggesting you go to the trouble (usually), but that you
not apply a double standard.

Closed loop is cause-effect-cause-effect-cause-effect . . .

No Jeff, it's not. Each variable in the loop is both cause and
effect _at the same time_. Try simulating a sequential state
control system on a computer and see what happens.

I have. It works.

Next you will ask to see it. This too is perfectly legitimate.
Unfortunately, it is embedded in the simulation that I am working on for
research and not easily teased apart. I plan to eventually post this work
to this net, but I am not yet ready.

We've been over these and other basic point 100s of times in the
last 7 years, Jeff. What's the problem?

I think there are two problems. One, the meaning of cause is multiple.
Cook and Campbell talk at length about this in their book. How you use the
word and how conventional psychology uses the word are different. Hence,
this is another semantic problem. One the other hand, many conventional
psychologists appear to not understand the limitations of the meaning they
use, so it is also a conceptual problem.

The other problem in terms of my understanding is that you all have focused
on ECUs with continuously available information about the state of the CV
and continuously available degrees of freedom for affecting the CV. In
those types of situations, the simultaneity of causality is critical. I
tend to deal with controlled perceptions that do not have those properties.
Hence, one can look at the chain of causality because the loop is
"temporarily" broken. I put quotes around temporarily because it is a
problematic word. It would be better to say that lags exist which allow
the measurement of o's and d's and p-primes (external observers' perception
of p, i.e., cv). These lags and measurements allow us to follow the chain.
We can look at arcs in the loop. I do not understand why this idea is so
difficult for you and Bill to understand. I am sure you will tell me it is
because I do not get it.

So we have it. A stalemate. We have been here before.

It means that the behavior you see organmisms "producing" is a
side-effect of the process of control of perception. If conventional
psychology is not about control then it is not about the behavior
of living organisms.

This is your view. You are not PCT. Issac was right, PCT is not Bill. It
is not you; it is not anyone. It is an idea (actually, a set of ideas) to
which many might contribute.

Why not fill in the gaps for me? Why not tell me what conventional
psychologists have done to help us understand control? It's not
a secret, is it?

I have tried in the past. I pass.

Bill Powers (980223.1313 MST) --

Reading the posts for the past few days has left me with an
overwhelming sense of futility. It is futile even to try to
explain why.

This leads me to believe that my functions are not merely ineffective in
controling the perceptions I care about, but negatively effective. I am
making matters worse. Still, I will cling to the belief it is a relative
minima.

Sincerely,

Jeff

[From Rick Marken (980224.1040)]

Me:

No Jeff. The error variable is still in the loop. The stimulation
from the electrode is an independent variable, but the error variable
is not.

Jeff Vancouver (980224.0900 EST) --

You make a distinction between the manipulation and the effect on
the variable of interest. This is legitimate, but could be make
when talking about disturbances affect on the CV as well.

I do make it when talking about disturbances. But your reply misses
the point anyway. You gave the example of electrical stimulation
to argue that error can be an independent variable. It can't, just
as any variable in a control loop can't. The disturbance and
reference are the only independent variables in a control loop.
Electrical stimulation just adds another independent variable
into the loop; it doesn't turn the error variable into an
independent variable, any more than the distrubance turns the
controlled variable into an independent variable.

Me:

Try simulating a sequential state control system on a computer
and see what happens.

Jeff:

I have. It works.

If it is really a sequential state control system then the gain must
be less than 1.0 (see Powers, Psych Review, 1978). It works but it
is not much of a control system since it counters only a fraction
of the effect of distrubances to the cv.

So we have it. A stalemate. We have been here before.

Right. You don't want to accept the fact that, because people
are control systems, we have to look at and study their behavior
in a whole new way -- one oriented around the discovery of the
variables they control. I don't want to accept that fact that
control systems can be delt with as lineal causal systems.
Our research doesn't convince you of our point; your verbalisms
don't convince us of yours. Stalemate.

Me:

Why not fill in the gaps for me? Why not tell me what conventional
psychologists have done to help us understand control? It's not
a secret, is it?

Jeff:

I have tried in the past. I pass.

Perhaps I didn't notice then. Why not just tell me one or two
quick things that conventional psychologists have done to help
us understand control? I think it would be _very_ interesting.
We always tell you what we have done to help us understand
control (we even put stuff on the net so you can do the
experiments for yourself); and we also tell you (and show you)
what we have done to show why we suspect that the results of
conventional social science research is suspect. Why be so coy
about this?

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

[From Bruce Gregory (980224.1405 EST)]

Rick Marken (980224.1040)

to

Jeff Vancouver (980224.0900 EST) --

Right. You don't want to accept the fact that, because people
are control systems, we have to look at and study their behavior
in a whole new way -- one oriented around the discovery of the
variables they control. I don't want to accept that fact that
control systems can be delt with as lineal causal systems.
Our research doesn't convince you of our point; your verbalisms
don't convince us of yours. Stalemate.

You two perceive people in different ways. Rick perceives people
as living control systems and Jeff does not. I am not sure
exactly how Jeff perceives people, but I am sure that it is not
as LCS's. This is not to say that Jeff does not _believe_ that
human beings are living control systems, just that he does not
perceive them this way. In order to eliminate conflict with
Rick, Jeff would have to perceive human beings differently. In
order to eliminate conflict with Jeff, Rick would have to
perceive human beings differently. The stalemate arises because
Rick and Jeff live in different worlds as far as their
perceptions of human beings are concerned.

Bruce

[from Jeff Vancouver 980224.1420 EST]

[From Rick Marken (980224.1040)]

I do make it when talking about disturbances. But your reply misses
the point anyway. You gave the example of electrical stimulation
to argue that error can be an independent variable. It can't, just
as any variable in a control loop can't. The disturbance and
reference are the only independent variables in a control loop.
Electrical stimulation just adds another independent variable
into the loop; it doesn't turn the error variable into an
independent variable, any more than the distrubance turns the
controlled variable into an independent variable.

Talk about loops and circularity. I cannot convince you that something
other than the disturbance or the reference condition could be an
independent variable because that is by your definition the only things
that could be. What about the studies that timed how fast a signal passes
along a nerve. Stimulate one and observe the neurochemical release at the
other (or something like that, I do not remember exactly). You keep
talking simulations, but you know we are dealing with physical mechanisms.

Me:

Try simulating a sequential state control system on a computer
and see what happens.

Jeff:

I have. It works.

If it is really a sequential state control system then the gain must
be less than 1.0 (see Powers, Psych Review, 1978). It works but it
is not much of a control system since it counters only a fraction
of the effect of distrubances to the cv.

It is a loosy control system. Just like the one represented by my typing
here. It has little effect on the controlled perception. This to me is
indicative of the types of ECUs I deal with. The ones the constaintly need
reorganization and thinking.

Why not fill in the gaps for me? Why not tell me what conventional
psychologists have done to help us understand control? It's not
a secret, is it?

Jeff:

I have tried in the past. I pass.

Perhaps I didn't notice then. Why not just tell me one or two
quick things that conventional psychologists have done to help
us understand control? I think it would be _very_ interesting.
We always tell you what we have done to help us understand
control (we even put stuff on the net so you can do the
experiments for yourself); and we also tell you (and show you)
what we have done to show why we suspect that the results of
conventional social science research is suspect. Why be so coy
about this?

You too have a copy of our Psych Bull article and my behavioral science
article.

Sincerely,

Jeff

[from Jeff Vancouver 980224.14:30 EST]

[From Bruce Gregory (980224.1405 EST)]

You two perceive people in different ways. Rick perceives people
as living control systems and Jeff does not. I am not sure
exactly how Jeff perceives people, but I am sure that it is not
as LCS's. This is not to say that Jeff does not _believe_ that
human beings are living control systems, just that he does not
perceive them this way. In order to eliminate conflict with
Rick, Jeff would have to perceive human beings differently. In
order to eliminate conflict with Jeff, Rick would have to
perceive human beings differently. The stalemate arises because
Rick and Jeff live in different worlds as far as their
perceptions of human beings are concerned.

I completely reject this view of the differences in our views, respectively.

Sincerely,

Jeff

[From Bruce Gregory (980224.1508 EST)]

Jeff Vancouver 980224.14:30 EST]

I completely reject this view of the differences in our views, respectively.

Am I to take it that you believe that you perceive human
beings as living control systems?

Bruce

Yes.

[From Bruce Gregory (980224.1508 EST)]

Jeff Vancouver 980224.14:30 EST]

I completely reject this view of the differences in our views,

respectively.

Am I to take it that you believe that you perceive human
beings as living control systems?

Bruce

Sincerely,

Jeff

···

At 03:05 PM 2/24/1998 EST, you wrote: