Understanding control of behavior: Why it matters

csgnet@lists.illinois.edu's profile photo

[From Rick Marken (2014.11.21.2200)]

···

Martin Taylor (2014.11.21.23.01)–

MT: Look, I have no problem with accepting that when you control your

perception of an environmental variable well, that environmental
variable is stabilized. What I have a problem with is the result of
saying that it is the environmental variable that is controlled with
the implication that therefore the perception is stabilized.

RM: I didn’t know that was wrong? Let’s say the perception that is under control is p = a1x1+a2x2 where x1 and x2 are the environmental variables. So the aspect of the environment that is controlled is a linear combination of x1 and x2. So there is really nothing in the environment that corresponds to p; it is a function of environmental variables that corresponds to p. When this function of environmental variables is controlled p is controlled. So controlling p is really equivalent to controlling the aspect of the environment (a1x1+a2x2) that corresponds to p.

MT: The

causation is, of course, circular, and for life purposes the
stabilization of the environmental variable is more important than
that of the perception.

RM: I don’t see how perception and environmental variable can be controlled separately. When you control a perception you are controlling the aspect of the environment (what you call the “environmental variable”) that corresponds to the perception. How could one vary independently of the other?

MT: But that doesn't alter the fact that only

the perception can be controlled. Problems arise in situations like
this thread when the environmental variable is taken to be the
controlled variable. It isn’t, and the fact that you, as a
rubber-band subject, could have opted out of the situation or added
your own variations of the knot position without changing what E was
controlling, demonstrates that it isn’t.

RM: I don’t understand this at all. Are you saying that I could have stopped controlling for the coin being over the knot and E would have still been able to control his perception of my finger even though my finger was no longer under his control?

MT: Sometimes wording matters. I've taken your side on other occasions

by pointing out that when control is good, the environmental
variable is just as controlled as is the perception,

RM: But what you call the “environmental variable” (which I would prefer to call the “aspect of the environment” for the reasons I mention above) always corresponds to the perceptual variable, whether control is good or bad.

and it really

doesn’t matter much if you say the environmental variable is
controlled. After all, that’s teh foundation for the Test for the
Controlled variable. But when to make that point leads to a
fruitless discussion that could be stopped in its tracks by
insisting on the fact that it is the perception that is controlled,
I think it would be best to not make the point.

RM: I’m sorry, I don’t understand. Are you saying that it is possible to control a perception of behavior without actually controlling the behavior (the “environmental variable”) that is perceived?

Best

Rick


Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

David Goldstein (2014.11.23.1240)

I would like to continue with the tennis example. I may be able to control where I hit the ball, say to the baseline close to the right sideline. I cannot control whether the other person will hit it back. I cannot control how the other person will hit it back (down-the-line, crosscourt, shallow, deep, slice, topspin, flat). My control is very limited over the details of what happens when and if my ball arrives where I want it to go. I have no control over whether the other person calls my shot in or out. If I disagree with the call, I can ask to play the point over. The “knot” for each of us, is keep the ball in play. That is the reference perception. We can’t both achieve the reference perception on a specific point. It is a one person wins and the other person loses situation for a specific point. I can try a
nd make it harder for the other person to achieve the reference perception. If the other person has a weak backhand ( I am a lefty and let’s say the other person is a righty), I would play to his backhand. I cannot control whether or how the other person will return my shot. I am not able to control the other person except maybe where the other person is when my ball arrives there. And that is based on the other person wanting to achieve the reference perception.

David

···

Martin Taylor (2014.11.21.23.01)–

MT: Look, I have no problem with accepting that when you control your

perception of an environmental variable well, that environmental
variable is stabilized. What I have a problem with is the result of
saying that it is the environmental variable that is controlled with
the implication that therefore the perception is stabilized.

RM: I didn’t know that was wrong? Let’s say the perception that is under control is p = a1x1+a2x2 where x1 and x2 are the environmental variables. So the aspect of the environment that is controlled is a linear combination of x1 and x2. So there is really nothing in the environment that corresponds to p; it is a function of environmental variables that corresponds to p. When this function of environmental variables is controlled p is controlled. So controlling p is really equivalent to controlling the aspect of the environment (a1x1+a2x2) that corresponds to p.

MT: The

causation is, of course, circular, and for life purposes the
stabilization of the environmental variable is more important than
that of the perception.

RM: I don’t see how perception and environmental variable can be controlled separately. When you control a perception you are controlling the aspect of the environment (what you call the “environmental variable”) that corresponds to the perception. How could one vary independently of the other?

MT: But that doesn't alter the fact that only

the perception can be controlled. Problems arise in situations like
this thread when the environmental variable is taken to be the
controlled variable. It isn’t, and the fact that you, as a
rubber-band subject, could have opted out of the situation or added
your own variations of the knot position without changing what E was
controlling, demonstrates that it isn’t.

RM: I don’t understand this at all. Are you saying that I could have stopped controlling for the coin being over the knot and E would have still been able to control his perception of my finger even though my finger was no longer under his control?

MT: Sometimes wording matters. I've taken your side on other occasions

by pointing out that when control is good, the environmental
variable is just as controlled as is the perception,

RM: But what you call the “environmental variable” (which I would prefer to call the “aspect of the environment” for the reasons I mention above) always corresponds to the perceptual variable, whether control is good or bad.

and it really

doesn’t matter much if you say the environmental variable is
controlled. After all, that’s teh foundation for the Test for the
Controlled variable. But when to make that point leads to a
fruitless discussion that could be stopped in its tracks by
insisting on the fact that it is the perception that is controlled,
I think it would be best to not make the point.

RM: I’m sorry, I don’t understand. Are you saying that it is possible to control a perception of behavior without actually controlling the behavior (the “environmental variable”) that is perceived?

Best

Rick


Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

[From Rick Marken (2014.11.22.1910)]

···

Martin Taylor (2014.11.22.11.32)–

RM: You sent this just to me but I presume you meant to send it to CSGNet so I’m copying my reply there.

MT: When you use the same word in two different meanings as though the

two meanings were the same, you can get yourself in a lot of
trouble! Here, you use “control” to mean “acting so as to influence
a perceptual variable” in the first part of the sentence and as
“acting so as to successfully influence a perceptual variable so
that it actually stays near its reference value” in the second part.
The two are not the same. If they were, there would have been no
need for Bill to have spent so much time worrying about how to make
reorganization work.

RM: I still don’t understand. Perhaps that’s because you didn’t answer my question. So let me try asking again: Are you saying that I could have stopped controlling for the knot being over coin [as in the rubber band demo described in B:CP p. 245) and E would still have been able to control his perception of my finger even though my finger was no longer under his control?

MT: I'll turn the question back on you. Are you saying that if you (S)

added a small variability to your reference location for the knot, E
would or would not be controlling the location of your finger?

RM: E would unquestionably be able to control the location of my finger. That’s what I proved with the original “Control of Behavior” demo of 2003 (http://www.mindreadings.com/ControlDemosJava/Coercion.html) that I mentioned in an earlier post. It’s also demonstrated in the current “Control of Behavior” demo (http://www.mindreadings.com/ControlDemo/BehavioralControl.html) since the sheep’s reference for the variable it is controlling (distance between dog and grass shoot) is continuously varying.

MT: What

about if your reference location change was large?

RM: It would make E’s control of my finger slightly poorer, just as large changes in the disturbance to the cursor in a tracking task makes control of the cursor poorer.

MT: What about if you

controlled only for having the knot somewhere in a 1mm disk, a 1 cm
disk, a 1 meter disk, and didn’t care where in that disk it was?

RM: E could still control my finger just fine.

MT: At

what point would you say that E is not controlling E’s perception of
your finger position?

RM: Only when E stops controlling for keeping my finger on the target coin.

MT: I don’t know your answer,

RM: Now you do.

MT: but mine is that there is no such point so

long as E continues to act to change E’s perception of your finger
location.

RM: Same as mine answer. As long as E is controlling for his perception of the location of my finger he is controlling his perception of the location of my finger. And, of course, he is also controlling my finger.

MT: E is always _able_ to control E's perception of your

finger position, even if E can’t see your finger and must use
imagination.

RM: E cannot control his perception of my finger if he can’t see my finger. If E continues to imagine controlling my now unseen finger he is not controlling at all; he’s imagining. He is certainly not controlling a perception of my finger and, thus, he is not controlling my finger.

E: When E gives up in frustration and goes to have a much

needed drink, that’s when E is no longer controlling a perception of
your finger location.

RM: E stopped controlling my finger as soon as he started “controlling” in imagination (which, as I said, is not really controlling).

MT: Well, it's outside the realm of this thread, but yes, you can be

controlling a perception of someone’s behaviour while imagining that
it is different from what it is in the environment.

RM: That’s only if you consider controlling in imagination to be controlling a perception. I don’t. The is no perception when you control in imagination. In PCT a perception is always a function of sensory input. Even if there is a little “fill in” from imagination, it’s not a perception unless it is at least to some extent sensory based.

MT: The other person

may be deliberately deceiving you as to what the behaviour is. The
physical motions may be the same, but the behaviour is not. Actors
do it. That’s their professional job.

RM: I don’t see how this could possibly happen. It sounds to me that you are saying that E could be controlling his perception of S’s finger position but that S, a skillful actor, is only pretending to move his finger. But I can’t think what it is that S could “actually” be doing other than moving his finger.

MT: However, your wording really hinges on your mixing up "control" with

“successful control”. That’s what has been confusing the thread.

RM: Why is this so hard? The behavior of a control system can be controlled by disturbing a variable that the control system is controlling. This can be done without the control system’s permission, it can be done even if the reference for the controlled variable being disturbed is constantly varying, and when it is done; the behavior of the control system – the variations in the system’s outputs that keep its perception under controlled – and not just a perception of that behavior, is actually controlled. The fact that this true – that the behavior of a control system can be controlled – is demonstrated with the rubber band demo (as described on p. 245 of B:CP) and by my Control of Behavior demo(s).

RM: But if no one but me wants to believe that it is possible to control the behavior of a control system than I guess there is really no reason to go from there and talk about what might be wrong with controlling the behavior of a control system. After all, if it’s impossible to control the behavior of a control system than controlling the behavior of a control system can’t be a problem because there is no such thing. So I think we can end this thread now. I’ve presented what I think is overwhelming evidence that the behavior of a control system can be controlled; if the rubber band demo, the Control of Behavior demos, and even Bill Powers’ own words (p. 245 of B:CP) don’t convince you then nothing will.

So let’s just move on. I give! I don’t even remember why this came up.

Best

Rick

And yet, it can be controlled;-)

            MT: But that doesn't

alter the fact that only the perception can be
controlled. Problems arise in situations like this
thread when the environmental variable is taken to be
the controlled variable. It isn’t, and the fact that
you, as a rubber-band subject, could have opted out of
the situation or added your own variations of the knot
position without changing what E was controlling,
demonstrates that it isn’t.

          RM: I don't understand this at all. Are you saying that

I could have stopped controlling for the coin being over
the knot and E would have still been able to control his
perception of my finger even though my finger was no
longer under his control?

          RM: I'm sorry, I don't understand. Are you saying that

it is possible to control a perception of behavior without
actually controlling the behavior (the “environmental
variable”) that is perceived?

Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

[Martin Taylor 2014.11.22.23.59]

Thanks. It was a "reply" as used to be, versus "reply list" as is,

error.
Yes, but E would not have been very succcessful, would he?
Your finger was NEVER under his control. His perception of your
finger was, and so long as you wanted to and were able to keep the
knot accurately on the mark, his control of his perception would
have been good, and your finger would have been well controlled
inthe environment.
I agree.
I agree.
Really? Please explain how E controls your finger when you don’t
care where the knot is so long as it is within a 1 meter disk.
Martin

···

[From Rick Marken (2014.11.22.1910)]

              Martin Taylor

(2014.11.22.11.32)–

            RM: You sent this just to me but I presume you meant

to send it to CSGNet so I’m copying my reply there.

                            MT:

But that doesn’t alter the fact that
only the perception can be controlled.
Problems arise in situations like this
thread when the environmental variable
is taken to be the controlled variable.
It isn’t, and the fact that you, as a
rubber-band subject, could have opted
out of the situation or added your own
variations of the knot position without
changing what E was controlling,
demonstrates that it isn’t.

                          RM: I don't understand this at all. Are

you saying that I could have stopped
controlling for the coin being over the
knot and E would have still been able to
control his perception of my finger even
though my finger was no longer under his
control?

            RM: I still don't understand. Perhaps that's because

you didn’t answer my question. So let me try asking
again: Are you saying that I could have stopped
controlling for the knot being over coin [as in the
rubber band demo described in B:CP p. 245) and E would
still have been able to control his perception of my
finger even though my finger was no longer under his
control?

              MT: I'll turn the

question back on you. Are you saying that if you (S)
added a small variability to your reference location
for the knot, E would or would not be controlling the
location of your finger?

            RM: E would unquestionably be able to control the

location of my finger. That’s what I proved with the
original “Control of Behavior” demo of 2003 (http://www.mindreadings.com/ControlDemosJava/Coercion.html )
that I mentioned in an earlier post. It’s also
demonstrated in the current “Control of Behavior” demo (http://www.mindreadings.com/ControlDemo/BehavioralControl.html )
since the sheep’s reference for the variable it is
controlling (distance between dog and grass shoot) is
continuously varying.

              MT: What about if

your reference location change was large?

            RM: It would make E's control of my finger slightly

poorer, just as large changes in the disturbance to the
cursor in a tracking task makes control of the cursor
poorer.

              MT: What about if

you controlled only for having the knot somewhere in a
1mm disk, a 1 cm disk, a 1 meter disk, and didn’t care
where in that disk it was?

RM: E could still control my finger just fine.

[From Rick Marken (2014.11.22.0930)]

···

Martin Taylor (2014.11.22.23.59)–

MT: Your finger was NEVER under his control. His perception of your

finger was,

RM: Yes, I know that’s what you think. That’s what everyone on this list (except me) thinks. So let’s just drop it.

Best

Rick

Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

[Martin Taylor 2014.11.23.09.01]

I think the issue should be resolved. I think we agree on everything

except semantics, which doesn’t leave much of a gap to be bridged.
The issue is not whether and to what degree control of an
environmental variable matches control of the corresponding
perception. I thin we seem to agree on that. The issue is whether
someone is or is not controlling a perception when s/he is
unsuccessfully acting to influence it. As I use the word, there is a
continuum from unachievable perfect control through good control,
poor control, and ineffective control (control actions do not
influence the perception) to counterproductive control (control
actions move the perception away from its reference value). You say there is a cutoff point somewhere in this continuum.
Somewhere between what I would call poor control and ineffective
control you say control stops being control. If the actions don’t
influence the environmental variable, you say control doesn’t exist.
I say that if a person is acting to influence a perception that has
a reference value, control continues.
Let’s forget about the elastic band experiment and think of a purely
physical environmental variable. This is a real-life example. I see
an apparently small rock in a flowerbed I am digging, and try to
pick it up to put it aside. It doesn’t move at all, being actually
the small visible top of a large boulder. I get a spade and dig
around it but still can’t get it to move. I get friends to help me
and it still doesn’t move. Am I controlling a perception of its
location? I say yes. I would guess you would say no. I finally did get this large boulder moved, using a car jack and
various other techniques. When did my non-control become control
again? The first time the boulder actually moved?
Surely if control does stop being control at some point of
ineffectiveness, you should be able to give some criterion for
saying where that point is.
Martin

···

On 2014/11/23 12:32 AM, Richard Marken
( via csgnet Mailing List) wrote:

rsmarken@gmail.com

[From Rick Marken (2014.11.22.0930)]

            Martin Taylor

(2014.11.22.23.59)–

            MT: Your finger was NEVER under his control. His

perception of your finger was,

          RM: Yes, I know that's what you think. That's what

everyone on this list (except me) thinks. So let’s just
drop it.

[From Rupert Young (2014.11.23 17.30)]

(Rick Marken (2014.11.21.1730)]

RM: We're just demonstrating an interesting scientific fact about purposeful (control) systems: their behavior can be controlled, in this case by applying disturbances to a variable that is controlled by the control system.

This is introducing error into the system such that it will act to reduce that error. So I guess there are two ways that behaviour can be controlled, one, as you say, to disturb the perception, and the other is to change, or set, the reference.

For the former a simple way to try it is to, when in someone's home and they are not watching, tip up one of their pictures on the wall so that it is not straight. Sooner or later they will straighten the picture, if that is their bag.

The latter can be acheived through directed learning (reorganisation) so that the person (the controllee control system) ends up controlling goals that you, the controller, determined. Some forms of this "education" (religious indoctrination) are successful to the extent that the result is the destruction of the controllee themselves, e.g. suicide bombers.

Perhaps, though, the determination of reference goals is achieved by disturbing perceptions at a higher level, such as "acceptance" within a group.

Regards,
Rupert

[Martin Taylor 2014.11.23.12 48]

[From Rupert Young (2014.11.23 17.30)]

(Rick Marken (2014.11.21.1730)]

RM: We're just demonstrating an interesting scientific fact about purposeful (control) systems: their behavior can be controlled, in this case by applying disturbances to a variable that is controlled by the control system.

This is introducing error into the system such that it will act to reduce that error. So I guess there are two ways that behaviour can be controlled, one, as you say, to disturb the perception, and the other is to change, or set, the reference.

For the former a simple way to try it is to, when in someone's home and they are not watching, tip up one of their pictures on the wall so that it is not straight. Sooner or later they will straighten the picture, if that is their bag.

The latter can be acheived through directed learning (reorganisation) so that the person (the controllee control system) ends up controlling goals that you, the controller, determined. Some forms of this "education" (religious indoctrination) are successful to the extent that the result is the destruction of the controllee themselves, e.g. suicide bombers.

Perhaps, though, the determination of reference goals is achieved by disturbing perceptions at a higher level, such as "acceptance" within a group.

Your second-last paragraph describes how the organism becomes organized so that the last paragraph (without "Perhaps, though,") becomes effective in an ongoing dynamic situation.

More simply, unless you stick wires in the brain, as is done in some brain surgeries, the only way to change someone else's reference values is to disturb a perception at a higher level. What that actually does is, as you say, determined by reorganization.

Martin

[From Rick Marken (2014.11.23.1020)]

RM: Which sounds to me like you are saying that you can be in control of a perception without being in control of the environmental variable (the aspect of the environment) that corresponds to that perception. I completely disagree with the “semantics” that I get from what you say here. Unless the meaning I should be getting from that sentence is “Your finger is under E’s control when E’s perception of your finger is under control”. But if that is what you meant then you have to admit that your way of saying that is pretty odd.

Best

Rick

···

Martin Taylor (2014.11.23.09.01) –

RM: Well, OK. I’ve got some time to waste. Let’s try it. You said (regarding E controlling S’s finger in the rubber band demo, which Bill described as “E can control S’s finger” on p. 245 of B:CP – a description with which you took issue)

RM: The way I would say this is: My finger was**EVER under his control to the exact extent to which his perception of my finger was. It seems like there is a pretty big semantic gap between your way of saying it and mine. Indeed, given my understanding of the English language I would say that we are saying the exact opposite of each other. I’d like to see how you would bridge that semantic gap!

MT: The issue is not whether and to what degree control of an

environmental variable matches control of the corresponding
perception.

I thin we seem to agree on that.

RM: It sure seems like it is the issue. So, no, we don’t agree on that. You said that my finger (an environmental variable) was never under control in the rubber band demo but E’s perception of my finger was. So it sure sounds like you are saying that the degree of control of an environmental variable (finger) can be 0 (no control) when the degree of control of the perception of that variable is >0. I take issue with that, my position being that when there is control (to whatever degree) of a perception of an environmental variable (which I prefer to call an aspect of the environment) then the environmental variable is also controlled (to the same degree). So when E controls the perception of my finger he is also controlling my finger.

MT: The issue is whether

someone is or is not controlling a perception when s/he is
unsuccessfully acting to influence it.

RM: I don’t think so. The issue (for me) is this sentence:

MT: I think the issue should be resolved. I think we agree on everything

except semantics, which doesn’t leave much of a gap to be bridged.

MT: Your [S’s] finger was NEVER under his [E’s] control. His perception of your finger was,

MT: Your finger was NEVER under his control. His perception of your finger was,

As I use the word, there is a

continuum from unachievable perfect control through good control,
poor control, and ineffective control (control actions do not
influence the perception) to counterproductive control (control
actions move the perception away from its reference value).

You say there is a cutoff point somewhere in this continuum.

Somewhere between what I would call poor control and ineffective
control you say control stops being control. If the actions don’t
influence the environmental variable, you say control doesn’t exist.
I say that if a person is acting to influence a perception that has
a reference value, control continues.

Let's forget about the elastic band experiment and think of a purely

physical environmental variable. This is a real-life example. I see
an apparently small rock in a flowerbed I am digging, and try to
pick it up to put it aside. It doesn’t move at all, being actually
the small visible top of a large boulder. I get a spade and dig
around it but still can’t get it to move. I get friends to help me
and it still doesn’t move. Am I controlling a perception of its
location? I say yes. I would guess you would say no.

I finally did get this large boulder moved, using a car jack and

various other techniques. When did my non-control become control
again? The first time the boulder actually moved?

Surely if control does stop being control at some point of

ineffectiveness, you should be able to give some criterion for
saying where that point is.

Martin


Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

David,

as I’m tennis player too, I can say only that I fully agree with you, and probably will all other tennis player. Only consider one detail.

DG:

I am not able to control the other person except maybe where the other person is when my ball arrives there.

HB : I’m experiencing it diferently. I don’t control where the other person will be when ball arrive on the other side. I control rather for »where opponent will not be«, when ball arrive on other side. So if I want to win a point I will try ,beside placing the ball as far as possible from opponent, to spin the ball or give the ball as much as possible acceleration, and so on, so that it will be harder for the opponent to return the ball or control the shot, which he will choose. But i can’t control whether opponent will be on the ball or not when ball arrive in opponents part of the playground. But I can control what I will do with ball or whether i’ll catch it if opponent is precisely shoting the ball enough away from me.

Maybe we can see the example with drop-shot. My attempt is to place the ball in the field as much as possible far from the opponent, hoping that he will not be there on the ball, when ball arrives on te other side. Or passing-shot… So I’m not at all »controlling« for the opponent to be »on the ball«or to be there when the ball will fall in his part of the playground. And this is what I think most players are doing to win the point. It’s quite similar to tactics in table tennis, although there are differences. The bases of the tactics is to make control of returning the ball harder for the opponent. So if I would place the ball near opponent, it’ quite easy for him to control the return and thus the winning of the point. I hope we understand what I wanted to say.

Best,

Boris

image00115.png

···

From: csgnet-request@lists.illinois.edu [mailto:csgnet-request@lists.illinois.edu] On Behalf Of David Goldstein (davidmg@verizon.net via csgnet Mailing List)
Sent: Saturday, November 22, 2014 7:07 PM
To: rsmarken@gmail.com
Cc: csgnet@lists.illinois.edu
Subject: Re: Understanding control of behavior: Why it matters

David Goldstein (2014.11.23.1240)

I would like to continue with the tennis example. I may be able to control where I hit the ball, say to the baseline close to the right sideline. I cannot control whether the other person will hit it back. I cannot control how the other person will hit it back (down-the-line, crosscourt, shallow, deep, slice, topspin, flat). My control is very limited over the details of what happens when and if my ball arrives where I want it to go. I have no control over whether the other person calls my shot in or out. If I disagree with the call, I can ask to play the point over. The “knot” for each of us, is keep the ball in play. That is the reference perception. We can’t both achieve the reference perception on a specific point. It is a one person wins and the other person loses situation for a specific point. I can try a nd make it harder for the other person to achieve the reference perception. If the other person has a weak backhand ( I am a lefty and let’s say the other person is a righty), I would play to his backhand. I cannot control whether or how the other person will return my shot. I am not able to control the other person except maybe where the other person is when my ball arrives there. And that is based on the other person wanting to achieve the reference perception.

David

Sent from my iPad

On Nov 22, 2014, at 1:03 AM, Richard Marken (rsmarken@gmail.com via csgnet Mailing List) csgnet@lists.illinois.edu wrote:

csgnet@lists.illinois.edu's profile photo

[From Rick Marken (2014.11.21.2200)]

Martin Taylor (2014.11.21.23.01)–

MT: Look, I have no problem with accepting that when you control your perception of an environmental variable well, that environmental variable is stabilized. What I have a problem with is the result of saying that it is the environmental variable that is controlled with the implication that therefore the perception is stabilized.

RM: I didn’t know that was wrong? Let’s say the perception that is under control is p = a1x1+a2x2 where x1 and x2 are the environmental variables. So the aspect of the environment that is controlled is a linear combination of x1 and x2. So there is really nothing in the environment that corresponds to p; it is a function of environmental variables that corresponds to p. When this function of environmental variables is controlled p is controlled. So controlling p is really equivalent to controlling the aspect of the environment (a1x1+a2x2) that corresponds to p.

MT: The causation is, of course, circular, and for life purposes the stabilization of the environmental variable is more important than that of the perception.

RM: I don’t see how perception and environmental variable can be controlled separately. When you control a perception you are controlling the aspect of the environment (what you call the “environmental variable”) that corresponds to the perception. How could one vary independently of the other?

MT: But that doesn’t alter the fact that only the perception can be controlled. Problems arise in situations like this thread when the environmental variable is taken to be the controlled variable. It isn’t, and the fact that you, as a rubber-band subject, could have opted out of the situation or added your own variations of the knot position without changing what E was controlling, demonstrates that it isn’t.

RM: I don’t understand this at all. Are you saying that I could have stopped controlling for the coin being over the knot and E would have still been able to control his perception of my finger even though my finger was no longer under his control?

MT: Sometimes wording matters. I’ve taken your side on other occasions by pointing out that when control is good, the environmental variable is just as controlled as is the perception,

RM: But what you call the “environmental variable” (which I would prefer to call the “aspect of the environment” for the reasons I mention above) always corresponds to the perceptual variable, whether control is good or bad.

and it really doesn’t matter much if you say the environmental variable is controlled. After all, that’s teh foundation for the Test for the Controlled variable. But when to make that point leads to a fruitless discussion that could be stopped in its tracks by insisting on the fact that it is the perception that is controlled, I think it would be best to not make the point.

RM: I’m sorry, I don’t understand. Are you saying that it is possible to control a perception of behavior without actually controlling the behavior (the “environmental variable”) that is perceived?

Best

Rick

Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.

Now available from Amazon or Barnes & Noble

Rick,

My text is somewhere in the middle of the text

···

From: csgnet-request@lists.illinois.edu [mailto:csgnet-request@lists.illinois.edu] On Behalf Of Richard Marken (rsmarken@gmail.com via csgnet Mailing List)
Sent: Friday, November 21, 2014 7:23 AM
To: csgnet@lists.illinois.edu
Subject: Re: Understanding control of behavior: Why it matters

[From Rick Marken (2014.11.20.2220)]

Kent McClelland (2014.11.20.1140)]

KM: Several of the recent posts in this thread by Rick and Martin have clarified things enough for me that I am beginning to think that I finally understand what Rick has been trying to say.

KM: Is this your argument? (Please correct me if I’m wrong.)

KM: People [attempt to] control [some aspect, as they perceive it, of] other people’s behavior [or, in other words, their observable actions] [often successfully, but not always] {with or without the other people’s cooperation].

KM: If this is your argument, I agree with it fully.

RM: Great. That is certainly a way of making my argument. But actually my argument has been made, non-verbally, by the “Control of Behavior” demo (http://www.mindreadings.com/ControlDemo/BehavioralControl.html). The demo shows that a control system (you, in the role of the dog) can control the behavior of another control system (the sheep). That is my argument: that the behavior of a control system can be controlled by another control system.

RM:All the verbal counter-arguments to to my argument become moot, I believe, in the face of that demo. If the measures of “Sheepdog Control” displayed at the end of a run show that there was reasonably good control (a Stability measure > 2) then you (a control system) have controlled the behavior of another control system (the sheep). This is happening whether or not you call the movements of the sheep its “behavior”; it’s happening whether the sheep wants to make these movements or not; it’s happening whether you are controlling the position of the sheep well or poorly.

KM: When you state the argument, however, without noting the stipulations in brackets (or just leaving them implicit), you risk that readers will misunderstand you, both within the PCT community and outside of it. This thread has illustrated some of the ways in which people might misunderstand you.

RM: I try to do the best I can at describing what I know to be true from the modeling. But understanding based on verbal descriptions can never be as good as an understanding based on modeling. I think verbal descriptions are important because we have to try to communicate our understanding to one another. But when there are disagreements about what the correct understanding is, I think the final arbiter has to be the model.

HB : Rick it’s good that you start thinking about the problem of »control« and of other possibilities. It seems that finaly you are prepared to talk. I didn’t want to interfere anymore, because you were so firmly on your side of »river bank«. But now it seems to me, that you try to consider other oppinons. But I hope that Kent won’t mind becasue I jumped in. Â

I propose you to make some more investigating what is the »final arbiter«.

BP (Bill Powers not to be mixed with Barb) : »If the effects of the model are just as hypothetical as the model, we don’t have a model, because we can’t check it against direct experience. The ultimate authority is always direct experience, the real reality we are incapable of doubting…« (LCS II, p.185)

/span>

HB : So as I see it the final »arbiter« has to be »experience of reality«, which is defined in Bill’s work, not model or demo, which is imgined construct, which in you case has troubles with defining »control variables« for a sheep. And I could add that also important is conversation about different »experiences of reality«. Video is no doubt closer to »reality« than »model or demo of reality«, which you presented for dog and sheep. Demo or model has to feet to experiences in reality or video that shows reality. So videos are in the case of »sheepdog« and sheep »higher authority« than demo, which is lacking also of some physiological evidence that your thinking in demo is right. I can say only that »controlled variables« for sheep are wrong if i compare them to video.

Best,

Boris

RM: I think I understand where people are coming from on this “control of behavior” issue because I was there myself until relatively recently. I wanted to believe that people are autonomous and uncontrollable. But I knew from the “rubber band demo” described in B:CP (p. 245 in the 2nd edition) that it was possible to control the behavior of a human control system. [For those who take Bill’s words rather than the model as scripture, note that in the third paragraph on that page Bill says: “Therefore, if S wants to control the knot, E can control S’s finger”.]

RM: But I still wanted to believe that a living control system is autonomous in a way that makes its behavior fundamentally uncontrollable. And I believed this autonomy/uncontrollability would come from the fact that the reference signals in living control systems, which are really the source of autonomy in these systems since they are set by the system itself (rather than by an outside “user” as in artificial control systems), would be constantly varying (Powers has called it “secular variation” and I have followed suit). Since the control of behavior that is seen in the rubber band demo occurs when S (the controllee) adopts a fixed reference for the controlled variable (knot on coin), I thought the ability of E (the controller) to control S’s behavior (finger position) would be compromised if S’s reference for the controlled variable were constantly changing.

RM: It was actually Bruce Abbott who predicted that secular variation in the controllee’s reference would not affect E’s ability to control S’s behavior. I strongly rejected Bruce’s suggestion, inventing all kinds of mental scenarios to justify my belief that secular variations in the controllee’s reference would preserve S’s autonomy and make it impossible for E to control S’s behavior.

RM: But I finally calmed down and set up a computer model/demo to test my idea. What I set up was essentially the same as current Control of Behavior demo (but without the cute dog and sheep). This was done in 2003 and the demo is still up on the net as a java program, which you can run if your computer will let you run java (http://www.mindreadings.com/ControlDemosJava/Coercion.html).

RM: Well, it turns out that ol’ Bruce was right. The demo showed that secular variation in the controllee’s reference for the controlled variable had no effect at all on the controller’s ability to control the controlee’s behavior via disturbance to the controlled variable. This was kind of a disappointing finding, not least because I had to admit that Bruce was right and I was wrong;-) But facts is facts.

RM: But I now understand that this was a very important finding about the nature of living control systems. It shows that, while living control systems are autonomous (in the sense that they set their own goals for what they should be experiencing) they are also controllable. And what makes control systems controllable is the fact that they have references for the way things should be. Since references are basically the same as “wants”, another way to say this is that what makes us controllable is our wants.

RM: So to the extent that being controllable is not being free, the PCT model tells us that what deprives us of our freedom is the fact that we want (have references for the way our perceptions should be). So true freedom – perfect autonomy – would be achieved if we could stop wanting, which seems rather Buddhist, doesn’t it. Indeed, I think Bill has said that when he was first discovering MOL and was practicing it on himself himself he managed to get his consciousness to a level above he could not go. At this level he was above all want (above all references in the hierarchy). I think he described it as a “wantless” state – one of perfect freedom. PCT nirvana. I think I’ve managed to get there once or twice and it is lovely.

RM: So, surprise. This apparently awful discovery of the controllability of control systems may suggest a path to one’s Buddha nature. Who would have guessed;-)

Namaste

Rick

David,

as I’m tennis player too, I can say only that I fully agree with you, and probably will all other tennis player. Only consider one detail.

DG:

I am not able to control the other person except maybe where the other person is when my ball arrives there.

HB : I’m experiencing it diferently. I don’t control where the other person will be when ball arrive on the other side. I control rather for »where opponent will not be«, when ball arrive on other side. So if I want to win a point I will try ,beside placing the ball as far as possible from opponent, to spin the ball or give the ball as much as possible acceleration, and so on, so that it will be harder for the opponent to return the ball or control the shot, which he will choose. But i can’t control whether opponent will be on the ball or not when ball arrive in opponents part of the playground. But I can control what I will do with ball or whether i’ll catch it if opponent is precisely shoting the ball enough away from me.

Maybe we can see the example with drop-shot. My attempt is to place the ball in the field as much as possible far from the opponent, hoping that he will not be there on the ball, when ball arrives on te other side. Or passing-shot… So I’m not at all »controlling« for the opponent to bbe »on the ball«or to be there when the ball will fall in his part of the playground. And this is what I think most players are doing to win the point. It’s quite similar to tactics in table tennis, although there are differences. The bases of the tactics is to make control of returning the ball harder for the opponent. So if I would place the ball near opponent, it’ quite easy for him to control the return and thus the winning of the point. I hope we understand what I wanted to say.

Best,

Boris

···

From: csgnet-request@lists.illinois.edu [mailto:csgnet-request@lists.illinois.edu] On Behalf Of David Goldstein (davidmg@verizon.net via csgnet Mailing List)
Sent: Saturday, November 22, 2014 7:07 PM
To: rsmarken@gmail.com
Cc: csgnet@lists.illinois.edu
Subject: Re: Understanding control of behavior: Why it matters

David Goldstein (2014.11.23.1240)

I would like to continue with the tennis example. I may be able to control where I hit the ball, say to the baseline close to the right sideline. I cannot control whether the other person will hit it back. I cannot control how the other person will hit it back (down-the-line, crosscourt, shallow, deep, slice, topspin, flat). My control is very limited over the details of what happens when and if my ball arrives where I want it to go. I have no control over whether the other person calls my shot in or out. If I disagree with the call, I can ask to play the point over. The “knot” for each of us, is keep the ball in play. That is the reference perception. We can’t both achieve the reference perception on a specific point. It is a one person wins and the other person loses situation for a specific point. I can try a nd make it harder for the other person to achieve the reference perception. If the other person has a weak backhand ( I am a lefty and let’s say the other person is a righty), I would play to his backhand. I cannot control whether or how the other person will return my shot. I am not able to control the other person except maybe where the other person is when my ball arrives there. And that is based on the other person wanting to achieve the reference perception.

David

Sent from my iPad

On Nov 22, 2014, at 1:03 AM, Richard Marken (rsmarken@gmail.com via csgnet Mailing List) csgnet@lists.illinois.edu wrote:

<image001.png>

[From Rick Marken (2014.11.21.2200)]

Martin Taylor (2014.11.21.23.01)–

MT: Look, I have no problem with accepting that when you control your perception of an environmental variable well, that environmental variable is stabilized. What I have a problem with is the result of saying that it is the environmental variable that is controlled with the implication that therefore the perception is stabilized.

RM: I didn’t know that was wrong? Let’s say the perception that is under control is p = a1x1+a2x2 where x1 and x2 are the environmental variables. So the aspect of the environment that is controlled is a linear combination of x1 and x2. So there is really nothing in the environment that corresponds to p; it is a function of environmental variables that corresponds to p. When this function of environmental variables is controlled p is controlled. So controlling p is really equivalent to controlling the aspect of the environment (a1x1+a2x2) that corresponds to p.

MT: The causation is, of course, circular, and for life purposes the stabilization of the environmental variable is more important than that of the perception.

RM: I don’t see how perception and environmental variable can be controlled separately. When you control a perception you are controlling the aspect of the environment (what you call the “environmental variable”) that corresponds to the perception. How could one vary independently of the other?

MT: But that doesn’t alter the fact that only the perception can be controlled. Problems arise in situations like this thread when the environmental variable is taken to be the controlled variable. It isn’t, and the fact that you, as a rubber-band subject, could have opted out of the situation or added your own variations of the knot position without changing what E was controlling, demonstrates that it isn’t.

RM: I don’t understand this at all. Are you saying that I could have stopped controlling for the coin being over the knot and E would have still been able to control his perception of my finger even though my finger was no longer under his control?

MT: Sometimes wording matters. I’ve taken your side on other occasions by pointing out that when control is good, the environmental variable is just as controlled as is the perception,

RM: But what you call the “environmental variable” (which I would prefer to call the “aspect of the environment” for the reasons I mention above) always corresponds to the perceptual variable, whether control is good or bad.

and it really doesn’t matter much if you say the environmental variable is controlled. After all, that’s teh foundation for the Test for the Controlled variable. But when to make that point leads to a fruitless discussion that could be stopped in its tracks by insisting on the fact that it is the perception that is controlled, I think it would be best to not make the point.

RM: I’m sorry, I don’t understand. Are you saying that it is possible to control a perception of behavior without actually controlling the behavior (the “environmental variable”) that is perceived?

Best

Rick

Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.

Now available from Amazon or Barnes & Noble

Apparently my initial attempt to send this went only to Martin, not to CSGnet as a whole.

···

On Sun, Nov 23, 2014 at 9:44 AM, Martin Taylor csgnet@lists.illinois.edu wrote:

[Martin Taylor 2014.11.23.09.01]
On 2014/11/23 12:32 AM, Richard Marken
(rsmarken@gmail.com via csgnet Mailing List) wrote:

[From Rick Marken (2014.11.22.0930)]

I think the issue should be resolved. I think we agree on everything

except semantics, which doesn’t leave much of a gap to be bridged.

The issue is not whether and to what degree control of an

environmental variable matches control of the corresponding
perception. I thin we seem to agree on that. The issue is whether
someone is or is not controlling a perception when s/he is
unsuccessfully acting to influence it. As I use the word, there is a
continuum from unachievable perfect control through good control,
poor control, and ineffective control (control actions do not
influence the perception) to counterproductive control (control
actions move the perception away from its reference value).

You say there is a cutoff point somewhere in this continuum.

Somewhere between what I would call poor control and ineffective
control you say control stops being control. If the actions don’t
influence the environmental variable, you say control doesn’t exist.
I say that if a person is acting to influence a perception that has
a reference value, control continues.

Let's forget about the elastic band experiment and think of a purely

physical environmental variable. This is a real-life example. I see
an apparently small rock in a flowerbed I am digging, and try to
pick it up to put it aside. It doesn’t move at all, being actually
the small visible top of a large boulder. I get a spade and dig
around it but still can’t get it to move. I get friends to help me
and it still doesn’t move. Am I controlling a perception of its
location? I say yes. I would guess you would say no.

I finally did get this large boulder moved, using a car jack and

various other techniques. When did my non-control become control
again? The first time the boulder actually moved?

Surely if control does stop being control at some point of

ineffectiveness, you should be able to give some criterion for
saying where that point is.

Martin
            Martin Taylor

(2014.11.22.23.59)–

            MT: Your finger was NEVER under his control. His

perception of your finger was,

          RM: Yes, I know that's what you think. That's what

everyone on this list (except me) thinks. So let’s just
drop it.

[From Rick Marken (2014.11.23.1515)]

···

On Sun, Nov 23, 2014 at 2:36 PM, “Boris Hartman” csgnet@lists.illinois.edu wrote:

HB : Rick it’s good that you start thinking about the problem of »control« and of other possibilities. It seems that finaly you are prepared to talk.

RM: No, I’m pretty much done talking about “control of behavior”. I’ve done what I can – even reducing myself to quoting “scripture” (p. 245 of B:CP) – to show that control of behavior is an observable (and pretty ubiquitous) phenomenon that is explained by PCT. But no one’ on CSGNet is buying it (as you correctly predicted). So if people think “control of behavior” doesn’t exist then there is really nothing to talk about.

Best

Rick

Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

[From Rick Marken ( 2014.11.23.1640)]

···

Tracy Harms (2014.11.23.11:50 EST)–

TH: PCT has, I propose, been struggling with definitional troubles all along.

RM: Good point. One really does have to know what the meaning of certain words are inorder to understand (and communicate) PCT clearly and correctly.

TH: I know I’m not alone in thinking this, and I hasten to add that I don’t think these troubles could have been easily avoided by selecting different words. It’s more that we want to refine a technical sense of certain words that have broad non-technical meanings. Today I want to focus on “control.” Students of PCT all learn that control is systematic stabilization. Every English speaker also knows control to be effective manipulation. These are two different meanings of the word, I think. Hopefully we’ll be living in a world where both meanings are in frequent use, so we might as well learn to get adept at cutting through confusion that comes when topics of manipulation threaten to cloud our thinking on behavior.

RM: Yes, I think “control” is probably the crucial word to understand. As you note, the lay meaning of “control” is more like “cause” than “control” as we understand it. The “controls” on a control panel don’t control in the PCT sense; they just cause a variable, such as the the size of the opening of a value, to vary. The person using the “controls” on the control panel is actually the one doing the controlling (usually of some measurand, such as the rate of flow in a pipe).

TH: Yes, we should recognize the behavior of individuals as a subject of control by other individuals. Further, we should not lose track of the PCT meaning of control when it becomes easy to go astray by misconstruing part of the loop (output) as control, per se.

RM: I agree. And I think that whenever we talk about “control” on CSGNet we should always use the PCT meaning of that word: maintaining a variable in a preselected state, protected from disturbance. I always do, or at least I always try to. So when I talk about “control of behavior” I am always talking about keeping a behavioral variable, such as the position of the sheep relative to the herd in my “Control of Behavior” demo: http://www.mindreadings.com/ControlDemo/BehavioralControl.html) in a preselected state (close to the herd), protected from disturbances (the movements of the herd). Examples of objective measures of control, which show quantitatively how well some behavioral variable has been controlled (how well the variable has been protected from disturbances to its preselected state) are shown at the end of the demo. These are the measures of RMS error and stability under “Sheepdog control” which show how well the behavior of the sheep (location relative to the herd) was controlled by the sheepdog (you).

Best

Rick


Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

[From Erling Jorgensen (2014.11.24 1130 EST)]

I want to give my sense of the import of this discussion, just to clarify in
my own mind where I am coming down on the matter of “control of behavior�? by
an outside living control system. I am not sure whether my opinion is a
majority report or minority report, among CSGNet in general.

My main problem is with the word “behavior�? itself, because to the user it is
obvious what they mean and to the hearer it is far from clear. It is that
very disconnect that leads me to want to use a word like “moving�? or “acting�?
or “doing,�? because their ambiguity about what is meant almost forces a
follow-up query. Queries like, “Well, what part of their moving is being
controlled?�? or “What result of their acting is being controlled?�?

My primary interest here is controlling for communicating clearly, and that
includes which words best fit a given model. I am not too concerned about the
word “control,�? because that is pretty well defined and specified within PCT.
Nor am I especially concerned about some kind of political correctness that
supposedly people should not be controlled by other people. I try to have
science trump ideology, as best I can, within my system concepts. Part of my
expertise -- by profession and training as a counseling psychologist, and by
inclination and attachment as the son and grandson of two “word smith�? pastors
-- is that of communication.

As I use the words, a finger is not a behavior; a location is not a behavior;
distance is not a behavior. Behavior is used for those things in the PCT
hierarchy that we often call Transitions, and if those transitions are timed
in sequence we often call them Events. It includes a lot of the �ing words iin
the English language. So then, “moving�? is a behavior. But in order to give
any communicative specificity in any given instance, some kind of perceptual
result usually needs to be added. “Moving what?�? or “Moving what into what
position?�? It is that communicative necessity that drives us again (thanks be
to Bill Powers) to realize it is the perceptual results that get controlled
for. Behavior gets done, results get controlled.

I have a personal preference for saving the word behavior for the lowest level
of the hierarchy, where actual forces in the environment get generated. Every
output above that is a specification for a certain perceptual result. But I
realize I am not going to change how everyone else uses the word behavior.
And so I humbly bow to the community on that.

I do not think it works to call the different results “aspects�? of behavior.
They could as easily be called “aspects�? of the causal relationships in play,
or “aspects�? of the system concepts. It is true that such aspects are
variable in that they can take on different values as to what is their
preferred state. But to call that the “behavior�? of those aspects just makes
things way too ambiguous, with a word loaded with outside connotations. Far
better in those situations, to specify the aspects themselves that are being
controlled, (using HPCT as a useful guide for categorizing the perceptions in
play.)

That is my summary on this discussion, which is to say, my perceptions and
preferences.

All the best,
Erling

[From Fred Nickols (2014.11.24.1430 EST)]

I've not participated until now because I wasn't sure I could contribute. But now I think I can. I also think I'm still with Erling.

I went back through the thread and came across this snippet from Rick:
[From Rick Marken (2014.11.19.1500)]
RM: This is a crucial observation!! It's important to know what control is in order to understand when control of behavior is happening. Controlling behavior is not the same as influencing or causing behavior. Controlling involves bringing a variable, such as the distance between the sheep and the herd in my demo, to a predetermined state, such as "sheep close to herd", and keeping it there, protected from disturbances, which include the movements of the herd and of the sheep itself. Given this definition of control, the distance of the sheep from the herd can clearly be controlled in my demo. Since the distance between sheep and herd depends at least in part on the movements of the sheep this variable is a behavior of the sheep.

Rick says at the very end "this variable is a behavior of the sheep." I can't accept that. The distance between sheep and herd is a variable. The sheep's movements are behaviors (describable at various levels of specificity). The distance between sheep and herd varies as a function of the sheep's movements but that distance is not a behavior of the sheep. Nor is it an "aspect" of the behavior of the sheep.

Different example. Suppose you are sitting in chair by a table. You are my prisoner. I have a gun. You don't. On the table is a glass of water. I tell you to drink. You refuse. I take out my gun, put it to your head and tell you to drink. Will you pick up the glass and take a drink? Probably (but maybe not). Am I "controlling" your behavior? Most people wouldn't quarrel with that assertion. Certainly, I made you do what I wanted. But YOU picked up the glass and drank from it. So even though I coerced you into taking a drink you remained in control of the behavior necessary to do that.

I prefer to think of all this along these lines...

Behavior, as the behaviorists have long held is a term that refers to the activity of an organism.

Behavior also has effects, often referred to as outcomes, results, accomplishments and just plain effects. In PCT, or so I believe, the effects of interest typically relate to the value of some variable that has been targeted for control. When I get you to do what I want done I am rarely interested in your behavior per se; it's the outcome or result that matters.

Consider me in my Navy days as a Chief. If I happened to shout out "Attention on deck!" in a crowded mess room I wasn't interested in the many and varied behaviors all the sailors presents would have to engage in. I was concerned with the end result of whatever behaviors it took them to come to an erect position, feet together and hands at sides, eyes looking straight ahead. The position of "at attention" is a variable and one that I could reasonably successfully impose on whomever happened to be in the mess room. How they got to that position was under their control. Being at attention is not an "aspect" of behavior; it is an outcome of the behaviors necessary to reach that position from whatever starting point might be the case and for all those sailors in the mess room there were lots of different starting points.

So I think Rick's statement above seriously muddies the waters with respect to the distinction between behavior and a targeted or controlled VARIABLE.

Fred Nickols

[Martin Taylor 2014.11.24.14.41]

I like that date string, all 0,1,2,4 :slight_smile:

[From Erling Jorgensen (2014.11.24 1130 EST)]

I want to give my sense of the import of this discussion, just to clarify in
my own mind where I am coming down on the matter of “control of behavior�? by
an outside living control system. I am not sure whether my opinion is a
majority report or minority report, among CSGNet in general.

Thanks for this. I internote my take on it. I have no problem with first three paragraphs, elided in the quote below.

...

I have a personal preference for saving the word behavior for the lowest level
of the hierarchy, where actual forces in the environment get generated. Every
output above that is a specification for a certain perceptual result. But I
realize I am not going to change how everyone else uses the word behavior.
And so I humbly bow to the community on that.

I don't know how others feel, but I think of "closing the door" or "voting for my favoured candidate" as behaviours, but would seldom think of "tensing that muscle" as a behaviour. So I would take a complete opposite tack from what you say, and would reserve "behaviour" for higher-level action outputs, and not use them for bending a leg or lifting a finger as such, which will be done in exactly the same way in support of many wildly different perceptions being controlled.

I do not think it works to call the different results “aspects�? of behavior.
They could as easily be called “aspects�? of the causal relationships in play,
or “aspects�? of the system concepts.

Of course they could, and "aspect" can be used in many other ways as well, to restrict the reader's attention to what matters about the thing whose "aspect" is mentioned. I am always careful to use "aspects" to make it clear that I am not referring to control of all the behaviour of the other organism. I would prefer to use "one degree of freedom of" rather than "aspect", because in HPCT, controlled perceptions are scalar values. However, "one degree of freedom of" is both wordy and just as liable to misinterpretation as "aspect". One could say "component", but that again has unwelcome connotations. "Element" suggests the smallest indivisible component. Just as a chemist might want to say "aromatic compound" to say that it is built with one or more benzene rings, a benzene ring is not an "element"; carbon is. A benzene ring is a component, but in the PCT context, the aspect of behaviour being considered is being considered as an extract from the rich multilevel stew of ongoing actions, rather than as a contribution to the stew.

To me my "behaviour" in writing this message is action to oppose a disturbance to my perception of the way I would like the word "behaviour" to be used. I write something that is intended to expose my reference value for the use of that word, in the hope it will disturb some higher level perception in a way that moves your reference for it toward mine.

   It is true that such aspects are
variable in that they can take on different values as to what is their
preferred state. But to call that the “behavior�? of those aspects just makes
things way too ambiguous, with a word loaded with outside connotations.

I think that to say "the behaviour of an aspect of behaviour" would be a weird construct, like any other that uses the same letter string in different senses in close proximity. In such a context, I probably would use "the variation of an aspect of behaviour" or something like that, while in other contexts I might say "the behaviour of a waveform is very like a sawtooth" quite happily.

  Far
better in those situations, to specify the aspects themselves that are being
controlled, (using HPCT as a useful guide for categorizing the perceptions in
play.)

Yes, if the context is sufficient to specify an aspect. But when the context is more general, one may want to express that only a one-degree-of-freedom chunk of behaviour is being considered, without need to specify what that is, or at what level, much as one can say that in a circle x^2 + y^2 = r^2 without having to say that x = 2 and y = 3. That's how I have been using "aspect". When someone closes a door following my request "please could you close the door", I am controlling a perception of the "closing the door" aspect of the person's behaviour, not his movements used to do it.

That is my summary on this discussion, which is to say, my perceptions and
preferences.

Thanks for that. Your summary disturbed a perception I control, with a consequent action (an aspect of my behaviour) that I hope will have the effect of reducing the error in that controlled perception.

Martin

[Martin Taylor 2014.11.24.15.12]

[From Fred Nickols (2014.11.24.1430 EST)]

I've not participated until now because I wasn't sure I could contribute. But now I think I can. I also think I'm still with Erling.

I went back through the thread and came across this snippet from Rick:
[From Rick Marken (2014.11.19.1500)]
RM: This is a crucial observation!! It's important to know what control is in order to understand when control of behavior is happening. Controlling behavior is not the same as influencing or causing behavior. Controlling involves bringing a variable, such as the distance between the sheep and the herd in my demo, to a predetermined state, such as "sheep close to herd", and keeping it there, protected from disturbances, which include the movements of the herd and of the sheep itself. Given this definition of control, the distance of the sheep from the herd can clearly be controlled in my demo. Since the distance between sheep and herd depends at least in part on the movements of the sheep this variable is a behavior of the sheep.

Rick says at the very end "this variable is a behavior of the sheep." I can't accept that.

I agree; the sheep is not controlling a perception of its distance between the sheep and the herd, so actions that happen to change that distance are not behaviours of the sheep. The sheep's movements are the only aspect of its behaviour, because that's the only action possibility it has.

   The distance between sheep and herd is a variable.

Known to the analyst, and to the dog, but not to the sheep.

  The sheep's movements are behaviors (describable at various levels of specificity).

The sheep has only one behaviour, moving left and right, so there are no other levels.

  The distance between sheep and herd varies as a function of the sheep's movements but that distance is not a behavior of the sheep. Nor is it an "aspect" of the behavior of the sheep.

Quite so.

Consider me in my Navy days as a Chief. If I happened to shout out "Attention on deck!" in a crowded mess room I wasn't interested in the many and varied behaviors all the sailors presents would have to engage in. I was concerned with the end result of whatever behaviors it took them to come to an erect position, feet together and hands at sides, eyes looking straight ahead.

Wouldn't you say that part of their behaviour was "coming to attention"? I would. I would say that another aspect of their behaviour was controlling a perception of your level of pleasure or displeasure with them, an outcome of the behaviour of coming to attention (and other similar behaviours).

  The position of "at attention" is a variable and one that I could reasonably successfully impose on whomever happened to be in the mess room. How they got to that position was under their control. Being at attention is not an "aspect" of behavior; it is an outcome of the behaviors necessary to reach that position from whatever starting point might be the case and for all those sailors in the mess room there were lots of different starting points.

To me, these are all different aspects of behaviour, as is the behaviour of participating in Navy ritual, an outcome of which might perhaps be a reduction of error in controlling for perceiving "self to be serving my country", an aspect of behaviour.

Martin

···

So I think Rick's statement above seriously muddies the waters with respect to the distinction between behavior and a targeted or controlled VARIABLE.

Fred Nickols

[From Rick Marken (2014.11.24.1640)]

···

Erling Jorgensen (2014.11.24 1130 EST)–

I want to give my sense of the import of this discussion…

My main problem is with the word “behaviorâ€? itself,

RM:My main problem with this discussion is that all I’ve heard is several different arguments aimed at denying the existence of a the phenomenon of control of behavior. It is the phenomenon described on p. 245 of B:CP 2nd edition. What is being demonstrated is the fact that one person, E, can control the finger movements of another person,S (which most people would call S’s behavior) by applying disturbances to a variable (knot over coin) that S is controlling. As long as S controls for the knot being over coin E can control S’s finger, moving it to another target coin: E is controlling S’s behavior.Â

RM: I was all set to have an interesting discussion of this phenomenon from a PCT perspective when  I was inundated with several arguments for why there is no such thing as control of behavior. The arguments agaist the existence of control of behavior were:

  1. There is no such thing as “behavior” so there is no such thing as control of behavior.
  2. Behavior is the control of perception. Since S is controlling his perception – his behavior – E can’t be controlling it.
  3. E is only controlling a perception of Ss behavior, not S’s actual behavior. So there is no such thing as control of behavior. There is only the control of the perception of behavior.
  4. S’s behavior cannot be controlled without S’s cooperation. This means that E cannot reliably control S’s behavior so there is no such thing as control of behavior.Â
  5. When I play tennis I’m not controlling the opponent’s behavior.So there is no such thing as control of behavior.

RM: What all these arguments have in common (except for the last one) is that they are using theoretical concepts (derived fro PCT) to deny the existence of a readily observed phenomenon (control of behavior). I believe theoretical concepts should be used to explain rather than deny  phenomena. And PCT provides the theoretical concepts that  explain the phenomenon of control of behavior. We can’t really talk about what PCT has to say about control of behavior unless we can agree that such a phenomenon exists. But it looks to  me like no one on the “there is no such thing as control of behavior” side of this debate (which apparently is everyone but me;-) is prepared to admit that what is happening in the rubber band demo described on p.245 of B:CP is an example of the phenomenon of control of behavior. Â

RM: But I guess I’ll try one more time and see if we can get anywhere on this. So here’s a question:

RM: In the rubber band demo described on p. 245 of B:CP, how would you describe what E is doing when he places S’s finger on the target coin? I suggest that you do this yourself (with a friend or lover) in the role of both S and E. You could make it somewhat more interesting by having S’s finger trace the outline of a picture on the table rather than just putting the finger on a target coin. What would you say that you are doing when you are E? What is you partner doing when he or she is on the role of E? If you don’t see that E is controlling S’s finger movements and that Ss finger movements are S’s behavior then there is really no hope.Â

RM: If, however, you do see that E is controlling S’s finger positoin then we can start talking about what is going on here from a PCT perspective.Â

Best regards

Rick

Â

because to the user it is

obvious what they mean and to the hearer it is far from clear. It is that

very disconnect that leads me to want to use a word like “movingâ€? or “actingâ€?

or “doing,â€? because their ambiguity about what is meant almost forces a

follow-up query. Queries like, “Well, what part of their moving is being

controlled?â€? or “What result of their acting is being controlled?â€?

My primary interest here is controlling for communicating clearly, and that

includes which words best fit a given model. I am not too concerned about the

word “control,â€? because that is pretty well defined and specified within PCT.

Nor am I especially concerned about some kind of political correctness that

supposedly people should not be controlled by other people. I try to have

science trump ideology, as best I can, within my system concepts. Part of my

expertise – by profession and training as a counseling psychologist, and by

inclination and attachment as the son and grandson of two “word smithâ€? pastors

– is that of communication.

As I use the words, a finger is not a behavior; a location is not a behavior;

distance is not a behavior. Behavior is used for those things in the PCT

hierarchy that we often call Transitions, and if those transitions are timed

in sequence we often call them Events. It includes a lot of the –ing words in

the English language. So then, “movingâ€? is a behavior. But in order to give

any communicative specificity in any given instance, some kind of perceptual

result usually needs to be added. “Moving what?â€? or “Moving what into what

position?â€? It is that communicative necessity that drives us again (thanks be

to Bill Powers) to realize it is the perceptual results that get controlled

for. Behavior gets done, results get controlled.

I have a personal preference for saving the word behavior for the lowest level

of the hierarchy, where actual forces in the environment get generated. Every

output above that is a specification for a certain perceptual result. But I

realize I am not going to change how everyone else uses the word behavior.

And so I humbly bow to the community on that.

I do not think it works to call the different results “aspectsâ€? of behavior.

They could as easily be called “aspectsâ€? of the causal relationships in play,

or “aspectsâ€? of the system concepts. It is true that such aspects are

variable in that they can take on different values as to what is their

preferred state. But to call that the “behaviorâ€? of those aspects just makes

things way too ambiguous, with a word loaded with outside connotations. Far

better in those situations, to specify the aspects themselves that are being

controlled, (using HPCT as a useful guide for categorizing the perceptions in

play.)

That is my summary on this discussion, which is to say, my perceptions and

preferences.

All the best,

Erling


Richard S. Marken, Ph.D.
Author of  Doing Research on Purpose
Now available from Amazon or Barnes & Noble