VS: Conflict (was ... long live William T. Powers

[Eetu Pikkarainen 2017-07-04]

(Home again and I can comment between the original message. During summer I will post less and irregularly.)


[From Rick Marken (2017.07.01.1040)]

RM: I am certainly willing to be shown that I am wrong.
Are you?

EP: Oh yes, it was only way to learn PCT. Reading B:CP showed that I had modelled action very defectively. Then Martin, Boris and many others – also you – helped me me to see that I had got something wrong from my reading.

Isn’t it always so?

(EP: Nice sheepdog demo!)

RM: You are right, Eetu, that in the CSGNet debates that have happened since Bill passed away I just haven’t been convinced that I was wrong.
That’s because words alone don’t do it for me. One of the reasons I was attracted to PCT-- possibly the main reason – is that Powers demonstrated what he said with clear, easy to understand empirical tests and computer (or mathematical) models (see
the Experimental Methods chapter of B:CP).

EP: OK, now I understand better. This somewhat reminds me of some old stories where the one challenged to duel may select the weapons
:blush: . For you words are only for elegant prose? I disagree. The most fundamental questions are answered with words. I mean logic. Of course it can formalized and
expressed by letters and logical symbols, but basically also by words. Behind every empirical or mathematical model there is some logic. The core point of logic is consistency. For me the fundamental demand for both scientific and practical ethics is that
we use our words consistently (no matter if it leads to inelegant prose). Only if this demand is fulfilled there can be any reason to appeal empirical, mathematical or computer models and methods.

RM: If you (or anyone else) could come up with such demonstrations of your claims abou t PCT – claims with which I disagree – then I might be convinced that I am wrong. Off the top of
my head, here are some of the claims with which I have disagreed that I could be shown to be wrong about by appropriate empirical tests and/or computer/mathematical models:

  1. Control of behavior is impossible

  2. A mass-spring system does something like controlling.

  3. There is no conflict involved when a person applies a disturbance to a controlled variable either to do the TCV or to demonstrate control of that variable

  4. Only a perception is controlled by a control system

  5. Control of the environmental correlate of a controlled perception (the CEV) is a side effect of control of that perception

  6. The power law tells us something about control of curved movement.

  7. Control of a program perception has nothing to do with the idea that people are doing the TCV on each other when they communicate.

EP: Interesting list, but somewhat strangely ordered and formatted issues. I cannot say much about 2, 6 and 7. Mostly I agree with 4. The 1 seems like logical consequence of 4 and 5 seems like an explanation what is the
relationship between 4 and 1.

EP: Why I claim that a (living) control system controls only its perception, not behavior or CEV? The definition of control is “Achievement and maintenance of a preselected
perceptual state in the controlling system, through actions on the environment that also cancel the effects of disturbances (B:PC, 296) [italics by me].� The whole PCT is based on this definition! According to it control means (by the concept of control
we mean) just and only the “achievement and maintenance of a preselected perceptual state in the controlling system�. What happens in and to the environment and behavior is something else than control. They can be causes or effects of control but not
control itself.

EP: Why behavior or CEV are not controlled? Because there are no preselected reference states for them. They are and change and are changed to what they must be for the controlling system’s perceptual state to be near
it’s preselected reference state – the only preselected state there is. What you call a reference state of the CEV is not preselected but determined along the process. The cancelling of disturbances alone does not make control.

EP: If control system is not controlling the CEV or behavior, then what it is doing for them? Just for this concept (which is different concept from control) we need a word. And guess what? I reread B:CP and there it
was in Glossary:

“ADJUST: To affect in such a way as to achieve a preselected perceptual result (B:PC, 296).�

So a controlling system is adjusting the aspect of its environment when and because it is
controlling its perception of that aspect.

EP: As far as I understand, this kind of change of words does not affect or threaten your empirical-mathematical-computer research in any way. It just would make your logic more consistent and lessen the confusions of
your readers. Hopefully?



Please, regard all my statements as questions,

no matter how they are formulated.

RM: I imagine there are more but that will do for now. I believe all of these claims are inconsistent
with empirical fact and/or the PCT model of purposeful behavior. But, again, an empirical demonstration or computer.mathematical model could show that I am wrong. And I would accept that, just as I accepted the fact that I was wrong about behavior being
controllable even when the system’s reference specification for the controlled variable is autonomously varying.

Best regards



30.6.2017 9.00 ip. Richard Marken rsmarken@gmail.com kirjoitti:

[From Rick Marken (2017.06.30.1100)]

Dag Forssell (2017.06.29.2245 )

DF: I have to say that I appreciate your tone and effort. Comments inserted.

RM: And I have to say that I’m kind of sorry I brought the whole thing up.As I recall, this all started (for me, at least) when Bruce Nevin said that the TCV always involved conflict. At first I said it didn’t but then realized that it
did, at least for a moment, when the disturbance used in the TCV is applied by a person. This struck me as an interesting implication of PCT – one that I had never made before – but not a terribly important one since disturbances are rarely applied by
the person doing research involving the TCV and, as far as I know, no one on CSGNet is doing research involving the TCV anyway.

RM: Nevertheless, I thought that analyzing situations like the TCV and Dag-Bill demonstration, where one control system causes a transient disturbance to a variable controlled by another person, would be a good exercise in seeing how to
apply the PCT model. But the resistance to my analysis has been so intense that I think, given it’s lack of practical importance, I’ll just drop it and look forward to the next thing about PCT that everyone can disagree with me about.

RM: I am rather amazed by how much PCT I have apparently forgotten since Bill passed away. Or maybe I just never really got it and Bill was too polite to mention it to me. Ah well, there’s always the Trout Quintet.



[From Rick Marken (2017.06.29.1120)]

Dag Forssell (2017.06.28.11:30 PDT)

DF: I think your most confusing and internally inconsistent paragraph below would never have been written if you would study and focus on the physics involved instead of impenetrable and too often erroneous math.

RM: This is an improvement over saying it’s “utter nonsense”. But I think that, rather than insulting my physics and math skills, it would improve the tone of the discussion if you would just say what was confusing and internally inconsistent about my analysis,
how math got in the way and why an understanding of physics would have helped.Â

RM: But let try explaining my analysis again, throwing in what little I know about the physics of the situation, and maybe you can tell me exactly where the confusion and inconsistencies are. I think the main bone of contention is whether or not there was
a conflict involved when Bill pulled down on the steering wheel. I believe there was a conflict, if only for a brief moment; I think you believe there was not.

Correct :slight_smile:

I think there was a conflict because at the instant Bill applied the pull he and you had different references for the same controlled variable: the position of the wheel.

I’ll grant you that I had a reference for the position of the steering wheel as my way of keeping the car in its lane, but I do not agree that Bill had a reference for the position of the steering wheel. Somehow, your notion that the model makes this necessary
is to put the model ahead of reality.

I hope you’ll agree that when this happens – when two control systems have different references for the same controlled variable – no matter how long it lasts, there is a conflict.Â

RM: You say there was no conflict because “Bill set a reference for applying a force to the steering wheel. All he wanted was to apply a disturbance”. My argument is that Bill wanted to do more than just apply a disturbance. I believe he wanted to apply the
disturbance in order to demonstrate how well you were controlling. As you said in an earlier post:

Again correct. But your understanding of the model requires you to add a non-existent reference to Bill, accusing him of wanting to turn the wheel to another position. You deny that a human being can apply a force; tense some muscle, without an expectation
that of movement, much less a specific movement.

Dag Forssell (2016.06.27.0920 PDT)Â

DF: Bill asked politely if he might demonstrate a good control system to me.

RM: A good control system controls some variable – the controlled variable – and controls it well, keeping it in a reference state, protected from disturbance. So demonstrating a good control system means demonstrating a control system that controls some
variable well.

That would be myself.

He was demonstrating this by applying an abrupt and strong disturbance to that variable and showing that the disturbance had little effect.

Where do you get “abrupt and strong”? Especially the abrupt part?

Since the disturbance was a pull force applied to the steering wheel and since physics tells me that, without a driver present, such a pull would have moved the wheel quite a bit, I conclude that Bill was demonstrating control of steering
wheel position by showing that the pull had little effect on steering wheel position; that steering wheel position was protected from that disturbance.


RM: In order to apply a disturbance to steering wheel position Bill had to apply a force to the steering wheel that would change its position if there were not a good control system present. In order to do this, Bill had to control for
producing a force that would change the position of the steering wheel if the driver were not present;

Yes indeed. Bill had to control for producing a force.

that is, Bill had to control for changing the position of the steering wheel.

Balls. He had only to apply force. Your deduction does not follow from the physics of the situation. It may follow from your idea of some hierarchy that exists only in your imagination.

This means Bill had to have set a reference specification for the steering wheel being in a different position than it was currently in.

He did not.

Indeed, he apparently set a reference specification for the steering wheel being in a very different position since he apparently pulled down on the wheel with a force that would have resulted in a large change in position if a good control
system were not present to keep that from happening.

My, his reference signal is devious. Large change in position is what he wanted?

RM: So at the instant Bill pulled down on the wheel, he and Dag had different references for the same controlled variable.

We did not. Position and force are not the same thing. Just like space and time are not the same thing.

So at that instant there was a conflict. Of course, Bill abandoned his control of wheel position almost immediately so the conflict lasted only a moment and would not even be noticed, which is why it might not seem like there was a conflict.
But there was, and the conflict would have “presented” rather dramatically if Bill had maintained his reference for the changed position of the wheel.

Luckily, he did not have such a reference in the first place.

In that case, Bill and Dag would have eventually been pushing with equal and opposite force on the wheel so Dag’s control of the wheel would have been lost, making it impossible for him to vary wheel position to compensate for disturbances
to the position of wheel and, hence, to the position of the car on the road; control of the position of the car on the raod would also have been lost and we wouldn’t be having this discussion today.

Well, we do because there was no conflict ever. You are making it up, sir!

RM: One last point. I conclude that the good control system Bill was demonstrating was the one controlling steering wheel position and not position of the car in its lane. This is because the disturbance was applied directly to wheel position.

Fair enough.

Of course, failure to have good control of wheel position would have consequences on the car’s position in the lane (as noted above); but control of the position of the car in its lane was not directly demonstrated because a disturbance
was not directly applied to that variable. Indeed, it would have been impossible for Bill to apply a disturbance to demonstrate control of car position because this would have required applying a disturbance, such as a lateral force on the car itself, that
directly affected this variable and even Bill was incapable of doing that.Â



RM: Think about how you would model this situation using PCT. I think you would find that you would have to include in the model of Bill a control system controlling for steering wheel position.

I don’t think so. Why don’t you lay out the physics of that?

The model might, indeed, include a system controlling applied force but I think that system would serve as the means used by the steering wheel position control system to bring (or try to bring) the position of the steering wheel to the
specified reference state. So thinking in terms of models we can see that Bill may, indeed, have set a reference for applying a particular force, but he did this as the means of bringing steering wheel position to a new reference state – that is, he applied
the force as the means of “disturbing” steering wheel position.

More nonsense. You do have a way of making the simple very complicated indeed.

Over and out.

Best, Dag

Richard S. Marken

"Perfection is achieved not when you have nothing more to add, but when you

have nothing left to take away.�

                            --Antoine de Saint-Exupery

Richard S. Marken

"Perfection is achieved not when you have nothing more to add, but when you

have nothing left to take away.�

                            --Antoine de Saint-Exupery