FW: Demonstration of control of behavior

[_Boris_Hartman]

Well I looked at your demo and I must say that I’m dissapointed. It’s not even close to videos that showed real behavior of real animals (LCS). It looks like that your demo is giving generalized conclusion : on certain behavior of the dog or sheep, we get certain behavior of any of them. As I see it, the demo is showing deterministic relationship between certain disturbances (stimulus) and certain behavior (repsons) which is deterministically repeating. There is no natural varying of behavior as videos show.

So my oppinion about your »demo« is :

1.  Whatever is moving along both axis on first sight I don't see anything resembling to PCT or control. It seems like both pictures are moving simultaneously with each other in opposite directions what I see as »stimulus-respons« (S-R). Sheep determine certain behavior of the dog or dog determine certain behavior of the sheep (It's not clear what's really happening). It's deterministic relationship. And that's contrary to PCT, which imply vatiability in actions on the same »stimulus«.  Whatever is stimulus (dog or sheep) in your demo, it goes along simultaneously with respons of each other.
   

2.  But LCS operate differently. Sheep can hesitate, or can stand in defending position toward a dog or can even attack dog, or can wander away from dog and the herd… As wee saw it on videos. That's how we would expect sheep to behave in natural environment as we saw it on videos. If I wouldn't see videos then I wouldn't know what you wanted to show.  You didn't look and consider all videos about possible sheep behavior on the same dog behavior.  So it looks like to you that sheepdog with certain behavior (that works like stimulus from environment) is "controlling" sheep behavior. But in PCT environment (dog) can not control sheep behavior.
   

RM (once upon a time) : Organisms seem to behave on purpose. Psychologist before Powers had noticed the purposiveness of behavior. They saw, for example, that organisms produce consistent results using highly variable actions. But most psychologist ended up attributing this variability to “statistical noise”; Powers on the other hand, saw it as essential. If actions did not vary, behavioral results would repeat only by chance., fluctuating as a result of the random effects of environmental disturbances. Instead, actions vary to compensate for the effects of disturbances, producing consistent results in an incosistent world – a proccess called “control” .

RM (once upon a time) : Powers built a model of behavior based on control theory. Control theory is the wrong model of behavior if behavior is evoked motor output. But is the right model of behavior if behavior is control.

RM (once upon a time) : To understand the behavior of a living control system, the observer must learn what perceptions the system is controlling; what reference images the system is trying to match.

RM (once upon a time) : It takes a while to understand that control system compensate for disturbances rather than respond to stimuli; that stimuli are controlled and not in control ; that living control system control and cannot be controlled.

HB :

First I must say that I’m always impressed when I see you in the role of “dr. Jekyll”" as in above text. And I’m always depressed when I see you acting like “Mr. Hyde”. So I don’t understand why you want to act as “Mr. Hyde” ?

I don’t see in your demo how sheep is producing “highly” consistent results using highly variable actions. I see only how determined “disturbances” (dog or sheep) which work as “stimulus”, produce determined "behavior"or evoke “wanted” motor outputs. They act simultaneously as “stimulus-respons” all the time repeating the same behaviors on the same stimulus, thus neglecting the essentials of PCT, which you described as variability and LCS can not be controlled…

So if I try to make PCT interpretation of your demo. Dog as disturbance to some “controlled variable” of the sheep should elicitate some purpose in sheep not “controlled behavior” as a result. So I think that demo should look like that sheep is not acting on the dog’s same “stimulus” (disturbance) but that sheep can vary behavior to the same dog behavior. As we saw it on the videos. Sheep has to choose goals for behavior not dog (environment) instead of sheep, if you are right with your writings above. So sheep has to control it’s behavior not dog. And that should be visible from your demo as it is on all videos if you want it to be PCT demo.

3.  And I don't understand how can dog "control sheep behavior" if they are going in different directinos. Where do you see any control here ? First video showed that sheepdog somehow "intersept" sheep. So the moving of the "dog" and sheep should be different, if you tryed to demonstrate what we saw on the video (previous discussions). It's obvious that you neglected some other videos which could show you how to do the right PCT demo with sheep's varying behavior. But you acted like perfect PCT. Just took what was needed for your purposes. And I think you missed and that's why I think your demo is missleading showing "S-R".
   

4.  And can you show me in your demo, how sheep or dog are "protected from disturbances"  if they both control… ???
   

&nbspp;

5.  And I still don't understand how you define "control" in the sense of "dog controlling sheep" ?  What dog is doing to sheep : applying disturbances (stimuli) or setting references for the sheep behavior ?
   

As I see your demo, you used PCT wording to prove “S-R” logic.

Best,

Boris

···

On Fri, Nov 14, 2014 at 10:11 PM, Boris Hartman boris.hartman@masicom.net wrote:

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: Sunday, November 09, 2014 1:06 AM
To: csgnet@lists.illinois.edu
Subject: Demonstration of control of behavior

[From Rick Marken (2014.11.08.1610)]

I finally completed a demonstration of control of behavior that was motivated by a conversation we had on the net some time ago. It’s at

http://www.mindreadings.com/ControlDemo/BehavioralControl.html

It took a while to do it because it was a bit difficult to figure out how to make a simple demonstration of control of behavior in terms of a sheepdog controlling a sheep. The route I took may not accurately represent the variable controlled by the sheep but the demonstration does show that it’s possible to control the behavior of another control system sans conflict.

I’d appreciate getting any comments/suggestions/ corrections that you might have regarding this demo.

Thanks.

Best

Rick

–

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

Now available from Amazon or Barnes & Noble

–

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

Now available from Amazon or Barnes & Noble

[rsmarken]

[From Rick Marken (2014.11.15.1340)]

Quick note to CSGNetters: For some reason when you just “reply” to posts from CSGNet you are likely to be responding only to the address of the person who sent the post. I have no idea why CSGNet works this way now but I think the best way to deal with this, for now, is to reply to CSGNet posts that you intend for CSGNet with “reply to all”.Â

···

On Fri, Nov 14, 2014 at 10:11 PM, Boris Hartman boris.hartman@masicom.net wrote:

Â

BH: Well I looked at your demo and I must say that I’m dissapointed. It’s not even close to videos that showed real behavior of real animals (LCS).

RM: The demos of PCT that have been developed by Bill Powers, myself and others are meant to demonstrate the principles of control in a very simplified situation. My demo includes only those variables and functions that are relevant to understanding how one control system (the sheepdog) can control the behavior (outputs) of another control system. This means leaving out a lot of  complicating detail, such as the hundreds of other control systems that would make the behavior of the figures more realistic.Â

BH: It looks like that your demo is giving generalized conclusion : on certain behavior of the dog or sheep, we get certain behavior of any of them. As I see it, the demo is showing deterministic relationship between certain disturbances (stimulus) and certain behavior (repsons) which is deterministically repeating. There is no natural varying of behavior as videos show.

RM: It’s, it looks like the movements of the sheepdog are deterministically affecting the movements of the sheep. But that’s the behavioral illusion. The sheep is actually controlling for keeping the image of the dog close to the grass shoot and the movements of the dog are a disturbance to that variable, which is being compensated for by the movements of the sheep. If the sheep were not controlling for keeping the image of the dog near the grass shoot, the movement of the dog would have no effect on the movements of the sheep; the dog would then not be able to control the sheep’s position (keeping it near the herd).Â

Â

BH:  So my oppinion about your »demo« is :

Â

1.     Whatever is moving along both axis on first sight I don’t see anything resembling to PCT or control.

RM: Right, you can’t easily see that the sheep is keeping the variable it is controlling (distance of dog image from grass shoot) at its reference while you (as the dog) are keeping the variable you are controlling (distance of sheep from herd) at its reference. You can really only see this clearly in the graphs plotted at the end of a run.

BH: It seems like both pictures are moving simultaneously with each other in opposite directions what I see as »stimulus-respons« (S-R).

RM: Right. It looks like S-R but it is actually disturbance resistance. Again, it’s the behavioral illusion.

BH: Sheep determine certain behavior of the dog or dog determine certain behavior of the sheep (It’s not clear what’s really happening). It’s deterministic relationship. And that’s contrary to PCT, which imply vatiability in actions on the same »stimulus«. Whatever is stimulus (dog or sheep) in your demo, it goes along simultaneously with respons of each other.

RM: It’s actually not contrary to PCT at all; it is PCT. Both you (as dog) and the sheep are control systems. There is the appearance of S-R when the dog’s movements disturb  the variable controlled by the sheep. But there really is no S-R at all in this demo; it’s all control.

BH: 2.     But LCS operate differently.

RM: No, they operate just like the control systems that are simulated in this demo.

Â

BH: Sheep can hesitate, or can stand in defending position toward a dog or can even attack dog, or can wander away from dog and the herd…

RM: Sure they can. But this demo is a simplification which assumes that the sheep is only controlling for keeping the image of the dog near the grass shoot. Real sheep control for lots of other things too but those details are eliminated from the demo to demonstrate how control of a control system works in the simplest situation.

BH: As we saw it on videos. That’s how we would expect sheep to behave in natural environment as we saw it on videos. If I wouldn’t see videos then I wouldn’t know what you wanted to show.Â

 RM: I’m not trying to imitate the behavior seen in the videos. I’m trying to show how a control system (like the sheepdog) can control another control system (the sheep). In real life there are a lot more complicated things going on but the demonstration explains what I believe is the essential observation in the video: how the sheepdog can control the location of a sheep  by moving around. It does it by disturbing a perceptual variable that is controlled by the sheep.Â

BH: You didn’t look and consider all videos about possible sheep behavior on the same dog behavior. So it looks like to you that sheepdog with certain behavior (that works like stimulus from environment) is “controlling” sheep behavior. But in PCT environment (dog) can not control sheep behavior.

RM: In fact, my demonstration shows clearly that in PCT Â a dog can indeed control the behavior of a sheep. Simply saying that “in PCT a dog cannot control sheep behavior” is not very convincing. It would be better if you showed why this couldn’t be done.Â

RM: Now you give a bunch of quotes from me that you apparently approve of. I think you give these quotes to show that they are inconsistent with the idea that it is possible to control the behavior of a living control system. Perhaps the most relevant one is this:

RM (once upon a time) : It takes a while to understand that control system compensate for disturbances rather than respond to stimuli; that stimuli are controlled and not in control ; that living control system control and cannot be controlled.

RM: That last sentence is wrong. While it’s true that living control systems cannot be controlled by stimuli (because stimuli are not control systems) they certainly can be controlled by other control systems.Â

Â

BH: I don’t see in your demo how sheep is producing “highly” consistent results using highly variable actions.

RM: The highly consistent result is shown in the final graph, which shows the variable controlled by the sheep staying right at it’s reference specification. It does this by moving back and forth (the variable actions) in order to keep the controlled variable at its reference.Â

Â

BH: I see only how determined "disturbances" (dog or sheep) which work as “stimulus”,  produce determined "behavior"or evoke “wanted” motor outputs. They act simultaneously as “stimulus-respons” all the time repeating the same behaviors on the same stimulus, thus neglecting the essentials of PCT, which you described as variability and LCS can not be controlled…

RM: As I said, the variations in the sheep’s position is the “variability” that is producing the consistent result of keeping the controlled variable at the reference. This is perfectly consistent with PCT because the sheep in the demo is a perceptual control system.Â

Â

 BH: So if I try to make PCT interpretation of your demo. Dog as disturbance to some “controlled variable” of the sheep should elicitate some purpose in sheep not “controlled behavior” as a result.

RM:Change everything after “of the sheep” to  “that creates error that drives the sheep’s movements that compensate for the disturbance” and you’ve got it.Â

Â

BH: So I think that demo should look like that sheep is not acting on the dog’s same “stimulus” (disturbance) but that sheep can vary behavior to the same dog behavior.

RM: I don’t know what this would accomplish for the sheep. But if you write a model of the sheep that behaves this way it would help me understand what you mean. Â

Â

BH:As we saw it on the videos. Sheep has to choose goals for behavior not dog (environment) instead of sheep, if you are right with your writings above. So sheep has to control it’s behavior not dog. And that should be visible from your demo as it is on all videos if you want it to be PCT demo.

BH: 3.     And I don’t understand how can dog “control sheep behavior” if they are going in different directinos.

RM: All you have to see is that you (the dog) can keep the sheep aligned with the herd. That’s the dog controlling the sheep’s behavior. How it does it is by moving appropriately to disturb the sheep’s controlled variable in such a way that the sheep moves in the desired direction. The dog moves opposite to the sheep to get the desired result because the sheep has to move opposite to the dog (because of the optics of the situation) to get the image of the dog back close to the grass shoot.Â

HB: 4.     And can you show me in your demo, how sheep or dog are "protected from disturbances" if they both control… ???

RM: That’s shown in the graphs are the end of the demo. The controlled variables of both the sheep and the dog are shown along with their references. The fact that the plot of the controlled variables for each stay closely aligned with their respective references shows that their actions are keeping these controlled variables close to their reference states, protected from the effects of disturbances; the disturbance to the sheep’s controlled variable is the movement of the dog (you) and the disturbance to the dog’s controlled variable is the movement of the herd as well as secular variations in the sheep’s reference for the state of its controlled variable.Â

5.  BH: And I still don’t understand how you define “control” in the sense of “dog controlling sheep” ? What dog is doing to sheep : applying disturbances (stimuli) or setting references for the sheep behavior ?

RM: The dog is controlling the position of the sheep relative to the herd. The measures of control (presented at the end of the demo) show how well this is being done. The dog does this by moving left or right, which is a disturbance to the perception (of the dog relative to the grass shoot) Â controlled by the sheep.Â

 BH: As I see your demo, you used PCT wording to prove “S-R” logic.

RM: I’m confident that you will see that I did nothing of the kind once you have studied the situation a bit more thoroughly. What I have done is use a PCT model of a sheep to show that another control system (you, in the role of a sheepdog) use variations in a stimulus (the dog’s position) to control the sheep’s position. The demo also shows that the behavior of a control system – in this case, the movements of the sheep – can appear to be a response to stimuli if one ignores the fact that the system is actually controlling some perception. And in this demo it is hard to see that the sheep is controlling a perception (of the distance between the image of the dog and the grass shoot), which is why I plot the graph at the end of a run, which shows that the sheep has been controlling a perception relative to a variable reference. The demo shows not only that the behavior of a control system can be controlled, it also shows how easy it is to mistake a control system for an S-R system.Â

BestÂ

Rick

Â

Â

Best,

Â

Boris

Â

Â

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: Sunday, November 09, 2014 1:06 AM
To: csgnet@lists.illinois.edu
Subject: Demonstration of control of behavior

Â

[From Rick Marken (2014.11.08.1610)]

Â

I finally completed a demonstration of control of behavior that was motivated by a conversation we had on the net some time ago. It’s atÂ

Â

http://www.mindreadings.com/ControlDemo/BehavioralControl.html

It took a while to do it because it was a bit difficult to figure out how to make a simple demonstration of control of behavior in terms of a sheepdog controlling a sheep. The route I took may not accurately represent the variable controlled by the sheep but the demonstration does show that it’s possible to control the behavior of another control system sans conflict.Â

Â

I’d appreciate getting any comments/suggestions/ corrections that you might have regarding this demo.Â

Â

Thanks.Â

Â

BestÂ

Â

Rick

–

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

Now available from Amazon or Barnes & Noble

Â

–

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

Now available from Amazon or Barnes & Noble

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

[_Boris_Hartman]

Rick,

the main problem in your explanation of sheepdog controlling the behavior of the sheep« is manipulation and »cheating«. So I will not answer all your asnwers as it’s useless beacuse your answers are arranged to suit your purposese, so you are varying your output to »reach your goals – images« with your demo.

<

In your demo you are supposing that »dog controls behavior of the sheep«, and you say you are obviously showing that sheep behavior is controlled »by dog’s behavior (environment)«. So by determining the sheep output on environmental stimuli (dog’s behavior) you neglected »control of perception in sheep« and possibility that sheep is that who is setting references for it’s behavior« not do ,because output (behavior) is known in advance. So sheep is not that who »chooses« behavior (setting references),  dog is in your version of »control«.

But dog can not control sheep behavior, because dog can not predict what behavior (reaction) sheep will choose on the dog’s stimuli. If you would consider sheep’s control then you would have to show how sheep can vary references and »controlled variable« to establish desired perception. And that is not always »behavior« which dog »wishes« as you showed in your determined demo. With determining sheep output from the environment (dog), you neglecte »sheep’s control« and you make an ilussion of »dog control the sheep behavior« by evoking »wanted« motor output. But that is possible only in behaviorism with S-R logic, where »stimuli deterministically control behavior«, evoke determined output. And that’s what your demo shows.  If you are 100% predicting output (behavior) on certain stimuli, you are totaly neglecting control process in LCS, and than we can probably really talk about control of LCS.

But tell me something, if you were running from the dog that tries to attack you, would you think of »eating baken chicken«. I think not. But you can probably try it.

So when the dog reduce the distance to the sheep, sheep is probably not thinking on »eating a grass«, but on it’s safety as you would. So »controlled variable« sheep »eating a grass« when dog is moving« toward it is practically impossible. So I advise you to see videos again and use them as basis for analysis, if you want to have at least a little scientific credibility.

The only real bases in this moment to analyse dog-sheep relationship are videos, unless we go to see it in the nature how dogs »control sheep behavior«. So we have to come as much as possible close to explanation what is happening on videos (in reality). But you didn’t see all the videos, to see what really happens. You used your imagination what should be happening to suit your purposes, and so you had troubles with finding sheep’s »controlled variable« as you said before. I understand why you had troubles. Becuase you had to find suitable controlled perceptions for the sheep so the demo would work, to realize your manipilation. There are many controlled perceptions for the sheep, seen on the video , but non is what you demonstrated in your demo. So your demo is imagined »construct« which suit your manipulative purposes.

As I said, the video is the only relevant bases for analysis, so we have to respect what is happening in »reality«, not in your mind as »perceptual illusion«.

1.  So the dog's controlled perception seen from the video is »distance of the sheep to the spot (herd)« and
   

2.   Sheep's controlled perceptions (or images as you said) are (seen on videos) :
   

•      »distance to the dog«,
  

•      »going away from the herd"
  

•      "defending from the dog"
  

•      "attacking the dog".
  

We must clearly see how sheep is controlling different perceptions on the same stimuli coming from environment (dog’s behavior), so to see that sheep is varying output to establish wanted perception or reliaze it’s different purposes (sheep is setting the references for it’s behavior, not dog). So we have to see that sheep is in »control« not »dog controlling sheep behavior« in relation »determined stimuli – determined behavior«. And again there is no »control of behavior« in PCT, it’s only control of perception. »Control of behavior« exist only in behavirosm and self-regulation theories.

In your demo you didn’t use any of those controlled perceptions for the sheep which are “really happening” because your manipulation probably wouldn’t work, as it would be obvious that sheep is controlling for it’s perception and dog for it’s perception which can vary. So why you don’t sit down again to your computer and make demos on the bases of video evidence for all the cases of sheep’s controlled pereptions on the same stimuli from environment (dog’s behavior approaching to the sheep). If you are talking the truth, you will prove it for any sheep controlled perception, not just for your choosen. I’m sure that we’ll have to see how dog and sheep are both controlling for it’s perceptions and there is no “control of behavior of LCS”. Anyway there si no “control of behavior”. It’s just “control of perception”.

And when you are sitting by the computer, making those demos, would you be so kind and demostrate us how student is not “controlled by the environment” (not reacting on stimuli) as you showed it to Barb. So demosntate us, when the dog is moved allong the axe, how student is not moving deterministically with the dog’s behavior (environmental stimuli). Let us say that dog is controlling perception of "attacking the student«. Then make us a plot of graph showing how »dog (environment) is not controlling student (LCS) behavior«.

So :

1. dog’s controlled perception is “attack the student” or »bite the student« and

2. student “controlled perception” is keeping the »safe distance to the dog" or any other controlled perception that would suit student reference »not to be bite by the dog« or other »safety perception«. The only reference that student can’t use in this situation is »keeping image« of eating.

Show us how student behavior is not »controlled by dog« or any other stimuli from environment ?

And please use all your advises you used to support Barb’s finding : »We have a weekly meeting, and today’s guest was from the counseling center (as a fly to the web, in my mind)! Of course, when she started talking about students reacting to environmental stimulus, I cringed. This affects me like chewing on aluminum foil«.

This is the same as sheep (LCS) is “not reacting to the environmental stimuli (dog)” with some in advance choosen behavior. Sheep and student are not controlled by the environmental stimuli or are not »reacting« to environmental stimuli. Barb’s conclusion is right for all LCS.

RM :

What we often notice about behavior are responses to disturbances to controlled perceptions and fail to notice that the responses only occur because people are controlling those perception. So we see a person get mad when someone butts in a line but fail to notice (or think it is too obvious to mention) that this happens only because the person has the goal of keeping their place in line. People who don’t care about keeping their place in line will not get mad given the same butting in line (“stimulus”).

HB :

Yuppi, you got it. »BEHAVIOR IS NOT RESPONS TO DISTURBANCES TO CONTROLLED PERCEPTION« So you see it’s your »perceptual illusion« that dog is »disturbing controlled perception« of the sheep and that sheep’s behavior is »being controlled« by a dog. You failed to notice that sheep behavior only occurs because sheep is controlling it’s perception, not being »controlled by a dog«. So you admitt it yourself that you were in »behavioral iluusion« not me. Or you want to contradict yourself.

RM :

Or we see increased purchases of an item when it is offered with an “incentive” like a reduced price but fail to notice (or ignore the fact) that this only happens if the person has the goal of paying as little as possible. A person who doesn’t care about the price of an item will not appear to be caused to buy it given the same incentive (“stimulus”).

HB :

So show us also with your demo how student is not “controlled by environment” with commercials by bying food with reduced price and so on.

If try to conclude. Dog is controlling it’s own perceptions, sheep is controlling it’s own perceptions and student is cotnrolling it’s own perceptions. It is as it has to be. No »control of behavior« because there can’t be any control of behavior. It’s behavioristic, S-R- term and you are using it in your demo. If anybody is living in any iluusion, than you are, manipulating demos and others to suit your purposes of proving your stand point. You live in some kind of percpetual illusion that LCS can be controled by environment although you said it very clearly :

RM (once upon a time) : …LCS control and can not be contrrolled.

HB :

And probably you will notice one thing more. When you are talking to Barb, you talk as dr. Jekyll, but when you are talking to me, you talk as »Mr. Hyde«. Two diferent approaches, two different parts of your personality, sorry LCS. So be always »dr. Jekyll«.

Best,

Boris

P.S. And if you once again change my initials from HB to BH (as that is the same as Bob Hintz) I’ll come to America and buy you a bear JJJ. This is a »threat« (attempt of control) as when you’ll drink one bear I’ll »force« you as »environmental stimuli« to drink some more. I will »control you«, of course if you will want so. JJ. It’s up to you what you will do on environmental »stimuli«, isn’t it ? But invitation stands. As you see I can »try to control« you with different means, but it’s up to you whether you will »accept control« or not.  So is with other LCS (sheep for example, although it is probably not drinking bear) JJ.