Behav. illusions

CSGnet Archive CSG2018
41

(From A. Matic 20.3.2018)

(from Rick Marken 2018-03-19_11:51:06)
RM: But as I have noted before, the power law is a different kind of
behavioral illusion than the S-R version of the illusion described in
Powers(1978).

···

-----
AM:
Oh. Sure, I agree it is not the kind of illusion described in Powers(1978).
Glad we agree on that one.

I don't think we should expand the meaning of the term "behavioral illusion"
to include all incorrect inferences about control. There are different kinds
of incorrect inferences, like assuming that a side effect is the controlled
variable. That one can stay in its own category.

42

[Rick Marken 2018-03-20_18:33:02]

(From A. Matic 20.3.2018)

RM: But as I have noted before, the power law is a different kind of
behavioral illusion than the S-R version of the illusion described in
Powers(1978).
-----
AM: Oh. Sure, I agree it is not the kind of illusion described in Powers(1978).
Glad we agree on that one.

AM: I don't think we should expand the meaning of the term "behavioral illusion"
to include all incorrect inferences about control. There are different kinds
of incorrect inferences, like assuming that a side effect is the controlled
variable. That one can stay in its own category.

RM: Actually, I'd be happy to dispense with the term "behavioral illusion" altogether. I think it can be taken as an insult. I think all of these "illusions" and "side-effects" can be safely placed in a single category called "irrelevant side-effects" of control. This is certainly true of the S-R illusion described in Powers (1978); the observed relationship between S and R is a side effect that is irrelevant to the fact that a variable (the contrlled variable) is being kept under control. Same is true of "operant behavior"; the observed relationship between reinforcement and responses is irrelevant to the fact that a variable (reinforcement rate) is being kept under control. And the same is true of the power law; the observed relationship between curvature and velocity is irrelevant to the fact that a variable (the trajectory itself) is being kept under control.
RM: So what do you think? I propose that we stop telling behavioral scientists that they are succumbing to an illusion and just say they are paying attention the the wrong things: irrelevant (but compelling) side effects of control rather than the variables being controlled.
Best
Rick

···

--
Richard S. MarkenÂ
"Perfection is achieved not when you have nothing more to add, but when you
have nothing left to take away.�
                --Antoine de Saint-Exupery

43

[From A. Matic 21.3.2018]

(Rick Marken 2018-03-20_18:33:02)
RM: So what do you think? I propose that we stop telling behavioral scientists
that they are succumbing to an illusion and just say they are paying attention
the the wrong things: irrelevant (but compelling) side effects of control
rather than the variables being controlled.

···

---

AM:
Well, there are already two camps, one looking at the power law as part of the
plan, the other looking at it as a side effect. A lot of computational models
are published in the 'plan trajectory -> inverse kinematics & dynamics ->
muscle activation' camp, and many pieces of evidence are believed to support
this reasoning. The other camp are mostly equilibrium point guys, who maintain
that forces are not planned, and also publish models. Like that Gribble and
Ostry paper.

Going by the 'extraordinary claims, extraordinary evidence' line of thinking,
if there is something to hierarchical control structures, then we should be
able to build a model that can trace an ellipse shape (or other shapes)
without being told at what speed it should go at which point, and it must
reproduce the power law as a side-effect. (and fail to produce the power law
in the situations where humans don't produce it).

44

[From Fred Nickols (2018.03.21.0702 ET)]

I think what you propose, Rick, is better than telling them they are succumbing to an illusion. However, telling them they are paying attention to the wrong things isn’t much better. In the first instance, they’re being fooled; in the second, they are misguided. Neither is likely to sit well. So how about saying simply that there’s more there than meets the eye.

Fred Nickols

···

From: Richard Marken rsmarken@gmail.com
Sent: Tuesday, March 20, 2018 9:34 PM
To: csgnet@lists.illinois.edu
Subject: Re: Re: Re: Behav. illusions

[Rick Marken 2018-03-20_18:33:02]

(From A. Matic 20.3.2018)

RM: But as I have noted before, the power law is a different kind of
behavioral illusion than the S-R version of the illusion described in
Powers(1978).

AM: Oh. Sure, I agree it is not the kind of illusion described in Powers(1978).
Glad we agree on that one.

AM: I don’t think we should expand the meaning of the term “behavioral illusion”
to include all incorrect inferences about control. There are different kinds
of incorrect inferences, like assuming that a side effect is the controlled
variable. That one can stay in its own category.

RM: Actually, I’d be happy to dispense with the term “behavioral illusion” altogether. I think it can be taken as an insult. I think all of these “illusions” and “side-effects” can be safely placed in a single category called “irrelevant side-effects” of control. This is certainly true of the S-R illusion described in Powers (1978); the observed relationship between S and R is a side effect that is irrelevant to the fact that a variable (the contrlled variable) is being kept under control. Same is true of “operant behavior”; the observed relationship between reinforcement and responses is irrelevant to the fact that a variable (reinforcement rate) is being kept under control. And the same is true of the power law; the observed relationship between curvature and velocity is irrelevant to the fact that a variable (the trajectory itself) is being kept under control.

RM: So what do you think? I propose that we stop telling behavioral scientists that they are succumbing to an illusion and just say they are paying attention the the wrong things: irrelevant (but compelling) side effects of control rather than the variables being controlled.

Best

Rick

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

45

[From Fred Nickols (2018.03.21.0739)]

I think what you propose, Rick, is better than telling them they are succumbing to an illusion. However, telling them they are paying attention to the wrong things isn’t much better. In the first instance, they’re being fooled; in the second, they are misguided. Neither is likely to sit well. So how about saying simply that there’s more there than meets the eye. I think they’ll be more responsive to the notion that they’re missing something than they will to being told they’re wrong.

Fred Nickols

[Rick Marken 2018-03-20_18:33:02]

[FWN] SNIP

RM: So what do you think? I propose that we stop telling behavioral scientists that they are succumbing to an illusion and just say they are paying attention the the wrong things: irrelevant (but compelling) side effects of control rather than the variables being controlled.

Best

Rick

···

–

Richard S. Marken

46

[Bruce Nevin 2018-03-21_12:34:13 UTC]

Perfect. Couldn’t agree more. Neutral terminology and a generative model.

As to the generative model … Is it an effect of controlling movement in orthogonal directions simultaneously, or is it an effect of the physics of the moving limbs, or both? Including larvae as an example may eliminate the physics-only option. So: is (more or less) straight-line movement to an end point preferential, and are x- and y-axis straight-line motion (with the starting point as a ‘weird’ kind of end point) distinct perceptual variables, such that simultaneous control of them is more difficult the tighter the curvature? If so, is conflict between them the source of difficulty? Or is the relationship between x-velocity and y-velocity controlled, and that control more difficult as the tighter curvature makes change in their relative velocities more rapid?

···

On Wed, Mar 21, 2018 at 6:43 AM, Adam Matic adam.matic@gmail.com wrote:

[From A. Matic 21.3.2018]

(Rick Marken 2018-03-20_18:33:02)

RM: So what do you think? I propose that we stop telling behavioral scientists

that they are succumbing to an illusion and just say they are paying attention

the the wrong things: irrelevant (but compelling) side effects of control

rather than the variables being controlled.

AM:

Well, there are already two camps, one looking at the power law as part of the

plan, the other looking at it as a side effect. A lot of computational models

are published in the 'plan trajectory → inverse kinematics & dynamics →

muscle activation’ camp, and many pieces of evidence are believed to support

this reasoning. The other camp are mostly equilibrium point guys, who maintain

that forces are not planned, and also publish models. Like that Gribble and

Ostry paper.

Going by the ‘extraordinary claims, extraordinary evidence’ line of thinking,

if there is something to hierarchical control structures, then we should be

able to build a model that can trace an ellipse shape (or other shapes)

without being told at what speed it should go at which point, and it must

reproduce the power law as a side-effect. (and fail to produce the power law

in the situations where humans don’t produce it).

47

[From Fred Nickols (2018.03.21.0836 ET)]

My posts do not seem to be getting through to the list. This is a test.

Fred Nickols

···

On Mar 21, 2018, at 8:34 AM, Bruce Nevin bnhpct@gmail.com wrote:

[Bruce Nevin 2018-03-21_12:34:13 UTC]

Perfect. Couldn’t agree more. Neutral terminology and a generative model.

As to the generative model … Is it an effect of controlling movement in orthogonal directions simultaneously, or is it an effect of the physics of the moving limbs, or both? Including larvae as an example may eliminate the physics-only option. So: is (more or less) straight-line movement to an end point preferential, and are x- and y
-axis straight-line motion (with the starting point as a ‘weird’ kind of end point) distinct perceptual variables, such that simultaneous control of them is more difficult the tighter the curvature? If so, is conflict between them the source of difficulty? Or is the relationship between x-velocity and y-velocity controlled, and that control more difficult as the tighter curvature makes change in their relative velocities more rapid?

/Bruce

On Wed, Mar 21, 2018 at 6:43 AM, Adam Matic adam.matic@gmail.com wrote:

[From A. Matic 21.3.2018]

(Rick Marken 2018-03-20_18:33:02)

RM: So what do you think? I propose that we stop telling behavioral scientists

that they are succumbing to an illusion and just say they are paying attention

the the wrong things: irrelevant (but compelling) side effects of control

rather than the variables being controlled.

AM:

Well, there are already two camps, one looking at the power law as part of the

plan, the other looking at it as a side effect. A lot of computational models

are published in the 'plan trajectory → inverse kinematics & dynamics →

muscle activation’ camp, and many pieces of evidence are believed to support

this reasoning. The other camp are mostly equilibrium point guys, who maintain

that forces are not planned, and also publish models. Like that Gribble and

Ostry paper.

Going by the ‘extraordinary claims, extraordinary evidence’ line of thinking,

if there is something to hierarchical control structures, then we should be

able to build a model that can trace an ellipse shape (or other shapes)

without being told at what speed it should go at which point, and it must

reproduce the power law as a side-effect. (and fail to produce the power law

in the situations where humans don’t produce it).

48

[From A Matic 21.3.2018]

[Bruce Nevin 2018-03-21_12:34:13 UTC]
As to the generative model .. Is it an effect of controlling movement in
orthogonal directions simultaneously, or is it an effect of the physics of the
moving limbs, or both? Including larvae as an example may eliminate the
physics-only option. So: is (more or less) straight-line movement to an end
point preferential, and are x- and y-axis straight-line motion (with the
starting point as a 'weird' kind of end point) distinct perceptual variables,
such that simultaneous control of them is more difficult the tighter the
curvature? If so, is conflict between them the source of difficulty? Or is the
relationship between x-velocity and y-velocity controlled, and that control
more difficult as the tighter curvature makes change in their relative
velocities more rapid?

···

------------

AM:
I'm not sure, for humans, the movement of the pen should be more or less the
same as the 'reference shape', so an ellipse for example. I haven't thought
much about larvae so far.

49

[Rick Marken 2018-03-21_14:44:20]

[From A. Matic 21.3.2018]

(Rick Marken 2018-03-20_18:33:02)
RM: So what do you think? I propose that we stop telling behavioral scientists
that they are succumbing to an illusion and just say they are paying attention
the the wrong things: irrelevant (but compelling) side effects of control
rather than the variables being controlled.
---

AM:
Well, there are already two camps, one looking at the power law as part of the
plan, the other looking at it as a side effect. A lot of computational models
are published in the 'plan trajectory -> inverse kinematics & dynamics ->
muscle activation' camp, and many pieces of evidence are believed to support
this reasoning. The other camp are mostly equilibrium point guys, who maintain
that forces are not planned, and also publish models. Like that Gribble and
Ostry paper.

RM: I agree that the "side effects camp" may by looking at the power law as a side effect but based on the models produced by this camp, they are not looking at the power law as a side effect of control.Â
Â

AM: Going by the 'extraordinary claims, extraordinary evidence' line of thinking,
if there is something to hierarchical control structures, then we should be
able to build a model that can trace an ellipse shape (or other shapes)
without being told at what speed it should go at which point, and it must
reproduce the power law as a side-effect. (and fail to produce the power law
in the situations where humans don't produce it).

RM: I demonstrated at least part of this with the model that accounts for the data in Figure 1 of our rebuttal to the rebuttals. The top level system controls for drawing an elliptical ellipse by varying its output in a way that matches the average observed trajectory produced by the person. The lower level system controls the position of the pen relative to the elliptically varying reference from the high level system. The ellipse that is drawn (Figre 1A in the rebuttal) follows the power law because in elliptical trajectories, or approximations thereto, there is a low correlation between affine velocity and curvature.Â
RM: The outputs of the lower level system-- the ones that keep the pen moving in the trajectory specified by the higher level reference variations  while protecting it from disturbances-- do not follow the power law because their trajectory has a relatively high correlation between affine velocity and curvature. If the disturbance to pen position were created in such a way that the outputs of the lower level system were themselves elliptical, -- or if you simply eliminated the disturbance so that the outputs of the lower level system traced out the same trajectory as the one specified by the output of the higher level system -- then you would find that both the drawn ellipse and the movements used to produce it follow the power law. This would show rather conclusively that whether or not a movement trajectory follows the power law depends on mathematical properties of the movement itself rather than on how the movement was produced. I think it would be a great little experiment for your lab to carry out!
Best
Rick

···

--
Richard S. MarkenÂ
"Perfection is achieved not when you have nothing more to add, but when you
have nothing left to take away.�
                --Antoine de Saint-Exupery

50

[From Adam Matic 22.3.2018]

(Rick Marken 2018-03-21_14:44:20)
RM: I agree that the "side effects camp" may by looking at the power law as a
side effect but based on the models produced by this camp, they are not
looking at the power law as a side effect of control.

···

----------

AM:
Side note: some versions of the lambda hypothesis are like those Braitenberg
vehicles - they call them 'equilibrium point', but in implementation and in
effect, they are angular control systems. Some of the published model, at
least.

More to the point - PL is not "a side effect of control". It is a side effect
of controlling a specific variable using a specific output function, in
specific situations. All those things needs to be identified - controlled
variables, means of control, and environmental factors.

---
RM: I demonstrated at least part of this with the model that accounts for the
data in Figure 1 of our rebuttal to the rebuttals. The top level system
controls for drawing an elliptical ellipse by varying its output in a way that
matches the average observed trajectory produced by the person. The lower
level system controls the position of the pen relative to the elliptically
varying reference from the high level system.

AM:
Maybe that model is a good start. You haven't published any trajectories
produced by the model, on any other details, but looks like it is a single
level model with a pre-calculated reference trajectory.

------------------------
RM: This would show rather conclusively that whether or not a movement
trajectory follows the power law depends on mathematical properties of the
movement itself rather than on how the movement was produced. I think it would
be a great little experiment for your lab to carry out!

AM:
Nope. Looks pointless to me. A better line of action is looking at tasks and
situations where the power law has been found in published research,
replicating the tasks and trying to identify controlled variables.

51

[Rick Marken 2018-03-22_08:20:42]

[From Adam Matic 22.3.2018]

AM: More to the point - PL is not "a side effect of control". It is a side effect
of controlling a specific variable using a specific output function, in
specific situations. All those things needs to be identified - controlled
variables, means of control, and environmental factors.

 RM: Yes, I can buy that. If you do some research where you actually do some testing to determine the variable(s) being controlled when organisms move that would be a huge step forward.Â

---
RM: I demonstrated at least part of this with the model that accounts for the
data in Figure 1 of our rebuttal to the rebuttals. The top level system
controls for drawing an elliptical ellipse by varying its output in a way that
matches the average observed trajectory produced by the person. The lower
level system controls the position of the pen relative to the elliptically
varying reference from the high level system.

AM:
Maybe that model is a good start. You haven't published any trajectories
produced by the model, on any other details, but looks like it is a single
level model with a pre-calculated reference trajectory.

RM: You're right that the reference to the cursor position control system -- the lower level system -- was pre-calculated based on the observed cursor trajectory. Therefore the model-produced elliptical trajectory of the cursor is not very interesting. But the model also perfectly reproduced the trajectory of the mouse movements that resulted in the elliptical cursor trajectory. So the model produced a mouse movement trajectory that looks just like the actual mouse movement trajectory shown in Figure 1B in our rebuttal to the rebuttal. That was a movement trajectory that didn't follow the power law at all and it was perfectly accounted for by a simple control model. The next step would be to figure out what the next higher level system is controlling that leads it to produce the elliptical trajectory of references to the lowest level systems. >

------------------------
RM: This would show rather conclusively that whether or not a movement
trajectory follows the power law depends on mathematical properties of the
movement itself rather than on how the movement was produced. I think it would
be a great little experiment for your lab to carry out!

AM:
Nope. Looks pointless to me.

RM: Then maybe it's something I should do;-)
Â

AM: A better line of action is looking at tasks and
situations where the power law has been found in published research,
replicating the tasks and trying to identify controlled variables.

RM: That would be great. I think you should replicate tasks where the power law has been found and tasks where it hasn't been found. I'd be happy to help you out with the design of experiments that would allow you to identify controlled variables.Â
Â
BestÂ
Rick

···

--
Richard S. MarkenÂ
"Perfection is achieved not when you have nothing more to add, but when you
have nothing left to take away.�
                --Antoine de Saint-Exupery

52

[From A. Matic 23.3.2018]

···

On Thu, Mar 22, 2018 at 4:20 PM, Richard Marken rsmarken@gmail.com wrote:

AM:

Of course the mouse trajectory correlates well with the model mouse trajectory, as in any tracking model. Try drawing a big ellipse in 1 second or less, and see how the model accounts for that, and what happens with the power law, for humans and for the model.

AM: Yep. You can make the target (or the reference) move at any speeds along an elliptical trajectory. Try high speeds and non-power law elliptical trajectories.

AM:

That’s the easier part. Coming up with hypotheses on what is controlled is the difficult part.

RM: You’re right that the reference to the cursor position control system – the lower level system – was pre-calculated based on the observed cursor trajectory. Therefore the model-produced elliptical trajectory of the cursor is not very interesting. But the model also perfectly reproduced the trajectory of the mouse movements that resulted in the elliptical cursor trajectory. So the model produced a mouse movement trajectory that looks just like the actual mouse movement trajectory shown in Figure 1B in our rebuttal to the rebuttal. That was a movement trajectory that didn’t follow the power law at all and it was perfectly accounted for by a simple control model. The next step would be to figure out what the next higher level system is controlling that leads it to produce the elliptical trajectory of references to the lowest level systems.

RM: Then maybe it’s something I should do;-)

RM: I’d be happy to help you out with the design of experiments that would allow you to identify controlled variables.

53

Down

···

From: Richard Marken [mailto:rsmarken@gmail.com]
Sent: Wednesday, March 21, 2018 2:34 AM
To: csgnet@lists.illinois.edu
Subject: Re: Re: Re: Behav. illusions

[Rick Marken 2018-03-20_18:33:02]

(From A. Matic 20.3.2018)

RM: But as I have noted before, the power law is a different kind of
behavioral illusion than the S-R version of the illusion described in
Powers(1978).

AM: Oh. Sure, I agree it is not the kind of illusion described in Powers(1978).
Glad we agree on that one.

AM: I don’t think we should expand the meaning of the term “behavioral illusion”
to include all incorrect inferences about control. There are different kinds
of incorrect inferences, like assuming that a side effect is the controlled
variable. That one can stay in its own category.

RM: Actually, I’d be happy to dispense with the term “behavioral illusion” altogether. I think it can be taken as an insult. I think all of these “illusions” and “side-effects” can be safely placed in a single category called “irrelevant side-effects” of control.

HB : As ussual you are wrong Rick because you are mixing PCT and RCT where your matra is : »Behavior is control« and you think that there is some »controlled varaible« being kep tunder control.

You still think that LCS can produce some controlled effects to environment and some irrelevant side effects«. There is no »controlled variable in environemnt« (see diagram LCS III) and nevromuscular connection can not produce controlled and uncontrolled effects .Nevromuscular connection works in one way and it’s not producing »controlled effects«. There are just effects to environment.

Bill P (B:CP):

OUTPUT FUNCTION : The portion of a system that converts the magnitude or state of a signal inside the system into a corresponding set of effects on the immediate environment of the system…<

HB : So we need an avidence that LCS can produce »controlled effects to environment which could keep some »controlled variable« under control. You can see and make your own interpretation what is happening in environment (it’s your construct), but it has to be in accordance to evidences from those (for ex. physiologist) who can confirm that »controlled effects« to environemnt can be produced. Should we believe your word just like that ? Because you are Richard Marken ???

Who are you anyway that we should beleive you that you are telling the truth ?

So it sems that you are back in bussines again. How do you know that there are any »side effects« outside ? From perception ? Anyway everything is perception. which is controlled. And in PCT that is the only thing that matters if organism want to survive. Â

RM : This is certainly true of the S-R illusion described in Powers (1978); the observed relationship between S and R is a side effect that is irrelevant to the fact that a variable (the contrlled variable) is being kept under control.

HB : Where did Bill exactly wrote that ? And where exactly did he wrote that »(the contrlled variable) is being kept under control« ???

How many times do I have to prove to you that there is generaly no »controlled variable« in environment of LCS. And there is no »side effects« or »right« effects. There are just effects to outer environment. Stop misleading people !!! Again …

Bill P (B:CP):

OUTPUT FUNCTION : The portion of a system that converts the magnitude or state of a signal inside the system into a corresponding set of effects on the immediate environment of the system…

RM : Same is true of “operant behavior”; the observed relationship between reinforcement and responses is irrelevant to the fact that a variable (reinforcement rate) is being kept under control.

HB : Nothing in environment of LCS is being kept under control. It’s just stability in enviroment that is the result of control in organism (24/7) where predefined state is kept under control.

Bill P (B:CP):

CONTROL : Achievement and maintenance of a preselected state in the controlling system, through actions on the environment that also cancel the effects of disturbances.

Hb : In PCT only internal enviroment is being kep tunder control what enables survival.   Whatever happens in outer environemnt can be maybe treated as »side effect« of internal control. Â

RM : And the same is true of the power law; the observed relationship between curvature and velocity is irrelevant to the fact that a variable (the trajectory itself) is being kept under control.

HB : Trajectory is being kept under control ??? What kind of RCT is this ? Is this a variation of RCT and »Behavior is control« ?

So what is keeping »the fact that a variable (the trajectory itself) is being kept under control ?What is keeping zhe trajectory under control ? In PCT we are talking about »Perception under control«. Which theory do you use ?

I hope you’ll not say that PCT is »Behavior control theory« or worse that »Telekinesis« is possible. You try to prove as many times before that there exists »control of behavior« and there are some »controlled effects to environment« and there is some »controlled variable and of course consequently there is also some »Controlled Perceptual Variable« or PCV. There si no such things in PCT.

RM: So what do you think? I propose that we stop telling behavioral scientists that they are succumbing to an illusion and just say they are paying attention the the wrong things: irrelevant (but compelling) side effects of control rather than the variables being controlled.

HB : This can be usefull. I also don’t understand why Bill talked baout »behaviorla illusion« if perception is all there is. Everything can be treated as »perceptual illusion«.

Perception can be interpreted as worse or better for control and that’s all. If you have falls perceptual »informations« it’s hard to survive (Maturana’s experiments with frogs and turned eyes). Because wrong perception  (perceptual illusion) leads to wrong control . And that’s probably what is happening to you Rick. It’s »perceptual illusion« that »behavior as control« is producing some »controlled effects«. Â

There are no »variables« in outer environment being kept under control in general sense because there is only control process going on in organism 24/7 which is really »keeping intrimsic variables« under control near genetic reference values all the time. If you want to call some process control it has to work all the time in control manner. And this kind of permanent control garantee survavial of Living beibgs. See diagram LCS III.

Where do you see some permanent or fixed »controlled variable« in environment ? Maybe you mixed something with »controlled quantitiy« which is totaly different concept in PCT. But I wrote so many times about these problems that you could remember something. The only »controlled variable« in PCT control loop is perceptual signal. In RCT it can be anything.Â

Boris

Best

Rick

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