Environment and Perception (was Re: Behavioural Illusion)

[rsmarken]

[From Rick Marken (2017.10.20.2150)]

···

Martin Taylor (2017.10.18.18.47)–

                BA: In some cases, however, the perceptual

signal is a function of two or more environmental
variables. In that case there exists no single
controlled quantity in the environment that
corresponds to the perceptual signal. Instead, the
controlled quantity is that combination of
environmental variables that determines the
perceptual signal, as defined by the input
function. For example, a receiver system may be
attempting to maintain a certain signal to noise
ratio. That ratio depends on both the signal
intensity and the noise intensity, both of which are
environmental variables. A third variable
corresponding to the ratio of the two does not exist
in the environment.Â

MT: That seems a very weird thing to say. The signal intensity exists in

the environment, the noise intensity exists in the environment, but
their ratio (an equally abstract quantity) does not. That makes no
sense to me at all.

RM: I agree with Bruce; a ratio of variables does not really exist in the environment. Same as area and perimeter in my “What is Size” demo (www.mindreadings.com/ControlDemo/Size.html). These perceptions are all different functions of lower level perceptions that are ultimately functions of environmental variables. But Bruce’s example made me realize that even variables that are presumed to be environmental variables, like signal and noise intensity, are actually perceptual variables themselves in the sense that they are functions of variables in what is called the “environment” in PCT diagrams.

RM: For example, the intensity of acoustic noise is actually a function of an actual environmental variable, which is pressure variations at the sensory surface (hair cells). The hair cells convert these pressure variations into neural impulses, which represent the intensity of the pressure variations in terms of firing rate. But what is the intensity of a time varying pressure variation? There are different ways to compute intensity. One involves integrating the effects of pressure variations over some time interval; this is a measure of intensity in terms of energy. Another possible way of measuring intensity is by averaging the value of the “peaks” in the pressure variations over some time period. When I was in grad school there was some controversy over how the hair cells produced a neural measure of intensity of sound; whether the hair cells computed a measure of energy or computed the average of the pressure peaks (peak picker). Maybe they’ve figured it out by now.Â

RM: But the point is that, whatever the hair cells are measuring as sound intensity, these cells compute a function (integral of pressure, average of peaks, or something else) of what physics tells us is actually going on in the environment (pressure variations in the air) and the results of this computation is a neural signal that is a perception of intensity. These are the “intensity” perceptions that are presumed to be at the lowest level of the control hierarchy; the level that most closely interfaces with the environment. And so we see that even at this level, the perceptual variable (the neural signal) represents an aspect (function) of a variable (pressure variations) in the environment.Â

RM: What this means is that when, in PCT, we talk about perceptions that are a function of environmental variables, the environmental variables we use as examples are themselves perceptual variables. So in the “What is size” demo, when I say that the perception of area is a function of the height and width of the sides of a rectangle, the implication is that the height and width of the sides are environmental variables. But, in fact, they are perceptual variables (that we can measure using instruments – the computer, in this case) that are themselves functions of perceptual variables (the intensity of the pixels that make up the display) which are themselves a function of physical variables (spatial variation in the amplitude of photons variations emanating from the screen). This fact may make for confusing philosophical discussions but it is not a problem for researchers who are mainly interested in finding the best description of the functions of lower level variables (be they called environmental variables or lower level perceptions) that define the perceptions that are being controlled when organisms are seen carrying out various behaviors.Â

RM: So does this mean that living control systems don’t control environmental variables? If by “environmental variables” we mean the variables that physics tells us are in the environment – like variations in air pressure – then the answer is “yes”, we don’t control these variables. What we control are aspects of these environmental variables – aspects that are defined by our perceptual functions. Then does this mean that we only control the perceptions defined by our perceptual functions and not the aspects of the environment that correspond to these perceptions. The answer to this is an emphatic “no”. This is demonstrated in Powers’ MulltiControl demo. The demo is here:Â

https://www.dropbox.com/s/2u00ac87bix2sjv/MultiControlPrj.exe?dl=0

and the write up is here:

https://www.dropbox.com/s/rwoqfa8v96g62ob/multiple_control.pdf?dl=0.

RM: In this demo, N different control systems are controlling N different linear combinations of the same environmental variables. The individual environmental variables themselves are not controlled (as can be seen in the demo as the wandering purple circles) but the aspects of these variables – the linear combinations of these variables that are the controlled perceptions – are controlled (as can be seen in the demo as the red circles staying close to the white (reference) circles).Â

RM: But while control systems control aspects of environmental variables and not the environmental variables themselves, I think it’s misleading to say (as it does in the Wikipedia write up on PCT) that PCT says that organisms don’t control their environment (or environmental variables), just their perceptions. For those who don’t know what environmental variables are (they are not the things we see and hear as being in the environment) or what perception means in PCT (it’s not an interpretation of the environment we see and hear) it can sound like PCT is solipsistic; that the theory says that organisms behavior in a world of make-believe. And this, of course, is not the case. The environment is an important part of the PCT model and perception is presumed to be a function of that environment.Â

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

[wmansell]

Thanks Rick, that’s a really helpful explanation.

···

Martin Taylor (2017.10.18.18.47)–

                BA: In some cases, however, the perceptual

signal is a function of two or more environmental
variables. In that case there exists no single
controlled quantity in the environment that
corresponds to the perceptual signal. Instead, the
controlled quantity is that combination of
environmental variables that determines the
perceptual signal, as defined by the input
function. For example, a receiver system may be
attempting to maintain a certain signal to noise
ratio. That ratio depends on both the signal
intensity and the noise intensity, both of which are
environmental variables. A third variable
corresponding to the ratio of the two does not exist
in the environment.

MT: That seems a very weird thing to say. The signal intensity exists in

the environment, the noise intensity exists in the environment, but
their ratio (an equally abstract quantity) does not. That makes no
sense to me at all.

RM: I agree with Bruce; a ratio of variables does not really exist in the environment. Same as area and perimeter in my “What is Size” demo (www.mindreadings.com/ControlDemo/Size.html). These perceptions are all different functions of lower level perceptions that are ultimately functions of environmental variables. But Bruce’s example made me realize that even variables that are presumed to be environmental variables, like signal and noise intensity, are actually perceptual variables themselves in the sense that they are functions of variables in what is called the “environment” in PCT diagrams.

RM: For example, the intensity of acoustic noise is actually a function of an actual environmental variable, which is pressure variations at the sensory surface (hair cells). The hair cells convert these pressure variations into neural impulses, which represent the intensity of the pressure variations in terms of firing rate. But what is the intensity of a time varying pressure variation? There are different ways to compute intensity. One involves integrating the effects of pressure variations over some time interval; this is a measure of intensity in terms of energy. Another possible way of measuring intensity is by averaging the value of the “peaks” in the pressure variations over some time period. When I was in grad school there was some controversy over how the hair cells produced a neural measure of intensity of sound; whether the hair cells computed a measure of energy or computed the average of the pressure peaks (peak picker). Maybe they’ve figured it out by now.

RM: But the point is that, whatever the hair cells are measuring as sound intensity, these cells compute a function (integral of pressure, average of peaks, or something else) of what physics tells us is actually going on in the environment (pressure variations in the air) and the results of this computation is a neural signal that is a perception of intensity. These are the “intensity” perceptions that are presumed to be at the lowest level of the control hierarchy; the level that most closely interfaces with the environment. And so we see that even at this level, the perceptual variable (the neural signal) represents an aspect (function) of a variable (pressure variations) in the environment.

RM: What this means is that when, in PCT, we talk about perceptions that are a function of environmental variables, the environmental variables we use as examples are themselves perceptual variables. So in the “What is size” demo, when I say that the perception of area is a function of the height and width of the sides of a rectangle, the implication is that the height and width of the sides are environmental variables. But, in fact, they are perceptual variables (that we can measure using instruments – the computer, in this case) that are themselves functions of perceptual variables (the intensity of the pixels that make up the display) which are themselves a function of physical variables (spatial variation in the amplitude of photons variations emanating from the screen). This fact may make for confusing philosophical discussions but it is not a problem for researchers who are mainly interested in finding the best description of the functions of lower level variables (be they called environmental variables or lower level perceptions) that define the perceptions that are being controlled when organisms are seen carrying out various behaviors.

RM: So does this mean that living control systems don’t control environmental variables? If by “environmental variables” we mean the variables that physics tells us are in the environment – like variations in air pressure – then the answer is “yes”, we don’t control these variables. What we control are aspects of these environmental variables – aspects that are defined by our perceptual functions. Then does this mean that we only control the perceptions defined by our perceptual functions and not the aspects of the environment that correspond to these perceptions. The answer to this is an emphatic “no”. This is demonstrated in Powers’ MulltiControl demo. The demo is here:

https://www.dropbox.com/s/2u00ac87bix2sjv/MultiControlPrj.exe?dl=0

and the write up is here:

https://www.dropbox.com/s/rwoqfa8v96g62ob/multiple_control.pdf?dl=0.

RM: In this demo, N different control systems are controlling N different linear combinations of the same environmental variables. The individual environmental variables themselves are not controlled (as can be seen in the demo as the wandering purple circles) but the aspects of these variables – the linear combinations of these variables that are the controlled perceptions – are controlled (as can be seen in the demo as the red circles staying close to the white (reference) circles).

RM: But while control systems control aspects of environmental variables and not the environmental variables themselves, I think it’s misleading to say (as it does in the Wikipedia write up on PCT) that PCT says that organisms don’t control their environment (or environmental variables), just their perceptions. For those who don’t know what environmental variables are (they are not the things we see and hear as being in the environment) or what perception means in PCT (it’s not an interpretation of the environment we see and hear) it can sound like PCT is solipsistic; that the theory says that organisms behavior in a world of make-believe. And this, of course, is not the case. The environment is an important part of the PCT model and perception is presumed to be a function of that environment.

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

[MartinT]

[Martin Taylor 2017.10.21.08.28]

      [From

Rick Marken (2017.10.20.2150)]

Pressure ***variation*** is a function of the ratio of the

pressure at one moment to the pressure at another moment (or the
difference, which is equally problematic in your view). It doesn’t
exist in the environment, either. Certainly it doesn’t at any one
moment. By your logic, nothing whatever that we perceive exists in
the environment.

I do not claim that everything we perceive does exist in the

environment. Perceptual functions get built and modified by
reorganization all the time (according to PCT). All I claim is that
perceptual functions that survive the rough and tumble of
reorganization in our environment are likely to correspond to real
properties of the environment.

Yep. Just like every other perception we have. They are all

variables that are the results of applying some function to the
values of other variables over space and time. At least we can agree
on that.

Yes. Maybe we use different words. What you call "aspect" I have

been calling “property”, but I think they mean the same.

Yep.

Until you start getting serious and interested in what might be

going on inside someone’s head. An “implication” is itself a rather
explicit function, so I guess it doesn’t exist either. Probably it
doesn’t because not everyone would take the same implication from
the same data. But maybe it does, and those who take the wrong
implication won’t keep their intrinsic variables in as happy a state
as those who take the correct implication.

Yep, it's functions all the way. So far, I can't think what you are

trying to get across that you imagine I wouldn’t agree with.

Yep.

I haven't actually seen the demo, but I thought the write-up was

great. It made the same point I tried to get across with my
“stochastic collective control” demo several months ago, which
caused you some angst. I guess it’s OK when Bill says it, but not
when I say that the perceptual representations of environmental
variables can be distributed rather than localized in specific
“wires” in the brain.

Now consider the evolutionary implications of what you said above.

If the “aspects” themselves are not individually important for
survival, but their combination represents a tiger about to jump, a
cliff one might walk over, a rock coming at your head, is it not
more likely for organisms that act as though those combinations
represented something in the environment to survive than for those
that act as thought the “aspects” of yellow and black patches, a
green-blue edge (of a grassy meadow with a cliff down to the sea),
or a looming spatial pattern (a function in space-time) where the
variables important to control. Yet you say that Bill’s demo shows
we don’t control the individual environmental variables, which in
some way is important in proving that the tiger does not exist in
the environment.

We can never know whether what we see represents a real environment

at all. Maybe there is an army of gnomes that produce every
individual firing of a sensory nerve in order to make things look
and feel as they do. Can we control the gnomes? Yes, I suppose so,
because they keep track of every muscle twitch and carefully compute
precisely which nerves should fire and when, in the simulated
environment they have collectively decided we should live in. But
it’s much simpler to hypothesize that our sensory nerve firings are
based on a world at least something like the one we perceive to be
there.

Bill's demo exists in no world important to any intrinsic variables

except the one he specified – control quality. His reorganizing
system is not going to die if he doesn’t keep, say, the ratio
between variable N and variable N+1 equal to the value of variable
N+2. His environment does not contain constraints imposed
independently of his reorganizing perceptual functions, as is the
case in the environment we seem to live in, an environment in which
chair legs are more often placed between the chair seat and a solid
surface than in any other relation to the chair seat. But what he
does show is important, and could be seen as a basic requirement for
the existence of a hierarchy of controllable perceptions – a bunch
of perceptions at one level can be a distributed representation of a
single-valued perception at a higher level.

I think you are logically inconsistent, or else you are omitting

something important, because we agree on all (I think “all”) the
mechanisms, but draw radically different conclusions from them.
Obviously I am correct – that’s a property (aspect)
of being an “I”. We all are, though “I” tomorrow may think “I”
yesterday was incorrect. “I” today is/am always correct. In other
words, as Bill often said, all we can ever be quite sure of is what
we perceive right now.

In previous iterations of this same argument, I have emphasized the

point that a function of variables is itself a variable. Other times
(as on this occasion) I have tried to make the point that it doesn’t
matter what you perceive out there, it is real reality that
determines what actually will happen when you act on the
environment, and if what you perceive isn’t there, controlling a
perception of it won’t do your intrinsic variables much good.

Maybe the sticking point for you is that we can (and a lot of the

time do) perceive things that aren’t there. But as J.G.Taylor
pointed out so long ago (55 years ago now) those perceptions don’t
tend to stick around as long as do those that matter to our ability
to act effectively in the world. The taste of lemonade, for example,
matters because it allows you to reduce the likelihood you are
drinking poison (or orange juice).

The perceptual functions that matter are more likely to represent

real aspects of the environment, whether they are as simple as a
line that separates the view of a grassy meadow and the sea far
below or as complicated as a political system. If you perceive
yourself living in a democracy whereas you are actually living in
North Korea, and you criticize the Dear Leader, you are just as dead
as you would be if you failed to duck a rock that you wrongly
perceive to be missing your head. Does the dictatorship exist in the
environment? I would say it does.

The level of the hierarchy doesn't matter. It's functions all the

way down (not turtles), but if those functions don’t represent
things in the environment at least approximately, they are a waste
of resources when it comes to survival long enough to propagate your
genes.

···

                Martin

Taylor (2017.10.18.18.47)–

                                BA:

In some cases, however, the
perceptual signal is a function of
two or more environmental
variables. In that case there
exists no single controlled
quantity in the environment that
corresponds to the perceptual
signal. Instead, the controlled
quantity is that combination of
environmental variables that
determines the perceptual signal, as
defined by the input function. For
example, a receiver system may be
attempting to maintain a certain
signal to noise ratio. That ratio
depends on both the signal intensity
and the noise intensity, both of
which are environmental variables.
A third variable corresponding to
the ratio of the two does not exist
in the environment.

              MT: That seems

a very weird thing to say. The signal intensity exists
in the environment, the noise intensity exists in the
environment, but their ratio (an equally abstract
quantity) does not. That makes no sense to me at all.

      RM: I agree with Bruce; a ratio of variables does not really

exist in the environment. …

      RM: For example, the intensity of

acoustic noise is actually a function of an actual
environmental variable, which is pressure variations at the
sensory surface (hair cells).

…

      RM: But the point is that, whatever the

hair cells are measuring as sound intensity, these cells
compute a function (integral of pressure, average of
peaks, or something else) of what physics tells us is actually
going on in the environment (pressure variations in the air)
and the results of this computation is a neural signal that is
a perception of intensity.

      These are the "intensity" perceptions

that are presumed to be at the lowest level of the control
hierarchy; the level that most closely interfaces with the
environment. And so we see that even at this level, the
perceptual variable (the neural signal) represents an aspect (function)
of a variable (pressure variations) in the environment.

      RM: What this means is that when, in

PCT, we talk about perceptions that are a function of
environmental variables, the environmental variables we use as
examples are themselves perceptual variables.

      So in the "What is size" demo, when I

say that the perception of area is a function of the height
and width of the sides of a rectangle, the implication is that
the height and width of the sides are environmental variables.

      But, in fact, they are perceptual

variables (that we can measure using instruments – the
computer, in this case) that are themselves functions of
perceptual variables (the intensity of the pixels that make up
the display) which are themselves a function of physical
variables (spatial variation in the amplitude of photons
variations emanating from the screen). This fact may make for
confusing philosophical discussions but it is not a problem
for researchers who are mainly interested in finding the best
description of the functions of lower level variables (be they
called environmental variables or lower level perceptions)
that define the perceptions that are being controlled when
organisms are seen carrying out various behaviors.

      RM: So does this mean that living

control systems don’t control environmental variables? If by
“environmental variables” we mean the variables that physics
tells us are in the environment – like variations in air
pressure – then the answer is “yes”, we don’t control these
variables. What we control are aspects of these
environmental variables – aspects that are defined by our
perceptual functions.

      Then does this mean that we only

control the perceptions defined by our perceptual functions
and not the aspects of the environment that correspond to
these perceptions. The answer to this is an emphatic “no”.
This is demonstrated in Powers’ MulltiControl demo. The demo
is here:

https://www.dropbox.com/s/2u00ac87bix2sjv/MultiControlPrj.exe?dl=0

and the write up is here:

https://www.dropbox.com/s/rwoqfa8v96g62ob/multiple_control.pdf?dl=0.

      RM: In this demo, N different control

systems are controlling N different linear combinations of the
same environmental variables. The individual
environmental variables themselves are not controlled (as can
be seen in the demo as the wandering purple circles) but the
aspects of these variables – the linear combinations of these
variables that are the controlled perceptions – are
controlled (as can be seen in the demo as the red circles
staying close to the white (reference) circles).

RM: But while control systems control * aspects* of environmental variables and not the environmental
variables themselves, I think it’s misleading to say (as it
does in the Wikipedia write up on PCT) that PCT says that
organisms don’t control their environment (or environmental
variables), just their perceptions. For those who don’t know
what environmental variables are (they are not the things we
see and hear as being in the environment) or what perception
means in PCT (it’s not an interpretation of the environment we
see and hear) it can sound like PCT is solipsistic; that the
theory says that organisms behavior in a world of
make-believe. And this, of course, is not the case. The
environment is an important part of the PCT model and
perception is presumed to be a function of that environment.

[rsmarken]

[From Rick Marken (2017.10.21.1530)]

···

Martin Taylor (2017.10.21.08.28)–

MT: I do not claim that everything we perceive does exist in the

environment. Perceptual functions get built and modified by
reorganization all the time (according to PCT). All I claim is that
perceptual functions that survive the rough and tumble of
reorganization in our environment are likely to correspond to real
properties of the environment.

RM: I think all "properties of the environment " are equally real in the sense that they are all perceptual variables that are functions of real environmental variables. I think we experience the world in terms of intensities, sensations, configurations, and so on because these are the aspects of physical reality that must be controlled in order for us to survive. I think the perceptual functions that allow us to perceive the world as we do are evolved properties of our brain. I don’t think reorganization within an individual childhood (or lifetime) could possibly come up with the neural structures that compute these different types of perceptions.

MT: I haven't actually seen the demo, but I thought the write-up was

great. It made the same point I tried to get across with my
“stochastic collective control” demo several months ago, which
caused you some angst. I guess it’s OK when Bill says it, but not
when I say that the perceptual representations of environmental
variables can be distributed rather than localized in specific
“wires” in the brain.

RM: The demo I’m talking about has nothing to do with showing that “perceptual representations of environmental variables can be distributed rather than localized in specific “wires” in the brain”. It showed that N different control systems can successfully control N different aspects (linear combinations) of the same environmental variables. If that’s what your “stochastic collective control” demo showed than I have no idea why it appeared to cause me angst. But my money is on it having shown nothing like that.

Â

MT: Now consider the evolutionary implications of what you said above.

If the “aspects” themselves are not individually important for
survival, but their combination represents a tiger about to jump, a
cliff one might walk over, a rock coming at your head, is it not
more likely for organisms that act as though those combinations
represented something in the environment to survive than for those
that act as thought the “aspects” of yellow and black patches, a
green-blue edge (of a grassy meadow with a cliff down to the sea),
or a looming spatial pattern (a function in space-time) where the
variables important to control. Yet you say that Bill’s demo shows
we don’t control the individual environmental variables, which in
some way is important in proving that the tiger does not exist in
the environment.

RM: Actually, I never said that. Bill’s demo doesn’t say anything about the adaptive significance of the linear combinations of environmental variables that are controlled. It just shows that a set of functions of the same set of environmental variables can produce perceptions that can be controlled simultaneously. The functions in the demo are created randomly; in real life these functions would have evolved to control the aspects of the environment that we need to be able to control in order to survive, such as our distance from tigers.Â

Â

MT: We can never know whether what we see represents a real environment

at all.

RM: I think the success of the physical sciences gives us good reason to believe that what we experience is a function of a real environment. Our perceptual experience doesn’t represent what science tells us is the real environment --Â tigers, for example, are not part of the physics model of the real environment; rather I think our perceptual experience is made up of perceptual variables that represent different aspects of that real environment.

MT: Bill's demo exists in no world important to any intrinsic variables

except the one he specified – control quality. His reorganizing
system is not going to die if he doesn’t keep, say, the ratio
between variable N and variable N+1 equal to the value of variable
N+2. His environment does not contain constraints imposed
independently of his reorganizing perceptual functions, as is the
case in the environment we seem to live in, an environment in which
chair legs are more often placed between the chair seat and a solid
surface than in any other relation to the chair seat. But what he
does show is important, and could be seen as a basic requirement for
the existence of a hierarchy of controllable perceptions – a bunch
of perceptions at one level can be a distributed representation of a
single-valued perception at a higher level.

RM: Right.Â

MT: Maybe the sticking point for you is that we can (and a lot of the

time do) perceive things that aren’t there.

RM: I think we can perceive things that aren’t there. I think these perceptions are called hallucinations. These perceptions are not controllable, much to the dismay of those who have them, because they are only partially functions of environmental variablesÂ

Â

MT: The perceptual functions that matter are more likely to represent

real aspects of the environment, whether they are as simple as a
line that separates the view of a grassy meadow and the sea far
below or as complicated as a political system.

RM: I don’t think any function of environmental variables can be considered more real than another. I think it’s better to say that the perceptual functions that matter – those that we have inherited through evolution – are the ones that represent aspects of the environment that we have to be able to control in order to survive.Â

Â

MT: If you perceive

yourself living in a democracy whereas you are actually living in
North Korea, and you criticize the Dear Leader, you are just as dead
as you would be if you failed to duck a rock that you wrongly
perceive to be missing your head. Does the dictatorship exist in the
environment? I would say it does.

RM: I would say that the basis for that perception exists in the environment. But if you want to think that the correlates of perceptual variables exist in the environment, that’s fine. It’s not really an important thing to worry about when you do research. For example, I don’t worry about whether vertical optical velocity is in the environment or not when modeling object interception. I just model it as a function of variables that are in the environment and the model behaves just like the real thing. The only problem I see with thinking of controlled variables as existing in the environment is when this leads to the notion (which I believe you expressed) that it is only the perception of the environmental variable that is controlled and not the environmental variable itself. As I’ve said, this is not only a complete misunderstanding of the concept of “control of perception”, if taken seriously it would mean that the study of the controlling done by living systems is impossible.Â

MT: It's interesting that a disagreement about whether one branch of

experimental psychology constitutes all of experimental psychology,
mixed in with a difference of opinion as to the meanings of “basis”
and “revolution”, provides evidence that I am an enemy of (or fail
to understand) PCT. I’m not interested in getting into that
argument. I will let what I say about PCT speak for itself. There
are things that we do and will disagree on with respect to PCT.
That’s healthy in any science. But a persistent tendency to claim
that any disagreement with your opinion is tantamount to treason or
heresy is neither scientific nor healthy.

RM: All I claim is that you have a fundamentally different idea of what PCT is about than I do. This difference doesn’t mean you are an enemy or traitor. Though I think you do an injustice to Bill’s legacy by treating the truly revolutionary implications of his work – that it invalidates the foundations of scientific psychology-- as merely a polemical device. And I do think you criticisms of my explanation of the power law of movement reflected a rather profound difference between your understanding of PCT and mine. But that’s not heretical or traitorous. No, my main complaint about you (and Bruce A) is that you are just not helping me do the work that I think needs to be done, and that is to figure out how to study behavior with an understanding of the fact that the systems under study are control systems. I’ve done a little of this figuring out on my own but I could have made a lot more progress (given my obvious limitations) if I had help from more people who have the same goal as I have.Â

      RM: In this demo, N different control

systems are controlling N different linear combinations of the
same environmental variables. The individual
environmental variables themselves are not controlled (as can
be seen in the demo as the wandering purple circles) but the
aspects of these variables – the linear combinations of these
variables that are the controlled perceptions – are
controlled (as can be seen in the demo as the red circles
staying close to the white (reference) circles).Â

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

[MartinT]

[Martin Taylor 2017.10.23.23.10]

[From Rick Marken (2017.10.21.1530)]

It shows both at the same time. That's obvious in the write-up.

Maybe the demo doesn’t make it so obvious. But you can see it for
yourself by ignoring all but one of the environmental variables and
seeing whether that one is controlled. Since they all are, I think
you will find that the one you look at is. Its internal
representation is distributed over all the perceptual variables.

Both show that a single environmental variable can look as though it

was being controlled through one simple control loop controlling one
single perceptual variable, even though the Analyst’s view shows
that there is no such single perceptual variable, but instead a lot
of different ones that are being controlled independently.

I'm glad you see that the same way I do. I would add "and

reorganization" after “through organization”, but the extent to
which that is important is presumably rather species-specific.

We agree on that, too, apart from the fact I would add "may" before

“exist”. In fact I was planning to write a message concentrating on
it in the hope of ending the multi-year proliferation of this kind
of thread.

As I explained in detail in a recent message [Martin Taylor

2017.10.18.16.18], over the last couple of years I have come to a
much more nuanced view of that issue. I concluded by saying:
“Perception is all we have that we can control. The environment is
what we really need to control. We can control the environment only
to the extend that it behaves as though our perceptions correspond
to it reasonably well. So for most practical purposes other than
careful theory or metaphysical philosophy, it doesn’t matter a whit
whether we say that perception or the CEV is controlled.”

When I criticize your explanation of the power law of movement, it

has nothing whatever to do with PCT. It has to do with your misuse
of mathematics, and falling for the old “divide by zero” trick used
to prove nonsense to unsuspecting fifth-graders, albeit yours is in
an unfamiliar guise. But of course you know this.

As for doing an injustice to Bill's legacy, I think you do him an

injustice by cramping the application of PCT into the little realm
of simple control loops, examined usually singly. (I exclude MoL
from this comment).

There's a reason why the book I am trying to write has a deliberate

pun in the working title “Powers of Perceptual Control” with no
initial “The”. It is because I feel that the power and grand sweep
of what follows from an deep understanding of PCT in the study of so
many domains of human experience (and life more generally) has been
vastly short-changed, at least on CSGnet. This, I think, has been
largely because the revolutionary aspects of Bill’s great insight
risk getting lost in squabbles about minutiae, squabbles that I am
sure put off quite a few people who could greatly help in advancing
the larger project – understand the living world in the same sense
that understanding atomic interactions permits an understanding of
the chemical world.

PCT should be the foundation not just of experimental psychology,

but of topics with more direct application to everyday life –
sociology and economics, for example, both of which are, to my mind,
currently in states like the Ptolomaic cycles phase of astronomy.
Even “behavioural economics” uses mainstream psychology and methods
rather than PCT.

That's hilarious. I have the same complaint about your

obstructionism, for which I could use the same words: “* you are
just not helping me do the work that I think needs to be done, and
that is to figure out how to study behavior with an understanding
of the fact that the systems under study are control systems* .”
It is largely this obstructionism, which goes back 2 1/2 decades,
that inhibits me from posting many of my ideas to CSGnet, instead
circulating them to one or two people at a time as I now do. It’s
why ECACS was founded, to allow ideas about “Complex Adaptive
Control Systems” to be “Explored”.

If I thought I would get reasoned criticism on CSGnet without it

being swamped by a flood of nonsense (see the first part of this
message for a good example), I would ask for help with ideas, some
of which are half-baked, some of which are ready for public tasting,
some of which are mere speculation about possibility.

Here's an example of the last class (mere speculation about

possibility). I submit it here as a test case. If anyone on CSGnet
can shed any light on it (or show how it fails) I would be
delighted:

Control is an emergent property of a control loop that depends on

four components being connected in a certain structure. When applied
to that structure, two external influences cause things to happen
that we see as control. Looking at larger sets of interconnected
control structures that deal with external variables that are
mutually associated or correlated, hierarchic control systems seem
to have an emergent property analogous to tensegrity, giving the
larger structures strength and flexibility, but also laying them
open to what in 1993 I first called “The Bomb in the Hierarchy”, an
explosive collapse of control.

I believe the emergent tensegrity property of physical structures in

3-D requires at least 9 components. Since complex control structures
exist in potentially very high-dimensionality spaces, is there any
such basic minimal structure for the control analogue of tensegrity
to emerge, or does the analogy fail when examined more closely.
Within the high dimensionality structure, one could examine a 3-D
subspace in which an apparent tensegrity analogue structure could
form, but would it tend to collapse or be strengthened by the larger
hyperspace in which it is embedded?

Martin

···

Martin Taylor (2017.10.21.08.28)–

…

            MT: I haven't actually seen the demo, but I

thought the write-up was great. It made the same point I
tried to get across with my “stochastic collective
control” demo several months ago, which caused you some
angst. I guess it’s OK when Bill says it, but not when I
say that the perceptual representations of environmental
variables can be distributed rather than localized in
specific “wires” in the brain.

          RM: The demo I'm talking about has nothing to do with

showing that “perceptual representations of environmental
variables can be distributed rather than localized in
specific “wires” in the brain”. It showed that N different
control systems can successfully control N different
aspects (linear combinations) of the same environmental
variables.

                    RM: In this demo, N

different control systems are controlling N
different linear combinations of the same
environmental variables. The individual
environmental variables themselves are not
controlled (as can be seen in the demo as the
wandering purple circles) but the aspects of
these variables – the linear combinations of
these variables that are the controlled
perceptions – are controlled (as can be seen in
the demo as the red circles staying close to the
white (reference) circles).

          If that's what your "stochastic collective control"

demo showed than I have no idea why it appeared to cause
me angst. But my money is on it having shown nothing like
that.

…

          RM: I don't think any function of environmental

variables can be considered more real than another. I
think it’s better to say that the perceptual functions
that matter – those that we have inherited through
evolution – are the ones that represent aspects of the
environment that we have to be able to control in order to
survive.

            MT: If you perceive yourself living

in a democracy whereas you are actually living in North
Korea, and you criticize the Dear Leader, you are just
as dead as you would be if you failed to duck a rock
that you wrongly perceive to be missing your head. Does
the dictatorship exist in the environment? I would say
it does.

          RM: I would say that the basis for that perception

exists in the environment. But if you want to think that
the correlates of perceptual variables exist in the
environment, that’s fine. It’s not really an important
thing to worry about when you do research.

          For example, I don't worry about whether vertical

optical velocity is in the environment or not when
modeling object interception. I just model it as a
function of variables that are in the environment and the
model behaves just like the real thing. The only problem I
see with thinking of controlled variables as existing in
the environment is when this leads to the notion (which I
believe you expressed) that it is only the perception of
the environmental variable that is controlled and not the
environmental variable itself. As I’ve said, this is not
only a complete misunderstanding of the concept of
“control of perception”, if taken seriously it would mean
that the study of the controlling done by living systems
is impossible.

            MT: It's interesting that a disagreement about whether

one branch of experimental psychology constitutes all of
experimental psychology, mixed in with a difference of
opinion as to the meanings of “basis” and “revolution”,
provides evidence that I am an enemy of (or fail to
understand) PCT. I’m not interested in getting into that
argument. I will let what I say about PCT speak for
itself. There are things that we do and will disagree on
with respect to PCT. That’s healthy in any science. But
a persistent tendency to claim that any disagreement
with your opinion is tantamount to treason or heresy is
neither scientific nor healthy.

      RM: All I claim is that you have a fundamentally different

idea of what PCT is about than I do. This difference doesn’t
mean you are an enemy or traitor. Though I think you do an
injustice to Bill’s legacy by treating the truly revolutionary
implications of his work – that it invalidates the
foundations of scientific psychology-- as merely a polemical
device. And I do think you criticisms of my explanation of the
power law of movement reflected a rather profound difference
between your understanding of PCT and mine.

      But that's not heretical or

traitorous. No, my main complaint about you (and Bruce A) is
that you are just not helping me do the work that I think
needs to be done, and that is to figure out how to study
behavior with an understanding of the fact that the systems
under study are control systems.

[rsmarken]

[From Rick Marken (2017.10.26.2230)]

···

Martin Taylor (2017.10.23.23.10)–

MT: It shows both at the same time. That’s obvious in the write-up.

RM: The environmental variables in the write up are q1, q2 and q3. Figure 5 shows that what are controlled are three different linear combinations of these environmental variables – three different perceptions of the same environmental variables. So what is obvious to me in the write up is that the linear combinations of variables – the perceptions – are controlled, not the environmental variables. This is also obvious when you run the demo; the perceptions stay close to their references while the environmental variables vary as they are pushed about by disturbances and the effects of the outputs of the control systems.

Â

MT: Maybe the demo doesn't make it so obvious. But you can see it for

yourself by ignoring all but one of the environmental variables and
seeing whether that one is controlled.

RM: I presume you mean that I can “see it” in my mind’s eye. But actually I can’t. But when I run the demo I can see with my own eyes that none of the environmental variables are controlled in the sense that they are not being maintained in constant reference states…

Â

MT: Since they all are, I think

you will find that the one you look at is. Its internal
representation is distributed over all the perceptual variables.

RM: In the demo none of the N environmental variables are controlled; what are controlled are linear combinations (aspects) of these environmental variables. We control aspects (functions) of environmental variables – perceptions – not the “raw” environmental variables that physics and chemistry tell us are out there.

MT: I’m glad you see that the same way I do.

 RM: Me too.Â

MT: As I explained in detail in a recent message [Martin Taylor

2017.10.18.16.18], over the last couple of years I have come to a
much more nuanced view of that issue. I concluded by saying:
“Perception is all we have that we can control. The environment is
what we really need to control. We can control the environment only
to the extend that it behaves as though our perceptions correspond
to it reasonably well. So for most practical purposes other than
careful theory or metaphysical philosophy, it doesn’t matter a whit
whether we say that perception or the CEV is controlled.”

RM: The PCT view of the matter is that the controlled perception and what you call the CEV are the same variable seen from two different points of view. The controlled perception, p, is the controlled aspect of the environment from the perspective of the control system and the CEV (called q.i in PCT) is the same controlled aspect of the environment from the point of view of the observer of the control system.

MT: When I criticize your explanation of the power law of movement, it

has nothing whatever to do with PCT. It has to do with your misuse
of mathematics, and falling for the old “divide by zero” trick used
to prove nonsense to unsuspecting fifth-graders, albeit yours is in
an unfamiliar guise. But of course you know this.

RM: Your mathematics are irrelevant because the equations for velocity and curvature that we present in the paper are the ones power law researchers use to calculate these variables from the observed movement data. The calculated values of velocity and curvature are then used in a regression analysis to determine how well the relationship between this variables is fit by a power function. Your criticism would be pertinent if you could show that the equations we used are not the ones power law researchers actually use to calculate the instantaneous velocity and curvature of movement. And you should also show that when the measures of velocity and curvature obtained with what you say are the correct computational equations are regressed on each other the best fit regression line is a power function with a coefficient close to 1/3 or 2/3.Â

RM: But you don’t need math to see that the power law of movement is an example of a behavioral illusion. All you have to know is how PCT explains intentionally produced curved movements; our mathematical analysis just shows why it is consistently found that the relationship between curvature and velocity in such movement (as well as in unintentionally produced curved movements) is a power law with a coefficient near 1/3 or 2/3.Â

MT: That's hilarious. I have the same complaint about your

obstructionism, for which I could use the same words: “* you are
just not helping me do the work that I think needs to be done, and
that is to figure out how to study behavior with an understanding
of the fact that the systems under study are control systems*.”

Â

RM: Yes, that’s because we have such different views of what PCT Is about.Â

MT: If I thought I would get reasoned criticism on CSGnet without it

being swamped by a flood of nonsense (see the first part of this
message for a good example), I would ask for help with ideas, some
of which are half-baked, some of which are ready for public tasting,
some of which are mere speculation about possibility.

MT: Here's an example of the last class (mere speculation about

possibility). I submit it here as a test case. If anyone on CSGnet
can shed any light on it (or show how it fails) I would be
delighted:

MT: Control is an emergent property of a control loop that depends on

four components being connected in a certain structure. When applied
to that structure, two external influences cause things to happen
that we see as control. Looking at larger sets of interconnected
control structures that deal with external variables that are
mutually associated or correlated, hierarchic control systems seem
to have an emergent property analogous to tensegrity, giving the
larger structures strength and flexibility, but also laying them
open to what in 1993 I first called “The Bomb in the Hierarchy”, an
explosive collapse of control.

MT: I believe the emergent tensegrity property of physical structures in

3-D requires at least 9 components. Since complex control structures
exist in potentially very high-dimensionality spaces, is there any
such basic minimal structure for the control analogue of tensegrity
to emerge, or does the analogy fail when examined more closely.
Within the high dimensionality structure, one could examine a 3-D
subspace in which an apparent tensegrity analogue structure could
form, but would it tend to collapse or be strengthened by the larger
hyperspace in which it is embedded?

RM: Maybe someone on CSGNet could help you out with this. But I’m interested in how PCT explains actual behavior. This looks to me like a question about mathematical properties of the model itself with no reference to how these properties relate to some observed behavior. That doesn’t mean there is anything wrong with your approach, just that it’s not interesting to me. This is just another way we differ in our views of what PCT is about. I view PCT as a theory that is designed to explain observable behavior; you seem to view PCT as an interesting mathematical model in itself. This is another reason why we you can’t help me do my work on PCT and I cant help you do yours.Â

          RM: The demo I'm talking about has nothing to do with

showing that “perceptual representations of environmental
variables can be distributed rather than localized in
specific “wires” in the brain”. It showed that N different
control systems can successfully control N different
aspects (linear combinations) of the same environmental
variables.

          RM: I don't think any function of environmental

variables can be considered more real than another. I
think it’s better to say that the perceptual functions
that matter – those that we have inherited through
evolution – are the ones that represent aspects of the
environment that we have to be able to control in order to
survive.Â

          RM: For example, I don't worry about whether vertical

optical velocity is in the environment or not when
modeling object interception. I just model it as a
function of variables that are in the environment and the
model behaves just like the real thing. The only problem I
see with thinking of controlled variables as existing in
the environment is when this leads to the notion (which I
believe you expressed) that it is only the perception of
the environmental variable that is controlled and not the
environmental variable itself. As I’ve said, this is not
only a complete misunderstanding of the concept of
“control of perception”, if taken seriously it would mean
that the study of the controlling done by living systems
is impossible.

      RM: All I claim is that you have a fundamentally different

idea of what PCT is about than I do. This difference doesn’t
mean you are an enemy or traitor. Though I think you do an
injustice to Bill’s legacy by treating the truly revolutionary
implications of his work – that it invalidates the
foundations of scientific psychology-- as merely a polemical
device. And I do think you criticisms of my explanation of the
power law of movement reflected a rather profound difference
between your understanding of PCT and mine.

      RM: But that's not heretical or

traitorous. No, my main complaint about you (and Bruce A) is
that you are just not helping me do the work that I think
needs to be done, and that is to figure out how to study
behavior with an understanding of the fact that the systems
under study are control systems.

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

[rupert]

[From Rupert Young (2017.10.29 16.40)]

  (Rick Marken

(2017.10.20.2150)]

Is the source code for this around somewhere?

Rupert

···

https://www.dropbox.com/s/2u00ac87bix2sjv/MultiControlPrj.exe?dl=0

[rsmarken]

[From Rick Marken (2017.10.29.1005)]

MultiControlUnit.pas.txt (7.34 KB)

···

Rupert Young (2017.10.29 16.40)

  (Rick Marken

(2017.10.20.2150)]

Is the source code for this around somewhere?

RM: Here’s a text version of the main program.Â

https://www.dropbox.com/s/2u00ac87bix2sjv/MultiControlPrj.exe?dl=0

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