Bringing it all back home

[_bara0361]

Yes, thank you for tying it all together so neatly…

clap clap clap

···

On Thu, Apr 24, 2014 at 3:42 PM, Fred Nickols fred@nickols.us wrote:

Nice job, Rick.

Fred Nickols

Performance Improvement Professional

Distance Consulting LLC

fred@nickols.us

www.nickols.us

Richard Marken wrote:

[From Rick Marken (2014.04.24.1310)]

This is a reply to three threads, the one’s called “Words”, "What’s wrong

with this picture" and “Self-Regulation”. I think these topics have close

ties to one another so I’m going reply to all three at the same time and

try to show how they are all related to one another and, hopefully, bring

it all back home to Perceptual Control Theory.

Here are the issues as I see them:

The “Words” thread seems to be about whether anything is lost if we avoid

the word “control” in favor terms like “self-regulation” in the the hopes

of making PCT sound more attractive to potential “buyers”; the "What’s

wrong with this picture" thread is about what might be wrong with the

control system diagram of a fly ball governor from a PCT perspective;

and the “Self-Regulation” thread seems to be about whether self-regulation

theories are equivalent to PCT.

I think avoiding the term “control” when discussing PCT is a bad idea for

many reasons not the least of which being that PCT is based on recognition

that the behavior of living organisms is control. Powers’ fundamental
insight was that control is a phenomenon – a fact – and that the

behavior

of living organisms is an example of that phenomenon: the production of

consistent results under circumstances that should produce inconsistency.

Understanding the phenomenon of control and how it is seen in the behavior

of living organisms is central to a PCT-based understand of living

systems.

So using a term other than “control” to describe PCT may be a good idea if

you’re trying to sell PCT but it’s a bad idea if you want to keep the

central focus of the theory – the phenomenon of control as it is seen in

the behavior of living systems-- front and center. I prefer to use labels

in a way that gives an accurate picture of what is being described rather

than as propaganda.

Keeping the word “control” in the description of PCT is also important

because Powers did not invent control theory. What he “invented” was the

correct application of control theory to the behavior of living systems.

And Bill was able to do this because of his recognition of the fact that

the behavior of living systems is control. It was recognition of the* fact

of control* that allowed Powers to apply the already existing *theory of

control *correctly to the behavior of living organisms. Which brings us to
what is wrong, form a PCT perspective, with the control system diagram of

the Watt governor:

[image: Inline image 1]

There is really nothing wrong with this diagram from the point of view of

an engineer who is trying to build a stable control system. The engineer

wants to build a system that keeps shaft speed, y(t), nearly equal to the

set point speed, r(t), with very little oscillation (instability) around

it. But from the point of view of a psychologist trying to understand how

an existing stable control system (a living organism) works this diagram

is

misleading and leaves out the most important thing we want to know about

that control control system in order to understand its: what it is

controlling.

Because Powers understood that behavior is control he knew that when we

look at the behavior of a living system we are looking at the behavior of

a

control system from the outside. And what we see are it’s actions – which

are equivalent to the movements of the collar of the flyball governor in

the diagram above – and the results of those actions – some of which are

equivalent to the controlled shaft speed, y(t), in the diagram above. So

the psychologist looking at the behavior of a living control system

doesn’t

know what the engineer looking at the behavior of an artifactual control

system knows: what variable it is controlling. The engineer working to

improve the stability of the shaft speed control system knows that the

system is controlling shaft speed because that’s what it’s been built to

control; the psychologist looking at the behavior of a living control

system can only guess at what that system is controlling. PCT shows how

the

psychologist can test that guess; and it shows it by adding some features

to the control diagram that were not really needed in the diagram used by

the engineer.

The main thing that is missing from the diagram (from a PCT perspective)

is

an indication that a controlled variable (y(t) in the diagram) has to be

perceived in order to be controlled. The engineer automatically includes a

mechanisms for perceiving the variable to be controlled – such as the

spinning flyball that “perceives” the the speed of shaft rotation – so an

engineering diagram typically leaves off a box for the sensor element

because it’s taken for granted. But Powers understood that including the

perceptual function box in the control diagram of a living system was

essential because determining what a living system is controlling is a

matter of determining what perceptions it is controlling. So that’s the

main problem that I see with the control diagram above; it takes for

granted that there is a mechanism for perceiving y(t) and that it is the

perceive value of y(t) that is controlled. If course, another problem is

that r(t) can be manipulated from outside the system with artifactual

control systems but not with living control systems. This is another

aspect

of the remapping of the control model to the behavior of living system

when

that behavior is understood to be an example of control.

So Powers’ recognition that behavior is control resulted in the proper
application (or mapping) of control theory to behavior, a mapping that we

here on CSGNet call PCT. And this proper mapping-- whatever you want to

call it, PCT, self-regulation, goal tending – results in a picture of

behavior as the control of perception. And when behavior is understood to

be the control of perception, understand the behavior is, first and

foremost, a matter of determining what perceptions the behaving system is

controlling. So if “self-regulation” theory is the same as PCT (if it maps

control theory to behavior in the same way as specified in B:CP) then

research based on self-regulation should be all about figuring out what

perceptual variables organisms are controlling. To my knowledge there is

not much research on self-regulation that has the determination of

controlled perceptual variables as its main aim. But I may be wrong about

this. Of course, determining what variables are controlled is not the only

aim of research based on PCT (or whatever you want to call theories that

properly map control theory to behavior). It’s also interesting to find

out

how people control the variables they control – and I suspect that this

is

mainly what Jeff’s research and modeling is about. But I do think that

determining what variables organisms control should really be the first

order of business in research based on PCT.

But my point here is simply that the way to judge whether any theory is

(or

is not) PCT (a theory that is based on a proper mapping of control theory

to behavior) is to look at the research, not the flow diagrams. I say this

based on my experience reading the first book by Carver and Scheier (I

think it was published in 1982). I was very excited when I discovered this

book because I thought I was one of the only people in the world (along

with Tom Bourbon, David Goldstein and Bill, of course) doing research on

PCT (it was just called control theory back then). I remember the first

couple of chapters of that book as giving an excellent account of Powers’

theory. But the wheels fell of in later chapter when they described their

research, which had nothing to do with determining controlled variables.

So

I learned way back then that it was possible to get all the diagrams right

and still not understand what the theory that we now call PCT is all

about.

It’s about understanding the controlling done by living systems which,

when control theory is properly applied to their behavior, means

*determining

the perceptual variables* that living systems control.

Best

Rick

–

Richard S. Marken PhD

www.mindreadings.com

It is difficult to get a man to

understand<http://en.wikiquote.org/wiki/Understand> something,
when his salary depends upon his not understanding it. – Upton Sinclair

[FredNickols]

Nice job, Rick.

Fred Nickols
Performance Improvement Professional
Distance Consulting LLC
fred@nickols.us
www.nickols.us

Richard Marken wrote:

···

[From Rick Marken (2014.04.24.1310)]

This is a reply to three threads, the one's called "Words", "What's wrong
with this picture" and "Self-Regulation". I think these topics have close
ties to one another so I'm going reply to all three at the same time and
try to show how they are all related to one another and, hopefully, bring
it all back home to Perceptual Control Theory.

Here are the issues as I see them:

The "Words" thread seems to be about whether anything is lost if we avoid
the word "control" in favor terms like "self-regulation" in the the hopes
of making PCT sound more attractive to potential "buyers"; the "What's
wrong with this picture" thread is about what might be wrong with the
control system diagram of a fly ball governor from a PCT perspective;
and the "Self-Regulation" thread seems to be about whether self-regulation
theories are equivalent to PCT.

I think avoiding the term "control" when discussing PCT is a bad idea for
many reasons not the least of which being that PCT is based on recognition
that the behavior of living organisms *is* control. Powers' fundamental
insight was that control is a phenomenon -- a fact -- and that the
behavior
of living organisms is an example of that phenomenon: the production of
consistent results under circumstances that should produce inconsistency.
Understanding the phenomenon of control and how it is seen in the behavior
of living organisms is central to a PCT-based understand of living
systems.
So using a term other than "control" to describe PCT may be a good idea if
you're trying to sell PCT but it's a bad idea if you want to keep the
central focus of the theory -- the phenomenon of control as it is seen in
the behavior of living systems-- front and center. I prefer to use labels
in a way that gives an accurate picture of what is being described rather
than as propaganda.

Keeping the word "control" in the description of PCT is also important
because Powers did not invent control theory. What he "invented" was the
correct application of control theory to the behavior of living systems.
And Bill was able to do this because of his recognition of the fact that
the behavior of living systems is control. It was recognition of the* fact
of control* that allowed Powers to apply the already existing *theory of
control *correctly to the behavior of living organisms. Which brings us to
what is wrong, form a PCT perspective, with the control system diagram of
the Watt governor:

[image: Inline image 1]

There is really nothing wrong with this diagram from the point of view of
an engineer who is trying to build a stable control system. The engineer
wants to build a system that keeps shaft speed, y(t), nearly equal to the
set point speed, r(t), with very little oscillation (instability) around
it. But from the point of view of a psychologist trying to understand how
an existing stable control system (a living organism) works this diagram
is
misleading and leaves out the most important thing we want to know about
that control control system in order to understand its: what it is
controlling.

Because Powers understood that behavior is control he knew that when we
look at the behavior of a living system we are looking at the behavior of
a
control system from the outside. And what we see are it's actions -- which
are equivalent to the movements of the collar of the flyball governor in
the diagram above -- and the results of those actions -- some of which are
equivalent to the controlled shaft speed, y(t), in the diagram above. So
the psychologist looking at the behavior of a living control system
doesn't
know what the engineer looking at the behavior of an artifactual control
system knows: what variable it is controlling. The engineer working to
improve the stability of the shaft speed control system knows that the
system is controlling shaft speed because that's what it's been built to
control; the psychologist looking at the behavior of a living control
system can only guess at what that system is controlling. PCT shows how
the
psychologist can test that guess; and it shows it by adding some features
to the control diagram that were not really needed in the diagram used by
the engineer.

The main thing that is missing from the diagram (from a PCT perspective)
is
an indication that a controlled variable (y(t) in the diagram) has to be
perceived in order to be controlled. The engineer automatically includes a
mechanisms for perceiving the variable to be controlled -- such as the
spinning flyball that "perceives" the the speed of shaft rotation -- so an
engineering diagram typically leaves off a box for the sensor element
because it's taken for granted. But Powers understood that including the
perceptual function box in the control diagram of a living system was
essential because determining what a living system is controlling is a
matter of determining what perceptions it is controlling. So that's the
main problem that I see with the control diagram above; it takes for
granted that there is a mechanism for perceiving y(t) and that it is the
perceive value of y(t) that is controlled. If course, another problem is
that r(t) can be manipulated from outside the system with artifactual
control systems but not with living control systems. This is another
aspect
of the remapping of the control model to the behavior of living system
when
that behavior is understood to be an example of *control*.

So Powers' recognition that behavior *is* control resulted in the proper
application (or mapping) of control theory to behavior, a mapping that we
here on CSGNet call PCT. And this proper mapping-- whatever you want to
call it, PCT, self-regulation, goal tending -- results in a picture of
behavior as the control of perception. And when behavior is understood to
be the control of perception, understand the behavior is, first and
foremost, a matter of determining what perceptions the behaving system is
controlling. So if "self-regulation" theory is the same as PCT (if it maps
control theory to behavior in the same way as specified in B:CP) then
research based on self-regulation should be all about figuring out what
perceptual variables organisms are controlling. To my knowledge there is
not much research on self-regulation that has the determination of
controlled perceptual variables as its main aim. But I may be wrong about
this. Of course, determining what variables are controlled is not the only
aim of research based on PCT (or whatever you want to call theories that
properly map control theory to behavior). It's also interesting to find
out
how people control the variables they control -- and I suspect that this
is
mainly what Jeff's research and modeling is about. But I do think that
determining what variables organisms control should really be the first
order of business in research based on PCT.
But my point here is simply that the way to judge whether any theory is
(or
is not) PCT (a theory that is based on a proper mapping of control theory
to behavior) is to look at the research, not the flow diagrams. I say this
based on my experience reading the first book by Carver and Scheier (I
think it was published in 1982). I was very excited when I discovered this
book because I thought I was one of the only people in the world (along
with Tom Bourbon, David Goldstein and Bill, of course) doing research on
PCT (it was just called control theory back then). I remember the first
couple of chapters of that book as giving an excellent account of Powers'
theory. But the wheels fell of in later chapter when they described their
research, which had nothing to do with determining controlled variables.
So
I learned way back then that it was possible to get all the diagrams right
and still not understand what the theory that we now call PCT is all
about.
It's about understanding the *controlling* done by living systems which,
when *control theory is properly applied* to their behavior, means
*determining
the perceptual variables* that living systems control.

Best

Rick

--
Richard S. Marken PhD
www.mindreadings.com

It is difficult to get a man to
understand<http://en.wikiquote.org/wiki/Understand&gt; something,
when his salary depends upon his not understanding it. -- Upton Sinclair

[MartinT]

[Martin Taylor 2014.04.25.00.19]

It's an unusual scientist who knows that a particular theory is

actually correct.
Congratulations!
Martin

···

On 2014/04/25 12:11 AM, Richard Marken
wrote:

[From Rick Marken (2014.04.24.2110)]

        ... What is important about PCT is

that it is the first correct use** of control
theory as a model of behavior; correct in the sense that
the variables and functions in the engineering control
model are correctly mapped to their counterparts in the
behaving system.

[rsmarken]

[From Rick Marken (2014.04.24.1310)]

This is a reply to three threads, the one’s called “Words”, “What’s wrong with this picture” and “Self-Regulation”. I think these topics have close ties to one another so I’m going reply to all three at the same time and try to show how they are all related to one another and, hopefully, bring it all back home to Perceptual Control Theory.

Here are the issues as I see them:

The “Words” thread seems to be about whether anything is lost if we avoid the word “control” in favor terms like “self-regulation” in the the hopes of making PCT sound more attractive to potential “buyers”; the “What’s wrong with this picture” thread is about what might be wrong with the control system diagram of a fly ball governor from a PCT perspective; and the “Self-Regulation” thread seems to be about whether self-regulation theories are equivalent to PCT.

I think avoiding the term “control” when discussing PCT is a bad idea for many reasons not the least of which being that PCT is based on recognition that the behavior of living organisms is control. Powers’ fundamental insight was that control is a phenomenon – a fact – and that the behavior of living organisms is an example of that phenomenon: the production of consistent results under circumstances that should produce inconsistency. Understanding the phenomenon of control and how it is seen in the behavior of living organisms is central to a PCT-based understand of living systems. So using a term other than “control” to describe PCT may be a good idea if you’re trying to sell PCT but it’s a bad idea if you want to keep the central focus of the theory – the phenomenon of control as it is seen in the behavior of living systems-- front and center. I prefer to use labels in a way that gives an accurate picture of what is being described rather than as propaganda.

Keeping the word “control” in the description of PCT is also important because Powers did not invent control theory. What he “invented” was the correct application of control theory to the behavior of living systems. And Bill was able to do this because of his recognition of the fact that the behavior of living systems is control. It was recognition of the fact of control that allowed Powers to apply the already existing theory of control correctly to the behavior of living organisms. Which brings us to what is wrong, form a PCT perspective, with the control system diagram of the Watt governor:

There is really nothing wrong with this diagram from the point of view of an engineer who is trying to build a stable control system. The engineer wants to build a system that keeps shaft speed, y(t), nearly equal to the set point speed, r(t), with very little oscillation (instability) around it. But from the point of view of a psychologist trying to understand how an existing stable control system (a living organism) works this diagram is misleading and leaves out the most important thing we want to know about that control control system in order to understand its: what it is controlling.

Because Powers understood that behavior is control he knew that when we look at the behavior of a living system we are looking at the behavior of a control system from the outside. And what we see are it’s actions – which are equivalent to the movements of the collar of the flyball governor in the diagram above – and the results of those actions – some of which are equivalent to the controlled shaft speed, y(t), in the diagram above. So the psychologist looking at the behavior of a living control system doesn’t know what the engineer looking at the behavior of an artifactual control system knows: what variable it is controlling. The engineer working to improve the stability of the shaft speed control system knows that the system is controlling shaft speed because that’s what it’s been built to control; the psychologist looking at the behavior of a living control system can only guess at what that system is controlling. PCT shows how the psychologist can test that guess; and it shows it by adding some features to the control diagram that were not really needed in the diagram used by the engineer.

The main thing that is missing from the diagram (from a PCT perspective) is an indication that a controlled variable (y(t) in the diagram) has to be perceived in order to be controlled. The engineer automatically includes a mechanisms for perceiving the variable to be controlled – such as the spinning flyball that “perceives” the the speed of shaft rotation – so an engineering diagram typically leaves off a box for the sensor element because it’s taken for granted. But Powers understood that including the perceptual function box in the control diagram of a living system was essential because determining what a living system is controlling is a matter of determining what perceptions it is controlling. So that’s the main problem that I see with the control diagram above; it takes for granted that there is a mechanism for perceiving y(t) and that it is the perceive value of y(t) that is controlled. If course, another problem is that r(t) can be manipulated from outside the system with artifactual control systems but not with living control systems. This is another aspect of the remapping of the control model to the behavior of living system when that behavior is understood to be an example of control.

So Powers’ recognition that behavior is control resulted in the proper application (or mapping) of control theory to behavior, a mapping that we here on CSGNet call PCT. And this proper mapping-- whatever you want to call it, PCT, self-regulation, goal tending – results in a picture of behavior as the control of perception. And when behavior is understood to be the control of perception, understand the behavior is, first and foremost, a matter of determining what perceptions the behaving system is controlling. So if “self-regulation” theory is the same as PCT (if it maps control theory to behavior in the same way as specified in B:CP) then research based on self-regulation should be all about figuring out what perceptual variables organisms are controlling. To my knowledge there is not much research on self-regulation that has the determination of controlled perceptual variables as its main aim. But I may be wrong about this. Of course, determining what variables are controlled is not the only aim of research based on PCT (or whatever you want to call theories that properly map control theory to behavior). It’s also interesting to find out how people control the variables they control – and I suspect that this is mainly what Jeff’s research and modeling is about. But I do think that determining what variables organisms control should really be the first order of business in research based on PCT.

But my point here is simply that the way to judge whether any theory is (or is not) PCT (a theory that is based on a proper mapping of control theory to behavior) is to look at the research, not the flow diagrams. I say this based on my experience reading the first book by Carver and Scheier (I think it was published in 1982). I was very excited when I discovered this book because I thought I was one of the only people in the world (along with Tom Bourbon, David Goldstein and Bill, of course) doing research on PCT (it was just called control theory back then). I remember the first couple of chapters of that book as giving an excellent account of Powers’ theory. But the wheels fell of in later chapter when they described their research, which had nothing to do with determining controlled variables. So I learned way back then that it was possible to get all the diagrams right and still not understand what the theory that we now call PCT is all about. It’s about understanding the controlling done by living systems which, when control theory is properly applied to their behavior, means determining the perceptual variables that living systems control.

Best

Rick

···

–
Richard S. Marken PhD
www.mindreadings.com
It is difficult to get a man to understand something, when his salary depends upon his not understanding it. – Upton Sinclair

[rsmarken]

[From Rick Marken (2014.04.24.2110)]

···

On Thu, Apr 24, 2014 at 5:52 PM, bara0361@gmail.com bara0361@gmail.com wrote:

Yes, thank you for tying it all together so neatly…

clap clap clap

RM: Thanks Fred and Barb. I’m glad you were able to make through all the typos and arcane grammatical structures. Much of what I said in that post I say much more clearly (I hope, since I spent much more time preparing it) in the paper I wrote for the memorial volume of LCS IV. One of the main points I make in that paper (and that hopefully came through in my post) is that what is important about PCT is neither the first nor the only use of control theory as a model of behavior (Bill himself would have been the first to agree with that; he was very well acquainted with the history of the application of control theory in psychology). What is important about PCT is that it is the first correct use**of control theory as a model of behavior; correct in the sense that the variables and functions in the engineering control model are correctly mapped to their counterparts in the behaving system.

Best regards

Rick

–
Richard S. Marken PhD
www.mindreadings.com
It is difficult to get a man to understand something, when his salary depends upon his not understanding it. – Upton Sinclair

On Thu, Apr 24, 2014 at 3:42 PM, Fred Nickols fred@nickols.us wrote:

Nice job, Rick.

[rsmarken]

[From Rick Marken (2014.04.24.2150)]

Martin Taylor (2014.04.25.00.19)--

Rick Marken (2014.04.24.2110)

... What is important about PCT is that it is the first correct use of control theory as a model of behavior; correct in the sense that the variables and functions in the engineering control model are correctly mapped to their counterparts in the behaving system.

MT: It's an unusual scientist who knows that a particular theory is actually correct.

RM: An excellent point. But I will point out that I didn't say that the theory is correct; I said it's the mapping of theory (control theory) to phenomenon (control behavior) that is correct.
RM: It's actually pretty easy show that PCT maps control theory to behavior correctly while other applications of control theory in psychology don't. For example, you can use control theory to predict behavior in a tracking task using target position as the value of the reference signal, r(t), in the control equations. And your prediction of the behavior in that task will be as precise as the prediction using PCT, which maps the reference signal to an internally set efferent neural signal in the brain of the subject.
RM: You can show that the PCT mapping is correct by simply repeating a tracking run after asking the subject to keep the cursor 1/2 inch to the left of the target. Now your predictions of behavior using the target as the reference signal are 1/2 inch off while the PCT prediction, using a new estimate for the internally set reference, are still right on. So mapping the reference signal in control theory to an external "stimulus", as is typically done in non-pCT applications of control theory, is clearly incorrect.
RM: So PCT may eventually be shown to be wrong (though what replaces it will not be like any non-PCT theory that is currently around). But until it is shown to be incorrect (in the sense that it is unable to account for the results of a test), it currently represents the correct way to map the control model to control behavior and it is currently the only theory that can explain the controlling done by living systems.
Best regards
Rick

···

--
Richard S. Marken PhD
<http://www.mindreadings.com>www.mindreadings.com

It is difficult to get a man to <http://en.wikiquote.org/wiki/Understand&gt;understand something, when his salary depends upon his not understanding it. -- Upton Sinclair

[MartinT]

[Martin Taylor 2014.04.25.09.20]

While I personally believe most of your conclusions are valid, I

don’t think your argument is.
The problem with this is that other control theories that don’t show
control of input in their diagrams nevertheless wind up with exactly
the same mathematics, and hence are indistinguishable from PCT by
any conceivable experiment of that kind.
Again, one only has to postulate a theory in which the reference
value is determined by some external stimulus to produce
mathematically identical results.
For me, the big problem isn’t with single control loops, but with
how they can be combined into an effective structure of many
interacting loops. Bill produced one such structure, for which the
key is indeed the mapping you describe, because it permits the
reference input to be inside the organism, whereas so many of the
mathematically equivalent competing structures put the reference and
the disturbance in the external environment. Since they are the only
two external inputs to a control loop, if they are both outside the
organism (other than possible controls such as on the gain), the
organism cannot contain complex structures of interacting simple
loops, though I suppose that in principle loops could be described
with other inputs – but what would those inputs do?
As I said, I agree with your conclusion. But I would arrive at that
conclusion differently.
I would start with the fact (if the laws of thermodynamics are to be
taken as correct) that any entity will decay entropically over time
if nothing is done to repair damage inflicted by external effects.
There are three ways to delay or temporarily reverse entropic decay:
(1) the object might have an internal structure strong enough that
the energy distribution of external influences is too low to break
internal bonds (most of the time – there are always outliers); (2)
A shield or skin may be placed around the object strong enough in
the first sense to protect the state of the organism; (3) the
organism may act to counter external influences that might influence
its internal structure (or to repair damage already done).
Possibility 1 is the case for rocks, metal constructions, and so
forth. For them, the decay is usually slow on the time-scale of
lifetimes. It is not true for living things, for which external
energy levels are quite high enough to break many of the internal
bonds, though it may be true for bones and seashells.
Possibility 2 is the case for cells, most organisms, and
particularly for organisms with exoskeletons, as well as for less
obvious cases such as an aquarium, for which the glass shields the
water, keeping it in a shape that may change because of evaporation,
but which otherwise doesn’t change.
Possibility 3 is PCT. It matters because the ambient energy levels
are sufficient to rot the organism, whether it be a tree, a
bacterium, or a person, if that organism doesn’t use a flow-through
energy supply to counter external influences or to repair damage.
Possibility 3 is PCT because in order to counter an external
influence (a “disturbance” in PCT language), some state internal to
the organism must exist to direct the energy flow in such a way as
to counter the energy supplied by the disturbance. That internal
state is defined in PCT as a “perception”, and there must
necessarily be some state against which the perception is compared,
a “reference”, as well as some kind of operation that produces the
perception as a function of past and present sensory values. Without going through the long trail from “intrinsic variables”,
which the above covers, to the control of arbitrary perceptions as a
way to ensure control of intrinsic variables, I think that the
necessity of some form of PCT is provable. Moreover, the form that
is required is one in which control loops have a reference value
supplied from inside, and have the possibility of generating actions
on the external environment. I therefore believe that the mapping
you describe as Bill’s big contribution is necessary to any
successful control theory of the functioning of living organisms –
and hence of psychology.
Martin

···

[From Rick Marken (2014.04.24.2150)]

            Martin Taylor

(2014.04.25.00.19)–

Rick Marken (2014.04.24.2110)

                    ... What is important

about PCT is that it is the first correct use**
of
control theory as a model of behavior; correct
in the sense that the variables and functions
in the engineering control model are correctly
mapped to their counterparts in the behaving
system.

            MT: It's an unusual scientist who knows that a

particular theory is actually correct.

          RM: An excellent point. But I will point out that I

didn’t say that the theory is correct; I said it’s the *
mapping* of theory (control theory) to phenomenon (control
behavior) that is correct.

          RM: It's actually pretty easy show that PCT maps

control theory to behavior correctly while other
applications of control theory in psychology don’t. For
example, you can use control theory to predict behavior in
a tracking task using target position as the value of the
reference signal, r(t), in the control equations. And your
prediction of the behavior in that task will be as precise
as the prediction using PCT, which maps the reference
signal to an internally set efferent neural signal in the
brain of the subject.

          RM: You can show that the PCT mapping is correct by

simply repeating a tracking run after asking the subject
to keep the cursor 1/2 inch to the left of the target. Now
your predictions of behavior using the target as the
reference signal are 1/2 inch off while the PCT
prediction, using a new estimate for the internally set
reference, are still right on. So mapping the reference
signal in control theory to an external “stimulus”, as is
typically done in non-pCT applications of control theory,
is clearly incorrect.

          RM: So PCT may eventually be shown to be wrong (though

what replaces it will not be like any non-PCT theory that
is currently around). But until it is shown to be
incorrect (in the sense that it is unable to account for
the results of a test), it currently represents the
correct way to map the control model to control behavior
and it is currently the only theory that can explain the
controlling done by living systems.

[rsmarken]

[From Rick Marken (2014.04.27.1245)]

Martin Taylor (2014.04.25.09.20)--

RM: It's actually pretty easy show that PCT maps control theory to behavior correctly while other applications of control theory in psychology don't.

MT: The problem with this is that other control theories that don't show control of input in their diagrams nevertheless wind up with exactly the same mathematics, and hence are indistinguishable from PCT by any conceivable experiment of that kind.

RM: I agree that other control theories that don't show control of input can make exactly the same predictions as PCT in some circumstances but I pointed out one simple experiment that makes it possible to clearly distinguish PCT from other applications of control theory (the one where the person is asked to keep the cursor a fixed distance from the target; another one would be asking the person to move the cursor to different positions relative to the target whenever they felt like it during an experimental run). There are many other experiments -- including variations on the the ones I described -- that one can conceive that make it possible to distinguish PCT from other applications of control theory to behavior.

RM: You can show that the PCT mapping is correct by simply repeating a tracking run after asking the subject to keep the cursor 1/2 inch to the left of the target.

MT: Again, one only has to postulate a theory in which the reference value is determined by some external stimulus to produce mathematically identical results.

RM: But there is no external stimulus that could be the reference for the cursor being 1/2 inch. I think what you describe here is putting your conclusions into your explanation; theoretically "begging the question.

MT: For me, the big problem isn't with single control loops, but with how they can be combined into an effective structure of many interacting loops.

RM: It seems to me the question of how you combine multiple control loop is secondary to the question of how a single control loop is mapped to behavior. According to the conventional application of control theory to behavior, which places reference signals in the environment, it is actually the environment that is doing the controlling, acting via the organism to keep variables in the state it wants them in. This mistake certainly precedes any problems that might result from trying to combine multiple control systems.

RM: So PCT may eventually be shown to be wrong...it currently represents the correct way to map the control model to control behavior and it is currently the only theory that can explain the controlling done by living systems.

MT: As I said, I agree with your conclusion. But I would arrive at that conclusion differently.

RM: OK. I like my route better than yours but as long as we get to the same place it's fine with me. Take any route you like.
Best
Rick

···

--
Richard S. Marken PhD
<http://www.mindreadings.com>www.mindreadings.com

It is difficult to get a man to <http://en.wikiquote.org/wiki/Understand&gt;understand something, when his salary depends upon his not understanding it. -- Upton Sinclair

[MartinT]

[Martin Taylor 2014.04.28.10.52]

I suspect we were thinking of different scenarios. You are quite

correct for a scenario in which the subject autonomously decides
what the reference offset should be, given an instruction such as
“Choose a distance and keep the cursor that far to the left of the
target.” But you actually said the the instruction would specify
that distance. To me, that offers an opening for someone to argue
that this reference value is provided from outside.
Perhaps so, but I think the two are necessarily intertwined, and the
answer to how they can be combined forces the “correct” mapping,
rather than the “correct” mapping being provable by itself, even if
your argument as amended above is a sufficient proof.
Here’s the core of my supportive argument in a few sentences.
An organism has an inside and an outside, with a physical skin,
membrane, shell … that separates inside from outside. A control
loop has two and only two signal inputs from outside the loop, in
PCT labelled “reference” and “disturbance”. Other inputs to the loop
modulate its behaviour, but are not directly introduced signals. The
question at issue is whether both reference and disturbance inputs
to the loop can be outside the organism.
Suppose there exists a control loop for which the reference and
disturbance inputs are both outside the organism. Such a control
loop cannot interact with other control loops except by acting on
the external environment and trusting that its effects on the
external environment will influence the signal inputs to the other
control loops. Since the external environment changes, such an
organization, while possible, is unlikely to be very stable over
evolutionary time. Since the disturbance is by definition created by
changes in the environment, it follows that teh reference input must
be from inside the organism.
I’m happy with your route as amended to allow the subject to choose
the offset autonomously. and I accept that mine is no more (and no
less) mathematically provable than yours. Together, I think they
make the point better than either does alone.
Martin

···

[From Rick Marken (2014.04.27.1245)]

            Martin Taylor

(2014.04.25.09.20)–

…

                        RM: You can show that the PCT mapping is

correct by simply repeating a tracking run
after asking the subject to keep the cursor
1/2 inch to the left of the target.

            MT: Again, one only has to postulate a theory in which

the reference value is determined by some external
stimulus to produce mathematically identical results.

          RM: But there is no external stimulus that could be the

reference for the cursor being 1/2 inch. I think what you
describe here is putting your conclusions into your
explanation; theoretically "begging the question.

            MT: For me, the big

problem isn’t with single control loops, but with how
they can be combined into an effective structure of many
interacting loops.

          RM: It seems to me the question of how you combine

multiple control loop is secondary to the question of how
a single control loop is mapped to behavior.

          MT: As I said, I agree

with your conclusion. But I would arrive at that
conclusion differently.

          RM: OK. I like my route better than yours but as long

as we get to the same place it’s fine with me. Take any
route you like.