What is PCT a Model of?

[From Bruce Gregory (2004.01.12.1150)]

Bill Powers (2004.01.12.0843 MST)

PCT is a model of human and animal organization. You say I don't need
models of that model. I don't even know what that means.

I thought PCT was a model of human and animal behavior.

Bruce Gregory

"Everything that needs to be said has already been said. But since no
one was listening, everything must be said again."
                                                                                Andre Gide

"What is hateful to you, do not to your fellow men. That is the entire
Law; all the rest is commentary."

                                                                                The Talmud

[From Bill Powers (2004.01.12.0956 MST)]

Bruce Gregory (2004.01.12.1150) --

PCT is a model of human and animal organization. You say I don't need
models of that model. I don't even know what that means.

I thought PCT was a model of human and animal behavior.

I see it as a model of the organization of the neuromotor system that
explains behavior. A simulation based on the model imitates some behavior
of the real organism.

Best,

Bill P.

[From Bruce Gregory (2004.01.12.1210)]

Bill Powers (2004.01.12.0956 MST)

Bruce Gregory (2004.01.12.1150) --

PCT is a model of human and animal organization. You say I don't need
models of that model. I don't even know what that means.

I thought PCT was a model of human and animal behavior.

I see it as a model of the organization of the neuromotor system that
explains behavior. A simulation based on the model imitates some
behavior
of the real organism.

I trust that it will cause nobody any difficulties if I continue to
regard PCT as a model of purposeful behavior. As far as I know, all the
tests of the model have involved behavior rather than the organization
of the neuromotor system.

Bruce Gregory

"Everything that needs to be said has already been said. But since no
one was listening, everything must be said again."
                                                                                Andre Gide

"What is hateful to you, do not to your fellow men. That is the entire
Law; all the rest is commentary."

                                                                                The Talmud

[From Bill Powers (2004.01.12.1107 MSAT)]

Bruce Gregory (2004.01.12.1210) --

I see it as a model of the organization of the neuromotor system that
explains behavior. A simulation based on the model imitates some
behavior of the real organism.

I trust that it will cause nobody any difficulties if I continue to
regard PCT as a model of purposeful behavior.

I think of behavior as what we can see a person doing with (for
example) his arms and legs. A model of that would just move the arms and
legs the way the real person does, as in an animation.

A model of the organization of the organism, on the other hand, says that
regardless of how the person moves his arms and legs, behind these
movements is a relatively fixed structure consisting of input transducers,
input functions, comparators, output functions, and actuators, all
connected by signals in a specific way. This same organization is proposed
to exist no matter what behavior we see going on.

In the tracking model I just posted to my web site (which I hope you will
try), one model (with adjustable parameters) is used to produce all the
different behaviors we see, when 10 different disturbances are applied with
3 different degrees of difficulty. The patterns of mouse movements are
different in all of these 30 different cases (and by erasing the file
"Distable" you can always cause another randomly generated set of 30 cases
to be created when you start the program). All of these different behavior
patterns are generated by just one structural model of the hypothesized
control system.

As far as I know, all the tests of the model have involved behavior rather
than the organization of the neuromotor system.

I hope you can see that while it is the behavior of the model that is
matched to the behavior of the real person, it is the functional
organization of the model that is being tested as a proposed model of the
functional organization of the real person. This eliminates a large number
of possible functional organizations, although it does not select one
organization from the set of behaviorally equivalent functional
organizations. Other approaches are need to narrow the possibilities
further -- some of which I showed in B:CP.

I am attaching the files needed for a Vensim simulation using the data from
an experimental run. For Vensim PLE you have to set the model parameters by
hand; to do automatic optimization, you need Vensim DSS. Some of you have
copies of this advanced version, donated by Bob Eberlein, originator of
Vensim and CEO of Ventura Systems, Inc. I'll give more detailed
instructions for how to do the optimization to anyone who wants them. You
can at least show how this model works with Vensim, and verify that it
works properly using the results of my optimization method. The main thing
you need to do is import the data from the data file you generated during
the experimental run.

Best,

Bill P.

ctrlsys3.mdl (7.35 KB)

track3a.vgd (368 Bytes)

track3.vpd (38 Bytes)

track3.voc (327 Bytes)

[From Rick Marken (2004.01.12.1210)]

Bruce Gregory (2004.01.12.1210)]

I trust that it will cause nobody any difficulties if I continue to
regard PCT as a model of purposeful behavior.

It certainly doesn't cause me any difficulties but I think it could limit
the depth of your appreciation of the PCT model.

I think of PCT as a model of purposeful behavior in the sense that it
represents the organization of a neuromotor system that could produce such
behavior. The nature of this organization is rather easy to describe.
According to PCT the neuromotor system is organized to control the level of
firing of afferent neurons. It effects this control by causing variations in
the output systems --muscles and glands -- that have an effect on the world
outside the afferent neurons. Variations in the output systems are caused by
efferent neural signals, the amplitudes of which are proportional to the
difference between other efferent neural signals functioning as reference
signals and the afferent neural signals that are under control.

This model of the organization of the neuromotor system is quite different
than the conventional model. The main difference is that, in PCT, descending
efferent signals function as reference signals (at all levels except the
lowest level of the nervous system) while in the conventional model these
signals function as efferent signals. A simple diagram of the two
organizations would be as follows:

PCT:

Afferent -->synapse<--Efferent
               >
               > Efferent
               v
          Motor output

Conventional:

Afferent -->synapse
               >
               > Efferent
               v
          Motor output

The organizational difference is in the arrangement of the arrows indicating
the presumed direction of effect of of the neural components.

As far as I know, all the
tests of the model have involved behavior rather than the organization
of the neuromotor system.

True. But the have nervous system components of the models that simulated
the observed behaviors have been organized -- and had to be organized -- as
per the PCT rather than the Conventional organization above.

It would be nice to see some direct tests of neuromotor organization. I
expect to see some in the near future, as we get more PCT aware researchers
working in neurophysiology.

Best

Rick

···

--
Richard S. Marken
MindReadings.com
Home: 310 474 0313
Cell: 310 729 1400

[From Bruce Gregory (2004.01.12.1525)]

Rick Marken (2004.01.12.1210)

True. But the have nervous system components of the models that
simulated
the observed behaviors have been organized -- and had to be organized
-- as
per the PCT rather than the Conventional organization above.

Yes, but for all I know the neuromotor control system is a computer
program or spreadsheet that functions in the way you describe. Since I
am ignorant of the details of the "wet" system, I'm just as happy
sticking to behavior.

Bruce Gregory

"Everything that needs to be said has already been said. But since no
one was listening, everything must be said again."
                                                                                Andre Gide

"What is hateful to you, do not to your fellow men. That is the entire
Law; all the rest is commentary."

                                                                                The Talmud

from [Marc Abrams (2004.01.12.1828)]

[From Bill Powers (2004.01.12.1107 MSAT)]

I am attaching the files needed for a Vensim simulation using the data

from

an experimental run. For Vensim PLE you have to set the model parameters

by

hand; to do automatic optimization, you need Vensim DSS. Some of you have
copies of this advanced version, donated by Bob Eberlein, originator of
Vensim and CEO of Ventura Systems, Inc.

It's _Ventana_ Systems.

Marc

[From Bill Powers (2004.01.12.1930 MST)]

Marc Abrams (2004.01.12.1828)--

It's _Ventana_ Systems.

You are SO right! I must have been thinking of where Dag used to live. Thanks.

Best,

Bill P.

[From Rick Marken (2004.01.12.2000)]

Bruce Gregory (2004.01.12.1525)--

Since I am ignorant of the details of the "wet" system,
I'm just as happy sticking to behavior.

You're a behaviorist? I thought you were a fan of models. PCT is a
model of the organization of a nervous system that can produce the kind
of purposeful behavior we observe.

Best

Rick

···

---
Richard S. Marken
marken@mindreadings.com
Home 310 474-0313
Cell 310 729-1400

[From Bruce Gregory (2004.01.13.0707)]

[From Rick Marken (2004.01.12.2000)]

Bruce Gregory (2004.01.12.1525)--

Since I am ignorant of the details of the "wet" system,
I'm just as happy sticking to behavior.

You're a behaviorist?

No.

I thought you were a fan of models.

A fan of models? No, I'm a fan of evidence.

PCT is a
model of the organization of a nervous system that can produce the kind
of purposeful behavior we observe.

Fine. I'll leave the assessment of that model to neuroscientists. I do
not feel competent to evaluate models of the nervous system.

Bruce Gregory

"Everything that needs to be said has already been said. But since no
one was listening, everything must be said again."
                                                                                Andre Gide

"What is hateful to you, do not to your fellow men. That is the entire
Law; all the rest is commentary."

                                                                                The Talmud

[From Rick Marken (2004.01.13.0850)]

Bruce Gregory (2004.01.13.0707)--

Rick Marken (2004.01.12.2000)--

PCT is a
model of the organization of a nervous system that can produce the kind
of purposeful behavior we observe.

Fine. I'll leave the assessment of that model to neuroscientists. I do
not feel competent to evaluate models of the nervous system.

PCT is a model of the nervous system in the sense that it is a model of the
functional and structural organization of a nervous system that can produce
behavior. So you don't need to directly study the nervous system to test the
PCT any more than you have to directly study the structure of matter to test
atomic theory. The PCT model can be evaluated using behavioral tests.
That's what my research on PCT, which is described in _Mind Readings_, _More
Mind Readings_ and the soon to be released _Even More Mind Readings_
(kidding), has been about: testing a model of the nervous system by
observing the behavior produced by the actual nervous system. It's like
testing the atomic model of matter by observing the behavior of chemical
reactions and crystallographic refraction. I imagine that that's what you
mean when you say your focus is behavior. Mine is too. But while me focus is
on overt behavior, I assume that the components of the PCT model that lets
me account with great accuracy for this behavior correspond to components of
the nervous systems of the agents doing the behaving.

We know enough about the nervous system to see plausible correspondences
between nervous system components and PCT model components. As Bill noted,
these correspondences were pointed out in B:CP. It would be nice if we could
get some researchers to start looking at these correspondences at the
physiological level. For example, it would be nice to show, using electrode
stimulation studies, that descending efferents from the cerebellum function
as references rather than effectors. But until such work is done we'll just
have to go with the behavioral evidence for the model.

Best

Rick

···

--
Richard S. Marken
MindReadings.com
Home: 310 474 0313
Cell: 310 729 1400

[From Bruce Gregory (2004.01.13.1427)]

[From Rick Marken (2004.01.13.0850)]

Bruce Gregory (2004.01.13.0707)--

Rick Marken (2004.01.12.2000)--

PCT is a
model of the organization of a nervous system that can produce the
kind
of purposeful behavior we observe.

Fine. I'll leave the assessment of that model to neuroscientists. I do
not feel competent to evaluate models of the nervous system.

PCT is a model of the nervous system in the sense that it is a model
of the
functional and structural organization of a nervous system that can
produce
behavior. So you don't need to directly study the nervous system to
test the
PCT any more than you have to directly study the structure of matter
to test
atomic theory.

Well that's a different view than any I've heard before. I'm used to
models that are "about" whatever they are modeling. Models of galactic
evolution are tested by observing galaxies. Models of stellar evolution
are tested by observing stars. Apparently models of the nervous system
are not tested by observing the nervous system, even though there are
ways to observe the nervous system. I'm sure this is why it has taken
me so long to acquire even an imperfect understanding of PCT.

Bruce Gregory

"Everything that needs to be said has already been said. But since no
one was listening, everything must be said again."
                                                                                Andre Gide

"What is hateful to you, do not to your fellow men. That is the entire
Law; all the rest is commentary."

                                                                                The Talmud

[From Rick Marken (2004.01.13.1200)]

Bruce Gregory (2004.01.13.1427)]

Rick Marken (2004.01.13.0850)--

PCT is a model of the nervous system in the sense that it is a model
of the functional and structural organization of a nervous system
that can produce behavior. So you don't need to directly study system to
the nervous test the PCT any more than you have to directly study the
structure of matter to test atomic theory.

Well that's a different view than any I've heard before. I'm used to
models that are "about" whatever they are modeling. Models of galactic
evolution are tested by observing galaxies. Models of stellar evolution
are tested by observing stars.

Yes. But don't these models make assumptions about the underlying mechanisms
that would produce the observed phenomena? For example, doesn't the galactic
evolution model make assumptions about the nature of the space, forces and
matter that produce the observed phenomena of galactic evolution? If so,
they are models like PCT. But if they are curve fitting models then they are
not like PCT.

Apparently models of the nervous system
are not tested by observing the nervous system, even though there are
ways to observe the nervous system. I'm sure this is why it has taken
me so long to acquire even an imperfect understanding of PCT.

I think whatever problems you are having understanding PCT stem from trying
to understand it in terms of verbal phrasing rather than in terms of actual
working models. The phrase "PCT is a model of the nervous system" is too
ambiguous to be useful for understanding what the PCT model is "about". If
you look at an actual, working PCT model you will see functions that must be
carried out, in real organisms, by the nervous system. If you know something
about the nervous system, you can see that the transformation of physical
energy into a perception must be carried out by sensory receptors and
afferent neural networks, the perceptual signal must be carried by afferent
neural fibers, the reference signals must be carried by efferent fibers, the
output signals must be carried by efferent fibers that connect to muscles
and glands. So this is the implicit model of the nervous system that
produces purposeful behavior. It is not a model of how the nervous system
does the computations, carries the signals, and so on. Such a model would,
I think, more appropriately be called a model of the nervous system. But
PCT is a model of the functional and structural characteristics of a nervous
system that can produce purposeful behavior.

Best

Rick

···

--
Richard S. Marken
MindReadings.com
Home: 310 474 0313
Cell: 310 729 1400

[From Bruce Gregory (2004.01.13.1705)]

Rick Marken (2004.01.13.1200)

So this is the implicit model of the nervous system that
produces purposeful behavior. It is not a model of how the nervous
system
does the computations, carries the signals, and so on. Such a model
would,
I think, more appropriately be called a model of the nervous system.
But
PCT is a model of the functional and structural characteristics of a
nervous
system that can produce purposeful behavior.

Fine. We have no disagreement then.

Bruce Gregory

"Everything that needs to be said has already been said. But since no
one was listening, everything must be said again."
                                                                                Andre Gide

"What is hateful to you, do not to your fellow men. That is the entire
Law; all the rest is commentary."

                                                                                The Talmud