Powers, 2007: I didn't apply control theory to homeostasis in B:CP

[Matti]

[From MK (2015.07.31.2340 CET)]

···

----
Subject: Re: Intrinsic reference conditions

[From Bill Powers (2007.12.24.0910 MST)]

Martin Taylor 2007.12.23.23.43

I guess I wasn't clear. An "error" in perceptual control
theory is the difference between a reference value and a
perceptual value in a control unit. The reason there is no
"intrinsic error" is that there is no reference value for an
intrinsic variable. If there is nothing for the value of a
variable to be compared against, the concept of "error" does
not apply.

[My comment continued]

On awakening this morning I got out B:CP and looked through it,
with some trepidation, for the discussion of homeostasis as it
relates to intrinsic reference signals and error signals.

Sure enough, it isn't there. Neither "homeostasis" nor "Cannon"
appears in the index nor, as far as I can find, in the text.

I was so focused on the connection between intrinsic error signals
and reorganization that I simply passed over the homeostatic
systems in which the reference signals and error signals appear.
I'm sure I must have written many times about homeostasis (I know
I reported to CSGnet upon discovering Mrosovsky's "Rheostasis"),
but I can't find anything about it in B:CP, even though I was
quite aware of that subject at the time of writing and considered
it to show a level of biochemical (and autonomic, as others have
reminded me) control systems. I can see that if another edition of
B:CP ever appears, it is going to require an added chapter on this
subject, or a large revision of the chapter on learning and
reorganization. I tell you, discovering a blind spot that large is
very painful.

One painful aspect of it is remembering how, when Gary Cziko wrote
about Bernard and Cannon in Without Miracles, I wondered why he
didn't credit me with applying control theory to homeostasis. The
reason is now quite clear: I didn't. I only thought I had done so.

So: my somewhat perfunctory mention of the possibility of a lack
of clear communication on my part turns out to be a very likely
explanation for why you. Martin, and probably many others don't
realize that the intrinsic control systems of which I spoke were
the same homeostatic systems that Bernard and then Cannon
recognized, and that led Arturo Rosenbleuth, a student of
Cannon's, to bring this subject to Norbert Wiener's attention,
thus giving rise to cybernetics. My only addition was to propose
that large enough error signals (how I wish I had termed them
homeostatic error signals) cause reorganization of the behavioral
systems to begin. In my diagram of the relationship of the
reorganizing system to the behavioral hierarchy (Fig. 14.1) I show
ONLY the reorganizing effects of intrinsic error signals. The gap
left by omitting the local output functions that normally correct
intrinsic errors is now the most prominent feature of that diagram
in my mind. How could I not have seen what I was leaving out?

Dag Forssell, since it was you who drew the latest and clearest
version of Fig. 14.1, perhaps you could undertake to add those
missing output functions that convert intrinsic error signals into
physiological effects in that part of the diagram. But read on
first.

Writing this, I now realize that the "ignoration" of the
homoeostatic control systems was more than a simple omission. I
failed to see a principle that becomes obvious when the
homeostatic systems are added in all their glory as complete
control systems. When the physiological loops are added, we see
that reorganization is triggered by excessive and prolonged error
signals in somatic control systems -- just as it is triggered by
excessive neural error signals in the behavioral systems of the
brain. This quickly brings in another consideration that I have
looked at and mentioned, which is that "pain" in many cases (if
not all) is simply an ordinary perceptual signal that is excessive
in magnitude, meaning that it is causing very large error signals.
Any perception, when carried to an extreme magnitude, is painful -
- we try very hard to make it smaller. We can now say that any
error signal, whether in a biochemical, autonomic, or behavioral
control system, will, when large enough and protracted enough, be
experienced as pain and will cause reorganization to begin.

This tells us that the reorganizing system must be a distributed
system that brings reorganization to all levels of control systems
from bottom to top. At the level of DNA, it exists in the form of
repair enzymes. The immune system is a higher-order version of
repair enzymes. Reorganization exists at every level and acts
locally to that level. So we arrive at the question, "what about
amoebae?" And the answer, too.
Reorganization is simply an aspect of any level of biological
control systems.

And that brings up a realization delayed by some 35 years because
of that blind spot: every level of organization has ITS OWN
reorganizing system that senses excessive error and applies its
reorganizing actions to that level. So the diagram of Fig. 14.1 is
probably wrong. It is not error at the physiological level, but
only error at the behavioral level, that leads to reorganization
at the behavioral (neural, brain) level. Reorganization does
result from excessive error at the homeostatic level, but its
effects happen at that level. If we reorganize our behavior
because of physiological problems, we do so only because those
physiological problems are not corrected by reorganization at the
physiological level, and lead to excessive errors in the
behavioral systems. It is the latter kind of error that leads to
reorganization at the behavioral level. So now we see that every
new level has to deal with whatever errors the levels below it
can't handle, with reorganization happening just as control of any
kind happens: locally.

I don't know how well this revision will survive aging, but it's
pretty clear that it wouldn't have occurred to me if you, Martin,
hadn't made the inflammatory proposal that there are no intrinsic
reference signals.

Best,

Bill P.

----

M

[DagArchive]

[From MK (2015.07.31.2340 CET)]

---

Subject: Re: Intrinsic reference conditions

[From Bill Powers (2007.12.24.0910 MST)]

Martin Taylor 2007.12.23.23.43
[From Dag Forssell (20150731 20:30 PDT)]

I scanned this post and noted that Bill gave me an assignment. I overlooked it at the time, but got to it years later. See http://www.livingcontrolsystems.com/intro_papers/Reorg_evolution.pdf

Best, Dag

···

At 02:43 PM 7/31/2015, you wrote:

[MartinT]

[Martin Taylor 2015.08.01.00/36]

That's an interesting find. I would have preferred to argue with Bill the point he makes in this quoted passage and in the PDF Dag linked, as I did in the thread from which this passage was taken. But that's not possible. If you do want to rehash it now after 8 years, I'd be happy to get lightly involved -- I'm trying to put my efforts in several other places for the next little while. But if I do, it must be on the basis of what is scientifically probable or testable, not on a parsing of Bill's words as though they are divinely inspired.

The positions I would take are (1) homeostasis does not imply either control or the existence of reference levels, and (2) If intrinsic variables actually are controlled as shown in the "final" version of Fig 14.1, wouldn't one expect that control to be connected through the regular perceptual control hierarchy in the same way as "regular" perceptual variables rather than being linked with the "other" control system by a cumbersome reorganization process that results in the intrinsic variable control being affected by the side-effects of perceptual control? Wouldn't the side-effects of perceptual control as shown in the figure act as disturbances to the intrinsic control systems rather than as the means of keeping the intrinsic variables near the levels at which the physiological functions of the organism work well?

Martin

···

[From MK (2015.07.31.2340 CET)]

----
Subject: Re: Intrinsic reference conditions

[From Bill Powers (2007.12.24.0910 MST)]

Martin Taylor 2007.12.23.23.43

I guess I wasn't clear. An "error" in perceptual control
theory is the difference between a reference value and a
perceptual value in a control unit. The reason there is no
"intrinsic error" is that there is no reference value for an
intrinsic variable. If there is nothing for the value of a
variable to be compared against, the concept of "error" does
not apply.

[My comment continued]

On awakening this morning I got out B:CP and looked through it,
with some trepidation, for the discussion of homeostasis as it
relates to intrinsic reference signals and error signals.

Sure enough, it isn't there. Neither "homeostasis" nor "Cannon"
appears in the index nor, as far as I can find, in the text.

I was so focused on the connection between intrinsic error signals
and reorganization that I simply passed over the homeostatic
systems in which the reference signals and error signals appear.
I'm sure I must have written many times about homeostasis (I know
I reported to CSGnet upon discovering Mrosovsky's "Rheostasis"),
but I can't find anything about it in B:CP, even though I was
quite aware of that subject at the time of writing and considered
it to show a level of biochemical (and autonomic, as others have
reminded me) control systems. I can see that if another edition of
B:CP ever appears, it is going to require an added chapter on this
subject, or a large revision of the chapter on learning and
reorganization. I tell you, discovering a blind spot that large is
very painful.

One painful aspect of it is remembering how, when Gary Cziko wrote
about Bernard and Cannon in Without Miracles, I wondered why he
didn't credit me with applying control theory to homeostasis. The
reason is now quite clear: I didn't. I only thought I had done so.

So: my somewhat perfunctory mention of the possibility of a lack
of clear communication on my part turns out to be a very likely
explanation for why you. Martin, and probably many others don't
realize that the intrinsic control systems of which I spoke were
the same homeostatic systems that Bernard and then Cannon
recognized, and that led Arturo Rosenbleuth, a student of
Cannon's, to bring this subject to Norbert Wiener's attention,
thus giving rise to cybernetics. My only addition was to propose
that large enough error signals (how I wish I had termed them
homeostatic error signals) cause reorganization of the behavioral
systems to begin. In my diagram of the relationship of the
reorganizing system to the behavioral hierarchy (Fig. 14.1) I show
ONLY the reorganizing effects of intrinsic error signals. The gap
left by omitting the local output functions that normally correct
intrinsic errors is now the most prominent feature of that diagram
in my mind. How could I not have seen what I was leaving out?

Dag Forssell, since it was you who drew the latest and clearest
version of Fig. 14.1, perhaps you could undertake to add those
missing output functions that convert intrinsic error signals into
physiological effects in that part of the diagram. But read on
first.

Writing this, I now realize that the "ignoration" of the
homoeostatic control systems was more than a simple omission. I
failed to see a principle that becomes obvious when the
homeostatic systems are added in all their glory as complete
control systems. When the physiological loops are added, we see
that reorganization is triggered by excessive and prolonged error
signals in somatic control systems -- just as it is triggered by
excessive neural error signals in the behavioral systems of the
brain. This quickly brings in another consideration that I have
looked at and mentioned, which is that "pain" in many cases (if
not all) is simply an ordinary perceptual signal that is excessive
in magnitude, meaning that it is causing very large error signals.
Any perception, when carried to an extreme magnitude, is painful -
- we try very hard to make it smaller. We can now say that any
error signal, whether in a biochemical, autonomic, or behavioral
control system, will, when large enough and protracted enough, be
experienced as pain and will cause reorganization to begin.

This tells us that the reorganizing system must be a distributed
system that brings reorganization to all levels of control systems
from bottom to top. At the level of DNA, it exists in the form of
repair enzymes. The immune system is a higher-order version of
repair enzymes. Reorganization exists at every level and acts
locally to that level. So we arrive at the question, "what about
amoebae?" And the answer, too.
Reorganization is simply an aspect of any level of biological
control systems.

And that brings up a realization delayed by some 35 years because
of that blind spot: every level of organization has ITS OWN
reorganizing system that senses excessive error and applies its
reorganizing actions to that level. So the diagram of Fig. 14.1 is
probably wrong. It is not error at the physiological level, but
only error at the behavioral level, that leads to reorganization
at the behavioral (neural, brain) level. Reorganization does
result from excessive error at the homeostatic level, but its
effects happen at that level. If we reorganize our behavior
because of physiological problems, we do so only because those
physiological problems are not corrected by reorganization at the
physiological level, and lead to excessive errors in the
behavioral systems. It is the latter kind of error that leads to
reorganization at the behavioral level. So now we see that every
new level has to deal with whatever errors the levels below it
can't handle, with reorganization happening just as control of any
kind happens: locally.

I don't know how well this revision will survive aging, but it's
pretty clear that it wouldn't have occurred to me if you, Martin,
hadn't made the inflammatory proposal that there are no intrinsic
reference signals.

Best,

Bill P.

----

M

[Matti]

[From MK (2015.08.01.2300 CET)]

Martin Taylor (2015.08.01.00/36)--

If you do want to rehash it now after 8 years

I actually don't as it was Powers' realization that he hadn't
mentioned either homeostasis or Cannon in B:CP that I found
interesting about that particular post. That is a serious omission,
and glaringly obvious once it has been pointed out. What else should
have been discussed in the book but wasn't?

Apparently I haven't yet reached the level of understanding required
to really care about the points raised by you and Boris.

M

[rsmarken]

[From Rick Marken (2015.08.01.1400)]

···

Dag Forssell (20150731 20:30 PDT)

I scanned this post and noted that Bill gave me an assignment. I
overlooked it at the time, but got to it years later. See

http://www.livingcontrolsystems.com/intro_papers/Reorg_evolution.pdf

RM: The revised Figure is really nice. It’s the one Tim and I used in our paper “Understanding the Change Process Involved in Solving Psychological Problems: A Model-based Approach to Understanding How Psychotherapy Works” (reprint available at https://www.dropbox.com/s/tc1wyrfdk375iee/SolvingPsychProbs.pdf?dl=0). I would have just added the following:

–

RM: I added the effect of average error in the control hierarchy on intrinsic state; (the bracket represents a sensor that integrates all the error in the control hierarchy into an “Average Error in Hierarchy” that has a direct effect on intrinsic quantities, like Adrenalin level). Without this connection the diagram implies that the only time reorganization works on the hierarchy is when actions on the outside environment drive intrinsic variables from their references. I think psychotherapists (certainly those of an MOL persuasion) agree that chronic error in the control hierarchy is itself a cause of intrinsic error (which we experience as anxiety or stress or depression – ie. emotion). In other words, there is a direct effect of error on the hierarchy on intrinsic error, even if there is no environmental effect of error that results in physiological processes that create intrinsic error.

RM: For example, without the direct connection I show between error in the hierarchy and intrinsic error, someone with chronic error in their hierarchy of control would not necessarily start reorganizing. For example, suppose you had chronic error in a relationship control system, the one that controls the relationship between you and your spouse. If there were no direct effect of this error on intrinsic state then reorganization would kick in only if this poor control had effects on intrinsic error via the environment – for example, your spouse throws a pan at your hear, creating considerable intrinsic error. But I think it’s more realistic to recognize that chronic error in the hierarchy like this can have direct, physiological effects on intrinsic variables, even if such error has no effect on intrinsic error via the environment (no pans thrown). In other words, error in the hierarchy of control has direct physiological effects on intrinsic state (which we experience as emotions) and is thus a cause of intrinsic error resulting in reorganization.

RM: By the way, the control loop shown in the diagram – the one that controls the intrinsic quantities (such as blood glucose level and core temperature) that make up the intrinsic state represents the many physiological control systems that are the “homeostatic” systems that control these variables. Actually, the diagram shows that homeostasis – the control of intrinsic quantities – is carried out by physiological mechanisms as well as by the action of the control hierarchy. For example, blood glucose level is maintained by physiological control processes (involving the secretion of insulin, for example) as well as by the behaviors produced by the control hierarchy (such as stuffing food into your mouth).

Best regards

Rick

Richard S. Marken

www.mindreadings.com
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

[rsmarken]

[From Rick Marken (2015.08.01.1725)]

···

MK (2015.08.01.2300 CET)–

MT: If you do want to rehash it now after 8 years

MK: I actually don’t as it was Powers’ realization that he hadn’t

mentioned either homeostasis or Cannon in B:CP that I found

interesting about that particular post. That is a serious omission,

and glaringly obvious once it has been pointed out. What else should

have been discussed in the book but wasn’t?

RM: I think the omission is far less serious than you do. It might have been nice to mention homeostasis in a brief paragraph in the chapter in B:CP where Bill describes the reorganization system (the Learning chapter). But I think leaving out mention of homeostasis in B:CP may actually have been intentional (though unconscious).

RM: Homeostasis is about the controlling done by physiological systems – systems that control physiological variables by physiological means. B:CP is about the fact that behavior itself – everything we do – IS control – the control of variable aspects of the environment by physical means. This was the revolutionary idea presented (and explained) in B:CP.

RM: B:CP was aimed at psychologists who understood behavior (and cognition) in cause-effect terms. These psychologists had no problem accepting the idea that the body contained many physiological control systems – homeostatic systems. What they didn’t know – and what B:CP explained to them – was that behavior itself was a process of control.

RM: So I think it was best for B:CP to ignore homeostasis since it would have been a diversion; since psychologists already accepted the existence of physiological control – homeostasis – and going on about it in B:CP might have just encouraged more of the “nothing but” response to B:CP (which Bill got plenty of anyway). Psychologists might have concluded that B:CP was about nothing more than the fact that the body is a homeostatic system. They might not have noticed that B:CP was about the fact that behavior itself is a homeostatic process. Homeostasis describes what is going on when blood glucose level is maintained at ~80 mg/d but it also describe what is going on when a car is maintained at a constant distance from the car in front. But talking about “homeostasis” might have obscured the second meaning – which is the one B:CP was about.

RM: So from my point of view there was very little lost (an interesting historical tidbit, perhaps) and quite a bit gained (emphasis on the fact that behavior IS control) by leaving a discussion of homeostasis out of B:CP.

Best

Rick

–
Richard S. Marken

www.mindreadings.com
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

[MartinT]

[Martin Taylor 2015.08.01.23.09]

In other words, you are asserting that the intrinsic variables are

not disturbed by influences from the environment? And that ALL the
control system errors are combined into just one variable that acts
as a second source of disturbances to the intrinsic state control
systems? Both assertions seem rather dubious to me.
Although I object to treating Bill’s words a holy writ, I would
point out that many times he expressed interest in investigating the
modularity and (my words) fractal nature of reorganization. Also
that he treated control error as though it itself was an intrinsic
variable rather than as a disturbance to some nebulous “intrinsic
state”. In such demonstrations as the Arm2 14 degree-of-freedom
reorganization, control error is the ONLY intrinsic variable, if I’m
not mistaken.
Both of those positions seem rather more reasonable than what you
seem to be proposing in this diagram revision.
They would if you accepted Bill’s concept of local or regional error
in the control hierarchy as an intrinsic variable that influenced
the reorganization rate in that part of the hierarchy. Anyway, is it
not one of the assumptions of e-coli reorganization that the
reorganization rate never actually goes to zero?
That’s a useful clarification, and it also points to the probable
modularization of the reorganization processes. (Incidentally, I
assume that many of the homeostatic processes involved are indeed
control systems; but many have much longer loops and complex
networks of influences that I would not call control systems but
that serve to stabilize the variables.)
What you haven’t explained (and neither did Bill) is why the final
revision of Figure 14.1 that created a control loop internal to the
intrinsic system that would be disturbed by the side-effects of
perceptual control would work better than the earlier version in
which the intrinsic variable control worked by means of using the
side-effect of perceptual control as its action output. But maybe I misunderstand what the revised revised diagram is
supposed to show.
It really doesn’t matter much what Bill said or thought, other than
the fact that a high proportion of what he said turned out either to
seem correct or to seem to lead toward a correct appreciation. But
he did have his blind spots where he was definitely and demonstrably
wrong, as do we all. Bill was just a genius, not a god, and every
genius makes mistakes. What matters is what Nature tells us when we
poke her. (My personal meta-theory of science, is that EVERY theory
– including this one – will eventually be shown to be wrong).
Martin

···

[From Rick Marken (2015.08.01.1400)]

Dag Forssell (20150731 20:30 PDT)

            I scanned this post and noted that Bill gave me an

assignment. I
overlooked it at the time, but got to it years later.
See

http://www.livingcontrolsystems.com/intro_papers/Reorg_evolution.pdf

          RM: The revised Figure is really nice. It's the one

Tim and I used in our paper “Understanding the Change
Process Involved in Solving Psychological Problems: A
Model-based Approach to Understanding How Psychotherapy
Works” (reprint available at https://www.dropbox.com/s/tc1wyrfdk375iee/SolvingPsychProbs.pdf?dl=0 ).
I would have just added the following:

        RM: I added the effect of average

error in the control hierarchy on intrinsic state; (the
bracket represents a sensor that integrates all the error in
the control hierarchy into an “Average Error in Hierarchy”
that has a direct effect on intrinsic quantities, like
Adrenalin level). Without this connection the diagram
implies that the only time reorganization works on the
hierarchy is when actions on the outside environment drive
intrinsic variables from their references.

        RM: For example, without the direct

connection I show between error in the hierarchy and
intrinsic error, someone with chronic error in their
hierarchy of control would not necessarily start
reorganizing.

        RM: By the way, the control loop

shown in the diagram – the one that controls the intrinsic
quantities (such as blood glucose level and core
temperature) that make up the intrinsic state represents the
many physiological control systems that are the
“homeostatic” systems that control these variables.
Actually, the diagram shows that homeostasis – the control
of intrinsic quantities – is carried out by physiological
mechanisms as well as by the action of the control
hierarchy. For example, blood glucose level is maintained by
physiological control processes (involving the secretion of
insulin, for example) as well as by the behaviors produced
by the control hierarchy (such as stuffing food into your
mouth).

[rsmarken]

[From Rick Marken (2015.08.02.0930)]

···

Martin Taylor (2015.08.01.23.0)–

MT: In other words, you are asserting that the intrinsic variables are

not disturbed by influences from the environment?

RM: Not at all. The diagram shows that intrinsic variables are disturbed by influences from the environment as well by the outputs of the physiological control system. For example, blood sugar level is influenced both by the insulin produced by the physiological (homeostatic) blood sugar level control system as well as by the environmental inputs (food) resulting from the outputs of the behavioral hunger control system.

MT: And that ALL the

control system errors are combined into just one variable that acts
as a second source of disturbances to the intrinsic state control
systems? Both assertions seem rather dubious to me.

RM: I agree that overall error in the hierarchy may not be the error that contributes to intrinsic error. That was just a simplification. All I wanted to include is the fact that error in the control hierarchy is itself likely to be a source of intrinsic error and this was not shown in the original version of Figure 14.1.

MT: Although I object to treating Bill's words a holy writ, I would

point out that many times he expressed interest in investigating the
modularity and (my words) fractal nature of reorganization. Also
that he treated control error as though it itself was an intrinsic
variable rather than as a disturbance to some nebulous “intrinsic
state”. In such demonstrations as the Arm2 14 degree-of-freedom
reorganization, control error is the ONLY intrinsic variable, if I’m
not mistaken.

RM: Yes, I think there is no question that reorganization can work on the specific places where there are control problems. But how it gets directed to those places and what it does when it gets there is not really well understood. The people who might be best at investigating this are those involved in therapy. David Goldstein just posted some ideas about how consciousness might be involved in this process. What we need is fewer “expressions of interest” and more actual investigation of what you call the “modularity” of reorganization.

MT:What you haven't explained (and neither did Bill) is why the final

revision of Figure 14.1 that created a control loop internal to the
intrinsic system that would be disturbed by the side-effects of
perceptual control would work better than the earlier version in
which the intrinsic variable control worked by means of using the
side-effect of perceptual control as its action output.

RM: It’s not so much a matter of “working better” as of properly representing what we know of how intrinsic variables, like blood glucose level, are controlled. In the original 14.1 it looked like these variables were controlled only via the control hierarchy. That is, it looked like the only way an intrinsic variable like blood sugar level was controlled was by ones overt behavior, like eating, that is produced by the control hierarchy. But we know that there are physiological processes involved in controlling these variables as well – the “homeostatic”, physiological control systems that are completely separate from our control hierarchy. It’s these homeostatic control processes that are now added to 14.1, not in order to make the model “work better” but to make it more consistent with what we know about how the body controls what Bill called “intrinsic variables” in B:CP.

Best

Rick

–
Richard S. Marken

www.mindreadings.com
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

[MartinT]

[Martin Taylor 2015.08.02.15.18]

I have no problem with most of your words in this message, but I

don’t think the diagram would lead the viewer to understand the
system as your words describe it. It certainly did not lead me to
imagine what you talk about. Quite the contrary, indeed. I think a
new figure or series of figures might be more effective than further
revisions of Figure 14.1.
Why not simply use Bill’s approach of taking perceptual control
error averaged over space and time to BE an intrinsic variable, or
rather a whole set of intrinsic variables? What was wrong with that?
I have argued in the past, and would still maintain, that it is
almost a requirement for any kind of perceptual control theory that
the quality of control should be a criterion for reorganization, and
that it cannot be the only criterion (i.e. it must be an intrinsic
variable in this model of the reorganization process, but it cannot
be the only intrinsic variable, as in the reorganization
demonstrations).
I’m fine with all that. So long as homeostatic processes do their
job, the physiological intrinsic variables won’t contribute much
error to increase the reorganization rate. One way of looking at the
entire system is that perceptual control exists only in order for
the side-effects of perceptual control to influence the intrinsic
variables. As an aside, it is interesting that these side-effects of perceptual
control, which would be disturbances to homeostatic systems
stabilized by a standard PCT-type controller, need not be
disturbances to more complex homeostatic structures that have inputs
other than the canonical “reference” and “sensory input”.
As I said, I have no problem with the intention. I have a problem
with the execution, in the form of a Figure which can mislead – at
least it misled me about what you (and possibly Bill) were getting
at.
Martin

···

[From Rick Marken (2015.08.02.0930)]

            Martin Taylor

(2015.08.01.23.0)–

            MT: In other words,

you are asserting that the intrinsic variables are not
disturbed by influences from the environment?

          RM: Not at all. The diagram shows that intrinsic

variables are disturbed by influences from the environment
as well by the outputs of the physiological control
system. For example, blood sugar level is influenced both
by the insulin produced by the physiological (homeostatic)
blood sugar level control system as well as by the
environmental inputs (food) resulting from the outputs of
the behavioral hunger control system.

            MT: And that ALL the

control system errors are combined into just one
variable that acts as a second source of disturbances to
the intrinsic state control systems? Both assertions
seem rather dubious to me.

          RM: I agree that overall error in the hierarchy may not

be the error that contributes to intrinsic error. That was
just a simplification. All I wanted to include is the fact
that error in the control hierarchy is itself likely to be
a source of intrinsic error and this was not shown in the
original version of Figure 14.1.

…

          RM: It's not so much a matter of "working better" as of

properly representing what we know of how intrinsic
variables, like blood glucose level, are controlled. In
the original 14.1 it looked like these variables were
controlled only via the control hierarchy. That is, it
looked like the only way an intrinsic variable like blood
sugar level was controlled was by ones overt behavior,
like eating, that is produced by the control hierarchy.
But we know that there are physiological processes
involved in controlling these variables as well – the
“homeostatic”, physiological control systems that are
completely separate from our control hierarchy.

          It's these homeostatic control processes that are now

added to 14.1, not in order to make the model “work
better” but to make it more consistent with what we know
about how the body controls what Bill called “intrinsic
variables” in B:CP.

[rsmarken]

[From Rick Marken (2015.08.03.0800)]

···

Martin Taylor (2015.08.02.15.18)

MT: I have no problem with most of your words in this message, but I

don’t think the diagram would lead the viewer to understand the
system as your words describe it. It certainly did not lead me to
imagine what you talk about. Quite the contrary, indeed. I think a
new figure or series of figures might be more effective than further
revisions of Figure 14.1.

RM: Great idea. Why don’t you take a first crack at it.

MT: Why not simply use Bill's approach of taking perceptual control

error averaged over space and time to BE an intrinsic variable, or
rather a whole set of intrinsic variables?

RM: Why not, indeed! It’s certainly the easiest way do it in a computer model. It’s the approach I’ve used as well when I have incorporated reorganization algorithms into my models. I just had an arrow pointing from the control hierarchy to the “Intrinsic State” bubble in Figure 14.1 to show that error in the control hierarchy could be considered a controlled intrinsic quantity. In the published version of Figure 14.1 (in B:CP 2nd Ed) there is no indication at all that error in the control hierarchy could be a controlled intrinsic quantity. My revision of 14.1 (shown again below) adding the “Average Error in Hierarchy” line connected to “Intrinsic State” was meant only to show that chronic error in the control hierarchy could be considered a controlled intrinsic quantity. If you have a better way of represetning that please feel free to post your diagram. I’d love to see it.

MT: One way of looking at the

entire system is that perceptual control exists only in order for
the side-effects of perceptual control to influence the intrinsic
variables.

RM: Yes, that’s a nice way of putting it.

Best

Rick

–
Richard S. Marken

www.mindreadings.com
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

            MT: In other words,

you are asserting that the intrinsic variables are not
disturbed by influences from the environment?

          RM: Not at all. The diagram shows that intrinsic

variables are disturbed by influences from the environment
as well by the outputs of the physiological control
system. For example, blood sugar level is influenced both
by the insulin produced by the physiological (homeostatic)
blood sugar level control system as well as by the
environmental inputs (food) resulting from the outputs of
the behavioral hunger control system.

          RM: I agree that overall error in the hierarchy may not

be the error that contributes to intrinsic error. That was
just a simplification. All I wanted to include is the fact
that error in the control hierarchy is itself likely to be
a source of intrinsic error and this was not shown in the
original version of Figure 14.1.

[rupert]

[From Rupert Young (2015.08.05 13.00)]

(Rick Marken (2015.08.01.1725)]

RM: Homeostasis is about the controlling done by physiological systems -- systems that control physiological variables by physiological means. B:CP is about the fact that behavior itself -- everything we do -- IS control -- the control of variable aspects of the environment by physical means. This was the revolutionary idea presented (and explained) in B:CP.

I think I'm coming down with a bad case of deja vu, but could you clarify? Are you saying that these two statements are equivalent:

"Behaviour is the control of perception"

"Behaviour is the control of variable aspects of the environment"

Rupert

[rsmarken]

[From Rick Marken (2015.08.05.0900)]

···

RM: Homeostasis is about the controlling done by physiological systems – systems that control physiological variables by physiological means. B:CP is about the fact that behavior itself – everything we do – IS control – the control of variable aspects of the environment by physical means. This was the revolutionary idea presented (and explained) in B:CP.
Rupert Young (2015.08.05 13.00)

RY: I think I’m coming down with a bad case of deja vu, but could you clarify? Are you saying that these two statements are equivalent:

“Behaviour is the control of perception”

“Behaviour is the control of variable aspects of the environment”

RM: Yes. There may be better ways to say it. But the best way is by actually implementing perception in a control model (a computer model or, even better, a robot model, like yours), in which case the perception is some function of physical variables.

Best

Rick

–

Richard S. Marken

www.mindreadings.com
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

[rupert]

[From Rupert Young (2015.08.05 20.00)]

(Rick Marken (2015.08.05.0900)]

Yes, and if we look at the live block demo from LCSIII I would have

thought that perception refers to the perceptual signal box and
variable aspects of the environment is referred to by the input
quantity; according to the system-environment boundary.

Initially both will have the same value, even if you change the

reference to 10, say.

However, if you set the input gain to something other than 1 then

the input quantity will change, but the only the perception will
remain equal to the reference. So, even though the perception is a
function of physical variables it is only the perception that is the
controlled variable, so these two statements are not equivalent.

Is this not the case? Or am I missing something?

Rupert

···

          Rupert

Young (2015.08.05 13.00)

                        RY: I think I'm coming down with a bad case of deja

vu, but could you clarify? Are you saying that these two
statements are equivalent:

          "Behaviour is the control of perception"



          "Behaviour is the control of variable aspects of the

environment"

          RM:  Yes. There may be better ways to say it. But the

best way is by actually implementing perception in a
control model (a computer model or, even better, a robot
model, like yours), in which case the perception is some
function of physical variables.

[_Erling_Jorgensen]

[From Erling Jorgensen (2015.08.05 1420 EDT)]

Rupert Young (2015.08.05 20.00)

Hi Rupert,

I think you’ve nailed the distinction perfectly, that “control of perception” is not the same as “control of variable aspects of the environment.” You anchor it in the LiveBlock simulation from LCSIII very well:

Yes, and if we look at the live block demo from LCSIII I would have thought that perception refers to the perceptual >signal box and variable aspects of the environment is referred to by the input quantity; according to the system->environment boundary.
Initially both will have the same value, even if you change the reference to 10, say.

However, if you set the input gain to something other than 1 then the input quantity will change, but only the >perception will remain equal to the reference. So, even though the perception is a function of physical variables it is >only the perception that is the controlled variable, so these two statements are not equivalent.

As you note, when the input function in the organism makes the perception something different than the input quantity itself, only the perception accurately tracks the value of the reference, which is the definition of control.

I get confused whether Rick M. really believes something in the environment is controlled, or whether he only means functions of things in the environment. He keeps using the word “aspects,” and I think that is where the ambiguity comes in, because “aspect” gives too much of the implication that it is “out there.” It would help me enormously if this replacement phrase of his were changed to: “Behavior is the control of variable functions of the environment.” Because, in my mind, that places the emphasis on: “Oh?? Which functions?” – (at which point we’re back inside the organism again.)

Thank you for your efforts at keeping these notions distinct.

All the best,

Erling

[<img src="cid:AOSZDIOAUXFH.537b60bf.jpg" border="0">
](https://urldefense.proofpoint.com/v2/url?u=http-3A__www.riverbendcmhc.org&d=AwMFAg&c=8hUWFZcy2Z-Za5rBPlktOQ&r=-dJBNItYEMOLt6aj_KjGi2LMO_Q8QB-ZzxIZIF8DGyQ&m=RBO4bd_AfaSvJaDLPkmHy39of2SzgUzyAy81DUwZ38Q&s=kmulrEotR0EfVhiUflZPyorrEhHNZ5azngNQiL2pJOM&e=)
  NOTICE: This e-mail communication (including any attachments) is CONFIDENTIAL and the materials contained herein are PRIVILEGED and intended only for disclosure to or use by the person(s) listed above. If you are neither the intended recipient(s), nor a person responsible for the delivery of this communication to the intended recipient(s), you are hereby notified that any retention, dissemination, distribution or copying of this communication is strictly prohibited. If you have received this communication in error, please notify me immediately by using the "reply" feature or by calling me at the number listed above, and then immediately delete this message and all attachments from your computer. Thank you.

[_Erling_Jorgensen]

[From Erling Jorgensen (2015.08.05 1420 EDT)]
I didn't see my initial submission of this show up at CSGNet, so I'm resending
it.

Rupert Young (2015.08.05 20.00)

Hi Rupert,
I think you've nailed the distinction perfectly, that "control of perception"
is not the same as "control of variable aspects of the environment." You
anchor it in the LiveBlock simulation from LCSIII very well:

Yes, and if we look at the live block demo from LCSIII I would have thought
that perception refers to the perceptual signal box and variable aspects of

the >environment is referred to by the input quantity; according to the
system- >environment boundary.

Initially both will have the same value, even if you change the reference to
10, say.
However, if you set the input gain to something other than 1 then the input
quantity will change, but only the perception will remain equal to the
reference. So, even though the perception is a function of physical variables

it >is only the perception that is the controlled variable, so these two
statements >are not equivalent.

As you note, when the input function in the organism makes the perception
something different than the input quantity itself, only the perception
accurately tracks the value of the reference, which is the definition of
control.

I get confused whether Rick M. really believes something in the environment is
controlled, or whether he only means functions of things in the environment.
He keeps using the word "aspects," and I think that is where the ambiguity
comes in, because "aspect" gives too much of the implication that it is "out
there." It would help me enormously if this replacement phrase of his were
changed to: "Behavior is the control of variable functions of the
environment." Because, in my mind, that places the emphasis on: "Oh?? Which
functions?" -- (at which point we're back inside the organism again.)

Thank you for your efforts at keeping these notions distinct.

All the best,
Erling

[rsmarken]

[From Rick Marken (2015.08.05.1500)]

···

Rupert Young (2015.08.05 20.00)

RY: Yes, and if we look at the live block demo from LCSIII I would have

thought that perception refers to the perceptual signal box and
variable aspects of the environment is referred to by the input
quantity; according to the system-environment boundary.

RY: Initially both will have the same value, even if you change the

reference to 10, say.

RY: However, if you set the input gain to something other than 1 then

the input quantity will change, but the only the perception will
remain equal to the reference. So, even though the perception is a
function of physical variables it is only the perception that is the
controlled variable, so these two statements are not equivalent.

RY: Is this not the case?

RM: It is the case that what a control system controls is the perceptual signal, which I presume, based on your analysis of the LiveBlock demo, is what you are referring to as the perception. But when the system is controlling the perceptual signal, keeping it matching the reference signal, it is also controlling the aspect of the environment to which that signal corresponds, which in the LiveBlock demo is the input quantity times the gain (q.i*g).

RM So in this very simple LiveBlock set up, q.i is the only variable in the system’s environment and the value of g defines the aspect of that environment that corresponds to the perception, p. When p is controlled (made to equal r) the aspect of the environment that is controlled depends on the value of g. If g = 1 then q.i1 = q.i is controlled. If g = .5 then q.i.5 is controlled.

RM: Since p = q.i * g, when the perception is brought into a match with r, the value of q.i will depend on g. If r is 10 and g is 1 then when p = r, q.i will = 10. If r is again 10 and g is .5 then when p = r, q.i will be 20. The perceptual signal is the same in both cases because the control system makes p = r. But the value of q.i that results in those perceptions being the same is different because a different aspect of q.i is being perceived (and controlled) in each case.

RY: Or am I missing something?

RM: I think so but it’s hard to say what it is. For some reason you (and several others) seem to think that control of perception is in some way separate from control of aspects of the environment. Actually, I don’t really understand what your point is.

RM: To me, “perception” refers to some kind of representation inside of a device of a state of affairs outside (in the environment of) the device. In PCT a perception is a perceptual signal – an afferent neural current – that varies as a function of variations (ultimately) in physical variables: p = f(x.1,x.2…x.n), where x.i are different physical variables.

RM: What you may be “missing” is the fact that all perceptual signals are the same – they are all just variations in the rate of afferent neural firing. The only difference between one perceptual signal and another is in how it’s derived from lower level perceptual inputs and, ultimately, sensed effects of physical variables. The “derivation” is the perceptual function that defines whether variations in a perceptual signal correspond to variations in the size of a book, the shape of a letter, the relationship between the book and a table top, etc. That is, the only way to discriminate between the different perceptual signals we control is in terms of the (variable) aspect of the environment to which that perception corresponds – or, better, the aspect of the environment from which that perceptual signal is derived…

RM: This applies to both human and artefactual control systems (the latter like your robots). In a robot controller, for example, the perceptual signals are all electrical currents. What these perceptions correspond to depends on how they are derived ultimately from measures of physical variables.

RM: Does this help?

          RY: I think I'm coming down with a bad case of deja

vu, but could you clarify? Are you saying that these two
statements are equivalent:

          "Behaviour is the control of perception"



          "Behaviour is the control of variable aspects of the

environment"

RM: Yes.

Best

Rick

Richard S. Marken

www.mindreadings.com
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

[_Boris_Hartman]

Hi Warren,

Nice message.

···

From: Warren Mansell (wmansell@gmail.com via csgnet Mailing List) [mailto:csgnet@lists.illinois.edu]
Sent: Thursday, August 06, 2015 10:09 PM
To: rsmarken@gmail.com
Cc: csgnet@lists.illinois.edu
Subject: Re: Powers, 2007: I didn’t apply control theory to homeostasis in B:CP

Maybe the only reason experimenters can test PCT is that they tend to derive similar perceptual variables from the environment as the participants. This helps the experimenter set up hypotheses for what the CV might be. The experimenter is not necessarily measuring the environment directly.

HB : I I understand you right, you are talking that everyone ha sit’s own perceptual world and on that bases everyone is able to conclude or at least try to conclude what other people are controlling or thinking. It’s of course difficult to »see« into others head but beside introspection there are interviews as scientific method, qualitative researches, etc. So there are quite number of scinetific methods with which we can try to conclude what people really control inside or think. As this is what is important. People are very different. Some are very oppened and we can make fast conclussions about his personality. Some are very closed and it’s hard to get a single clue what he is thinking about. And one day he »explode« and kill half of the school. It’s difficult to predict amd research what people really think. They can lie, manipulate, and so on.

WM : I really like Rick’s explanation of what aspects of environment contribute to a perception, but I really like

HB : Sorry to disapoint you Warren but I can’t any argument or support for Rick in later books, shall we say from 1998. I can’t find »controlled aspect of environment« in his generic diagrams nor in his text. I can’t find any citation that could support what Rick is talking about. As I see it Bill’s literature is talking something else. There was quite the same situation with Martin and Rick, when Riick was promoting »Perception : control of behavior« or some »stimulus-respons theory«. I found imidiatelly a dozen of thougts from Bill’s literature that were supporting Martin view : »Behavior : Control of perception«. But I couldn’t find any argumetn in Bill’s book that could support Rick’s view. So maybe we can repeat the whole procedure.

RM : Rupert’s insistence that the aspects of the environment that are affected by behaviour as part of the closed loop are not controlled, but manipulated.

HB : I think there is no doubt that Rzpert id right. I can find thousands of citations from Bill’s literature which are suporting Ruperts view, including all Bill generic diagrams, which show »Control of perception«.

If you want, we can make a competition. Let us try to read B:CP, 2005 and LCS III and compare for whom we will find more Bill’s arguments, including diagrams and so on. I invite all the members to help in this judgement. This is the only way I see we can once for all solve the problem, whether Bill talked about »Control of perception« or »Control of behavior« or »controlled aspect of environment«.

I’ll start with »feed-back« function. Feed-back function is the only thing I see in Bill’s diagram in outer environment and it is as Rupert discribe it. It says :

FFEDBACK FUNCTION (LCS III, p. 28) : Physical properties that convert action or behavior into effect on input quantity.

That’s all what Bill is saying about events in external environment. I’m sorry Warren but I can’t find any comparator, reference or »error« signal that Bill would indicate that there is control process going on in environment and that Rick could be right. He is obviously wrong and he is using some RCT diagram with »controlled aspect of environment«, which is in Bill’s diagram non existant. It seems that he is imagining things that are not present.

And I hope Warren we agree that Bill is reference for who is talking PCT and who is not.

We see that only thing Bill told about anything happening in environment of organism is »feed-back« function or effect that output has on input. He is simply talking about effect, not controlled effect or »control of behavior«, simply effect. And I hope we agree that this is »GENERAL DIAGRAM«, not some »SPECIAL«. It’s modeling every behavior, not just some specific ones.

So he is not talking about »controlled aspect of environment« or »Control of behavior« or else. Just effect of output on input. That’s it. We don’t need to make »Control behavior« disertation.

And beside that he is also saying : »Notice that we clarify the controlled variable as an input variable, not an output variable« (LCS III, p. 32).

If »output variable« is not controlled, what is making then »controlled aspect of environment« ? What is controlling outside in environment ???

WM : They seem to disagree with one another, but I just see enormous value from both perspectives.

HB : Yes they seem to disagree, but the only arbiter here I see is Bill’s work. We are here to promote his work, to upgrade, to make interperetations and so on. Everything is suppose to be about Bill and his PCT theory. Not some personal propmotions in the sense RCT (Rick’s control theory) and his perosonal diagrams. Everything has to be argumented with Bill’s work or any other evidence from the final arbiter : nature.

Best

Boris

Warren

On 6 Aug 2015, at 20:54, Richard Marken (rsmarken@gmail.com via csgnet Mailing List) <csgnet@li sts.illinois.edu> wrote:

[From Rick Marken (2015.08.06.1255)]

Rupert Young (2015.08.06 20.00)–

RM: When you make lemonade – control for the perception of the taste of lemonade – you are manipulating variables in the environment – the amount of lemon juice (x.1), water (x.2) and sugar (x.3) that you put in the pitcher – in such a way that they can be constructed into a perception of the taste of lemonade. So you are controlling an aspect of the environment – the relative proportion of different chemicals in the pitcher – when you control for the perception of “the taste of lemonade”. This is why I say that when you control a perception you are also controlling an aspect of the environment – the aspect of the environment from which that perception is constructed.

RY: Yes, you manipulate variables in the environment, but I don’t think that is the same as controlling an aspect of the environment. To control the perception of the sweetness of your lemonade you vary the amount of sugar until the desired sweetness is realised.

RM: Well I’m clearly not going to bring you to my point of view. And you are clearly not going to bring me to yours. The main reason is that I couldn’t do my work in PCT if I adopted your point of view. None of the research described in Mind Readings, More Mind Readings_ and Doing Research on Purpose could (or would) have been done if it were true that people are not controlling an aspect of the environment when they control a perception. I couldn’t have done any testing for controlled variables – or build models of people controlling those variables – if people were not controlling aspects of the environment when they were controlling perceptions.

RM: The whole point of The Test for Controlled Variables(TCV) is to figure out what aspect of the environment a person is controlling. If, as you and others say, no aspects of the environment are being controlled when people ar e controlling their perceptions then it would be impossible to scientifically study control of perceptions. The perceptions people control would be completely private and there could be no science of living control systems because you could never figure out what they are controlling. But we are able to determine what perceptions people are controlling using the TCV because when people are controlling perceptions that are controlling aspects of the the environment occupied by the person doing the TCV.

RM: If you interpret PCT to mean that all that is controlled are perceptual variables that don’t correspond to variable aspects of the environment then you are taking PCT to be theory that denies the possibility of being tested. Not very scientific. So this interpretation of PCT seems to me to be not only patently ridiculous but flatly contradicted by all the PCT research I’ve done and all the demonstrations of PCT principles I have developed.

RM: So I guess we’ll just have to agree to disagree.

Best

Rick

–

Richard S. Marken

www.mindreadings.com
Author of Doing Research on Purpose.

Now available from Amazon or Barnes & Noble

[rsmarken]

[From Rick Marken (2015.08.06.1000)]

···

Rick Marken (2015.08.05.1500)

RM: Since this is a rather verbal group I think it’s better to say that that perceptions (perceptual signals) are constructed, not derived, from lower level perceptual inputs and, ultimately, sensed effects of physical variables. This is a better description of the process implied by the equation p = f(x.1,x.2…x.n) – the function f() constructs a perception from the “raw material” provided by the physical variables x.1,x.2…x.n. Construction is also a better description of the perceptual process in PCT since Powers was more comfortable with “constructivist perception” types like von Glasersfeld than with “direct perception” types like J. J. Gibson.

RM: I just realized that this may be the reason we are arguing about whether control systems control an “aspect of the environment” when they control a perceptual signal. According to the PCT “constructivist” model of perception, a perception is a construction that does not necessarily have a counterpart in the environment. The perception of the “taste of lemonade” is a good example. This perception exists nowhere in the environment; it is constructed from the sensed state of various chemical variables.These chemicals are the arguments, x.1,x.2…x.n, of the perceptual function that constructs the perception of the taste of lemonade.

RM: The perception constructed from x.1,x.2…x.n will only be the “taste of lemonade” if each x.1,x.2…x.n has the correct value (the proper proportions of lemon juice and sugar, for example). But I can see how one can think of the perception of the taste of lemonade as not existing in the environment, because it doesn’t; it only exists as a perception in the person drinking the lemonade. But this perception won’t exist without the proper values of the physical variables, x.1,x.2…x.n, that exist out in the environment. The perceptual function, f(), determines what values of the environmental variables will result in a perception of the taste of lemonade. In this sense, the perceptual function defines the “aspect of the environment” – the configuration of values of x.1,x.2…x.n – that correspond to the perception of the “taste of lemonade”.

RM: When you make lemonade – control for the perception of the taste of lemonade – you are manipulating variables in the environment – the amount of lemon juice (x.1), water (x.2) and sugar (x.3) that you put in the pitcher – in such a way that they can be constructed into a perception of the taste of lemonade. So you are controlling an aspect of the environment – the relative proportion of different chemicals in the pitcher – when you control for the perception of “the taste of lemonade”. This is why I say that when you control a perception you are also controlling an aspect of the environment – the aspect of the environment from which that perception is constructed.

Best

RM: To me, “perception” refers to some kind of representation inside of a device of a state of affairs outside (in the environment of) the device. In PCT a perception is a perceptual signal – an afferent neural current – that varies as a function of variations (ultimately) in physical variables: p = f(x.1,x.2…x.n), where x.i are different physical variables.

RM: What you may be “missing” is the fact that all perceptual signals are the same – they are all just variations in the rate of afferent neural firing. The only difference between one perceptual signal and another is in how it’s derived from lower level perceptual inputs and, ultimately, sensed effects of physical variables.

Rick

Richard S. Marken

www.mindreadings.com
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

[rupert]

[From Rupert Young (2015.08.06 20.00)]
(Rick Marken (2015.08.05.1500)]
Surely q.i is the environmental aspect not q.ig? It is q.i which is
on the environment side of the boundary not q.ig.
q.ig IS p, which, as you say, is being controlled to keep it to
match the reference. But q.i is not.
I can’t speak for the others, but here’s a few of my points.
Sure, at lower levels, but at higher levels they can be imagined
representations of things that do not actually exist outside. Sure, a perceptual signal (q.ig) may correspond to, or be a
function of, variable aspects of the environment (q.i) but it is the
perceptual signal that is controlled not the variable aspects of the
environment.
Rupert

···

            Rupert Young

(2015.08.05 20.00)

                        RY: I think I'm

coming down with a bad case of deja vu, but
could you clarify? Are you saying that these
two statements are equivalent:

                        "Behaviour is the control of perception"



                        "Behaviour is the control of variable

aspects of the environment"

RM: Yes.

            RY: Yes, and if we look at the live block demo .....

However, if you set the input gain to something other
than 1 then the input quantity will change, but the only
the perception will remain equal to the reference. So,
even though the perception is a function of physical
variables it is only the perception that is the
controlled variable, so these two statements are not
equivalent.

            RY: Is this not the case?

          RM: It is the case that what a control system controls

is the perceptual signal, which I presume, based on your
analysis of the LiveBlock demo, is what you are referring
to as the perception. But when the system is controlling
the perceptual signal, keeping it matching the reference
signal, it is also controlling the aspect of the
environment to which that signal corresponds, which in the
LiveBlock demo is the input quantity times the gain
(q.i*g).

            RY: Or am I missing

something?

          RM: I think so but it's hard to say what it is. For

some reason you (and several others) seem to think that
control of perception is in some way separate from control
of aspects of the environment. Actually, I don’t really
understand what your point is.

•     The concept "aspect" of the environment seems a bit vague and
  

  could refer to many things; though you may be using it in a
  specific sense. For example, when we are about to have a shower
  we hold our hand under the water stream and turn the heat dial
  until we feel the heat is to our liking. The heat dial is an
  aspect of the environment but, I am sure you will agree, it is
  not that which is the controlled variable, even though it is an
  environmental variable. The heat dial is varied until the
  perceptual goal is reached, and is part of the feedback
  function, I would say. It is our perception which is being
  controlled, which is internal to our nervous system, and,
  therefore not an environmental variable or aspect. I think what
  you may identify as the aspect of the environment is the actual
  temperature of the water (which could be objectively measured)
  and this may come to a particular value which corresponds to our
  desired perception. However, it is not the temperature of the
  water that is the controlled variable, but our perception
  of the temperature of the water. (Incidentally I would be very
  interested to know if anyone here gets into a dry shower, stands
  under it and then turns on the water; without testing
  the temperature).

• The phrase "Behaviour is the control of
  variable aspects of the environment " suggests that it
  is something out in the environment that is controlled and
  neglects the fundamental point that it is internal perceptions
  that are controlled. It may be the case that variable
  aspects of the environment appear to be controlled, like the
  temperature of the water, but these are as a side-effect of
  the control of perceptions.

• Lower down in the hierarchy it does seem that variable
  aspects of the environment are involved in control (even if
  not actually controlled). However, as we go up the levels
  perceptions become more conceptual and abstract. At the
  higher levels it questionable whether
  variable aspects of the environment are even
  involved in control. How, for example, could someone’s
  control of their perception of god be said to be the same as
  control
  of variable aspects of the environment, or even
  involve
  control of variable aspects of the
  environment; likewise with justice or someone’s perception
  that Mexican migrants are rapists or that a woman’s word is
  worth half of that of a man’s. The statement

                "Behaviour is the control of perception" is a general
  

  principle applicable to all levels of behaviour; the
  statement “Behaviour is the control of variable
  aspects of the environment” may be applicable to some lower
  level systems but is not a general principle.

          RM: To me, "perception" refers to some kind of

representation inside of a device of a state of affairs
outside (in the environment of) the device. In PCT a
perception is a perceptual signal – an afferent neural
current – that varies as a function of variations
(ultimately) in physical variables: p = f(x.1,x.2…x.n),
where x.i are different physical variables.

          RM: What you may be "missing" is the fact that all

perceptual signals are the same – they are all just
variations in the rate of afferent neural firing. The only
difference between one perceptual signal and another is in
how it’s derived from lower level perceptual inputs and,
ultimately, sensed effects of physical variables. The
“derivation” is the perceptual function that defines
whether variations in a perceptual signal correspond to
variations in the size of a book, the shape of a letter,
the relationship between the book and a table top, etc.
That is, the only way to discriminate between the
different perceptual signals we control is in terms of the
(variable) aspect of the environment to which that
perception corresponds – or, better, the aspect of the
environment from which that perceptual signal is derived…

[rupert]

[From Rupert Young (2015.08.06 20.00)]

(Rick Marken (2015.08.06.1000)]

Yes, you *manipulate*     variables in the environment, but I

don’t think that is the same as controlling an aspect of
the environment. To control the perception of the sweetness of your
lemonade you vary the amount of sugar until the desired
sweetness is realised.

Rupert

···

RM: When you make lemonade – control for the
perception of the taste of lemonade – you are
manipulating variables in the environment – the amount of
lemon juice (x.1), water (x.2) and sugar (x.3) that you
put in the pitcher – in such a way that they can be
constructed into a perception of the taste of lemonade. So
you are controlling an aspect of the environment – the
relative proportion of different chemicals in the pitcher
– when you control for the perception of “the taste of
lemonade”. This is why I say that when you control a
perception you are also controlling an aspect of the
environment – the aspect of the environment from which
that perception is constructed.