circular causation

Bruce

Interesting, thanks

AJ: But what about the (billions or trillions) of living organisms inside the body? Are they not applying PCT according to their lights? I am
also curious about intermediate more complex organisations, i.e. the organs.

BN: Yes, one of my abiding questions about the nervous system is: what’s in it for the nerve cell? / Clearly, the rate of firing in a neuron
is not controlled by the neuron; if it were, control by the neuron would inevitably come in conflict with control by the organism

AJ: I do not know enough about the detail, but arguing from general principles, I should say that the means whereby the organism as a whole influences
the behaviour of the nerve cell in its own control behaviour is by changing the cue signals (by changing electrical, chemical and other characteristics of the environment).

BN: As to the more general question, what advantages ‘motivate’ the long evolution of multi-celled organisms from isolated cells to colonies,
predator/prey, parasites, symbiotes, etc., the obvious answer, or part of an answer, seems to be stabilization of the cell’s environment, facilitating the cell’s control of its inputs.

AJ: I have been reluctant to buy into this argument for decades. Firstly, I know too much about the historical origin of this line of thinking and I
am not convinced that a tautology constitutes a theory, useful as it may be. I am also aware that after decades of attempting to explain the production of life initially, we do not yet have an answer. I am also conscious that the dominant methodology has
been reductionist and positivist. There is very little attention being given to the structural aspects of the organism except by a few biologists who have been considered rather brilliant but eccentric. I note that there is a movement in physiology away
from the bottom up approach to the structuring of the organism as a whole, one of which consists of middle-up/down. I also subscribe to Goethe’s theory of morphology, at least as far as botany, it has a substantial following. So I am inclined to think that
the jury is out in terms of understanding. You do not need to tell me that this is a somewhat dissident view.

···

…………………¦â€¦â€¦â€¦â€¦………………………………….

Angus Jenkinson

On 27/12/2017, 15:07, “Bruce Nevin” bnhpct@gmail.com wrote:

[From Bruce Nevin (2017.12.27.09:44 ET)]

Angus Jenkinson, 20.10, Singapore 27.10.17 –

AJ: But what about the (billions or trillions) of living organisms inside the body? Are they not applying PCT according to their lights? I am also curious about intermediate more complex organisations, i.e.
the organs.

Yes, one of my abiding questions about the nervous system is: what’s in it for the nerve cell?

Clearly, the rate of firing in a neuron is not controlled by the neuron; if it were, control by the neuron would inevitably come in conflict with control by the organism of whose behavioral control hierarchy it is a functional part. The
rates of firing, the perceptual, reference, and error signals in our model, can only be byproducts of whatever values the neuron is controlling as it carries on its little cellular life. The rate of firing probably
must be imperceptible as such to the cell, ensuring that the cell cannot control it.

As to the more general question, what advantages ‘motivate’ the long evolution of multi-celled organisms from isolated cells to colonies, predator/prey, parasites, symbiotes, etc., the obvious answer, or part of an answer, seems to be stabilization
of the cell’s environment, facilitating the cell’s control of its inputs.

One wonders about the failure of individual cells to be ‘good citizens’, resulting in cancerous tumors.

I wonder, too, about possible ‘byproducts’ of our participation in collective control. Do unintended side effects have functions, necessarily imperceptible to cellular us, in some larger structure? Such effects would have to be imperceptible
to us as such, ensuring that we cannot bring them under our control.

/Bruce

On Wed, Dec 27, 2017 at 7:12 AM, Angus Jenkinson angus@angusjenkinson.com wrote:

Angus Jenkinson, 20.10, Singapore 27.10.17

Thank you Bruce. We are converging.

But what about the (billions or trillions) of living organisms inside the body? Are they not applying PCT according to their lights? I am also curious about intermediate more complex organisations, i.e. the organs.

Mobile message

On 27 Dec 2017, at 00:31, Bruce Nevin bnhpct@gmail.com wrote:

[From Bruce Nevin (2017.12.26.11:17 ET)]

Angus Jenkinson (Tue, Dec 26, 2017 at 3:31 AM) –

Yes, I fully agree about ‘external determinism’, and as we know PCT demonstrates how environmental factors do not govern behavior. This has got confused because of how difficulties controlling perceptions of environmental factors can influence
what perceptions are controlled, by what feedback paths, and in accord with what reference values. However, I was referring to what one by contrast might term ‘internal determinism’ within a given negative-feedback control loop. Substituting ‘determine’ in
place of ‘cause’, I rephrase:

By physical properties of the environment, output q.o determines (with environmental disturbances as other determinants) the perceived state of perceived environmental variables {v1…vn}; which by physical
properties of the environment such as transmission of photons determine stimulation to input sensors {s1…sn}; which by chemical and electrical properties of the nervous system and its biochemical environment determine the value (the firing rate) of an inhibitory
‘neural current’ in a perceptual input function; synapsing of which with an excitatory reference signal determines the value of an error signal neural current; which (often along with other such error signals) determines the value of an excitatory reference
signal; which determines output q.o; closing the loop of circular determination through the environment.

Looking forward with you to a good year ahead.

/Bruce

Martin

No butting in. I think it is a collegiate field.

I followed your line of enquiry, and your comment that it was something new in this group. (I have left your comment below untouched since it is basically
the whole argument that I think needs to hold together and to which I am responding).

There are several factors that I think would need to be taken account of. I imagine you were trying to keep things simple.

In the first instance, you talk about an arbitrary network, but of course the essence of an organism is that it is not an arbitrary network. And then
your discussion is primarily about flows of energy, but I think in order to be able to adequately deal with the organisation of a living system, we need to deal with information, regarded as the basis of order. Maintenance of various information states will
be critical to the homeostatic aspect and it will be part of the control system. I would go further and add imparity, from ternary cybernetics. And then the process in which membranes come about, or islands, is an organizational closure. That is the production
of the island is not arbitrary but brought forth by the island itself, if we stay with the metaphor. It is precisely in respect of this that ternary cybernetics introduces both information and imparity. Essentially, an island is not merely something that
happens to have a boundary, but rather something that has its own organism organisation in which the physical boundary of the membrane is only one aspect of its closure.

We are also dealing with recursive organization, so that an organ is a “middle unit� in a recursive hierarchy. I think it might be interesting to draw
an analogy between this and the control hierarchy. Each lower unit has to belong to the whole organism in terms of an identity fit, which is part of the information loops, and as shows up in for example the immune system.

Notwithstanding these remarks, I would be interested to see how you develop your thinking in this area.

···

…………………………¦â€¦â€¦â€¦â€¦â€¦â€¦………………………….

Angus Jenkinson

On 27/12/2017, 16:32, “Martin Taylor” mmt-csg@mmtaylor.net wrote:

[Martin Taylor 2017.12 27.09.54]

Angus Jenkinson, 20.10, Singapore 27.10.17

Thank you Bruce. We are converging.

But what about the (billions or trillions) of living organisms inside the body? Are they not applying PCT according to their lights? I am also curious about intermediate more complex organisations, i.e. the organs.

Sorry to butt in, but this is interesting.

First, let’s make a distinction between negative feedback loops and the subset of them that are control loops. Think of a large arbitrary network consisting of links that can transport energy. If there is any circuit A-B-C…-Z-A within the network, there is
a feedback loop in which variations of energy flow from A to B have some influence on the variations of energy flow from Z to A. If that influence is such as to decrease the variation in the energy flow A-B, the loop is a negative feedback loop, otherwise
it is a positive feedback loop.

A feedback loop has a loop gain. If the loop gain is positive and greater than 1.0 the energy variation will grow exponentially, until some limit is reached either in the available energy supply from an external source (which must exist) or some part of the
link is broken physically (melting, explosion, burn-out). Positive feedback loops with a loop gain greater than unity inside an organism cannot last for long unless the loop gain is non-linear as a function of energy flow in at least one of the links, so that
the loop gain goes to or below unity at a sufficiently low energy flow rate.

Many interlocking negative feedback loops can survive within this arbitrary network. Some are long, some short. Some are symmetrical in the sense that energy flow rates are similar all around the loop, some are not. A control loop is a specific kind of short
negative feedback loop characterized by strong asymmetry and the existence of at least one source of energy flow variability from outside the loop (but still possibly within the network). A typical control loop has two, with one node (the A’s and B’s above)
in the energy flow in each direction between the two nodes at which energy variation from outside enters the loop. We conventionally call these two nodes R (reference) and D (disturbance), and the energy flow on the side of the loop between R and D is much
greater than between D and R.

Without this kind of asymmetry, the negative feedback loop is a homeostatic loop rather than a control loop. A control loop is a specific variety of homeostatic loop. Its main thermodynamic function is to use a through energy flow to export entropy from its
inside and distribute it to the outer world, in effect acting like a refrigerator for a single degree of freedom.

Returning to the large arbitrary network, many control loops and homeostatic loops can be sustained within it. Around any part of it, one can draw a boundary, arbitrarily dividing the Universe defined by the entire network into segments that may overlap or
may be isolated within islands defined by these boundaries. Large regions of the network defined by wide ranging boundaries may contain many small network islands that contain smaller ones ad infinitum (or until they are so small that they cannot contain a
loop of any kind). Energy flows cross these boundaries in both directions, with equal flows in and out of any boundary-defined region. If the flows were not equal, the part of the network within the boundary would either boil, burn, or freeze.

In some cases, we choose to define boundaries in the network using features that are not part of the network, such as physical membranes such as skin or the exterior of an organic cell. These membranes are physically defined by their restriction on or dissipation
of energy flows across them. Cells have pores they can open and shut to allow specific kinds of molecules in and out, multicellular organisms have sensors that allow low-energy variations in across their skin and effectors that allow high-energy variations
out, as well as specialized places to accept high-level energy sources and export high-entropy waste.

When a control loop has links that cross a boundary, whether over a membrane or not, it must have one link that outputs energy variation and one that inputs energy variation, and these are asymmetric, the input being at a much lower energy level than the output.
This asymmetry is not necessary for a generic homeostatic loop, but for most organisms, even bacteria, the energy levels of the inputs allowed by a membrane are far lower than for the outputs, in which the membrane itself may contribute through shape change
of the organism. This being the case, of loops that cross membranes, the control type is appreciably more likely than other kinds of homeostatic loops.

Control systems are “network islands” that may have many low-level inputs and high-level outputs whose effects on the “outside” influence the variations in the low-level inputs. A control system may consist of a single control loop, many control loops, or some
complex in which no physical entities can be associated with any particular unitary control loop. The hierarchical perceptual control system hypothesized by Powers is a specific example of a control system of a kind that might be present within every organism,
no matter how small, or might be present in none. A “network island” that contained only the perceptual side of the hierarchy would not be a control system, though it might be part of a homeostatic system, but it is arguable whether a “network island” that
contained only the output side of the hierarchy might be a control system, since its inputs are all low-level “error” variables while its outputs are high-level and do influence the values of its inputs.

I suspect this difference between the perceptual and output sides of the control hierarchy might be responsible for the one-sided behaviourist and cognitive approaches to psychology. defining the “network island” boundary by the visible skin of an organism
cannot admit those approaches.

If we consider organisms as “network islands”, this argument suggests that homeostatic loops will be relatively common in societies relative to control loops, as they will be in networks that do not cross physical boundaries inside organisms (e.g. hormonal
and other physiochemical loops that work within a fluid medium). But in closed societies that enforce a boundary between inside and outside (group members and non-members) the question is the extent to which the group allows the input of low-level variation
to affect high-level output. Some may, some may not. Ones for which an observed or imagined skin colour (low-level) affects high-level action output that might reduce the likelihood of observing that skin colour do have the characteristics of control systems,
no matter how their energy flows are structured internally. So are sports teams.

Angus asks about organs within bodies. What defines them? An organ has a boundary that can be specified without reference to energy flows. A surgeon can say that this cell is part of a heart or of a liver or kidney whereas that cell is not. From outside the
organ we can see that it has a function within the organism, and its function helps keep the rest of the organism functioning stably. So we can suggest that an organ within an organism is likely to be part of at least one homeostatic loop when it is functioning
properly. But what about from the inside of the membrane that defines the boundary of the organ? What energy variations flow into and out of it? Are they symmetric? Are the inputs at low energy levels and the outputs at higher levels? For the heart, they certainly
are, and the same is true of any muscular organ, including individual muscles and muscle fibres.

The question is whether some low-level input is compared with another to produce an influence on the high-level output. If so, and if the high-level output influences the variation of the low-level input, then the organ constitutes a control system (possibly
consisting of many control loops and internal feedback loops of various kinds). If not, it is not a control system.

I don’t think I have seen control systems described from this viewpoint on CSGnet, though I imagine Bill Powers probably thought of it and maybe wrote about it somewhere I haven’t seen or have forgotten. But it is often useful to look at things from a different
angle if you want to make sense of them. I hope this has that function.

Martin

[Martin Taylor 2018.01.02.12.24]

True. The interesting question is "why is it not?.

Yes, I don’t think that viewpoint has previously been used on
CSGnet, and I think that it is valuable to see things from many
different viewpoints. Energy involves thermodynamic considerations
that apply very generally, to networks, whether arbitrary or not, as
well as to continuous materials.
Absolutely, but when I tried to make this point on CSGnet a
quarter-century ago, it caused such a string disturbance as to swamp
the mailing list with diatribes from Rick about there being “no
information from the perception in the output. None. Nada”, and
since he was as impervious to mathematical proof than as he is now,
I have carefully avoided introducing concepts from information
theory into PCT discussions (at least those on CSGnet; there are
others).
I don’t know what you mean by “imparity” apart from “difference or
inequality”, and from what a quick google suggests, “ternary
cybernetics” sounds like an unsophisticated stab at the idea of
hierarchical control. I presume you mean neither of these. Could you
explain?
I think I tried to say much the same, but in case I didn’t, I will
simply say that I agree.
One of the questions I was addressing is whether this is actually
true. I am not sure it is. And I am not sure that it is not. Maybe
it is true of some organs and not of others.
I think Bill Powers would have referred to this as “word salad” in
the absence of a more precise description of such concepts as
“identity fit”, “the information loops”, and so on. Moreover, is it
reasonable to draw an analogy between the place of an organ and the
control hierarchy or its organism, when the question is whether it
is an actual part of the mechanism of the control hierarchy?
I would be happy to discuss it with you, but probably not on CSGnet,
for reasons mentioned above.
As a postscript to my message earlier today [[Martin Taylor
2017.18.02.11.22], I wonder if Gödel’s proof applies in fuzzy maths.
Martin

···

On 2018/01/2 11:49 AM, Angus Jenkinson
wrote:

Martin

Â

        No

butting in. I think it is a collegiate field.

Â

        I

followed your line of enquiry, and your comment that it was
something new in this group. (I have left your comment
below untouched since it is basically the whole argument
that I think needs to hold together and to which I am
responding).

Â

        There

are several factors that I think would need to be taken
account of. I imagine you were trying to keep things
simple.

Â

        In

the first instance, you talk about an arbitrary network, but
of course the essence of an organism is that it is not an
arbitrary network.

        And then your discussion is primarily about flows of energy,
        but

I think in order to be able to adequately deal with the
organisation of a living system, we need to deal with
information, regarded as the basis of order. Maintenance of
various information states will be critical to the
homeostatic aspect and it will be part of the control
system.

        I would go further and add imparity, from ternary

cybernetics.

        And then the process in which membranes come about, or

islands, is an organizational closure. That is the
production of the island is not arbitrary but brought forth
by the island itself, if we stay with the metaphor. It is
precisely in respect of this that ternary cybernetics
introduces both information and imparity. Essentially, an
island is not merely something that happens to have a
boundary, but rather something that has its own organism
organisation in which the physical boundary of the membrane
is only one aspect of its closure.

Â

        We

are also dealing with recursive organization, so that an
organ is a “middle unit� in a recursive hierarchy.

        I think it might be interesting to draw an analogy between

this and the control hierarchy. Each lower unit has to
belong to the whole organism in terms of an identity fit,
which is part of the information loops, and as shows up in
for example the immune system.

Â

        Notwithstanding

these remarks, I would be interested to see how you develop
your thinking in this area.

Â

[From Bruce Nevin (2017.01.02.13:40Â ET)]

Angus Jenkinson (Tue, Jan 2, 2018 at 11:49 AM) –

Martin Taylor 2018.01.02.12.24 –

MMT:Â I think that it is valuable to see things from many different viewpoints.Â

Indeed. But it is necessary to show that the diverse viewpoints are viewpoints on the same thing.

AJ: I think in order to be able to adequately deal with the organisation of a living system, we need to deal with information, regarded as the basis of order. Maintenance of various information states will be critical to the homeostatic aspect and it will be part of the control system.

Are you asserting that information is a controlled perception? That’s what “maintenance” sounds like. Or is it the case that when a system is so structured as to control its inputs in a stable way one can observe ‘information states’, and that when its control is no longer stable, or when it is no longer so structured, those ‘information states’ cannot be observed?

MMT:Â when I tried to make this point on CSGnet a quarter-century ago, it caused such a [strong] disturbance as to swamp the mailing list with diatribes … about there being “no information from the perception in the output…” … [consequently] I have carefully avoided introducing concepts from information theory into PCT discussions

If we accept Bateson’s definition of the elementary unit of information as a difference that makes a difference, then changes in the variables that are successively transformed each into the next around the loop are information. (By this I mean of course changes in physical quanta transformed to changes in neural rates of firing = perceptual signal p; changes in the difference resulting from r-p = error signal transformed to changes in reference inputs at higher levels of the hierarchy and ultimately to motor signals at the lowest level; changes in motor signals transformed to changes in muscular contraction/relaxation; that transformed to changes in limb movement and force; that transformed to changes in effect upon perceived physical quanta, merrily we go around.) But I am not confident that this is what you mean by information.Â

I recall with some trepidation that Shannon was equivocal about choosing the term ‘information’ for his measure of channel efficiency, and unhappy with it in the years that followed. In any other usage, ‘information’ has to do with meaning, and the resulting equivocation has been pernicious. Equally pernicious is equivocation between the alternative choice ‘entropy’ and the meaning of that term in thermodynamics, from which has devolved the conception of information as structure. Angus, is this your usage when you talk of “information, regarded as the basis of order”?

So I would ask you to be explicit and clear what you mean by the term ‘information’, as well as about its relevance to PCT.

MMT:Â I would be happy to discuss it with you, but probably not on CSGnet, for reasons mentioned above.

Obviously, I am urging you not to decamp.

Refs:

Logan, Robert K. (2012) What Is Information?: Why Is It Relativistic and What Is Its Relationship to Materiality, Meaning and Organization. Information 2012, 3, 68-91; doi:10.3390/info3010068

Encyclopedia of Information Science and Technology, Fourth Edition. (The Shannon story is on p. 4428 https://goo.gl/4aWWTW.)

···

On Tue, Jan 2, 2018 at 12:45 PM, Martin Taylor mmt-csg@mmtaylor.net wrote:

[Martin Taylor 2018.01.02.12.24]

  On 2018/01/2 11:49 AM, Angus Jenkinson

wrote:

Martin

Â

        No

butting in. I think it is a collegiate field.

Â

        I

followed your line of enquiry, and your comment that it was
something new in this group. (I have left your comment
below untouched since it is basically the whole argument
that I think needs to hold together and to which I am
responding).

Â

        There

are several factors that I think would need to be taken
account of. I imagine you were trying to keep things
simple.

Â

        In

the first instance, you talk about an arbitrary network, but
of course the essence of an organism is that it is not an
arbitrary network.

True. The interesting question is "why is it not?.
        And then your discussion is primarily about flows of energy,
Yes, I don't think that viewpoint has previously been used on

CSGnet, and I think that it is valuable to see things from many
different viewpoints. Energy involves thermodynamic considerations
that apply very generally, to networks, whether arbitrary or not, as
well as to continuous materials.

        but

I think in order to be able to adequately deal with the
organisation of a living system, we need to deal with
information, regarded as the basis of order. Maintenance of
various information states will be critical to the
homeostatic aspect and it will be part of the control
system.

Absolutely, but when I tried to make this point on CSGnet a

quarter-century ago, it caused such a string disturbance as to swamp
the mailing list with diatribes from Rick about there being “no
information from the perception in the output. None. Nada”, and
since he was as impervious to mathematical proof than as he is now,
I have carefully avoided introducing concepts from information
theory into PCT discussions (at least those on CSGnet; there are
others).

        I would go further and add imparity, from ternary

cybernetics.

I don't know what you mean by "imparity" apart from "difference or

inequality", and from what a quick google suggests, “ternary
cybernetics” sounds like an unsophisticated stab at the idea of
hierarchical control. I presume you mean neither of these. Could you
explain?

        And then the process in which membranes come about, or

islands, is an organizational closure. That is the
production of the island is not arbitrary but brought forth
by the island itself, if we stay with the metaphor. It is
precisely in respect of this that ternary cybernetics
introduces both information and imparity. Essentially, an
island is not merely something that happens to have a
boundary, but rather something that has its own organism
organisation in which the physical boundary of the membrane
is only one aspect of its closure.

I think I tried to say much the same, but in case I didn't, I will

simply say that I agree.

Â

        We

are also dealing with recursive organization, so that an
organ is a “middle unit� in a recursive hierarchy.

One of the questions I was addressing is whether this is actually

true. I am not sure it is. And I am not sure that it is not. Maybe
it is true of some organs and not of others.

        I think it might be interesting to draw an analogy between

this and the control hierarchy. Each lower unit has to
belong to the whole organism in terms of an identity fit,
which is part of the information loops, and as shows up in
for example the immune system.

I think Bill Powers would have referred to this as "word salad" in

the absence of a more precise description of such concepts as
“identity fit”, “the information loops”, and so on. Moreover, is it
reasonable to draw an analogy between the place of an organ and the
control hierarchy or its organism, when the question is whether it
is an actual part of the mechanism of the control hierarchy?

Â

        Notwithstanding

these remarks, I would be interested to see how you develop
your thinking in this area.

Â

I would be happy to discuss it with you, but probably not on CSGnet,

for reasons mentioned above.

As a postscript to my message earlier today [[Martin Taylor

2017.18.02.11.22], I wonder if Gödel’s proof applies in fuzzy maths.

Martin

[Martin Taylor 2018.01.02.17.19]

[From Bruce Nevin (2017.01.02.13:40 ET)]

Angus Jenkinson (Tue, Jan 2, 2018 at 11:49 AM) –

        Martin Taylor

2018.01.02.12.24 –

MMT: I think that it is
valuable to see things from many different viewpoints.

        Indeed. But it is necessary

to show that the diverse viewpoints are viewpoints on the
same thing.

That's quite a different issue. "Indeed" suggest that you agree with

what I said. Does the following comment indicate that you did not
understand that I was talking about control networks?

          AJ: I think in order to be

able to adequately deal with the organisation of a living
system, we need to deal with information, regarded as the
basis of order. Maintenance of various information states
will be critical to the homeostatic aspect and it will be
part of the control system.

        Are you asserting that

information is a controlled perception? That’s what
“maintenance” sounds like.

I don't know what Angus means by "Information states", but for

myself I would say that “information” could label a controlled
variable, as in “I will try to get the information for you”, but
when you are using it in describing a network, it has the same
status as any other variable in the network. If you consider that
they could be controlled variables in the mind of a designer of the
network, then they can be controlled. If you are analyzing
mathematically a well specified network and are holding a variable
to some value, then it is a controlled perception in the analyst.

I don't think Angus is sufficiently careful in distinguishing

homeostasis from control, but I could be wrong, because he uses the
words “homeostatic aspect” and indeed a control loop is a type of
homeostatic system, depending on whether you consider that change of
the homeostatic state according to the value of a particular
variable (reference, in a control loop) is permitted for a
homeostatic system.

Martin

[From Rick Marken (2017.01.02.1920)]

···

 Bruce Nevin (2017.10.17.08:50 PT)

BN: I see circular causation in a control loop as follows. By physical properties of the environment, output q.o is a cause (with environmental disturbances as other causes) of the perceived state of perceived environmental variables {v1…vn};

RM: I think you have too many "perceived"s in there: what is controlled is the perceived state of environmental variables?

Best

Rick

Â

which by physical properties of the environment such as transmission of photons are a cause of stimulation to input sensors {s1…sn}; which by chemical and electrical properties of the nervous system and its biochemical environment are the cause of an inhibitory ‘neural current’ in a perceptual input f
unction; synapsing of which with an excitatory reference signal is the cause of an error signal neural current; which is a cause (often along with other such error signals) of an excitatory reference signal; which is a cause of output q.o; closing the loop of circular causation.

Substitute “determines” or “is a determinant of” in place of “cause”. You perhaps may not regard these as synonymous; I do. I’m sure there are niceties of philosophy that I have not considered.

/Bruce

Richard S. MarkenÂ

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

[From Bruce Nevin (2017.01.02.22:00 ET)]

Bruce Nevin (2017.10.17.08:50 PT) –

BN: I see circular causation in a control loop as follows. By physical properties of the environment, output q.o is a cause (with environmental disturbances as other causes) of the perceived state of perceived environmental variables {v1…vn};

Rick Marken (2017.01.02.1920) –

RM: I think you have too many "perceived"s in there: what is controlled is the perceived state of environmental variables?

E.g. a light source is a perceived environment variable. Its brightness is an environment variable that I was referring to as the perceived state of the light source. My apology if that’s careless.

···

On Tue, Jan 2, 2018 at 9:20 PM, Richard Marken rsmarken@gmail.com wrote:

[From Rick Marken (2017.01.02.1920)]

 Bruce Nevin (2017.10.17.08:50 PT)

BN: I see circular causation in a control loop as follows. By physical properties of the environment, output q.o is a cause (with environmental disturbances as other causes) of the perceived state of perceived environmental variables {v1…vn};

RM: I think you have too many "perceived"s in there: what is controlled is the perceived state of environmental variables?

Best

Rick

Â

which by physical properties of the environment such as transmission of photons are a cause of stimulation to input sensors {s1…sn}; which by chemical and electrical properties of the nervous system and its biochemical environment are the cause of an inhibitory ‘neural current’ in a perceptual input f
unction; synapsing of which with an excitatory reference signal is the cause of an error signal neural current; which is a cause (often along with other such error signals) of an excitatory reference signal; which is a cause of output q.o; closing the loop of circular causation.

Substitute “determines” or “is a determinant of” in place of “cause”. You perhaps may not regard these as synonymous; I do. I’m sure there are niceties of philosophy that I have not considered.

/Bruce

Richard S. MarkenÂ

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

Angus Jenkinson 2018 16:26 UK

Martin

Thought I would respond to this separately.

I am in agreement that there is

no information from the perception in the output. In formation does not flow like water through a pipe. It is created by the shape of actions. The agent is active in reading the world and filtering information from it. This is processed according to the intentional
goals and action follows as another autonomous action. One might say that this is a transformational mapping but I think that would be wrong. All the initial perceptual information vanishes into memory or loss and new information is formed.

Mobile message

···

On 2 Jan 2018, at 17:45, Martin Taylor mmt-csg@mmtaylor.net wrote:

but I think in order to be able to adequately deal with the organisation of a living system, we need to deal with information, regarded as the basis of
order. Maintenance of various information states will be critical to the homeostatic aspect and it will be part of the control system.

Show Quoted Content

but I think in order to be able to adequately deal with the organisation of a living system, we need to deal with information, regarded as the basis of
order. Maintenance of various information states will be critical to the homeostatic aspect and it will be part of the control system.

Absolutely, but when I tried to make this point on CSGnet a quarter-century ago, it caused such a string disturbance …about there being “no information from the perception in the output.”

Angus Jenkinson (2018-01-03)

Martin

See the exchange below

Ternary cybernetics is I think an important addition to the understanding of organisation or systems. It is the theory of JD Stewart who is another brilliant but sidelined
scientist. What he argues is that in the explanation of the formation, maintenance and behaviours of (at least) organic systems (organisms), the two domains of energy and information are insufficient. A third domain is required. This is what he calls imparity.
In brief, an action involves a decision and in that decision information difference is not enough to explain the action. The difference in information between options (where this is any behaviour at any level of system) does not in itself produce the decision;
there has to be an imparity between the information for the agent. A lady goes into a shop and sees two curtains one with a red and the other with a blue pattern. Difference in information. So what? She buys the red. He works it out all rather thoroughly and
has a useful website that is easily googled. He is, or was, a British academic.

Best wishes

···

……………………………………………………………………………….

Angus Jenkinson

On 02/01/2018, 17:45, “Martin Taylor” mmt-csg@mmtaylor.net wrote:

AJ: I would go further and add imparity, from ternary cybernetics.

MT: I don’t know what you mean by “imparity” apart from “difference or inequality”, and from what a quick google suggests, “ternary cybernetics” sounds like
an unsophisticated stab at the idea of hierarchical control. I presume you mean neither of these. Could you explain?

[Martin Taylor 2018.01.03.16.18]

···

I will have to have a look. but from
what you describe, it suggests that Stewart has noticed that
stimulus-response behaviourism is inadequate to deal with human
behaviour, and that one has to take into account discrepancies
between what one wants and what is. If it isn’t that, and
“imparity” means simply that the perceptual world is not uniform,
I will have to find out by reference to the original.

  By the way, in case you were wondering, I wonder if anyone thinks

energy and information are sufficient to explain any, let alone
all, of the formation, maintenance, and behaviour of organisms?
Isn’t that suggestion simply setting up a straw man to tilt at?

  Martin
        Angus

Jenkinson (2018-01-03)

Â

Martin

Â

        See

the exchange below

Â

        Ternary

cybernetics is I think an important addition to the
understanding of organisation or systems. It is the theory
of JD Stewart who is another brilliant but sidelined
scientist. What he argues is that in the explanation of the
formation, maintenance and behaviours of (at least) organic
systems (organisms), the two domains of energy and
information are insufficient. A third domain is required.
This is what he calls imparity. In brief, an action involves
a decision and in that decision information difference is
not enough to explain the action. The difference in
information between options (where this is any behaviour at
any level of system) does not in itself produce the
decision; there has to be an imparity between the
information for the agent. A lady goes into a shop and sees
two curtains one with a red and the other with a blue
pattern. Difference in information. So what? She buys the
red. He works it out all rather thoroughly and has a useful
website that is easily googled. He is, or was, a British
academic.

        Best

wishes

………………………………………………………………………….

Angus Jenkinson

Â

Â

Â

          On 02/01/2018, 17:45, "Martin Taylor"

<mmt-csg@mmtaylor.net >
wrote:

Â

          AJ:

I would go further and add imparity, from ternary
cybernetics.Â

                  MT:

I don’t know what you mean by “imparity” apart from
“difference or inequality”, and from what a quick google
suggests, “ternary cybernetics” sounds like an
unsophisticated stab at the idea of hierarchical control. I
presume you mean neither of these. Could you explain?

Â

Martin, to your question, I don’t think so. A good many physicists, cyberneticians, chemists, general systems theorists and complexity scientists (and
others) would seek to explain everything in terms of information and energy, energy effectively taking various forms of in-formation.

···

………………………………………………………………….

Angus Jenkinson

On 03/01/2018, 21:26, “Martin Taylor” mmt-csg@mmtaylor.net wrote:

[Martin Taylor 2018.01.03.16.18]

I will have to have a look. but from what you describe, it suggests that Stewart has noticed that stimulus-response behaviourism is inadequate to deal with human behaviour, and that one has to take into account
discrepancies between what one wants and what is. If it isn’t that, and “imparity” means simply that the perceptual world is not uniform, I will have to find out by reference to the original.

By the way, in case you were wondering, I wonder if anyone thinks energy and information are sufficient to explain any, let alone all, of the formation, maintenance, and behaviour of organisms? Isn’t that suggestion simply setting up a straw man to tilt at?

Martin

Angus Jenkinson (2018-01-03)

Martin

See the exchange below

Ternary cybernetics is I think an important addition to the understanding of organisation or systems. It is the theory of JD Stewart who is another brilliant but sidelined
scientist. What he argues is that in the explanation of the formation, maintenance and behaviours of (at least) organic systems (organisms), the two domains of energy and information are insufficient. A third domain is required. This is what he calls imparity.
In brief, an action involves a decision and in that decision information difference is not enough to explain the action. The difference in information between options (where this is any behaviour at any level of system) does not in itself produce the decision;
there has to be an imparity between the information for the agent. A lady goes into a shop and sees two curtains one with a red and the other with a blue pattern. Difference in information. So what? She buys the red. He works it out all rather thoroughly and
has a useful website that is easily googled. He is, or was, a British academic.

Best wishes

…………¦â€¦â€¦………………………………………€¦â€¦â€¦â€¦â€¦â€¦â€¦â€¦â€¦â€¦….

Angus Jenkinson

On 02/01/2018, 17:45, “Martin Taylor” mmt-csg@mmtaylor.net wrote:

AJ: I would go further and add imparity, from ternary cybernetics.

MT: I don’t know what you mean by “imparity” apart from “difference or inequality”, and from what a quick google suggests, “ternary cybernetics” sounds like
an unsophisticated stab at the idea of hierarchical control. I presume you mean neither of these. Could you explain?

[From Bruce Nevin (2017.10.17.08:50 PT)]

Angus, you might do a search on the word “determine”. This was often Bill’s preferred way of speaking of the relation between successive functions in the loop.

I have used the term ‘circular causation’, frankly adopting it from my reading of Buddhist psychology, with which I am in sympathy.

I see circular causation in a control loop as follows. By physical properties of the environment, output q.o is a cause (with environmental disturbances as other causes) of the perceived state of perceived environmental variables {v1…vn}; which by physical properties of the environment such as transmission of photons are a cause of stimulation to input sensors {s1…sn}; which by chemical and electrical properties of the nervous system and its biochemical environment are the cause of an inhibitory ‘neural current’ in a perceptual input f
unction; synapsing of which with an excitatory reference signal is the cause of an error signal neural current; which is a cause (often along with other such error signals) of an excitatory reference signal; which is a cause of output q.o; closing the loop of circular causation.

Substitute “determines” or “is a determinant of” in place of “cause”. You perhaps may not regard these as synonymous; I do. I’m sure there are niceties of philosophy that I have not considered.

···

/Bruce

Bruce

My sincere apologies for my late response. And thank you for your explanation. I too have some sympathies with Buddha and quoted him extensively in a
book I wrote.

I am of course familiar with “determine� and it was the very possibility of a deterministic view built into “circular causation� that troubled me. I am
convinced not only that there is no determinism in human action (the closest to it is habit), but that in fact PCT demonstrates that humans (at least) act to negate external determinism. Moreover the perceptual control properties of living organism is functional
at all levels of human system organism.

Very best wishes for 2018

···

………â€â€¦â€¦………………………………………………………………….

Angus Jenkinson

On 18/10/2017, 00:10, “Bruce Nevin” bnhpct@gmail.com wrote:

[From Bruce Nevin (2017.10.17.08:50 PT)]

Angus, you might do a search on the word “determine”. This was often Bill’s preferred way of speaking of the relation between successive functions in the loop.

I have used the term ‘circular causation’, frankly adopting it from my reading of Buddhist psychology, with which I am in sympathy.

I see circular causation in a control loop as follows. By physical properties of the environment, output q.o is a cause (with environmental disturbances as other causes) of the perceived state of perceived environmental variables {v1…vn};
which by physical properties of the environment such as transmission of photons are a cause of stimulation to input sensors {s1…sn}; which by chemical and electrical properties of the nervous system and its biochemical environment are the cause of an inhibitory
‘neural current’ in a perceptual input f unction; synapsing of which with an excitatory reference signal is the cause of an error signal neural current; which is a cause (often along with other such error signals) of an excitatory reference signal; which is
a cause of output q.o; closing the loop of circular causation.

Substitute “determines” or “is a determinant of” in place of “cause”. You perhaps may not regard these as synonymous; I do. I’m sure there are niceties of philosophy that I have not considered.

/Bruce

[From Bruce Nevin (2017.12.26.11:17 ET)]

Angus Jenkinson (Tue, Dec 26, 2017 at 3:31 AM) –

Yes, I fully agree about ‘external determinism’, and as we know PCT demonstrates how environmental factors do not govern behavior. This has got confused because of how difficulties controlling perceptions of environmental factors can influence what perceptions are controlled, by what feedback paths, and in accord with what reference values. However, I was referring to what one by contrast might term ‘internal determinism’ within a given negative-feedback control loop. Substituting ‘determine’ in place of ‘cause’, I rephrase:

By physical properties of the environment, output q.o determines (with environmental disturbances as other determinants) the perceived state of perceived environmental variables {v1…vn}; which by physical properties of the environment such as transmission of photons determine stimulation to input sensors {s1…sn}; which by chemical and electrical properties of the nervous system and its biochemical environment determine the value (the firing rate) of an inhibitory ‘neural current’ in a perceptual input function; synapsing of which with an excitatory reference signal determines the value of an error signal neural current; which (often along with other such error signals) determines the value of an excitatory reference signal; which determines output q.o; closing the loop of circular determination through the environment.

Looking forward with you to a good year ahead.

···

On Tue, Dec 26, 2017 at 3:31 AM, Angus Jenkinson angus@angusjenkinson.com wrote:

Bruce

Â

My sincere apologies for my late response. And thank you for your explanation. I too have some sympathies with Buddha and quoted him extensively in a
book I wrote.

Â

I am of course familiar with “determineâ€? and it was the very possibility of a deterministic view built into “circular causationâ€? that troubled me. I am
convinced not only that there is no determinism in human action (the closest to it is habit), but that in fact PCT demonstrates that humans (at least) act to negate external determinism. Moreover the perceptual control properties of living organism is functional
at all levels of human system organism.

Â

Very best wishes for 2018

Â

………€¦â€¦……………………………………………………… ¦â€¦â€¦â€¦â€¦.

Angus Jenkinson

Â

Â

Â

On 18/10/2017, 00:10, “Bruce Nevin” bnhpct@gmail.com wrote:

Â

[From Bruce Nevin (2017.10.17.08:50 PT)]

Â

Angus, you might do a search on the word “determine”. This was often Bill’s preferred way of speaking of the relation between successive functions in the loop.

Â

I have used the term ‘circular causation’, frankly adopting it from my reading of Buddhist psychology, with which I am in sympathy.

Â

I see circular causation in a control loop as follows. By physical properties of the environment, output q.o is a cause (with environmental disturbances as other causes) of the perceived state of perceived environmental variables {v1…vn};
which by physical properties of the environment such as transmission of photons are a cause of stimulation to input sensors {s1…sn}; which by chemical and electrical properties of the nervous system and its biochemical environment are the cause of an inhibitory
‘neural current’ in a perceptual input f unction; synapsing of which with an excitatory reference signal is the cause of an error signal neural current; which is a cause (often along with other such error signals) of an excitatory reference signal; which is
a cause of output q.o; closing the loop of circular causation.

Â

Substitute “determines” or “is a determinant of” in place of “cause”. You perhaps may not regard these as synonymous; I do. I’m sure there are niceties of philosophy that I have not considered.

Â

/Bruce

Angus Jenkinson, 20.10, Singapore 27.10.17

Thank you Bruce. We are converging.

But what about the (billions or trillions) of living organisms inside the body? Are they not applying PCT according to their lights? I am also curious about intermediate more complex organisations, i.e. the organs.

Mobile message

···

On 27 Dec 2017, at 00:31, Bruce Nevin bnhpct@gmail.com wrote:

[From Bruce Nevin (2017.12.26.11:17 ET)]

Angus Jenkinson (Tue, Dec 26, 2017 at 3:31 AM) –

Yes, I fully agree about ‘external determinism’, and as we know PCT demonstrates how environmental factors do not govern behavior. This has got confused because of how difficulties controlling perceptions of environmental factors can influence what perceptions
are controlled, by what feedback paths, and in accord with what reference values. However, I was referring to what one by contrast might term ‘internal determinism’ within a given negative-feedback control loop. Substituting ‘determine’ in place of ‘cause’,
I rephrase:

By physical properties of the environment, output q.o determines (with environmental disturbances as other determinants) the perceived state of perceived environmental variables {v1…vn}; which by physical properties of the
environment such as transmission of photons determine stimulation to input sensors {s1…sn}; which by chemical and electrical properties of the nervous system and its biochemical environment determine the value (the firing rate) of an inhibitory ‘neural current’
in a perceptual input function; synapsing of which with an excitatory reference signal determines the value of an error signal neural current; which (often along with other such error signals) determines the value of an excitatory reference signal; which determines
output q.o; closing the loop of circular determination through the environment.

Looking forward with you to a good year ahead.

/Bruce

[From Bruce Nevin (2017.12.27.09:44 ET)]

Angus Jenkinson, 20.10, Singapore 27.10.17 –

AJ: But what about the (billions or trillions) of living organisms inside the body? Are they not applying PCT according to their lights? I am also curious about intermediate more complex organisations, i.e. the organs.

Yes, one of my abiding questions about the nervous system is: what’s in it for the nerve cell?

Clearly, the rate of firing in a neuron is not controlled by the neuron; if it were, control by the neuron would inevitably come in conflict with control by the organism of whose behavioral control hierarchy it is a functional part. The rates of firing, the perceptual, reference, and error signals in our model, can only be byproducts of whatever values the neuron is controlling as it carries on its little cellular life. The rate of firing probably must be imperceptible as such to the cell, ensuring that the cell cannot control it.

As to the more general question, what advantages ‘motivate’ the long evolution of multi-celled organisms from isolated cells to colonies, predator/prey, parasites, symbiotes, etc., the obvious answer, or part of an answer, seems to be stabilization of the cell’s environment, facilitating the cell’s control of its inputs.

One wonders about the failure of individual cells to be ‘good citizens’, resulting in cancerous tumors.

I wonder, too, about possible ‘byproducts’ of our participation in collective control. Do unintended side effects have functions, necessarily imperceptible to cellular us, in some larger structure? Such effects would have to be imperceptible to us as such, ensuring that we cannot bring them under our control.

···

On Wed, Dec 27, 2017 at 7:12 AM, Angus Jenkinson angus@angusjenkinson.com wrote:

Angus Jenkinson, 20.10, Singapore 27.10.17

Thank you Bruce. We are converging.

But what about the (billions or trillions) of living organisms inside the body? Are they not applying PCT according to their lights? I am also curious about intermediate more complex organisations, i.e. the organs.

Mobile message

On 27 Dec 2017, at 00:31, Bruce Nevin bnhpct@gmail.com wrote:

[From Bruce Nevin (2017.12.26.11:17 ET)]

Angus Jenkinson (Tue, Dec 26, 2017 at 3:31 AM) –

Yes, I fully agree about ‘external determinism’, and as we know PCT demonstrates how environmental factors do not govern behavior. This has got confused because of how difficulties controlling perceptions of environmental factors can influence what perceptions
are controlled, by what feedback paths, and in accord with what reference values. However, I was referring to what one by contrast might term ‘internal determinism’ within a given negative-feedback control loop. Substituting ‘determine’ in place of ‘cause’,
I rephrase:

By physical properties of the environment, output q.o determines (with environmental disturbances as other determinants) the perceived state of perceived environmental variables {v1…vn}; which by physical properties of the
environment such as transmission of photons determine stimulation to input sensors {s1…sn}; which by chemical and electrical properties of the nervous system and its biochemical environment determine the value (the firing rate) of an inhibitory ‘neural current’
in a perceptual input function; synapsing of which with an excitatory reference signal determines the value of an error signal neural current; which (often along with other such error signals) determines the value of an excitatory reference signal; which determines
output q.o; closing the loop of circular determination through the environment.

Looking forward with you to a good year ahead.

/Bruce

[Martin Taylor 2017.12 27.09.54]

Angus Jenkinson, 20.10, Singapore 27.10.17

Thank you Bruce. We are converging.

    But what about the (billions or

trillions) of living organisms inside the body? Are they not
applying PCT according to their lights? I am also curious about
intermediate more complex organisations, i.e. the organs.

Sorry to butt in, but this is interesting.

First, let's make a distinction between negative feedback loops and

the subset of them that are control loops. Think of a large
arbitrary network consisting of links that can transport energy. If
there is any circuit A-B-C…-Z-A within the network, there is a
feedback loop in which variations of energy flow from A to B have
some influence on the variations of energy flow from Z to A. If that
influence is such as to decrease the variation in the energy flow
A-B, the loop is a negative feedback loop, otherwise it is a
positive feedback loop.

A feedback loop has a loop gain. If the loop gain is positive and

greater than 1.0 the energy variation will grow exponentially, until
some limit is reached either in the available energy supply from an
external source (which must exist) or some part of the link is
broken physically (melting, explosion, burn-out). Positive feedback
loops with a loop gain greater than unity inside an organism cannot
last for long unless the loop gain is non-linear as a function of
energy flow in at least one of the links, so that the loop gain goes
to or below unity at a sufficiently low energy flow rate.

Many interlocking negative feedback loops can survive within this

arbitrary network. Some are long, some short. Some are symmetrical
in the sense that energy flow rates are similar all around the loop,
some are not. A control loop is a specific kind of short negative
feedback loop characterized by strong asymmetry and the existence of
at least one source of energy flow variability from outside the loop
(but still possibly within the network). A typical control loop has
two, with one node (the A’s and B’s above) in the energy flow in
each direction between the two nodes at which energy variation from
outside enters the loop. We conventionally call these two nodes R
(reference) and D (disturbance), and the energy flow on the side of
the loop between R and D is much greater than between D and R.

Without this kind of asymmetry, the negative feedback loop is a

homeostatic loop rather than a control loop. A control loop is a
specific variety of homeostatic loop. Its main thermodynamic
function is to use a through energy flow to export entropy from its
inside and distribute it to the outer world, in effect acting like a
refrigerator for a single degree of freedom.

Returning to the large arbitrary network, many control loops and

homeostatic loops can be sustained within it. Around any part of it,
one can draw a boundary, arbitrarily dividing the Universe defined
by the entire network into segments that may overlap or may be
isolated within islands defined by these boundaries. Large regions
of the network defined by wide ranging boundaries may contain many
small network islands that contain smaller ones ad infinitum (or
until they are so small that they cannot contain a loop of any
kind). Energy flows cross these boundaries in both directions, with
equal flows in and out of any boundary-defined region. If the flows
were not equal, the part of the network within the boundary would
either boil, burn, or freeze.

In some cases, we choose to define boundaries in the network using

features that are not part of the network, such as physical
membranes such as skin or the exterior of an organic cell. These
membranes are physically defined by their restriction on or
dissipation of energy flows across them. Cells have pores they can
open and shut to allow specific kinds of molecules in and out,
multicellular organisms have sensors that allow low-energy
variations in across their skin and effectors that allow high-energy
variations out, as well as specialized places to accept high-level
energy sources and export high-entropy waste.

When a control loop has links that cross a boundary, whether over a

membrane or not, it must have one link that outputs energy variation
and one that inputs energy variation, and these are asymmetric, the
input being at a much lower energy level than the output. This
asymmetry is not necessary for a generic homeostatic loop, but for
most organisms, even bacteria, the energy levels of the inputs
allowed by a membrane are far lower than for the outputs, in which
the membrane itself may contribute through shape change of the
organism. This being the case, of loops that cross membranes, the
control type is appreciably more likely than other kinds of
homeostatic loops.

Control systems are "network islands" that may have many low-level

inputs and high-level outputs whose effects on the “outside”
influence the variations in the low-level inputs. A control system
may consist of a single control loop, many control loops, or some
complex in which no physical entities can be associated with any
particular unitary control loop. The hierarchical perceptual control
system hypothesized by Powers is a specific example of a control
system of a kind that might be present within every organism, no
matter how small, or might be present in none. A “network island”
that contained only the perceptual side of the hierarchy would not
be a control system, though it might be part of a homeostatic
system, but it is arguable whether a “network island” that contained
only the output side of the hierarchy might be a control system,
since its inputs are all low-level “error” variables while its
outputs are high-level and do influence the values of its inputs.

I suspect this difference between the perceptual and output sides of

the control hierarchy might be responsible for the one-sided
behaviourist and cognitive approaches to psychology. defining the
“network island” boundary by the visible skin of an organism cannot
admit those approaches.

If we consider organisms as "network islands", this argument

suggests that homeostatic loops will be relatively common in
societies relative to control loops, as they will be in networks
that do not cross physical boundaries inside organisms (e.g.
hormonal and other physiochemical loops that work within a fluid
medium). But in closed societies that enforce a boundary between
inside and outside (group members and non-members) the question is
the extent to which the group allows the input of low-level
variation to affect high-level output. Some may, some may not. Ones
for which an observed or imagined skin colour (low-level) affects
high-level action output that might reduce the likelihood of
observing that skin colour do have the characteristics of control
systems, no matter how their energy flows are structured internally.
So are sports teams.

Angus asks about organs within bodies. What defines them? An organ

has a boundary that can be specified without reference to energy
flows. A surgeon can say that this cell is part of a heart or of a
liver or kidney whereas that cell is not. From outside the organ we
can see that it has a function within the organism, and its function
helps keep the rest of the organism functioning stably. So we can
suggest that an organ within an organism is likely to be part of at
least one homeostatic loop when it is functioning properly. But what
about from the inside of the membrane that defines the boundary of
the organ? What energy variations flow into and out of it? Are they
symmetric? Are the inputs at low energy levels and the outputs at
higher levels? For the heart, they certainly are, and the same is
true of any muscular organ, including individual muscles and muscle
fibres.

The question is whether some low-level input is compared with

another to produce an influence on the high-level output. If so, and
if the high-level output influences the variation of the low-level
input, then the organ constitutes a control system (possibly
consisting of many control loops and internal feedback loops of
various kinds). If not, it is not a control system.

I don't think I have seen control systems described from this

viewpoint on CSGnet, though I imagine Bill Powers probably thought
of it and maybe wrote about it somewhere I haven’t seen or have
forgotten. But it is often useful to look at things from a different
angle if you want to make sense of them. I hope this has that
function.

Martin