Entropy and the necessity of PCT

[Martin Taylor 2004.05.13.2310]

[From Bill Powers (2004.05.13.1007 MST)]

Martin Taylor 2004.05.12.16.37 --

My argument for the necessity of PCT is based on the
very wide-ranging CLAIM of classical physics that entropy increases
over time in closed systems. It doesn't make accurate prediction of
behaviour, but it makes some prediction, and it does so at zero cost
in added hypothesized degrees of freedom for anyone who has accepted
the basic laws of classical physics. It is therefore very strngly
supported. To deny it, one has either to rework classical physics, or
invalidate the argument that classical physics applies when the
entity being examined is a living organism.

I reserve a little corner of skepticism about that claim, too. It is
possible to design a control system on paper which works without any
mention of entropy, so the basic principles of negative feedback control
apply independently of the role of entropy.

I guess I haven't made myself as clear as I might have done. I
thought it was only Marc who didn't understand. So I'm reposting here
the relevant passage from my Editorial in the PCT issue of IJHCI (in
which issue, as some contributors may not know, PCT was applied at
many perceptual levels and different application areas, from simple
tracking through general Human-ComputerInteraction to self-image and
crowd behaviour). I've added a couple of short paragraphs at the
bottom to what was originally published, because the extract as
published leads on to other material. (The following was previously
placed on the ECACS Forum at
<http://www.ecacs.net/cgi-bin/discus/show.cgi?tpc=192&post=527#POST527>).

···

--------------------
The Necessity of Perceptual Control Theory

There is one overwhelming fact about life: to survive, any organism
must in some way stabilize its essential internal chemistry in the
face of disturbances from a turbulent outer world. At least one major
theory proposes that the species that survived the mass extinction
that ended the Cretaceous were preceisely those best able to achieve
this stabilization in the face of large disturbances. Our bodies are
thermodynamically unstable, and decay away as soon as we cease to act
to counter the influences that would destroy us. Every living thing
has ancestors all of whom behaved so that they stabilized their
internal chemistry at least long enough to propagate their genes.

Protection against the buffeting of the outer world can be done in
two ways, and life employs them both. The first way is to develop
passive armour, such as a membrane, skin, or shell. But the armour
cannot be so perfect as to isolate the organism completely from the
outer world; if it did, the organism would die an entropic death.
Just to sustain a minimal internal organization, any living thing
must at least take in high-quality energy from the outer world, and
excrete less organized waste energy. Most do much more.

The second, and more important, way an organism can protect itself
against the disturbances of the world is to counter them as they
occur, actively and powerfully. To do this, the organism must be able
to sense important states of the outer world; it must be able to
compare the sensed states with desirable conditions for those states;
and it must be able to act to influence them so that it can bring
about and maintain the desirable conditions.

"To sense" means to alter some internal state, such as a chemical
concentration or a neural firing rate, in correspondence with changes
of something in the outer world. In Perceptual Control Theory, such
an internal state is called a "perceptual signal," and the value of a
perceptual signal is a "perception." To stabilize a state near some
reference condition is the technical definition of "control." Hence,
when an organism is countering the disturbances form the outer world,
it is controlling its perceptions.

"Perception," in PCT, carries no connotation of consciousness. It is
just the value of a signal. A perception may be related directly to
the current state of some property of the outer environment, but very
rarely is the connection so direct. The value of a perceptual signal
in most cases depends not only on the current state of the physical
variables that affect sense organs, but also on the current values of
other internal variables, which often depend on the history of the
organism.

The external variables that affect the value of a perceptual signal
might be very complex and context-dependent-such as those that relate
to the perception of the commitment of one's government to democratic
values-or they might be as simple as the rate at which visible
photons impinge on the retina, which relates to the perception of
brightness. Both perceptions-democratic values and brightness-
according to Perceptual Control Theory, are just values of a signal,
and neither need be conscious to be effective.

Bringing a perception of some state to a desired (reference or goal)
value with which it is compared, and maintaining it there, is control
in the strict engineering sense of the word. The perception of the
external state is what is stabilized, not the external state itself,
and still less the action that the organism uses to influence the
external state-hence "Perceptual" Control Theory. For this reason,
PCT has a core tenet: "All behaviour is the control of perception."
The actions that stabilize the perception may vary dramatically as
the environmental influences change, but a well controlled perception
varies only when its reference value changes.

Not all perceptions can be controlled. Many, such as the perception
of the height of the sun in the sky, exist as signals that one's
actions cannot influence. Other perceptual signals, such as a
perception of the democratic commitment of the government, can be
influenced only slightly by one's action. Yet other perceptions can
be controlled with ease, such as, for most people, the perceived
location of one's hand.

Perceptual control systems: engineered and human

It is not only living things that contain signal values. Inanimate
objects such as computers do, too. If those values are consequent on
states of the world outside the computer, they may legitimately be
called "perceptions." If a computer acts so as to maintain at a
reference level a signal value that corresponds to a state of the
outer world, it is controlling a perception, just as a living
organism might do. This does not mean that the computer is considered
to be alive, but it does mean that the analytic techniques of control
can be applied to the relevant operations of the computer, just as
they can to the relevant processes of a living thing. Human and
computer both control perceptions by sensing some aspect of the world
and acting to influence the resulting signal toward a reference
value, in the face of external influences that would alter its value.

From a thermodynamic point of view, a control system acts as a
refrigerator for its immediate environment. This means that it has to
use a flow of energy that allows it to export entropy to a part of
the environment unlikely to affect its controlled perceptions. In
colloquial language, no perceptual control system should "foul its
own nest."

Another way of looking at the situation from a physics point of view
is that this necessary energy flow almost certainly will give rise to
some kind of self-organized system with internal feedback loops
(implicit and continuous, as in the structure of a vortex, or
explicit and possibly even discrete). Furthermore, self-organized
systems of this kind tend to evolve toward the edge of chaos, and we
should expect the Perceptual Control networks that constitute an
individual organism to do likewise.

Martin

From[Bill Williams 13 May 2004 10:40 PM CST]

[Martin Taylor 2004.05.13.2310]

>[From Bill Powers (2004.05.13.1007 MST)]
>
>Martin Taylor 2004.05.12.16.37 --
>
>> My argument for the necessity of PCT is based on the
>>very wide-ranging CLAIM of classical physics that entropy increases
>>over time in closed systems. It doesn't make accurate prediction of
>>behaviour, but it makes some prediction, and it does so at zero cost
>>in added hypothesized degrees of freedom for anyone who has accepted
>>the basic laws of classical physics. It is therefore very strngly
>>supported. To deny it, one has either to rework classical physics, or
>>invalidate the argument that classical physics applies when the
>>entity being examined is a living organism.
>
>I reserve a little corner of skepticism about that claim, too. It is
>possible to design a control system on paper which works without any
>mention of entropy, so the basic principles of negative feedback control
>apply independently of the role of entropy.

I guess I haven't made myself as clear as I might have done.

I recommend the essay Martin attached. But, there is also a book by
George Evelyn Hutchinson 1965 _The Ecological Theater and the
Evolutionary Play_ New Haven: Yale University Press that expands
upon the perception that control requires expenditure of energy.
Hutchinson was somewhat unusual for a field biologists in that he
was also interested in theoretical issues.

I think when Bill Powers says,

It is possible to design a control system on paper which works without
any mention of entropy.

He is over looking, for the moment, the fact that no control system can
operate without the expenditure of energy and thus the increase of
entropy.

Bill Williams

[From Bill Powers (2004.05.14.0349 MST)]

Martin Taylor 2004.05.13.2310--

The Necessity of Perceptual Control Theory

There is one overwhelming fact about life: to survive, any organism
must in some way stabilize its essential internal chemistry in the
face of disturbances from a turbulent outer world.

This whole thing is a beautiful essay, whether or not I agree with every word.
When you say that living systems evolve toward the edge of chaos, I can
easily translate that into the idea that reorganization reduces error until
any further refinement would cross the boundary of instability. I see it
happen every time I simulate some form of reorganization.

This doesn't really involve entropy directly, since such a process can be
simulated without requiring any physically meaningful dissipation of
energy. However, because real systems always do need energy to behave and
must get it from somewhere in usable form, energy and entropy are at least
entailed in the operation of real systems, though not in any tightly
deterministic way. This is what I was trying to say in my previous post.
Although the idea of entropy is unavoidable, and although it sets limits on
possible behavioral organizations, it is not deterministic in the sense
that only one sort of entropy change goes with one kind of behavioral
action. A rested muscle and a fatigued muscle are in quite different states
of entropy, yet the behavior produced by the muscle is almost independent
of its entropy, because of the way negative feedback control systems work.
What does change is the amount of energy wasted without doing useful work,
but that is irrelevant to the operation of the neuromotor control system as
long as the energy supply lasts.

Best,

Bill P.

[From Bjorn
Simonsen(2004.05.21,13:31 EuST)]

[From Bill
Powers (2004.05.14.0349 MST)]

This doesn’t really involve entropy directly,
since such

a process can be simulated without requiring any
physically

meaningful dissipation of energy. However, because
real

systems always do need energy to behave and must
get it

from somewhere in usable form, energy and entropy
are at

least entailed in the operation of real systems,
though

not in any tightly deterministic way.

I limit the definition of entropy (PCT connexion) as a measure of the amount of energy in a
physical system that cannot be used to do work.

I perceive our nervous systems as not perfect. There is some “leakage” of
energy if we follow an Output signal from a higher level downwards in the
system.

In B:CP you p.53 you describe that “the lowest-level systems have the
fastest response” and “

that the higher in the hierarchy one looks at behavior, the longer becomes
the averaging time and the longer time a disturbance may act without beeing
corrected.”

At p. 282 you demonstrate that if the speed is decreased
you achieve a stable system.

At p. 245 you say: “In a hierarchy of control systems,
lower-order systems must necessarily have a faster response than higher-order
systems. That is a basic stability requirement”.

And your wonderful “holding ones arm straight out”
experiment exemplifies what you write at p. 53.

I understand this as:

  1.  There
    

is a “leakage”.

  1.  Using
    

more time compensates the ”leakage”.

  1.  There
    

is entropy when a perception is controlled. (The amount of energy that is
leaked away cannot be used to do work.)

How can you
above say: “This doesn’t really involve entropy directly, since such a
process can be simulated without requiring any physically meaningful
dissipation of energy”.


bjorn

I limit the definition of entropy
(PCT connexion) as
[From Bill Powers (2004.05.21.1303 MDT)]

I think I finally got off of standard time.

Bjorn Simonsen(2004.05.21,13:31
EuST)]

a measure of
the amount of energy in
a physical system that cannot be used to do
work.

How
can you above say: “This doesn’t
really involve entropy directly, since such a process can be simulated
without requiring any physically meaningful dissipation of
energy”.
If entropy determines or strongly influences control, then we
would expect changes in entropy to lead to changes in control, or for a
given degree of control to entail a specific change in entropy. This is
not the case. Of course all actions involve entropy changes, but no
particular amount of entropy changes is tied to any particular manner of
control. I can pick up a card and put it into order among the other cards
in my hand, which reduces the entropy of the arrangement of cards. But I
can do this with my arm muscles very tense so they fight against each
other all the way, or in a relaxed way which involves the least possible
expenditure of energy. The entropy change in my muscles is very different
in the two cases, yet the same control action takes place with hardly any
visible difference.

In a simulation of control the situation is even more extreme. I can
model a stable control system or an unstable one, or even a
stimulus-response system, and the energy used by the computer, and
changes of entropy in the computer, can be essentially identical in every
case. The variable under control may be only a number generated by
another part of the program, or it can be a number generated by a sensor
detecting the state of a physical variable, a variable that the computer
actually affects with an output transducer. Yet the process of control as
it takes place in the computer program can be completely identical
whether there is an actual physical variable or not.

My view is that energy, entropy, and other such terms are abstract
descriptions, not causal entities. Calculations of their states may
change as we measure a real process going on, but those calculations are
not causing the states to change or even constraining them. To give them
any kind of causal role is, I think, a mistake.

Best,

Bill P.

Re: Entropy and the necessity of
PCT
[Martin Taylor 2004.05.21.21.14]

[From Bill Powers (2004.05.21.1303
MDT)]
I think I finally got off of standard
time.
If entropy determines or strongly influences control, then we
would expect changes in entropy to lead to changes in control, or for
a given degree of control to entail a specific change in entropy. This
is not the case. Of course all actions involve entropy changes, but no
particular amount of entropy changes is tied to any particular manner
of control. I can pick up a card and put it into order among the other
cards in my hand, which reduces the entropy of the arrangement of
cards. But I can do this with my arm muscles very tense so they fight
against each other all the way, or in a relaxed way which involves the
least possible expenditure of energy. The entropy change in my muscles
is very different in the two cases, yet the same control action takes
place with hardly any visible difference.

That isn’t really the point about entropy. A heat engine can be
highly efficient or very inefficient, but entropy considerations limit
what it can do.

Even that isn’t the point when we come to control. The real point
about entropy in respect of control systems is that a control system
is a refrigerator. Any control system is. Why is that? It is because
the job of the control system is to reduce the variability of one
variable as compared to another. The values of the perceptual variable
have a reduced variance as compared to what they would have if the
control system were to be turned off. That’s cooling, in the general
sense.

How much energy it uses, and how much it increases the entropy of
the world that includes the whole control loop, is irrelevant. But
there’s a limit to each of those quantities. Just as we now can buy
refrigerators that heat the room less than the refrigerators of 25
years ago did, a nice smoothly operating control system will use less
energy, and contribute less entropy to the world than will one with
lots of side-effects or that involves the kind of conflict you
suggest.

My view is that
energy, entropy, and other such terms are abstract descriptions, not
causal entities. Calculations of their states may change as we measure
a real process going on, but those calculations are not causing the
states to change or even constraining them. To give them any kind of
causal role is, I think, a mistake.

They are indeed abstract descriptions and not causal ntities, but
real live entities can’t avoid the natural laws that describe the
behaviour of those descriptions. Energy (or rather, mass-energy) WILL
be conserved, no matter how clever your control system design. And
even though the control system will reduce the entropy of a small part
of the world, its actions will never reduce the entropy of the world
as a whole, and unles the control system takes infinite time to
operate, its actions will increase the entropy of the world. One of
those statements is a design equality, the other a design limit.
Neither is causal.

Martin

Re: Entropy and the necessity of
PCT
[Martin Taylor 2004.05.21.22.29]

[From Bill Powers (2004.05.21.1303
MDT)]

My view is that
energy, entropy, and other such terms are abstract descriptions, not
causal entities. Calculations of their states may change as we measure
a real process going on, but those calculations are not causing the
states to change or even constraining them. To give them any kind of
causal role is, I think, a mistake.

In respect of the necessity of PCT, the
entropy point is that a structure subject to interaction with the
world outside itself will change because of the random outside
influences. There are two ways this can be avoided: (1) either wall
itself off entirely from outside influences, in which case it dies an
entropic death, coming to the state of highest entropy within its
little closed univers, or (2) act on the outer world to counter those
influences that would change its structure. Possibility 2 requires
that there be some representation internal to the structure of just
which aspects of the structure need to be maintained that way, and a
means to ward off or to counter the external influences.

Possibility 2 is PCT. Clearly, PCT
cannot by itself sustain the entire structure, because the
representation would be as complex as the structure while being
included as part of the structure, an impossibility for a finitie
structure. So, some aspects of teh structure can be passively
protected by a shell or membrane or equivalent, without running into
an entropic death, while other aspects use PCT to sustain the
structure, in particular those aspects of the structure that involve
the import-export of energy and the export of entropy.

The refrigerator view of PCT represents
the “cooling” of its structure by cooling the controlled
aspects of the environment. The entropy reduction implied by the PCT
interaction must be compensated by the entropy rise involved in the
energy import-export flow (in a biological organism, eating and
excretion, breathing, radiation, and so forth).

Martin

Re: Entropy and the necessity of PCT
From[Bill Williams 21 May 2004 11:30 PM CST]

···

[Martin Taylor 2004.05.21.22.29]

The refrigerator view of PCT represents the “cooling” of its structure by cooling the controlled aspects of the environment. The entropy reduction implied by the PCT interaction must be compensated by the entropy rise involved in the energy import-export flow (in a biological organism, eating and excretion, breathing, radiation, and so forth).

I find your argument appealing. As long as we don’t treat entropy as some teleological force, it seems reasonable and even preferable to explicitly recognize the obvious-- that control systems “use” energy.

Bill Williams

[From Bjorn Simonsen (2004.05.22,08:55 EuST)]

[Martin Taylor 2004.05.13.2310]

Protection against the buffeting of the outer
world can be done in

two ways, and life employs them both. The first
way is to develop

passive armour, such as a membrane, skin, or
shell. But the armour

cannot be so perfect as to isolate the organism
completely from the

outer world; if it did, the organism would die an
entropic death.

Just to sustain a minimal internal organization,
any living thing

must at least take in high-quality energy from the
outer world, and

excrete less organized waste energy.
Most do much more.

When you say “…the organism would die an entropic
death. …”, can I in accordance with PCT say; If there are no disturbances from
the outer world, all control systems in an organism will have an error like
zero because we perceive what we wish to perceive.

When all control systems in an organism have an error
like zero the organism is brain death.

Not all perceptions can be controlled. Many, such
as the perception

of the height of the sun in the sky, exist as
signals that one’s

actions cannot influence.

In another connexion Rick
also expressed that living
organisms cannot control all perceptions. I am inclined to disagree. If any ECS
is controlling a perception (the signal after PIF) the error (after the
Comparator) will be zero. If it takes time Reorganization will start. When the
intrinsic error is zero again, our organism maybe control other perceptions.

I don’t fall for your example. I don’t control the
height of the sun. I don’t have any wish to perceive that the sun shall have a
certain height. But other people in Norway control the height of the sun. They travel
to Spain in the winter and the error again becomes zero. (I guess you think the
angle over the horizon? – If you think upon the distance to the sun I could
start earning money for a rocket trip. I know I will die before I can afford
such a trip. But I could start controlling that perception too. – I will not.)

I have not terminated my discussion with Rick, but I
will appreciate to hear your comments.

Perceptual control systems: engineered and human.

I see a difference
between an engineered computer and a human (I think you do too).

A thermostat is an
engineered computer. The program (the brain) serves the sensor (the organism
cell), but the sensor can’t reorganize.

bjorn

···

[From
Bjorn Simonsen (2004.05.22,09:30EuST)]

[From
Bill Powers (2004.05.21.1303 MDT)]

If entropy determines or strongly influences control, then we would expect
changes in

entropy to lead to changes in control,
or for a given degree of control to entail a specific

change in entropy. This is not the case.

When
I change the slowing factor, s in
Rick’s hier.exl

        O(t+1) = o(t) +s[g(r-p) – 0(t)]

it
leads to changes in control.

My
basis for saying this is:

  1.  There is a
    

“leakage” when brain signals move from one level to another.

  1.  Using more time
    

compensates the ”leakage”. Changing
the slowing factor to a less value expresses this leakage.

  1.  There is entropy
    

when a perception is controlled. (The amount of energy
that is leaked away cannot be used to do work.)

Is
it possible to reverse your first sentence? Because we get changes in control
when we change the slowing factor, entropy influences control?

bjorn

Re: Entropy and the necessity of
PCT
[Martin Taylor 2004.05.22.0905]

[From Bjorn Simonsen (2004.05.22,08:55
EuST)]
[Martin Taylor 2004.05.13.2310]

Protection against the buffeting of the outer
world can be done in
two ways, and life employs them both. The first
way is to develop
passive armour, such as a membrane, skin, or
shell. But the armour
cannot be so perfect as to isolate the organism
completely from the
outer world; if it did, the organism would die an
entropic death.
Just to sustain a minimal internal organization,
any living thing
must at least take in high-quality energy from the
outer world, and
excrete less organized waste energy. Most do much
more.

When you say “…the organism would die an entropic
death. …”, can I in accordance with PCT say; If there are no
disturbances from the outer world, all control systems in an organism
will have an error like zero because we perceive what we wish to
perceive.

No. Not at all. And when I talked about dieing and entropic
death, I was not talking about a situation in which by chance there is
a period in which no perceptions were being disturbed, but about an
organism that has no contact with the world outside itself. Zero. No
sensory input, no energy input, no waste disposal… And if such an
organism has anything we would call perception, that, too, would die
away into random noise eventually.

When all control systems in an organism have an error
like zero the organism is brain death.

No. Quite apart from the fact that it’s highly unlikely that an
sufficiently large structure of control systems could have completely
non-conflicting control systems, it takes energy and the export of
entropy to keep errors near zero in the presence of contact with the
outer world.

Not all perceptions can be controlled. Many, such
as the perception
of the height of the sun in the sky, exist as
signals that one’s
actions cannot influence.

In another connexion Rick also expressed that living organisms cannot control all perceptions.
I am inclined to disagree. If any ECS is controlling a perception (the
signal after PIF) the error (after the Comparator) will be
zero.

How do you figure that? Controlling means acting so as to alter the
perception, not reducing the error to zero. No perceptual control
system can reduce error to zero, but in principle, one can get
arbitrarily close.

If it takes time Reorganization will start. When
the intrinsic error is zero again, our organism maybe control other
perceptions.

??? I think you have an entirely different conception of PCT or
HPCT than I do. In my conception, this pair of sentences has as much
sense as “Colourless green ideas sleep furiously”. I can’t
formulate a response to them.

I don’t fall for your example. I don’t control the
height of the sun.

What example? And why should I think I could control the height
of the sun, or any other of a host of environmental conditions? Where
did that notion come from? Out of the blue sky?

Perceptual control systems: engineered and
human.

I see a difference between an engineered computer and
a human (I think you do too).
A thermostat is an engineered computer. The program
(the brain) serves the sensor (the organism cell), but the sensor
can’t reorganize.

I fail to see the
relevance. I’m dealing with the boudary conditions that apply to all
control system, engineered or human.

Martin

PS. I’m away for a week,
so may well not be in a position to respond further, depending on what
other traffic occurs during the week.

[From Bill Powers (2004.05.22.0900 MDT)]

Martin Taylor 2004.05.21.21.14 --

Even that isn't the point when we come to control. The real point about
entropy in respect of control systems is that a control system is a
refrigerator. Any control system is. Why is that? It is because the job of
the control system is to reduce the variability of one variable as
compared to another.

So if a blacksmith takes one horseshoe out of a pile and heats it to a
carefully-controlled orange-red, he is refrigerating it? One has to be
careful not to let metaphors get out of hand.

I do understand your point about entropy being associated with limits of
performance. But performance (if I may repeat my own point) is limited by
irereversible processes that prevent recovery of all the work done on a
system, and that is why entropy effects are seen -- not the other way
around. I don't believe that entropy is some overarching principle that
magically governs the way things work. It's the other way around: because
of the way things work, we can derive a generalized concept which we call
entropy.

I think this is important because it is too easy to fall into a way of
speaking about generalizations that makes them seem more real than the
observations they are derived from, or as if they explain something that
the observations don't explain. To be honest, I don't really know where I'm
going with this idea -- it's probably not formed fully enough to be talking
about in public. But does there seem to be a splinter of sense in what I'm
saying?

Best.

Bill P.

If entropy determines or
strongly influences control, then we would expect changes in

entropy to lead to changes in control, or for a given degree of
control to entail a specific

change in entropy.
[From Bill Powers (2004.05.22.0917 MDT)]

Bjorn Simonsen (2004.05.22,09:30EuST)

···

This is not the
case.

When I change the slowing factor, s in Rick’s hier.exl

O(t+1) = o(t) +s[g(r-p) – 0(t)]

it leads to changes in
control.

That’s an excellent example. There are two ways to implement a slowing
factor in a control system. One is to use a dashpot (piston in a cylinder
filled with a viscous fluid) attached to thye load (the controlled
variable), and the other is to use a rate-of-change sensor that adds a
velocity component to the perception of the controlled variable. If the
piston is used, there is a large increase in entropy involved in every
action, because a significant part of the energy used for output is
dissipated as heat and is not recoverable. On the other hand, if rate
feedback is used to get the same effect (mathematically and in reality),
the energy losses are actually reduced, and the entropy increase is
much smaller. The rate feedback causes the output to get smaller before
the error has reached zero, and this means that the action relaxes sooner
before the goal is reached, and overshoots and undershoots are prevented
– they use energy, too.

So there are two ways to stabilize a control system, one entailing large
increases in entropy and the other much smaller increases – and
behaviorally, they are indistinguishable. It is probably true that we
learn to control in ways that are not excessively wasteful of energy, but
there is no law that says we have to.

As Martin Taylor said, entropy concerns limits of performance. But it
says nothing about performance that is well below those limits (as is
most often the case). It also says there is no free lunch: the entropy we
reduce is paid for by the entropy we increase by a larger amount. But I
don’t subscribe to the school of thought that says the behavior of living
systems has been optimized by evolution – I’m more a member of the
“barely good enough to survive” school. So I don’t think that
entropy or even energy considerations come into play very often as
limiting factors.

Best,

Bill P.

Re: Entropy and the necessity of
PCT

[From Bill Powers (2004.05.22.0900
MDT)]

Martin Taylor 2004.05.21.21.14 –

Even that isn’t the point when we come to
control. The real point about

entropy in respect of control systems is that a control system is
a

refrigerator. Any control system is. Why is that? It is because the
job of

the control system is to reduce the variability of one variable as

compared to another.

So if a blacksmith takes one horseshoe out of a pile and heats it to
a

carefully-controlled orange-red, he is refrigerating it? One has to
be

careful not to let metaphors get out of hand.

This emphatically isn’t a metaphor. With respect to the
macroscopic property (properties) for which the blacksmith is
controlling, yes, he is refrigerating the metal of the horseshoe. The
fact that your kitchen refrigerator is dealing in the average variance
of the molecular motion degrees of freedom is irrelevant to the
variance reduction in particular degrees of freedom the blacksmith is
concerned with.

I do understand your point about entropy
being associated with limits of

performance. But performance (if I may repeat my own point) is limited
by

irereversible processes that prevent recovery of all the work done on
a

system, and that is why entropy effects are seen – not the other
way

around. I don’t believe that entropy is some overarching principle
that
magically governs the way things work.
It’s the other way around: because

of the way things work, we can derive a generalized concept which we
call
entropy.
I think this is important because To be
honest, I don’t really know where I’m

going with this idea – it’s probably not formed fully enough to be
talking

about in public. But does there seem to be a splinter of sense in what
I’m
saying?

Yes, there’s sense in it. But as with most things related to
control theory, it works both ways. I go along with much of what you
say, except when you add “not the other way round.” Often,
“The other way round” is equally true, as it is in this
case.

You correctly say: “because of the way things work, we can
derive a generalized concept which we call entropy.” But I would
also say “because of the way things work, we can derive a
generalized concept which we call irreversible processes”.

“Entropy” and “irreversible processes” are
abstractions at the same level, and either can be used to explain the
other, if you first believe in the one you are using as an explanatory
principle. They are connected in a conceptual mesh that includes a lot
of other stuff, such as it being more probable that a random influence
will move the phase point of a system out of than into a precisely
defined tiny region of the phase space. If you believe that, then you
believe in entropy (Boltzmann version), from which you can get the
notion of an irreversible process.

I definitely agree with: "it is too easy to fall into a way
of

speaking about generalizations that makes them seem more real
than the

observations they are derived from, or as if they explain
something that

the observations don’t explain." I run into this a lot, and
it confuses the issue no end when one tries to use the terms and
concepts in a more rigorous sense. People tend to think one is being
metaphoric (as I did about “control” when I first
encountered PCT).

However, I do think you have to be careful when you use the word
“explain.” That term carries a lot of baggage, and leads to
a lot of misunderstanding. What constitutes an explanation lies in the
mind of the one receiving the explanation, not in the acts (words) of
the explainer.

···

To Bjorn, you wrote: “I don’t think that
entropy or even energy considerations come into play very often as
limiting factors.”

No, they probably don’t. However, the necessity argument doesn’t
have any connection with this fact. Once the necessity of PCT has been
established, considerations of control efficiency bring this statement
into play.


I’m just about to get on the road, and will be out of touch for a
week.

Martin

[From Bjorn Simonsen (2004.5.31,09:42 EuST)]

I know I am late

[From Bjorn Simonsen
(2004.05.22,08:55 EuST)]

[Martin Taylor 2004.05.22.0905]

If it takes time
Reorganization will start. When the

intrinsic error is zero again, our organism maybe

control other perceptions.

??? I think you have an
entirely different conception of PCT or HPCT than I do.

In my conception, this pair of sentences has as much sense as
"Colourless

green ideas sleep furiously". I can’t formulate a response to them.

I think you formulated a response. Let me try again.

Conscious controlling some perceptions can be time-consuming. Depending on how my organism behaves or
fails to behave there will be effects on its intrinsic state. Reorganizing may
start. It is unanticipated which
control activities reorganizing will have, but they will contribute to keep the
intrinsic error to be zero.

        >>I

don’t fall for your example. I don’t control the height of the sun.

What example? And why should I
think I could control the height of the sun,

or any other of a host of environmental conditions? Where did that notion

come from? Out of the blue sky?

Maybe we talk at cross-purposes. But I read your text
below as an example (one of many) of perceptions you can’t control.

[Martin Taylor 2004.05.13.2310]

Not all perceptions can be controlled. Many, such
as the perception

of the height of the sun in the sky, exist as
signals that one’s

actions cannot influence. Other perceptual
signals, such as a

perception of the democratic commitment of the
government, can be

influenced only slightly by one’s action. Yet
other perceptions can

be controlled with ease, such as, for most people,
the perceived

location of one’s hand.

Please
point out what is wrong when a Norwegian in November-January controls his
perception of the height of the sun in the sky and he doesn’t perceive it in
the way he desires. Then he changes his place of residence to Spain or to the
Canary Islands. Now he perceive the height of the sun in the sky in a way he
wishes. (A is said
to control B for every disturbing influence acting on B, A generates an action
that tends strongly to counteract the effect of the disturbing influence on B.)

Connected
with this example I think some people control the perception of having adequate
body temperature than the height of the sun in the sky.

And I
think some people, when the sun is in zenith, generate an action that change
from lying nude on the beach to standing position with a broad brimmed hat on
his head.

bjorn

···

Re: Entropy and the necessity of
PCT
[Martin Taylor 2004.05.31.10.19]

[From Bjorn
Simonsen (2004.5.31,09:42 EuST)]

I know I am
late

[From Bjorn
Simonsen (2004.05.22,08:55 EuST)]
[Martin Taylor
2004.05.22.0905]

If it takes
time Reorganization will start. When the

intrinsic error is zero again, our organism maybe

control other perceptions.

??? I think you
have an entirely different conception of PCT or HPCT than I do.

In my conception, this pair of sentences has as much sense as
"Colourless

green ideas sleep furiously". I can’t formulate a response to
them.

I think you
formulated a response. Let me try again.
Conscious
controlling some perceptions can be
time-consuming.

So far as I know, neither PCT nore HPCT implies a position on
consciousness. At varioustimes, different people (including me) have
made proposals about it, but none, so far as I know, have related
consciouslness to reorganization, as you seem to do. Maybe you could
describe your thoughts about how the two connect in a bit more detail?
There might be something to discuss.

My personal take on consciousness (with only subjective evidence
to support it) is that consciousness relates to the fact that we have
too few output degrees of freedom to allow all perceptions to be
controlled at once. Consciousness then relates to the multiplexing
problem. That is true also when the control actions for a
particular perception are not being very effective. Other lower-level
means might be appropriate, which has a similar consequence (a change
from controlling one perception to controlling another at the same
level). If there is a connection to reorganization, it could be there,
provided that one of the intrinsic variables is related directly to
persistent or increasing error in the direct control hierarchy. But I
think that possible relationship is rather tenuous.

Maybe your conception of the relationship between consciousness
and reorganization is more direct?

Depending on
how my organism behaves or fails to behave there will be effects on
its intrinsic state. Reorganizing may start.

In “classic” HPCT, reorganizing is going on all the
time. It happens with lower probability when the intrinsic errors are
less, but it doesn’t stop.

It is
unanticipated which control activities reorganizing will have, but
they will contribute to keep the intrinsic error to be
zero.

Reorganizing can equally well result in an increase in intrinsic
error, but then it is more probable that another quick reorganization
event will occur. Again, I’m talking about “classical” HPCT
as I understand it.

I don’t fall for your example. I don’t control the height of
the sun.
What example?
And why should I think I could control the height of the sun,

or any other of a host of environmental conditions? Where did that
notion

come from? Out of the blue sky?

I understood you to be declaring that you don’t control the
height of the sun. Now I understand that you intended to quote me to
that effect. And you are correct that we do have means to control it
to some extent. But even in th tropics, we can’t keep the sun near the
zenith all the time. Our planes are not (yet) fast enough to
circumnavigate the Earth in 24 hours.

However, that’s irrelevant to the argument. My point was that not
all perceptions are equally easily controlled. They range from
complete inability to influence the perception, to almost
perfect control. Perceptions we can’t (or don’t at the moment)
control, nevertheless can contribute to teh construction of a
perception that we do control.

The issue of entropy is that if we don’t
control at least some perceptions (meaning internal states that are
affected by the outer world), then external events will disrupt our
internal structure irreparably. The only alternative is to shield our
internal structure from all interaction with the outer world – to
create an isolated universe. In an isolated universe, entropy tends to
increase to a maximum, a state usually called “entropic
death”.

The point is that any structured system
either must decay inexorably, or must use perceptual control to
maintain itself. Perceptual control means to act so as to maintain
some internal state near a reference value by acting to counter
whatever external factors might influence that internal
state.

Martin

[From Bjorn Simonsen (2004.06.03, 11:40 EuST)]

[Martin Taylor 2004.05.31.10.19]

I think our discussion slides into a new theme. We may
continue, but first I will end the conversation about my first theme.

[From Bjorn Simonsen
(2004.5.31,09:42 EuST)]

[From Bjorn Simonsen
(2004.05.22,08:55 EuST)]

[Martin Taylor
2004.05.22.0905]

If it takes time
Reorganization will start. When the

intrinsic error is zero again, our organism maybe

control other perceptions.

??? I think you have an
entirely different conception of PCT or HPCT than I do.

In my conception, this pair of sentences has as much sense as
"Colourless

green ideas sleep furiously". I can’t formulate a response to
them.

I think you formulated a
response. Let me try again.

Conscious controlling some
perceptions can be
time-consuming.

So far as I know, neither PCT
nore HPCT implies a position on consciousness. At varioustimes,

different people (including me) have made proposals about it, but none, so
far as I know, have

related consciouslness to reorganization, as you seem to do. Maybe you
could describe your

thoughts about how the two connect in a bit more detail? There might be
something to discuss.

I should not
have mixed in the concept “Conscious”. My point is that controlling certain
perceptions can be time-consuming.

When we control certain perceptions that are
time-consuming (p#r gives e#zero and actions/disturbances give a new p still
different from r and actions/disturbances and so on ), reorganizing may form
new attractors. We can’t plan which attractors, but some will be formed. Now we
control a perception we couldn’t control earlier.

I know you may say that we now are controlling another
perception. But my point is that living organisms control their perceptions.
Sometimes it looks like we can’t control all perceptions. Then Reorganizing gives
us opportunities. And we continue controlling our perceptions. So is life.

I agree that

In “classic” HPCT,
reorganizing is going on all the time. It happens with lower

probability when the intrinsic errors are less, but it doesn’t stop.

[From Bjorn Simonsen
(2004.05.22,08:55 EuST)]

[Martin Taylor 2004.05.13.2310]

In another connexion Rick
also expressed that living organisms cannot control all

perceptions. I am inclined to disagree. If any ECS is controlling a
perception (the

signal after PIF) the error (after the Comparator) will be zero.

How do you figure that? Controlling means acting so as to alter the
perception, not reducing

the error to zero. No perceptual control system can reduce error to zero,
but in principle, one

can get arbitrarily close.

Linguistic precision is necessary. I absolutely agree
with your last sentence.

My personal take on
consciousness (with only subjective evidence to support it) is that

consciousness relates to the fact that we have too few output degrees of
freedom to

allow all perceptions to be controlled at once. Consciousness then relates
to the multiplexing

problem. That is true also when the control actions for a particular
perception are not

being very effective. Other lower-level means might be appropriate, which
has a similar

consequence (a change from controlling one perception to controlling
another at the

same level). If there is a connection to reorganization, it could be there,
provided that

one of the intrinsic variables is related directly to persistent or
increasing error in the

direct control hierarchy. But I think that possible relationship is rather
tenuous.

Maybe your conception of the
relationship between consciousness and reorganization is more direct?

I can imagine that we have too few output degrees of
freedom to allow all perceptions to be controlled at once. I associate both a
conflict and reorganizing, but will not continue this trend now.

I am more involved in thinking upon consciousness just
as a control of perception.

First my definition: Consciousness is the expressing
of one or more perceptions.

Background comments: Looking out my window I can see
hills, houses, the sea, the sky and more. When I focus on three houses in my
field of view I can describe their colour, the roof colour, their chimney and
more. When I stop focusing, I discovered the chimney on the nearest house. It
was in my field of view, but I didn’t include this chimney in my description.
Of course the chimney on the nearest house was portrayed on my retina. And of
course this independent variable was an element in my dependent perceptual signal.
And of course this perceptual signal was included in my behavior,

I have experimented with such visual impressions and corresponding
auditory experiments listening to a symphony and tasting experiments tasting
wine.

Driving a certain route I pass a tunnel. There I have
focused on the two red rear lights from the car in front, white lights passing
on my both sides when I drove and also remembering the appearance of my holiday
cottage in the mountain. I had to
stop this experimenting when I drove the car because I remembered a view of my
holiday cottage very strong. Afterwards I remembered that in a moment it was
only my holiday cottage I could describe. And that was the reason I almost
drove in the rock wall. In that short time I was not able to describe the two red
lights ahead and the white lights passing by at my sides. During this short
experiment I was not able to describe the yellow lights in the tunnel roof passing
by.

Of course there were a picture on my retina of the
objects I couldn’t describe etc.

Conclusion: Consciousness is the expressing of one or
more perceptions. This definition works for me.

I have done experiments with formulating a learned
sentence of 48 words. I have silent formulated the sentence about 20 –30 times
a day for more than a month (not so many times in the end of the 40 days). And the
reciting time I used to formulate the content of the sentence was shortened
from about 30 seconds till less than a second.

I mention this experiment to explain what I mean by
describing what I focus.

I have done experiments controlling the configuration
of some of my body parts. I have done experiments controlling events as regards
my body parts (walking, moving my right arm backwards and forwards and my left
arm sideways in and out, my head from one side to the other, my eyes open and
closed and my tongue in and out (I do this where people don’t see me).

I have done tests telling my capacity to control up to
5 abstract processes.

I control conscious more elements at the lowest levels in HPCT and I
have problems with controlling more than one System Concept.

End of background comments.


** ** >Maybe your conception of the
relationship between consciousness and reorganization is more direct?


For me consciousness is just Control of perceptions.
And the perceptions I control conscious are the perceptions I can describe.

Bill has a equal definition in B:CP saying:
Consciousness consist of perception and awareness.

I don’t think I can control reorganizing, but I am
sure I have controlled perceptions after Reorganizing.

bjorn