We are the problem

[From Rick Marken (990205.0800)]

Kenny Kitzke (990204.2100EST)]

Now you know why you and PCT are so unpopular. And, with
any luck you will be abandoning a systems concept of science
that is inconsistent with human nature.

I still have no idea why I'm so unpopular (I'm so adorable;
just ask my Mom;-)) but I have always known why PCT is
unpopular; PCT is unpopular because it's a disturbance to
the prevailing system concept in the life sciences; the
concept of living systems as causal, rather than control,
systems. The PCT control of input concept of life is being
resisted by those scientists who have a vested interest in
maintaining the current causal conception of how organisms
"work". But this causal system concept will eventually be
abandoned (it may take a couple of generations) because it
is inconsistent with the data, such as that which can be
obtained in any of the demos at:

http://home.earthlink.net/~rmarken/demos.html

I say that science (as an enterprise) is inconsistent with
human nature because human nature is to control (make our
perceptions match references for what those perceptions
should be). Science asks us to _give up control_ to some
extent; it asks us to be willing to revise our references
for what our perceptions _should be_ based on careful
observation of what our perceptions _are_. This is the
opposite of control and I can tell you from personal
experience that it is _very_ hard to do. For example, for
a long time scientists wanted to believe that genes mix like
paint. When Mendel came along with evidence that genes mix
in discrete units and are not "diluted" after each mating
his results were ignored or dismissed because they were a
disturbance to the prevailing conception of gene mixing.

Scientists are human; they want to maintain their perceptions
at their references for those perceptions. But scientists do,
eventually, (sometimes it takes a generation or two) change
their references when the evidence (the disturbance to the
prevailing point of view) becomes overwhelming or when a
model is developed that simply and elegantly accounts for
more of the existing data than the existing model (as with
Einstein's theory).

According to PCT, the cause of human problems is not the
particular system concepts we adopt or the moral principles
we defend or the type of society we live in. The cause
of human problems is _human nature_; we are the problem. It
is our controlling nature that is the problem. What makes
our controlling nature a problem is that, among the
perceptions we want to control are perceptions of the
behavior of other people. Our efforts to control our
perceptions of the behavior of other people typically
result in interpersonal conflict.

This problem cannot be solved by selecting better goals
(better morals, better system concepts, etc.). Nor can
it be solved by arranging society in a particular way
(anarchic, socialistic, etc.). The problem can only be
solved by recognizing that the problem exists and then
being willing to try to abandon (as Bill says in the
now famous Ch. 17 in B:CP) even the _desire_ to control
other people's behavior. This is a recommendation to
_overcome_ one's own human (controlling) nature. This is
not an easy thing to do and there is no recipe for how
to do it. But this is the conclusion we come to from
looking at human problems from a PCT perspective:

We (our controlling natures) are the problem and _we_
(whatever it is about us that allows us to overcome the
desire to control other people) are the solution.

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

[From Bruce Gregory (990205.1130 EST)]

Rick Marken (990205.0800)

This problem cannot be solved by selecting better goals
(better morals, better system concepts, etc.). Nor can
it be solved by arranging society in a particular way
(anarchic, socialistic, etc.). The problem can only be
solved by recognizing that the problem exists and then
being willing to try to abandon (as Bill says in the
now famous Ch. 17 in B:CP) even the _desire_ to control
other people's behavior. This is a recommendation to
_overcome_ one's own human (controlling) nature. This is
not an easy thing to do and there is no recipe for how
to do it. But this is the conclusion we come to from
looking at human problems from a PCT perspective:

If this is what is called for, the program seems doomed to failure. I've
never had much success in giving up my desires. In fact, I don't know
_how_ to control these perceptions. (I don't know how to set my
reference level for chocolate to a very low level.) It seems to me that
a much more achievable goal is to control for not interfering with the
ability of others to exercise control. This I think I know how to do.

Bruce Gregory

[From Rick Marken (990205.1010)]

Me:

The problem can only be solved by recognizing that the
problem exists and then being willing to try to abandon
(as Bill says in the now famous Ch. 17 in B:CP) even the
_desire_ to control other people's behavior.

Bruce Gregory (990205.1130 EST) --

If this is what is called for, the program seems doomed
to failure.

What program?

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

[From Bill Powers (990205.1019 MST)]

Bruce Gregory (990205.1130 EST)--

I've
never had much success in giving up my desires. In fact, I don't know
_how_ to control these perceptions. (I don't know how to set my
reference level for chocolate to a very low level.) It seems to me that
a much more achievable goal is to control for not interfering with the
ability of others to exercise control. This I think I know how to do.

The problem with "giving up" desires is that you are trying to give just
one small group of systems at one or two intermediate levels total control
over the whole system. Your cognitive systems may think they "should" give
up chocolate or sex or breathing, but the rest of the systems have their
own reference levels and resist.

Thought and reasoning are not the highest levels in the hierarchy. This is
why they have so little effect on changing anything important. To make any
significant changes, you must change your principles, and to change your
principles you must change your system concepts. When we're considering
system concepts, we're looking at things higher than logic, science,
methods, language, and so on. We can only experience at this level,
considering how the world looks in terms of its organization, internal
harmony, beauty, and consistency, and be pleased or dissatisfied with it.
If we are dissatisfied enough, changes will begin, and with luck the
changes will be for the better, eventually. If our system concepts change,
so will our principles change to become more mutually consistent, and as
our principles change so will our thoughts, reasoning, and language.

There is no way to change just one thing in the hierarchy.

Best,

Bill P.

[From Bruce Gregory (990205.1315 EST)]

Rick Marken (990205.1010)

Me:

> The problem can only be solved by recognizing that the
> problem exists and then being willing to try to abandon
> (as Bill says in the now famous Ch. 17 in B:CP) even the
> _desire_ to control other people's behavior.

Bruce Gregory (990205.1130 EST) --

> If this is what is called for, the program seems doomed
> to failure.

What program?

The program of recognizing that the problem exists and then being
willing to try to abandon even the _desire_ to control other people's
behavior.

Bruce Gregory

[From Bruce Gregory (990205.1322 EST)]

Bill Powers (990205.1019 MST)

When we're
considering
system concepts, we're looking at things higher than logic, science,
methods, language, and so on. We can only experience at this level,
considering how the world looks in terms of its organization, internal
harmony, beauty, and consistency, and be pleased or
dissatisfied with it.
If we are dissatisfied enough, changes will begin, and with luck the
changes will be for the better, eventually. If our system
concepts change,
so will our principles change to become more mutually
consistent, and as
our principles change so will our thoughts, reasoning, and language.

Since these systems perceptions reference levels are at the highest
conjectured level in the hierarchy, I assume they are hard-wired
(otherwise I have no idea how they got there). If we have such a lofty
wiring maybe Ken is right after all.

Bruce Gregory

[From Rick Marken (990205.1230)]

Me:

What program?

Bruce Gregory (990205.1315 EST)

The program of recognizing that the problem exists and then
being willing to try to abandon even the _desire_ to control
other people's behavior.

As you can tell from my post [Rick Marken (990205.0800)] and
Bill's [Bill Powers (990205.1019 MST)], I am a lot better at
the "describing the problem" part of the program than at
the "explaining how to solve it" part.

Best

Rick

···

---
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

[From Kenny Kitzke (990205.1600 EST)]

<Bruce Gregory (990205.1322 EST)>

<Since these systems perceptions reference levels are at the highest
conjectured level in the hierarchy, I assume they are hard-wired
(otherwise I have no idea how they got there). If we have such a lofty
wiring maybe Ken is right after all.>

Hey Bruce! You are a funny guy some times. :sunglasses:

If I am right? Why I know I'm right! Humans are a lofty capable species
not dependent just on hard wire. You know, like scientists have discovered
TV and radio and light. :sunglasses:

Nah, no way. Ken is always wrong and Rick is always right. I just got
carried away in my imagination level; wherever that is in the HPCT
hierarchy?

But, as soon as you and Rick discover that I'm right, I am going to start
selling Bibles (possibly give them away) on my PCT business web site. It
sure would beat trying to get managers to accept Perceptual Control Theory
as a basis to operate their organizations. It has been a very disturbing
struggle so far. :sunglasses:

kenny

[From Bill Powers (990205.1539 MST)]

Bruce Gregory (990205.1322 EST)]

Since these systems perceptions reference levels are at the highest
conjectured level in the hierarchy, I assume they are hard-wired
(otherwise I have no idea how they got there). If we have such a lofty
wiring maybe Ken is right after all.

This is just another version of the "where does the highest level come
from?" problem. The highest level, since it always exists by definition,
comes from somewhere. It could be genetically coded in. It could be the
result of reorganization. It could be adopted from remembered instances of
highest-level perceptions. It could be installed by some supernatural
technician. It could be the result of natural selection.

I would prefer an answer that will apply throughout life, even before the
system concept level is reached. After all, at all times there is a highest
operating level to which a new level will eventually be added, up to some
maximum level.

My vote is for reorganization.

Best,

Bill P.

[Martin Taylor 990205 15:58]

[From Bruce Gregory (990205.1322 EST)]

Since these systems perceptions reference levels are at the highest
conjectured level in the hierarchy, I assume they are hard-wired
(otherwise I have no idea how they got there). If we have such a lofty
wiring maybe Ken is right after all.

Not at all. According to the notions of reorganization that were current
when I started on CSGnet, reorganization builds new control units as well
as reconnecting and reweighting existing ones.

A new control unit necessarily takes its inputs from somewhere and sends
its outputs somewhere. Its inputs come from perceptual signals that
are constructed by the Perceptual Input Functions (PIFs) of existing
control systems (or from the sensors), and its outputs go to the
reference input functions of existing control systems (not necessarily
the same ones as feed the PIFs) or to muscle fibres.

If the new control system does something useful (meaning that it
contributes to the stability of intrinsic variables), it is less likely
to be eliminated by subsequent reorganization than if it interferes with
the existing control hierarchy.

Forgetting for the moment the "levels" hypothesized by Powers, think of
the "degrees of freedom" problem. Imagine that there are just 3 existing
elementary control units (ECUs), which output to 3 different muscles
(never mind where their inputs come from). Now add two new ECUs that
output to the reference inputs of these three (like a small chunk of
Rick's spreadsheet demo), and that take their inputs from the perceptual
signals of the original three. Provided that the two new ones have
linearly independent PIFs, and that the original 3 ECUs were able to
control their perceptual signals, the new ones can control their
perceptual signals. So could a third new one with the same connections.

But if a fourth is added, conflict is mathematically guaranteed among
the four--not necessarily between any two, but within the group. It is
simply not possible to control four perceptual signals that independently
are derived from the same three input values.

Now think about these three original ECUs and reorganization. If it
turns out that intrinsic variables are better maintained near their
evolved reference values by adding new ECUs that take as inputs the
original three perceptual signals, and feed their outputs to the
original three reference inputs, reorganization can, and may, construct
as many as three of these higher-level ECUs. But it will not construct
a fourth--at least not unless there is some mechanism for switching one
of the four off so that the other three can do their work--because a
fourth would induce conflict, and one of the four would be reorganized
out of existence one way or another.

Now extend your thinking. Imagine that reorganization is continually
making new ECUs, linked _anywhere_ in the hierarchy (provided that
perceptual input functions are always fed from perceptual signals and
their output signals go only to existing reference inputs).

What kind of new ECUs will (a) control well, and (b) have a reasonable
probability of enhancing the stability of the intrinisic variables? Those
are the ones that will survive. Actually, the converse is an easier
question--what kind are unlikely to enhance the stability of intrinsic
variables?

One kind that is unlikely is a N+1st member of a set for which the
outputs go to the N reference inputs of the N members of a lower set.
In the absence of a switching mechanism, the N+1st forces conflict, and
at least one of the N (not necessarily the new one) will soon be
reorganized away because of the imposed conflict.

Another unlikely kind is one that connects to a lower level output
directly as well as through another ECU--i.e. A outputs to C and also
to the reference input of B which also outputs to C. This is like trying
to control at the same time both the angle of the steering wheel and
independently to placement of the car within its lane. Such a connection
is likely (though not guaranteed) to result in bad control, so
further reorganization is likely to eliminate it. Bill P has produced
good arguments that such "level-jumping" connections on the output side
are unlikely to be helpful (though on the input side, there seems to
be no _a prior_ argument against them).

Without going more deeply into it, the easiest place for a new ECU to
survive is likely to be on the top of the existing set, until a whole
"layer" has been built. Levels are easily built on top of levels, but
unless the new top-level control systems help to stabilize the intrinsic
variables, they probably will be evanescent. What I think this implies
is that we may, but probably will not, build tall thin spires. More
probably we are likely to build wide layers, that perhaps can be
identified with Bill's levels--though there may be several layers within
each level. The possibility of building tall spires does suggest a
possible reason for multiple personality syndrome.

Throughout the above, I referred to the inevitability of conflict if
N+1 ECUs at one level feed the reference inputs of N ECUs at a lower
level. But this is true only if they all try to control simultaneously.

If there is a mechanism for switching among them so that no more than N
try to control at once, there is no problem. One perception may be watched
but not controlled, while the others retain control. What we are talking
about is the shifting of attention.

Once there is a mechanism for avoiding conflict among N+1 ECUs working
through N lower level ECUs, it is possible to build N+K ECUs at the
higher level. At the lowest level, humans have no more than 100-120
individual muscles that might be moved independently, and 20-40 is
probably a more reasonable estimate, given how independently mobile our
toes and our top finger joints are. But we control far more than that
number of perceptions at any intermediate level, not all at once, but
at different times, by shifting attention to where control is required
and away from where it is not so much required at any given moment.

If this "reorganization" explanation for the top-level control systems is
valid, it also argues for why people might have very different upper
level reference levels. So long as it works for keeping the intrinsic
variables stable, anything will do, whether it be a belief in God or
simply a set of principles of fairness and tolerance, or even principles
of selfishness and dominance over other people (or species). If it ain't
broke, reorganization won't fix it (probably).

One argument in favour of this "reorganization" explanation is that young
children don't seem to have control above the lowest levels, but add
the ability to control at ever higher levels over a period of many years.

Martin

Martin

From Bruce Gregory (990206.0752 EST)]

Martin Taylor 990205 15:58

Thanks for this remarkably clear and helpful post. You cast light on a
number of issues I have been wrestling with. I really appreciate the time
and care you took to formulate it.

Bruce Gregory

[From Bill Powers (990206.1003 MST)]

Martin Taylor 990205 15:58 --

A new control unit necessarily takes its inputs from somewhere and sends
its outputs somewhere. Its inputs come from perceptual signals that
are constructed by the Perceptual Input Functions (PIFs) of existing
control systems (or from the sensors), and its outputs go to the
reference input functions of existing control systems (not necessarily
the same ones as feed the PIFs) or to muscle fibres.

etc....

Martin, this is the most lucid and complete discussion of the development
of the hierarchy to date. While it's strictly a logical development from
premises, it is simple and clear, with not much space for trickery.

I would like to suggest that there may be some special properties for
control systems which send reference signals to the same systems from which
they derive their perceptual signals. Imagination, to be specific, may work
the most like a model of the environment in that case. Just a hunch.

Thanks for a valuable post.

Best,

Bill P.

···

If the new control system does something useful (meaning that it
contributes to the stability of intrinsic variables), it is less likely
to be eliminated by subsequent reorganization than if it interferes with
the existing control hierarchy.

Forgetting for the moment the "levels" hypothesized by Powers, think of
the "degrees of freedom" problem. Imagine that there are just 3 existing
elementary control units (ECUs), which output to 3 different muscles
(never mind where their inputs come from). Now add two new ECUs that
output to the reference inputs of these three (like a small chunk of
Rick's spreadsheet demo), and that take their inputs from the perceptual
signals of the original three. Provided that the two new ones have
linearly independent PIFs, and that the original 3 ECUs were able to
control their perceptual signals, the new ones can control their
perceptual signals. So could a third new one with the same connections.

But if a fourth is added, conflict is mathematically guaranteed among
the four--not necessarily between any two, but within the group. It is
simply not possible to control four perceptual signals that independently
are derived from the same three input values.

Now think about these three original ECUs and reorganization. If it
turns out that intrinsic variables are better maintained near their
evolved reference values by adding new ECUs that take as inputs the
original three perceptual signals, and feed their outputs to the
original three reference inputs, reorganization can, and may, construct
as many as three of these higher-level ECUs. But it will not construct
a fourth--at least not unless there is some mechanism for switching one
of the four off so that the other three can do their work--because a
fourth would induce conflict, and one of the four would be reorganized
out of existence one way or another.

Now extend your thinking. Imagine that reorganization is continually
making new ECUs, linked _anywhere_ in the hierarchy (provided that
perceptual input functions are always fed from perceptual signals and
their output signals go only to existing reference inputs).

What kind of new ECUs will (a) control well, and (b) have a reasonable
probability of enhancing the stability of the intrinisic variables? Those
are the ones that will survive. Actually, the converse is an easier
question--what kind are unlikely to enhance the stability of intrinsic
variables?

One kind that is unlikely is a N+1st member of a set for which the
outputs go to the N reference inputs of the N members of a lower set.
In the absence of a switching mechanism, the N+1st forces conflict, and
at least one of the N (not necessarily the new one) will soon be
reorganized away because of the imposed conflict.

Another unlikely kind is one that connects to a lower level output
directly as well as through another ECU--i.e. A outputs to C and also
to the reference input of B which also outputs to C. This is like trying
to control at the same time both the angle of the steering wheel and
independently to placement of the car within its lane. Such a connection
is likely (though not guaranteed) to result in bad control, so
further reorganization is likely to eliminate it. Bill P has produced
good arguments that such "level-jumping" connections on the output side
are unlikely to be helpful (though on the input side, there seems to
be no _a prior_ argument against them).

Without going more deeply into it, the easiest place for a new ECU to
survive is likely to be on the top of the existing set, until a whole
"layer" has been built. Levels are easily built on top of levels, but
unless the new top-level control systems help to stabilize the intrinsic
variables, they probably will be evanescent. What I think this implies
is that we may, but probably will not, build tall thin spires. More
probably we are likely to build wide layers, that perhaps can be
identified with Bill's levels--though there may be several layers within
each level. The possibility of building tall spires does suggest a
possible reason for multiple personality syndrome.

Throughout the above, I referred to the inevitability of conflict if
N+1 ECUs at one level feed the reference inputs of N ECUs at a lower
level. But this is true only if they all try to control simultaneously.

If there is a mechanism for switching among them so that no more than N
try to control at once, there is no problem. One perception may be watched
but not controlled, while the others retain control. What we are talking
about is the shifting of attention.

Once there is a mechanism for avoiding conflict among N+1 ECUs working
through N lower level ECUs, it is possible to build N+K ECUs at the
higher level. At the lowest level, humans have no more than 100-120
individual muscles that might be moved independently, and 20-40 is
probably a more reasonable estimate, given how independently mobile our
toes and our top finger joints are. But we control far more than that
number of perceptions at any intermediate level, not all at once, but
at different times, by shifting attention to where control is required
and away from where it is not so much required at any given moment.

If this "reorganization" explanation for the top-level control systems is
valid, it also argues for why people might have very different upper
level reference levels. So long as it works for keeping the intrinsic
variables stable, anything will do, whether it be a belief in God or
simply a set of principles of fairness and tolerance, or even principles
of selfishness and dominance over other people (or species). If it ain't
broke, reorganization won't fix it (probably).

One argument in favour of this "reorganization" explanation is that young
children don't seem to have control above the lowest levels, but add
the ability to control at ever higher levels over a period of many years.

Martin

Martin

[From Rick Marken (90206.0955)]

Martin Taylor (990205 15:58) --

According to the notions of reorganization that were current
when I started on CSGnet, reorganization builds new control
units as well as reconnecting and reweighting existing ones...

Bruce Gregory (990206.0752 EST) --

Thanks for this remarkably clear and helpful post.

Bill Powers (990206.1003 MST) --

Martin, this is the most lucid and complete discussion of the
development of the hierarchy to date.

Allow me to add my voice to this love fest; I think it was a
really terrific post. Nice work, Martin!

By the way, I am taking Kenny Kitzke's advice today and studying
the Bible instead of PCT. I am learning a _lot_. I hope I'm
using the right version: King James. I will post what I have
learned this evening.

Best

Rick

···

--

Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken/

[From Bruce Gregory (930209.1150 EST)]

Martin Taylor 990205 15:58

If this "reorganization" explanation for the top-level
control systems is
valid, it also argues for why people might have very different upper
level reference levels. So long as it works for keeping the intrinsic
variables stable, anything will do, whether it be a belief in God or
simply a set of principles of fairness and tolerance, or even
principles
of selfishness and dominance over other people (or species).
If it ain't
broke, reorganization won't fix it (probably).

I have had more than one occasion over the past several days to recall
this pithy observation. It certainly suggests reasons for what otherwise
appears to make no sense. I think I'll post it on my wall. Thanks
again, Martin .

Bruce Gregory

[Bruce Gregory (990211.1010 EST)]

Martin Taylor 990205 15:58

Without going more deeply into it, the easiest place for a new ECU to
survive is likely to be on the top of the existing set, until a whole
"layer" has been built. Levels are easily built on top of levels, but
unless the new top-level control systems help to stabilize
the intrinsic
variables, they probably will be evanescent.

To me, this suggests why "new knowledge" is often evanescent. When
confronted with information that conflicts with their beliefs, students
seem to construct an "understanding" that resembles what we would expect
from a new ECU on top of their existing set. It seems to work for the
moment but is likely to be inconsistent with existing control systems.
In order to earn a permanent place in the hierarchy, "true"
reorganization (changes in the existing hierarchy) would have to occur.
Absent this reorganization, the new higher level ECU is unlikely to help
stabilize intrinsic variables and therefore does not persist.

If there is a mechanism for switching among them so that no
more than N
try to control at once, there is no problem. One perception
may be watched
but not controlled, while the others retain control. What we
are talking
about is the shifting of attention.

I suspect this switching is the mechanism that allows inconsistent
"beliefs" to co-exist in the hierarchy.

If this "reorganization" explanation for the top-level
control systems is
valid, it also argues for why people might have very different upper
level reference levels. So long as it works for keeping the intrinsic
variables stable, anything will do, whether it be a belief in God or
simply a set of principles of fairness and tolerance, or even
principles
of selfishness and dominance over other people (or species).
If it ain't
broke, reorganization won't fix it (probably).

Yes. And this fact explains why it is almost impossible to alter
reference levels (beliefs) that are integral to system stability. Like
the impeachment trial, full of sound and fury, changing no one's mind.

One argument in favour of this "reorganization" explanation
is that young
children don't seem to have control above the lowest levels, but add
the ability to control at ever higher levels over a period of
many years.

Yup. Makes perfect sense.

Bruce Gregory

[From Kenny Kitzke (90212.0955EST)]

<From Rick Marken (90206.0955)>

<By the way, I am taking Kenny Kitzke's advice today and studying
the Bible instead of PCT. I am learning a _lot_. I hope I'm
using the right version: King James. I will post what I have
learned this evening.>

So, would you say that I got you to read the Bible? I made Rick do what I
wanted him to do? Was it by coercion or persuasion?

I'm not surprised about learning a lot. It happens to everyone who studies
the Bible, especially when they can read it with an open mind. Once I
started reading it, I could not put it down. Some passages that I have
read 50 times, I suddenly obtain a whole new idea when reading it quietly
and contemplatively.
It is pretty clear to me that after a lifetime of studying every day, I
will still not comprehend all there is there. Compared to the Bible,
understand PCT and behavior is a piece of cake.

KJV is fine. I use several because I think there are some errors in most
versions, including the KJV. I await your learning from the Bible.

best wishes,

kenny