A couple questions for the physics-adept on CSGnet

[From Rick Marken (2010.04.16.1010)]

Bill Powers (2010.04.16.0818 MDT)]

Bruce Gregory (2010.04.16.0707 EDT) --

BG: A point of clarification. Can we now assume that Rick Marken will cease
to make declarative statements that might be taken as descriptions of
directly observed processes, though they are really predictions from a
hypothetical theory? I say hypothetical because there are no quantitative
models of higher order control. The clarity would be most appreciated.

I'm not against using declarative statements that assume the truth of a
theory; I do it myself. But I try to include a reasonable number of
disclaimers, such as "according to PCT" and "The standard PCT diagram shows"
and so on. I think disclaimers help newcomers sort out what can actually be
observed from what is only imagined. They remind us, too, of what remains to
be accomplished...

Marvelous post Bill. I also always try to be clear about when I am
talking theory and when I am talking observation. This distinction is
particularly important to me since, as you know, I am a big fan of
observation (and theory -- modeling -- too, but what's cool about the
theory for me is how well if predicts what is observed). But to the
extent that I have failed in this regard (making clear the distinction
between theory --like "reference signal" -- and observation -- like
"stability factor" -- I will try to improve. In order to improve I
have to be able to see where I have failed. So I would appreciate it
if BG would point to the places where I have particularly egregiously
made "declarative statements that might be taken as descriptions of
directly observed processes" -- as he did when describing QM and the 2
slit experiment.

Best

Rick

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[From Bruce Gregory (2010.04.16.1844 EDT)]

[From Bill Powers (2010.04.16.0818 MDT)]

Instead of defending quantum mechanics, why not just take it for what it
is, an attempt to make sense of observations by imagining processes and
entities that we can’t observe? Wavefronts, photons, probabilities,
entanglements, and all the rest of such paraphernalia of quantum (and
standard) mechanics are inventions of the human imagination, and are not
observable. Like electrons, protons, neutrons, neutrinos, and quarks,
they belong to a model of that which we can’t observe. If reality
contained such things and all the proposed relationships among them, then
it would necessarily react to our experimental probes by making our
perceptions change in the specific ways that we observe. But while 2 + 2
= 4, it also is true that 3 + 1 = 4, so the same predicted outcome can
occur in ways different from what we imagine.

BG: I am not sure where you got the idea that I am defending quantum mechanics. I have been trying, unsuccessfully it seems, to explain where certain ideas such as the wave-particle duality arose from attempts to make sense of the successful predictions of quantum mechanics. The history of quantum mechanics contains examples of apparently disparate approaches, e.g., Schroedinger’s wave mechanics and Heisenberg’s matrix mechanics, provide to share a common underlying mathematical structure.

Maybe the world of the small does work according to the baroque ad-hoc
rules of quantum mechanics. But it’s also possible that the way it works
has a far simpler explanation that we just haven’t stumbled across yet.

BG: I don’t know what you mean by “baroque ad-hoc rules” Quantum mechanics, at least in Feynman’s formulation, is remarkably straight forward. It is difficult for me to imagine a “far simpler explanation.”

I
remember that my friend Walter Weller, while a graduate student at the
Dearborn Observatory, used the new computer in the back room, an IBM 650,
to compute the positions of planets in the sky using epicycles, and they
worked very nicely (though Wally incurred Dr. Hynek’s wrath). All those
crystalline spheres spinning within spheres were equivalent to a Fourier
series. We know now, as Ptolemy didn’t, that any waveform however complex
can be expressed as the sum of a series of sine and cosine waves, and
that is what Wally did. Ptolemy was basically right, although the
imagined crystalline spheres didn’t have counterparts in reality. The
explanation we accept today, which has pretty much been proven by sending
spacecraft and people where theory says the planets and satellites are,
remains a theory, but hardly ever requires a disclaimer, except perhaps
“at velocities small compared with the velocity of light.” And
it’s a whole lot simpler than the crystalline spheres idea, with its
plenitude of unexplained properties.
It’s always been something of a disappointment to me that the observable
aspects of PCT get less attention than the theoretical aspects. I have
suggested, for example, that there are 11 levels of perception and
control which are related to one another in very specific,
observable, ways. I found them mostly by observing, not
theorizing. Even though it took me decades to discover those eleven (nine
without anyone else’s help), I’m sure they can’t be exactly right or in
exactly the right order or the only levels there are. If other people
would examine these proposals and compare them with their own
perceptions, some of the defects and lacks might be remedied and we could
be more sure that these are the right levels, and real, meaning common to
all of us. We have ways of setting up models and testing them, at least
for one level at a time, and with enough people doing the investigating I
think we could really get somewhere. But how many people are looking at
their own perceptions with the intent of checking out my descriptions?
And how many have just thanked me politely for the nice theory and
adopted it as a belief system?

BG: As I look out my window I see trees, a marsh, boats at a marina, and hills beyond. I don’t perceive anything corresponding to the upper levels of the hierarchy as you describe it. I am not suggesting that your description lacks merit, just that it does not match my experience.

The observed levels have nothing to do with PCT, which is an attempt to
explain how such observations might be accounted for by a model of what
we (carefully) imagine goes on inside the brain.

BG: I find that statement very strange. If the higher levels have nothing to do with PCT, why are we discussing them? Considering my inability to perceive them, it would be a lot simpler for me to ignore them.

Fortunately, I’m now working with Henry Yin, a real neuroscientist, and
there is a good chance we will actually find things inside brains that
both account for what we observe in direct experience and correct the
interpretations that are off the mark. I doubt that this task will be
completed by only two of the hundreds of thousands of people who might
also be looking into this matter, but maybe we can get the process
started.

Bruce, every theory, including quantum mechanics and PCT, should be
assumed wrong until proven right. We shouldn’t accept any theory until we
have run out of ways to prove that it can’t be true. It’s trying
systematically, experimentally, to prove the theory wrong that teaches us
what its strengths are, and weeds out the nonsense (and in some cases, as
with phlogiston, the entire theory)

BG: Bill, I completely agree with that claim.

Bruce

[From Bruce Gregory (2010.04.18.1846 EDT)]

[From Rick Marken (2010.04.16.1010)]

So I would appreciate it
if BG would point to the places where I have particularly egregiously
made “declarative statements that might be taken as descriptions of
directly observed processes” – as he did when describing QM and the 2
slit experiment.

BG: I seem to remember you claiming that tracking experiments involve everything we need to model human behavior. If I am recalling incorrectly feel free to set the record straight.

Bruce

[From Rick Marken (2010.04.16.1645)]

Bruce Gregory (2010.04.18.1846 EDT)]

Rick Marken (2010.04.16.1010)]

So I would appreciate it
if BG would point to the places where I have particularly egregiously
made "declarative statements that might be taken as descriptions of
directly observed processes" -- as he did when describing QM and the 2
slit experiment.

BG: I seem to remember you claiming that tracking experiments involve
everything we need to model human behavior. If I am recalling incorrectly
feel free to set the record straight.

I think the issue is whether I do what you did, which is to talk about
_theoretical_ processes as though they were observable processes. I
have no idea what the claim that you attribute to me has to do with
that issue. But I can't imagine having ever made the claim (that
"tracking experiments involve everything we need to model human
behavior") anyway. It just makes no sense.

What I might have said is that the basic tracking experiment is a
nice, clear example of the _observable_ phenomenon of control; all the
essential elements of control are clearly visible. Since controlling
is what people do, tracking experiments are a nice way to study human
behavior. In a tracking experiment, for example, the controlled
variable as well as the actions and disturbances that affect it can be
readily monitored (observed) and measured.

The control we see in the tracking task is not theoretical; it is
observable. The theory of how this controlling happens is theoretical;
the perceptual function, perceptual signal, comparator, reference
signal, error signal and output function -- all theory.

The first chapter of _Mind Readings_ is a reprint of a paper called
"The Nature of Behavior: Control as Fact and Theory". I think it makes
a pretty good distinction between the phenomenon of control that we
observe and the theory of control (control theory) we use to explain
that phenomenon.

I try not to make the mistake of conflating observation and theory but
if I ever do please feel free to point it out, just as Bill and I feel
free to point out when you do it.

Best

Rick

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[From Bill Powers (2010.04.16.1710 MDT)]

Bruce Gregory (2010.04.16.1844 EDT) --

BG: I don't know what you mean by "baroque ad-hoc rules" Quantum mechanics, at least in Feynman's formulation, is remarkably straight forward. It is difficult for me to imagine a "far simpler explanation."

Here is what Einstein had to say about quantum mechanics:

···

============================================================================

"On the basis of quantum theory there was obtained a surprisingly good representation of an immense variety of facts which otherwise appeared entirely incomprehensible. But on one point, curiously enough, there was failure: it proved impossible to associate with these Schrodinger waves definite motions of the mass points - and that, after all, had been the original purpose of the whole construction. The difficulty appeared insurmountable until it was overcome by Born in a way as simple as it was unexpected. The de Broglie-Schrodinger wave fields were not to be interpreted as a mathematical description of how an event actually takes place in time and space, though, of course, they have reference to such an event. Rather they are a mathematical description of what we can actually know about the system. They serve only to make statistical statements and predictions of the results of all measurements which we can carry out upon the system. (Albert Einstein, 1940)

That's the only position that makes sense to me. Quantum mechanics -- like all of physics -- is about human perceptions, not Reality. I had a vague and uninformed premonition about this when I took undergraduate courses in physics. It made me uneasy; it was as if someone was expressing a point of view in great detail without realizing that it was something separate from what was being viewed and interpreted. They seemed to be saying that because we can't measure position or momentum without disturbing them, there is something uncertain about them when they're not being disturbed.

This is what PCT is going to bring to physics, ultimately. Maybe -- I hope this is actually possible. If we can take into account the properties of the levels of perception, when we finally figure out what they really are, we can try to factor them out of our observations of physical phenomena, so we can at last discover what we are seeing that is attributable to the world outside, and what is being added and filtered by the world inside. The whole argument about whether uncertainty is a feature of reality or of the observer shows that we have not yet found a way to do this.

BG: As I look out my window I see trees, a marsh, boats at a marina, and hills beyond. I don't perceive anything corresponding to the upper levels of the hierarchy as you describe it.

Tree, marsh, and boat are names of configurations. "Trees" and '"boats," however, are categories as they include very different configurations under the same name. "Beyond" is a relationship. Are the trees stationary, or bending back and forth as if in a wind? (transitions) Are they farther away than the boats, the same distance, or closer? (relationship) Are the boats in the marina all pointing in the same direction? (relationship)Is anything happening in this scene -- that is, does anyone park a car, does a cloud appear or dissolve, or does a boat come to a dock? (event) Do you ever think things like "If it looks any stormier in the next few minutes, I'm going to have to go home and shut the windows." You know, logical reasoning, plans.

I am not suggesting that your description lacks merit, just that it does not match my experience.

That's only because there are parts of your experience that you exempt from the label "perception", that you don't even think of as perceptions. You perceive a relationship between my levels and your experience that you call a "mismatch," but you say you don't perceive relationships. A specific mismatch is a relationship, the term "mismatch" is the name of a category of similar relationship perceptions. "The upper levels of your hierarchy" is a generalization, I would say a principle-level perception, and "your hierarchy" suggests a relationship between me and the hierarchy.

Maybe you're unaware of the experiences to which all these words are pointing, but that would surprise me. Most people, I think, experience more than the words; if asked, they can point to where they see the thing in question, and even give details about it (lower-order perceptions). At higher levels, they can point to examples of what they mean. I assume that when you referred to the trees, boats, and hills you could see out your window, you were either experiencing them in real time, or imagining/remembering them -- I assume that a boat is more than b-o-a-t to you.

The observed levels have nothing to do with PCT, which is an attempt to explain how such observations might be accounted for by a model of what we (carefully) imagine goes on inside the brain.

BG: I find that statement very strange. If the higher levels have nothing to do with PCT, why are we discussing them?

Because they are observations that require an explanation, not theories offered as explanations. PCT is a body of theories. That's what the T stands for.

Considering my inability to perceive them, it would be a lot simpler for me to ignore them.

When you say you can't perceive principles, for example, you're telling me that you don't know what the word principle means -- it's just a shape or a sound that you have learned to write or say in the appropriate sentences, without any meaning except perhaps, by substitution, other words. You're saying you don't know what the principle of least effort means, or the principle of tit for tat, or of conservation of energy. That is, you may admit to using those words, but they don't stand for anything but the words. If someone demonstrated the principle of returning good for evil (without using the words), you wouldn't recognize it. You're saying you can't find square roots using Newton's Method, or do long division, or find the average of a list of numbers, or tell when a person is being cautious.

I can believe you when you say you don't have vivid three-dimensional pictures in color and sound that you call principles. Those attributes are lower-level perceptions. But when you say that concepts like relationship or category or sequence or logic or principle aren't in your perceptions, I think the real problem is how you categorize your experiences, not any inability to experience things at what I call the higher levels.

If you want to understand my levels, you have to start by deciding that everything you experience, without exception, is a perception. Everything. When you think, that's a perception. When you are startled, that's a perception. When you doubt something, that's a perception. No exceptions. That's the only way to avoid setting some perceptions aside and thinking of them as "just the way things are," and thus missing their significance in the hierarchy entirely.

I understand that this is difficult. Why do you think it took me thirteen years to come up with only four more levels after the five in the 1960 paper, and ten years to add two more? I spent a lot of time on this and missed many things that were right under my nose. I'm probably still missing obvious things. But I haven't assumed that there's nothing there to find, as you appear to be proposing.

Fortunately, I'm now working with Henry Yin, a real neuroscientist, and there is a good chance we will actually find things inside brains that both account for what we observe in direct experience and correct the interpretations that are off the mark. I doubt that this task will be completed by only two of the hundreds of thousands of people who might also be looking into this matter, but maybe we can get the process started.

Bruce, every theory, including quantum mechanics and PCT, should be assumed wrong until proven right. We shouldn't accept any theory until we have run out of ways to prove that it can't be true. It's trying systematically, experimentally, to prove the theory wrong that teaches us what its strengths are, and weeds out the nonsense (and in some cases, as with phlogiston, the entire theory)

BG: Bill, I completely agree with that claim.

So we're not permanently separated by an impenetrable barrier. Our difficulties are largely semantic. This has most often been the case, ever since PCT was known as "feedback theory" and had five fuzzily-defined levels. As you found out in your work at the Smithsonian on how people learn scientific concepts, nobody comes to PCT without a theory of behavior already in place. To learn PCT is to unlearn everything incompatible with it. The unlearning is the hardest part; that's why people without a lot of training find little difficulty in understanding PCT: half the job is already done. It's hardest for highly trained people from other fields in the life sciences. Neuroscience, for example, is solidly stuck in S-R theory and Skinnerian behaviorism and they are spinning their wheels deeper and deeper into the mud as they extend this kind of interpretation to everything they find. This is a great pity because in neuroscience there are incredibly talented people doing experiments one would have thought are physically impossible to do. The wrong theoretical superstructure can ruin such an endeavor, leading it down one blind alley after another. I fervently hope that when neuroscience decides to see what is down our alley, it will not turn out to be another futile excursion.

Best,

Bill P.

[From Bruce Gregory (2010.04.16.2202 EDT)]

[From Rick Marken (2010.04.16.1645)]

Bruce Gregory (2010.04.18.1846 EDT)]

Rick Marken (2010.04.16.1010)]

So I would appreciate it
if BG would point to the places where I have particularly egregiously
made "declarative statements that might be taken as descriptions of
directly observed processes" -- as he did when describing QM and the 2
slit experiment.

BG: I seem to remember you claiming that tracking experiments involve
everything we need to model human behavior. If I am recalling incorrectly
feel free to set the record straight.

I think the issue is whether I do what you did, which is to talk about
_theoretical_ processes as though they were observable processes. I
have no idea what the claim that you attribute to me has to do with
that issue. But I can't imagine having ever made the claim (that
"tracking experiments involve everything we need to model human
behavior") anyway. It just makes no sense.

BG: I couldn't agree more.

What I might have said is that the basic tracking experiment is a
nice, clear example of the _observable_ phenomenon of control; all the
essential elements of control are clearly visible. Since controlling
is what people do, tracking experiments are a nice way to study human
behavior. In a tracking experiment, for example, the controlled
variable as well as the actions and disturbances that affect it can be
readily monitored (observed) and measured.

BG: Of course. But as Bill points out, we cannot rule out the possibility that another model could equally well explain the behavior in the same way that something "simpler" than quantum mechanics could explain the two-slit experiment.

The control we see in the tracking task is not theoretical; it is
observable. The theory of how this controlling happens is theoretical;
the perceptual function, perceptual signal, comparator, reference
signal, error signal and output function -- all theory.

BG: Just to be a little more precise, the behavior in the tracking is observable. The claim that it represents control is not observable. It is based on a theory.

The first chapter of _Mind Readings_ is a reprint of a paper called
"The Nature of Behavior: Control as Fact and Theory". I think it makes
a pretty good distinction between the phenomenon of control that we
observe and the theory of control (control theory) we use to explain
that phenomenon.

I try not to make the mistake of conflating observation and theory but
if I ever do please feel free to point it out, just as Bill and I feel
free to point out when you do it.

BG: It would help me if you would point out were I conflated observation and theory when I wrote:

BG earlier: The "act of observation" enters the picture because of the probability amplitudes responsible for the interference. We never observe these amplitudes but only individual electrons or photons. So long as an irreversible process has not occurred, the probability amplitudes can produce interference. When an irreversible process occurs the probability amplitude "collapses" and a particle is observed in some given state. We go through all this because we cannot say that the electron is in given state prior to the observation. If we make this assumption, the statistics comes out wrong. So in certain experiments we say that an electron is in in a "superposition of states", e.g. spin up and spin down. When we observe the electron, we always find it either spin up or spin down. The problem arises when we say that the detector is also a quantum system. The theory then predicts that the detector is in a superposition of states corresponding to the superposition of states of the electron. Since this is not what we observe, one "solution" is to say that consciousness collapses the state vector. Schroedinger did not think much of this "solution", hence the infamous Schroedinger's cat thought experiment. According to the orthodox interpretation of quantum mechanics the cat is in a superposition of live/dead states until the box is opened and the cat is observed. Schroedinger used this example to point out the absurd consequences of the orthodox interpretation. The orthodox, however, did not find the argument to be absurd. Thus we still have consciousness playing a role in some interpretations of quantum mechanics.

Thanks.

Bruce

[From Bill Powers (2010.04.16.2032 MDT)]

Bruce Gregory (2010.04.16.2202 EDT) --

BG: Just to be a little more precise, the behavior in the tracking is observable. The claim that it represents control is not observable. It is based on a theory.

I have to disagree with that. Control is a very precisely defined observational phenomena. It is visible from outside the organism we guess to be doing the controlling.

To identify control, you must find a controlled variable. A controlled variable is a physical variable, in the space common to us all, which is maintained constant, or in a constant pattern of change, against disturbances. If you can't find any variable fitting this description, the process terminates here.

Since there are many possible reasons for a variable to remain constant against disturbances (it might be part of a system too massive to be altered detectable by an applied disturbance), we have to add some more conditions.

First, we must know that when no other variable forces (influences) are applied to the variable under examination, the disturbance we are using will change the variable in a consistent and predictable way.

Second, we must demonstrate that any failure of the disturbance to have its full predicted effect is due to the presence of another identifiable force or influence that tends to disturb the same variable in a way nearly equal and opposite to the effect of the disturbance we are applying.

These steps will give us a fairly high probability that some control system is present that is keeping the variable nearly constant. If no such source can be found, we have not found the control system and can stop here.

We find the source of the opposition by tracing the opposing force to its source, in some other system. This would be an output of a control system if the rest of it can be shown to exist.

Finally, we need to find out how that other system is sensing the state of the controlled variable (the assumption that some system is sensing it is a form of proof that I can't remember the name of). If no such system can be found, we have to conclude that we haven't proven the existence of a control system or explained the negatively-aimed relationship of the output we found to the disturbance being applied.

We can find the means of sensing by interrupting paths that could be used for sensing. When all modalities of detection have been tested, and interrupting any one or more of them does not cause the opposing output to cease, again we have the Scottish Verdict, not proven. If the opposite does cease, we can move step by step along the discovered path to the sensor and identify it.

If we reach this final point, we have proven that negative feedback control exists, and we have located the region in which the system doing the controlling probably exists. We will have done this entirely by making observations of the physical environment outside any potential control system. No theory is involved, only empirical experimentation.

The theory comes in when we now ask what happens between the sensor and the output that we have found, to generate the control phenomenon we have observed. That is where PCT and the models we use come in. We don't need PCT to discover that something is being controlled, and to locate the system doing the controlling. We do need PCT to explain how that system does it.

Best,

Bill P.

···

>
> The first chapter of _Mind Readings_ is a reprint of a paper called
> "The Nature of Behavior: Control as Fact and Theory". I think it makes
> a pretty good distinction between the phenomenon of control that we
> observe and the theory of control (control theory) we use to explain
> that phenomenon.
>
> I try not to make the mistake of conflating observation and theory but
> if I ever do please feel free to point it out, just as Bill and I feel
> free to point out when you do it.

BG: It would help me if you would point out were I conflated observation and theory when I wrote:

BG earlier: The "act of observation" enters the picture because of the probability amplitudes responsible for the interference. We never observe these amplitudes but only individual electrons or photons. So long as an irreversible process has not occurred, the probability amplitudes can produce interference. When an irreversible process occurs the probability amplitude "collapses" and a particle is observed in some given state. We go through all this because we cannot say that the electron is in given state prior to the observation. If we make this assumption, the statistics comes out wrong. So in certain experiments we say that an electron is in in a "superposition of states", e.g. spin up and spin down. When we observe the electron, we always find it either spin up or spin down. The problem arises when we say that the detector is also a quantum system. The theory then predicts that the detector is in a superposition of states corresponding to the superposition of states of the electron. Since this is not what we observe, one "solution" is to say that consciousness collapses the state vector. Schroedinger did not think much of this "solution", hence the infamous Schroedinger's cat thought experiment. According to the orthodox interpretation of quantum mechanics the cat is in a superposition of live/dead states until the box is opened and the cat is observed. Schroedinger used this example to point out the absurd consequences of the orthodox interpretation. The orthodox, however, did not find the argument to be absurd. Thus we still have consciousness playing a role in some interpretations of quantum mechanics.

Thanks.

Bruce

[From Bill Powers (2010.04.16.2110 MDT)]

Bruce Gregory (2010.04.16.2202 EDT)

BP: I'm usurping Rick Marken's replies, but I'm sure he will have similar comments to offer.

BG: It would help me if you would point out were I conflated observation and theory when I wrote:

BG earlier: The "act of observation" enters the picture because of the probability amplitudes responsible for the interference. We never observe these amplitudes but only individual electrons or photons.

BP: We do not observe individual electrons or photons. We infer their existence from the measurements we do make (like where we can see a bright spot on a screen) and their theoretical relationship to those imagined entities.

BG earlier: So long as an irreversible process has not occurred, the probability amplitudes can produce interference. When an irreversible process occurs the probability amplitude "collapses" and a particle is observed in some given state.

We do not observe anything listed in those sentences. We don't observe probability amplitudes, interference, "collapse" of anything, particles, or states of particles. All those things are related through a theory, or many theories, to what we do observe with our eyes and instruments.

BG earlier: We go through all this because we cannot say that the electron is in given state prior to the observation.

We do not observe electrons or states of electrons before or after an observation. Those imaginary entities are related through a theory to what we do observe.

I think that's enough. The whole paragraph is full of imaginary unobservable entities, described as if we could observe them. Perhaps this makes clearer what I mean by "observing." I mean, basically, sense data, not inference. Every inference rests on a theory, and doesn't count as an observation. Observations are what we can directly experience happening (and note that this does not mean "what is really happening").

BG earlier: If we make this assumption, the statistics comes out wrong. So in certain experiments we say that an electron is in in a "superposition of states", e.g. spin up and spin down. When we observe the electron, we always find it either spin up or spin down. The problem arises when we say that the detector is also a quantum system. The theory then predicts that the detector is in a superposition of states corresponding to the superposition of states of the electron. Since this is not what we observe, one "solution" is to say that consciousness collapses the state vector. Schroedinger did not think much of this "solution", hence the infamous Schroedinger's cat thought experiment. According to the orthodox interpretation of quantum mechanics the cat is in a superposition of live/dead states until the box is opened and the cat is observed. Schroedinger used this example to point out the absurd consequences of the orthodox interpretation. The orthodox, however, did not find the argument to be absurd. Thus we still have consciousness playing a role in some interpretations of quantum mechanics.

I couldn't have organized this response without the excellent organization of your comments.

Best,

Bill P.

[From Rick Marken (2010.04.16.2040)]

Bill Powers (2010.04.16.2032 MDT)--

>Bruce Gregory (2010.04.16.2202 EDT) --

BG: Just to be a little more precise, the behavior in the tracking is
observable. The claim that it represents control is not observable. It is
based on a theory.

BP: I have to disagree with that. Control is a very precisely defined
observational phenomena. It is visible from outside the organism we guess to
be doing the controlling...

Thanks. Saved me a lot of work. You might also have pointed out that
your latest (and greatest) book on control theory is subtitled "The
Fact of Control". Control theory exists to explain the fact of
control, just as QM exists to explain the fact of interference
patterns that occur when single electrons are shot through two
adjacent slits at a photographic plate.

So the _fact_ of control is not something we (PCTers) have just
noticed. We've been saying this ("control is a fact") for a long time.
Indeed, the reason psychology (and the other life sciences) have not
gotten interested in PCT (an explanation of the phenomenon of control)
is because psychologists don't know that control is a fact and the
behavior of living system is an example of this fact. Psychologists
think that the fact to be explained is the observed (and apparently
causal) relationship between environmental and behavioral variables
(S-R relationships). And they see control theory as one possible
explanation of _that_ fact. But control theory explains both observed
S-R relationships (as resistance to disturbance of a controlled
variable -- the "behavioral illusion") and control (which conventional
psychological theories don't try to explain because conventional
psychologists don't know that control exists).

Understanding that control is a fact is essential to understanding why
PCT is such an important (and revolutionary) theory of living systems.
It's a theory that explains a fact about living systems that is still
unknown to all but a _very_ small group of behavioral scientists: the
fact of control.

Best

Rick

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[From Bruce Gregory (2010.04.17.0646 EDT)]

[From Bill Powers (2010.04.16.1710 MDT)]

Bruce Gregory (2010.04.16.1844 EDT) --

BG: I don't know what you mean by "baroque ad-hoc rules" Quantum mechanics, at least in Feynman's formulation, is remarkably straight forward. It is difficult for me to imagine a "far simpler explanation."

Here is what Einstein had to say about quantum mechanics:

Quantum Physics: Niels Bohr: Explaining Bohr's Model of the Atom. Niels (Neils) Bohr Biography, Quotes, Pictures

"On the basis of quantum theory there was obtained a surprisingly good representation of an immense variety of facts which otherwise appeared entirely incomprehensible. But on one point, curiously enough, there was failure: it proved impossible to associate with these Schrodinger waves definite motions of the mass points - and that, after all, had been the original purpose of the whole construction. The difficulty appeared insurmountable until it was overcome by Born in a way as simple as it was unexpected. The de Broglie-Schrodinger wave fields were not to be interpreted as a mathematical description of how an event actually takes place in time and space, though, of course, they have reference to such an event. Rather they are a mathematical description of what we can actually know about the system. They serve only to make statistical statements and predictions of the results of all measurements which we can carry out upon the system. (Albert Einstein, 1940)

That's the only position that makes sense to me. Quantum mechanics -- like all of physics -- is about human perceptions, not Reality. I had a vague and uninformed premonition about this when I took undergraduate courses in physics. It made me uneasy; it was as if someone was expressing a point of view in great detail without realizing that it was something separate from what was being viewed and interpreted. They seemed to be saying that because we can't measure position or momentum without disturbing them, there is something uncertain about them when they're not being disturbed.

BG: Not quite. They are saying that if we assume that electrons have a definite position and momentum, even when we do not observe them we are unable to make sense of the patterns in the the two-slit experiment. If we close one slit, we observe a pattern. If we close the other slit we observe a similar pattern. In each case we believe the electrons passed through the open slit. Since this passage is not observed, it is an assumption of the theory. When we open both slits, however, we observe a pattern that is not the sum of the patterns produced when each slit alone was open. We observe what we call interference, and interference is a phenomenon associated with waves not with particles. It matters how many paths an electron might have taken even though the electron is always observed as an extremely-small well-defined particle.

This is what PCT is going to bring to physics, ultimately. Maybe -- I hope this is actually possible. If we can take into account the properties of the levels of perception, when we finally figure out what they really are, we can try to factor them out of our observations of physical phenomena, so we can at last discover what we are seeing that is attributable to the world outside, and what is being added and filtered by the world inside. The whole argument about whether uncertainty is a feature of reality or of the observer shows that we have not yet found a way to do this.

BG: I am afraid you have lost me there. There is no need to invoke an observer in Feynman's description of the two-slit experiment. There is nothing subjective about the predictions or the observations. They are as Real as anything gets.

If you want to understand my levels, you have to start by deciding that everything you experience, without exception, is a perception. Everything. When you think, that's a perception. When you are startled, that's a perception. When you doubt something, that's a perception. No exceptions. That's the only way to avoid setting some perceptions aside and thinking of them as "just the way things are," and thus missing their significance in the hierarchy entirely.

BG: Fine, so long as it is understood that this is part of the theory, not an observation. In PCT, anything that is experienced is treated as a perception. In addition there are numerous things that I am unaware of but could become aware of that are also perceptions, such as the way I am moving the wheel to keep the car in its lane. So in PCT all experiences both potential and actual are perceptions. This extends to imaginary objects and situations as well. God is a perception, as is the possibility of life on Mars. I can see why you are a dualist in the tradition of Descartes.

I understand that this is difficult. Why do you think it took me thirteen years to come up with only four more levels after the five in the 1960 paper, and ten years to add two more? I spent a lot of time on this and missed many things that were right under my nose. I'm probably still missing obvious things. But I haven't assumed that there's nothing there to find, as you appear to be proposing.

BG: I wasn't aware that I was proposing any such thing. What did I say that led you to that conclusion?

So we're not permanently separated by an impenetrable barrier. Our difficulties are largely semantic. This has most often been the case, ever since PCT was known as "feedback theory" and had five fuzzily-defined levels. As you found out in your work at the Smithsonian on how people learn scientific concepts, nobody comes to PCT without a theory of behavior already in place. To learn PCT is to unlearn everything incompatible with it. The unlearning is the hardest part; that's why people without a lot of training find little difficulty in understanding PCT: half the job is already done. It's hardest for highly trained people from other fields in the life sciences. Neuroscience, for example, is solidly stuck in S-R theory and Skinnerian behaviorism and they are spinning their wheels deeper and deeper into the mud as they extend this kind of interpretation to everything they find. This is a great pity because in neuroscience there are incredibly talented people doing experiments one would have thought are physically impossible to do. The wrong theoretical superstructure can ruin such an endeavor, leading it down one blind alley after another. I fervently hope that when neuroscience decides to see what is down our alley, it will not turn out to be another futile excursion.

BG: I don't see that the task is all that difficult. Everything that I would call a signal in the brain, you would call a perception. At least as far as I can tell.

Bruce

[From Bruce Gregory (2010.04.17.0703 EDT)]

[From Bill Powers (2010.04.16.2032 MDT)]

Bruce Gregory (2010.04.16.2202 EDT) –

BG: Just to be a little more precise, the behavior in the tracking is observable. The claim that it represents control is not observable. It is based on a theory.

I have to disagree with that. Control is a very precisely defined observational phenomena. It is visible from outside the organism we guess to be doing the controlling.

BG: O.K. So the term control applies only to observed behavior. When I add a column of figures in my head, I am not engaging in control. Control is only involved when I write down the result or tell you the result.

To identify control, you must find a controlled variable. A controlled variable is a physical variable, in the space common to us all, which is maintained constant, or in a constant pattern of change, against disturbances. If you can’t find any variable fitting this description, the process terminates here.

BG: Again, thought is not controlled behavior.

Since there are many possible reasons for a variable to remain constant against disturbances (it might be part of a system too massive to be altered detectable by an applied disturbance), we have to add some more conditions.

First, we must know that when no other variable forces (influences) are applied to the variable under examination, the disturbance we are using will change the variable in a consistent and predictable way.

Second, we must demonstrate that any failure of the disturbance to have its full predicted effect is due to the presence of another identifiable force or influence that tends to disturb the same variable in a way nearly equal and opposite to the effect of the disturbance we are applying.

These steps will give us a fairly high probability that some control system is present that is keeping the variable nearly constant. If no such source can be found, we have not found the control system and can stop here.

We find the source of the opposition by tracing the opposing force to its source, in some other system. This would be an output of a control system if the rest of it can be shown to exist.

Finally, we need to find out how that other system is sensing the state of the controlled variable (the assumption that some system is sensing it is a form of proof that I can’t remember the name of). If no such system can be found, we have to conclude that we haven’t proven the existence of a control system or explained the negatively-aimed relationship of the output we found to the disturbance being applied.

BG: From this I conclude that “control in imagination” is a theoretical construct and may always remain so.

We can find the means of sensing by interrupting paths that could be used for sensing. When all modalities of detection have been tested, and interrupting any one or more of them does not cause the opposing output to cease, again we have the Scottish Verdict, not proven. If the opposite does cease, we can move step by step along the discovered path to the sensor and identify it.

If we reach this final point, we have proven that negative feedback control exists, and we have located the region in which the system doing the controlling probably exists. We will have done this entirely by making observations of the physical environment outside any potential control system. No theory is involved, only empirical experimentation.

BG: Again, mental activity is not behavior so whether or not control is involved in mental behavior is purely conjectural. In fact, mental behavior is more or less an inconvenient fact that we do not have to worry about as far as PCT is concerned.

The theory comes in when we now ask what happens between the sensor and the output that we have found, to generate the control phenomenon we have observed. That is where PCT and the models we use come in. We don’t need PCT to discover that something is being controlled, and to locate the system doing the controlling. We do need PCT to explain how that system does it.

BG: PCT describes the behavior of humans and “zombies” (beings with no interior life) in exactly the same way. I understand.

Bruce

[From Bruce Gregory (2010.04.17.0805 EDT)]

[From Bill Powers (2010.04.16.2110 MDT)]

Bruce Gregory (2010.04.16.2202 EDT)

BP: We do not observe individual electrons or photons. We infer their existence from the measurements we do make (like where we can see a bright spot on a screen) and their theoretical relationship to those imagined entities.

BG earlier: So long as an irreversible process has not occurred, the probability amplitudes can produce interference. When an irreversible process occurs the probability amplitude “collapses” and a particle is observed in some given state.

We do not observe anything listed in those sentences. We don’t observe probability amplitudes, interference, “collapse” of anything, particles, or states of particles. All those things are related through a theory, or many theories, to what we do observe with our eyes and instruments.

BG earlier: We go through all this because we cannot say that the electron is in given state prior to the observation.

We do not observe electrons or states of electrons before or after an observation. Those imaginary entities are related through a theory to what we do observe.

BG: It would seem that light is also an imaginary entity, while seeing is a perception. Beyond our perceptions, everything else is imaginary. Is that fair?

I think that’s enough. The whole paragraph is full of imaginary unobservable entities, described as if we could observe them. Perhaps this makes clearer what I mean by “observing.” I mean, basically, sense data, not inference. Every inference rests on a theory, and doesn’t count as an observation. Observations are what we can directly experience happening (and note that this does not mean “what is really happening”).

BG: Again, I suspect Descartes would be very happy with this characterization. As far as I am concerned, your thoughts and emotions are imaginary unobservable entities. It sounds almost like Skinner updated to include control circuitry.

Bruce

[From Bruce Gregory (2010.04.17.0931 EDT)]

[From Rick Marken (2010.04.16.2040)]

So the fact of control is not something we (PCTers) have just
noticed. We’ve been saying this (“control is a fact”) for a long time.
Indeed, the reason psychology (and the other life sciences) have not
gotten interested in PCT (an explanation of the phenomenon of control)
is because psychologists don’t know that control is a fact and the
behavior of living system is an example of this fact. Psychologists
think that the fact to be explained is the observed (and apparently
causal) relationship between environmental and behavioral variables
(S-R relationships). And they see control theory as one possible
explanation of that fact. But control theory explains both observed
S-R relationships (as resistance to disturbance of a controlled
variable – the “behavioral illusion”) and control (which conventional
psychological theories don’t try to explain because conventional
psychologists don’t know that control exists).

Understanding that control is a fact is essential to understanding why
PCT is such an important (and revolutionary) theory of living systems.
It’s a theory that explains a fact about living systems that is still
unknown to all but a very small group of behavioral scientists: the
fact of control.

BG: One possibility that you seem to give short shrift is that there may be many questions about behavior that the control perspective does little to illuminate. PCT focusses on how living systems achieve their goals. It lacks convincing models of why individuals have particular goals in particular circumstances. Basically PCT seems to punt these questions by invoking upper levels in a hierarchy. There are claims about these higher levels, but little or no evidence that they exist or how they work. When this shortcoming is remedied, psychologists may be more interested in think in PCT terms.

While thinking of humans in terms of their goals can very helpful, the mechanisms by which they attempt to achieve those goals is often less so. Controlling a perception of having more money may well describe a Wall Street Banker’s motivation, but does not illuminate the workings of a collateralized debt obligation in any way that I can see. But perhaps I am just missing something both obvious and important.

Bruce

[From Rick Marken (2010.04.17.1120)]

Bruce Gregory (2010.04.17.0931 EDT)--

Rick Marken (2010.04.16.2040)--

Understanding that control is a fact is essential to understanding why
PCT is such an important (and revolutionary) theory of living systems.
It's a theory that explains a fact about living systems that is still
unknown to all but a _very_ small group of behavioral scientists: the
fact of control.

BG: One possibility that you seem to give short shrift is that there may be
many questions about behavior that the control perspective does little to
illuminate.

Well, I'll extend my shrift a bit and say that there may, indeed, be
many questions about behavior that the control perspective doesn't
illuminate. I just haven't heard them. You got any?

PCT focusses on _how_ living systems achieve their goals.

That's too informal for me. I would say that PCT explains the what,
how, why of the controlling done by living systems.

It [PCT] lacks convincing models of _why_ individuals have particular goals in
particular circumstances.

I think of a "convincing" model as one that accounts for data better
than other models or that is the only currently known model that
accounts for the data. So is there some data you have in mind that is
related to "why individuals have particular goals in particular
circumstances" that PCT can't account for? If so I'd really like to
see it.

Basically PCT seems to punt these questions by
invoking upper levels in a hierarchy. There are claims about these higher
levels, but little or no evidence that they exist or how they work.

Well, maybe just a "little" evidence but certainly not "no" evidence.
There are some tests of multi-level models at
Levels of Control,
http://www.mindreadings.com/ControlDemo/HP.html and
http://www.mindreadings.com/Coordination.html.

When this shortcoming is remedied, psychologists may be more interested
in think in PCT terms.

I doubt it. Psychologists will not take PCT seriously (let alone
understand it) until they discover the phenomenon that PCT explains:
control.

While thinking of humans in terms of their goals can very helpful, the
mechanisms by which they attempt to achieve those goals is often less so.

Thinking of humans in terms of goals (references for the state of
perceptual representations of environmental variables) is not
"helpful"; It's the only currently known way we can account for what
we see people doing: controlling. Scientists have to learn to see that
behavior _is_ control before they can really appreciate PCT. I
recommend my paper "Looking at behavior through control theory
glasses" (APA PsycNet) to see that
the first step in understanding control theory is understanding the
phenomenon it explains: control. In PCT we're very big on the idea
that _phenomena come phirst_.

Controlling a perception of having more money may well describe a Wall
Street Banker's motivation, but does not illuminate the workings of a
collateralized debt obligation in any way that I can see. But perhaps I am
just missing something both obvious and important.

I think you are missing the fact that "making money" is an example of
control while a collateralized debt obligation is something that
simply exists in the environment. The development (and use) of such
obligations is, however, an example of control.

Best

Rick

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[From Bruce Gregory (2010.04.17.1528 EDT)]

[From Rick Marken (2010.04.17.1120)]

Thinking of humans in terms of goals (references for the state of
perceptual representations of environmental variables) is not
“helpful”; It’s the only currently known way we can account for what
we see people doing: controlling. Scientists have to learn to see that
behavior is control before they can really appreciate PCT.

BG: I think I’ve got. (1) It’s all perception. (2) It’s all control. Have I missed anything?

Bruce

[From Rick Marken (2010.04.17.1545)]

Bruce Gregory (2010.04.17.1528 EDT)--

> Rick Marken (2010.04.17.1120)]
>
> Thinking of humans in terms of goals (references for the state of
> perceptual representations of environmental variables) is not
>"helpful"; It's the only currently known way we can account for what
> we see people doing: controlling. Scientists have to learn to see that
> behavior _is_ control before they can really appreciate PCT.

BG: I think I've got. (1) It's all perception. (2) It's all control. Have I
missed anything?

Just what "it" is in these two sentences. (Hint: "It" is not the same
thing in both).

Best

Rick

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com