PCT Training & Education

[From Fred Nickols (2003.07.20.0925 EDT)] --

I like the idea of providing ways and means for people to assess their own
grasp of PCT and additional ways and means of improving said grasp of
PCT. Rick's questions are a start. If we get serious about this, I would
make one suggestion right away: There should be several forms of any
"test." One can be used as an initial assessment, a diagnostic if you
will. The results from this can point the test taker to the material from
which they might remedy any knowledge gaps. Other forms of the test or
assessment can be used to determine how well any knowledge deficiency might
have been remedied.

Another assessment issue will tie to the difference between assessing
knowledge and assessing skill or competency. The assessment instrument
might ask questions and the answers (in a multiple-choice, constructed
response or essay format) will suffice. The assessment instrument might
also present certain tasks (e.g., build a spreadsheet or Vensim or other
kind of dynamic model of some behavioral phenomenon). And, of course, the
assessment items themselves can call forth quite a bit of creativity and
imagination on the part of the test taker (e.g., one long essay item could
consist of writing rebuttals to selected criticisms of PCT).

By the way, Rick's first question intrigued me so, without benefit of
referring to any "open book" (or closed ones that I know of), here's my
response:

Q. What is a controlled variable?

A. A controlled variable can be just about anything. The term refers
to some perceived or at least perceivable aspect of our world that we wish
to control. In the course of driving to work, to draw on an often-used
example, keeping my car in whatever lane I've selected is one example of a
controlled variable, a variable that might be better named "position of
auto with respect to lane." The speed at which I'm traveling offers
another example of a controlled variable. Both of those examples are at
what many people might call the "conscious" level. There are controlled
variables of which I am not ordinarily so consciously aware. My heartbeat
and respiration rates are controlled variables but I am not aware of
controlling them and don't believe I can consciously do so (although I am
told that Yoga masters can, within limits, control both). For a more down
to earth example, I just took a sip of coffee from the mug on my
desk. That involved grasping the cup, raising it to my lips, opening my
mouth, touching it to my lips, tilting it, taking a swallow of now lukewarm
coffee and then tiling it back the other way, closing my lip and returning
the mug to its previous position on my desk. There are many "controlled
variables" in that simple act; consider just a few: the position of my
fingers in the handle of the mug, the position of the mug with respect to
my lips, the angle and rate of the tilt (up and down), and the position of
the mug on the desk upon its return. I did not think consciously about
controlling these things, I simply took a sip of coffee. Controlled
variables can also be much "bigger" as it were than taking a sip of coffee
or driving to work. As a manager or executive, I might be bent on
controlling some aspect of the performance of my organization (e.g., sales,
profit, expenses, productivity, quality, unit costs, etc.). Unlike taking
a sip of coffee or driving to work, these controlled variables are not
under my direct control; that is, my behavior does not affect them
directly. To control variables such as those just mentioned, my actions
must be viewed as interventions in some larger structure. I cannot change
expenses directly, but I can hire (or fire) people in a pretty direct
manner and, as a consequence, affect expenses. I can direct that cheaper
materials be substituted for those now used and thus affect expenses. I
can direct that prices be lowered (and hope that sales will increase). In
short, when I seek to control variables that are far removed from me in
space and time, I need a map of the architecture in which those variables
are embedded. Why? Because in order to reliably control those distant
variables, I need to know the intervening variables that must also be
controlled. Change, in the case of sipping coffee or driving to work, is
pretty direct; I grasp the cup, turn the wheel, apply the brakes, take a
sip, accelerate and slow down and put down the cup (as a matter of fact, I
can be spotted on occasion driving and drinking coffee at the same
time). But, change in controlled variables like sales and expenses is
indirect; I act in one place at one time and I intend to realize certain
effects elsewhere later on.

The preceding is a reasonable reflection of my current grasp of the meaning
of the term "controlled variable." Having written it out, it prompts some
questions of my own. Is a controlled variable anything we attempt or
intend to control or is it only those things we can in fact control? Is
there a point where disturbances are so overwhelming that any pretension at
control of a particular variable is impractical (e.g., when product
failures drive expenses through the roof, or someone jostles my elbow and
the mug winds up in a position that spills hot coffee in my lap or when
another car veers into my lane forcing me out of mine) or, do these kinds
of disturbances mean merely that I have altered my reference level for the
controlled variable in question? I guess what I'm asking is does a
controlled variable depend on the existence of control or intent to control
enough to define it as such?

Good question, Rick.

Regards,

Fred Nickols
"Assistance at a Distance"
Distance Consulting
nickols@safe-t.net
www.nickols.us

[From Rick Marken (2003.07.21.1210)]

Fred Nickols (2003.07.20.0925 EDT) --

These are great points, Fred. The test's the thing... We'll definitely discuss
this at the conference.

I'll reply to your controlled variable answer as soon as I feel a moment of calm
(ie. when the meeting is over). Thanks for giving it a try. A cursory glance
suggests that you did very well. I will just quickly say that I think of a
"controlled variable" as a variable that _is_ (not potentially or hopefully) under
control. Other than that, I think we agree. But I'll try to give it a more careful
look when I get a chance.

Best regards

Rick

···

I like the idea of providing ways and means for people to assess their own
grasp of PCT and additional ways and means of improving said grasp of
PCT. Rick's questions are a start. If we get serious about this, I would
make one suggestion right away: There should be several forms of any
"test." One can be used as an initial assessment, a diagnostic if you
will. The results from this can point the test taker to the material from
which they might remedy any knowledge gaps. Other forms of the test or
assessment can be used to determine how well any knowledge deficiency might
have been remedied.

Another assessment issue will tie to the difference between assessing
knowledge and assessing skill or competency. The assessment instrument
might ask questions and the answers (in a multiple-choice, constructed
response or essay format) will suffice. The assessment instrument might
also present certain tasks (e.g., build a spreadsheet or Vensim or other
kind of dynamic model of some behavioral phenomenon). And, of course, the
assessment items themselves can call forth quite a bit of creativity and
imagination on the part of the test taker (e.g., one long essay item could
consist of writing rebuttals to selected criticisms of PCT).

By the way, Rick's first question intrigued me so, without benefit of
referring to any "open book" (or closed ones that I know of), here's my
response:

Q. What is a controlled variable?

A. A controlled variable can be just about anything. The term refers
to some perceived or at least perceivable aspect of our world that we wish
to control. In the course of driving to work, to draw on an often-used
example, keeping my car in whatever lane I've selected is one example of a
controlled variable, a variable that might be better named "position of
auto with respect to lane." The speed at which I'm traveling offers
another example of a controlled variable. Both of those examples are at
what many people might call the "conscious" level. There are controlled
variables of which I am not ordinarily so consciously aware. My heartbeat
and respiration rates are controlled variables but I am not aware of
controlling them and don't believe I can consciously do so (although I am
told that Yoga masters can, within limits, control both). For a more down
to earth example, I just took a sip of coffee from the mug on my
desk. That involved grasping the cup, raising it to my lips, opening my
mouth, touching it to my lips, tilting it, taking a swallow of now lukewarm
coffee and then tiling it back the other way, closing my lip and returning
the mug to its previous position on my desk. There are many "controlled
variables" in that simple act; consider just a few: the position of my
fingers in the handle of the mug, the position of the mug with respect to
my lips, the angle and rate of the tilt (up and down), and the position of
the mug on the desk upon its return. I did not think consciously about
controlling these things, I simply took a sip of coffee. Controlled
variables can also be much "bigger" as it were than taking a sip of coffee
or driving to work. As a manager or executive, I might be bent on
controlling some aspect of the performance of my organization (e.g., sales,
profit, expenses, productivity, quality, unit costs, etc.). Unlike taking
a sip of coffee or driving to work, these controlled variables are not
under my direct control; that is, my behavior does not affect them
directly. To control variables such as those just mentioned, my actions
must be viewed as interventions in some larger structure. I cannot change
expenses directly, but I can hire (or fire) people in a pretty direct
manner and, as a consequence, affect expenses. I can direct that cheaper
materials be substituted for those now used and thus affect expenses. I
can direct that prices be lowered (and hope that sales will increase). In
short, when I seek to control variables that are far removed from me in
space and time, I need a map of the architecture in which those variables
are embedded. Why? Because in order to reliably control those distant
variables, I need to know the intervening variables that must also be
controlled. Change, in the case of sipping coffee or driving to work, is
pretty direct; I grasp the cup, turn the wheel, apply the brakes, take a
sip, accelerate and slow down and put down the cup (as a matter of fact, I
can be spotted on occasion driving and drinking coffee at the same
time). But, change in controlled variables like sales and expenses is
indirect; I act in one place at one time and I intend to realize certain
effects elsewhere later on.

The preceding is a reasonable reflection of my current grasp of the meaning
of the term "controlled variable." Having written it out, it prompts some
questions of my own. Is a controlled variable anything we attempt or
intend to control or is it only those things we can in fact control? Is
there a point where disturbances are so overwhelming that any pretension at
control of a particular variable is impractical (e.g., when product
failures drive expenses through the roof, or someone jostles my elbow and
the mug winds up in a position that spills hot coffee in my lap or when
another car veers into my lane forcing me out of mine) or, do these kinds
of disturbances mean merely that I have altered my reference level for the
controlled variable in question? I guess what I'm asking is does a
controlled variable depend on the existence of control or intent to control
enough to define it as such?

Good question, Rick.

Regards,

Fred Nickols
"Assistance at a Distance"
Distance Consulting
nickols@safe-t.net
www.nickols.us

--
Richard S. Marken, Ph.D.
Senior Behavioral Scientist
The RAND Corporation
PO Box 2138
1700 Main Street
Santa Monica, CA 90407-2138
Tel: 310-393-0411 x7971
Fax: 310-451-7018
E-mail: rmarken@rand.org

[From Bryan Thalhammer (2003.07.21.1855 CDT)]

Dear All,

A quote from one of the 50 I do believe: "Sounds like a really good solution
to me. --Fred Good" Hi Fred Good! I agree with Rick and Fred Nickols that
the questions (and answers) are proceeding well in their first mockup of
what such a self-test or consistency test would be like. I am sorry I could
not be in LA with the rest of you, but I assure you that with this teamwork
evidenced by what I have been reading in THIS thread, we cannot fail!!!

Cordially,

--Bryan Thalhammer
Norridge, IL (a CTA ride from O'Hare, a lot more from the point of view of
having a ticket, etc.)

···

[From Rick Marken (2003.07.21.1210)]

> Fred Nickols (2003.07.20.0925 EDT) --

These are great points, Fred. The test's the thing... We'll
definitely discuss
this at the conference.

I'll reply to your controlled variable answer as soon as I feel a
moment of calm (ie. when the meeting is over). Thanks for giving
it a try. A cursory glance suggests that you did very well.
...I think we agree. But I'll try to give it a more careful
look when I get a chance.

Best regards

Rick

>
>
> I like the idea of providing ways and means for people to
> assess their own grasp of PCT and additional ways and means
> of improving said grasp of PCT. Rick's questions are a start.
> If we get serious about this, I would make one suggestion right
> away: There should be several forms of any "test." One can
> be used as an initial assessment, a diagnostic if you will.
> The results from this can point the test taker to the
> material from which they might remedy any knowledge gaps.
> Other forms of the test or assessment can be used to determine
> how well any knowledge deficiency might have been remedied.....

[From Bruce Nevin (2003.07.28 14:38 PDT)]

I think of a "controlled variable" as a variable that _is_ (not potentially or hopefully) under control.

There's an odd wrinkle or two here to winkle out.

An observer can know that a variable is controlled only by observing resistance to a disturbance that changes the state of the variable to an undesired state.

The fact that a variable is under control is unknown and unknowable to an observer if there is no disturbance that changes it to an undesired state.

The fact that such disturbance 'elicits' a 'response' that restores the desired state is evidence that it is a controlled variable at that time.

It may also be taken as evidence that it was a controlled variable all along, during the time preceding the disturbance. But, in fact, in the absence of observed resistance to disturbance, the time at which the variable started being controlled is unknown and unknowable to an observer.

The controller can be aware of 'watching' for disturbances. An observer can note this alertness too, but attributing the cause is dicey. The controller can even be mistaken about the purpose of her alert watchfulness, and can be unaware of controlling a variable until a disturbance 'elicits' a 'response' that restores the variable to a desired state; indeed, may be up to then unaware of the desired state (reference value). But this is a matter of awareness, not of control.

The observer can take testimony from the controller about what is being controlled. But one can testify only to that of which one is aware. It follows from the above (and other considerations) that such testimony may be neither accurate nor complete. The only way to be sure that it is accurate is by observing resistance to disturbance. There is no way to be sure that it is complete. So self assessments and verbal reports are of limited value and can be misleading.

This way too fussy and detailed for the basic Q&A definitions, of course. But come to think of it maybe the basic stuff should include links that drill down to caveats and details.

         /Bruce

···

At 12:12 PM 7/21/2003, Richard Marken wrote:

[From Rick Marken (2003.07.29.0850)]

Bruce Nevin (2003.07.28 14:38 PDT)

>I think of a "controlled variable" as a variable that _is_ (not
>potentially or hopefully) under control.

There's an odd wrinkle or two here to winkle out....

These are all good points but they seem to me to go beyond the basic question:
What is a controlled variable? I think a question like this is useful to the
extent that it allows students of PCT to demonstrate that they know what we are
talking about when we discuss a "controlled variable"; that they understand, in
other words, what the term "controlled variable" means. What I would be looking
for in an answer to this question is some understanding of the idea that a
controlled variable is a _perceptual_ variable that is a function of elementary
environmental variables (such as masses, forces, positions, etc), that it is a
_variable_ aspect of the environment that is brought to and maintained in a
reference state by the actions of a control system, and that examples of such
variables are 1) the lane in which a car is traveling (which can vary from left to
right) 2) the position of the car in its lane (which can vary from left of center
lane to right of center lane), 4) the word one is saying (which can vary from
aardvark to zebra) 4) one's political position (which can vary from left of center
to right of center), and so on.

Best regards

Rick

···

At 12:12 PM 7/21/2003, Richard Marken wrote:

--
Richard S. Marken, Ph.D.
Senior Behavioral Scientist
The RAND Corporation
PO Box 2138
1700 Main Street
Santa Monica, CA 90407-2138
Tel: 310-393-0411 x7971
Fax: 310-451-7018
E-mail: rmarken@rand.org

[From Bill Powers (2003.07.28.1455 MDT)]

Just got back from the CSG meeting -- Rick, we made it to Flagstaff the
first day. Those who had to miss the annual meeting missed something
special this year -- see Rick's summary post. Dag Forssell, yet again,
recorded the whole conference on video tape so that others can attend
vicariously. DEag's efforts will become more and more valuable as time goes

I think of a "controlled variable" as a variable that _is_ (not
potentially or hopefully) under control.

There's an odd wrinkle or two here to winkle out.

An observer can know that a variable is controlled only by observing
resistance to a disturbance that changes the state of the variable to an
undesired state.

The fact that a variable is under control is unknown and unknowable to an
observer if there is no disturbance that changes it to an undesired state.

The fact that such disturbance 'elicits' a 'response' that restores the
desired state is evidence that it is a controlled variable at that time.

This is a misconception that seems to come up periodically. In fact, it is
almost never necessary to alter the controlled quantity by a significant
amount in order to do the test. The test, to be valid, must be conducted in
such a way that the controlling system maintains the controlled variable
very close to its reference level at all times. In other words, the
controlling system continues to control successfully.

An easy way to illustrate this is with the rubher bands. With person A
keeping the knot over the dot, person B varies the pull on the appropriate
end of the rubber bands If person A is controlling the position of the knot
relative to the dot, person A's end of the rubber band will move so as to
cancel the variations in person B's pull, with the result that the knot
remains over the dot. Of course person B can pull so hard and so rapidly
that the knot is caused to depart significantly from the position of the
knot, but in that case the test is invalid, because person A has lost control.

In cases involving discrete variables, some thought may be required to find
a way of disturbing the proposed controlled variable that the other person
can prevent from making any important change in that variable. Keep in mind
that the whole point of the test is to apply a disturbing influence that
the controlling system can SUCCESSFULLY counteract.

You may want to modify the development in your post in the light of this
basic correction.

Best,

Bill P.

···

on -- he is the memory of the CSG. Bruce Nevin (2003.07.28 14:38 PDT) -- >At 12:12 PM 7/21/2003, Richard Marken wrote:

[From Bruce Nevin (2003.07.28 14:38 PDT)]

Yes, the essential thing is that the disturbance should change the state
of the variable but doesn’t.

Bill Powers (2003.07.28.1455 MDT) –

Bruce Nevin (2003.07.28 14:38 PDT) –

An observer can know that a variable is
controlled only by observing

resistance to a disturbance that changes the state of the variable to
an

undesired state.

… by observing a disturbance that should change the state of the
CV but doesn’t (or doesn’t as much as it should, and even that effect is
countered).

The fact that a
variable is under control is unknown and unknowable to an

observer if there is no disturbance that changes it to an undesired
state.

… that should change it to an undesired state.

The fact that such
disturbance ‘elicits’ a ‘response’ that restores the

desired state is evidence that it is a controlled variable at that
time.

… that maintains the desired state

But the only evidence of the ‘response’ may be the failure of the
variable to be disturbed. Hence the almost irresistable desire to
strengthen the disturbance until you observe some loss of control.

Those who think of communication as disturbing others’ controlled
variables seem often to fall into this.

    /Bruce

Nevin

···

At 05:15 PM 7/29/2003, Bill Powers wrote:

[From Bill Powers I2003.07.30.1734 MDT)]

Bruce Nevin (2003.07.28 14:38 PDT)--

>Those who think of communication as disturbing others' controlled
variables seem often >to fall into this.

I assume that your meaning is not just "some people seem to feel they have
to disturb other people's controlled variables to feel they are
communicating." If that was actually your meaning the following may be of
some interest anyway.

I think we need some advice here from a language expert. The problem is
that "disturbance" has two meanings: (1) the cause of a perturbation or (2)
the perturbation itself.We need a way to distinguish, in this case, a
communication that would be the cause of a perturbation of some controlled
variable _if the person receiving it did not quickly do something to
counteract the effect_ or _keep the effect from getting any larger_.

When we say that a communication is a disturbance, we mean that there is
some perception (other than the communication itself) that would be
expected to change as a consequence of receiving the communication. To
control that perception would mean saying or doing something that would
have the opposite effect on the same perception, For example, I might say
to someone, "I think that what you just said was mistaken." If saying that
caused a perturbation of the person's perception of being right, we might
expect the person to reply with some attempt to show that the statement is
incorrect, or to produce more evidence in support of the criticized idea,
or in general to say or do something that would tend to restore the sense
of being right to its original level (the action might not even be
observable -- the person might simply dismiss the criticism, thinking
privately "He doesn't know what he's talking about").

It's possible that a simple statement of fact can be a disturbance -- that
is, a potential cause of perturbations. If I say to a psychiatrist that
behavior is the control of perception, and the sentence is correctly
understood, the psychiatrist might find the _meaning_ of that sentence far
from what he or she believes to be true. Whether there is any response to
the statement will depend on whether it has succeeded in creating any
deviation of a perception from the psychiatrist's reference levels. Oddly
enough, there will be no resistance if my statement is NOT correctly
understood, but is misinterreted as, for example, "perception controls
behavior" or somnething else the hearer agrees with.

I think that a general case can be made for saying that communications are
disturbing variables (another way to refer to the causal meaning). If a
communication leaves all of your perceptions exactly as they were before,
has anything actually been communicated? An information theorist would say
"no." But I think we can say "no" or at least "probably not" simply on the
grounds that we _expect_ to perceive something new when someone talks to
us, even if it's only "dinner is served." We wouldn't want to leave that
dinner cooling on the table, would we?

Best,

Bill P.

Bruce Nevin (2003.07.30 19:49 PDT)

Bill Powers I2003.07.30.1734 MDT)–

I think we need some advice here from a
language expert. The problem is

that “disturbance” has two meanings: (1) the cause of a
perturbation or (2)

the perturbation itself.

If we mean d in a control-loop diagram, then we mean (1) and we
have no distinct referent for (2). But d is as perceived by the
observer, not by the controller, whose control actions resist (2). The
controller can observe (notice, be aware of) only (2), not d, and
indeed easily misattributes the cause of (2) to sources other than
d.
When we talk about subjective effects, as in communication, we must
assume the point of view of the controller rather than the observer. But
the controller has no reliable point of view on disturbances d nor
even on perturbations of control (2), but only at best on outputs
(control actions and qo) and imagined attributions about d.
Wait a minute, I can feel that rock sliding under my foot. Right, and the
world is filled with myriad disturbances in the realm of billiard-ball
physics, Bateson’s Pleroma. I can track some of them, predict a few of
them, and that tells us nothing about communication.
OK then, I can see you moving that end of the rubber band, doesn’t that
mean I can observe d at least in some cases? Well, yes, as far as
that goes, but it really goes no farther than your control outputs, and
that’s on the same level of forces and impacts as that unstable stone
underfoot. What I attribute, or misattribute, as d is what I
imagine to be your controlled variables and your references for
them.

We need a way to distinguish, in this case,
a

communication that would be the cause of a perturbation of some
controlled

variable _if the person receiving it did not quickly do something to

counteract the effect_ or keep the effect from getting any
larger
.

“Quickly”, you say. How remarkably elastic this notion has
become!

Is it not, rather, that some perception is maintained in its preferred
state internally to the controller, and communication is about informing
others that this is so?

When we say that a communication is a
disturbance, we mean that there is

some perception (other than the communication itself) that would be

expected to change as a consequence of receiving the communication.
To

control that perception would mean saying or doing something that
would

have the opposite effect on the same perception, For example, I might
say

to someone, “I think that what you just said was mistaken.” If
saying that

caused a perturbation of the person’s perception of being right, we
might

expect the person to reply with some attempt to show that the statement
is

incorrect, or to produce more evidence in support of the criticized
idea,

or in general to say or do something that would tend to restore the
sense

of being right to its original level (the action might not even be

observable – the person might simply dismiss the criticism,
thinking

privately “He doesn’t know what he’s talking
about”).

In each of these cases alike there is an internal process of maintaining
control. The only difference between the last and the others is in
whether the person subsequently tells others about it.
That’s to control others’ perception of oneself, no?
So the idea that you were talking about is a perception, and the
person’s control of it was disturbed by your saying that it was mistaken.
Why do they then refute that allegation, rather than just shrugging and
saying to themselves “He doesn’t know what he’s talking about”?
The person is also controlling how they are perceived by others, and the
allegation that they are mistaken (and might be mistaken and unreliable
in other matters?) disturbs their control of their perception of other
peoples’ perception of them.

I know, that bounces back and forth a bit, like a hall of mirrors. But if
you’re going to talk about communication that’s where you have to go,
into that hall of mirrors.

I think that a general case can be made for
saying that communications are

disturbing variables (another way to refer to the causal meaning).

Does recognition disturb a controlled variable?
I think what is disturbed here is control for things recognized to fit
together with one another. Some people are less tolerant than others of
things not fitting together into a consistent whole yet.
This is an internal processing that we haven’t talked about much. I’m
sure (article of confidence in theory?) that ratiocination involves
control processes and associative memory, but I’m blatterbugged if I
could say what the variables are and how they’re connected. I am
confident that Frege was talking about laws for thought when he
wrote his book Laws of Thought. As has been said in another
connection, if we innately controlled that way there wouldn’t be any need
to prescribe it.

If a

communication leaves all of your perceptions exactly as they were
before,

has anything actually been communicated?

There’s the communication - a bit of information - and there’s what you
do with it, how you fit it into what you already know. The latter is the
governing factor. The difference between saying and telling. The bit of
information may not have been a disturbance in itself at all. A hierarchy
of recognizers recognized the words, word-dependencies, etc., recognizers
associated them with ‘concepts’ already known. Then some deductive
ramification is a disturbance - at its simplest, something believed to be
the case, and the disbelief that it is the case, juxtaposed, assertion
and denial. Maybe the conclusion pops out that solipsism follows
necessarily from a bunch of other things that were previously said, and
the new communication only reminded you of those previously recognized
notions, and without awareness your control processes and associative
memory processes had been deducing that conclusion and with the present
reminder the disturbing conclusion comes to your awareness. Did the bit
of information that someone said, which served to awaken that
awareness, actually tell the person all that, and was it really
the disturbance?

If you’re seriously interested in talking about communication, you must
first recognize how very far from rubber bands you are. I’m not saying
it’s impossibly complicated and we shouldn’t try. I’m only saying we
shouldn’t fool ourselves that it’s so simple as disturbance and
resistance to disturbance.

Bon appétit!

    /Bruce

Nevin

···

At 08:00 PM 7/30/2003, Bill Powers wrote:

[From Rick Marken (2003.07.30.2010)]

Bill Powers I2003.07.30.1734 MDT)

I think that a general case can be made for saying that communications are
disturbing variables (another way to refer to the causal meaning). If a
communication leaves all of your perceptions exactly as they were before,
has anything actually been communicated? An information theorist would say
"no."

I think you meant to say that an information theorist would say "yes". In
information theory a communication is a stimulus that reduces "uncertainty".
The measure of information reduction is proportional (base 2 logarithmically)
to the number of possible communications that might have occurred.

But I think we can say "no" or at least "probably not" simply on the
grounds that we _expect_ to perceive something new when someone talks to
us, even if it's only "dinner is served." We wouldn't want to leave that
dinner cooling on the table, would we?

Yes. As you proved long ago, there is no information in a controlled
perception.

Welcome back!

Best

Rick

···

--
Richard S. Marken
MindReadings.com
marken@mindreadings.com
310 474-0313

[From Rick Marken (2003.07.31.0850)]

Bruce Nevin (2003.07.30 19:49 PDT)

If you're seriously interested in talking about communication, you must first
recognize how very far from rubber bands you are.

If communications were really very far from rubber bands then it would be very far
from control theory. In fact, a communication -- something someone says -- is
precisely equivalent to the disturbance (d) produced by the experimenter's finger
in the rubber band demo. A communication is a disturbance if, like movement's of
the experimenter's finger, it is independent of the actions of the recipient of
the communication (who is equivalent to the subject in the rubber band demo), and
if it influences the state of a controlled variable (just as movements of the
experimenter's finger influence, via the rubber band connection, the position of
the knot in the rubber band demo). The main difference I can see between
communication and the rubber band demo is that an observer can see the controlled
variable far more readily in the latter (the knot) than the former (the controlled
perception influenced by the communication).

This leads me to suggest another question for the PCT exam: "What is a disturbance
variable"?

This question should be easy to answer based on my comments above. This is an open
post test;-)

Best regards

Rick

···

--
Richard S. Marken, Ph.D.
Senior Behavioral Scientist
The RAND Corporation
PO Box 2138
1700 Main Street
Santa Monica, CA 90407-2138
Tel: 310-393-0411 x7971
Fax: 310-451-7018
E-mail: rmarken@rand.org

[From Bill Powers (2003.07.31.0835 MDT)]

Bruce Nevin (2003.07.30 19:49 PDT)

This conversation began with your reply to a comment by Rick Marken and an
answer by me to your comment:

Bruce Nevin (2003.07.28 14:38 PDT) --

I think of a "controlled variable" as a variable that _is_ (not
potentially or hopefully) under control.

There's an odd wrinkle or two here to winkle out. An observer can know
that a variable is controlled only by observing resistance to a
disturbance that changes the state of the variable to an undesired state.
The fact that a variable is under control is unknown and unknowable to an
observer if there is no disturbance that changes it to an undesired state.
The fact that such disturbance 'elicits' a 'response' that restores the
desired state is evidence that it is a controlled variable at that time.

I wrote:

This is a misconception that seems to come up periodically. In fact, it is
almost never necessary to alter the controlled quantity by a significant
amount in order to do the test. The test, to be valid, must be conducted in
such a way that the controlling system maintains the controlled variable very
close to its reference level at all times. In other words, the controlling
system continues to control successfully.

This was followed by a series of comments on my post that I understood to
indicate your agreement with the above.

Now I'm not so sure you were agreeing, but the following seems to
illustrate the point:

Bill Powers I2003.07.30.1734 MDT)--

I think we need some advice here from a language expert. The problem is
that "disturbance" has two meanings: (1) the cause of a perturbation or (2)
the perturbation itself.

If we mean d in a control-loop diagram, then we mean (1) and we have no
distinct referent for (2). But d is as perceived by the observer, not by
the controller, whose control actions resist (2). The controller can
observe (notice, be aware of) only (2), not d, and indeed easily
misattributes the cause of (2) to sources other than d.

It would appear that you found my comment to be off the mark in some way,
and that you were supplying a correction to whatever error you detected. In
short, my communication resulted in some error for you, which you are
trying to correct by the above. This, of course, is in line with the idea
that communications are disturbances (d, as you say) which move some
perception detectably away from its reference level, calling for an action
to move it back.

Perhaps this also illustrates your statement that the controller, while
able to detect the action opposing the effect of the disturbance, may not
see exactly what variable was disturbed, and may attribute his own action
(in this case, a communication intended to correct an error) to causes
other than the actual disturbance d (here, my communication). This of
course is only a conjecture on my part, which you are in a better position
to check than I am.

Let's skip toward the end of your post to avoid getting repetitive.

I said:

If a
communication leaves all of your perceptions exactly as they were before,
has anything actually been communicated?

You said:
There's the communication - a bit of information - and there's what you do
with it, how you fit it into what you already know. The latter is the
governing factor. The difference between saying and telling. The bit of
information may not have been a disturbance in itself at all.

I think I'd count the information as a "d" if the result is that you want
to correct the change, and would not have done so if the "d" had not
occurred. I don't know what you mean by "a disturbance in itself." Is
anything a disturbance "in itself?" That is, independently of anything
being disturbed by it?

A hierarchy of recognizers recognized the words, word-dependencies, etc.,
recognizers associated them with 'concepts' already known. Then some
deductive ramification is a disturbance - at its simplest, something
believed to be the case, and the disbelief that it is the case,
juxtaposed, assertion and denial. Maybe the conclusion pops out that
solipsism follows necessarily from a bunch of other things that were
previously said, and the new communication only reminded you of those
previously recognized notions, and without awareness your control
processes and associative memory processes had been deducing that
conclusion and with the present reminder the disturbing conclusion comes
to your awareness. Did the bit of information that someone said, which
served to awaken that awareness, actually tell the person all that, and
was it really the disturbance?

I get the distinct impression here that you are attempting to refute the
idea that communications are disturbances. This tells me that the
particular statement, "communications are disturbances," is disturbing
something that you feel needs to be corrected -- and that you're attempting
to correct it. Is this impression correct?

If you're seriously interested in talking about communication, you must
first recognize how very far from rubber bands you are. I'm not saying
it's impossibly complicated and we shouldn't try. I'm only saying we
shouldn't fool ourselves that it's so simple as disturbance and resistance
to disturbance.

That, of course, tends to support my hypothesis that you are resisting some
implication of the statement that communications are disturbances. I
wouldn't claim that this proves the general case, but it certainly looks
like a positive instance. What do you think?

Best,

Bill P.

···

At 12:12 PM 7/21/2003, Richard Marken wrote:

[From Bill Powers (2003.07.31.0902 MDT)]

Rick Marken (2003.07.30.2010)--

I think you meant to say that an information theorist would say "yes". In
information theory a communication is a stimulus that reduces
"uncertainty".The measure of information reduction is proportional (base 2
logarithmically)to the number of possible communications that might have
occurred.

I meant to say "no." If a signal causes no change of information at the
destination, it carries no information. For example, if I say "My name is
Bill," and then immediately say "My name is Bill" again in exactly the same
way, no information about my name is carried by the second statement (and
perhaps not by the first if you already knew my name). I know, I don't
believe it either, but that's the idea I was conveying.

Best,

Bill P.

[Bruce Nevin (2003.07.31 9:40 PDT)]

Bill Powers (2003.07.31.0835 MDT)

[...]
This was followed by a series of comments on my post that I understood to
indicate your agreement with the above.

It was. No disturbance in what you wrote, but I wrote anyway because I had some things that I was thinking about that I wanted to add to the conversation because I thought they were germane and I thought they might interest you.

Now I'm not so sure you were agreeing, ... the following seems to
illustrate the point:

Bill Powers I2003.07.30.1734 MDT)--

I think we need some advice here from a language expert. The problem is
that "disturbance" has two meanings: (1) the cause of a perturbation or (2)
the perturbation itself.

If we mean d in a control-loop diagram, then we mean (1) and we have no
distinct referent for (2). But d is as perceived by the observer, not by
the controller, whose control actions resist (2). The controller can
observe (notice, be aware of) only (2), not d, and indeed easily
misattributes the cause of (2) to sources other than d.

It would appear that you found my comment to be off the mark in some way,
and that you were supplying a correction to whatever error you detected.

No, I was just adding to what you had said, elaborating on it. It was not due to a disturbance caused by what you said, it was a product of thought processes that have been going on in me for some time prior to this discussion, and phrased now in terms of this thread because it provides a venue for talking about it in public. The core of it is the limitation of terminology to what is in a control-system diagram, the problems of ambiguity etc. when words refer to things not in the diagram, and concern to always to verify that we're not omitting things of importance from the diagram and the model.

In
short, my communication resulted in some error for you, which you are
trying to correct by the above.

As indicated. But it wasn't your words that constitute the disturbances to variables that I am controlling in my reply.

If I walk into a room and set my cup down on a desk, the desk doesn't disturb a perception that I am controlling, rather, I perceive it as means to control certain perceptions. Just so, the text that I was replying to did not disturb any perceptions that I am controlling, rather, I perceived it as means to control perceptions in public conversation that I had been controlling in the privacy of my thought processes.

Certainly provocative things can be said and people can be provoked by them, but it is important, I think, to recognize that communication can be a dance of mutual opportunity rather than a sparring match.

Perhaps this also illustrates your statement that the controller, while
able to detect the action opposing the effect of the disturbance, may not
see exactly what variable was disturbed, and may attribute his own action
(in this case, a communication intended to correct an error) to causes
other than the actual disturbance d (here, my communication). This of
course is only a conjecture on my part, which you are in a better position
to check than I am.

It doesn't illustrate it, for reasons said, but I do agree with what you are saying. I should, since it is your paraphrase of what I said, and, being a paraphrase, indicates to me that you have understood me. The only dissention here is that B's response to what A has said is not necessarily resistance to disturbance caused by something that A has said, it may simply be B agreeing and having more to say on the subject. If I agree and say "Yes" I am controlling some variables such as wanting you to know that I agree, but I am not resisting any disturbance caused by what you have said.

I said:

If a
communication leaves all of your perceptions exactly as they were before,
has anything actually been communicated?

You said:
There's the communication - a bit of information - and there's what you do
with it, how you fit it into what you already know. The latter is the
governing factor. The difference between saying and telling. The bit of
information may not have been a disturbance in itself at all.

I think I'd count the information as a "d" if the result is that you want
to correct the change, and would not have done so if the "d" had not
occurred.

You are assuming that the purpose of the response is to correct the bit of information. That's a possible communicative intention, but not a necessary one.

I don't know what you mean by "a disturbance in itself."

All I meant was that the bit of information was not a disturbance. The reason for responding was not a desire to correct that bit of information, but rather to add to it, to elaborate on it, to tell you something that it reminds me of, etc.

Is
anything a disturbance "in itself?" That is, independently of anything
being disturbed by it?

Nothing technical or particularly astute was intended by this phrase.

I get the distinct impression here that you are attempting to refute the
idea that communications are disturbances. This tells me that the
particular statement, "communications are disturbances," is disturbing
something that you feel needs to be corrected -- and that you're attempting
to correct it. Is this impression correct?

Yes, except that I am saying that communications are not always and only disturbances. Certainly mutually provocative exchanges occur.

What I am controlling here is a disparity between what I intended and (my understanding of) what you understood. That is indeed a disturbance to my perception that you have understood what I said, and I am responding in this message to resist that disturbance. But in the prior message, which this is about, I was elaborating on what you had said, with no intention of disagreeing with it, only adding more to it in what I thought was the same vein.

In the course of doing this I have asserted that communication is not limited to tit-for-tat provocation (disturbance and resistance to disturbance which in turn causes disturbance). This assertion seems to disturb some variable that you are controlling. I am guessing that it suggests that communication does not have as simple an explanation in the model as you thought. You might think that I am saying that the model is inadequate to explain it and needs extending in some way, and if you are I could well understand why you would resist that. But in fact I think all of this can be accounted for within the model without extension. It's a bit of a mystery what perceptions we are controlling when we are 'thinking', particularly the kind of back-burner thinking that comes to the fore when something reminds us, and the only means for controlling them that has been proposed is the imagination connection. We already knew that this is largely uncharted territory. Maybe that in itself is a disturbance. I shouldn't be surprised if it were. But these disturbances, if such indeed they are, are not due to what I said, but rather by what you have made of it. So you have an example before you as intimately near as is possible.

If not, then please help me out and tell me what variables you find disturbed by what I have said, how your response has resisted the disturbance, and how my words are the source of disturbance. I very much would like to know, because I really prefer more cooperative modes of communication.

         /Bruce Nevin

···

At 11:02 AM 7/31/2003, Bill Powers wrote:

[From Rick Marken (2003.07.31.1400)]

Bill Powers (2003.07.31.1258 MDT)--

Argument: an extended mutual disturbance applied through the medium of words.

I think you could define a conversation in exactly the same way:

Conversation: an extended mutual disturbance applied through the medium of words.

I think the difference between what we call a "conversation" and what we call an
"argument" is only in the gain with which the participants are controlling for the
variables (ideas) being discussed.

Best

Rick

···

--
Richard S. Marken, Ph.D.
Senior Behavioral Scientist
The RAND Corporation
PO Box 2138
1700 Main Street
Santa Monica, CA 90407-2138
Tel: 310-393-0411 x7971
Fax: 310-451-7018
E-mail: rmarken@rand.org

[From Rick Marken (2003.07.31.1405)]

Bruce Nevin (2003.07.31 12:50 PDT)

Rick Marken (2003.07.31.0850)--

>If communications were really very far from rubber bands then it would be
>very far from control theory. In fact, a communication -- something
>someone says -- is precisely equivalent to the disturbance (d) produced by
>the experimenter's finger in the rubber band demo.

I hope that my conversation with Bill on this topic has made it clearer to
you what I meant, and why this may be true some of the time but is not a
necessary attribute of communication all of the time.

Not really. When is a communication not a disturbance?

I should be explicit about what I'm aware of controlling here. I really
don't have any response to what you've said, I've already said it, but I'm
controlling a perception (in imagination) of you feeling annoyed that you
said something to me and I didn't answer you.

I was not annoyed at all. Don't worry. I really don't get annoyed at all when
people don't answer me.

Best

Rick

···

--
Richard S. Marken, Ph.D.
Senior Behavioral Scientist
The RAND Corporation
PO Box 2138
1700 Main Street
Santa Monica, CA 90407-2138
Tel: 310-393-0411 x7971
Fax: 310-451-7018
E-mail: rmarken@rand.org

[From Rick Marken (2003.07.31.1445)]

[From Peter J. Burke UCR 7/31/2003 2:07PM PST]

Rick Marken (2003.07.31.1400)
Conversation: an extended mutual disturbance applied through the medium of
words.
I think the difference between what we call a "conversation" and what we
call an "argument" is only in the gain with which the participants are
controlling for the variables (ideas) being discussed.

I think it is important to remember that there are hundreds of controlled
variables in a conversation. As Bruce pointed out in his presentation at the
conference, people control their sounds, as well as their words, and I would
add, the meanings that are produced. The multiple meanings serve to both
create "differences" as well as to reduce "differences" for the perceiver,
across all of the meanings that are perceived.

I agree.

It's important to remember that a "disturbance" is just a variable (such as a
statement in conversation) that influences the state of a controlled perception.
As in a tracking task (or the rubber band demo), a disturbance can push the
controlled variable toward or away from its reference state. What is important
about a disturbance is that it is an influence on a controlled variable that is
_independent_ of the actions of the controlling agent. As Bill noted, a
disturbance does not necessarily have an undesirable influence on the controlled
variable. If someone came to my door and gave me $100,000 cash, that would be
disturbance to one of my controlled variables (my cash reserves) but it would be a
disturbance that pushes that variable in a desirable direction (toward my
reference for cash reserves). If, during a conversation, someone told me that
they loved my "PCT Glasses" paper, that would also be a disturbance, one that
pushes another controlled variable (my perception of how much people like my work)
toward my reference for that variable.

Best regards

Rick

···

--
Richard S. Marken, Ph.D.
Senior Behavioral Scientist
The RAND Corporation
PO Box 2138
1700 Main Street
Santa Monica, CA 90407-2138
Tel: 310-393-0411 x7971
Fax: 310-451-7018
E-mail: rmarken@rand.org

[From Peter J. Burke UCR 7/31/2003 11:56AM PST]

   [From Rick Marken (2003.07.31.0850)]
     If communications were really very far from rubber bands then it would
be very far
     from control theory. In fact, a communication -- something someone
says -- is
     precisely equivalent to the disturbance (d) produced by the
experimenter's finger
     in the rubber band demo. A communication is a disturbance if, like
movement's of
     the experimenter's finger, it is independent of the actions of the
recipient of
     the communication (who is equivalent to the subject in the rubber band
demo), and
     if it influences the state of a controlled variable (just as movements
of the
     experimenter's finger influence, via the rubber band connection, the
position of
     the knot in the rubber band demo).

Communication is an interesting one because it serves both as disturbance
and often as a counter-disturbance. Many of the communications we receive
have with others serve to verify our identities, which is to say they help
bring our perceptions of who we are into alignment with our reference of who
we are. At the same time, some disturbances are also introduced because in
one communication so many meanings are altered in the situation. This makes
it very difficult to track exactly what is happening in social interaction
and communication.

Peter
Peter J. Burke
Professor and Chair
Department of Sociology
University of California
Riverside, CA 92521-0419
Phone: 909/827-7198
Fax: 909/787-0333
peter.burke@ucr.edu
http://wat1203.ucr.edu

[From Bill Powers (2003.07.31.1258 MDT)]

Bruce Nevin (2003.07.31 9:40 PDT)--

Looks as though we're converging. I should have said, of course, that a
communication is a disturbance if and only if the person receiving it feels
moved to do or say something to correct any perceptions it has affected. It
is certainly not the case that all communications are disturbances all of
the time. I understand that back-and-forth conversation can serve to build
up concepts on both sides, and is not usually intended to cause trouble for
the other person.

There may have been an unintended implication that disturbances are always
undesirable. If I help you carry your suitcase, the disturbance of my added
lift is prevented from making the suitcase accelerate upward by the
lessening of your effort. Since my intention was to lessen your effort, and
presumably you are not sorry to be carrying less weight, the disturbance
and the adjustment that counteracts it's effect are quite benign, If you
are struggling to find the right word, and I communicate it to you, your
struggle can cease and your error is corrected. Disturbances are simply
external influences that act on the same variables you are controlling, so
they may either aid or oppose your own efforts. If you maintain control
they are neutral in value, as neutral as a crosswind that springs up -- in
any direction -- while you're steering around a curve.

Argument: an extended mutual disturbance applied through the medium of words.

Best,

Bill P.

[From Bruce Nevin (2003.07.31 12:41 PDT)]

Bill Powers (2003.07.31.1258 MDT)--

Looks as though we're converging.

Agreed. I'm glad because there does seem to me to have been a commonly made assumption on the net that conversational exchanges can only proceed by disturbance and resistance to disturbance, which in turn constitute a disturbance.

There may have been an unintended implication that disturbances are always
undesirable.

An excellent point! The technical terms are unhappily tendentious. A disturbance must be disturbing. An error must be wrong. I have thought for some time that it might be wiser to refer to the delta between perceptual input and reference input as delta rather than error. The term is commonly enough used now to be readily understood. Maybe there's a similarly neutral term for disturbance. Influence?

         /Bruce Nevin

···

At 03:10 PM 7/31/2003, Bill Powers wrote: