Control of behavior

CSGnet Archive CSG1996
21

From [ Marc Abrams (980830.1241) ]

I am using Word97. and Outlook Express in Windows. Clicking
on the attached word doc freezes my computer, without
opening word. When I copy the file to disk & then open the
file I have no problems. Very Strange. First time this has
happened to me.

Marc

[From Rick Marken (980830.0910)]

Tim Carey (980830.1505) --

If this doesn't work I'll try something else.

I am using Word 6. This time I saw the text description
of the variables but no diagram. Sorry. I have no
idea what the problem is. How about Ascii?

Best

Rick
--

Richard S. Marken Phone or Fax: 310

474-0313

Life Learning Associates e-mail:

rmarken@earthlink.net

···

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

22

From [Bruce Gregory (980830.1340 EDT)]

Marc Abrams (980830.1241)

I am using Word97. and Outlook Express in Windows. Clicking
on the attached word doc freezes my computer, without
opening word. When I copy the file to disk & then open the
file I have no problems. Very Strange. First time this has
happened to me.

I am using Word97 and Outlook in Windows95. No problemo. I'm suspicious of
Outlook Express.

Bruce Gregory

p.s. Nice job Tim.

23

[From Rick Marken (980830.1220)]

Bruce Gregory (980829.2120 EDT)--

Can you give me an example of something that can be controlled
"arbitrarily"?

The location of a rock can be controlled arbirarily in the sense
that I can put the rock anywhere I want and it will stay there;
my control of this variable will not be hindered by any of _the
rock's_ intentions because the rock has no intentions. The
location of the subject's finger in the finger tracking demo
_cannot_ be controlled arbitrarily; my control of this variable
can be lost when the subject's intention changes or when my
disturbances force the subject to do something that conflicts
with an existing intention.

Bruce Gregory (980829.2130 EDT)--

So you are saying that you can _try_ to control another person,
but you will only succeed if the other person is controlling
perceptions compatible with your desires or if you are willing
to use unlimited force.

In the kind of control we are discussing (control via disturbance
to a controlled variable), yes. In other kinds of control of
behavior (such as control through the use of force) this
consideration is irrelvant.

···

-----

I picked up "Making Sense of Behavior" yesterday afternoon and
randomly picked, as a starting point, page 92 or so. Coincidentally,
this turns out to be in the middle of the chapter on "Conflict
Between People" in which Bill discusses control of behavior (in
those words, for those who seem to think that talking about
"control of behavior" violates some tenet of PCT). I recommend
it to everyone who is interested in this issue of "control of
behavior", especially the discussion from p. 98 on.

While reading it, I had an idea about why we might be having
all these problems in the discussion of control of behavior.
The problem seems to center around the idea of whether living
control systems _can_ be controlled. Tim Carey and Bruce Gregory
seem to think that living control systems _cannot_ be controlled
and that PCT shows that this is the case. I think that living
control systems _can_ be controlled and that PCT shows that this
is the case.

The idea I had while reading Bill's chapter on "controlling
people" was this:

The message of PCT is not: Behavior _cannot_ be controlled.
The message of PCT is: Human attempts to control human behavior
are the primary cause of conflict between humans.

So whether behavior can or cannot be controlled is really
irrelevant. The fact is that behavior can be controlled but
it's not easy to do, it doesn't work forever and, most importantly,
it creates interpersonal conflict. So I don't feel that
uncomfortable about people saying behavior can't be controlled;
there is certainly a sense in which it can't. What I do think is
important, however,is recognizing that people _do_ act as controllers
of behavior.

I think some of the arguments against the idea that people can be
controlled have included the correllary conclusion that, therefore,
people cannot act as controllers of other people; the teacher is
not a controller if the kid wants to do what the teacher wants, for
example. This is an idea that I find not only incompatible with
PCT but also inconsistent with the message of PCT regarding control
of behavior; that human attempts to control human behavior are
the primary cause of conflict between humans. How could human
attempts to control behavior be a problem if people can't
control behavior? I think PCT explains how.

I think there is no question that people do act to control other
people's behavior; whether these other's people's behavior is
_really_ controllable is not important. But I think it's important
to recognize that other people _do_ act as controllers of other
people's behavior, whether we believe this controlling could be
successful or not. Because it is only by recognizing when people
(others and ourselves) are acting as controllers of human behavior
that we can consider _other_ ways of dealing with human behavior --
other than controlling it.

So I suggest directing the discussion of "control of behavior"
away from the question of whether behavior can or can't be
controlled and towards the question of how we know when people
(including ourselves) are controlling other people's behavior
(or whatever).

Best

Rick
--

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

24

[From Tim Carey (980831.0600)]

[From Rick Marken (980830.1220)]

While reading it, I had an idea about why we might be having
all these problems in the discussion of control of behavior.
The problem seems to center around the idea of whether living
control systems _can_ be controlled. Tim Carey and Bruce Gregory
seem to think that living control systems _cannot_ be controlled
and that PCT shows that this is the case. I think that living
control systems _can_ be controlled and that PCT shows that this
is the case.

For me, Rick the problem is not whether or not a living system can or can't
be controlled but _what_ is actually being controlled. I think we have
already demonstrated that it is the _actions_ of living control systems
that can be externally controlled. Unless I've misinterpreted the meaning
of behaviour in PCT terms, I don't think control of actions equates to
control of behaviour. In fact, the reason I think it's possible to control
someone else's actions (at least for a little while) is because this is one
aspect of the individual's behaviour that they are _not_ controlling.

The message of PCT is not: Behavior _cannot_ be controlled.
The message of PCT is: Human attempts to control human behavior
are the primary cause of conflict between humans.

Agreed.

So whether behavior can or cannot be controlled is really
irrelevant. The fact is that behavior can be controlled but
it's not easy to do, it doesn't work forever and, most importantly,
it creates interpersonal conflict.

Here is where I have the problem. When I read "behaviour", I interpret that
to mean: the working of the entire loop, that is acting, perceiving and
comparing. Bill goes over this in chapter one of his new book (at least he
does in the old prepublished copy I have ... I don't have the new one yet).
The only one of these three things that I think can be externally
controlled is the acting part.

I think some of the arguments against the idea that people can be
controlled have included the correllary conclusion that, therefore,
people cannot act as controllers of other people; the teacher is
not a controller if the kid wants to do what the teacher wants, for
example.

I think the _only_ thing people are is controllers.

I think there is no question that people do act to control other
people's behavior; whether these other's people's behavior is
_really_ controllable is not important. But I think it's important
to recognize that other people _do_ act as controllers of other
people's behavior, whether we believe this controlling could be
successful or not.

I think people act as controllers of other people's actions. I think people
aren't really concerned about the rest of other people's behaviour (i.e.,
comparing and perceiving) all the care about is the actions of the other
person, probably because this is the aspect of the other person's behaviour
that _they_ can perceive.

If you're meaning "actions" when you use the term "behaviour" then I agree
completely. If you're meaning "acting, perceiving, and comparing" then I
disagree completely :slight_smile:

Regards,

Tim

25

[From Tim Carey (980831.0630)]

[From Rick Marken (980830.0910)]

I am using Word 6. This time I saw the text description
of the variables but no diagram. Sorry. I have no
idea what the problem is. How about Ascii?

I can't find Ascii. I'll try Rich Text Format.

Regards,

Tim

PCT.rtf (56 Bytes)

26

[From Bruce Gregory (980830.1645 EDT)]

Rick Marken (980830.1220)

I think there is no question that people do act to control other
people's behavior; whether these other's people's behavior is
_really_ controllable is not important. But I think it's important
to recognize that other people _do_ act as controllers of other
people's behavior, whether we believe this controlling could be
successful or not. Because it is only by recognizing when people
(others and ourselves) are acting as controllers of human behavior
that we can consider _other_ ways of dealing with human behavior --
other than controlling it.

Fascinating idea. Since we are essentially controllers, our options seem
rather limited. We can attempt to control covertly (manipulation); we can
refuse to develop reference levels for the perceived behavior of others
(indifference or sainthood); or we can set our reference levels in ways that
do not conflict with their (permissiveness). Any others?

So I suggest directing the discussion of "control of behavior"
away from the question of whether behavior can or can't be
controlled and towards the question of how we know when people
(including ourselves) are controlling other people's behavior
(or whatever).

Whenever we or others have a reference level for the behavior of others we
are likely to attempt to control their behavior.

Bruce Gregory

27

[From Bruce Gregory (980830.1655 EDT)]

Tim Carey (980831.0630)

> [From Rick Marken (980830.0910)]

> I am using Word 6. This time I saw the text description
> of the variables but no diagram. Sorry. I have no
> idea what the problem is. How about ASCII?

I can't find ASCII. I'll try Rich Text Format.

ASCII is txt or DOS txt. You will loose your very nice diagram.

Bruce Gregory

28

[From Bill Powers (980830.1555 MDT)]

Tim Carey (980830.0940)--

Again, my problem is with your use of the term "behaviour". If you are
meaning "externally observable output" or "actions" then I don't have a
problem with that. If you're meaning the "working of the entire loop" then
I don't understand that at all.

I agree. "The working of the entire loop" doesn't refer to anything,
because each part of the loop does something different. The loop maintains
the perception near the reference signal by acting on the environmental
input quantity that is being perceived, opposing disturbances at the same
time. If anyone refers to this whole process with the word "behavior" why
can't we just assume that the person doesn't want to get any more specific
and let it go at that? I may say "I'm surprised at your behavior" without
meaning to get technical about just what it is I'm surprised at.

Best,

Bill P.

29

[From Rick Marken (980830.1550)]

Tim Carey (980831.0630) --

Let's try another approach on the diagram. Maybe the problem
is your mailer's encoding system. Since Bruce Gregory got the
diagram OK maybe he could post them. OK, Bruce?

I think we're making some real progress on this control of
behavior issue.

Tim Carey (980831.0600)--

I think we have already demonstrated that it is the _actions_
of living control systems that can be externally controlled.

I thought we have also demonstrated that _controlled variables_
can also be controlled. If we haven't, I will soon.

the reason I think it's possible to control someone else's
actions (at least for a little while) is because this is one
aspect of the individual's behaviour that they are _not_
controlling.

Actions are not controlled from the perspective of the control
system that is _varying_ those actions as the means of controlling
it's perception. But actions are often themselves controlled
variables. For example, in the finger tracking demo, the subject's
finger position is an action that is varied to control the distance
between fingers. But finger position is also a controlled variable;
you can demonstrate this by pushing gently on the subject's
hand while she is moving it; the subject will push back against
your push in order to protect finger position from the disturbance.

Here is where I have the problem. When I read "behaviour", I
interpret that to mean: the working of the entire loop, that
is acting, perceiving and comparing. Bill goes over this in
chapter one of his new book

I don't think Bill ever says (or intends to say) that behavior is
the working of the entire loop. Behavior is what we see people
doing. PCT tells us that some of these doings are actions aimed at
producing intended results, some are intended results themselves
and many are both. Take a behavior like "swimming" for example;
this behavior involves actions (arm, leg and head movements) and
intended results (crawl, butterfly, backstroke). In fact, the
actions are also intended results (produced by muscle tensions)
and the intended results are also actions (aimed at controlling
for winning the race, impressing the girls, etc).

I think a good way to translate the phrase "behavior is the
control of perception" is as follows: the observable behavior
of a living control system is a visible side effect of a control
process that is aimed at controlling the system's private,
internal representations (perceptions) of the world in which
its behavior occurs.

I think people act as controllers of other people's actions.

OK. That's progress. Now, when you give a child a choice of
eating spinich or brocoli are you controlling the child's
actions or its controlled variables? Is eating stuff an action
or an intended result or both (hint: both)? Is eating a behavior
(hint: yes) ?

Me:

Because it is only by recognizing when people (others and
ourselves) are acting as controllers of human behavior
that we can consider _other_ ways of dealing with human
behavior -- other than controlling it.

Bruce Gregory (980830.1645 EDT)

Fascinating idea. Since we are essentially controllers, our options
seem rather limited.

Yes!! That's the problem! And you give a good account of the
options:

We can attempt to control covertly (manipulation); we can refuse
to develop reference levels for the perceived behavior of others
(indifference or sainthood); or we can set our reference levels
in ways that do not conflict with their (permissiveness). Any
others?

Yes. A very important one. Negotiation. It's the option that Bill
discusses in MSB (Making Sense of Behavior).

Whenever we or others have a reference level for the behavior
of others we are likely to attempt to control their behavior.

Right! (Even better without the "likely"). When we have a
reference for the behavior of others _and_ we are organized to
control our perception of that behavior then we are controlling
their behavior. As you (I think) said it's probably impossible
to not have references for how people should behave. The trick
(and it's a trick that would be worth figuring out) is to lose
the references or keep them and lower the gain on the behavior
control systems.

Best

Rick

···

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

30

[From Bruce Gregory (980830.1935 EDT)]

Rick Marken (980830.1550)

Tim Carey (980831.0630) --

Let's try another approach on the diagram. Maybe the problem
is your mailer's encoding system. Since Bruce Gregory got the
diagram OK maybe he could post them. OK, Bruce?

Attached is Tim's document. It is a two page Word 6/95 document.

Bruce Gregory

PCT diagram 2A.doc (67 Bytes)

31

[From Bruce Gregory (980830.1935 EDT)]

Rick Marken (980830.1550)
>
> Tim Carey (980831.0630) --
>
> Let's try another approach on the diagram. Maybe the problem
> is your mailer's encoding system. Since Bruce Gregory got the
> diagram OK maybe he could post them. OK, Bruce?

Attached is Tim's document. It is a two page Word 6/95 document.

Interesting. The diagram was lost! I'll try forwarding Tim's original
diagram, also two pages.

Bruce Gregory

PCT diagram 21.doc (67 Bytes)

32

[From Bruce Gregory (980830.2005)]

Rick Marken (980830.1550)

Yes. A very important one. Negotiation. It's the option that Bill
discusses in MSB (Making Sense of Behavior).

The negotiation between the classroom teacher and the student with the plan
in RTP seems to me to qualify as an attempt in this direction.

> Whenever we or others have a reference level for the behavior
> of others we are likely to attempt to control their behavior.

Right! (Even better without the "likely"). When we have a
reference for the behavior of others _and_ we are organized to
control our perception of that behavior then we are controlling
their behavior. As you (I think) said it's probably impossible
to not have references for how people should behave. The trick
(and it's a trick that would be worth figuring out) is to lose
the references or keep them and lower the gain on the behavior
control systems.

My wild speculation is that control at the program level, and possibly
above, involves just such alterations of the gain of control systems at
lower levels. I'll have to wait until the course in modeling is over before
I tackle this, however.

Bruce Gregory

33

[From Tim Carey (980831.2005)]

[From Rick Marken (980830.1550)]

> I think we have already demonstrated that it is the _actions_
> of living control systems that can be externally controlled.

I thought we have also demonstrated that _controlled variables_
can also be controlled. If we haven't, I will soon.

Hmmm. I'm not sure what you mean on this one. Who is contolling the
controlled variable?

Actions are not controlled from the perspective of the control
system that is _varying_ those actions as the means of controlling
it's perception. But actions are often themselves controlled
variables. For example, in the finger tracking demo, the subject's
finger position is an action that is varied to control the distance
between fingers. But finger position is also a controlled variable;
you can demonstrate this by pushing gently on the subject's
hand while she is moving it; the subject will push back against
your push in order to protect finger position from the disturbance.

How do you separate this from the cv of "distance between fingers"? Isn't
the reason they push back because you are disturbing the distance cv? Is
there anyway you can disturb the position cv and not disturb the distance
cv?

I don't think Bill ever says (or intends to say) that behavior is
the working of the entire loop. Behavior is what we see people
doing. PCT tells us that some of these doings are actions aimed at
producing intended results, some are intended results themselves
and many are both. Take a behavior like "swimming" for example;
this behavior involves actions (arm, leg and head movements) and
intended results (crawl, butterfly, backstroke). In fact, the
actions are also intended results (produced by muscle tensions)
and the intended results are also actions (aimed at controlling
for winning the race, impressing the girls, etc).

So since actions are both actions and intended results are you saying that
"behaviour = actions" in PCT talk? I'm actually a bit confused by the
second sentence. How can we observe an intended result? Wouldn't you just
be making a guess (perhaps an educated guess, but a guess nevertheless)
that the particular actions you see are in fact also intended results?

I think a good way to translate the phrase "behavior is the
control of perception" is as follows: the observable behavior
of a living control system is a visible side effect of a control
process that is aimed at controlling the system's private,
internal representations (perceptions) of the world in which
its behavior occurs.

OK, so in Bill's new book where he talks about acting, perceiving, and
comparing, am I correct in saying that in PCT the term "behaviour" only
refers to the acting component.

OK. That's progress. Now, when you give a child a choice of
eating spinich or brocoli are you controlling the child's
actions or its controlled variables? Is eating stuff an action
or an intended result or both (hint: both)? Is eating a behavior
(hint: yes) ?

Hmmmm. From what you say above, I would say that I was controlling a
perception of seeing the kid eat a green vegetable. I've worked in special
schools and I'm fairly confident that I can't control a kids eating actions
if they don't want to eat. I still don't know what you mean by an external
person controlling someone else's controlled variables.

In the special issue of American Behavioral Scientist (Vol 34, No. 1)
behaviour is defined in the glossary as follows: Controlled perceptual
experience. From an observer's point of view, behaviour corresponds to
environmental variables that are controlled by the system and the response
outputs taht are used to control these variables. Lifting a cup involves
producing a particular perception of the cup; to an observer, the behaviour
is movement of the cup and the arm movements that produce this result.

I find this definition very strange. The first sentence seems to refer to
the "working of the entire loop", but the second part seems to refer to
"actions". Also, in this definition, behaviour seems to be only considered
from the observer's perspective. Is this an accpeted PCT convention?

In your paper The Nature of Behavior: Control as fact and theory you make
the point that behaviour _is_ control. Again this would seem to suggest
that the term is at least sometimes used to refer to the working of the
entire loop unless this statement also means "observable output is
control".

Is there some reason that behaviour has never been used as a term on any of
the diagrams that are used to illustrate PCT, or have I just missed some?

Regards,

Tim

34

[From Rick Marken (980831.1100)]

re: Tim Carey's diagram:

I was able to view your diagram of the finger pointing demo
this morning and it was, indeed, right on target. Nice job.

Me:

I thought we have also demonstrated that _controlled variables_
can also be controlled. If we haven't, I will soon.

Tim Carey (980831.2005)

Hmmm. I'm not sure what you mean on this one. Who is contolling the
controlled variable?

The controller is controlling the "victim's" controlled variable.
Look at your diagram; if A decided to control B's cv', instead
of just disturbing it to control B's qo', then you would have
the situation I described; a controlled variable (cv') that is
controlled by the controller. Try altering your diagram to show
this.

Me:

But finger position is also a controlled variable; you can
demonstrate this by pushing gently on the subject's hand while
she is moving it; the subject will push back against your push
in order to protect finger position from the disturbance.

Ye:

How do you separate this from the cv of "distance between
fingers"? Isn't the reason they push back because you are
disturbing the distance cv? Is there anyway you can disturb
the position cv and not disturb the distance cv?

Good observation. I don't know of any way to disturb these
variables independently.

I think one way to determine that two variables (finger-finger
distance and finger position) are being controlled would be to
keep a temporal record of variations in the distance disturbance,
the push disturbance, the distance output (finger position) and
the push back output variables. The (negative) correlation between
distance disturbance variations and distance output variations
should be much larger than that between push disturbance variations
and distance output variations; similarly, the (negative) correlation
between distance disturbance variations and push back output
variations should be much lower than that between push disturbance
variations and push back output variations.

So since actions are both actions and intended results are you
saying that "behaviour = actions" in PCT talk?

No. I'm saying what I said: that behavior is both actions and
intended results.

How can we observe an intended result?

Ah. Glad you asked! Check out my amazing (and much neglected :-()
"Test for the Controlled Variable" demo at:

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

and watch as intended results appear before your very eyes. PCT
magic!

OK, so in Bill's new book where he talks about acting, perceiving,
and comparing, am I correct in saying that in PCT the term
"behaviour" only refers to the acting component.

No. When people talk about "behavior" they are talking about
observable side effects of control; actions (qo), intended
results (qi) and unintended side effects. Bill is talking
about a _model_ that explains the behavior we _observe_.

Also, see Bill Powers (980830.1555 MDT).

I still don't know what you mean by an external person controlling
someone else's controlled variables.

One variable people control is their orientation with respect to
the ground. If I control another person's orientation (I control
whether they are lying down or standing up), then I (an "external
person") am controlling someone else's controlled variable.

Is there some reason that behaviour has never been used as a term
on any of the diagrams that are used to illustrate PCT

No. I guess it just seemed unnecessary. Maybe it would help to
put the word "behavior" on the environment side of the control
diagram with arrows pointing from "behavior" to the variables
qo and qi.

Best

Rick

···

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

35

[From Rick Marken (980831.1105)]

Bruce Gregory (980830.2005)--

The negotiation between the classroom teacher and the student
with the plan in RTP seems to me to qualify as an attempt in
this direction.

Perhaps. Could you explain why you think having a kid develop
a plan in order to get back into class qualifies as a form of
negotiation?

Best

Rick

···

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

36

[From Bruce Gregory 9980831.1411 EDT)]

Rick Marken (980831.1105)

Bruce Gregory (980830.2005)--

> The negotiation between the classroom teacher and the student
> with the plan in RTP seems to me to qualify as an attempt in
> this direction.

Perhaps. Could you explain why you think having a kid develop
a plan in order to get back into class qualifies as a form of
negotiation?

My understanding of the process is that the kid develops the plan and uses
it has input in his negotiations with the classroom teacher to re-enter the
classroom.

Rick Marken (980831.1100)

When people talk about "behavior" they are talking about
observable side effects of control; actions (qo), intended
results (qi) and unintended side effects. Bill is talking
about a _model_ that explains the behavior we _observe_.

Isn't it more accurate to say that qi is the input quantity and the
reference signal the intended results?

Bruce Gregory

37

[From Rick Marken (980831.1210)]

Bruce Gregory 9980831.1411 EDT)--

My understanding of the process is that the kid develops the
plan and uses it has input in his negotiations with the
classroom teacher to re-enter the classroom.

What is being negotiated? In particular, what is the teacher supposed
to be willing to "give" in the "give and take" of the negotiation?

Isn't it more accurate to say that qi is the input quantity and the
reference signal the intended results?

Yes. I'll buy that. But maybe the most accurate way to say it is:
The reference state of qi is the intended result.

Best

Rick

···

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

38

[From Bruce Gregory (980831.1510 EDT)]

Rick Marken (980831.1210)

Bruce Gregory 9980831.1411 EDT)--

> My understanding of the process is that the kid develops the
> plan and uses it has input in his negotiations with the
> classroom teacher to re-enter the classroom.

What is being negotiated? In particular, what is the teacher supposed
to be willing to "give" in the "give and take" of the negotiation?

Tim is the one to answer this question. All my knowledge is second hand.

Bruce Gregory

39

[From Tim Carey (980901.0540)]

[From Bruce Gregory (980831.1510 EDT)]

> What is being negotiated? In particular, what is the teacher supposed
> to be willing to "give" in the "give and take" of the negotiation?

Tim is the one to answer this question. All my knowledge is second hand.

It really depends on the situation. I don't think I could give a blanket
answer. If the kid wanted to sit in a different position in the class
because that would help them be less distracted or whatever the teacher
usually goes along with that. If they wanted to sit near a friend who could
help them stay on task by reminding them perhaps, the teacher would usually
go along with that. If the kid wanted the teacher to check in on him
individually at regular time intervals to see how he was doing at his work,
the teacher would usually go along with that. If the kid was finding the
work too difficult and needed different work the teacher would usually set
some, or if they needed to go to a support teacher to have more intensive
help with their work the teacher would arrange that.

I've run out. As I said at the top it really depends on the individual kid
and what they think they need to succeed in class. The only proviso is that
to "succeed in class" means to succeed at the work that is being done in
class and to succeed in such a way that it doesn't interfere with the
learning of other students. If they wanted to practice juggling in class,
for example, because they thought that would help them succeed, I don't
think the teacher would go along with that. The teacher might, on the other
hand, come to some arrangement where they have free time so that they could
leave the class to juggle.

Does this help?

Regards,

Tim

40

[From Tim Carey (980901.0545)]

[From Rick Marken (980831.1100)]

I was able to view your diagram of the finger pointing demo
this morning and it was, indeed, right on target. Nice job.

Thanks.

The controller is controlling the "victim's" controlled variable.
Look at your diagram; if A decided to control B's cv', instead
of just disturbing it to control B's qo', then you would have
the situation I described; a controlled variable (cv') that is
controlled by the controller. Try altering your diagram to show
this.

I'm really struggling with this one and perhaps this is behind my
difficulty in understanding other things you've explained. How do I
represent this on the diagram? In order to control their cv, do I make
their qo my cv? I'm starting to think of problems with this even as I ask
the question but I'll wait for your answer and we can go from there.

Ah. Glad you asked! Check out my amazing (and much neglected :-()
"Test for the Controlled Variable" demo at:
and watch as intended results appear before your very eyes. PCT
magic!

Yep, done it. It's a great demo, I really like it. It's one I sometimes
show my friends when I'm trying to explain the theory (being the fun kind
of guy that I am ;-)). If, while doing this demo, my reference was to keep
two squares "clear" would the demo still work? I just thought of that while
I was doing it. It works great if I'm controlling the position of one of
the squares ... before too long that square goes black. But if I was just
controlling for two squares staying clear then it could be the case that,
at times, the other square would go black and in this instance that would
seem to be an unintended side effect wouldn't it?

No. When people talk about "behavior" they are talking about
observable side effects of control; actions (qo), intended
results (qi) and unintended side effects. Bill is talking
about a _model_ that explains the behavior we _observe_.

Yep. Isn't the model though, the entire loop? If I see someone sitting on a
park bench and I observe him to be stationary. Do I say that he is just
sitting? If he is running through different scenes in his imagination (say
the romantic dinner he's looking forward to tonight, the great movie he saw
on the weekend, the argument he got involved in, the tennis tournament he
won when he was 12, etc) from the PCT perspective none of this would be
considered "behaviour" would it because none of it is observable. Am I
right in assuming this?

One variable people control is their orientation with respect to
the ground. If I control another person's orientation (I control
whether they are lying down or standing up), then I (an "external
person") am controlling someone else's controlled variable.

OK. So this could be like ballerinas or babies. If a baby cries and I've
picked it up then I've controlled it's orientation, is that what you're
saying? If it was controlling for being picked up, does this make me part
of the babies feedback function?

No. I guess it just seemed unnecessary. Maybe it would help to
put the word "behavior" on the environment side of the control
diagram with arrows pointing from "behavior" to the variables
qo and qi.

OK, so behaviour is only considered to be what you can see. Does that mean
that once qi becomes p it is no longer behaviour? Is behaviour then,
anything that's not a neural signal?

Regards,

Tim