More Control of Behavior

CSGnet Archive CSG1998
21

[From Bruce Abbott (980913.1320 EST)]

Bruce Gregory (980913.1105 EDT) --

I think I'll heed Bill's admonition. When it comes to the speed of my car,
only I am in a position to control it. The state may have sanctions designed
to influence the way I establish the reference levels I choose for my speed,
but these sanctions do not constitute control in the sense that PCT
describes control.

Are you sure? Here is what Bill said:

To define a variable that someone is controlling, you have to define an
input quantity and an output quantity. The input quantity is the physical
situation that is being perceived, and the output quantity is the means by
which the control system has an effect on the input quantity. Disturbances
are independent variables in the environment that also influence the input
quantity.

If the person does perceive the input quantity, and does have an effect on
it by acting on the environment, and if the signs are all right and the
loop gain is sufficient, then the person is controlling the input quantity.

Let's see how closely the actual system conforms to these criteria. The
state legislature and execuitive branch have built a control system for
motorist speed. It includes an array of sensors (radar-equipped state
troopers), perceptual signal (the indicated speeds of motorists), a
reference speed (the posted speed limit, or less), a comparator (the
troopers again), an error signal (error is nonzero if motorist speed is
sufficiently above posted speed limit), an output function (the troopers
again), whose actions tend to bring the input variable (perceived motorist
speeds) back to reference (posted speed limit or less). It seems to work,
too (in that most motorists obey the speed limit when they perceive that
state troopers are monitoring their speeds).

That sounds like a perfectly good control system to me, in that it seems to
me to meet all of Bill's criteria. If you disagree, please explain where
and how it fails.

Regards,

Bruce

22

[From Rick Marken (980913.1140)]

Me:

If the trooper is organized as a control system with respect to
a particular environmental variable (like the speed of your
car) and if the loop is closed and the loop gain is high enough
then control is taking place -- even if the speed of your car
matches the trooper's reference for your speed so that the
trooper need do no more than sit quietly on his motorcycle to
keep the speed of your car at his reference level.

Bruce Gregory (980913.1345 EDT) --

O.K. If you want to believe in magic, go right ahead. The trooper
is _not_ organized like a control system with respect to the speed
of my car.

If, as you say, the trooper is not organized as a control system
with respect to the speed of your car, then, indeed, he cannot
control the speed of your car. You will note that in my statement
above I said "If the trooper is organized as a control system with
respect to a particular environmental..." That's IF as in "IF we
_assume_ that the trooper is organized to control speed then we
can conclude that he is controlling speed even when he is not
_acting_ to change your speed".

You counter my conclusion by changing the assumption; you just
assume that the trooper is _not_ organized as a control system
with respect to the speed of your car so he cannot be controlling
the speed of your car. Given the changed assumption, I agree with
this conclusion. Of course, the trooper is not and cannot be
controlling your speed if he is not organized as a control system
with respect to your speed. Note, however, this trooper is not
going to do anything about your speed _regardless_ of how fast
you drive the car. So if you assume that the trooper is not a
control system with respect to the speed of your car then have a
problem explaining why the trooper goes after your if you _are_
breaking the speed limit. The trooper you describe is one who
will not try to change your speed whether you are going over
the speed limit or not.

Are you saying, perhaps, that the trooper simply cannot be organized
as a control system with respect to the speed of your car? That the
trooper cannot possibly control the speed of your car because that
would violate some kind of natural law? Is this what you meant to
imply by your comments about "thought control" and "magical
thinking"? If so, I can tell you now that there is no such law; the
trooper can organize him (or her) self as a control system with
respect to your speed if s/he wants to.

Bruce Gregory (980913.0935 EDT) --

At the moment I lack the modeling skills needed to convert
this story into a working model

Me:

This seems like a serious shortcoming in a person who says that
we can only understand what we can model.

Bruce:

I have many serious shortcomings. Fortunately magical thinking
is not one of them.

I was merely suggesting that you might not have so much difficulty
with this "control of behavior" issue (revealed by your comment
about "magical thinking") if you could turn your imaginings into
a working model. And you can do this with a _little_ thought now
that we have Vensim available.

Best

Rick

···

--

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

23

[From Bruce Gregory (980913.1450 EDT)]

Bruce Abbott (980913.1320 EST)

Let's see how closely the actual system conforms to these criteria. The
state legislature and executive branch have built a control system for
motorist speed.

No, they have designed a system to _limit_ a motorist's speed.

It includes an array of sensors (radar-equipped state
troopers), perceptual signal (the indicated speeds of motorists), a
reference speed (the posted speed limit, or less), a comparator (the
troopers again), an error signal (error is nonzero if motorist speed is
sufficiently above posted speed limit), an output function (the troopers
again), whose actions tend to bring the input variable (perceived motorist
speeds) back to reference (posted speed limit or less). It seems to work,
too (in that most motorists obey the speed limit when they perceive that
state troopers are monitoring their speeds).

That sounds like a perfectly good control system to me, in that
it seems to
me to meet all of Bill's criteria. If you disagree, please explain where
and how it fails.

A very interesting idea. I am not sure Bill would like it though, since it
is a control system that transcends the individuals who make up the society.
I think it is an analogy. I agree that the individual trooper is a control
system designed to issue tickets to drivers who are exceeding the posted
limit by some amount. The troopers, however, are autonomous, their reference
systems are established by themselves, not by the state legislature or
governor. Troopers are free to pursue speeders or to ignore them. (If the
troopers were _really_ concerned with limiting the speed of offenders, they
would follow them to determine if the offenders resumed speeding; this the
troopers do not, in my experience, do.) If the troopers really want to limit
the speed of the drivers, I think your rolling roadblock is the only way to
accomplish this. The other alternative is to install Rick's throttle
controls in all cars and allow the troopers to really act as systems
organized to control the speed of cars.

Bruce Gregory

24

[From Bruce Gregory (980913.1500 EDT)]

Rick Marken (980913.1140)

Are you saying, perhaps, that the trooper simply cannot be organized
as a control system with respect to the speed of your car?

Yes.

That the
trooper cannot possibly control the speed of your car because that
would violate some kind of natural law?

No. Just that the trooper lacks a mechanism such as your radio controlled
throttle system that would allow him to control the speed of my car.

Is this what you meant to
imply by your comments about "thought control" and "magical
thinking"? If so, I can tell you now that there is no such law; the
trooper can organize him (or her) self as a control system with
respect to your speed if s/he wants to.

The only way I see this happening if the trooper turns off the engine of my
car and tows me at whatever speed he desires. This I agree would allow him
to control the speed of my car. Otherwise I don't see how this is possible.
How, for example, can a trooper control my speed to equal 50 mph in a 65 mph
zone? If he puts his car in front of mine, as some have done, I can, and
have, simply passed the trooper. A trooper, as _you_ pointed out, is
probably controlling a program for issuing tickets. I have no doubt that
such organization is possible.

I was merely suggesting that you might not have so much difficulty
with this "control of behavior" issue (revealed by your comment
about "magical thinking") if you could turn your imaginings into
a working model. And you can do this with a _little_ thought now
that we have Vensim available.

I would be grateful if _you_ would produce such a model since I obviously do
not understand how it is even organized.

Bruce Gregory

25

[From Rick Marken (980913.1250)]

Me:

Are you saying, perhaps, that the trooper simply cannot be
organized as a control system with respect to the speed of
your car?

Bruce Gregory (980913.1500 EDT) --

Yes.

But this is incorrect. You have already agreed (see below) that
a trooper _can_ control the speed of a car; he can do it in what
you call the "Marken-Abbott fashion" (in the way the controllee's
output is controlled in Bruce Abbott's Vensim model).

Me:

[Are you saying] That the trooper cannot possibly control the
speed of your car because that would violate some kind of
natural law?

You:

No. Just that the trooper lacks a mechanism such as your radio
controlled throttle system that would allow him to control the
speed of my car.

But this is only a technical difficulty. My radio controlled
throttle system is not currently in used (thank goodness). I
only mentioned it because you had said:

Until I see a model of a trooper controlling the speed of my
car without blocking the road ahead of me in the Marken-Abbott
fashion I will assume that it is not possible and will do my
best to ignore claims to the contrary.

In other words, you wanted to see a model of a system (trooper)
that controls your speed directly. So I provided a model that
does this. In fact, real troopers can control speed directly
without using the Marken-Abbott technique (which works only
if the driver is controlling for distance from the trooper) _or_
the radio controlled accelerator lever. In high speed chases
(where the trooper's goal is clearly to control the chasee's
speed) the trooper can follow the chasee until he runs out of
gas, until his tires are shot out or until other cops physically
obstruct his path with some kind of barrier (like another car).
Real troopers can and sometimes _will_ directly control the speed
of your car.

The only way I see this [control of my speed] happening [is] if
the trooper turns off the engine of my car and tows me at whatever
speed he desires. This I agree would allow him to control the
speed of my car.

There you go!

How, for example, can a trooper control my speed to equal 50 mph
in a 65 mph zone?

The fact that the controller can't keep the controlled variable
in all of its possible states doesn't mean that the controller
is not controlling. I can't get my Geo to go 130 mph on the
straightaway (I could if it were a Mercedes) but I can control
its speed. The police in the car chase can only keep the chasee's
car speed in two states: 0 mph or>0 mph but they are still
controlling car speed.

I would be grateful if _you_ would produce such a model since I
obviously do not understand how it is even organized.

Ok. But I can't learn PCT for you.

Best

Rick

···

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

26

[From Bruce Abbott (980913.1550 EST)]

Bruce Gregory (980913.1450 EDT) --

Bruce Abbott (980913.1320 EST)

Let's see how closely the actual system conforms to these criteria. The
state legislature and executive branch have built a control system for
motorist speed.

No, they have designed a system to _limit_ a motorist's speed.

They have designed a _control_ system to limit a motorist's speed.

That sounds like a perfectly good control system to me, in that
it seems to
me to meet all of Bill's criteria. If you disagree, please explain where
and how it fails.

A very interesting idea. I am not sure Bill would like it though, since it
is a control system that transcends the individuals who make up the society.
I think it is an analogy.

If it has all the elements of a control system, connected up in the right
way, and behaving as a control system behaves, it is the real thing. This
has all right parts, connected up in the right way, and acts to restore the
controlled variable to its reference range when that variable rises above
it. It's a real control system, not merely analogous to one.

I agree that the individual trooper is a control
system designed to issue tickets to drivers who are exceeding the posted
limit by some amount.

If the intent of the legislature in designing this system and having it
implemented was not to limit the speeds of motorists, then it is hard to
understand why the troopers have been instructed to issue tickets to
speeders (unless this is just a pretext for relieving the motorists of their
money so as to enrich the government!). Of course, the individual trooper
may only be controlling the truth-status of "speeders get tickets,
nonspeeders do not." Nevertheless, the legislature and the governor have
set the trooper to this task because they assume that these actions on the
part of the trooper will tend to bring motorists' speeds down to values not
exceeding the speed limit.

The troopers, however, are autonomous, their reference
systems are established by themselves, not by the state legislature or
governor.

Autonomous, yes. Independent of outside influence, no. How else can you
can explain why so many autonomous control systems (state troopers) decide
to go out there and ticket speeders, just as the governor and the
legislature want them to?

Regards,

Bruce

27

[From Bruce Gregory (980913.1650 EDT)]

Rick Marken (980913.1250)

Me:

> Are you saying, perhaps, that the trooper simply cannot be
> organized as a control system with respect to the speed of
> your car?

Bruce Gregory (980913.1500 EDT) --

> Yes.

But this is incorrect. You have already agreed (see below) that
a trooper _can_ control the speed of a car; he can do it in what
you call the "Marken-Abbott fashion" (in the way the controllee's
output is controlled in Bruce Abbott's Vensim model).

I was under the mistaken assumption that we were talking about whether or
not troopers _do_ control the speed of my car on the turnpike. Apparently we
were talking about how troopers conceivably _might_ control the speed of my
car. Yes, troopers conceivably might do this, although they have not been
known to do so.

In other words, you wanted to see a model of a system (trooper)
that controls your speed directly. So I provided a model that
does this. In fact, real troopers can control speed directly
without using the Marken-Abbott technique (which works only
if the driver is controlling for distance from the trooper) _or_
the radio controlled accelerator lever. In high speed chases
(where the trooper's goal is clearly to control the chasee's
speed) the trooper can follow the chasee until he runs out of
gas, until his tires are shot out or until other cops physically
obstruct his path with some kind of barrier (like another car).
Real troopers can and sometimes _will_ directly control the speed
of your car.

Look at it another way. I can control the behavior of a trooper simply by
stepping on the gas as I approach him. I can cause the trooper to turn on
his lights and siren and to follow me! I do this by disturbing a perception
that he is controlling ala Abbott. Fair enough?

Ok. But I can't learn PCT for you.

See my above application of the Abbott model. Does this demonstrate an
understanding of PCT?

Bruce Gregory

28

[From Bruce Gregory (980913.1800 EDT)]

Bruce Abbott (980913.1550 EST)

If it has all the elements of a control system, connected up in the right
way, and behaving as a control system behaves, it is the real thing. This
has all right parts, connected up in the right way, and acts to
restore the
controlled variable to its reference range when that variable rises above
it. It's a real control system, not merely analogous to one.

True, so long as the reference speed is in the trooper's head and not simply
posted on the highway. As I noted, it results in tickets being given. It is
not clear that it restores the controlled variable to its reference range. I
still observe speeders even though tickets are being given. In fact, the
trooper who is writing a ticket is unlikely to pursue you if you speed up
after you pass him by. A best people slow down for a mile or two after they
see a ticket being issued. Some, as I say, speed up.

If the intent of the legislature in designing this system and having it
implemented was not to limit the speeds of motorists, then it is hard to
understand why the troopers have been instructed to issue tickets to
speeders (unless this is just a pretext for relieving the
motorists of their
money so as to enrich the government!). Of course, the individual trooper
may only be controlling the truth-status of "speeders get tickets,
nonspeeders do not." Nevertheless, the legislature and the governor have
set the trooper to this task because they assume that these
actions on the
part of the trooper will tend to bring motorists' speeds down to
values not
exceeding the speed limit.

True, but intentions of legislators are not relevant to the operation of the
control system. Prohibition was passed with the intention of stopping the
consumption of alcohol. It didn't.

>The troopers, however, are autonomous, their reference
>systems are established by themselves, not by the state legislature or
>governor.

Autonomous, yes. Independent of outside influence, no. How else can you
can explain why so many autonomous control systems (state troopers) decide
to go out there and ticket speeders, just as the governor and the
legislature want them to?

I never claimed they were independent of outside influence. If they have a
quota for tickets, and they are not meeting it, they will probably set their
reference speed closer and closer to the posted limit.

Bruce Gregory

29

[From Rick Marken (980913.1600)]

Bruce Gregory (980913.1650 EDT) --

Yes, troopers conceivably might do this [control the speed
of cars], although they have not been known to do so.

Well, that's progress.

Look at it another way. I can control the behavior of a trooper
simply by stepping on the gas as I approach him. I can cause the
trooper to turn on his lights and siren and to follow me! I do
this by disturbing a perception that he is controlling ala Abbott.
Fair enough?

Right. You can control a variable (the lights and siren being on
or off) the trooper is using to control a variable he is
controlling (like the state of the relationship between your
speed and your receipt of a ticket) if you can disturb this
variable (disturb the state of the relationship between your
speed and your receipt of a ticket).

Does this demonstrate an understanding of PCT?

Sure. But it would have helped if you had described the
_variables_ you think you and the trooper are controlling
and how you each control these variables.

Bruce Gregory (980913.1750 EDT) --

In my house, the thermostat is set to 65 F. During the summer,
the temperature in the house never drops below 70 F. Is the
thermostat controlling the temperature?

What do you think? Your answer will be another indication of how
well you understand control theory.

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 (980913.1900 EDT)]

Bill Powers (980913.0732 MDT)

Also, it is good to remember, now and then, that people do not control
things, they control variables. So one person does not control another
person, any more a person controls a car.

When I am driving on the turnpike, the speed of my car can be modeled by a
simple control loop that includes a reference for 70 mph. There is no need
to introduce the idea that my speed is being controlled, or even effected,
by the presence or absence of a state trooper. Why then introduce the notion
that my speed is being controlled by the trooper? This addition seems
totally gratuitous and does not improve one iota your ability to predict the
speed of the car.

Bruce Gregory

31

[From Bruce Gregory (980913.1915 EDT)]

Rick Marken (980913.1600)

Bruce Gregory (980913.1750 EDT) --

> In my house, the thermostat is set to 65 F. During the summer,
> the temperature in the house never drops below 70 F. Is the
> thermostat controlling the temperature?

What do you think? Your answer will be another indication of how
well you understand control theory.

In my view the question is, what are we trying to model? If we are trying to
model the temperature of the house, there is no need to introduce the
thermostat and the furnace into the model. To say that the thermostat is
controlling the temperature in this case smacks more of belief than it does
of science. The thermostat and furnace play no role in any model I would put
forth until they are required to explain the changes in temperature. I
realize that this must make me a PCT pariah but Occam's razor is too
important a tool to put aside for reasons of theological purity.

Bruce Gregory

32

From [ Marc Abrams (980913.1948) ]

I think each of you ( Bruce G & A and Rick ) are correct.
The control process entails a number of points where someone
might try to _affect_ our ability to control something. I
can try to provide a "reference" level for you that _you_
will encorporate _into_ your control loop. I believe Rick
was addressing that. You can try to affect someones actions
( output ) and have _that_ incorporated into your control
loop. t I believe that is what Bruce Gregory was arguing
against ( i.e. the ability of _anyone_ to _physically_
control the speed of the car beside the driver, except maybe
for a crazed passenger in the front seat who puts his/her
foot on the accelerator :slight_smile: ) The other way that someone
_might_ affect the control process is through qi ( i.e.
either a disturbance or by somehow changing the action going
into the qi ( possibly by deceit, or a lie ).

The important thing to remember is that it is the _TOTALITY_
of the loop that matters. Not just one individual part. As
long as the loop remains _intact_ control is going on. So
someone very well might _affect_ and hence cause _effects_
at various points with or without control being "broken".

In working through a few problems and hence the model I will
be posting to the net I have seen some pretty outrageous
"examples" of S -> R Psych modeled in SD. What was crazy was
how the models _supported_ these theories. Well folks,
models can _support_ _ANY_ argument. The trick is to
understand the underlying logic and math to understand _how_
the model was formulated. Pictures ( Diagrams ) just aren't
enough. Neither are descriptive words. Geesh, I'm becoming a
real cynic. Thanks Bill :slight_smile:

Bruce Gregory (980913.1750 EDT) --

> In my house, the thermostat is set to 65 F. During the
> summer, the temperature in the house never drops
>below 70 F. Is the thermostat controlling the
>temperature?

Rick asked and said:

What do you think? Your answer will be another indication
of how well you understand control theory.

Bruce G's response:

In my view the question is, what are we trying to model?

Great answer, I agree. The question as originally
formulated by you is incomplete. Are people in the room?.

If we are trying to model the temperature of the house,

there >is no need to introduce the thermostat and the
furnace into >the model.

Here I disagree. It might be inappropriate for _your_ model
but maybe not mine. You said "we are trying..." implying
more then one perspective of the problem.

To say that the thermostat is controlling the temperature

in >this case smacks more of belief than it does of science.
The >thermostat and furnace play no role in any model I
would put

forth until they are required to explain the changes in
temperature.

I totally agree and support your right to include and
exclude variables as _you_ see fit, for _your_ system
definition and boundries. As Bill stated, lets establish the
facts first _then_ build the model.

I realize that this must make me a PCT pariah but Occam's
razor is too important a tool to put aside for reasons of
theological purity.

I think your argument has more to do with establishing
_your_ system boundaries and components then with PCT and I
agree with your position on that. I don't understand the
connection to PCT though.

Marc

33

[From Bruce Abbott (980913.2020 EST)]

Bruce Gregory (980913.1915 EDT) --

Bruce Gregory (980913.1750 EDT)

In my house, the thermostat is set to 65 F. During the summer,
the temperature in the house never drops below 70 F. Is the
thermostat controlling the temperature?

In my view the question is, what are we trying to model? If we are trying to
model the temperature of the house, there is no need to introduce the
thermostat and the furnace into the model. To say that the thermostat is
controlling the temperature in this case smacks more of belief than it does
of science. The thermostat and furnace play no role in any model I would put
forth until they are required to explain the changes in temperature. I
realize that this must make me a PCT pariah but Occam's razor is too
important a tool to put aside for reasons of theological purity.

In the activity you describe as modeling the temperature of the house, you
are trying to explain why the variable, "temperature of the house" varies as
it does. If the furnace never kicks on (because the house temperature never
falls below set point), then there is no need to bring the house's one-way
temperature control system into the explanation, because the temperature of
the house is never affected by action resulting from the error between the
set-point temperature and the actual temperature. Under the conditions you
describe, the system may as well not be there.

Nevertheless, if the system is operating properly, it is still controlling
the temperature of the room against excursions below set-point, by
monitoring the room temperature, comparing it to the reference of 65
degrees, and being constructed so as to turn on the furnace if the room
temperature as sensed falls below set point. It is not controlling against
excursions of room temperature above set point.

In a two-way system (one that takes action to oppose error resulting from
deviations from reference in either direction) with no deadband, and
deviation from set-point would be opposed. But control systems can also act
merely to limit excursions of the controlled variable, with no action being
called for until deviations from reference exceed a certain value. The
result is a deadband area within which variations in the input will not
produce any corrective action. In the case of one-way control systems, no
corrective action will be taken until the deviation is in a particular
direction. In the case of your home heating system, deviations above
set-point are not a cause for corrective action; the system does not control
against deviations in this direction. Yet it is always controlling against
deviations below set-point, whether these occur or not.

If you wished to describe the system of variables so that you could
accurately predict the changes in room temperature under all possible
conditions, you would need to include the control system in your
description. If you wished merely to explain the changes in room
temperature during the summer, when the room temperature never drops below
70 degrees, you wouldn't, even though the control system designed to keep
room temperature from falling much below set-point is there and functioning.

Regards,

Bruce

34

[From Bill Powers (980913.2036 MDT)]

Rick Marken (980913.1000)--

Bruce Gregory (980913.1108 EDT)--

Thanks, Bill.

I'm glad you liked it. If you subtitute "trooper" for "person"
and "speed of your car" for "input quantity" you will see that
Bill is saying just what Bruce Abbott and I have been saying: that,
if the signs are all right and the loop gain is sufficient, then
the trooper is controlling the speed of your car when the speed
of your car is the controlled input variable.

Can't we just transcend this stupid example and go on to something else? I
couldn't care less who was right or wrong.

35

[From Bill Powers (980913.2043 MDT)]

Bruce Abbott (980913.1320 EST)--

Let's see how closely the actual system conforms to these criteria. The
state legislature and execuitive branch have built a control system for
motorist speed. It includes an array of sensors (radar-equipped state
troopers), perceptual signal (the indicated speeds of motorists), a
reference speed (the posted speed limit, or less), a comparator (the
troopers again), an error signal (error is nonzero if motorist speed is
sufficiently above posted speed limit), an output function (the troopers
again), whose actions tend to bring the input variable (perceived motorist
speeds) back to reference (posted speed limit or less). It seems to work,
too (in that most motorists obey the speed limit when they perceive that
state troopers are monitoring their speeds).

That sounds like a perfectly good control system to me ...

I see only one problem with it: it is not designed to perceive YOUR speed.
As you say, it is set up to perceive "motorists'" speed. The Department of
Transportation does not receive a report that Bruce Abbott was speeding. It
knows only that motorists drove, on I-95, at an average speed of 82 mph,
and that 1,255 speeding citations were issued from April 1 to May 1. While
your own performance contributed to that average speed, it did so only to a
minute degree; the loop gain must have been extremely low.

There is some leeway in deciding just how much loop gain there has to be to
permit us to identify a control system, but I suspect that the threshold
gain will be something more than 0.001.

Best,

Bill P.

36

[From Bill Powers (980913.1903 MDT)]

Bruce Gregory (980913.1900 EDT)--

When I am driving on the turnpike, the speed of my car can be modeled by a
simple control loop that includes a reference for 70 mph. There is no need
to introduce the idea that my speed is being controlled, or even effected,
by the presence or absence of a state trooper. Why then introduce the notion
that my speed is being controlled by the trooper? This addition seems
totally gratuitous and does not improve one iota your ability to predict the
speed of the car.

Why do you set the reference for your speed at 70 mph? This setting is the
net action of all the higher-level control systems on whose perceptions
speed has some influence. One of them might be "I can't afford to get
another speeding ticket." When you can't see any troopers ahead or behind,
or lurking behind road signs, or bleeping your radar detector, you zoom up
to your chosen speed of 70 mph because you imagine no chance of being
stopped, and experience no error with respect to getting a speeding ticket.
But if you spot a trooper ahead, this produces a disturbance of that
control system, which you counteract by resetting your speed reference
signal to 55 mph (the local limit). As a result, of course, the car slows
to 55 mph, and your perceived danger of getting a ticket drops to zero. So
YOU are controlling the likelihood of getting a ticket by reducing your
reference level for speed, and THE TROOPER, by displaying himself along the
road, is controlling your speed by (he assumes) disturbing your perception
of the likelihood of getting a ticket. Thus each of you controls what
matters most to him without preventing the other from doing the same.

There are other assumptions about the nature of the controlled variables
that might lead to different outcomes.

Best,

Bill P.

37

[From Rick Marken (980913.2030)]

Bruce Gregory (980913.1750 EDT) --

In my house, the thermostat is set to 65 F. During the summer,
the temperature in the house never drops below 70 F. Is the
thermostat controlling the temperature?

Rick Marken (980913.1600)

What do you think? Your answer will be another indication of how
well you understand control theory.

Bruce Gregory (980913.1915 EDT)

In my view the question is, what are we trying to model? If we
are trying to model the temperature of the house, there is no
need to introduce the thermostat and the furnace into the model.
To say that the thermostat is controlling the temperature in
this case smacks more of belief than it does of science. The
thermostat and furnace play no role in any model I would put forth
until they are required to explain the changes in temperature.\

Compare your answer to the correct one:

Bill Powers (980913.2056 MDT)

Yes.

See how much simpler things get when you understand control
theory.

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 Bill Powers (980913.2117 MDT)]

Bruce Gregory (980913.1915 EDT)--

In my view the question is, what are we trying to model? If we are trying to
model the temperature of the house, there is no need to introduce the
thermostat and the furnace into the model. To say that the thermostat is
controlling the temperature in this case smacks more of belief than it does
of science.

In particular, we are trying to devise a model that will account for the
behavior of the house's temperature over time, given that the furnace is
going on and off, people are walking in and out of doors and opening and
closing windows, and winds of varying temperatures and sun and cloud
phenomena are changing the heat losses and gains through the walls of the
house. What we observe is that the net effect of all the heat losses and
gains through doors, windows, and walls just happens to be offset by
variations in the on-time of the furnace, so that the temperature near a
small box on the wall stays remarkably close to 70 degrees F. And we also
observe that after a person goes to the box and fiddles with it, all the
same relationships go on happening, but the temperature is now constant at
65 degrees F.

This is the phenomenon for which we need a model.

I realize that this must make me a PCT pariah but Occam's razor is too
important a tool to put aside for reasons of theological purity.

And it is too sharp for us to encourage its use by the unskilled.

Best,

Bill P.

39

[From Bruce Gregory (980914.0631 EDT)]

Rick Marken (980913.2030)]

See how much simpler things get when you understand control
theory.

Indeed. Almost as Einstein said, "A pillow on which true believers rest
their heads so their sleep is not disturbed."

Bruce Gregory

40

[From Bruce Gregory (980914.0633 EDT)]

Bill Powers (980913.1903 MDT)

So
YOU are controlling the likelihood of getting a ticket by reducing your
reference level for speed, and THE TROOPER, by displaying himself
along the
road, is controlling your speed by (he assumes) disturbing your perception
of the likelihood of getting a ticket. Thus each of you controls what
matters most to him without preventing the other from doing the same.

There are other assumptions about the nature of the controlled variables
that might lead to different outcomes.

I'll accept that as the last word.

Bruce Gregory