Taxonomy Revisited :-)

From [ Marc Abrams (090911.0752) ]

In my post on taxonomy I serendipitously tripped over
something I think is important to be aware of. In replies to
my post on a taxonomy, both Bruce Abbott and Rick Marken
suggested that the CV be dropped and qi used instead. Both
argued ( correctly, from a theoretic view point ) that the
CV was unnecessary and was fully represented by qi. ( or was
it qo, Rick :slight_smile: ). I disagree with both and the the "State
trooper" model ( among others, like the coercion
discussion ) shows why. First a little background. In my
taxonomy post ( Marc Abrams 980909.0353 ) I used Rick's
reply to Bruce Gregory (Rick Marken (980908.1525) as an
_example_ of a modeling _style_ I thought would be useful. I
did _not_ go back to the beginning of the Marken-Gregory
thread because I was _not_ interested in _accurately_
modeling the discussion. I used Rick's pot because he
_defined_ or Mapped the elements (r2, output, etc. ) into a
_story about the state trooper. In trying to understand the
thinking I used Rick's definitions to map the problem. It
seemed like a good way to _read_ a model. Interestedly
enough, what I got back was a critique of my story from
Bruce's G and A and Rick. Which was _Totally_ beside _my_
point in the taxonomy post :slight_smile: Which brings me to my point
about _this_ post. :slight_smile:

What often happens on CSGnet is that discussions flip flop
back and forth between _A_ theoretic PCT ( general ) model
and a real (specific ) _example_ using PCT to understand it.
_One_ of the reasons ( not the only one ) this happens is
that the _CV_ is _implicitly_ defined in qi in the theoretic
model. For theoretic models this is fine. As long as we are
_consistent_ in it's use. But when we use _real_ examples
and _change_ the kinds of "disturbances", and/or the kinds
of "output" ( hence, _changing_ the perception and CV ).
.That is no longer good. The model then becomes _very_
slippery. Thats what happened with the coercion discussion.
Rick who usually masters in the theoretic, but _loves_ the
gossipy real stuff :slight_smile: starts "magically" changing parts of
the input function. To most of us this seems like magic, but
from Rick's standpoint it's just 'ol theoretic qi. No big
deal. Well we are both right. Rick has found out that his
"intuition" ( theoretic based assumptions ) don't always
hold up with "real" examples. Not because there is anything
wrong with the theory. But because of inconsistent
_application_ of it. The CV is the central "plot" to any PCT
model. A CV may and can contain _any_ number of disturbances
and outputs. _theoretically_ combining all these does not
make a difference. As they say it _all_ becomes perception.
But with real examples _everything_ else in the model is
_based_ on that so we really have to be _explicit_ with it.

Any comments?

Marc

[From Rick Marken (980911.0900)]

Marc Abrams (090911.0752)

Both [Bruce A. and myself] argued ( correctly, from a theoretic
view point ) that the CV was unnecessary and was fully
represented by qi.

Not quite right. I think we both agree that the CV (controlled
variable) is definitely _necessary_ in our models of behavior;
indeed, it is _essential_ for understanding behavior. Bruce A.
and I were making a _notational_ point. The notation qi (input
quantity) has already been used as the symbol for "controlled
variable" in B:CP and it is consistent with the convention of
using lower case for variables and upper case for functions. You
can call the controlled variable CV but that's the same as calling
it qi.

I used Rick's pot

Shhhh. Starr is running the country now. Drugs, sex and rock
and roll are no longer immature recreations; they are high
crimes and misdemeanors.

because he _defined_ or Mapped the elements (r2, output, etc. )
into a _story about the state trooper.

Actually, Bruce A. and I mapped variables in the real world to
variables in the model.

Interestedly enough, what I got back was a critique of my
story from Bruce's G and A and Rick.

We just disagreed with the real world variables you mapped to the
variables in the model.

Which was _Totally_ beside _my_ point in the taxonomy post :slight_smile:

Which was?

What often happens on CSGnet is that discussions flip flop
back and forth between _A_ theoretic PCT ( general ) model
and a real (specific ) _example_ using PCT to understand it.

I would say that we are always trying to map variables in the
model to variables in the world, and vice versa.

_One_ of the reasons ( not the only one ) this happens is
that the _CV_ is _implicitly_ defined in qi in the theoretic
model.

qi _is_ the CV in the model. It's as _explicit_ as the Starr
report will be (ain't it cute how all those Republican's
are rushing to _distribute_ pornograpy on the internet;-).

For theoretic models this is fine. As long as we are
_consistent_ in it's use. But when we use _real_ examples
and _change_ the kinds of "disturbances", and/or the kinds
of "output" ( hence, _changing_ the perception and CV ).
.That is no longer good. The model then becomes _very_
slippery. Thats what happened with the coercion discussion.

I have absolutely no idea what you are talking about here?
To apply the control model appropriately, the CV (also called qi)
in the model must map to the CV of the system being modelled.
In order to know the variable controlled by the real system
(the real system's CV) we have to do _the Test. In our discussions
on the net, we are usually just guessing what variable (CV) a
real system (like a trooper or a driver) _might_ be controlling;
then we can map the behavior of the model to the behavior of
teh real system. If we _imagine_ that a trooper is controlling
for the speed at which a car is being driven (driving speed is
the trooper's CV) and that the driver of that car is controlling
for his distance from the trooper (distance from trooper is the
driver's CV) then we can see (via Bruce A's model) how the
trooper can control the driver's speed by disturbing the distance
(by speeding up or slowing down) between himself and the driver.

Rick who usually masters in the theoretic, but _loves_ the
gossipy real stuff :slight_smile: starts "magically" changing parts of
the input function.

Any master of the theory is also a master of the reality. There
is no master of the reality who has not mastered the theory. I
would like to see where you think I "magically" changed part of
the input function. I have no idea what you might mean by this.

To most of us this seems like magic, but from Rick's standpoint
it's just 'ol theoretic qi.

Again. qi is just a symbol for the controlled environmental
variable (or controlled _quantity_ since it may not be a
simple physical variable). qi is part of the model but its
"reality" (like the "reality" of mass) can be established by
experimental test. For example, there is very good evidence
that the qi for fielders trying to catch a fly ball is optical
velocity.

Rick has found out that his "intuition" ( theoretic based
assumptions ) don't always hold up with "real" examples.

It wasn't my assumptions that didn't "hold up". It was my
mental calculation of the implications of those assumptions.
What didn't hold up was my intuitive predictions of what
would happen under certain circumstances.

The CV is the central "plot" to any PCT model.

That's for sure!

A CV may and can contain _any_ number of disturbances
and outputs. _theoretically_ combining all these does not
make a difference. As they say it _all_ becomes perception.

I don't know what you mean by "does not make a difference".
The disturbances and outputs that affect a CV certainly do
make a difference in how the system behaves. Also, the
controlled variable (CV or qi or whatever you want to call it)
is the exact correlate (in the observer's perception) of the
perceptual variable that is controlled by the control system.
That is: p = qi. Yes. The controlled variable does "become
perception"; indeed, it _is_ a perception, for both the observer
and the control system. But it's still important to know what
variable a system is controlling -- what it's CV (or qi) is -- in
order to know what it is _doing_ (controlling qi).

But with real examples _everything_ else in the model is
_based_ on that so we really have to be _explicit_ with it.

We _always_ use real examples when we talk about the model;
in the simplest case, we talk about the model in terms of
real _tracking_ behavior. When you do any of the demos at
my website you are producing _real_ behavior; it may not
seem like "important" behavior; but it's certainly real. The
model (in those demonstrations where I have included model
behavior) is being fit to the _real_ behavior that you have
produced.

Any comments?

Do these help?

Best

Rick

路路路

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

[From Bruce Gregory (980911.1235 EDT)]

Rick Marken (980911.0900)

Marc Abrams (090911.0752)

> because he _defined_ or Mapped the elements (r2, output, etc. )
> into a _story about the state trooper.

Actually, Bruce A. and I mapped variables in the real world to
variables in the model.

Actually, neither you nor Bruce A. (nor I for that matter) have access to
variables in the "real world" when we are talking about an imaginary
scenario. Even when we _have_ data, unless we perform the Test, we are
making up a story about the data.

Any master of the theory is also a master of the reality. There
is no master of the reality who has not mastered the theory. I
would like to see where you think I "magically" changed part of
the input function. I have no idea what you might mean by this.

I don't know what Marc has in mind, but you changed from arguing that the
trooper was controlling my speed to arguing that the trooper was controlling
a program for issuing tickets. (I happen to agree with the second argument.)

Bruce Gregory

[From Rick Marken (980911.1240)]

Bruce Gregory (980911.1235 EDT)

I don't know what Marc has in mind, but you changed from arguing
that the trooper was controlling my speed to arguing that the
trooper was controlling a program for issuing tickets. (I happen
to agree with the second argument.)

I must not have made myself clear. I wasn't arguing about what
variable the trooper was _actually_ controlling (it was a _story_,
remember; you have to do _experiments_ to find out what variables
people are actually controlling). All I (and Bruce Abbott) was
doing was noting that your original story about the trooper and
the driver did not fit the model of control of behavior that
Bruce Abbott developed. Bruce A. and I noted that a story that
maps correctly onto Bruce's "Control of Behavior" demo is the
one where the trooper (who is the controller in the model)
controls the speed of a driver who is controlling his distance
from the trooper. Actually, I think I told this story (trooper
controlling speed by disturbing distance controlled by driver)
before Bruce A. did but you seemed to like Bruce. A's version
better, which is fine with me.

Best

Rick

路路路

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

[From Bruce Gregory (980911.1551 EDT)]

Rick Marken (980911.1240)

Bruce Gregory (980911.1235 EDT)

I must not have made myself clear. I wasn't arguing about what
variable the trooper was _actually_ controlling (it was a _story_,
remember; you have to do _experiments_ to find out what variables
people are actually controlling). All I (and Bruce Abbott) was
doing was noting that your original story about the trooper and
the driver did not fit the model of control of behavior that
Bruce Abbott developed. Bruce A. and I noted that a story that
maps correctly onto Bruce's "Control of Behavior" demo is the
one where the trooper (who is the controller in the model)
controls the speed of a driver who is controlling his distance
from the trooper. Actually, I think I told this story (trooper
controlling speed by disturbing distance controlled by driver)
before Bruce A. did but you seemed to like Bruce. A's version
better, which is fine with me.

No matter who proposed the story, it is a lot more believable than the story
in which the trooper sitting in his car by the side of the role is
controlling my speed. Since he has no way of affecting my speed short of
positioning himself in front of me, it is difficult for me to imagine a
model that allows the sedentary trooper to control my speed. However, if you
can propose one, I'd love to see it.

On a related topic. If I decide, at the principle or system level, to allow
the world to be exactly as it is, i.e., to adjust my reference levels to
match my perceptions, am I thereby controlling the world? (Sorry Phil.)

Bruce Gregory

[From Rick Marken (980911.1400)]

Bruce Gregory (980911.1551 EDT)

No matter who proposed the story, it is a lot more believable
than the story in which the trooper sitting in his car by the
side of the role is controlling my speed.

You can't tell what a person is doing (controlling) by just
watching what they are doing (acting). Just because a trooper
is acting in a particular way (sitting in his car) doesn't mean
that he _is_ or _is not_ doing something (controlling speed).

Perhaps you find it difficult to believe that a trooper sitting
in his car _could_ be controlling your speed because you look at
behavior as visible actions and their consequences. This is the
behaviorist view of behavior and it's obviously a tough view to
overcome (in others and in oneself). We tend to think that a
person isn't doing anything if he isn't _acting_ in a particular
way or producing obvious consequences. But when you are able to see
behavior as the control of perception, then you will see that what
a person is doing (how they are _acting_ ) at a particular time
is not necessarily a good indication of what they are doing
(controlling).

A good demonstration of the fact that you can't tell what peole
are doing (controlling) by looking at what they are doing (actions
and the consequences of those actions) is my "Test for the
Controlled Variable" demo at:

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

Since he has no way of affecting my speed short of positioning
himself in front of me, it is difficult for me to imagine a
model that allows the sedentary trooper to control my speed.

When your speed is at or lower than his reference then he need
do nothing more than sit there to keep your speed under control.
If your speed is above his reference he'll chase you down or
produce whatever action is necessary to slow you down you. If he
wants to stay seated he can just shoot out your tires with a high
powered rifle.

On a related topic. If I decide, at the principle or system
level, to allow the world to be exactly as it is, i.e., to
adjust my reference levels to match my perceptions, am I
thereby controlling the world? (Sorry Phil.)

This is an interesting question. As you've described it you
are not controlling the world (or your perceptions of that
world). What you are controlling is the perception of a
relationship between perceptual variables and reference variables
and (according to your description) you are doing this only by
adjusting the references. So you are adjusting what you want
(references) to match what you are getting (perceptions).
This is a control process, one that requires a control system
(in you) that can control a perception of the relationship
between a perception and its reference.

I think we _do_ try to adjust our goals to match our perceptions
_sometimes_; but I don't think it can work as a general principle
for dealing with the world. Eventually you're going to find
yourself "wanting" a perception that is not wanted by other
control systems. For example, you can _want_ to get laid off
(because you perceive that you have been laid off) but then you
can't buy food. I suppose you can now _want_ no food (because you
perceive no food) but then you perceive the pain of hunger. I
guess you can then try to want that pain and the death that
comes eventually. I think it's clear from this example, though,
why adjusting references to match perceptions didn't catch on;
it's because the people who delt with the world this way rarely
made it past being a blastocyte.

Best

Rick

路路路

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

[From Bruce Gregory (980911.1735 EDT)]

Rick Marken (980911.1400)

Bruce Gregory (980911.1551 EDT)

> No matter who proposed the story, it is a lot more believable
> than the story in which the trooper sitting in his car by the
> side of the role is controlling my speed.

You can't tell what a person is doing (controlling) by just
watching what they are doing (acting). Just because a trooper
is acting in a particular way (sitting in his car) doesn't mean
that he _is_ or _is not_ doing something (controlling speed).

Perhaps you find it difficult to believe that a trooper sitting
in his car _could_ be controlling your speed because you look at
behavior as visible actions and their consequences.

It's not his lack of action that bothers me, but his lack of an external
linkage capable of altering my speed. A fisherman is fishing even though he
is "not doing anything". Once the fish bites the fisherman can (usually)
control the fish's location because of the line connecting the two of them.
The trooper however is in a different situation. He can either leave me to
my own devices or he can pull me over and give me a ticket. I, on the other
hand, can maintain any speed that my car is capable of producing. This is
the reason that I have trouble accepting that he is controlling the speed of
my car (other than by employing the Marken-Abbott device of getting three of
his comrades to ride with him blocking all four lanes of the turnpike and
slowing down to some speed they agree on).

On a related topic. If I decide, at the principle or system
level, to allow the world to be exactly as it is, i.e., to
adjust my reference levels to match my perceptions, am I
thereby controlling the world? (Sorry Phil.)

This is an interesting question. As you've described it you
are not controlling the world (or your perceptions of that
world). What you are controlling is the perception of a
relationship between perceptual variables and reference variables
and (according to your description) you are doing this only by
adjusting the references. So you are adjusting what you want
(references) to match what you are getting (perceptions).
This is a control process, one that requires a control system
(in you) that can control a perception of the relationship
between a perception and its reference.

I raised this question because it contains a major element of my unhappiness
with the trooper controlling my speed. He is really capable of controlling
my speed _only_ if he wants it be zero so that he can give me a ticket. That
is, he has no freedom to set his reference level for my speed to be anything
other than (1) whatever it happens to be, or (2) 0 mph. I grant this is a
form of control, but a very constrained form of control.

Bruce Gregory

[From Rick Marken (980911.1515)]

Bruce Gregory (980911.1735 EDT)

The trooper however is in a different situation. He can either leave me to
my own devices or he can pull me over and give me a ticket. I, on the other
hand, can maintain any speed that my car is capable of producing. This is
the reason that I have trouble accepting that he is controlling the speed
of my car

This is getting surrealistic. The trooper thing was a _story_. You made
it up! Did you make up a trooper who was controlling the speed of your
car? If so, then any trouble you have accepting that the trooper is
controlling the speed of your car is something you will have to take up
with your own imagination. If you are having trouble accepting that some
_real_ trooper you know is controlling your speed, then you can only
reduce your anxiety about this by Testing to determine whether he
is controlling this variable or not.

I raised this question because it contains a major element of my
unhappiness with the trooper controlling my speed. He is really
capable of controlling my speed _only_ if he wants it be zero

If you say so.

so that he can give me a ticket. That is, he has no freedom to set
his reference level for my speed to be anything other than (1)
whatever it happens to be, or (2) 0 mph.

Actually, he has the freedom to set the reference for your speed to
anything he needs it to be to make the higher level controlled
perception (speed AND ticket) TRUE. This is the way hierarchical
control works; a goal at level N specifies the perception to be
perceived at level N but it doesn't specify a state for the goals
(or actions, if N = 1) at level N-1 that produce that perception.

Best

Rick

路路路

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

[From Bruce Gregory (980911.2240 EDT)]

Rick Marken (980911.1515)

This is getting surrealistic.

I'm glad I'm not the only one who feels that way.

Bruce Gregory