Essay from Bill Williams

[From Rick Marken (2001.10.22.1130)]

Bill Williams (19 October 20001 5:00 CST) --

There is no more justification for the economists GNP numbers
than there is for the psychologists I.Q. numbers. Mises, however,
is not alone in recognizing this problem.

What is the problem with GNP numbers (or with IQ numbers, for that matter)?
GNP certainly seems like a reasonable (and justifiable) aggregate measure of
economic performance. What's the problem? What's the solution?

And, if the economy is
a closed-loop control system, I suppose it is fair to ask, then
"What happened to the loopgain in 1929 and in the following
great depression?"

Yes, it certainly is fair to ask that question. But I think one should first
ask whether there is evidence that 1929 resulted from a change in loop gain.
T. C. Powers presents evidence that 1929 was not a loop gain phenomenon but,
rather, a simple response to disturbance. The variable disturbed, according
to TCP's model, was aggregate income (GNI). It was disturbed by leakage
created by severe maldistribution of wealth. The stock sell off was a
compensating action that replaced income that investors had lost to leakage.

I'm not sure what, if any answer there is to the problem of
economic data. The numbers are in some sense just made up.

I agree that it's difficult to know what the best aggregate measures of an
economy are. But models will help; models are based on existing data but the
models can also suggest new kinds of data that should be collected. Of
course, that doesn't answer the question of _how_ to measure the variables.
It's hard to know how best to measure aggregate variables like GNP. But even
if economic data collection is difficult I still think economics should be
based on data. Without data against which to test models all we have in
economics are baseless conclusions handed down by "authorities" like von
Mises.

Best regards

Rick

···

--
Richard S. Marken, Ph.D.
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 Williams 22 October 2001 5:00 CST]

[From Rick Marken (2001.10.22.1130)]

> Bill Williams (19 October 20001 5:00 CST) --

> There is no more justification for the economists GNP numbers
> than there is for the psychologists I.Q. numbers. Mises, however,
> is not alone in recognizing this problem.

What is the problem with GNP numbers (or with IQ numbers, for that matter)?

I certainly not an expert on IQ numbers, but I thought Steve Gould's
_Mis-Measure of Man_ book, especially the revised edition expressed a
correct criticism of the methods used in constructing I.Q. tests. Perhaps
someone else can explain the arguments I've forgotten.

The GNP numbers are put together using a kind of utilitarian weighting,
with new goods entered at current prices, to correct for problems in the
GNP numbers alternative indexes such as the UN index of human development
measures have been developed. The Liebhausky book in the refs to the
essay has a good, if dated, discussion of the problem under the caption
"index numbers."

GNP certainly seems like a reasonable (and justifiable) aggregate measure of
economic performance. What's the problem? What's the solution?

An instance of the problem can be illustrated by comparing a poor person in
a slum in the united states having a life that is inferior in many ways to
a very much poorer ( GNP measure ) person living in a community in India.
By inferior, I mean living a shorter life, having a higher chance of being
killed or beaten, being less able to, and in fact reading less, being more
likely to live in a disordered, or broken family, etc ...

Despite the problems, I use GNP numbers. But, I cross my fingers when I do
so, because the process by which the numbers are generated assumes a
utilitarian consumer.

> I'm not sure what, if any answer there is to the problem of
> economic data. The numbers are in some sense just made up.

I agree that it's difficult to know what the best aggregate measures of an
economy are. But models will help; models are based on existing data but the
models can also suggest new kinds of data that should be collected.

This would certainly be the preferrred solution. But, I am doubtful about the
prospects for constructing a method for going from individual preferences to a
cummulative index.

Of

course, that doesn't answer the question of _how_ to measure the variables.
It's hard to know how best to measure aggregate variables like GNP. But even
if economic data collection is difficult I still think economics should be
based on data. Without data against which to test models all we have in
economics are baseless conclusions handed down by "authorities" like von
Mises.

I am fully in aggreement that Mises does not provide a foundation for a
control theory reconstruction of economics. However, some of his critical
arguments appear to be worth considering. The actual usefullness of the
existing data, may be better than the theory which is involved in its
construction, but its nothing to cheer about. Sen has some pretty good
things to say about some of these issues.

Best
  Bill Williams

···

______________________________________________________________________
Do you want a free e-mail for life ? Get it at http://www.email.ro/

[From Fred Nickols (2001.10.21.0615)] --

>Rick Marken (2001.10.17.1400)]
>
>Fred Nickols (2001.10.17.0958) --
>
> > I'm certainly not all the way through his book by any
> > means, but it seems to me that the view of economics set
> > forth by Ludwig von Mises comes awfully close to being a
> > control theory view of people and markets. See, for
> > example, his book: Human Action - A Treatise on
> > Economics.
>
>What is it about von Mises' economics that leads you to this conclusion?
>I've read some things by and about von Mises and I didn't get the
>impression that von Mises came close to having a "control theory view"
>of economics.

Well, as is often the case, my enthusiasm got the better of me. In my
remark above, I probably should have stopped after people and not included
"markets." von Mises clearly puts economics in the context of a science of
human action instead of situating human action in the context of economics
and so I am incorrect if I implied that he was taking a control theory view
of economics (which I never exactly said anyway). That aside, I'd also be
incorrect if I said that von Mises took a control theory view of anything.

I think Fred may be conceeding too much here in saying, that von MIses [never]
took a control theory view of anything. It seems to me that any conception of
human experience that is of any consequence must include within it some of the
elements of control theory. Economics makes some stringent demands upon a
theory of behavior. The simple requirement that economics makes that a theory
of behavior ought to be able to account for supply and demand, the interactions
of price and quanities traded is beyond most psychological theories. So, far
only orthodox economics has presented anything approaching an analysis which is
capable of explaining supply and demand. The Austrian school of which Mises is
considered a part, may come closer than the Anglo-Saxon version of orthodoxy to
having provided a psychological foundation for economics. Von MIses published a
book 1962 _The Ultimate Foundations of Economic Science: an Essay on Method_. I
wouldn't at all deny that it contains a lot of stuff that is, or at anyrate
suggests a control thoery analysis. But, I would agree with Rick, that only
control theory provides a consistent foundation for reconstructing economics or
social theory.

best
  Bill Wllliams

···

______________________________________________________________________
Do you want a free e-mail for life ? Get it at http://www.email.ro/

[From Rick Marken (2001.10.22.2300)]

Me:

What is the problem with GNP numbers (or with IQ numbers, for that matter)?

Bill Williams (22 October 2001 5:00 CST) --

I certainly not an expert on IQ numbers, but I thought Steve Gould's
_Mis-Measure of Man_ book, especially the revised edition expressed a
correct criticism of the methods used in constructing I.Q. tests.

The GNP numbers are put together using a kind of utilitarian weighting,
with new goods entered at current prices, to correct for problems in the
GNP numbers alternative indexes such as the UN index of human development
measures have been developed. The Liebhausky book in the refs to the
essay has a good, if dated, discussion of the problem under the caption
"index numbers."

These are two different _kinds_ of criticism of these numbers. Gould's
criticism of IQ numbers concerns their use as a measure of an imagined
characteristic of the mind called "intelligence". Liebhausky's criticism
of GNP numbers concerns their validity as an estimate of a true value,
the actual GNP. I don't think anyone has criticized the validity of IQ
numbers as an estimate of one's true score on a test. Gould's criticism
of IQ numbers is more like Sen's criticism of GNP numbers. Gould says IQ
numbers don't really measure intelligence and Sen says GNP numbers don't
really measure prosperity. But both would agree that the IQ and GNP
numbers themselves are reasonably good estimates of one's actual test
score and the nation's actual GNP, respectively.

Me:

GNP certainly seems like a reasonable (and justifiable) aggregate measure of
economic performance. What's the problem? What's the solution?

Bill:

An instance of the problem can be illustrated by comparing a poor person in
a slum in the united states having a life that is inferior in many ways to
a very much poorer ( GNP measure ) person living in a community in India.

Yes. I agree. GNP may not be the right measure of "wealth" (this is
precisely the Sen criticism of GNP). But I think GNP is a reasonably
good measure of what it purports to measure: the total cost of
production of all goods and services produced by a nation in a
particular time interval.

Despite the problems, I use GNP numbers. But, I cross my fingers when I do
so, because the process by which the numbers are generated assumes a
utilitarian consumer.

Yes. This could certainly affect the accuracy of the estimate of true
GNP. But the GNP numbers are pretty big; even if they're off by a few
billion dollars.

Me:

I agree that it's difficult to know what the best aggregate measures of an
economy are. But models will help; models are based on existing data but the
models can also suggest new kinds of data that should be collected.

Bill:

I am doubtful about the prospects for constructing a method for going from
individual preferences to a cummulative index.

I don't understand why you are doubtful. We have shown how aggregate
(cumulative) control phenomena, like an aggregate "virtual" preference
for certain goods, can emerge from the behavior of a collection of
individual control system acting simultaneously in the same environment.
I think you have shown something quite similar with your conflict
program. I think economics will eventually be able to derive laws of the
collective based on an understanding of how groups of individual control
systems interact.

Best regards

Rick

···

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

[From Bill Williams 23 October 20001 8:30 CST]

[From Rick Marken (2001.10.22.2300)]

I don't understand why you are doubtful. We have shown how aggregate
(cumulative) control phenomena, like an aggregate "virtual" preference
for certain goods, can emerge from the behavior of a collection of
individual control system acting simultaneously in the same environment.

This is a very different sort of system than I was thinking about. And,
if such measures were constructed I think some of the results might
generate significantly different numbers than the existing conventional
measures. While I don't doubt that it could be done, It wouldn't be a trival
job constructing such measures.

I think you have shown something quite similar with your conflict
program. I think economics will eventually be able to derive laws of the
collective based on an understanding of how groups of individual control
systems interact.

OK, it may help to distinguish two sides of the problem. Perhaps the cost side
of the GNP estimates are fairly good. But, from the demand or welfare side,
there may be a big difference in what's been called "household production." If
you consider a family who spend a big part of their budget on whisky, cigaretes,
high fat, high salt, high sugar food, with a Seventh Day Adventist family-- no
whisky, no smoking, high quality vegetarian diet there's going to be a big
difference in their health status. Sen, I am confident , would prefer some
"real" measure of health as a "output" over an "input" measure such as the
dollar value of the consumption goods consumed. However, going to a measure of
"welfare" in terms of health status is subject to criticism by someone who will
say that they like their whiskey, cigars, and high fat, etc food. THey would
feel deprived if they didn't have such a life style. So, who are we to tell
people how they should live their lives? This isn't "objective science" some
will say, and they'd be right I suppose. But, would it be "good" science? I
tend to think so. If they want to make comments about "life style" police
that's OK. In practice I think care would have to be exercised in how such
measures were implemented. A good deal of care and caution.

Maybe the above goes off in a different direction than your concerns. But, it
seems consistent with the direction Sen's work has taken.

Best
  Bill Williams

···

______________________________________________________________________
Do you want a free e-mail for life ? Get it at http://www.email.ro/

[From Bill Powers (2001.10.23.0922 MDT)]

Bill Williams 23 October 20001 8:30 CST--

Perhaps the cost side
of the GNP estimates are fairly good. But, from the demand or welfare side,
there may be a big difference in what's been called "household

production." If

you consider a family who spend a big part of their budget on whisky,

cigaretes,

high fat, high salt, high sugar food, with a Seventh Day Adventist

family-- no

whisky, no smoking, high quality vegetarian diet there's going to be a big
difference in their health status.

This is another case where thinking in terms of levels of control may
resolve some problems or at least put them in proper perspective. As you
put it, the "cost side" of GNP estimates may be fairly good. But they do
not belong at the same level as the "health status" considerations. From
the standpoint of the consumer, the cost side is determined in part by the
particular goods and services that consumers want and are willing to spend
their money (credit for labor or investment) on. But it is also determined
by prices, which are raised and lowered by producers (and distributors) who
are controlling for income.

It may be that if consumers buy all the goods and services they want at
prices they can afford, the result is a growing economy measured in GNP,
but one that produces neither physical nor mental health.The latter is a
higher-order concern, independent of the mechanics of market transactions.
"What people want" has two effects: one is the effect on themselves of
getting (or not getting) what they want, and the other is the effect on the
producers and sellers of goods and services, who have to provide what the
consumer wants at an affordable price in order to stay in business. Just to
give an extreme example, suppose people come to like the sweet taste of
lead in paint. The market for leaded paint will adjust as suppliers vie for
position, and a balance will be reached at which leaded paint for human
consumption is provided at just the price the consumers are willing and
able to pay for it. Of course this will lead to a nation suffering from
lead poisoning, but that has nothing to do with market forces. The people
will suffer from lead poisoning not because of the market but because of
what they want to eat.

Just as in working with the behavior of a single person, we can separate
low-order aspects of market interactions from higher-order aspects. We can
analyze lower orders of organization without necessarily being able to
analyze the higher orders. I can rather easily (after all this time)
analyze the control loops with which you move a mouse to track a target,
but I have no techniques yet for analyzing how or why you cooperate with me
by serving as a participant in a tracking experiment. In the same way, we
ought to be able to analyze the mechanics of buying and selling, producing
and consuming, without yet being able to analyze the choices of what is
bought and what is sold in relation to human welfare in the larger sense.
My little program for simulating bargaining (rounds of bidding and asking)
does not specify what is wanted or given by either side; it's a low-level
module that would be part of any market model in which bargaining is to occur.

I think it would be ambitious enough for us to aim at analyzing and
modeling the mechanics
of production and consumption, leaving the question of whether the goods
and services involved are to our benefit for the time when we've solved the
easier problems.

Best,

Bill P.

[From Bruce Gregory (2001.1023.1349)]

Bill Powers (2001.10.23.0922 MDT)

The people
will suffer from lead poisoning not because of the market but because of
what they want to eat.

What consumers want is shaped by interactions with others. Seeing others
enjoy the sweet taste of lead in paint and being bombarded by adds for
good-tasting lead paint seems likely to me at least to have some affect on
consumer preferences.

[From Bruce Gregory (2001.1023.1400)]

[From Rick Marken (2001.10.22.2300)]

These are two different _kinds_ of criticism of these numbers. Gould's
criticism of IQ numbers concerns their use as a measure of an imagined
characteristic of the mind called "intelligence". Liebhausky's criticism
of GNP numbers concerns their validity as an estimate of a true value,
the actual GNP. I don't think anyone has criticized the validity of IQ
numbers as an estimate of one's true score on a test. Gould's criticism
of IQ numbers is more like Sen's criticism of GNP numbers. Gould says IQ
numbers don't really measure intelligence and Sen says GNP numbers don't
really measure prosperity. But both would agree that the IQ and GNP
numbers themselves are reasonably good estimates of one's actual test
score and the nation's actual GNP, respectively.

To anyone who is interested in what IQ does and does not estimate, I
recommend Arthur Jensen's _The g Factor_. Jensen lays out all the studies
and what they allow and do not allow one to infer. (The evidence for
Spearman's "g" -- IQ is heavily g-loaded--seems to me at least to be beyond
doubt.) Jensen has been the subject of much criticism, but as far as I can
tell, this is yet another example of our penchant for punishing the bearer
of bad news. (On a personal note, I find very little evidence to contradict
the offensive hypothesis that what we call the educational system is
actually nothing more or less than a g-filter.)

[From Bill Powers (2001.10.231633 MDT)]

Bruce Gregory (2001.1023.1349)--

What consumers want is shaped by interactions with others. Seeing others
enjoy the sweet taste of lead in paint and being bombarded by adds for
good-tasting lead paint seems likely to me at least to have some affect on
consumer preferences.

Possibly. I wasn't trying to account for consumer preferences (reference
signals), however. I trust you agree that once the reference signal (the
want) is in place, it determines what the person will try to attain.

Actually, the people who used to eat leaded paint the most were children,
and I doubt that they were bombarded by ads for it or saw a lot of others
enjoying the sweet taste. I think it's one of those things you discovered
for yourself.

I think that what consumers want is probably not "shaped by interactions
with others." That's a little too much of a shortcut, and it ignores many
other reasons people have for wanting things. One thing that can give the
appearance of such shaping is telling people that some unsatisfied want
they have will be taken care of by X. If the statement were true, then this
kind of shaping would be beneficial, though I wouldn't use the term
shaping. Unfortunately, in the context suggested by calling people
"consumers," it probably not true much of the time.

Best,

Bill P.

[From Bruce Gregory (2001.1024.1018)]

Bill Powers (2001.10.231633 MDT)

Actually, the people who used to eat leaded paint the most were children,
and I doubt that they were bombarded by ads for it or saw a lot of others
enjoying the sweet taste. I think it's one of those things you discovered
for yourself.

Perfectly reasonable, just not the scenario you described.

I think that what consumers want is probably not "shaped by interactions
with others."

Oh? I wonder where the wants for the latest technology and luxury cars come
from? They must be the result of reorganization. I wonder how hunter
gathers come to reorganize to want Bose speakers and Lexus cars?

That's a little too much of a shortcut, and it ignores many
other reasons people have for wanting things. One thing that can give the
appearance of such shaping is telling people that some unsatisfied want
they have will be taken care of by X.

Such as telling a Bushman that his unsatisfied want for a luxury car will
be taken care of by buying a Lexus?

If the statement were true, then this
kind of shaping would be beneficial, though I wouldn't use the term
shaping. Unfortunately, in the context suggested by calling people
"consumers," it probably not true much of the time.

I agree. I was simply suggesting an alternative.

[From Bill Powers (2001.10.24.0903 MDT)]

Oh? I wonder where the wants for the latest technology and luxury cars come
from? They must be the result of reorganization. I wonder how hunter
gathers come to reorganize to want Bose speakers and Lexus cars?

When I said that you were taking a shortcut, what I meant was that there's
probably more to getting people to buy things than just implanting a
reference signal for a specific item in their heads (which I don't think is
possible). Bose speakers and Lexus automobiles are only means to ends, and
people have many ends they pursue at the same time. If a person wants
affordable high-fi, and Bose speakers satisfy that want, then clearly it is
not the Bose speakers the person wants, but the high fidelity and the low
price, by any name. The origins of wanting to hear high fidelity sound and
wanting to conserve buying power are not necessarily found in influences
from others. If I like music I am likely to prefer hi- over lo-fi speakers,
and if my budget is limited I am likely to prefer less- over more-expensive.

Of course there are prestige (and similar) factors to consider, in some but
not all people. Perhaps you have a tin ear, and the main effect of having
Bose speakers in your bookcase is to impress the kinds of friends you
choose; what you want is to see admiration, approval, or envy in their
reactions. But even in that case, it is not the Bose speakers (or the Lexus
automobile) you want per se, but the appearance of admiration etc. in your
visitors. You'd buy anything you thought might provoke that reaction. And
the buck doesn't even stop there, because we can wonder why it is you need
admiration etc. so much; in the case of the Lexus, why you are willing to
sacrifice so much of your buying power to get them if it's not the car
itself you want for your own reasons. I won't conjecture about the reasons,
but clearly they would be things that would involve much more than simple
"influences of others."

The influence of others on yourself, I think, turns out to be based mainly
on what you want from others; you'll do what it takes to get it. And if you
keep asking why you want that last thing that turned up, sooner or later
you'll come to a want that is strictly your own private property, that you
can't blame on anyone else. To me, this calls into question the quick
judgment that people want things just because others cause them to want them.

Best,

Bill P.
'

[From Bruce Gregory (2001.1024.1200)]

Bill Powers (2001.10.24.0903 MDT)

The influence of others on yourself, I think, turns out to be based mainly
on what you want from others; you'll do what it takes to get it.

I agree.

And if you
keep asking why you want that last thing that turned up, sooner or later
you'll come to a want that is strictly your own private property, that you
can't blame on anyone else. To me, this calls into question the quick
judgment that people want things just because others cause them to want them.

I would hope so. It never occurred to me that you want things because
others cause you to want them. My reasoning was identical to yours: the
influence of others on yourself depends mainly on what you want from
others. Advertising offers, but does not compel, solutions to the problem
of getting what you want from others.

[From Bruce Nevin (2001.10.24 13:28 EDT)]

Rick Marken (2001.10.18.2220)--
>I just built a very simple example of a virtual control system of
>this sort. I set up a set of 10 real control systems all controlling the
>same perception relative to different reference specifications for the
>variable they control. This collection of control systems ends up
>keeping the common controlled variable in a virtual reference state that
>happens to correspond to the average reference specification for the
>collective.

I don't think this is quite the right structure. You have each controlling a value of the same variable, and therefore each in conflict with all the others as to that value. Rather, each controls a value of a different variable (or of a variable locally shared by the other participant[s] in a given transaction). Other functions in the environment (not all of which are within peer control systems) relate these local variables to one another, whence your aggregate values. Any peer-to-peer conflicts as to the value of a controlled variable are local, not aggregate. They may also be in conflict with the aggregated value. I want real estate values to go down if I'm buying or being assessed for taxes, up if I'm selling or borrowing against equity. But these aggregate values of variables cannot be modelled by putting the individuals in conflict and calling the average of their conflict an aggregate value.

  Bruce Nevin

···

At 22:20 10/18/2001 -0700, Richard Marken wrote:

[From Bruce Nevin 2001.10.24 14:28 EDT]

Rick Marken (2001.10.19.1200) --

> Bill Powers (2001.10.19.0554 MDT)--

> Economic laws must be like gas laws
> in physics, where laws connecting temperature, pressure, and volume (while
> perfectly valid) can be shown to arise from the interactions of moving,
> spinning, semi-elastic molecules.

Yes. But the gas laws (the analog of the aggregate control phenomena) are
still useful in themselves. [...] they would certainly want to
understand the latter as validation of the former).

> In studying Lovelock's Daisy World, I
> realized that it is possible for a global regulatory system to arise from a
> collection of local control systems.

This is a perfect analog of the aggregate economic "virtual control
system" that, I think, arises from a collection of individual control systems (controlling for consumption by contributing to production).

An appealing analogy. Notice that the model articulated by the gas laws is different in form from the model of molecular movement. Likewise, the model of the local control systems is different from the structure of a global regulatory system that might emerge from it. There is *no* reason a priori to suppose that an economy is homomorphous with the control systems that constitute it, and these appealing analogies strongly suggest that they are different in form.

         Bruce Nevin

···

At 11:59 10/19/2001 -0700, Richard Marken wrote:

[From Rick Marken (2001.10.24.1400)]

Me:

>I just built a very simple example of a virtual control system of
>this sort. I set up a set of 10 real control systems all controlling the
>same perception relative to different reference specifications for the
>variable they control. This collection of control systems ends up
>keeping the common controlled variable in a virtual reference state that
>happens to correspond to the average reference specification for the
>collective.

Bruce Nevin (2001.10.24 13:28 EDT)

I don't think this is quite the right structure.

It's a structure of control systems that demonstrates the concept of a virtual
reference state for a variable. That's what I wanted it to demonstrate. Why
do you think it's not quite right?

Best

Rick

···

--
Richard S. Marken, Ph.D.
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 (2001.10.24.1350)]

Me:

>This is a perfect analog of the aggregate economic "virtual control
>system" that, I think, arises from a collection of individual control
>systems (controlling for consumption by contributing to production).

Bruce Nevin (2001.10.24 14:28 EDT)

An appealing analogy. Notice that the model articulated by the gas laws is
different in form from the model of molecular movement.

It depends on what you mean by "different in form". Different equations describe
the laws. But the gas and molecular laws are of the same "form" in the sense that
they are both cause-effect laws. Same is true for actual and virtual control
systems. Different equations describe the models; but the actual and virtual
(aggregate) control models are of the same form: closed loop input control models.

Best regards

Rick

···

--
Richard S. Marken, Ph.D.
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 Bruce Nevin (2001.10.24 21:09 EDT)]

Rick Marken (2001.10.24.1400) --

···

At 13:39 10/24/2001 -0700, Richard Marken wrote:

Me:

> >I just built a very simple example of a virtual control system of
> >this sort. I set up a set of 10 real control systems all controlling the
> >same perception relative to different reference specifications for the
> >variable they control. This collection of control systems ends up
> >keeping the common controlled variable in a virtual reference state that
> >happens to correspond to the average reference specification for the
> >collective.

Bruce Nevin (2001.10.24 13:28 EDT)

> I don't think this is quite the right structure.

It's a structure of control systems that demonstrates the concept of a virtual
reference state for a variable. That's what I wanted it to demonstrate. Why
do you think it's not quite right?

If we're talking about how the appearance of an aggregate value can arise from autonomous control systems each controlling different but environmentally related variables, then it's the wrong model. My mistake. I didn't realize that you preferred to model conflict among many control systems over the same variable.

         Bruce Nevin

[From Bill Williams 24 October 2001 9:00 CST]

[From Bruce Gregory (2001.1024.1200)]

> Bill Powers (2001.10.24.0903 MDT)

> sooner or later
>you'll come to a want that is strictly your own private property, that you
>can't blame on anyone else. To me, this calls into question the quick
>judgment that people want things just because others cause them to want them.

I would hope so. It never occurred to me that you want things because
others cause you to want them. My reasoning was identical to yours: the
influence of others on yourself depends mainly on what you want from
others. Advertising offers, but does not compel, solutions to the problem
of getting what you want from others.

It seems to me that Bruce's statement concerning advertising could be extended
to include the relationship between the self and a school, a church, or what
ever. Choices made concerning the problem of "what to do" are rendered by, and
only by, the self. But, the contexts in which the choices which a self makes
are in large part offered or determined by others. So, if you don't wear
control theory glasses, its going to look as if those who have controlled the
context in which the choice is made have controlled the choice. So if a bar
heavily salts the food, it is still the customer who decides what and how much
to drink, but the choice is being made in a context of the consumption of
heavily salted food. If the customers wish to "blame" someone maybe they ought
to blame themselves for hanging out in such an establishment.

An insistence that the context in which a choice is made is vitally important
has so frequently been argued from the mistaken psychological premise of
behaviorism that it may sound as if it is inherently unsound. But it is not
neccesarily the result of "a quick judgement" to the effect that other people's
actions "cause", in a control theory sense of the term, us to do something or
anything.

thanks Bruce
   Bill Williams

···

______________________________________________________________________
Do you want a free e-mail for life ? Get it at http://www.email.ro/

[From Bill Williams 24 October 2001 10:30 CST]

[From Bill Powers (2001.10.23.0922 MDT)]

suppose people come to like the sweet taste of lead in paint.

Just suppose that there is a vast advertising industry putting up billboards,
handling out taste samples, creating icons through the use of popular
"personalities", vending hip gear complete with logos, and all the associated
mythology that the most creative minds money can buy can generate. Just suppose
they find one of the most addictive products around to focus their energies
upon. It seems to me that the question of "how come" some many people come to
like the taste is a question worth exploring. ( See the recent book _Ashes to
Ashes_ on the tobacco industry. )

I think it would be ambitious enough for us to aim at analyzing and
modeling the mechanics
of production and consumption, leaving the question of whether the goods
and services involved are to our benefit for the time when we've solved the
easier problems.

I think what you may not be considering in your judgement about which problems
to work on first is the issue of how many "we" there are to work on problems
may depend upon which problems "we" choose to work on. The "life style"
questions are vital questions in which a fair number of people are intensely
interested. According to control theory the good news is that people have the
capacity to resist efforts to "shape" their choices in ways that are harmful to
their own welfare. Glasser's little book along these lines_Positive Addictions_
has a small but fairly devoted following. Thus the question of "benefits" it
seems to me is a matter that is worth exploring now, rather than waiting until
a time when we have run out of easier problems. The idea that we have the
capacity to "choose" what we are going to want, it seems to me is a "fact" that
is worth emphasizing. And, it is a question that can be explored both as a
"practical matter" in terms of what works and what doesn't in terms of life
style programs and coaching as well as in terms of a "purer" research involving
control theory mechanics as in the work we've been collaborating upon in rather
abstract models of interaction between goals and levels of goals.

Best
  Bill Williams

···

______________________________________________________________________
Do you want a free e-mail for life ? Get it at http://www.email.ro/

[From Bill Powers (2001.10.25.0556 MDT)]

Choices made concerning the problem of "what to do" are rendered by, and
only by, the self. But, the contexts in which the choices which a self makes
are in large part offered or determined by others. So, if you don't wear
control theory glasses, its going to look as if those who have controlled the
context in which the choice is made have controlled the choice.

True. Allow me to paraphrase and expand on this theme.

This same consideration extends to _all_ contexts of experience. If you
don't wear control-theory glasses, it will seem that both the physical and
the social world determine what animals and people choose to do. Each of us
lives in a world that has certain properties. Whatever we end up wanting to
experience, we must act on the world as it is, including the living systems
in it, to get what we want. Further, since we must never forget the
hierarchical nature of the human goal structure, when we look for means of
achieving a goal at one level, we are also thereby looking for an
experience to want at a lower level. We must pick those lower-level wants
that will, if achieved, accomplish the higher-level goal in the world as it
is. This is how we become entangled with other people, for other people
must use the same environment (which includes ourselves) to get what they
want, and the means they select often interferes with our own selection
(the "occupied" sign in that little door in the airplane). Furthermore,
other people often _are_ the means we choose: "Please pass the salt."

We also have to remember that many people make their livings by trying to
control what other people choose, and enjoy doing so. By means both clever
and crude, they get control over things that other people need, or convince
other people that what they already need or want, but can't get for
themselves, can be obtained just by buying something or hiring some
service. Sometimes, of course, this is true -- you can definitely avoid bad
breath, and its effects on people around you, by buying some kind of
toothpaste and brushing your teeth with it. You can even attract girls -- a
particular sort of brainwashed, ignorant, and deluded girls, perhaps, but
girls nonetheless -- by driving the right kind of car. Advertising has to
contain a grain of truth if it's to work for very long and if the
advertiser is to go on convincing people that it's generally worth buying
things from that person.

But the point we have to keep reminding ourselves of is that nobody can
reach into anyone else's head and tweak the reference signals there, and
even if anyone could do this, systems at a higher level would tweak them
right back again. Any arbitrary change in any reference signal at any level
(including the highest -- work that one out) amounts to a disturbance,
because it changes what is going on in another system somewhere, and will
elicit resistance.

Another thing -- this word "choice." I don't think it's a robust enough
term to carry the load in this discussion. It's hard to be sure what a
person means by this term, and often what is meant is contrary to fact.

First, it can mean acting voluntarily rather than automatically or because
of coercion. I think that's how Glasser uses it in "choice theory."

Second, it can mean selecting one object or course of action among several.
Implied but seldom discussed is the idea that the alternatives are mutually
exclusive -- you can select _only one_, and when it's selected the other
possibilities are ruled out.

Third, because of the point above choice implies conflict. because if one
could choose _both_ or _all_ of the alternatives, "choice" wouldn't mean
anything. You have to make a choice only when you want _both_ alternatives
at the same time, but it is impossible to _have_ both. I have to choose
between having my cake later, and eating it now.

Fourth, it's used whenever there is more than one possible outcome or
alternative, WHETHER OR NOT THE ACTOR CONSIDERS ANY ALTERNATIVE OTHER THAN
THE ONE "CHOSEN." When, most mornings, Mary walks down the path to the main
road, she turns left at the main road and goes to the place where the
newspaper is delivered in its little box. I could say, "Ah, I see you have
chosen to turn left this morning." To say that is to imply that she was
presented with the alternatives of turning left and turning right, and that
after considering both possibilities, she chose to turn left. After all,
_I_ could plainly see the alternatives; surely she could too.

But no. When I ask her a little time from now, I'm sure she will say that
she does not choose left over right. She never even considers turning
right, because that's not the way to where the newspaper is. To get to
where the newspaper is, you turn left from the front door down the
driveway, right down the path, and left toward the newspaper box. The
control systems involved do not choose between alternatives. They simply
convert any remaining error into the appropriate action, and turn in the
directuion toward the newspaper box.

The point is, of course, that we read "choice" into situations where no
choice is actually involved, just because _we_ can see alternatives.

I wouyld much prefer to use the term choice _only_ when when there is a
conflict. Normally, we avoid having to make choices. We simply do what
works to get us what we want.

Best,

Bill P.