PCT as physiology-psychology link

CSGnet Archive CSG2002
21

[From Fred Nickols (2002.09.24.0400)] --

I'm not about to mix into the exchange between Bill Williams and Rick
Marken, however, Bill's closing question to Rick raises a notion I'd like
to pursue.

[From Bill Williams UMKC 24 September 2002 12:33

<snip>

So, my question for you would be how would you justify from a control theory
perspective an argument that external circumstance is the cause of behavior.

This whole cause and effect issue has always troubled me in discussions on
this list. Here's why...

Cause and effect has always seemed to me to be a linear construct,
something like the following:

         cause ---> effect

That linearity can be strung out in longer series where the effect of a
cause becomes the cause of the next effect in the series. This view, I
think, is what leads to the search for so-called "root" causes.

         cause --------> effect
                         cause --------> effect
                                         cause --------> effect

In any event, it has always seemed to me that cause and effect are relative
instead of absolute notions. Nothing, taken out of context, can be
classified as cause or effect; this can only be done with respect to one
thing in relation to another.

When I fold this serial view around in a loop, I get something like the
following:

                 cause --------> effect
                 effect cause
                     ^ |
                     > >
                     > v
                 cause effect
                 effect <-------- cause

That's why when I wrote the systems view poem so many years ago (right
after reading B:CP for the first time) it contained this line: " . . .
there is no beginning and no end, no cause, no effect, not one can be found
in closed loops going around and around."

Now, back to Bill's question to Rick: "...how would you justify from a
control theory perspective an argument that external circumstance is the
cause of behavior."

If behavior is part of a closed loop, it seems to me that any and all
cause-and-effect views of behavior, whether they posit external or internal
"causes" will invoke a linear instead of a loop view and thus continue to
miss a point that is brought up on this list on a fairly regular basis;
namely, that all parts of a closed loop are active all the time.

I don't think list members are likely to argue that the reference signal
causes behavior, nor that perceptions cause behavior, nor even that the
error signal causes behavior (although that one's tempting; almost as
tempting as "disturbance"). Why? Because of the closed-loop view I
think. It's rather like trying to identify the starting point of a
circle. You can designate one, to be sure, but it's an arbitrary exercise.

But, "disturbance" is a different matter. If I borrow the closed loop
depicted above and add "disturbance," I get something like the following:

                 cause --------> effect
                 effect cause
                     ^ |
                     > >
                     > v
                 cause effect
                 effect <-------- cause <-------- disturbance

A disturbance is outside the loop, isn't it? Are disturbances the S in the
old S --> R formulation?

Fred Nickols
nickols@safe-t.net

22

[From Bill Powers (2002.09.24.0903 MDT)]

Bill Williams UMKC 24 September 2002 12:33 --

>The question ... is consistent with a belief that an environmental

stimulus, or input can generate causal forces within me which result in my
becoming depressed or "down." And, in fact I'm not depressed ... If I
were depressed, which I'm not, from a control theory persepctve ( the
magic glasses ) it would be the result of my own internal organization,
and in a condition which I have choosen rather than some external cause--

Yes. Nice commentary. This is what we have to keep coming back to. If we
had all been raised this way from infancy, it would be easy, but I think
that all of us get blind-sided by our own reactions that we have practiced
unceasingly for decades. I know that I am causing my own feelings by what I
want in this situation, but that's in my forebrain. In fact I wish you'd
quit making me feel so bad.

The other side, however, is that once we take responsibility for our own
feelings, we still have the problem of what to do when people say and do
things that, in our opinion, damage things or people we value. We may still
get angry and want to do something energetic to make the other person shape
up -- without blaming that person for our feelings, but still blaming the
other for doing something we think is wrong. The issue here is not taking
responsibility for our feelings, but how to handle serious and even
emotional disputes. Taking reponsibility for our own feelings does not
automatically lead to agreement about the right way to do things,
correctness of a theory, or validity of an approach. We still have to
figure out what to do about that.

Best,

Bill P.

23

[From Bill Powers (2002.09.24.0932 MDT)]

Fred Nickols (2002.09.24.0400) --

>Cause and effect has always seemed to me to be a linear construct,

something like the following:

        cause ---> effect

That linearity can be strung out in longer series where the effect of a
cause becomes the cause of the next effect in the series. This view, I
think, is what leads to the search for so-called "root" causes.

        cause --------> effect
                        cause --------> effect
                                        cause --------> effect

In any event, it has always seemed to me that cause and effect are relative
instead of absolute notions. Nothing, taken out of context, can be
classified as cause or effect; this can only be done with respect to one
thing in relation to another.

You're right about there being difficulties with the idea of cause and
effect. One version you've left out is _multiple_ causation:

cause 1 ------->effect<--------cause 2

This shows up in your last diagram where both the system's action and an
independent disturbance contribute to the state of the controlled quantity.

In my view, causation is simply too old-fashioned a term to help us much
any more. The way modern engineers and scientists think about nature is not
in terms of causes and effects, but in terms of relationships among
variables. As one side of the seesaw goes up, the other goes down: H(left)
= H0 - H(right). Notice that this also works the other way around: H(right)
= H0 - H(left), where H is height above the ground and H0 is the height of
the fulcrum. There are no "events" involved in this relationship, and
without events it's very hard to think in terms of causes and effects. Of
course you can create an event by making one side of the seesaw move
suddenly upward and stop; in that case, the other side will move suddenly
downward and stop. But it's not the moving-up event that "causes" the
moving-down event: it's the relationship expressed by either one of the
above equations that makes the events appear to be related. The equation
remains in effect no matter how the ends move, or whether they move at all.
That's how the physical sciences handle cause and effect now: you'll rarely
see those words used.

Basically, cause and effect are magical terms because they do not provide
mechanisms to explain what's going on. Flipping the light switch up causes
the light to turn on -- that's what we say when we don't want to be
explicit about how this works. It's also what we say when we _don't know_
how it works. Somehow, moving this thingie from the lower position to the
upper position causes the filament in that globular object to get
white-hot. Weird.

With the seesaw, we can say that moving one end causes the other end to
move, but here we can see the mechanism and there's nothing to explain.
We're just looking at a long plank which can tilt to different angles from
the horizontal. If one end goes up the other end HAS TO go down, unless the
plank breaks.

Causation is a term for empiricists investigating events. When you model
things, you substitute real or proposed relationships among variables for
the concept of causation.

Best,

Bill P.

···

When I fold this serial view around in a loop, I get something like the
following:

                cause --------> effect
                effect cause
                    ^ |
                    > >
                    > v
                cause effect
                effect <-------- cause

That's why when I wrote the systems view poem so many years ago (right
after reading B:CP for the first time) it contained this line: " . . .
there is no beginning and no end, no cause, no effect, not one can be found
in closed loops going around and around."

Now, back to Bill's question to Rick: "...how would you justify from a
control theory perspective an argument that external circumstance is the
cause of behavior."

If behavior is part of a closed loop, it seems to me that any and all
cause-and-effect views of behavior, whether they posit external or internal
"causes" will invoke a linear instead of a loop view and thus continue to
miss a point that is brought up on this list on a fairly regular basis;
namely, that all parts of a closed loop are active all the time.

I don't think list members are likely to argue that the reference signal
causes behavior, nor that perceptions cause behavior, nor even that the
error signal causes behavior (although that one's tempting; almost as
tempting as "disturbance"). Why? Because of the closed-loop view I
think. It's rather like trying to identify the starting point of a
circle. You can designate one, to be sure, but it's an arbitrary exercise.

But, "disturbance" is a different matter. If I borrow the closed loop
depicted above and add "disturbance," I get something like the following:

                cause --------> effect
                effect cause
                    ^ |
                    > >
                    > v
                cause effect
                effect <-------- cause <-------- disturbance

A disturbance is outside the loop, isn't it? Are disturbances the S in the
old S --> R formulation?

Fred Nickols
nickols@safe-t.net

24

[From Bruce Nevin (2002.09.24 11:59 EDT)]

Fred Nickols (2002.09.24.0400) --

                cause -------->effect
                effect cause
                    ^ |
                    > >
                    > v
                cause effect
                effect <-------- cause <-------- disturbance

A disturbance is outside the loop, isn't it? Are disturbances the S in the
old S --> R formulation?

Yup. And, furthermore, control outputs closely mirror disturbances (in order to resist them), so in that narrow and usually invisible sense the disturbances (S) do cause the control outputs (R) that cancel them out. The rub is that the specific outputs that resist disturbances are not what psychologists count as "responses" (R). With successful control, maintaining the (usually unidentified) CV at the preferred value or course of values, disturbance-resisting outputs are usually unobserved because resistance and disturbance cancel each other--no observable "behaviors" to count.

         /Bruce Nevin

···

At 04:43 AM 9/24/2002 -0400, Fred Nickols wrote:

25

[From Bill Williams UMKC 24 September 2002 10:48 CST]

[From Fred Nickols (2002.09.24.0400)] --

I'm not about to mix into the exchange between Bill Williams and Rick
Marken, however, Bill's closing question to Rick raises a notion I'd like
to pursue.

Fred, I think the point you make may be a good one. Maybe the use of the cause
and effect nominclature should be carefully reconsidered when talking about
control systems. Rephrased my question to Rick might be restated in more PCT
correct lingo as "How come you assume the loop gain starting from the standpoint
of an external circumstance is higher than the loop gain starting from the
standpoint of an internal reference level." But, then there are also questions
about how the loops are connected up.

Rick has described his loop as being connected up so that an expression of glee
over the Trade Center and Pentagon attacks causes him pain. For better or
worse I'm not connected that way. And, Rick seems to assume that his
inventiveness and sneering would somehow through an environmental loop result in
my being "down" or depressed. But, actually the effect is just the opposite. I
encounter people daily that retain even if they're not aware that they have
acquired it, the world view of behaviorism. People they say are "shaped" by
their cultural. In many cases the people that use this language of "cultural
conditioning" are not genuinely open to reconsidering the assumptions they make
about human behavior. So there is in a practical sense a rather sharp limit to
exploring the way that they are attached to such a worldview. I go a bit beyond
what is considered proper-- there's ways that they can get as they say back at
me. Like not giving me anymore paper, or screwing with my email, or deciding my
former office is needed for storge and stuff like that. So, when Rick
mistakenly assumes that what he does "causes" ( already I'm being more careful,
the scare quote surrounding the "" about nominclature ) to be down, the reverse
is actually true. Rick obviously is open to a control theory perspective, he's
been a long time advocate of control theory. But, when he's not being careful
he slips into the world view, or at anyrate the language of behaviorism. And,
this is very interesting. Maybe RIck can tell us how it happens. The answer
may be just that it is a matter of an incompletely extinguished habit. Maybe we
can help in an exploration of this issue by asking him questions. Fortunately
Rick doesn't control my supply of paper, so I can, without fear, ask any
questions I want to.

So, Fred, I think I agree with the point you made with diagram. But, its
somewhat awkward in argumentation have to draw a picture and then point to it.
Is the "Where's the loop gain?" question the verbal equlivant?

Cordially yours

Bill Williams

···

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

26

[From Fred Nickols (2002.09.24.1325)] --

[From Bill Williams UMKC 24 September 2002 10:48 CST]

<snip>

Fred, I think the point you make may be a good one. Maybe the use of the
cause
and effect nominclature should be carefully reconsidered when talking about
control systems. Rephrased my question to Rick might be restated in more PCT
correct lingo as "How come you assume the loop gain starting from the
standpoint
of an external circumstance is higher than the loop gain starting from the
standpoint of an internal reference level." But, then there are also
questions
about how the loops are connected up.

I don't understand "loop gain" yet so I can't comment. More below.

<snip>

So, Fred, I think I agree with the point you made with diagram. But, its
somewhat awkward in argumentation have to draw a picture and then point to it.
Is the "Where's the loop gain?" question the verbal equlivant?

As noted above, I don't understand "loop gain" comments yet. But I do
think that in discussing and speculating about behavior, we are probably
better off using some language and terminology that is a better fit with
PCT than some older terms such as "cause and effect," etc. I try not to
bandy the PCT terms about too much because I'm not that confident in my
grasp of them. That said, I can see utility in comments tied to controlled
variables, disturbances, error signals, reference conditions (or signals --
See? I don't have that straight in my mind(?) yet) and, of course,
everyone's favorite: perceptions. I'm fond of "intentions," too, but it
seems to me that that term can refer to only a limited subset of reference
conditions (or reference signals--which is it)?

Every now and then I see exchanges on this list (and on lots of other lists
as well) where one person posts something and another person "goes off" in
replying to the first post. I assume the first post disturbed some
variable the second poster is controlling for. On lists other than CSG, I
don't expect any subsequent postings to delve into what's going on from a
PCT view; those folks simply don't know or don't care about PCT. But, on
PCT, it seems to me that analyses of certain exchanges could benefit (all
of us) by looking at it through PCT glasses. On the other hand, mindful of
what I said about "intentions" above, I suspect that many of the variables
we attempt to control are not known to us in a conscious, rational,
deliberate way.

Suppose, for example, that someone writes something that I take as a
personal affront. Some controlled variable, let's call it my sense of
personal dignity, has been disturbed. I respond in a way that restores my
perception of that variable to its reference condition (and that can
require nothing more than a matter of me formulating and posting what I
believe is a suitable retort). But, my retort disturbs some controlled
variable on the part of the other person. He or she responds, so do I, so
does he or she, and the cycle continues. It's easy for me to generalize
that two living control systems are caught up in what we would ordinarily
call a spat or a tiff of some kind because the way they each perceives the
other's message disturbs one or more controlled variables. It gets a lot
trickier when you go digging for the controlled variables in question (and,
Yes, I know about "the test"). I'm reasonably sure in my own mind that
most of don't have a clue regarding much of what we're controlling. That's
not because we're dolts or ninnies but because so many of our reference
conditions aren't established through conscious, rational, deliberate
processes. I don't know about you but, speaking for myself, I don't recall
ever sitting down and saying to myself, "I will control for ensuring that I
am accorded the respect and dignity owed me as a human being" or any other
such possible reference condition. Yet, I will also wager that I have some
reference condition similar to that because of the hurt I feel and the
offense I take when I am subjected to words and deeds that violate that
standard or reference condition.

Now I happen to believe that human beings have a great deal in common. I
suspect that many of us have very similar reference conditions for lots and
lots of situations and circumstances. I also happen to believe that, by
and large, human beings are pretty darn effective. But I believe the upper
limits of their effectiveness are constrained and restrained by their
methods and tools, including their theories and their so-called mental
models. PCT, it seems to me, holds the promise of what I used to call
"popping the constraints," that is, breaking through to a new level of
effectiveness. To do that, we have to figure out how to apply it and, in
my view, efforts to figure out how to apply it don't have to wait on
further elaboration and proof of the theory via experimentation (although I
will happily acknowledge that efforts to apply PCT must stay aligned with
the current state of that theory).

Here's a question for you: What reference condition and what perceptual
signal would have produced the kind of error signal that I'm attempting to
cancel out with this post?

Fred Nickols
nickols@safe-t.net

27

[From Rick Marken (2002.09.24.1520)

Bill Williams UMKC 24 September 2002 12:33

> [From Rick Marken (2002.09.23.2220)]
>
> Is it the inventing or the sneering (or both) that gets you down?

Who said anything about me being down? More inventiveness on your part.

I'm glad to hear you're not down. It seemed like you were a bit down on me when
you said:

You have an exceptional capacity for inventing reasons to sneer at
people.

But I would appreciate it if you would tell me what you think I was "inventing"
when I said:

they had nothing else to do but argue since they were not busy
building and testing models of organisms?

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

28

[From Rick Marken (2002.09.24.1550)]

Bill Williams (UMKC 24 September 2002 12:33)

So, my question for you would be how would you justify from a control theory
perspective an argument that external circumstance is the cause of behavior.

I would justify it on the basis of the fact that control systems act to resist
disturbances to controlled variables. Disturbances are external circumstances
that have an effect on variables controlled by the control system. When control is
good (when a controlled variable, qi, is being kept in its reference state, r),
variations in behavior (qo) are caused, via the feedback function, f, relating
behavior to the controlled variable, by variations in disturbances (external
circumstances, d) so:

qo~ -1/f (d)

An external circumstance, d, is the cause of behavior, qo. QED

What control theory glasses let us see is that the causal connection between
external circumstance and behavior is through the _environment_ (the feedback
function, f), not the organism.

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

29

[From Rick Marken (2002.09.24.1600)]

Bill Williams (UMKC 24 September 2002 10:48 CST) --

Rick has described his loop as being connected up so that an expression of glee
over the Trade Center and Pentagon attacks causes him pain. For better or
worse I'm not connected that way.

That just means we're controlling different perceptions. The Sept. 11 attacks
caused me great pain because they were an insuperable disturbance to many
perceptions I was and am still controlling. All those perceptions were driven far
from their references by the disturbance of Sept. 11 and many have still not
returned to normal.

So, when Rick
mistakenly assumes that what he does "causes" ( already I'm being more careful,
the scare quote surrounding the "" about nominclature ) to be down, the reverse
is actually true.

So I caused you to be up? That's great!

Rick obviously is open to a control theory perspective, he's
been a long time advocate of control theory. But, when he's not being careful
he slips into the world view, or at anyrate the language of behaviorism. And,

this is very interesting. Maybe RIck can tell us how it happens.

Sure. The answer is simply that I find it awkward to keep saying things like
"reacted to a disturbance to a controlled variable" instead of "got you down (or
up)". I should have asked "was it my inventiveness or my sneering (or both) that
was a disturbance to whatever variables you are controlling". Your comment to me
-- "You have an exceptional capacity for inventing reasons to sneer at people" --
was an obvious reaction to my comment about a set of philosophers having "nothing
else to do but argue since they were not busy building and testing models of
organisms"? I asked you whether it was it my inventiveness or my sneering that
got you down in an effort to identify the controlled variable I had disturbed. I'm
still curious. That's why I want to know what you think I "invented".

The answer
may be just that it is a matter of an incompletely extinguished habit.

Well, "habit" would be an S-R explanation, wouldn't it? I think the answer is that
I am controlling for variables like "conciseness or expression" and "common
understanding of language".

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

30

[From Bill Williams UMKC 24 September 2002 4:16 PM CST]

[From Fred Nickols (2002.09.24.1325)] --

Here's a question for you: What reference condition and what perceptual
signal would have produced the kind of error signal that I'm attempting to
cancel out with this post?

Fred,

I think your question is a good one. And, I think it is a question which I
ought to take some time so that I can answer deliberately and after some
reflection. It won't take me that long, I've been thinking about the sort of
issues which I believe you are concerned-- so expect a reply but with a delay,
no later than Tuesday afternoon.

cordially yours

Bill Williams

···

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

31

[From Fred Nickols (2002.09.24.1752)] --

Okay; but, if you're going to give this some serious thought, here are a
few additional questions in the same vein.

         Does the error signal owe to a newly-established or altered
reference condition or to updated perceptual signals or to a "disturbance"?

         What kind of perceptual signal will reduce or null out the error
signal?

         How can we put all this in plain language so ordinary folks (i.e.,
non-PCT specialists) can make use of it?

Cheers,

Fred Nickols
nickols@safe-t.net

···

At 12:35 AM 9/25/2002 +0300, you wrote:

[From Bill Williams UMKC 24 September 2002 4:16 PM CST]

> [From Fred Nickols (2002.09.24.1325)] --

>
> Here's a question for you: What reference condition and what perceptual
> signal would have produced the kind of error signal that I'm attempting to
> cancel out with this post?
>
> Fred,

I think your question is a good one. And, I think it is a question which I
ought to take some time so that I can answer deliberately and after some
reflection. It won't take me that long, I've been thinking about the sort of
issues which I believe you are concerned-- so expect a reply but with a delay,
no later than Tuesday afternoon.

cordially yours

Bill Williams

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

32

[From Bill Powers (2002.09.24.1904 MDT)]

Rick Marken (2002.09.24.1550)--

> Disturbances are external circumstances

that have an effect on variables controlled by the control system. When
control is
good (when a controlled variable, qi, is being kept in its reference
state, r),
variations in behavior (qo) are caused, via the feedback function, f, relating
behavior to the controlled variable, by variations in disturbances (external
circumstances, d) so:

qo~ -1/f (d)

An external circumstance, d, is the cause of behavior, qo. QED

This equation holds only when the reference signal is constant. You can
also show that when the disturbance is constant, behavior is caused by
changes in the reference signal. In fact, behavior is jointly caused, with
the reference signal and the disturbance both contributing to it, and (when
loop gain is high) the form of the dependence being set by the external
functions: Fd and Ff.

What control theory glasses let us see is that the causal connection between
external circumstance and behavior is through the _environment_ (the feedback
function, f), not the organism.

The behavioral "reaction" to a disturbance depends on where the reference
signal is set. It can be set so a given disturbance increases the error, or
decreases it. Thus the sign of the reaction is determined by the reference
setting even for a constant reference signal. The same action we call a
"reaction" when there is a disturbance can result from a change in
reference level when there is no disturbance.

So a person's reaction to a disturbance is not totally determined by
environmental factors. It is equally determined by the goals of the person
being disturbed. The magnitude of the reaction also depends on the loop
gain, which is determined by the part of the loop internal to the person
just as much as by the part that is external. A person can reduce the loop
gain to zero independently of the gain in the external feedback loop,
because loop gain is the product of the gains in each part of the loop. As
a result, we are free to choose not to react by reducing the importance of
an error to zero (I no longer care what you do), or by setting the
reference signal to the same value as the post-disturbance perception
(Whatever you call me, I'll just agree with it).

A strict relationship between disturbance and action occurs only when the
reference signal and the loop gain are being held constant.

Best,

Bill P.

33

[From Bill Williams UMKC 24 September 2002 7:16 PM CST]

[From Fred Nickols (2002.09.24.1752)] --

Okay; but, if you're going to give this some serious thought, here are a
few additional questions in the same vein.

I think I can answer one of the questions now.

         How can we put all this in plain language so ordinary folks (i.e.,
non-PCT specialists) can make use of it?

  I think the idea that we can express control theory in "plain language" is
misleading. It seems to me that our current langauge, plain or otherwise, is
bound up in all sorts of ways with behaviorism, and with notions about
causation that don't work when applied to human beings. The term "cause" as
you've point out is a notion that isn't neccesarily helpful in trying to talk
about what is going on in a living system. I once spent an after noon in a
library going through texts in medical physiology. THe sort of text used in a
junior or senior course often taken by pre-med students. THe term "reflex"
changed its meaning over the years from 1920 to 1980. In 1920 reflex was used
for a stimulus-Response linkage. By 1980 the term had changed to a usage which
signified a control system which resists disturbances. THe treatment of the
Knee-jerk reflex was treated correctly by the 1980's. Over the years the nerve
net diagrams that accompanied the text also changed. But, they still called it
a "reflex."

I've come to the conclusion that rather than trying to get the words right,
we'd be better off experiencing lots of demos of various sorts which would give
us an "intuitive" or non-verbal understanding of the meaning of a control
process. I think must have spent 5 years working with an osciliscope and a
simple electronics bench building control circuits before I talked to anyone
about control theory-- and then it was to build a control unit for a bean
harvester, and do electrical maintaince for an airline.

It seems to me that at least some of the disputes on the CSGnet may not be
about control theory as much as about the words we use and the different
meanings that we assign to those words. Sometimes disputes seem to center upon
who has the right to define what the words "really" mean. I'm talking about
the disputes between people whom we can assume "know" control theory to some
extent. There are, however, people who appear on the CSGnet who seem to find
throwing words about fun and seem to have no concern what-so-ever with the
phenomena of control itself. So, mix genuine disputes and misunderstandings
with misconceptions and sheer ignorance and the combination can be expected to
generate confusion and pointless disputes.

So my answer would be that experiencing control is a far better way to learn
about control than through language whether plain or fancy. Its not really
responsive to your question, but it seems to me that the only language that is
sufficiently clear to accurately convey the meaning of control theory is a
programing language.

There's a moderately well known economist, Geoff Harcourt, who sometimes asks
at conferences, but "What do we tell the children on Sunday morning?" In a way
its a good question, but perhaps more important than what we "tell" the
children is what they see day by day. I think eventually the language problem
will eventually get sorted out when the community understands control theory.
But, it doesn't seem to me that now or in the near future very many people are
going to learn control theory, at least not very well, by way of an ordinary
language-- whether plain or not. Nearly Everyone that I can think of that seems
to really know something about control theory has learned control theory by
working with actual control systems or by programing. This may change, but as
long as the residue of behaviorism is with us, to learn control theory requires
learning it in a way that can resist all of the ways that we are exposed to
behaviorism in everyday life.

When I first encountered Norbert Weiner's Cybernetics, I quickly decided that I
wasn't going to learn cybernetics by studying his equations-- given the
limitations in my education I decided that I'd have to learn by experience
working with actual but toylike devices. Lots of chips ended up being fried in
the course of my learning something about control theory. I think most people
are more like me than they are like Norbert Weiner.

cordially yours

  Bill Williams

PS

Rick, I plan to respond to Fred's question ( since he was first ) but I do plan
to get back to your question/objections.

···

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

34

[From Dan Palmer (2002.09.25.1131 Melbourne Time)]

Howdy Rick,

Rick Marken (2002.09.22.2210):

...Isn't it likely that the three sets of scholars
mentioned above ended up arguing that the "organism literally extends
beyond the sensory and motor surfaces" because they had nothing else to do
but argue since they were not busy building and testing models of
organisms?

Actually, Maturana and Varela did build and test a working model of an
organism (which was sophisticated enough to self-emerge from a simulated
primordial soup). But much more interesting, I think, is the case of W Ross
Ashby, who in his _Design for a Brain_ (1960) was extremely busy building and
testing (cybernetic) models of organisms, one of which he called the
homeostat (see Chapter 8). His position with respect to the
organism-environment boundary?:

"As the organism and its environment are to be treated as a single system,
the dividing line between 'organism' and 'environment' becomes partly
conceptual, and to that extent arbitrary. Anatomically and physically, of
course, there is usually a unique and obvious distinction between the two
parts of the system; but if we view the system functionally, ignoring purely
anatomical facts as irrelevant, the division of the system into 'organism'
and 'environment' becomes vague. Thus, if a mechanic with an artificial arm
is trying to repair an engine, then the arm may be regarded either as part of
the organism that is struggling with the engine, or as part of the machinery
with which the man is struggling" (p. 40).

"These divisions [organism-environment], though arbitrary, are justifiable
because we shall always treat the system as a whole, dividing it into parts
in this unusual way merely for verbal convenience in description" (p. 41).

"It should be noted that from now on 'the system' means not the nervous
system but the whole complex of the organism and its environment" (p. 41).

Cheers,
Dan Palmer

Ashby, W. R. (1960). Design for a brain (2nd ed.). London: Chapman & Hall.
Varela, F., Maturana, H. R. & Uribe, R. (1974). Autopoiesis: The organization
of living systems, its characterization and a model. Biosystems, 5, 187-196.

ps. Thanks muchly to Bill Powers for his illuminating comments re continuous
loop operation, fleshing out the chameleon example, and organization as
immaterial :wink:

35

[From Rick Marken (2002.09.24.1950)]

Rick Marken (2002.09.24.1550)--

>qo~ -1/f (d)
>
>An external circumstance, d, is the cause of behavior, qo. QED

Bill Powers (2002.09.24.1904 MDT)]

This equation holds only when the reference signal is constant.

Right. The complete equation is, approximately:

qo ~ r - 1/f (d)

where r and d have _independent_ effects on qo. So variations in r have no effect
on the causal relationship between qo and d, a fact that is demonstrated in my
"Control of Behavior" demo at http://home.earthlink.net/~rmarken/demos.html. So it
is still quite correct to say that disturbances to controlled variables are an
independent cause of behavioral variations, qo.

>What control theory glasses let us see is that the causal connection between
>external circumstance and behavior is through the _environment_ (the feedback
>function, f), not the organism.

The behavioral "reaction" to a disturbance depends on where the reference
signal is set.

Not true. The effects of disturbance and reference on qo are _independent_. This
was the point Bruce Abbott made some time ago and it's the reason why I developed
the demo mentioned above: in order to prove Bruce wrong (because at the time I
thought, like you, that the behavioral reaction would depend on variations in r; I
should have seen from the mathematics that it does not). In fact, the demo proved
that I was wrong and Bruce was right. Hooray Bruce!

It can be set so a given disturbance increases the error, or
decreases it. Thus the sign of the reaction is determined by the reference
setting even for a constant reference signal. The same action we call a
"reaction" when there is a disturbance can result from a change in
reference level when there is no disturbance.

I think you are trying too hard to rescue an idea whose time has gone;-)

So a person's reaction to a disturbance is not totally determined by
environmental factors.

That's true.

The magnitude of the reaction also depends on the loop
gain, which is determined by the part of the loop internal to the person
just as much as by the part that is external.

Now you're talking. Loop gain variations _will_ affect the causal relationship
between d and qo.

As a result, we are free to choose not to react by reducing the importance of
an error to zero (I no longer care what you do)

Well, free within the context of one's other goals. You can choose not to react by
functionally relinquishing control of a perception. But control of that perception
may matter to other systems inside you. So your choice to not react may result in
other systems trying to unchoose your choice, ie. internal conflict.

A strict relationship between disturbance and action occurs only when the
reference signal and the loop gain are being held constant.

I believe this is true for loop gain but not for the reference signal.

Best regards

Rick

···

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

36

[From Bill Powers (2002.09.25.0100 MDT)]

Rick Marken (2002.09.24.1950)--

> This equation holds only when the reference signal is constant.

Right. The complete equation is, approximately:

qo ~ r - 1/f (d)

where r and d have _independent_ effects on qo. So variations in r have no
effect
on the causal relationship between qo and d, a fact that is demonstrated
in my
"Control of Behavior" demo at
http://home.earthlink.net/~rmarken/demos.html. So it
is still quite correct to say that disturbances to controlled variables are an
independent cause of behavioral variations, qo.

> >What control theory glasses let us see is that the causal connection
between
> >external circumstance and behavior is through the _environment_ (the
feedback
> >function, f), not the organism.
>
> The behavioral "reaction" to a disturbance depends on where the reference
> signal is set.

Not true. The effects of disturbance and reference on qo are _independent_.

I see your point: in terms of changes, the change in action opposes the
change in disturbance independently of the change in reference signal.
However a positive change in action can show up as a decrease in an action
in the opposite direction. Suppose you're steering a car around a left bend
in the road. Your action is to apply a force twisting the steering wheel
to the left. Now if a crosswind arises that is blowing toward the left, the
car would turn too fast, and you would have to _reduce_ your steering
effort to keep the car on the road. If the wind blew just right, you could
let go of the steering wheel (for a second or two).

Of course if you were on Bonneville Salt Flats, you might be intending to
turn either to the right or to the left. In that case, whether the
crosswind blowing to the left resulted in a _decrease_ or an _increase_ in
your steering effort would depend on whether you wanted the car to be
turning to the left or to the right. While the algebraic change in steering
effort would be the same, the absolute magnitude of the effort could either
increase or decrease, depending on the reference condition.

This sort of situation arises whenever the action that brings a controlled
variable to its reference condition must be sustained in order to maintain
the state of the controlled variable. The car, when caused to deviate from
straight-line motion, is subject to a centrifugal force caused by the
deviation, and the driver has to maintain a steering force just to overcome
it. This is not an independent disturbance, since it is a function of the
steering effort. In fact it's better just to think of the force as what is
required to maintain the controlled variable in a given condition -- part
of the definition of the external feedback function.

This can also apply in situations where physical dynamics are irrelevant.
Suppose we have a joint bank account, and the bank is applying service
charges every month because the balance is too low. If I am paying those
charges, you could say I am exerting a certain amount of effort to maintain
the account at a certain level. If you start paying those charges, I will
decrease my payments to keep the level the same. The change in my action
has an effect that is opposed to your disturbance, but in fact my action
decreases when you apply the disturbance. Of course if I wanted the bank
account to decrease, I might switch to making withdrawals to counteract
your payments of the service charges. Silly example.

Let's try for a more directly relevant example, one in which my reaction to
your comments will depend on my reference levels for various controlled
variables. Let's say I'm trying to impress someone with my knowledge of
control theory, so I am putting out a lot of effort to answer questions,
correct misinformation, and so forth. I want correct information to be
given, but I also want to be the one who gives it. You tell me, "That's not
the right explanation, you've made a mistake." Of course this goes against
one of the things I'm trying to do, present myself as an authority, so I
will increase my effort to give correct answers and set the record
straight, and I will probably also argue with you that I was correct all
the time, even if I end up changing my explanation in the way you pointed out.

But suppose I reduce the setting of the reference level for being an
authority, so I am not aiming to give such an impression of correctness and
authority and so on. Suppose I reduce that particular reference level to
zero. Now what will happen when you come up with a criticism of something
I've said? Now my desire to see a correct answer given will be satisfied by
your correction, and I will step aside and let you explain in my place. It
will not matter to me which of us gives the explanation as long as it is
correct. In fact, if you now try to defer to me and say I'm really the
authority and so on, I will resist your suggestions because my reference
level for being the authority is now zero. So I still resist your
disturbance, but now by decreasing my effort to be the authority and in
fact exerting a _negative_ effort. You're helping me toward my old goal of
being the authority, but now I actively resist because my goal has changed.

I've probably brought in considerations that make this not quite a parallel
to the original case, but you see what I'm getting at.

Setting the authority reference signal to zero does not mean I stop
controlling that variable. I will resist if you try to make me seem either
less of an authority (an anti-authority trying to give false information)
or more of an authorithy. But I can, alternatively, cease to consider being
an authority to be of any importance one way or the other. I could say,"If
I know more about a specific subject than anyone else, for example what I
ate for breakfast, then that makes me the authority on that matter. On the
other hand, if someone else knows more than I do -- say they have a
videotape of what I actually ate -- then they become the authority. So
what?" In that case I will not oppose anyone's effort to increase or
decrease my standing as an authority -- it won't matter to me what anyone
thinks on that score. My loop gain for that variable will be zero.

> As a result, we are free to choose not to react by reducing the
importance of

> an error to zero (I no longer care what you do)

Well, free within the context of one's other goals. You can choose not to
react by
functionally relinquishing control of a perception. But control of that
perception
may matter to other systems inside you. So your choice to not react may
result in
other systems trying to unchoose your choice, ie. internal conflict.

True. Exactly. If you're trying to rid yourself of reactions that have
become unwanted, you have to explore those other reasons for continuing to
have them. A young man trying to get a PhD has every reason to want at
least certain people to recognize his authoritative knowledge about
_something_. And those reasons can persist long after they've become
irrelevant.

> A strict relationship between disturbance and action occurs only when the
> reference signal and the loop gain are being held constant.

I believe this is true for loop gain but not for the reference signal.

I think I've introduced a second factor, which is absolute magnitude of
effort as opposed to sign of effort. To sum up, in algebraic terms, the
change in effort remains opposed to the change in size of the disturbance
regardless of the setting of the reference signalm, as you say (I'm
ignoring quibbles about nonlinearities). However, the absolute magnitude of
the effort may either increase or decrease in response to a given
disturbance, depending on where the reference signal is set.

I am very appreciative of Bill Williams' efforts to find and stick to the
optimum path through these thickets. I am sure he is reviewing his remarks
on various subjects as the rest of us are trying to do.

Best,

Bill P.

37

[From Bill (Imsomniac) Powers (2002.09.25.0320 MDT)]

Bill Williams UMKC 24 September 2002 7:16 PM CST]

>It seems to me that at least some of the disputes on the CSGnet may not be

about control theory as much as about the words we use and the different
meanings that we assign to those words. Sometimes disputes seem to center
upon who has the right to define what the words "really" mean.

Too true. You may remember a paper I wrote (I think it's in one of the LCS
volumes) called "The X phenomenon." This was after one of the debates in
which I was accused of trying to tell people what words "really" mean. In
the X paper, I defined the meaning for which we need a word that does not
inadvertently steer us off into different meanings. The word I wanted to
use, of course, was "control." But at the time, people were objecting that
this word was used in a lot of ways, and who was I to tell anyone they were
wrong?

Obviously, I didn't want to tell them they were wrong in using the word
control, but in confusing the phenomenon which I indicate with that word
with the very different phenomena that go with the other meanings. All that
I say about control systems applies ONLY to the type of system I call a
control system, and not to any other kind of system. So I said, here is the
"X phenomenon," defined as a negative feedback loop in which certain
relationships exist and which behaves in relation to its environment in
specific ways. And I said that I would gladly give up use of the word
"control" if someone would suggest a different word for X which we could
agree we would always use to mean the action of an X system and no other kind.

That didn't work, either. I think that basically there are people who don't
want to give up the other _meanings_. This manifests itself as an argument
over words, but they really want us to include other kinds of systems in
which they believe as valid models of behavior, and including them under
the term control is one way to sneak them in. If planning actions that will
have specific effects and then executing them is a valid model of behavior,
then we can say it is an example of "control" just as much as the operation
of a negative feedback loop. And since we're all interested in control
systems, we have to support the plan-and-execute model, right? I think
that's what we've been up against more than a mere argument over words.

Best,

Bill P.

38

[From Bruce Nevin (2002.09.23 17:00 EDT)]

···

At 12:35 PM 9/23/2002 -0500, Richard Marken wrote:

[From Rick Marken (2002.09.23.1230)]

Fred Nickols (2002.09.23.0800) –

So, if it can be argued that we act to control our perceptions instead of

what it is we perceive (which some are sure to view as a very technical

point even if it is literally true), why can’t it be equally validly argued

that “we” extend beyond our skin?

Because I think it unnecessarily obfuscates an important distinction made by

PCT between functions of the organism (controller) and functions of the

environment in which the organism controls. […] Without this boundary, there is control theory, no behavioral illusion and

nothing particularly interesting about PCT.

It’s an interesting question where, and on what grounds, you draw the boundary. One of Bateson’s examples is a blind person tapping down the sidewalk with a cane. Where do you draw the boundary between sensory input function and environment? At the fingertips? At the tip of the cane? Partway down the cane? On a strict biological interpretation (the skin, that’s it, no farther!) it must be where the hand touches the cane. The blind person might not agree. Is this no more than the way when you drive your car you feel differences of road texture, you sense the extension of fenders and bumpers with something like proprioception, you cringe when another car gets too close, or when you ‘squeeze’ through a narrow passage? Or, conversely, rather than no more, are those examples no less than the blind person’s experience?

It is possible and sensible to model at a higher level of control as though the environment consisted of the lower orders of perception. The loop through the ‘real environment’ is represented in those terms. Does this move the boundary back ‘inside’ the skin and the lowest-level sensory inputs?

We project these higher-level reference perceptions into the environment and by the very nature of control do everything in our power to erase every discrepancy between the projection and the actuality. Remember Martin Taylor’s proposals a few years ago about the projection of the perceptual hierarchy into the environment. What we perceive as being in the environment is in this very real sense a reflection of what is in the organism.

When we close a loop through the environment the entire loop is our creation. I don’t mean that when I place a cup on a horizontal surface I create the cup or the surface, but rather that the control loop in which doing so served my purpose (preparing to fill the cup, freeing my hand for something else, whatever) is my creation. I think this is the kernel of truth in Gibson’s idea of an ‘affordance’ in the environment. For me it’s a place to set the cup down. For my wife it’s the television, or a suitcase, or the dining table, or the hatbox decorated with decoupage many years ago by her grandmother, and the extent to which the cup there constitutes a disturbance to some CV of hers varies accordingly.

OK, you say, there you have it. A conflict demonstrates that you two are in the same environment, because you want some aspect of it in one state, and she wants that same aspect of the environment in another state. I agree with the first part of this, to this extent, that interpersonal disagreements and agreements are our means of verifying knowledge. However, the second part is a bit wobbly. It is true that when you apply the Test for the controlled variable to a person it is an application of interpersonal conflict. But one resisted disturbance is rarely enough to identify a CV. In other words, just because an observer can analyze our actions and assert that her control of her perceptions and my control of my perceptions have some particular variable in common, it does not follow that either of us is in fact controlling that particular variable. Control may have unintended side effects, and conversely resistance to a side effect may interfere with what is being controlled.

39

[From Bruce Nevin (2002.09.24 12:59 EDT)]

Bruce Nevin (2002.09.24 12:59 EDT)–

in that narrow and usually invisible sense the disturbances (S) do cause the control outputs (R) that cancel them out.

I should have put “cause” in scare quotes.
One of the ‘fallacies of relevance’ is “false cause”, post hoc ergo propter hoc (B followed A, therefore A caused B). Even if B always follows A, that does not necessarily mean that A causes B.
However (Bill’s point), it does show a relationship between A and B.
The relationship between disturbances and the control outputs that resist them (S and R in the narrow sense above) is expressed by the control system equations, in which more than two variables and constants are involved.
In B:CP Bill showed how the statement of this relationship can be simplified algebraically (p. 277). If R=q0-q0* (where q0* is the output with no disturbance and q0 is the output resisting disturbance) then the “stimulus-response law” (the effect of stimulus inputs on behavioral outputs) is expressed by a simple ratio between two aspects of the environment, the environment function (units of input per output unit) divided by the disturbance function (input units per disturbance unit). It is possible to say that R is “output units”, but S could be understood as “disturbance units” or “input units”.
ni/o nid
---- = —
mi/d mio
where
m = the number of input units per output unit
n = the number of input units per disturbance unit

i = the measured number of input units

o = the measured number of output units

d = the measured number of disturbance units

    /Bruce Nevin
···

At 11:59 AM 9/24/2002 -0400, Bruce Nevin wrote: