Rick's Discover Posting

[From Fred Nickols (970924.1645 ET)]

Rick Marken (970923.0900)

The October issue of _Discover_ magazine is all about "Behavior".
I bought a copy because I though it might be fun to read through
it to see if there was any trace of an understanding of behavior
as a purposeful process of controlling perceptions relative to
internal references. So far, no luck. But I have learned that the
new, trendy approach to understanding behavior is in terms of
genes (combined with influences from the environment) and evolution.
In the first article I read, the author makes the point that genes
don't determine behaviors; rather they determine _tendencies_ to
behave in certain ways in the presence of environmental events. So
the _new_ view of human nature is that we behave as we do because
our genes determine _tendencies_ to respond in various ways to
environmental stimuli. I'm probably biased, but this doesn't seem
like a huge step beyond S-R psychology to me.

On the one hand, I can see where the view expressed above might have some
appeal in accounting for "instinctive" reactions; on the other hand, I doubt
it does. The scary part is that we'll probably soon be hearing a courtroom
defense claiming that, "My genes made me do it!"

Acts and Results

What is missing from everything I have looked at in the _Discover_
issue so far is any understanding of the fact that when we talk
about "behavior" we are talking about both _actions_ and the
_results_ of those actions; we are talking about both hand
movements (actions) and the signature on the contract (the result).

Here's where non-PCTers who have been mucking around in the areas of
behavior and performance have a problem. A distinction is typically drawn
between an action and the result it produces. In human performance
technology circles, the action is usually labeled "behavior" and the result
of action is labeled an "accomplishment." (See Tom Gilbert's book, Human
Competence.) I assume that the PCTer insistence on having the term behavior
include the action and its outcome owes to the closed-loop view of control.
Do I have this right?

Also missing is the understanding that most consistently produced
results are produced by different actions every time -- and
necessarily so or the results (due to disturbances to those results)
would not repeat. So the same "behavior" (result) is (and must always
be) produced by different behaviors (actions).

Most people I know would say, "Yeah, you're right, but so what?" to the
comment immediately above. I have written dozens of articles and my actions
from article to article have only superficial resemblance to one another.
That said, the general pattern of behavior is fairly consistent: I write a
little, rewrite a lot, go away and come back, write some more and rewrite
even more. Then I get someone else to read and comment and the cycle starts
all over again. Then it's a lot of go away and come back; an extended
period of tinkering. Finally, when I think I'm about finished tinkering, I
send it off to an editor somewhere for consideration.

Control

This kind of "behavior" (variables actions producing consistent
results) can only be accomplished by a system that perceiives the
state of the result (the signature), compares this percpetion to a
reference for what that perception _should be_, and continuously
acts (moves the hand) to keep the discrepency between perception
and reference near zero. The behavior of this system -- the
relationship between acts, results and disturbances to those
results -- is called _control_. This control system generates
neither actions or results; rather, it generates a _perception_
(of the intended result) that meets the specifications provided
by the reference for the perception.

I don't know about you, Rick, but if you looked at my signature, you'd say I
have a pretty loose set of reference conditions. Speaking of signatures,
what about reference conditions that might be embedded in muscles and joints
(e.g., as is the case if I sign my name with my eyes closed). Visual
perceptions are only part of the picture, right?

With this understanding of the natujre of behavior, the authors of
the _Discover_ articles on "Behavior" would realize that, if
anything about behavior is inherited it is reference specifications
for the states of perceptions -- not specific ways of
generating actions. They would also undertsand that the apparent
influence of the environment on responses is not an effect of the
environment at all; it is the control system acting to protect
controlled perceptions from environmental disturbance.

Anyone who's ever toyed with making change in organizations ought to
appreciate the paragraph above.

Extremism

Reading the _Discover_ issue made me realize why a PCT revolution
is such a remote possibility in the near future. The problems with
the discussions of behavior in that issue are fundamental. I see
two ways for control theorists to deal with these problems: 1) try
to critique the papers carefully and try to explain the flaws in
detail or 2) explain the fundamental problem (that they are based
on an incorrect understanding of thenature of behavior itself) and
reject them out of hand, calling for a new start for behavioral
science.

The first approach seems like the most reasonable; but it has not
been very effective because when you get into the details you tend
to lose the big picture. This is what is happening in the discussion
with Bruce Abbott; when we get into the details of the problems
with the EAB view of behavior we (and the people listening) tend to
miss the big picture -- which is that events in the environment
don't cause or control behavior (actions or results); rather,
behavior (actions) control environmantal results (as perceived).

The second approach focuses on the big picture but it can't help
but come off as extremist. When the big picture is "everything
you are reading about behavior in the psych journals, textbooks
and reputable national science magazines (like _Discover_) is
based on a fundamental misconception about the nature of behavior
and is, thus, wrong or misleading" you can't help sounding like an
extremsist. Nobody wants to imagine that behavioral science has been
built, for the last century, on a basic misconception about behavior.
No one wants to think that nearly all the research done to date is
virtually useless. People want to take a moderate position on issues;
they want to believe that the old psychology can't just be all wrong.
The problem for PCT is that science is not politics; the old
psychology simply is _demonstrably_, fundamentally wrong.

I see no solution to this problem. I don't see how either approach
to presenting the PCT view of behavior can "work". I think all we
can do is hope that people of courage and integrity will discover
PCT, be convinced of its merits based on the _evidence_, and start
studying and dealing with other people as what they are; living
control systems.

I don't know what the solution is, either, Rick, but I do know one thing:
PCT won't get very far with most people if PCTers insist on defining terms
that are in common usage in ways that are inconsistent with that common
usage. The case in point, of course, is behavior. Don't misunderstand; I'm
not taking issue with your definition of behavior or your right to redefine
it, I'm simply saying that, as a practical matter, you won't get far seizing
terms that are in common use, redefining them, and then attempting to
communicate using your definition. Nor will you get very far with a
concerted effort to prove everyone else wrong. Even if you're right, so
what? As I understand it, the objective is to gain acceptance for the PCT
point of view. I might be alone in this, but I don't think gaining
acceptance for PCT requires actively discrediting other points of view.
That sounds more like a jihad. Gaining acceptance for PCT means introducing
it in ways that people can understand and make use of it. Frankly, I think
most people do deal with other people as "living control systems." They
might lack the theoretical base offered by PCT, and know next to nothing
about the nature of closed-loop control systems, but they darn well know
about people and how to deal with them; they've been doing it for thousands
of years, and passably well, I might add. More important, if PCT doesn't
lead to a better way of dealing with other people, it's not likely to gain
any acceptance at all because it adds no value; it's just one more in a long
line of theories, none of which have led to much of anything except for an
occasional disaster or two.

Mercy! Let me get down off this soapbox before I fall and hurt myself :slight_smile:

Regards,

Fred Nickols
Senior Consultant
The Distance Consulting Company
nickols@worldnet.att.net

[From Rick Marken (970924.1840)]

Fred Nickols (970924.1645 ET) --

Thanks for the nice comments on the Discover post. I only
take exception to your suggestions at the end of the post.

PCT won't get very far with most people if PCTers insist on
defining terms that are in common usage in ways that are
inconsistent with that common usage.

Great. Then PCT will get far because we don't insist on defining
terms (like "behavior") is ways that are inconsistent with that
common usage. What we are trying to do is communicate an
idea using language. The idea is that what is commonly called
"behavior" is a process of controlling perceived consequences
of action. Nobody's trying to redefine "behavior"; I'm trying
to call attention to the fact that the perceptions that are pointed
to by the word "behavior" (perceptions referred to by words like
"walking", "acting weird", "tensing muscles", etc) are either
controlled results of action ("walking"), uncontrolled side effects
of action ("acting weird"), or actions used to produce
controlled results ("tensing muscles").

Nor will you get very far with a concerted effort to prove
everyone else wrong.

I agree. But it's hard for people to avoid noticing that, if
PCT is right, then one or another of their cherished ideas is
wrong. Look at the problem with reinforcement; PCT shows that
there is simply no such thing; the concept of reinforcement
is wrong. What do you do, then, when someone wants to explain
reinforcement with PCT. Just say "OK, here's how you do it"?
I think the only thing you can honestly say to a person who
wants to use PCT to explain reinforcement is what you would
say to a person who wants to use Newton's laws to explain
"impetus"; "get a life".

As I understand it, the objective is to gain acceptance for
the PCT point of view.

Well, that's not my main objective. My main objective is help
people _understand_ the PCT model (it's far deeper than a point
of view). I would like people to accept PCT -- but only if
acceptance is based on _evidence_. People who accept PCT because
it sounds good or because it seems to support their existing
biases or because some person they respect accepts it are not
going to accept it for long. I want people who accept PCT
because they can see (by doing the basic demos of control
phenomena) that no other model can explain what is happening;
that the only explanation of what's going on is that the behavior
they are seeing is the control of perception.

Frankly, I think most people do deal with other people as
"living control systems."

Maybe so. But I've seen many parents and teachers who seem to
go out of their way to ignore this fact about people. And at
a social level there seems to be absolutely no serious
understanding of people as control systems.

They might lack the theoretical base offered by PCT, and know
next to nothing about the nature of closed-loop control systems,
but they darn well know about people and how to deal with them;
they've been doing it for thousands of years, and passably well,
I might add.

Tell that to the Palestinians, the Bosnians, the people living
in South Central, almost anyone in India, European Jewry, etc.
No, I would say that people have been doing a rather piss poor
job of dealing with each other for thousands of years.

More important, if PCT doesn't lead to a better way of dealing
with other people, it's not likely to gain any acceptance at all
because it adds no value; it's just one more in a long line of
theories, none of which have led to much of anything except for
an occasional disaster or two.

Again, I only care about people who accept or reject PCT on the
basis of _scientific evidence_ . People who accept PCT on any
other basis are going to run into the same disappointments
they would encounter after signing up for any religion. People
should _only_ accept PCT when they 1) understand the model and
2) are convinced by the evidence that _only_ the PCT model can
explain the behavior of living systems.

Showing the practical value of any theory is an iffy situation.
I believe that PCT can be of enormous practical value. But
what if someone tries to apply PCT in some practical situation
and finds that there is no improvement or, even, that things
get worse? A person who accepts PCT on the basis of practical
value (rather than on the basis of points 1)and 2) above) will
probably reject PCT on the basis of these results. But a person
who does understand PCT and is convinced by the evidence might
be inclined to look deeper (as Tom ourbon does when looking into
the application of Ed Ford's RTP program in schools) to see if
what was really applied was PCT. What Tom has been finding is
that every time the RTP program seems to be failing, it's because
teachers (the people applying PCT) have abandoned the principles
of PCT. So they say they are applying PCT but they are not.

So, again, I think it's _far_ more important for people to
accept PCT based on 1) an understanding of the model and
2) _evidence_ that only the PCT model can explain the behavior
of living systems. Don't we expect engineers to understand and
accept (on the basis of evidence) Newton's laws before we turn
them loose on practical applications, like building bridges?
I think we should expect the same understanding and acceptance
(on the basis of evidence) of PCT before we turn practitioners
loose on practical applications, like building healthy psyches.

You can have your soap box back now, Fred;-)

Best

Rick

···

--

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

[Hans Blom, 970925c]

(Rick Marken (970924.1840))

Look at the problem with reinforcement; PCT shows that there is
simply no such thing; the concept of reinforcement is wrong.

The fact that there is no such thing as reinforcement _in PCT_ does
not logically imply that the concept itself is wrong. Unless PCT
presents itself as an all-compassing "theory of everything". Which it
doesn't (I hope ;-).

What do you do, then, when someone wants to explain reinforcement
with PCT.

Simply saying "PCT cannot explain it" might be an alternative...

I think the only thing you can honestly say to a person who wants to
use PCT to explain reinforcement is what you would say to a person
who wants to use Newton's laws to explain "impetus"; "get a life".

That's not an attitude that will help to make PCT popular...

Greetings,

Hans

[From Rick Marken (970925.0800)]

Hans Blom (970925c) --

The fact that there is no such thing as reinforcement _in PCT_ does
not logically imply that the concept itself is wrong.

Right. But the fact that there is no such thing as reinforcement,
period, does.

Me:

I think the only thing you can honestly say to a person who wants to
use PCT to explain reinforcement is what you would say to a person
who wants to use Newton's laws to explain "impetus"; "get a life".

Hans:

That's not an attitude that will help to make PCT popular...

You could be right. But, as I said in my previous post, I only
care that PCT be popular with people who _understand_ the model
and accept it on the basis of evidence. You do neither. I won't be
thrilled if PCT becomes popular for religious reasons -- that is,
if it becomes popular because people like our attitude or because
they like the way we talk.

If people don't want to bother learning the model and evaluating
it based on evidence then that's the way it goes; then PCT will
never be popular in the way I would want it to be popular and it
might fade away when Bill, Tom and I die. So? It was just an idea;
a beautiful idea that precisely and quantitatively explains the
behavior of living organsism, but, still, just a thought;-)

Best

Rick

···

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

[From Bruce gregory 970925.1200 EDT)]

Rick Marken (970925.0800)

Hans Blom (970925c) --

> The fact that there is no such thing as reinforcement _in PCT_ does
> not logically imply that the concept itself is wrong.

Right. But the fact that there is no such thing as reinforcement,
period, does.

Sigh. I sometimes wish we could tilt at things other than
windmills, but I have had as little success converting Rick as
he has had converting Hans. There must be a message here. After
this final effort, I promise to give up. The conceptual
framework that includes reinforcement is deeply flawed. It is
charitable to say that it has even minimal value. Saying that it
is "wrong" adds little to this as far as I am concerned, but
Rick and I differ. He is certain that it _is_ wrong and that the
_right_ thing to do is to keep calling this to the attention of
its "true believers". Those windmills look like enemies of the
king to me Don Quixote, lets mount up and dispatch them...

Sancho

[From Rick Marken (970925.1100)]

Me:

But the fact that there is no such thing as reinforcement,
period, does.

Sancho:

He [me;-)] is certain that it [reinforcement] _is_ wrong and
that the _right_ thing to do is to keep calling this to the
attention of its "true believers".

I thought you were supposed to be the down to earth one, Sancho.
But you seem to oppose my saying "reinforcement does not exist"
(to anyone who will listen, by the way, not just the "true
believers") for the most arcane reasons. Since consequences don't
strengthen responses but are, rather, controlled by responses,
what possible objection could you have to my saying that
reinforcement doesn't exist? It doesn't, does it?

Bruce Gregory (970925.1210 EDT) --

I'm in your [Bruce Abbott's] camp on this one. When it comes to
the utility of the notion of a reinforcer, I'm on the other side.

Guess I'll be lookin'for a new sidekick;-)

I just can't see any way to understand control or the control model
and, at the same time, see the notion of a reinforcer as anything
but a horrendous mistake.

Best

The Don

···

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

[From Bruce Gregory (970925.1745)]

Rick Marken (970925.1100)

Me:

> But the fact that there is no such thing as reinforcement,
> period, does.

Sancho:

> He [me;-)] is certain that it [reinforcement] _is_ wrong and
> that the _right_ thing to do is to keep calling this to the
> attention of its "true believers".

I thought you were supposed to be the down to earth one, Sancho.
But you seem to oppose my saying "reinforcement does not exist"
(to anyone who will listen, by the way, not just the "true
believers") for the most arcane reasons. Since consequences don't
strengthen responses but are, rather, controlled by responses,
what possible objection could you have to my saying that
reinforcement doesn't exist? It doesn't, does it?

Whether or not something "exists" depends on your theoretical
framework. (Quine called it the ontological commitment of
discourse.) "Action-at-a distance" exists in a Newtonian
framework, but not in an Einsteinian framework. Einstein's
framework is superior to Newton's because it covers every case
that Newton's does and situations where Newton's breaks down. In
this sense "action-at-a-distance" doesn't "exist" -- it isn't
part of our best theoretical apparatus. Nevertheless, we still
use Newton's framework to solve many problems. Whenever we do
so, we act _as though_ "action-at-a-distance" does exist.
EAB is analogous to Newton's framework and PCT to Einstein's,
with the _major_ exception that EAB has little, if any, utility.
I would prefer to challenge EAB enthusiasts to present
quantitative predictions of the simplest purposeful behavior
rather than to argue with them about whether one of their
theoretical constructs "exists". To answer your question, no,
reinforcement does not exist. But let's keep that between
ourselves :wink:

Bruce Gregory (970925.1210 EDT) --

> I'm in your [Bruce Abbott's] camp on this one. When it comes to
>the utility of the notion of a reinforcer, I'm on the other side.

Guess I'll be lookin'for a new sidekick;-)

Talk about a foul weather friend! I don't see why it is
desirable to argue that a food pellet is a controlled variable.
As Freud said, "Sometimes a cigar is just a cigar." I accept
Bruce Abbott's statement that a food pellet is part of a control
loop and what the animal controls (eating the pellet, making it
appear in the cup) is the conrolled variable. I can't see what
the fuss is about. There is no evidence that the process called
"reinforcement" exits, on that I certainly agree. It'll take a
larger disagreement than this to rid you of your faithful
sidekick.

I just can't see any way to understand control or the control model
and, at the same time, see the notion of a reinforcer as anything
but a horrendous mistake.

Well, I suppose Einstein could say that action-at-a-distance is
a "horrendous mistake", but he never did. He just produced a
more encompassing theory. EAB is not Newtonian physics, but the
idea of reinforcement has found its way into everyday speech.
Saying it is a horrendous mistake to those who do not understand
PCT may not be terribly productive. I prefer to ignor the term
altogether. It raises far more problems than it "solves".

Sancho

[From Rick Marken (970925.1840)]

Bruce Gregory (970925.1745)

I don't see why it is desirable to argue that a food pellet
is a controlled variable...I accept Bruce Abbott's statement
that a food pellet is part of a control loop and what the
animal controls (eating the pellet, making it appear in the
cup) is the conrolled variable. I can't see what the fuss
is about.

What the fuss is about:

1. There is already a name for the part of the control loop
occupied by the food pellet -- the part called qi in B:CP
(p. 274). qi is called the _controlled quantity_; it is the
environmental correlate of the controlled perception. If the
controlled perception is pellets/unit time then qi is
1/(average time between food pellets); the individual food
pellet that Bruce wants to call a "reinforcer" is part of the
definition of qi. If the components of a controlled quantity
are to be called "reinforcers" then (as Bill noted) the
height and width of the rectangle in my size control
experiment (ME, p.52) are reinforcers. This is, at best, a
peculiar use of the word "reinforcer".

2. Controlled variables are the central concept in PCT;
everything an organism does (all of its behavior, as that
term is commonly understood) is aimed at bringing controlled
variables to reference states and protecting then from
disturbance. Understanding behavior, according to PCT, is,
therefore, largely a matter of understanding what variables
an organism is controlling -- its controlled variables. Yet
controlled variables are completely ignored by conventional
psychology in general and EAB in particular. You will
look in vain through the _Discover_ issue on behavior
for any mention of controlled variables (or for any words
that could possibly be construed as referring to controlled
variables). Therefore, I think it is _very_ important to
clearly point out the controlled variables that are involved
in any behavior (like operant behavior).

Something about food pellets is almost certainly one of the
variables controlled in operant behavior. But food pellets have
been identified as "reinforcers" by EABers under the assumption
that they control behavior. If we want people to understand
that an aspect of the pellets is controlled, it seems to
me that the best way to do this is to refer to the pellets
as what they are -- a controlled variable. Calling the pellets
"reinforcers" as a means of saying that the pellets are
really just a component of a controlled variable is, at best,
bad terminology (since these "reinforcers" don't _strengthen_
anything) and, at worst, misleading (diverting attention from
the fact that the pellets are under control).

It'll take a larger disagreement than this to rid you of your
faithful sidekick.

Ohhhh Pancho!

(Oops. Wrong latin stereotype;-))

Best

Cisco

···

--

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

[From Bruce Gregory (970926.1225 EDT)]

Rick Marken (970925.1840)]

What the fuss is about:

1. There is already a name for the part of the control loop
occupied by the food pellet -- the part called qi in B:CP
(p. 274). qi is called the _controlled quantity_; it is the
environmental correlate of the controlled perception. If the
controlled perception is pellets/unit time then qi is
1/(average time between food pellets); the individual food
pellet that Bruce wants to call a "reinforcer" is part of the
definition of qi.

Fine.

If the components of a controlled quantity
are to be called "reinforcers" then (as Bill noted) the
height and width of the rectangle in my size control
experiment (ME, p.52) are reinforcers. This is, at best, a
peculiar use of the word "reinforcer".

I agree. And we both agree that the term is misleading.

Therefore, I think it is _very_ important to
clearly point out the controlled variables that are involved
in any behavior (like operant behavior).

Count me in!

Something about food pellets is almost certainly one of the
variables controlled in operant behavior. But food pellets have
been identified as "reinforcers" by EABers under the assumption
that they control behavior. If we want people to understand
that an aspect of the pellets is controlled, it seems to
me that the best way to do this is to refer to the pellets
as what they are -- a controlled variable.

Or, part of the definition of qi.

Calling the pellets
"reinforcers" as a means of saying that the pellets are
really just a component of a controlled variable is, at best,
bad terminology (since these "reinforcers" don't _strengthen_
anything) and, at worst, misleading (diverting attention from
the fact that the pellets are under control).

Again, I agree completely.

Ohhhh Pancho!

Ohhhh Cisco!

Sancho