EAB and PCT: Not So Different?

<[Bill Leach 941027.18:29 EST(EDT)]

[Bruce Abbott (941027.1000 EST)]

In the interest of brevity I did not include changes in feedback
function along with ordinary disturbances, but the conclusion is the
same. If you do not agree with me that environmental disturbances alter
behavior in predictable ways, then you do not believe that the high
correlations between output and disturbance shown, say, in the DEMO1
demonstration, have been demonstrated. And yes, those relationships
emerge only when you strictly control extraneous variables (you won't
see those correlations in DEMO1 if you allow me to randomly block your
view of the screen, either). If my office gets cold I put on my
coat--you can bet on it.

"Me thinks that maybe" I might be about to learn something about EAB!

...placed in its proper context as a response to Rick's assertion that

the radical behaviorist's search for functional relationships between
environmental conditions and behavior amounts to animism-- ...
The argument you are advancing is that such relationships emerge only
because they are observed under highly controlled environmental
conditions. I agree with you! Good grief, I am not trying to DEFEND the
Skinnerian model, but to DESCRIBE it and note some points of possible
contact with PCT. In fact, I agree with most of your comments.

Picky, picky!

I should have let that one alone... I think I knew what you meant and
that is what I should have focused upon. My appology sir!

How so? It's N-of-1 research... Of course, people can and do misapply
these methods by treating everyone as if the same, but this is not an
inherent fault of the methods.

I may still be off-base here but I still think that the fundamental
problem is probably not necessarily a part of EAB but rather a result of
the background that most EAB researchers would likely have. I think that
it is of vital importance to realize that multiple subjects with the
"same" goal can and at least occassionally will, exhibit radically
different behaviour for similar disturbances. In a like manner, multiple
subjects exhibiting the "same" behaviour may in fact be controlling for
somewhat different perceptions.

If anything, the fact that a large number of repeatable experiments
provide consistant results (particularly when the disturbance affects an
error value for an intrinsic variable) will tend to lead to some very
incorrect conclusions about the method's applicability in general.

Unfortunately I have not read any of Ed Ford's books, but from your
description perhaps I should. How about providing the references?

I believe that Ed's books are listed in the introductory message on CSG-L
but if not, I'll see what I can do.

I agree. Have your read Alfie Kohn's book entitled "Punished by
rewards: the trouble with gold stars, incentive plans, A's, praise, and
other bribes" (Houghton Mifflin, 1993)?

No, I have not but it is "on the list" already.

-bill

[From Bruce Abbott (941027.1000 EST)]

So far I have heard from Bill Leach, Tom Bourbon, and Bill Powers in reply to
my posting, EAB and PCT: Not So Different. I'll deal with each in turn,
starting with Bill Leach.

Me (941025.1025 EST)

[Bill Leach 941025.21:16 EST(EDT)] --

What nonsense. Translate EAB into PCT and you have: environmental
disturbances alter [note: control in the EAB sense of the term] behavior
in predictable ways. Where is your animism?

I am afraid that the problem I see with this statement is that it has not
been demonstrated that "environmental disturbances alter behaviour in
predictable ways". I suspect that if one controls the environment
strictly enough and then disturbs an intrinsic variable sufficiently that
one can obtain "predictable" results.

In the interest of brevity I did not include changes in feedback function
along with ordinary disturbances, but the conclusion is the same. If you do
not agree with me that environmental disturbances alter behavior in
predictable ways, then you do not believe that the high correlations between
output and disturbance shown, say, in the DEMO1 demonstration, have been
demonstrated. And yes, those relationships emerge only when you strictly
control extraneous variables (you won't see those correlations in DEMO1 if you
allow me to randomly block your view of the screen, either). If my office
gets cold I put on my coat--you can bet on it.

At issue here is not that operant conditioning can be used to "modify"
behaviour but rather that the analysis and basis for such a process is
faulty.

That may be AN issue, but is was not MY issue. My argument must be placed in
its proper context as a response to Rick's assertion that the radical
behaviorist's search for functional relationships between environmental
conditions and behavior amounts to animism--the belief that inanimate objects
have desires and purposes that they attempt to carry out. The argument you
are advancing is that such relationships emerge only because they are observed
under highly controlled environmental conditions. I agree with you! Good
grief, I am not trying to DEFEND the Skinnerian model, but to DESCRIBE it and
note some points of possible contact with PCT. In fact, I agree with most of
your comments.

In this view reinforcement "selects" certain behaviors out of all those
that have been emitted in exactly the same way that survival "selects"
organisms in Darwinian evolution. The result is that adaptive behavior
emerges.

I can only ask, how exactly is it that "survival" selects organisms and
just what is this "survival" that is doing the selecting.

Picky, picky! This is only a metaphor, not more "evidence" of animism. If
you really want me to explain Darwin's theory of evolution to you, I can, but
I suspect you know it at least as well as I do.

One of the goals of EAB is to establish more effective and humane
control over behavior. I sense that you have strong objections to this
goal because the methods developed thus far appear to require some kind
of deprivation (i.e., disturbance) to achieve that control.

Not to mention that the methods ignore the unique individuality of the
subjects.

How so? It's N-of-1 research... Of course, people can and do misapply these
methods by treating everyone as if the same, but this is not an inherent fault
of the methods.

. . . The fact is that people often try to exert control in
ways that are often contrary to the interests of others. Ultimately you
must either prevent the other guy from having his way or somehow change
his goals. Either way you are controlling.

In this you raise questions that are of burning interest to many of us.
You are also implicitly introducing the issue of "rights" at the same
time. I would suggest that you might want to look at several of Ed
Ford's books. I will be so bold as to venture that his suggestions (and
indeed implementations) might well be called "operant conditioning" but I
think that one will see subtle differences. Differences that I believe
are critical in such matters. If you have (or do) read Ed's work, I
would be most interested in your comments in that regard.

Unfortunately I have not read any of Ed Ford's books, but from your
description perhaps I should. How about providing the references?

That the problems exist is not, I think, subject to debate. I do think
that both the nature of the problem and the nature of the solution is
debatable. Given that we accept that there are certain "rights" that
every person should enjoy, then there must also be a means of compelling
others to not violate those rights. Compulsion should apply however in
situations of the extreme. Otherwise, non-compelling means should be
attempted to prevent violation of the rights of one by another. As an
aside, a very hard look should be taken at just what should consist of a
"right" in a society.

I agree. Have your read Alfie Kohn's book entitled "Punished by rewards: the
trouble with gold stars, incentive plans, A's, praise, and other bribes"
(Houghton Mifflin, 1993)? It's all about how extrinsic rewards undermine
rather than promote the behaviors they are designed to encourage. It provides
a strong argument against the kind of control envisioned by Skinner simply on
the basis that it doesn't work as advertised when attempted in the real world.

ยทยทยท

------------------------------------------------------------------------------

Tom Bourbon [941025.1155] --

Tom's response was brief but provided the following summary of my post:

Bruce replied to Rick in some detail, with an account of how PCT is
in accord with radical behaviorism: consequences select behavior; PCT can
show behaviorists how to more effectively control the behavior of people;
when behaviorists control people more effectively, the people who are
controlled will _believe_ that they (the ones controlled) are controlling
more effectively; there are no vestiges of S-R thinking in contemporary
behaviorism, and so on.

There are some interesting distortions here: (a) I never said the PCT is "in
accord with" radical behaviorism, (b) I never said that people who are
controlled will _believe_ that they are controlling more effectively, and (c)
I never said that there are no vestiges of S-R thinking in contemporary
behaviorism. On the other points I plead guilty.

What I DID say was that there are some points of contact between EAB and PCT
(which I attempted to explain). I DID say that consequences select behavior
(although not just any behavior, but those that reduce the error signal) and I
concede this as an unfortunate lapse. However, if you substitute Bill
Powers's reformulation my point survives, which is that in both views learning
involves selection by consequences (even if they differ as to what is
selected). I DID say that people will attempt to control others and that PCT
offers an alternative to the use of brute force (carrots and sticks). In
fact, as I wrote that section I had in mind your studies on conflict and
cooperation. And I DID say that S-R psychology is dead. I was speaking, of
course, of the theoretical view that began with Thorndike and ended with
Hull/Spence, which holds that all behavior arises in response to specific,
immediately prior stimuli. It IS dead. Whether or not there are "vestiges"
of S-R thinking is a different issue.

I think most of our disagreements arise from problems of definition--we use
the same terms to mean different things. In comparing EAB and PCT I have had
to use EAB terms, some of which happen to mean other things within PCT. Where
this has happened I have tried to provide the EAB definition so as to avoid
confusion. Control is one such term, S-R psychology another. In my view,
your use of the latter term is much too broad, for it includes ANY case of
linear causality. That is not "S-R thinking" as I defined it.

Finally, I think you are overreacting (stimulus-response? (;->] ). In
asserting that EAB and PCT are "not so different," what I had in mind was that
they are not AS different as they appear to be at first glance--that there is
some common ground. In the same spirit I might claim that turtles and oysters
are not so different: among other things they both live in shells. But this
is not to say that they are the same, or even highly similar. I fully
appreciate that in other ways the differences between EAB and PCT are vast.
---------------------------------------------------------------------------

Bill Powers (941025.1625) --

Rick Marken is arguing, of course, from the PCT definition of control,
which is very narrow and precise. But I think people tend to use the
term control correctly when something about the situation brings out the
difference between controlling and either "influencing" or
"determining."

Yes, and I like your distinctions among those three terms. But in describing
EAB objectives in EAB terms, I am not free to redefine them in my own way.
That is why I took the trouble to explain what the word "control" means to
radical behaviorists. Given that I took the trouble to make this meaning
clear, it just blows me away to hear Rick interpret my words as if I had meant
"control" in its narrower PCT dress.

What is not generally appreciated outside the field of learning and
behavior is that S-R psychology--the notion that all behavior consists
of responses to prior stimuli--died a quiet death a long time ago.
Nobody believes in it, least of all Skinnerian radical behaviorists.

Bruce, I claim that this is a myth, at least outside radical behaviorism
(I'll take your word for that). Take practically any experimental
journal in the life sciences, and you will find experiments in the form
of "the effect of manipulated variable A on behavior B." That is what we
mean when we talk about S-R theory.

And this is a very different definition from the narrow one I provided, so we
are not talking about the same thing. The distinction is important, because
it is not appreciated in the "TUMBLE" e. coli demonstration in which the user
is asked to indicate whether the simulation should use "reinforcement" or not.
That reinforcement works this way is precisely the S-R view I was talking
about that nobody believes in.

The point of PCT is not to make recommendations about controlling the
behavior of others, people or animals, but to deduce from basic
principles what will happen under various conditions where control of
others might occur.

The point of EAB is not to make recommendations about controlling the behavior
of others, people or animals, either. Its point is to understand how
environmental conditions of the past and present influence behavior (i.e., the
search for functional relationships). In either case (EAB and PCT), apparent
success will lead to applications, like it or not. Several examples of PCT
applications have recently been described on CSG-L.

What PCT adds is a new definition of what gets selected--responses that
reduce error between reference and perception.

In the light of what I said above, you can see that this isn't quite
right. What's learned is not a specific response to a specific input or
error. It's a _relationship_ between the state of an input or error and
the state of a response variable. What's learned is a control system,
not an activity.

(Blush) Yes, of course. What I was trying to convey--and conveyed badly--was
that the relationship between behavioral output and error was learned, not a
specific response TO error. In the balance of your reply (not copied here)
you add that other components of the loop can be acquired, too, which is
something I neglected. Yet the mechanism by which this relationship is built
may yet involve "selection by consequences."

The remainder of your post is directed to the general issue of control-by-
others and shows how PCT can explain why behavior modification works--when it
does work--and when these techniques would be expected to break down. It
seems to be a general discussion of these issues rather than an argument
against my position; much of it nicely summarizes B:PC Chapter 17.

Cheers,

Bruce