[From Bruce Abbott (2009.02.18.1115 EST)] --
Bill Powers (2009.02.18.1603 MST) --
Bruce Abbott (2009.02.18.1510 EST)
Do you agree with that diagnosis and proposal for treatment? Did the
articles say that behavioral therapy (if you want to eat, start
talking) works better than MOL?
>Bill, in defense of David, your characterization of behavioral therapy
>(or "behavior analysis" as this field of clinical application is now
>called) is seriously misleading. The behavioral approach to treating an
>anxiety disorder is not to coerce the client, through bribes and
>punishments, into doing something the therapist wants him or her to do,
>but to arrange an environment in which the client's anxieties with
>respect to some situation will be eliminated. Even before that stage of
>treatment is begun, the techniques of behavior analysis would be
>applied to determine the conditions under which the child talks versus
>remains mute (what are the discriminative stimuli that "control"
>[strongly influence] this
>behavior?) and to assess whether "anxiety" plays a role in it.
Right. Well, you know that I don't believe stimuli can do any
discriminating, or that arranging an environment is likely to do anyone
much good (except in obvious cases like putting the tiger back in its
cage). Behavior modification requires that you find something that a
person wants, and then withhold it until the person produces the
behavior desired by the experimenter, or ceases the behavior that the
experimenter doesn't want to see. You can also start doing something to
a person that he doesn't want done, and then stop doing it on condition
that the person behave or not behave according to what you desire. That
pretty much covers what's actually done, doesn't it?
No, that pretty much mischaracterizes what's actually done. First, I don't
know of anyone in the field of behavior analysis who believes that stimuli
do any discriminating. According to the behaviorist view, people and animals
learn to discriminate based on stimuli that correlate with the effectiveness
of certain of their actions. The hungry pigeon learns that a keypeck will
produce grain only when the key is illuminated red, and therefore pecks at
the key only when it wants grain AND the key is red; the child learns to ask
Mommie for an ice cream only when she perceives that Mommie isn't angry with
her. As for the term "stimulus control," I deliberately placed quotation
marks around the term "control" and suggested "strongly influence" as an
alternative because the stimulus doesn't control anything as we in PCT
define the term "control." Even operant conditioners define the
discriminative stimulus as a stimulus that "sets the occasion for a
response," a clear indication that it is not seen as an event that "makes"
the individual do anything.
There are lots of way to talk about this, some of which sound less nice
than others. "[arranging] an environment in which the client's
anxieties with respect to some situation will be eliminated" sounds
very nice -- nobody could be against eliminating anxieties -- but
exactly what are those arrangements? Don't they all involve either
positive or negative reinforcement? And how are they different from
reward for good behavior or punishment for bad behavior?
One technique for reducing and potentially eliminating a specific anxiety is
systematic desensitization, developed by Joseph Wolpe. This technique
incorporates several elements, including relaxation training, establishment
of an anxiety hierarchy, extinction, and counter-conditioning. The client is
asked to create in imagination a hierarchy of scenarios involving the
anxiety-producing situation, ranging from those that produce little anxiety
to those that produce strong anxiety. The person is also asked to assign a
number -- say, between 1 and 10 -- to the level of anxiety currently being
experienced, where 1 would be little or none and 10 would be intense. The
client is also given training in full-body muscle relaxation, in which the
person attends to different muscles (such as those in the forehead or jaw)
and attempts to relax each group in turn until all muscles are relaxed.
The therapeutic session then begins by having the client imagine the scene
lowest on the anxiety hierarchy and to report a number reflecting the level
of anxiety this image arouses. The person is then asked to erase this image
and replace it with a calming scene while getting the muscles relaxed. These
two steps are repeated until the client can imagine the first scene without
experiencing an increase in perceived anxiety. The therapist then asks the
client to imagine the scene next higher in the anxiety hierarchy. The
procedure is repeated with each level until the client is able to imagine
even the scene that was highest in the anxiety hierarchy without
experiencing an increase in anxiety.
The theory behind this approach assumes that the client has associated the
situation in question with consequences of an anxiety-provoking nature
(e.g., a child has become intensely afraid of dogs, even though most dogs
are friendly). The degree of anxiety is out of proportion to the actual
danger. Having the client imagine the situation lowest on the anxiety
hierarchy, in a setting where nothing bad happens, constitutes an extinction
trial for that anxiety. Extinction to this imaginary scene generalizes to
the similar but more intensely-provoking scenes higher up on the hierarchy,
weakening the association of the situation with the anxiety state. Having
the client relax immediately after imagining the scene associates the scene
with a physical state that is partly incompatible with the anxiety state, in
that the latter produces muscle tension. That's the counter-conditioning
element.
I didn't notice any positive or negative reinforcement being mentioned in my
description of systematic desensitization.
Going back to something else you said earlier, "I don't believe . . . that
arranging the environment is likely to do anyone much good . . ." Just what
do you think the MOL does, if not to arrange an environment that facilitates
certain behaviors on the part of the client?
>If the consensus is that the behavioral approach works best of those
>tried (and I doubt that MOL has been tried), then we PCTers might try
>to understand how such successes can be understood when analyzed from a
>control-systems perspective. For example, do effective behavioral
>techniques for reducing anxiety arrange conditions that promote
>reorganization in a specific direction, for example such that certain
>situations that now act as strong disturbances to some variable that
>the client is controlling, no longer do so?
If it works, it works. Can't argue with that. But does it work? That
depends a lot on who you ask. And for whom does it work? That's a
little trickier, especially in a situation like David Goldstein's youth
center which is really a place where adolescents are sent by New Jersey
courts until they shape up. Behavior mod in such situations, of which
there are many, is not done primarily for the benefit of the resident
population, or at least not for the benefit of the one who is being
given the reinforcements if he or she is good.
Somebody else decides what "good" is. In David's case, the state or the
law decides.
Again, the idea that modern behavior analysis is simply a matter of applying
schedules of rewards and punishments to hapless "clients" seriously distorts
how this work is carried out by competent professionals. Also, the
techniques involved require and include close monitoring to determine what
effects the treatment may be having (if any), so their efficacy (or lack of)
in a given case is not a matter of mere opinion.
>Behavioral analysis grew out of a research tradition that emphasizes
>the influence of the environment on the individual's behavior.
>Consequently it developed a technology for bringing about predictable
>behavioral change that relies on altering in specific ways the
>individual's personal and social environment.
Well, so they say. That's a pretty idealized theoretical view, isn't
it? And I don't think they say "influence." They speak of "stimulus
control," don't they?
See my earlier comment about this above.
What we really need are studies of various approaches to human problems
by people who have no stake in any approach. Would you take my word for
how effective MOL is? Would you have taken Skinner's word for the
effectiveness of behavior modification?
> In contrast, PCT emphasizes the individual's internal structure of
> control and posits that much of this structure developed through
> random changes that were retained to the extent that they reduced
> error in systems controlling essential variables. It is less clear
> from this point of view how to bring about changes in the client's
> control structure that the client and therapist would perceive as a
> successful outcome of the therapy. Perhaps this is one reason why
> practitioners have been slow to adopt PCT as the theoretical model on
> which to base their work.
I agree. Conventional therapists more than once have said that MOL
leaves them feeling that they aren't doing anything. A therapist wants
to feel that he can plan and carry out a successful program of
treatments that will make his patient better. The idea that patients
get themselves better is a bit of a come-down, though of course when
the therapist really gets into MOL, it becomes plain that there are
ways to assist in this process.
Behavior mod is a lot like most other therapies, in that the therapist
is in charge and decides what the problem is and how to fix it. I know
that there is a show of asking the patient's opinion in some of these
methods, but the therapist is still running the show.
That is a very hard thing to give up.
That's not always true, maybe not even typical. Often it is a problem that
is giving the client considerable grief and it is the client who wants help
in overcoming it.
"Prediction and control of behavior toward socially desirable norms"
is still the basic definition of the science of psychology, in the
minds of many psychologists -- I hope not most.
I don't recall the "toward socially desirable norms" part as being in the
definition of psychology, at least not since I've been in it. (In fact, a
current definition states that psychology is the science of behavior and
mental processes.) As for prediction and control, one would like to know
whether there are identifiable conditions associated with given
psychological outcomes. Is it an accident that street gangs tend to develop
in impoverished inner city neighborhoods or do the conditions of life in
those neighborhoods somehow lead many kids into gang affiliation and gang
behaviors? Are there conditions one could establish -- interventions, if you
will -- that would lead to a different outcome, less damaging to the
individuals involved and to the rest of us? Those are questions of
prediction and control. I don't see why one should find pursuing the answers
to them objectionable.
Bruce A.