Behavior Mod

[Richard_Marken]

[From Rick Marken (951214.1300)]

Dennis Delprato (951214) --

Kwee and Lazarus...may be receptive to an account of behavior that explains
why "resistance is bound to occur..."

I agree. I believe clinical psychologists are often the most receptive to
PCT, possibly because they are less "ideological" and, thus, more pragmatic
than so-called "scientific" psychologists. Kwee and Lazarus may be in
a position where they can not only accept PCT but, more importantly,
_understand_ it.

Since we're still on the subject, I think I will take up Bruce Abbott's
(951213.1730 EST) invitation to explain "the nuances of the practice of
behavior modification for all of us here on CSG-L". When I finish, Bruce
is more than welcome to critique it and/or "invite some behavior analysts in
to give us their expert opinion of PCT".

Behavior modification is an application of the "natural selection" model of
behavior which assumes that behavior is selected _by_ its consequences. The
goal of behavior modification is to change the "topology" of behavior by
increasing the frequency of occurrance of desirable behaviors and decreasing
the frequency of occurrance of undesirable behaviors.

The frequency of occurrance of a behavior increases when that behavior is
followed (selected) by a reinforcer and decreases otherwise (the behavior is
"extinguished", ie., it becomes "extinct"). So behaviors are modified (the
relative frequency of different behaviors is changed) by making sure that
only certain behaviors (the "desirable" ones) are followed by reinforcers.

What constitutes a desirable behavior can be determined by the behavior
modifier, the behavior modifiee or by both in consultation. A desirable
behavior is typically one that is thought to be incompatible with some
undesirable behavior; the undesirable behavior cannot be done at the same
time as the desirable one. Thus, increasing the frequency of occurrance of
the desirable behavior automatically decreases the frequency of occurrance of
the undesirable one. If the undesirable behavior is eating between meals, the
desirable behavior might be singing opera betwen meals (assuming that you
can't sing and eat at the same time).

The first step in behavior modification is the identification of a
reinforcer; this is not always easy but we'll just assume it can be done.
The next step is the identification of the desireable behavior whose
frequency of occurrance is to be increased. The next step is to "shape" (if
necessary) the desireable behavior by reinforcing successive approximations
to it; shaping is only needed if the behaving system can't be asked to
produce the desired behavior "voluntarily".

Once you've got the behaving system regularly emitting the desired behavior
and being reinforced for it, you start providing reinforcers less frequently,
because intermittant reinforcement maintains behavior better than continuous
reiforcment; that is, intermittant reinforcement produces behavior that is
highly resistant to extinction.

These are the procedures of behavior modification that are derived from
the principles of reinforcement theory -- the theory that says behavior is
selected by its consequences. Other procedures and goals of behavior
modification - - such as:

teaching people to discriminate along appropriate demensions, and to vary
their behavior in appropriate ways accordingly so as to minimize conflict,
eliminate self-defeating actions, deal effectively with the environment and
with others, and so on.

which were mentioned by Bruce Abbott (951213.1730 EST), are not based on
reinforcement theory in any obvious way (indeed, it is not even clear what
these procedures and goals are). I suspect that these procedures and goals,
such as they are, have been added to the behavior modification toolkit as a
result of experience with real live control systems.

Best

Rick

[Bruce_Abbott]

[From Bruce Abbott (951215.1905 EST)]

Rick Marken (951214.1300) --

Since we're still on the subject, I think I will take up Bruce Abbott's
(951213.1730 EST) invitation to explain "the nuances of the practice of
behavior modification for all of us here on CSG-L". When I finish, Bruce
is more than welcome to critique it and/or "invite some behavior analysts in
to give us their expert opinion of PCT".

O.K., you're on!

Behavior modification is an application of the "natural selection" model of
behavior which assumes that behavior is selected _by_ its consequences. The
goal of behavior modification is to change the "topology" of behavior by
increasing the frequency of occurrance of desirable behaviors and decreasing
the frequency of occurrance of undesirable behaviors.

Very nice, Rick, so long as it is recognized that the phrase "selected _by_
its consequences" is not intended to mean that consequences are active
agents doing the selecting. According to the theory, because organisms are
constructed as they are, certain consequences, under certain conditions,
affect the organism in such a way that the activities which produced them
become more likely (or less likely) to occur under those conditions in the
future.

What constitutes a desirable behavior can be determined by the behavior
modifier, the behavior modifiee or by both in consultation.

O.K. . . .

A desirable
behavior is typically one that is thought to be incompatible with some
undesirable behavior; the undesirable behavior cannot be done at the same
time as the desirable one. Thus, increasing the frequency of occurrance of
the desirable behavior automatically decreases the frequency of occurrance of
the undesirable one. If the undesirable behavior is eating between meals, the
desirable behavior might be singing opera betwen meals (assuming that you
can't sing and eat at the same time).

This is a strange definition for what constitutes a "desirable" behavior.
One might try to suppress an "undesirable" behavior by response competition
with a "desirable" one, but that certainly isn't the usual procedure.

The first step in behavior modification is the identification of a
reinforcer; this is not always easy but we'll just assume it can be done.
The next step is the identification of the desireable behavior whose
frequency of occurrance is to be increased. The next step is to "shape" (if
necessary) the desireable behavior by reinforcing successive approximations
to it; shaping is only needed if the behaving system can't be asked to
produce the desired behavior "voluntarily".

The first step in behavior modification would be identification of what
behaviors need to be changed, in what direction. Where "undesirable"
behaviors occur relatively frequently, the next step would be to identify
the sources of reinforcement that seem to be maintaining the behavior at
those levels. Where "desirable" behaviors occur relatively infrequently,
one would identify sources of reinforcement that are effective for the
individual in question. One would also want to identify what other
effective sources of reinforcement are available to the individual and what
the contingencies are. The same would be true of any sources of punishment
and of any noncontingent sources. In addition, the problem to be resolved
is not always one of getting some behavior to occur more or less frequently,
but getting it under appropriate stimulus control. I could keep going, but
I think you get the picture: there's much more to do than just identify an
effective reinforcer. This phase is called "assessment."

Once you've got the behaving system regularly emitting the desired behavior
and being reinforced for it, you start providing reinforcers less frequently,
because intermittant reinforcement maintains behavior better than continuous
reiforcment; that is, intermittant reinforcement produces behavior that is
highly resistant to extinction.

That might be one technique, but in most situations the goal is to get the
behavior maintained by its own natural sources of reward. Even behavior
maintained by intermittant reinforcement is going to extinguish sooner or
later after external reinforcement is withdrawn, unless it offers its own
intrinsic sources of reward.

These are the procedures of behavior modification that are derived from
the principles of reinforcement theory -- the theory that says behavior is
selected by its consequences.

These procedures are only two very basic ones, and do not by any means
exhaust the tools available to the practitioner which are based on operant
principles.

Other procedures and goals of behavior
modification - - such as:

teaching people to discriminate along appropriate demensions, and to vary
their behavior in appropriate ways accordingly so as to minimize conflict,
eliminate self-defeating actions, deal effectively with the environment and
with others, and so on.

which were mentioned by Bruce Abbott (951213.1730 EST), are not based on
reinforcement theory in any obvious way (indeed, it is not even clear what
these procedures and goals are). I suspect that these procedures and goals,
such as they are, have been added to the behavior modification toolkit as a
result of experience with real live control systems.

You haven't even mentioned shaping, discrimination training, extinction and
counterconditioning procedures, and many other techniques based on operant
and classical conditioning principles. It's a poor shadow of the real thing
that you present. There is a great deal more to behavior modification than
the principle of reinforcement. All in all, Rick, you've done a nice job of
describing some basic parts of it, but you've left out a good deal as well.
What you did describe doesn't sound at all like Clockwork Orange, either.

Regards,

Bruce