Changing Another Person LO20152

Jon Krispin (
Thu, 10 Dec 1998 11:48:53 -0500

[Linked by your host to LO20128 --]

Responding to LO20098 and LO20128

Hello Fred and other list members,


Thanks for your comments (in LO20098 and LO20198) on my recent post
(LO20081) regarding changing another person. I would like to respond to
some of your comments, although, as Rick mentioned, I don't think that you
will need any armor.

First, I want to apologize for seemingly putting words in your mouth when
I wrote:

>Here is a perfect example of an antecedent (gun to head with demand being
>made) that is very effectively paired with a consequence (a negative,
>IMMEDIATE, certain consequence). As Fred points out, the likelihood that
>this antecedent will be successful at influencing behavior is very high.

referring to your comment in LO20060

>>That said, can someone put a gun to my head and coerce me into doing
>>something I wouldn't ordinarily do? Of course, but in the last analysis, but
>>in the last analysis I am the one doing the doing.

I wanted to use your illustration because it demonstrated a concept that I
have been trying to communicate, the antecedent, its role in promting
behavior, and the idea that the ability of an antecedent to influence
behavior is much greater when it is effectively paired with a consequence.
You certainly did not say that the likelihood that the antecedent would
likely influence your behavior was very high. I was drawing my inference
from the fact that you answered your own question with the term "of
course", implying that you were indeed likely to respond to the situation
by complying with the demand. The actual inference was mine, and mine

The second comment to which I would like to respond is the following
section of your post, also in LO20098:

>There is no way consequences strengthen or weaken any preceding behavior
>or actions. Consequences might be conveniently spoken of as "shaping" future
>actions in similar situations, but the behavior and actions that occurred
>previously are gone off into the past where they are not influenced by
>anything, not even the vagaries of memory. ...snip...

Your observations here are correct, and, reviewing the section of my post
to which you were responding, I did not adequately clarify that the
behaviors that were influenced by the consequences were, in fact, future
behaviors. I did however write later in the post:

>...consequences are both the outcome that results from the behavior
>(the *effect*) and that which influences the future occurrence of the
>behavior (the *cause*).

In addition, my post on the ABC model in LO19894 stated this relationship
more clearly. Specifically, I wrote:

>The consequences following the behavior are what will ultimately determine
>whether or not the behavior will continue as before, strengthen in terms of the
>rate or likelihood that it will occur again in the future, or weaken.

So Rick's suggestion in his host's note in your post that the future
behaviors were actually that to which I was referring was correct.

In LO20098, you drew a distinction between actions and behaviors, derived
from Perceptual Control Theory (developed and articulated by William T.

>...But my behavior is under my control. It grows increasingly
>important in my own thinking to draw a distinction between my behavior and
>my actions. My actions, and frequently their effects, are observable and,
>under suitable conditions, controllable by others. My behavior is not
>entirely observable and it is not at all amenable to control by others.
>Indeed, attempts to control my behavior will engender conflict whereas my
>actions are quite often very negotiable. ...snip...

I am not familiar with the intricacies of Perceptual Control Theory, but,
if I understand your distinction correctly, I don't know that I agree with
it. I am not sure how to specifically address the distinction that you
are drawing, so, instead, I will try to explain the behaviorist
perspective further. I believe that consequences for our actions (your
definition) shape our future actions, but the influence doesn't stop
there. Our perceptions, mental models, attributions, inferences,
attitudes, etc... are also influenced by these same consequences. If
these "internal" activities are part of what you have distinguished as
"behaviors", then I would have to disagree that these behaviors are not
amenable to influence, although I might agree that they are not amenable
to control (more on this later).

I suspect that one of the areas where we may not be connecting here has to
do with the idea of causation and cause-effect relationships. Much of the
thinking that has been done on causation has debated such ideas as whether
the concept of causation is an ontological phenomenon, an epistomological
phenomenon, or both. In the behaviorist concept of causation, this type
of debate has little relevance. This is because the behaviorist
perspective does not adhere to any traditional definition of causation.
Coincidentally, this debate has little relevance in systems thinking for
the same reason.

One definition of causation which typifies many of the attempts to define
it was proposed by the philosopher Immanuel Kant. He suggested three
conditions (physical and logical) that must be present in order for
causation to be inferred. They are contiguity (the *cause* and the
*effect* are closely related in time and space), succession (the *effect*
always appears to follow the *cause* in space and time), and
irreversibility (the *effect* cannot possibly appear before the *cause* in
a logical sense. It appears if and only if the "cause" has appeared first
- the order of their appearance is irreversible). If these three
conditions are present, then we will begin to infer production of the
*effect* by the *cause*.

While I may sometimes use causal language to describe the
interrelationship that can be observed between a behavior and its
consequences, this is not meant to suggest that the relationship is the
same as is defined above. I usually try to talk about "influence",
"shaping" or "interrelationship" in order to avoid leading someone into
this quagmire, but I sometimes slip. Obviously, as I have described it,
the behaviorist perspective concept of influence, or *causality* directly
violates 1 of these three conditions (The cause and the effect are often
not very closely related in time and space, so the premise of contiguity
is often violated), and it is almost impossible to reconcile the other 2
conditions (succession and irreversibility) with the behaviorist account.
This does not, however, cause a problem for the behaviorist perspective in
any way. In fact, it is necessary to suspend most of this type of
thinking in order to view the world from the behaviorist perspective. The
consequence of the behavior plays the dual role of effect/result and the
cause of behavior (but not in the traditional sense in which *cause* is
usually thought).

This same linear understanding of cause-effect (or stimulus-response)
relationships described above also epitomizes the perspective from which
we must free ourselves to begin to exercise the discipline of systems
thinking as well. In LO20081, I quoted Senge in the Fifth Discipline (page
73) to support this:

>The essence of the discipline of systems thinking lies in a shift of mind:
>-seeing interrelationships rather than linear cause-effect chains, and
>-seeing processes of change rather than snapshots

I do think that it is important to note that the behaviorist perspective
does not necessarily invalidate the more typical discussions of
cause-effect relationships. The antecedents to behavior in the ABC model
might possibly be construed as "causes" in the more traditional sense, and
they play an important role in understanding behavior. Antecedents (the
conditions and events present that set the stage for behavior) are what
allow us to generalize or discriminate between one situation and another.
Antecedents signal to us times at which some reinforcers for a given
behavior are available and when they aren't. In a traditional causal
model, these conditions and events may be interpreted as *causes* (notice
that they satisfy Kant's conditions).

So why break out of this linear perception of influence to adopt the
behaviorist perspective? The best answer that I can give is because it is
practical. Focusing on antecedent events and conditions as causes does
not allow us to "see" the key interrelationships to understand behavior or
change it to nearly the same extent that the behaviorist/systems
perspective does. The behaviorist perspective is often a magnitude of
order greater in its ability to help us achieve these ends.

Fred, in LO20128 you wrote:

>If in Situation A, where I desire Outcome B, patterned set of Actions C
>produces Outcome B, then if I find myself again in a situation similar to
>A with the aim of obtaining an outcome similar to B I might find a
>set of actions similar to C of use to me and I might engage in
>roughly that same pattern of actions.

>What I've written above might sound roughly the same as saying that
>consequences shape behavior, however, it is not. What is stated above
>includes intention or purpose, which is not a part of reinforcement
>theory. ...snip...

Your first paragraph above does indeed sound very much like the
interrelationship that I have been advocating between consequences and
behavior. In your second paragraph above, you bring intention into the
picture as a point of differentiation between your perspective and mine.
You are generally correct in stating that reinforcement theory does not
consider intention. I would, however, like to point out that, while
reinforcement theory and the behaviorist perspective does not tend to
consider intention, it does not necessarily deny the concept of intention.
In many situations, however, bringing intention into the picture can
complicate the picture unnecessarily.

Intention as a concept again begs the debate of whether it is an
ontological phenomenon, an empistomological phenomenon, or both. One
understanding of intention suggests that it is largely an epistomological
attempt, after the fact, to construct a rationale for why we behaved the
way that we did (similar to Karl Weick's concepts of bracketing and
enactment). Even if you don't subscribe to this perspective, including
intention in our present context is plagued with problems and generally
doesn't do much to help us predict or explain the behavior OF OTHERS. You
and I may be able to understand our own intentions, and this may very well
be important for us, but, it is virtually impossible for another person to
understand our own intentions reliably, even if we try to articulate them.
The reason for this is because, so often, what we say we are going to do
(our intentions) and what we actually do don't match up (Chris Argyris has
described this as the gap between our espoused theories and our theories
in practice).

>From a behavioral perspective, this can be explained by examining the
differences between the consequences that follow our verbal behavior (such
as the expression of intentions) and those that follow our overt actions.
For example, the consequences which immediately reinforce what we say
(socially administered reinforcers such as the approval of others present
when we are articulating our intentions) are often absent when we are
acting. In an attempt to gain this social approval (or avoid possible
social disapproval) we may even lie about our intentions. Our actions, on
the other hand, are often reinforced immediately by a completely
independent set of consequences. And, as I mentioned in LO20098, the
immediately available reinforcers much more powerfully influence our
behavior. So, despite our intentions, we may end up doing something
completely different. At a distance (removed in time and space from the
decision point), when we are articulating our intentions, we can adopt a
"rational" perspective where we give appropriate weights to the possible
consequences, but this still does not mean that we will act "with
integrity" to our expressed intentions in the moment of choice (although,
it is a step in the right direction).

Because of these problems, when I am coaching others in the application of
the ABC model, I always, always, always try and keep the discussion and
attention focused on the observable behaviors (or, what you have termed
actions). We can't possibly know with any reliability exactly what
another person is thinking, but we can observe what they do. Often,
trying to infer what the other person is thinking and what their motives
are will get us into trouble. This is the point that is made in the
ladder of inference discussion in the Fifth Discipline and what I was
further trying to illustrate in LO19894. Dealing only with observable
behaviors still gives us plenty to work with in the application of
immediate feedback and contingent reinforcement for desired behaviors as a
method to affect change and improvement.

I have heard some argue that simply trying to change overt behavior is too
limiting in scope, and will ultimately fail because the *thinking*
behaviors are never addressed. While this might appear to be the case, it
actually is not as true as it appears. This is because our thinking
behaviors (mental models, perceptions, inferences, attitudes, etc...) are
also shaped by the consequences of our observable behaviors in the same
manner that our observable actions are shaped. Our mental models, being
very complex and the result of our reinforcement history (the patterns of
reinforcement to which we have been exposed) over our lifetimes, change at
a much slower rate than our observable actions in many cases.

Our mental models certainly impact/influence our intentions, and they
affect how we interpret the actions of others that we see and the outcomes
of our own behavior. They also fit the criteria for *cause* in the
traditional defintion given above. The combination of these factors often
leads to the errant conclusion that the place where we have the greatest
leverage to change behavior (in the action sense) is by trying to address
the mental models first. I believe that the exact reverse is true. The
fastest way to change mental models is to change the overt behaviors
first. Why is this the case? Because getting an individual to change
their overt behaviors opens them up to obtaining new outcomes and
experiencing new reinforcers. If these new reinforcers support the
behavior (demonstrate that the new behavior is more effective at producing
desired outcomes than the behavior that it replaced), the new behavior
will be learned and incorporated into the mental model, thus influencing
it. If we never try the behavior, we will never experience the

If all consequences were experienced simultaneously, all we would have to
do in many cases is prompt the behavior and allow the consequences to
"speak for themselves". Almost any type of antecedent that could get the
behavior to occur once would suffice. In the real world, this approach is
far from adequate because the positive outcomes that may be realized do
not occur immediately, they are far removed in time and sometimes space.
Even worse, the immediate consequences for trying the new behavior are
punishing (increased effort, increased time required up front, acquisition
of a new skill set, etc...). The immediate perception that results is
that the new behaviors provide no benefit and require more effort.
Without the support of other reinforcers to supplement the naturally
occurring consequences, the new behavior will die out (extinction) before
the alternative, desired outcomes can be realized. We must "build the

To illustrate this, think of teaching a child to ride a bicycle. If we
said to our 5 year old one day, "I think that you would really like riding
a bike. Here's a new bike. Why don't you go and try it?" and sent them
out into the driveway to learn on their own, how likely do you think they
would be to ever learn? I don't think that it would be very probable at

Now suppose that they went out and tried, and came back and said, "I
couldn't get it to work". If you responded by saying, "the real key is in
the balance, and this is done most easily if you get the bike moving. In
order to do this, you have to turn the pedals - you know, like on your
tricycle. Now why don't you go back out and try some more?". Would you
have done much to increase the probability that they would learn how to
ride? Again, I don't think so. What would much more likely happen is
that they would fall, hurt themselves and become very afraid of bicycles.
In other words, the immediate consequences available to the new bike rider
for trying bike riding behavior are very punishing. The learning of the
balancing and other behaviors necessary to ride a bike will most likely
never be positively reinforced by the naturally occurring consequences.

So, what do we do in order to teach a child to ride a bike? We shape the
behavior by taking measures to lessen the likelihood that the punishers
will be experienced (for example, training wheels and helmet, and holding
on to the back of the seat to prevent them from falling over), and we get
out there with them and provide immediate feedback on their performance
and immediate positive reinforcement (praise, cheering, affirmation) when
they show that they are "getting it". We gradually move the position of
the training wheels, let go of the seat for short periods, etc... (knowing
that this increases the likelihood that they may experience some of the
punishing consequences) as they improve. At some point, the positive
reinforcers that they have experienced/are experiencing will become
enough that, even if they do lose their balance and fall, they will want
to try again. The positive reinforcers that were delayed from the time
that the initial bike riding behavior was attempted will be experienced
and will be enough to continue to sustain the behavior over time.

To summarize, we build the bridge between the behavior and its ultimate
consequences by providing appropriate feedback and contingent
reinforcement. This is the backbone of the behavioral approach to
affecting change. Cognitive explanations, verbal instructions, good
intentions will never accomplish this goal on their own, although they may
indeed be helpful.

Now to our perceptions and attitudes. In LO19894, I began to describe
some of the adverse "metamorphic" effects that using coercive methods
(specifically punishment, negative reinforcement, and even extinction) to
influence behavior can have on the influencer. Being subjected to those
types of consequences also have similar adverse effects on the target.
This is true whether the aversive consequences are naturally occurring or
socially administered (for example, used as a means of management
control). In the latter situation, the resulting negative perceptions and
attitudes, and the subsequent resistance to further influence will have
the added characteristic of being focused on those who are delivering the
socially administered consequences.

In a laboratory environment, the behavior of, for example, rats will be
vastly different based on the type of reinforcement schedule that is used
to shape and "motivate" their behavior. Rats that have been subjected to
positive reinforcement (e.g., bar pressing behavior reinforced by
receiving food) tend to exhibit several characteristics. First, they may
press the bar as fast as they can for as long as they can. The effort
that they give to their task is often limited only by physical
constraints. They tend to respond well to human handling and they "play
well" with other rats. As a result, they appear friendly. In contrast,
rats that have been subjected to coercive techniques such as negative
reinforcement (e.g., bar pressing behavior reinforced by the removal of a
mild electric shock) behave very differently. They will most often learn
to press the bar and avoid experiencing the shock at all. But they only
press the bar the minimum amount required to avoid the shock or terminate
it when it occurs. In addition, their "disposition" is very different.
They will tremble when handled. They will take measures to avoid
unnecessary contact with humans, and often, other rats as well. And they
may even become overtly aggressive towards humans and other rats.

Humans tend to respond in a manner parallel to that which I have described
for the rats when they are subjected to different kinds of reinforcement.
If negative reinforcement is predominant, cynicism, fear, stress, burn
out, etc... are typical. People in this type of environment will do only
as much as is required to get by and avoid any negative consequences. The
body shows up, but the heart and mind are left at the door. In situations
such as this, people will often give their actions, but they will never
give their soul. Maybe this is what you mean, Fred, when you say that
your behavior is not amenable to control, but your actions are often
negotiable. And it may also explain why you stated that attempts to
control your behavior will engender conflict. On the other hand, people
have been reinforced with positive consequences feel pride and joy in
their work, feel empowered, take ownership and give their discretionary
performance (above and beyond the minimum required). These
characteristics "spontaneously" emerge and can only occur when positive
reinforcement is at work.

Finally, the real success and effectiveness of the use of behaviorist
methods to influence any behavior, whether it is overt action or internal
behaviors like attitudes, perceptions, or mental models, is determined,
not by the "influencer", but by the target. This is because they are the
ones who will make the judgement as to whether or not any attempts to
reinforce are positive or negative, not the influencer. The influencer
may be able to discern how it was received by the "target" by observing
the patterns of the behavior that is shaped by the reinforcers, but they
cannot decide for the performer if it is positive or negative. The
performer will decide if they are only going to give enough action to
avoid a negative outcome (like losing their job), or if they will go
beyond that, and how they fell about it.


Jon Krispin

"You cannot break natural laws, you can only break yourself against them." Cecil B. DeMille
The same can be said of behavioral laws.


"Jon Krispin" <>

Learning-org -- Hosted by Rick Karash <> Public Dialog on Learning Organizations -- <>