An excellent paper by Dennis

$$$$$$ FROM CHUCK TUCKER 951025 $$$$$$

    This is an excellent paper. I find the overall summary of PCT
    very useful and helpful as an attempt to tell others what about
    PCT. I would hope that they could be used to develop more
    precise and detailed instructions on "How to do PCT research."
    Does anyone want to try their hands and fingers and brain at
    such a task? I sure would welcome it? Regards, Chuck



                  Beyond Murray Sidman's Coercion and Its Fallout

                                 Dennis J. Delprato

                            Eastern Michigan University


            In Coercion and Its Fallout, Sidman (1989) proposes that
            coercion is at the root of many harmful and ineffective
            interpersonal and social practices. After critically
            examining coercion, he offers an alternative in the form of
            positive reinforcement based on naturally-occurring
            deprivations. This paper proposes that Sidman's thesis can
            be extended by the incorporation of Goldiamond's
            constructional approach and Powers's perceptual control
            theory. It argues that the latter provides a coherent
            framework for understanding the effects of coercion and how
            there might even be negative side effects of the use of
            positive reinforcement.

             [The 14 points under PCT are abridged in the Psychological
                             Record, 1995, 45, 339-347]

                 Murray Sidman's aim in Coercion and Its Fallout (1989,
            Authors Cooperative, Inc., Boston, MA) is to "indicate a
            critical kind of change that will have to take place in our
            social interactions if we are ever to do something
            constructive about the miseries we currently inflict on each
            other" (p. ix). He proposes that coercion is at the root of
            many harmful and ineffective practices, and that behavior
            analysis provides a framework for noncoercive ways of
            dealing with human behavior.

                 To many psychologists and nonpsychologists Sidman's aim
            may appear incongruous insofar as behavior analysis
            ("behavior modification") seems the epitome of coercive
            control. This is because behavior modifying techniques
            based on punishment, threat of punishment, termination of
            aversive conditions (negative reinforcement), and positive
            reinforcement requiring prior deprivation have become
            identifying marks of the application of behavior analysis to
            social problems. However, although this form of behavior
            analysis is much associated with B. F. Skinner, Skinner
            (e.g., 1948, 1968, 1974) long recommended against "aversive
            control," that is, punishment, threats, and negative
            reinforcement. Sidman, to some extent, seeks to rescue
            applications of behavior analysis from those who ignore
            Skinner's warning of the deleterious consequences of
            aversive control.

            The Case Against Aversive Control

                 Sidman reiterates and elaborates Skinner's arguments
            against attempts to control the behavior of others by
            punishment and other aversive means. Scientific study
            reveals grounds beyond moral and humanitarian considerations
            that question the long-term utility of aversive control.
            Because coercive control does not help individuals develop
            socially acceptable ways of obtaining positive reinforcers,
            they are likely to use the behaviors subject to aversive
            control whenever the opportunity allows, as in recidivism.
            And when the only way individuals have of meeting their
            needs is coercively suppressed, they may develop new
            undesirable ways of coping. Organisms adapt to physical
            punishers, they lose their ability to suppress behavior;
            thus, suppression may require stronger and stronger
            punishers. When used in attempts to suppress behavior,
            painful stimuli can actually come to function as positive
            reinforcers. In other words, self-injurious (e.g.,
            masochistic) behavior can develop whereby the person
            actively seeks out painful physically threatening stimuli.
            New sources of aversive stimuli in the individual's
            environment can be created to expand the range of their
            sources of misery. A common reaction to a punishing
            environment is to do as little as possible resulting in
            withdrawal and inaction. Aversive conditions and
            environments increase the likelihood of attempts to escape.
            The individual may flee, tune out, or drop out. After
            initial exposure to aversive conditions, organisms are prone
            to avoid them. The child is not inclined to attend school,
            the worker is tardy or absent, somatic symptoms may develop
            to support avoidance, bizarre or "neurotic" defense
            mechanisms may arise.

                 One of the most common protections we have against
            coercion is counter-coercion. Persons are contemptuous of
            coercive agents. They retailiate, aggress, counterattack:
            "The best defense is a good offense" (Sidman, 1989, p. 187).
            Finally, the principle of countercontrol covers general ways
            in which we counter coercion. People who are subject to
            coercive control are provoked to control the controllers.
            In addition to several means already mentioned,
            countercontrol can range from very subtle strategems such as
            "turning on the charm" to dramatic ones such as
            assassination and murder.

            Coercion and Positive Reinforcement

                 In the introductory chapter, Sidman defines coercion as
            the "use of punishment and the threat of punishment to get
            others to act as we would like, and ... our practice of
            rewarding people just by letting them escape from our
            punishments and threats" (p. 1). Sidman finds coercion the
            most common way in which people try to influence each other:

            "Make him squirm until he does it right," or "Give him some
            candy, but if he doesn't do what you want, take it away" (p.

                 Like Skinner, Sidman's alternative to coercive
            practices is "positive reinforcement," by which he means
            rewarding people "by letting them produce something good"
            (p. 6). But here things begin to get complicated. In the
            most straightforward sense, application of positive
            reinforcement requires that the individual only can gain
            access to "something good" by performing an action desired
            by someone else who is in charge of the "something good."
            Thus, positive reinforcement requires that the behaver be
            deprived of the reinforcer ("something good") and usually
            this is accomplished by (powerful) others.

                 Certainly deprivation seems coercive, and Sidman does
            add deprivation to his definition of coercion. Sidman's
            solution to coercive practices is positive reinforcement
            ("the hallmark of [his approach to] behavior analysis," p.
            8), and most (or many) applications of positive
            reinforcement use deprivation. This means Sidman must
            formulate a version of positive reinforcement bereft of its
            most common prerequisite, viz., deprivation.

            Naturally-Occurring Deprivation

                 The magnitude of Sidman's problem is highlighted if we
            consider a few uses of positive reinforcement that he
            disavows due to coercive deprivation. We might place
            prisoners in solitary confinement and make social contact
            available if they behave docilely. One may remove a child's
            possession and give it back in return for good behavior. A
            wife may withhold affection and offer it only upon her
            husband's performance of an act she prefers.

                 Sidman's solution to the problem of using positive
            reinforcement that does not require coercive deprivation is
            based on his distinction between artificially-imposed
            (human-made) deprivations and naturally-occurring
            deprivations. All instances of deprivation in the above
            examples are of the former class. They are deliberately
            imposed on the behaver in order to allow for the use of
            positive reinforcers. As such the deprivation involves
            coercion. The controller is using a position of power to
            affect the life of controllee.

                 Naturally-occurring deprivations exist "without social
            intervention; that is the way the the world works" (p. 221).
            According to Sidman, even in the case of individuals who
            initially seem responsive only to a very small number of
            reinforcers, one of which is food, therapists can begin
            teaching skills by recognizing that food deprivation does
            occur in the normal course of events and using food as a
            reinforcer at mealtimes.

                 Apparently because the main purpose of his book is to
            critically examine coercion, Sidman devotes only the final
            two of its 17 chapters to noncoercive positive reinforcement
            as an alternative way of interacting with others. He does
            not provide a framework for generally applying positive
            reinforcement without deprivation. Perhaps the clearest
            guideline he offers is to use social reinforcers whenever
            possible. These include attention, approval, praise, and
            recognition. Nonetheless, Sidman's recommendation to
            exploit naturally-occurring deprivations should allow the
            thoughtful reader to identify some applications on their

            Beyond Sidman's Solution

                 I was drawn to Sidman's Coercion out of a longstanding
            concern with pervasive attempts by individuals and society
            to "strongarm" others to behave in certain ways. In my
            view, Sidman does a marvelous job of arguing against
            aversive control, including artificial deprivation, as a way
            of arranging society, of providing psychological services,
            of managing people, and of interacting with others on a
            daily basis. Remaining to be done here is more complete
            formal scientific documentation of the many principles and
            findings contraindicating coercive control--if moral,
            ethical, and humanitarian ones are insufficient. But the
            data are in the literature.

                 Some might deprecate Sidman's efforts from the point of
            view that he treats coercion rather superficially. That is,
            he does not delve into philosophical, legal, constitutional,
            political, and scientific aspects of coercion as did, for
            example, writers in Behaviorism and Ethics (Krapfl & Vargas,
            1977). But I believe such criticism misses the mark.
            Sidman's sphere is "grass roots." In Coercion he is talking
            to behavior specialists, as well as to concerned
            nonspecialists. If anything, the impact of Sidman's efforts
            likely will be exhibited more by the inspiration it provides
            to the general public to begin changing how they evaluate
            their own and specialists' attempts to "control" others than
            by direct effects on professionals.

                 If Coercion has shortcomings, I suggest they derive
            from other sources. Although the author has argued
            persuasively against coercive methods and has pointed the
            way to alternatives, his theoretical framework does not move
            his solution as far as it might go in the direction of
            large-scale, practical alternatives. We can look beyond the
            behavior analytic perspective Sidman uses to identify two
            supplements that may be indispensable for widespread
            adoption of noncoercive behavioral control.

            Constructional Social Interaction Strategies

                 Sidman comes close to one important supplement to his
            position when he asks to see therapists "not just stopping
            behavior, but constructing behavior" [italics added] (p. 8),
            and when he urges that we teach parents, teachers, police,
            and therapists to "look for desirable actions" (p. 213), not
            for undesirable actions. However, this is as far as he goes
            toward any systematic incorporation of Goldiamond's (1974)
            well-formulated constructional alternative to eliminative
            (coercive) strategies.

                 Goldiamond (1974) distinguished between eliminative and
            constructional ways of interacting with others. Eliminative
            approaches are coercive and in clinical work, tend to treat
            the target of complaints as pathological, emphasize what the
            identified client should not do, focus on eliminating
            behaviors, and do not explicitly identify and develop
            socially acceptable behaviors. Constructional approaches do
            not have these characteristics, instead, the focus is on the
            construction of behaviors. This entails identifying what
            the client should be doing if they are not to be the target
            of complaints and searching for what positive behaviors they
            currently exhibit from which further building can proceed.
            As is the case with Sidman's alternative to coercive
            control, the constructional approach offers positive
            reinforcement as the fundamental procedure. Like Sidman,
            Goldiamond (1974) finds deprivation coercive and he de-
            emphasizes extrinsic reinforcers such as points or tokens.

                 Goldiamond's constructional approach adds support to
            Sidman's case by providing a coherent noncoercive framework
            as an alternative to coercive practices. Furthermore, a
            constructional point of view, expands the range of specific
            noncoercive procedures beyond that of nondeprivational
            positive reinforcement, and it highlights the coercive
            nature of certain other procedures. Modeling, when used to
            demonstrate desired behavior, greatly broadens the
            possibilities for noncoercive influence. And, because they
            do not show the individual what to do, taken alone the
            differential reinforcement of other behavior (presenting
            reinforcers contingent on nonoccurrences of undesired
            behavior) and extinction (withholding reinforcers following
            occurrences of undesired behavior) are eliminative and

            Coercion and Perceptual Control Theory: When Control Systems

                 Despite the encouragement a constructional approach
            provides to Sidman's hope for universal adoption of
            noncoercive control, there is reason to believe this goal
            will not be met unless we move even further away from
            assumptions about behavior made by those who promote
            coercive control. Briefly put, to coercively control the
            behavior of others is to hold to the fundamental assumption
            that psychological behavior is naturally under a particular
            type of control. Despite widespread differences in theories
            of behavior, up to the present the most influential theories
            have in one way or another equated control of behavior with
            lineal cause and effect sequences. Seemingly incompatible
            views have agreed that behavior is the outcome of causal
            controlling forces be they mental, cognitive, genetic,
            physiological, environmental, or some combination of these
            sources. According to such views, all behavior is coerced,
            i.e., compelled by force. The procedures Sidman warns
            against are only the most obvious instances of control by
            lineal causal sequences. From this standpoint, authentic
            alternatives to coercive control would not fit the basic
            explanatory requirements for the science of behavior.

                 In the face of the enormous impact of the causality of
            classic mechanics on psychology, a few theorists have argued
            that living systems are more appropriately explained in
            other ways. With his operant construct and "new" causal
            mode of selection by consequences, Skinner (1953, 1981)
            departed from mechanistic causality. Much earlier, Dewey
            (1896) moved even further away than did Skinner from cause
            -->effect chains when he offered an alternative to the
            mechanistic conception of the reflex arc. Dewey rejected
            external and internal lineal causal sequences and proposed
            that the nominal stimulus and nominal response of reflexes
            were better described as operating continuously or as a
            single unit, i.e, as a circuit. In the same vein, Kantor
            (1959) developed a framework for the whole of psychology
            based on the mutuality of putative stimulus and response.
            Kantor's fundamental unit, the behavioral segment, is
            conceptualized as stimulus<--->response to reflect
            interdependencies among the constituents of psychological

                 Although Skinner, Dewey, and Kantor took important
            steps in the direction of explanations not based on coercive
            cause--->effect processes, the pervasiveness of coercively
            oriented theories and procedures indicates more remains to
            be done if we are to have a convincing perspective from
            which to understand the contraindications of coercive
            control and to develop effective alternatives. Dewey (1896)
            provides but a first glance at nonmechanistic explanations.
            Kantor's (1959) theory lacks demonstrative predictive power.
            And among other considerations, Skinner's (1953, 1981)
            solution based on positive reinforcement that follows from
            the operant formulation and consequential selection is not
            free of negative side effects (Balsam & Bondy, 1983).

                 The possibility of noncoercive behavioral control is
            greatly enhanced by the availability of a theory of wide
            scope that incorporates and expands fundamental insights of
            other attempts to devise explanations not requiring lineal
            causes and effects. According to perceptual control theory
            (Powers, 1973, 1978), our attempts to formulate an orderly
            conception of the universe by making living systems
            continuous with nonliving systems have been marked by
            overstating continuity. Theorists' emphasis on continuity
            between inanimate and animate objects and events no doubt
            contributes greatly to psychologists' continued acceptance
            of lineal cause and effect explanations of their subject
            matter, even though physical science itelf no longer
            compares the world to a machine (Feigl, 1953, Holton, 1973,
            Russell, 1953). Perceptual control theory takes into
            account distinctive characteristics of living systems while
            keeping explanations thoroughly naturalistic, hence within
            the boundaries of science.

                 Because perceptual control theory has its origins in
            cybernetics which has given us modern servomechanisms, one
            might question it as merely another attempt to draw an
            anology between machines and living organisms, with
            attendant mechanistic implications. However, Powers (1973)
            pointed out that that the analogy is the reverse of this.
            Control engineers developed servomechanisms to imitate human
            behavior. They observed that humans could keep an external
            variable at a predetermined state against perturbing
            influences and constructed systems with sensors and feedback
            loops to simulate the human operator. Servomechanisms are
            approximations of human adjustmental behavior.

                 To decrease the likelihood that perceptual control
            theory will be confused with attempts to force control
            systems into psychology without allowing for distinctly
            psychological processes (e.g., Poulton, 1974), I refer to
            Powers's basic starting point as the psychological or
            behavioral control system. Powers now uses the term
            perceptual control system for the same purpose. Here I
            restrict the presentation to only those aspects of
            psychological control systems that are most important for
            showing how this theory handles coercion.

                 Powers's psychological control system is a model for
            behavior. Some of the fundamental assumptions are:

                 1. Physical behavior (the topography of behavior) is
            important but not to be confused with psychological
            behavior. The actions of muscles and organs are not the
            essence of psychological behavior. Behavior is an
            abstraction ultimately known in particular cases by way of
            relationships between variables. Behavior is what the
            organism is doing in terms of meaning or purpose. The
            physical behavior that is observed directly or acts on
            transducers is not itself meaningful.

                 2. Psychological behavior refers to the operation of a
            natural control system that is organized according to the
            effects or consequences its physical behavior has on its
            inputs. Because the organism knows the consequences of its
            physical behavior (inputs) via sensors, we can state that
            psychological behavior is distinguished by relationships
            between physical behavior and perception.

                 3. Psychological control systems are a set of physical
            signals (variables). Control systems (organisms) only
            control quantifiable signals. Thus, perceptual input is a
            measurable quantity. To the behavior analyst as outside
            observer, behavior is an abstraction, but to the behaver it
            is directly experienced.

                 4. Psychological control systems control perceptual
            signals. They control only nonperceptual variables such as
            aspects of the environment secondarily in accordance with
            the perceptual impact of the "external" variables.

                 5. The basic characteristic of all psychological
            control systems is the maintenance of controlled perceptual
            variables at stable levels. The stabilized value of a
            controlled perceptual variable is referred to as a reference
            level or "goal." Theorists who have attempted to adapt
            control systems to psychological events often refer to the
            reference level as the set point, and they often use heating
            and cooling systems as examples of control systems.
            Although the analogy between temperature control systems and
            psychological control systems is legitimate with certain
            caveats, this analogizing invariably fails to make a
            convincing connection to actual psychological events.
            Powers's perceptual control theory seems to overcome this

                 6. A great advance of Powers's psychological control
            system conception over that of earlier adaptations of
            control theory is the inclusion of effects on the controlled
            perceptual variable other than those produced by the control
            system itself. These effects are known as disturbances.
            Disturbances perturb the control system and often contribute
            to moving the controlled perceptual variable away from the
            reference level. Although we may think of disturbances as
            "environmental" effects, the conventional meaning of
            environment as outside the organism is not implied. The
            source of disturbances is simply outside the control system
            under analysis. For example, intraorganismic processes are
            important sources of disruption to controlled perceptual
            signals. Of no little consequence is that the disturbance
            construct opens up the possibility of systematically
            including in a model of behavior what we have typically
            written off as uncontrolled sources of variability and
            called extraneous variables.

                 7. When disturbances move the controlled perceptual
            variable away from the reference level, the control system
            engages in corrective action. In quantitative, working
            models of perceptual control theory the momentary difference
            between the value of the controlled variable and the value
            of the reference signal is quantified as an error signal.
            Physical behavior is the means by which the system alters
            perceptual input to reduce error. As here defined, physical
            behavior includes overt motoric movements, as well as other
            organismic action systems such as the glands and smooth

                 8. The organism need not directly know disturbing
            influences so as to be able to predict them or to verbally
            describe them, in the case of humans. What the organism
            needs to adjust to disturbances is an error-sensor and it
            need not verbally describe the error signal either. Indeed,
            because disturbances typically are continually varying,
            hence the error signal is likewise varying, adjustments are

                 9. Psychological control systems function according to
            negative feedback control. This means that the fundamental
            control or causal process is not lineal and coercive in the
            form of sequences of causes (forces) and effects. Instead
            all physical signals of the control system are
            simultaneously interrelated. The controlled perceptual
            variable, reference signal, and error signal are integrated
            along with the physical actions of the system. While one
            variable is changing, all others are changing. The
            relationships among the system's elements are reciprocal.
            In simple terms, the world acts on the organism and
            simultaneously the organism acts on the world. In
            quantitative models of perceptual control theory, systems
            are described by simultaneous differential equations.
            Feedback control means there is no push-pull, instigators,
            or initiating agents. Psychological control systems do not
            adjust by having one part change independently of other
            parts. Adjustment to changing conditions (disturbances)
            occurs differently. As we will discover below, the nature
            of the adjustment process gets at the heart of why coercion
            is a hazardous way to go about controlling behavior.

                 10. Another noteworthy advance of Powers's theory over
            that of earlier versions of behavioral control systems is
            that it recognizes that real-life behavior is never
            adequately described by a single control system or by one
            level of control systems. Multiple goals (reference levels)
            are potentially operative. In order to describe this in the
            naturally-behaving organism, a hierarchy of control systems
            is required. For example, in a given case the psychological
            behavior of getting to work may be accomplished by driving a
            car. Driving requires putting the key in the ignition,
            turning the key, depressing the accelerator, etc. These
            actions are ultimately carried out by various muscular
            tensions. Getting to work is the goal of a higher-order
            control system that does not directly do anything
            physically. This higher-order system sends outputs to
            lower-level systems, in this example, those involved with
            driving a car. The outputs from the higher-order system
            serve as reference levels for lower-order systems. The
            "bottom line" of this example is the lowest level systems
            controlling muscular tensions whose goals are the outputs
            from control systems immediately above in the hierarchy,
            i.e. car driving control systems. When the entire
            hierarchical system is operating properly, error signals
            produced by disturbances of the higher-level systems are
            corrected by changes in outputs of the systems. These
            result in variations in goals (reference signals) to lower-
            level systems.

                 11. Because it is normal for control systems at all
            levels of the hierarchy to be subject to disturbances, goals
            are constantly changing throughout a behavioral episode.
            The physical behavior detected by outside observers that
            gets the organism from point A to point B, for example, is
            the means by which the behaver "gets what it wants" (i.e.,
            matches perceptual inputs to reference levels, corrects
            error signals, meets goals). The psychological behavior of
            the episode is inseparable from the goals. The hierarchical
            system adjusts to unpredictable disturbances by changing
            goals and, to the external observer, by varying physical
            behavior. If the person is confronted with a dead battery,
            the car-starting control system cannot meet the goal of the
            higher-order getting to work system. The output (goal) of
            the latter may change to calling a friend from starting the
            car; this is met by particular physical behavior, and the
            person proceeds on to work.

                 12. The circumstance just described in which the
            person experiences an unexpected dead battery in the getting
            to work hierarchical system illustrates a mild form of an
            important principle for understanding coercion, that of
            reorganization. We can think of the change to the atypical
            goal of calling a friend as a reorganization of the system.
            Reorganization was brought into play by persistent intrinsic
            error in the system. The rather quick way in which the
            person corrected the error signal associated with the dead
            battery makes it debatable as to whether or not the
            hierarchical control system even reorganized in this case.
            The change of the goal from starting car to calling friend
            was smoothly adapted to. Reorganization has more serious
            consequences when alternative lower-level goals do not
            contribute to resolving error in the higher control system.
            If, for example, getting to work cannot be accomplished
            within a reasonable time, reorganization may result in
            unpredictable and abnormal adjustments manifested by
            physical behavior such as aggression, organic disturbances,
            crying, "giving up," or bizarre perceptions.

                 13. Reference levels (goals) are properties of
            individuals. With the proper observational conditions, an
            external observer can identify someone else's goals.
            However, in ordinary practice, we are rather insensitive to
            others' goals and frequently inadvertantly impose our goals
            on others. When control systems outside an individual
            encounter an individual, when control systems interact, the
            situation is ripe for conflicting goals. For one system to
            reduce error, the other system must increase error. The
            interacting systems are both attempting to control the same
            perceptual quantities, but with respect to different
            reference levels. The result is that the individual is in a
            state of unresolvable intrinsic error. Reorganizing will
            continue as long as the intrinsic error remains. In
            harmonious interactions between interpersonal control
            systems, each may "give," i.e., alter their goals in
            directions that mimimizes error for both. (Of course, the
            interaction is called harmonious because the systems behave
            this way.) In coercive control, the coercive agent does not
            give. The coercee is placed in a state of chronic intrinsic
            error; hence, reorganizing persists with attendant
            countercontrolling consequences that may be exhibited in
            physical behavior that is unpredictable, socially
            undesirable, aggressive, and so on.

            Putting It Together: The Fallout of Coercion, Constructional
            Social Interactions, and Perceptual Control Theory

                 In Coercion and Its Fallout, Sidman makes a major
            contribution by documenting the deleterious effects of
            attempts to coercively control behavior. With his
            suggestion of relying on positive reinforcement based on
            naturally-occurring deprivation, Sidman takes a constructive
            step away from coercion, but does not adopt a framework that
            either makes fallout a natural consequence of coercive
            control or shows how all attempts to control the behavior of
            others, in whatever way, are hazardous. Goldiamond's (1974)
            constructional approach is a straightforward expansion of
            Sidman's thesis, and it contains some implications that are
            basic from the standpoint of perceptual control theory. A
            helping agent operating from a constructional orientation is
            always concerned with "what the patient is after" (their
            goals) (Goldiamond, 1974, p. 30), negotiates mutually
            agreed-upon goals (thus minimizing the likelihood of
            conflicting goals between helping agent and patient), and
            avoids obviously coercive practices.

                 Perceptual control theory provides a coherent
            foundation for understanding how individuals adjust to
            coercion and why there might even be negative side effects
            of positive reinforcement (Balsam & Bondy, 1983). Attempts
            by one agent to control the behavior of another are
            disturbances that establish goals that bring forth
            reorganization until intrinsic error resulting from
            conflicting goals is corrected. Ten years before Balsam and
            Bondy (1983) marshalled evidence for negative side effects
            of reward, Powers (1973) warned against using even
            naturally-occurring rewards to control behavior:

                 If what is withheld is truly needed to correct
                 intrinsic error, then throughout the period of
                 withholding, reorganization will be occurring. The
                 only thing that will stop reorganization is for the
                 person suffering the error to correct or at least
                 minimize the intrinsic error. He will reorganize until
                 he is able to circumvent the withholding--get control
                 of what he needs so that he can prevent intrinsic error
                 before it can trigger off reorganization. In short, he
                 will learn to cheat the system. He cannot help it. He
                 will automatically reorganize anything, including his
                 most precious beliefs, if what he is doing is resulting
                 in uncorrected intrinsic error. However he must
                 rationalize, however guilty he feels, however he must
                 distort his perceptions, however he must become
                 unconscious of his own motives or split himself into
                 independent subsystems, he will cheat and get the
                 reward for himself while everyone else (he supposes) is
                 working to get it in the approved way. (pp. 268-269)

                 In terms of perceptual control theory, coercion is
            arbitrary control: "attempts to make behavior conform to one
            set of goals without regard to other goals (and control
            systems) that may already be controlling that behavior--that
            must already exist, since behavior exists" (Powers, 1973, p.
            259). The possibilities for coercion are even greater than
            noted by Sidman, for arbitrary control is not restricted to
            between-individual interactions. Individuals can subject
            themselves to coercive control and exhibit the same fallout
            as when others are directly involved.

                 The message of perceptual control theory is that
            anything we do to others or to ourselves that creates
            intrinsic error will result in reorganization with attendant
            fallout. Admittedly, perceptual control theory has no set
            of procedures to offer as alternatives to coercive control.
            The basic implication in cases of interpersonal interactions
            is to minimize arbitrary control by doing as Goldiamond
            (1974) suggests--take seriously what others are after,
            negotiate, and work with them in a cooperative way. Others
            should be as informed as possible about where those in
            positions of authority are "coming from." The bottom line
            is to "let others alone" unless there is convincing evidence
            that our intrusions are necessary. As Powers (1973) put it,
            "Control of behavior is not wrong or sinful or irrational or
            evil. It is simply inconsistent with the facts of human
            nature" (p. 271).


            Balsam, P. D., & Bondy, A. S. (1983). The negative side
               effects of reward. Journal of Applied Behavior Analysis,
               16, 283-296.

            Dewey, J. (1896). The reflex arc concept in psychology.
               Psychological Review, 3, 357-370.

            Feigl, H. (1953). Notes on causality. In H. Feigl & M.
               Brodbeck (Eds.), Readings in the philosophy of science
               (pp. 408-418). New York: Appleton-Century-Crofts.

            Goldiamond, I. (1974). Toward a constructional approach to
               social problems. Behaviorism, 2, 1-84.

            Holton, G. (1973). Introduction to concepts and theories in
               physical science (2nd ed.). Reading, MA: Addison-Wesley.

            Kantor, J. R. (1959). Interbehavioral psychology (2nd ed.).
               Granville, OH: Principia Press.

            Krapfl, J. E., & Vargas, E. A. (Eds.) (1977). Behaviorism
               and ethics. Kalamazoo, MI: Behaviordelia.

            Poulton, (1974). Tracking skill and manual control. New
               York: Academic Press.

            Powers, W. T. (1973). Behavior: The control of perception.
               New York: Aldine.

            Powers, W. T. (1978). Quantitative analysis of purposive
               systems: Some spadework at the foundations of scientific
               psychology. Psychological Review, 85, 417-435.

            Russell, B. (1953). On the notion of cause, with
               applications to the free-will problem. In H. Feigl & M.
               Brodbeck (Eds.), Readings in the philosophy of science
               (pp. 387-407). New York: Appleton-Century-Crofts.

            Sidman, M. (1989). Coercion and its fallout. Boston, MA:
               Authors Cooperative, Inc.

            Skinner, B. F. (1948). Walden two. New York: Macmillan.

            Skinner, B. F. (1953). Science and human behavior. New
               York: Free Press.

            Skinner, B. F. (1968). The technology of teaching. New
               York: Appleton-Century-Crofts.

            Skinner, B. F. (1974). About behaviorism. New York: Knopf.

            Skinner, B. F. (1981). Selection by consequences. Science,
               213, 501-504.