Beyond Murray Sidman's 'Coercion and Its Fallout.'
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 countercoercion. Persons are contemptuous of coercive agents. They retaliate, 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. 1).
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 can only 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, namely, 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 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 by 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 2 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 thoughtful readers to identify some applications on their own.
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 clients 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 deemphasizes 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 coercive.
Coercion and Perceptual Control Theory: When Control Systems Interact
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, that is, compelled by force. The procedures Sidman warns against are only the most obvious instances of control guided by assumptions of lineal causality. 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 [right arrow] 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, that is, 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 [equivalence] response to reflect interdependencies among the constituents of psychological events.
Although Skinner, Dewey, and Kantor took important steps in the direction of explanations not based on coercive cause [right arrow] 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 may not be 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 greatly expands fundamental insights of earlier attempts to devise explanations not requiring lineal causes and effects. Perceptual control theory (Powers, 1973, 1978) formulates a psychological (perceptual) control system as 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. Psychological behavior is what the organism is doing in terms of meaning or purpose.
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 perceptual inputs.
3. Psychological control systems are a set of physical signals (variables).
4. Psychological control systems control perceptual signals. They control 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."
6. An important 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.
7. When disturbances move the controlled perceptual variable away from the reference level, the control system engages in corrective action.
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.
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.
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.
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.
12. Hierarchical control systems reorganize (change goals) in the face of persistent intrinsic error. Reorganization has serious consequences when alternative lower-level goals do not contribute to resolving error in higher control systems. 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 inadvertently 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," that is, alter their goals in directions that minimize 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) marshaled 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).
References
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.
GOLDIAMOND, I. (1974). Toward a constructional approach to social problems. Behaviorism, 2, 1-84.
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.
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.
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.
 Printer friendly
 Cite/link
 Email
 Feedback
| |
| Author: | Delprato, Dennis J. |
|---|---|
| Publication: | The Psychological Record |
| Date: | Jun 22, 1995 |
| Words: | 3680 |
| Previous Article: | Problem Drinkers: Guided Self-Change Treatment. |
| Next Article: | Explanatory and predictive roles of inner causes: a reply to Overskeid. |
| Topics: | |