Reinforcement Theory


Responsible Thinking Process (RTP) «




Home Page Free Book on RTP First Book to Read School Statistics Current RTP Trainers Contact Us

Joe Sierzenga Memorial




Perceptual Control Theory, Reinforcement Theory,
Countercontrol, and the Responsible Thinking Process «

By Tom Bourbon (
Houston, Texas, USA, 17 October 1997
Copyrightę 1997 by W. Thomas Bourbon

From their first undergraduate courses on school discipline and classroom management, through their evaluations by administrators when they are seasoned veterans, teachers hear that they are responsible for the behavior of students in their classrooms.

They also hear that they can use reinforcements, especially "positive reinforcements," to control students' behavior. In this document I analyze B. F. Skinner's reinforcement theory, focusing on some important parts of the theory that are unknown to most people. (I know that Skinner said he did not have a theory, but his assertion that reinforcement controls behavior is a theoretical statement, not a statement of fact.)

I show how Skinner's theory led to discipline programs in which educators try to control students' behavior with reinforcements and I describe why those programs fail. I also discuss the ways that perceptual control theory (PCT) and Ed Ford's responsible thinking process (RTP) are different from the idea that reinforcement controls behavior.

I was trained as a research psychologist (see "biographical information" at the end of this document). I believe that many psychological theories, including Skinner's reinforcement theory, have inspired applications that can be harmful to people, especially to children. My critical analysis of Skinner's theory is not a criticism of educators who have used reinforcement. I admire and respect people who have dedicated their lives to education.

I hope that my analysis of Skinner's theory, and of the earlier cause-and-effect theories from which it came, will help educators better understand some of the things that have happened to them and to their students. I also hope my efforts will begin to atone for the harm done to many students and educators by some members of my former profession.

This document has five parts.

In Part 1, I briefly describe cause-and-effect theories about behavior; Skinner's is a cause-and-effect theory. I also analyze Skinner's idea that reinforcement controls behavior, and I show how that idea has been translated into discipline programs for schools. In Part 2, I discuss Skinner's ideas about a phenomenon he called "countercontrol," which I believe is never mentioned by people who say that teachers can use reinforcement to control students' behavior. In Part 3, I discuss a few additional topics from behaviorism, as it is applied to discipline programs in schools. In Part 4, I describe core principles in perceptual control theory. I show how those principles explain Skinner's control and countercontrol, and how they invalidate reinforcement theory as a foundation for discipline programs in schools. In Part 5, I briefly describe how Ed Ford's responsible thinking process differs from discipline programs that embody reinforcement theory.

At no time in this document do I address specific details from the various discipline programs in which educators use reinforcement to control students' behavior. There is no need for me to do that. Instead, I focus on the principles that are common to all reinforcement-based programs. No matter how different they might seem to be from one another, and no matter how different their creators might say they are from one another, every discipline program that relies on reinforcements suffers from the same conceptual and methodological problems that are found in Skinner's reinforcement theory. There are no exceptions.

Part 1: The Theory that the Environment Controls Behavior:

A Journey Through the World of Cause and Effect

Many traditional discipline programs rely on practices where one person, usually a teacher or administrator, tries to control the behavior of another, usually a child. Some programs use brute force, such as spanking, or hitting with an object like a ruler, a paddle, or a cane. Often, when a particular level of physical force does not work, the hitter increases the force by hitting harder, or more often.

Those practices reveal a belief on the part of the hitters that behavior is caused by prior physical events in the environment. Their belief is a direct carryover from theories that explain how physical forces cause non-living things to change: pushing a cup causes it to move across the table, and the harder you push, the faster and farther it moves; applying heat to a rock causes its temperature to rise, and the greater the heat, the faster and farther its temperature rises.

That kind of direct relationship between the magnitude or strength of a physical force and the magnitude of its effects on inanimate things is an example of the physical laws of cause-and- effect (C--E). Discipline through brute force is a direct attempt to apply C--E laws to people. Hitters believe that the application of force will change the child, who will then behave the way the hitter wants.

When they are applied to living things, especially to people, the traditional laws of cause and effect look almost magical. For example, it is as though the hitter believes, "If I just hit this child hard enough, or often enough, then somehow, some time, I will cause the child to behave the way I want."

A variation on hitting is shouting or yelling. Like the hitter, the shouter seems to believe that the key to controlling a child's behavior is to apply enough physical force; "If I just shout at this child loudly enough and often enough, then I will cause the child to behave the way I want." Another variation on C--E discipline requires the teacher to stare at a disrupting student until the student settles down.

I suppose that if a little bit of staring does not work, then the teacher must stare harder and harder, until finally the student cannot resist the "force" or "power" of the stare! Most people do not realize that the same model of cause-and-effect that is used in physical science is part of every discipline program that uses reinforcement, or punishment, to control children's behavior.

In this paper, we will examine several variations on the theme of lineal cause and effect, as it is applied in discipline programs. ("Lineal" means acting in a straight line, as in the theory that a causal chain runs directly from environmental causes to behavioral effects.)

Hitting and yelling are not the only ways that adults try to control children's behavior. Other programs rely on locking students in school (detention), or locking them out of school (expulsion or suspension). Many people who practice discipline by lock and key seem to assume that the mere act of locking students in or out of school will cause them to change in unspecified ways that result in "good behavior."

Discipline through lock and key is another example of the belief that behavior is an effect, determined by prior causes in the environment. Just like hitters and shouters, people who rely on detention and suspension seem to believe that if a little bit of a treatment does not change a child's behavior, all they need to do is apply more of it; if after one day of being detained or suspended the student still "misbehaves," then try a few more days.

If a few more days don't work, try a few weeks. Followed to its logical conclusion, eventually a child might be banished from school altogether. Just like hitting and yelling, a reliance on detention and suspension is an example of magical cause-and-effect thinking -- magical because people are not inanimate objects that always change in direct proportion to physical forces that are applied to them. Instead, people often act to eliminate or cancel the effects of things that are done to them, a topic I discuss in my section on perceptual control theory.

Some of the more sophisticated discipline programs use techniques inspired by psychology and behavioral science, techniques like those in behavioristic psychology.

Radical behaviorism, developed by B. F. Skinner, is the school of psychology that says one person can use "reinforcement" (sometimes incorrectly called "reward") to control the behavior of another.

In schools, behaviorism turns into programs where teachers and administrators try to "reinforce" students every time they "emit" certain "behaviors" the adults want to see. In those schools, educators often withhold "reinforcement" when students "misbehave."

Popular reinforcers used in schools include things like candy, "nice" words, smiles, "positive" affirmations, gold stars, happy-face stickers, parties at the end of a specified time, and pizza in the classroom. Let us look closely at some of the theoretical assumptions and methodological practices in behaviorism, before we too hastily accept the idea that the environment controls behavior.

John B. Watson's Behaviorism:

In the United States, behaviorism began with the work of John B. Watson, who viewed psychology as a science of reflexive stimulus-response (S--R) relationships. In 1913, Watson wrote that, "Psychology as the behaviorist views it is a purely objective branch of natural science. Its theoretical goal is the prediction and control of behavior." He asserted that, "In a system of psychology completely worked out, given the response the stimuli can be predicted; given the stimuli the response can be predicted."

(Quotes from, "Psychology as a behaviorist views it," Psychological Review, 1913, Vol. 20, pages 158-177) Watson explained all behavior as responses to stimuli, with a specific response following a specific stimulus in a purely reflexive manner. Watson's S--R behaviorism perfectly embodied the model of mechanical cause and effect from the physical sciences.

Learning was called "conditioning" and it was said to entail nothing more than the development of new associations between stimuli and responses -- between physical causes in the environment, and behavioral effects.

Later, Watson wrote that, "It is the business of behavioristic psychology to be able to predict and control human activity." (1930, Behaviorism, Revised Edition, Chicago: Phoenix Books, page 11) In the "Introduction" to the same book, Watson wrote, "The raw fact that you, as a psychologist, if you are to remain scientific, must describe the behavior of man in no other terms than those you use in describing the behavior of the ox you slaughter, drove and still drives many timid souls away from behaviorism."

Pretty heady stuff, Watsonian behaviorism; the environment controls behavior, whether in the slaughterhouse, or in the classroom. Right now, almost all research in cognitive science and neuroscience uses the same description of classical cause-and-effect conditioning developed by Watson, nearly a century ago. Obviously, many contemporary cause-and-effect techniques of discipline (like hitting, shouting, staring, and restraining) are also close to Watson's C--E model of behavior. That might be fine, if people behaved like inanimate rocks and raindrops, but they don't.

B. F. Skinner's Radical Behaviorism:

B. F. Skinner is generally credited with establishing "neo-behaviorism" as the successor to Watson's S--R behaviorism. Skinner emphasized the theoretical idea that "reinforcement" from the environment controls behavior. In that sense, Skinner's ideas are more closely related than Watson's to most contemporary discipline programs in schools, but they share the mechanical model of lineal cause and effect. For example, on the idea that the environment controls behavior.

Skinner said:

"The external variables of which behavior is a function provide for what may be called a causal or functional analysis. We undertake to predict and control the behavior of the individual organism.

This is our 'dependent variable' -- the effect for which we are to find the cause. Our 'independent variables' -- the causes of behavior -- are the external conditions of which behavior is a function. Relations between the two -- the 'cause-and-effect relationships' in behavior -- are the laws of a science.

A synthesis of these laws expressed in quantitative terms yields a comprehensive picture of the organism as a behaving system."

(B.F. Skinner, 1953, Science and human behavior, New York: The Free Press, page 35) Later, in a dialogue with R. I. Evans (Evans, 1968, B. F. Skinner: The Man and His Ideas, N.Y.: Dutton), Skinner said: "If I can't give a clean-cut statement of a relationship between behavior and antecedent variables, it is no help to me to speculate about something inside the organism which will fill the gap."

"We need a complete account at the external level. After all, the organism cannot initiate anything, unless you assume it is capable of whimsical changes. As a determinist, I must assume that the organism is simply mediating the relationship between the forces acting upon it and its own output." (pages 22, 23) In other words, Skinner said that an organism, including a person, is merely an object that connects mechanical causes to their effects, and the effects are the organism's behavior.

In those passages, we see that Skinner's theory is every bit as dependent on the idea of linear cause and effect as is Watson's. Both of them describe living things as though their actions are like those of inanimate objects, buffeted here and there by the mechanical forces of nature. Both of them describe behavior as an effect, a consequence of causes in the environment, which would mean, of course, that the environment in the classroom controls a student's behavior.

Deprivation and Positive Reinforcers:

Experts who say educators can use positive reinforcers ("positive consequences") to control students' behavior usually don't tell the educators they must first deprive students of whatever they plan to use as a reinforcer.

As far back as 1938, Skinner described how deprivation is necessary when a person wants to use reinforcers to control another organism. That was when he wrote, in the "preface" to his book, The Behavior of Organisms: An Experimental Analysis (N.Y.: Appleton-Century-Crofts): "The simplest contingencies involve at least three terms -- stimulus, response, and reinforcement -- and at least one other variable (the deprivation associated with the reinforcement) is implied." I think it is interesting that, even relatively early in his work, Skinner was waffling on his acknowledgment of the central role deprivation plays in operant conditioning.

Notice how he says there are at least three terms, "and at least one other variable (the deprivation associated with the reinforcement) is implied." Make no mistake about it, deprivation is not merely "implied;" it is the essential action that allows behaviorists to create the illusion that reinforcement controls behavior, whether the behavior occurs in the experimental laboratory, the classroom, or the slaughterhouse.

Watson's S-R behavioristic psychology described behavior as reflexive responses to stimuli. Skinner's radical behaviorism described behavior as responses in the presence of stimuli that were associated with reinforcers, where the reinforcers worked because the organism was deprived.

The "formula" in Skinner's psychology is stimulus-response-reinforcement (+deprivation): S-R-R (+D). (Sometimes the formula is written, Antecedent-Behavior-Consequence, or A-B-C. I would add +Deprivation , making the formula A-B-C (+D).) Without deprivation, or denial of access to something an organism wants or needs, there is no reinforcement. If a rat eats pellets of food until it stops of its own accord, then the behaviorist cannot use food to "reinforce" the rats actions.

If a child has all of the attention he or she needs or wants, then a teacher cannot use attention to "reinforce" the child's behavior. No deprivation, no control of behavior. No control of behavior, no discipline program.

The following passage, from one of the "Bibles" of laboratory research methods for radical behaviorists, describes how completely they relied on deprivation to create the illusion that reinforcers control behavior. In 1957, C. B. Ferster and B. F. Skinner wrote that: "Birds are sometimes matched on deprivation level as measured by the rate of responding under a variable-interval schedule. Each bird is run for a session of fixed length each day.

A criterion rate of responding is selected to which each bird is to be adjusted. This is expressed as a number of responses per session. The weight to which each bird is fed is adjusted upwards or downwards in terms of the difference between the number of responses emitted on that day and the criterion. A scale of adjustment is chosen as the number-of-grams added or subtracted from the weight for each 1,000 responses shown in that difference." (Ferster & Skinner, 1957, Schedules of Reinforcement, N.Y. Appleton-Century-Crofts, pages 29, 30)

Let me translate that passage. Each bird is deprived of food until it is somewhere around 80% of its natural body weight, then it obtains additional food only if it pecks a key that produces pellets of food after variable intervals of time have passed.

The experimenter decides, in advance, how many times he or she wants to see every bird peck the key during an experimental session of a standard length.

By some means, the experimenter must "adjust" each bird until it produces that rate of pecking. The adjustment is made by noticing each day whether the bird pecked more times, or fewer times, than the experimenter wanted it to peck.

Next, depending on the direction and magnitude of any difference between what the experimenter wants to see, and what he or she does see, the experimenter adds to the amount of food the bird can eat in its home cage, or subtracts food from what is available in the home cage. In other words, the experimenter adjusts the degree to which the animal is deprived of food, until the animal pecks as often as the experimenter wants it to peck. Of course, all the while, the animal is doing whatever it can to obtain as much food as it needs or wants.

There you have radical behaviorism in a nutshell -- the "science of behavior" that many authorities say holds the key to successful discipline in the classroom. Anyone who tells educators that they can use reinforcers to control students' behavior should also tell them just how much operant conditioning depends on prior deprivation, of the persons who are to be controlled, by the person who would control them.

Either the controller does that, or the controller must find someone who is already deprived, and deliberately leave the controllee in that deprived state, until the controllee does what the controller wants him or her to do.

Sometimes a controller even creates in the controllee a desire or demand for "the reinforcer," as when a teacher gives students candy, then lets them know they can get more, but only if they perform the actions the teacher wants to see.

The least I would say is that any expert on discipline-through-reinforcement who leaves out that interesting little detail about the need for prior deprivation is careless. I could say more than that.

It doesn't sound nice to tell educators that they must first deprive students of desirable things in order to control them with "reinforcers," therefore many experts simply don't tell them. Another thing experts often forget to mention is that any time one person (let's say a teacher) controls another person's behavior (let's say a student's behavior), the controllee (the student) can control the controller (the teacher) in return, and the controller cannot prevent it.

That is the phenomenon Skinner called "countercontrol." If I were a betting man, I would wager that hardly any educator who has heard about controlling behavior through reinforcement has ever heard about countercontrol. Let me tell you a little about it, rather, let B. F. Skinner tell you about it.

Part 2: B. F. Skinner on "Countercontrol."

If we return to Evans's dialogue with Skinner (R. I. Evans, 1968, B. F. Skinner: The Man and His Ideas, N.Y.: Dutton), we see the following passage: "When people are being pushed around, controlled by methods which are obvious to them, they know who and what is controlling them, and when things become too aversive, they turn against the controller." "When you know what is being done to you, you know where to turn in order to escape.

But some kinds of drugs and some kinds of positive reinforcement can be used without identifying the controller. Even though you are inclined to revolt, you don't know whom to revolt against." (page 53). In other words, people who are being controlled will try to escape, or to turn on the controller, but if the controller is sneaky enough, then even though the controllee will want to run away or to attack someone, maybe the controllee won't be able to figure out from whom to run, or on whom to turn.

That certainly sounds like a good foundation for discipline in the classroom, doesn't it? Skinner had a name for the phenomenon he described in that passage: countercontrol. Was countercontrol an idea Skinner came across only later in his career? Did Skinner think it was a rare phenomenon? No, and no.

I believe Skinner's most extensive discussion about countercontrol was in 1953, in his book, Science and Human Behavior. (N.Y.: The Free Press) That was the book in which Skinner demonstrated how to apply the principles of operant conditioning, which he had developed while working with animals, to human beings.

Skinner made very clear his idea that a person's behavior is totally controlled by that person's history of exposure to reinforcing stimuli in the environment.

A person can initiate nothing. Everything is determined by the person's environment, which might include a controller who is determined to control the person. If that sounds like a paradox, don't worry about it; Skinner says it is not a paradox.

Then there is the following material about countercontrol

". . . control is frequently aversive to the controllee. Techniques based upon the use of force, particularly punishment, or the threat of punishment, are aversive, by definition, and techniques which appeal to other processes are also objectionable when, as is usually the case, the ultimate advantage to the controller is opposed to the interest of the controllee." "One effect upon the controllee is to induce him to engage in countercontrol. He may show an emotional reaction of anger or frustration including . . . behavior which injures or is otherwise aversive to the controller."

"Because of the aversive consequences of being controlled, the individual who undertakes to control other people is likely to be countercontrolled by all of them." "The opposition to control is likely to be directed toward the most objectionable forms -- the use of force and conspicuous instances of exploitation, undue influence, or gross misrepresentation -- but it may extend to any control which is 'deliberately' exerted because of the consequences to the controller."

"The countercontrol exercised by the group and by certain agencies may explain our hesitancy in discussing the subject of personal control frankly and in dealing with the facts in an objective way. But it does not excuse such an attitude or practice.

This is only a special case of the general principle that the issue of personal freedom must not be allowed to interfere with the scientific analysis of human behavior. As we have seen, science implies prediction and, insofar as the relevant variables can be controlled, it implies control. We cannot expect to profit from applying the methods of science to human behavior if for some extraneous reason we refuse to admit that our subject matter can be controlled." (pages 321, 322)

I will translate that passage.

Control is often aversive to the person who is controlled, even if the controller thinks it won't be. In any case, control is usually to the advantage of the controller, over the interest of the controllee. 

Control often induces a controllee to countercontrol the controller, by doing something the controller does not like, or by doing the controller bodily harm. Because being controlled is aversive to the controllee, every person who tries to control other people is likely to be countercontrolled by every person he or she tries to control.

(What a remarkable thought!) While it is most likely that countercontrol will be directed at obviously objectionable forms of control, it may extend to any control that the controllee dislikes, and that could well be any control, period.

To continue my translation of Skinner, perhaps the countercontrol they have faced from controllees accounts for why behaviorists are sometimes reluctant to talk openly about what they do. But just because they can expect to be opposed by everyone they try to control, that is no excuse for behaviorists to shrink from their task of controlling human behavior. How else can "we" expect to profit from controlling people, if, just because they object to being controlled, we behaviorists fail to control them? (I wonder to whom Skinner refers when he says that "we" will benefit from this exercise in control over those who object to being controlled. Is it behaviorists, or humankind?)

There, in the other half of the nutshell, is a part of radical behaviorism, and operant conditioning, and control of behavior by reinforcement, that is probably known to only very few people. Even if we allow that reinforcement controls behavior, which I do not believe is true, then any person who tries to control another is likely to be countercontrolled in return.

Once a controllee figures out that someone is controlling him or her, and once the controllee decides to countercontrol the controller, then there is nothing the would-be controller can do to escape from countercontrol. I do not know the proportion of controlled people that Skinner really thought would use annoyance or injury to countercontrol their controllers. If we take him at his word, then nearly all controllees will act that way.

What is certain is that control through reinforcement creates the opportunity for every controllee to engage in countercontrol, and that opportunity exists for so long as the controller persists in using reinforcements and their associated deprivation. The only way the controller can escape from the possibility of countercontrol is to give up trying to control the actions of the controllee.

I have examined many discipline programs that say educators can control student behavior through positive reinforcement. I have not found one program where the authority on discipline even mentions the subject of countercontrol. Surely, experts on the application of operant conditioning in education know what Skinner said about countercontrol in his classic book on science and human behavior. There must be good reasons why the experts withhold that information from educators.

During my visits to schools, whenever I have talked about countercontrol I have learned that all teachers were trained to believe they could use reinforcement to control students' behavior. At one time or another, practically every teacher has been evaluated for how well he or she controls students' behavior.

Virtually every educator has reported the frustration of being countercontrolled by students, yet only one educator has reported ever hearing about countercontrol, whether in university classes on behavior management, or in training sessions from experts on discipline. (She heard it 20 years ago, in an obscure university, from a professor who was 80 years old.) How could that be?

One reason for the remarkable silence about countercontrol in discipline programs could be that very few behaviorists have written about it. After 1953, even Skinner did not write about it extensively. Considering that countercontrol is an important and unavoidable consequence of trying to control other people, I would expect to see it discussed in every major article and book about behaviorism, and learning theory, and operant conditioning, and practical applications of behaviorism, like discipline programs. Most often, it is not even mentioned in those sources. I wonder why?

Part 3: Some Unfinished Business on the Subject of Controlling Others

Reinforcement, Punishment and the Endless Search: We need to look a little more closely at the behaviorists' concepts of "reinforcement" and "punishment." Those words refer to the "C" in the behaviorists' formula, A-B-C +D. Reinforcement and punishment are supposed to be consequences of behavior that come to control behavior. In a behavioristic discipline program, reinforcement and punishment are allegedly the means by which the educator controls a student's behavior. Using the definitions from Skinner's behaviorism, a reinforcement is a consequence of behavior that increases the frequency or strength of the behavior, whereas a punishment decreases the strength or frequency of the behavior. Notice the cause-and effect quality of the definitions: A reinforcement or punishment is something a teacher does to a student (as a cause), in order to change the student's behavior (as an effect). It is as though reinforcers and punishers possess special "power" to reinforce or punish; behaviorists often describe reinforcers as having such "power." The behaviorists' attribution of reinforcing power to inanimate objects and events is another idea from C -- E theory that has an almost magical quality.

Behaviorists often describe how difficult it can be for a controller to find a reinforcer or punisher that seems to "control" the behavior of a particular person (the controllee). Often the controller must test and discard many likely candidates before locating one that seems to work. Sometimes behaviorists call that process of trial and error testing "the search for the reinforcer (or the punisher)." The literature for various behavioristic discipline programs rarely discusses "the search" in detail. The silence is unfortunate, in that "the search" gives us a hint that the controllee determines whether any particular consequence of behavior is important or not; the controller cannot make that determination in advance. The silence is one more example of how developers of programs for discipline through reinforcement (and punishment) have left educators unequipped to evaluate those programs. For example, a teacher might think that spanking is a punisher -- a consequence that will decrease the frequency of a student's disruptive behavior. What if, after being spanked, a student disrupts more and more? In that case, a behaviorist would say that spanking is a reinforcer that increases the frequency of disruption. It might be the case that, for whatever reason, the student wants to be spanked. Perhaps the student uses spanking as a way to prove to peers that he or she is "tough."

For another example, a teacher might discover that happy-face stickers seem to work as reinforcers -- consequences that seem to increase the frequency of certain actions by students. The teacher could easily believe that stickers have the power to control students' behavior. But what does it mean if students begin to ignore the stickers? What if the frequency of behavior the teacher wants to see begins to decrease, while behavior the teacher does not like begins to increase? Have the stickers "lost their power?" No. They never had any power. Instead, for whatever reasons, for a little while students wanted stickers, then they decided they didn't want stickers any more. The students always determine what they want and don't want. For so long as they want to perceive themselves getting a particular thing, like a sticker, then students will behave to control for their own perceptions of having stickers and from the teachers' point of view, stickers will seem to work as reinforcers. But as soon as students don't want stickers any more, that inning in the game is over. On the issue of what they want and don't want, students are always in charge. If educators insist on using "reinforcers" and "punishers" to "control" students' behavior, then they must be prepared to renew "the search for the reinforcer or punisher" every time the students change their minds about what they want and don't want. Even if educators succeeded in the most recent round of "the search," they will still be inescapably susceptible to another challenge -- the possibility that students will realize the educators are controlling them and then will decide to countercontrol the educators in return. The only way educators can escape from endless repetitions of "the search" and from the threat of being countercontrolled is to give up their attempts to control students' behavior.

Points and levels: In many discipline programs, teachers use elaborate systems of "points" or "levels" to deliver reinforcers and punishers that are supposed to control students' behavior. In the spirit of all C -- E programs, giving and taking away points or levels are things that teachers do to students in the hope of causing students' behavior to change. Some programs use "good" points, others use "bad" points. Good points are used like tokens, or other positive reinforcers. Teachers give students specified numbers of points for committing approved actions, then students exchange their points for goods or privileges they want. Of course, at the start of the program teachers must deprive students of those goods or privileges, or else they cannot serve as "reinforcers." Bad points are used like punishers. Teachers award them when students commit unapproved actions. A student who accumulates a specified number of bad points loses certain rights or privileges.

Point systems reveal the convoluted logic of behavioristic discipline programs, a C -- E logic that requires educators to perform impossible tasks. To receive good points is supposed to be positive reinforcement (students' actions are followed by the delivery of something) that causes good behavior to increase. Sometimes a student who accumulates good points, then does something bad, must give up some of the good points. The loss of good points is supposed to be a punishment that causes bad behavior to decrease. On the other hand, bad points are supposed to be punishers that cause bad behavior to decrease. If a student who accumulates bad points does something good, then some of the bad points might be taken away. The loss of bad points is supposed to be negative reinforcement (students' actions are followed by the removal of something) that causes good behavior to increase.

If you are confused by this logic, please don't think I created it. I am simply reporting it and it confuses me. In fact, many authors of textbooks of psychology make a mess of the relationships between allegedly positive and negative reinforcement, and punishment. Even if we understand the logic of reinforcement and punishment, and even if we believe it is sound, we still must face the fact that students alone determine what they want and do not want. They alone determine whether points will seem to be good or bad. Even if we use points and see the behavior we want from students, we remain vulnerable to countercontrol.

In discipline programs that use levels, there are good levels and bad levels. In programs that use good levels, when students perform good actions, they are allowed to advance up the levels, from a state of relative deprivation and restriction, to one of privilege and freedom. If they "slip" and commit bad actions, then they must slide back down the levels, from a state of relative privilege and freedom to one of relative deprivation and restriction. Moving up the good levels is supposed to be positive reinforcement (students' actions are followed by the delivery of something) and to cause good behavior to increase. Sliding down the good levels is supposed to be punishing and to cause bad behavior to decrease. In programs that use bad levels, when students perform bad actions, they move down from levels or relative privilege and freedom, to levels of diminished freedom and privilege. Conversely, when they perform good actions, students move up from levels of reduced freedom and privilege to levels of higher freedom and privilege. Moving down bad levels is supposed to be punishing and cause bad behavior to decrease, while moving up bad levels is supposed to be negatively reinforcing (students' actions are followed by the removal of something) and cause good behavior to increase. (This is not my logic!) Everything that I said about points is true of levels.

Institutionalization: What happens if a behavioristic discipline program seems to work the way people expect? Educators deliver reinforcers and punishments and students behave the way educators want. Everything looks fine. Educators conclude that students have truly learned to behave as they should. Everyone concludes that reinforcers and punishers really do control students' behavior.

Imagine that the students are labeled as emotionally disabled or disturbed or impaired (labels vary from state to state, and country to country) and are from a specialized unit in the school. The discipline program seems to work and the students are sent back to regular classrooms. From what I have seen in many schools, within a short time large numbers of those students return to the special unit. In some schools, all of them return. Similarly, students in special schools, or in correctional facilities, often seem to "acquire appropriate behaviors" under a behavioristic discipline program. They are released from the facility, but large numbers of them return within a short time.

The high probability that students who exit a program or facility will return to that place is often called the recidivism problem. I believe that recidivism often occurs when, prior to their release from the program, students were merely complying with educators' expectations. For a while, the things students wanted and did not want were similar to the things educators wanted to use as reinforcements and punishments. For a while, students refrained from countercontrolling educators. Solely as a matter of convenience to themselves, students complied with educators' expectations for them -- students were "going along with" educators to get what they (students) really wanted. When release from a program or facility is the ultimate reinforcer in a point system, or the highest level in a level system, students will often willingly "exhibit appropriate behaviors," and "suppress inappropriate behaviors," for however long they must until they earn their release.

Once they are released, those students enter a world where the institutional schedules of reinforcers and punishers no longer apply, and where no one is watching and waiting to reward or punish their every move. The students have not learned to think for themselves or to set their own standards for conduct. They have not learned to set and achieve their own goals without unnecessarily disturbing other people. In their new settings "on the outside," some of those conditioned students will perform appropriate acts, then ask the person in charge, "What do I get for that?" When they do not receive the reinforcers they expect, many students disrupt their new surroundings and are sent back to the special program or institution -- back to where educators thought they had conditioned the students to make the proper responses, all the time.

Students who are subjected to discipline through reinforcers and punishers often become "institutionalized." That is an old term that has always signified that students learn how to show the authorities exactly the behavior they want to see, until the authorities mistakenly conclude that the students have undergone a fundamental change, like a cure or a rehabilitation. Compliance is not a cure.

Schools that use behavioristic discipline programs, like those with points or levels, unwittingly teach a skill, compliance, that many adults use later to the detriment of society. In many states, including my own, laws from the legislatures, or arbitrary rulings from the courts, require that prisons use systems of points or levels to reward prisoners for "good behavior." If prisoners "exhibit appropriate behaviors" and "suppress inappropriate behaviors," they earn relative privileges and freedom within the prison. They also accumulate points to exchange for time off their sentences. Many prisoners who buy time and obtain early release immediately resume their careers in specializations like serial rape and serial child abuse or molestation. Society is outraged, but the legislatures and courts believe everything is as it should be. After all, behavioral scientists, from Skinner to the very latest authorities on the subject, say that programs of reinforcement and punishment, often packaged as systems of points and levels, are exactly what we should use in our prisons and schools.

Boxes and trust: In spite of all the fallacies and failings I just described, cause-effect theory and the applications it spawned still dominate the behavioral and social sciences and the brain sciences, as they have for most of the past century. How could that be, given that absolute, unfailing control of one person by another is an illusion?

For one thing, most young children probably trust most adults. Therefore, most children accept as gifts the "rewards" that adults give them. If those rewards were genuine gifts, there would be no problem, but many adults really intend for the "gifts" to be their keys to gaining control of children's behavior. As soon as the first child recognizes that the "gifts" are actually bribes, the dance of control and countercontrol begins. Adults, who had their brief and shining experience of what they thought was control of another person, frantically search for new rewards that will restore their control over children. The dance continues.

For another thing, the kind of control advocated by Skinner and the radical behaviorists seems to work -- under the right conditions. The "scientific" radical control of behavior begins in an animal laboratory. A person houses a small experimental animal in a small wire cage, then deprives it of food until is loses at least twenty-percent of its body weight. Then the person lifts the animal from its home cage and places it in a smaller box with a lid that locks from the outside. (Why must there be a lid and a lock?) Only then does food seem to control the animal's behavior. I respectfully suggest that, when it is applied to people, "control of behavior through reinforcement" seems to work best in circumstances that most closely approximate those in the laboratory -- cages and locks (prisons and other institutions where people are confined); controllers who are larger and stronger and more experienced than those who are controlled; controllers who dominate access to resources or conditions that are important to those who are controlled. In a quest to use reinforcers to control students, do we want to approximate those conditions in our homes and schools?

Cause-effect and the decade of the brain: Cause-effect theory permeates all sciences of behavior and of life, including the brain sciences. A few years ago, we entered a presidentially-decreed "Decade of the Brain." The idea was that we had learned so much about how the brain works in recent years that, with one final decade-long push, we would unlock the last of the great mysteries about our minds and brains. Well, the great decade ended at the final second of 1999. What do we have to show for it? From what I can see, we have a lot of suggestions and mandates from brain scientists, telling educators how to reform education so as to take advantage of all that we know about the brain.

I respectfully suggest that brain science tells us hardly anything about how the physiological activity in our brain is related to our behavior. Brain science certainly cannot support sweeping reforms in education. Why do I say that? The answer is simple. Virtually all research on brain function uses cause-effect procedures from behavioral science. Nearly all brain scientists try to explain how the brain connects stimuli to responses, but that is not how we behave. I do not know of a major program of brain research that looks for information about how the brain participates in the process by which we behave to control our own perceptions.

So what? Let me give you one example. Many authorities on the subject of "attention deficit \ hyperactivity disorder" (ADHD) say that the brains of ADHD children are "defective." Consequently, the children "cannot control their own behavior." But people do not control their behavior; they control their perceptions. Whatever is happening when a child is labeled with ADHD, it is not that the child cannot control his own behavior.

If an expert believes that an ADHD child cannot control her behavior, then what "treatment" does the expert recommend? Some widely-respected experts on brain science and ADHD say that teachers must use radical control of behavior by reinforcement for "right" behavior, combined with precisely-targeted punishment for "wrong" behavior. If the child "cannot control her own behavior," then the teacher must control it for her. That's it. So much for the Decade of the Brain.

Part 4: Perceptual Control Theory

A detailed description of PCT is in my document that accompanies this one on Ed Ford's web site, The companion document is titled, "Perceptual Control Theory, Reality Therapy & the Responsible Thinking Process." I recommend that you read it for more details about PCT. Here, I reproduce some of "Part 1," the basic introduction to PCT, from that document, then I describe how we use PCT to explain Skinner's phenomena of control and countercontrol.

Bill Powers and PCT: In the early 1950s, William T. Powers made the brilliant observation that people behave to deliberately control many, but not all, of their own perceptions of the world. A person who acts on the world to control his or her own perceptions must affect parts of the world. From outside that person, we observers can see some of the environmental variables that the person controls.

From our vantage point outside the other person, we see events and relationships and processes that would ordinarily vary, but that are controlled by the person, which is to say the person keeps those events and relationships and processes at some predetermined state or condition.

A car hurtling along the road at a high speed would soon end up in a ditch, or against a tree, or crashing into another car, except for the actions of the driver.

The driver keeps the car moving toward the destination he selects, along the route he selects, at the speed he selects, in the lane he selects, at his selected distance behind a car ahead.

Think of all of the perceptions the driver controls while he is driving from one destination to another, and think of how very different those events in the world would be if the driver were not controlling those perceptions.

To explain how people control their perceptions, Powers developed "control system theory (CST)," which was the early name for what is now called perceptual control theory. The new name was adopted early in the 1990s, to distinguish Power's theory from the many fallacious ideas that some people had come to call "control theory." Powers said that people specify (select, determine, intend) part of what they perceive happening in the world.

They compare what they actually perceive against what they intend to perceive. If there is no discrepancy or difference (called "perceptual error") between actual and intended perceptions, the person does not act to change the world; but if there is a discrepancy (a "perceptual error), the person acts to eliminate the error.

People behave to eliminate, or prevent, differences between actual and intended perceptions. People behave to cancel out the effects of anything in the environment that disturbs the perceptions they are trying to control. (Much of Powers' earlier writing is available in two edited collections: William T. Powers (Ed.), Living Control Systems I ( Previously Published Papers), 1989; and William T. Powers (Ed.), Living Control Systems II (Previously Unpublished Papers), 1992. Both books are available from, Benchmark Publications, New Canaan, Connecticut.)

If you look back to the section of the present document where Ferster and Skinner described how to use deprivation in operant conditioning, you will see a perfect description of perceptual control.

Remember that they said the experimenter must (a) have a specific idea about how many pecks he or she wants to see a bird make in a session, (b) observe how many times the bird actually pecks, (c) "adjust" the amount of food the bird gets in its home cage until (d) he or she sees the bird pecking as often as is desired. Compare that description by Ferster and Skinner with the my paragraph immediately above this one.

That is how close Skinner came to understanding the universal phenomenon of perceptual control. That is how close Skinner came to greatness. Powers got it right.

Powers developed an elegantly simple mathematical model to explain the process of perceptual control. In 1973, more than 20 years after he began his work on PCT, Powers published a book, Behavior: The Control of Perception (known as BCP and published by Aldine, in Chicago), and a companion article, "Feedback: Beyond Behaviorism," in the journal, Science.

In 1973, I read BCP and the article in Science, and I became part of the small group of behavioral scientists who have worked to develop PCT, by behavioral research, and by computer modeling. PCT is a formal, quantitative theory of behavior. It is radically different from any widely accepted theory in the behavioral, social, or cognitive sciences, or in the life sciences.

At the core of PCT is a testable model of behavior, not just a system of ideas that Powers "believes." Any time we think there is a way to change the theory to make it better, we test the change, to see if they produce the specific results we thought they would. If the results are not what we expected, we must reject the proposed change, no matter how appealing it seemed -- no matter how much we might have liked it. Proposed changes to the basic PCT model are accepted only if they produce demonstrated improvements in the way the model works.

PCT Explains Control and Countercontrol:

In 1973, Bill Powers wrote the following passages in, Behavior: The Control of Perception. "I think it can be shown that the idea of controlling behavior -- one's own or that of other people -- stems from an old but incorrect concept of human nature, incorrect because it fails to recognize the control-system properties of human nature." (page 251)

One of Power's insights into behavior is that, "Organisms do not care about how they act as long as the actions do not disturb the perceptions they do care about." (page 264-5) Consequently, if one person (A) wants to control the actions of another person (B), "The best A can hope for is to get B to behave in ways that satisfy A's goals without preventing B from satisfying his own." (page 262)

The idea that one person can control the behavior of another reduces to something as simple as that -- something that can be illustrated with nothing more complicated than a pair of rubber bands, joined by a knot.

Imagine a room with a blackboard. On the blackboard, person A from Power's example above draws a dot. Person B grasps one end of the knotted rubber bands, and person A holds the other end. B tries to keep the knot in the rubber bands positioned over the dot on the board, while A moves her end of the bands slowly up and down and away from the dot.

For every move A makes, B must move his end of the bands exactly as far in the opposite direction, to cancel the effects of A's moves and keep the knot over the dot.

That is a simple example of control. B knows what he wants to see; A disturbs it; B does whatever he must, to cancel the effects of A on what he wants to see. What looks to all the world like a S-R "reflex" is in fact a clear example of perceptual control. That is true of all "reflexes."

Now imagine that A draws a second dot, one foot above the level of the first one, and close to where B is standing. The two people repeat the same simple demonstration described above, but half way through it, A moves her hand down and toward herself, watching B's hand while he moves it, as he must, to eliminate A's effect on the position of the knot relative to the dot. This time, because A pulls her end of the bands down and to the left, B must pull his end up and to the right, until suddenly A stops moving her hand, and B's hand comes to rest -- exactly over the second dot that A drew on the blackboard.

Because A knew what B wanted to perceive, it was trivially easy for her to disturb B's perception in such a way that his (B's) actions became what A wanted them to be.

That is an example of how one person can control the behavior of another, exactly as Powers described the phenomenon in 1973. Just as Powers said, A can control B's actions in that way only so long as she does not make it impossible for B to control the perception he wants -- the knot over the dot. If A makes that task impossible for B, then she will suddenly loose her ability to control his actions. In one sense, A's actions are already countercontrolled by B, in that A must moderate her own actions so as to avoid excessively disturbing B's control, and B alone determines what is excessive to him. But the rubber bands allow an even more dramatic demonstration of countercontrol.


A and B repeat the previous demonstration, with the second dot near to B.

This time, B knows that A "made" him put his hand over the second dot, while he kept the knot over the first dot.

This time, B decides to have fun at the expense of A. At first, B keeps the knot over the dot, as he did in the previous two demonstrations.

He even goes along when A starts her moves that require him to move his hand over the second dot.

Then, when his hand is above the second dot, B abandons his goal of keeping the knot over the original dot and watches A's hand; B has decided to make A move her hand wherever he (B) wants to see it on the blackboard -- he has decided to make her "climb the wall."

First, B moves his hand down slightly below the position where A wants him to keep it. Now A, who does not know the game has changed, must move her hand down, a move that would have disturbed the previous perception B controlled (knot on dot) so that B would move his hand up.

But now, B sees exactly what he wants to see -- A is moving her hand down the blackboard. Next, B moves his hand slightly above the second dot. In her attempt to "make" B move his hand back over the second dot, all that A knows to do is to move her hand far up on the blackboard, in an attempt to make B move his hand back down to the second dot.

Under the original rules of the first demonstration, B would move his hand down, to make the knot go back over the dot. But now, B is truly countercontrolling A; he is turning her attempt to control him back on her in a way she cannot avoid; for so long as A continues trying to control him, B is able to countercontrol her. What is more, the harder A tries to control B, the easier it is for B to countercontrol her -- the less he needs to move his hand away from the second dot, in order to see her move frantically up the board to get him back on that dot. Now, B has made A "climb the wall." In the classroom, teachers often discover that, the harder they try to control students' behavior, the easier it is for students to make them (teachers) "climb the wall." Often, all it takes is for a student to lift one corner of his or her mouth, ever so slightly, in "that kind of smile."

Of course, B cannot make the task impossible for A, or she will stop her actions, which is exactly the situation Powers described. To keep A, "in the game," B should occasionally allow A to position his (B's) hand over the second dot, then he can move it slightly in another direction and watch A move, frantically trying to lead him back to the mark. Now, each person controls the perception he or she wants, and each of them controls the other's actions.

The intricate dance of control and countercontrol is easy to explain as an example of perceptual control, where the participants' actions occur to eliminate the effects of disturbances that each person causes to the other person's perceptions. Skinner described the unavoidable phenomenon of countercontrol, but he could not explain it; PCT explains it, and PCT provides clues about how to avoid it. That is the subject of my next section.

What does this analysis of control and countercontrol mean for educators? Behaviorists, from Skinner to the present, have known that for a controller (educator) to control a controllee (student), the controller must deprive the controllee of something important, then use the important something as "reinforcement."

Behaviorists have also known for more than forty years that all controllees might feel distress over being controlled. Some controllees (students) might want to get away from their controllers (educators), and all controllees have the opportunity to countercontrol their controllers, either by making them uncomfortable, or by injuring them.

Perceptual control theorists have shown that control and countercontrol are special instances of one person controlling his or her own perceptions of what another person is doing. Once it starts, there is no logical end to the escalating dance of control and countercontrol, or to the violence it might breed.

How can educators avoid the negative features of control and countercontrol? The answer is simple: they can abandon their attempts to control students' behavior. Nothing else will work.

Do you mean I can't be nice to students anymore? Many educators who hear a discussion about control and countercontrol become dismayed and ask if they are supposed to stop doing nice things for students, or stop giving them nice things.

Their confusion is understandable. From the beginning of their training to be teachers, through however long they have been in the profession, educators have heard a constant story. They have been told that they are responsible for controlling students' behavior and that they should use reinforcers to accomplish that control.

The experts in behavioristic behavioral science, and their followers who sell discipline programs to schools, have told educators that "reinforcers" have the power to control behavior. Educators have been led down a torturous path, by advocates of cause-and-effect theories.

There is no need for educators to abandon saying nice things to students, or doing nice things for them. Teachers can do nice things any time they want. All they need to do to avoid the endless rounds of control and countercontrol is to expect nothing in return for the things they do. If teachers continue to act in certain (nice) ways, with the expectation that students will necessarily act a certain way in return, then teachers will forever remain susceptible to being disappointed by students, and their disappointment is the seed from which countercontrol can grow.

Students who perceive that they can disappoint educators often progress to the discovery that they can play off that disappointment to manipulate and control educators. Some students become experts at playing that game.

It is a game that might continue indefinitely, unless students become bored and move on to something new, or unless educators stop trying to control students. The instant a teacher gives up trying to control students' behavior, students are no longer able to disappoint the teacher, and the teacher cannot be controlled.

What is required is that the educator stop acting toward students with the expectation that they will react in a particular way.

Educators who do nice things with no expectation of any particular reaction from students cannot be disappointed when that reaction does not occur. Educators who act without expectations of specific reactions from students cannot be countercontrolled.

All a teacher needs to do to end the cycle of control and countercontrol is move away from discipline programs inspired by the lineal model of cause-and-effect.

Everything is a matter of the perceptions that a teacher intends to experience. It is all a matter of intent -- nothing more, nothing less. By way of his or her own actions, does a teacher intend to control students' actions? In response to his or her own actions, does the educator expect to see children behave in a certain way?

Does the educator expect students to express thanks, or to be grateful, or to show love in return? If the answer to any of those questions is "yes," then the relationship between educator and student can lead to the cycles of control and countercontrol. If the answer is "no," and the educator gives all of the good things with no expectation of anything in return, then the good things are a gift, pure and simple. When they are given in that spirit, good things and good acts become the means for celebration, not for control and countercontrol.

Must educators become saints, in order to avoid the dance of control and countercontrol? Not at all! What they need is a discipline program that teaches students how to control their own experiences without unnecessarily disturbing other people, and to resolve the conflicts that occur when people accidentally disturb one another. They need a process that allows educators to support students who work to develop their own solutions to problems. Ed Ford's "Responsible Thinking Process" (RTP) is that kind of program. I describe RTP briefly in my next section.

Part 5: Ed Ford's Responsible Thinking Process «

Ed Ford calls his discipline program the responsible thinking process and he describes it in two books, Discipline for Home and School, and Discipline for Home and School, Book 2: Program Standards for Schools. (Both books are available from Brandt Publishing, Scottsdale, Arizona.)

Ford developed RTP with the idea in mind that all of us behave to control some of our own perceptions, and that whenever several of us share an environment, sooner or later, our actions to control our own perceptions will disturb others while they try to control their perceptions. Sometimes disturbances like those are unintended, other times they are deliberate. Whether accidental or deliberate, when we disturb others we often find ourselves in conflict. When disturbances and conflicts occur in schools, the offenders are often subjected to discipline.

I described Ed Ford's RTP at length in my companion document, "Perceptual Control Theory, Reality Therapy & the Responsible Thinking Process," which I mentioned earlier. When you read that document, or Ford's books, you will see that RTP is not like programs in which educators try to control the actions of students. In RTP, student's are responsible for the consequences of their actions, whether those consequences are intended, or unintended. Educators are not asked to do something that is impossible, or something that sets them up to be countercontrolled by their students.

Educators do not attempt to deprive students of things they need or want, in an attempt to control student's behavior with positive reinforcement or positive consequences. Educators do not bribe students with offers of candy and activities that students can obtain only by performing actions educators demand of them. When they follow the principles of RTP, teachers treat students with respect, rather than like inanimate objects that are controlled by their environments, therefore students no longer have any "power" to countercontrol teachers and to "manipulate the system."

A common result when a majority of the faculty in a school begins to successfully use RTP is that most people say the school is much calmer, and that students and educators alike say they are treated with respect. The school is light years away from the hopeless dance of control and countercontrol that Skinner described. That is exactly the dance students and educators were doing in many schools before they began to use RTP.


Biographical information

Tom Bourbon

I am a behavioral scientist who, since 1973, has helped to develop perceptual control theory (PCT) as a formal, quantitative, scientific theory of behavior. I have done that in collaboration with Bill Powers, who developed PCT in the 1950's, and with several of Bill's colleagues, especially Rick Marken.

In my work with PCT, I have conducted behavioral research and computer modeling. For 25 years, I taught in a university department of psychology; for 19 of those years, I taught PCT as the major topic in all of my courses.

For eleven years, some of them overlapping my years in the university, I was affiliated with neurosurgery departments in medical schools, where I conducted research on relationships between brain functions and behavior in humans. One area of psychology that I studied as a student, and taught as a professor, was the history of psychology.

Consequently, I am familiar with B. F. Skinner's operant conditioning theory. I am also familiar with the ways Skinner's reinforcement theory relates to psychology in general, and to PCT in particular.





WARNING: Some are teaching RTP but are neither accredited or qualified.

Both in the U.S. and in other countries, there are some educators teaching RTP
and some schools claiming to use RTP, that are not accredited by RTP, Inc.

Also, if a person were to give a presentation on RTP without permission,
they would be in violation of the Lanham Act.