Chapter Six

Aversive Counterconditioning


In the last chapter on desensitization it was seen how counterconditioning can be used to reduce aversive-avoidance reactions to situations. In this chapter is the opposite, use of counterconditioning to reduce unwanted positive-approach reactions. This is the procedure of aversive counterconditioning, the countercond itioning of positive reactions using the response to an aversive (unpleasant) situation as the incompatible response (Feldman, 1 966; HaIlam & Rachman, 1976; Rachman & Teasdale, 1969). For example, a person addicted to some drug (e.g., heroin, alcohol, tobacco) has positive associations to many aspects of taking the drug, including such things as pleasant associations to a particular bar and drinking friends, a calming effect associated with lighting up a cigarette, a reduction of withdrawal symptoms after taking more heroin, or socially approved relaxing of inhibitions associated with drinking alcohol. These types of positive associations continually make it more probable the person will again use the drug, thus strengthening the addiction even though the long range effects of Using the drug are undesirable and even aversive. The logic of aversive counterconditioning is to pair situations that elicit the undesired positive response (e.g., the handling and taste of a cigarette) with stimuli that elicit a dominant, incompatible, aversive response (e.g., the reaction to electric Shock) as a way of reducing the positive reaction.


Aversive counterconditioning follows the general approach of counter- conditioning described in Chapter 3: First, we identify those stimuli that elicit the undesired response, in this case a pleasant-approach response. Next, we determine or establish ways to produce an incompatible response, in this case an aversion response. Then we gradually apply our counterconditioning, using a hierarchy to control response dominance. In aversive counterconditioning, a hierarchy is often unnecessary because the aversion response may always be dominant to the undesired response. However, this is not always the case and practitioners sometimes overlook the importance of a hierarchy in some aversive counterconditioning. Finally, remember that in counterconditioning we can stop at any point along the way between the two incompatible responses. Thus in aversive counterconditioning we can stop while the client still has a positive reaction to the stimuli, continue on and stop when the client feels neutral to the stimuli, or continue further until the client feels aversion to the stimuli.


Aversive counterconditioning is primarily used with self-rewarding behaviors. The person smoking an undesired amount of marijuana is reinforced (rewarded) for smoking by the results of smoking. The person who is sexually aroused by specific stimuli (e.g., people of the same sex, young children, certain types of clothes) is reinforced by the sexual arousal and resulting sexual fantasies and behaviors. This self-reinforcing aspect of the behaviors makes them difficult to treat by most counseling-therapeutic approaches. Behavioral treatment involves aversive counterconditioning to reduce some of the positive associations resulting from the natural source of reinforcement, as well as helping the client develop alternative reinforcing behaviors. For example, aversive counterconditioning may involve electric shock paired with photos of young children that elicit undesired sexual arousal.


This is a good place to see the importance of temporal contingencies on behavior. Consider an alcoholic who often gets drunk, feels happy, has a good time with his drinking friends, and withdraws from his daily problems. As a result of this, he later wakes up with hangovers, is slowly losing his job, and is involved in marriage problems, all related to his drinking. Now why do these bad effects from drinking not have a greater influence on his drinking? Part of the reason is that they are too distant in time to the actual drinking. The positive associations, which occur close in time to the drinking, have a stronger effect on the drinking behavior than the more distant negative associations In fact the negative associations may act as stimuli for more drinking. Thus a purpose of aversive counterconditioning is to bring to bear immediate aversive associations to stimuli associated with the undesired behavior.


Behavior modifiers frequently use a combination of procedures to increase treatment effectiveness. This is particularly true for the types of behaviors treated with aversive counterconditioning. It would be unusual if aversive counterconditioning were all that was necessary. Usually treatment would involve other components such as desensitization, training in social skills, or vocational training. For example, it would be a disservice to a homosexual to merely help him reduce his sexual arousal to people of the same sex. In addition, it may be desirable to help him develop his heterosexual arousal and social skills, as well as deal with many problems associated with the change in life-style.


Because aversive counterconditioning is often unpleasant, it is generally restricted to behaviors that are difficult to treat by other means and to situations in which the advantage of cure more than offsets any disadvantages of procedure. In reality, many clients have reported they found aversive counterconditioning less unpleasant than interpersonal probings, interpretations, and evaluations they experienced in some other forms of therapy. However, the use of aversive events certainly raises ethical questions and creates p roblems such as some clients disliking treatment, some clients becoming more aggressive or more anxious, and the fact that some procedures cannot be used with some clients (such as electric shock treatments with some cardiac patients).




In most of the current literature, the material covered in this chapter is subsumed under the term aversion therapy, which includes two distinct, and often confused, models: aversive counterconditioning and operant punishment. Aversive counterconditioning, discussed in this chapter, is based on respondent conditioning, the systematic pairing of two sets of stimuli, one which elicits the undesired response and one which elicits the aversion response. The stimuli are paired independent of what responses the client makes. The client’s responses are only a measure of the progress of the conditioning. Operant punishment, discussed in the next chapter, is based on operant conditioning. Here the aversive stimulus is presented contingent upon a particular response of the client and usually occurs if and only if the client makes the response. Consider the use of an unpleasant odor in the treatment of over-eating. In aversive counterconditioning, the odor would be paired with cues that tend to elicit or encourage over-eating, as a way of weakening the support for over-eating. In operant punishment, the odor Would be paired with the response of over-eating as a way to suppress the act of over-eating.


Keeping in mind the distinction between the two approaches is important for a number of reasons: The aversive stimulus is contingent on stimuli in One case and responses in the other. This leads to significant differences in the occurrence and timing of the aversive stimulus. The types of results and changes in behavior that follow from the two procedures are different. And the literatures on respondent and operant conditioning suggest different optimal ways of conditioning.


In practice or in reading many of the reports of aversion therapy, it is often difficult to separate aversive counterconditioning from operant punishment, with many situations having components of both. In reality, it is impossible to have one without the other. For whenever the aversive stimulus occurs the person is doing something (doing “nothing” is doing something; a live person is always behaving) and hence is punished for what he is doing. And whenever the aversive stimulus occurs, it will be associated with whatever stimuli are present at that time. The key then becomes the contingencies the practitioner emphasizes in his treatment program.




Examples of aversive counterconditioning are found in a variety of settings. A common way to break a dog of killing chickens is to hang a dead chicken around the dog’s neck for a while. Pliny the Elder suggested part of treatment for alcoholism might involve putting spiders in the bottom of the person’s glass. And aversive associations are often established at the verbal level as when a person says “Bob is a gossip.”


Most of the early clinical reports and experimental studies of aversive counterconditioning used drugs or electric shock to produce aversion. Much of the early treatment of alcoholism used chemically induced nausea (Lemere & Voegtlin, 1950). Treatment may consist of pairing the taste of the person’s favorite alcohol with a drug that makes him nauseous. A common practice today is to keep an alcoholic on the drug antabuse (disulfiram), which, combined with even a small amount of alcohol, causes nausea and vomiting that may last a couple of hours. However, the use of antabuse is seldom sufficient by itself and generally needs to be part of a more comprehensive treatment program.


Raymond (1964) reported several cases of aversive counterconditioning using apomorphine-induced nausea. One case was a 63-year-old alcoholic man who was given two treatments a day for 31 days. This was sufficient to produce abstinence from alcohol for the three years of follow-up. A second case involved a 14-year-old boy addicted to cigarettes. After three treatments he stopped smoking and even the smoke from his father’s cigarette made him feel ill. Thompson and Rathod (1968) employed drugs that would partially stop respiration for a short time. They were able to reduce the use of heroin by associating this aversion with the taking of heroin.


Many practitioners and researchers came to prefer electric shock to drugs as the aversive stimulus. One reason is that with electric shock it is easier to control many of the variables which are important in conditioning, such as onset, offset, duration, and intensity of the aversive event. Fewer trials per day can generally be given with drugs than shock. Drugs, such as those that produce vomiting, are often more unpleasant to the client and staff than electric shock. Greater individual differences and side effects occur with drugs. And some of the drugs depress the central nervous system and hence retard learning and treatment effectiveness.


Lubetkin and Fishman (1974) used shock in the treatment of a 2 3-year- old heroin addict. The client imagined and described behavior sequences leading up to and including heroin intake, while receiving electric shocks along the way. Shifting to imagining a scene of a drug-free situation was paired with the offset of the shock. Following treatment the client remained drug free during the eight-month follow-up.


It is also possible to use portable shock units, which then permit in vivo aversive counterconditioning (McGuire & Vallance, 1964). This permits counterconditioning to important environmental stimuli, allows the client to carry-out much of the treatment himself, and makes it possible to gear part of the treatment to internal cues, such as a craving for a drug, which may be difficult to reproduce in the clinic.


Aversive counterconditioning has become broader and more flexible as a wider range of aversive stimuli have been employed. Other sources of aversion include cigarette smoke, unpleasant odors, aversive pictures or sounds, social and personal criticism, and aversive imagined scenes.

Lichtenstein and his associates (1973) reduced smoking by having a machine blow warm cigarette smoke in their clients’ faces while they were smoking. This produced a significant reduction at the six-month follow-up with about 60 percent of the clients totally abstinent.


Morganstern (1974) took advantage of the fact that smoking a cigarette is aversive to many non-smokers. His client was a 24-year-old obese female who ate much in addition to her regular meals (pizza and ice cream one to three times per day, almost 200 pieces of candy and dozens of cookies and doughnuts per week). Treatment, in the clinic and self-treatment at home, consisted of pairing eating junk food (e.g., candy) with taking a drag of a cigarette and then spitting out the food and exclaiming, “Eating this junk makes me sick.” Through this the client was able to loose 53 pounds.






The onset of an aversive stimulus is unpleasant by definition. It is the onset that is paired with other stimuli in aversive counterconditioning and with a response in punishment. The offset of the aversive stimulus is pleasant by contrast. This offset can be associated with stimuli (called relief stimuli) to which we wish to respondently condition a positive effect. Or the offset can be used to reward a desirable behavior. In treating alcoholism, the drinking of alcohol can be paired with shock followed by drinking of orange juice during the relief time (McBrearty et al., 1968).


Feldman (Feldman, 1966; Feldman & MacCulloch, 1965) argues that the offset of the aversive stimulus should be used to reward a desired behavior such as avoidance of the inappropriate object and/or approach to the appropriate object. The main reason is that the learning under the operant model (particularly avoidance responses) generally takes longer to extinguish than learning under the respondent model.


Although the various suggestions for use of the offset of the aversive stimulus are reasonable, to date there is a lack of research systematically evaluating the effect of this variable. There is also a need to separate respondent from operant variables. For example, Rachman and Teasdale (1969, p. 137) argue that the effective process in Feldman’s procedure is not the development of an avoidance response, but rather the respondent conditioning of anxiety.




An interesting question often raised is why should the effects of aversive counterconditioning generalize to situations in which clients know they will not encounter the aversive stimulus. If people are undergoing aversive counterconditioning in a clinic for alcoholism, why should this affect them when they are at the corner pub away from the clinic and practitioner? The facts are that aversive counterconditioning does generalize, in varying degrees, to non-clinic situations. One possibility is that the conditioning takes place at a level outside of cognitive control (see Chapter 9). That is, as a result of conditioning, alcohol-related cues elicit specific responses in or out of the clinic. Although people are aware they are not in the clinic, this awareness has a minimal effect on their conditioned reactions. In this sense, it is like a phobia in which specific situations trigger anxiety even though the person knows the anxiety is unreasonable and undesired. Bandura (1969, chap. 8), on the other hand, accounts for the generalization in terms of self-control. He suggests that as a result of aversive counterconditioning clients can stop the unwanted behavior in vivo by reinstating in imagination the aversive reactions they experienced in treatment.


But whatever the explanation, several things can be done to facilitate generalization: The clinic can be arranged to match the real world as much as possible. Many behavioral treatments of alcoholism use simulated bars set up in the clinic. Any possible in vivo conditioning should be emphasized. Part of the treatment should help establish self-control skills, such as mentioned above relative to Bandura and discussed below relative to covert sensitization. Finally, it is often desirable to have the client return periodically to the clinic for booster sessions, additional conditioning sessions to minimize loss of treatment effects.




Several problems exist in developing a theoretical account of the effects of aversive counterconditioning (Hallam & Rachman, 1972, 1976). One problem, discussed earlier, is that aversive counterconditioning and operant punishment are often confused and confounded in practice and in explanations. A second problem is that most of the literature consists of case studies and treatment programs in which aversive counterconditioning is just one component. This is useful for developing effective treatment approaches, but more research is necessary for factoring out the relative importance of different components. There has also been little parametric research, systematically varying different variables such as the onset and intensity of the aversive stimulus. Of the many applied counterconditioning approaches, aversive counterconditioning lends itself better than most for such parametric studies, which would help factor out such things as the operant and respondent components. There is also a need for research related to many other variables, including the client’s motivation to change and the expectancies the client has for the effects of the treatment. Given these qualifications let us consider four basic theories, counterconditioning, extinction, state theory, and cognitive theory.




According to the counterconditioning explanation (Chapter 3), aversion reactions gradually become respondently conditioned to stimuli that previously elicited the positive or approach responses. Thus through aversive counterconditioning the taste and smell of a cigarette, which were pleasing and reinforcing, are gradually neutralized or made aversive. There are currently two problems with this theory. First is that laboratory studies on conditioning emotional reactions in humans suggest that these conditioned reactions extinguish quickly. This could mean that the counterconditioning in aversive counterconditioning also extinguishes quickly and thus long-term treatment effectiveness may depend on other treatment components. Or it may be that, for reasons yet to be determined, the aversive counterconditioning in clinical situations is longer lasting than the quite different laboratory studies.


The second problem with the counterconditioning explanation is to account for the generality of results that often follows fairly specific conditioning. For example, treatment of alcoholism may involve aversive counterconditioning in the clinic to a specific set of stimuli related to specific drinks and the cues of the simulated bar. It would not be unusual if the treatment effects carried over to a wide range of different drinks and situations. The question is how to account for such a generality of results. A simple counter- conditioning explanation would account for these results in terms of generalization, the carry-over of the conditioning from specific stimuli to similar stimuli. However, what constitutes “similar” probably depends on cognitive processes of the client, rather than simple physical properties of the stimuli. Critics of the counterconditioning theory say the theory cannot adequately account for the generality of the results.




As discussed earlier (Chapter 3), anything that can be interpreted as counter- conditioning can also be interpreted as extinction, although I am not familiar with anyone yet advocating an extinction theory for what is covered in this chapter. According to an extinction theory, stimuli are presented during treatment without letting them be paired with the pleasant result, thus producing extinction. For example, a person might taste a martini, but spit it out before getting effects from the alcohol. Such a hypothesis would need to account for the effect of the aversive event. One possibility is that the aversive event prevents the usual pleasant reaction from occurring and hence facilitates extinction.


State theory


Hallam and Rachman (1976) have proposed a tentative “state theory,” which is based on a change in general responsiveness (sensitization), rather than changes in specific stimulus-response associations. According to this theory, aversive counterconditioning has two effects: (1) a general sensitization, a tendency simply to be more reactive to the types of stimuli encountered in treatment; and (2) a suppression of the undesired behavior, with the suppression diminishing over time. The success of the treatment depends on how long the suppression lasts, what alternative reinforcing behaviors are developed during suppression, and the amount of reinforcement from the success in suppressing the undesired behavior.


Cognitive theory


According to cognitive theories of aversive counterconditioning (e.g., Bandura, 1969), the effects of treatment are based on such elements as expectancy of treatment results and the learning of self-control. Cognitive theorists essentially suggest learning is more cognitive or central than is their conceptualization of conditioning theories such as counterconditioning. The generality of results following specific treatment is often offered as support of this position. Another possible cognitive aspect of aversive counterconditioning is based on cognitive dissonance. The argument is that because the treatment is aversive clients will change their behavior following treatment to “justify” (avoid dissonance) going through such aversion (Carlin & Armstrong, 1968). A critical test for many cognitive theories would be to see if aversive counterconditioning generalizes to situations where the client knows he won’t encounter the aversive stimulus and is motivated away from such generalization.


Because of the many confounding variables discussed at the beginning of this section, it is not possible to currently separate the various effects. Probably different treatment programs are based on and emphasize different combinations of the various factors suggested by the different theories.




A variation of aversive counterconditioning is covert sensitization (also called verbal aversion and aversive imagery) in which the stimuli to be counterconditioned and the aversive event are imagined scenes (Cautela, 1966b, 1967, 1970c; Cautela & Wisocki, 1971). Here a person imagines a situation in which the undesired behavior would occur, imagines beginning to do the undesired behavior or intending to do it, and then imagines a scene that is aversive, such as vomiting, falling into a cesspool, or social criticism. For example, Polakow (1 975) used covert sensitization as part of a program in

treating a 24-year-old, female barbituate addict. Imagined scenes of thinking about barbituates, making contact with a dealer, and ingesting pills were associated with images of being attacked by hordes of large sewer rats (Which she had indicated on the Fear Survey Schedule as being a strong Source of fear).


Although little research has been done on optimizing covert sensitization, Cautela, who has done much of the work in this area, suggests several guidelines for effective treatment: The emphasis of the treatment should be on the intent to do the behavior, rather than the behavior itself. This catches the problem earlier in the behavior chain, and Cautela believes it minimizes overgeneralization. That is, the covert sensitization may involve conditioning related to the client’s intent on drinking alcohol, rather than anyone drinking alcohol or alcohol in general. Cautela also suggests alternating the aversive scenes with scenes in which the client performs an alternative desired behavior and experiences relief. The imagery may be enhanced through the various senses, as by having the client smell or hear something related to what he is imagining. Finally, as is true of most aversive counterconditioning, covert sensitization is often most effective when combined with other procedures such as relaxation training. The following is part of Cautela’s (1970c) treatment of smoking for a client:


You are sitting at your desk in the office preparing your lectures for class. There is a pack of cigarettes to your right. While you are writing, you put down your pencil and start to reach for a cigarette. As soon as you start reaching for the cigarette, you get a nauseous feeling in your stomach. You begin to feel sick to your stomach, like you are about to vomit. You touch the package and bitter spit comes into your mouth. When you take the cigarette out of the pack, some pieces of food come into your throat. Now you feel sick and have stomach cramps. As you are about to put the cigarette in your mouth, you puke all over the cigarette, all over your hand, and all over the package of cigarettes. The cigarette in your hand is very soggy and full of green vomit. There is a stink coming from the vomit. Snots are coming from your nose. Your hands feel all slimy and full of vomit. The whole desk is a mess. Your clothes are full of puke. You get up from the desk and turn away from the vomit and cigarettes. You immediately begin to feel better being away from the cigarettes. You go to the bathroom and wash up and feel great being away from the cigarettes.


There are many advantages to covert sensitization, in general and in comparison to other forms of aversive counterconditioning. One advantage is that it requires no apparatus or drugs. A second advantage is that it can be made very specific, such as gearing it toward eating the second piece of pie or drinking the second manhattan when the treatment is geared toward reducing over-eating or excessive drinking. Perhaps most important is that the client can utilize the procedure on himself, permitting in vivo counter- conditioning and the development of a powerful self-control skill.


Consider a female college student who has trouble with over-eating. The cafeteria, snack bar, vending machines, and food in her room all provide cues that cause her to eat excessively and unwisely. Part of her treatment may involve covert sensitization, perhaps with aversive scenes such as seeing her boyfriend with his arm around another girl, laughing at how fat the client is. Now after a few treatment sessions, the client may use these scenes in a self-control fashion. When a candy bar in a vending machine calls out to her, she can switch to an imagined scene to stop her desire for, or intent to get, the candy bar. With practice, she may find that eventually she does not need to call up the whole aversive scene, but can do it indirectly through a subjective feeling of “willing” not to buy or eat something. With more time, she may find that her willing becomes automatic and the whole process slides out of consciousness.


There are many problems, primarily lack of research, in experimentally evaluating covert sensitization and deciding among various theoretical explanations (see Mahoney, 1974a, p. 93—103). Most of the literature is case studies reporting varying degrees of success. There are few systematic controlled studies; and covert sensitization is often used with, and hence confounded with, other procedures. The fact that the whole treatment takes place in the client’s imagination makes evaluation of exactly what is going on very difficult. All the problems and theories related to aversive counter- conditioning, discussed earlier, apply here. In this context, I have described covert sensitization as an example of aversive counterconditioning, while Cautela sees it as an example of operant punishment. In practice it is usually possible to find elements of both of these. The issue is which you emphasize.


The following is a sample of some of the mixed reports on covert sensitization: In cases of overeating covert sensitization is often part of a successful treatment program (e.g., Cautela, 1966b). But some controlled studies have questioned its usefulness (e.g., Diament & Wilson, 1975). Foreyt and Hagen (1973) compared covert sensitization and a placebo control (imagining a pleasant scene rather than an aversive one) for weight reduction. Both groups showed a significant decrease in their perceived palatability of the foods imagined during treatment, but no significant weight loss. The results were explained in terms of factors such as suggestion and attention. Janda and Rimm (1972) found covert sensitization effective in reducing eating and weight, but the results at the end of treatment were only significant for those subjects reporting the highest degree of arousal when presented with the aversive scenes. At a six-week follow-up the changes for the entire group of covert sensitization subjects were significantly greater than for the controls. Barrett and Sachs (1974) studied the effects on smoking of covert sensitization. They compared four groups, a forward group (smoking scene, then aversive scene), a backward group (aversive scene, then smoking scene), a backward interval group (backward plus 60 seconds between scenes), and a non-associative group (aversive scene only). They found all treatments to be equally effective. They suggested their results were best explained in terms of such variables as motivation or cognitive changes. In evaluating all such research studies, it is important to keep in mind that covert sensitization is Probably most effective when coupled with other treatment procedures, particularly when dealing with complex self-reinforcing behaviors such as Over-eating, smoking, and excessive drinking.


Overall, aversive counterconditioning appears to be a potentially useful Component in the treatment programs for some difficult behaviors. Evaluation, explanation, and improvement of aversive counterconditioning awaits further research.




Procedurally, aversive counterconditioning is the counterconditioning of situations that elicit positive-approach behavior using an aversive event, such as unpleasant reactions to certain drugs, electric shock, pictures, odors, sounds, and responses of other people. Covert sensitization is aversive counterconditioning in which the stimuli to be conditioned and the aversive event are imagined scenes. In aversive counterconditioning, the aversive event is primarily respondently associated with stimuli that tend to elicit aspects of the undesired behaviors; while in operant punishment, the aversive event is primarily associated with the occurrence of the undesired behaviors and is usually used to suppress these behaviors. The offset of the aversive event may be used to respondently condition positive effect to some stimuli and/or operantly reinforce some desired behavior. Theories of aversive counterconditioning include counterconditioning, extinction, state theory, and cognitive theory. Although considerable more research is needed, aversive counterconditioning appears to be an effective component of many change programs, particularly when coupled with other approaches. This is especially true in dealing with self-reinforcing behaviors that are difficult to reduce by other means.




Give three examples of aversive counterconditioning occurring naturally, that is, without someone specifically setting up the contingencies.
Give two different hypothetical clinical examples of aversive counterconditioning using aversive events other than those mentioned in the chapter.
Outline an aversive counterconditioning program for smokers who come to your clinic. How would you maximize generalization to situations outside the clinic?
Outline an aversive counterconditioning program for a hypothetical case of heroin addiction, using hierarchies, in vivo conditioning, and respondent use of the offset of the aversive event.
Outline a covert sensitization program for a hypothetical problem drinker. What other approaches (e.g., desensitization) may be part of your overall program? Why?
Using a table or diagram, show the relationships among respondent conditioning, counterconditioning, desensitization, and aversive counterconditioning.
What are the implications of the fact that whenever you have aversive counterconditioning you also have operant punishment? Give an example in which this would be a serious problem.
Are there any problem behaviors that are not self-reinforcing which you would change by aversive counterconditioning? Explain your answer.
How important in aversive counterconditioning is the client’s motivation to change? How would a counterconditioning theorist answer this question differently from a cognitive theorist?
There have been several times in which a sex offender is offered the choice of going to jail or going through aversive counterconditioning. Discuss the ethical and practical issues involved.
Describe an experiment that would distinguish, at least for one situation, between a counterconditioning and state theory explanation of aversive counterconditioning.
Construct a theory of aversive counterconditioning that incorporates counterconditioning, sensitization, suppression, and self-control.





Davison II, W. S. Studies of aversive conditioning for alcoholics: A critical review of theory and research methodology. Psychological Bulletin, 1 974, 8 1, 571—581.

Hallam, R. S. & Rachman, S. Current status of aversion therapy. In Hersen, M., Eisler, R. M., & Miller, P. M. (eds.), Progress in behavior modification. Vol. 2. New York: Academic Press, 1976.

Rachman, S. & Teasdale, J. Aversion therapy and behaviour disorders: An analysis. Coral Gables, Fla.: Univ. Miami Press, 1969.