Journal of Consulting and Clinical Psychology 1978, Vol. 46, No. 5, 1016-1022

Avoidance-Approach:

The Fifth Basic Conflict

Seymour Epstein University of Massachusetts at Amherst The basic conflicts are almost always listed as approach-approach, approachavoidance, avoidance-avoidance, and double approach-avoidance conflict. The possibility of avoidance-approach conflict, in which a steeper gradient of approach intersects a gradient of avoidance, is ignored because it is assumed that approach gradients cannot be steeper than avoidance gradients, and that even if avoidance-approach conflict could exist, it would be of no interest, as the individual would simply slay away from the conflicting goal. It is demonstrated that there is no good reason for assuming that approach gradients cannot be steeper than avoidance gradients, and there is considerable evidence that they often are. It is further noted that individuals can be placed in situations not of their choosing. If an individual with an avoidance-approach conflict were placed on the goal side of the intersection of the gradients, the person would enthusiastically approach a goal that had previously been avoided. Thus, avoidanceapproach conflict can account for ego-alien behavior, such as when a shy, sexavoidant "model" boy commits a violent crime of passion. The implications and causes of avoidance-approach conflict in everyday life are discussed. In almost all textbooks of introductory and abnormal psychology, four basic conflicts are listed with accompanying diagrams. These are approach-approach, avoidance-avoidance, approach-avoidance, and double approachavoidance conflict. Particular attention is then given to approach-avoidance conflict, because it can account for a variety of clinical phenomena, such as that an individual in such a conflict tends to become entrapped at a midpoint from a goal and to suffer, as a consequence, from a state of heightened drive. The entrapment is explained by the assumption that the gradient of avoidance is steeper than the gradient of approach. No consideration is given to the possibility of a conflict, which I shall refer to as "avoidance-approach" conflict, in which a steeper gradient of approach intersects a gradient of avoidTbis article was supported by National Institute of Mental Health Research Grant MH-01293. I wish to express my appreciation to James Avcrill for his helpful criticism of an earlier version of this article. Requests for reprints should be sent to Seymour Epstein, Department of Psychology, University of Massachusetts, Amherst, Massachusetts 01003.

ance. There are probably two reasons why avoidance-approach conflict has been ignored. One if that it is assumed that the gradient of avoidance is necessarily steeper than the gradient of approach. The other is that it is assumed that even if a case could be made for such conflict on theoretical grounds, it would be of no practical interest, as an individual with a flatter gradient of avoidance than of approach would simply stay away from the conflicted goal. It will be demonstrated in this article that approach gradients can be steeper than avoidance gradients, and that there are important clinical phenomena, such as ego-alien behavior, that can be accounted for once this is recognized. Lewin (1935), who originally formulated the concept of approach-avoidance conflict, simply stated without evidence that "the negative vector usually increases gradually in strength and finally becomes stronger than the positive" (p. 90). It is not clear whether he meant the word usually to indicate that given a negative incentive, there is usually a goal gradient, which would appear to be selfevident, or whether it was to indicate that the negative vector usually overtakes the posi-

Copyright 1978 by the American Psychological Association, Inc. 0022-006X/78/4605-1016$OO.V5

1016

AVOIDANCE-APPROACH CONFLICT

tive. In the latter case, it would indicate that Lewin did not believe that the positive vector is always steeper than the negative one, but when it is, some interesting consequences follow. Miller (1944) also initially assumed without explanation that avoidance gradients are steeper than approach gradients. More recently, Miller (19S9) revised his position and noted that there is no intrinsic reason for avoidance gradients to be steeper than approach gradients, and that whether they are steeper or less steep depends on the extent to which the gradients are based on inner relative to outer cues. He observed that in many of the studies in which steeper gradients of avoidance were found, the avoidance gradients were based on the delivery of punishment, an external source of stimulation, whereas the approach gradients were based on hunger, an internal source of stimulation. It appears that there is no logical basis for assuming that approach gradients cannot be steeper than avoidance gradients. As for empirical evidence on steepness of approach and avoidance gradients, it has been demonstrated that a number of factors (cf. review in Heilizer, 1977), such as stimulus differentiation (e.g., Bugelski & Woodward, 1951; Hearst, 1962; Saltz, Whitman, & Paul, 1963), number of trials (e.g., Elder, Kuehne, Clarke, & Larre, 1970; Elder, Kuehne, & Moriarty, 1970; Schroeder & Gerjuoy, 1965; Weiss, 1960), runway length (e.g., Clifford, 1973), and mental age (Tempone, 1965), are directly related to steepness of gradients. Among these, stimulus differentiation appears to be the most important factor, as it can account for the others. Thus, the finding that runway length is an important factor can be accounted for by the consideration that the greater the runway length, the more easily the end and the beginning of the runway can be differentiated (Saltz et al., 1963). The influence of number of reinforced trials can be attributed to an increase in stimulus discrimination that occurs over trials, and the influence of mental age can be attributed to the more accurate discrimination of children with higher mental age. Stimulus differentiation can also account for Miller's observation that avoidance gradients have

1017

been found to be steeper than approach gradients when avoidance is contingent on external cues and approach on internal cues, as internal cues may not be as easily discriminated as external cues. It follows from the above that depending on such factors as training in stimulus discrimination, number of reinforced trials, and degree to which motives are based on external relative to internal stimuli, approach gradients can be less steep, more steep, or no different in steepness from avoidance gradients. This conclusion is supported by a number of studies that have directly compared the steepness of approach and avoidance gradients. Some have reported avoidance gradients to be steeper than approach gradients (e.g., Brown, 1948; Miller & Kraeling, 1952; Miller & Murray, 1952; Murray & Berkun, 1955), some have reported approach gradients to be steeper than avoidance gradients (e.g., Hearst, 1960, 19,62; Smith, 1965, 1969), and some have reported no difference in steepness of approach and avoidance gradients (e.g., Desiderato, Foldes, & Gockley, 1966; Gjesme, 1974; Hearst, 1960; Rigby, 1954). Following an analysis of the relative steepness of approach and avoidance gradients, Maher (1966) noted that there is no firm support for the assumption that avoidance gradients tend to be steeper than approach gradients. He observed that the classic studies of Brown (1948), widely cited as evidence for the assumption, have serious methodological flaws. In a study by Maher and Nuttall (in Maher, 1966), order of testing was found to influence the relative steepness of gradients, and Maher concluded that Brown's finding of a steeper gradient of avoidance than of approach could be attributed to fatigue effects associated with order of testing. Even in studies that report no significant difference in steepness of avoidance and approach gradients, it should be considered that some subjects produced steeper gradients of approach and others of avoidance. That is, there are individual differences with respect to relative steepness of approach and avoidance gradients. In a study by Gewirtz (1959) that directly dealt with this issue, reliable individual differences were found in the relative steepness of approach and avoidance gradients. In a

SEYMOUR EPSTEIN

1018

\

APPROACH STRONGER THAN AVOIDANCE

AVOIDANCE STRONGER THAN APPROACH

\ \ \

EQUILIBRIUM POINT

-APPROACH \ ""•-•GRADIENT \

AVOIDANCE GRADIENT!

FEARED GOAL

NEAR

DISTANCE

FAR

Figure, 1. Approach-avoidance conflict. (The individual approaches the goal when at a distance, vascillates at the intersection of the gradients, and avoids the goal when closer to it. The result is that the individual becomes trapped by his or her own molives, neither being able to fulfill nor abandon the approach motive.)

recent study in our own laboratory (Losco & Epstein, 1977), relative steepness of avoidance and approach gradients was investigated while holding magnitude and quality of incentive constant. Although a weak, significant tendency was found for the avoidance gradient to be steeper than the approach gradient, inspection of the results for individuals revealed that the phenomenon was far from uniform among subjects, with some producing much steeper approach than avoidance gradients. It was further noted that the mean tendency was not very robust and was readily canceled out by incidental factors, such as motor exertion. Considering all the evidence together, it can safely be concluded that it is possible for approach gradients to be steeper than avoidance gradients. As already noted, Miller has come to the same conclusion. He has indicated that the only reason that he did not pursue its implications is that he found a steeper gradient of approach than of avoidance to be "less perspicuous than the other pattern" (Miller, 19S9, p. 2 2 2 ) . A pattern in which a steeper gradient of approach intersects a gradient of avoidance not only has significant implications in its own

right, but it is important to recognize it in order to avoid confusion between approachavoidance and avoidance-approach conflict, it is instructive, in this respect, to contrast the major features of the two types of conflict. There are three major behavioral consequences for an individual with an approachavoidance conflict. A person with such a conflict approaches the conflicted goal at a distance, is blocked and vascillates at an intermediate point at which the gradients intersect, and retreats when closer to the goal (see Figure 1). According to Miller, the most efficacious treatment for approach-avoidance conflict is to reduce the avoidance drive, as attempting to force the individual to reach the goal can induce unmanageable levels of anxiety, and would very likely fail, since avoidance tendencies might well mount more rapidly than the combined approach tendencies. Avoidance-approach conflict also has three behavioral consequences. The individual in an avoidance-approach conflict avoids the goal when on the far side of the intersection of the gradients, is hesitant and blocked, at least momentarily, when at the point of intersection of the gradients, and exhibits an accelerated approach response if placed on the goal side of the intersection point (see Figure 2 ) . Thus, in both approach-avoidance and avoidance-approach conflict, there are manifestations of strong avoidance reactions and of blocking and hesitation, although in the case of avoidance-approach conflict the blocking is of short duration, as movement in either direction breaks the deadlock. Considering that what is near and far from a goal and what is brief and enduring blocking are matters of judgment, it is quite possible to confuse approach-avoidance conflict with avoidance-approach conflict, unless sufficient information is obtained by observing behavior at different points along time, distance, or cue dimensions. Perhaps the most interesting difference between approach-avoidance and avoidanceapproach conflict is that only in the former case is the individual propelled into the conflict area by his or her own volition and remains trapped there. As a result, in the absence of outside pressure, the person experiences continuous tension. No such stable

AVOIDANCE-APPROACH CONFLICT

equilibrium exists in avoidance-approach conflict.1 If left to his or her own volition, the individual with an avoidance-approach conflict would simply remain out of the realm of influence of the conflict and would therefore not experience enduring stress. As already noted, this is probably one of the reasons why avoidance-approach conflict has not attracted any attention. Further consideration, however, reveals that avoidance-approach conflict is not as benign as the above analysis suggests. First, consider that the area of conflict to be avoided could be a highly significant one, such as the experience of close relationships with others. The individual with such a conflict could avoid tension only by imposing severe restrictions on his or her realm of experience. It might be argued that the limitations are imposed only by the avoidance gradient, and nothing is added by postulating an avoidance-approach conflict. After all, what function can the approach gradient play if an individual will not voluntarily expose himself or herself to its influence? The answer is that any extrinsic motive or circumstance beyond the individual's control could propel the individual to a point beyond the intersection of the two gradients. In such a circumstance, the person would exhibit behavior that is radically out of character with his or her normal behavior. The person would show an accelerating approach reaction to a goal that had been assiduously avoided up to then. If the approach motivation involved antisocial behavior, the behavior would appear as an ego-alien breakthrough of a destructive impulse. An example of such a case is the shy, inhibited "model" boy who, suddenly faced with a temptation that he had previously succeeded in avoiding, commits a bizarre crime of passion. On the other hand, if the approach motive were a constructive one, such as the expression of socially acceptable heterosexual feelings and had been avoided because of an excessively broad fear gradient, then an accidental encounter from which withdrawal was difficult could be therapeutic. Once the individual was on the near side of the goal, there would be no further conflict, and the individual would eagerly approach a goal that had previously been avoided. With

1019

APPROACH STRONGER THAN AVOIDANCE

AVOIDANCE STRONGER THAN APPROACH

-^-EQUILIBRIUM POINT

APPROACH GRADIENT-j_N

FEARED GOAL

NEAR

DISTANCE

Figure 2. Avoidance-approach conflict. (The individual avoids the goal when at a distance, is momentarily blocked at the point of intersection of the gradients, but, if placed by circumstances beyond his or her control at a closer point will rapidly and intensely approach the goal and exhibit ego-alien behavior.)

repetitive exposure to the same situation, and consequent goal attainment, the conflict would gradually be extinguished. Given the significant consequences of pathological levels of avoidance-approach conflict, it is important to consider the conditions in real life that could give rise to it. As noted earlier, according to Miller, a critical condition for determining the relative steepness of approach and avoidance gradients is the degree to which response tendencies are influenced by inner relative to external cues. When a high proportion of inner cues is involved in an avoidance gradient, the gradient will be relatively flat and conducive to the development of avoidance-approach conflict. In humans, inner cues frequently consist of thoughts and images. To the extent that an individual is trained to believe that the expression of an impulse in any form is bad, the individual will have a broad avoidance gradient. Now consider that the same individual's approach tendencies are minimally mediated by inner responses, so that if impulses to 1 This, of course, is also true of approach-approach conflict and of avoidance-avoidance conflict when there are no restraints against leaving the field.

1020

SEYMOUR EPSTEIN

approach arc aroused, it is apt to be because of the presence of external cues. Such an individual would have a steeper approach than avoidance gradient, because the avoidance tendencies are internalized to a greater extent than the approach tendencies, and this could even be the case if the approach tendencies were activated by a combination of physiological and external cues, as in the sex drive. Given sufficiently broad injunctions against the expression of an impulse in any circumstances, there is obviously no need to discriminate between the conditions in which the impulse should or should not be expressed. It will be recalled that poor stimulus discrimination is a major factor in producing broad generalization gradients. Thus, whether one wishes to analyze the comparative steepness of gradients by considering the relative contribution of inner to outer cues, or by considering the role of stimulus discrimination, one arrives at the same conclusion, namely, that strong categorical prohibitions against the expression of an impulse foster broad avoidance gradients and are conducive to the development of an avoidance-approach conflict. The relative strengths, or heights, of approach and avoidance gradients are obviously critical factors in determining whether an avoidance-approach conflict will occur. If the approach motive is weak enough compared to the avoidance motive, the avoidance motive will, of course, dominate it at all levels, and the individual will simply exhibit a generalized avoidance tendency to all motive-relevant stimuli. The opposite will be the case if the approach motive is sufficiently strong compared to the avoidance motive. Conflict can only exist within a restricted range of relative strengths of the gradients that allows them to intersect. Given such conditions, the stronger the approach and avoidance drives, the more extreme the conflict-related phenomena that will be exhibited, and thus the greater the potential for extreme ego-alien behavior. Tt follows that the conditions for producing pathological levels of avoidanceapproach conflict are those that foster the internalization of intense and broad prohibitions against inherently strong approach molives.

Given the above analysis of avoidanceapproach conflict, it follows that overcontrolled persons are more apt to exhibit bizarre ego-alien behavior than undercontrolled persons. This conclusion is supported by studies of criminal violence, which show that the most extreme crimes are committed by individuals whose hostility is characteristically overcontrolled (cf. Blackburn, 1968; Haven, 1973; Megargee, 1966, 1971; White, McAdoo, & Megargee, 1971). There are some interesting implications for psychotherapy that follow from an analysis of avoidance-approach conflict. For one, if the approach impulses are nondestructive, as in the case of certain sexual feelings, then persuasion, coercion, and confrontation should serve a useful role, as once the individual is brought to the goal side of the point of intersection of the gradients, the conflict will be resolved. This can be contrasted with approach—avoidance conflict, in which the same techniques are contraindicated, as pressure to approach what is feared is apt to induce unmanageable levels of anxiety and therefore cause the patient to withdraw from therapy (cf. Bollard & Miller, 1950). Thus, it is important for the therapist to discriminate between avoidance—approach and approachavoidance conflict. It also follows that the therapist must distinguish an avoidance-approach conflict in which the approach motive is socially acceptable from one in which it is destructive. It is obviously unwise to encourage an individual with the latter type of conflict to proceed to the goal side of the intersection point of the gradients or to lower the fear gradient, as either would result in the expression of destructive behavior. Tt is noteworthy, in this respect, that lowering the fear gradient is the procedure recommended by Dollard and Miller (1950) for treating approach-avoidance conflict. The method of choice for treating an avoidance-approach conflict when the approach motive is potentially destructive is to increase discrimination of the avoidance dimension so that selective avoidance reactions against extreme and inappropriate expression of the approach motive can be substituted for

AVOIDANCE-APPROACH

a blanket prohibition against its expression in any form. It is interesting to consider that the differences in therapeutic approaches recommended for approach-avoidance and avoidance-approach conflicts with socially acceptable and unacceptable motives correspond to major emphases in different schools of psychotherapy. For approach-avoidance conflict, the method of choice recommended by Bollard and Miller (1950) is reduction of fear, or lowering of the avoidance motive. This appears to be appropriate when the approach motive involves socially acceptable tendencies. For avoidance-approach conflict, when the approach motive is socially acceptable, the method of choice is to induce the individual to arrive at the goal side of the intersection of the gradients. Although reduction of fear might ultimately work, other procedures, such as encouragement, coercion, and environmental manipulation provide a simpler and more efficient form of treatment. It is noteworthy that such procedures are frowned on by adherents of depth psychology, such as psychoanalysts, and by adherents of accepting forms of treatment, such as clientcentered therapists, but are routinely practiced by counseling psychologists, rationalemotive psychologists, and psychodramatists. For both approach-avoidance and avoidanceapproach conflicts in which the inhibited drive is destructive, the method of choice is to teach discrimination, both with respect to the stimulus dimension and the response dimension, with the aim of ultimately replacing an all-or-none system of total avoidance or total impulse expression with a system that facilitates modulated control. Schools of therapy that emphasize such an approach among other techniques include psychoanalysis, rational-emotive therapy, and, in some cases, behavior therapy. References Blackburn, R. Personality in relation to extreme aggression in psychiatric offenders. British Journal of Psychiatry, 1968, 114, 1301-1302. Brown, J. S. Gradients of approach and avoidance responses and their relation to motivation. Journal of Comparative and Physiological Psychology, 1948, 41, 450-465.

CONFLICT

1021

Bugelski, B. R., & Woodward, D. P. The effect of distinctive cues on the spatial gradient of avoidance. Journal of Comparative and Physiological Psychology, 1951, 44, 450-456. Clifford, T. Between-chain and within-chain speed gradients as functions of length. Learning and Motivation, 1973, 4, 74-90. Desiderate, O., Foldes, J. M., & Gockley, J. S. Incubation of anxiety: Effect on temporal generalization. Psychonomic Science, 1966, 6, 139-140. Dollard, J., & Miller, N. E. Personality and psychotherapy. New York: McGraw-Hill, 1950. Elder, S. T., Kuelme, T. E., Clarke, N., & Larre, E. E. Approach-avoidance conflict: I. Empirically determined speed-of-locomotion approach gradients. Psychological Reports, 1970, 27, 623-628. Elder, S. T., Kuehne, T. E., & Moriarty, D. D., Jr. Approach-avoidance conflict: II. Role of olfactory cues. Psychological Reports, 1970, 27, 631-638. Gewirtz, H. K. Generalization of children's preferences as a function of reinforcement and task similarity. Journal of Abnormal and Social Psychology, 1959, 58, 111-118. Gjcsme, T. Goal distance in time and its effects on the relations between achievement motives and performance. Journal of Research in Personality, 1974, 8, 161-171. Haven, H. J. Descriptive and developmental characteristics of chronically ovcrcontrolled hostile prisoners. (Doctoral dissertation, Florida State University, 1972.) Dissertation Abstracts International, 1973, 33, 5S15B. Hearst, E. Stimultancous generalization gratients for appetitive and aversive behavior. Science, 1960, 132, 1769-1770. Hearst, E. Concurrent generalization gradients for food-controlled and shock-controlled behavior. Journal of the Experimental Analysis of Behavior, 1962, 5, 19-31. Heilizer, F. A review of theory and research on the assumptions of Miller's response competition (conflict) models: response gradients. Journal of General Psychology, 1977, 97, 17-71. Lcwin, K. A dynamic theory of personality:Selected papers. New York: McGraw-Hill, 1935. Losco, J., & Epstein, S. Relative steepness of approach and avoidance gradients as a function of magnitude and valence of incentive. Journal of Abnormal Psychology, 1977, 86, 360-368. Maher, B. A. Principles of psychopathology: An experimental approach. New York: McGraw-Hill, 1966. Megargee, E. I. Undercontrolled and overcontrolled personality types in extreme anti-social aggression. Psychological Monographs, 1966, 80(3, Whole No. 611). Megargee, E. I. The role of inhibition in the assessment and understanding of violence. In J. E. Singer (Ed.), The control of aggression and violence: Cognitive and physiological factors. New York: Academic Press, 1971. Miller, N. E. Experimental studies of conflict. In H. McV. Hunt (Ed.), Personality and the behavior disorders. New York: Ronald Press, 1944.

1022

SEYMOUR EPSTEIN

Miller, N. E. Liberalization of basic S-R concepts: Extensions to conflict behavior, motivation and social learning. In S. Koch (Ed.), Psychology: A study of a science (Vol. 2 ) : General systematic formulations, learning, and special processess. New York: McGraw-Hill, 19S9. Miller, N. E., & Kracling, D. Displacement: Greater generalization of approach than avoidance in a generalized approach-avoidance conflict. Journal of Experimental Psychology, 1952, 43, 217-221. Miller, N. E., & Murray, E. J. Displacement and conflict: Learnablc drive as a basis for the steeper gradient of avoidance than of approach. Journal of Experimental Psychology, 1952, 43, 227-231. Murray, E. J., & Berkun, M. M. Displacement as a function of conflict. Journal of Abnormal and Social Psychology, 19SS, 51, 47-56. Rigby, W. K. Approach and avoidance gradients and conflict behavior in a predominantly temporal situation. Journal of Comparative and Physiological Psychology, 1954, 47, 83-89. Saltz, E., Whitman, R. N., & Paul, C. Performance in the runway as a function of stimulus-differenti-

ation, American Journal of Psychology, 1963, 76, 124-127. Schroeder, S. R., & Gerjuoy, H. Running-speed goal gradient with massed training. Psychological Record, 1965, 15, 425-433. Smith, N. W. GSR measures of cigarette smokers' temporal approach and avoidance gradients. Psychological Record, 1965, 15, 261-268. Smith, N. W. Spatial conflict gradients of cigarette smokers: A methodology. Journal oj General Psychology, 1969, 80, 287-290. Tempone, V. J. Stimulus generalization as a function of mental age. Child Development, 1965, 36, 229235. Weiss, R. F. Deprivation and reward magnitude effects on speed throughout the goal gradient. Journal of Experimental Psychology, 1960, 60, 384390. White, W. C. Jr., McAdoo, W. G., & Megargcc, E. I. Personality factors associated with over- and under-controlled offenders. FCI Research Reports, 1971, 3(5), 1-6.

Received July 29, 1977 •

Avoidance--approach: the fifth basic conflict.

Journal of Consulting and Clinical Psychology 1978, Vol. 46, No. 5, 1016-1022 Avoidance-Approach: The Fifth Basic Conflict Seymour Epstein Universi...
554KB Sizes 0 Downloads 0 Views