An emotion-attribution approach to moral behavior: interfacing cognitive and avoidance theories of moral development.

A theory is presented concerning the impact of attributions about the causes of emotional responses as they infl uence self-control in temptation situations. Research is reviewed indicating a high level of adult sensitivity to external infl uence in making such causal attributions. Two studies are presented in which the post transgression emotions of second-grade children are labeled shame (because of being found out) or guilt (due to the transgression itself); when a similar situation was subsequently represented as safe from detection, shame-condition children transgressed 60–80% more than guilt-condition subjects. It is suggested that emotional arousal elicited in temptation situations because of past punishment or options that are inconsistent with the self-image is necessary for inhibition but not suffi cient unless attributed to a relevant cause. The literature on the relative effectiveness of moral socialization techniques is discussed with respect to the theory, and the relevance to cognitive dissonance and to overjustifi cation approaches to motivation is discussed. An integration of social-learning and cognitive-developmental theories is approached through explicating the translation of moral decision into behavior by focusing on the ways that cognition may exert partial control over the impact of less fi nely differentiated emotional response, allowing cognitive overrides of contradictory emotional dispositions without eliminating the emotion. 300 DIENSTBIER ET AL. IN PSYCHOLOGICAL REVIEW 82 (1975) the resultant behavior may relate to higher orders of moral judgment as defi ned by modern cognitive approaches to morality. Although the association of these negative emotional states with decisional and behavioral processes may take place in part through mechanisms of conditioning and learning specifi ed by traditional avoidance learning theorists (e.g., Mowrer, 1950), these associations depend heavily on the causal attributions that are made about the source of the negative emotions during socialization experiences. Social and situational infl uences (often temporary) play an important role in forming causal attributions; different physical and verbal socialization techniques particularly provide different information relevant to the child’s causal attributions. Later, in the face of temptation, the impact of the emotion response on behavior will be heavily infl uenced by the still malleable beliefs held about the causal origins of the emotional response. When the individual believes detection of transgression is not possible, a negative emotional response in the face of temptation is necessary but not suffi cient for the inhibition of transgression. In order for that emotional arousal to serve an inhibitory function, the individual must identify his emotional discomfort as due to a relevant cause, such as the transgression per se (and the implications of the transgression for self-image, etc.), rather than as due to an irrelevant cause, such as the fear of punishment. With maturity, control of emotional attributions passes from the socializing agent to the individual and becomes related to the level of moral development, although situational and social cues continue to play an important role. EM O T I O N AT T R I B U T I O N The work of Schachter and Singer (1962) demonstrated that when subjects injected with adrenalin did not anticipate that arousal effects would result, they tended to attribute their emotional state to the situation, feeling happier or angrier depending on the specifi c post injection experimental context. That study was the fi rst in a series by Schachter, his students, and others that indicated that people are quite fl exible in the manner in which they make causal inferences about the source and meaning of their emotional arousal. Subsequent research exploring the implications of those fi ndings for moral behavior followed the work of Lykken (1957). Lykken had shown that psychopathic individuals seemed not to experience normal emotion-induced inhibitions. Schachter and Latané (1964) demonstrated that psychopathic criminals (compared to normal criminals) would learn to avoid shocked errors (compared to nonshocked errors) in a lever-maze task only after sustaining arousal from an adrenalin injection. Schachter and Ono (cited in Schachter & Latané, 1964) found increased cheating by college students who had received the tranquilizer chlorpromazine (compared to placebo control groups). These studies can be viewed as evidence of the necessity of emotional arousal as an important component in the avoidance of responses with aversive consequences, with an increase in arousal facilitating that avoidance and a decrease attenuating avoidance. However, since only the degree of arousal was manipulated, and not interferences the subject might make concerning causal attribution, the necessity of arousal was demonstrated but the issue of the suffi ciency of arousal was not addressed. By illustrating the role of emotional arousal, these data provided support for Mowrer’s (1950) two-factor avoidance model. Mowrer’s model emphasized that the emotional response that the organism developed in response to the cues signaling an aversive event would motivate the avoidance response, which, when successful, would lead to the reinforcing reduction of the negative emotional state.1 Emotional arousal in anticipation of punishment for a potential transgression should, by this interpretation, lead to the avoidance of the temptation behavior. But Schachter and Latané (1964) speculated that for the psychopath, the avoidance of behavior with aversive consequences depended on more than the mere presence of arousal. 1 More modern theorists have extended this perspective, suggesting that anxiety becomes conditioned specifi cally to the “behavioral and cognitive precursors of the act” (Aronfreed, 1968, pp. 54–55). EMOTION-ATTRIBUTION APPROACH TO MORAL BEHAVIOR 301 Psychopaths experience a high level of physiological arousal with no apparent subjective emotional distress in response to cues that would evoke an emotional response in normal individuals. Schachter and Latané suggested that it was the chronic inability of the psychopath to experience his physiological arousal as emotion that accounted for his moral insensitivity.2 These suggestions approach the perspective proposed in this article, but our view differs by suggesting that even in temptation situations where arousal is experienced as emotion, specifi c causal attributions are crucial in determining the behavioral outcome. It was the Schachter and Latané (1964) research that led to a series of studies by the fi rst author indicating that college-age men and women more often succumbed to the temptation to cheat when they could attribute the emotional arousal associated with cheating to a placebo pill, which supposedly had associated emotional side effects (Dienstbier, 1972; Dienstbier & Munter, 1971). Whereas Schachter and Singer (1962) had demonstrated that external cues are important in allowing the attribution of and experiencing of artifi cially induced arousal as specifi c emotion, the cheating research demonstrated the reverse—that naturally induced specifi c emotions could be interpreted as nonspecifi c arousal when subjects attributed the associated arousal to a nonemotionally relevant source such as the placebo pill. Since these cheating studies are important for the theoretical perspective developed in this article, they are reviewed here in some detail. Recruited as participants in a study on the effects of a vitamin supplement on vision, subjects were given a vocabulary test as a “delay task while waiting for the pill to take effect.” Besides anticipating visual effects from the pill, a random half of the subjects were led to anticipate arousal effects of “a pounding heart, hand tremor, sweaty palms, a warm or fl ushed face, and a tight or sinking feeling in the stomach” (following Nisbett & Schachter, 1966). Control subjects anticipated benign side effects irrelevant to arousal. After the vocabulary test, to validate the potency of the pill, subjects viewed the auto-kinetic illusion, and were led to understand that the pill was responsible for their seeing the immobile light as moving. Thereafter, subjects were given the opportunity to check over their vocabulary test answers and to see the correct answers. They were informed that the department developing the vocabulary test was interested in using it for future prediction of college success, and that the “board of psychologists” responsible for it would like to interview any students who did particularly poorly on the test (less than 20 right). The test was designed so that less than 1% of college freshmen could attain that score without cheating. The period during which the individual could look over the correct answers to the vocabulary test and his own answers was also introduced as a “delay period,” in this case prior to the rating of side effects because “it was important that the subjects rate the side effects they might be experiencing just 9 minutes after they had experienced the visual effects.” The salience of the side effects expectation was increased for that period by having subjects look over a side effects checklist at the onset of that period. Any cheating that took place during the delay period could be detected because pressure-sensitive paper (making a second copy of the answers) was placed under the answer paper during the taking of the test; the answer paper was removed by the subjects prior to the cheating period, so that thereafter any changes made on the real answer paper would not correspond to those on the pressuresensitive paper . 2 Suggestions by Mandler (1964) and the theoretical structure proposed by Lazarus (1968) are useful in understanding possible mechanisms to account for the psychopath’s lack of emotional experience in the face of arousal. Mandler suggested that the childhood of the typical psychopath is characterized by frequent emotional upheaval~ and that the individual might learn to minimize the emotional discomfort of these frequent traumas by ignoring the physiological signals from his body, which, if allowed to contribute to a strong emotional feeling, would greatly accentuate his distress. Mandler’s suggestions fi t well with Lazarus’s (1968) concept of benign reappraisal, the process by which the stress reaction is reduced by a cognitive reappraisal involving a lowered assessment of the potential danger in a situation after the individual has adjudged that he is powerless to affect the outcome. 302 DIENSTBIER ET AL. IN PSYCHOLOGICAL REVIEW 82 (1975) In the fi rst study (Dienstbier & Munter, 1971) 49% of the subjects who anticipated arousal side effects from their placebo pills cheated, compared to 27% of the control condition, with cheating defi ned as changing any answers, This signifi cant difference became more impressive when it was noted that the entire effect was due to the male subjects, whose cheating rates were 56% and 17% in the two conditions, respectively. Follow-up research (Dienstbier, 1972) indicated that women, too, respond to the arousal placebo with increased cheating if the threat of the consequences for failure is made less frightening (it appeared that in the original study the more intense anxiety experienced by women in response to the threat of being called by the “board of psychologists” interfered with their attending to the pill side effects). In a second study reported in the same article, it was further determined that the effect of the arousal placebo in reducing resistance to temptation was indeed due to the subject’s attributing specifi cally to the pill—merely attending to arousal feelings without having taken the placebo did not produce the effect. It appeared that all subjects experienced emotional arousal while confronting the temptation to cheat, largely as a result of considering the act of cheating and the potential negative consequences. But when subjects could attribute the cause of their arousal to the pill, that arousal ceased to have an inhibiting effect. It is not, in other words, that the subjects in either pill condition experienced more or less arousal. but rather that the attribution of the arousal to a neutral source (in the arousal pill condition) defused that arousal, preventing it from having emotional relevance.3 3 There are two lines of evidence suggesting that subjects make the attributions of emotional causality at a less than verbally conscious level. First, when subjects were interviewed, none ever indicated conscious awareness of such a decisional process. (Subjects were completely debriefed and questioned concerning possible suspiciousness following the cheating research.) Second, at the end of the cheating period, when arousal subjects rated the degree to which they experienced the arousal side effects, the rate of reporting such side effects Whereas Schachter and Latané (1964) had speculated that this process in chronic form accounted for the moral insensitivity of the psychopath, the cheating studies suggested that the same process could be demonstrated to produce episodes of moral insensitivity in normal subjects. The work of Solomon and Wynne (1954) and subsequent researchers emphasized that although emotional arousal was necessary in the acquisition of avoidance responses, the elimination of overt signs of arousal was often observed once such responses had been well learned. Except for the Schachter and Ono data (cited in Schachter & Latané 1964), which demonstrated increased cheating with chloropromazine, we would not have anticipated, therefore, that emotional arousal played any role in moral decision making for normal adults. Yet the cheating results suggested that even for 19-year-olds in temptation situations, the avoidance of transgression is still heavily dependent on emotional arousal. Additionally, the studies demonstrated that although 19-yearold adults should have had considerable experience in learning which features of their own behavior and of the environment lead to negative emotional arousal, they were surprisingly susceptible to external infl uences in attributing arousal experienced in the face of temptation. But however powerful these data are, they do not provide the theoretical bridge to our fi nal point. Our thesis requires that we demonstrate more than the defusing of emotional arousal through attribution to a non-emotional source in a context as artifi cial as that associated with the placebo pill studies, We must demonstrate instead that emotional arousal may be defused was surprisingly low. (There was, unfortunately, no control group with which they could be compared on this dimension, since the benign symptom subjects rated themselves on the benign symptoms.) It appears, therefore, that although the information concerning expected symptoms is given verbally, thereafter the very signifi cant impact that these symptom expectations have upon cheating behavior is achieved by processes not apparent to the subject. EMOTION-ATTRIBUTION APPROACH TO MORAL BEHAVIOR 303 for moral behavior even though still experienced as negative emotion, through the attribution to an emotion-inducing source that has no direct relevance to the moral decision. SOCIALIZATION AND EMOTION ATTRIBUTION We speculated that children would be even more receptive than adults had proven to be to cues that would suggest different causal attributions for emotional arousal, and that these cues would play an important role in determining the impact of socialization processes. Specifi cally, if a socialization pattern left the child with the impression that the cause of his emotional arousal or discomfort was his own misbehavior, the impact of this internal orientation on the child’s behavior would be very different from the impact of an external orientation in which equally intense emotion was thought to be caused by fear of punishment. The latter attribution would be particularly ineffective in situations in which the tempted individual could be certain that detection was impossible. Even early in their socialization history, most normal children will experience some emotional arousal following transgression and detection by a disapproving adult. Several features of the situation naturally lead to that arousal: the child’s knowledge that the adult disapproves, previous scoldings now recalled, anxiety over the uncertainty of the outcome, and the likelihood that the adult will emphasize the discrepancy between the child’s behavior and higher standards. The manner in which the socializing agent responds to the child’s transgression will infl uence the causal attributions the child will make about his emotional discomfort. Socialization procedures that draw the child’s attention to the transgression rather than to the aftermath (confrontation and punishment) facilitate the attribution of arousal to the act of transgression. Generally, such procedures involve discussions of the transgression in a manner that does not threaten the child, so that the child is able to dwell on the transgression itself and the behavioral standards he has violated, rather than on the threat that confrontation offers. Often the self-concept is manipulated (e.g., “you are a bad boy”) in relation to the standard or behavior (“because you didn’t listen and lied”). The attribution of arousal to the anticipated transgression later leads the child to avoid the transgression itself. On the other hand, responses by socializing agents that threaten the child or that become salient through pain draw suffi cient attention away from the transgression so that the child is likely to attribute his emotional discomfort exclusively to confrontation with the socializing agent. When later facing a similar temptation situation, a child treated in the latter manner might experience a high level of emotional arousal, but by attributing his arousal to fear of detection and punishment he would tend to resist temptation only if he believed that detection was likely. In our research with children we wished to buttress our theory by establishing a temptation and transgression situation in which the emotional arousal following detection would be equal for the two groups of children, with the children assured that future transgressions could not be detected. We then imparted to the children an internal (due to own behavior) or an external (due to confrontation) causal attribution concerning the source of that emotional arousal. It was hypothesized that the children who believed that they “felt bad” for their previous transgression (internal orientation, or guilt) would misbehave far less than children in the same situation who believed that they “felt bad” because their previous transgression had been detected (external orientation, or shame).