Journal of Abnormal and Social Psychology 1963, Vol. 67, No. 2, 95-104 MENTAL ILLNESS, PREDICTABILITY, AND AFFECTIVE CONSEQUENCES AS STIMULUS FACTORS IN PERSON PERCEPTION* KENNETH J. GERGEN AND EDWARD E. JONES Duke University Before and after attempting to predict the behavior of a stimulus person (SP) in a choice situation, 64 Ss were asked to record their impressions of him on a trait rating scale. By design (a) the SP was presented as normal or mentally ill, (b) his behavior was predictable or unpredictable, and (c) errors of prediction were or were not signaled by a noxious buzzer. As pre- dicted, evaluative judgments of the mentally ill SP did not vary as a function of predictability unless predictive failures were accompanied by unpleasant consequences, and unless the SP himself had control over the delivery of these consequences. Judgments of the normal SP were solely a function of his predictability, regardless of the consequences of predictive failure. The planning of the present study pro- ceeded from the assumption that persons who are categorized as mentally ill tend to arouse a basic emotional ambivalence in those who come in contact with them. Thus far, the psychological literature has empha- sized the negative components of this ambiva- lence and has highlighted the derogatory semantic associates given by the man in the street to such terms as "neurotic" and "psy- chotic" (Nunnally, 1961). There are many obvious reasons for assigning negative at- tributes to those who are mentally ill: they are often obstructive, demanding, sullen, self- centered, and/or irrational. At the same time, however, the mentally ill person may be perceived as pathetic, a tragic figure victimized by a cruel environment. Following the logic of Heider's (19S8) comments on phenomenal causality, we would have to ex- pect a certain inhibition of negative feelings toward the mentally ill person, at least to the extent that this person does seem de- feated by a net of circumstances not of his own making. 1 The present experiment was conducted while Kenneth J. Gergen held a research traineeship at the Durham, North Carolina, Veterans Administra- tion Hospital. We are especially indebted to Paul Daston, Malcolm Meltzer, Russell Tomlinson, and H. R. Silberman for their strongly facilitating administrative help, and to other members of the hospital staff for making it possible to conduct the experiment in that setting. The study also is part of a research program supported by National Science Foundation Grant No. G-88S7. But there is more to the positive side of the ambivalence toward the mentally ill than a mere suspension or inhibition of negative feelings. The neurotic or psychotic person is typically in a position which arouses some degree of nurturance in others. In a way, the sicker, the more incapacitated, the more ir- rational a person becomes the more he clearly becomes an object of our pity and sympathy. While we may be disturbed by, and avoidant toward the schizophrenic, at the same time we are sympathetically involved in his tragedy and emotionally committed to his rehabilitation. It was the purpose of the present study to examine some of the conditions under which this emotional ambivalence might be "split" so that either the positive or the negative component would be dominant. If we con- sider the plight of a perceiver faced with the task of judging the characteristics of a mentally ill person, we might predict that the attribution of negative characteristics would be inhibited under conditions where the his- torical and situational factors contributing to the illness are made salient, and the perceiver is in a detached position and his personal needs and aspiration are not affected by the actions of the sick person. In the present study the salience of historical and situational factors was held constant and at a relatively high level, and the degree of affective con- sequences was systematically varied. That is, half of the subject perceivers were exposed 95