someone], sonkei [feeling of respect for someone]) and thus were designated as positive and other focused. In contrast, some negative emotions are typically derived from one's failure to offer or reciprocate favors to relevant others and thus to fully participate in the relationship. They are thus closely linked to disturbance to interdependence and a consequent desire to repair the disturbance. They include oime [feeling of indebtedness], shame, and guilt. Finally, as noted before, interdependent selves are likely to tolerate ambivalence regarding one's interdependent status with some relevant others. Interestingly, some emotions are uniquely linked to this interpersonal ambivalence. Three such emotions (amae [hopeful expectation of others' indulgence or favor], tanomi [feeling like relying on someone], and sugari [feeling like leaning on someone]) were examined. Japanese respondents reported how frequently they experienced each of the 15 emotions listed in Table 2. The five-factor structure implied by the theoretical designation of the 15 emotions to one of the five types was verified in a confirmatory factor analysis (JOreskog, 1969). A correlation matrix for the five types is given in Table 3. There was a strong correlation between positive and negative ego-focused emotions, as may be expected if both of them are derived from and also foster and reinforce an independent construal of self. Furthermore, these ego-focused emotions are clearly distinct from the other-focused emotions. Thus, neither positive nor negative ego-focused emotions had any significant relationship with other-focused, positive emotions. Interestingly, however, these ego-focused emotions were significantly associated with the ambivalent and, to a larger extent, with the negative other-focused emotions, suggesting that the experience of ego-focused emotions, either positive or negative, is readily accompanied, at least in Japanese culture, by the felt disturbance of a relationship and, thus, by a strong need to restore harmony. Alternatively, being embedded in a highly reciprocal relation and feeling obliged to contribute to the relationship may sometimes be perceived as a burden or pressure, hence rendering salient some of the ego-focused emotions. Finally, the three types of other-focused emotions (positive, ambivalent, and negative) are all positively correlated. Can the frequency of experiencing the five types of emotions be predicted by one's predominant construal of self as independent or interdependent? To address this issue, Kitayama and Markus (1990) also asked the same respondents eight questions designed to measure the extent to which they endorse an independent construal of self (e.g., "Are you a kind of person who holds on to one's own view?"; "How important is it to hold on to one's own view?") and eight corresponding questions designed to measure the extent to which they endorse an interdependent construal of self (e.g., "Are you the kind of person who never forgets a favor provided by others?"; "How important is it to never forget a favor provided by others?"). Consistent with the current analysis, the frequency of experiencing both positive and negative ego-focused emotions significantly increased with the independent construal of self. They were, however, either negatively related (for positive emotions) or unrelated (for negative emotions) to the interdependent construal of self. In marked contrast to this pattern for the ego-focused emotions, all three types of other-focused emotions were significantly more frequently experienced by those with more interdependent construals of self. These emotions, however, were either unrelated (for positive and negative other-focused emotions) or negatively related (for the ambivalent emotions) to the independent construal of self. Consequences for Motivation The study of motivation centers on the question of why people initiate, terminate, and persist in specific actions in particular circumstances (e.g., Atkinson, 1958; Mook, 1986). The answer given to this question in the West usually involves some type of internal, individually rooted need or motive-the motive to enhance one's self-esteem, the motive to achieve, the motive to affiliate, the motive to avoid cognitive conflict, or the motive to self-actualize. These motives are assumed to be part of the unique, internal core of a person's self-system. But what is the nature of motivation for those with interdependent self-systems? What form does it take? How does the ever-present need to attend to others and to gain their acceptance influence the form of these internal, individual motives? Are the motives identified in Western psychology the universal instigators of behavior? As with cognition and emotion, those motivational processes that implicate the self depend on the nature of the self-system. If we assume that others will be relatively more focal in the motivation of those with interdependent selves, various implications follow. First, those with interdependent selves should express, and perhaps experience, more of those motives that are social or that have the other as referent. Second, as we have noted previously, for those with independent selves, agency will be experienced as an effort to express one's internal needs, rights, and capacities and to withstand undue social pressure, whereas among those with interdependent selves, agency will be experienced as an effort to be receptive to others, to adjust to their needs and demands, and to restrain one's own inner needs or desires. Motives related to the need to express one's agency or competency (e.g., the achievement motive) are typically assumed to be common to all individuals. Yet among those with interdependent selves, striving to excel or accomplish challenging tasks may not be in the service of achieving separateness and autonomy, as is usually assumed for those with independent selves, but instead in the service of more fully realizing one's connectedness or interdependence. Third, motives that are linked to the self, such as self-enhancement, self-consistency, self-verification, self-affirmation, and self-actualization, may assume a very different form depending on the nature of the self that is being enhanced, verified, or actualized. More interdependent motives? Murray (1938) assembled what he believed to be a comprehensive list of human motivations (see also Hilgard, 1953, 1987). Many of these motives seem most relevant for those with independent selves, but the list also includes some motives that should have particular salience for those with interdependent selves. These include deference, the need to admire and willingly follow a superior, to serve gladly; similance, the need to imitate or emulate others, to agree and believe; affiliation, the need to form friendships and associations; nurturance, the need to nourish, aid, or protect another; succorance, the need to seek aid, projection, or sympathy and to be dependent; avoidance of blame, the need to avoid blame, ostracism, or punishment by inhibiting unconventional impulses and to be well behaved and obey the law; and abasement, the need to comply and accept punishment or self-deprecation. Many of the social motives suggested by Murray seem to capture the types of strivings that should characterize those with interdependent selves. When the cultural imperative is to seek connectedness, social integration, and interpersonal harmony, most of these motives should be typically experienced by the individual as positive and desirable. In contrast, when the cultural task centers on maintaining independence and separateness, holding any of these motives too strongly (e.g., similance and succorance) often indicates a weak or troubled personality. Thus, Murray, for example, gave the need to comply the pejorative label of need for abasement. The limited evidence for the idea that those with interdependent selves will experience more of the social or interdependent motives comes from Bond (1986), who summarized several studies exploring the motive patterns of the Chinese (see also McClelland, 1961). He found that the level of various motives are a fairly direct reflection of the collectivist or group-oriented tradition of the Chinese. Thus, Chinese respondents show relatively high levels of need for abasement, socially oriented achievement, change, endurance, intraception, nurturance, and order; moderate levels of autonomy deference, and dominance, and succorance; and low levels of individually oriented achievement, affiliation, aggression, exhibition, heterosexuality, and power. The socially oriented achievement motive has, as its ultimate goal, a desire to meet expectations of significant others, whereas the individually oriented achievement motive implies a striving for achievement for its own sake (discussed later). Hwang (1976) found, however, that with continuing rapid social change in China, there is an increase in levels of exhibition, autonomy, intraception, and heterosexuality and a decrease in levels of deference, order, nurturance, and endurance. Interestingly, it appears that those with interdependent selves do not show a greater need for affiliation, as might at first be thought, but instead they exhibit higher levels of those motives that reflect a concern with adjusting oneself so as to occupy a proper place with respect to others. The motive for cognitive consistency. Another powerful motive assumed to fuel the behavior of Westerners is the need to avoid or reduce cognitive conflict or dissonance. Classic dissonance occurs when one says one thing publicly and feels another, quite contrasting thing privately (Festinger & Carlsmith, 1959). And such a configuration produces particular difficulty when the private attitude is a self-defining one (Greenwald, 1980). One might argue, however, that the state of cognitive dissonance arising from counterattitudinal behavior is not likely to be experienced by those with interdependent selves. First it is the individuals' roles, statuses, or positions, and the commitments, obligations, and responsibilities they confer, that are the constituents of the self, and in that sense they are self-defining. As outlined in Figure 1, one's internal attributes (e.g., private attitudes or opinions) are not regarded as the significant attributes of the self. Furthermore, one's private feelings are to be regulated in accordance with the requirements of the situation. Restraint over the inner self is assigned a much higher value than is expression of the inner self. Thus, Kiefer (1976) wrote: Although Japanese are often acutely aware of discrepancies between inner feelings and outward role demands, they think of the latter... as the really important center of the self. Regarding feelings as highly idiosyncratic and hard to control, and therefore less reliable as sources of self-respect than statuses and roles, the Japanese tends to include within the boundaries of the concept of self much of the quality of the intimate social group of which he is a member. (R. J. Smith, 1985, p. 28) More recently, T Doi (1986) has argued that Americans are decidedly more concerned with consistency between feelings and actions than are the Japanese. In Japan there is a virtue in controlling the expression of one's innermost feelings; no virtue accrues from expressing them. Triandis (1989), for example, reported a study by Iwao (1988), who gave respondents a series of scenarios and asked them to judge which responses would be appropriate for the person described in the scenario. In one scenario, the daughter brings home a person from another race. One of the possible responses given was "thought that he would never allow them to marry but told them he was in favor of their marriage" This answer was rated as best by only 2% of Americans. In sharp contrast, however, it was rated as best by 44% of the Japanese. Among the Americans, 48% thought it was the worst response, whereas only 7% of the Japanese rated it as the worst. Common motives in an interdependent context. of those motives assumed by Murray (1938) and Hilgard (1987) to be universally significant, the achievement motive is the most well-documented example. Variously defined as the desire to overcome obstacles, to exert power, to do something as well as possible, or to master, manipulate, or organize physical objects, human beings, or ideas (Hall & Lindzey, 1957; Hilgard, 1987), the achievement motive is thought to be a fundamental human characteristic. However, the drive for achievement in an interdependent context may have some very different aspects from the motive for achievement in an independent cultural context. In a recent analysis of the content and structure of values in seven cultures (i.e., Australia, United States, Spain, Finland, Germany, Israel, and Hong Kong), S. H. Schwartz and Bilsky (1990) found a conflict between values that emphasize independent thought and action and those that emphasize restraining of one's own impulses in all samples except Hong Kong. In the Hong Kong sample, self-restraint appeared to be quite compatible with independent thought and action. Although all individuals may have some desire for agency or control over their own actions, this agency can be accomplished in various ways (Maehr, 1974). Pushing oneself ahead of others and actively seeking success does not appear to be universally valued. An illuminating analysis of control motivation by Weisz et al. (1984) suggests that acting on the world and altering the world may not be the control strategy of choice for all people. Instead, people in many Asian cultures appear to use what is termed secondary control. This involves accommodating to existing realities "sometimes via acts that limit individualism and personal autonomy but that enhance perceived alignment or goodness of fit with people, objects, or circumstances" (Weisz et al, 1984, p. 956). The American notion of achievement involves breaking away, pushing ahead, and gaining control over surroundings. How do selves concerned with fitting in and accommodating to existing realities achieve? The question of achievement motive in an interdependent context is all the more compelling because many of the most collective societies of the world currently appear extremely preoccupied with achievement. In an analysis of Chinese children's stories, for example, Blumenthal (1977) found that the most common behavior was achievement-oriented in nature, the second most frequent was altruism, and the third was social and personal responsibility. Among junior high school students in Japan, the motto "pass with four, fail with five" is now common. This refers to the fact that if one is sleeping 5 hr a night, he or she is probably not studying hard enough to pass exams. It appears, however, that this strong emphasis on achievement motivation is, in part, other motivated. It is motivated by a desire to fit into the group and to meet the expectations of the group. In the child's case, the group is the family, and the child's mission is to enhance the social standing of the family by gaining admission to one of the top universities. The motive to achieve need not necessarily reflect a motive to achieve for "me" personally (Maehr & Nicholls, 1980). It can have social or collective origins. Children are striving to achieve the goals of others, such as family and teachers, with whom they are reciprocally interdependent. Consistent with this notion, Yu (1974) reported that the strength of achievement motivation was correlated positively with familism and filial piety Striving for excellence necessarily involves some distancing or separating from some others, but the separation allows the child to properly accomplish the task of the student and thus to fulfill his or her role within the family. Several studies by Yang (Yang, 1982/1985; Yang & Liang, 1973) have sought to distinguish between two types of achievement motivation: individually oriented and socially oriented. Individually oriented achievement motivation is viewed as a functionally autonomous desire in which the individual strives to achieve some internalized standards of excellence. In contrast, socially oriented achievement motivation is not functionally autonomous; rather, individuals persevere to fulfill the expectations of significant others, typically the family (Bond, 1986). With socially oriented achievement, when the specific achievement goal is met, the intense achievement motivation formerly evident may appear to vanish. This analysis indeed fits many anecdotal reports indicating that once admitted into the college of their choice, or hired by their preferred company, Japanese high school and college students are no longer particularly interested in achievement. Once a new goal is established, of course, the socially oriented achievement motive may he easily reengaged by any figure who can serve as a symbolic substitute for family members. A longitudinal survey conducted in Japan over the last 30 years (Hayashi, 1988) has repeatedly shown that approximately 80% of the Japanese, regardless of sex, age, education, and social class, prefer a manager with a fatherlike character (who demands a lot more than officially required in the work, yet extends his care for the person's personal matters even outside of work) over a more Western-type, task-oriented manager (who separates personal matters from work and demands as much as, yet no more than, officially required). In a large number of surveys and experiments, Misumi and his colleagues (summarized in Misumi, 1985) have demonstrated that in Japan a leader who is both demanding and personally caring is most effective regardless of the task or the population examined (e.g., college students, white-collar workers, and blue-collar workers). This is in marked contrast to the major conclusion reached in the leadership literature in the United States, which suggests that leadership effectiveness depends on a complex interaction between characteristics of leaders, characteristics of followers, and, most important, on the nature of the task (Fiedler, 1978; Hollander, 1985). According to our analysis, in Japan as well as in other interdependent cultures, it is the personal attachment to the leader and the ensuing obligation to him or her that most strongly motivate people to do their work. Motivation mediated by a strong personal relationship, then, is unlikely to be contingent on factors associated with the specific task or environment. The self-related motive. The motive to maintain a positive view of the self is one motive that psychologists since James (1890) through Greenwald (1980), Hailer (1983), Steele (1988), and Tesser (1986) have assumed to be universally true. What constitutes a positive view of self depends, however, on one's construal of the self.' For those with independent selves, feeling good about oneself typically requires fulfilling the tasks associated with being an independent self; that is, being unique, expressing one's inner attributes, and asserting oneself (see Table 1). Although not uncontested, a reasonable empirical generalization from the research on self-related motives is that Westerners, particularly those with high self-esteem, try to enhance themselves whenever possible, and this tendency results in a pervasive self-serving bias. Studies with American subjects demonstrate that they take credit for their successes, explain away their failures, and in various ways try to aggrandize themselves (e.g., Gilovich, 1983; Lau, 1984; J. B. Miller, 1986; Whitley & Frieze, 1985; Zuckerman, 1979). Maintaining self-esteem requires separating oneself from others and seeing oneself as different from and better than others. At 4 years old, children already show a clear self-favorability bias (Harter, 1989). When asked to compare themselves with others with respect to intelligence, friendliness, or any skill, most children think they are better than most others. Wylie (1979) reported that American adults also consider themselves to be more intelligent and more attractive than average, and Myers (1987), in a national survey of American students, found that 70% of students believe they are above average in leadership ability, and with respect to the 'ability to get along with others," 0% thought they were below average, 60% thought they were in the top 10%, and 25% thought they were in the top l Moreover, as documented by Taylor and Brown (1988), among Americans, most people feel that they are more in control and have more positive expectations for themselves and their future than they have for other people. This tendency toward false uniqueness presumably derives from efforts of those with independent selves to maintain a positive view of themselves. The motive to maintain a positive view of the self may assume a somewhat different form, however, for those with interdependent selves. Feeling good about one's interdependent self may not be achieved through enhancement of the value attached to one's internal attributes and the attendant self-serving bias. Instead, positive feelings about the self should derive from fulfilling the tasks associated with being interdependent with relevant others: belonging, fitting in, occupying one's proper place, engaging in appropriate action, promoting others' goals, and maintaining harmony (see Table 1). This follows for at least two reasons. First, people with interdependent selves are likely to be motivated by other-focused emotions, such as empathy and oime (i.e., the feeling of psychological indebtedness) and to act in accordance with the perceived needs and desires of their partners in social relations, and this may produce a social dynamic where individuals strive to enhance each other's self-esteem. In such reciprocal relationships, other enhancement could be more instrumental to self-enhancement than direct attempts at self-enhancement because the latter are likely to isolate the individual from the network of reciprocal relationships. second, self-esteem among those with interdependent selves may be based in some large measure on their capacity to exert control over their own desires and needs so that they can indeed belong and fit in. As noted earlier (see also Weisz et al., 1984), such self-control and self-restraint are instrumental to the ability to flexibly adjust to social contingencies and thus are highly valued in interdependent cultures. Indeed, self-restraint together with flexible adjustment is often regarded as an important sign of the moral maturity of the person. A developmental study by Yoshida, Kojo, and Kaku (1982, Study 1) has documented that self-enhancement or self-promotion are perceived quite negatively in Japanese culture. Second (7-8 years old), third (8-9 years old), and fifth graders (10-11 years old) at a Japanese elementary school were asked how their classmates (including themselves) would evaluate a hypothetical peer who commented on his own superb athletic performance either in a modest, self-restrained way or in a self-enhancing way. The evaluation was solicited on the dimension of personality ("Is he a good person?") and on the dimension of ability ("Is he good at [the relevant athletic domain]?"). As shown in Figure 4A, the personality of the modest peer was perceived much more positively than was that of the self-enhancing peer. Furthermore, this difference became more pronounced as the age (grade) of the respondents increased. A similar finding also has been reported for Chinese college students in Hong Kong by Bond, Leung, and Wan (1982), who found that individuals giving humble or self-effacing attributions following success were liked better than those giving self-enhancing attribution. The most intriguing aspect of the Yoshida et al. (1982) study, however, is their finding for the ability evaluation, which showed a complete crossover interaction (see Figure 48). Whereas the second graders took the comment of the peer at face value, perceiving the self-enhancing peer to be more competent than the modest peer, this trend disappeared for the third graders, and then completely reversed for the fifth graders. Thus, the fifth graders perceived that the modest peer was more competent than the self-enhancing peer. These findings indicate that as children are socialized in an interdependent cultural context, they begin to appreciate the cultural value of self-restraint and, furthermore, to believe in a positive association between self-restraint and other favorable attributes of the person not only in the social, emotional domains but also in the domains of ability and competence. Although it is certainly possible for those with independent selves to overdo their self-enhancement (see Schlenker & Leary, 1982), for the most part, the American prescription is to confidently display and express one's strengths, and those who do so are evaluated positively (e.g., Greenwald, 1980; Mullen & Riordan, 1988). Self- or other-serving bias. Given the appreciation that those with interdependent selves have for self-restraint and self-control, the various self-enhancing biases that are common in Western culture may not be prevalent in many Asian cultures. In an initial examination of potential cultural variation in the tendency to see oneself as different from others, Markus and Kitayama (in press) administered questionnaires containing a series of false-uniqueness items to large classes of Japanese college students in Japan and to large classes of American college students in the United States. In both cases, the classes were chosen to be representative of university students as a whole. They asked a series of questions of the form "What proportion of students in this university have higher intellectual abilities than yourself?" There were marked differences between the Japanese and the American students in their estimations of their own uniqueness; the Americans displayed significantly more false uniqueness than the Japanese. American students assumed that only 30% of people on average would be better than themselves on various traits and abilities (e.g., memory, athletic ability, independence, and sympathy), whereas the Japanese students showed almost no evidence of this false uniqueness. In most cases, the Japanese estimated that about 50% of students would be better than they were or have more of a given trait or ability. This is, of course, the expected finding if a representative sample of college students were evaluating themselves in a relatively nonbiased manner. In a recent series of studies conducted in Japan with Japanese college students, Takata (1987) showed that there is no self-enhancing bias in social comparison. In fact, he found just the opposite-a strong bias in the self-effacing direction. Participants performed several anagram problems that were alleged to measure memory ability. After completion of the task, the participants were presented with their actual performance on some of the trials and also the performance of another person picked at random from the pool of subjects who had allegedly completed the study. The direction of the self-other difference was manipulated to be either favorable or unfavorable to the subject. The dependent measures were collected in a private situation to minimize self-presentational concerns. Furthermore, because it was considered possible that the subjects might still believe they had a chance of seeing the other person afterward, in a followup study the "other person" was replaced with a computer program that allegedly simulated the task performance of the average college student. Several studies (e.g., Goethals, 1989; Marks, 1984; Wylie, 1979) reveal that with respect to abilities, Americans typically give themselves higher ratings than they give to others. Thus,when a comparison with another is unfavorable to the self, the self-enhancement hypothesis predicts that Americans should show little confidence in this estimate of their ability and seek further information. This, in fact, was the case in an American study by J. M. Schwartz and Smith (1976), which used a procedure very similar to Takata's (1987). When subjects performed poorly relative to another person, they had very little confidence in their own score. These American data contrast sharply with the Japanese data. Takata's study shows a tendency exactly the opposite of self-enhancement. Furthermore, the pattern did not depend on whether the comparison was made with another person or with the computer program. The Japanese subjects felt greater confidence in their self-evaluation and were less interested in seeking further information when they had unfavorable self-evaluations than when they had favorable ones. Similarly, Wada (1988) also reported that Japanese college students were convinced of their level of ability on a novel, informationintegration task after failure feedback, but not after success feedback. These data suggest what might be called a modesty bias or an other-enhancement bias in social comparison. A similar modesty bias among those with interdependent selves has also been suggested by Shikanai (1978), who studied the causal attribution for one's own success or failure in an ability task. Typically American subjects believe that their internal attributes such as ability or competence are extremely important to their performance, and this is particularly the case when they have succeeded (e.g., Davis & Stephan, 1980; Gilmor & Reid, 1979; Greenberg, Pyszczynski, & Solomon, 1982; Weiner, 1986). In the Shikanai study, Japanese college students performed an anagram task. Half of them were subsequently led to believe that they scored better than the average and thus "succeeded," whereas the other half were led to believe that they scored worse than the average and thus "failed" Subjects were then asked to choose the most important factor in explaining the success or the failure for each of 10 pairs made from the 5 possible causes for performance (i.e., ability, effort, task difficulty [or ease], luck, and mental-physical "shape" of the day). Shikanai analyzed the average number of times each cause was picked as most important (possible minimum of 0 and maximum of 4). As shown in Figure 5, a modesty bias was again obtained, especially after success. Whereas failure was attributed mainly to the lack of effort, success was attributed primarily to the ease of the task. Furthermore, the potential role of ability in explaining success was very much downplayed. Indeed, ability was perceived to be more important after a failure than after a success, whereas task difficulty (or its ease) was regarded to be more important after a success than after a failure. Subsequent studies by Shikanai that examined attribution of success and failure of others did not find this pattern (Shikanai, 1983, 1984). Thus, the pattern of "modest" appraisal seems to be specific to the perception and the presentation of the self and does not derive from a more general causal schema applicable to both self and others. For others, ability is important in explaining success. Yoshida et al. (1982, Studies 2 and 3), who studied explanations of performance in a Japanese elementary school, found the tendency to dc-emphasize the role of ability in explaining success as early as the second grade. Observations of a tendency to self-efface, and not to reveal the typical American pattern of blaming others or the situation when explaining failure, have been made outside of the experimental laboratory as well. In a study by Hess et al. (1986), Japanese mothers explained poor performance among their fifth graders by claiming a lack of effort. In marked contrast, American mothers implicated effort in their explanations but viewed ability and the quality of the training in the school as equally important. This study also required the children to explain their own poor performance by assigning 10 points to each of five alternatives (ability, effort, training at school, bad luck, and difficulty of math). Japanese children gave 5.6 points to lack of effort, but American children gave 1.98 points. H. Stevenson (personal communication, September 19, 1989) noted that in observations of elementary school classrooms, Japanese teachers, in contrast to American teachers, rarely refer to differences in ability among their students as an explanation for performance differences, even though the range of ability as assessed by standardized tests is approximately the same. Those with interdependent selves thus seem more likely to view intellectual achievement not as a fixed attribute that one has a certain amount of, but instead as a product that can be produced by individual effort in a given social context. The nature of modesty. The exact nature of these modesty, self-effacing, or other-enhancing biases has yet to be specified. Perhaps those from interdependent cultures have simply learned that humility is the desired response, or the culturally appropriate response, and that it is wise not to gloat over their performance or to express confidence in their ability. This interpretation implies that the modesty biases observed in the studies described herein are primarily the result of impression management and that the subjects involved actually could have held different, perhaps opposite, beliefs about themselves and their ability. However, it is also possible that these other-enhancement biases reflect, or are accompanied by, psychologically authentic self-perceptions. There are two related possibilities consistent with this suggestion. First, given the press not to stand out and to fit in, people in interdependent cultures may acquire through socialization a habitual modest-response tendency. In large part, it may be a function of the need to pay more attention to the other than to the self, just as the self-serving bias is believed to result from a predominant focus on the self (see Ross & Fletcher, 1985). Consequently for those with interdependent selves, whenever certain aspects of self need to be appraised in public, a modest, self-effacing pattern of responses may occur spontaneously. Furthermore, this modesty can be motivated by many other-focused emotions that are central to the construal of self as an interdependent entity. From an independent viewpoint, such modesty seems false and the result of suppressing a "natural" pride in one's attributes. Yet, such pride is only natural within a view of the self as an independent entity. From an interdependent view, modest responses may be experienced quite positively and engender the pleasant, other-focused feelings that are associated with connecting and maintaining interdependence. Such positive, other-focused feelings also may be responsible for the finding that Japanese students are more convinced of and more confident in their ability after failure than success. The satisfaction of doing well that can accompany good performance on a novel, decontextualized task may be mitigated by the threat of potential uniqueness and uncertainty over how to respond to it. Moreover, if a predominant basis of self-esteem is how well one fits in and preserves relationships and interpersonal harmony, then failing to distinguish oneself with a highly successful performance may not be particularly devastating. Of course, because those in Asian cultures believe high ability to be a result of effort does not mean that they do not differentiate between ability and effort. In all likelihood, they believe that effort and ability are related in a multiplicative fashion to determine performance. Thus, for instance, in a recent study by Stipek, Weiner, and Li (1989), Chinese respondents reasoned, just as their American counterparts did, that if a person shows the same level of performance with much less effort expended on the task, the person must have a high level of the relevant ability. Our point is simply that those in Asian cultures believe that abilities are relatively more changeable over a long span of time through the effort the person expends: Certainly it will not be as devastating as it is to the person whose self-esteem rests primarily on doing well individually and on separating oneself from others. Second, among those with interdependent selves, there may not be an awareness of one's own ability in general or in the abstract. Instead, one's own ability in a given task under a given condition may be inferred from whatever cues are available in the specific situation in which the task is performed. And whatever is inferred in this way may be experienced as authentic and genuine. For example, upon receipt of feedback about their ability, interdependent selves may first attend and think not so much about their ability as about the approval or disapproval of the person who gives the feedback. If approval or disapproval can be strongly and unambiguously inferred, then the perception of approval or disapproval may provide a strong heuristic clue about ability; if one receives approval, one must have high ability in this situation, whereas if one receives disapproval, then one must have low ability in this situation. In the absence of a strong, enduring belief about one's ability in the abstract, such a heuristic may provide a subjectively genuine self-appraisal. This analysis also suggests why those with interdependent selves may be convinced of their low ability after a failure feedback to a much greater extent than they are convinced of their high ability after a success feedback. Because of the prevalent social norms for polite behavior in interdependent cultures, disapproval can be more unequivocally inferred from negative feedback than approval can be inferred from positive feedback. These suggestions about the source of a modest self-appraisal have yet to be empirically tested, but they are worthy of careful inquiry because these forms of self-appraisal may be quite unique to interdependent cultures. On the basis of empirical evidence, however, this much seems clear Those with interdependent selves will typically not claim that they are better than others, will not express pleasure in the state of feeling superior to others, and indeed may not enjoy it. A strong, pervasive motive for self-enhancement through taking personal credit for success, denying personal responsibility for failure, and believing oneself to be better than average may be primarily a Western phenomenon. It is akin to being the nail that stands out. So far, the empirical evidence on cultural variation in self-related motives is limited largely to differences in self-enhancement versus other enhancement. However, other self-related motives, such as self-affirmation (Steele, 1988), self-verification (Swann & Read, 1981), and self-actualization (Maslow, 1954), may also differ across cultures in similar ways. A series of studies by Steele has shown that the negative psychological impact of one's own misdeed, blunder, or public embarrassment can be reduced once another, significant aspect of the self is activated and affirmed. Thus, one's threatened self-worth can be restored by a reminder of another, unthreatened aspect of the self (e.g, "I may not be athletic, but at least I'm creative"). To the extent that very different aspects of self are highly valued among those with interdependent selves, this process of self-affirmation may also differ. For those with independent selves it will be the internal attributes of self that may most effectively offset each other and reestablish threatened self-esteem, whereas for those with interdependent selves it may be the more public aspects of the self, like one's significant social roles, statuses, and important interpersonal relations, that must be focal in self-esteem maintenance. Thus, self-affirmation for an interdependent self will require an opportunity to ensure that one is fitting in and engaging in proper action in a given situation. In a similar vein, exactly what is verified in self-verification and what is actualized in self-actualization may also differ considerably across cultures. Currently, it is common to assume that individuals are motivated to verify and actualize an internally coherent set of attributes that they regard as significant. Our present analysis would imply, however, that people with interdependent selves may strive to verify and actualize the more public qualities of the self-the ones that allow them to conceive of themselves as respectable and decent participants in significant interpersonal relationships. Furthermore, among those with interdependent selves, selfverification and self-actualization may even be achieved through the realization of some more general, abstract forms of relation, that is, one's relationship to or one's role in society or even in the natural or cosmic system. The self-description studies reviewed earlier suggest this possibility. In general, the selfdescriptions of those with interdependent selves have been found to be quite concrete and situation specific (see Cousins, 1989). There is, however, one interesting, reliable exception to this. Subjects from Asian cultural backgrounds (presumably those with predominantly interdependent selves) often provide extremely global self-descriptions, such as "I am a unique creation," "I am a human being," "I am an organic form," and "I am a product of my environment?' It could appear that these statements are too abstract to be informative in any pragmatic sense (Rosch, 1978). The lack of information contained in these descriptions, however, may be more apparent than real. Note that these global statements presuppose a view of the world as an encompassing whole in which these subjects perceive themselves to be a part or a participant. And for these subjects, it may be these relationships that must be verified and actualized. We have suggested the different forms that some self-related motives might assume if they are based in an interdependent rather than an independent construal of self. Further empirical work is required to determine whether the types of self-related motives described herein are indeed as prevalent in Eastern interdependent cultures as they have been found to be in Western, particularly American, cultures. It could be that these selfrelevant motives are not part of the set of universal individual strivings,' but instead an outgrowth of an independent self-system rooted in the press for separation and individuation. Conclusions We have described two divergent construals of the self-an independent view and an interdependent view. The most significant differences between these two construals is in the role that is assigned to the other in self-definition. Others and the surrounding social context are important in both construals, but for the interdependent self, others are included within the boundaries of the self because relations with others in specific contexts are the defining features of the self. In the words of Lebra (1976), the individual is in some respects "a fraction" and becomes whole when fitting into or occupying one's proper place in a social unit. The sense of individuality that accompanies an interdependent self includes an attentiveness and responsiveness to others that one either explicitly or implicitly assumes will be reciprocated by these others, as well as the willful management of one's other-focused feelings and desires so as to maintain and further the reciprocal interpersonal relationship. One is conscious of where one belongs with respect to others and assumes a receptive stance toward these others, continually adjusting and accommodating to these others in many aspects of behavior (Azuma, 1984; Weisz et al, 1984). Such acts of fitting in and accommodating are often intrinsically rewarding, because they give rise to pleasant, other-focused emotions (e.g., feeling of connection) while diminishing unpleasant ones (e.g., shame) and, furthermore, because the self-restraint required in doing so forms an important basis of self-esteem. Typically, then, it is others rather than the self that serve as the referent for organizing one's experiences. With an independent construal of the self, others are less centrally implicated in one's current self-definition or identity. Certainly others are important for social comparison, for reflected appraisal, and in their role as the targets of one's actions, yet at any given moment, the self is assumed to be a complete, whole, autonomous entity, without the others. The defining features of an independent self are attributes, abilities, traits, desires, and motives that may have been social products but that have become the "property" of the self-contained individual (see Sampson, 1989) and that are assumed to be the source of the individual's behavior. The sense of individuality that accompanies this construal of the self includes a sense of oneself as an agent, as a producer of one's actions. One is conscious of being in control over the surrounding situation, and of the need to express one's own thoughts, feelings, and actions to others, and is relatively less conscious of the need to receive the thoughts, feelings, and actions of others. Such acts of standing out are often intrinsically rewarding because they elicit pleasant, egofocused emotions (e.g., pride) and also reduce unpleasant ones (e.g., frustration). Furthermore, the acts of standing out, themselves, form an important basis of self-esteem. The Role of the Self The relative importance that is accorded to others in these two construals has a wide range of psychological implications. In this article, we have outlined some of the cognitive, emotional, and motivational consequences of holding a view of the self that includes others and that requires others to define the self. Although a rapidly expanding volume of studies suggest that some aspects of cognitive functioning are relatively hardwired, many features of the way people perceive, categorize, or assign causality are probably not basic processes that derive in any straightforward way from the functioning of the human machinery or "hardware:' Rather, these processes are to a large extent personal, reflecting the nature of the self that anchors them. Thus, they reflect all of those factors, including cultural aspects, that jointly determine the self. If one perceives oneself as embedded within a larger context of which one is an interdependent part, it is likely that other objects or events win l perceived in a similar way. For example, a given event involving a particular actor will be perceived as arising from the sitd tional context of which this actor is an interdependent part, rather than as stemming solely from the attributes of the actor, Or, in answering any question, one's first tendency may be to consider the particular social situation that is defined by the current interaction (e.g., teacher-student, worker -co-worker, and younger-elder) and then to gauge the range of responses that are most appropriate to this situation. These construals of self are probably abstracted through early patterns of direct interactions with parents and peers. The way people initially, and thus thereafter, most naturally or effortlessly perceive aid understand the world is rooted in their self-perceptions and' self-understandings, understandings that are themselves constrained by the patterns of social interactions characteristic of the given culture. Consequences for Self-Processes Our discussion of the cognitive, emotional, or motivational consequences has by no means exhausted the range of potential consequences of holding an independent or interdependent construal of the self. Consider first the set of processes connected by a hyphen to the self. It is reasonable to assume that all of these phenomena (e.g., self-affirmation [Steele, 1988], self-verification [Swami, 19831, self-consciousness [Fenigstein, Scheier, & Buss, 19751, self-control [Carver & Scheier, 1981], self-actualization [Maslow, 1954], or self-handicapping [Jonà & Berglas, 1978]) could assume a somewhat different form depending on how interdependent the self is with others. . Self-esteem for those with an independent construal of the self depends on one's abilities, attributes, and achievements. The most widely used measure of self-esteem, the Rosenberg Self-Esteem Scale, requires the endorsement of items like "I am a person of worth" or "I am proud of my abilities:' Self-esteem associated with an interdependent self could include endorsement of similar items, although what it means to be, for example, a person of worth could well have a different meaning. Or high self-esteem may be more strongly associated with an dorsement of items that gauge one's ability to read the situation and to respond as required. If this is the case, a threat or a challenge to the self may not come in the form of feedback that one is unlike a cherished conception of the inner or dispositional Self (dumb instead of smart; submissive rather than dominant) but instead in terms of a threat of a disruption of, or a disconnection from, the relation or set of relations with whichone forms an interdependent whole. The focus on the distinction between independent versus interdependent selves has the potential to provide a means of integrating research on a large number of separate personaldY constructs. One of the significant distinctions that appears re, peatedly throughout Western psychology reflects a variation among individuals in how tuned in, sensitive to, oriented toward, focused on, or concerned they are with others. The introversion-extraversion dimension reflects this difference, as does the inner-directed-outer-directed distinction (Reisman, Den- 1 ney, & Glazer, 1950). Other related distinctions include high versus low self-monitoring (Snyder, 1979), personal identity versus social identity (Cheek, 1989; Hogan, 1975), public versus private self-consciousness (Fenigstein, 1984), social orientation versus individual orientation (Greenwald, 1980), collectivist-mind ividualism (Hui, 1988; Triandis, 1989), and field independence-field dependence (Witkin & Goodenough, 1977). In fact, Witkin and his colleagues described a field-dependent person as one who includes others within the boundaries of the self and who does not make a sharp distinction between the self and others. Many of the empirical findings (described in Witkin & Goodenough, 1977; Witkin, Goodenough, & Oltman, 1979) about the interpersonal expertise and sensitivities of field-dependent people are similar to those described herein for people with interdependent selves. Consequences for Social Psychological Phenomena Other social behaviors may also depend on one's mediating model of the self (see Triandis, 1989, for a recent analysis of some of these effects). Thus, for one with an interdependent self, conformity may not reflect an inability to resist social pressure and to stick by one's own perceptions, attitudes, or beliefs (the defining features of the self). Instead, conformity to particular others with whom the other is interdependent can be a highly valued end state. It can signify a willingness to be responsive to others and to adjust one's own demands and desires so as to maintain the ever-important relation. The conformity observed for these subjects with interdependent selves when surrounded with others who form part of an important social unit, could well be much higher than typically observed. However, conformity to the desires and demands of those outside the important social unit or the self-defining in-group may not be required at all. Thus, for those with interdependent selves, a typical Aschtype conformity paradigm involving subjects and strangers as confederates may result in less conformity than typically observed in American studies. Studies of other phenomena such as social facilitation or social loafing could also produce differential effects, depending on the self-systems of the subjects. Should those with interdependent construals of the self show pronounced social facilitation compared with those with individual selves? Or should those with interdependent selves be less susceptible to social loafing (decrements in performance when one's individual contribution to the group product cannot be identified; see Harkins, Latané, & Williams, 1980)? Our analysis is also relevant to two of the central problems in Western psychology-the inconsistency between attitudes and behavior and the inconsistency between personality and behavior. As we have noted, interdependent selves do not prescribe or require such a consistency between one's internal attributes and one's actions. Consequently, the press for consistency should be much less important and much less bemoaned when not observed. In fact, consistency from an interdependent perspective may reflect a lack of flexibility, insensitivity to the context, rigidity, or immaturity. Further analysis of the consequences of different construals of the self may also prove fruitful in understanding some basic social psychological questions. Social psychologists report that people are enormously influenced by others, often to an extent that the investigators and certainly individuals themselves, find unbelievable. People conform, obey, diffuse responsibility in a group, allow themselves to be easily persuaded about all manner of things, and become hopelessly committed to others on the basis of minimal action (e.g., see Myers, 1989). Even within highly individualist Western culture, most people are still much less self-reliant, self-contained, or self-sufficient than the prevailing cultural ideology suggests that they should be. Perhaps Western models of the self are quite at odds with actual individual social behavior and should be reformulated to reflect the substantial interdependence that characterizes even Western individualists. Sampson (1989) has recently argued that the reality of globalization and a shrinking world will force just such a rethinking of the nature of the individual. C onstruals of the Self and Gender Many important gender differences may also be linked to divergent construals of the self. Recent feminist theory on empathy suggests that relations have a power and a significance in women's lives that have gone largely unrecognized (e.g., Belenky, Clinchy, Goldberger, & Taruie, 1986; Jordan & Surrey, 1986; J. B. Miller, 1986; Stewart & Lykes, 1985). An awareness of and sensitivity to others is described as one of most significant features of the psychology of women. If this is the case, then selfesteem and self-validation should depend not only on being able to do a job well, but on fostering and sustaining relationships. As Gilligan (1986) claimed, a willingness and an ability to care are standards of self-evaluation for many women. This theoretical work is forging a new vision of dependence, one that is similar in many ways to some Eastern views. Being dependent does not invariably mean being helpless, powerless, or without control. It often means being interdependent. It thus signifies a conviction that one is able to have an effect on others and is willing to be responsive to others and to become engaged with them. In other words, there is an alternative to selfishness (which implies the exclusion of others) besides selflessness (which is to imply the exclusion of the self or self-sacrifice): There is a self defined in relationship to others (see Chodorow, 1978; Gilligan, 1982; Markus & Oyserman, 1988). Difficult Questions Carrying out the research necessary to systematically investigate the range of basic consequences of having one or another construal of the self raises several complex questions. Some of these we have only touched on. For example, a persistent issue is how deep or pervasive are these cultural differences? Are the observed differences primarily a reflection of differences in styles of behavioral expression, or do they also reflect differences in the phenomenology accompanying the behavior? If there are norms against the display or expression of anger, what happens to the nature of the felt anger! In other words, is it the case, as we suggest here, that these norms can sometimes be internalized to the extent that they determine the nature of one's experience? For example, a recent study by Bontempo, Label, and Triandis (1989) compared the public and private responses of individuals from a collectivist culture with those of individuals from an individualist culture. The researchers asked respondents to indicate how enjoyable it would be to engage in a time-consuming, individually costly behavior such as visiting a friend in the hospital. Only in the public condition did individualists claim that the behavior would be enjoyable. The collectivists, in contrast, claimed that the behavior would be enjoyable even when their responses were private. The view that altruistic behaviors are only seemingly altruistic and that they are public actions without any subjective, private foundation can perhaps be traced to the insistence of Western psychologists on the internal attributes (feeling, thought, and traits) as the universal referents for behavior. They have thus understandably failed to attend to the possibility of the other as a referent for behavior, and thus to the possibility of other-focused emotions. There is, however, the possibility that such emotions can motivate genuine, other-oriented, altruistic behaviors, without any conscious, or even unconscious, calculation of individual payoff, and as such serve as the important glue of interdependent relationships. Another thorny issue centers on the assessment of cultural differences. The use of introspective reports, for example, which are typically quite useful in the study of cognition, emotion, and motivation, may be problematic in cross-cultural research because within a given cultural context, people have little access to the absolute extent of their attention or responsiveness to others. This may explain, for example, why Triandis et al. (1988) found that those with collective selves do not report a greater than average awareness of or concern for the demands of others. Another persistent issue is that of translation and equating stimuli and questionnaires. Can psychologists readily assume that when an American and a Japanese use the word embarrass it indicates a similar emotional experience? Can they hypothesize, for example, that those with interdependent selves should show more high self-monitoring (i.e., attention to the behavior of others) than those with independent selves, and then assume that a translation ofSnyder's (1979) scale into Japanese or Chinese will be sufficient to reflect these differences? One may even ask to what extent a construct such as self-monitoring can be unequivocally defined -across different cultures with remarkably different construals of self. In sum, we have argued that the view one holds of the self is critical in understanding individual behavior and also in understanding the full nature of those phenomena that implicate the self. A failure to replicate certain findings in different cultural contexts should not lead to immediate despair over the lack of generality of various psychological principles or to the conclusion of some anthropologists that culturally divergent individuals inhabit incomparably different worlds. Instead, it is necessary to identify the theoretical elements or processes that explain these differences. We suggest that how the self is construed may be one such powerful theoretical element.