Hazel Rose Markus and Shinobu Kitayama, "Culture and the Self: Implications for Cognition, Emotion, and Motivation"

In America, "the squeaky wheel gets the grease." In Japan, "the nail that stands out gets pounded down." American parents who are trying to induce their children to eat their suppers are fond of saying "think of the starving kids in Ethiopia, and appreciate how lucky you are to be different from them" Japanese parents are likely to say "Think about the farmer who worked so hard to produce this rice for you; if you don't eat it, he will feel bad, for his efforts will have been in vain" (H. Yamada, February 16, 1989). A small Texas corporation seeking to elevate productivity told its employees to look in the mirror and say "I am beautiful" 100 times before coming to work each day. Employees of a Japanese supermarket that was recently opened in New Jersey were instructed to begin the day by holding hands and telling each other that "he" or "she is beautiful" ("A Japanese Supermarket." 1989).

Such anecdotes suggest that people in Japan and America may hold strikingly divergent construals of the self, others, and the interdependence of the two. The American examples stress attending to the self, the appreciation of one's difference from others, and the importanc eof asserting the self. The Japanese examples emphasize attending to and fitting in with others and the importance of harmonious interdependence with them. These construals of the self and others are tied to the implicit, normative tasks that various cultures hold for what people should be doing in their lives (cf. Cantor & Kihlstrom, 1987; Erikson, 1950; Veroff, 1983). Anthropologists and psychologists assume that such construals can influence, and in many cases determine, the very nature of individual experience.

Despite the growing body of psychological and anthropological evidence that people hold divergent views about the self, most of what psychologists currently know about human nature is based on one particular view-the so-called Western view of the individual as an independent, self-contained, autonomous entity who (a) comprises a unique configuration of internal attributes (e.g., traits, abilities, motives, and values) and (b) behaves primarily as a consequence of these internal attributes (Geertz, 1975; Sampson, 1988,1989; Shweder & LeVine, 1984). As a result of this monocultural approach to the self (see Kennedy, Scheier, & Rogers, 1984), psychologists' understanding of those phenomena that are linked in one way or another to the self may be unnecessarily restricted (for some important exceptions, see Bond, 1986,1988; Cousins, 1989; Fiske, in press; Maehr & Nicholls, 1980; Stevenson, Azuma, & Flakuta, 1986; Triandis, 1989; Triandis, Bontempo, Villareal, Asai, & Lucca, 1988). In this article, we suggest that construals of the self, of others, and of the relationship between the self and others may be even more powerful than previously suggested and that their influence is clearly reflected in differences among cultures. In particular, we compare an independent view of the self with one other, very different view, an interdependent view. The independent view is most clearly exemplified in some sizable segment of American culture, as well as in many Western European cultures. The interdependent view is exemplified in Japanese culture as well as in other Asian cultures. But it is also characteristic of African cultures, Latin-American cultures, and many southern European cultures. We delineate how these divergent views of the self-the independent and the interdependentcan have a systematic influence on various aspects of cognition, emotion, and motivation.

We suggest that for many cultures of the world, the Western notion of the self as an entity containing significant dispositional attributes, and as detached from context, is simply not an adequate description of selfhood. Rather, in many construals, the self is viewed as interdependent with the surrounding context, and it is the "other" or the "self-in-relation-to-other" that is focal in individual experience. One general consequence of this divergence in self-construal is that when psychological processes (e cognition, emotion, and motivation) explicitly, or even quite implicitly, implicate the self asa target or as a refer ent, the nature of these processes will vary according to the exact form or organization of self inherent in a given construal. With respect to cognition, for example, for those with interdependent selves, in contrast to those with independent selves, some aspects of knowledge representation and some of the processes involved in social and nonsocial thinking alike are influenced by a pervasive attentiveness to the relevant others in the social context. Thus, one's actions s are more like likely to be seen as situationally bound, and characterizations of the individual will include this context Furthermore, for those with interdependent construals of the self, both the expression and the experience of emotions and motives may be significantly shaped and governed by a consideration of the reactions of others. Specifically, for example, some emotions, like anger, that derive from and promote an independent view of the self may be less prevalent among those with interdependent selves, and self-serving motives may be replaced by what appear as other-serving motives. An examination of cultural variation in some aspects of cognition, emotion, and motivation will allow psychologists to ask exactly what is universal in these processes, and it has the potential to provide some new insights for theories of these psychological processes.

In this analysis, we draw on recent research efforts devoted to characterizing the general differences between American or Western views of personhood and Eastern or Asian perspectives (e.g., Heelas & Lock, 1981; Hofstede, 1980; Marsella et al., 1985; Roland, 1988; Schwartz & Bilsky, 1990; Shweder, 1990; Shweder & LeVine, 1984; Stigler, Shweder, & Herdt, 1990; Triandis, 1989; Triandis & Brislin, 1980; Weisz et al., 1984). We extract from these descriptions many important differences that may exist in the specific content, structure, and functioning of the self-systems of people of different cultural backgrounds. The distinctions that we make between independent and interdependent construals must be regarded as general tendencies that may emerge when the members of the culture are considered as a whole. The prototypical American view of the self, for example, may prove to be most characteristic of White, middle-class men with a Western European ethnic background. It may be somewhat less descriptive of women in general, or of men and women from other ethnic groups or social classes.' Moreover, we realize that there may well he important distinctions among those views we discuss as similar and that there may be views of the self and others that cannot easily be classified as either independent or interdependent.

Our intention is not to catalog all types of self-construals, but rather to highlight a view of the self that is often assumed to be universal but that may be quite specific to some segments of Western culture. We argue that self-construals play a major role in regulating various psychological processes. Understanding the nature of divergent self-construals has two important consequences. On the one hand, it allows us to organize several apparently inconsistent empirical findings and to pose questions about the universality assumed for many aspects of cognition, emotion, and motivation (see Shweder, 1990). On the other hand, it permits us to better specify the precise role of the self in mediating and regulating behavior.

The Self: A Delicate Category

Universal Aspects of the Self

In exploring the possibility of different types of self-construals, we begin with Hallowell's (1955) notion that people everywhere are likely to develop an understanding of themselves as physically distinct and separable from others. Head (1920), for example, claimed the existence of a universal schema of the body that provided one with an anchor in time and space. Similarly, Allport (1937) suggested that there must exist an aspect of personality that allows one, when awakening each morning, to be sure that he or she is the same person who went to sleep the night before. Most recently, Neisser (1988) referred to this aspect of self as the ecological self, which he defined as "the self as perceived with respect to the physical environment: 'I' am the person here in this place, engaged in this particular activity" (p. 3). Beyond a physical or ecological sense of self, each person probably has some awareness of internal activity, such as dreams, and of the continuous flow of thoughts and feelings, which are private to the extent that they cannot be directly known by others. The awareness of this unshared experience will lead the person to some sense of an inner, private self.

Divergent Aspects of the Self

Some understanding and some representation of the private, inner aspects of the self may well be universal, but many other aspects of the self may be quite specific to particular cultures. People are capable of believing an astonishing variety of things about themselves (cf. Heelas & Lock, 1981; Marsella et al., 1985; Shweder & LeVine, 1984; Triandis, 1989). The self can be con strued, framed, or conceptually represented in multiple ways. A cross-cultural survey of the self lends support to Durkheim's (1912/1968) early notion that the category of the self is primarily the product of social factors, and to Mauss's (1938/1985) claim that as a social category, the self is a "delicate" one, subject to quite substantial, if not infinite, variation.

The exact content and structure of the inner self may differ considerably by culture. Furthermore, the nature of the outer or public self that derives from one's relations with other people and social institutions may also vary markedly by culture. And, as suggested by Triandis (1989), the significance assigned to the private, inner aspects versus the public, relational aspects in regulating behavior will vary accordingly. In fact, it may not be unreasonable to suppose, as did numerous earlier anthropologists (see Allen, 1985), that in some cultures, on certain occasions, the
individual in the sense of a set of significant inner attributes of thePerson, may cease to be the primary unit of consciousness. instead, the sense of belongingness to a social relation may become so strong that it makes better sense to think of the relationship as the functional unit of conscious reflection.

The current analysis focuses on just one variation in what people in different cultures can come to believe about themselves. This one variation concerns what they believe about the relationship between the self and others and, especially, the degree to which they see themselves as separate from others or as connected with others. We suggest that the significance and the exact functional role that the person assigns to the other when defining the self depend on the culturally shared assumptions about the separation or connectedness between the self and others.

Two Construals of the Self: Independent

and Interdependent

The Independent Construal

In many Western cultures, there is a faith in the inherent separateness of distinct persons. The normative imperative of this culture is to become independent from others and to discover and express one's unique attributes (Johnson, 1985; Marsella et al, 1985; J. G. Miller, 1988; Shweder & Bourne, 1984). Achieving the cultural goal of independence requires construing oneself as an individual whose behavior is organized and made meaningful primarily by reference to one's own internal repertoire of thoughts, feelings, and action, rather than by reference to the thoughts, feelings, and actions of others. According to this construal of self, to borrow Geertz's (1975) often quoted phrase, the person is viewed as "a bounded, unique, more or less integrated motivational and cognitive universe, a dynamic center of awareness, emotion, judgment, and action organized into a distinctive whole and set contrastively both against other such wholes and against a social and natural background" (p. 48).

This view of the self derives from a belief in the wholeness and uniqueness of each person's configuration of internal attributes (Johnson, 1985; Sampson, 1985, 1988, 1989; Waterman, 1981). It gives rise to processes like "self-actualization," "realizing oneself," "expressing one's unique configuration of needs, rights, and capacities," or "developing one's distinct potential?' The essential aspect of this view involves a conception of the self as an autonomous, independent person; we thus refer to it as the

independent construal of the self Other similar labels include individualist, egocentric, separate, autonomous, idiocentric, and self-contained. We assume that, on average, relatively more individuals in Western cultures will hold this view than will individuals in non-Western cultures. Within a given culture, however, individuals will vary in the extent to which they are good cultural representatives and construe the self in the mandated way.

The independent self must, of course, be responsive to the social environment (Fiske, in press). This responsiveness, however, is fostered not so much for the sake of the responsiveness itself. Rather, social responsiveness often, if not always, derives from the need to strategically determine the best way to express or assert the internal attributes of the self. Others, or the social situation in general, are important, but primarily as standards of reflected appraisal, or as sources that can verify and affirm the inner core of the self.

The Interdependent Construal

In contrast, many non-Western cultures insist, in Kondo's (1982) terms, on the fundamental connectedness of human beings to each other. A normative imperative of these cultures is to maintain this interdependence among individuals (De Vos, 1985; Hsu, 1985; Miller, 1988;Shweder& Bourne, 1984). Experiencing interdependence entails seeing oneself as part of an encompassing social relationship and recognizing that one's behavior is determined, contingent on, and, to a large extent organized by what the actor perceives to be the thoughts, feelings, and actions of others in the relationship. The Japanese experience of the self, therefore, includes a sense of interdependence and of one's status as a participant in a larger social unit (Sampson, 1988). Within such a Construal, the self becomes most meaningful and complete when it is cast in the appropriate social relationship. According to Lebra (1976) the Japanese are most fully human in the context of others.

This view of the self and the relationship between the self and others features the person not as separate from the social context but as more connected and less differentiated from others. People are motivated to find a way to fit in with relevant others, to fulfill and create obligation, and in general to become part of various interpersonal relationships. Unlike the independent self, the significant features of the self according to this construal are to be found in the interdependent and thus, in the more public components of the self. We therefore call this view the interdependent Construal of the self The same notion has been variously referred to, with somewhat different connotations, as sociocentric, holistic, collective, allocentric, ensembled, constituitive con contextualist, connected, and relational. As with the independent self, others are critical for social comparison and self-validation, yet in an interdependent formulation of the self, these others become an integral part of the setting, situation, or context to which the self is connected, fitted, and assimilated. The exact manner in which one achieves the task of connection therefore, depends crucially on the nature of the context, particularly the others present in the context. Others thus participate actively and continuously in the definition of the interdependent self.

The interdependent self also possesses and expresses a set of internal attributes, such as abilities, opinions, judgments, and personality characteristics. However, these internal attributes are understood as situation specific, and thus as sometimes elusive and unreliable. And, as such, they are unlikely to assume a powerful role in regulating overt behavior, especially if this behavior implicates significant others. In many domains of social life, one's opinions, abilities, and characteristics are assigned only secondary roles-they must instead be constantly controlled and regulated to come to terms with the primary task of interdependence. Such voluntary control of the inner attributes constitutes the core of the cultural ideal of becoming mature. The understand of one's autonomy as secondary to, and constrained by, the primary task of interdependence distinguishes interdependent selves from independent selves for whom autonomy and its expression is often afforded primary significance. An independent behavior (e.g., asserting an opinion) exhibited by a person in an interdependent culture is likely to be based on the premise of underlying interdependence and thus may have a somewhat different significance than it has for a person from an independent culture... The fundamental units of the self-system, the core conceptions, or self-schemata are thus predicated on significant interpersonal relationships.

An interdependent self cannot be properly characterized as a bounded whole, for it changes structure with the nature of the particular social context. Within each particular social situation, the self can be differently instantiated. The uniqueness of such a self derives from the specific configuration of relationships that each person has developed. What is focal and objectified in an interdependent self, then, is not the inner self, but the relationships of the person to other actors (Hamaguchi 1985)

The notion of an interdependent self is linked with a monistic philosophical tradition in which the person is thought to be of the same substance as the rest of nature (see Bond, 1986; Phillips, 1976; Roland, 1988; Sass, 1988). As a consequence, the relationship between the self and other, or between subject and object, is assumed to be much closer. Thus, many non-Western cultures insist on the inseparability of basic elements (Galtung, 1981), including self and other, and person and situation. In Chinese culture, for instance, there is an emphasis on synthesizing the constituent parts of any problem or situation into an integrated or harmonious whole (Moore, 1967; Northrop, 1946). Thus, persons are only parts that when separated from the larger social whole cannot be fully understood (Phillips, 1976; Shweder, 1984). Such a holistic view is in opposition to the Cartesian, dualistic tradition that characterizes Western thinking and in which the self is separated from the object and from the natural world.

Examples of the interdependent self An interdependent view of the self is common to many of the otherwise highly diverse cultures of the world. Studies of the mainland Chinese, for example, summarized in a recent book by Bond (1986), show that even among the most rapidly modernizing segments of the Chinese population, there is a tendency for people to act primarily in accordance with the anticipated expectations of others and social norms rather than with internal wishes or personal attributes (Yang, 198 I b). A premium is placed on emphasizing collective welfare and on showing a sympathetic concern for others. Throughout the studies of the Chinese reported by Bond, one can see the clear imprint of the Confucian emphasis on interrelatedness and kindness. According to Hsu (1985), the supreme Chinese virtue, lenj implies the person's capability to interact with fellow human beings in a sincere, polite, and decent fashion (see also Elvin, 1985).

Numerous other examples of cultures in which people are likely to have some version of an interdependent self can also be identified. For example, Triandis, Mann, Lisansky and Betancourt (1984) have described the importance of simpatico among Hispanics. This quality refers to the ability to both respect and share others' feelings. In characterizing the psychology of Filipinos, Church (1987) described the importance that people attribute to smooth interpersonal relations and to being "agreeable even under difficult circumstances, sensitive to what others are feeling and willing to adjust one's behavior accordingly" Similarly Weisz (in press) reported that Thais place a premium on self-effacement, humility, deference, and on trying to avoid disturbing others. Among the Japanese, it is similarly crucial not to disturb the Wa, or the harmonious ebb and flow of interpersonal relations (see also Geertz, 1974, for characterizations of similar imperatives among the Balinese and Moroccans).

Beattie (1980) claimed that Africans are also extremely sensitive to the interdependencies among people and view the world and others in it as extensions of one another. The self is viewed not as a hedged closure but as an open field. Similarly Marriott (19 76) argued that Hjndu conceptions assume that the self is an open entity that is given shape by the social, context. In his insightful book, Kakar (1978) described the Hindu's ideal of interpersonal fusion and how it is accompanied by a personal, cultural sense of hell, which is separation from others. In fact, Miller, Bersoff, and Harwood (1990), in a recent, carefully controlled study on moral reasoning, found that Indians regard responsiveness to the needs of others as an objective moral obligation to a far greater extent than do Americans. Although the self-systems of people from these cultures are markedly different in many other important respects, they appear to be alike in the greater value (when compared with Americans) that is attached to proper relations with others, and in the requirement to flexibly change one's own behavior in accordance with the nature of the relationship.

Even in American culture, there is a strong theme of interdependence that is reflected in the values and activities of many of its subcultures. Religious groups, such as the Quakers, explicitly value and promote interdependence, as do many small towns and rural communities (e.g., Bellah, Madsen, Sullivan, Swidler, & Tipton, 1985). Some notion of a more connected, ensembled, interdependent self, as opposed to a self-contained, independent self, is also being developed by several of what Sampson (1989) calls "postmodern" theorists. These theorists are questioning the sovereignty of the American view of the mature person as autonomous, self-determined, and unencumbered. They argue that psychology is currently dominated by a view of the person that does not adequately reflect the extent to which people everywhere are created by, constrained by, and respon sive to their various interpersonal contexts (see Gergen & gen, 1988; Gilligan, 1982; Miller, 1986; Tajfel, 1984).

Further definition of the interdependent self

Theorists of Japanese culture are beginning to characterize the interdependent self much more specifically than was previously attempted. These descriptions offer some more refined ideas of how an interdependent view of self can depart markedly from an inde.. pendent view of self (see Nakane, 1970; Plath, 1980; R. J. Smith, 1983). For example, building on a study of L. T. Doi (1973); Bachnik (1986) wrote:

(in Japanese society) rather than there being a single social reality, a number of possible perspectives of both self and social life are acknowledged. Interaction in Japanese society then focuses on the definition of the appropriate choice, out of all the various possibilities. This means that what one says and does will be different in different situations, depending on how one defines one's particular perspective versus the social other. (p. 69)

In Japan, the word for jibun, refers to "one's share of the shared life space" (Hamaguchi 1985). The self, Kimura (cited in Hamaguchi, 1985) claimed, is "neither a substance nor an attribute having a constant oneness" (p. 302). According to Hamaguchi (1985), for the Japanese, "a sense of identification with others (sometimes including conflict) pre-exists and selfness is confirmed only through interpersonal relationships
.... Self-ness is not a constant like the ego but denotes a fluid concept which changes through time and situations according to interpersonal relationships" (p. 302).

The Japanese anthropologist Lebra (1976) defined the essence of Japanese culture as an "ethos of social relativism" This translates into a constant concern for belongingness, reliance, dependency, empathy, occupying one's proper place, and reciprocity. She claimed the Japanese nightmare is exclusion, meaning that one is failing at the normative goal of connecting to others. This is in sharp contrast to the American nightmare, which is to fail at separating from others, as can occur when one is unduly influenced by others, or does not stand up for what one believes, or when one goes unnoticed or undistinguished.

An interdependent view of self does not result in a merging of self and other, nor does it imply that one must always be in the company of others to function effectively, or that people do not have a sense of themselves as agents who are the origins of their own actions. On the contrary, it takes a high degree of self-control and agency to effectively adjust oneself to various interpersonal contingencies. Agentic exercise of control, however, is directed primarily to the inside and to those inner attributes, such as desires, personal goals, and private emotions, that can such disturb the harmonious equilibrium of interpersonal transaction. This can be contrasted with the Western notion of control, which primarily implies an assertion of the inner attributes and a consequent attempt to change the outer aspects, such as one's public behaviors and the social situation (see also Weisz et al., J984).

Given the Japanese notion of control that is inwardly directed, the ability to effectively adjust in the interpersonal domain may form an important basis of self-esteem, and individualized styles of such adjustment to social contingencies may contribute to the sense of self-uniqueness. Thus, Hamaguchi (1985). for example, reported that for the Japanese, "the straightforward claim of the naked ego" (p. 303) is experienced as childish. Self-assertion is not viewed as being authentic, but instead immature. This point is M. White and LeVine's (1986) description of the meaning of sunao, a term used by Japanese parents to characterize what they value in their children:

A child that is sunao has not yielded his or her personal autonomy for the sake of cooperation; cooperation does not suggest giving up the self, as it may in the West; it implies that working with others is the appropriate way of expressing and enhancing the self. Engagement and harmony with others is, then, a positively valued goal and the bridge-to open-hearted cooperation, as in sunaois through sensitivity, reiterated by the mother's example and encouragement. (p. 58)

Kumagai (1981) said sunao "assumes cooperation to be an act of affirmation of the self" (p. 261). Giving in is not a sign of weakness; rather, it reflects tolerance, self-control, flexibility, and maturity.

The role of the other in the interdependent self

In an interdependent view, in contrast to an independent view, others will be assigned much more importance, will carry more weight, and will be relatively focal in one's own behavior. There are several direct consequences of an interdependent construal of the self. First, relationships, rather than being means for realizing various individual goals, will often be ends in and of themselves. Although people everywhere must maintain some relatedness with others, an appreciation and a need for people will be more important for those with an interdependent self than for those with an independent self. Second, maintaining a connection to others will mean being constantly aware of others and focusing on their needs, desires, and goals. In some cases, the goals of others may become so focal in consciousness that the goals of others may be experienced as personal goals. In other cases, fulfilling one's own goals may be quite distinct from those of others, but meeting another's goals, needs, and desires will be a necessary requirement for satisfying one's own goals, needs, and desires. The assumption is that while promoting the goals of others, one's own goals will be attended to by the person with whom one is interdependent. Hence, people may actively work to fulfill the others' goals while passively monitoring the reciprocal contributions from these others for one's own goalfulfillment. Yamagishi (1988), in fact, suggested that the Japanese feel extremely uncomfortable, much more so than Americans, when the opportunity for such passive monitoring of others' actions is denied.

From the standpoint of an independent, "self-ish" self, one might be led to romanticize the interdependent self, who is ever attuned to the concerns of others. Yet in many cases, responsive and cooperative actions are exercised only when there is a reasonable assurance of the "good-intentions" of others, namely their commitment to continue to engage in reciprocal interaction and mutual support. Clearly, interdependent selves do not attend to the needs, desires, and goals of all others. Attention to others is not indiscriminate; it is highly selective will be most characteristic o relationships with "in-group" members. These are others with whom one shares éi common fate such as family members or members of the same lasting social group, such as the work group. Out-group members are typically treated quite differently and are unlikely to experience either the advantages or disadvantages of interdependence. Independent selves are also selective in their association with others but not to the extent of interdependent selves because much less of their behavior is directly contingent on the actions of others: Given the importance of others in constructing reality and regulating behavior, the in-group-out-group distinction is a vital one for interdependent selves and the subject eboudaiy of one's "in-group" may tend to be narrower for the interdepen dent selves than for the independent selves (Triandis, 1989).

To illustrate the reciprocal nature of interaction among thosewith interdependent views, imagine that one has a friend over for lunch and has decided to make a sandwich for him. The conversation might be: "Hey, Tom, what do you want in your sandwich? I have turkey, salami, and cheese?' Tom responds, "Oh, I like turkey." Note that the friend is given a choice because the host assumes that friend has a right, if not a duty, to make a choice reflecting his inner attributes, such as preferences or desires. And the friend makes his choice exactly because of the
belief in the same assumption. This script is "natural," however, only within the independent view of self. What would happen if the friend were a visitor from Japan? A likely response to the question "Hey, Tomio, what do you want?" would be a little moment of bewilderment and then a noncommital utterance like "I don't know" This happens because under the assumptions of an interdependent self, it is the responsibility of the host to be able to "read" the mind of the friend and offer what the host perceives to be the best for the friend. And the duty of the guest, on the other hand, is to receive the favor with grace and to be prepared to return the favor in the near future, if not right
at the next moment. A likely, interdependent script for the same situation would be: "Hey, Tomio, I made you a turkey sandwich because I remember that last week you said you like turkey more than beef. And Tomio will respond, "Oh, thank you, I really like turkey" -

The reciprocal interdependence with others that is the sign of the interdependent self seems to require constant engagement of what Mead (1934) meant by taking the role of the other. It involves the willingness and ability to feel and think what others are feeling and thinking, to absorb this information without) being told, and then to help others satisfy their wishes and realize their goals. Maintaining connection requires inhibiting the"!" perspective and processing instead from the "thou" perspective (Hsu, 1981). The requirement is to "read" the other's mind and thus to know what the other is thinking or feeling. In contrast, with an independent self, it is the individual's responsibility to "say what's on one's mind" if one expects to be attended to or understood.

Consequences of an Independent or an Interdependent

View of the Self


In the current analysis, we hypothesize that the independent versus interdependent construals of self are among the most general and overarching schemata of the individual's self-system. These construals recruit and organize the more specific self-regulatory schemata.' We are suggesting here, therefore, that the exact organization of many self-relevant processes and their outcomes depends crucially on whether these processes are rooted in an independent construal of the self or whether they are based primarily on an interdependent construal of the self. For example, in the process of lending meaning and coherence to the social world, we know that people will show a heightened sensitivity to self-relevant stimuli. For those with an independent view of self, this includes information relevant to one's self-defining attributes. For one with an interdependent view of self, such stimuli would include information about significant others with whom the person has a relationship or information about the self in relation to another person.

Affect regulation involves seeking positive states and avoiding negative ones. Positive states are those that enhance or promote one's view of the self, and negative states are those that challenge this view. For a person with an independent view of self, this involves seeking information that confirms or enhances one's internal, private attributes. The most desirable situations are those that allow one to verify and express those important internal attributes and that convey the sense that one is appropriately autonomous. In contrast, for a person with an interdependent view of self, one might expect the most desirable states to be those that allow one to be responsive to one's immediate context or that convey the sense that one is succeeding in his or her interdependent relationships or statuses.

A third importnat function of the self-concept suggested by Markus and Wurf (1987) is that of motivating persons, of moving them to action. The person with an independent view of self should be motivated to those actions that allow expression of one's important self-defining, inner attributes (e.g., hardworking, caring, independent, and powerful), whereas the person with an interdependent view of self should be motivated to those actions that enhance or foster one's relatedness or connec tion to others. On the surface, such actions could look remarkably similar (e.g., working incredibly hard to gain admission to a desirable college), but the exact source, or etiology, of the energizing motivation may be powerfully different (De Vos, 1973; Maehr & Nicholls, 1980).

In the following sections, we discuss these ideas in further detail and review the empirical literature, which suggests that there are significant cognitive, emotional, and motivational consequences of holding an independent or an interdependent view of the self.

Consequences for Cognition

If a cognitive activity implicates the self, the outcome of this activity will depend on the nature of the self-system. Specifically, there are three important consequences of these divergent self-systems for cognition. First we may expect those with interdepcpdcnt selves to be more attentive and sensitive to others than those with independent selves. The attentiveness and sensitvity to others, characterizing the interdependent selves, will result in a relatively greater cognitive elaboration of the other or of the self-in-relation-to-other. Second, among those with interdependent selves, the unit of re sentation of both the self and the other will include a relatively specific social context in which the self and the other are embedded. This means that knowledge about persons, either the self or others, will not be abstract and generalized across contexts, but instead will remain specific to the focal context. Third, a consideration of the social context and the reactions of others may also shape some basic, nonsocial cognitive activities such as categorizing and counterfactual thinking.

In exploring the impact of divergent cultural construals on thinking, we assume that how people think (the process) in a social situation cannot be easily separated from what they think about (the content; Shweder, 1990; Shweder & Bourne, 1984). Extensive research on social cognition in the past decade has suggested the power of content in social inference (e.g., see Fiske & Taylor, 1984; Markus & Zajonc, 1985, for reviews). It is the nature of the representation (e.g., self, another person, a weed, or clam chowder) that guides attention, and that determines what other relevant information will be retrieved to fill in the gap of available sense data. For example, investigations by D)ndrade (1981) and Johnson-Laird (1983) indicate that the greater the familiarity with the stimulus materials, the more elaborate the schemata for framing the problem, and the better the problem solving. In general, then, how a given object is culturally construed and represented in memory should importantly influence and even determine how one thinks about the object. Accordingly, the divergent representations of the self we describe should be expected to have various consequences for all cognition relevant to self, others, or social relationships.

More interpersonal knowledge. If the most significant elements of the interdependent self are the self-in-relation-toothers elements, there will be a need, as well as a strong normative demand, for knowing and understanding the social surrounding, particularly others in direct interaction with the self. That is, if people conceive of themselves as interdependent parts of larger social wholes, it is important for them to be sensitive to and knowledgeable about the others who are the coparticipants in various relationships, and about the social situations that enable these relationships. Maintaining one's relationships and ensuring a harmonious social interaction requires a full understanding of these others, that is, knowing how they are feeling, thinking, and likely to act in the context of one's relationships to them. It follows that those with interdependent selves may develop a dense and richly elaborated store of information about others or of the self in relation.

Kitayama, Markus, Tummala, Kzrokawa, and Kato (1990) examined this idea in a study requiring similarity judgments between self and other. A typical American finding i that the self is judged to be more dissimilar to other than other is to the self (Holyoak & Gordon, 1983; Srull & Gaelick, 1983). This finding has been interpreted to indicate that for the typical American subject, the representation of the self is more elaborated and distinctive in memory than the representation of another person. As a result, the similarity between self and other is judged to be less when the question is posed about a more distinctive object (Is self similar to other?) than when the question is posed about a less distinctive object (Is other similar to self?). If, however, those with interdependent selves have at least as much knowledge about some others as they have about themselves, this American pattern of findings may not be found.

To test these predictions, Kitayama et al. (1990) compared students from Eastern cultural backgrounds (students from India) with those from Western cultural backgrounds (American students). As shown in Figure 2, for the Western subjects, Kitayama et al. replicated the prior findings in which the self is perceived as significantly more dissimilar to the other than is the other to the self. Such a finding is consistent with a broad range of studies showing that for individuals with a Western background, supposedly those with independent selves, selfknowledge is more distinctive and densely elaborated than knowledge about other people (e.g., Greenwald & Pratkanis, 1984). This pattern, however, was nonsignificantly reversed for the Indian subjects, who judged the self to be somewhat more similar to the other than is the other to the self. It appears, then, that for the latter, more interdependent subjects, knowledge about others is relatively more elaborated and distinctive than knowledge about the self. Asymmetry in similarity judgments is an indirect way to evaluate knowledge accessibility, but a more direct measure of cross-cultural differences in knowledge of the other should reveal that those with interdependent selves have more readily accessible knowledge of the other.

Context-specific knowledge of self and other.

A second consequence of having an interdependent self as opposed to an independent self concerns the ways in which knowledge about self and other is processed, organized, and retrieved from memory. For example, given an interdependent self, knowledge about the self may not be organized into a hierarchical structure with the person's characteristic attributes (e.g., intelligent, competent, and athletic) as the superordinate nodes, as is often assumed in characterizations of the independent self. In other words, those with interdependent selves are less likely to organize knowledge about the "self in general" or about the "other in generaL" Specific social situations are more likely to serve as the unit of representation than are attributes of separate persons. One learns about the self with respect to a specific other in a particular context and, conversely about the other with respect to the self in a particular context.

In exploring variations in the nature of person knowledge, Shweder and Bourne (1984) asked respondents in India and America to describe several close acquaintances. The descriptions provided by the Indians were more situationally specific and more relational than those of Americans. Indian descriptions focused on behavior; they described what was done, where it was done, and to whom or with whom it was done. The Indian respondents said, "He has no land to cultivate but likes to cultivate the land of others," or "When a quarrel arises, he cannot resist the temptation of saying a word," or "He behaves properly with guests but feels sorry if money is spent on them" It is the behavior itself that is focal and significant rather than the inner attribute that supposedly underlies it. Notably this tendency to provide the specific situational or interpersonal context when providing a description was reported to characterize the free descriptions of Indians regardless of social class, education, or literacy level. It appears, then, that the concreteness in person description is not due to a lack of skill in abstracting concrete instances to form a general proposition, but rather a consequence of the fact that global inferences about persons are typically regarded as not meaningful or informative.

Americans also describe other people in terms of the specifics of their behavior, but typically this occurs only at the beginning of relationships when the other is relatively unknown, or if the behavior is somehow distinctive and does not readily lend itself to a trait characterization. Rather than saying "He does not disclose secrets," Americans are more likely to say "He is discreet or principled?' Rather than "He is hesitant to give his money away," Americans say "He is tight or selfish?' Shweder and Bourne (1984) found that 46% of American descriptions were of the cntext free variety, whereas this was true of only 20% from the Indian sample - -

A study by J. 0. Miller (1984) on patterns of explanation among Indian Hindus and Americans revealed the same tendency for contextual and relational descriptions of behavior among Indian respondents. In the first phase of her study, respondents generated two prosocial behaviors and two deviant behaviors and then explained why each behavior was undertaken. For example, in the prosocial case, respondents were asked to "describe something a person you know well did re cently that you considered good for someone else." Miller coded the explanations for reference to dispositional explanations; for reference to social, spatial, temporal location; and for reference to specific acts or occurrences. Like Shweder and Bourne (1984), she found that on average, 40% of the reasons given by American respondents referred to the general dispositions of the actor. For the Hindu respondents, dispositional explana tions constituted less than 20% of their responses.

In a second phase of the study, Miller (1984) asked both American and Indian respondents to explain several accounts of the deviant behaviors generated by the Indian respondents. For example, a Hindu subject narrated the following incident:

This concerns a motorcycle accident. The back wheel burst on the motorcycle. The passenger sitting in the rear jumped. The moment the passenger fell, he struck his head on the pavement. The driver of the motorcycle-who is an attorney-as he was on his way to court for some work, just took the passenger to a local hospital and went on and attended to his court work. I personally feel the motorcycle driver did a wrong thing. The driver left the passenger there without consulting the doctor concerning the seriousness of the injury-the gravity of the situation-whether the passenger should be shifted immediately-and he went on to the court. So ultimately the passenger died. (p. 972)

Respondents were asked why the driver left the passenger at the hospital without staying to consult about the seriousness of the passenger's injury On average, Americans made 36% of their attributions to dispositions of the actors (e.g., irresponsible, pursuing success) and 17% of their attributions to contextual factors (driver's duty to be in court). In comparison, only 15% of the attributions of the Indians referred to dispositions, whereas 32% referred to contextual reasons. Both the American and the Indian subjects focused on the state of the driver at the time of the accident, but in the Indian accounts, the social role of the driver appears to be very important to understanding the events. He is obligated to his role, he has a job to perform. Actions are viewed as arising from relations or interactions with others; they are a product of obligations, responsibilities, or commitments to others and are thus best understood with respect to these interpersonal relations. This preference for contextual explanations has also been documented by Dalal, Sharma, and Bisht (1983).

These results call into question the exact nature of the fundafundamentla attribution error (Ross, 1977). In this error, people, in their efforts to understand the causes of behavior, suffer from an inescapable tendency to perceive behavior as a consequence of the internal, personal attributes of the person. Miller's (1984) Indian respondents also explained events in terms of properties or features of the person, yet these properties were their role relationships-their socially determined relations to specific others or groups. Because role relationships necessarily implicate the social situation that embeds the actor, it is unclear whether the explanations of the Indian respondents can be viewed as instances of the fundamental attribution error. It may be that the fundamental attribution error is only characteristic of those with an independent view of the self.

The tendency to describe a person in terms of his or her specific behavior and to specify the context for a given behavior is also evidenced when those with interdependent selves provide self-descriptions. Cousins (1989) compared the self-descriptions of American high school and college students with the self-descriptions of Japanese high school and college students. He used two types of free-response formats, the original Twenty Statements Test (TST; Kuhn & McPartland, 1954), which simply asks "Who Am I?" 20 consecutive times, and a modified TST, which asks subjects to describe themselves in several specific situations (me at home, me with friends, and me at school). When responding to the original TST, the Japanese self-descriptions were like those of the Indians in the Shweder and Bourne (1984) study. They were more concrete and role specific ("I play tennis on the weekend"). In contrast, the American descriptions included more psychological trait or attribute characterizations ("l am optimistic:' and "lam friendly"). However, in the modified TST, where a specific interpersonal context was provided so that respondents could envision the situation (e.g., me at home) and presumably who was there and what was being done to whom or by whom, this pattern of results was reversed. As shown in Figure 3, the Japanese showed a stronger tendency to characterize themselves in psychological trait or attribute terms than did Americans. In contrast, Americans tended to qualify their self-descriptions, claiming, for example, "I am sometimes lazy at home?'

Cousins (1989) argued that the original TST essentially isolates or disembeds the "I" from the relational or situational context, and thus self-description becomes artificial for the Japanese respondents, who are more accustomed to thinking about themselves within specific social situations. For these resPondents, the contextualized format "Describe yourself as you are with your family" was more "natural" because it locates the self in a habitual unit of representation, namely in a particular terpersonal situation. Once a defining context was specified, Japanese respondents were decidedly more willing to make generalizations about their behavior and to describe themselves abstractly using trait or attribute characterizations.

American students, in contrast to their Japanese counterparts, were more at home with the original TST because this test elicits the type of abstract, situation-free self-descriptions that form the core of the American, independent self-concept. Such abstract or global characterizations, according to Cousins (1989), reflect a claim of being a separate individual whose nature is not bound by a specific situation. When responding to the contextualized self-description questions, the American students qualified their descriptions as if to say "This is how I am at home, but don't assume this is the way I am everywhere?' For American respondents, selfness, pure and simple, seems to transcend any particular interpersonal relationships.

Basic cognition in an interpersonal context.

One's view of self can have an impact even on some evidently nonsocial cognitive activities. I. Liii (1986) described the emphasis that the Chinese place on being loyal and pious to their superiors and obedience to them, whether they are parents, employers, or government officials. He claimed that most Chinese adhere to a specific rule that states "If your superiors are present, or indirectly involved, in any situation, then you are to respect and obey them" (I. Liu, 1986, p. 78). The power and the influence of this rule appear to go considerably beyond that provided by the American admonition to "respect one's elders?' I. Liu (1986) argued that the standard of self-regulation that involves the attention and consideration of others is so pervasive that it may actually constrain verbal and ideational fluency. He reasoned that taking account of others in every situation is often at odds with individual assertion or with attempts at innovation or unique expression. This means, for example, that in an unstructured creativity task in which the goal is to generate as many ideas as possible, Chinese subjects may be at a relative disadvantage. In a similar vein, T. Y Liu and Hsu (1974) suggested that consideration of the rule "respect and obey others" uses up cognitive capacity that might otherwise be devoted to a task, and this may be the reason that Chinese norms for some creativity tasks fall below American norms.

Charting the differences between an independent self and interdependent self may also illuminate the controversy surrounding the debate between Bloom (1981, 1984) and Au (1983, 1984) over whether the Chinese can reason counterfactually (for a thorough review of this debate, see Moser, 1989). Bloom's studies (1981) on the counterfactual began when he asked Chinese-speaking subjects questions like "If the Hong Kong government were to pass a law requiring that all citizens born outside of Hong Kong make weekly reports of their activities to the police, how would you react?" Bloom noted that his respondents consistently answered "But the government hasn't," "It can't:' or "It won't?' Pressed to think about it anyway, the respondents became frustrated, claiming that it was unnatural or un-Chinese to think in this way. American and French respondents answered similar questions readily and without complaint. From this and subsequent studies, Bloom (1981, 1984) concluded that Chinese speakers "might be expected typi cally to encounter difficulty in maintaining a counterfactual perspective as an active point of orientation for guiding their cognitive activities" (1984, p. 21).

Au (1983) challenged Bloom's conclusions. Using different stimulus materials and also different translations of the same stimulus materials, she reported that Chinese subjects performed no differently from their Western counterparts. The controversy continues, however, and many investigators remain unconvinced that the differences Bloom and others have observed in a large number of studies on counterfactual reasoning are solely a function of awkward or improper translations of stimulus materials.

Moser (1989), for example, discussed several of Bloom's (1981, 1984) findings that are not easily explained away. He described the following question that Bloom (1981, pp. 53-54) gave to Taiwanese, Hong Kong, and American subjects in their native language.

Everyone has his or her own method for teaching children to respect morality. Some people punish the child for immoral behavior, thereby leading him to fear the consequences of such behavior. Others reward the child for moral behavior, thereby leading him to want to behave morally Even though both of these methods lead the child to respect morality, the first method can lead to some negative psychological consequences-it may lower the child's self-esteem.

According to the above paragraph, what do the two methods have in common? Please select only one answer.

A. Both methods are useless.

B. They have nothing in common, because the first leads to negative psychological consequences.

C. Both can reach the goal of leading the child to respect morality

D. It is better to use the second.

E. None of the above answers makes sense. (If you choose this answer, please explain

Bloom (1984) reported that 97% of American subjects responded C, but that only 55% of the Taiwanese and 65% of the Hong Kong respondents answered C. In explaining his results, he wrote:

Most of the remaining Chinese-speaking subjects chose D or E and then went on to explain, based on their own experience and often at great length and evidently after much reflection, why, for instance, the second method might be better, or why neither method works, or why both methods have to be used in conjunction with each other, or perhaps, why some other specified means is preferable. For the majority of these subjects, as was evident from later interviewing, it was not that they did not see the paragraph as stating that both methods lead the child top morality, but they felt that choosing that alternative and leaving it at that would be misleading since in their experience that response was untrue. As they saw it, what was expected, desired, must be at a minimum an answer reflecting their personal considered opinion, if not a more elaborated explanation of their own experiences relevant to the matter at hand. Why else would anyone ask the question? American subjects, by contrast, readily accepted the question as a purely "theoretical" exercise to be responded to according to the assumptions of the world it creates rather than in terms of their own experiences with the actual world. (Bloom, 1981, p. 54)

It is our view that the differences in response between the Americans and the Chinese may be related to whether the respondent has an independent or interdependent construal of the self. If one's actions are contingent on, determined by, or made meaningful by one's relationships and social situations, it is reasonable to expect that respondents with interdependent selves might focus on the motivation of the person administering the question and on the nature of their current relationship with this person. Consequently in the process of responding, they might ask themselves, "What is being asked of me here? What does this question expect of me or require from me? What are potential ramifications of answering in one way or another in respect to my relationship with this person?" In Lebra's (1976) terms, what is "my proper place?" in this social interaction [i.e., me and the interviewer], and what are the "obligations attached to [it?]" (p. 67). To immediately respond to the question as a purely abstract or theoretical exercise would require ignoring the currently constituted social situation and the nature of one's relationship with the other. This, of course, can be done, but it does not mean that it will be easily, effortlessly or automatically done. And this is especially true when the pragmatics of a given context appears to require just the opposite. It requires ignoring the other's perspective and a lack of attention to what the other must be thinking or feeling to ask such a question. One's actions are made meaningful by reference to a particular set of contextual factors. If these are ignored or changed, then the self that is determined by them changes also. Those with relatively unencumbered, self-contained, independent selves can readily, and without hesitation, entertain any of a thousand fanciful possible worlds because there are fewer personal consequences-the bounded, autonomous self remains essentially inviolate.

One important implication of this analysis is that people with interdependent selves should have no disadvantage in counterfactual reasoning if the intent of the questioner and the demand of the situation is simply to test the theoretical reasoning capacities of the person. One such situation would involve an aptitude test such as the Scholastic Aptitude Test (SAT). Indeed, on the quantitative portion of the SAT that requires substantial hypothetical and counterfactual reasoning (e.g, "If Tom walked 2 miles per hour, then how far will he have walked in 4 hours?"), both Taiwanese and Japanese children perform considerably better than their American peers (Stevenson et al., 1986).

It would appear important, therefore, to distinguish between competence and performance or between the presence of particular inference skills and the application of these skills in a particular pragmatic context (see also Laboratory of Comparative Human Cognition, 1982). The discussion thus far implies that regardless of the nature of the self-system, most people with an adequate level of education possess the skills of hypothetical reasoning and the ability to think in a counterfactual fashion. Yet, the application of these skills in a particular situation varies considerably with the nature of the self-system. Some people may invoke these skills much more selectively. For those with interdependent selves, in contrast to those with independent selves, a relatively greater proportion of all inferences will be contingent on the pragmatic implications of a given situation, such as the perceived demands of the interviewer, the convention of the situation, and the rules of conversation.

Do styles of thinking and inference vary above and beyond those that derive from the pragmatic considerations of particular social situations? This question has yet to be more carefully addressed. However, given the tendency to see people, events, and objects as embedded within particular situations and relationships, the possibility seems genuine. Chiu (1972), for example, claimed that the reasoning of American children is characterized by an inferential-categorical style, whereas the reasoning of Taiwanese Chinese subjects displays a relationalcontextual style. When American children described why two objects of a set of three objects went together, they were likely to say "because they both live on a farm" In contrast, Chinese children were more likely to display a relational-contextual style, putting two human figures together and claiming the two go together "because the mother takes care of the baby" In the latter case, the emphasis is on synthesizing features into an organized whole. Bruner (1986) referred to such differences as arising from a paradigmatic versus a narrative mode of thought. In the former, the goal is abstraction and analyzing common features, in the latter, establishing a connection or an interdependence among the elements.

Consequences for Emotion

In psychology, emotion is often viewed as a universal set of largely prewired internal processes of self-maintenance and self-regulation (Buck, 1988; Darwin, 1896; Ekman, 1972; LeDoux, 1987). This does not mean, though, that emotional experience is also universal. On the contrary, as suggested by anthropologists Rosaldo (1984), Lutz (1988), and Solomon (1984), culture can play a central role in shaping emotional experience. As with cognition, if an emotional activity or reaction implicates the self, the outcome of this activity will depend on the nature of the self-system. And apart from the fear induced by bright lights and loud sounds, or the pleasure produced by a sweet taste, there are likely to be few emotions that do not directly implicate one's view of the self. Thus, Rosaldo (1984) contended "feelings are not substances to be discovered in our blood but social practices organized by stories that we both enact and tell. They are structured by our forms of understanding" (p. 143), and we would add, specifically, by one's construal of the self. In an extension of these ideas, Lutz (1988) argued that although most emotions are viewed as universally experienced "natural" human phenomena, emotions are anything but natural. Emotion, she contended, "can be viewed as cultural and interpersonal products of naming, justifying, and persuading by people in relationship to each other. Emotional meaning is then a social rather than an individual achievement-an emergent product of social life" (Lutz, 1988, p. 5).

Among psychologists, several cognitively oriented theorists of emotion have suggested that emotion is importantly implicated and embedded in an actual social situation as construed by the person (e.g., De Riviera, 1984; Roseman, 1984; Scherer, 1984). Accordingly, not only does the experience of an emotion depend on the current construal of the social situation (e.g Frijda, Kuipers, & ter Schure, 1989; Shaver, Schwartz, Kirson, & O'Connor, 1987; C. Smith & Ellsworth, 1987), but the experienced emotion in turn plays a pivotal role in changing and transforming the very nature of the social situation by allowing anew construal of the situation to emerge and, furthermore, by instigating the person to engage in certain actions. From the current perspective, construals of the social situation are constrained by, and largely derived from, construals of the self, others, and the relationship between the two. Thus, emotional experience should vary systematically with the construal of the self.

The present analysis suggests several ways in which emotional processes may differ with the nature of the self-system. First, the predominant eliciting conditions of many emotions may differ markedly according to one's construal of the self. Second, and more important, which emotions will be expressed or experienced, and with what intensity and frequency, may also vary dramatically.

Ego-focused versus other-focused emotions.

The emotions systematically vary according to the extent to which they follow from, and also foster and reinforce, an independent or an interdependent construal of the self. This is a dimension that has largely been ignored in the literature. Some emotions, such as anger, frustration, and pride, have the individual's internal attributes this or her own needs 1oals, desires, or abilities) as the primary referent. Such emotions may be called ego focused. They result most typically from the blocking (e.g., "I was treated
unfairly"), the satisfaction, or the confirmation (e.g., "I performed better than others") of one's internal attributes. Experiencing and expressing these emotions further highlights these self-defining, internal attributes and leads to additional attempts to assert them in public and confirm them in private. As a consequence, for those with independent selves to operate effectively, they have to be "experts" in the expression of these emotions. They will manage the expression, and even the experience, of these emotions so that they maintain, affirm, and bolster the construal of the self as an autonomous entity. The public display of one's own internal attributes can be at odds with the maintenance of interdependent, cooperative social interaction, and when unchecked can result in interpersonal confrontation, conflict, and possibly even overt aggression. These negative consequences, however, are not as severe as they might be for interdependent selves because the expression of one's internal attributes is the culturally sanctioned task of the independent self. In short, the current analysis suggests that, in contrast to those with more interdependent selves, the ego-focused emotions will be more frequently expressed, and perhaps experienced, by those with independent selves.

In contrast to the ego-focused emotions, some other emotions, such as sympathy, feelings of interpersonal communion and shame, have another person, rahter than one's internal attricbutes, as the primary rferent. Such emotions may be called other focused They typically result from being sensitive to the other, taking the perspective of the other, and attempting to promote interdependence. Experiencing these emotions highlights one's interdependence, facilitates the reciprocal ex-
changes of well-intended actions, leads to further cooperative social behavior, and thus provides a significant form of self-validation for interdependent selves. As a consequence, for those with interdependent selves to operate effectively, they will have to be "experts" in the expression and experience of these emotions. They will manage the expression, and even the experience, of these emotions so that they maintain, affirm, and rein-
force the construal of the self as an interdependent entity. The other-focused emotions often discourage the autonomous expression to inhibition and ambivalence. Although among inde selves these consequences are experienced negatively (e.g., as timidity) and
can, in fact, have a negative impact, they are tolerated, among interdependent selves, as the "business of living" kakar1978 p. 34). Creating and maintaining a connection to others is the primary task of the interdependent self. In short, this analysis suggests that, in contrast to those with more independent selves, these other-focused emotions will be more frequently expressed and perhaps even experienced among those with interdependent selves.

Ego-focused emotions-emotions that foster and create independence.

In a comparison of American and Japanese undergraduates, Matsumoto, Kudoh, Scherer, and Wallbott (1988) found that American subjects reported experiencing their emotions longer than did Japanese subjects, even though the two groups agreed in their ordering of which emotions were experienced longest (i.e., joy = sad > anger guilt > fear = shame = disgust) Americans also reported feeling these emotions more intensely than the Japanese and reported more bodily symptoms (e.g., lump in throat, change in breathing, more expressive reactions, and more verbal reactions) than did the Japanese. Finally, when asked what they would do to cope with the consequences of various emotional events, significantly more of the Japanese students reported that no action was necessary.

One interpretation of this pattern of findings may assume that most of the emotions examined, with the exception of shame and possibly guilt, are what we have called ego-focused emotions. Thus, people with independent selves will attend more to these feelings and act on the basis of them, because these feelings are regarded as diagnostic of the independent self. Not to attend to one's inner feelings is often viewed as being inauthentic or even as denying the "real" self. In contrast, among those with more interdependent selves, one's inner feelings may be less important in determining one's consequent actions. Ego-focused feelings may be regarded as by-products of interpersonal relationships, but they may not be accorded privileged status as regulators of behavior. For those with interdependent selves, it is the interpersonal context that assumes priority over the inner attributes, such as private feelings. The latter may need to be controlled or dc-emphasized so as to effectively fit into the interpersonal context.

Given these differences in emotional processes, people with divergent selves may develop very different assumptions about the etiology of emotional expressions for ego-focused emotions. For those with independent selves, emotional expressions may literally "express" or reveal the inner feelings such as anger, sadness, and fear. F those with interdependent selves, however. an emotional expression may be more often regarded as a public instrumental action that may or may not be related di- with this analysis, Matsumoto (1989), using data from 15 cultures, reported that individuals from hierarchical cultures (that we would classify as being generally interdependent; see Hofstede, 1980), when asked to rate the intensity of an angry, sad, or fearful emotion displayed by an individual in a photograph, gave lower intensity ratings than those from less hierarchical cultures. Notably, although the degree of hierarchy inherent in one's cultures was strongly related to the intensity ratings given to those emotions, it was not related to the correct identification of these emotions. The one exception to this finding was that people from more hierarchical cultures (those with more interdependent selves) were less likely to correctly identify emotional expressions of happiness. Among those with interdependent selves (often those from hierarchical cultures), positive emotional expressions are most frequently used as public actions in the service of maintaining interpersonal harmony and, thus, are not regarded as particularly diagnostic of the actor's inner feelings or happiness.

For those with interdependent selves (composed primarily of relationships with others instead of inner attributes), it may be very important not to have intense experiences of eüo-focused emotions, and this may be particularly true for negative emotions like anger. Anger may sriously threaten an interdendent self and thus may be highly dysfunctional In fact, some anthropologists explicitly challenge the universalist ilist view that all people experience the same negative emotions. Thus, in Tahiti, anger is highly feared, and various anthropological accounts claim that there is no expression of anger in this culture (see Levy, 1973; Solomon, 1984). It is not that these people have learned to inhibit or suppress their "real" anger but that they have learned the importance of attending to others, considering others, and being gentle in all situations, and as a consequence very little anger is elicited. In other words, the social reality is construed and actually constructed in such a way that it does not lend itself to the strong experience, let alone the outburst, of negative ego-focused emotions such as anger. The same is claimed for Ukta Eskimos (Briggs, 1970). They are said not to feel anger, not to express anger, and not even to talk about anger. The claim is that they do not show anger even in those circumstances that would certainly produce complete outrage in Americans. These Eskimos use a word that means "childish" to label angry behavior when it is observed in foreigners.

Among the Japanese, there is a similar concern with averting anger and avoiding a disruption of the harmony of the social situation. As a consequence, experiencing anger or receiving anger signals may be relatively rare events. A study by Miyake, Campos, Kagan, and Bradshaw (1986), which compared Japanese and American infants of 11 months of age, provides suggestive evidence for this claim. These investigators showed each infant an interesting toy and paired it with a mother's vocal expression of joy, anger, or fear. Then they measured the child's latency to resume locomotion toward the toy after the mother's utterance. The two groups of infants did not differ in their reactions to expressions of joy or fear. But, after an angry vocal expression of the mother, there was a striking difference between the two groups. The Japanese children resumed locomotion toward the toy after 48 s, American children after only 18 s. It may be that the Japanese children are relatively more traumatized by their mother's anger expressions because these are such rare events.

Notably, in the West, a controversy exists about the need, the desirability, and the importance of expressing one's anger. Assuming a hydraulic model of anger, some argue that it is necessary to express anger so as to avoid boiling over or blowing up at a later point (Pennebaker, 1982). Others argue for the importance of controlling one's anger so as not to risk losing control. No such controversy appears to exist among those in predominantly interdependent cultures, where a seemingly unchallenged norm directs individuals to restrain their inner feelings and particularly the overt expression of these feelings. Indeed, many interdependent cultures have well-developed strategies that render them expert at avoiding the expression of negative emotions. For example, Bond (1986) reported that in China discussions have a clear structure that is explicitly designed to prevent conflict from erupting. To begin with, discussants present their common problems and identify all the constraints that all the participants must meet. Only then do they state their own views. To Westerners, such a pattern appears as vague, beating around the bush, and not getting to the heart of the matter, but it is part of a carefully executed strategy of avoiding conflict, and thus perhaps the experience of negative emotions. Bond, in fact, noted that among school children in Hong Kong and Taiwan, there is a tendency to cooperate with opponents even in a competitive reward structure and to rate future opponents more positively than others who will not be opponents (Li, Cheung, & Kau. 1979, 1982).

In a recent cross-cultural comparison of the eliciting conditions of several emotions, Matsumoto et al. (1988) also found that Japanese respondents appear to be avoiding anger in close relations. Specifically, for the Japanese, closely related others were rarely implicated in the experience of anger. The Japanese reported feeling anger primarily in the presence of strangers. It4 thus appears that not only the expression but also the experience of such an ego-focused emotion as anger is effectively averted within an interdependent structure of relation. When anger arises, it happens outside of the existing interdependence, as in confrontation with out-groups (e.g., Samurai warfare in feudal Japan). In _contrast Americans and Western Europeans report experiencing anger primarily in the presence of closely r-elate others. This is not surprising, given that expressing and experiencing ego-focused, even negative emotions, is one viable way to assert and affirm the status of the self as an independent entity. Consistent with this analysis, Stipek, Weiner, and Li (1989) found that when describing situations that produce anger. Chinese subjects were much more likely than American subjects to describe a situation that happened to someone else ("a guy on a bus did not give up a seat to an old woman"). For Americans, the major stimulus to anger was the situation where the individual was the victim ("a friend broke a promise to me").

Other emotions, such as pride or guilt, may also differ according to the nature of the mediating self-system. As with anger, these expressions may be avoided, or they will assume a somewhat different form. For example, if defined as being proud of one's own individual attributes, pride may mean hubris, and its expression may need to be avoided for those with interdependent selves.' Consistent with the idea that pride in one's own performance may be inhibited among those with interdependent selves, Stipek et al. (1989) found that the Chinese were decidedly less likely to claim their own successful efforts as a source of pride than were Americans. These investigators also reported that the emotion of guilt takes on somewhat different connotations as well. Among those with independent selves, who are more likely to hold stable, cross-situational beliefs and to consider them self- definitional. "violating a law cause of guilt. Among however, the most com reported source of guilt was "hurting others psychologically."

Other-focused emotions-emotions that create and foster interdependence. Those with interdependent selves may inhibit the experience, or at least the expression, of some ego-focused emotions, but they may have a heightened capacity for the experience and expression of those emotions that derive primarily from focusing on the other. In Japan and China, for example, there is a much greater incidence of cosleeping, cobathing, and physical contact between mother and child than is typically true in most Western countries. The traditional Japanese mother carries the child on her back for a large part of the first 2 years. Lebra (1976) claimed that Japanese mothers teach their children to fear the pain of loneliness, whereas Westerners teach children how to be alone. Japanese and Chinese socialization practices may help the child develop an interdependent self in the first place, and at the same time, the capacity for the experience of a relatively greater variety of other-focused emotions.

The greater interdependence that results between mothers and their children in Japan is reflected in the finding that the classification of infants according to the nature of their attachments to their mothers (i.e., secure, ambivalent, and avoidant) departs markedly from the pattern typically observed in Western data. Specifically, many more Japanese infants are classified as "ambivalently attached" because they seem to experience decidedly more stress following a brief separation from the mother than do American infants (Ainsworth, Bell, & Stayton, 1974; Miyake, Chen, & Campos, in press). This finding also indicates that a paradigm like the typical stranger situation is inhere linked to and independent view of self and thus may not be appropriate for guaging attachment in non-Western cultures.

In Japan, socialization practices that foster an intense closeness between mother and child give rise to the feeling of amae. Amae is typically defined as the sense of, or the accompanying hope for, being lovingly cared for and involves depending on and presuming another's indulgence. Although, as detailed by Kumagai and Kumagai (1985), the exact meaning of amae is open to some debate, it is clear that "the other" is essential. When a person experiences amae, she or he "feels the freedom to do whatever he or she wills" while being accepted and cared for by others with few strings attached. Some say amae is a type of complete acceptance, a phenomenal replication of the ideal mother-infant bond (L. T Doi, 1973). From our point of view, experiencing amae with respect to another person maybe inherent in the formation and maintenance of a mutually reciprocal, interdependent relationship with another person. If the other person accepts one's amae the reciprocal relationship is symbolically completed, leading to a significant form of self-validation. If, however, the other person rejects one's amae, the relationship will be in jeopardy.

For the purpose of comparing indigenous feelings, such as amae, with the more universal ones, such as anger and happiness, Kitayama and Markus (1990) used a multidimensional scaling technique which allows the identification of the dimensions that individuals habitually or spontaneously use when they make judgments about similarities among various emotions. Recent studies have demonstrated that people are capable of distinguishing among various emotions on as many as seven or eight cognitive dimensions (Mauro, Sato, & Tucker, 1989; C. Smith & Ellsworth, 1987). In these studies, however, the dimensions have been specified a priori by the experimenter and given explicitly to the respondents to use in describing the emotions. When the dimensions are not provided but allowed to emerge in multidimensional scaling studies, two dimensions are typically identified: activation (or exciteint) id pleasantness (e.g., Russell, 1980). And it appears that most Western emotions can be readily located on a circumplex plane defined by these two dimensions. Thus, although people are capable of discriminating among emotions on a substantial number of dimensions, they habitually categorize the emotions only on the dimensions of activation and pleasantness.

More recently Russell (1983; Russell, Lewicka, & Nut, 1989) applied the same technique to several non-Western cultural groups and replicated the American findings. He thus argued that the lay understanding of emotional experience may indeed be universal. Russell used, however, only those terms that have clear counterparts in the non-Western groups he studied. He did not include any emotion terms indigenous to the non-Western groups such as amae It is possible that once terms for such indigenous feeling states are included in the analysis, a new dimension, or dimensions, may emerge. To explore this possibility, Kitayama and Markus (1990) sampled 20 emotions from the Japanese language. Half of these terms were also found in English and were sampled so that they evenly covered the circumplex space identified by Russell. The remaining terms were those indigenous to Japanese culture and those that presuppose the presence of others. Some (e.g., fureai [feeling a close connection with someone else]) refer primarily to a positive association with others (rather than events that happen within the individual, such as success), whereas others refer to interpersonal isolation and conflict (e.g., oime [the feeling of indebtedness]). Japanese college students rated the similarity between 2 emotions for each of the 190 pairs that could be made from the 20 emotions. The mean perceived similarity ratings for these pairs were then submitted to a multidimensional scaling.

Replicating past research, Kitayama and Markus (1990) identified two dimensions that closely correspond to the activation and the pleasantness dimensions. In addition, however, a new dimension emerged. This third dimension represented the extent to which the person is engaged in or disengaged from an interpersonal relationship. At the interpersonal engagement end were what we have called other-focused emotions, such as shame, fureai [feeling a close connection with somebody else], and shitashimi [feeling familiar], whereas at the disengagement end were found some ego-centered emotions, such as pride and tukeagari [feeling puffed up with the sense of self-importance], along with sleepiness and boredom. This interpersonal engagement-disengagement dimension also differentiated between otherwise very similar emotions. Thus, pride and elation were equally positive and high in activation, yet pride was perceived as considerably less interpersonally engaged than elation. Furthermore, anger and shame were very similar in terms of activation and pleasantness, but shame was much higher than anger in the extent of interpersonal engagement.

More important, this study located the indigenous emotions within the three-dimensional structure, permitting us to understand the nature of these emotions in reference to more universal emotions. For instance, amae was low in activation, and neither positive nor negative, fairly akin to sleepiness, except that the former was much more interpersonally engaged than the latter. This may indicate the passive nature of amae, involving the hopeful expectation of another person's favor and indulgence without any active, agentic solicitation of them. Completion of amae depends entirely on the other person, and, therefore, amae is uniquely ambivalent in its connotation on the pleasantness dimension. Another indigenous emotion, oime, involves the feeling of being psychologically indebted to somebody else. Oime was located at the very negative end of the pleasantness dimension, perceived even more negatively than such universal negative emotions as anger and sadness. The extreme unpleasantness of oime suggests the aversive nature of unmet obligations and the press of the need to fulfill one's obligations to others and to return favors. It also underscores the significance of balanced and harmonious relationships in the emotional life of those with interdependent selves.

The finding that the Japanese respondents clearly and reliably discriminated between ego-focused emotions and otherfocused emotions on the dimension of interpersonal engagement versus disengagement strongly suggests the validity of this distinction as an essential component of emotional experience at least among Japanese and, perhaps, among people from other cultures as well. In a more recent study Kitayama and Markus (1990) further tested whether this theoretical dimension of emotion also underlies and even determines how frequently people may experience various emotions and whether the frequency of emotional experience varies with their dominant construal of self as independent or interdependent.

Kitayama and Markus (1990) first sampled three emotions common in Japanese culture that were expected to fall under one of the five types theoretically derived from the current analysis. These types are listed in Table 2. Ego-focused positive emotions (yuetukan [feeling superior], pride, and tukeagari [feeling puffed up] are those that are most typically associated with the confirmation or fulfillment of one's internal attributes, such as abilities, desires, and needs. Ego-focused, negative emotions (anger, futekusare [sulky feeling], and yokyufuman [frustration]) occur primarily when such internal attributes are blocked or threatened. Also included were those correspondingly positive or negative emotions associated with the maintenance or enhancement of interdependence. Thus, three emotions are commonly associated with the affirmation or the completion of interdependent relationships (fureai [feeling of connection with someone], shitashimi [feeling of familiarity to