17
Culture and Social Cognition
Toward a Social Psychology of Cultural Dynamics
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
YOSHIHISA KASHIMA
Social cognition, broadly defined as human thought about social behavior, has received considerable attention in the literature since the cognitive revolution of the
1960s and, indeed, has become one of the most important areas of study in mainstream psychology. Within this large area, cross-cultural research on social cognition
has come to play an extremely important role in defining issues and in influencing research and theory.
In this chapter, Kashima presents a comprehensive overview of the area of culture
and social cognition. He first begins with an excellent discussion of the concept of
culture in psychology, distinguishing the concept of culture as meaning from cultural
dynamics. As Kashima suggests, cultural dynamics has to do with the paradoxical
phenomenon of cultural stability and change, which arises from two contemporary
views of culture: system oriented and practice oriented. These definitions and discussions about the concept of culture are essential to Kashima’s later points about the necessity for the development and creation of theories and research on cultural dynamics, which represent a further evolution of research and thinking about social
cognition, and an integration of approaches and knowledge from various disciplines.
The bulk of Kashima’s chapter is devoted to a state-of-the-art review of research
on culture and social cognition. This review promises to be one of the most comprehensive reviews on this topic. He begins with a treatment of the historical context of
early social cognition research and with a presentation of background studies in the
area. His detailed review spans such topics as availability of concepts, causal attributions, self-concepts, social and personal explanation, self-evaluation, and others. He
delineates many of the issues that are highlighted through his thorough evaluation of
the research literature, pointing out both what we know and what we do not in each
area. The reader is sure to view this area of his chapter as an important resource for
this line of inquiry.
Using his review of the literature as a platform, Kashima delineates his ideas concerning future research and theoretical work in the area. With regard to future empirical work, he suggests that two topics in the area of culture and social cognition—the
explanation of social action and the maintenance of self-regard—deserve closer scrutiny and further research in the future. In particular, while much is known about
what North Americans tend to do with regard to these topics, relatively much less
325
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
326
CULTURE AND SOCIAL BEHAVIOR
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
is known about other people around the world, leaving this area ripe for investigation. In particular, the holistic approach and worldview perspective of East Asians
may bring insights into this area of psychological functioning that heretofore were
unconsidered.
Clearly, however, the major thrust of Kashima’s argument for future work concerns
the creation of what he terms the social cognition of cultural dynamics. As he explains at the beginning of his chapter and throughout his literature review, much of
the early social cognition research and theories were characterized by an individualistic conception of meaning, according to which meaning is constructed solely within
an individual person’s mind. There are many reasons for these biases in the literature, including the fact that most research was done in the United States by American researchers. Even research that was conducted outside the United States was often conducted by researchers who were trained in the United States (and thus
influenced by Western educational dogma) or influenced by these factors. In the future, however, greater emphasis will need to be placed on the development of a theoretical framework that incorporates both cognitive and communicative processes in
understanding cultural dynamics—that is, the processes by which cultural meanings
are constructed in ongoing social activities among multiple individuals, as well as
within an individual’s mind. This view of social cognition is inherently more complex, involving relational, collective, and individual issues, including the incorporation of context and history, as well as future and present time orientations. For these
reasons, the development of such a theoretical viewpoint will necessitate fundamental changes in the ways in which we do research, which will ultimately lead to ways
in which we understand human behavior in potentially profoundly different ways
than now. This development of new theories and methodologies to ensure the continued evolution of knowledge in this area of psychology is commensurate with a message given by all authors throughout this volume.
Until recently, culture has been a neglected
concept in social cognition. Most theories, at
worst, have ignored culture entirely or, at best,
assumed that culture is connected unproblematically to the traditional social psychological
concepts such as attributions and attitudes. To
wit, the first edition of the Handbook of Social
Cognition (Wyer & Srull, 1984) has no entries of
culture, and this marginal status of the culture
concept continued until the 1990s, as seen in
the absence of culture in the second edition of
the Handbook (Wyer & Srull, 1994). However,
culture emerged recently as a major theme
in social cognition. There is an increase in
publication on culture and social cognition according to my recent search of the literature
from 1989 to 1997 of the computer database
PSYCINFO (Y. Kashima, 1998b).
The main aim of this chapter is to make a
case for a perspective that I call a social psychology of cultural dynamics. It attempts to
understand global dynamics of culture as generated from cognitive and communicative processes of individuals in interaction with each
other in social contexts. The chapter is divided
into four sections. In the first section, the concept of culture is examined, and major meta-
theoretical tenets of a social psychology of
cultural dynamics are derived. In the second
section, traditional metatheoretical and theoretical characteristics of social cognition are reviewed. The third section reviews the recent
explosion of research on culture and social cognition that past reviews (e.g., Fletcher & Ward,
1988; J. G. Miller, 1988; Semin & Zwier, 1997;
Zebrowitz-McArthur, 1988) did not cover. In
the last section, empirical and theoretical directions of future research are suggested.
Culture Concept
in Psychology
To clarify the perspective of a social psychology
of cultural dynamics, it is necessary to clarify
the concept of culture. The culture concept,
despite its popularity and long history in social
sciences, is multifaceted, and often ambiguous.
Culture as Meaning
Culture is analytically separable from concepts
such as society and social system (e.g., Giddens,
1979; Parsons, 1951; Rohner, 1984; for a more
recent discussion, see Y. Kashima, 2000a). On
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
CULTURE AND SOCIAL COGNITION
one hand, society is an organized collection
of individuals and groups, and social system
refers to an enduring pattern of interpersonal,
intergroup, and person-group relationships
within a society. On the other hand, culture
is a set of meanings shared, or at least sharable, among individuals in a society. Therefore,
questions regarding power, resources, and
friends have to do with social systems. In contrast, culture has to do with questions about
what it means to have power and resources and
what it means for a person to be a friend of
another.
The concept of meaning, however, is complex. At this stage, let us approximate meaning
to the use of symbols, that is, material objects
(including sound, light, and other chemical
characteristics that are discernible by human
senses) that are used to stand for something
else. Obviously, words have meanings in this
sense. Nevertheless, this sense of meaning goes
beyond linguistic meaning. When a nonverbal
gesture stands for other ideas (e.g., vertically
stretched index and middle fingers standing
for victory), this involves a meaning. When a
toddler uses a round object as a steering wheel
of a car, the child is engaged in a meaningful
activity.
Nonetheless, what it stands for does not exhaust the meaning of a symbol. The denotative
(extensional) meaning is that to which a symbol
refers (i.e., its referent). However, there is more
to meaning than reference. As Frege (1984)
noted long ago, if the referent of a phrase such
as morning star or evening star is all there is
to meaning, then a statement like “The morning
star is the evening star” is a meaningless tautology. Yet, this statement can have a rich meaning
given that humans had not known for a long
time that the morning star and the evening star
referred to the same object, Venus. Frege called
this extra component of meaning sense. Meaning thus has at least two aspects, reference and
sense.
It is important to note that referential meaning should include not only literal meaning,
but also figurative meaning. For instance, Lakoff and Johnson (1979) noted that a number
of abstract concepts in English were based on
metaphors. English sentences such as, “That
meeting was a waste of time,” can be understood in terms of a metaphor that likens time
to money. Just as money is wasted, time can be
wasted, too. In 1994, Y. Kashima (also see Y.
Kashima & Callan, 1994; Shore, 1996) argued
that cultural metaphors provide rich meanings
327
for the experience of mental and social activities. In addition, narratives may also play an
important role in the production and maintenance of cultural meanings (Bruner, 1990; Y.
Kashima, 1998a).
Cultural Dynamics
Cultural dynamics has to do with the paradoxical phenomenon of cultural stability and change,
that is, how some aspects of a culture are maintained in the midst of constant change, and
cultural change continues despite strong forces
of cultural maintenance. This question arose
from a tension between two contemporary
views of culture, system oriented and practice
oriented (Y. Kashima, 2000a; also see Matsumoto, Kudoh, & Takeuchi, 1996). A systemoriented view treats culture as a relatively
enduring system of meaning. Culture is conceptualized as a repository of symbolically coded
meanings shared by a group of people, which
provides structure to their experience. In contrast, a practice-oriented view regards culture
as signification process in which meanings are
constantly produced and reproduced by concrete individuals’ particular activities in particular situations. The system-oriented view highlights the stability of culture, whereas the
practice-oriented view focuses on the fluid nature of culture in flux.
The culture-as-meaning-system view was
expressed by a number of cross-cultural psychologists and anthropologists. Most notably,
when Triandis (1972) defined subjective culture as a “cultural group’s characteristic way
of perceiving the man-made part of its environment” (p. 4), he was highlighting the enduring
and systemic aspect of culture. A well-known
anthropologist, Geertz (1973), characterized
culture as “interworked systems of construable
signs . . . something within which [social events,
behaviors, institutions, or processes] can be intelligibly . . . described” (p. 14). Geertz’s formulation, called symbolic anthropology, likens
culture to a text, which is publicly accessible
and in need of reading and interpretation. Despite a difference between the views of culture
of Triandis and Geertz, there is an underlying
similarity. They both treat culture as a system
of meanings that is shared within a group of
people.
For example, theorists who take this perspective often characterize a culture by using
a global concept such as individualism or collectivism (Hofstede, 1980; Triandis, 1995), implying that a relatively stable system of beliefs
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
328
CULTURE AND SOCIAL BEHAVIOR
and values is shared in a society. Similarly,
when Geertz (1984, p. 126) characterized the
Western conception of the person as “a bounded, unique, more or less integrated motivational
and cognitive universe,” he implied that this
conception was shared by people in the West.
The culture-as-signification-process view was
put forward by a variety of psychologists influenced by Vygotsky (1978; for an explication of
Vygotsky, see Wertsch, 1985) and other thinkers of the Russian cultural-historical school.
These include Cole (1996), Greenfield (1997),
Lave and Wenger (1991), Rogoff (1990), Valsiner (1989), and Wertsch (1991). Although
their theory of culture has progressed beyond
Vygotsky’s original formulation, they view culture as a collection of concrete everyday practices that occur in everyday life (e.g., basket
weaving, estimating amounts of rice). Boesch’s
(1991) symbolic action theory and Poortinga’s
(1992) context-specific cross-cultural psychology are similarly concerned with concrete activities as they occur within symbolic, physical,
and social contexts. In anthropology, researchers influenced by Bourdieu (1977; habitus) and
Giddens (1979; structuration) or by contemporary Marxist thoughts often take a similar view.
Ortner (1984), a neo-Geertzian, also approaches
culture from a similar viewpoint.
An example of this approach is provided by
a conceptualization of schooling (for a recent
review, see Rogoff & Chavajay, 1995). For instance, Cole (1996) views schooling as a collection of context-specific and domain-specific
cognitive and motor activities (e.g., reading and
writing, remembering a list of words) that influence children’s cognitive task performance, such
as recall and syllogistic reasoning. In other
words, instead of explaining cultural differences in syllogistic reasoning performance in
terms of differences in cognitive style (e.g., logical versus prelogical reasoning), this approach
suggests that people from Western cultures
tend to perform syllogistic reasoning tasks better than illiterate people because the reasoning
tasks resemble activities that the former are
used to at school.
The two conceptions of culture differ on a
number of metatheoretical dimensions. First,
they differ in time perspective. The systemoriented view tends to see culture from a longterm perspective and attempts to capture stable
aspects of a culture within a historical period
(decades or centuries). In contrast, the practiceoriented view tends to construe culture from
a short-term perspective and tries to identify
activities (that is, people doing things together
with tools) that recur in specific contexts. In
other words, a unit of time is longer for a system-oriented investigation than for a practiceoriented analysis.
Second, they differ in context specificity and
domain specificity. The system-oriented view
is generally concerned with culture viewed as
a whole, as a context-general and domain-general meaning system that is carried and realized
by a group of individuals. Culture, then, is abstracted from specific contexts of social action.
Culture is often regarded as present, although
it may lay dormant, in all contexts of social
activities and all domains of life. The practiceoriented view, on the other hand, is interested
in culture as particular activities that use particular artifacts (i.e., tools and other material objects) in particular contexts. This is a view of
culture as a collection of context-specific signification activities. To the extent that a domain
of meaning is often associated with a particular
context (e.g., things to do at school or at home),
this view tends toward a view that cultural
meanings are domain specific.
Third, they differ in unit of analysis. The
system-oriented view takes a group of individuals as a unit of analysis, and culture is a phenomenon closely associated with the collectivity. In a way, culture is regarded as a property
of the group. In contrast, the practice-oriented
view takes a practice (a pattern of activities
carried out by people) as a unit of analysis. In
this perspective, culture is a property of situated activities, that is, people acting in context.
It should be noted that this notion of practice
and situated activities includes not only individuals, but also routine activities that take
place in space and time.
Neither view alone can provide a complete
picture about cultural dynamics. One view’s
strength is the other’s weakness. On one hand,
the system-oriented view takes culture as given
for a collective in a historical period. Culture in
this sense becomes a “cause” or an independent
variable in a quasi-experimental design of typical cross-cultural studies. In fact, comparative
investigations must by necessity treat culture
as stable systems and compare the slices of cultural traditions. However, this view often looks
for factors external to culture as engines of cultural change (e.g., technology, material wealth,
and ecology). Creative activities within a culture as a basis for cultural change tend to fall
outside the scope of this perspective.
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
CULTURE AND SOCIAL COGNITION
On the other hand, the practice-oriented
view takes culture as constantly produced and
reproduced. As such, both stability and change
are part and parcel of culture. Developmental psychologists, who are concerned with
how children are enculturated to become fullfledged participants of a culture, are necessarily
interested in context-specific activities. After
all, children must learn culture not by osmosis,
but from concrete everyday activities. However, it is unclear in this view how one can
determine theoretically which aspects of situated activities are to persist and which are to
change. Furthermore, while this view provides
detailed analyses of particular activities, it fails
to shed light on a general pattern, a cultural
theme, or something like a context-general
meaning system that seems to cut across a number of domains of activities (e.g., see Jahoda’s
1980 criticism of Cole’s 1996 approach).
Thus, the system-oriented and practice-oriented views of culture provide complementary
perspectives on cultural dynamics. The culture-as-system view highlights the persistence
of culture over time, whereas the culture-aspractice view focuses on the fluctuation of cultural meaning across contexts and over time.
Nonetheless, both local fluctuations and global
stability characterize culture. My contention
is that we must investigate how both can be
true. From the present perspective, the central
question of cultural dynamics is how individuals’ context-specific signification activities can
generate, under some circumstances, something stable that may be called a context-general
meaning system and, under other circumstances, a rapid and even chaotic change.
Culture and Social Cognition:
Historical Context of Early
Social Cognition Research
Despite some early attempts at incorporating
culture into human psychology (e.g., Wundt’s
Völkerpsychologie), culture, broadly defined
as shared meanings, has been outside the scope
of academic psychology for much of the first
half of the 20th century under the dogmatic
and restrictive reign of logical positivism as a
philosophy of science and behaviorism as its
psychological counterpart. Behaviorism, in
particular, banished any talk of human thought
from the academic discourse of psychology.
The cognitive revolution of the 1960s, in which
human thought was reclaimed as a central con-
329
cern of psychology, failed to bring meaning,
and therefore culture, back into the mainstream
of academic psychology (Bruner, 1990).
Social cognition emerged as an attempt at
bringing cognition into social psychology. The
1960s saw publications of classic texts in attribution theories (e.g., Jones & Davis, 1965; Kelley, 1967), and social psychology was flooded
with research on attribution processes in the
1970s. A more self-conscious effort to draw on
cognitive psychology began as well, making use
of prototypically cognitive psychological methods such as recall and recognition memory, reaction time, and the like (e.g., Hastie et al., 1980,
on person memory). The significance of social
cognition in social psychology is undeniable.
Some have gone as far as to claiming that social
psychology is largely represented by social cognition (H. Markus & Zajonc, 1985). All the while,
however, social cognition research emulated
the cognitive psychology, pursuing a universal
model of human cognitive processes at the expense of culture.
It is intriguing to note that social cognition
of the 1970s and 1980s was characterized by
its dual emphasis on the individual person as
a central focus. On one hand, much of the work
was largely concerned about the process by
which people form cognitive representations
about themselves and other individuals. One
enduring question has been how one comes to
construe a person (either another individual or
oneself) in terms of his or her dispositional
characteristics, such as personality traits (e.g.,
as described by adjectives such as introverted
and extraverted) or attitudes (e.g., stances with
regard to social issues such as Castro’s Cuba or
abortion; for a review, see S. T. Fiske & Taylor,
1991). On the other hand, theories of social
cognition paid exclusive attention to the individual person’s cognitive processes, that is, the
encoding of incoming information into cognitive representations, and the storage and retrieval of them for further use. These theories
were social only to the extent that they dealt
with social stimuli (i.e., other people). In other
words, social cognition then exemplified the
individualist conception of the person in terms
of its subject matter and theoretical assumption.
What underlay the early social cognition
research was an individualist conception of
meaning. According to this view, the individual constructs meaning by operating on cognitive representations stored in his or her own
mind. To be sure, an individual person equipped
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
330
CULTURE AND SOCIAL BEHAVIOR
with the capacity to encode perceptual information into cognitive representations can also
decode such individual representations into
symbolic codes that are understandable to
other individuals. Nonetheless, this individualist model of cognition makes for a model of
communication that regards interpersonal communication as mere transmission of information (Clark, 1985). At an extreme, social cognitive minds can be likened to computers that
send signals back and forth through rules of
syntax and semantics. In this case, culture can
be reduced to a “codebook” in which rules can
be found to translate between cognitive codes
and symbolic codes.
During the Great Leap Forward of social cognition, however, metatheoretical, theoretical,
and empirical challenges to the mainstream social psychology began to cumulate. Metatheoretically, Gergen (1973) argued that social psychology cannot hope to “discover” natural laws
of social behavior, but only acquire historically
contingent knowledge. Although social psychology may develop a theory of social behavior at one point in time, once it is disseminated
to the general public, people can try to develop
patterns of behaviors that differ from, or even
contradict, it. In other words, humans are selfconstituting in that our collective attempt at
characterizing ourselves can end up influencing ourselves. Gergen’s argument that humans
are self-constitutive and the products of history
that we ourselves have created echoes the point
made by the counter-Enlightenment thinkers,
such as Vico and Herder, who opposed the Enlightenment thought that regarded human nature as largely fixed and governed by universal
natural laws (for a more detailed discussion,
see Y. Kashima, 2000a).
Theoretically, some of the central concepts
in social psychology began to be scrutinized.
For instance, social psychologists began to examine concepts such as personality traits and
social attitudes, which were presumed to describe the underlying dispositions of people or
consistency in their behavior. Most fundamentally, the capacity of the dispositional characteristics to predict specific behavior was questioned (see Mischel, 1968, for personality traits
and Wicker, 1969, for attitudes). Based on these
challenges, cognitively oriented researchers
such as Shweder and D’Andrade (1979) and
Cantor and Michel (1979) began to formulate
theories of personality traits that treated them
as indicating perceived, as opposed to actual,
consistency in behavior. In other words, an in-
dividual does not necessarily possess a disposition, but merely appear to do so. This trend
reached its peak when Ross (1977) used the
term fundamental attribution error to refer to
North American participants’ tendency to attribute personality dispositions to an actor despite contextual information suggesting otherwise. When attribution of a disposition is
regarded as an error, social psychology can
hardly take its dispositional concepts seriously.
Empirically, drawing on the past literature
on culture and social behavior (e.g., for a review, see Triandis & Brislin, 1980), cross-cultural psychology began to mount empirical
challenges to social psychology by presenting
evidence that there is some significant cultural
variability in social behavior (Bond, 1988).
Amir and Sharon’s (1987) was among the most
memorable contributions. They sampled several North American studies published in major
journals of social psychology and systematically replicated the experimental procedures
in Israel. They reported that, although main
effects could be replicated, some of the finegrain interaction effects could not be, despite
their importance for the main theoretical claims
of the original papers. Within the context of the
globalization of economy and the rapid change
in the world order, such as the political and
economic collapse of the Communist bloc and
the emergence of newly industrializing nations
(e.g., Japan, South Korea, Taiwan, Hong Kong,
and Singapore), these cross-cultural challenges
began to attract the attention of mainstream researchers.
Cross-Cultural Research in
Social Cognition
Background
The current popularity of culture and social
cognition research owes much to Hofstede
(1980) and Shweder and Bourne (1984). Hofstede’s research was based on his work value
surveys around the world. In this massive, empirically driven work, he extracted dimensions
on which cultures can be placed. The individualism dimension attracted the greatest attention
partly because of the importance of the individualism concept in social sciences in general.
A number of social scientists (e.g., Tönnies’
(1955), Gemeinschaft and Gesellschaft; Durkheim’s (1964) mechanical and organic solidarity) used related concepts to characterize the
transformation of Continental Europe from its
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
CULTURE AND SOCIAL COGNITION
medieval past to the modern era. Close-knit
communities in which everyone had known
everyone else broke down, and there emerged
modern nation states in which a central government controls the trade, police, and military
might. The emphasis shifted from the community to the individual, with the gradual strengthening of individual rights. Collectivism characterizes the traditional sociality, whereas the
modern social relationship is individualistic.
An empirical finding that fueled the interest
was probably its correlation with 1970 per capita gross national product. Country-level individualism positively correlated with per capita
gross national product at .82. Richer countries
in North America and western Europe are individualist, whereas poorer countries in Asia and
South America tend to be collectivist.
Hofstede’s (1980) finding clearly showed
that there was a significant relationship between cultural values and economic activities,
which is generally consistent with the accepted
view of modernization, that is, from traditional
communities to modern societies. His work
provided a conceptual framework in which to
interpret and understand myriad cross-cultural
studies on beliefs, attitudes, and values. The
concepts of collectivism and individualism
also refocused theoretical attention on a central
issue of social sciences, that is, the relationship
between the collective and the individual.
Shweder and Bourne’s study (1982), in contrast, was an ambitious, theoretically driven
project. They posited three major theoretical
orientations in interpreting cross-cultural diversity. Universalism looks for human universals in diversity by attempting to identify a
higher order generality or by concentrating on
a clearly defined band of data. Evolutionism
rank orders cultural patterns relative to a normative model (e.g., the cannon of propositional
calculus, Bayes’ rule of probabilistic reasoning)
in terms of their deviation from the norm. It
typically treats cultures as progressing toward
the normative ideal. Relativism seeks to interpret each cultural pattern as an inherently
meaningful pattern by itself and to maintain
the equality among them.
Against the background of the literature
arguing for western Europeans’ abstractness
relative to other cultures, such as Bali and Gahuku-Gama of New Guinea, Shweder and Bourne
(1984) conducted interviews and showed that
middle-class Euro-American participants, compared to Oriyan participants from a traditional
city of Bhubaneswar in India, tended to use
331
abstract dispositional characterizations (e.g.,
“He is a leader” as opposed to “He lends people
money”), and that their descriptions tended not
to be put into context (e.g., “He is verbally abusive” as opposed to “He is verbally abusive to
his father-in-law whenever they meet at his
home”). Oriyas’s contextualized person description, they argued, is a sign of their holistic,
sociocentric conception of the relationship of
the individual to society.
In adopting this relativist view, Shweder
and Bourne (1984) argued against evolutionist
interpretations. They showed that Oriyas adopted a contextual person description regardless
of formal education, literacy, or socioeconomic
status. According to them, this provides evidence against evolutionist explanations. Evolutionists would explain relative concreteness in
Oriyas’s person description in terms of some
cognitive deficit associated with a lack of education, literacy, or socioeconomic background.
Oriyas do have abstract traitlike words in their
language and are capable of generating those
abstract concepts in an interview. This argues
against the possibility that Oriyas lack abstract
categories with which to describe people abstractly or lack a general capacity to do so. It is
unlikely that North Americans encounter their
target persons in more diverse settings and
therefore are more likely to be able to abstract
their dispositional characteristics. There is no
evidence to suggest that North Americans live
in a more heterogeneous social environment to
prompt more abstract patterns of thinking.
These two lines of work were drawn together
into a single focus around 1990 by two major
papers on culture and self, which triggered the
avalanche of cross-cultural research in social
cognition. Triandis (1989) theorized about cultural antecedents of the prevalence and access
of self-concepts. He postulated that there are
three types of self-concepts: private, public,
and collective. The private self-concept is concerned with people’s conceptions about their
own personal goals; the public self-concept has
to do with people’s concerns about how others
view them; and the collective self-concept is
about people’s involvement in their in-groups.
Every culture contains these different selfconcepts, but characteristics such as individualism, cultural complexity, and affluence
determine the prevalence of the three types of
self-concepts. Private self-concepts may be
prevalent in individualist cultures, whereas
public and collective self-conceptions may be
prevalent in collectivist cultures. Cultural com-
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
332
CULTURE AND SOCIAL BEHAVIOR
plexity and affluence may promote the prevalence of private self-conceptions as well. Furthermore, different self-concepts may be more
accessible in different social situations in different cultures according to Triandis.
In 1991, H. R. Markus and Kitayama proposed a theory about psychological consequences
when different self-concepts are accessed. In
this influential formulation, they postulated
that there is the universal aspect of the self,
which is the self as a physically distinct body
in time and space. However, the self can be
construed in two different ways, independent
and interdependent. Independent self-construals are characterized by their emphasis on the
uniqueness and separateness of the individual
self in contrast to others. In contrast, interdependent self-construals are characterized by
their interpenetrations with significant others.
That is, selves are conceived to be in interdependent social relationships with other people.
They suggested that these self-construals, when
accessed, would influence cognitive, affective,
and motivational processes.
Although Hofstede’s (1980) and Shweder
and Bourne’s (1984) studies were not without
their critics (for instance, see Y. Kashima, 1987,
on Hofstede and see Spiro, 1993, on Shweder
and Bourne), their contributions suggested that
there are significant cultural differences, which
may be examined empirically in terms of the
contrast between worldviews that emphasize
sociality (e.g., collectivist, sociocentric, interpersonal, and interdependent) and those that
emphasize individuality (e.g., individualist,
egocentric, personal, and independent). Triandis (1989) and H. R. Markus and Kitayama
(1991) focused research attention on the self.
Generally drawing on the then-current literature on social cognition of self-processes and
cross-cultural psychology, they launched a theory that suggested that self-concepts mediate
the effect of culture on psychological processes.
Nonetheless, there were important differences. First, they differed in the unit of analysis.
Hofstede (1980) used countries or cultures as
the unit of analysis, computing cultural averages on surveys, whereas Shweder and Bourne
(1984), as well as H. R. Markus and Kitayama
(1991), treated individuals as the unit of analysis. Triandis (1989) attempted to connect the
two using the self-concept as a central mediator. They also differed in their focus on self or
person in general. Triandis’s and Markus and
Kitayama’s contributions were concerned with
self-concepts. However, Shweder and Bourne’s
contribution was about conceptions of the person observed from people’s verbal descriptions
of their acquaintances. Finally, they differed in
operationalization of the constructs. Hofstede
made individualism and collectivism operational in terms of importance of personal independence from the organizational context, while
Shweder and Bourne made egocentric and sociocentric views of the person operational in
terms of abstractness of person descriptions.
Triandis and Markus and Kitayama treated them
as reflecting the same underlying self-conceptions. In the current literature of culture and
social cognition, these differences are generally
glossed over or even ignored. But, is it warranted?
Conceptual Advances
Some theoretical advances since the early
1990s have significant implications for culture
and social cognition.
Availability, Accessibility,
and Applicabilty of Concepts
Higgins (1996) defined availability, accessibility, and applicability of concepts and provided
a comprehensive review and discussion of the
literature on knowledge activation. The availability of a concept refers to whether the
concept is stored in an individual’s memory,
whereas the accessibility of a concept means
the “activation potential” of the available concept or the ease with which the concept already
available in the mind is activated for use. The
accessibility of a concept may vary chronically
or temporarily due to factors such as motivation
and frequency and recency of activation. Bargh,
Bond, Lombardi, and Tota (1986) showed that
chronic and temporary sources of accessibility
are additively combined to produce effects.
When accessible concepts are applicable to a
given stimulus, the concepts are applied to the
stimulus to interpret it. Much of recent research
shows that concepts may be activated automatically and used without conscious awareness
(Bargh, 1996).
Hong, Chiu, and Kung (1997) provided an
example of priming cultural concepts, in which
the accessibility of concepts was increased temporarily. Hong et al. showed Hong Kong Chinese students pictures of objects that symbolized either Chinese or American culture and
had the participants answer short questions
such as, “What does this picture symbolize?”
A short while later, in an allegedly unrelated
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
CULTURE AND SOCIAL COGNITION
study, the participants rated the importance of
traditional Chinese values. The participants endorsed the traditional Chinese values more in
the Chinese picture condition than in the American picture condition. Pictures that symbolized a culture may have activated concepts and
knowledge structures associated with the culture, which influenced subsequent cognitive
processes in the experiments.
The availability and accessibility of concepts may be associated closely with the language people use. This is one way of interpreting what is known as the Whorfian hypothesis,
the idea that language determines thought
(Chiu, Krauss, & Lee, 1999; for a recent review
on the Whorfian hypothesis, see Hunt & Agnoli, 1991). Hoffman, Lau, and Johnson (1986)
provided an example consistent with this
thinking. They identified English and Chinese
terms for which there were no equivalent economical words or phrases in the other language.
Behavioral descriptions were developed for
each term. English monolingual and ChineseEnglish bilingual individuals were given these
person descriptions with the aim of forming
distinct impressions. The bilingual individuals
read them either in English or in Chinese. The
bilingual individuals’ impressions of the target
individuals and recognition memory were influenced by the concepts available in the language used in the experiment, although recall
was not.
In examining cultural differences in social
cognition, it is important to consider the availability, accessibility, and applicability of a relevant concept in cultures concerned. If a culture
does not provide a concept of importance, people from that culture could not use it (unless
they invent it on the spot); if a concept is available in the cultures concerned, they may differ
in accessibility and therefore may result in differences in cognitive processes; even if a concept is equally available and accessible in the
cultures concerned, it may not be equally applicable in both. What is an intriguing possibility
is that cultural concepts that are not consciously available, accessible, or applicable
may still exert influences on social cognitive
processes without awareness of the members
of the culture. When a culture is going through
a major change and its members are actively
attempting to forget or discredit its past cultural
practices, there may emerge a discrepancy between conscious awareness about concepts (explicit cognition) and automatically activated
concepts (implicit cognition; see also Hetts, Sa-
333
kuma, & Pelham, 1999). A good example of this
is perhaps once-prevalent stereotypes in the
contemporary culture of political correctness
(e.g., Devine, 1989).
A Variety of Causal
Attributions
Another class of theoretical advances in social
cognition has to do with the meaning of causal
attribution. In the classical attribution theories,
it has commonly been assumed that personal
and situational attributions perfectly correlate
negatively. That is, attributing a behavior to a
person means that the context of the behavior is
not causally implicated. Alternatively, saying
that a behavior is situationally caused means
that the person is not causally responsible. This
hydraulic assumption of personal and situational causation (Heider, 1958) has been called
into question by some empirical studies (F. D.
Miller, Smith, & Uleman, 1981). At least North
American participants may regard personal
and situational causation as two independent
forces.
In line with this, current theories of attribution (e.g., Gilbert & Malone, 1995; Krull, 1993;
Trope, 1986) draw a distinction between the
attribution of a disposition and the adjustment
of a dispositional inference on the basis of the
contextual factors that may constrain the action. To put it differently, the cognitive process
responsible for a dispositional inference is distinguished from the contextualization of the
dispositional inference. These theories developed in North America suggest that, when faced
with information about a behavioral episode,
people first categorize the action into a dispositional category (e.g., personality trait, attitudes), and the implication of this categorization is then adjusted in light of the information
about the contextual constraints. If the context
is likely to hinder the enactment of the action,
the dispositional inference is curtailed, albeit
insufficiently, by some normative standards.
Obviously, whether the same processes apply
around the world needs to be examined by
cross-cultural investigations. Nonetheless, the
conceptual distinction between disposition
and contextualization is highly pertinent to the
discussion of cross-cultural research on social
cognition, as discussed below.
Two types of personal attribution, dispositional attributions (using personality traits to
describe and explain a behavior) and agentic
attributions (saying that a person is responsible
for the behavior), have been assumed to be
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
334
CULTURE AND SOCIAL BEHAVIOR
equivalent conceptually in the classical attribution literature. However, recent research has
shown that a distinction needs to be made between them. Semin and Marsman (1994) and
D. J. Hilton, Smith, and Kin (1995) showed that
attributing to a person abstract dispositional
characteristics is psychologically different from
attributing agency to the person. To put it differently, to describe a person by a certain personality trait on the basis of an observed behavior
is not the same as saying that this person is
responsible for the behavior.
that they are indeed conceptually separable.
Brewer and Gardner (1996) also theorized about
the conceptual separation among individual,
relational, and collective aspects of the self.
In line with the tripartite distinction among
individual, relational, and collective aspects of
the self, cross-cultural research on social cognition is reviewed according to whether the individual, relationship, or group is the target of
conception.
Individual as Target
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
Individual, Relational, and
Collective Selves
Although the pioneering work in culture and
social psychology contrasted the individualcentered and sociocentered worldviews (individualist, egocentric, personal, independent vs.
collectivist, sociocentric, interpersonal, interdependent), Y. Kashima’s (1987) and Oyserman’s (1993) exploratory factor analyses (also
see Trafimow, Triandis, & Goto, 1991), as well
as Singelis’s (1994) confirmatory factor analysis, suggested that the individual-centered and
sociocentered conceptions of the self are two
independent concepts.
More recently, Y. Kashima et al. (1995) further differentiated the sociocentered self into
relational and collective facets, making distinctions among individual, relational, and collective self-conceptions. The existing theories of
culture and self often conflate two types of sociality: one primarily concerned with the selfother relationship and the other about the relationship of the self with the in-group. Whereas
the interpersonal relationship between the self
and other individuals may provide a significant
basis of sociality, the relationship between the
self and its in-group constitutes another social
aspect of the self, which requires separate treatment. They showed that measures of these three
aspects of the self had relatively small correlations among themselves, and that, more importantly, individual and collective self-conceptions differentiated East Asian (Japanese
and Korean) and English-speaking (Australian,
American) cultures with Hawaiians in between, but relational self-conceptions differentiated men and women regardless of their cultural background. The finding that collective
and relational self-conceptions had different relationships with culture and gender suggests
Social cognition researchers have traditionally
maintained a clear distinction between cognitions about the self (e.g., H. Markus, 1977) and
those about others. This conceptual separation was to some extent based on the preconception that the self is a special psychological
phenomenon that is uniquely different from
any other psychological phenomena, as seen,
for instance, in the assertion of Descartes about
one’s privileged access to one’s self-knowledge
(i.e., “Cogito ergo sum”). This assumption has
been reinforced further by empirical findings
that emphasize a difference between self-perception and perception of others. For instance,
the classical research on actor-observer bias in
attribution suggests that at least North Americans explain the behaviors of others in terms
of dispositional characteristics more than their
own behaviors (e.g., Nisbett, Caputo, Legant, &
Maracek, 1973). Watson’s (1982) review showed
its robustness, but interestingly, also revealed
that North Americans tend to explain both themselves and others more in terms of personality
dispositions than their circumstances. In other
words, the literature on the actor-observer bias
showed a significant similarity between selfcognition and cognition of others.
In discussing a cultural difference in selfcognition and cognition of others, the oft-cited
passage of Geertz (1984) provides a useful starting point:
The Western conception of the person 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 its social and
natural background, is, however incorrigible
it may seem to us, a rather peculiar idea
within the context of the world’s cultures.
(p. 126)
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
CULTURE AND SOCIAL COGNITION
The anthropological insight of Geertz can be
abstracted into two component ideas. First, a
person is attributed psychological agency,
which is a “more or less integrated motivational
and cognitive universe, a dynamic center of
awareness, emotion, judgment, and action.”
Second, the individual is a figure against the
background of the social and natural context,
that is, the individual is “set contrastively both
against other wholes and against its social and
natural background.”
This last idea needs further explication.
Geertz is not saying that the Western conception ignores the social and natural context, but
is asserting that the most prominent part of the
phenomenal field is the individual person, and
that the social and natural context in which the
person is embedded lies in the background.
This is often assumed to mean that the Western
conception of the person is abstract or dispositional (e.g., John is friendly), and that nonWestern conceptions are concrete and contain
action descriptions (e.g., John plays with children even if he doesn’t know them well). Furthermore, dispositional descriptions of a person mean that the person is decontextualized.
Taken together, it is commonly believed that
past cross-cultural studies of self-cognition and
cognition of others showed that Western conceptions of the self and other are more agentic,
dispositional, and decontextualized than their
East Asian counterparts.
However, in light of the recent theoretical
developments in social cognition, the crosscultural studies require a more nuanced interpretation. As pointed out before, social cognition research suggests that the attribution of
agency should be distinguished conceptually
from the attribution of disposition, and the cognitive process for attributing a dispositional
characteristic (e.g., a personality trait) to a person is distinguished from the cognitive process
for contextualizing the dispositional attribution. In other words, agency attribution, dispositional attribution, and contextualization are
all separable psychological processes in North
America and probably in European cultures.
However, there is no reason to expect a priori
prevalence of these psychological processes
should covary with culture.
Explaining and
Describing Others
How Shall a Person Be Described? In 1984
and in 1987, J. G. Miller examined descriptions
of acquaintances’ positive and negative behav-
335
iors given by North Americans and Hindu Indians of four age groups (8, 11, and 15 years old
and adults).The study found that, in general,
North American participants gave more dispositional and fewer contextual explanations than
Hindu Indians. However, the tendency to give
dispositional explanations increased with age
for North Americans, but not for Hindu Indians.
In contrast, the tendency to give contextual explanations increased with age for Hindu Indians, but not for North Americans. In addition,
J. G. Miller reported that, of the four Indian
adult groups examined, an Anglo-Indian group
showed a preference for dispositional explanations relative to three Hindu Indian groups, although some of the Hindu Indian groups had
more exposure to the Western-type education
and way of life. Also, J. G. Miller gave a different
group of North American participants English
translations of behavioral narratives given by
the Hindu Indian participants. The North American explanations again were oriented more toward disposition and less oriented toward context, although the behaviors to be explained
originated from the Hindu Indian narratives.
This last finding clearly showed that it was
not the nature of the narratives that caused the
cultural difference in explanatory style. All in
all, J. G. Miller’s study showed clearly that
North Americans generated explanations different from those generated by Hindu Indians.
Her finding that a cultural explanation style
became more pronounced for older participants
provides strong evidence for the cultural explanation of the difference in explanatory style.
Morris and Peng’s (1994) Study 2 provided
additional support for a cultural difference in
the type of explanations generated for individual behaviors. They coded newspaper articles
about mass murderers (one American and one
Chinese) in the United States; the articles appeared in English-language (New York Times)
and Chinese-language (World Journal) newspapers published in New York and circulated
worldwide. The proportion of segments that
signified dispositional or contextual explanation was computed for each article. They found
that English-language newspaper articles tended
to explain the behaviors in dispositional terms
more than Chinese-language newspaper articles for both cases (F. Lee, Hallahan, & Herzog
showed a similar trend in their 1996 study).
A reliable difference in contextual explanation
was not found at the .05 level for either case.
Using a different method, Morris and Peng’s
(1994) Study 3 asked American and Chinese
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
336
CULTURE AND SOCIAL BEHAVIOR
(including Hong Kong, People’s Republic of
China, and Republic of China) physics graduate
students to rate the importance of a variety of
dispositional and contextual explanations of
the two murder cases (instead of generating explanations). On average, American participants
rated the importance of dispositional causes
as greater than Chinese participants for the
Chinese murderer, although they did not show
a reliable difference for the American murderer.
In contrast, American participants rated the importance of contextual causes lower than Chinese participants for both cases. Taken together, Morris and Peng found some evidence
that Americans may generate dispositional explanations more than Chinese, although there
may not be differences in generation of contextual explanations. In contrast, Chinese may
evaluate contextual explanations as more important than Americans, although there may
not be a strong difference in evaluating the importance of dispositional causes (Morris and
Peng made a similar point).
It is interesting to note that Morris and Peng
(1994) reported a pattern of findings consistent
with this interpretation in their Study 1. They
showed computer-generated movements of a
black circle in reaction to a square (inanimate
objects) and those of fish in reaction to a school
of fish (animate objects) to American and Chinese high school and graduate students. Participants were asked to rate the extent to which
these movements were due to dispositional or
contextual forces. As expected, there was no
cultural difference in the evaluation of causality for inanimate objects. For animate objects,
high school students exhibited an expected pattern, although graduate students did not show
any cultural difference. Chinese high school
students rated contextual forces as more important than their American counterparts for all
types of movements. However, Chinese students rated dispositional forces as less important than Americans only for one of three.
Why People Describe Others the Way They Do
Why do Asians and North Americans describe
others the way they do? A popular answer is
that it is due to a cultural difference in individualism or collectivism or due to independent
or interdependent self-construal. Nonetheless,
there is a surprising paucity of supportive evidence for this explanation. If we were to attribute a causal role to independent and interdependent self-construal, that is, if the prevalence
and activation of self-schemata in an Asian or
Western individual are to explain the cultural
differences in person descriptions and explanations, we should be able to measure self-construals and show that the cultural differences
disappear if we statistically control for the selfconstruals (self-concept mediation hypothesis). However, as Matsumoto (1999) noted,
studies that took this approach did not find
empirical support for a mediation effect of selfconstruals. Clearly, the self-concept mediation
hypothesis needs to be examined more fully
with more sophisticated measures. An alternative hypothesis is that there may be a cultural
theory that affects the psychological process
involved in both self-cognition and cognition
of others (cultural theory hypothesis).
Although it is not easy to separate these two
viewpoints empirically, a cultural theory approach has received some attention in the past.
In 1992, Y. Kashima, Siegal, Tanaka, and Kashima examined the role played in Australia
and Japan by people’s implicit theory about
attitude-behavior relationship. They reasoned
that, in English-speaking countries, the values
of sincerity and authenticity (Trilling, 1972)
encourage people to make their feelings and
avowals consistent with each other, whereas
the Japanese notions of omote and ura (front
and back, respectively) suggest that people
should express their feelings appropriately in
suitable contexts (Doi, 1986). Accordingly,
Australians would have a stronger belief in attitude-behavior consistency than Japanese. Japanese and Australian students’ attitude attributions were examined in a paradigm used by
Jones and Harris (1967), in which participants
were asked to read a hypothetical actor’s essay
about environmental issues. It was found by
Y. Kashima et al. that Australians attributed
corresponding attitudes more than Japanese
overall. Nonetheless, this cultural difference
was mediated by the extent to which Australians and Japanese differed in beliefs in attitude-behavior relationship. When the effect of
attitude-behavior relationship beliefs was statistically controlled, the cultural effect on attitude attribution became nonsignificant. It is interesting to note that subsequent studies on
attitude attributions comparing Americans
with Koreans (Choi & Nisbett, 1998) or with
Taiwanese (Krull et al., 1999) showed that there
was no cultural difference in the extent to
which participants attributed attitudes corresponding to the actor’s behavior.
In a recent study, Chiu, Hong, and Dweck
(1997) showed that people’s implicit theory of
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
CULTURE AND SOCIAL COGNITION
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
personality is related to the tendency to make
dispositional attributions. According to Dweck
(1999), people hold an implicit theory about the
nature of personality. Some believe personality
consists of fixed and unchangeable traits (entity
theory), whereas others believe personality is
a dynamic quality that can be developed and
changed. In four studies, Chiu et al. showed
that, when compared to incremental theorists,
entity theorists are more likely to generalize an
individual’s behavior from one specific situation to another (Study 1), to predict an individual’s behavior from his or her personality trait
(Study 2), and to attribute trait dispositions
from a single behavior (Study 3). In Study 5, a
manipulation of people’s implicit theory also
produced a similar result. In their Study 4, the
relationship between implicit theory and tendency to make trait dispositional judgments
was examined in Hong Kong and the United
States. In both cultures, implicit theory predicted dispositional attributions, and Americans showed a stronger tendency to make
dispositional attributions than Hong Kong Chinese; however, this cultural difference was not
related to implicit theory of personality. Both
samples showed a similar level of entity theory.
Taking Situational Constraints into Account
Even if cultures differ in dispositional attribution, this does not always mean that they differ
in the extent to which situational constraints
are taken into consideration. The participants
in Study 2 of Y. Kashima et al. were told, in
one condition, that the writer wrote the essay
freely, but in the other condition, they were
told that the writer wrote the essay because he
was instructed to do so by his teacher. Jones
and Harris (1967) argued that, when the essay
writer’s behavior was constrained by an authority’s instruction, the behavior should not be
diagnostic of the writer’s underlying attitudes,
and therefore a rational observer should not
attribute an attitude in correspondence to the
behavior.
However, their study (and others; see Jones,
1979) found that people tended to attribute attitudes despite the situational constraints. This
tendency to give insufficient weight to situational constraints has been called a correspondence bias (Gilbert & Malone, 1995). Replicating the work of Jones and Harris (1967), Y.
Kashima et al. (1992) found that both Japanese
and Australians failed to take into account the
situational constraint on the essay writer. Consistent with this, Choi and Nisbett (1998; Study
337
1) in the United States and Korea, and Krull et
al. (1999; Study 1) in the United States and
Taiwan found little difference in correspondence bias using the attitude attribution paradigm; the last researchers found no cultural
difference in their replication of the work of
Ross, Amabile, and Steinmetz (1977) in the
United States and Hong Kong (Study 2).
Nevertheless, Choi and Nisbett (1998)
showed that, when situational constraints of
behavior are made salient, Koreans take into
account the information about situational constraints more than Americans. In their Study
2, they used the attitude attribution paradigm
of Jones and Harris (1967). That is, participants
were told that a student wrote an essay under
choice and no choice conditions, and the
salience of situational constraints was manipulated at two levels. In one condition, the participants experienced the same situational constraints as the essay writer (i.e., they were given
no choice in writing an essay), and in the other
condition, the participants experienced the
constraints and were given a set of arguments
for them to use in writing their essays. Choi and
Nisbett combined the data from the no choice
condition in their Study 1 with the Study 2
data and found that the salience manipulation
decreased the amount of correspondence bias
for Koreans, but had no effect for Americans.
This may mean that Koreans (and possibly East
Asians in general) are more sensitive to situational constraints under some circumstances
than Americans. Alternatively, Koreans may be
more empathetic than Americans. Note that, in
their experiment, the participants were required to experience situational constraints
and transpose this experience onto the essay
writer. Some evidence corroborates this interpretation. In the work of Choi and Nisbett, as
well as that of Y. Kashima et al. (1992), the
participants’ own attitudes predicted the attitudes attributed to the essay writer more
strongly in the Korean or Japanese sample than
in the U.S. or Australian sample.
Conceptualizing the Self
Open-Ended Self-Descriptions The results of
cross-cultural studies of open-ended self-descriptions largely mirror those of descriptions
of others. When asked to describe themselves,
North Americans tend to use more abstract, decontextualized words and phrases. Studies reviewed typically made use of the Twenty Statements Test (TST; Kuhn & McPartland, 1954) or
its variants, in which people are asked to an-
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
338
CULTURE AND SOCIAL BEHAVIOR
swer the question, “Who am I?” by completing
20 sentences that start with “I am . . . ”
Bond and Cheung (1983), in their pioneering
study, examined Hong Kong Chinese, Japanese,
and American students’ self-descriptions on
the TST and found that Japanese self-descriptions included fewer general psychological attributes (typically personality trait words) than
American ones, though they failed to find a
difference between Japanese and Chinese participants in this regard. Subsequent studies using
a similar technique showed that in both Malaysia (Bochner, 1984) and India (Dhawan, Roseman, Naidu, Thapa, & Rettek, 1995), self-descriptions tended to have lower percentages of
personality traitlike descriptions than in English-speaking countries (Australia and Britain
for Bochner; United States for Dhawan et al.).
However, English-speaking Indian participants
could show a level of personality trait use similar to British and Bulgarian participants (Lalljee & Angelova, 1995).
Cousins (1989) used a variant of this method.
He first used the TST and examined all the selfdescriptions, as well as five self-descriptions
that the participants selected as most important. He reported the results of the five most
important self-descriptions as there was only a
small difference. They found that U.S. students’
self-descriptions included a greater proportion
of personality traitlike descriptions (58%) than
their Japanese counterparts (19%); however,
Japanese students used a greater proportion
(27%) of social descriptions, such as social
roles, institutional memberships, and the like
than American students (9%).
Immediately after the typical TST, Cousins
(1989) asked his participants to “Describe yourself in the following situations:” followed by
the phrases “at home,” “at school,” and “with
close friends” (p. 126). The exact format of this
“contextualized” version of the self-description
task is unclear from his writing, however. For
instance, it is unclear how many times the participants were to write their self-descriptions,
whether they were told to describe themselves
a set number of times for each of the three settings listed (i.e., at home, at school, and with
close friends), or if none of these things were
explicitly stated. Nonetheless, the findings are
intriguing. Cousins reported the reversal of the
TST finding: That is, the Japanese participants
mentioned pure attributes more (41%) than the
Americans (26%). In this contextualized version, the Americans qualified their traitlike
self-descriptions more (35%; e.g., “I am usually
open with my brother,” p. 129) than the Japanese (22%). Leuers and Sonoda (1996; Sonoda & Leuers, 1996) largely replicated Cousins’s findings using data from Japanese and
Irish subjects.
Following Cousins (1989; also see Shweder
& Bourne, 1984), these findings can be interpreted as showing that culturally constituted conceptions of the person are different between the
United States and Japan. The Japanese tendency to describe themselves using abstract
traitlike terms in the contextualized format
indicates that they are as capable of abstract
self-descriptions as their American or Irish
counterparts. Cousins argued that the Japanese
conception of the self is more situated and contextualized.
Rhee, Uleman, Lee, and Roman (1995) examined self-descriptions of Koreans, Asian Americans, and European Americans on the TST.
They also divided the Asian American group
into three groups: those who mentioned both
ethnicity (e.g., Asian American) and nationality
(e.g., Chinese, Indian; doubly identified), those
who mentioned either ethnicity or nationality
(singly identified), and those who mentioned
neither (unidentified). The percentage of trait
self-descriptions increased from Koreans (17%),
to doubly identified Asian Americans (24%),
to singly identified Asian Americans (31%),
and to European Americans (35%), as expected.
Surprisingly, unidentified Asian Americans
had the greatest percentage of trait self-descriptions (45%). The authors then computed the
percentage of autonomous, social, abstract, and
specific self-descriptions. The percentage of abstract descriptions was greatest for unidentified
Asian Americans, followed by Euro-Americans, singly identified and doubly identified
Asian Americans, and Koreans (the percentage
of specific descriptions was opposite to this
trend). Likewise, the percentage of autonomous
self-descriptions followed the exact pattern as
that of abstract ones (the percentage of social
descriptions showed a reverse pattern). As Triandis, Kashima, Shimada, and Villareal (1986)
suggested, those who are extremely acculturated into the host culture (unidentified Asian
Americans) may have become even more Americanized than the majority of the host culture.
Intriguingly, the correlation between abstract
and autonomous self-descriptions was highest
among Euro-Americans and unidentified and
singly identified Asian Americans (.77 to .74),
but was lower among the other groups (.58 to
.34). The cross-cultural variation in the correla-
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
CULTURE AND SOCIAL COGNITION
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
tions implies a conceptual distinction between
abstract self and agentic self.
Structured Measures of Self-Conceptions There
is less direct evidence for cultural difference
in self-conceptions when structured measures
are used (Takano, 1999). Singelis (1994) was
probably the first to provide this type of evidence. He showed that Asian Americans scored
higher on the interdependent and lower on the
independent self-construal than European
Americans in Hawaii, and that a difference between the tendency of Asian and European
Americans to make situational explanations
can be explained by the interdependent selfconstrual. Nonetheless, the data came from
American participants from different cultural
backgrounds.
Kashima et al. (1995) provided further evidence for cultural differences in self-conception by examining two East Asian countries
(Japan and Korea), two Western countries (Australia and the United States), and Hawaii. They
devised measures of four different aspects of
self-conception. Two of the four pertained to
agency and assertiveness, that is, the extent to
which the self is perceived to be a goal-oriented
agent or an assertive individual (individualist);
one had to do with the extent to which the self
is conceptualized in relation to another individual (relational); and the last aspect was concerned with the self as a member of one’s ingroup (collective). The four self-aspects were
shown to have only moderate correlations in
Australia, the United States, Hawaii, Japan, and
Korea. A major cultural difference was found
for the two individualist aspects of the self.
Australian and American students rated higher
on these measures than Japanese and Korean
students, with the Hawaiian sample in between. Although there was a small cultural difference for the collective aspect of the self, cultural differences in relational self showed an
unexpected pattern, with the Korean and
Japanese samples marking the highest and lowest scores, respectively. Instead, there was a
stronger gender difference on relational self:
Women were more relational than men in most
samples.
Self in Context Theorists have suggested that
individualist or collectivist tendencies (e.g.,
Triandis, 1995), and indeed accessible self-conceptions (e.g., Triandis, 1989), are context dependent. Although the context sensitivity of
individualism and collectivism has been shown
339
(e.g., Fijneman, Willemsen, & Poortinga, 1996;
Matsumoto, Takeuchi, Andayani, Kouznetsova,
& Krupp, 1998; Rhee, Uleman, & Lee, 1996) and
the context-sensitive measures of these constructs have been developed (e.g., Matsumoto,
Weissman, Preston, Brown, & Kupperbusch,
1997), it is unclear how context-sensitive cultural differences in the self-concepts are. One
possibility is that collective selves are more
accessible and individualistic selves are less
accessible across all contexts in collectivist cultures such as East Asia than in individualist
cultures such as North America (generality hypothesis). Another possibility is that individualistic and collective selves can be accessible
in different contexts in different cultures (Culture × Context Interaction Hypothesis). In the
domain of resource exchange behavior, Poortinga and colleagues (Fijneman et al., 1996;
Poortinga, 1992; van den Heuvel & Poortinga,
1999) postulated a universalist interaction hypothesis. According to these researchers, there
is a universal pattern of resource exchange so
that certain types of resources are more likely
to be exchanged with certain types of others;
however, there are some cultural variations in
specific contexts (also see Kroonenberg & Kashima, 1997).
Uleman, Rhee, Bardoliwalla, Semin, and
Toyama’s (1999) study extended this line of
reasoning to the domain of self. They constructed a new measure of relational self in
which a respondent is asked to indicate how
close the self is to specific others such as immediate family, relatives, and close friend in terms
of global closeness, emotional closeness, mutual support, identity, reputation, similarity,
and harmony. The degree of closeness was indicated by the amount of overlap between two
circles, as in a Venn diagram. The data were
collected from Euro-Americans, Asian Americans, Dutch, Turkish, and Japanese university students. Their Culture × Gender × Target ×
Closeness type analysis of variance (ANOVA)
suggested that, of all the two-way interaction
effects, Target × Closeness type effect was the
largest, which suggests the importance of context specificity of relational self. To explore a
significant Culture × Target × Closeness type
interaction, they computed the deviation scores
of each culture’s Target × Closeness type means
from the averages across all cultural groups and
conducted a cluster analysis on these deviation
scores. This showed that the European American and Dutch groups formed a tight individualist cluster, and the Turkish and Japanese
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
340
CULTURE AND SOCIAL BEHAVIOR
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
groups formed another looser collectivist cluster, with the Asian Americans joining the latter.
Cognitive Representations of the Self Crosscultural research has been conducted not only
on self-descriptions, but also on cognitive representations of the self. Trafimow et al. (1991)
postulated that different types of self-cognitions may be stored in different storage places.
In particular, they tested between two models,
one suggesting that individualist and collective
self-cognitions are stored in one location (onebasket model) and the other model suggesting
they are stored in two separate locations (twobasket model). The two-basket theory predicts
that priming one type of self-conceptions would
increase the accessibility of only the same type
of self-conceptions. However, one-basket theory predicts that both individual and collective
selves would be more accessed when either
individualist or collectivist concepts are primed.
According to Trafimow et al., the one-basket
theory predicts that the retrieval of one type of
self-cognition is equally likely to be followed
by any other type of self-cognition. However,
the two-basket theory predicts that the retrieval
of one type of self-cognition is more likely to
be followed by the same type of self-cognition.
To put it differently, if the two-basket theory
is true, the conditional probability of retrieving
an individual self-cognition given that an individual self-cognition has been retrieved immediately before is greater than the conditional
probability of retrieving a collective self-cognition in the same condition or vice versa, that
is, p(I*I) > p(I*C) and p(C*C) > p(C*I).
In Study 1, Trafimow et al. (1991) primed
both European American students and students
who had Chinese family names and whose native language was not English by having them
think what made them different from their families and friends (individual prime) and what
they had in common with their families and
friends (collective prime) for 2 minutes and
then had the participants respond to the TST
for 5 minutes. The experiment was conducted
in English. Although all participants reported
more individual self-cognitions than collective
ones, this tendency was greater for North American students than for Chinese students. Consistent with the two-basket theory, the priming
manipulation had differential effects on individual and collective self-cognitions. In addition, the conditional probabilities computed
from the TST showed the pattern that the authors argued was consistent with the two-
basket theory. In Study 2, they primed individualist and collectivist concepts by a different
method (having participants read a story that
emphasized either personal characteristics or
family relationships) using American students
and replicated the Study 1 findings. Trafimow,
Silverman, Fan, and Law (1997) conducted a
comparable experiment using English and Chinese in Hong Kong with bilingual Chinese students. The two priming methods in Trafimow
et al. (1991) were both used in this experiment,
including an additional no priming control
condition. In the English language condition,
the results were largely consistent with those of
Trafimow et al. (1991). However, in the Chinese
language condition, the priming manipulation
had no effect, although the conditional probabilities showed the pattern consistent with the
two-basket model.
Gardner, Gabriel, and Lee (1999) extended
the work of Trafimow and colleagues (1991) by
examining the effects of priming individual and
collective selves not only on self-descriptions, but also on value and morality judgments. In Experiment 1, they showed that two
methods of priming individual and collective
selves (Trafimow et al., 1991, and Brewer &
Gardner, 1996) affected in expected ways North
American students’ TST responses, as well as
endorsement of individualist and collectivist
values and the extent to which the responsibility to help needy others was seen to be a universal obligation, and that the effect of priming
on value endorsement was mediated by selfdescriptions. Experiment 2 was conducted with
American and Hong Kong students using English. Trafimow et al.’s (1991) priming manipulation was followed by a value questionnaire.
In the no prime control condition, Americans
endorsed individualist values more strongly
than collectivist values, but this was reversed
in Hong Kong. However, when Americans’ collective selves were primed, Americans endorsed collectivist values more than individualist values; the priming of individual selves
did not affect value endorsement. In contrast,
when Hong Kong Chinese students’ individual
selves were primed, they endorsed individualist values more than collectivist values; the
priming of collective selves did not change
the value endorsement pattern. It is yet to be
seen whether this is replicated using Chinese
(see Trafimow et al., 1997).
Issues Associated with the Research on SelfRepresentation There are both methodological and theoretical problems about self-repre-
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
CULTURE AND SOCIAL COGNITION
sentations. Methodologically, the TST has many
problems despite its popularity. Wylie (1974)
expressed some doubts about the construct validity of its coding schemes. In the contemporary uses of TST, a variety of coding schemes
has been suggested. Triandis (1995) suggested
the percentage of social items S% as a measure
of allocentrism (individual-level construct of
collectivism). Trafimow et al. (1991) classified
TST responses into two categories, individual
and collective. Bochner (1984) used a tripartite
scheme of individual, relational, and collective
responses. Watkins, Yau, Dahlin, and Wondimu (1997) suggested a four-part scheme: idiocentric (individual-level construct for individualism), large group (e.g., gender, occupation),
small group (e.g., family), and allocentric (e.g.,
I am sociable). Other researchers (e.g., Cousins,
1989; Dhawan et al., 1995) used more complex
coding schemes. Rhee et al. (1995; also see
Parkes, Schneider, & Bochner, 1999) devised a
scheme in which many categories are used to
code self-descriptions, but they are then aggregated to construct two indices, abstractness (as
opposed to specific) and autonomy (as opposed
to social). This coding scheme is consistent
with the current theory of attributions that distinguishes attributions of trait dispositions and
those of agency (as discussed above). Some researchers used only a subset of 20 statements
(e.g., Bochner, 1984; Cousins, 1989), whereas
others used all 20. These methodological differences may not affect conclusions about cultural
differences according to Watkins et al. (1997),
however.
The TST has other problems as well. Not
only its context-free nature (e.g., Cousins, 1989),
but also its use of the word “I” as a cue may
be problematic. As E. S. Kashima and Kashima
(1997, 1998) noted, different languages have
different sets of first-person pronouns, with
some languages having multiple first-person
pronouns (e.g., Japanese). This raises a difficult
question of which personal pronoun to use in
cross-cultural comparisons (also see Leuers &
Sonoda, 1999). No systematic investigation has
been conducted on this issue. Finally, Triandis,
Chan, Bhawuk, Iwao, and Sinha (1995) reported that, within a U.S. sample, a measure
of allocentrism (individual-level collectivism)
based on TST (S%) did not correlate with structured measures of allocentrism. This last finding suggests that psychological processes that
lead to the use of socially relevant descriptors
in self-descriptions may not be related to the
341
attitudes and values measured by structured
questionnaires.
Theoretical issues have been raised about
self-representations. From the perspective of
Deaux’s model of social identity (1993; Deaux,
Reid, Mizrahi, & Ethier, 1995), Reid and Deaux
(1996) challenged the two-basket theory of selfrepresentation. According to Deaux, self-cognitions are organized in a more integrated manner
than the two-basket theory suggests. A woman
may represent herself in terms of her role as a
sister/daughter in her family, her occupation
as a lawyer, or as the partner of her significant
other. Associated with each social identity (collective self) may be a set of psychological attributes (individual selves), such as relaxed and
smart for the sister/daughter identity and hardworking, active, and smart for the lawyer identity. The integrated model suggests, then, that
collective selves are associated with each other
and individual selves are associated with each
other, just as the two-basket theory implies. In
addition, the former postulates that some individual selves are associated with collective
selves as well.
Reid and Deaux (1996) examined Deaux’s
model in an elaborate recall experiment in three
sessions that spanned several weeks. In the first
session, each of the 57 participants in the interview listed self-defining characteristics that are
social categories or groups to which they belong
(see S. Rosenberg & Gara, 1985). One week later,
in the second session, each participant rated
the importance of the individualized list of collective and individual self-descriptions. After
a 5-minute distracter task, each participant recalled items from the list. Several weeks after
the second session, 29 of the original 57 rated
the extent to which the individual selves (psychological attributes) were associated with
each collective self (social identities). Reid and
Deaux showed that the conditional probability
measures followed a pattern similar to that of
Trafimow et al. (1991) and examined the adjusted ratio of clustering (ARC) scores, which
measure the extent to which recalled items are
clustered around a theme (Roenker, Thompson, & Brown, 1971). The ARC score for clustering around individual versus collective themes
(consistent with the two-basket theory) was .23,
and the ARC score for social identities (consistent with the integrated model) was .34. Both
scores were significantly greater than zero. Although the latter score is numerically greater
than the former, no statistical test was reported.
The conditional probabilities of recalling an
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
342
CULTURE AND SOCIAL BEHAVIOR
individual self given that a collective self has
been recalled previously and the conditional
probabilities of recalling a collective self given
that an individual self has been retrieved previously closely matched what was expected
from the degree of association between collective and individual selves as examined by a
cluster analysis (DeBoeck & Rosenberg, 1988).
The authors concluded that the results are generally more consistent with the integrated
model than with the two-basket model.
Nonetheless, the data of Trafimow and colleagues (1991, 1997) and Reid and Deaux may
need to be interpreted with caution. Recall that
both teams used conditional probabilities as
measures of memory association between
individual and collective self-cognitions. Skowronski and colleagues (Skowronski, Betz,
Sedikides, & Crawford, 1998; Skowronski &
Welbourne, 1997), however, showed that conditional probabilities may provide biased estimates of memory associations. This is because
expected conditional probabilities depend on
the total number of individual and collective
self-cognitions. The concerns of Skowronski
and colleagues’ should be addressed in future
studies. Although the ARC results obtained in
the Reid and Deaux study may not be affected
by this concern, a critical statistical test was
not reported in their study, as noted above. It
is too early to conclude definitively the validity
of these models.
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
Self-Evaluation
Positive Self-Regard Despite the centrality of
the self-esteem concept in North America, it
may not occupy as central a place for people
from East Asian cultures. Heine, Lehman, Markus, and Kitayama (1999) argued that East
Asians and Japanese in particular do not have
a strong need for positive self-regard as it is
usually conceived within contemporary social
psychology. According to them, North Americans are self-enhancing, whereas Japanese are
self-improving. North Americans seek to identify positive attributes of the self (positive abilities in particular) and attempt to maintain and
enhance self-esteem by affirmation when their
self-esteem is under threat (e.g., failing in a
task). In contrast, Japanese seek to identify discrepancies between what is ideally required of
them and what they perceive themselves to be
and attempt to improve those failings. In other
words, both North Americans and Japanese try
to reach the ideal, but the former focus on positives and try to move toward the ideal, whereas
the latter focus on negatives and try not to fall
behind.
Heine et al. (1999) suggest that self-enhancement and self-improvement are functional in
independent and interdependent cultures, respectively. In cultures in which people view
themselves as independent agents and seek to
distinguish themselves from others, it is functional to emphasize one’s uniqueness by insisting that one is above average. In contrast, in
cultures in which people view themselves as
interdependent with others and seek belongingness with their in-groups, one gains a sense
of belongingness by trying to attain the ideal
that is shared by the members of one’s significant in-group. It is adaptive in this type of
culture to try not to fall behind others rather
than to try to go beyond them. In line with
this argument, samples of Japanese individuals
have consistently revealed lower levels of selfesteem than their North American counterparts,
as gauged by M. Rosenberg’s (1965) self-esteem
measure. Corroborating this is the finding that
the self-esteem of Japanese visiting North America tends to increase, while the self-esteem of
North Americans visiting Japan tends to decline. Indeed, a number of studies suggest that
Japanese students do not exhibit the tendency
to maintain their self-esteem that their North
American counterparts do.
One such instance is the so-called unrealistic optimism bias (for reviews, see Greenwald,
1980; Taylor & Brown, 1988). Heine and Lehman (1995) showed that European Canadian
students exhibited a greater degree of optimism
relative to Japanese students. In their Study 1,
they used two methods for examining optimism
bias. One (within-group) method was to have
participants rate the likelihood of positive and
negative life events (e.g., enjoying one’s career,
becoming an alcoholic) happening to them relative to average students of the same sex in their
university. In the other (between-group) method,
one group of participants estimated the percentage chance of the events happening to
themselves without any reference to average
students, and the other group estimated the percentage chance of the same events happening
to the average same-sex student from their university. On both measures, the Canadian students showed an optimism bias by estimating
the likelihood of their enjoying positive events
to be greater and that of their suffering from
negative events to be smaller than the average
student. The Japanese students, however, did
not exhibit this pattern. In Study 2, Heine and
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
CULTURE AND SOCIAL COGNITION
Lehman examined the extent to which Canadian and Japanese students are optimistic about
negative life events relevant to independent
(e.g., becoming alcoholic) and interdependent
(e.g., making your family ashamed of you) aspects of the self. Again, Canadians showed optimism regardless, but Japanese displayed less
optimism on the within-group measure or even
pessimism on the between-group measure.
That East Asians show a lower level of
optimism than North Americans has been corroborated by other studies. In Y. T. Lee and
Seligman’s (1997) study, European Americans
showed the highest level of optimism, followed
by Chinese Americans, with mainland Chinese
exhibiting the lowest level of optimism. In a
related vein, Y. Kashima and Triandis (1986)
also showed that Japanese students exhibited
less self-serving attributions following success
and failure experiences than their American
counterparts. The Japanese students studying
in the United States attributed their failures in a
task purportedly related to intelligence to their
lack of ability more than American students.
Nevertheless, it is important to note that less
optimism among East Asians does not necessarily generalize to all collectivist cultures. Chandler, Shama, Wolf, and Planchard (1981) examined attribution styles in five cultures (India,
Japan, South Africa, the United States, and Yugoslavia) and found that self-serving attribution
patterns were present in all cultures, including
collectivist cultures such as India.
One possible interpretation of the Japanese
self-improving tendency is a modesty bias. That
is, Japanese would be privately self-enhancing
as much as Americans, but they publicly present themselves as modest by being self-effacing.
Existing evidence does not seem to support this
argument, however. For example, Kashima and
Triandis (1986) examined Japanese and American students’ attributions of success and failure
in both public and private conditions and detected no difference between them. Japanese
were self-deprecating in both conditions. This
study, nonetheless, is a weak test of this hypothesis due to its small sample size. More
recently, Heine, Takata, and Lehman (2000)
showed that the Japanese tendency to focus on
their negative performance seems to be present
even in an experimental setting in which selfpresentation appears to be difficult.
It is interesting that even Japanese people
appear to have some implicit positive selfregard. Kitayama and Karasawa (1997) showed
that Japanese students exhibited a name-letter
343
effect, in which a letter used in their own names
was regarded more positively than those letters
not used in their names, suggesting that they
have some positive regard for things associated
with themselves. Hetts et al. (1999) also reported a series of experiments in which selfregard was measured implicitly. In Study 1,
they measured the regard of Asian Americans,
European Americans, and recent Asian immigrants for their individual self and collective
self. The task was to decide whether a word
followed by a prime word such as “me” or “us”
is “good” or “bad.” The faster a good response
is relative to a bad response, the more positively
regarded is the prime word. There was a large
difference among the three groups on the positivity of “me,” with the positivity of the Asian
American group being the greatest, followed by
that of the European Americans. The regard of
the recent Asian immigrants regard for “me”
was even negative. In contrast, the order of the
groups was reversed for the positivity of “us”
(see Rhee et al., 1995, for a similar finding about
strongly acculturated Asian Americans).
In Study 2 of Hetts et al. (1999), Asian Americans, European Americans, and recent Asian
immigrants did a word completion task. Word
fragments were to be completed while the participants responded to another task designed to
prime either individual or collective self. The
number of completed word fragments that implied positive meanings relative to that which
implied negative meanings was used to measure implicit self regard. The results replicated
those of Study 1.
In Study 3 of Hetts et al. (1999), the selfregard of Japanese students was measured using the method used in Study 1. The Japanese
words “watashi” (I), “watashitachi” (we), and
“jibun” (self), were used as a prime. The results
showed that the Japanese students who had
never lived outside Japan showed a positive
response to “watashitachi” and “jibun,” but a
neutral response to “watashi.” In contrast, this
was reversed for the students who had lived
for at least 5 consecutive years in the United
States or Canada. They showed a positive regard for “watashi,” but had negative responses
to “watashitachi” and “jibun.” Throughout the
studies, however, explicit measures of selfregard (e.g., M. Rosenberg’s 1965 self-esteem
measure) did not vary across groups or covary
with implicit measures, suggesting a dissociation between implicit and explicit cognition.
Implicit measures of self-regard, however, cor-
Matsumoto, D. (2001). The handbook of culture and psychology. Retrieved from http://ebookcentral.proquest.com
Created from waldenu on 2018-04-17 14:27:43.
Copyright © 2001. Oxford University Press, Incorporated. All rights reserved.
344
CULTURE AND SOCIAL BEHAVIOR
related strongly with the duration of residence
in the United States in Studies 1 and 2.
Two types of explanations have been offered
for the cultural difference in self-regard. One
proximal explanation was provided by Higgins’
(1987; Higgins, Klein, & Strauman, 1985) selfdiscrepancy theory. One regards oneself well
to the extent that one’s perceptions of oneself
(actual self) are close to one’s ideal (ideal self).
A large discrepancy between the actual self and
ideal self causes depression and dejection. Consistent with this, Heine and Lehman (1999)
showed that Japanese students had a greater
actual-ideal discrepancy than European and
Asian Canadian students. Nevertheless, this explanation may need to be examined more fully
as they found that the correlation between actual-ideal discrepancies and Zung’s (1965) SelfReport Depression Inventory differed across
cultures, with the correlation highest for European Canadians, followed by Asian Canadians
and Japanese. This may imply that measures
have different reliability across cultures, that
other aspects of the theory can explain this cultural difference, or that the applicability of selfdiscrepancy theory differs across cultures.
A more distal explanation was provided by
Kitayama, Markus, Matsumoto, and Norasakkunkit (1997). They argued that each culture
provides situations that afford people to behave
in a way that is typical of their culture. It is
the situations available in given cultures that
perpetuate psychological differences between
them. To support this, they adopted a novel
situation-sampling method. This method involved two steps. First, Japanese and Americans described situations in which they felt
their self-esteem increased or decreased. There
were 50 descriptions randomly selected from
the descriptions generated by a (male or female)
sample from each culture (Japan or North
America) under each of the two instruction conditions (self-esteem increasing or decreasing),
for a total of 400 situation descriptions. Note
that these situations embedded a Gender ×
Culture × Condition factorial design with 50
situations in each cell. Bilingual individuals
translated Japanese situation descriptions into
English and American ones into Japanese and
edited them for cross-cultural intelligibility. In
the second step, these 400 situation descriptions served as stimuli for different samples
of Japanese (Japanese in Japan and Japanese
studying in the United States) and European
American students. Participants were asked (a)
whether their self-esteem would be affected,
(b) if so, whether it would increase or decrease,
and (c) by how much did it increase or decrease.
First, the extent to which situations were perceived to influence self-esteem was examined
by two methods. In the across-participant analysis, the proportion of situations selected out
of 50 was computed for each participant, and
an analysis was conducted using participants
as a unit of analysis. In the across-situation
analysis, the proportion of participants who selected a given situation as affecting their selfesteem was computed for each situation, and
an analysis was conducted using situations as
a uni...
Purchase answer to see full
attachment