The Nonverbal Expression of Pride: Evidence for Cross-Cultural
Jessica L. Tracy
University of British Columbia
Richard W. Robins
University of California, Davis
The present research tests whether recognition for the nonverbal expression of pride generalizes across
cultures. Study 1 provided the first evidence for cross-cultural recognition of pride, demonstrating that
the expression generalizes across Italy and the United States. Study 2 found that the pride expression
generalizes beyond Western cultures; individuals from a preliterate, highly isolated tribe in Burkina Faso,
West Africa, reliably recognized pride, regardless of whether it was displayed by African or American
targets. These Burkinabe participants were unlikely to have learned the pride expression through
cross-cultural transmission, so their recognition suggests that pride may be a human universal. Studies 3
and 4 used drawn figures to systematically manipulate the ethnicity and gender of targets showing the
expression, and demonstrated that pride recognition generalizes across male and female targets of
African, Asian, and Caucasian descent. Discussion focuses on the implications of the findings for the
universality of the pride expression.
Keywords: pride, self-conscious emotion, cross-cultural, universal, emotion recognition
Of all the complex emotions . . . pride, perhaps, is the most plainly
—Darwin, The Expression of the Emotions in Man and Animals
One of the major findings in the behavioral and social sciences
is the discovery that a small set of basic emotions—anger, disgust,
fear, happiness, sadness, and surprise— have distinct, universally
recognized nonverbal expressions (Ekman & Friesen, 1971; Ek-
man, Sorenson, & Friesen, 1969; Izard, 1971). This finding, which
emerged from Ekman and colleagues’ (Ekman et al., 1969; Ekman
& Friesen, 1971) seminal research on highly isolated preliterate
tribal groups in Papua New Guinea, led many researchers to accept
Darwin’s (1872) view that emotions and their expressions are a
universal, evolved part of human nature.
Over the past few decades, several attempts have been made to
identify additional universally recognized emotion expressions,
but, with a few possible exceptions (contempt, embarrassment, and
shame; Ekman & Friesen, 1986; Haidt & Keltner, 1999; Izard,
1971; Russell, 1991), these attempts have failed. As a result, the
six basic emotions have a special status in the emotion literature.
In contrast, emotions assumed to lack universal expressions have
received considerably less attention within the field of affective
science. In particular, researchers have largely neglected the self-
conscious (or social) emotions, such as embarrassment, pride, and
shame. These emotions involve complex self-evaluative processes
(Tangney & Dearing, 2002; Tracy & Robins, 2004a), emerge later
in development than do basic emotions (M. Lewis, Alessandri, &
Sullivan, 1992), are thought to be unique to humans and possibly
the great apes (Hart & Karmel, 1996), and play a central role in
status seeking, dominance, and other fundamental social behaviors
(Keltner & Buswell, 1997; Tracy & Robins, 2007c).
However, recent research suggests that the self-conscious emo-
tion of pride may have a recognizable nonverbal expression. The
pride expression, shown in Figure 1, includes the body (i.e.,
expanded posture, head tilted back) as well as the face (i.e., low
intensity, non-Duchenne smile); it has been shown to be reliably
recognized and distinguished from similar emotions (e.g., happi-
ness) using forced-choice and open-ended response methods; and
it is reliably recognized by American adults and children as young
as 4 years old (Tracy & Robins, 2004b; Tracy, Robins, & Lagat-
tuta, 2005). Pride recognition rates are comparable with recogni-
tion rates for the basic emotions, and, like the basic emotions, pride
can be recognized from a single snapshot image very quickly and
efficiently (Tracy & Robins, in press). Moreover, similar displays
have been documented in spontaneous nonverbal behaviors shown
in response to a pride-eliciting event, such as successful comple-
tion of a task (Belsky & Domitrovich, 1997; M. Lewis et al., 1992;
Stipek, Recchia, & McClintic, 1992; Weisfeld & Beresford, 1982)
and victory at the Olympic Games (Tracy & Matsumoto, 2007).
Together, these findings suggest that pride may merit inclusion
in the class of emotions thought to be universal and to have
Jessica L. Tracy, Department of Psychology, University of British
Columbia, Vancouver, British Columbia, Canada; Richard W. Robins,
Department of Psychology, University of California, Davis.
We thank the National Institute of Mental Health (Grant MH20006), the
National Institute of Aging (Grant AG022057), and the Social Sciences and
Humanities Research Council of Canada (File #410-2006-1593) for their
support. We also thank Jill Bendziewicz, Simona Ghetti, Jean Traore, and
Maggie Traore for their assistance in collecting the data for Studies 1 and
2; Michael Biehl and Dory Schachner for their assistance in Studies 3 and
4; and Erika Rosenberg for her assistance in FACS coding the stimuli. In
addition, Jessica L. Tracy thanks the members of her dissertation commit-
tee for their helpful comments on the article, research methods, and study
design: Paul Ekman, Bob Emmons, Phil Shaver, and Dean Simonton.
Correspondence concerning this article should be addressed Jessica L.
Tracy, Department of Psychology, University of British Columbia, 2136
West Mall, Vancouver, British Columbia V6T 1Z4, Canada. E-mail:
Journal of Personality and Social Psychology Copyright 2008 by the American Psychological Association
2008, Vol. 94, No. 3, 516–530 0022-3514/08/$12.00 DOI: 10.1037/0022-35126.96.36.1996
evolved to serve specific adaptive functions, such as signaling
survival-relevant messages to one’s kin. Pride, in particular, is
likely to play a functional role in the maintenance and enhance-
ment of social status, an essential component of a social animal’s
fitness. Individuals experience pride after a socially valued
achievement, and these feelings may alert them that their behavior
(or self) is valued by others and that they are thus unlikely to be
rejected by the group and may deserve increased status (Leary,
Tambor, Terdal, & Downs, 1995). These feelings may also rein-
force the socially valued behaviors that generated the emotion
(Hart & Matsuba, 2007; Herrald & Tomaka, 2002; Weiner, 1985).
At the same time, the interpersonal nonverbal expression of pride
may serve a complementary adaptive function by alerting others
that the proud individual merits increased acceptance and status.
Despite theoretical reasons for expecting pride and its associated
expression to be an adaptive part of human nature, however,
studies have not yet addressed the critical question that lies at the
heart of this issue: Does the pride expression generalize across
cultures? In contrast, four lines of research have been used to
support the claim of universality in the basic emotions. First,
Ekman and colleagues (1969; Ekman & Friesen, 1971) demon-
strated that expressions documented in the United States were
reliably recognized by highly isolated, preliterate tribal individuals
from Papua New Guinea. This finding is considered by many to be
one of strongest pieces of evidence supporting the case for uni-
versality and the concept of innate human nature, because Ekman’s
preliterate, culturally isolated participants could not have learned
the emotion expressions through cross-cultural transmission (e.g.,
films, television, magazines; D. E. Brown, 1991; D. E. Buss, 1992;
Pinker, 2002; but see Russell, 1994). Second, at the same time
Ekman’s groundbreaking work was being conducted in Papua
New Guinea, other studies were being conducted in nations all
over the world demonstrating that individuals speaking a range of
languages and steeped in a wide variety of cultural traditions
reliably recognized the basic emotion expressions (Ekman et al.,
1969, 1987; Izard, 1971). Third, more recent studies have demon-
strated that emotion recognition generalizes across targets of dif-
ferent cultures and ethnicities (e.g., Beaupre & Hess, 2005; Biehl
et al., 1997). These studies manipulated the ethnicity of targets
showing the expression and found reliable recognition of the basic
emotion expressions across all ethnicities examined (African, Cau-
casian, Chinese, Japanese); these findings also held across several
perceiver cultures (e.g., Canada, Hungary, Japan, Poland, Sumatra,
United States, Vietnam; Beaupre & Hess, 2005; Biehl et al., 1997;
Matsumoto & Ekman, 1989). Fourth, a growing number of studies
have examined the spontaneous production of emotion expressions
across cultures and found that individuals from a diverse range of
cultures (e.g., North American, East Asian, South and Central
American, Western European, Eastern European) tend to show
similar emotion expressions in similar situations, such as watching
a stressful video or winning or losing a judo match in the Olympic
games (e.g., Ekman, 1971; Matsumoto & Willingham, 2006).
These studies suggest that emotion expressions are recognized
across cultures because they are experienced and displayed across
However, with the exception of production studies examining
the spontaneous display of pride, researchers have yet to conduct
similar cross-cultural or cross-ethnicity studies on recognition of
the pride expression. In fact, in a comprehensive meta-analysis of
emotion recognition studies, Elfenbein and Ambady (2002) iden-
tified cross-cultural studies examining over 36 different emotion
expressions, but none searching for a pride expression. Thus, we
do not know whether pride is recognized outside of the United
States, let alone by nonliterate, highly isolated individuals who
have not been exposed to Western culture. We also do not know
whether pride is recognized across targets of different ethnic or
racial groups; all previous studies on the pride expression have
used only Caucasian American targets. If the pride expression
evolved to communicate fitness-promoting messages about an
individual’s status, then we would expect to see cross-cultural
evidence for recognition. In fact, if the pride expression is an
evolved part of human nature, then it should be recognized across
cultures that have had no contact with each other. If individuals in
such disparate cultures can recognize pride, that would provide
strong support for the claim that the pride expression is a human
universal, because, in the absence of cross-cultural contact, the
most likely reason these cultures would share the ability to recog-
nize pride is that this ability is part of human nature.
It is also possible, however, that the pride expression is univer-
sal but its specific components show minor variations across
cultures and that these variations produce lower recognition rates
when the expression is transplanted, in an etic fashion, from one
culture to another. Some authors have argued that even the basic
emotion expressions are influenced to some degree by culture-
specific processes, which are said to produce dialects in expres-
sions that are otherwise universal. According to this perspective,
which has received mixed support, these dialects seem to influence
cross-cultural recognition levels of basic emotion expressions,
such that expressions are recognized at higher rates when shown to
individuals who belong to the culture from which the expression
emerged (Elfenbein & Ambady, 2002; but see Matsumoto, 2002).
Figure 1. The prototypical pride expression. The expression includes a
small smile, head tilted slightly (approximately 20°) back, expanded pos-
ture, and arms akimbo with hands on hips. The individual pictured in the
photo gave consent for his likeness to be published in this article.
A third possibility is that pride will not be recognized in certain
cultures because the experience and expression of pride vary
across cultures in fundamental ways. As a self-conscious emotion,
pride is a highly social emotion elicited by complex self-evaluative
processes and thus may show greater cross-cultural variability than
emotions like fear and happiness, which are intrinsically less social
in nature. Research suggests that at least some self-evaluative
processes vary across cultures; for example, there is evidence that
cultures with a collectivistic orientation place less value on explicit
self-enhancement than do more individualistic cultures (Heine,
Kitayama, & Hamamura, 2007; but see Sedikides, Gaertner, &
Vevea, 2005). In fact, studies have found that pride is viewed less
positively in collectivistic than in individualistic cultures (Eid &
Diener, 2001) and is less typically experienced in response to
successes of the personal self than the collective self (Stipek,
1998). Consequently, in collectivistic cultures, the pride expres-
sion may be infrequently displayed in everyday social contexts and
more highly regulated when it is displayed. Thus, it is unclear
whether the pride expression documented in the United States
would generalize to more collectivistic cultures.
Finally, a fourth possibility is that the pride expression docu-
mented thus far is actually a culture-specific gesture, much like the
“thumbs up” sign, which serves a learned communicative function
within the United States but is not a universal part of the human
If this is the case, then the pride expression
may generalize to cultures that have some exposure to American
culture but not to highly isolated cultures that have little contact
with American culture or media.
The present research tests these competing possibilities. If we
find evidence that pride can be recognized across cultures, it would
strengthen the case that pride should be placed within a function-
alist framework, given the longstanding assumption that emotions
evolved, in part, to serve communicative functions through their
associated, ritualized nonverbal signals (Ekman, 1992). This find-
ing would also challenge two extant assumptions in the emotion
literature: (a) All positive emotions share a single nonverbal ex-
pression (Ekman, 1992; Fredrickson & Branigan, 2001; Izard &
Haynes, 1988), and (b) cognitively complex, phylogenetically
recent emotions, such as the self-conscious emotions, are not
associated with distinct nonverbal expressions.
In Study 1, we tested whether individuals born and raised
outside the United States could recognize the pride expression.
Specifically, we examined recognition among a sample of Italians
who neither spoke nor understood English. In Study 2, we exam-
ined pride recognition among a sample of preliterate West Afri-
cans living in rural villages in Burkina Faso who had no known
exposure to Western culture. In this study, we also tested whether
shame, another fundamental self-conscious emotion previously
found to have a nonverbal expression that is recognized within the
United States (Izard, 1971; Keltner, 1995), is recognized in an
isolated, preliterate culture. In Studies 3 and 4, we turned our
attention to the cross-ethnic generalizability of individuals who
display the pride expression (i.e., targets). These two studies tested
whether pride is recognized in male and female targets from three
ethnic groups (African, Asian, and European) and whether the
level of pride recognition varies by target and perceiver gender and
The present research extends previous findings in several ways.
First, this is the first set of studies to test whether the pride
expression is reliably recognized outside the United States. Sec-
ond, this is the first research to test whether shame is recognized
by preliterate individuals from a highly isolated tribal culture. All
previous research on the shame expression has examined recogni-
tion among participants who are at least somewhat exposed to
Western culture (e.g., have a high school education, live in or a
near a populous city, etc.). Third, this is the first research to
examine recognition of the basic emotion expressions in a prelit-
erate, highly isolated non-Western culture using a modified forced-
choice format that includes response options such as none of these
is correct, which addresses several concerns associated with the
more traditional strict forced-choice response format (Frank &
Stennett, 2001). Furthermore, our preliterate participants were also
given the option of identifying expressions in an open-ended
manner rather than choosing any of the responses provided; the
open-ended format further addresses concerns associated with the
forced-choice method (Russell, 1994). Fourth, this is the first
research to test whether pride is reliably recognized when shown
by male and female targets of varying ethnicities and to examine
whether target gender and ethnicity affect the level of pride rec-
Participants. Twenty-eight individuals (50% women) ranging
in age from 22 to 38 years (M⫽27 years) living in several
community locations in Bologna, Italy, were recruited through an
advertisement at a research institution in Bologna. Interested in-
dividuals were told that they must have been born and raised in
Italy, and must not be able to speak or understand English. The
study was conducted at the research institution by an Italian
Stimuli. Four targets, a male and female Asian-American and
a male and female Caucasian American (age range ⫽22–32 years)
posed the pride expression shown in Figure 1 as well as expres-
sions of contempt, happiness, and surprise. All photos were taken
from the waist up. Targets wore identical white shirts and posed in
front of a plain blue background. Posing instructions were based
on the Directed Facial Action task (DFA; Ekman, Levenson, &
Friesen, 1983) for all emotions other than pride. Posing instruc-
tions for pride were based on previous research on the pride
expression (Tracy & Robins, 2004b).
Procedure. Participants were seated in front of a laptop com-
puter and were shown each photo on the 14-inch (35.56-cm)
monitor for 30 s. For each photo, participants were asked to choose
the emotion that “best matches the emotion expressed by the
person in the photo” from the following options: contempt
(disprezzo), boredom (noia), excitement (eccitazione), happiness
(gioia and felicita`), pride (fierezza and orgoglio), surprise (sor-
presa), none of these is correct (nessuna di queste emozioni
descrive correttamenta la fotografia), and other: ___ (altro: si
prega di specificare ___). Boredom and excitement were included
because these words are rarely but occasionally applied to the
One example of a learned gesture that communicates a distinct emotion
within a particular culture but has failed to pass the universality test is the
“tongue bite” display of shame in India (Haidt & Keltner, 1999).
518 TRACY AND ROBINS
pride expression in open-ended recognition studies of pride (Tracy
& Robins, 2004b). The none of these is correct option was in-
cluded because it has been shown to address concerns associated
with the standard forced-choice format (Frank & Stennett, 2001).
Specifically, participants will use this option to label nonsense
expressions that have no known emotion word label and to label
known expressions when the correct emotion word is not provided.
The other option was included to allow participants to respond in
an open-ended manner, to further address concerns associated with
the forced-choice format (Matsumoto et al., 2002; Russell, 1994).
The rating form was translated into Italian by a bilingual Italian
psychologist and was independently confirmed by back-translating
the response options using an Italian–English dictionary (Dizionari
Garzanti, 1997). Pride has two translations in the Italian language,
fierezza and orgoglio, and both were included as options.
avoid drawing attention to pride, two translations for happiness,
felicita` and gioia, were also included.
Results and Discussion
All four pride photos included were identified as pride (either
fierezza or orgoglio) at greater-than-chance frequencies (M⫽
78%, range ⫽61%– 89%; p⬍.05), based on binomial tests with
chance set at 33%.
Judges were slightly more likely to label pride
photos as fierezza (56%) than orgoglio (23%; p⬍.05), but they
tended to use the two words interchangeably across photos. Figure
2 shows mean recognition rates for pride, happiness, surprise, and
contempt. Mean pride recognition did not differ from mean rec-
ognition rates for happiness or surprise, t(27) ⫽0.97 and t(27) ⫽
0.83, ns, respectively, but was significantly greater than mean
recognition for contempt, t(27) ⫽5.24, p⬍.05. For a cross-
cultural comparison, Figure 2 also presents recognition rates from
an American sample of 75 undergraduate research participants
(69% women) who responded to the same stimuli following the
same procedures. As can be seen, the pride recognition rate did not
differ between the two groups, t(48) ⫽0.86, ns.
We found no effect of target gender or ethnicity on level of pride
recognition, t(27) ⫽1.49 for ethnicity and t(27) ⫽0.30 for gender,
both ns, but we cannot draw any conclusions about target effects
from this null finding, given the small number of targets included.
When pride was not correctly recognized, it was identified as
happiness (22% of errors, 4% of overall responses), contempt (9%
of errors, 2% overall), excitement (4% of errors, 0.90% overall),
none of these (4% of errors, 0.90% overall), and other (60% of
errors, 12% overall). When participants chose other for the pride
expression, they were most likely to label it as “satisfied” (4%),
“carefree” (3%), “challenging” (2%), and “determined” (2%); sev-
eral other labels were applied a single time by a single observer
(e.g., “superior,” “self-certain,” “wisdom,” etc.). Participants
rarely labeled other emotion expressions as pride; false-alarm rates
were 5% (happiness), 5% (contempt), and 0.90% (surprise).
The findings from Study 1 demonstrate that the pride expression
generalizes outside American culture. Italian participants recog-
nized pride at rates similar to those found in the United States.
Combined with previous findings from U.S. samples, this study
suggests that there is a reliably recognized pride expression within
Western culture. However, it remains possible that the pride ex-
pression documented in these studies is not a universal aspect of
human nature, but rather is a learned gesture specific to Western
cultures, similar to the “thumbs up” sign. To test this possibility,
we must examine whether pride is recognized by individuals who
have not been exposed to Western culture. Given the global
ubiquity of Western culture and Western media, it is important to
examine pride recognition among individuals who have little or no
access to Western media, to ensure that they could not have
learned the expression through cross-cultural transmission (i.e.,
television, films, magazines, newspapers). Study 2 provides the
first such test by examining whether pride is recognized by non-
literate individuals from rural and remote tribal villages in Burkina
Faso, West Africa.
Burkina Faso is an ideal location to examine questions of
universality because it is one of the poorest and most isolated
countries in the world. It has the world’s second highest illiteracy
rate (75%) and was recently ranked by a United Nations Human
Development Report (2005) as the third least-developed country in
the world, based on life expectancy, educational attainment, and
income. At the time the research was conducted, there was only
one television station in the country and no electricity (and thus no
Italian-English dictionaries translate both terms as “pride.” According
to the Italian psychologist who conducted the research, the two terms have
overlapping meanings in many contexts, but in some contexts fierezza
includes a sense of dignity and honor whereas orgoglio implies excessive
pride. It is thus possible that the terms reflect the distinct meanings of
authentic versus hubristic pride (Tracy & Robins, 2007b), or beta versus
alpha pride (Tangney, Wagner, & Gramzow, 1989), but further research is
needed to support this speculation.
This chance rate is more stringent than that typically accepted within
the emotion literature, which is based on the number of options presented
(e.g., 17% in the present study; Hertenstein, Keltner, App, Bulleit, &
Jaskolka, 2006). It is also more stringent than the rate proposed by critics
of forced-choice emotion recognition studies; these critics have argued that
participants in such studies do not guess randomly among the options
presented but rather choose from four true emotion options defined by the
two orthogonal dimensions of arousal and valence (Russell, 1994), sug-
gesting a “true” chance guessing rate of 25%.
Pride Happiness Surprise Contempt
Figure 2. Mean recognition rates for pride, happiness, surprise, and
contempt expressions in Italian (N⫽28) and American (N⫽75) samples
for Study 1.
televisions) in the research participants’ villages, which are not
located near tourist destinations. All of these factors contribute to
participants’ isolation from Western culture, allowing for an ef-
fective test of the universality question. If these individuals rec-
ognize the pride expression, then it is unlikely to be because they
learned it through cross-cultural transmission.
Burkina Faso is also an ideal culture to test the generalizability
of the pride expression because African countries tend to have
highly collectivistic cultural values (Hoftstede, 1984), which con-
trast sharply with the more individualistic values of most Western
cultures. Perceptions of emotions and self-processes relevant to
pride (e.g., self-enhancement) differ across these two types of
cultures (Heine et al., 2007; Markus & Kitayama, 1991). In par-
ticular, collectivistic cultures tend to promote the group over the
individual, such that individuals in these cultures are more prone to
accept status differences rather than try to change them and assert
the self (Hoftstede, 2001; Rossier, Dahourou, & McCrae, 2005).
These values may be inconsistent with pride, an emotion typically
geared toward enhancing and affirming the self; this may be one
reason why individuals from these cultures tend to place less value
on pride (Eid & Diener, 2001). Thus, evidence for pride recogni-
tion in Burkina Faso would suggest that the emotion transcends a
fundamental cultural difference.
We also tested whether Burkinabe participants could recognize
the nonverbal expression of a second self-conscious emotion,
shame. Previous studies suggest that individuals from the United
States, England, France, Germany, Greece, Japan, Sweden, and
Switzerland can recognize a shame expression (Izard, 1971; Kelt-
ner, 1995), but there have been no studies examining this issue
among nonliterate, highly isolated non-Western individuals. More-
over, in one of the few studies conducted in a non-Western, less
industrialized culture (India), Haidt and Keltner (1999) failed to
find better-than-chance recognition for the most commonly studied
version of the shame expression, despite the fact that their partic-
ipants had some formal education and over half were university
Thus, it remains unclear whether the expression of
shame, an emotion considered to be the antithesis of pride (Dar-
win, 1872), generalizes to preindustrialized (and therefore gener-
ally less Westernized) cultures.
Participants. Thirty-nine individuals (68% women) ranging in
age from 20 to 75 years (Mdn ⫽46 years) were recruited by word
of mouth. All participants lived in small rural settlements approx-
imately 5 miles (8 km) from the village of Toussianna, where the
research took place. The majority of participants were subsistence
farmers, and all participants inhabited mud huts with no electricity
or plumbing. None of the participants had ever attended school,
none could read or write in any language, and none could speak or
understand any language other than the language of their tribe,
Dioula, which has no formal written form. Each participant was
paid 5 kilograms of rice and the monetary equivalent of $2.00
Interview procedure. Participants were interviewed individu-
ally by one of four interviewers, always of the participant’s gender.
All interviewers were born in Toussianna but had left the village to
attend school in a larger town or city in Burkina Faso. Interviewers
were literate in French (the language taught in Burkinabe schools)
and were trained on the research procedures in French before the
study was conducted. The interview protocol was translated from
English to French by a bilingual (French–English) researcher from
Burkina Faso, who ensured that all phrasing matched Burkinabe
linguistic customs, and was back-translated by a bilingual
(English–French) psychologist. During the interview, the inter-
viewers read the instructions in French and translated each ques-
tion aloud into Dioula. Interviewers recorded participants’ verbal
responses on a written form.
Prior to conducting the research, all four interviewers and our
Burkinabe collaborator, a California-university-educated political
leader who led the research team, discussed and reached consensus
on the correct Dioula translations for each emotion term and for
the general instructions. Interviewers were trained to remember the
particular wording of these translations, given that Dioula lacks a
written form. We observed this training session with a translator
and were satisfied that complete agreement was reached.
Test of familiarity with Western culture. To test participants’
familiarity with Western culture, we showed them five 4-inch ⫻
6-inch (10.16-cm ⫻15.24-cm) laminated photographs of well-
known Western individuals (David Beckham, Tony Blair, George
W. Bush, Tom Cruise, Michael Jordan) who represent different
cultural and national backgrounds and ethnicities and who ac-
quired their fame in diverse ways (i.e., sports, politics, and film).
We also included laminated 4-inch ⫻6-inch (10.16-cm ⫻15.24-
cm) photographs of two well-known individuals from Burkina
Faso (President Blaise Campore and former revolutionary leader
Thomas Sankara) to test participants’ familiarity with their own
culture (pictures of these two individuals are displayed in villages
throughout the country) and to ensure that they understood the
recognition task. For each photograph, participants were asked,
“Who is the person in this photograph?” Responses were scored as
correct if participants named the person or their title (e.g., “Blaise
Campore” or “President of Burkina Faso”).
Test of emotion recognition. Four targets—a male and female
West African and a male and female Caucasian American—were
photographed from the waist up posing the pride expression shown
in Figure 1. This expression was found in previous research to be
the best recognized, or most prototypical, version of pride (Tracy
& Robins, 2007a). Two of these targets (the male Caucasian
American and the female West African) were also photographed
posing anger, disgust, fear, happiness, sadness, shame, and sur-
Posing instructions were based on the DFA (Ekman et al.,
1983) for the basic emotions and on previous research (Heerey,
Haidt and Keltner (1999) did find reliable recognition for an alternate
version of the shame expression, the “face cover.” In the present study,
limited access to participants reduced the number of stimuli we could
include, so we assessed shame recognition with the more frequently studied
variant of the expression (described in the Method section of Study 2),
found to be reliably recognized in several cultures.
Limited access to participants reduced the number of stimuli we could
include, so we opted to include only two exemplars of each nonpride
520 TRACY AND ROBINS
Keltner, & Capps, 2003) for shame.
A leading expert in the Facial
Action Coding System (FACS), Erika Rosenberg, assisted in the
posing session (which occurred in the United States) and verified
that all expressions included the correct facial muscle movements
according to the Emotion Facial Action Coding System (EM-
FACS; Rosenberg & Ekman, 1995) and no other movements.
To address the main research question, we showed participants
each photo and asked, “Which of the emotions listed below best
matches the emotion expressed by the person in the photo?” They
then were read the following list of options: anger,disgust,fear,
happiness,pride,sadness,shame, and surprise, as well as the
options I don’t know and another emotion. The last two options
were modifications of the procedure used in Study 1 to address
concerns about participants’ lack of familiarity with research pro-
tocol. Choosing the none of these are correct option requires
participants to understand that experimenters may not provide the
correct answer or that there is no correct answer, and our
Burkinabe English–French translator suggested that these partici-
pants would feel more comfortable choosing I don’t know in cases
where they had no answer.
Familiarity with Western culture. None of the participants
accurately identified any of the five well-known Western individ-
uals, but 74% of participants recognized at least one of the two
well-known Burkinabes: President Campore (69%) and Thomas
Sankara (51%). Thus, participants’ failure to recognize the West-
erners was likely due to their lack of familiarity with Western
culture rather than to a failure to understand the task. Combined
with other unique features of this sample (i.e., their illiteracy, lack
of formal education, inability to speak or understand French),
participants’ lack of familiarity with famous Western figures
makes it very unlikely that they had any familiarity with the pride
expression as shown by Western figures in the media. The only
remaining possibility of Western exposure is through tourism, but
Toussianna is not a tourist destination (in fact, the only bus that
goes near the village has no official stop there), and participants
lived in villages located at least several miles outside of Tous-
sianna that were not accessible by public transportation (partici-
pants arrived at the research site by foot).
Emotion recognition. Figure 3 shows mean recognition rates
for all emotion expressions. Participants recognized every self-
conscious and basic emotion expression at greater-than-chance
frequencies ( p⬍.05), based on binomial tests with chance set at
12.5% (based on the number of emotion response options pro-
vided). In this study, we followed the protocol set out by Ekman et
al. (1969) for studying preliterate cultures: We based chance on the
number of options provided, such that a better-than-chance re-
sponse indicates that participants were not guessing randomly
among the options. Several unique features of this study, and of
Ekman et al.’s study, support this decision: participants’ complete
lack of familiarity with research procedures, the potential for
inaccuracies when translating to a spoken dialect, the difficulty of
the task for individuals who have no formal education and are not
accustomed to viewing photographs of strangers, and, in the
present study, the fact that we included response options that
directly address several of the limitations of the standard forced-
choice response format. However, we also conducted binomial
tests using the more stringent rate of 33%, which was used in
Study 1. Using this more stringent chance rate, the recognition
rates for disgust, happiness, pride, sadness, and surprise remained
significantly greater than chance ( p⬍.05).
The pride and shame recognition levels did not differ signifi-
cantly from recognition levels for any of the basic emotions except
happiness (84%), which, consistent with previous cross-cultural
research (Elfenbein & Ambady, 2002), was better recognized than
every other emotion (all ps⬍.05). The mean recognition rate
averaged across the two self-conscious emotions (46%) was com-
parable with the mean recognition rate averaged across the six
basic emotions (50%).
Recognition levels did not differ significantly for male (48%)
versus female (52%) participants, t(36) ⫽1.45, ns, or for male
(55%) versus female (49%) targets, t(38) ⫽1.96, ns. However,
expressions posed by Caucasian American targets (54%) were
better recognized than expressions posed by West African targets
(45%), t(38) ⫽3.78, p⬍.05. Given the limited number of targets
used, this finding is likely due to differences in the facial physi-
ognomy or posing skills of these particular individuals, especially
because there is no evidence for an out-group bias in emotion
recognition when the perceivers are not members of a minority
group (Beaupre & Hess, 2005; Elfenbein & Ambady, 2002). There
were no target gender or ethnicity effects for the self-conscious
emotions, t(38) ⫽0.68 and t(38) ⫽0.56, ns, respectively, and no
interactions between participant gender and target gender or target
ethnicity, F(1, 36) ⫽0.60 and F(1, 36) ⫽1.24, ns, respectively.
When pride was not correctly recognized, it was identified as
happiness (52% of errors, 22% of overall responses), sadness (15%
of errors, 6% overall), surprise (13% of errors, 6% overall), fear
(4% of errors, 2% overall), anger (3% of errors, 1% overall),
shame (3% of errors, 1% overall), disgust (1% of errors, 0.64%
The shame expression that was posed for Study 2 showed a head tilt
downward and eye-gaze downward; this expression has previously been
associated with shame (Keltner, 1995).
34% 33% 30%
Happiness Surprise Pride Sadness Dis
ust Shame An
Figure 3. Mean recognition rates for self-conscious (gray bars) and basic
emotion expressions in Burkina Faso, Africa (Study 2; N⫽39). The
dashed line represents the recognition rate that would occur by chance, if
participants guessed randomly among the emotion label response options
presented (12.5%). Recognition rates for all emotions were significantly
greater than chance, p⬍.05.
overall), other (4% of errors, 2% overall), and I don’t know (3% of
errors, 1% overall). When shame was not correctly recognized, it
was identified as sadness (64% of errors, 42% of overall re-
sponses), disgust (8% of errors, 5% overall), anger (6% of errors,
4% overall), fear (6% of errors, 4% overall), surprise (4% of
errors, 2% overall), and other (10% of errors, 6% overall).
The false-alarm rates for both pride and shame were fairly low.
Participants occasionally labeled other emotions as pride: happi-
ness (13% of responses to happy expressions), fear (11%), sadness
(5%), disgust (4%), surprise (2%), and anger (1%). False alarms
for shame were lower: 3% of responses to disgust, 3% to anger,
and 1% to pride. Figure 4 shows the hit rate (number of times an
expression was correctly identified) divided by the base rate (over-
all number of times that response option was used) for each
emotion. Both pride and shame hit rate/base rate proportions were
among the three highest. Thus, when individuals used the pride
and shame response options, they tended to use them correctly.
The findings of Study 2 suggest that pride is reliably recognized
and distinguished from related emotions, including happiness, by
nonliterate, culturally isolated, non-Western individuals. Further-
more, these findings suggest that shame, too, is reliably recognized
by these individuals. This study thus provides the strongest evi-
dence to date that the self-conscious emotions of pride and shame
meet the primary criterion for universality that exists within the
emotion literature (Ekman, 1992).
Nonetheless, it is noteworthy that recognition rates for all emo-
tions were lower than rates typically found in educated Western
samples (Elfenbein & Ambady, 2002). This difference, which also
holds for the seminal studies of Ekman et al. (1969) and for other
recognition studies conducted on individuals from Africa (e.g.,
Beaupre & Hess, 2005), may be due to complexities associated
with implementing the research procedure in a preliterate culture,
such as translating an interview protocol into a spoken dialect, the
fact that preliterate participants must remember response options
rather than read them, and the overall novelty of the task.
However, lower recognition rates may also reflect important
cultural differences. Recent research suggests that there may be
cultural dialects, or subtle variations, in otherwise universal emo-
tion expressions, and these dialects may facilitate recognition in
the culture from which the particular expression emerged but
hinder it elsewhere (Elfenbein & Ambady, 2002; but see Matsu-
moto, 2002). Supporting this interpretation, other studies have
found particularly low recognition of the Western-derived fear
expression in non-Western cultures; for example, fear recognition
rates in educated Japanese samples are similar to those found in
our isolated Burkinabe sample (Matsumoto et al., 2002), and
similar rates have been found among educated Africans living in
Canada (Beaupre & Hess, 2005). Dialect theory would suggest that
cultural factors have influenced the way fear is expressed, such
that prototypical versions of fear differ across cultures. Future
studies should explore this possibility by complementing our etic
approach (i.e., transporting research procedures and stimuli from
one culture to another) with an emic approach that tests whether
pride and shame expressions derived in Burkina Faso are better
recognized than the Western-derived stimuli used here; such stud-
ies must fully cross target and perceiver ethnicity to directly test
dialect theory (Matsumoto, 2002).
Such an approach is particularly important for the case of
shame, given its lower level of recognition in the present study and
the previous finding that it is more reliably identified in at least one
culture as a “face cover” than in its typical form in Western culture
(Haidt & Keltner, 1999). Together, these findings suggest that the
shame expression critically requires hiding of the face— either
turning it down or covering it with the hands—which fits with
longstanding theoretical views that shame promotes a desire to
hide or escape (H. B. Lewis, 1971; Tangney & Dearing, 2002).
Future research should examine whether distinct variants of the
expression are associated with meaningfully distinct aspects of the
emotion (e.g., shame vs. embarrassment) or whether cultural dia-
lects emerge from the unique social rules and prescriptions for
displaying shame in each culture.
The findings of Study 2 also provide further support for the
universality of the basic emotion expressions. No previous study in
a preliterate, highly isolated non-Western culture has found rec-
ognition for the basic emotions using any assessment method other
than the traditional strict forced-choice format. Our use of the
modified forced-choice format addresses several major limitations
We also calculated unbiased hit rates, which take into account both
error and base rates (Wagner, 1993). Unbiased hit rates were 0.45 (pride),
0.45 (happiness), 0.29 (disgust), 0.29 (shame), 0.29 (surprise), 0.16 (sad-
ness), and 0.10 (fear).
It is interesting to note that the self-conscious emotion expressions,
which all seem to require gross body or head movements, are generally
more “crude” than the basic emotion expressions, which require fine-
grained facial musculature movements. This distinction appears to be
somewhat paradoxical, given that self-conscious emotions are assumed to
have emerged more recently in evolutionary history than the basic emo-
tions and require more complex cognitive processes, a higher level of
self-complexity, and so on. The explanation may lie in the relative com-
municative range of the two types of expressions. Gross body-based
expressions are visible from a farther distance than face-only expressions,
and self-conscious emotion expressions may have evolved for the direct
purpose of communication (e.g., signaling status), whereas basic emotion
expressions are typically assumed to have evolved for other original
functions but to have become ritualized for the purpose of communication
(Eibl-Eibesfeldt, 1989; Tracy & Robins, 2007a).
Figure 4. Mean hit (accuracy) rates divided by base rates for each
emotion in Burkina Faso, Africa (Study 2; N⫽39).
522 TRACY AND ROBINS
of the traditional format, including the possibility that the strict
forced-choice format inflates agreement because participants must
choose one of the options provided. Study 2 demonstrated that
highly isolated individuals will agree on labels for the basic
emotion expressions even when given the option to say “I don’t
know” or to say that the expression conveys some “other” emotion.
In other words, these participants’ agreement on the correct emo-
tion labels for anger, disgust, fear, happiness, sadness, and sur-
prise, as well as pride and shame, is unlikely to be inflated by an
artificial restriction of choices.
Studies 1 and 2 demonstrated that pride recognition generalizes
across perceivers of vastly different cultures and, as was shown in
Study 2, this effect is unlikely to result from cross-cultural trans-
mission. Furthermore, other research has found that the recogniz-
able pride expression is spontaneously displayed by individuals
across highly divergent cultures in response to success (Tracy &
Matsumoto, 2007). In this study, individuals from 36 nations,
representing three distinct culture groups, were shown to display
the pride expression in response to the same pride-eliciting
event—winning a judo match in the Olympic Games; this finding
held within all three cultures. As a result, the primary remaining
question regarding the cross-cultural generalizability of the pride
expression is whether recognition generalizes across targets of
different cultures or, given that culture is not readily perceived by
simply observing targets, different ethnicities. Can observers iden-
tify the pride expression in others, regardless of those others’
ethnicity and gender? Does target ethnicity or gender influence the
level of pride recognition? Study 3 addresses these questions.
We had several hypotheses for this study. First, according to the
universality hypothesis, if the pride expression is a universal part
of human nature that functions to send an adaptive signal, then it
should be recognized when shown by all humans of both genders
and all ethnicities. In fact, this seems to be the case for the basic
emotions: Studies that have systematically varied target gender
and ethnicity have typically found few, if any, effects on recogni-
tion (Boucher & Carlson, 1980; Keltner, 1995; Massaro & Ellison,
1996; Matsumoto & Ekman, 1989).
Second, it is possible that pride recognition will be significantly
greater than chance for all targets, supporting the universality
hypothesis, but mean recognition levels will vary by target char-
acteristics. Specifically, according to the perceived status hypoth-
esis, there should be meaningful differences in recognition levels
for emotions that signal status when shown by targets perceived to
differ in status (Keltner, 1995). If pride functions to convey high
status, then it is likely to be displayed more frequently by high-
status individuals. Thus, familiarity between high-status targets
and pride may lead observers to have an elevated expectancy, or
bias, to infer pride in individuals who are members of high-status
ethnic or gender groups, regardless of the expression they are
actually displaying. If this is the case, then we should find higher
pride recognition rates and higher pride base rates (rates of using
the pride label across expressions) for targets belonging to higher
status groups (e.g., Caucasians and men) than for targets belonging
to lower status groups (e.g., African Americans and women).
Supporting this account, Keltner (1995) found higher recognition
rates for embarrassment in African American than Caucasian tar-
gets and interpreted this finding as a result of embarrassment
signaling low status combined with observers’ bias toward per-
ceiving African Americans as low status. Other studies support the
more general idea that target ethnicity and gender can influence the
way observers perceive targets’ emotions, especially when emo-
tion expressions are ambiguous or when observers have race-based
biases (Hess, Adams, & Kleck, 2004; Hess, Blairy, & Kleck, 1997;
Hugenberg & Bodenhausen, 2004). As Macrae and Bodenhausen
(2000) explained, “Activated categorical representations . . . pro-
vide the perceiver with expectancies that can guide the processing
of subsequently encountered information . . . [These expectancies
can] lead the perceiver to emphasize stereotype-consistent infor-
mation” (p. 103).
According to the cultural value hypothesis, pride should be more
readily identified in targets belonging to ethnic groups that are
stereotypically associated with pride or thought to value pride. For
example, in the 1960s, African Americans strove to reinforce
“Black pride,” and elements of the pride expression were incor-
porated into calls for ethnic solidarity, exemplified by the “Black
Power” stance. In contrast, Asian Americans are typically associ-
ated with the cultural value of modesty and, in fact, have been
shown to have negative views of self-esteem, self-enhancement,
and pride (Eid & Diener, 2001; Heine et al., 1999). These diver-
gent cultural values may lead to divergent associations between
each of these ethnic groups and pride, such that African Americans
may be assumed to frequently experience and display pride,
whereas Asian Americans may be assumed to do so only infre-
quently. If observers are aware of these presumed (or actual)
cultural differences, then this may influence pride recognition rates
and base rates of pride identification (i.e., false alarms) in these
targets. Thus, the cultural value hypothesis suggests a prediction
that directly competes with the perceived status hypothesis: higher
pride recognition and base rates of pride identification in African
American versus Asian American targets.
A potential confound to any study examining the effects of
target gender or ethnicity on emotion recognition is the possibility
that effects found are the result of other characteristics that vary
across targets (e.g., attractiveness, facial physiognomy, etc.). Sev-
eral studies have addressed this issue by including a large number
of targets of each ethnicity and gender, thereby decreasing the
likelihood that, on average, the targets of one ethnic– gender group
will differ from the targets of another ethnic– gender group on any
characteristic of relevance other than ethnicity and gender (e.g.,
Biehl et al., 1997). However, the only way to completely rule out
this potential confound is to find targets that vary only in ethnicity
For this reason, we developed a set of drawn figure targets that
are identical (all are modifications of the same drawing) except for
features stereotypically associated with gender and ethnicity. In
Study 3, we first established that each drawn figure is reliably
One interesting exception to this general rule emerged in Elfenbein and
Ambady’s (2002) meta-analysis of cross-cultural emotion recognition stud-
ies. They found that, within cultures, minority group members show higher
recognition for emotion expressions displayed by majority out-group mem-
bers than majority group members do for the expressions of minority
associated with the intended ethnic group and then tested whether
pride recognition is influenced by target gender or ethnicity.
Participants. Eighty-three undergraduate students (65%
women) from the University of California, Davis participated in
exchange for course credit. Participants self-identified their race as
Asian (50%), Caucasian (31%), Latino (10%), African American
(3%), and Other or Mixed (6%).
Stimuli. A professional artist who specializes in drawn figures
was recruited to draw human characters from the waist up showing
anger, contempt, disgust, fear, happiness, pride, sadness, and sur-
prise. Nonpride emotion expressions were based on the DFA
(Ekman et al., 1983), and the pride expression was based on
previous research on pride (Tracy & Robins, 2007a). The artist
drew a single human character eight times, once portraying each
emotion expression. Expressions were evaluated by a leading
expert in FACS, Erika Rosenberg, and two rounds of modifica-
tions were made on the basis of her suggestions, until a final set of
eight expressions were verified to include each of the movements
relevant to each expression (according to EM-FACS) and no other
movements. The artist next scanned these images to produce three
copies of each target and then modified each to produce six drawn
targets showing all eight expressions: three men and three women
(age 20 –25 years) of African, Asian, and European–Caucasian
ethnic descent. Target gender was manipulated by changing hair-
style, hair length, and shape of chest. Target ethnicity was manip-
ulated by changing the targets’ skin tone (with shading; all figures
were drawn in black and white), hairstyle, hair thickness, hair
color, and shape and size of nose, mouth, and eyes (Blair, Judd,
Sadler, & Jenkins, 2002; Hugenberg & Bodenhausen, 2004; Liv-
ingston & Brewer, 2002).
Other than ethnic and gender characteristics, the drawn targets
were identical. All were drawn wearing the same clothing (a white
t-shirt), were the same size, and had the same basic facial physi-
ognomy. Figure 5 shows three of these targets (one from each
ethnic group) displaying the pride expression. By varying ethnicity
and gender using these drawn figure targets, we ensured that any
target effects found can be attributed to differences in target gender
or ethnicity. Previous judgment studies using drawn figure targets
have demonstrated the effectiveness of this method for comparing
judgments of traits from body movements (e.g., Schwartz, Tesser,
& Powell, 1982; Spiegel & Machotka, 1974).
Procedure. Participants viewed each of the six drawn figure
targets showing each emotion expression, projected onto a large
4-ft ⫻6-ft (1.2-m ⫻1.8-m) screen, for 30 s each. To verify that
targets represented the intended ethnic groups, we asked partici-
pants to, for each image, rate the “extent to which the person in this
slide appears to be African-American” on a 7-point scale ranging
from 0 (not at all)to6(extremely). Participants were then asked to
provide the same rating for Caucasian American and Asian Amer-
ican images. Thus, each target was rated for the extent to which
he/she was representative of each ethnic group.
After rating targets’ ethnicity, participants were asked to
“choose the emotion that best matches the emotion expressed by
the person in the slide” from the following list of options: anger,
none of these are correct. Participants then rated the intensity of
the emotion expressed on a 7-point scale ranging from 0 (not at all
intense)to6(extremely intense). We assessed perceptions of
intensity as a second gauge of how target gender or ethnicity might
influence perceptions of emotions across targets that are, in actu-
ality, showing the same emotions at the same level of intensity.
Target ethnicity. We conducted a series of repeated-measures
analyses of variance (ANOVAs) predicting ratings for each ethnic
group from the intended target ethnic group. Results showed that
African targets showing pride were more likely to be identified as
African American (M⫽5.40) than as Asian American (M⫽0.42)
or Caucasian American (M⫽0.30), F(2, 78) ⫽725.32, p⬍.05.
Similarly, Asian targets showing pride were more likely to be
identified as Asian American (M⫽4.84) than as African Amer-
ican (M⫽0.20) or Caucasian American (M⫽0.76), F(2, 80) ⫽
600.98, p⬍.05. Finally, Caucasian targets showing pride were
more likely to be identified as Caucasian American (M⫽5.25)
than as African American (M⫽0.35) or Asian American (M⫽
1.29), F(2, 80) ⫽612.45, p⬍.05. These effects held across all
other emotions: F(2, 77) ⫽2,298.38 for African targets (Ms⫽
5.26, 0.37, and 0.14 for African, Asian, and Caucasian, respec-
tively); F(2, 77) ⫽766.16 for Asian targets (Ms⫽5.06, 0.43, and
0.61 for Asian, African, and Caucasian, respectively); and F(2,
68) ⫽1,312.70 for Caucasian targets (Ms⫽5.38, 0.41, and 0.83
for Caucasian, African, and Asian, respectively; all ps⬍.05).
Given the magnitude of these effects, we felt confident that our
drawn figure targets manipulated ethnicity as intended.
Pride recognition. On the basis of binomial tests, recognition
rates for all six pride expressions (M⫽74%, range ⫽52%– 86%)
were significantly greater than chance ( p⬍.05), with chance set
conservatively at 33%. When pride was not accurately recognized,
it was labeled as contempt (49% of errors, 13% overall), happiness
(11% of errors, 3% overall), disgust (4% of errors, 1% overall),
anger (0.82% of errors, 0.10% overall), and none of these (34% of
errors, 8% overall).
We next ran a Target Gender ⫻Ethnicity ANOVA predicting
pride recognition and found main effects for ethnicity, F(2,81) ⫽
9.83, and gender, F(1,82)⫽19.60 (both ps⬍.05). Pride recogni-
tion was higher for female (M⫽82%) than for male (M⫽70%)
targets and for Caucasian and Asian (Ms⫽80% and 78%, respec-
tively) than for African (M⫽65%) targets. We also found a Target
Researchers who wish to use the drawn figures stimulus set for
research purposes should contact the first author.
Figure 5. The prototypical pride expression shown by drawn figure
targets of Asian (male), Caucasian (male), and African (female) descent
and used in Studies 3 and 4.
524 TRACY AND ROBINS
Gender ⫻Target Ethnicity interaction, F(2, 81) ⫽4.41, p⬍.05,
suggesting that the gender difference was largest for African
targets (Ms⫽78% for the female vs. 52% for the male).
We next replicated these analyses predicting emotion intensity
ratings for participants who correctly identified the pride expres-
sion (to ensure that we were predicting pride intensity ratings,
specifically). We found an effect of target ethnicity, F(2, 46) ⫽
3.44, p⬍.05, on pride intensity, suggesting that Caucasian targets
were perceived as displaying pride most intensely (M⫽3.75),
followed by African (M⫽3.50) and Asian (M⫽3.46) targets;
only the difference between Caucasian and African targets was
significant,F(1, 81) ⫽5.37, p⬍.05 (other Fs were 0.16 and 3.59,
ns). We also found an effect of target gender; pride was perceived
as slightly more intense when shown by women (M⫽3.63) than
by men (M⫽3.51), F(1, 47) ⫽3.01, p⬍.05, one-tailed (predic-
tion based on the finding of higher pride recognition for female
than male targets).
There was no effect of perceiver gender, F(1, 76) ⫽0.00, or
perceiver ethnicity (Caucasian vs. Asian), F(1, 61) ⫽0.31, on
pride recognition, nor were there any Perceiver Gender ⫻Per-
ceiver Ethnicity interactions, F(1, 59) ⫽0.02. There were also no
Perceiver ⫻Target interactions; the interactions between perceiver
gender and target gender, F(1,76) ⫽0.34, perceiver gender and
target ethnicity, F(2, 75) ⫽2.30, perceiver ethnicity and target
ethnicity, F(2, 60) ⫽0.61, and perceiver ethnicity and target
gender, F(1, 61) ⫽0.00, were all nonsignificant. There also were
no three- or four-way interactions.
Next, we examined whether these findings applied to pride
false-alarm rates—that is, the tendency to use the pride label for
nonpride emotion expressions—in addition to recognition rates
(the tendency to use the pride label for the pride expression). The
pride label was rarely but occasionally applied to other expres-
sions: disgust (2% of responses to disgust expressions), contempt
(1%), surprise (1.3%), fear (0.80%), anger (0.60%), sad (0.60%),
and happiness (0.40%). We ran ANOVAs predicting the tendency
to label all other emotions as pride from target gender and ethnic-
ity. We found no overall (across emotions) effects of target eth-
nicity or gender, nor any effects within any particular emotion.
Thus, the effects of target gender and ethnicity on pride recogni-
tion applied only to participants’ accuracy for pride and not to
participants’ general tendency to use the pride label across emo-
These findings provided mixed support for our three hypotheses.
First, we found clear support for the universality hypothesis; pride
was recognized well above chance in targets of both genders and
all three ethnic groups. Second, the ethnicity effect that emerged
supports the perceived status hypothesis and disconfirms the cul-
tural value hypothesis. Pride was most readily recognized in Cau-
casian and Asian targets, the two groups of highest status. Fur-
thermore, Asian targets, who are not typically associated with
pride-related values, were more easily identified as showing pride
than were African targets, who are more likely to be stereotypi-
cally associated with pride.
In contrast, the gender effect that emerged is contrary to what
we would expect from the perceived status hypothesis (and, to the
extent that it is relevant, the cultural value hypothesis). Pride was
better recognized in women than in men, and this effect was
particularly strong among African targets—the group that may be
perceived as lowest status. There may, however, be a somewhat
straightforward explanation for this finding. If pride is not ex-
pected in women because of the perceived higher status of men,
then it may be all the more noticeable when female targets display
pride—leading to paradoxically higher pride recognition in these
targets. In contrast, when stereotypically high-status male targets
show pride, the expression may be perceived as part of the target’s
gender characteristics (e.g., men have broader chests than women,
so expanded chest may be perceived simply as “male”). Macrae
and Bodenhausen (2000) explained this possible outcome of ste-
reotype categorization: “Expectancies also sensitize the perceiver
to unexpected data, leading to a greater emphasis on stereotype-
inconsistent information following category activation” (p. 103).
This explanation suggests that the present findings may also sup-
port the cultural value hypothesis. If African American men are
associated with the cultural value of pride, then perceivers may
view these targets’ pride expressions as another feature of their
ethnicity and thus show lower recognition for pride displayed by
the African male target.
This interpretation of the target gender effect is supported by the
fact that we did not find higher base rates of pride identification
(i.e., pride false alarms) in female targets—we only found higher
accuracy rates for pride recognition. This suggests that these
effects were not due to participants making associations between
women and pride, regardless of the emotion displayed; rather, the
effects may have been driven by an assumption that women do not
typically display pride, so, when they do, it must, in fact, be pride.
These findings also may be driven by participants making direct
comparisons among targets on the basis of ethnicity and gender.
Such a response style may, in fact, have been encouraged by the
experimental design, given that participants were shown many
targets who clearly differed in ethnic and gender characteristics,
and were asked to identify targets’ emotions immediately after
identifying their ethnicity. Thus, it is important to test whether the
present findings hold when participants are less likely to make
direct comparisons among groups. In real-life settings, individuals
do not typically determine that a target feels pride only after
comparing his/her expression with another target from a different
gender or ethnic group. For this reason, we conducted a replication
study in which each participant viewed only a single target.
In Study 4, we used a between-subjects design to test the effects
of target gender and ethnicity on pride recognition.
Participants. Participants were 211 undergraduate students
(60% women) from the University of California, Davis, who
participated in exchange for course credit. Participants self-
identified their race as Asian (47%), Caucasian (31%), Latino
(9%), African American (2%), and other or mixed (11%).
Stimuli. The set of drawn figure targets that was developed for
Study 3 was used.
Procedure. Participants (N⫽211) were randomly assigned to
one of six ethnic– gender conditions. Participants in each condition
viewed only a single target (African man, African woman, Asian
man, Asian woman, Caucasian man, or Caucasian woman) show-
ing all eight emotion expressions. Participants viewed each of the
eight images, projected onto a large 4-ft ⫻6-ft (1.2-m ⫻1.8-m)
screen, for 30 s. The order of the images was randomized. As in
Study 3, participants were asked, for each image, to “choose the
emotion that best matches the emotion expressed by the person in
the slide” from the following list of options: anger,contempt,
disgust,fear,happiness,pride,sadness,surprise, and none of these
are correct. Participants were also again asked to rate the intensity
of the emotion expressed on a 7-point scale ranging from 0 (not at
all intense)to6(extremely intense).
On the basis of binomial tests, recognition rates for all six pride
expressions (M⫽78%, range ⫽70%– 87%) were significantly
greater than chance ( p⬍.05), with chance set at 33% (see Figure
6 for recognition rates by target). When the pride expression was
not accurately recognized, it was labeled as contempt (59% of
errors, 13% overall), happiness (15% of errors, 3% overall), dis-
gust (2% of errors, 0.37% overall), surprise (2% of errors, 0.37%
overall), and none of these (22% of errors, 5% overall).
We next ran a Target Gender ⫻Target Ethnicity ANOVA
predicting pride recognition and found a main effect of gender,
F(1, 205) ⫽3.47, p⬍.05 (one-tailed; effect was predicted on the
basis of the findings of Study 3), but no effect of ethnicity, F(2,
205) ⫽0.13, ns, and no Gender ⫻Ethnicity interaction, F(2,
205) ⫽0.17, ns. Pride recognition was higher for female (M⫽
83%) than for male (M⫽73%) targets. We next replicated these
analyses predicting pride intensity ratings only for participants
who accurately recognized pride. There was an effect of target
gender, F(1, 159) ⫽16.10, p⬍.05, on intensity, suggesting that
pride in female targets (M⫽3.99) was viewed as more intense
than pride in male targets (M⫽3.25). There was no effect of target
ethnicity on intensity, F(2, 159) ⫽0.02, nor was there any Eth-
nicity ⫻Gender effect on intensity, F(2, 159) ⫽1.60, both ns.
As in Study 3, we did not find perceiver effects. There was no
effect of perceiver gender, F(1, 192) ⫽0.00, or perceiver ethnicity
(Caucasian vs. Asian American) on pride recognition (t⫽0.11;
both ns), nor did perceiver gender interact with either target
gender, F(1, 192) ⫽0.06, or target ethnicity, F(2. 192) ⫽1.44,
Next, we examined target effects on pride false-alarm rates.
There were only four pride false alarms (i.e., occasions in which
other emotion expressions were labeled as pride): three for con-
tempt and one for happiness.
Not surprisingly, these numbers
were too low to produce any general trends; there was no target
gender or target ethnicity effect on pride base rate, F(2, 204) ⫽
2.45 and F(4, 410) ⫽1.12, respectively; both ns. Thus, target
gender influenced participants’ accuracy in recognizing pride but
did not influence their general tendency to use the pride label
Study 4 replicated the finding of Study 3 that pride recognition
is higher for female than for male targets, suggesting that this
effect was not due to participants making direct comparisons
between male and female targets prior to judging targets’ emo-
tions. Alternatively, gender may be such a ubiquitous construct in
social judgments that these comparisons are made automatically,
regardless of the presence of stimuli of the opposite gender. In
contrast, these automatic comparisons may be less likely, or less
clear cut, for a multicategory construct like ethnicity. Regardless,
the findings from both studies suggest that, although pride is
reliably recognized across targets of both genders and all three
ethnicities examined, it is recognized with greater accuracy when
shown by female as opposed to male targets.
The absence of an effect of target ethnicity on pride recognition
suggests that the ethnicity effect that emerged in Study 3 may have
resulted from participants making direct comparisons among tar-
gets of different ethnic groups. Given that participants in Study 3
spent the majority of their time responding to questions about
targets’ ethnicity, they may well have assumed that the goal of the
experiment was to study the effects of ethnicity on emotion iden-
tification and thus varied their responses accordingly. Differential
levels of power in Study 3 (which was within-subjects) versus
Study 4 (between-subjects) may also have contributed to the
different findings, but this seems unlikely given that there was not
even a trend toward greater pride recognition in Caucasian or
Asian targets, compared to African Americans, in Study 4 (see
The present research addressed what may be the most critical
question for our understanding of the nonverbal expression of
pride: Is the pride expression likely to be universally recognized,
or is it more likely to be a culture-specific gesture, similar to the
The confusion between contempt and pride, seen in error rates and
false alarm rates in Studies 1, 3, and 4, may be due to the semantic overlap
between the two emotions. Contempt involves moral superiority, which
may overlap with the superiority associated with pride—particularly hu-
bristic pride, which is reliably conveyed by the same nonverbal expression
as authentic pride (Tracy & Robins, 2007a). The confusion may also result
from the fact that one of the components of the pride expression— head tilt
back— has also been associated with contempt (Rosenberg & Ekman,
African Asian Caucasian
Figure 6. Recognition rates for the pride expression by target gender and
ethnicity in Study 4 (N⫽211).
526 TRACY AND ROBINS
Western “thumbs up” sign or the Indian tongue-bite display of
shame? Study 1 showed that the expression is not unique to
American culture; perceivers outside the United States in a non-
English-speaking culture with divergent cultural values were able
to recognize it. Study 2 provided considerably more compelling
evidence for the expression’s cross-cultural generalizability. Pride
was reliably recognized by nonliterate perceivers who are almost
entirely isolated from Western culture. This finding, which paral-
lels Ekman et al.’s (1969; Ekman & Friesen, 1971) finding that the
basic emotions are recognized by preliterate members of the Fore
tribe in Papua New Guinea, suggests that the pride expression
should be added to the small pantheon of emotions thought to be
universal. In fact, with the sole exception of happiness, recognition
rates for both self-conscious emotions included in Study 2—pride
and shame— did not differ from recognition rates for the basic
emotions in this sample, and these rates were highly similar to
those that have been found previously in nonliterate, preindustri-
alized non-Western cultures using the forced-choice method (e.g.,
Ekman et al., 1969). It is important to note that the findings from
Study 2 are not simply evidence of another culture recognizing
pride. Rather, we specifically chose the Burkina Faso sample
because these participants are extremely unlikely to have learned
the pride expression from exposure to Western media or Western
individuals (i.e., cross-cultural transmission). Our evidence that
Burkinabe participants recognize pride is thus not simply evidence
that individuals from another culture (outside the United States)
recognize the expression, but rather suggests that the expression
may be universal. That is, if the Burkinabe did not learn the pride
expression through Western media, then they most likely know it
because it is part of human nature.
Finally, Studies 3 and 4 demonstrated that the pride expression
generalizes across the gender and ethnicity of targets as well as
perceivers. Although we did not examine all possible target eth-
nicities, we examined the three ethnic groups that represent the
majority of people in the world. Using an experimental design that
controlled for potentially confounding factors, such as target at-
tractiveness or other differences in facial physiognomy, we found
that pride is reliably recognized in male and female targets of
African, Asian, and European (Caucasian) descent. We also found
that the level of pride recognition is affected by these target
characteristics; recognition is higher for female than for male
targets and, when participants are encouraged to directly compare
targets of varying ethnicities, for Caucasian and Asian than for
It has previously been assumed that self-conscious emotions
differ from basic emotions because they lack universally recog-
nized expressions (Ekman, 1992). The present research challenges
this assumption. Our findings suggest that even highly social,
cognitively complex self-evaluative emotions can be universal.
This finding may require a change in our notion of what a univer-
sal basic emotion is.
If pride and shame are universal, then humans may have evolved
to communicate social messages about dominance and submission
through transitory emotions. In other animals, the communication
of status seems to occur through similar kinds of expressions (e.g.,
a chimpanzee’s “cocky” gait; de Waal, 1989), but these displays
may be associated more with stable traits than momentary feelings.
An alpha chimpanzee is typically the dominant individual across
most interactions and over long periods of time (de Waal, 1989).
In humans these displays can be stable over time (there are reliable
individual differences in pride-proneness; Tracy & Robins,
2007b), but they may also shift in response to particular situations.
Humans repeatedly negotiate and renegotiate status during new
interactions, in part because nontransitive status hierarchies,
whereby an individual may be higher status than others in some
contexts but not others, are the norm in human societies. Pride and
shame may be the emotions that fuel these status exchanges. By
displaying the pride expression after a success, an individual may
inform his/her social group of his/her achievement and convey the
message that he/she deserves respect and higher status. In our
evolutionary history, this kind of interpersonal message was likely
critical for the promotion and maintenance of social alliances,
which were (and are) essential to human survival.
If pride functions to convey status, then it makes sense that
target characteristics relevant to status, such as gender and ethnic-
ity, influence the level of pride recognition. Observers in Studies 3
and 4 were better able to recognize pride when it was shown by
women, a paradoxical effect that may have emerged because
female pride is somewhat incongruous. In Western (and many
non-Western) societies, men typically have higher status than
women, and women across the globe have lower self-esteem than
men (Robins, Trzesniewski, Tracy, Gosling, & Potter, 2002).
Thus, women may experience pride less frequently and intensely
than men, leading to an association between maleness and pride.
The absence of a gender difference in pride false-alarm rates (i.e.,
labeling targets as showing pride when they were displaying other
emotion expressions) supports this interpretation, but it should be
further probed in future research. Studies are needed to manipulate
status more directly (e.g., comparing recognition for pride dis-
played by a boss vs. an employee) and test whether similar effects
emerge. If pride recognition is higher in targets that are more
clearly defined as high status, then it would raise questions about
our interpretation of the present findings. If, however, pride is
more readily recognized when shown by explicitly low-status
targets, then it would support our account.
The present findings may have implications for the function of
the pride experience, as well as its expression. Evidence consistent
with a universal pride expression indirectly suggests that the
experience of pride may also be universal. If pride were not
cross-culturally experienced and displayed, then how could indi-
viduals across such diverse cultures agree on its expression? The
pride experience thus may have evolved to serve an adaptive
function as well. Researchers have suggested that self-esteem
functions as a social barometer, or sociometer, to inform individ-
uals of their social status and ensure that they behave in ways that
A third alternative exists as well: some features common to Western
and Burkinabe culture could have allowed the pride expression to emerge
separately in both cultures as a social construction. We view this expla-
nation as highly unlikely, given both the large disparity between Western
and Burkinabe culture and anecdotal evidence for similarities between the
pride expression and dominance displays seen in nonhuman primates (de
Waal, 1989). However, future research should address this issue by con-
ducting additional pride recognition studies in a range of cultures.
will maintain their status and the acceptance of others (Leary et al.,
1995). Given that pride is the emotion most strongly related to
self-esteem (J. D. Brown & Marshall, 2001), it may be the affec-
tive motivator behind the maintenance and enhancement of self-
esteem. Specifically, when individuals experience a success, they
feel pride in response, and these feelings promote positive thoughts
and feelings about the entire self—leading to the high self-esteem
that informs individuals of their increased social value. Interest-
ingly, recent research suggesting that there are two distinct facets
of pride—authentic and hubristic—implies that pride may boost
self-worth through two distinct personality processes, depending
on which facet is experienced. Although “authentic” pride is
positively correlated with genuine self-esteem, the more self-
aggrandizing, “hubristic” facet is negatively related to self-esteem,
but positively related to narcissism, especially when shared vari-
ance with self-esteem is statistically removed (Tracy & Robins,
2007b). Thus, a hubristic pride response to success may promote
narcissism, rather than genuine self-esteem, which in turn may
boost status, but by increasing others’ admiration for the proud
individual, rather than their liking of him or her (Robins, Tracy, &
Shaver, 2001). Both facets of pride are reliably associated with the
same nonverbal expression, at least within Western cultures (Tracy
& Robins, 2007b), suggesting that they serve the same interper-
sonal function of signaling increased status to others.
Additional questions regarding the pride expression, some of
which were generated by this research, remain unanswered. First,
although individuals in Burkina Faso were able to accurately
recognize pride, it remains unclear whether they imbue the Dioula
word for pride with the same meaning that we associate with the
word pride. In fact, even in Western cultures, the word “pride”
does not have a single, consistent meaning, as can be seen from the
two different Italian translations for pride and the two semantically
distinct facets of pride in English (i.e., authentic and hubristic;
Tracy & Robins, 2007b). Of importance, the finding that the pride
expression generalizes across cultures does not imply that the ways
in which individuals value, experience, and respond to the emotion
are similar across cultures. An important direction for future re-
search is to examine cultural differences in beliefs about pride and
the situations in which it is expressed versus regulated. Future
studies should examine, for example, whether the two-facet struc-
ture of pride generalizes across cultures, or whether it is an
outcome of the way that pride has been elaborated in Western
cultures, where considerable emphasis is placed on understanding
and enhancing the self. Anecdotal evidence, based on informal
conversations with our Burkinabe collaborators, suggests that if
only one facet of pride exists in Burkinabe culture, the hubristic
facet is the more likely candidate.
Second, although the prototypical pride expression was devel-
oped from previous research on preverbal children’s behavioral
responses to success (Belsky & Domitrovich, 1997; M. Lewis et
al., 1992; Stipek et al., 1992; Weisfeld & Beresford, 1982) and
new findings suggest that this expression is spontaneously dis-
played by Olympic athletes from a wide range of cultures in
response to victory (Tracy & Matsumoto, 2007), we do not know
whether the pride expression typically co-occurs with actual pride
experiences in everyday life. An important future study will ma-
nipulate pride and test whether the expression and subjective
feeling co-occur. Such research should also examine whether pride
has distinct physiological or neural correlates, as has been found
for several basic emotions (e.g., Ekman et al., 1983; LeDoux,
1996; Panksepp, 1998). Recent studies have found preliminary
evidence for distinct neural patterns associated with self-processes
(see Beer, 2007, for a review); these patterns may be relevant to
the display and experience of pride. Furthermore, given recent
evidence of distinct physiological correlates of shame (Dickerson,
Gruenewald, & Kemeny, 2004), it is plausible that pride also has
a unique biological substrate.
Third, although pride may have evolved to communicate status-
related messages, this presumed adaptive function has not been
empirically tested. In future research, we hope to examine whether
the pride expression does in fact promote the status of those who
show it. In addition, although there is no evidence that any non-
human animals show pride in response to success, it would be
worth empirically exploring this possibility among other primates
that are genetically similar to humans, such as chimpanzees or
bonobos. Recent research provides a nonverbal coding scheme for
assessing the pride expression (Tracy & Robins, 2007a), which
comparative psychologists could use to code for the presence of
this expression in other animals as they respond to status-
enhancing achievements (e.g., overcoming an attacker).
In summary, we view the present findings as the first evidence
in support of Darwin’s (1872) claim about pride. Pride is uniquely
different, in many ways, from the basic emotions, and even its
expression is somewhat different, requiring the body as well as the
face. However, these findings suggest that, at least from a behav-
ioral perspective, pride may be as much a part of human nature as
any other emotion.
Beaupre, M. G., & Hess, U. (2005). Cross-cultural emotion recognition
among Canadian ethnic groups. Journal of Cross-Cultural Psychology,
Beer, J. S. (2007). Neural systems for self-conscious emotions and their
underlying appraisals. In J. L. Tracy, R. W. Robins, & J. P. Tangney
(Eds.), The self-conscious emotions: Theory and research (pp. 53– 67).
New York: Guilford.
Belsky, J., & Domitrovich, C. (1997). Temperament and parenting ante-
cedents of individual difference in three-year-old boys’ pride and shame
reactions. Child Development, 68, 456 – 466.
Biehl, M., Matsumoto, D., Ekman, P., Hearn, V., Heider, K., Kudoh, T., &
Ton, V. (1997). Matsumoto and Ekman’s Japanese and Caucasian facial
expressions of emotion (JACFEE): Reliability data and cross-national
differences. Journal of Nonverbal Behavior, 21, 3–21.
Blair, I. V., Judd, C. M., Sadler, M. S., & Jenkins, C. (2002). The role of
Afrocentric features in person perception: Judging by features and
categories. Journal of Personality and Social Psychology, 83, 5–25.
Boucher, J. D., & Carlson, G. E. (1980). Recognition of facial expression
in three cultures. Journal of Cross-Cultural Psychology, 11, 263–280.
Brown, D. E. (1991). Human universals. Philadelphia: Temple University
Brown, J. D., & Marshall, M. A. (2001). Self-esteem and emotion: Some
thoughts about feelings. Personality and Social Psychology Bulletin, 27,
Buss, D. E. (1992). Is there a universal human nature? Contemporary
Psychology, 37, 1262–1263.
Darwin, C. (1872). The expression of the emotions in man and animals (3rd
ed.). New York: Oxford University Press.
528 TRACY AND ROBINS
de Waal, F. B. M. (1989). Chimpanzee politics: Power and sex among
apes. Baltimore, MD: Johns Hopkins University Press.
Dickerson, S. S., Gruenewald, T. L., & Kemeny, M. E. (2004). When the
social self is threatened: Shame, physiology, and health. Journal of
Personality, 72, 1191–1216.
Dizionari Garzanti. (1997). Inglese–Italiano, Italiano–Inglese [English–
Italian, Italian–English]. Milan, Italy: Garzanti Editore.
Eibl-Eibesfeldt, I. (1989). Human ethology. New York: De Gruyter.
Eid, M., & Diener, E. (2001). Norms for experiencing emotions in different
cultures: Inter- and intranational differences. Journal of Personality and
Social Psychology, 81, 869 – 885.
Ekman, P. (1971). Universals and cultural differences in facial expression
of emotion. In J. Cole (Ed.), Nebraska Symposium on Motivation: Vol.
19. Perspectives on motivation (pp. 207–283). Lincoln: University of
Ekman, P. (1992). An argument for basic emotions. Cognition and Emo-
tion, 6, 169 –200.
Ekman, P., & Friesen, W. V. (1971). Constants across cultures in the face
and emotion. Journal of Personality and Social Psychology, 17, 124 –
Ekman, P., & Friesen, W. V. (1986). A new pan-cultural facial expression
of emotion. Motivation and Emotion, 10, 159 –168.
Ekman, P., Friesen, W. V., O’Sullivan, M., Chan, A., Diacoyanni-
Tarlatzis, I., Heider, K., et al. (1987). Universals and cultural differences
in the judgments of facial expressions of emotion. Journal of Personality
and Social Psychology, 53, 712–717.
Ekman, P., Levenson, R. W., & Friesen, W. V. (1983, September 16).
Autonomic nervous system activity distinguishes among emotions. Sci-
ence, 221, 1208 –1210.
Ekman, P., Sorenson, E. R., & Friesen, W. V. (1969, April 4). Pan-cultural
elements in facial displays of emotion. Science, 164, 86 – 88.
Elfenbein, H. A., & Ambady, N. (2002). On the universality and cultural
specificity of emotion recognition: A meta-analysis. Psychological Bul-
letin, 128, 203–235.
Frank, M. G., & Stennett, J. (2001). The forced-choice paradigm and the
perception of facial expressions of emotion. Journal of Personality and
Social Psychology, 80, 75– 85.
Fredrickson, B. L., & Branigan, C. (2001). Positive emotions. In T. J.
Mayne & G. A. Bonanno (Eds.), Emotions: Current issues and future
directions (pp. 123–151). New York: Guilford.
Haidt, J., & Keltner, D. (1999). Culture and facial expression: Open-ended
methods find more expressions and a gradient of recognition. Cognition
and Emotion, 13, 225–266.
Hart, D., & Karmel, M. P. (1996). Self-awareness and self-knowledge in
humans, apes, and monkeys. In A. E. Russon, K. A. Bard, & S. T. Parker
(Eds.), Reaching into thought: The minds of the great apes (pp. 325–
347). Cambridge, England: Cambridge University Press.
Hart, D., & Matsuba, M. K. (2007). The development of pride and moral
life. In J. L. Tracy, R. W. Robins, & J. P. Tangney (Eds.), The self-
conscious emotions: Theory and research (pp. 114 –133). New York:
Heerey, E. A., Keltner, D., & Capps, L. M. (2003). Making sense of
self-conscious emotion: Linking theory of mind and emotion in children
with autism. Emotion, 3, 394 – 400.
Heine, S. J., Kitayama, S., & Hamamura, T. (2007). The inclusion of
additional studies yields different conclusions: A reply to Sedidikes,
Gaertner, & Vevea (2005), Journal of Personality and Social Psychol-
ogy. Asion Journal of Social Psychology, 10, 49 –58.
Herrald, M. M., & Tomaka, J. (2002). Patterns of emotion-specific ap-
praisal, coping, and cardiovascular reactivity during an ongoing emo-
tional episode. Journal of Personality and Social Psychology, 83, 434 –
Hertenstein, M. J., Keltner, D., App, B., Bulleit, B. A., & Jaskolka, A. R.
(2006). Touch communicates distinct emotions. Emotion, 6, 528 –533.
Hess, U., Adams, R., & Kleck, R. E. (2004). Facial appearance, gender,
and emotion expression. Emotion, 4, 378 –388.
Hess, U., Blairy, S., & Kleck, R. E. (1997). The intensity of emotional
facial expressions and decoding accuracy. Journal of Nonverbal Behav-
ior, 21, 241–257.
Hoftstede, G. (1984). National cultures revisited. Behavioral Science Re-
search, 18, 285–305.
Hoftstede, G. (2001). Culture’s consequence: Comparing values, behav-
iors, institutions, and organizations across nations (2nd ed.). Thousand
Oaks, CA: Sage.
Hugenberg, K., & Bodenhausen, G. V. (2004). Ambiguity in social cate-
gorization: The role of prejudice and facial affect in race categorization.
Psychological Science, 15, 342–345.
Izard, C. E. (1971). The face of emotion. East Norwalk, CT: Appleton-
Izard, C. E., & Haynes, O. M. (1988). On the form and universality of the
contempt expression: A challenge to Ekman and Friesen’s claim of
discovery. Motivation and Emotion, 12, 1–16.
Keltner, D. (1995). Signs of appeasement: Evidence for the distinct dis-
plays of embarrassment, amusement, and shame. Journal of Personality
and Social Psychology, 68, 441– 454.
Keltner, D., & Buswell, B. N. (1997). Embarrassment: Its distinct form and
appeasement functions. Psychological Bulletin, 122, 250 –270.
Leary, M. R., Tambor, E. S., Terdal, S. K., & Downs, D. L. (1995).
Self-esteem as an interpersonal monitor: The sociometer hypothesis.
Journal of Personality and Social Psychology, 68, 518 –530.
LeDoux, J. (1996). The emotional brain: The mysterious underpinnings of
emotional life. New York: Touchstone.
Lewis, H. B. (1971). Shame and guilt in neurosis. New York: International
Lewis, M., Alessandri, S. M., & Sullivan, M. W. (1992). Differences in
shame and pride as a function of children’s gender and task difficulty.
Child Development, 63, 630 – 638.
Livingston, R. W., & Brewer, M. B. (2002). What are we really priming?
Cue-based versus category-based processing of facial stimuli. Journal of
Personality and Social Psychology, 82, 5–18.
Macrae, C. N., & Bodenhausen, G. V. (2000). Social cognition: Thinking
categorically about others. Annual Review of Psychology, 51, 93–120.
Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications
for cognition, emotion, and motivation. Psychological Review, 98, 224 –
Massaro, D. W., & Ellison, J. W. (1996). Perceptual recognition of facial
affect: Cross-cultural comparisons. Memory & Cognition, 24, 812– 822.
Matsumoto, D. (2002). Methodological requirements to test a possible
in-group advantage in judging emotions across cultures: Comment on
Elfenbein and Ambady (2002) and evidence. Psychological Bulletin,
128, 236 –242.
Matsumoto, D., Consolacion, T., Yamada, H., Suzuki, R., Franklin, B.,
Paul, S., et al. (2002). American–Japanese cultural differences in judg-
ments of emotional expressions of different intensities. Cognition and
Emotion, 16, 721–747.
Matsumoto, D., & Ekman, P. (1989). American Japanese cultural differ-
ences in intensity ratings of facial expressions of emotion. Motivation
and Emotion, 13, 143–157.
Matsumoto, D., & Willingham, B. (2006). The thrill of victory and the
agony of defeat: Spontaneous expressions of medal winners of the 2004
Athens Olympic Games. Journal of Personality and Social Psychology,
91, 568 –581.
Panksepp, J. (1998). Affective neuroscience: The foundations of human and
animal emotions. New York: Oxford University Press.
Pinker, S. (2002). The blank slate: The modern denial of human nature.
New York: Viking.
Robins, R. W., Tracy, J. L., & Shaver, P. R. (2001). Shamed into self-love:
Dynamics, roots, and functions of narcissism. Psychological Inquiry, 12,
Robins, R. W., Trzesniewski, K., Tracy, J. L., Gosling, S., & Potter, J.
(2002). Global self-esteem across the lifespan. Psychology and Aging,
17, 423– 434.
Rosenberg, E. L., & Ekman, P. (1995). Conceptual and methodological
issues in the judgment of facial expressions of emotion. Motivation and
Emotion, 19, 111–138.
Rossier, J., Dahourou, D., & McCrae, R. R. (2005). Structural and mean-
level analyses of the five-factor model and locus of control: Further
evidence from Africa. Journal of Cross-Cultural Psychology, 36, 227–
Russell, J. (1991). Negative results on a reported facial expression of
contempt. Motivation and Emotion, 15, 281–291.
Russell, J. A. (1994). Is there universal recognition of emotion from facial
expressions? A review of the cross-cultural studies. Psychological Bul-
letin, 115, 102–141.
Schwartz, B., Tesser, A., & Powell, E. (1982). Dominance cues in non-
verbal behavior. Social Psychology Quarterly, 45, 114 –120.
Sedikides, C., Gaertner, L., & Vevea, J. L. (2005). Pancultural self-
enhancement reloaded: A meta-analytic reply to Heine (2005). Journal
of Personality and Social Psychology, 84, 539 –551.
Spiegel, J., & Machotka, P. (1974). Messages of the body. New York: Free
Stipek, D. (1998). Differences between Americans and Chinese in the
circumstances evoking pride, shame, and guilt. Journal of Cross-
Cultural Psychology, 29, 616 – 629.
Stipek, D., Recchia, S., & McClintic, S. (1992). Self-evaluation in young
children. Monographs of the Society for Research in Child Development,
57(1, Serial No. 226).
Tangney, J. P., & Dearing, R. L. (2002). Shame and guilt. New York:
Tangney, J. P., Wagner, P., & Gramzow, R. (1989). The Test of Self-
Conscious Affect (TOSCA). Fairfax, VA: George Mason University.
Tracy, J. L., & Matsumoto, D. (2007). More than a thrill: Cross-cultural
evidence for spontaneous displays of pride in response to victory at the
Olympic Games. Manuscript submitted for publication.
Tracy, J. L., & Robins, R. W. (2004a). Putting the self into self-conscious
emotions: A theoretical model. Psychological Inquiry, 15, 103–125.
Tracy, J. L., & Robins, R. W. (2004b). Show your pride: Evidence for a
discrete emotion expression. Psychological Science, 15, 194 –197.
Tracy, J. L., & Robins, R. W. (2007a). The prototypical pride expression:
Development of a nonverbal behavioral coding system. Emotion, 7,
789 – 801.
Tracy, J. L., & Robins, R. W. (2007b). The psychological structure of
pride: A tale of two facets. Journal of Personality and Social Psychol-
ogy, 92, 506 –525.
Tracy, J. L., & Robins, R. W. (2007c). Self-conscious emotions: Where
self and emotion meet. In C. Sedikides & S. Spence (Eds.), The self in
social psychology. Frontiers of social psychology series. New York:
Tracy, J. L., & Robins, R. W. (in press). The automaticity of emotion
Tracy, J. L., Robins, R. W., & Lagattuta, K. H. (2005). Can children
recognize the pride expression? Emotion, 5, 251–257.
United Nations human development report. (2005). Retrieved December 5,
2005, from http://hdr.undp.org/reports/global/2005/pdf/presskit/
Wagner, H. L. (1983). On measuring performance in category judgment
studies of nonverbal behavior. Journal of Nonverbal Behavior, 17, 3–28.
Weiner, B. (1985). An attributional theory of achievement motivation and
emotion. Psychological Review, 92, 548 –573.
Weisfeld, G. E., & Beresford, J. M. (1982). Erectness of posture as an
indicator of dominance or success in humans. Motivation and Emotion,
Received December 8, 2005
Revision received October 17, 2007
Accepted November 2, 2007 䡲
E-Mail Notification of Your Latest Issue Online!
Would you like to know when the next issue of your favorite APA journal will be available
online? This service is now available to you. Sign up at http://notify.apa.org/ and you will be
notified by e-mail when issues of interest to you become available!
530 TRACY AND ROBINS