Are facial displays social? Situational influences in the attribution of emotion to facial expressions.
ABSTRACT Observers are remarkably consistent in attributing particular emotions to particular facial expressions, at least in Western societies. Here, we suggest that this consistency is an instance of the fundamental attribution error. We therefore hypothesized that a small variation in the procedure of the recognition study, which emphasizes situational information, would change the participants' attributions. In two studies, participants were asked to judge whether a prototypical "emotional facial expression" was more plausibly associated with a social-communicative situation (one involving communication to another person) or with an equally emotional but nonsocial, situation. Participants were found more likely to associate each facial display with the social than with the nonsocial situation. This result was found across all emotions presented (happiness, fear, disgust, anger, and sadness) and for both Spanish and Canadian participants.
- [Show abstract] [Hide abstract]
ABSTRACT: This article examines the importance of semantic processes in the recognition of emotional expressions, through a series of three studies on false recognition. The first study found a high frequency of false recognition of prototypical expressions of emotion when participants viewed slides and video clips of nonprototypical fearful and happy expressions. The second study tested whether semantic processes caused false recognition. The authors found that participants made significantly higher error rates when asked to detect expressions that corresponded to semantic labels than when asked to detect visual stimuli. Finally, given that previous research reported that false memories are less prevalent in younger children, the third study tested whether false recognition of prototypical expressions increased with age. The authors found that 67% of eight- to nine-year-old children reported nonpresent prototypical expressions of fear in a fearful context, but only 40% of 6- to 7-year-old children did so. Taken together, these three studies demonstrate the importance of semantic processes in the detection and categorization of prototypical emotional expressions.Emotion 09/2008; 8(4):530-9. · 3.88 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Free download on the link: http://airccse.org/journal/jma/5513ijma05.pdf DynEmo is a database available to the scientific community (https://dynemo.upmf-grenoble.fr/). It contains dynamic and natural emotional facial expressions (EFEs) displaying subjective affective states rated by both the expresser and observers. Methodological and contextual information is provided for each expression. This multimodal corpus meets psychological, ethical, and technical criteria. It is quite large, containing two sets of 233 and 125 recordings of EFE of ordinary Caucasian people (ages 25 to 65, 182 females and 176 males) filmed in natural but standardized conditions. In the Set 1, EFE recordings are associated with the affective state of the expresser (self-reported after the emotion inducing task, using dimensional, action readiness, and emotional labels items). In the Set 2, EFE recordings are both associated with the affective state of the expresser and with the time line (continuous annotations) of observers’ ratings of the emotions displayed throughout the recording. The time line allows any researcher interested in analysing non-verbal human behavior to segment the expressions into emotions. Free download here: http://airccse.org/journal/ijma_current13.htmlThe International Journal of Multimedia and Its Applications. 10/2013; 5(5):61-80.
- [Show abstract] [Hide abstract]
ABSTRACT: Collective rituals are biologically ancient and culturally pervasive, yet few studies have quantified their effects on participants. We assessed two plausible models from qualitative anthropology: ritual empathy predicts affective convergence among all ritual participants irrespective of ritual role; rite-of-passage predicts emotional differences, specifically that ritual initiates will express relatively negative valence when compared with non-initiates. To evaluate model predictions, images of participants in a Spanish fire-walking ritual were extracted from video footage and assessed by nine Spanish raters for arousal and valence. Consistent with rite-of-passage predictions, we found that arousal jointly increased for all participants but that valence differed by ritual role: fire-walkers exhibited increasingly positive arousal and increasingly negative valence when compared with passengers. This result offers the first quantified evidence for rite of passage dynamics within a highly arousing collective ritual. Methodologically, we show that surprisingly simple and non-invasive data structures (rated video images) may be combined with methods from evolutionary ecology (Bayesian Generalized Linear Mixed Effects models) to clarify poorly understood dimensions of the human condition.Frontiers in Psychology 01/2013; 4:960. · 2.80 Impact Factor
The Spanish Joumal of Psychology
2m2, Vol. 5, No. 2,119-124
CopyrigN 2002 by Thc Spanish Joun,aI of Psychology
Are Facial Displays Social? Situational Influences in the
Attribution of Emotion to Facial Expressions
José-Miguel Femández-DoIs’, Pilar Carrera’, aud James A. Russell2
‘Universidad Autónoma de Madrid
Observers are remarkably consistenr in attributing particular emotions lo particular facial
expressions, at least in Weslern societies. Here, we suggest that this consistency is an
insíance ofte fundamental altribution error. We rherefore hypothesized that a small
variation in the procedure of the recognition study. which emphasizes situational
information, would change Ihe participanís’ attributions. In two studies, participants were
asked ro judge whetheraprorotypical “emotional facial expression” was more plausibly
associated with a social-communicative situation (one involving conirnunication lo anoiher
person) or with an equally emotional buí nonsocial, situation. Participants were found
more likely to associate each facialdisplay with ihe social iban with the nonsocial situation.
This resulí was found across alí emotions presented (happiness, fear, disgust, anger, and
sadness) and for both Spanish and Canadian participants.
Key words:facial expression, recognition of amonan, amellan, amellan axprassion
La atribución de emociones a determinadas expresiones faciales es un fenómeno
notablemente robusto, al menos en las sociedades occidentales. En este artículo
proponemos que la consistencia de dichas atribucioneses un caso de error fundamental
de atribución. Si nuestra hipótesis escorrecta, pequeñas variaciones en el procedimiento
de los estudios típicos sobre reconocimiento de emociones (dando un mayor énfasis a
la información situacional) cambiarán de manera sustantiva la forma en la que los
perceptores atribuyen emociones a las expresiones faciales. Para comprobar dicha
hipótesis hemos llevado a cabo dos estudios en los que los participantes deben decidir
si una expresión prototípica de emoción se asocia con una situación social (que implica
comunicación con otra persona) o con una situación emocional pero no social. Nuestros
suietos asociaron la expresión facial con la situación social, en lugar de la no social. Los
resultados fueron los mismos para todas las emociones consideradas (alegría, miedo,
asco, enfado y tristeza) y tanto para su¡etos españoles como para los canadienses.
Palabras clave: expresión facial, reconocimiento de emociones, emoción, expresión
Wc thank Lara Weick and Lisa Wong for their excellent work on Ibis study.
Correspondence concerning tbis anide sbould be addressed ¡o José-Miguel Fernández-Dols, Facultad de Psicologia. Universidad
Autónoma de Madrid. Ciudad Universitaria de Cantoblanco. 28049 Madrid (Spain). E-mail: firstname.lastname@example.org
FFRNÁNDEZ-DOLS, CARRERA. AND RUSSELL
Are Facial Displays Social?
The pirase “facial cxpression of emotion” captures the
common-sense idea thaI certain facial movernenís are closely
tied with emotion. Influential theories in Ihe psychology of
emotion maintain fis tie by assuming that facial displays
are a spontaneous result of, part of, or precursor to the
occurrence of emotion (e.g., Buck, 1984; Ekman, 1972).
According to such views, human beings universally recognize
rhc emotion signalled by the facial display. (For a review
of sorne alternative views, see Russell, Bachorowski, &
Femández-Dois, in press).
Consensus on the “recognition” (attribution) of emotion
from a few facial expressions is well docurnented with
Westem samples, but it is far from being as consistení and
extensive in non-Western, especially illiterate societies (see
Russell & Fernández-DoIs, 1997). Figure
summary of relevant rcsults from cross-culrural judgrnent
studies. TEe figures given are average “recognition seores”
(the percentage of participanís who select fe predicted
emotion) for six “basic expressions” (happiness, surprise,
anger, sadness, fear, and disgust). The f¡rst bar comes from
Westem literate societies; the magnitude is impressive. Ihe
secoad bar comes fi-orn nun-Western literate societies; this
second score is reliably lower than the first but sílíl high.
The third bar comes from more isolated, illiterate samples;
these participants attributed happiness to smiles but yielded
noticeably lower recognition scores with alí other expressions
(Russell, 1994; see also Elfenbein & Ambady, 2002).
1 provides a
U Amount of “recognition”
Figure 1. Average recognition scores for six facial expressions of
emotion (happiness, surprise, anger, sadness, fear, and disgust)
across Western linerane, non-Western literate, and isolated illiterate
cultures. Recognition scores are proportional lo Ihe amount of
Western influence (adapted from Russell & Femández-DoIs. 1997).
In this article, we suggest that receivers’ remarkable
consistency when attributing a particular emotion to a
particular expression is an example of the fundamental
atíribution error: People, particularly those from Wcstern
cultures, underestimate situational infitiences, v,tile
inferring stable abilities, altitudes, and personality traits
that lead them to expect “consistency in behavior or
outcomes across widcly disparate situations and contexts”
(Ross, 1977, p. 184).
Indeed fe design of the recognition siudies encourages
or presupposes fe fundamental altribution error by
ignoring the context in which thc facial display takes place.
People guide their artributions about facial expression and
emotion by ideal, simplified, and consistent representations
of situations in which they lump together emotional
experience and facial behavior that actually happened al
different times (Fernández-DoIs, 1999). For example,
affiliative smiling and happiness are usually observed in
Ihe same interactive situatiofis, which leads people to
believe that srniling is a sign of happiness across ah kinds
of situations and contexís.
In the experiments reponed here, we test whether people
are capable of differentiaíing the actual immediate experience
of emotion from the surrounding interaction.
When are Facial Displays Thought to Occur?
Fridlund (1994) suggested that the social motive, rather
than the emotion, elicited by the presence of anotherperson
would be fe mosí powerful determinaní of Ihe occurrence
of a facial display. In contrasí, on the standard emotional
accouní of facial displays, the presence of another person
is secondary, in íhaí it might, buí need not, invoke a
“display rule” (Ekman, 1972; Klineberg, 1940). thereby
inhibiting, exaggeraíing, or masking ihe spontaneous facial
A simple variation of the usual procedure in recognition
studies will allow us to test whether, when provided with
situational information, people link “expressions ofemotion”
to thc experience of ernotion (“immediate experience”
condition) or rather to olber processes such as Une
communication of social motives.
Participants were shown two photographs of the saíne
person: one with a proíoíypical “emotional” expression aud
one wiíhout (“neutral”). Participants werc also given twa
sentences. both expressing the corresponding emotion
coníent.in one seníence (nonsocial condition),
protagonist was rhinkiu¡g Ihe sentence. In the other (Ihe
social-communicative coi~dition), the protagonist was
speaking ihe sentence. For instance, in one trial concerning
happiness, the facial expression was a smile and the
sentences concerned winning the lottery. In ihe nonsocial
condition, fe protagonist thought, “We have won ihe prize!”
In the social condition, the protagonisí spoke to someone,
saying “John, we have won ihe prize!” The participant’s
task was Lo pair each situation with a photograph.
If Ihe emotion-based account of facial displays is correct,
participants would consider íwo altematives: (a) Iffe social
situation suggested a display rule, thai is, the inhibition of
te exprcssion, the participapí would associate the social
situation with ihe neutral face. The nonsocial situation would
ihen be paired with ihe intense, fulí expression; (b) lf ihe
participant did not perceive any reason for inhibiíing ihe
facial display in ihe situation, ten the choice would be
random. Overalí, ten, 50% or fewer subjects would pair te
facial display with ihe social situation, 50% or more would
pair it with the nonsocial situation. On ihe other hand, if our
hypothesis is corred, participants’ choices should be ruled
by an opposite set of assumptions. Facial displays would be
judged as communicative behavior and most participanís
would associate fe facial display with the social situation.
Participanis were 20 adults, of whom half were men,
half women, and alí students of te Universiíy of British
Columbia (Canada). Participants were selected from Ihe
volunteer list of first and second year psychology siudenis,
and given credit toward iheir course marks.
There were 11 pairs of photographs. Each pair consisted
of ihe same actor with a neutral and with one prototypical
“expression of emotion.” There were two each of fear,
happiness, sadness, and disgust. and three of anger. One
expression of fear carne from Camras’ collection (Camras,
Grow, & Ribordy, 1983), one of happiness from Russell’s
collection (Russell & Bullock. 1985), and ihe rest from
Ekman and Friesen’s set (Ekman & Friesen, 1976).
There were 11 pairs of sentences (two each of fear,
happiness, sadness, and disgusí. and three of anger), with
both sentences in each pair describing te same emotional
content. In one sentence, Ihe person was th¡nking about Ihe
situation (cg., “So ihis means thaI fis food is spoiled!”),
whereas in ihe other sentence, the person was speaking to
someone about it (e.g., “Look al ihis food. It’s spoiled!”).
To emphasize fe difference between the two conditions,
Wc borrowed aconvention froin cartoons, as shown in Figure
2. Ihe social/spoken versions of fe 11 sentences are given
It might be argued that even though the contení of the
seníences was te same in Ihe social and the nonsocial
situaíions, the social situations are inherently more emotional.
If so, participanís here simply associated the facial displays
with Ihe more intense emotional experience. For sorne
seníences, the social version indeed might be somewhat more
Figure 2. Sentence pair offered lo parricipanis in Trial 3 (Anger). Study 1.
Display rules allow not only inhibition buí also masking nr exaggeration. These possibililies were noí considered to be among Ihe
participanis’ choices here.
FERNÁNDEZ-DOLS, CARRERA, AND RUSSELL
emotional; for instance, “You harmed our son!” implies direcí
confrontation, whereas “Then, my own husband harmed my
son!” does nol. Never[heless, [he opposite is probably the
case for other sentences. For instance, discovering thaI
someone is hidden in fe kiíchen or being followed by a
sírange man would likely creaíe more fear if alone.
Subjec[s were tesíed individually. Ihere were II trials.
Bach [rial concerned one emotion. Tlie experimenter
presented ihe first pair of photographs and corresponding
pair of sentences. The participan[ was asked to pair faces
and seníences in [he “more plausible” or “more natural”
way. The experimenter poiníed out thaI one sentence was
íhough[, Ihe o[her spoken, and Iha[ Ihere were no right or
wrong answers. A separate random order of the 11 trials
was created for each participaní.
Experimení 2 was a replication of Experimení 1 wiíh a
different ser offacial expressions (Malsumoto& Ekman. 1988)
and a sample thai included Canadian and Spanish participanís.
The method was identical lo thaI of Experimení 1, with
[he following exceptions. Participanís were 40 adults, 20
Spanish aud 20 Canadian flrst- and second-year psychology
siudenis, of whom half were men, half wornen. There were
10 trials instead of 11, and ihe coníení of several of the
sentences was modified slightly, as shown in Table 2.
Photographs were of 10 different models (alí Westem and
female) taken from Maisurnoto and Ekman (1988).
Table 1 shows Ihe percentage of par[icipants associa[ing
ihe facial displays with [hespoken sentence, thai is, ihe social
situalion. Overalí, facial expression was associated wi[h ihe
spoken sentence on 178 of [he 220 trials (80.9%). For each
type ofemotion, a majoiity ofpafflcipants (from 75% [o92.5%)
placed Ihe facial expression with ihe spoken version of Ihe
sen[ence. A Wilcoxon signed ranks tesí, comparing Ihe
magnitude and direc[ion of Ihe differences wi[hin pairs of
possible choices for the 11 trials (facial expression associaied
with spoken sentence vs. neutral face associated with te spoken
sentence) was highly significan[ (z = 2.94, p < .003). Moreover,
every par[icipant individually showed Ihis pat[ern; across the
11 trials, [he median number of spoken sen[ences associated
wi[h emotion expression was 9, wi[h a range of 6 to 11.
Table 2 shows [he percentage of par[icipanís associating
[he facial display wi[h ihe spoken sentence (social situa[ion).
Overalí. te resulis were similar lo Ihose obíained in Síudy
1. Facial displays were associated wiíh a social situalion on
288 of ihe 400 Irials (72%). For each Iype of emotion, a
majori[y of participanís (from 62.5% to 86.2%) placed the
facial display with [he spoken version of Ihe sentence. A
Wilcoxon signed ranks [est, companng the magni[ude and
direction of [he differences within pairs ofpossible choices
for ihe 10 trials (facial expression associated wiíh spoken
sentence vs. neu[ral face associated with Ihe spoken sentence)
was highly significan[ (z = 2.66, p < .008, for Ihe whole
sample; z = 2.68, p .< .007, for the Canadian subsample;
and z = 2.43, p < .015, for the Spanish subsample). Over
ihe 10 [rials, ihe median number of spokcn seníences
of Suhjects Associating Facial Displays with a Spoken, Social Situation
Spoken Version of Sentence
“John, we have won the prize!”
“John, our son will come back from [hewar tomorrow!”
“You harmed our son!”
“Peter, you cheated me!”
“Lionel, you have made me miss ihe bus!”
“John, 1 miss my son so much!”
‘Sweetheart, 1 have not passed <he exam!”
‘Please, help me! There is somebody hidden in my kitchen!”
“John. please help me! A sirange man is followinr me!”
“Look at this dead rat!”
“Look al <his food, it’s spoiled!”
Note. Canadian sample, N = 20.
Percentage cf Subjects Associating Facial Displays with a Spoken, Social Situation
EmotionSpoken Version of Sentence Spanish % Canadian %
Happiness“John, we have won <he prize!”
“John, our brother wilI come back fi-orn <he war tomorrow!”
“You, my own husband, are playing a trick on me!”
“Lionel, you have made me miss the bus!”
‘John, 1 miss my sonso much!”
“Sweetheart, 1 have noí passed ihe exam!”
‘Please, help me! There is somebody hidden in my kitchen!”
‘John, please help me! A sirange man is following me!”
“Look at this dead ral!”
‘Look al this food, it’s spoiled!”
Note. Spunish saniple, IV = 20, Canadian 5ampie, IV = 20.
associated with facial displays was 7, wi[h a range of 6 [otO
for fe Canadians, and 6.5 for fe Spaniards, with a range
of 3 [o 10. Ah in ah, na[ional¡[y
change [he basic pa[tern of responses already described.
An al[emative explanation [o [hese resul[s might be as
follows: <a) Par[icipan[spaired te intense facial display
wiíh Ihe spoken sentence because [he other possible pairing
(display plus though[ sentence) seerned sornehow unna[ural,
or (b) par[icipan[s paired ihe neutral display wiúi ihe though[
sentence because ihe combination of neutral expression wi[h
Ihe spoken seníence seerned somehow unnatural.
To explore Ihese possibilities, a small study was
conduc[ed. The 40 possible combinations of each pair of
facial displays (neutral plus expressive) and [he sentences
(Ihough[ plus spoken) used in Experirnení 2 were formed.
Four groups of 10 combinations each were formed so as lo
avoid repe[ítions of either facial display or sentence. Forty
paríicipan[s were assigned to one of te four independení
groups. (Bach group íherefore judged
cornbinations.) Por alí cornparisons ofin[erest, [he data were
[herefore between-subjec[s. Their task consísted of judging
“how na[ural” was each combination on a seale from O
(nzaxinzally ¡innatural) to 10 (naxirnally natural).
No combination of the facial display wi[h [he thoughí
sentence was considered significar¡tly less nalural Ihan thai
display with the spoken sentence. In three cases, however, [he
combinaúon of fe neutral face with fe spoken sentence was
considered significan[ly less natural Ihan the combination of
[heneutral face wif the [houghísentence: [hefirsí, fifth, and
sixth sentences (t = —2.28,p < .05; t = 2.39, p < 05; t = —3.24,
p < .01, respecúvely) Usted in Table2, ‘¡be resul[s of Experimení
2 were re-analyzed wiíh these three sentences ornilted. A
Wilcoxon signed ranks tesí, cornparing ihe magnilude and
direc[ion ofthe differences wiíhin pairs ofpossible judgrnenís
for [he seven trials (facial expression associated with spoken
sentence vs. neutral face associated wiíh [he spoken sentence)
was highly significaní (z = 2.36, p c .018). Facial expression
was found lo be associated with Ihe social si[ua[ion.
of paríicipan[s did no[
In [he s[udies repor[ed here, we have found [ha[
par[icipants were able Lo differen[iaíe ihe actual immediate
experience of emotion from [he surrounding interaction.
A simple manipulalion, ernphasizing sítua[ional ínforma[ion,
allowed us to show [hal people do no[ necessarily línk
expressions of emotion [o [he experience of emotion
(immedíate experience condi[ion),
actions. The usual “recognition” studies do nol [ake mío
accoun[ ihese other ac[ions (e.g., communica[ive ac[ions).
Thus, subjec[s are forced into making raw, large-scale
correlationsbe[ween facial behavior and emotion
Sorne addi[ional evidence supports Ihese findings.
Carroll and Russell (1996) carried ouí a s[udy
recognition of ernotion in which they provided par[icipan[s
with sorne sítua[ional informa[ion
explain [he physical features of ihe proío[yp¡cal
expressions of emotion (e.g., staring eyes, furrowed
eyebrows, lips pressed tigh[ly
nonernolional ac[ions included in Ihe si[ua[íon
person peering a[ a distant message and having difficujíy
in making a decision about it). They found [hat people
judged facial expressions in terrns of [he situation instead
of the expected emolion, reversing [he usual [rend in
recognition siudies. Por exarnple, what is usually Jabeled
an “expression of anger” became an “expression of
puzzlemení” when it was displaying physical ac[ions Iha[
were plausible for nonernotional reasons in a given
coniexí. ¡‘copIe could “recognize” [ha[ a “universal
expression” is nol necessarily related [o an emotion. Sorne
saliení situational hints helped Ihem to overcorne [he bias
[hat usually leads Ihem ío “recognize” universal
Therefore, ourresuhis encourage pursuit of [he idea tha[
people “recogni[ion” of ernotion is an instance of a<[ributional
bu[ raiher fo other
tha[ allowed Ihem [o
toge[her in anger) by
FERNÁNDEZ-DOLS, CARRERA, AND RUSSELL
fluck, R. (1984). The communicorion of cinchan. New York:
Camras, L.A., Grow, V.G., & Ribordy, SC. (1983). Recognition
of emotional expression by abused children. Journal of Clinicol
Child Psychology 12, 325-328.
Carroll, J.M., & Russell, lA. (1996). Do facial expressions signal
specific emotions? Judging the face in context. Jaurnal of
Pcrsonality and Social Psycbology, 70, 205-218.
Ekrnan, P. (1972). Universals and cultural differences in facial
expressions of emotions. In J, Cole (al.), Nebraska Sytnposiuni
en Motivation: Vol. 19. 1971 (pp. 207-283). Lincoln, NE:
Universi<y of Nebraska Press.
Ekman, E, & Friesen, W.V. (¡976). Pictures offacial affecr. Palo
Alto, CA: Consulting Psychologists Press.
Elfenhein, HA., & Ambady, N. (2002). On <he universality and
cultural specificity of emotion recognition : A meta-analysis.
Psychological Bulletin, 128, 203-235.
Fernández-Dois, J.M. (1999). Facial expression and emo<ion: A
si<uationis< view. lo E Philippot, R.S. Feldrnari, & EJ. Coats
(Eds.), fije social context of nonverbal heliavior (PP. 242-261).
New York: Carnbridge University Press.
Fridlund, Al. (1994). Human facial expression: Aa evolutionary
view. NewYork: Academic Eress.
Klineberg, 0. (¡940). Socialpsycltology. New York: Holt.
Marsumoto. D., & Ekman, P. (1988). Japanese and Caucasion
facial expressionsof enwtion (JACFFE~. Slide and brochure.
San Francisco, CA: San Francisco State Universi<y. (Available
from the first aulbor.)
Ross, L. (1977). The intuitive psychologistand bis shortconiings:
Distortions u ihe attribuíion process. Advances in Experimental
Social Psvcl;ology 10, 174-221.
Russell, JA. (1994). Ls there universal recognition ofemotion fi-orn
facial expressions? A review of cross-cul<ural studies.
Psycbo¿ogical Bufletin, 110,426-450.
Russell. JA., Bachorowski, J-A.. & Femández-DoIs, J.M. (lii press).
Facial and vocal expressions of emorion. Annual Review of
Russell, JA., & Bullock, M. (1985). Multidimensional scaling of
emotional facial expressions: Similarity from preschoolers (o
adults. Joarnal of Personalihanó SocialPsychology 48, 1290-
Russell JA., & Fernández-DoIs, J.M. (1997). Wha< does a facial
expression mean? In J.A. Russell & J.M. Fernández-DoIs (Eds.),
Tl2e psychology offacial expression (PP. 3-30). New York:
Cambridge University Press.
Received September 2, 2002
Revision received September 16, 2002
Accepted September 30, 2002