1 © 2011 American Psychological Association Emotion ● , 000–000 1528-3542/11/$12.00 DOI: 10.1037/a0026119 , No. ●● 2011, Vol. Can An Anger Face Also Be Scared? Malleability of Facial Expressions Sherri C. Widen and Pamela Naab Boston College Do people always interpret a facial expression as communicating a single emotion (e.g., the anger face as only angry ) or is that interpretation malleable? The current study investigated preschoolers’ ( N 60; 3–4 years) and adults’ ( N 20) categorization of facial expressions. On each of five trials, participants selected from an array of 10 facial expressions (an open-mouthed, high arousal expression and a closed-mouthed, low arousal expression each for happiness, sadness, anger, fear, and disgust) all those that displayed the target emotion. Children’s interpretation of facial expressions was malleable: 48% of children who selected the fear, anger, sadness, and disgust faces for the “correct” category also selected these same faces for another emotion category; 47% of adults did so for the sadness and disgust faces. The emotion children and adults attribute to facial expressions is influenced by the emotion category for which they are looking. facial perception, emotion, child development, malleability, categorization Keywords: categorization tasks, children include the target face in the cate- In their day-to-day experiences, people see and interpret others’ emotional reactions, including emotional facial expressions. In gory, but also include other faces of the same valence and similar traditional emotion theory, such facial expressions are assumed to levels of arousal. As age increases, children are more likely to communicate single, discrete emotions (e.g., Ekman, 1980). Other exclude “incorrect” facial expressions, indicating that their emo- evidence suggests that interpretation of facial expressions is more tion concepts are becoming narrower and more adult-like. For malleable—that is, a person may interpret the same face as ex- example, when young preschoolers are asked to find the angry pressing a variety of emotions in different contexts (e.g., Carroll & faces on a categorization task, they will include many negative Russell, 1996; Aviezer et al., 2008). This study is part of a larger valence emotions: The angry face and also the disgust, fear, and project investigating children’s understanding of emotion. It asked sad faces (Widen & Russell, 2008a). As age increases, children how children understand facial expressions: Do facial expressions exclude the low arousal sad face first from the high arousal anger communicate only one emotion for children? Or are the emotions category. Then they begin to exclude the other higher arousal that a facial expression communicates more malleable for them? faces—the fear face, and finally the disgust face. The malleability of facial expressions suggests that, while faces An additional factor that may influence which faces children remain preeminent, the information they communicate is limited “incorrectly” include in an emotion category, or which faces are (e.g., where the person is looking or staring, whether the person is more malleable, is perceptual similarity. Facial expressions that crying, smiling, grimacing, etc., and the person’s levels of va- look most similar, although assumed to express different emotions, lence—feels good vs. feels bad—and arousal; Carroll & Russell, are more likely to be included in the same category. For example, 1996). For example, when presented with an “anger” face, young anger and disgust look similar though they share only incidental preschoolers interpret it as “feels bad” rather than “feels good.” muscle movements such as wrinkles around the brow and possibly They may also interpret this face as having high rather than low raising of the upper lip (Widen & Russell, 2008b). arousal levels. The valence and arousal dimensions help to explain 1 which faces children “incorrectly” include in an emotion category (e.g., Bullock & Russell, 1984; Widen & Russell, 2008a). On Overview of the Study This study investigated how children and adults categorized high and low arousal facial expressions of happiness, sadness, Sherri C. Widen, Department of Psychology, Boston College; and anger, fear, and disgust. There were two facial expressions for each Pamela Naab, Department of Counselling, Developmental, & Education emotion: an open-mouthed (high arousal) expression and a closed- Psychology, Boston College. mouthed (low arousal) expression. To test the assumption that the We thank the staff, parents, and children at the daycares that participated in this study. Without their help, this research could not be done. We thank open-mouthed expressions were perceived as having higher Jennifer Gallucci, Nicole Nelson, Mary Kayyal, and Kerrie Pieloch for their comments on earlier versions of this article. This study was funded by a grant from the National Science Foundation (0421702). 1 On our perspective, we believe that children’s nontarget responses Correspondence concerning this article should be addressed to Sherri C. reveal a great deal about their understanding of emotion categories and that Widen, Department of Psychology, McGuinn Hall, 140 Commonwealth it is important to analyze all of children’s responses on emotion tasks, both Avenue, Boston College, Chestnut Hill, MA 02467. E-mail: “correct” and “incorrect.” [email protected] 1
2 WIDEN AND NAAB 2 2 20) arousal than the closed-mouthed expressions, adults ( N Our third prediction was that these inclusions would be medi- ated by perceptual similarity. For example, disgust and anger have rated (on a 7-point Likert scale: 0 extreme sleepiness to 7 similar levels of arousal as well looking similar; these faces should extremely high arousal ) the arousal level of the photographs used be “incorrectly” included in the other’s category with high fre- in the current study. Open-mouthed facial expressions were judged quency. as more highly aroused ( 4.65) than closed-mouth facial M expressions ( t 3.57), M p 8.75, .001. This pattern held for 19 4.75, low: M 3.25), each of the pairs: happiness (high: M Method t 8.82, 4.00, low: M .001; sadness (high: p M 19 2.60), t 4.70, low: M p 3.56, M .002; anger (high: 19 Participants 3.05), t 5.15, low: 5.05, p .001; fear (high: M M 4.70), 19 Participants were 60 children, all proficient in English and t M 4.65, p 2.65, .02; and marginally for disgust (high: 19 enrolled in daycares in the Greater Boston area. There were 15 t 4.25), M low: 2.03, p .06. The rank order of each of the 19 boys and 15 girls in each age group: 3-year-olds (36 to 47 months, 10 facial expressions (from high to low arousal) was: open- SD 42.1 month, mean age 3.5 months) and 4-year-olds (48 to mouthed fear, open-mouthed happiness, closed-mouthed fear, 5.0 months). Of the SD 65 months, mean age 57.0 months, open-mouthed anger, open-mouthed disgust, closed-mouthed dis- total sample, 71% were Caucasian, 13% Hispanic, 8% Asian, 8% gust, open-mouthed sadness, closed-mouthed happiness, closed- other ethnicities. A group of 20 university-aged adults (mean mouthed anger, and closed-mouthed sadness. The midpoint of the 18.9 years; 15 female) was also included as a comparison age scale (3.5) was used as the neutral point between high and low group; course credit was given in exchange for participation. Of arousal. Thus, three faces in this array were low arousal (closed- the total adult sample, 55% were Caucasian, 25% Asian, 10% mouthed happiness, closed-mouthed anger, and closed-mouthed Hispanic, and 10% other ethnicities. sadness) and the other seven were high arousal. On each of five trials (happy, sad, angry, scared, disgusted), the Materials child saw all 10 facial expressions at once and was asked to find the one(s) that felt the target emotion (e.g., Which one of these 5” pho- The facial expressions were 10 black-and-white 3” people feels happy?). The child continued to make selections until tographs of women posing prototypical facial expressions of emo- he or she indicated that none of the remaining faces displayed the tion, selected from Ekman and Friesen’s (1976) Pictures of Facial target emotion (e.g., “Does anyone else feel happy? Or did you get Affect. Only photographs of Caucasian women were included them all?” This phrasing was intended to reduce demand chara- because we did not want to confound differences in children’s cteristics on the child to select more faces if he or she felt that all attributions of emotion to males and females (e.g., Widen & the target faces had been found). We also included the instruction, Russell, 2002) or race (e.g., Tuminello & Davidson, in press) with at the beginning of the procedure, “Look at these faces. See how the malleability of their emotion concepts. There were two pho- some of them feel the same but some of them feel different?” to tographs for each emotion: One set was high arousal (open- encourage children to notice the similarities and differences be- 18: 25; sadness C1 mouthed) (happiness JM1 4: AUs 6 12 tween the faces before the trials began. Together, these instructions AUs 1 4 11 25 23 26; anger MF2 7: AUs 4 5 reduced the likelihood that children would select all the faces 25 31; disgust 26; fear C1 23: AUs 1 2 5 11 25 while allowing them the flexibility to include different faces in NR2 7: AUs 9 25) and one set was low arousal (closed- each emotion category. mouthed) (happiness PF1 6: AU 12; sadness A2-6: AUs 1 4 Adults were included as a comparison group to illustrate the 24; fear MF1 5 2 30: AUs 1 12: 4 18; anger C2 11; end-point of development. That is, if adults included faces in 17). 8: AUs 9 disgust JM2 addition to the target face for any of the emotion categories, then it would be unreasonable to expect children to include only the Procedure target faces for those categories. The experimenter spent the first visit at each preschool getting Our first prediction was that children would include both the to know each child who had received parental consent. On a high arousal (open-mouthed) and low arousal (closed-mouthed) subsequent visit, the experimenter invited each child, individually, target facial expressions on all trials. That is, when asked find the to play a game (participate in the study) with her. happy people, children would include both the open- and closed- mouthed happiness facial expressions. Similar categorization stud- ies have reported children’s inclusion of target faces (e.g., Bullock Choice-From-Array Task & Russell, 1984; Widen & Russell, 2008a). In the choice-from-array task, the 10 facial expressions were Our second prediction was that nontarget facial expressions presented at once, and the child was encouraged to look closely at would be included in each category: High arousal (open-mouthed) each one. The experimenter introduced the faces by saying, “Look expressions would be included in the high arousal categories of at these people [pointing]. See how some of them feel the same but happiness, anger, disgust, and fear; low arousal (closed-mouthed) expressions would be included in the low arousal category of sadness (e.g., Widen & Russell, 2008a). For example, for the anger 2 The adult sample was composed of five males and 15 females. The category (a high arousal emotion), the high arousal (open- SD mean age of was, mean age 18.9 years, .9 years. Of the total mouthed) sadness face would be a more likely selection than sample, 55% were Caucasian, 25% Asian, 10% Hispanic, and 10% other ethnicities. would the low arousal (closed-mouthed) sadness face.
3 MALLEABILITY OF FACIAL EXPRESSIONS 3 .01) p .003) and scared (3 years: 3.37, 4 years: 1.97; p sons, some of them feel different? I’m going to ask you to pick the people that feel a certain way. Only pick the people that feel that categories and marginally significant for the anger category (3 way. Remember: not all the people feel the same.” The experi- .06). The differences between age years: 3.63, 4 years: 2.60; p menter then began the first trial by asking the child, “Which one of groups for sad (3 years: 2.83, 4 years: 2.27) and disgust (3 years: these people feels X? (happy, sad, angry, scared, disgusted).” 3.03, 4 years: 2.73) were not significant but followed the same When the child had made a selection, he or she was mildly praised, trend. For each category, 4-year-olds included fewer faces than and the experimenter removed that photograph. The child was then 3-year-olds, indicating that emotion categories narrowed with age. asked, “Does anyone else feel X? Or did you get them all?” This The Age Sex Emotion-Category interaction was also procedure was repeated until the child indicated that no one else 2.81, F significant, p .02, but only one comparison (4, 224) felt the target emotion (or until there were no photographs left), between sexes was significant: For the 3-year-olds, boys included and then the next trial was introduced with a new set of photo- more faces than girls in the happy category (boys: 4.87, girls: 3.07; graphs. Thus, throughout the experimental trials every effort was p .02). In addition, one comparison was marginally significant: For made to both let the child select all the faces he or she felt fit in a the 4-year-olds, boys included more faces than girls in the disgust particular category and also to be clear that not all faces felt the p .09). There was no general trend category (boys: 3.40, girls: 2.07; same and that another selection was not required. The experi- for the boys to include more faces than the girls and the main effect menter shuffled the photographs, displayed them on the floor, for sex was not significant, F (1, 56) p 1.51, .22. selected a different emotion word (“Which one of these people feels Y?”), and continued the procedure until all of the emotion Setting a Criterion for “Incorrect” Inclusions trials were completed. The order of the trials was random as was the order of the facial expressions within the array. Our primary interest was the malleability of children’s interpre- tations of facial expressions—that is, whether they included the Adult Comparison Group same facial expression in more than one category. The first step was to set a criterion against which children’s “incorrect” inclu- The adults completed the choice-from-array task in a question- sions could be compared. Table 1 shows the percentage of children naire format. The 10 faces for each trial were displayed in an array 3 who included each face in each category. Each face was of two rows. On each trial (happy, sad, angry, scared, disgusted), “incorrectly” included in a category by at least one child. To participants were asked to put an X in the circle(s) for the face(s) identify the base rate of incorrect inclusions for each face, the that displayed the target emotion. The order of the trials was mean incorrect inclusion for that face was tallied. For example, for random as was the order of the facial expressions within the array. the high intensity happy face, the base rate was 9.2% 5.0% 15.0% ([5.0% 9.2%). 11.7%]/4 Scoring The second step was to compare the percentage of “incorrect” inclusions for each face on each trial to the base rate for that face On the choice-from-array task, all the faces that each participant using dependent samples tests. Four of the 40 “incorrect” inclu- t selected on a given trial were given a score of 1, all remaining sions (8 possible incorrect inclusions 5 emotion trials) were faces were given a 0. significantly higher than that facial expression’s base rate (see Table 1): Including the low arousal sad face in the scared category, Results and Discussion t .009, the high arousal disgust face in the angry 2.69, p 59 category, t 8.69, p .001, the low arousal disgust face in the 59 Preliminary Analysis of Category Breadth t angry category, p 7.46, .001, and the high arousal anger 59 face in the disgust category, t 3.07, p .003. Two inclusions To investigate the effects of age and sex on the breadth of 59 were marginally significant: Including the high arousal sad face in children’s emotion categories, the number of faces each child the scared category, t 1.70, .09, and the low arousal scared p included on each trial was totaled. In a mixed-design ANOVA 59 t face in the disgust category, .07. These six p 1.86, ( .05), age group (2 levels: 3 years, 4 years) and sex (2 levels) 59 “incorrect” inclusions were included in the analysis of malleabil- were between-subjects factors, and emotion-category (5 levels: ity. happy, sad, angry, scared, disgusted) was a within-subject factor. This process was repeated with the faces that the adults “incor- The dependent variable was the total number of faces included on rectly” included in each category. Two inclusions were signifi- each trial (range 0–10). cantly higher than the base rate for adults (see Table 1): Including The main effect for emotion-category was marginally signifi- t the high arousal sad face in the scared category, p 2.39, p 2.20, (4, 224) F cant, .07. Children included the most faces 19 in the happy (3.12) and angry categories (3.12), followed by disgusted (2.88), scared (2.67), and sad (2.55). The main effect for 3 When the percentage of children who selected each face on their first age was also significant, .03. Both of these 4.81, p F (1, 56) choice was analyzed, the same pattern as is shown in Table 1 was found, main effects were qualified by the significant Age Emotion- though the percentages were lower. Specifically, children’s modal re- p (4, 224) F Category interaction, .03, which indicated 2.75, sponses for the happy, sad, and scared trials were the target expressions that the main effect for age was due primarily to differences faces (range: 23% 47%). For the disgust trial, children chose anger faces between age groups on the happy and scared categories. The 38%) than the disgust faces (13% 18%). For the more frequently (27% difference between age groups was significant for the happy (3 anger trial, they chose the two disgust faces and the high arousal anger face years: 3.97, 4 years: 2.27; Least Significant Difference compari- 22%) than the low arousal anger face (7%). more frequently (13%
4 WIDEN AND NAAB 4 Table 1 Proportion of Participants Who Included Each Face in Each Category Category Base rate “incorrect” Disgusted Scared Facial expression Intensity Angry Sad Inclusions Happy Children 5.0 Happiness High 88.3 15.0 5.0 11.7 9.2 6.7 7.1 5.0 8.3 Low 86.7 8.3 66.7 20.0 High Sadness 33.3 22.9 11.7 26.7 18.3 Low 13.3 76.7 18.3 6.7 35.0 Disgust High 11.7 10.0 40.0 76.7 16.7 28.8 27.9 8.3 45.0 Low 71.7 18.3 13.3 10.0 Anger High 8.3 40.0 78.3 23.3 20.4 20.0 25.0 35.0 31.7 28.3 Low 16.7 11.7 31.7 53.3 High Fear 16.7 22.1 28.3 20.4 58.3 8.3 31.7 25.0 16.7 Low Adults Happiness 0.0 0.0 0.0 0.0 0.0 100.0 High Low 90.0 5.0 0.0 5.0 0.0 2.5 Sadness 0.0 35.0 8.8 0.0 High 0.0 90.0 0.0 0.0 15.0 3.8 Low 0.0 90.0 0.0 Disgust High 70.0 55.0 0.0 0.0 17.5 95.0 3.8 15.0 0.0 Low 0.0 0.0 95.0 0.0 High 2.5 0.0 0.0 Anger 10.0 0.0 0.0 Low 40.0 6.3 10.0 15.0 0.0 85.0 0.0 0.0 0.0 0.0 High Fear 0.0 0.0 Low 0.0 0.0 0.0 85.0 Note . “Correct” responses are in bold. “Incorrect” inclusions were significantly higher .10). p “Incorrect” inclusions were marginally higher than that facial expression’s base rate (.05 p than that facial expression’s base rate ( .01). 2 focuses on the six facial expressions that were “incorrectly” .03, and the high arousal disgust face in the angry category, t 19 4.99, .001. p included above the base rate for that face. For each of these faces, it shows the percentage of children who “correctly” selected a face on the target trial and then also “incorrectly” selected the same Malleability of the Interpretation of face on another trial. For example, 40 children selected the high Facial Expressions arousal sadness face as sad. Of these, 14 (35.0%) also selected this face as scared. On average, when these six faces were included in Next, we focused on the malleability of children’s interpreta- the “correct” category, they were also included in an “incorrect” tions of facial expressions. First, we looked at whether children included the same face in more than one emotion category. Table category on 48.1% of the trials. These results indicate that children Table 2 Percentage of Adults and Children Who Included Faces in the “Correct” Category and Another Category “Correct” Category Face included in “Incorrect” Sad trial Intensity Angry trial Scared trial category Happy trial Disgusted trial Children 35.0 (14/40) High Sadness 34.8 (16/46) Low Disgust High 75.0 (18/24) Low 70.3 (19/27) 39.1 (18/46) High Anger 34.1 (13/35) Low Fear Adults High 38.9 (7/18) Sadness High 54.5 (6/11) Disgust Note. Only those faces that were included in an “incorrect” category above the base rate for that face are included in this table. In parentheses is the number of times a face was selected for an “incorrect” category over the number of times it was selected for the “correct” category.
5 MALLEABILITY OF FACIAL EXPRESSIONS 5 1994). Instead, the emotion that facial expressions communicate is were interpreting these facial expressions anew on each trial: The influenced by their context. In the current study, that context was emotion that facial expressions communicate was influenced by the emotion category for which a person was looking. This result the emotion category for which the child was looking. supports Carroll and Russell’s (1996) suggestion that the informa- The one unexpected “incorrect” inclusion was the low arousal tion people read most easily from facial expressions is not a sad face in the scared category. Adults rated this sad face as having specific discrete emotion. Instead, the specific emotion attributed the lowest arousal (2.6/7) of the 10 faces. Fear is typically con- to a facial expression is malleable because people most easily read sidered to be a high arousal emotion (e.g., Russell, 1980). To levels of pleasure and arousal as well as physical information (e.g., further investigate this inclusion, we compared each age group’s gaze direction, whether the person is crying, talking, smiling, inclusions of this face in the scared category. There were 21 shouting, grimacing, etc.; Carroll & Russell, 1996; Aviezer et al., inclusions: 14 (46.7%) by 3-year-olds and seven (23.3%) by 2008). 4-year-olds. This difference was significant, independent groups The malleability of children’s interpretation of facial expres- t 8.69, p .001. Thus, the inclusion of the low arousal sad 59 sions was evident in the finding that at least one child included face in the scared category was due primarily to the 3-year-olds every face in each “incorrect” category, and that these inclusions broad fear category. were significantly above base rates for four faces and marginally Table 2 illustrates another point: Some categories were broader so for two others. About one third of children who selected the fear than others. The happy and sad categories included no nontarget face as fear, the anger face as anger, and the sad faces as sad also faces—at least not above the base rate. The remaining three selected these same faces for another emotion category; up to three categories each included two faces other than the target faces: quarters did so for disgust faces. Thus, even children who had Angry included both of the disgust faces, which is a common categorized facial expressions “correctly” also categorized these “error” for both children and adults. Scared included both sadness faces “incorrectly” as another emotion. faces. And disgusted included the high arousal anger face and the The majority of children and adults included the high arousal low arousal fear face. This varying breadth corresponds to the angry face in the angry category but excluded the low arousal one. emotion categories that children acquire early (happiness, sadness) This finding suggests that judgments about the anger facial ex- versus late (fear, disgust) (Widen & Russell, 2008a). (Anger is an pression may be influenced by whether participants (children or early acquired category, and so did not fit the pattern. The results adults) are presented with the high arousal (open-mouthed) anger of this trial are discussed further below.) face or the low arousal (closed-mouthed) version. This possibility Adults also showed some malleability in their interpretations of requires further investigation and should also alert researchers to facial expressions, at least on the disgusted and scared trials (see be aware of the specific faces they choose when investigating Table 2). They selected the high arousal sadness face as scared and people’s understanding of the anger face—using a low arousal the high arousal disgust face as angry. On average, on 46.7% of the anger face may result in the underestimation of children’s (and trials, when these two faces were included in the “correct” cate- adults’) understanding of anger. gory, they were also included in an “incorrect” category. One explanation for the faces that children “incorrectly” include in different emotion categories is that they base their judgments on Angry Trial the broad dimensions of valence and arousal (e.g., Bullock & Russell, 1984; Widen & Russell, 2003). The number of faces that As predicted, children included both of the target faces for each children admitted to each category varied with emotion. They category—except on the anger trial (see Table 1). When children included the most faces when asked to find the happy and angry were asked to find the angry expression(s), significantly more faces and the fewest faces when asked to find the sad people. In selected the high arousal anger expression than the low arousal addition, the number of faces that children included in each cate- t one, dependent measures 5.80, .001. p (59) gory decreased with age, especially for happy, scared, and angry. In addition, the angry trial was the only one on which the The current study provides some support for this view, but also frequency of “incorrect” inclusions of any of the faces was higher raises some questions especially about the development of chil- than the inclusions of a target face. Compared to the low arousal dren’s understanding of arousal. First, children did not “incor- anger face, significantly more children included the high arousal rectly” include faces of the opposite valence above base rate levels. t disgust face, .001, and the low arousal disgust p 5.37, (59) Instead, children “incorrectly” included faces of similar valence. face, t .001. p 5.55, (59) These faces also had similar levels of arousal to the target emotion Adults also showed this pattern (see Table 1): Significantly (except for the inclusion of the low arousal sad face in the scared more selected the high arousal anger expression than the low category). Children included both of the disgust faces in the angry arousal one, t p 4.82, .001. Compared to the low arousal (19) category, the high arousal anger face in the disgust category, the anger face, more adults included the high arousal disgust face but high arousal sadness face in the scared category, and the low this difference was not significant, t 1.55, .13. p (19) arousal fear face in the disgust category. Both of the faces adults “incorrectly” included were of the same valence and similar Conclusion arousal to the category in which they were included. Four of the children’s “incorrect” inclusions were also influ- In the current study, the emotion that facial expressions com- enced by perceptual similarity. The high arousal sad and fear faces municated was malleable: Children (and even adults) included the shared a number of muscle movements (both had raised inner same faces in different emotion categories. Facial expressions of brows, upper lip pulled laterally [nasolabial deepener], and lips basic-level emotions do not communicate a single discrete emotion parted) making the perceptual similarity of these two faces readily as assumed by the traditional theory of emotion (e.g., Ekman,
6 WIDEN AND NAAB 6 tions are even more likely to be included in more than one emotion evident. The perceptual similarity of the anger and disgust faces— category. though they share only incidental muscle movements—was al- The current study established that children’s interpretation of ready mentioned. It is not possible, in the current study, to tease facial expressions is malleable but did not investigate how this arousal (open vs. closed mouth) and perceptual similarity apart to malleability might interact with other areas of development. An- assess which is stronger for these four inclusions. other avenue for future research is the potential interaction of The inclusion of the low arousal sad face in the scared category children’s language, cognitive, or social development with the was not predicted by either of our hypothesis that related to malleability of their interpretation of facial expressions. Prior nontarget target inclusions. First, this was a low arousal face— research has shown links between language development and indeed, the lowest arousal face of the 10 faces presented—included emotion understanding (e.g., Pons, Lawson, Harris, de Rosnay, in a high arousal category. Second, the low arousal sad face did not 2003) and between theory of mind development and emotion share perceptual similarity with either of the fear faces. Thus, this understanding (e.g., Lagattuta & Wellman, 2001; Rieffe, Meerum inclusion was not predicted but neither was it unprecedented. In Terwogt, & Cowan, 2005). prior categorization studies, young preschoolers have included low The current study contributes to the view that young children’s arousal faces (sad, sleepy, serene) in high arousal categories understanding of emotion is different from adults’. While chil- (scared, surprised) and vice versa (Bullock & Russell, 1984; Bull- dren’s understanding of emotion is systematic and, at all levels, ́kely et al., 2011; Widen & Russell, ock & Russell, 1985; Sze describes the entire emotion domain, the youngest children have a 2008a). In each study, this effect decreased by four years. The valence-based understanding of emotion (feels good vs. feels bad; same pattern occurred in the current study: 3-year-olds made the Widen & Russell, 2003, 2008a). Somewhat older children divide majority of these inclusions; 4-year-olds were significantly less their “feels bad” category into two categories: a high arousal likely to do so. It is not clear why young preschoolers include low category labeled angry and a low arousal one labeled sad—but arousal faces in high arousal categories. This question provides is these categories remain broader than the adult categories of the an avenue for future research. same names. The process of differentiation continues, as children The inclusion of disgust faces in the anger category suggests acquire more categories and labels, but even at the end of pre- that for children, both of the disgust faces are better exemplars of school, children’s understanding of emotion does not yet corre- anger than is the prototypical low arousal anger face. Thus, the spond to our adult understanding of emotion (Widen & Russell, current study joins with prior studies (e.g., Bullock & Russell, 2003, 2008a). 1984; Camras & Allison, 1985; Gosselin, 1995; Gosselin & La- roque, 2000; Markham & Adams, 1992; Russell & Widen, 2002; Widen & Russell, 2002, 2003, 2004, 2008a, 2008b, 2010) in References finding that children interpret the prototypical disgust face as Aviezer, H., Hassin, R. R., Ryan, J., Grady, C., Susskind, J., Anderson, anger. On this trial, adults also included the high arousal disgust Bentin, S. (2008). Angry, disgusted, or afraid? Studies on the A.,... face more frequently than the low arousal anger face. Thus, with Psychological Science, 19, 724–732. malleability of emotion perception. development, children should become less likely to interpret the doi:10.1111/j.1467-9280.2008.02148.x low arousal disgust face as anger, but continue to do so for the high Bullock, M., & Russell, J. A. (1984). Preschool children’s interpretation of arousal disgust face at least when asked to look for anger expres- International Journal of Behavioral De- facial expressions of emotion. sions. velopment, 7, 193–214. doi:10.1016/S0163-6383(84)80255-6 The sample in the current study had limited ethnic diversity. Bullock, M., & Russell, J. A. (1985). Further evidence on preschooler’s Although there no reason to expect that other groups of children interpretations of facial expressions of emotion. International Journal of might be more or less malleable in their categorization of facial 15–38. doi:10.1016/S0163- Behavioral Development, 8, 6383(85)80013-8 expressions, this is an empirical question. One exception that has Camras, L. A., & Allison, K. (1985). Children’s understanding of emo- been demonstrated is due to neither race nor culture but to indi- tional facial expressions and verbal labels. Journal of Nonverbal Behav- vidual differences: Children who have been abused are more likely ior, 9, 84–94. doi:10.1007/BF00987140 to attribute anger to a variety of facial expressions (e.g., Pollack, Carroll, J. M., & Russell, J. A. (1996). Do facial expressions signal specific Cicchetti, Hornung, & Reed, 2000; Pollak, & Sinha, 2002). Journal of Per- emotions? Judging emotion from the face in context. All of the facial expressions used in the current study were 205–218. doi:10.1037/0022- sonality and Social Psychology, 70, posed by Caucasian women. This selection of stimuli was a con- 3522.214.171.124 scious decision to prevent any confounds of children’s categori- The Face of Man. Ekman, P. (1980). New York, NY: Garland Publishing, zation of facial expression with gender or race—both of these Inc. effects have been demonstrated (e.g., Tuminello & Davidson, in Ekman, P. (1994). Strong evidence for universals in facial expressions: A Psychological Bulletin, 115, 268– reply to Russell’s mistaken critique. press; Widen & Russell, 2002). Future research might investigate 287. doi:10.1037/0033-2909.115.2.268 the effects of poser gender or different ethnicities on the mallea- Palo Alto, Pictures of facial affect. Ekman, P., & Friesen, W. V. (1976). bility of children’s (and adults’) emotion categories. CA: Consulting Psychologists Press. The facial expressions used in this study were prototypical facial Gosselin, P. (1995). The development of the recognition of emotional expressions of basic-level emotions and therefore negative and Canadian Journal of Behavioral Science facial expressions in children. high arousal emotions were overrepresented. Future studies might Revue, 27, 107–119. doi:10.1037/008-400X.27.1.107 include more positive and low arousal facial expressions to inves- Gosselin, P., & Laroque, C. (2000). Facial morphology and children’s tigate whether such faces have different effects on the malleability categorization of facial expressions of emotions: A comparison between of children’s and adults’ interpretation of them. It seems likely that Journal of Genetic Psychology, 161, 346– Asian and Caucasian faces. facial expressions that are not assumed to represent specific emo- 358. doi:10.1080/00221320009596717
7 MALLEABILITY OF FACIAL EXPRESSIONS 7 Tuminello, E. R. & Davidson, D. (in press). What the face and body reveal: Lagattuta, K. H., & Wellman, H. M. (2001). Thinking about the past: Early In-group emotion effects and stereotyping of emotion in African Amer- knowledge about links between prior experience, thinking, and emotion. ican and European American children. Journal of Experimental Child 82–102. doi:10.1111/1467-8624.00267 Child Development, 72, . Advance online publication. doi:10.1016/j.jecp.2011 Psychology Markham, R., & Adams, K. (1992). The effect of type of task on children’s .02.016 Journal of Nonverbal Behavior, 16, identification of facial expressions. Widen, S. C., & Russell, J. A. (2002). Gender and Preschoolers’ Perception 21–39. doi:10.1007/BF00986877 of Emotion. 248–262. doi:10.1353/ Merrill-Palmer Quarterly, 48, Pollack, S. D., Cicchetti, D., Hornung, K., & Reed, A. (2000). Recognizing mpq.2002.0013 Emotion in Faces: Developmental Effects of Child Abuse and Neglect. Widen, S. C., & Russell, J. A. (2003). A closer look at preschoolers’ freely Developmental Psychology, 36, 679–688. doi:10.1037/0012- produced labels for facial expressions. Developmental Psychology, 39, 16126.96.36.1999 114–128. doi:10.1037/0012-16188.8.131.52 Pollak, S. D., & Sinha, P. (2002). Effects of early experience on children’s Widen, S. C., & Russell, J. A. (2004). The relative power of an emotion’s recognition of facial displays of emotion. Developmental Psychology, facial expression, label, and behavioral consequence to evoke preschool- 38, 784–791. doi:10.1037/0012-16184.108.40.2064 Cognitive Development, 19, ers’ knowledge of its cause. 111–125. Pons, F., Lawson, J., Harris, P. L., & de Rosnay, M. (2003). Individual doi:10.1016/j.cogdev.2003.11.004 differences in children’s emotion understanding: Effects of age and Widen, S. C., & Russell, J. A. (2008a). Children acquire emotion catego- language. Scandinavian Journal of Psychology, 44, 347–353. doi: Cognitive Development, 23, 291–312. doi:10.1016/ ries gradually. 10.1111/1467-9450.00354 j.cogdev.2008.01.002 Rieffe, C., Meerum Terwogt, M., & Cowan, R. (2005). Children’s Under- Widen, S. C., & Russell, J. A. (2008b). Children’s and adults’ understand- standing of Mental States as Causes of Emotions. Infant and Child ing of the “Disgust Face.” 1513–1541. Cognition and Emotion, 22, Development, 14, 259–272. doi:10.1002/icd.391 doi:10.1080/02699930801906744 Russell, J. A. (1980). A circumplex model of affect. Journal of Personality Widen, S. C., & Russell, J. A. (2010). Children’s scripts for social emo- and Social Psychology, 39, 795–804. doi:10.1037/h0077714 tions: Causes and consequences are more central than are facial expres- Russell, J. A., & Widen, S. C. (2002b). A Label Superiority Effect in sions. 565–581. doi: British Journal of Developmental Psychology, 28, children’s categorization of facial expressions. Social Development, 11, 10.1348/026151009X457550d 30–52. doi:10.1111/1467-9507.00185 ́kely, E., Tiemeier, H., Arends, L. R., Jaddoe, V. W. V., Hofman, A., Sze Received December 16, 2010 Verhulst, F. C., & Herba, C. M. (2011). Recognition of facial expres- Revision received August 22, 2011 425–435. doi:10.1037/ Emotion, 11, sions of emotions by 3-year-olds. a0022587 Accepted September 15, 2011
D E A R T M E N T O F D E F E N S E P N A L O F W A R M A W U A L J U N E 2 0 1 5 O F F I C E O F G E N ER A L C O U N S E L D P A R T M E N T E O F D E F E N S EMore info »
Customary International Humanitarian Law Henckaerts Volume I: Rules and In 1996, the International Committee of the Red Cross, alongside a range of Doswald-Beck Customary International renowned expert...More info »
IMPLEMENTATION HANDBOOK FOR THE CONVENTION ON THE RIGHTS OF THE CHILD FULLY REVISED THIRD EDITION IMPLEMENTATION HANDBOOK IMPLEMENTATION HANDBOOK FOR THE CONVENTION ON THE FOR THE CONVENTION ON THE RI...More info »
MANAGING THE RISKS OF EXTREME EVENTS AND DISASTERS TO ADVANCE CLIMATE CHANGE ADAPTATION SPECIAL REPORT OF THE INTERGOVERNMENTAL PANEL ON CLIMATE CHANGEMore info »
The Health Consequences of Smoking—50 Years of Progress A Report of the Surgeon General U.S. Department of Health and Human ServicesMore info »