Facial Expressions – Early Work

Across several industrialized and isolated cultures, Ekman has found people universally recognize the happiness facial expression. Similar to findings on fear (Ekman et al., 1969, click here for review), the majority of participants correctly identified happiness, including the Fore and Borneo participants (see Table 7).

In his third study (for a review, go here), Ekman and colleagues (1987) tested facial expression identification across 10 countries.  Again, happiness was the only positive emotion utilized in this study.  As show in Table 10, in the single-judgment task, participants in all countries correctly identified happiness when shown the happy photo.  This suggests happiness may be universally recognized, even in the isolated Minangkabu tribe (located in Sumatra). But remember this study included only one positive emotion – so it would probably be easy to label a photo displaying a smile as happiness. The multiple-judgment task also suggested happiness was universally identified.  In this task, participants rated the intensity of the six emotions shown in the happy photo.  For the happy photo, participants rated this photo as most intensely happiness.

Some other work has confirmed universality in happiness facial expressions.  In one study (Matsumoto, 1992), participants saw 48 photos of six emotional expressions (anger, disgust, fear, happiness, sadness, and surprise).  Each emotion was displayed in 8 different photos.  Participants viewed all 48 photos one at a time.  While viewing each photo, participants picked an emotion label from the following seven emotions: anger, contempt, disgust, fear, happiness, sadness, and surprise. 97.62% of American participants and 98.30% of Japanese participants correctly labeled the happiness expression as happiness across the 8 happy photos.  Further analysis revealed identification did not differ across cultures.  A similar study by Matsumoto and Ekman (1989) found that 97.97% of Americans and 97.59% of Japanese correctly identified happiness across eight photos.

In the fear chapter, we discussed a study by Crivelli et al. (2016), which can be reviewed  here.  In this study, participants were randomly assigned to one of five emotion label conditions (happy, sad, anger, fear, disgust). From six different photos of facial expressions, participants were asked to pick the photos that displayed their assigned emotion from six photos. Thus, in the happiness condition, participants were asked to point to the photo that displayed the happy facial expression. In the happy condition, a significantly greater percentage of Spaniards (100%) than Trobrianders (58%) correctly selected the smiling expression. The Trobrianders also tended to pick the neutral face (23%), the sad pout (8%), the fear gasp (8%), and the angry scowl (4%) as representing happiness. These findings might suggest that the happiness expression was universally identified.  But note that happiness is again the only positively valenced emotion in this study – thus again making the correct answer easier to identify.

In study 2 (Crivelli et al., 2016), participants were children and adolescents recruited from Mwani.  Participants were asked to match emotion words to emotion facial expressions displayed in the photo or video clip. All participants completed their tasks for all five emotions – happy, sadness, anger, fear, and disgust (a within-subjects variable). The facial expressions displayed were the same as in Study 1. The photo facial expressions were still photos taken from the video clip. All six facial expressions, including neutral, were shown at the same time for the photos and the video clips.

When instructed to select the happy facial expressions, 58% in the static and 53% in the dynamic condition correctly selected the smiling face. (There were no significant differences in the proportion of participants who selected the correct answer in the dynamic or static condition.) In general, participants also selected the neutral expression.

Recall that Lisa Barrett and her colleagues (Gendron et al., 2014b) asked Boston and Himba participants to sort facial expressions into similar piles based on similarity.  They found both groups sorted smiles into one pile, suggesting joy was universally recognized. For a review of this study, go here.