Running Head: POSED AND SPONTANEOUS FACIAL EXPRESSION
A Comparison Between Posed and Spontaneous Facial Expressions: Expression intensity and Experienced Intensity Loek Peelen Social Psychology Xia Fang Tuesday, July 7, 2016 University of Amsterdam
Abstract
The difference between posed and spontaneous facial expression was examined with 100 participants. The aspects that were taken into consideration were based on the perception of facial expression, namely expression intensity and experienced intensity. In the first experiment 50 participants rated their own posed facial expression intensity for the emotions anger and disgust. In the second experiment 50 participants rated their own spontaneous facial expression intensity for the emotions anger and disgust. It turned out that there was no significant difference between the expression intensity for posed and spontaneous facial expression. Both the participants in experiment 1 and 2 also rated the experienced intensity for intended and non-intended emotion. It turned out that for the intended emotions, anger and disgust, there was no significant difference between posed and spontaneous facial expression. For the non-intended emotions there was a significant difference between posed and spontaneous facial expression, spontaneous facial expression elicits more experienced intensity for non-intended emotions.
Introduction
Human communicate their emotions to others, through verbal and non-verbal forms (Scherer, Mortillaro & Mehu, 2013). Among humans a form of non-verbal communication is, conveying emotion through facial expressions. This form of non-verbal communication is the most common way of emotion expression (Wang, Liu, Lv, Lv, Wu, Peng, Chen & Wang, 2010). This is the case because by moving certain combinations of facial muscles, different emotions can be displayed. So humans can communicate by using emotions, procuced by facial expressions, this means that there is a production and perception process of emotions in humans (Vigil, 2009). To study the production and perception of facial expression of emotion, two different kinds of manners can be used to elicit facial expressions. Namely posed facial expression and spontaneous facial expression. To elicit posed facial expression, typically the experimenter instructs actors to pose designated emotional expressions for a photographer, and uses the photographs as stimuli in emotion perception tasks (e.g., Ekman, 1973; Ekman & Friesen, 1975; Leathers & Emigh, 1980; Woodall, Burgoon, & Markel, 1980). Many studies have demonstrated that posed facial expressions of emotion are quite unambiguous and easily identified (Motley & Camden, 1988). To elicit spontaneous facial expression, participants (expressers) are not told to deliberately show a facial expression of emotion. On the contrary, the spontaneous facial expression is usually a reflection of how a person feels. This is because participants are instructed to tell a story that might elicit a emotion that will express in the face.
So there are different ways to elicit a facial expression and these different ways to obtain the facial response can differ from each other. Hunt (1941) argued that posed facial cues constitute a socially learned code that is unrelated to spontaneously occurring cues. Ekman (1972) suggested that the eliciting circumstances for posed and spontaneous cues are different. If there is a difference between posed and spontaneous stimuli there is a possibility
that there is a difference between the response, namely the production and perception of facial expression. Because the posed facial expression arises from a reenacted emotion while the spontaneous facial expression arises from a real emotion, the posed facial expression could be seen as a more fake emotion and the spontaneous facial expression could be seen as a more real emotion (Tcherkassof, Bollon, Dubois, Pansu & Adam, 2007). The distinction between posed and spontaneous facial expression is rarely addressed by communication research, however, where instead it has become convenient for studies of nonverbal behavior to ignore the fact that there are two different ways of elicit facial expression. It is convenient to use posed stimulu and not spontaneous stimuli because of a couple of reasons. One of the major problems is gathering natural facial expressions displaying various emotions, that is, gathering realistic material to be judged by observers. Recording naturally occurring emotional
experiences is empirically not easy. Moreover, inducing strong emotional states for experimental purposes gives rise to ethical questions (cf. Philippot, 1993). Therefore the scientific study of emotion mainly rests on facial expressions posed or simulated by actors, despite the lack of spontaneity and naturalness of this material (Tcherkassof et al., 2007). Thus, research on the production and perception of facial expression with only the use of posed stimuli is taken, implicitly at least, to generalize to their spontaneous counterparts. However to ignore the distinction is appropriate only to the extent that the two kinds of facial expressions are similar (Motley & Camden, 1988).
It is questionable, whether results revealed by research on posed facial expressions can be generalized to spontaneous facial expressions. Relating to this question, the main aim of the present research is to explore whether there is a difference between posed and
spontaneous facial expressions.
Specifically in the present research, the perception of facial expression will be examined. Two aspects that lie in the perception area will be compared between posed and
spontaneous facial expressions, (a) the intensity of the facial expressions and (b) the intensity of the underlying experiences. First the posed facial expressions may be different from the spontaneous facial expressions in terms of expression intensity. This difference might be explained by the social situations in where facial expression is elicited. Ekman and Friesen (1969) have articulated a set of norms that supposedly govern, how, when, and where specific emotions should be expressed within a particular culture. Three of these four -display rules- involve the modulation of the intensity of a facial expression (i.e., exaggeration, attenuation, and neutralization). Similarly, Buck (1984, 1988) argued that a substantial part of the
socialization process is geared towards the habitual inhibition and control of feelings and facial expression. Posed facial expressions are not exposed to any display rules. The posed expressions are the purest form of the facial expression of an emotion (Tcherkassof et al., 2007) . In other words, these posed expressions are usually exaggerated and thereby may differ in intensity from spontaneous expressions (Wang et al., 2010). Consequently, in the present research we expected that intensities of posed facial expressions would be higher than intensities of spontaneous facial expressions.
The second comparison we are drawing between posed and spontaneous facial expressions in the present research pertains to the underlying experiences. Specifically, the emotions in posed conditions are artificial whereas the emotions in spontaneous conditions are real emotions, felt by expressers (Motley & Camden, 1988). In posed conditions, expressers mainly try to display the correct facial expressions without the requirements of experiences, whereas in spontaneous conditions, expressers mainly try to experience specific emotions without the requirements of the external expressions (Motley & Camden, 1988). Thus, in the present research we expected that the experienced intensity of intended emotion is higher when a spontaneous facial expression is elicited than when a posed facial expression is elicited.
In addition to the intended emotion, expressers may also experience other
non-intended emotions in spontaneous condition (i.e., real-world) conditions. For example, people may feel alternative emotions such as anger, when telling a sad story. Conversely, people are less likely to experience non-intended emotions when instructed to pose a facial expression of a specific emotion (Tcherkassof et al., 2007). Therefore, in the present research, we expected that the experienced intensity of non-intended emotion is higher when a spontaneous facial expression is elicited than when a posed facial expression is elicited.
Methods
Overview
Three emotion terms were adopted in the present study: Neutral, Anger and Disgust. Participants in the posed condition were instructed to display the three facial expressions of emotion, whereas participants in the spontaneous condition were instructed to tell three personal experiences which are related to the three emotions. After that the participants judge their own facial expression on expressive intensity and clarity and they judge the level of experienced intensity for intended and non-intended emotions.
Materials
To measure the intensity of the facial expression of the emotion condition a
questionnaire was used with the question, (How strong is the intensity of the expression?) on a scale of 0 to 10 (0=not at all, 10 = extremely). Also the clarity of the emotion condition was measured, with the question (How clearly does the expression signal the target emotion?) on a scale of 0 to 10 (0=not clearly at all, 10 = very clearly).
To measure the experienced intensity of the intended and the non-intended emotions a questionnaire was used. The extent to which the five basic emotions were experienced while posing or telling a story was rated with the question (How strongly did you experience anger,
disgust/gross, fear, sad, happy during the pose?) or the question (How strongly did you experience anger, disgust/gross, fear, sad, happy during the story?). Experienced intensity was rated on scales of 0 to 10 (0=not at all, 10 = extremely). Also the difficulty of posing the intended emotion was measured with the question, (How difficult you feel to pose the target emotion?) on a scale of 0 to 10 (0=very easy, 10 = very difficult). The comfortability of sharing a story was measured with the question, (How comfortable you feel to tell a story with the target emotion?) on a scale of 0 to 10 (0=very easy, 10 = very difficult).
Procedure
Posed condition. Participants in the posed condition were asked to reenact the respective emotional expression such that “their friends would be able to understand easily how they feel.” For each emotion term that participants need to reenact, participants received the emotion label, the definition of the emotion term (Ekman & Cordaro, 2011), and an example of the emotion term (Matsumoto, 2003), (see Appendix A). Participants tested various expressions of the same emotion with the camera. Once they were satisfied with their pose, they pressed a button that rang a bell, and the experimenter provided them with the next label. Participants were filmed continuously throughout the session, A video camera was used in both posed and spontaneous conditions to record participants’ facial expressions of
emotion. Experimenters kept out of view of the participants during the posing phases and did not provide feedback or recommendations of any kind. After each pose, participants answer the question about the experienced intensity of the intended and non-intended emotion and also answer the question about the difficulty of the pose.
Using this protocol, each participant posed 3 states, neutral was always the first one to be posed, helping participants to get used to the procedure. The order of anger and disgust was counterbalanced across participants.
After they had posed the emotion condition participants looked at their video in the same order as posing and chose the frame which they considered their clearest expression of the respective intended emotion. Based on their choice, they rated the clarity and the intensity of the expression in the specified frame.
Spontaneous condition. Participants in the spontaneous condition were instructed to recall a past life event involving the respective intended emotion (i.e., neutral, anger, disgust), and recounted their experience in detail to a robot (participants were led to believe that they were going to help to develop an emotional robot who can understand human’s emotions in the future).
After each emotion induction, participants answer the question about the experienced intensity of the intended and non-intended emotion and also answer the question about the how comfortable they were sharing the story. In addition, participants were filmed throughout the entire session.
Using this protocol, each participant recounted three stories, neutral was always the first one to be induced, helping participants to get used to the procedure. The order of anger and disgust was counterbalanced across participants. After the induction session, participants looked at their own video and completed the same questionnaires as those in posed condition.
Results Manipulation checks
Participants rated the clarity of their facial expressions in both posed (M = 6.84, SD = 1.84) and spontaneous (M = 6.43, SD = 1.96) conditions, and no difference was found
between these two conditions, F(1,98) = 1.82, p = .181, ŋp2 = .018. In addition, participants in the posed conditions rated the difficulty of posing expressions of emotion, while participants in the spontaneous conditions rated the comfortability of story telling. For the posed
conditions, the difficulty of posing anger expressions (M = 5.22, SD = 2.17) was not different from that of posing disgust expressions (M = 4.80, SD = 2.11), t(49) = 1.33, p = .19. For the spontaneous condition, the comfortability of telling angry stories (M = 4.98, SD = 2.16) was not different from that of telling disgusting stories (M = 5.30, SD = 2.32), t(49) = -1.68, p = .10.
Main analysis
We were interested in to what extent the posed facial expressions would be different from the spontaneous facial expressions. The two types of facial expressions were compared in terms of expression intensity (relating to Hypothesis 1) and emotion intensity (relating to Hypotheses 2 and 3).
To test Hypothesis 1 that the intensities of posed facial expressions would be higher than the intensities of spontaneous facial expressions, we examined the expression intensity ratings in a 2 (Condition: Posed, Spontaneous) × 2 (Emotion: Anger, Disgust)
mixed-design ANOVA, with participant’s expression intensity ratings being the dependent variable. Condition was a between-subjects variable and Emotion was a within-subjects variable. All effects can be found in Table 1.
Table 1
Condition × Emotion Mixed-Design Analysis of Variance for the Expression Intensity Ratings
Effects Expression Intensity df F p Condition (C) (1, 98) 0.51 ns .005 Emotion(E) (1, 98) 19.11 < .001 .163 C × E (1, 98) 4.78 .031 .046
Note. ns represents a non-significant effect (p > .05)
Regarding Hypothesis 1, the main effect of Condition was not significant, F(1, 98) = .51, p = .48, ŋp2 = .005. The expression intensities were not different between posed (M = 6.43, SD = 1.79) and spontaneous (M = 6.19, SD = 2.05) conditions. The interaction of Condition and Emotion was significant, F(1,98) = 4.78, p = .031, ŋp2 = .046. A simple effect analysis was followed. For anger, the intensities of posed expressions (M = 6.22 ; SD = 1.89), were marginally higher than those of spontanneous expressions (M = 5.56, SD = 2.02), F(1,98) = 2.85, p = .095, ŋp2 = 028; for disgust, no differences were found between the intensities of posed expressions (M = 6.64, SD = 1.70) and those of spontaneous expressions (M = 6.82, SD = 2.09), F(1,98) = .224, p = .64, ŋp2 = .002. It suggests that for expressers themselves, the expression intensities are different between spontaneous and posed expressions of anger/disgust
Both Hypotheses 2 and 3 were related to the underlying experiences of expressers, with Hypothesis 2 predicting that intended emotions would be experienced more intensively in the spontaneous conditions than in the posed conditions, while Hypothesis 3 predicting that non-intended emotions would also be experienced more intensively in the spontaneous
conditions than in the posed conditions. We examined the experience intensity ratings of intended and non-intended emotions in a 2 (Condition: Posed, Spontaneous) × 2 (Emotion: Anger, Disgust) × 2 (Rating Type: Intended, Non-intended) mixed-design ANOVA, with
participant’s experience intensity ratings being the dependent variable. Condition was a between-subjects variable, and Emotion and Rating Type were within-subjects variables. The results pertaining to the hypotheses are presented below. A complete overview of effects can be found in Table 2.
Table 2
Condition × Emotion × Rating Type Mixed-Design Analysis of Variance for the Experience Intensity Ratings Effects Experience Intensity df F p Condition (C) (1, 98) 4.16 .044 .041 Emotion (E) (1, 98) 533.08 < .001 .845 Rating Type (R) (1, 98) 0.30 ns .003 C ×E (1, 98) 0.30 ns .003 C × R (1, 98) 0.07 ns .001 E × R (1, 98) 21.69 < .001 .181 C × E × R (1, 98) 0.06 ns .001
Note. ns represents a non-significant effect (p > .05)
Regarding Hypothese 2 and 3, the main effect of Condition was significant, F(1,98) = 4.16, p = 0.44, ŋp2 = .041, with experience intensity ratings in the spontaneous conditions (M = 4.07, SD = 1.80) being higher than those in the posed conditions (M = 3.54, SD = 1.71). No interaction effect involving Condition was significant (see Table 2). It suggests that
individuals experience more intensively when displaying spontaneous expressions of emotion compared to displaying posed expressions of emotion.
Contrary to Hypothesis 2, for intended emotions, no differences were found between the posed (M = 5.85, SD = 2.24 ) and spontaneous (M = 6.27, SD = 2.26 ) conditions, F(1,98) = 1.10, p = .297, ŋp2 = .011. It suggests individuals who are asked to pose expressions of
emotion feel the intended emotions to a similar extent to individuals who display spontaneous expressions of emotion.
Consistent with Hypothesis 3, for non-intended emotions, there was a main effect of Condition, F(1,98) = 8.15, p < .01, ŋp2 = .077, with experience intensity ratings in the
spontaneous conditions (M = 1.87, SD = 1.34) being higher than those in the posed conditions (M = 1.23, SD = 1.17). This shows that individuals who display spontaneous expressions of emotion feel the non-intended emotions to a larger degree compared with individuals who are asked to pose expressions of emotion.
Discussion
In this research we examined in what aspects and to which degree that the posed facial expressions of emotion were different from the spontaneous facial expressions of emotion. Two aspects were compared: Expression intensity and experience intensity. The expression intensity was related to the outward appearances of expressions while the experience intensity was associated with the underlying experiences of expressions. The results show that the posed and spontaneous facial expressions differed from each other, but also shared
similarities. Specifically, the expression intensities of posed expressions were not different from those of spontaneous expressions; however, the experience intensities of spontaneous expressions were higher than those of posed expressions, expressions were also different from those of spontaneous expression.
Hypothesis 1 predicted that the expression intensities would be higher for posed facial expression than for spontaneous facial expression. However this prediction was not supported in the research, the expression intensities of posed facial expression were comparable to those of spontaneous facial expressions. This contradiction to our expectations may be explained by the display rules, which refer to a set of norms that supposedly govern, how, when, and where specific emotions should be expressed within a particular culture. The posed facial
expressions are not regulated by the display rules. They are usually exaggerated and thereby may be more intense (Wang et al., 2010). To the opposite, the spontaneous expressions of emotion are typically regulated by the display rules. In the current study, the participants in the spontaneous conditions were asked to tell a story to the robot Eva. Participants were led to believe that Eva is a robot who is developed to communicate with human and understand humans’ emotions. We therefore expected that participants would use similar interaction strategies when talking with Eva. Two reasons may account for the result that there were no differences between posed and spontaneous conditions. First, a robot is different from a real person, so the participants were the only human beings in the test area. Consequently, with no other human in the room the display rules didn’t apply to this situation, and so there were no differences between posed and spontaneous facial expression intensity. This means that for future research it is of great importance to really imitate a social interaction more precise in where spontaneous facial expression is elicited. Second, the participants rated the expression intensity themselves, they are expected to give high value for their emotional expressions, and they are biased in judging their own expressions.
Experience intensity was, in line with the predications, different between posed and spontaneous facial expression. However the dissimilarity was mainly found within non-intended emotions. We thus should interpret, the results with caution. Hypothesis 2 stated, that the intended experience intensities would be higher in the spontaneous conditions than in the posed conditions. It turns out that the experience intensities were the same for both types of facial expression. So it is possible that the statement, that in posed conditions, expressers mainly try to display the correct facial expressions without the requirements of experiences, whereas in spontaneous conditions, expressers mainly try to experience specific emotions without the requirements of the external expressions, is slightly incorrect. It is possible that posed facial expression, only for intended emotions, also is accompanied by an underlying
feeling of the emotion that is expressed by the facial muscles. An explanation can may be found in the Facial Feedback Hypothesis, this hypothesis holds that facial expressions play a causal role in the experience of emotions; emotional experience of the expresser is in part determined by his or her own facial expressions (Ekman, Levenson & Friesen, 1983).
Hypothesis 3 predicted that, the experienced intensities of non-intended emotion would be higher in the spontaneous conditions compared to the posed conditions, which was supported by the current research. So in line with the theory, spontaneous facial expression triggers experiences of non-intended emotions, because spontaneous facial expressions are real emotions (Motley & Camden, 1988) and an experience of a real emotion brings other emotions along. However it is possible that the Facial Feedback Hypothesis can also apply for the spontaneous facial expression. This suggests that, because an emotion was expressed in the face, this emotion was felt by the expresser, in this case the spontaneous facial
expression. But the Facial Feedback Hypothesis is not sufficient to explain why expressers experience higher intensities of non-intended emotions. This is due the fact that expressers in the spontaneous condition do not show facial expressions of the non-intended emotions, and thus receive no feedback from their facial expressions on their experienced intensities.
Limitations and Future research
In this study we looked mainly at how humans see their own facial expressions and how they experience their own facial expressions. It turns out that the differences between the posed and spontaneous facial expressions are limited, and especially for how people see their own facial expressions. The overarching purpose of the study was to examine if the trend of using only posed facial expression in emotion research is righteous, before this question could be answered correctly there are some aspects that should be taken into consideration for future research. This research proves that posed and spontaneous differ from each other, but these
differences are not visible in the face, they are only underlying experiences. To really get an unambiguous answer it is of interest to see how people perceive posed and spontaneous facial expressions other than their own. So in the future research judges should estimate the
expression intensities of the posed and spontaneous facial expressions. Further more, in this research we looked at 1 aspect of facial expressions, namely the expression intensity. It is valuable to look at different aspects of facial expression in future research. It may be possible that there is a difference in facial muscle activity between posed and spontaneous facial expression (Berenbaum & Roter, 1992). With Facial Action Coding System (FACS) muscle activity can be measured and so in future research it can be examined if there is a difference between muscle activity in posed and spontaneous facial expression (Mavadat, Mahoor, Zhao, & Ji, 2015). Only when these aspects are explored it might be concluded if posed facial expression can be generalized to spontaneous facial expressions in emotion research.
Conclusion
For now it can be concluded that the differences between posed and spontaneous facial expression are not as profound as expected. The perception of the facial expressions is the same in terms of expression intensity and experienced intensity this for intended emotions. Only when non-intended emotions are included, a difference is found between posed and spontaneous facial expression. Spontaneous facial expressions will elicit more versatile emotions.
References
Adam, J., Bollon, T., Dubois, M., Pansu, P., & Tcherkassof, A. (2007). Facial expressionof emotions: A methodological contribution to the study of spontaneous and dynamic emotional faces. European Journal of Social Psychology, 37, 1325-1345.
Barr, C. L., & Kleck, R. E. (1995). Self-other perception of the intensity of facial expressions of emotion: Do we know what we show? Journal of Personality and Social
Psychology, 4, 608-618.
Buck, R. (1984). The communication of emotion. New York: Guilford Press.
Buck, R. (1988). The perception of facial expression: Individual regulation and social
coordination. In T. R. Alley (Ed.), Social and applied aspects of perceiving faces (pp. 141-165). Hillsdale, NJ: Erlbaum.
Camden, C, T., & Motley, M, T. (1988). Western of Journal of Speech Communication, 52, 1-22.
Campos, J. J., & Barrett, K. C. (1984). Toward a new understanding of emotions and their development. In C. E. Izard, J. Kagan, & R. B. Zajonc (Eds.), Emotions, cognition, and behavior (pp. 229-264). Cambridge, England: Cambridge University Press. Chen, F., Liu, Z., Lv, S., Lv, Y., Peng, P., Wang, S., Wang, X., & Wu, G. (2010). A natural
visible and infrared facial expression database for expression recognition and emotion inference. IEEE Transactions on Multimedia, 7, 682-691.
DeFrank, R, S., Hall, J, A., Rosenthal, R., & Zuckerman, M. (1976). Encoding and decoding of spontaneous and posed facial expressions. Journal of Personality and Social Psychology, 5, 966-977.
Ekman, P. (1972). Universal and cultural differences in facial expressions of emotion. Nebraska Symposium on Motivation, 20.
Ekman, P. (1973). Cross-cultural studies on facial expression. In P. Ekman (Ed.), Darwin and facial expression: A century of research in review (pp. 169 222). New York:
Academic Press.
Ekman, P., & Friesen, W. V. (1969b). The repertoire of nonverbal behavior: Categories, origins, usage, and coding. Semiotica, 1, 49-98.
Ekman, P., & Friesen, W. V. (1975). Unmasking the face: A guide to recognizing emotions from facial cues. Englewood Cliffs, NJ: Prentice-Hall.
Ekman, P., Friesen, W. V., & Levenson, R. W. (Autonomic nervous system activity distinguishes among emotions. Science, 221, 1208-1210.
Hunt, W. (1941). A recent development in the field of emotion. Psychological Bulletin, 38, 249-276.
Ji, Q., Mahoor, M. H., Mavadati, S. M., & Zhoa, Y. (2015). Measuring the intensity of spontaneous facial action units with dynamic Bayesian network. Pattern Recognition, 48, 3417-3427.
Leathers, D. G., & Emigh, T. H. (1980). Decoding facial expressions: A new test with decoding norms. Quarterly Journal of Speech, 66, 418-436.
Malatesta, C. Z. (1985). Developmental course of emotion expression in the human infant. In G. Zivin (Ed.), The development of expressive behavior: Biology-environment
interactions (pp. 183-213). New York: Academic Press.
Mehu, M., Mortillaro, M., & Scherer, K. R. (2013). Understanding the mechanism underlying the production of facial expression of emotion: A componential perspective. Emotion Review, 1, 47-53.
Poyatos, F. (1983), New perspectives in nonverbal communication: Studies in cultural anthropology, social psychology, linguistics, literature, and semiotics. Oxford: Pergamon Press.
Saarna, C. (1979). Children’s understanding of display rules for expressive behavior. Developmental Psychology, 15, 424-429.
Woodall, W. G., Burgoon, J K., & Markel, N. N. (1980). The effects of facial-head cue combinations on interpersonal evaluations. Communication Quarterly, 28, 47-55.
Appendix
Emotion terms posed and spontaneous condition (in Dutch) Neutral: betekent kalm en geen emotie werd gevoeld.
Een voorbeeld, een persoon vertelt over de routine van zijn normale dag.
Boosheid: betekent een gevoel van misgenoegen ontstaan door letsel, mishandeling, verzet en uit zichzelf vaak in een velangen om terug te vechten tgen het object wat voor deze boosheid zorgt.
Een voorbeeld, een persoon wordt door iemand anders beledigd. Walging: betekent een misselijkmakende tegenzin, of afkeer.
Een voorbeeld, een persoon raakt per ongeluk een hondendrol aan.
Timeframes posed and spontaneous condition, neutral, anger and disgust
For each participant in the spontaneous condition, disgust is about physical disgust or moral disgust.
Posed condition Neutral
Posed condition Disgust
Spontaneous condition Neutral
