“Am I happy or sad here?”: A study on human recognition of emotion via facial expression

     In 1980, Seyfarth, Cheney, & Marler published their findings on vervet monkeys alarms calls signaling the presence of different predators. Three distinctive calls were emitted for three different predators which allowed other vervet monkeys in the group to flee to the form of safety best fit against the predators abilities. For example one of the vervet’s predators is an eagle which hunts by sighting prey as it flies over trees. If a vervet spots an eagle it would make the eagle alarm signal and the rest of the group would be signaled to move into the safety of the leave coverage to prevent the eagle from seeing and preying upon them. This study paved the way for research on communication among animals focusing on both verbal (spoken word or grunts) and nonverbal (body posture and facial expression) expressions. One such study took place in 2018 by Lo, Li, Lee, & Yeung, who investigated human ability to differentiate emotions based on facial expressions.

     Lo, Li, Lee, & Yeung were looking to see if two emotions displayed through human facial expression would be deemed different if the two emotions presented differed in valence and arousal. Valence of an emotion was defined as “an indication of pleasantness/unpleasantness of an emotional signal” (Lo, Li, Lee, & Yeung 2018), or whether or not the emotion hints at positivity/welcomeness or negative/avoidance. Arousal was defined as a “matter of vigourness of emotion and their expression” (Lo, Li, Lee, & Yeung 2018), or how engaged the person displaying the emotion appeared. For example happiness would have a positive (+) valence because the person to whom the expression is directed would be more likely to approach the person making the expression, and happiness would also have a positive (+) arousal because the person making the facial expression would be seen as fully engaged in the display and more likely to react positively to someone engaging with them. On the other hand sadness would have a (-) valence because the person to whom the emotion is directed would be prompted to avoid the person making the expression, and sadness would also have a negative (-) arousal because the person making the expression would be seen as unlikely to engage with someone. The emotions were paired by Lo et al, based on their complete contrast in valence and arousal, or their partial contrast where the emotions differed in either their valence or their arousal. For example a complete contrast pair would be happiness (+valence, +arousal) and sadness (-valence, -arousal). Where as a partial contrast would be happiness (+valence, +arousal) and anger (-valence + arousal).   

     Lo, Li, Lee, & Yeung hypothesized that when presented with two pictures of a face each displaying a different emotion the participant would respond quicker and more accurately to facial expression displaying completely contrasting emotions rather than partially contrasting emotions. They also hypothesized that discrimination of valence contrast pairs would result in quicker and more accurate matching than those of arousal contrast pairs.

     In the experiment 30 undergraduate students used a computer to rated two emotions as identical, as the pictures were presented to them. The accuracy and speed of the participant’s response was recorded to see how the emotions differed in their discriminability Unfortunately the data collected did not provide evidence to support the idea that a person will determine an emotional pairing as differning completely or partially.

     It was believed that this finding was attributed to human’s ability to differentiate between all emotions, as this is needed to best survive in an environment. We can tell happiness for anger and happiness from excitement, so when they are paired a person will rate them as being different or somewhat similar without considering the deeper aspects of the emotion as defined by valence or arousal levels.  

     One significant finding of the study was the speed to which the participants reacted to pairings featuring fearful expression. Participants were quicker to respond to these pairings than those without fear. This was thought to be due to the “process of detecting fearful call or expression having to be fast and with no additional thinking processing needed for response” (Lo, Li, Lee, & Yueng 2018). So fear response is automatic because one needs to quickly engage in fight or flight behaviors to ensure survival.

     While there was no evidence to support the idea that humans are capable of quickly picking up on the similarities of different facial expressions of emotions that differ completely or share similar valence or arousal scores, Lo, Li, Lee, & Yeung do introduce the concept of emotions having deeper degrees of variability amongst each other. Previously researcher focused solely on our ability to differentiate between two emotional displays, but now emotions can be compared using their level of valence and/or arousal. It was suggested that level of dominance (harmfulness) be used in place of arousal level, because it to can be used to gauge the producers intentions. Using dominance as an dimension of emotion would change how the receiver would rate the emotion since those who appear more likely to cause harm would be more likely to be avoided when displaying negative emotions (eg anger) and conversely those appearing to be less likely to cause harm would be less likely to be avoided when displaying anger because they do not pose a threat. While Lo,Li, Lee, & Yueng did not find evidence to support their hypothesis, they did introduce new concepts to the study of emotion and future studies will expand the available information regarding emotion using their study as a starting point.

References

Seyfarth, R. M., Cheney, D. L., & Marler, P. (1980). Monkey responses to three different alarm calls: evidence of predator classification and semantic communication. Science, 210(4471), 801-803.

Lo, L. Y., Li, W. O., Lee, L. P., & Yeung, P. S. (2018, 05). Running in fear: An investigation into the dimensional account of emotion in discriminating emotional expressions. Cognitive Processing,19(4), 505-515. doi:10.1007/s10339-018-0868-3.