My cousin alerted me of this paper about the cross-cultural perception of facial expressions, by Rachel Jack and colleagues. The study uses an innovative method to uncover how we perceive facial expressions, and more specifically whether `Westerners’ (Europe, North America) and ‘East Asians’ (China, Japan, Korea, Mongolia, Thailand, Taiwan) use the same criteria.
The authors find several differences, such as that Westerners pay more attention to the lower part of the face, while East Asians to the eyes. Also, East Asians attach more significance to the immediate signs of emotion, while Westerners pay more attention to later parts of facial displays. While more work is neededd to understand the reasons for these differences (the authors have some ideas in their Discussion), the data show clear variation in how emotions are perceived (and, presumably, produced) across cultures. The implication is that our emotion recognition mechanism cannot be wholly innate, but it must be open to learning the specifics of each culture.
This conclusion agrees with the fact that other aspects of facial perception, such as the perception of attractiveness, may vary across cultures. Many years ago (on my timescale, it was 2002), some colleagues and I tried to figure out how much nature and how much nurture we should expect in the perception of attractiveness. We concluded, based on what we know about perceptual mechanisms and about the evolution of biological signals, that people’s criteria of attractiveness should be mostly learned, thus leaving space for cross-cultural variation also in this domain. The paper is here.
Method. In case your are still reading, the method employed to reconstruct the criteria of Westerners and East Asians is as follows (simplifying a bit). A computer software generates short (1.25 s) animation of faces by combining randomly selected “facial action units” (AUs), which represent how different muscles move parts of the face. These movies (lots of them) were rated by observers for intensity and quality (sadness, happiness, anger, etc.). The idea is that, although the movies show random combinations of AUs, statistical analysis of people’s responses allow the reconstruction of each AU’s role. For example, the analysis might pick up that expressions including AU 6 (raising the cheeks) are typically rated as happier than expressions without AU6.