Perceiving emotional and semantic similarity between facial expressions of emotions

Research output: Contribution to journalAbstractpeer-review


Facial expressions of emotion have been assumed to be innate, discrete, and universal across cultures. Recent studies challenge this idea and propose that facial and vocal expressions of emotions are blended, gradient and sensitive to cultural contexts. Studies of facial emotion often use an emotion categorization task, which may conceal the nature of facial emotion perception and its sensitivity to cultural context. In the present study, Asian and European participants firstly rated how much a facial expression is in agreement with a set of emotion categories or semantic concepts (i.e., an emotion- and a semantic-profiling task), and then they made judgments about the similarity between pairs of facial expressions of emotions. We found that the perceptual similarity between facial expressions of emotion can be predicted by the profiling responses to individual facial expressions and by stimulus-based similarity. This result suggests we may use both physical and conceptual properties of facial emotions to determine their similarity. European and Asian participants showed a similar pattern of perceptual similarity between facial expressions of emotion, nonetheless, their responses to the emotional and semantic profiling tasks were different. These results favour a blended representation of facial emotion, demonstrating that the same facial configuration conveys information about multiple categories of emotion and links to a set of interconnected semantic concepts. How we perceive these emotional and semantic aspects of facial expression is affected by both culture-general and culture-specific processes. [This study was supported by a Royal Society Research Grant (RGS\R2\202066) to MZ and a grant from National Natural Science Foundation of China (32071048) to GZ.]
Original languageEnglish
Pages (from-to)226
Number of pages1
Issue number1_SUPPL
Publication statusPublished - 1 Dec 2021

Cite this