How expression, context and perspective determine judgments of emotion

Journal article

Within the field of Affective Computing, facial expressions were historically assigned emotion labels by annotators without knowledge of the context in which the expressions were produced. Often these 3rd-person impressions were used instead of 1st-person judgments about the target’s feelings. The field now appreciates that 1st- and 3rd-person judgments can differ dramatically and that context plays a central role in explaining these differences. More recently, there is growing appreciation of the importance of a 2 nd-person perspective. When people are engaged in a social task, impressions of their partner’s emotions differ from the impressions of detached bystanders. In this paper, we explore how facial expressions, contexts and perspective (1st, 2nd-, and 3rd-person) interact to determine judgments of emotion. We explore this using automatic facial expression analysis of natural expressions produced in the context of a social task (the prisoner’s dilemma). Our findings suggest that expression and context combine to determine judgments, but the way they combine differs depending on perspective. We discuss the implication of these findings for automatic emotion recognition methods and the inferences such methods can support

Authors: Han, B; Hoegen, J; Lucas, G; Shore, D; Parkinson, B

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: IEEE
• Journal: IEEE Transactions on Affective Computing
• Publication date: 2025-10-07
• Acceptance date: 2025-10-03
• DOI: 10.1109/taffc.2025.3618588
• EISSN: 1949-3045
• Language: English