We often rely on our instincts when interpreting others’ facial expressions.
However, new research suggests that our interpretation of emotions might be less accurate—and more influenced by language—than we realize.
A recent study reveals that merely labeling facial expressions with emotion words can significantly distort how we later remember those expressions.
The research, led by Kaitlyn M. McMullen and colleagues from the University of New Hampshire, St. Mary’s College of Maryland, and Yale University, was published earlier this year in the journal Emotion.
Facial expressions of emotion, which we encounter and interpret daily, are often subtle and ambiguous.
Our brains typically simplify this complexity by categorizing these expressions, such as distinguishing between “angry” and “fearful.”
But the process of categorization itself can introduce biases into how we perceive and recall these emotions.
To investigate this phenomenon, the researchers conducted two separate studies involving more than 600 participants. Participants were shown computer-generated facial expressions that blended two emotions (anger and fear) in equal increments.
Initially, participants learned to distinguish between two ambiguous expressions labeled explicitly with emotion words, such as “more angry” or “more fearful.”
Later, they were tested to see if they could correctly recognize these learned expressions among other similarly ambiguous faces.
The researchers found a consistent bias: participants frequently misremembered expressions as more extreme than they actually were.
For instance, expressions labeled as “angry” were recalled as closer to a stereotypical angry face, even though they were only mildly angry in reality.
Crucially, this bias significantly diminished or disappeared altogether when emotion labels were replaced with neutral labels like “Face A” or “Face B.”
These findings emphasize the powerful role language plays in shaping our perceptions. When we attach labels to emotions, we unintentionally reinforce stereotypes associated with those categories.
Moreover, the study suggests that personal beliefs about emotions also influence memory biases.
Participants who held stronger essentialist beliefs – thinking emotions are biologically innate and clearly distinct – exhibited stronger memory biases, particularly when explicit emotion labels were absent.
This research is significant because it highlights how deeply our language and cognitive beliefs influence emotional perception, which could have practical implications for mental health and social interactions.
For example, consistently misremembering someone’s expression as more angry or fearful than it truly was might lead to misunderstandings, increased stress, or unnecessary conflict in relationships.
Yet, the authors acknowledge that their study has limitations.
The participants were primarily young college students, potentially limiting the generalizability of the findings to broader populations.
Future research could explore whether these memory biases occur similarly across diverse groups, using more varied emotional expressions and participants from different age groups and cultural backgrounds.
Despite these limitations, the study provides compelling evidence about the subtle yet profound ways in which emotion words shape our memories.
It encourages mindfulness in how we label emotions – not only in others but also in ourselves – as these labels might affect our emotional health, interpersonal relationships, and overall understanding of human emotions.
Citation
McMullen, K. M., MacVittie, A., Leffers, J. S., Betz, N., & Wormwood, J. B. (2025). Categorization processes in emotion expression recognition: The roles of language and essentialism. Emotion, 25(5), 1151–1168. https://doi.org/10.1037/emo0001483
