Constructing Emotional Memories: Influence of Emotion Labels on Perceptual Memory Biases for Ambiguous Facial Expressions

Across 4 studies, we examined memory biases for ambiguous facial expressions of emotion. In all studies, participants first learned to discriminate between two ambiguous target expressions drawn from the middle of a continuous array of blended expressions (i.e., an array created from morphing images of two facial expressions together, with still images selected in equidistant increments). Participants were then asked to remember which specific expressions they learned to discriminate. Across all 4 studies, we observed evidence of categorical processing biases in participants’ memories for ambiguous expressions labeled with emotion category words. In all studies, target expressions labeled with emotion words (e.g., angry’ or ‘fearful’) were misremembered as more perceptually distinct from one another and as more perceptually similar to the stereotypical expression for their labeled emotion category than they were in reality. Critically, memory biases were significantly reduced when the provided labels indicated non-membership in a category (e.g., not angry’; Studies 1 and 2), were arbitrary (e.g., ‘Face A’; Study 2), or were other socially relevant but non-emotion category words (e.g., ‘masculine’; Study 3). Finally, contrary to predictions, we also found that the biasing influence of emotion category words on memory for ambiguous facial expressions was driven by the words being presented during retrieval (i.e., memory) and not encoding (i.e., learning) of the target expressions (Study 4). Findings highlight the critical role of language in constructing memories of facial expressions of emotion and provide evidence that words indicating category membership can promote categorical processing biases when remembering ambiguous stimuli more generally.