Assessing the Power of Words to Facilitate Emotion Category Learning

Previous research suggests that labels shape the categorization of emotional stimuli such as facial configurations, yet the strongest evidence of labels' influence on category learning comes from work on object categories. In particular, Lupyan et al. (Psychol Sci 18(12):1077-1083, 2007) found that novel categories of aliens were learned faster by participants provided with nonsense labels during feedback. We summarize a series of five studies in which we examined whether this word-enhancement effect on learning would extend to novel categories of emotion. These studies were conceptual replications of the paradigm used by Lupyan et al. (Psychol Sci 18(12):1077-1083, 2007) designed so that participants would associate novel expressive behaviors with situated experiences. We hypothesized that participants would learn to categorize exemplars of novel emotion categories over the duration of the task, and that categorization would be facilitated for participants who were presented with category labels during learning. We simultaneously analyzed data from all five studies in an integrative data analysis, allowing us to test the effects of learning over time and label condition with increased statistical power. Across all five studies, we found that, while participant performance did improve over time, in no case was it facilitated by including emotion labels at feedback. These results join others in suggesting that the effect of labels on emotion categorization may be more context-dependent than previously supposed-varying by the type of category learning task as well as the specific categories being learned and their relationship to previously acquired knowledge-such that there may be multiple pathways for emotion category learning.

The online version contains supplementary material available at 10.1007/s42761-021-00084-4.