Events

Psychological concepts are not static: their meanings, moral significance, and social functions can shift across historical periods and cultural contexts. This talk presents a common natural language processing framework for studying such changes in large-scale historical text corpora. Using U.S. Congressional speeches and China’s People’s Daily, we applied sentence-embedding methods to quantify the extent to which texts from each year expressed theoretically defined psychological constructs. I illustrate this approach through two related applications. The first examines changing relational norms—care, hierarchy, transaction, and mating—within discourse about different social relationships. The resulting patterns recovered theoretically expected relational signatures, such as stronger mating norms in discussions of romantic relationships and stronger hierarchy norms in discussions of employer–employee relationships, while also revealing how these expectations shifted historically. The second application examines changing moral conceptions of suffering, including concern for suffering, belief in adversity as a source of growth, and lack of empathy. The analyses revealed culturally distinct trajectories and different relationships between economic conditions and the moralization of suffering. Together, these projects demonstrate how historical text, psychological theory, and modern language models can be combined to study the cultural evolution of psychological concepts at a scale difficult to achieve with conventional methods.