Abstract
Individuals frequently adopt others' perspectives both when interpreting language and when formulating their own responses in conversation. This experiment tested how participants used perspective information to resolve references for bare nouns in Mandarin. Specifically, it explored whether, when faced with two interlocutors, participants distinguished between each individual's perspective or considered both as a whole. Using a classical referential game, the study manipulated the visual perspectives of two partners. In Experiment 1, both speakers had the same seating direction and visual field, and the results showed that participants equally took their perspectives into account above chance levels, providing a baseline finding for referential resolution of Mandarin bare nouns in perspective-taking studies. In Experiment 2, both speakers had the same seating direction but one of them shared the larger portion of visual field with the participants. The results showed that participants took the perspectives of the two speakers independently, while also comparing the perspectives of both interlocutors to facilitate quicker and more accurate referential resolution. These findings demonstrate that perspective-taking is a complex and dynamic process, providing evidence for the study of perspective-taking in Mandarin and contributing insights into comprehension processing in multiparty conversations.