Abstract
This study adapts the social network model to build structured representations of conversational interactions in the Tang yulin . In these dialogue networks, nodes represent interlocutors, and edges between nodes indicate the frequency and the direction of words from speakers to listeners. Algorithms calculate a series of statistical measurements that gauge the local and global importance of each node. The study highlights an influential, global “Xuánzong factor” across the dialogue network, and argues that Empress Wu Zetian is depicted as a more passive figure, more of a listener than a speaker, when compared with the male rulers of the Tang.