Network structure shapes consensus dynamics through individual decisions.

Abstract

How do shared narratives emerge in decentralized online networks? Prior research using simplified group coordination tasks (e.g., face-naming) shows network structure shapes group consensus, but the underlying cognitive mechanisms remain unclear. Here, we examine how network structure influences the emergence and semantic content of shared narrative beliefs in experimental online social networks, using natural language processing measures and agent-based modeling. Media content with complex causal structure attenuates network structure effects by encouraging longer exploration of background knowledge. Yet network structure still shapes the narrative content communicated. An embedding-based narrative alignment measure shows that fully connected groups orient their interactions more toward communicating causes of an event, whereas locally connected networks emphasize the event's effects. A group's network structure also influences representational and language change in personal narratives: participants in fully connected networks showed the largest increase in causal language in personal narratives written after interaction, which also orient more around the narrative's causal events.