Fictional narrative as a variational Bayesian method for estimating social dispositions in large groups

Area of Science:

• Cognitive Science
• Social Psychology
• Evolutionary Anthropology

Background:

• Modelling intentions in large groups requires tracking complex belief hierarchies, which is cognitively demanding.
• Large groups offer coordination advantages through specialization, creating a tension between cognitive costs and benefits.

Purpose of the Study:

• To propose fictional narrative as a cultural tool for managing large group interactions.
• To explain how narrative structures facilitate inference of unknown social motives.

Main Methods:

• Literature review of narrative cognition, framing it within a variational Bayesian inference model.
• Agent-based modeling to simulate social category convergence.

Main Results:

• The proposed variational Bayesian model aligns with existing findings in narrative cognition.
• Agent-based simulations naturally produced narrative-like social categories.

Conclusions:

• Fictional narratives serve as cultural tools that reduce the cognitive load of social interaction in large groups.
• Prototypical narrative roles act as interpretive priors, enabling efficient estimation of social motives.