New paper unifies opinion dynamics via strategic communication game theory

A new paper submitted to IEEE (arXiv:2607.16036) provides a game-theoretic foundation for opinion dynamics. The authors—Hassan Munif, Anthony Couthures, Vineeth S. Varma, Samson Lasaulce, and Tamer Başar—show that common communication mappings (linear exchange, saturated signaling, and discrete public actions) can all be derived from a single strategic communication game played on a weighted influence network. Each agent acts both as a receiver (estimating neighbors’ opinions) and as a sender (broadcasting a public signal to influence its audience). The key innovation: the network’s effect on a sender reduces to a scalar "network-induced exaggeration factor," allowing the game to decouple into independent scalar cheap-talk problems. This framework unifies seemingly disparate models: aligned incentives recover linear averaging, persuasive senders facing naive receivers produce saturated signaling, and persuasive senders facing Bayesian receivers cannot credibly reveal opinions, yielding interval quantizers that match CODA (continuous-opinion discrete-action) dynamics.

The work also embeds these communication rules into repeated opinion updates, creating strategic-CODA dynamics that preserve opinion clustering and exclude extremism. A striking prediction emerges: the more disconnected an audience is from a speaker (i.e., the less influence the speaker holds), the more the speaker exaggerates their public signal. Under strong exaggeration, credible public expression collapses to a simple binary (yes/no) stance even while private opinions remain continuous. This has implications for understanding polarization, social media echo chambers, and the collapse of nuanced discourse into binary positions. The paper offers a rigorous mathematical framework and suggests that network topology directly amplifies or dampens extreme public expressions.