CogSci 2025

•

July 31, 2025

•

San Francisco, United States

keywords:

group behaviour

agent-based modeling

bayesian modeling

philosophy

Formal models in social epistemology explore why rational agents might polarize. While paradigmatic models focus on binary topics, e.g., ”Is H true or false?”, many real-world issues involve multi-option topics: ”Which of n > 2 options is true/best?” This paper introduces a model of rational deliberation on multi-option topics to address the following question: As a group discusses more options, should we expect their beliefs to polarize more or less? We find a dual effect: as the number of options increases, agents are more likely to disagree on which option is most likely correct. This makes it harder to reach consensus on a single position. At the same time, their beliefs—and thus their disagreements—become less extreme. Hence, while agents are more likely to disagree, these disagreements are less intense. Since each trend aligns with a familiar concept of polarization, more options can increase or decrease polarization, depending on one’s measurement.

Downloads

PaperTranscript English (automatic)

Next from CogSci 2025

Investigating the Relationship Between Rumination and Executive Functions: The case of Inhibition and Switching
poster

Investigating the Relationship Between Rumination and Executive Functions: The case of Inhibition and Switching

CogSci 2025

Georgios Boulouzos and 1 other author

31 July 2025

Similar lecture

Rational Polarization: Sharing Only One’s Best Evidence Can Lead to Group Polarization
technical paper

Rational Polarization: Sharing Only One’s Best Evidence Can Lead to Group Polarization

CogSci 2024

Leon Assaad

27 July 2024