Spontaneous Symmetry Breaking, Group Decision Making and Beyond 1. Echo Chambers and Random Polarization

TL;DR

This paper models social decision-making using an adapted Ising model, revealing that initial opinions and history influence the formation of echo chambers and polarization, challenging the idea that these are solely due to network structure.

Contribution

It demonstrates that social equilibrium states depend on initial conditions and history, highlighting the role of dynamics in the emergence of echo chambers and polarization.

Findings

Different initial opinions lead to distinct equilibrium states.

Echo chambers form as a result of opinion dynamics, not initial network structure.

Social media algorithms can influence the speed of opinion consensus.

Abstract

Starting from a symmetrical multiple choice individual I build a sociophysics model of decision making. Reducing the choices to two and interactions to pairs recover the Ising model from physics at zero temperature. The associated equilibrium state is a spontaneous symmetry breaking with the whole group sharing a unique choice selected at random. However, my focus departs from physics, which aims at identifying the true equilibrium state discarding any possible impact of the initial conditions, the size of the sample and the used update algorithm. Any memory of the past history is erased. In contrast, I claim that dealing with a social system, the history of the system must be taken into account in identifying the relevant social equilibrium state. Using Monte Carlo simulations I explore the spectrum of non universal equilibrium states of the Ising model at zero temperature. I show that different initial conditions with the same value of the order parameter lead to different equilibrium states. The same applies for different sizes and different update algorithms. The results indicate that in the presence of a social network composed of agents sharing different initial opinions, it is their interactions that lead them to share a unique choice and not their mere membership in the network. This finding sheds a new light on the emergence of echo chambers, which appear to be the end of a dynamical process of opinion update and not its beginning with a preferential attachment. Furthermore, polarization is obtained as a side effect of the random selection of the respective unanimous choices of the various echo chambers within a social community. The study points to social media exchange algorithms, which are purely technical levers independent of the issue and opinions at stake, to tackle polarization by either hindering or accelerating the completion of symmetry breaking between agents.