[Paper Review] Why more contact may increase cultural polarization

Andreas Flache, Michael W. Macy|ArXiv.org|Apr 24, 2006

Opinion Dynamics and Social Influence15 references22 citations

TL;DR

This paper challenges the conventional view that increased social contact reduces cultural polarization by showing that under realistic conditions—continuous cultural traits and negative social influence (heterophobia)—greater interaction range can actually increase polarization. Using agent-based modeling with modified Axelrod dynamics, the authors demonstrate that homophily and rejection of dissimilarity amplify fragmentation, reversing the expected homogenizing effect of contact.

ABSTRACT

Following Axelrod's model of cultural dissemination, formal computational studies of cultural influence have suggested that more contact between geographically distant regions may increase overall cultural homogeneity and reduce societal polarization. In the present paper, we show that two plausible modifications of Axelrod's original mechanism turn the effect of range of communication upside-down. We assume a continuous rather than a nominal state space and we add the negative side of social influence, heterophobia and rejection. Computational analyses of the resulting model demonstrate that now a larger range of contact can increase rather than decrease the extent of polarization in the population. Further experiments identify the window of conditions under which the effect obtains.

Motivation & Objective

• To investigate whether increased communication range in cultural interaction models leads to greater cultural homogeneity or polarization.
• To examine the impact of continuous cultural trait spaces versus discrete (nominal) ones on societal polarization.
• To explore the role of negative social influence (heterophobia) in shaping cultural dynamics.
• To identify the conditions under which increased contact leads to higher polarization rather than convergence.
• To extend Axelrod's model of cultural dissemination by incorporating realistic social mechanisms like rejection of dissimilarity.

Proposed method

• Adapted Axelrod's agent-based model of cultural dissemination to use a continuous state space instead of discrete cultural traits.
• Incorporated heterophily and rejection behavior (heterophobia) as negative influence mechanisms, where agents actively resist or reject dissimilar others.
• Simulated agent interactions over time in a spatially structured network with variable communication ranges.
• Used computational simulations to analyze long-term cultural dynamics under different parameter settings.
• Tracked cultural diversity and polarization levels using metrics such as the number of distinct cultural clusters and the variance of cultural profiles.
• Conducted sensitivity analyses across a range of communication ranges and heterophobia intensities to identify critical thresholds.

Experimental results

Research questions

• RQ1 Does increasing the range of social contact in a cultural interaction model lead to greater cultural homogeneity or increased polarization?
• RQ2 How does replacing discrete cultural traits with continuous traits affect the outcome of cultural convergence?
• RQ3 Under what conditions does the inclusion of negative social influence (heterophobia) reverse the expected homogenizing effect of contact?
• RQ4 What is the critical window of parameters where increased contact leads to higher polarization instead of lower?
• RQ5 How do the interplay of homophily and rejection shape the emergence of cultural clusters in a population?

Key findings

• Increasing the range of communication leads to higher cultural polarization when agents exhibit heterophobia and cultural traits are continuous.
• The reversal of the homogenizing effect occurs only within a specific range of heterophobia intensity and communication range.
• With continuous traits and no negative influence, increased contact reduces polarization—confirming Axelrod’s original result.
• When heterophobia is introduced, the system transitions from convergence to fragmentation as contact range increases.
• The model identifies a critical threshold where the balance between similarity-seeking and dissimilarity-avoiding behaviors determines polarization outcomes.
• The number of distinct cultural clusters increases significantly with higher contact range under heterophobia, indicating stronger fragmentation.