Abstract
This study investigates how belief dynamics and social network structures generate different patterns of social change and diversity. The two belief dynamics studied here are indirect minority influence and random drift; the former is parameterized by a leniency threshold ($\lambda$) and the later by an error rate ($\epsilon$). The patterns of social change are examined in terms of magnitude, speed, and frequency. Diversity and polarization are examined in terms of global belief variation (inverse Simpson index) and local neighborhood difference (Hamming distance). Key findings are that indirect minority influence robustly produces a gradual, small, yet frequent social change across various network structures. However, random drift produces a rapid punctuated social change especially in a society with high connectivity such as complete, scale-free, or random networks but gradual changes in lattice or small world networks. When a society has a modular community structure, indirect minority influence generates a diversity regime whereas random drift generates a polarized regime. Finally, distinct tipping points for social change were identified in different network structures.\end{abstract}
