Dr. Ian Lustick, a Professor and Bess W. Heyman Chair of Political Science at the University of Pennsylvania, leads CSED scholars and affliates in a discussion of his collaborative work on the importance of local knowledge in influencing behaviors.
Recent work on anti-regime mobilization, protest behavior, ethnic polarization, and language adoption, has focused on the crucial role of cascading mechanisms and the tipping points that may set them in motion or sustain their operation. Much of the formal or semi-formal work in this area has relied on game theoretic or rational choice approaches that make radical assumptions about individuals and the populations they belong to in order to achieve algebraic tractability. One of these assumptions is the effective absence of spatiality, or local knowledge.
In this paper, authors Ian Lustick and Dan Miodownik conduct agent-based modeling experiments to investigate the implications for cascading patterns of change in a population of varying “local knowledge.” Lustick and Miodownik use the PS-I agent based modeling platform to experiment with “Tippy,” a virtual world with agents able to assume one of two states and initially perfectly divided between activation on each state. A tip is defined as having occurred if all agents in the array arrange themselves to be activated on one of the two states. Expanding the size of each agent’s sight radius operationalizes the relaxation of limits on the amount of local knowledge available to individuals. Results suggest increasing cascading behavior with increases in agent sight radius, leading the authors to conclude that the literature tends to substantially over predict cascades.