Learning Theory and the Logic of Critical Mass

The Oliver-Marwell theory of is a prominent solution to the 'free-rider problem that plagues action in large groups. I reformulate the theory as a stochastic learning model in which cooperative responses are shaped by the social sanctions and cues generated by the responses of others. This relaxes four assumptions in the original formulation: that the actors are rational, decisions are isolated events, outcomes are deterministic, and public goods have pure jointness of supply and collective profit. Computer simulations then show how adaptive actors become trapped in a suboptimal equilibrium and how they escape through attainment of critical mass. Start-up problems arise from accommodation to social costs and not from low returns to early contributors as previously believed. Simulations also identify a new dilemma of group size and show why solidarity tends to emerge in response to crisis. Finally, normative solidarity appears to be a consequence rather than cause of critical mass but may promote recovery from random deviance.