One of the classic problems in social science is understanding how self-enforcing norms emerge within large groups. While many different models have been developed to understand these collective behaviors, ranging from linguistic conventions to cultural practices to ideological consensus, studying these dynamics empirically has remained intractable. Using an experimental online environment, we study how the interaction of individual mechanisms with social networks drives the emergence of consensus even in the absence of the forces traditionally argued to be necessary, such as centralized authority, common knowledge, and focal points. Our results inform the effectiveness of various models of collective behavior for understanding this social process.