I address the question of what holds cities—particularly early cities—together. I first discuss the question of urban order, or social order in cities. Archaeologists are starting to come around to the standard answer to that question in the social sciences: urban order is created and maintained by the operation of two types of social forces: institutions (top-down forces) and generative processes (bottom-up forces). I review the changes in archaeological thinking that led from obsolete and inadequate models of social order (statism and agency) to the emerging current understanding. I discuss the concept of institutions, including a new archaeologically-useful definition, and describe five types of early urban institutions (political, economic, political economy, social, and religious). I then present three types of generative process: population/demography, self-governance, and self-coordination. I also urge archaeologists to stop using idiosyncratic definitions of concepts (like institution) that serve to isolate archaeology from wider scholarship in the social and natural sciences.