The purpose of this chapter is to put Onuf’s constructivism into the context of the agent-structure debate. Returning to the debate, Hollis and Smith felt that Carlsnaes failed to overcome the agent-structure problem merely by allowing for time as a variable. In effect, Carlsnaes claimed to solve the agent-structure problem by treating agents and structures as if they take turns affecting the social world. The positivist focus on units as positivities brings the discussion of levels back to its ontological core, an effort well served by analyzing the relation between levels and social constructions as an ontological issue. It is essential to note at the outset that all positivities are always simultaneously parts and wholes. The concluding discussion will look at the levels of analysis debate in International Relations, the relation of which to the agent-structure debate formed an important part of Wendt’s several exchanges with Martin Hollis and Steve Smith.