This paper is concerned with the power of social science and its methods. We first argue that social inquiry and its methods are productive: they (help to) make social realities and social worlds. They do not simply describe the world as it is, but also enact it. Second, we suggest that, if social investigation makes worlds, then it can, in some measure, think about the worlds it wants to help to make. It gets involved in 'ontological politics'. We then go on to show that its methods - and its politics - are still stuck in, and tend to reproduce, nineteenth-century, nation-state-based politics. How might we move social science from the enactment of nineteenth-century realities? We argue that social-and-physical changes in the world are - and need to be - paralleled by changes in the methods of social inquiry. The social sciences need to re-imagine themselves, their methods, and their 'worlds' if they are to work productively in the twenty-first century where social relations appear increasingly complex, elusive, ephemeral, and unpredictable. There are various possibilities: perhaps, for instance, there is need for 'messy' methods. But in the present paper we explore some implications of complexity theory to see whether and how this might provide productive metaphors and theories for enacting twenty-first-century realities.