In this thesis, I explore how a fictionalist approach can be applied to social ontology, especially for conferred kinds, in the context of action explanation and prediction by comparing this view with various realist views. Chapter 1 is concerned with clarifying constitution accounts and the criteria for comparison. Chapter 2 gives three different kinds of problems against this realist view, which makes this view less credible. Chapter 3 introduces conferralist accounts. Chapter 4 will deal with the question of whether conferralist (...) accounts can solve these problems. I will argue that even if the second problem is solved, the first and third problems still arise in conferralism. In Chapter 5, I propose a fictionalist social ontology. Thus, in Chapter 6, I argue that social fictionalism solves the problems in question. (shrink)