I argue that the ontological status of fictional characters is determined by the beliefs and practices of those who competently deal with works of literature, and draw out three important consequences of this. First, heavily revisionary theories cannot be considered as ‘discoveries’ about the ‘true nature’ of fictional characters; any acceptable realist theory of fiction must preserve all or most of the common conception of fictional characters. Second, once we note that the existence conditions for fictional characters are extremely minimal, (...) it makes little sense to deny the existence of fictional characters, leaving anti-realist views of fiction unmotivated. Finally, the role of ordinary beliefs and practices in determining facts about the ontology of fictional characters explains why non-revisionary theories of fiction are bound to yield no determinate or precise answer to certain questions about fictional characters, demonstrating the limits of a theory of fiction. (shrink)