1. INTRODUCTIONHow should we make sense of ordinary talk about the mind? When we say that someone has a particular belief or desire, are we claiming that they have a certain sort of causal state inside their heads? If we are, might future cognitive science show that we are wrong? Might it turn out that mental states, like beliefs and desires, do not exist? In this paper, I explore one way to approach these longstanding questions concerning the status of folk psychology: mental fictionalism. Put simply, mental fictionalism is the view that, even if mental states do not exist, it is useful to talk as if they do. Mental states are useful fictions.Recent philosophy of mind has seen a growing interest in mental fictionalism (e.g., Wallace 2007; Demeter 2013a). To date, much of the discussion has concerned the general features of this approach and the difficulties it faces.1 In this paper, I develop a specific form of mental fictionalism by drawing on Kendall Walton's work on make-believe (Walton 1990, 1993). According to the account I will put forward, when we say someone has a particular belief or desire, we are not making a claim about their inner machinery. Instead, talk about mental states is a useful pretence for describing people and their behaviour. As a result, the legitimacy of ordinary talk about the mind does not depend upon whether future cognitive science will discover beliefs or desires inside our heads.The structure of the paper is as follows. First, in Section 2,1 give a brief outline of Walton's theory, focusing on his discussion of metaphor and prop-oriented make-/ -believe. In Section 3,1 show how we may draw on Walton's analysis to develop a fictionalist analysis of folk psychology. In Section 4, I further clarify and motivate this account by comparing it to a number of well-known alternative positions, including behaviourism, instrumentalism, and eliminativism, as well as alternative forms of fictionalism. Finally, in Section 5, I consider some important challenges facing mental fictionalism.2. PROP ORIENTED MAKE BELIEVEImagine some children playing with a doll. They pick the doll up and hold it in their arms, push it down the hallway in a pushchair, cover it with a blanket, and so on. As they do so, the children imagine themselves picking up a baby and cradling it, taking it to the shops and lying it down in its cot to go to sleep. In Walton's theory, dolls and other objects used in make-believe are called props and the rules that govern their use in the game are called principles of generation (Walton 1990). Within a game of make-believe, the properties of the props, together with the relevant principles of generation, make propositions fictional. To say that a proposition is fictional, in Walton's sense, is simply to say that participants in the game are prescribed to imagine it. For example, if the doll's are closed, the children are to imagine that the baby is asleep. It is fictional that the baby is asleep. Notice that, since the content of a game of make-believe depends only on the props and principles of generation, it possesses a certain kind of objectivity: if the doll's eyes are closed then it is fictional that the baby is asleep, even if none of the children happen to notice this.One important feature of games of make-believe is participation. Children playing with a doll do not simply sit and look at it. Instead, they carry out various actions with the doll (e.g., putting a bottle to its mouth) and these actions generate imaginings within the game (e.g., that the children are feeding the baby). Like the doll, the children themselves become props in the game. The children also participate verbally. For example, David might sing a lullaby as he rocks the doll in his arms, thereby making it fictional that he is singing a lullaby to get the baby to sleep. Importantly, acts of pretence can be used to make genuine assertions. Suppose that Anna looks at the doll and says She's sleeping now! …