The practice of make-believe has traditionally been viewed as valuable by psychologists and philosophers more or less exclusively as a mechanism for social initiation, a tool for everyday problem solving, or a method for children to learn about adult responsibilities. This instrumentalist approach has influenced the development of a wide variety of approaches to the pedagogical use of make-believe, from classroom experiments in play therapy to the influential "New Games" movement. This article examines the alternative hypothesis that games of make-believe should be conceived of as ends in themselves. It is suggested that a general explanation of the inherent value of imaginative play would provide support for Kendall Walton's claim that to describe any artwork as representational is just to treat it as a prop in some game of make-believe. Also discussed are attempts made by Ron Edwards and Bernard Suits to provide general accounts of the value of make-believe. Their work hints at the existence of some surprising connections between this issue and certain broader and more fundamental philosophical concerns about the human good. (Contains 25 notes.)