Abstract
Since Descartes, mathematics has been dominated by a reductionist tendency, whose success would seem to promise greater certainty: the fewer basic objects mathematics can be understood as dealing with, and the fewer principles one is forced to assume about these objects, the easier it will be to establish a secure foundation for it. But this tendency has had the effect of sharply limiting the expressive power of mathematics, in a way that is made especially apparent by its disappointing applications to the social sciences. We should move in the opposite direction: toward a mathematics that deals in general with constructed objects, and whose scope includes fictional, poetic characters as much as numbers and sets