This article examines a peculiarity dating from Classical times, namely, that democracy may be achieved, in practice, independently of and prior to its articulation as theory. This peculiarity has implications for the way in which the history of democratic theory is understood, and also for the place of the democratic theorist in society. Paul Feyerabend, Richard Rorty, Chantal Mouffe and John Keane are theorists of democracy, but they all depart, first, from the commitment to the universal truth?claims that underpin other (...) schools of democratic thought, and, second, from the concomitant belief in the priority of theory over practice. In doing so, they make it difficult to theorise how democracy might be brought about, except, circularly, where it already exists. On the one side, Feyerabend, Rorty and, to a lesser extent, Keane, assume that ?democracy? already exists, so that its realisation requires no theory. On the other, Mouffe argues that ?democracy? does not yet exist in practice, but her attempts to theorise its realisation are not convincing. (shrink)