The ever-proliferating views of Chaucer's texts amount in part to disagreements about who or what determines his narratives: lifelike characters, doctrinal principles, the cycles of history, material conditions, the prototypical subject, the reader, even the text itself. In Chaucer's Agents, Carolynn Van Dyke shifts our focus from any particular kind of cause to the representation of cause itself that is, to agency. 'Agency' is widely used but seldom defined. Indeed, academic writers use it in contrary ways. To linguists, philosophers, and most social scientists, it means the power to initiate actions, but economists and legal scholars define it as delegated power. Defining 'agency' broadly as the capacity to cause action, Van Dyke argues that the words opposing uses reveal a fundamental ambiguity: agency is always double, autonomous and subordinate. That doubleness was particularly evident in late-medieval England. Political and ecclesiastical rulers aggrandized power with instruments that weakened it. Philosophers denied reality of universal ideas but acknowledged their force as mental representations. Textual scholars and poets simultaneously downplayed and emphasized human authorship. Chaucer responded to those fluctuations by modeling them. His works deploy an exceptional range of agents, from lifelike peasants to transcendent personifications, and the kind of agency continually changes both within and among individual texts. Chaucer's Agents draws on medieval and modern theories of agency to provide fresh readings of the major Chaucerian texts. Collectively, those readings aim to illuminate Chaucer's responses to two great problems of agency: the degree to which human beings and forces qualify as agents, and the equal reference of 'agent' to initiators and instruments. Each chapter surveys medieval conceptions of the agency in question: allegorical realities, intelligent animals, pagan gods, women, and the author and then follows that kind of agent through representative Chaucerian texts.