This article considers a theoretical problem at the center of historical research on literacy, the solution to which has implications for studies of contemporary literacies as well. Four models of literacy are identified. The (1) traditionally received skills model still flourishes in educational policy, but is increasingly rejected by literacy scholars who employ either (2) a functional model of literacy as an instrument of power relations or (3) a semiotic “marker” model of literacy. While traditional communications history has evinced little interest in historical studies of literacy, (4) a highly visible grand theory model of literacy associated with McLuhan, Innis, and Ong has its origins in the received model. Where the received model is optimistic about the long-term social and political effects of literacy, however, the grand theory model is pessimistic. Although a growing body of scholars has argued that the definition of any literacy must be located in its actual practices, and although many scholars now believe that oral-literate dichotomies are overly simple historical categories, this article takes that thinking farther and argues that literacy does not merely coexist or interact with oral practices and skills, but includes them. That is, the definition of literacy consists in the written and oral practices and skills, but includes them. That is, the definition of literacy consists in the written and oral practices organized around texts in a particular culture. Support for this argument is taken from pertinent evidence in the history of literacy.