Harnessing the tacit knowledge latent in communities of practice in organizations is a major impetus in knowledge management research and practice. The concept of practice itself is closely associated with activity that is below-view such as intuition and tacit knowing. Indeed, the features binding the members of the communities are often tacit in nature, including things such as rules of thumb, ideologies, embedded habits, or predispositions. Much research in knowledge management posits a dichotomy of doing and saying: what we can do is, in these frameworks, necessarily distinct from what we can say. Polanyi’s (1966) idea that “we know more than we can tell” (p. 4) is often cited to affirm this differentiation. This article seeks to review this relation with tacit knowledge as a focus and suggests that the skillful practice of communities of practice is carried in the discourse which they produce. We adopt a functional approach to discourse, drawn from Systemic Functional Linguistics, that suggests a realization relationship between doing, meaning, and saying rather than a series of dichotomies involving these three semiotic modes. According to this view, what we can say embodies what we can mean which in turn embodies what we can do (Halliday, 1975). This approach is in accord with Wenger’s (1998) opposition to formalist dichotomies when theorizing social action. This entry is structured to present the potential of discourse analysis as an analytical tool to understand the tacit component of participation in communities of practice. The background section details the issues which theories of practice have raised for knowledge management and information systems research. We then review the analytical tools which the field of linguistics offers to uncover implicit knowledge and assumptions in communities (e.g., Iedema, 2003; Jorderns & Little, 2004; Zappavigna-Lee & Patrick, 2004; Zappavigna-Lee et al., 2003). We conclude by arguing that the nature of our skillful practice may be carried in language: we articulate what we know through patterns and features of language of which we are not consciously aware. Analysis of this kind of language aims to elicit implicit meaning and is allied with psychoanalytical methods that attempt to understand implicit aspects of social experience.