It is argued in this article that the concept of practice is one of the key concepts in Wittgenstein's later philosophy. It partly replaces his earlier talk about the inexpressible. ?The practice has to speak for itself, as Wittgenstein succinctly puts it. The concept of practice not only points to the ways in which the unity of our concepts are underpinned, as Gordon Baker has it, it also comprises the skills involved in handling the conceptualized phenomena, our prereflective familiarity with (...) them, expressed in the sureness in our behaviour towards them, and the judgmental power exercised in applying or withholding a given concept on a particular occasion. These factors are all relevant to the establishment of knowledge, but they cannot themselves be fully and straightforwardly articulated by verbal means. Nevertheless, they represent what we go by when we apply concepts and other types of rules. To follow a rule is what Wittgenstein calls a practice. The sketched analysis of this concept makes us understand better how it is possible to apply a rule without the support of another rule. It also makes us realize in what sense one is justified in talking about tacit knowledge in connection with the application of concepts and rule?following in general. Quite a lot hangs on seeing the world aright at this point. (shrink)

This article examines the learning of a scientific procedure, and its connection to the greater scientific community through the notion of Wittgensteinian rules. The analysis reveals this connection by demonstrating that learning in interaction is largely grounded in rule-based community descriptions and judgments rather than any inner process. This same analysis also demonstrates that learning processes are particularly suited for such an analysis because rules and concomitant phenomena comprise a significant portion of any learning interaction. This analysis further reveals the (...) elucidating merit of Wittgensteinian rules, their relation to community and the concept of practice, and promotes the efficacy of participant-generated rule-formulations as analytic descriptors. (shrink)