Conceptual Role Semantics

This article treats of a family of theories variously known also as “inferentialism,” “inferential/functional/cognitive/causal/computational role semantics,” “functionalism,” and “use-theory of meaning.” However, it will throughout use only “CRS,” short for “Conceptual Role Semantics,” to be understood in a suitably wide sense. This kind of theory has been propounded and discussed in very different theoretical contexts, including the philosophy of mathematics, formal logic, the philosophies of mind and language, and cognitive science. In the philosophies of mind and language, it has been discussed in rather disparate kinds of literature, ranging from discussions about later Wittgenstein to causal theories of mental content. It is a theory of linguistic meaning and/or mental content (depending on which is taken as basic), and holds that what an expression means (or what makes a given concept the concept it is), is determined by the expression’s (concept’s) psychological or inferential role. This role can be identified either in wholly descriptive (e.g., causal) terms or partly in normative terms, and concerns the expression’s or concept’s behavior in inferences. Inferences, in turn, can be understood either in the usual, narrow sense or as involving also perception and action. Examples of a CRS account of the meaning of the word “and” may be that it is determined by the inferences from “A and B” to both “A” and “B” and from the latter back to “A and B” (what determines the meaning of “and” will be either the correctness of these rules [normative CRS] or the fact that they are actually followed [non-normative CRS]). The concept of redness could similarly be taken as determined by transitions from certain perceptions to beliefs involving the concept, and from such beliefs to other beliefs. CRS is the most common approach to meaning in cognitive science and linguistics (with the notable exception of formal semantics). The main alternative is the view that meanings or concepts should be understood in terms of reference, satisfaction, and truth, but ecumenical views are common. CRS can be traced as far back as to the Associationism of the British Empiricists and to Kant, but more relevantly, to verificationism, the later Wittgenstein, Kazimierz Ajdukiewicz, and Gerhard Gentzen (whose interests, however, were restricted to the logical constants). In post-Enlightenment, non-analytic philosophy, similar ideas have been expressed by Friedrich Hegel, Ferdinand de Saussure, and Martin Heidegger. This article, however, focuses exclusively on CRS in analytic philosophy from Ajdukiewicz and Gentzen and onward. The first example of a full attempt at a generalized CRS is typically thought to be found in the works of Wilfrid Sellars, published during the 1950s–1970s. Since then, new works defending some version of CRS, as well as discussions about them, have appeared regularly until the early 21st century. The most important defenders of CRS since Sellars include (in roughly chronological order) Gilbert Harman, Ned Block, Christopher Peacocke, Robert Brandom, and Paul Horwich, while its most influential critic is arguably Jerry Fodor.