Marina Terkourafi
In its most inclusive, "social meaning" is an umbrella term for inferences about the speaker, irrespective of how these inferences arise, are represented, or change over time/place. However, once these differences are heeded, what various authors have called "social meaning" turns out to share little beyond this general definition. An important difference stems from the observation that, while for Labov truth-conditional equivalence provided the opportunity for social meaning to arise, in experimental pragmatic accounts it is the difference in truth-conditional content that enables it. Consequently, the two types of social meaning are: a) represented differently (indexically vs. propositionally), b) constituted differently (via association/enregisterment vs. by Gricean inference), and c) differ in their arbitrariness or alienability, with the latter remaining stable across time/place, such that they cannot result in language change. Yet, in both cases, there are at least two competing variants and it is the use of one rather than the other that generates social meaning for the hearer capable of identifying the relevant alternatives. However, even this expanded view of social meaning fails to account for social meaning generated in the absence of any alternatives, when expressions devoid of denotational meaning perform a socially binding function for their users. To take in also these cases of social meaning par excellence, where the sole meaning of an expression is its social one, I propose a conceptualization of social meaning as input and as output to the inferential process that allows us to capture its essentially performative, multi-modal, and hearer-dependent nature. • “Social meaning” has varied, sometimes conflicting interpretations. • Recent views link social meaning to differences, rather than equivalence, in truth conditions. • Conceptualising social meaning as inferential input/output captures its complexity. • This view captures its performative, multimodal, hearer-dependent nature.