CONCEPT
Inferential Role Semantics
Sellars’s account of meaning as the role a symbol plays in a web of inferences—not the thing it names, not the glow inside a mind—and the most prescient theory of language anyone produced before the technology arrived that would make it concrete.
Meaning, on
Wilfrid Sellars’s account, is not a relation between a word and a thing, and it is not an inner mental glow behind the symbol. It is a role—the position a symbol occupies in a structured web of moves. To grasp the meaning of an expression is to know how to use it: what licenses you to assert it, what it licenses you to infer, what it commits you to, how it connects to other expressions and to the world and to action. Sellars called this view inferential-role semantics, decomposing language into three kinds of move: language-entry transitions (the world prompts utterance), intra-linguistic transitions (utterances license other utterances by inference), and language-exit transitions (utterances issue in action). To mean something by a word is to be a player in all three. The most startling fact about modern AI is that this theory of meaning is the one these machines actually implement.