Coercion (linguistics)

In linguistics, coercion is a term applied to a process of reinterpretation that is triggered by a mismatch between the semantic properties of a selector and the semantic properties of the selected element.[1] The term was first used in the semantic literature by Marc Moens and Mark Steedman, who adopted it on "loose analogy with type-coercion in programming languages".[2] In the framework of the generative lexicon (a formal compositional approach to lexical semantics), Pustejovsky (1995:111) defines coercion as "a semantic operation that converts an argument to the type which is expected by a function, where it would otherwise result in a type error." "Coercion" in the Pustejovsky framework includes complement coercion and aspectual coercion.[3]

An example of complement coercion is the sentence "I began the book", where the predicate "began" is assumed to be a selector which requires its complement to denote an event, but "a book" denotes an entity, not an event. So, on the coercion analysis, "begin" coerces "a book" from an entity to an event involving that entity, allowing the sentence to be interpreted to mean, e.g., "I began to read a book," or "I began to write a book." [3]

An example of aspectual coercion involving temporal connectives is "Let's leave after dessert" (Pustejovsky 1995:230). Another example of aspectual coercion from psycholinguistics research includes sentences such as "The tiger jumped for an hour," where the prepositional phrase "for an hour" coerces the lexical meaning of "jump" to be iterative across the entire duration.[4]

Coercion is closely related to the notions of active zone, construal/conceptualization, and syntactic accommodation known from various schools within the cognitive linguistics movement.