Embedded Metaphor & Subsentential Pragmatics: Revisiting the Scope Argument

The so-called ‘scope argument’ challenges Gricean theories of metaphor by claiming that metaphorical readings are directly expressed. That is, because metaphorical readings survive under the scope of logical and intensional operators, they figure in what is said/explicitly communicated. In this article, I resist that conclusion. I show that other putative implicatures pass the scope test to motivate the idea that _at least _some implicated content arises within embedded contexts while resisting the claim that such content is what is said. To deal with such content, I argue that local, pragmatically inferred content is truth-conditionally relevant without thereby being a part of what is said. This move carries important consequences for how to draw the boundary between semantics and pragmatics. It raises two additional challenges for a theory of metaphor: the calculation and compositionproblem. I address these challenges by sketching a subsentential Gricean model whereby embedded metaphor is treated as a local implicature triggered by pressures on Gricean maxims, compositionally integrated by type, whose meaning is predictably indeterminate and defeasible, unlike said content. Building on work on embedded implicatures, this model preserves Grice’s cooperative architecture while explaining metaphor’s truth-conditional ‘effects’. The result is a lean semantics with a principled account of embedded metaphorical meaning.