On the role of generalized conversational implicature in semantic change

Abstract

Levinson (1995, 2000) and Traugott & Dasher (2002) suggest that semantic change proceeds from particularized conversational implicature via generalized conversational implicature to coded meaning. However, this model is ultimately neither theoretically nor empirically tenable. Our alternative proposal builds on the assumption that PCI are in the communicative foreground of a message while GCI are in its background. The following sequences therefore seem to be possible: (1) a PCI semanticizes directly: PCI (* → GCI) → coded meaning; (2) a PCI turns into a GCI, but is not fully semanticized: PCI → GCI (*→ coded meaning); (3) a GCI semanticizes, but only after being foregrounded as a PCI: GCI → PCI → coded meaning.