Towards a Blackboard Model of Accenting

This paper sketches the outlines of a model of accenting that is inspired by the so-called`blackboard' architectures used in artiicial intelligence. This simple model, which is based on a combination of syntactic/metrical and se-mantic/pragmatic considerations, tries to explain how a conspiracy of various factors (namelyòbstacles' as well as`triggers' for accenting) determines the locations of sentence accents in speech. The interactions between factors related to novelty and constrastivity are studied in some detail. The paper focuses primarily on the problem of accent location in the prosodic component of a concept-to-speech system, while also devoting some space to the converse problem of accent interpretation in speech understanding.