Phrase structure and prosodic phrasing
Much current work is based on the rather simplistic assumption that higher-level prosodic boundaries such as intonation unit boundaries are derivative of syntactic boundaries (syntactic phrase boundaries serve as input for the computation of prosodic phrase boundaries). One corollary of such an approach is the widespread practice of including prosodic phenomena in the definition of syntactic constructions, a practice particularly common for constituents at either edge of a clause such as topics and afterthoughts.
This talk questions this practice and asks whether it makes sense to include prosody in the definition of syntactic units. It is based on the hypothesis that prosodic and syntactic phrasing provide essentially independent ways of chunking linguistic information. There probably are historical interdepencies between the two phrase types, but in the opposite direction of current assumptions (syntactic configurational structure being historically derivative of prosodic grouping). Importantly, the two phrasing types become actually only distinguishable once they are at least partially independent of each other. Synchronically observable overlaps have the character of strong tendencies rather than fully grammatical regularities. Speakers are essentially free as to form the higher-level prosodic units they need for informational reasons.
Professor Nikolaus Himmelmann, Universität zu Köln