Prosodic analysis : the next generation

In many text-to-speech systems, sentence prosody is derived by string-oriented rules, which are often rather ad hoc and linguistically unsound. By contrast, a system which employs a syn tactic parser allows for highly general and computationaily simple procedures defining the mapping between syntax and prosody. The program described in this paper, PROS2, is a first attempt towards a "next generation". First, the input sentence is parsed by applying phrase structure rules. Next, the syntactic representation is mapped onto a metrical tree, augmented with "focus" markers. The metrical tree provides an abstract characterization of accent and phrasing. While the program, at the time of writing, is still at the prototype stage, it holds considerable promise for further development and can be expanded in several directions.