Research being carried out in the area of intonation assignment in text-to-speech synthesis is reported. An attempt to provide a flexible framework to allow for experiments is described, and output from the system is compared with data from a corpus of spoken English currently being analysed.