A dialogue system with contextually appropriate spoken output intonation