Although vocal expressivity is a major component of human communication, it has not been a major topic in linguistics in the 20th century. The situation is radically different now with the worldwide development of many scientific programs focusing on the vocal expression of emotions. In most of the studies, however, the vocal encoding of emotions is processed according to perceptual and/or acoustic correlates only. The goal of this paper is to show the contribution of a unified approach based on the combination of prosodic cues and syntactico-semantic constraints for a global understanding of the different linguistic mechanisms involved in the expression of emotion in speech.