Recent and future technological progress in the fields of signal processing, algorithms and artificial intelligence foresee communication between men and machines becoming increasingly close to that between two human beings, thus revealing a strong tendency towards anthropomorphism. In the field of telecommunications, the principal medium of communication is still the voice. Today, machines can speak (through vocal synthesis) and hear (using word recognition). They even "reason" and are capable of holding a discussion by means of software agents for intelligent dialogue.
However, although they carry out the principal functions of vocal communication, it is tempting to say that, pragmatically, machines are still deaf and dumb. We should remember that pragmatism is to a certain extent the "function of form", that is to say, all of the information borne by the form of the signal (vocal, in this instance) which may modify the initial sense of the semantic content. Irony is a good example: a certain tone of the voice will lead to an altogether different understanding of a message like "The weather is still fine today", compared to a neutral tone, especially if the weather is rainy when this sentence is uttered.
This chapter presents the bases for studies concerned with the emotions carried by vocal interaction and attempts to illustrate to what extent these studies lead to manmachine interfaces, especially with regard to communicating objects that are simpler and more pleasant to use by users.