Contributions to the contextualization of human-machine spoken interaction