How does the mediation mechanism define what can and cannot be said? What do science communication stakeholders say when they speak? Can they speak from their own point of view (Gilligan, 1982), and if not, whose voices are they embodying? Can they, for example, position themselves on the basis of their experience of research or contact with research? Or, on the contrary, do they step aside to speak "in the name of science"? What sciences does mediation talk about and what aspects does it omit? Under what conditions can science communication be a source of reflexivity? Science-society mediation efforts contribute in particular to the circulation of knowledge, the construction of a certain image of science and the opening up of spaces for speaking and discussion between research actors and civil society. These forms of oral or written communication make it possible to establish a diversity of relationships with the audiences they address, and can open up the space for reflection, potentially allowing science communication actors to impose their presence in the mediation context to explicitly activate their relationship with science and to bringing their attachments to practices into the relationship with the audiences. Although science communication actors are used to and trained to take audiences into account, in a communication context, they often forget that they themselves are a non-neutral actor in the context, with a direct effect on the context owing to their involvement, by means of their way of being. As such, the interests, concerns, ideologies and alliances of the speaker are rarely questioned, even though they form the basis of the discourse and, to an even greater extent, the mediator's relationship with their audience. Georgio Agamben (2014) gives a definition of the mechanism (see also Foucault's works, 1971; 1975) that allows us to closely examine what is at stake in scientific mediation: "anything that has, in one way or another, the capacity to capture, orient, determine, intercept, shape, control and ensure the movements, behaviours, opinions and discourses of living beings." As such, a mediation mechanism does not allow one to behave freely, nor to say everything that could be said about science, and often restricts mediators to what is expected of their movements, their behaviours and their discourse, for the benefit of science and the institutions (laboratories, universities, etc.) they represent. The mechanism does not allow their own voice to be heard: it is simply not built for that. It is embedded in a fabric of implicit norms and values, both in the way science mediation is conceived and in science as it should be and as it should be represented. Mediators and researchers construct their narrative and their discursive identity by adhering to these norms and values, or on the contrary in contrast thereto when the mediation mechanism is constructed to accommodate a critical discourse. The mechanism thus determines the way in which the actors who participate in it will consider what matters and how they will speak. Would understanding the mediation mechanisms in which we are caught up enable us to activate other forms of communication, other relationships and other narratives on and with science? At this stage, I can rephrase my questions: how can one speak within this mechanism? On whose behalf do PhD students speak? What voices are put forward? How does the researcher approach the subject? What are they narrating? What sciences do they talk about? Which sciences are represented through such a mechanism? [[How to share the sciences]] Who speaks? What is the relationship? Thinking about mediation in terms of presence and relationship - going beyond the diffusionist model Speaking about science < [[Speaking about science - reading]] Baudouin Jurdant chooses to question popularisation as a way of speaking about science. Not by speaking "in the name of science", but by reintegrating a viewpoint of its one's own, a particular perspective, in order to repair the removal of the speaker and the deficit of reflexivity characteristic of scientific writing and of objective pretension. Is it a question of this primary orality, of speech without a speaker, or on the contrary, of a possible place of emergence or rebirth of the speaking subject, which initiates the reflexive, critical and political subject?