The recent emergence of “work discussion spaces” represents a major opportunity to better integrate “real work” within businesses’ processes of action and decision-making. In the desirable hypothesis where discussion operates in an autonomous way among actors, the engineering of these spaces has to be designed in such a way as to manage without an external third party during the discussion. But one still has to be sure that discussion is genuinely about real work. This text makes a theoretical-methodological proposal to answer this need: access to real work within discussion can be envisaged thanks to a methodology of eliciting activity among participants who are professionals possessed of knowledge about their own work activity, which needs to be brought to light. Activity thus elicited should be able to be “embodied” by mobilizing the eliciting subject’s activity, in order that the group can grasp it, analyze it and draw out teachings in order to act upon situations. Different ways of eliciting, already used in research and intervention structures (confrontation, explicitation, simulation), can usefully be mobilized in this direction.
Structures thus built up allow a quite original process of activity-analysis by eliciting to be undertaken. We propose modelling on the basis of four discussion-structures to which ergonomists have contributed. This process of analysis benefits from being followed by an action process aimed at transforming situations in line with the objectives of ergonomics. It also necessarily has to be preceded by a profound diagnosis of the business’s problematics, so as to orient the issues connected to discussion spaces. Although the ergonomist is in fact absent from the discussion, s/he should on the contrary be particularly present in the construction of the initial diagnosis and the design of the discussion structure arising from it.