In his social theory, Latour lays emphasis on description rather than explanation. He advocates for the largest description possible that follows the tracing of the relations between actors^[1]1Bruno Latour, Reassembling the Social. An Introduction to Actor-Network-Theory. (New York: Oxford University Press, 2007).. The benefits of this approach is to force the researcher as cartographer to look again and again, to avoid neglecting seemingly secondary clues and connections. Yet as all predefined sociological notions such as society or class are suspended, this approach is in danger of losing sense of priorities and urgencies in a quest for exhaustivity. Confronted with the same question in his discussion of Foucault as a cartographer, Gilles Deleuze argues that weighing relations always involves a theory of power^[2]2Gilles Deleuze, “A New Cartographer (Discipline and Punish),” in Michel Foucault: Critical Assessments, ed. Barry Smart (Routledge, 1988), 284.. The diagram is a map of the relations of forces understood in a broad sense. Diagramming is the perilous exercise of finding the balance between the imperatives of looking anew, suspending judgement and a sensitivity to the tensions that manifest the field of struggle.

A priori, the object of a diagram, in my case of a controversy, is presumed to be already charged with tensions and oppositions. The practice of diagramming requires the unlearning of expected positions, working against my initial assumptions. To undo a certain image of the controversy that I have in mind when I encounter a situation. To show for instance how much actors need to collaborate and agree to be able to fight/ oppose each other. This unlearning is the measure of its success. The nature of diagramming is then a relation with an image that it re-maps, re-describes. Diagramming is always about making these two images (how I see agents present themselves to me initially and the map of their relations that I discover) available to the reader/viewer. This tension is what makes it interesting.

The danger is to consider the diagram as something that definitely illuminates a controversy. To a certain extent, the mapping metaphor carries with it expectations of definite knowledge and runs the risk of leaving no space for the logic of practice that resists a full explicitation. It is crucial that a diagram contains elements that remain self-defeating. It is important that it deals with the expectations it inevitably generates. A good diagram must remain somehow scrambled.