Through the description of a clinical case of a child and a session incident, the author shows how the processes of cross-identification intervene at several levels in child analysis : first, between the analyst and the parents to establish a therapeutic alliance that is indispensable for carrying out the analysis, then between the analyst and the child in the intrinsic work of the sessions, where primary and archaic identifications can be played out again, but also between the child, their parents and the outside world in therapeutic consultations.