This paper advocates that some limits of the rational agent hypothesis result from the improper assumption that one individual should be modeled as a single rational agent. We model an individual composed of two autonomous and interacting structures, conscious and unconscious. Each agent utility form depends both on external signals and other structures' actions. The perception of the signal depends on its recipient and its grid of interpretation. We study both the static and dynamic version of this interaction mechanism. We show that the dynamics may display instability, depending on the structures interactions'strength. However, if unconscious has a strategic advantage, greater stability is reached. By manipulating other structures'goals, the strategic agent can lead the whole system to an equilibrium closer to its own optimum. This result shows that some switch in the conscious'objective can appear. Behaviors that can't be explained with a single utility can thus be rational if we add a rational unconscious agent. Our results justify our hypothesis of a rational interacting unconscious. It supports the widening of the notion of rationality to multi-rationnality in interaction.