摘要 : ? 2022Humanoid robots might become more and more present in the most ordinary contexts of millions of people worldwide. Humans reason about these artificial agents mainly through the attribution of human characteristics, a process... 展开 ? 2022Humanoid robots might become more and more present in the most ordinary contexts of millions of people worldwide. Humans reason about these artificial agents mainly through the attribution of human characteristics, a process called anthropomorphism. However, despite number of studies, how we develop and structure the representation of non-human agents is still an open question. In the present paper, we aim at integrating the anthropomorphism into the cognitive control theory, a construct from cognitive neuroscience that refers to information processing and cognitive resources managing that varies adaptively to the situation. In three experiments we manipulated the cognitive load of participants during the observation of an active robot to investigate how the load could impact the online structuration of participants’ mental representation of the robot. The two first experiment converged in arguing for a control process resource-demanding to switch from the social cognition to the physical cognition inhibiting anthropomorphic inferences. The third experiment investigated the influence of the “what” and “why” observation goals on the cognitive load effect arguing that an explicit focus on intentionality attribution bias the automatic process of anthropomorphism. The representation and perception of robots are further discussed in term of cognitive control theory and social cognition. 收起