Perspective taking, a cognitive process of understanding information from the others viewpoint, is essential for formingcommunication skills. Whereas this process is considered to involve detachment of the reference frame from the own eyeand attachment of it to the others eye, we instead hypothesized here that it is mediated by representing the others intrinsic(i.e., proprioceptive) coordinate frame, since our cognitive abilities often rely on the physical presence. To examinethis possibility, we asked the participants to learn to control avatars motion in the virtual reality space from the third-person perspective and sought interaction between the ability to represent avatars intrinsic coordinate systems via motoradaptation and the ability to take avatars spatial perspective. We found significant facilitation of perspective taking abilityby the motor adaptation experience, which supports our hypothesis that the perspective taking encompasses a process ofrepresenting the others intrinsic coordinate frame. We suggest that the perspective taking is an embodied cognitive processwhich underpins theory of mind and empathy.