Choice is an essential feature in personal construct psychology, and arguably the main task of the personal construct is to clarify why the person chooses one course of action rather than another. Yet the choice corollary appears to be in danger of being tautological, saying that whatever the person did constituted the elaborative choice. G. A. Kelly proposed the choice corollary as a useful heuristic device to counter external observers' beliefs that they can know what is best for the person. However, there is a danger in overcorrecting, in assuming that every action of the person constitutes an elaborative choice. This is most likely to occur in the neurotic paradox, where clients experience self-definition as problematic for them. I argue that we should emphasize construing as inseparable from action. This leads us to a phenomenological rather than a cognitive reading of construct theory. M. Merleau-Ponty's (1962) concept of sedimentation is then helpful in understanding the type of self-definition that is characterized in the neurotic paradox. The implications of this position for psychological reconstruction are briefly discussed