We tested the hypothesis that the process of listening and interpreting requires an adaptive regression. A 4-phase model of the underlying therapist's process was proposed. Previous studies showed (a) that reflective countertransference (CT) is related to adaptive defenses, whereas reactive CT is linked to inadaptive use of defenses and (b) that experienced therapists are more reactive and less reflective than are novices. This work was a partial replication. Two instruments were used: the Countertransference Rating System, which distinguishes 3 types of CT (reflective, reactive, and objective–rational), and an adaptation of Holt's system, which provides scores of the degree of regression and of the defense effectiveness. Results were consistent with those previously found. Reflective CT was associated with good defenses, and reactive CT was associated with a poorer quality of defenses. Objective–rational CT was not related to defense effectiveness. Experts also appeared to be more reactive than were novices and to express a greater level of regression.