ABSTRACT
This paper gains from being read together with ‘Withdrawal and Regression’, Winnicott’s other main paper on the topic (Winnicott, 1954b). Both describe a situation that may happen in certain analyses, what he calls ‘a regression to dependence’. In this situation he argues for an adaptation of classical technique and a prioritization of the importance of the setting in preference to the normal interpreting tools of analytic work. In certain cases of severe psychopathology this regression to dependence becomes a feature of the analysis. Both states are different from more transient states of regression that analysts recognize in their more intact patients.
