This article proposes that for humour to be effective as a therapeutic intervention in psychoanalytic treatment, it must approximate an affect releasing and growth promoting interpretation. The presence and psychological significance of surprise, which is common to both wit and 'good' interpretations, is taken as the point of departure and the importance of the psychoanalyst's spontaneity is discussed in relation to this.

The author explains his opposition to the use of humour as a contrived communication: as a parameter aimed at reviving a dying analysis or as an attempt to resolve an impasse. In particular the dangers of countertransference acting out and the provision of transference gratifications are explored. Equal emphasis is placed on recommending that the analysis of a patient's humorous responses should not be neglected and how an analyst without humour may negatively affect the treatment process.