Transcripts of therapy sessions, which capture the verbatim interactions between patient and analyst, are of limited value for process research (and therapy supervision) because they do not take into account the 'vocal channel' through which affective information is transmitted. Auditory reproduction and audiovisual aids eliminate this defect, but confront the investigator with a surfeit of information, so that selection criteria must be worked out. Data classification and coding then become major problems. The Chapple interaction chronograph paves the way for research, utilizing computers which render data storage, coding and analysis independent of the investigator. The process analyser connected to the language-analysis apparatus in Giessen permits the seismographic investigation of micro-phenomena of linguistic behaviour. It functions as a sort of perceptual amplifier, which enables the therapist to be confronted with his own style of communication vis-à-vis the patient, as well as alerting him to changes