This discussion follows from a key premise that there is potential clinical value in identifying and directly treating the cognitive abnormalities which for over a century have been seen as the hallmarks of schizophrenia. This is both an old and a relatively new idea. It is old, in that the psychodynamic approaches to psychotherapy, especially ego psychology, have long sought to modify cognition in schizophrenia, if unsystematically and indirectly. It is new, in that only in the last decade have both the contemporary technologies of the information processing laboratory and of cognitive-behaviour modification been applied to the problem of treatment. The failures of psychodynamic treatment approaches with schizophrenic patients have perhaps discouraged widespread experimentation with cognitive models in clinical assessment and treatment. This is unfortunate, as contemporary clinical procedures and laboratory technologies are hardly comparable to psychodynamic methods. In the past decade findings have accumulated, in the form of case reports, single-subject experiments and treatment trials, which support the potential usefulness of assessing and directly treating schizophrenic patients' cognitive abnormalities (reviewed by Spaulding et al, 1986).