Abstract
This paper critically examines the recent impact of cognitivism upon the field of clinical psychology, and concludes that certain criteria of scientific adequacy have been compromised. The argument is developed that the introduction of information processing constructs to theoretical models of psychopathology has made a potentially valuable contribution to the discipline; but that the acceptance of mental events as dependent measures has severely undermined the scientific credibility of experimental attempts to evaluate such models. It is proposed that future progress will require the adoption of a particular methodological constraint. Specifically, it is suggested that cognitive explanations of psychopathology can only be tested adequately by evaluating the validity of the behavioural predictions that they generate. Using examples of recent research that has investigated the cognitive characteristics of vulnerability to anxiety and depression, an attempt is made to demonstrate that adherence to this proposed constraint not only is possible, but actually provides a far greater degree of understanding than could be attained through the use of alternative methodologies. It is postulated that the future scientific status of clinical psychology may depend upon our collective response to the issues that are raised in this paper.