"Cognitive inhibition" is a concept that has found a firm place in the interpretation of performance by normal subjects on tasks involving adherence to a plan and suppression of incorrect responses to distractors. The presence of "negative priming" is the classical indicator of cognitive inhibition. Negative priming occurs when, in a sequence of stimuli each of which is composed of a target and a distractor, the distractor of the first stimulus becomes the target of the second stimulus: reaction time to the second stimulus is slowed because of the inhibition applied to the distractor of the first stimulus. The concept has been extended to the interpretation of pathological behavior and symptoms. Pathological subjects have been found to show deficient negative priming. Thus, negative ideation in depression as well as intrusive paranoid associations in schizophrenia have been related to a deficit in the capacity to inhibit inappropriate representations. In this paper, we briefly review some of the experimental evidence from normal subjects that has contributed to the acceptance of cognitive inhibition as a key process in the control of normal cognition, as well as more recent evidence that has led to a revision of the concept. Negative priming in normal subjects has been found to be dependent upon characteristics of the experimental situation as perceived by the subject. In particular, priming is observed when the subject anticipates difficulty in determining the response and proceeds with caution. Thus, inhibition is not an automatic "brake" applied to irrelevant material, but rather the product of strategic considerations within the experimental situation. This revision of the cognitive inhibition hypothesis leads to a re-interpretation of the apparently deficient cognitive inhibition seen in depressed or schizophrenic subjects. According to this more recent interpretation, deficient cognitive inhibition in pathological subjects can be seen as a less adaptive strategic adjustment to the task. The pathology seems to touch higher-level executive functions rather than a deficient inhibitory "brake". In depressed subjects, abnormal performance in selective attention tasks could be related to the underlying pathology in two ways: some depressed subjects show a marked lack of energy and a psychomotor slowing: these subjects do not exhibit normal negative priming, probably because of a reduction of cognitive resources. Other depressed subjects show abnormal performance as reflected by negative priming greater than normal: this result could be related to an exaggerated tendency to verify a correct response. Schizophrenic subjects show a lack of negative priming that seems most plausibly to be related to an ineffectual integration of the experimental instructions concerning both speed and accuracy in the response. This re-interpretation of the cognitive deficiency in pathological patients provides a better fit with recent experimental results from normal subjects, and with cognitive deficits measured in pathological subjects.