Pure cognition and hence pure cognitive dysfunction might be expected to have no direct relation to any specific emotion. Changes in cognitive processing will change the assessment of stimuli and thus could change emotional responses nonspecifically. However, neurology suggests a more direct relation between at least some aspects of cognition and emotion. The limbic system in general and the hippocampus in particular have been suggested at different times to be crucial for both memory and emotion. Even recently, O'Keefe and Nadel (The hippocampus as a cognitive map, Oxford University Press, 1978) proposed that the hippocampus is a spatial, or cognitive, map, while Gray (The neuropsychology of anxiety: An enquiry into the functions of the septo-hippocampal system. Oxford University Press, 1982) proposed that it is central to anxiety. This apparent incongruity can be resolved by combining recent developments in the psychology of anxiety (which emphasise changed processing biases), recent extensions of Gray's theory (which bring it closer to cognitive views), and recent theories of the role of the hippocampus in memory (which see it as controlling rather than storing information). This paper proposes that at least some instances of clinical anxiety could result from hyperactivity of the septo-hippocampal system, which would produce cognitive dysfunction in the form of increased negative association of stimuli with a consequential increase in anxiety when the stimuli are subsequently presented.