Research on temporal lobe epilepsy (TLE) and aggression is reviewed in order to learn why it has not contributed more to our understanding of the neural basis of aggression. The research problem can be resolved into two questions: Are temporal lobe epileptics more aggressive? What inferences can be drawn about the factors underlying aggression in TLE? Difficulties in finding suitable operational definitions for TLE and for aggression have, to date, received insufficient attention. In published studies, sample bias and lack of regard for the validity and reliability of behavioral assessments prevent our deciding whether an association exists between interictal aggressive behavior and TLE. Even if this association were demonstrated, one could not draw direct neurobehavioral inferences from it, since a variety of social and psychological, as well as neurophysiological variables could contribute to the association. These would have to be controlled. Methodological refinements could lead to more satisfactory answers to the first question, but the second is considerably more difficult. Some sources of complexity and apparent contradictions in the experimental literature on the neurology of aggression are discussed. They consist of technical problems, the complexity of the neural substrate of aggression and the influence of environmental cues and learning. The combined use of electro- physiological and interview techniques has yielded some interesting results with TLE patients.