(Section continued from previous page)
Certain categories of aggression in previous classifications are not included here. Predatory aggression and quiet attack on another species are not included because they are not intraspecific aggression, and because they appear to be under the control of a different motivational system. Moyer (1968) includes a category of instrumental aggression, for which I do not know of any examples in the ethological or neuroscience literature. Wilson (1975) includes categories of sexual aggression, parental disciplinary aggression, weaning aggression, and moralistic aggression, which have not been extensively studied, or in some cases observed, in the species under consideration here.
Because the present classification is not based solely upon a set of logical distinctions and categories, such as response or stimulus categories, its conclusions may seem at times to be surprising, illogical, or, at least, not parsimonious. In particular, it is surprising that there should be so much overlap between the two motivational systems of defense and submission. In fact, in an earlier paper (Lehman & Adams 1977) I suggested that defense and submission should logically involve separate and non-overlapping patterns. Data from brain lesions and stimulation require a different conclusion, however. Freezing, fleeing, and upright postures, along with a number of other motor patterns, are exhibited by animals both before and after lesions of the medial hypothalamus that shift the animal from submission to defense. Only the lunge-and-bite attack appears for the first time after such lesions. Although we may wish to assume that the neural mechanisms of behavior are ultimately "logical," the logic may not be immediately obvious and may require a functional and evolutionary analysis that goes beyond the scope of the present review.
Although the emphasis has been placed here upon genetically-determined, "hard-wired" neural circuitry of social behavior, there is no doubt that important aspects of the functioning of this circuitry, and perhaps part of the circuitry itself, change as a function of the animal's experience. Some of these changes are due to the actions of hormones. Others are more properly considered as examples of learning and memory. I have argued elsewhere in detail (Adams, in press) that in muroid rodents there are seven points at which learning plays an important role in the motivational systems of offense, defense, and submission. 1) Previously-neutral stimuli may become conditional motivating stimuli for defense and submission if they are paired with unconditional motivating stimuli. 2) The directing stimuli for the routes followed during approach and escape locomotion in a familiar place may be learned by experience. 3) The stimuli, largely olfactory, by which an animal determines if its opponent is a consociate, are partly learned and may, in some cases, be considered as a type of imprinting. 4) The motivating stimuli of neophobia are of necessity dependent upon learning. 5) The ability of sudden movement and noise to serve as motivating stimuli for defense depends upon early experiential factors (Clark & Galef 1977). 6) The ability of restraint and dorsal tactile stimuli to serve as motivating stimuli may be lessened by experience with handling. 7) Visual stimuli can be conditioned to serve as releasing stimuli for the upright posture in the rat (Kanki & Adams 1978).
In keeping with the purpose of this paper, which is to help focus and direct research questions concerning brain mechanisms of aggression, the remainder of the discussion will concentrate on the types of experiments that might be performed to test and improve this hypothetical model.
One way in which the present model is undoubtedly oversimplified is its unidirectionality of neural projections -that is, information flow. As shown, information is received from stimuli, processed, and conveyed to motivational mechanisms that activate motor patterning mechanisms that organize motor patterns. No reverse flow is shown. Attentional mechanisms that might alter perceptual systems are not included. In brain-stimulation experiments concerning the interspecific attack of a cat upon a rat, it has been shown that perceptual systems are modified by the activation of motivational systems (Flynn et al. 1971). Although the disruption of intraspecific aggression following brain lesions has often been attributed to attentional losses (e.g. Sprague et al. 1961), a definitive study of the question would require brain stimulation experiments similar to those done on interspecific attack. It seems likely, from the fact that most neuro-anatomical connections are reciprocal, and from the findings on interspecific attack, that such reverse connections will be found when the appropriate experiments are done for offense, defense, and submission.
The location, and indeed the existence, of the hypothesized offense motivational mechanism needs to be determined. Probably the most efficient experimental strategy would involve reliable elicitation of offense by electrical stimulation of the lateral hypothalamus or by isolation-induced offense in the rat, and the systematic placement of midbrain lesions in order to find a location at which all the motor patterns for offense are abolished. Then lesions of that region would be tried on animals who show all the various types of offense - that is, inter-male fighting, offense by females (in the hamster and gerbil), and competitive fighting - to make sure that the lesions abolish all these types of offense.
(Section continued on next page)