Some alternative levels of analysis
Some alternative levels of analysis. The questions raised in my target article can be analyzed on many levels. I have chosen to use a neuroethological approach in this case. Other levels of analysis are also appropriate, however, as the commentators point out. These include a strictly ethological approach, an analysis of the role of learning, an analysis of the role of hormones, a pharmacological approach, and a regulatory systems approach.
Although I have used ethological concepts in the present paper, I have not presented much ethological data. Yet any complete description of behavioral systems must conform to the behavior of animals observed under natural conditions, as correctly emphasized by Koolhaas and Wiepkema. For this reason, I have recently reviewed the literature on aggressive behavior in naturalistic as well as laboratory settings for all species of muroid rodents (Adams, submitted for publication). The present neural analysis conforms to the results that derived from the comparative study of naturalistic behavior.
An ethological approach is important, among other reasons, for its insistence upon the accurate description of motor patterns of behavior. As the Blanchards, Waldbillig, and Rodgers point out, failure to provide such accurate descriptions has plagued much of the previous literature, especially in psychology and physiology. I propose that workers in the field of muroid rodent aggression should base their descriptions of behavior upon those of Grant and Mackintosh (1963) and those working with cats should use the descriptions of Leyhausen (1956, 1979a).
An analysis of the role of learning in aggressive behavior is also important, as many commentators point out. Andrew invokes learning to account for stability of behavior patterns; Baenninger suggests that aggressive behavior must be even more "plastic" than feeding; Isaacson asks regarding offense, defense, and submission, "Are they really hard-wired?"; Karli emphasizes learned changes in mouse-killing by rats; Koolhaas and Wiepkema emphasize the "flexibility" of behavior; Laborit believes that I do not give sufficient emphasis to the role of memory; and Panksepp speaks of submission as a "habit" based on "contiguous reinforcement." But, contrary to the impression one might gain from this, the literature on learning effects upon aggression, at least in cats and rodents, is remarkably sparse. In the target article, I list the points in the motivational systems where 1 think there are ontogenetic changes as a function of experience that might be called learning. This brief list is considered in greater detail in another recent. review (Adams, in press). In fact, I would challenge anyone to be more specific than I have been here as to how learning affects aggression and where its effects take place in the nervous system. Learning does not take place in a vacuum or in a Skinnerian "black box"; rather, it represents changes in the functioning of neural circuitry. Only when we have some notion of the neural circuitry can we begin to pin down these effects specifically.
The neural circuitry of aggression is greatly affected by hormones, as well as by learning, as pointed out by Brain, Gandelman, Koolhaas, and Laborit. Hormonal effects are even more profound than learning, in my opinion, and are so complex (See Brain and Gandelman commentaries) that I have chosen to review them in another paper, submitted for publication, entitled "Hormone influences on motivational systems of social behavior in muroid rodents and their significance for reproductive states." As in the case of learning, I submit that these hormonal effects may be analyzed most effectively in terms of their action upon specific types of neurons of the circuitry outlined in the target article.
The pharmacology of aggression has a rich literature, but because I have no experimental experience with it, I neglected it in the target article. As noted by the Blanchards, Decsi & Nagy, Eichelman, Laborit, Miczek, and Senault, an analysis of motivational systems could be strengthened greatly by consideration of these data. As in the case of learning and hormonal effects, ideally one would analyze the pharmacology of aggression in terms of actions at specific synapses and sets of neurons of the circuitry outlined here. Senault doubts the usefulness of this approach, however, and points to apomorphine-induced fighting as an example. In my opinion, the apomorphine-induced fighting in rats that he mentions is simply defense-activated upright posture and boxing. The relevant neural circuitry, as well as the motor pattern, resembles that of defense, and it is quite fortuitous, that the involvement of the globus pallidus parallels the involvement of that structure in the penile displays of squirrel monkeys. As in defense, apomorphine-induced fighting is not increased by septal, ventromedial, or olfactory lesions. These lesions can shift the behavior of an animal from submission to defense, but if an animal is already showing defense, no further change should he expected. On the other hand, lesions of the lateral hypothalamus and amygdala may decrease defense, presumably by interrupting the forebrain defense pathway. The involvement of the putamen, globus pallidus, and substantia nigra in apomorphine-induced fighting is its only unique feature, at least among those listed by Senault.
Another important approach to motivational systems is to consider them as regulatory and to analyze the temporal sequencing of their motor patterns. Wiepkema mentions several recent reviews that emphasize such an approach. I have made some preliminary observations relevant to the question of aggression as a regulatory system in a recent empirical paper (Lehman and Adams 1977). For example, we found that while most behavioral sequences can occur in either direction, that is, they are symmetrical, the bite-and-kick attack usually terminates an offense sequence, suggesting that it is a consummatory response. Unfortunately, there are insufficient data from any one mammalian species to make such an approach feasible at the present time, and before we can analyze regulatory systems across species, we must understand one better in a single species. Any such analysis also ought to account for the intriguing relation between aggressive behavior induced by brain stimulation and the phenomena of negative and positive rewarding characteristics of the stimulation, as discussed by Karli and Eichelman.
(End of section)