Brain Mechanisms for Offense, Defense, and Submission
Comments by R. J. Andrew
School of Biological Sciences, University of Sussex, Brighton BN1 9OG, England
Page 24

Title/Abstract page

Pages 1 - 2

Defense: motivational mechanism
Page 3

Defense: motivating stimuli
Pages 4 - 5

Defense: motor patterning mechanism
Page 6

Defense: releasing & directing stimuli
Page 7

Pages 8 - 9 - 10

Pages 11 - 12

Primitive mammals & primates
Page 13

Pages 14 - 15 - 16

Figure 1: Defense
Page 17

Figure 2: Submission
Page 18

Figure 3: Interaction
Page 19

Figure 4: Offense
Page 20

Figure 5: Composite
Page 21

Open Peer Commentary
Pages 22-49

Author's Response:
motivational systems

Pages 50 - 51 - 52

Author's Response:
alternative analyses

Page 53

Author's Response:
specific questions

Pages 54 - 55 - 56

Author's Response:

Page 57

References A-E
Page 58

References F-M
Page 59

References N-Z
Page 60


Page 61

Avian data on aggression. In the course of his article Adams effectively re-interprets some data of mine and DeLanerolle (Andrew 1973; Andrew and Delanerolle 1974) He argues that lesions in the central gray of the chick midbrain "depress defense vocalizations, defensive pecking and freezing in a novel environment." In fact, what we described was rather different and (we felt) rather more interesting. Our lesions destroyed the intercollicular area, whose mammalian homologue is probably the subcollicular area. Both could indeed be regarded as including lateral extensions of the midbrain central gray, but surprisingly, lesions (in work I have not yet published), confined to the midline central gray of the chick did not produce the characteristic effects such as muting, which follow from intercollicular lesions. It is interesting therefore that Adams notes that "part of the tegmentum adjacent to the lateral borders of the central gray" must be destroyed in order to produce the full syndrome that he observes in the rat following complete lesions of the central gray. In the chick (but not of course necessarily in mammals) it is enough to destroy structures on the lateral borders of the central gray to produce what may be a comparable syndrome.

However, it is the behavioural changes resulting from intercollicular lesions in the chick that I would like to discuss more fully. One initial interest was in the muting that results. All calls are lost, and not simply the peeps given by cold, frightened, or frustrated chicks; these latter might perhaps be termed defense calls, in that they are often given when fleeing and may precede or follow (but very rarely accompany) defensive pecking.

We then discovered, to our surprise, that all of the phases of behaviour that are normally accompanied by calling had also disappeared. It is necessary to use a word like "phase" because, for example, feeding was normal, except that the extended session of excited feeding (normally accompanied by twitter calls, which intact chicks give on finding a food source) was absent. Instead, the first bout was exactly the same as subsequent ones. Equally, small beads and similar objects that normally are attractive to chicks and evoke repeated pecking with twitters were ignored. It is important to note that such pecking differs from defensive pecking both in its form and in the stimuli that release it. Defensive pecking is typically delivered from a crouch with outspread wings, after sustained fixation in which the animal remains still; often it has backed away from the stimulus just before such fixation. Such a peck can be evoked by large stimuli as well as small and, if so, is aimed at the centre of the large stimulus rather than at some surface marking. A very similar difference holds for the pecking at large objects that is facilitated by testosterone; here the hormone only affects pecks at large stimuli (Andrew and Clifton, unpublished), We suspect that the same lower-level mechanism may be involved in both testosterone-facilitated pecking (which might be "offense" in Adams's terminology) and defensive pecking.

Intercollicular lesions do, as part of their global effects, also affect behaviour evoked by frightening stimuli (eg fleeing, hiding, and visual scanning in a novel environment are almost or quite abolished). However, we maintained (and still believe) that this is only one aspect of a more basic change that we characterised as a loss of the ability to respond to visual (and probably other) stimuli as if they were conspicuous or highly valent; one crucial aspect of such response is the appearance of "emotional" behaviour.

More extensive speculation on this hypothesis will be found in Andrew (1975); it would be out of place here, where indeed it could be argued that findings in chicks, which apparently disagree with hypotheses derived from mammalian data, may reasonably be ignored.

I would like to suggest, nevertheless, that some deductions may be drawn that are pertinent to the present discussion. Firstly, I would argue that ethologists have been much to blame in the rigidity with which they have clung to the categories of behaviour that were adopted in the infancy of the discipline. This rigidity has often resulted in the choice of a limited range of experimental situations, which of necessity confirm the categories. Thus (to take a real example) if a male fish is tested by the presentation of a sexually receptive female, or a territorial rival, or a place to lay eggs, it is not surprising that analysis of the types of behaviour shown should suggest that the components of behaviour can be classified as sexual, aggressive, or parental.

The same sort of constraint may well affect the study of defensive, submissive, and offensive behaviour. In the cat emotional responses, which are often listed as being the crucial indices distinguishing "affective" attack (which would be classified as defensive by Adams) from confident attack, probably also occur in a much wider range of situations. These deserve formal investigation. Frustration is one obvious example: in many primates the same set of calls is given in such situations (and in a number of others) as occur in social interactions involving defense (Andrew 1962).

There may be, in addition, a more difficult problem to solve, if we are properly to disentangle the causal mechanisms of affective behaviour. Learning may lead to the use in adults of stabilised patterns of behaviour as effective responses in social interaction. Defense and submission are obvious candidates. The overlap between the "motivational systems" for defense and submission noted by Adams as surprising might, on this argument, reflect the fact that these are not the best ways of categorising behaviour at (say) the midbrain level, but they do correspond to crystallised patterns used by the whole animal in very clear and specific situations.

I believe that a very important role of studies like Adams's is to make us dissatisfied with present causal models of behaviour. Clearly Adams feels this too and has taken some valuable and concrete steps in this direction. I am arguing only that we should push the process much further.

previous page
home page
next page