Aggression: the biology of war and peace

  • Details
  • Transcript
  • Audio
  • Downloads
  • Extra Reading

Aggression lies at the heart of biological competition but its consequence in both human and animal societies can be self-destructive and a threat to social stability.

Do other social species declare war on each other and can we learn anything from them about how to live together in peace?

Download Transcript


The Biology of War and Peace


Professor Keith Kendrick


The rate of violent crime in London has risen by an alarming 37% in the last year and is the highest in the UK. Increasing violence in human societies is often cited as being a major cause of concern and the priority for finding a cure ranks along-side, or even above, that for cardiovascular disease, cancer or AIDS. We still fight wars, even though by doing so we increase the risk that weapons of mass destruction will continue to be developed and that sooner or later this may lead to a deliberate or accidental compromise of our planet’s ability to support existing forms of life – including humans.

In the USA there are around 3 million violent crimes each year and a recent review paper has reported that, using self-reports, around one in four women claim they have experienced rape or serious sexual assault and many of these may involve physical violence beyond the sexual assault itself. Obviously the number of reported rapes and sexual assaults are far lower than this and some caution has to be applied to interpreting self-report data. However, an important consideration is that in around half of cases the rapist and/or the victim will have drunk significant amounts of alcohol (Abbey et al 2004). This link with alcohol consumption is also seen with other types of violent crime.

Why are we so aggressive and can we do anything about it?

In this lecture I will consider answers to the following questions:

    What is aggression and why has it evolved?
    Is aggression an inevitable trait in all species?
    How is aggression controlled by the brain?
    What hormonal, biochemical and genetic factors may contribute to excessive aggression?
    What effects does experience have on levels of aggression?
    Why do some species engage in group-based aggression in order to kill neighbours?
    Can we learn anything from other species about how to live more peacefully?


What is aggression and why has it evolved?

This is not quite as simple a question to answer as is might appear. True, it is easy to define aggression as any overt behaviour intended to inflict damage on another individual. However, even a cursory glance at the animal kingdom will lead you immediately to distinguish between predatory and affective aggression. As is suggested by this distinction, one would expect predatory aggression to be devoid of emotional experience and consequence. An animal that has evolved to kill members of other species to survive must be able to do so in as automated and unemotional way as possible. Indeed, as we will see, predators seem to experience few, if any of the hormonal or brain activity changes associated with some form of emotional response when engaged in aggression towards prey. On the other hand, affective aggression, as the term suggests, involves emotional and physical changes both before, during and after it is exhibited and is primarily only associated with aggression directed at other members of your own species or during defence against predators from other species.

As humans we are also capable of affective and predatory aggression with the former tending to be more reactive and the latter more proactive and planned (see review by McEllistrem, 2004). In humans predatory aggression is no longer, of course, primarily about putting food on the table but refers to premeditated acts of aggression towards other people or property which tend to have been reinforced by learning. It is well recognised that pathological killers often seem to have lost the ability to experience any negative emotional response to killing another individual and may actually even experience pleasure from the act. To all intents and purposes therefore they have become virtual full time predators! It is notable that one of the more robust observations concerning habitual criminals, murderers and psychopaths is that they are physiologically hypo aroused with lower heart rates, skin conductance and slower frequency brain waves (EEG). One hypothesis that has been put forward is that this chronic state of under arousal is actually aversive and that criminality is simply one successful way of trying to get a buzz and increase arousal levels. Of course, turning to a life of crime is not the only way of achieving such a buzz and extreme sports activities such as bungee jumping, hang gliding and white-water rafting spring to mind. Having a chronic state of under arousal even seems to have a pay off in some professions. Apparently bomb disposal experts who are most frequently decorated for fearlessness and bravery have the lowest heart rates – presumably this helps reduce the emotional stress involved. Whether bombers also have the same condition of under arousal is not known.

So why have most species developed an affective aggression response? There would seem to be a number of answers to this. From a purely defensive point of view the generation of an associated affective response will mobilise stress hormones such as cortisol and adrenalin which both prepare the muscles for action and help protect against injury and even debilitating pain. These changes also impact on brain areas controlling emotional responses, such as fear, attention and memory as well as for aggression. In combination these changes help individuals make a rapid and effective response to threat, enhance their level of vigilance against any subsequent attack and promote formation of strong memories for the events that led up to an attack and how successful the response made was at countering it. However, it must be emphasised that another key adaptive advantage to adopting an “affective” aggression defensive response is that the strong emotional signals that you give out (vocalisations and facial and body threat postures) may be sufficient to put your attacker off having a go at you at all!

From an offensive viewpoint there are the same advantages of an affective response as I have outlined for a defensive situation. Not only do you get pumped up for action but your threatening emotional signals may be enough to kow-tow your selected opponent into submission without a fight. With offensive aggression there is usually far more opportunity for a cognitive contribution where the relative risks of attacking another individual can be weighed up. Indeed, achieving the goal of subjugating another individual may simply involve the use of dominance displays as opposed to direct physical violence. Social dominance and the propensity with which individuals are prepared to take risks in responding to environmental and social challenges – termed “impulsivity” – are key components of an aggressive personality.

It is important to emphasise that with social hierarchies it is often not aggression per se that establishes the pecking order as much as the perception of an individual’s likely level of aggressive response when threatened. It is this perception, and the alliances that result from it, that establish stable dominance hierarchies. In a stable hierarchy it is almost possible to be dominant without exhibiting much overt aggression.

We are all aware of what a destructive trait aggression can be, especially when it involves one group of individuals deliberately setting out to injure and subjugate, or even annihilate, another group – an act that we simply call “making war”. However, while in all contexts acts of real or threatened aggression are a key survival strategy for any species, even a cursory look at other species will tell you that actual physical violence between individuals is relatively rare although aggressive threat displays and the use of submissive gestures are far more common. It is these latter behaviours which serve to keep the peace within animal communities. Indeed, it is easy to see why any adaptive strategy which involves resorting constantly to physical violence will be potentially disastrous for survival. In most instances serious physical combat between evenly matched individuals will result in injury to both winner and loser and this will then affect the physical, and therefore reproductive, fitness of both.

Ultimately, of course, the purpose of affective aggression is to predominate over, or possess someone or something. Its use is therefore to obtain and control key resources that have particular survival or reward value. For the majority of animal species the key resources are access to food, water, mates and a protected environment. For humans we have vastly extended the number of resources which we may be prepared to exhibit aggressive behaviour to gain access too – even down to trivial things such as access to the television remote control! This is not to say that aggression is the only way we can obtain resources and rewards since we also use negotiation, cheating and deceit.

The philosopher Thomas Hobbes in “Leviathan” (1651) aptly summarises the main causes of interpersonal violence in humans:

“So that in the nature of man, we find three principal causes of quarrel; Firstly, competition; secondly, diffidence; thirdly, glory. The first maketh men invade for gain; the second, for safety; and the third, for reputation. The first use violence, to make themselves masters of other men's persons, wives, children, and cattle; the second, to defend them; the third, for trifles, as a word, a smile, a different opinion, and any other sign of undervalue, either direct in their persons or by reflection in their kindred, their friends, their nation, their profession, or their name.”

From the discussion above it should have become clear that all forms of aggression have evolved for the twin purposes of survival and reproduction whether it be used for obtaining food, access to a mate or protection for yourself or your family. However a key question is whether the fact that it tends to only be used as a last resort by most species means that it can be dispensed with altogether if more peaceful alternatives become available?


Is aggression an inevitable trait in all species?

There are two opposing views of aggression.

Aggression, like other motivated behaviours such as eating, drinking and sex, is an inherited spontaneous tendency – i.e. it is in our “nature” and therefore unavoidable without considerable genetic adaptation.

(This view has received support as a whole, or in part, by a wide range of individuals including Hobbes, Freud, Lorenz and, to some extent, Pinker)

Aggression is primarily the result of experience and external factors – i.e. it is all down to nurture.

(This is very much the popular “wish” view and reflects the kinds of approaches adopted by behaviourists such as Watson, child psychologists such as Bandura and a whole host of sociologists and psychologists culminating in the UNESCO “Seville Statement” produced in the 1980s:

The Seville Statement: It is scientifically incorrect to say:

We have inherited a tendency to make war from our animal ancestors

War or any other violent behaviour is genetically programmed

In the course of human evolution there has been a selection for aggressive behaviour more than other kinds of behaviour.

Humans have a “violent brain”.

War is caused by “instinct” or any single motivation.

As with most different behaviours the expression of aggression in any particular individual is almost certainly a combination of influences from both nature and nurture, although the idea that aggression is a primary drive akin to feeding, drinking and sex carries with it the inescapable conclusion that it is a behavioural trait that cannot be eradicated, only controlled. I will consider the view that it is all down to learning in a later section and will confine myself here to a brief consideration of the idea that aggression is an inevitable feature of our nature.

Freud was certainly of the opinion that aggression was a primary human drive in the same way as sex. The famous ethologist, Konrad Lorenz, developed an influential model that attempted to explain motivated behaviours, including aggression. This he called the “Hydraulic model” and is rather similar in concept to a pressure cooker. The idea was that feeding, drinking, sex and aggression were ultimately fixed action patterns of behaviour and that a combination of internal and external signals created mounting pressure within the system until they reached a critical level and blew a release valve (the blowing of the release valve was equated to the inevitable performance of a particular motivated behaviour in response to a releasing stimulus). While this was intended to represent an all or nothing response it also incorporated the idea that the strength of the stimulus required to elicit the fixed action response was a function of the current state of motivation (i.e. how much pressure had built up in the pressure vessel!). Put simply, the greater the head of steam that has already been built up the smaller the change (i.e. the external stimulus) that was required to trigger its release.

A whole range of experimental observations with many different animal species has provided confirmation for the general idea that the display of motivated behaviours in response to appropriate stimuli is inevitable and that intensity thresholds for triggering them vary as a function of recent experience. Of course the animal does not simply sit around waiting for the right stimulus to come along but will become more and more active in seeking out the appropriate stimulus as the drive to do so increases.

If Lorenz is right about aggression then this makes for some disturbing conclusions:

Aggression is inevitable – accumulating energy must find an outlet.
Humans and other animals will actively look for fights
After a fight has occurred individuals will become less aggressive
Even animals reared in isolation will show aggressive behaviour

While Lorenz’s model has been modified by Deutsch to incorporate aspects such as feedback from the environmental effects of aggression (i.e. you need to know to stop when aggression has achieved its aim) the above predictions are largely supported by observations of the behaviour of many different species. If it is correct then the obvious take home message for humans is that we must have socially acceptable outlets for aggression made available to us otherwise we will have little chance of controlling interpersonal violence and warfare.

This view that aggression must inevitably be expressed by humans and needs control has been voiced by many anthropologists, historians, psychologists and philosophers:

“Hereby it is manifest that during the time men live without a common power to keep them all in awe, they are in that condition which is called war; and such a war as is of every man against every man.” The result of this being: “…. continual fear, and danger of violent death; and the life of man, solitary, poor, nasty, brutish, and short.” (Thomas Hobbes – Leviathan)

“The story of the human race is war. Except for brief and precarious interludes there has never been peace in the world; and long before history began murderous strife was universal and unending”

(Winston Churchill).

Sex differences in aggression


The opinion that a world populated only by women would be a much more peaceful place has been expressed by many. I will consider the basis of brain and hormonal contributions to reduced levels of aggression in females compared with males in a moment. However, it is important to consider from an evolutionary point of view why in the vast majority of species both the incidence of physical aggression and injury and death as a result of it is very much a male domain.

Males in most species are both physically stronger than females and can reproduce their genes best by mating with as many females as possible. This simple scenario means that females cannot normally physically dominate over males, have only a limited potential for choosing a mate and will often live together with a number of different members of their own sex. Their best reproductive strategy is to nurture and defend their own offspring, to reproduce with males who can best protect them and to keep close social ties with other female counterparts. Female aggression therefore tends to be restricted to protecting young and competing for resources against other females – whether males, food or the safest spot to sleep! In this sense it has been argued that female aggression tends to focus on protection of their young and acquisition of resources to aid their own survival and that of their young. It has less to do with dominance which is far more important for males (see Campbell, 1999). Since the majority of male aggression is centred around establishing dominance and control over access to females this provides an immediate easy explanation for why males have evolved to be more physically aggressive than females.

A further important factor is the perceived risk of engaging in physical violence. For most mammalian species where offspring are provided with maternal care, females appear to be significantly more fearful of potential threats than males. This is not particularly surprising since their offspring are highly vulnerable and mothers are not only concerned with self-preservation. An important conclusion from this is that engaging in physical violence will normally carry a much greater perceived risk for a female than for a male. This provides yet another reason why females should have evolved a much higher threshold for resorting to physical violence than males. However, this does not exclude the possibility that females resort to more indirect forms of aggression and many would argue that human female skills in using social psychology to inflict suffering on others are considerably more advanced than those employed by males!

In the absence of males, or in cultures where male advantages in physical strength have been controlled in some way, a prediction could be that females might start to become more aggressive and develop more of an interest in social dominance. In many countries in the developed world there are some signs that this may be happening, as far as humans are concerned, with female involvement in violent crime seemingly increasing. Also with female equality becoming established in many cultures there have also been trends towards greater female interest in social dominance issues and recent reports suggest that female bosses are actually more prone to bullying their employees than their male counterparts are! This dominance change has parallels in Bonobo apes, where females are the dominant sex (even though not physically stronger than males) and are also seemingly more involved in social dominance issues. However, as I will discuss later, their strategy for achieving this does not normally involve physical aggression.

So all in all some caution may need to be applied to any conclusion that females are an inherently peaceful sex. This may owe more to social structure and political expediency than to biology.

In any event even if human females are less prone to the competitive urges that can completely dominate male behaviour they may also be responsible for propagating the numbers of dominant and aggressive males. This is one of the arguments put forward by Richard Wrangham in his book entitled: “Demonic Males: Apes and the Origins of Human Violence”. In my first lecture “Sex, hormones and animal passion: making a virtue out of necessity” (September 2002) I described experiments by Dave Perrett ’s group in the Psychology Department at St Andrews University which showed that human females normally prefer the faces of more feminine characteristic males for the majority of their menstrual cycle (good potential fathers) but at mid cycle when then are most likely to conceive they switch to preferring the faces of more masculine, macho individuals. Similarly, other research has shown that men with voices that women rate as being the most attractive also tend to be the most masculine and the individuals concerned have higher testosterone levels.


How is aggression controlled by the brain?

Not surprisingly research into how the brain controls different forms of aggression is restricted by ethical issues. The most revealing work using animal models with potential relevance to humans has focussed on species that both exhibit affective/defensive aggression and predatory aggression in order to try to understand the extent to which they are controlled by separate systems. The most successful models have also been species which exhibit easily measurable aggressive behaviour repertoires. Not surprisingly therefore cats have figured prominently. Defensive rage in cats includes behavioural signs such as growling, meowing, yowling, hissing, arching the back and aggressive paw strikes as well as sympathetic changes such as pupillary dilation, increased heart rate and blood pressure, urination and piloerection. On the other hand predatory aggression towards a rat or a mouse is distinctive in that it involves the virtual absence of all of these signs and is all about stalking, chasing, catching and biting.

Work carried out primarily by groups in the USA (see Gregg and Siegel, 2001) has identified the brain regions involved in both defensive and predatory aggression in cats. In general work on other species, notably rats, mice and hamsters, has shown that they have similar brain control systems as has the small amount of work on monkeys and studies of pathological aggression in humans. It is important to emphasise that for tests of predatory aggression, the prey (usually rats or mice) are presented to the predator fully anaesthetised. For tests of affective aggression the usual approach is to place a strange animal in the home cage of the experimental one for a brief period and this will almost always trigger an aggressive response from the latter (i.e. territorial defence).

Research has shown that, as with other motivational behaviours such as sex and hunger, the hypothalamus at the base of the brain is a key structure for mediating both kinds of aggression. However, whereas defensive aggression is controlled by the medial part of the anterior hypothalamus, predatory attack is controlled by the lateral part. These regions send projections into the brainstem arousal and motor control centres, notably the periaqueductal gray although here too there is anatomical and functional specificity (the defensive aggression projection is to the dorsal part of this region which links with structures that are responsible for producing sympathetic arousal – changes in heart rate, blood pressure etc; the predatory aggression projection is to the ventral part which does not connect with sympathetic arousal centres and is mainly involved in motor control and sending feedback information to the hypothalamus).

Activation of the two different hypothal ami c aggression control centres also involves inputs from different regions within what is loosely called the emotional brain (limbic system). The main structures involved are the amygdala, septum and hippocampus. A considerable amount of attention has been paid to the amygdala since it is known to control fear responses and to exert a strong regulatory control over both dominance and aggression in all mammals, including monkeys and humans. Some of the earliest experiments reported by Kluver and Bucey in 1939 showed that damage to some parts of the temporal lobe (probably mainly due to the amygdala) turned aggressive dominant individuals into non-aggressive subordinate ones. Conversely stimulation of other parts of this region can trigger aggression and at least two human mass murderers have been found to have had a tumour growing within this brain region which may well have caused it to become over stimulated (Charles Whitman who killed 16 people with a rifle from the top of the University of Texas Tower in 1966 and Ulrike Meinhof one of the leaders of the terrorist Bader-Meinhof group). The rabies virus in both dogs and humans also triggers aggression and rage by attacking regions of the temporal lobe (hippocampus and amygdala) involved in their control.

The amygdala receives inputs from all of the major sensory modalities but the parts which have a key involvement in social, affective aggression (medial and cortical amygdala) are particularly strongly innervated by the parts of the brain dealing with smell information. To this end it is interesting that for a number of smell-based species, like rodents, anosmia markedly reduces aggression. With maternal sheep, once they have given birth and bonded with their own lambs, they become highly aggressive towards all strange lambs. However, if they are unable to use their sense of smell this aggression is virtually completely suppressed and they will mother any lamb even though they can use their visual and auditory senses to recognise which lamb is their own. It seems that smell signals are strongly connected to brain aggression systems and in rodents this also applies to the control of sexual behaviour. Indeed, there is a very close overlap between the pathways from the amygdala to the hypothalamus controlling affective aggression and those controlling male sexual behaviour. They are actually so close that electrical stimulation of the medial part of the anterior part of the hypothalamus in male rats can simultaneously provoke both sexual and aggressive responses. The extent to which the two systems interact normally is unknown but their close proximity may explain why, at least for males, sex and aggression are often associated in a number of mammalian species, including humans.

Returning to the issue of links between the affective and predatory aggression systems, the work with cats provides strong evidence that they are reciprocally inhibitory to one another. Stimulation of the parts of the amygdala and anterior hypothalamus which trigger affective aggression will inhibit the expression of aggression towards prey. This makes good evolutionary sense since you are unlikely to want the have the two systems activated at the same time. It also perhaps provides a basis for one system becoming dominant over the other either as a result of abnormal circuitry or learning. In the cat, at least, the predatory system seems to have a high degree of innate hardwiring since even animals that have never seen a rat or had any hunting experience will respond to electrical stimulation of their lateral hypothalamus by killing one but will not be aggressive towards another cat or any form of inanimate object. Just how this system might be functioning in the modern human brain is not known although for our hunting ancestors one can imagine it was likely to have been very important. One can, however, imagine a scenario whereby in individuals exhibiting pathological levels of violence with no apparent emotion this predator aggression system has somehow become completely predominant over the affective one.

You might be wondering at this stage whether our advanced human brain has developed any additional control systems for regulating the expression of aggression by primitive brain areas that function in a broadly similar way even in reptiles. The answer is that we have. The areas of our highly developed brain neocortex which are involved in the control of emotions and for the making all kinds of judgements, including moral ones, can exert a direct control over the subcortical amygdala, hypothal ami c, hippocampal and septal regions that mediate aggression. The major areas involved are the orbitofrontal and prefrontal cortices and the cingulate cortex. The frontal cortex has, in particular, undergone a large increase in size during the course of primate evolution and is thought to play key roles not only in moral and rational judgements by also self awareness and memory. It seems to be particularly important in the control of impulsive risk taking behaviour and aggression.

Damage to the orbitofrontal cortex in particular is associated with individuals becoming aggressive and seemingly unable to make appropriate assessments of risk. One of the most famous cases from the nineteenth century illustrating this was a railway construction manager called Phineas Gage who had the misfortune to have a steel rod damage the frontal cortex as a result of an accident involving explosives. He changed afterwards from a responsible, reliable, calm individual to a completely irresponsible, unreliable and aggressive one prone to inveterate risk-taking behaviour. There seems to be dissociation however between personal action and the ability to judge the correct course of action on a hypothetical basis. A more recent case of an individual with a benign tumour in this region showed that while his personal judgements were impulsive, indecisive and unreliable he was perfectly able to decide the most appropriate theoretical courses of action that other individuals should take when given a set of choices.

Imaging studies in humans suggest that this region of the frontal cortex and also the cingulate cortex are less strongly activated in response to emotive stimuli in individuals that are prone to aggression. Reiss et al (1998) found for example that murderers of either the “affective” or “predatory” type had increased activity in subcortical brain structures associated with aggression (such as the amygdala). Only affective aggression murderers had reduced activity in the prefrontal cortex. Another interesting observation is that in the human female brain the orbitofrontal cortex (which suppresses aggression) is proportionately larger than the amygdala (which initiates it) compared with males. Perhaps there is therefore some anatomical support for the idea that females are better able to control their aggression than males!


What hormonal, biochemical and genetic factors may contribute to excessive aggression?

So is aggression just a testosterone-based thing? This sex hormone, along with other androgens, is produced by the male testes, and to a lesser by the adrenal glands. Androgens have important organisational effects on the brain during development that are of key importance for determining whether an individual will display male as opposed to female behaviours in adulthood. Animal studies have shown consistently that it is the high levels of testosterone produced either in the late pre-natal or early post-natal stages of development and then again at puberty, which determine whether individuals will display male sexual responses and male patterns of aggression in adulthood. In general, the presence of testosterone in adulthood is also important for normal levels of male sexual behaviour and aggression to be displayed following the organisational effects of the hormone in earlier development. Another important consequence of high testosterone secretion is that it also promotes muscle growth which is clearly a major contributor to the likely success of any aggressive encounter.

Testosterone is not however an exclusively male hormone. Females also produce testosterone and other androgens both from the ovaries and the adrenal glands. However circulating levels are considerably lower. A normal man, for example, has circulating levels of around 10 nanograms/ml whereas a woman has 7-times less. If testosterone really drives aggression then this would seem to provide a clear biological basis for the difference between male and female levels.

However, the role of testosterone as the biological driver for aggression is far from clear. On the one hand, for other animal species at least, castration in adulthood does reduce levels of male aggression and this is reversed by replacing testosterone. For seasonal breeding animals such as deer and sheep male testosterone levels decrease during the non-breeding season and this is associated with reduced aggression as well as a reduced sex drive. On the other hand having normal levels of testosterone does not universally equate with high levels of physical aggression and giving additional treatment with testosterone to such individuals does not make them more aggressive. Another seemingly universal observation is that testosterone levels tend to increase immediately prior to an aggressive encounter and afterwards the victor continues to have high levels whereas the loser’s are reduced. Also, socially dominant individuals have higher concentrations of the hormone than their subordinates. The most likely cause of lower testosterone levels in subordinate individuals and those that have just lost a fight is the depressive effects that adrenal stress hormones such as cortisol have on them.

In humans (see Mazur and Booth, 1998) prisoners with a prior record of violent and aggressive crimes have significantly higher levels of testosterone than those without such a history. Rapists who are most violent during the act also have higher testosterone levels than less violent rapists or normal individuals. Individuals exhibiting high levels of antisocial violence also have high testosterone levels.

Men with high basal testosterone levels are less likely to marry and once married are more likely to divorce due to extramarital sex and aggression towards their partner. Furthermore, testosterone levels actually increase in individuals prior to and after divorce.

Testosterone levels in human males also rise when they win in competitive events, such as sport, and decline when they lose. It is not actually necessary to take part in the competition for this to happen. Dedicated male football supporters, for example, also exhibit increased testosterone levels when their team wins and decreased ones when they lose.

However, increased testosterone levels are not necessary for triggering aggression any more than they are necessary for triggering sexual responses. If the relationship was this simple then being outside a football stadium at the end of a match would be a very dangerous place to be! Also, females for example do not consistently show altered levels of the hormone prior to or following physical aggression. When predators either prepare to kill prey or are successful in doing so this does not equate with increased testosterone levels either. Finally, it has to be pointed out that for all individuals, testosterone levels have a distinct circadian rhythm and reach their highest during the early hours of the morning when we are asleep and decidedly neither aggressive nor easily aroused sexually.

If one steps back for a minute and looks more closely at the way testosterone influences the activity of the brain then evidence strongly favours its primary action as an organiser and facilitator rather than as a direct effector of behaviour. Testosterone is well known to interact with growth factors in the brain to influence the growth of new connections. In the early part of my career I was the first to show in the rat that rather than directly altering the activity of parts of the brain dealing with the control of male sexual behaviour the presence of normal basal testosterone levels actually optimised the responsiveness of the nerve cells in these regions to the sensory and arousal cues they would receive from females (Kendrick and Drewett, 1979). The major effects were however between whether testosterone was present or absent and small changes in fluctuating levels would have little importance. It seems likely that the same is true for aggression with the absence of testosterone having the most significant effects in decreasing attention to salient cues and small increases or decreases having relatively minor ones.

Brain neurochemical systems and aggression


There are a number of different neurochemical systems and associated genes that have been implicated in the control of aggression. The most established chemical correlate of high aggression in humans, and in a number of other animal species, is reduced activity in the brain serotonin system. A considerable body of evidence suggests that this brain serotonin (5-HT) system, acting primarily through 5-HT1 receptors, can inhibit aggression. However, this same system is important for the control of impulsive and risk-taking behaviour which can in turn lead to aggression. Low levels of brain serotonin are also associated with social incompetence and, as I discussed in a previous lecture (Stress, Anxiety and Depression (SAD) – October 2004) with anxiety and depression.

In humans low levels of the serotonin metabolite (5-HIAA) in the cerebrospinal fluid have been associated with aggression and other forms of antisocial behaviour, including assault, arson, murder and child beating as well as in violent forms of suicide. A recent brain imaging study in humans has also shown that individuals with a serotonin transporter gene (5-HTT) that has at least one short allele are not only more prone to anxiety and depression but show greater levels of activation in the amygdala when asked to discriminate between angry and fearful faces (Hariri et al 2002).

Research with groups of monkeys in the wild has found that those individuals with low levels of 5-HIAA were more likely to die than those with higher levels since they engaged in far more high risk, impulsive behaviours, such as threatening dominant males. On the other hand treating males with a serotonin agonist drug actually made them more dominant whereas similar treatment with an antagonist did the opposite. This makes the important point that being aggressive per se does not necessarily result in you becoming dominant. Low brain serotonin activity is also associated with poor social skills in general and we all know that in most highly social groups dominance is often more about having good interpersonal skills than commanding respect through competitiveness and aggression.

Diets which are low in tryptophan, a precursor of serotonin which acts to increase its concentrations in the brain, have been shown to increase aggression in a number of species, including humans. Lithium has also been used successfully in humans to reduce aggression and one of the actions it has is to increase the uptake of tryptophan and produce more serotonin. One might expect therefore that serotonin selective reuptake inhibitor drugs (SSRIs) such as Prozac, which increase the availability of serotonin in the brain and are used commonly for treatment of depression and anxiety, would also be effective for treating pathological aggression. However, while some studies have claimed reductions in aggression following Prozac treatment other recent reports have cited increased aggression as a side effect of the drug.

The other brain transmitter substances that are broadly inhibitory for aggression are gamma-aminobutyric acid (GABA) and the endogenous opioids, endorphins. Drugs which act on these systems, such as barbiturates, benzodiazepines and morphine, all reduce general arousal as well as aggression. Work with cats has also established that GABA is a key transmitter in mediating the reciprocal inhibition between affective and predatory aggression in centres in the hypothalamus.

Of the other key transmitters involved in the control of aggression all are stimulatory. These include the classical transmitter systems acetylcholine, noradrenaline and dop ami ne and the neuropeptides substance P, cholecystokinin and arginine vasopressin. Of these, as far as humans are concerned, tricyclic antidepressants and mono ami ne oxidase inhibitors (used for treatment of depression), which increase noradrenaline concentrations, actually tend to increase aggression whereas beta-blockers decrease it. Dop ami ne antagonists such as haloperidol have also been used relatively successfully to control pathological aggression.

Of the neuropeptides associated with the control of aggression the most interesting is probably arginine vasopressin which is present in many of the brain areas controlling aggression. Research with hamsters and some other rodents has shown the infusions of this peptide into the brain induce social aggression whereas antagonists do the opposite. Mice and rats that either lack the genes for producing vasopressin or either of its receptors also showed reduced aggression. One of the effects of vasopressin is to modulate noradrenaline and dopamine release and this may be one of the ways that it induces aggression. However, research on hamsters has also shown that treatment with serotonin reuptake inhibitors (Prozac) blocks the pro-aggression actions of vasopressin suggesting that vasopressin may also normally trigger aggression by reducing brain serotonin levels.

Vasopressin is also important for facilitating social olfactory recognition and in species such as Prairie voles which exhibit social monogamy it facilitates both pair bond formation and paternal behaviour. Its effects on aggression are dependent on the presence of testosterone and this sex hormone has important effects on the synthesis and production of vasopressin as well. So here we have a direct link between a brain peptide that has evolved both to promote social bonding and aggression in males. Relatively little has been reported about the role of this peptide in aggression in humans although its levels are raised in the brain of highly aggressive individuals. An intriguing possibility is that it might represent a link between male-male bonding and group aggression. Interestingly, this peptide does not seem to promote bonding in females, where instead another related peptide called oxytocin is involved. Indeed vasopressin has yet to be shown to have a major role in regulating aggression in females and release of oxytocin actually suppresses aggression in females. So here we have yet another biological sex difference with control of social bonding being related to increased aggression in males but to reduced aggression in females!

Alcohol, drugs and aggressive behaviour

For humans, more than half of all violent crimes are associated with the perpetrators having consumed alcohol or drugs prior to the crime. While it is clearly difficult to establish a direct causal link, laboratory studies on both humans and other animals have shown repeatedly that alcohol and many other recreational drugs increase the likelihood of violent reactions to external stimuli that would normally be ignored.

Alcohol is both a stimulant and a depressant as far as the brain is concerned but probably its main effects on increasing aggression are through reducing the activity of the inhibitory transmitter systems which normally act to control both impulsivity and aggression. However, the general blunting of perceptual and cognitive skills produced by alcohol should also be considered as making an important contribution since they make it less likely that subtle social signals will be perceived and interpreted correctly and may also result in the individuals themselves inadvertently communicating signals they do not actually wish to. These latter kinds of issues may be especially pertinent in cases of rape and sexual assault since the vast majority of these occur between individuals that know each other and may be actively dating.

Amphetamine usage has also been linked with increased aggression and this is probably because its major action within the brain is to facilitate the release of pro-aggression neurotransmitters such as noradrenaline and dopamine.


What effects does experience have on levels of aggression?

This is potentially the most controversial aspect of aggression research, particularly as it relates to humans. Much is claimed and trumpeted in social and political statements, but actually very little is proven. For example in the last 15 years or so, increased exposure to violence in the media and computer games and availability of lethal weapons have all been blamed for escalating violent crime. However, for the majority of this period (i.e. during the 1990s) rates of violent crime have actually been in decline. A recent review on the effects of exposure to violence in the media and computer games found no consistent evidence that it increases frequencies of violent crime (Savage, 2004) although many studies have shown that it does tend to promote increased aggressiveness. Similarly, in the USA the development of the frontiersman, macho male culture has been blamed for promoting male violence. However, the USA is by no means the most violent culture in the modern world and indeed a number of other male dominated cultures, such as Japan, have relatively low levels of violent crime.

Twin studies in humans have reported between a 30 and 60% concordance in the aggression levels of monozygotic twins but it is difficult to partial out the relative contributions from genetic and experiential similarities. Children raised by aggressive parents do in general appear to have an increased risk of becoming aggressive themselves. Similarly a large amount of research has provided evidence that children suffering from early physical abuse are more likely to become abusers themselves.

Cross-fostering experiments in animals involving strains of rodents characterised by high or low levels of aggression have consistently shown strong parental influences. Thus, in a recent cross-fostering experiment between aggressive monogamous California mice and less aggressive promiscuous White-footed mice, fostered offspring took on the aggressive characteristics of their parents rather than their sub-species. They also developed the same patterns of vasopressin concentrations (Bester-Meredith and Marler, 2001). As already discussed this peptide is strongly associated with aggression as well as with monogamy and paternal care.

It is probably safe to conclude at this stage that patterns of aggressive behaviour can be significantly influenced by experience but the interactions are by no means simple and we still have a very poor understanding of the most significant risk factors and how they might combine with genotype factors.

Whatever research does ultimately manage to establish, one obvious key learning factor is always going to be the level of perceived risk of punishment associated with aggressive acts. In developed countries the rule of law is generally strong and therefore perceived risk is higher and rates of violent crime are low. In less developed cultures where the rule of law is weaker the perceived risk of aggression is lower and rates of violent crime are higher. If the rule of law is further backed up by social values which discourage aggression then this will reduce violence still further since individuals in social species generally fear social hostility and isolation as much as they do incarceration.


Why do some species engage in group-based aggression in order to kill neighbours?

So what is the purpose of aggression between groups of individuals and is it a particular human characteristic? For any group members of any species to survive they need year round access to a plentiful food supply and appropriate conditions to conceive and rear young. While some animal species, and some human cultures such as hunter gatherer societies, achieve these aims by constantly moving from one place to another, this can be a high risk strategy. It can also limit sophisticated use of environmental resources which can mainly only be achieved through considerable time-investment.

Most species, and most human cultures, have therefore adopted the fixed territory approach where an area sufficient to maintain a group of individuals is colonised and utilised by them and guarded against intrusion by other groups of the same species. In general the larger the territory is that you can control the better resourced you are as a group to deal with the various hardships that the environment can throw at you. For human cultures a fixed territory also allows cultivation of the land to improve resource provision.

However, once this territorial approach is adopted your neighbours quickly become your rivals. As you increase your numbers through successful reproduction, or experience hardship from unexpected environmental challenges, you need to expand your home territory and this can often only be achieved through taking territory from a neighbouring group. Furthermore, it is important to exercise some forward planning strategy which will make it more likely that when it is necessary to take some territory from another group it will be a relatively easy task that does not cause significant losses of members of your group.

This is a recipe for being on a permanent war footing with all your neighbours! The most simple, forward planning strategy is to patrol constantly the borders of your territory in sufficient numbers with the hope of trapping a member of a neighbouring group. Killing that individual can then be achieved with minimal risk to your own group since you will have the advantage of superior numbers. With this plan you can hopefully systematically reduce the numbers of individuals in a neighbouring group so that when you need to take over part of their territory they will be powerless to resist you. An added twist to this scenario is that it is better to deplete the neighbouring group only of its defending male members and so make it possible to capture their females when you take over the territory and thereby increase your reproductive potential.

In theory, all a group of individuals would need to do is to keep together within large groups and well within their defined territory. For most social animal species this seems to be possible but where social structure and acquisition of resources require small groups to engage in wide-ranging foraging attempts, or cultivation in human farming cultures, involving travel around and/or constant utilisation of the whole territorial space then this can lead to problems. The immediate solution to this is to create no-go areas between neighbouring territories which will act to minimise border skirmishes. This works well in the beginning, but these no-go areas rapidly become sources of rich-pickings for food. The increased numbers of prey and/or lush vegetation within these no-go areas tempt members from each group to forage in them even though this may be at great personal risk. The result of this strategy is therefore to create an active war zone between the territories of neighbouring groups.

Why can’t a policy of “love thy neighbour” be adopted so that groups can live in harmony and combine forces if required? The answer to this probably comes back to the simple Darwinian principle of wanting to promote your own genes. Independent groups of many animal species tend to centre on either a single male or a group of related males with exclusive access to a group of females. In terms of promoting their own genes the last thing they would want to do is to be cooperative with another group of unrelated males since cooperation would presumably also mean allowing reproductive access to the females in the group.

Of course with human cultures the need for large scale cooperation between groups has eventually been recognised and led to the creation of Nation states. However, in most instances this can easily be viewed as a simple organisational expediency and has shifted the emphasis away from largely opportunistic encounters between unevenly matched small groups of individuals to deliberate large scale conflicts between large numbers of two highly organised and more evenly matched ones. Only humans have the cognitive skills to organise deliberate warfare on this latter kind of scale but have any other animal species have adopted the kinds of opportunistic war-based strategies outlined above?

Most animal species are territorial and will respond aggressively to any strangers that intrude. However, deliberate or accidental incursions into a rival’s space rarely result in death since there is always a home advantage and a stranger will only make a serious attempt to take on a resident male if he considers that he has a significant strength advantage. This is largely a case of interpersonal aggression and does not generally involve groups. Indeed, up until 15 years or so ago it was generally thought that only humans amongst mammals deliberately engaged in strategies aimed at weakening rival groups by killing its members systematically.

We now know that humans are not unique in this respect. The work of Jane Goodall, in particular, on chimpanzees in the Gombe National Park has shown quite clearly that these apes can engage in systematic warfare against neighbouring groups by killing isolated members deliberately. The killing is only carried out by males and is highly organised with the victim being immobilised by individuals pinioning its arms and legs while others rip out its vital organs from the abdomen and/or tear its neck out, thereby severing the trachea, jugular vein and carotid arteries. Furthermore, groups of animals set out in small raiding parties which seem deliberately intent on finding a member of a neighbouring group that is isolated. An important qualification in all this is that the chimpanzees only attack when in a position of overwhelming advantage – i.e. they do not attempt to kill their rivals unless there is a very small risk of injury. In this important sense they seem eminently more intelligent than their human cousins who often engage in wars that are bound to inflict heavy loss of life on both sides!

The only other mammalian species that actively seeks to kill members of neighbouring groups is the wolf. Studies by David Mech and colleagues in Alaska ’s Denali National Park have established that they exhibit patterns of warlike behaviour towards neighbouring groups which are very similar to those seen in Chimpanzees (Mech et al, 1998).

Contrary to what many may think the incidence of death through violence around the world is almost certainly at its lowest in human history. Although there is considerable variation across different regions and cultures in the World on average the number of persons dying as a result of interpersonal violence is currently only between 0.1 and 0.3% even during the time of the two world wars. In some early human cultures this figure might have been as high as 50% when competition for scarce resources was intense. Interestingly estimates of deaths through interpersonal violence in both chimpanzees and wolves are fairly similar and range between 0.2 and 0.5%. However, a report in 1997 on chimpanzee groups whose ranges were drastically influenced by logging activity estimated that for a period, death rates increased to 50% since disturbance to the stability of the different groups resulted in a huge increase in warfare!


Can we learn anything from other species about how to live more peacefully?

While no one would seriously suggest for a moment that we try to adopt a lifestyle more akin to any other species, the animal kingdom does suggest some interesting pointers at what can be achieved:

Few other species, with the possible exception of the great apes, have a developed moral sense which can help suppress overt aggression and yet most seem able to control it.
Aggression is mainly controlled by highly developed social perception skills where individuals are in tune with the social positions and moods of others and adopt either submissive or threat postures to prevent escalation into physical violence.
Animals use rough and tumble play and group activities such as foraging and hunting not only to maintain social relationships but also to burn off pent-up aggression.
It seems that the more social you are, in terms of being motivated to form attachment bonds, the more potential you have for aggression.

From these observations we might tentatively conclude that the way forward to reduce interpersonal aggression is not simply to try to improve our moral sense but also to either improve our social perception and communication skills and engage in more fun pursuits involving physical activity and play fighting or alternatively work on becoming a less social species. However, the only way to avoid warfare would seem to be to have no territorial borders and for everyone to have access to the same level of resources!

Another option suggested by other animal species is the old adage “make love not war”. While our “Hippy” culture may have largely come and gone it is interesting to see what adopting this policy long-term can achieve. Bonobo apes, unlike their chimpanzee and human cousins, are relatively physically non-aggressive and do not seem to engage in any form of group warfare. Genetically they only differ from us and chimpanzees by a matter of a few percent and therefore one would assume have a similar complement of pro-aggression genes. How then have they achieved their relative state of peace?

The answer does seem to be that they do “make love not war” but a secondary consequence of the negation of male aggression has been that their society is dominated by females. In general whenever two individuals encounter a situation where competitive instincts are likely to kick in – such as a new source of food or a female in oestrous – rather than engaging in some form of aggression the potential protagonists engage in a quick bout of sex (by quick I mean around 15 seconds, the individuals in this species seem to go for quantity rather than quality). Sex can take place between all partner combinations (i.e. both heterosexual and homosexual) and ages. As an aside, Bonobos are the only other mammalian species where the sexual act normally takes place with partners facing each other.

The success of this make love not war strategy does however have an important caveat that I have already alluded to above. You can’t use it in a monogamous context! Male and female Bonobos do not form exclusive monogamous partnerships and males do not even get involved in parenting. The closest social relationships are between females and the females will therefore act as a group to restrain any male that lets his aggression get out of control. I am sure there are some important principles to be considered here in the context of evolving a more peaceful human culture. On the down side however it should be stated that of our three species it is the Bonobos that are under threat of extinction and while undoubtedly intelligent they do not seem to make much use of this in the wild. For example, unlike chimpanzees and humans, they do not seem to make sophisticated use of tools.

Work with other animal species has, of course, shown us specific ways that we can interfere with brain systems controlling aggression using drugs, surgery or genetic engineering in order to create more pacific individuals. However, this is more relevant to pathology than normality. The aggression system is so interlinked with all aspects of our emotional lives that it is almost impossible to target it without having a negative impact on our ability to experience normality in other key emotions and even cognitive function. Indeed, the key to individual control of aggression lies not just in moral education and learning, and the rule of law, but in finding appropriate creative outlets for its expression.

Arguably for many humans the most creative outlet for aggression is finding solutions to difficult challenges. People perhaps like to view characteristics such as ambition, dedication and commitment in problem solving as representative of some higher mental attribute but the driving force behind it must reside in our aggressive competitive nature. After all, human advances in knowledge, with the possible exception of offensive and defensive weapons, have occurred mainly during times of peace. Perhaps the major contributor to triggering human physical aggression is simply boredom and/or lack of opportunity to find an appropriate outlet for your skills and ambitions.


Some final conclusions:

    Aggression, dominance and impulsivity are related
    For males aggression is about social dominance and competition  
    For females it is about protection and acquisition of resources
    Affective and predatory aggression are distinctive
    Female brains have more cortical aggression control
    Testosterone makes an aggressive male brain but not necessarily aggression
    Females find high T males more attractive for reproduction but not parenting!
    Reduced brain serotonin is linked to increased aggression
    Alcohol removes inhibitory control of aggression and obscures perception of social signals  
    Control of male but not female bond formation is linked to aggression  
    Experience can blunt or accentuate aggression but is not responsible for it  
    Warfare is not an exclusive human invention  
    Make love not war may be an option for some species but not for us  
    We need creative outlets for aggression



© Keith Kendrick, 11 January 2005


Selected references:

Antonia Abbey et al. (2004) Sexual assault and alcohol consumption: what do we know about their relationship and what types of research are still needed. Aggression and Violent Behavior 9:271-303.

JK Bester-Meredith and CA Marler (2001) Vasopressin and aggression in cross-fostered California mice (peromyscus californicus) and white-footed mice (peromyscus leucopus). Hormones and Behavior 40: 51-64.

Christopher Boesch, Gottfried Hohmann and Linda Marchant (2002) Behavioural diversity in Chimpanzees and Bonobos, Cambridge University Press (available from Amazon).

Anne Campbell (1999) Staying alive: Evolution, culture, and women’s intrasexual aggression. Behavioural and Brain Sciences 22:203-252.

Joseph McEllistrem (2004) Affective and predatory violence: a bimodal classification system of human aggression and violence. Aggressive and Violent Behavior 10:1-30.

Thomas Gregg and Allan Siegel (2001) Brain structures and neurotransmitters regulating aggression in cats: implications for human aggression. Progress in Neuro-Psychopharmacology and Biological Psychiatry 25:91-140.

A. Hariri et al (2002) Serotonin transporter genetic variation and the response of the human amygdale. Science 297:400-403.

Keith Kendrick and Robert Drewett (1979) Testosterone reduces refractory period of stria terminalis neurons in the rat brain. Science204:877-879.

Konrad Lorenz (1966) On Aggression, Methuen (available from Amazon).

L. David Mech, Layne G Adams, Thomas J Meier and Bruce W Dale (1998) The Wolves of Denali, University of Minnesota Press, Minneapolis.

Allan Mazur and Alan Booth (1998) Testosterone and dominance in men.Behavioural and Brain Sciences 21:353-397

Stephen Pinker (2002) The Blank Slate: The Modern Denial of Human Nature, Allen Lane, Penguin Press (available from Amazon)

A Raine et al (1998) Reduced prefrontal and increased subcortical brain functioning assessed using positron emission tomography in predatory and affective murderers. Behavioral Sciences and the Law 16:319-332.

Joanne Savage (2004) Does viewing violent media really cause criminal violence? Aggression and Violent Behavior 10:99-128.

Richard Wrangham (1996) Demonic Males: Apes and the Origins of Human Violence.


This event was on Thu, 13 Jan 2005

professor keith kendrick

Professor Keith Kendrick

Professor of Physic

Professor Keith Kendrick is Systems and Behavioural Neuroscientist and was Gresham Professor of Physic between 2002 and 2006.

He has been a member of the...

Find out more

Support Gresham

Gresham College has offered an outstanding education to the public free of charge for over 400 years. Today, Gresham plays an important role in fostering a love of learning and a greater understanding of ourselves and the world around us. Your donation will help to widen our reach and to broaden our audience, allowing more people to benefit from a high-quality education from some of the brightest minds.