Portal de Periódicos da CAPES

The micro‐sociology of violence

2009; Wiley; Volume: 60; Issue: 3 Linguagem: Inglês

10.1111/j.1468-4446.2009.01256.x

1468-4446

Randall Collins,

Physical education and sports games research

In recent years, sociologists have acquired much better evidence of what happens in violent situations. The advent of video cameras in the 1990s shocked public sensibilities about police violence. Much more data is now available from CCTV surveillance film for violence of many kinds. Videos made by cell phones now record the realities of brawls in everyday life. Still photographers have become more ubiquitous and intrepid, and there is now an abundance of photography of the details of riots, demonstrations, and military violence. In addition to these new sources of visual evidence, forensic reconstructions of shootings show how many bullets are fired, how many hit, and which hit the wrong targets. Military battles have been recontructed with lasers, and the physical realities and impossibilites of classic battle accounts have been shown. No evidence speaks for itself, and one always needs a theoretical eye in order to see what is before us. Thus as micro-sociological theory comes to bear on violent situations, we become better able not only to interpret the details of visual evidence but also to use ethnographic observations and interviews with better focus on crucial processes. Ethnographers and interviewers have become more intrepid too, and we have gotten much closer to violence as it actually happens. Over the decades we have come a long way from reliance on official statistics, through victim surveys (mostly producing bare incidence numbers), interviews of incarcerated populations, interviews in the wild, and into the scene of violence itself. Explanations in terms of background variables – the usual suspects of class, ethnicity, age, gender, family, culture – are a long way from explaining whether violence happens or does not happen. Instead we now are able to focus on the dynamics of interaction in situations. We are getting nearer the holy grail of evidence on all situations, rather than sampling on the dependent variable. We need to explain why violence fails to come off in many – indeed a majority – of the situations in which it threatens, and what is distinctive about the dynamics of situations which proceed more and more deeply into the tunnel of violence. A key micro-sociological tool for dissecting violent situations is indicators of emotions. Psychologists have shown universal expressions of basic emotions, above all in faces but also in body postures (Ekman and Friesen 1975; Ekman 1985) Emotions are shown in three zones of the face: the mouth and jaw; the region around the eyes; and the forehead. Anger is displayed in strong muscular clenching of the jaw, either with lips tightly pressed or with a square open, shouting mouth; eyes tightly focused to stare down the opponent; and the unmistakable tell-tale sign of vertical wrinkles between the eyebrows. Fear contrasts with anger in every facial zone: the jaw is slack with the mouth falling weakly open; eyes staring widely with lids pulled back unable to resist what is being seen; eyebrows turned up at the inner corners producing horizontal waves across the forehead. Cultural clichés and common discourse assume that anger is the emotion of violence. But my examination of hundreds of photos of individuals in military combat, police arrests, riots, and robberies show that anger is rare on the faces of persons engaged in violence; by far the most common emotion is fear, or its milder version, tension. (Collins 2008; in what follows, this is the source for photographic or ethnographic evidence not otherwise cited.) The prevalence of fear and tension is supported by studies both of the subjective phenomenology of violence and its physiological correlates. Interviews with the police about occasions when they fired their guns typically describe visual distortions (tunnel vision), auditory distortions (being unable to hear gunshots, though these are normally painfully loud), time distortions (feelings of moving in slow motion) (Klinger 2004; Artwohl and Christensen 1997). Grossman (1995, 2004) documents ubiquitous symptoms of fear and stress in combat situations, above all in close contact with the enemy. The experience of a pounding heartbeat is common, produced by the elevation of adrenaline and cortisol (Mazur 2005). I call the emotion of violent situations confrontational tension/fear (ct/f). Its major effects are: first, to inhibit most violence from happening; and second, when ct/f is sufficiently overcome for violence to occur, violence is typically incompetent at hitting its desired targets. One of the earliest findings of military sociology was that only 15–25 per cent of front-line soldiers regularly fired their guns in combat (Marshall 1947); subsequent training efforts have raised the firing ratio somewhat, but at the cost of increasing the amount of wild firing which does not hit its target (Collins 2008: 43–59). I have calculated, on the basis of photographic evidence and available statistics, that virtually all of the violence in each conflict-engaging population – crowds in riots and brawls, police, gangs, criminals – is carried out by a small proportion, typically on the order of less than 15 per cent at maximum (Collins 2008: 370–4). And not only is violence confined to a small minority of groups which ostensibly set out to use violence, but violence itself when unleashed is widely incompetent. Soldiers and police hit their targets in actual combat much less frequently than in practice on firing ranges; missing ones targets is extremely common in violence by criminals (where we can calculate this from numbers of bullets fired); similar patterns appear from blows thrown in fights. Friendly fire – hitting members of ones own side, including shooting oneself by accident – is common in military zones, where accidents account for up to 15 per cent of casualties; friendly fire happens frequently (though at unknown rates) for gang violence, carousing brawls and other civilian violence. And it is not only military who hit civilian targets; such wild firing is common in criminal violence as well as in weaponless brawls causing bystanders to be hurt. The cause of all this incompetent violence is ct/f. Violence is difficult, not easy; we have been led astray in thinking otherwise, both by media violence (which typically portrays fighters as good or bad, but nevertheless as brave and competent) and by our criminological theories (which assume that background conditions in an individual's biography are sufficient to account for violence). On the basis of micro-situational evidence, I say not so; the very situation of confrontation, when violence is threatened, generates ct/f, and this creates a barrier to the effective use of violence against ones opponent. The perceptual distortions noted above make the effective performance of violence difficult. Grossman (2004: 31) notes that when heart rate rises above 115 beats per minute (bpm), fine motor skills deteriorate; above 145 bpm, complex motor skills deteriorate; above 175 bpm, one loses peripheral and depth perception, audition, and cognitive processing. These are the effects of fear, not merely of heartbeat elevated by bodily exercize. Under these emotional conditions, violence does not come off as intended. Why are humans not good at violence? One argument is that we are culturally socialized, as modern people, not to be violent. But the same patterns of incompetent violence are found in tribal warfare, where the culture does not prohibit killing neighbouring tribes and even enjoins it, and in gang fighting (which structurally and culturally resembles adjacent tribes); and the same psychological and physiological processes appear on both sides of the law, in legitimate and illegitimate violence. A second argument is rationalistic: people fear being hurt by the enemy they are fighting. But this is variable; fighters are much more fearful and tense in close confrontations, even though long-distance weapons such as bombs and artillery are much more dangerous; and it is especially difficult to commit violence against a person who is looking in ones eyes than from behind or with head hooded. (Grossman 1995; Collins 2008: 73–8) A Mafia boss compared the ease of killing an animal with the difficulty of killing a human being: When you aim at a man, your hands shake, your eyes twitch, your heart flutters, your mind interferes . . . If possible you should always touch the body with your gun to make sure the man is dead. Man is the hardest animal to kill. (Bonanno 1985) This is a high level of realistic observation by a professional killer. Conventional styles of discourse favour boasting and a self-portrayal of toughness which hides ct/f, although detailed interview questioning almost always turn up admissions of fear and physiological symptoms of tension. It is the moment of interactional confrontation, more than anything else, which causes the tension. I have presented evidence elsewhere (Collins 2004) that ordinary non-conflictual human interaction shows a tendency for persons who establish a common focus of attention to become mutually entrained in a common rhythm of speech and bodily movements, and to intensify a shared emotional mood, what I have called an interaction ritual. Here I would argue that ct/f arises precisely because violent interaction cuts against the grain of the micro-processes which make up an interaction ritual. The tension arises because one is struggling not just against the other person, but against the tendency within ones own nervous and endocrinological systems to become entrained in the rhythms and emotions of the other person. From this point of view, ct/f is the exception which proves the rule of the solidarity-producing tendency of focused micro-interaction. Where mutual entrainment is blocked by action at cross-purposes but the focus of attention is sustained – since violence-threatening interaction produces an extremely intense focus of attention on the other person – the human individual is subjected to deeply disturbing tension. It is also possible that an individual in a violence-threatening situation experiences objective fear of being hurt by the other, and the two sources of tension/fear may add together. Nevertheless, persons who objectively have the upper hand in the situation of power and have little chance of being hurt also tend to experience ct/f. This is why the bodily indicators of fear rather than anger are seen at the moment of violence. This is not to deny that anger has a strong physiological base in the so-called reptile brain which dates back deep in infra-human and indeed pre-mammalian evolution. But the human brain is dominated by cognitive areas which are entwined with emotional connections; and these emotions are easily communicated to other humans through finely tuned bodily organs (Turner 2002). It is this evolved hardwiring for cognitions-plus-socially-entraining-emotions which produce the solidarity of interaction rituals. Anger gives explosive muscular energy, useful chiefly for bursts of effort to overcome an obstacle; as a social communication of the intention to treat the other person as an obstacle to one's will, however, anger runs up against the tendency for interaction ritual entrainment, and hence the arousal of anger produces ct/f. Although anger can be a precursor of a fight, it tends to be shut off during the fight itself. The main display of anger is as bluster – a social performance which may begin spontaneously but can be consciously encouraged in an attempt to dominate the other person without fighting. Anger is by no means a sufficient condition for violence; to the extent that we see anger in photos of violent confrontations, it is almost entirely in safely controlled situations, such as a demonstrator making angry gestures while the rest of the line holds firmly in place. Anger does not make violence automatic or easy; it remains to overcome the barrier of ct/f. And anger is not a necessary condition for violence; much military, police, crowd, and criminal violence happens without anger. For violence to take place, persons must find a pathway around the barrier of ct/f. Examining the micro-situational evidence for many kinds of violence, I suggest five main pathways which these take (Collins 2008). Here I will concentrate on one such pathway: finding a weak victim, someone who situationally gives in to emotional dominance and thereby eliminates the action at cross-purposes which generates ct/f. There are several other pathways: audience-oriented staged and controlled fair fights; confrontation-avoiding by remote violence; confrontation-avoiding by deception; confrontation-avoiding by absorption in technique. With limited space here I will explicate only the micro-mechanisms of attacking the weak, which is particularly central in military and crowd violence, as well as in individual fights and criminal violence. The most striking photographic evidence of attacking the weak concerns violence in riots. Photos of hostile demonstrations and riots show two chief phases: either two sides are lined up in orderly masses, confronting one another, but without violence; or the crowd has broken up into small groups, which is when virtually all the violence happens. (I am ignoring here the occasional phase where a few individuals are in front of a crowd throwing rocks at a distance.) Moreover, this violence shows a distinct numerical pattern: one individual has been isolated from supporters and is running away or has fallen to the ground, and is being attacked by small cluster of between three and six attackers (the upper number sometimes rising to eight). This pattern is found in police attacking demonstrators and demonstrators attacking police; in whites attacking blacks and vice versa; in many different ethnic combinations. The underlying dynamic is not merely racism or ethnic hostility, since one can find it in intraethnic violence, including labor violence and sports violence. Nor is it masculinity that produces this unequal violence; although most crowd violence is male on male, there are also photos of female-on-female violence, again in the ratio ranging between 3-to-1 and 6-to-1 of attackers versus fallen victim. Women, in general, commit less violence than men; but when they do so in groups, they are especially likely to attack another woman, and they adhere to the usual numerical ratio. This is a truly Simmelian pattern: the most dangerous social number is a ratio in the range of 3-to-1 and 6-to-1. Crowd violence is thus very one-sided. At the point where violence actually happens, the victim is isolated from social support, passive, not resisting. This illustrates a key point about attacking the weak: dominance is not merely physical but emotional. Some photos of military violence show an isolated soldier or policeman being attacked; in a photo of the overthrow of Milosevic in Serbia in 2000, the soldier has his gun still in holster, while he holds his hands covering his head, trying to escape the blows of four men who are attacking him with sticks and tyre irons. Other photos, of US troops in Iraq, and of peace-keeping forces elsewhere, show an isolated soldier with a machine gun in hand, who is backing away from a crowd who pursues him with angry gestures. The effective use of force is not in the weapons per se, but in establishing emotional dominance. Emotion dominence opens the way to physical dominance; it is after one side has become emotionally weak, emotionally dominated in the interaction, that violence generally does its damage. Until this point, weapons are largely an item of theatrical display, an effort at bluster, since even if weapons are fired ct/f makes most of their firing miss. This fits with the micro-interactional pattern I have argued above: as long as individuals (or groups) confront one another at cross purposes, they are up against the barrier of ct/f, runnning contrary to the physiologically ingrained tendency to fall into a common rhythm with each other. Passivity on one side allows the other side to impose their own rhythm upon the interaction; on the detailed micro level, the victim falls into participation with the rhythm of violence being perpertrated upon him or herself. A spectacularly atrocious form of violence happens in a dynamic sequence which I have called forward panic. Here, two sides confront each other over a period of time, during which tension builds up. This is documented in video sequences of police chases, such as the famous Rodney King beating in Los Angeles in 1991, which came at the end of a chase at up to 115 miles per hour. At the end of the chase, a large number of police had gathered (in this case, 21 officers had cornered King when he left his car, four of whom did all the beating while the others circled and shouted support). The turning point is the transition from a tense confrontation to overwhelming advantage on one side and the sudden weakness of the opponent. This leads to hot rush– the emotional entrainment of the attackers in their own release of adrenaline; piling on– the one-sided attack of the many against the few, the temporarily strong against the temporarily weak; and overkill– the repeated beating or other acts of violence going far beyond what is ostensible necessary to establish dominance. Since the Rodney King beating consisted of white (including hispanic) officers attacking a black target, the culturally available interpretation of the beating was racism. But this cannot be the whole story, or even the most important part of the dynamic. Racial hostility, stereotyping, and fear can be part of the initial pattern which contributes to the building up of tension; however, racism by itself does not automatically overcome the barrier of ct/f; even racists need a micro-situational advantage which allows the release of violence. Moreover, racism is not a necessary condition for either the tension build-up or its release; forward panics occur in a variety of ethnic combinations including within the same race. Most severe atrocities, with the pattern of hot rush, piling on, and overkill, come at the end of a forward panic sequence. This is found in police violence, including highly publicized cases where a large number of bullets are fired; in such cases, it is striking how many bullets miss their target – typically half of more, even though fired at close range. Forward panics also generate atrocities in crowd violence and in military violence; I have adopted the term forward panic from the French military research, Ardant du Picq (1999 [1903]), who called it 'flight toward the front'– a pattern of emotional release like a panic retreat but in this case rushing towards a fallen enemy, no longer able to resist. In the era when battles were fought at close range, most major victories involved forward panics; a period of prolonged tension during which both sides did little damage – because of the incapacitating effects of ct/f– was followed by an event which made one side weak and unable to defend itself – typically by some accident such as a human traffic jam, where one side became unable to move or fell to the ground, setting off a frenzied slaughter by the now-dominant attackers. In military battle, emotional dominance precedes most of the physical destruction. Forward panics also make up the most severe kind of domestic abuse, with the pattern of hot rush, piling on, and overkill, both in adult violence against children (and against weak elders) and in spousal violence by men against women. There are several kinds of domestic abuse, some of which are not forward panics; the micro-situational sequences of each are important to distinguish in causality and in practical policy. (Collins 2008: 134–55) Major forms of violent crime typically have a pattern of attacking the weak. In this case, since the crime is planned (although not necessarily in much detail or far in advance), the sequence of prolonged tension and sudden weakness is absent; instead the craft of being a criminal involves finding weak victims (Wright and Decker 1997; Collins 2008: 174–86). To some extent, this involves finding a victim who is physically weaker: muggers typically seek victims who are smaller and weaker (males on females, young females on elderly females), and often with the advantage of two or three against one. Again, sheer physical dominance is rarely the full story; armed robbers, who have the advantage of a deadly weapon over unarmed targets, nevertheless choose their moment when their victim is off guard, and can be emotionally dominated. A key feature is upsetting the rhythm of activities by the victim, such as appearing at the moment when a store owner is closing up the shop door. The technique of the armed robber is to gain control of the micro-interactional situation, imposing his momentum upon the victim. The male pronoun here is deliberate; female robbers tend to work with male accomplices; when females rob males, it is most typically by a prostitute who has lured the victim into an indefensible situation, typically lying down with his pants off (Wright and Decker 1997: 69, 100, 111). Armed robbers prefer victims who are old, white in a black neighbourhood, drunk, or oblivious to their surroundings. Armed robbers are practicing micro-sociologists, careful observers of who appears both to have money (symbolic indicators of social class) and looks emotionally weak, a nervous look, walking too fast, etc. (Wright and Decker 1997: 74, 83–90). The most lucrative targets in poverty zones are likely to be drug dealers, and here the robber needs to have an especially good micro-interactional technique to catch their target off guard. Robbers in St. Louis use the term 'fronting' to refer to persons who are showing off, talking tough, but with little ability to back up their words; typically this is scorn expressed by older robbers for teenage drug dealers whom they consider easy victims (Wright and Decker 1997: 67–8). These robbers are almost literally Goffmanian sociologists, oriented to the subtle cues of interactional style. Again, it is not simply a matter of displaying a gun, but of choosing the right moment for maximal emotional dominance; when a gun is actually used in action, it is more likely to be used to hit the victim with than to shoot him or her (Morrison and O'Donnell 1994). Robbers emphasize the dramatics of dominance, including by a strategic use of racial epithets, using racial putdowns (including blacks on blacks) to add the shock of insult to the total dominance of the situation. This is similar to Jack Katz's (1999) famous argument 'Pissed off in L.A.' in which racial epithets uttered in road rage are not expressions of deep-seated racial attitudes but are ritualistic devices to make verbal sense of a situation of frustrated blockage of vehicalized bodily movement on the highway. In this case, the black robber who calls another black a 'nigger' is engaging in a strategic and situational use of racist ritual for local situational advantage (Wright and Decker 1997: 59, 105). Armed robbers, although they work hard to establish emotional dominance, are not immune to ct/f; they feel psychological discomfort in robbing upon unfamiliar turf, and tend to forego wealthier victims in unfamiliar settings for more modest income on their home turf (Wright and Decker 1997: 74). A large majority of robbers prefer to approach their victim from the rear, since they dislike being looked in the face. This might be interpreted as a utilitarian concern not to be identified, but it fits as well the pattern that hit men, torturers and executioners also prefer to approach from the rear and to have a hooded victim (indeed, the most dangerous situation is where hoods are worn, especially on both sides) (Grossman 1995). Inconsistently, the same robbers say they prefer to use a big gun to impress and scare the victim (Wright and Decker 1997: 97, 105), even though they prefer the victim not to see them; the inference is that the gun is a symbol of dominance inside their own minds, which is consistent with the ritualistic behaviour of robbers with their guns at times when they are not actually on the job (Morrison and O'Donnell 1994). Residential burglars, too, are tense and wary about breaking into someone else's home. They verbally rationalize this tension by imagining the scenario of encountering someone inside, and go to considerable lengths to ascertain that the house is empty at the moment. This sounds like a utilitarian consideration, but it has strong elements of anxious fantasy: it is often verbalized that the occupant may be waiting inside to ambush them in the dark by shooting them as the come in through the window. In effect this is turning the tables, catching them off guard at a vulnerable moment, precisely what a robber attempts to do with their victim; some robbers even describe their fear of this scenario as a reason for preferring armed robbery (on their own chosen turf) over burglary (entering someone else's turf) (Wright and Decker 1997: 52; 1994: 12–13). But this cannot be objectively a very common danger; among this sample of burglars, 75 per cent have never been convicted of any burglary during their career, and given that their number of lifetime burglaries is on the order of 10 to 100 or more (for 90 per cent of this sample of burglars), the chance of trouble in any particular burglary must be vanishingly small. These burglars give utilitarian explanations of their fear of encountering a person: avoiding more serious charges than burglary, plus the possibility of being recognized when committing burglaries in their own neighbourhood; nevertheless they prefer operating close to their home territory, as the tension rises when they are farther away from home. A micro analysis of the moments at which burglars feel maximal tension indicates that the strongest emotional pressure is while crossing the boundary and entering the target (Wright and Decker 1994: 103–5, 118, 128–31, 140). Once safely inside, the burglar tends to calm down, feeling less tense and agitated, especially upon checking the house and finding no one at home. This is a decrease in ct/f precisely because their is no human confrontation. But the anxiety generally remains high while inside someone else's turf; we see this through a number of indicators: burglars who have completed their job and are back at home on their own turf often go through a phase of euphoria including prolonged tension-releasing laughter (Wright and Decker 1994: 161); 93 per cent of burglars in this sample did very quick searches and rapidly left the house, even though a longer search would turn up more loot, and their experienced chances of being caught are very low (Wright and Decker 1994: 157). While inside they often feel the need to use the toilet, but many shit on the floor – not out of expressed motive of hostility, but because they fear being caught in a small toilet room with only one exit (Wright and Decker 1994: 159; Walsh 1980) This is similar to the proportion of soldiers (up to 20 per cent) who defecate in their pants in combat (Collins 2008: 46–47). Because burglars are so frequently nervous, shaking, and scared, many use heroin, crack, or marijuana to calm themselves, even though they have a realistic utilitarian concern that these drugs make them physically clumsy (Wright and Decker 1994: 129). The behaviour of burglars and armed robbers gives us a clue to the precipitation of violence. Both types of criminals experience a considerable amount of ct/f at the moment of their encounter with their target, even though they have trained themselves in techniques of minimizing resistance and/or contact and communication with victims. The greatest tension appears at the moment of initiating the actual commission of the crime; once past this threshold, the experienced criminal falls into a rhythm of his own making, and moves into a somewhat increased comfort zone. But ct/f is still there, and any sudden resistance on the part of the victim tends to set off a forward-panic type of reaction. Thus persons who resist a robbery have approximately a 1-in-3 chance of being injured, even though the majority of robbers dislike having to overcome resistance (Wright and Decker 1997: 7, 84; Reaves 1993); when a burglar is surprised by a home occupant – exactly what he most fears – the burglar will typically fight. I have used the analogy that violence is like going down into a tunnel (Collins 2008: 360–9); the most difficult part is the mouth of the tunnel, the barrier of ct/f; once past this barrier, getting into the rhythm of the contentious action itself becomes self-entraining; thus when a victim is encountered, even one putting up resistance, one is able to escalate further into repeated exertions of violence. The micro-timing of violence is crucial, all the way through the sequence of violent action. It is hard for humans to begin, but once begun it tends to set a rhythm of its own which is hard to stop. Here we see why crime is not easily predictable from background variables such as poverty, race, or family background. There are far more persons who fit the crime-prone pattern than those who actually commit such crimes, because initial motivations or opportunities for crime and violence are still a long way from the interactional competence to carry them out. A repetitive criminal is someone who has learned techniques which are successful a large proportion of the time. Those who do not learn these techniques, one can infer, soon desist from crime; this spontaneous desistance is probably greater than the amount which comes from being caught. The world of crime, and of casual violence as well, is itself a competitive world; those who are successful have learned techniques of emotional dominance not only over their victims, but to a lesser degree over their competitors as would-be criminals and tough guys. A small minority learn how to overcome their own ct/f or at least to manage their emotions sufficiently to carry out successful violence; they do this above all by becoming good observers of other persons' ct/f, which they turn to their own advantage. It is little exaggeration to say that violent persons are specialists in manipulating ct/f.