Portal de Periódicos da CAPES

The scientific significance of Collins's Violence 1

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

10.1111/j.1468-4446.2009.01258.x

1468-4446

Mark Cooney,

Crime, Deviance, and Social Control

In Violence: A Micro-sociological Theory, one of the world's greatest sociologists, applies his formidable intellectual powers to the problem of human aggression. But how far does Randall Collins's big and ambitious book advance the scientific understanding of violent behaviour? Addressing that question, I pinpoint three major strengths and three major weaknesses of Violence before offering, at the end, a summary evaluation. The three principal strengths of the book have one feature in common: they are, in large part, products of Collins's departure from a criminological perspective on violence. The first virtue of Violence is that it is about violence. By contrast, most theoretical work that purports to be about violence only addresses violent propensities or dispositions. The most popular criminological theories, for instance, all seek to explain why some individuals or aggregates (e.g., cities, states, countries) are more disposed to violence than others (see, e.g., Akers 1985, Gottfredson and Hirschi 1990; Sampson, Raudenbush, and Earls 1997; Agnew 2006). Violent acts themselves they relegate to the vague concept of 'situational opportunity.' Dispositional theories have a major problem: they vastly over-predict violence. As Collins (2008: 2–3) notes, offenders carry their violent dispositions with them at all times, yet commit violent acts only sporadically. Dispositions merely explain the inner motivation to commit violence. About the outward behaviour all they can predict is that the motivated person is disposed to engage in some violence, at some time, in some place, against some people, in some manner. Dissatisfied with that level of imprecision, Collins rightly shifts the focus away from the individual to the situation or encounter, seeking to discover what is distinctive about those that take a violent turn. Collins is not the first thinker to move from dispositions to situations. Donald Black (1990, 2004a, b) holds that the key to explaining violent acts – who is violent toward whom and in what way – is to discover what is distinctive about the social structure or geometry of violent and non-violent conflicts. Curiously, Collins (2008: 22) makes little reference to Black's work, dismissing it as focusing primarily on the societal reaction to violence, thereby overlooking Black's substantial body of theoretical work addressing the social geometry of terrorism, feuding, vengeance, discipline, and rebellion. Generality is a second striking and significant virtue of Violence. Collins looks at violence wherever, whenever, or at whatever structural level it occurs. Thus, dueling, domestic assault, police brutality, rock-throwing rioting, military massacres, drunken brawling, Wild West gun fighting, schoolyard fights, sports violence, on and off the field, mosh pitting, and other types of aggression all receive an airing in this book. Collins's mastery of an enormous literature on these subjects is all the more impressive given that a mere ten years previously he published an even longer book on an entirely different topic: philosophy (Collins 1998). The breadth of Collins's vision stands in sharp contrast to the exclusive focus on criminal violence of criminologists. Collins urges scholars to look at the behaviour itself, at the aggressive infliction of bodily harm, however it is defined by the powerful. In doing so, he demonstrates the importance of fresh sources of data – video evidence, biographies, battle reconstructions, close observations of violent encounters, and the like. That alone is an important contribution. The third major strength of Collins's book is that it formulates an original theory of violence rooted in a general sociological perspective. For Collins, social life is founded on situational interaction (see e.g., Collins 1998, 2004). Humans, he argues, are hardwired to get emotionally entrained in interaction, responding to the words, gestures, actions, emotions, expressions, and rhythms of others. Entrainment is pleasurable and comes easily. Violence, on the other hand, is difficult. Aggression constitutes a serious rupture of the rituals of interactional solidarity. Despite the message given by television, movies, or even crime statistics, it is not easy to beat or kill others. Most of us fear doing so and are incompetent if we try. What holds us back, however, is not primarily the fear of being injured, it is the profound unease we experience at the prospect of violating the propensity to reciprocal entrainment. Only rarely do the situational dynamics of encounters allow us to surmount the confrontational tension we face when contemplating such a glaring violation of how we should, and almost always do, conduct ourselves in face-to-face interaction. Combined, these three strengths amount to the most important phenomenological treatment of violence ever formulated. From now on, all discussion of how people experience confrontation and violence must start from Collins's work. For all its merits, however, Violence does have its limitations. The three I shall highlight arise, paradoxically, out of the very strengths I have identified. Each stems, in particular, from a failure on the author's part to develop his points fully. By transferring attention from individuals to situations, Collins makes a more precise theory of violence logically possible. Unfortunately, he does not deliver such a theory. He argues that, no matter how acrimonious the quarreling or aggressive the posturing, confrontations will not become violent unless one of two factors is present: 1. A weak victim, or 2. An audience encouraging a fight. But this is unsatisfactory. Dominance is inherent in many encounters that do not generate the 'forward panic' of frenzied violence toward a weak opponent that Collins describes (2008: ch. 3). The vast majority of conflicts between husbands and wives in traditional hierarchical households surely did not lead to beatings. Most prisoner–guard, parent–child, or teacher–child conflicts today do not turn violent. Neither do most group-on-individual conflicts. Violence may be more frequent in these hierarchical conflicts, but it too is a rarity. Collins's answer is that dominance is situational – it results not so much from background status characteristics as from negotiations between the parties in concrete encounters that could have take alternative pathways into the tunnel of violence (see, e.g., 2008: 154). But this undermines the predictive power of his explanation – what factors tend to propel unequal interacting parties down the violence pathway rather than the non-violent pathways? It also raises the spectre of circularity: How can we know if a victim is 'weak' unless he or she is beaten – precisely that which we are trying to understand? Hierarchy plays a more complex role in violent conflict than Collins indicates, as Donald Black's (1995; 1998) pure sociology perspective makes clear. For Black, violent conflict obeys universal social laws. Those laws are found in the location and direction of the conflict in a multi-dimensional social space, its social geometry. Are the disputing parties equal or unequal; is the grievance upward or downward, close or distant? Different case geometries thus yield different forms of conflict management, some violent. Lynching (violence against individuals by semi-organized perpetrators), for example, is most often a response to offences (murders, rapes) by lower status against higher status actors and hence typically has a markedly downward direction (Senechal de la Roche 1997). By contrast, terrorism (covert violence inflicted by highly organized civilians on other civilians) is more likely to be inflicted in the opposite direction, by lower status actors against higher status targets (Black 2004b). Criminal homicide, however, is not strongly directional, most often occurring among approximate equals at the bottom of a larger system of hierarchy (Cooney 1998: 22–44). But many other variables, apart from audiences, are important as well, including relational distance, the degree of involvement in the life of another (Black 1976: 40–1). For instance, Scott Phillips (2003) interviewed young men convicted of murder and aggravated assault in Georgia and Texas about the conflict underlying the offence for which they were incarcerated (e.g., a murder arising out of a drug debt) and about a parallel conflict at around the same time in their lives to which they had not responded with violence. Their responses confirmed Black's (1990) prediction that vengeance (i.e., aggression in response to aggression) increases with relational distance: the men were consistently more tolerant of aggression from family members and close friends than from those with whom they had little or no prior contact The second factor – audiences – cannot do all the work Collins assigns it either. Collins (2008: 198–200) notes that large, enthusiastic audiences who form for a previously scheduled fight and who know the parties – as in high schools – are best for overcoming the confrontational tension that inhibits most antagonists. But he does not specify when people become part of an enthusiastic audience in the first place. An audience is one type of third party – actors other than the principals – who become involved in conflict (Black and Baumgartner 1983). Only some third parties cheer the principals on to fight. Others may urge forbearance, pull the principals apart, or mediate a mutually acceptable solution. In Warriors and Peacemakers: How Third Parties Shape Violence (Cooney 1998), I use Blackian theory to analyse the third-party geometry of violent and non-violent conflicts. Drawing on modern and cross-cultural data, I find, for example, that the probability of violence declines when third parties are socially close to both sides. Conversely, the probability of violence increases when one or more of the sides consist of a group of third parties who are socially close to the principal and to each other. Hierarchy matters here too: conflicts are more likely to be settled without violence when third parties are moderately socially superior to the principals than when they are either inferior or greatly superior to the principals in status. Scott Phillips and I (2005) tested and confirmed some of these ideas with Phillips's interview data. My point is that the role of audiences needs to be – and can be explained – because third parties may promote non-violence, not just violence. Collins's explanatory project is hampered by the one significant lacuna in his bibliography: the anthropological literature. Immersion in high-quality anthropological data helped Black (2004a), for example, to identify the conditions under which disputants will use more or less lethal weapons in their conflicts, when violent conflict will take the form of feuding, and what distinguishes tribal feuding from modern gang warfare. Likewise, any progress I have made in my work on third parties owes much more to the rich cross-cultural data than to the sparse information in the modern violence literature, which depends so heavily on official accounts of crime, and which typically treats third parties as peripheral actors. Those data, moreover, enabled me not only to address the distinguishing features of violent conflicts but also to shed light on such larger issues as variation in violence across stateless societies and across lower-income communities within modern societies. In sum, while Collins's emphasis on situational dynamics is worthwhile, much more remains to be said about the factors that differentiate that small minority of situations that result in violence from the great majority that do not. Violence, as mentioned previously, is remarkable for the breadth of its scope, its generality. But, here again, Collins stops too soon. For having staked out the capacious contours of violence, he then needs to come back and sub-divide the field into meaningful parcels. For violence is not a single entity. Dueling over a matter of honour differs drastically from serial killing; beating and robbing a store owner is quite different from slamming into an opponent on a football field. Moralistic violence (committed in the course of a conflict) is distinct from predatory violence (committed for gain or pleasure) or recreational violence (committed for fun) (Black 2004a: 146). These types of violence may have some overlapping features in common, but they appear to arise under largely distinct circumstances. Thus, compared to moralistic violence, predatory violence is more likely to: • Involve strangers • Be committed across class and racial boundaries • Exhibit organizational asymmetry (group-on-individual or individual-on-group) • Display a larger age differential between the parties • Involve parties who are not strongly tied to a place, who are mobile. Even within these categories there is much variation. Moralistic violence, for example, includes such diverse forms as vigilantism, rioting, lynching, terrorism, feuding, and genocide that have overlapping, but ultimately distinct, social geometries (Senechal de la Roche 1996, 1997; Black 2004a, b; Campbell 2009). Collins does not distinguish these types of violence because he does not need to since, despite his sanguine statement on the opening page that 'this vast array can be explained by a relatively compact theory', his analysis is largely descriptive. Most of the book is taken up with demonstrating universal features of violence: its interactional nature, its infrequency, its domination by the violent few. There is no sustained explanation of why violence varies, not just by type but also across time, space, and social groups. The reader will not find in Violence any testable explanation of such familiar empirical patterns as why some societies have little violence and others a great deal, why European homicide rates have declined sharply over the past eight hundred years, why domestic violence represents a larger proportion of modern than pre-modern violence, or why lower income African American men are more violent than their middle-class white counterparts (see e.g., Eisner 2003; Cooney 2003). Still, there is an original theoretical perspective in Violence. That phenomenological standpoint is the third major strength of Collins's book, but it is also a source of its final major weakness. The central message of Collins's book is that violence is difficult. Violence is difficult because it defiles the rules of interactional ritual, the foundation of social life itself. And because it is difficult, violence is much rarer than most people realize. But is violence rare because it is difficult? In a 1948 paper, the English statistician Lewis F. Richardson showed that the frequency of lethal violence varies inversely with its size or intensity. Thus the most common act of lethal violence is the killing of a single individual, but, very occasionally, world wars occur, in which millions of people are slain. Richardson argued that the distribution of violence displays a power law. Suppose a lethal event twice as large is four times as rare. If this pattern holds for lethal events of all sizes, then the distribution is said to be a power law. The signature of a power law is a long tail caused by the low intensity of most events. In the great majority of worldwide terrorist attacks, for example, just a few people are injured or killed, but, in a very occasional incident, many thousands are victimized (Clauset, Young and Gleditsch 2007). Although he does not state that violence describes a power law, Collins's analysis implies that it at least approximates one: the vast majority of aggressive interactions fizzle out; some exhibit minor pushing and shoving; a small few, spectacular violence. The problem, though, is that it is not just violence that displays this tendency. It seems highly plausible that most instances of deviant behaviour of any type are minor, a few are of medium magnitude, and a very small number are major. Consider some examples: • Fraud. Most fraud is likely to be small-scale, such as the writing of bad checks for relatively small amounts. Larger frauds that entrap more victims for greater amounts are rarer. Still more rare are frauds that fleece enormous sums of money from large groups of people and institutions over many years (e.g., the Madoff Ponzi scheme) • Drug dealing: In the typical drug exchange, a small quantity of drugs and money changes hands (see, e.g., Jacques 2005). Some deals are larger; a few involve truly enormous transfers of drugs. • Speeding offences. While most speeding offences are likely to be small breaches of the speed limit, a tiny minority represent extremely large excesses, like the driver in Texas who was ticketed for driving 167 mph over the 75 mph speed limit Whether these patterns follow a power law distribution in the strict sense, they resemble one in that there is, fortunately for us humans, an inverse relationship between the frequency and intensity of the behaviour. If much deviant behaviour is like that, then we need a theory that extends beyond violence. Interactional theft, for instance, might rightly be seen as a breach of interactional solidarity, but what about long-range fraud? Or speeding offences? Or drug-dealing? An interaction-based theory can scarcely account for those deviant behaviours that mainly involve distant interaction, that may not occur in interaction at all, or that have willing partners. Power laws appear to characterize much human and non-human behaviour, including natural disasters such as earthquakes and forest fires (Buchanan 2001). Why, is unclear. Perhaps the more intense the event, the more triggering conditions it requires, and the lower the probability they have of coalescing at any one time or place (Black 2009). But if power laws govern our world to some extent, then we would expect that there would be correlates at the phenomenological level – that people would experience violence as difficult. That would explain the facts that Collins emphasizes – the grimaces, the rigid body postures, the tension most people report in violent situations, and the general ineptness that they display in inflicting violence. But these facts would all follow from the central fact of violence being rarely distributed in social life – they represent the micro-level manifestations of a macro-level pattern. Far from being the ultimate explanation of the rarity of violence; they are just a vehicle through which larger, impersonal forces are expressed. From this perspective, Collins is right that violence is rare, and that people experience it as difficult, but he is not right that it is rare because people experience it as difficult. It is rare because of more general principle of social life: intense instances of human behaviour are typically much less frequent than less intense instances. ***** What, then, is the scientific significance of Collins's Violence? The book has to get high marks for originality, for providing researchers with a fresh and fruitful subject matter: the micro-interaction of violence. It must also be applauded for its sheer scope, for the generality of its vision. But the ultimate goal of any science is explanation, and here the work is less accomplished. If, as George Homans (1967: 27) once remarked, theory is a game that is won by explaining the most facts with the fewest ideas, then Violence, for all its virtues, fails to bring home any prizes for, at the end of the day, strikingly little variation in violent behaviour – across time, situations, groups, or societies – is explained in its five hundred and twenty-six pages.