Portal de Periódicos da CAPES

Reply to Felson and Cooney

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

10.1111/j.1468-4446.2009.01259.x

1468-4446

Randall Collins,

Community Health and Development

I appreciate the comments of my distinguished colleagues, which are useful both in clarifying the micro-situational theory and in indicating where it needs to be synthesized with other bodies of theory. Felson [Felson 2009] complains of over-reliance on anecdotes as compared to quantitative research. What we need, however, is close detail on the sequence of events which take place in the natural world – in real life rather than in artificially contrived experiments, and such descriptions usually must come from someone's account of incidents they have seen or participated in. The term anecdote implies the narration is mere hearsay, stories embellished and recirculated as entertainment; such stories indeed tell us more about cultural clichés than what actually happens. My method has been to seek first-hand accounts of experience, and where possible to probe for detail about emotions, numbers of persons present, and sequence of events. Mere quantitative analysis, based on abundance of easily coded data, is not necessarily of great explanatory value if its variables are too remote from the scene of action and strip away too much of the relevant dynamics. The term ‘anecdote’ should be banished from our sociological vocabulary; what we want to exclude is unreliable hearsay, but we also need the most relevant data. My central difference with Felson is that he tends to see things from a more aggregate level and without the fine micro details. Micro-sociology looks at events at the level of fractions of seconds at key moments and turning points, and at the exact pattern of sequences over seconds and minutes. Most of our data on violence lumps what happens into much larger chunks. Thus when Felson says harm-doing is not rare, but is an accepted response to deviant behaviour, he is taking the lumps of aggegrated events (after all, there are millions of assaults, etc.) and overlooking that on the micro-level the number of incidents of conflict which get past micro-situational barriers and into violence is a rather small proportion (as one sees whenever we avoid sampling on the dependent variable; for data on how failure to do so has made us misperceive the relationship between alcohol and violence, see Collins 2008: 263–73). The fact that people find violence acceptable under certain rationales tells us more about their cultural discourse than about what they actually do. Similarly, I agree that behaviour is generally goal-oriented, but that does not make it easy, and crossing the situational barrier to violence is still the sine qua non of actual violence. Felson misstates my argument about forward panic. It is one version among several types of attacking the weak; it has a distinctive pattern – piling on and overkill, violence going on much longer than necessary to win the fight; it has a distinctive phenomenology, the feeling of being in a dreamlike state apart from one's usual self, while experiencing being caught up or entrained in one's own bodily rhythms such that the same violent action is repeated over and over without conscious control; and it has a distinctive causal sequence – a prolonged period of tension vis-à-vis an opponent which is suddenly released when the opponent turns to flee, becomes isolated from supporters, or falls down and gets tangled up. Forward panics happen chiefly in police chases, military battles, and riots, and in one type of domestic abuse; but there are other forms of domestic abuse such as prolonged psychological-plus-physical torture regimes which lack the features just mentioned. Bullying is an institutionalized form of attacking the weak which lacks both the temporal sequence and the overkill characteristic of forward panic. Predatory violence is not forward panic but generally depends upon techniques for locating weak victims and for attacking them at vulnerable moments. To distinguish forward panic from other kinds of attacking the weak, we look for evidence not in photographs of facial expressions at the moment of violence but for evidence on the entire sequence over time. It is also a mistake to confuse, as Felson does, forward panic with rage. As I have indicated from facial indicators, rage (extreme anger) is not generally displayed at the moment of violence itself, but while blustering confrontation is going on. If further micro-situational contingencies unfold so that violence comes off, rage is usually replaced by other emotions (chiefly expressions of tension and fear). It may be the case, however, that the physiological arousal which goes into rage carries over into the aggressive action even at the time when the emotional expression is tension and fear; this appears to be the case when a burglar suddenly meets the home occupant, and fits Felson and Messner's (1996) evidence that an assault turns deadly if the victim is considered threatening and dangerous, making the attacker feel it necesssary to finish them off. I doubt whether the actual sequence of thoughts at that moment in the killer's mind is a verbal statement that it is better to do this now to avoid future retaliation; instead the antagonists' attention are concentrated on the muscular efforts of a violent contest which becomes self-entraining – repeated and escalated by feedback from one's own prior bodily movements, and from the opponent's. As I have argued, the barrier of confrontational tension fear (ct/f) is most difficult at the outset; once it is broken through, the fighter is caught up in a new rhythm, and once inside it can be hard to get back out. Perhaps we need a new vocabulary for momentary emotional states and where they appear in the micro-sequence of conflict, bluster, and actual violence; and we need further micro-research on the details of how this happens. Felson, perhaps light-heartedly, suggests that anyone who challenges the assumptions of contemporary evolutionary psychology must be a creationist. The issue however is whether theories of the evolution of human physiology fit the facts. Evolutionary pyschologists appear to be attached to the utilitarian image of the selfish individual, a generalization of homo economicus. Without imposing these philosophical presuppositions about what humans are fundamentally, the ethnographic evidence shows that humans in all known societies carry out rituals which generate solidarity, and that persons easily become caught up in micro-interactional rhythms and emotions which they communicate to one another (see evidence summarized in Collins 2004; Turner 2000). It is a reasonable inference that humans are hard-wired for a high degree of subtle communication, especially via emotional channels. Evolutionary theory needs to accomodate this empirical pattern, rather than engage in a hypothetical exercise of explaining human beings as if they evolved from selfish individualists. Experiments based on game theory which show that tit-for-tat is a strategy could generate cooperation, do not demonstrate that this was in fact the path either of biological evolution or of human history. Game theory is a rationalistic model which draws its conclusions either by pure inference from a priori premises, or by experiments in highly constrained artificial circumstances; game theory does not much resemble the way people make decisions and interact with each other in observed micro-sociological reality, especially when people are allowed to interact face to face. In contrast, the theory of humans as hardwired for solidarity rituals fits well with an explanation of how human rationality evolved, via the internalization of social conversations into thinking in internal dialogue, and memories of emotionally evocative symbols created by shared communication (Collins 2004: 183–220; Mead 1934; Tomasello 1999). We do not need a crude image of humans either as naturally aggressive or pacifist (here Felson and I agree); it is the mechanism of interaction rituals which generates higher degrees of human planning and social coordination, and these in turn lead both to more idealized moral obligations, but also to human capacity to organize for warfare, to keep soldiers from running away from the front line, and to invent all the other ways in which ct/f can be circumvented. It would be valuable to promote the further integration of neuro-physiology, evolutionary biology, and the observational sciences of social interaction; but at this point a key step is for the evolutionists to catch up with the findings of modern sociology, anthropology and history, rather than remaining trapped by the premises of nineteenth-century utilitarianism. Felson's comments are most useful where he critiques the sub-theory of attacking a weak victim, based on evidence on aspects of situations in which violence is unleashed or inhibited. He points out that acts of submission can avoid escalation of a verbal quarrel into violence. This reminds us that violence comes at the end of a sequence of several steps, and various moves can head off conflict before a tense confrontation builds up; my emphasis is that the final step to violence depends upon the target showing him/herself emotionally weak in the face of a threatening gesture, and this generally happens in a few seconds or even in a flicker of fractions of seconds. Cooney makes a similar criticism: not all powerful people attack weaker persons at all times, but relatively seldom; my conclusion is that we need to pay attention to the sequence which leads to escalation, as well as to the micro-interactional details of what happens when antagonists do and do not proceed to violence. My chapter entitled ‘How Fights Start, or Not,’ (Collins 2008: 337–369) presents evidence that antagonists who mirror each other's threatening gestures stalemate and stabilize the conflict; the continued repetition of angry insults (which is what conversational recordings of angry quarrels generally show) eventually turns boring, bringing down the emotional level, and the conflict dissipates (barring audience intervention). This is true also of fights which end after a few desultory punches, or party-brawlers who wrestle clumsily on the floor while the audience, and eventually the protagonists, lose interest. In contrast, it is when the micro-interactional balance is upset that attacks are successful; photographic evidence of body postures and facial expressions show the victim taking part in his or her own beating by taking the posture of recipient of the attacking movements. Felson cites evidence from Tremblay (as I do also, in Collins 2008: 468) that the height of violence is among two-year olds, and declines thereafter. Is this because children learn politeness and self-inhibition? No doubt; but as indicated by Felson and Russo (1988), it is also because greater parental protection at that age makes younger and physically weaker children more willing to fight their adjacent siblings. Felson argues that this runs counter to the pattern of attacking the weak, since greater age gaps between siblings reduce violence. I think what we have here is a strong audience effect, with the audience intervening not so much to prevent the fight but to protect the younger, who becomes psychologically the stronger. I have suggested that even babies, whose aggressive tactic is limited to crying, sense an emotionally weak target – an adult easily controlled by crying (Collins 2008: 139–41; for a detailed observation see Katz 1999: 229–73). The details of these comparisons underscore the point that micro-interactional weakness is more emotional than physical. Felson notes that attacking a weak target is often ruled out because although it is physically easy the rewards are low, and one can even lose status among onlookers. This is true in certain social communities, and thus gang violence typically targets a rival gang member or surrogate of the proper age, gender and ethnic group; in honorific street fights, boys fight boys, girls fight girls; male–female violence is largely at home behind closed doors where reputations in the larger community are shut off. Thus the meso-level of the community structure and the physical setting are important parts of the earlier sequence in which antagonisms build up; but on the micro-interactional level, the last step into violence still involves finding dominance over the emotionally weak. This is shown in the techniques of violent predators; not only do they chose their victims apart from honorific considerations, but they choose the moment to attack when their target is emotionally off balance, and the attacker can impose his momentum on the micro-interactional details of the situation. Felson cites evidence that persons who commit violent crime have low self-control. But on the micro level, such persons succeed in violence by mastering subtle techniques of judging other persons' emotions and attentions; they are not merely blundering through the world in oblivious self-assertion. What we call low self-control may be our middle-class value judgment; in contrast, I suggest that persistent criminals are those who have learned a micro-interactional skill that we do not approve of. Cooney (Cooney 2009) justly faults me for omitting the evidence assembled in support of Donald Black's theory of social geometry. Thus my analyses of attacking the weak are underdetermined, and need further specification by group structure which correlate with upward, downward, and lateral violence in stratification hierarchies, and different types of conflict-management in distant and close relationships. I am aware of much of this research and teach it in my class on social conflict; my only excuse for not incorporating it in the book is my focus on micro-situational processes. The Black/Cooney/et al. data is meso-level, patterns spread out in time and space, whereas my data is micro-level, in sequences of minutes or seconds. It is hard to get data on both levels simultaneously. But now that we know what we should be looking for, we can move to the next step of bringing the two together. Cooney notes that my analysis of audience effects would be bolstered by the Blackian theory of conflict management by third parties. Indeed, my tables summarizing fight ethnographies show a strong relationship between the audience's support of, indifference to, or opposition to a fight and the severity of the fight (Collins 2008: 203, 204, 246); and I further show that crowds which are deliberately assembled (as for a concert or athletic event) are more likely to support fights than accidentally assembled crowds (as in traffic situations). Cooney goes further and shows that the social relationships and hierarchy between audience and principals determines what attitude the crowd will take. Here we get a useful cumulation of converging research. Cooney criticizes the scientific contribution of the micro-interactional theory on two grounds. One argument is that a power law governs all social phenomena, such that more extreme events are always rarer. But here I think Cooney is being too clever for his own goal, which is to provide a predictive explanatory theory. Richardson's power law tells us nothing about who will do what in which circumstances; furthermore, mere statistical patterns do not make a theory, since they lack a mechanism. In contrast, I provide direct evidence of the emotional difficulty persons have surmounting ct/f, a difficulty which exists for mild violence as well as for severe. Around this I infer various causal patterns: given particular kinds of micro-configurations, violence happens or does not happen. My theory does not lack causality, but it is on a level where we must collect data which are unusual in the social sciences. Cooney complains, justifiably, that this micro-theory does not make predictions about macro differences in aggregate patterns of violence: why violence changes historically, why some societies have more of particular kinds of violence than others, why there are race and class differences in incidence. We could continue to approach this on a purely macro level; here there have been some advances in explaining long term historical trends, notably the theory of state penetration and monopolization of violence, and changing household and community boundaries of the kind which Black and colleagues have been good at specifying. At the same time, there are paradoxes, such as the reversal of a long historical trend toward lessening violence and crime, a reversal which look place in many modern societies around 1960 (Eisner 2003). Micro causes of violence surely are implicated in the causal chain of any kind of violence, since every instance of violence must somehow get to the sticking point where individuals commit it, and the task of micro-macro synthesis remains an important one on the agenda of sociology. My book on micro-violence is the first of a two-part series, the second of which aims to contribute to this task. There is, however, a strong advantage for micro theory of violence on the practical level. Macro theories which attribute violence to the amount of poverty and discrimination, for instance, or decline in violence to the historical extension of the bureaucratic power of the state, have practical consequences only by calling for massive social engineering. If our aim is to end violence by ending social inequality, that implies that violence will not be remotely near to being eliminated in our lifetimes, if ever. Micro-interactional theory of violence, in contrast, provides immediately applications, not on the meso/macro level of government programmes, but on the level of persons learning techniques to keep violence-threatening encounters from escalating. I have suggested several (Collins 2008: 463–4) including micro-interactional techniques of not showing emotional weakness while allowing confrontations to die out through boring micro-repetitions of bluster. Micro-interactions are amenable to conscious action by their participants which have palpable consequences, in contrast to macro-social engineering which is hard to implement and hitherto largely ineffective. The implications of macro theories of violence are pessimistic; micro theory is optimistic.