The sociological eye means looking at things for what they are, as best we can given the blinders of interest and ideology, of cliché and ritualized belief. It is not an individual enterprise. Chaining our efforts together as a long-term network of theorists and researchers improves one’s own sociological vision, provided we make the effort. The sociological eye holds up a periscope above the tides of political and intellectual partisanship, spying out the patterns of social life in every direction.

Thursday, July 7, 2011


Bullying was once a fairly well-defined phenomenon. Recently the term has been expanded by journalists, politicians, and in popular expression. What difference does it make what we call these events? The word is being used to cover differing types of conflict, which have different causal paths, and thus very different implications for what to do about them, and for the damage done.

Traditional bullying is picking on network isolates-- victims who are lowest in the group status hierarchy, who lack friends and allies, and lack the emotional energy to defend oneself. Bullying is a repetitive relationship, the same bullies persistently domineering and tormenting the same victims. The classic version was in British boarding schools, where older boys were allowed to make a younger boy into a servant, carrying their books, cleaning their rooms, and generally deferring and taking orders. Nineteenth century school administrators regarded this system as a salutatory way for boys to learn discipline; but it often intensified into maliciousness, physical abuse, and commandeering the younger boy’s possessions. Some boys became school bullies. [sources in Collins, Violence: A Micro-sociological Theory, chapter 4.] The system was called fagging and the younger boys were called fags; this was the origin of the slang term for homosexuals, although that was not its original connotation.

Bullying is not a single event but an ongoing relationship, i.e. a network tie with asymmetrical content: one side bullies the other, never vice versa. It has a specific network location: bullies are not the top of the status hierarchy, but middle-ranking, not very popular themselves, but aggressors rather than victims. Bullying should not be confused with a dominance contest over who is the top-ranking male, which centers on the top contenders, and matches good fighters and leading personalities against each other. Bullying is exploitation by a particularly predatory type of individual from the middle against the bottom. In effect bullies make up for not very good social skills by picking on those who are even worse. Being a bully is not just anybody who fights; it is a specialized role in the status hierarchy, and not a very honorific one.

Classic bullying arises in total institutions like prisons, boarding schools or camps. Key conditions are: there is no escape from close contact with the same set of people; reputations are widely circulated; and the split between control staff and inmates creates a code of no snitching which cuts off victims from protection by authorities. The totalness of institutions is a continuum; as the strength of these variables increases, we may expect bullying relationships to be more frequent.

Classic bullying should be distinguished from scapegoating, where everyone in the group gangs up on a single victim. Usually this is someone who is blamed for a community catastrophe, or otherwise becomes the center of hostile attention. Scapegoating tends to be a single-shot event, rather than an ongoing relationship. The scapegoat might be low-ranking, an isolate, new arrival or cultural deviant; but scapegoats can also be selected from the elite. This happens in scandals, where the secondary scandal-- threatening supporters of the scandalous individual with contagious blame if they don’t join the condemnatory majority-- can rapidly strip even eminent persons of support.

Scapegoating is not carried out by bullies seeking individual dominance, but is genuinely mass-participation ritual of community solidarity, self-righteous Durkheimian unity at its least attractive. Scapegoating tends to arise in tightly integrated communities-- not the hierarchic ones characteristics of bullying; in complex societies, scapegoating requires a huge media frenzy to generate a comparable amount of focus and social pressure. On a smaller level, there is some evidence that girls focus their attacks (mainly verbal) on the lowest-ranking girl-- i.e. a collective action of the entire group against the bottom. This fits with females as being more solidarity-oriented than males, using verbal and emotional attack to keep up group unity at the expense of common target. In contrast, boys tend to fight it out over individual status at the top [studies reviewed in John Levi Martin, Social Structures, 2009, chapter 4]

What Isn’t Bullying?

It is misleading to refer to all kinds of personal conflict as bullying, even if it does happen in school or among young people. Bullying, as a repetitive, unequal relationship among individuals, where distinctive bullies target low-status isolates, has a very different structure and causality than two-sided fights. Among the latter are:

Individual honor contests: two rivals square off against each other, whether with fists, blades or guns, informally or under conventional rules like a duel. Honor contests are almost never top against the bottom, because there is no honor to be gained unless you show you can beat someone of considerable prowess, or at least stand in with them. This is a reason why bullies have mediocre status at best.

Intergroup fights: horizontal struggles between rival gangs, ethnic groups, schools, or neighbourhoods. These can be pretty nasty; in part, because the antagonists tend to be mutually closed Durkheimian communities, so they have no moral compunctions against vicious tactics; on the verbal level, they are prone to derogatory stereotyping, including racial slurs. And because confrontational tension makes fighting difficult to carry out in real life, groups are most successful when they engage in ambushes, drive-bys, or ganging up on outnumbered members of an opposing group who happen to stray into vulnerable territory. Thus actual incidents between gangs or ethnic groups may have something of the look of bullying, where a stronger group beats up on a weaker. News stories about a single incident cannot tell us whether it is bullying or not. Horizontal conflict is not a repetitive relationship of institutionalized inequality, but generally a sequence of alternating tactical advantages.

Another important difference is that inter-group violence chooses its targets as members of a group, not as low-status isolates. For this reason, intergroup violence is probably not as psychologically debilitating as being a bully victim, and may even give emotional energy and solidarity. In contrast to bullying, which leaves victims with very negative self-images, intergroup violence often gives members meaningful self-narratives-- one of the main attractions of belonging to a fighting group. [This is brought out vividly in Curtis Jackson-Jacobs, Tough Crowd: An Ethnographic Study of the Social Organization of Fighting; unpublished ms, UCLA.]

Some intergroup fights combine with aspects of bullying, where a weaker group is repeatedly attacked by a stronger. Instances include school majority black students attacking academically better-performing Asian minorities (e.g. in Philadelphia high schools in 2009-10). But although one side is dominant in the violence, there is an element of horizontal conflict as well, as the two groups compete with different resources-- violence vs. academic capital.

Insult contests: individuals bragging, boasting and making gestures about their alleged superiority to others. This can be done in a tone of entertainment and humor, or it can be hostile and malicious, attempting to establish emotional dominance; it can remain contained, or escalate in emotional tone and physical violence. Ethnographies of gangs and youth culture show a great deal of this. Although insults can be part of a bullying relationship, where they serve to maintain emotional dominance, or to provoke the victim into futile and humiliating outbursts; nevertheless much insults are not part of a unequal relationship. Moves in an insult contest are often reciprocal, and may be compatible with equality and even a ritualistic form of play producing solidarity. An observer cannot simply classify all insults as bullying, without seeing what kind of relationship it is.

Malicious gossip: This is a form of insult, but instead of being in your face, allowing the possibility of direct response, negative gossip is indirect. Gossip is felt to be more unfair, because it is harder to counteract. Nevertheless, malicious gossip is not necessarily bullying. It does not always, or even generally, take the form of attacks on those at the bottom; often it is an attack on those at the top, and on leaders of rival groups. Nor need gossip originate from bullying specialists (although it could-- persons who initiate malicious gossip might be structurally analogous to bullies, although we lack good data on this aspect of gossip networks). Most importantly, malicious gossip is often two-sided, between factions mutually attacking each other.

Research on children’s and adolescent status systems shows that girls tend to engage in more verbal attacks than boys. This is sometimes referred to as bullying, but before deciding, we need to examine the structure of relationships. Outcomes can be quite different, depending on whether the target is isolated, or herself a well-integrated member of a clique. Girls’ two-sided quarrels in the goldfish bowl of school or neighbourhood may well be the equivalent of gang fighting for boys, manufacturing a sense of excitement and meaningful narratives for their lives. How you experience this depends on your network location.

Homophobic attacks: conceptual confusions and real consequences

With increasing public focus on attacks based on sexual orientation, there has also been considerable muddying of what is actually going on. Homophobic attacks are bullying if there is a repetitive pattern of attacks by individuals or bully cliques on isolated, low status individuals; and their target is homosexual. What if the target is not an individual but an entire group of gay persons? Most of the reported evidence among school children is not about group confrontations, but attacks on isolated individuals, although group violence sometimes happens among adults when a gay bar or hangout is attacked by homophobic outsiders. [For examples see Elizabeth A. Armstrong and Suzanna Crage, 2006, “Movements and Memory: the Making of the Stonewall Myth,” American Sociological Review 71: 724-751.] What difference does it make, if it’s all bad? From the sociology of violence we can infer that isolates are much easier targets than groups; unless there is an extreme imbalance in numbers or weapons, attacks by one group on another usually abort. Homosexuals are more likely to be harassed and attacked in classic bullying conditions-- a reputational goldfish bowl of a quasi-total institution, with isolated individuals at the bottom of the status hierarchy. A reasonable hypothesis is: where there are groups (i.e. real networks) of homosexuals in schools, they are less likely to be attacked.

What if the attack is not repetitive but an ephemeral incident? Again, what difference does it make? But differences in degree do matter; subjectively, most of the damage of being on the receiving end of bullying relationship comes from the constant harassment, leaving a feeling of hopelessness and inability to act on one’s own volition.

Homophobic attacks where the entire community unites in ganging up on a victim, are not bullying, but scapegoating. In schools, this ranges from mocking and jeering, to pranks (playing keep-away with one’s things, stealing his possessions, locking a boy in his locker, dumping him into a trash bin) in a progression to varying degrees of violence. What difference does it make how we classify it? It makes a big difference in terms of practical counteractions. Scapegoating versus bullying is the difference between trying to change the entire culture and dynamics of a school, and trying to control or remove a small group of bullying specialists. It is the difference between a lynch mob (and the community structure and mentality that fosters it), and dealing with a small number of criminals, and not very popular ones at that. [On the structural conditions for lynch mobs, and the network relationship between them and their victims, see Roberta Senechal de la Roche, 1997. “The Sociogenesis of Lynching,” in W. Fitzhugh Brundage, Under Sentence of Death: Lynching in the South.] If the problem is homosexual bullying rather than homosexual scapegoating-- and most detailed evidence seems to indicate the former-- this is an easier problem, giving optimism for the future.

Another variant may be called pseudo-homophobic insult. Here homosexuals are not involved at all, but only invoked rhetorically. Such homosexual taunting operates as part of a repertoire of insult. It can be used both horizontally or vertically.

Insults are a major part of horizontal conflict, between rival groups or individuals [patterns summarized in Collins, Violence, chapter 9], and are more common than violence itself. Ethnographic literature on black gangs shows it is fairly routine to call someone ‘nigger’, either playfully or as a degrading expression. To call someone ‘gay’ can operate in a similar way. We lack good comparative evidence on which contexts are where this is likely to happen; but the following gives at example where pseudo-homophobic insult is routine. [personal communication June 2010, from Anthony King, Univ. of Exeter sociologist engaged in research on training and combat practices of US Marines and UK Marines.] In combat training to clear a building, British Marines line up closely one behind each other in a “stack”, ready to fan out once they are inside. American Marines deride the stack as “gay”. This does not mean they literally believe the Brits are gay; they are critical, in part on practical grounds that the stack makes them more vulnerable to all being shot in the doorway; and even more so out of inter-service rivalry between elite, high solidarity combat teams that are otherwise quite similar. The close bodily formation of the stack brings an ironic association with homosexuality as a readily available insult. The insult itself is a ritual of rivalry among equals.

Vertical pseudo-homophobic insult can be a way that boys mock low status isolates. To call someone gay is a form of rhetoric, an escalated insult meant to be especially wounding. The perpetrator may or may not believe the insult to be true (another detail awaiting good research). The consequences can be severe; there is evidence that in a considerable proportion of school rampage shootings, the shooters are striking back at those who insulted them in this fashion [Katherine Newman et al. 2004, Rampage: The Social Roots of School Shootings]. But investigating cases of school rampages are sampling on the dependent variable. Some recipients of pseudo-homosexual insults do not strike back, but commit suicide. But if calling someone gay is a popular-- and therefore frequent-- form of insult, in the great majority of instances it must be the case that the recipient neither strikes back with murder nor commits suicide. Here the pattern is similar to the sociology of violent interaction: most conflicts go no further than verbal bluster. It remains to be found: what are the conditions under which pseudo-homosexual insults result in escalation, or not?

In both non-sexual violence and homophobic conflicts, there are many contingencies between occasions for taking offense and subsequent escalation. Some persons are more touchy than others, and this touchiness is sociologically grounded in situational dynamics and network positions. Understanding what makes a chain of events worse, or better, is not simply a matter of a static culture of homophobia. The interactional patterns and locations of the individuals involved are what is fateful for the paths they follow.

Research Methodology Makes a Crucial Difference

There are widely disparate reports on the amount of bullying in schools. Some recent reports reach as high as 80% of students claiming they are victims of bullying-- if true, this would be a huge break from the well-documented pattern of low-status isolates as victims. Detailed studies of traditional bullying found about 16-18% of second graders as victims, and 3-5% of 9th graders, with girls always the lower number than boys (Dan Olweus, 1993, Bullying at School.) High estimates come from using survey questions that ask whether someone is subjected to being left out of activities, name-calling, rumours, teasing, sexual comments, threats, pushing or hitting [e.g. Bradshaw et al., “Assessing rates and characteristics of bullying through an internet-based survey system.” Persistently Safe Schools, 2006]. But we have no way of knowing from such answers whether these are two-sided fights, insult contests, or teasing games; or if they fit the bullying pattern of repeated, asymmetrical aggression between specialists in domineering and isolated low-status victims. We can only tell the dynamics of bullying-- and other varieties of violence-- if we explicitly ask whether these aggressive actions are reciprocated; if they are repeated, and between whom; and what the network positions are of these individuals are in the status hierarchy.

As it stands, there is no good evidence to suggest there is any more widespread bullying than in the past; conceivably real bullying could be lower, as schools have become more control-oriented. What seems certain is that the appearance of an epidemic of bullying has been created by inflating the definition, so that it now includes all kinds of horizontal fighting, and indeed any negative expressions at all among school children.


This is a recent journalistic term for insults, malicious rumours, and degrading images and videos spread via the Internet and other electronic media. The effects of high-tech character assassination can be very negative, including some wide-publicized suicides. Since such instances constitute sampling on the dependent variable, however, it is not clear what proportion of the presumably vast amount of negative postings lead to what kinds of results. If nasty cyber-communication is very widespread, (with surveys ranging from 20% to 50% of youth saying they have been targets: Sameer Hinduja, and Justin Patchin, Bullying Beyond the Schoolyard, 2009; National Crime Prevention Council, report Feb. 2007 ), the majority of victims may simply accept it as the new normal and shrug it off, or otherwise depending on their network position.

Is it bullying, or another kind of common conflict? Let us go through a brief check-list, reversing order and starting with the major types of two-sided fights:

Individual honor contests: conceivably the e-media can be used for public quarrels between two individuals, by means of which they get greater honor and eliteness. But the e-media are more widely participatory, giving audiences a chance to take part; unlike duels, where there is a sharp break between the audience who are supposed to keep their place, and the duelists in the center of attention, e-media audiences tend to spill into the fight itself, usually in an unrestrained and undignified way. Anonymity and lack of physical presence make it easy to do so. I conclude it is difficult to get elite status by fighting on the internet.

Intergroup fights: These horizontal brawls seem fairly common on the e-media. Much of what is called cyber-bullying may be of this sort.

Insult contests: In the early days of the Internet, so-called flame wars were common; given the opportunities for making long-distance connections with persons one is unlikely to ever meet, plus the use of pseudo-identities, insults were much more common than they are in everyday discourse. This is the opposite of what happens in real-life talk; conversational analysis (CA) concludes from a wide survey of evidence there is a preference for agreement in face-to-face encounters. The social media make it easier to spread negative messages. Here intergroup fights and insult contests collapse into the same thing, since there is little one can do in a fight in cyber space except make insults -- along with trying to release damaging information, or to hack the communications device itself, or its finances. IMPORTANT TANGENTIAL POINT: There is little in the pattern of hackers’ behavior to suggest they follow any of the conflict patterns I have listed. The topic of hackers remains a gaping hole in our sociological knowledge of the contemporary world.

Malicious gossip: Compared to direct word-of-mouth insult, indirect gossip has more of a Durkheimian quality, constructing the collective reality of a community defining an individual, especially if the network is large and its boundaries are vague, so circulated insults take on the anonymous and objective quality of “what everyone knows”. Gossip could be ganging up on an individual; but it is likely that much of the negative gossip spread by E-media is mutual recrimination among rival gossip networks. Probably most of this is horizontal, or even upward carping (as is most hacking), rather than the downward pattern of bullying.

Now for the forms of conflict that are asymmetrical, picking on an isolated individual:

Traditional bullying: there is little evidence of what proportion of cyber-negativity is repetitive aggression by habitual bullies against isolated, low-status victims too cowed to retaliate. Certainly there is no cyber equivalent of the original fagging pattern, where the bully made his victim into a servant, or a sexual slave as in prisons. Cyber attacks can hurt, but they seem incapable of forcing anyone to do their will. From particular cases, we know there is some genuine cyber-bullying; some of it adds cyber-mediated insults and rumours to face-to-face harassment, jeering, pushing and hitting. How much is bullying exclusively on-line, without personal contact? Research is needed to tell us which is worse, and what difference it makes.

Note that mediated bullying, or at least harassment, is not new. The tools of harassment include anonymous telephone calls (probably most prevalent in mid-20th century; some data is available since this is a category routinely collected in police reports, even now). This should remind us that all communications media, historically, could be used for harassment-- further back in time, it was poison-pen letters.

The chief difference with cyber media is that they are so widely networked that negative rumours can spread very rapidly, and leave permanent records, and hence the victim’s sense that a huge, impersonal collective consciousness has them skewered in its scornful attention. Thus the pressure of cyber-attacks may be stronger and more emotionally damaging than other kinds of mediated reputational attacks. But we don’t know this from systematic evidence; and it may not be true, since there are other kinds of exacerbating and mitigating factors in the realm of social relationships and resources. The research is yet to be done.

Scapegoating: the entire community, or a large segment of it, gangs up on a target of its outrage. This sounds like what happens when cyber-attacks go into a feeding frenzy, drawing in more and more participants. But maybe cyber-attacks only give the appearance of a community-wide feeding frenzy. Email can append long lists of recipients, and by carrying along a growing tail of previous messages, a dozen persons may generate the illusion of a huge number of messages, when in fact they are mostly recycling the same messages with additions.

I have personally observed such email-cascades develop on a half-dozen occasions during my presidency of the American Sociological Association in winter-spring 2011. (Most of these were campaigns from inside ASA membership; one was an attack on the ASA and its leaders from a right-wing political movement.) Not to say that these were all character assassination campaigns, but they shared the pattern of a flurry of messages being sent in a period of days, growing rapidly more importunate, denouncing a particular situation or policy and urgently demanding something be done. What I want to emphasize is a common pattern: once a critical mass was reached (sometimes after a slow start), messages came more rapidly and with more vehement content; but then the flurry dropped off again within 3-5 days. Upon careful inspection of each set of messages, I concluded that less than 20 people were sending and resending the great bulk of the messages. Especially during the up-phase, the process gave the impression that a huge and growing number of persons were involved, an exploding Durkheimian collective consciousness promising to engulf everything in its path. Where did this sudden bout of emotional enthusiasm and mutual entrainment come from, and why did it die off so rapidly? The timing of the emails (conveniently time-stamped for researchers) showed them getting closer and closer together, until the peak; thereafter, as the intervals between messages began to lengthen, the emotional urgency in their contents began to drop off precipitously, as did their numbers. At its height, a cyber-flurry exemplifies the “circular reaction” that Herbert Blumer and other collective behavior researchers have described for the flow of mutually supporting emotions in an excited crowd.

Although a set of people linked only through their computers or hand-held media devices lack the physical co-presence that I have argued is a precondition for a successful interaction ritual, it can generate a high level of collective effervescence when participants ramp up their sending and resending of messages to a rapid rate. At peak moments, I felt the excitement myself, even though I was more on the fending-off side than the side mobilizing the cascade, finding myself anticipating the period of hours, then minutes, when the next message would come in. When the pace slowed down, so did my excitement. Rhythmic entrainment generates emotional excitement, amplified by Durkheimian solidarity, as if we are all collectively pedaling a set of interlinked bicycles together, mutually rushing toward a speed-record. But as Durkheim noted, periods of collective effervescence are limited in time. The collective process of building excitement passes a second turning point (second after the critical mass of takeoff), where the sense of rushing forward together begins to flatten out, loses its enthusiasm and then begins to dissipate. The energy and entrainment depends on the sense that our collective enterprise is continually growing, adding more members (which, as noted, may well be an illusion of the cyber-format). This ephemeral community of communication relies on the sense of acceleration; when this palpably falls off, its jolt of emotional energy declines.

My inference is the following: cyber-bullying is really cyber-scapegoating, or rather a cyber-effervescent version in which a moderate size group of people become excitedly entrained in their common enterprise of trashing someone via E-media. Unlike bullying, where the chief link is between bully and victim, the former draining the latter of emotional energy and thereby getting a little status surge, in cyber-effervescent-scapegoating, the important emotional tie is among the posters of the negative messages. The victim is just a focal point, virtually a non-person, who serves only as content to circulate messages about.

A lurid example is the so-called “Kill Kylie” campaign. In 2004 a group of classmates ganged up on an eighth-grade girl in Vermont, proliferating websites and posts filled with homophobic attacks, and urging her to commit suicide. [, August 19, 2006; analysis in paper by Jason Haas, University of Pennsylvania, 2011.] In my terminology, this was probably pseudo-homophobic insult, since it is unclear that Kylie was gay; nor did she appear to be low status or a social isolate.

The middle-class kids who tried to drive her to her death might seem unusually vicious. But their behavior, I would suggest, is largely to be attributed to the collective effervescence of the cybernetwork experience, very likely a higher emotional rush than anything they had experienced previously. In this respect they are like members of lynch mobs, who often describe their experience retrospectively as unreal, a lapse from their normal consciousness. (In his lectures at Berkeley in 1964, Blumer described in this way a lynch mob that he observed in Missouri in the 1920s; see also “forward panic” in Collins, Violence, chapter 3). The perpetrators probably experienced their network scapegoating cascade more as antinomian than as evil, an exciting alternative reality, a festive holiday from morality (AKA “moral holiday”) but bound together in an emotional community of primitive Durkheimian solidarity. In short, they were doing it for the shared buzz. In lay terms, as some survey evidence suggests, kids do cyberattacks because they are fun.

Girls take part in cyber-bullying more than boys [Hinduja and Patchin, 2009]. This is not surprising, since girls do more scapegoating than bullying. An ironic conclusion is that girls’ greater concern for group solidarity makes them more attracted to the effervescence of cyber-scapegoating. The other major form of cyber-troublemaking, hacking, is much more the province of boys, and is oriented not to their group but to the disruption they can cause to authoritative organizations.

Bottom Line

Many different kinds of conflicts can take place in closed communities like schools, both in direct confrontation and via old and new media. Bullying has the most severe results for its victims, chiefly because they are in isolated network positions. Other kinds of conflict may actually generate a good deal of solidarity and meaningfulness for participants, albeit at the cost of some physical casualties and organizational disruption.

But bullying can only be recognized if one knows the location of participants in their social networks. Teachers may not have a very good sense of the network and status structure that is the context for any particular event of name-calling, exclusion or violence. It may seem that the best policy is simply to ban everything that is the slightest bit aggressive or negative. School administrators, who are even further from the action, are even less likely to know the social realities on the ground.

Kids themselves can generally tell the difference between the class bully being mean to an isolate, and playful teasing among friends, honor contests, or group rivalries. Officials trying to impose discipline by blanket orders, prohibiting all less-than-ideal-middle-class-politeness, may get a certain amount of surface compliance-- if they invest enough resources in monitoring. But such authorities also convince the kids that adults are rigid doctrinaires, clueless about what is really going on. The result may be nothing worse than to reinforce the normal suspiciousness on the part of the youth underground against official authorities. More seriously, it may make some kids feel they are being unjustly punished for acts misunderstood by self-righteous adults, reinforcing a spiral of alienation and defiance that is a component of criminal careers.

The practical advice may not be easy to carry out, but it is this: learn the network structure of the group, and judge all conflict in terms of its location.