Frenzied attacks. A micro-sociological analysis

Frenzied attacks. A micro-sociological analysis
of the emotional dynamics of extreme
youth violence1
Don Weenink
Inspired by phenomenological and interactionist studies of youth violence, this
article offers an empirical evaluation of Collins’s micro-sociological theory of
violence.The main question is whether situations of extreme violence have distinct
situational dynamics. Based on analyses of 159 interactions taken from judicial
case files, situations of extreme youth violence, here called frenzied attacks, were
identified on the basis of the state of encapsulation of the attackers (absorbed in
the violence, their sole focus is the destruction of the victim) and the
disproportionateness of the violence (the attackers continue to hurt the victims
even though they do not pose a threat or a challenge to them). Qualitative and
statistical analyses revealed that this emotional state results from a social figuration
in which the emotional balance shifts toward complete dominance of the
attackers. Thus, the occurrence of frenzied attacks is associated with the moment
victims hit the ground, paralyse and start to apologize, with the numerical dominance
of the attackers’ supportive group and with feelings of group membership,
in the form of solidarity excitement and family ties in the attackers’ group.Alcohol
intoxication is of influence as well, but contrary to the expectation, this effect was
independent from solidarity excitement.The article concludes that Collins’s theory
on the emotional dynamics of violence adds a new dimension to the phenomenological
and interactionist traditions of research.
Keywords: Violence; youth violence; micro-sociology; sociology of emotions
Following up on the review symposium on Collins’s (2008) micro-sociological
theory of violence in the British Journal of Sociology (Collins 2009; Cooney
2009; Felson 2009), this article offers an empirical evaluation of the theory as
Weenink (Rural Sociology Group, Wageningen University) (Corresponding author email:
© London School of Economics and Political Science 2014 ISSN 0007-1315 print/1468-4446 online.
Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden,
MA 02148, USA on behalf of the LSE. DOI: 10.1111/1468-4446.12088
The British Journal of Sociology 2014 Volume 65 Issue 3
applied to youth violence. More specifically, the research reported here
explores how and to what extent central concepts in Collins’s theory – forward
panics (here called frenzied attacks), weak victims, supportive audiences and
rapid shifts in the emotional balances – appear in youth violence. The main
question is whether situations of extreme youth violence have distinct properties
that sets them apart from other forms of youth violence. This question is
investigated through qualitative and quantitative analyses of 159 detailed
descriptions of violent interactions among Dutch youth, taken from judicial
case files. Inspired by a rich tradition of phenomenological and interactionist
studies, this article wishes to contribute to this knowledge by focusing on the
emotional dynamics of violence – an issue that has not been taken up explicitly
in this tradition so far (but see Luckenbill 1977).
The study at hand analyses violence in a narrow sense, as interactions in
which the physical damaging of human bodies by other human bodies is the
predominant focus of the action and attention. Narrowing the scope further,
the analyses are restricted to youth violence in public space. The plan is first to
explain Collins’s theory of the situational dynamics of violence, followed by a
summary of Cooney’s (2009) and Felson’s (2009) critical reviews and a discussion
of how the theory relates to prior phenomenological and interactionist
analyses of youth violence.The next section describes the data, the sample and
the analytical procedures. Subsequent sections present the results of qualitative
analyses, followed by a statistical assessment of the likelihood for violent
situations to turn extreme. Finally, the article highlights the main results and
offers new questions for future micro-sociological research on youth violence.
Collins’s micro-sociological theory of violence: concepts, critiques
and contextualization
In Violence.A Micro-sociological Theory, Collins (2008) lays out the emotional
dynamics of violence. Following Collins (2009: 569), the focus of attention on
the other that produces emotional attunement in normal interactions, is sustained
in antagonistic confrontations, but instead of attunement, now it results
in confrontational tension and fear (ct/f). As ct/f forms a barrier that inhibits
people from using violence, the overwhelming majority of confrontations
result in bluster and stalemates or people relieve ct/f by giving in, backing up
or leaving the scene (for more on ct/f, see Collins 2008 chapter 2). Violence
does occur when people circumvent the barrier of ct/f. This happens when
weak, passive victims give in to emotional dominance and/or when supportive
audiences help to build up emotional dominance. Collins’s (2008: chapter 3)
most pregnant explanation of the emotional dynamics of violence concerns his
conceptualization of forward panics. In these situations, attackers enter a state
of aggressive frenzy. Two features of forward panics are highlighted here. This
412 Don Weenink
© London School of Economics and Political Science 2014 British Journal of Sociology 65(3)
is first encapsulation: absorbed in the violent action, the destruction of the
opponents becomes the attackers’ sole focus of attention. Second, this form of
violence is disproportionate in the sense that the attackers seem unable to
stop: they continue to hurt the victims even though they do not pose a threat
or challenge to the attackers’ dominance anymore. Forward panics occur when
the emotional balance between two opponents suddenly and rapidly shifts, so
that the situation presents opportunities for total, uncontested domination by
one party. Collins regards these situations as panics, as they originate from
rapidly circumvented fear and confrontational anxiety.
The critical reviews by Cooney (2009) and Felson (2009) may help to specify
Collins’s conceptual model further. Note that the following discussion does not
provide a representative summary of their reviews, but only considers the
points that are relevant for the topic at hand.
First, Cooney and Felson do not agree with the idea that the occurrence of
violence is conditional on the presence of either weak victims or supportive
audiences alone. Cooney (2009: 588) argues that many antagonistic situations
in which these two conditions are met do not result in violence, and he raises
the question of how weak adversaries can be identified other than that they are
the defeated party. Felson (2009: 582–3) indicates that attacking others
depends on the rewards and costs involved. For instance, Jackson-Jacobs
(2004) shows that taking a beating as the underdog may result in ‘narrative
gratifications’: the construction of a story in which a lost fight evidences all
group members’ worthiness, as they dared to take on stronger opponents.
Furthermore, Felson (2009: 579, 584) touches upon the main question of this
article by asking whether and how forward panics can be distinguished from
other forms of violence. To summarize the argument so far, questions revolve
around the empirical appearances of forward panics and ct/f, weak victims and
supportive audiences, as well as with regard to the causal relationships
between them.
Relating these concepts to prior analyses of (youth) violence may help to
understand how they appear empirically. Consider first the role of ct/f . The
literature on gang violence indicates that in some deprived neighbourhoods,
‘respect’ (Anderson 1999: 10, 72; Jankowski 1991: 140) and ‘reputation’
(Sanders 1994: 71–5) based on instant aggressive responsiveness are part of
pre-emptive strategies to attain deference and personal safety (see also
Horowitz 1983: chapter 5; Jacobs and Wright 2006: 32–6; 41–2; 123–4). The
chronic threat of violence in these environments results into anxiety and fear,
an ‘ecology of danger’ where social interactions are perceived as threatening
or lethal (Wilkinson and Fagan 2001: 178), also among the ‘tough’ youth
(Trickett 2011). Jankowski (1991: 169) observed that as the level of fear and
anxiety rises among gang members, the severity and the scope of the violence
increases, which suggests that prolonged ct/f forms the emotional grounding of
extreme violence under these conditions.
Frenzied attacks 413
British Journal of Sociology 65(3) © London School of Economics and Political Science 2014
However, there are also indications that youth seek to experience violence
for the thrill of it, for instance among football hooligans (Dunning, Murphy
and Williams 1988: 209–10; Spaaij 2008) and in night life or leisure time
(Graham and Wells 2003; Katz 1988: chapter 3; Tomsen 1997). In these
instances of ‘recreational’ violence, prior building up of ct/f does not seem to
take place. Collins indicates that some situations of frenzied violence resemble
forward panics in form, although they have a different emotional grounding.
These pseudo forward panics – here called frenzied attacks – are driven by
intense feelings of group membership: ‘outbursts into collective violence, and
especially in the rhythmic, repetitive pattern that constitutes the overkill and
the atrocity, is so compelling to its participants because it constitutes an
extremely high degree of solidarity’ (Collins 2008: 130). Jankowski (1991:
171–2) explicitly notes the link between feelings of group membership and
violence that turns extreme:
when members [of a gang] act as part of a collective, they frequently go too
far, becoming caught up in the dynamics of group action rather than considering
the consequences of that action. Sometimes individual members
and the group find it difficult to determine when enough force has been
used, that is, when to quit. (See also Decker and Van Winkle 1996: 24)
To conclude: intense feelings of group membership, solidarity excitement, may
offer the alternative emotional pathway to frenzied attacks in youth violence.2
Feelings of group membership result from a shared focus of attention and
the experience of emotional attunement (cf. Collins 2004: chapter 2). To investigate
the relationship between the prior building up of solidarity excitement
and frenzied attacks, Hochstetler’s (2001) study of how robbery and burglary
groups develop a shared focus of attention is useful. Hochstetler (2001: 748–
52) identifies three ways of group alignment: ‘incremental signalling’ (the use
of small bodily or verbal moves to check whether the others are receptive to
the idea), ‘target convergence’ (mutual and instantaneous recognition of an
appealing target) and ‘establishing identity’ (recognizing and appealing to
group members’ reputations of criminal capability, thus turning past experiences
into expectations of upcoming action). In addition to developing a
shared focus of attention, the group members must also be emotionally
attuned to build up solidarity excitement. This happens when group members
start to express similar emotions, for instance humiliation or anger, or perhaps
joyous enthusiasm for violence as a possible line of action.
Turning now to how supportive audiences appear in prior studies of youth
violence. Studies of youth street violence (Jankowski 1991: chapter 5; Sanders
1994: 19, 56), research into violent football hooliganism and riots (Adang 1998:
18,40, 66) and analyses of violence during night life (Tomsen 1997: 98) indicate
that the violence is often committed by smaller subgroups of changing composition,
in the co-presence of the rest of their group, the audience. The
414 Don Weenink
© London School of Economics and Political Science 2014 British Journal of Sociology 65(3)
audience creates a stage by focusing the attention to the violence, by watching
it happen, or more intensively, by scolding and yelling or joining in (Felson
1982; Sanders 1994: 88–92; Tomsen 1997: 98). Felson (1982: 249, 52) found that
the presence of third parties alone, and also their encouragement, results into
more severe conflicts. In addition, third parties are more likely to take sides
when the opponents are younger (Felson, Ribner and Siegel 1984). Black
(1998: 125–43) argued that one important dimension that determines the
intensity of partisanship (taking sides) is the unequal relational distance of the
third parties vis-à-vis both opposing sides; the social closeness to one side and
the social remoteness from the other (Black 1998: 126; see also Cooney 1998:
69–70). Furthermore, unequal group ties (gangs or lineage), are associated
with stronger partisanship and more intense conflict as compared to unequal
individual ties (friends) (Cooney 1998: 132–4; Phillips and Cooney 2005: 351).
Considering group ties, Decker and Van Winkle (1996: 250–1, 255) reported
that most gang members found ‘blood’ (kinship) relations more important and
stronger than gang affiliations. Thus, it can be expected that family ties in
particular play a role in the occurrence of extreme violence. As the focus is
here on the role of third parties in gaining emotional dominance, this article
considers the dominance of the attackers’ supportive group versus that of the
victims.The supportive group consists of persons who are physically present at
the violent situation and who are members of the same group to which one of
the principals belongs. They do not engage in the physical violent action itself
but they may take sides verbally, as partisans, or they may watch and yell, as a
supportive audience that provides a stage.
As for weak victims, studies of violent football hooligans and street violence
demonstrate that youth seek to arrange situations in which they outnumber or
surprise the rival ‘firm’ or ‘gang’ (Jankowski 1991: chapter 5; Sanders 1994:
chapter 4 on drive-by shootings; Spaaij 2008). In addition, these literatures
suggest that beating up lonely or otherwise vulnerable victims by a small group
of attackers – ‘wolf packing’ (Keiser 1969: 50) – provides a way to attain a
reputation of toughness or ‘badness’ (Adang 1998 42–3; Katz 1988: chapter 3).
Jacobs and Wright (2006: 54) note that in ‘calculated retaliation’, the attackers
plan their moments of revenge when the victims are likely to be in a weaker
position, so that the duration and intensity of the violent punishment may be
Finally, one last aspect that should be taken into account in youth violence
is the role that alcohol plays.3 In their review of the literature on the relationship
between alcohol and violence (and drugs, but the former relationship is
much stronger), Parker and Auerhahn (1998: 303) indicate that, particularly
with regard to youth violence, the disinhibiting effects of alcohol are strongly
related to the places where alcohol is consumed (see also Graham, et al. 2006;
Graham and Wells 2001). Thus, the relationship between alcohol and youth
violence is particularly likely to appear in special zones – a time out, or moral
Frenzied attacks 415
British Journal of Sociology 65(3) © London School of Economics and Political Science 2014
holiday – in which conventional norms are suspended. Tomsen’s (1997)
account of a ‘top night’ highlights that in these zones of transgression, cultural
expectations link collective drinking to the experience of violence (see also
Graham and Wells 2003; Lister, Hobbs, Hall and Winlow 2000). Thus, the
relationship between alcohol intoxication and the occurrence of frenzied
attacks is here regarded as part of a collective ‘quest for solidarity excitement’,
to paraphrase Elias and Dunning (1986).
To conclude, prior work suggests that in areas where violence is not
endemic, youth violence is more likely to arise from solidarity excitement
rather than from circumvention of ct/f. Furthermore, the literature on youth
violence suggests that the other components of Collins’s theory – weak victims
and supportive groups – may turn out to be factors that contribute to the shift
toward emotional dominance that may result in the occurrence of frenzied
attacks. The effect of alcohol intoxication is expected to interact with that of
solidarity excitement. The argument is summarized in Figure I.
The following research questions attempt to evaluate Collins’s conceptual
framework, taking the above considerations into account. In the remainder of
this article, violent situations other than frenzied attacks are referred to as
controlled forms of youth violence. The term controlled here means that in
these types of violence, attackers are aware of their environment during the
attack and stop hurting the victim when they do not pose a threat to their
dominance any more.
1. How, and to what extent do frenzied attacks (as indicated by the attackers’
encapsulation and the disproportionateness of the violence) appear
in youth violence?
Figure I: Expected causal relationships between situational properties and the likelihood for
youth violence to turn into a frenzied attack, all relationships are positive
Weak victims
Shift toward emotional dominance Frenzied attack
Dominance of attackers’
supportive group
Family ties among attackers’ group
Solidarity excitement
Alcohol intoxication
416 Don Weenink
© London School of Economics and Political Science 2014 British Journal of Sociology 65(3)
2. What makes for weak victims in youth violence and to what extent do
they increase the likelihood of frenzied attacks?
3. What makes for the dominance of supportive groups in youth violence
and to what extent does it increase the likelihood of frenzied attacks?
4. What makes for rapid shifts in the emotional balances in youth violence
and to what extent do they increase the likelihood of frenzied attacks?
5. What makes for solidarity excitement prior to the attack (as indicated by
the attackers’ shared focus of attention and emotional attunement) and
to what extent does it increase the likelihood of frenzied attacks?
6. To what extent does the relationship between the attackers’ alcohol
intoxication and the likelihood of frenzied attacks interact with solidarity
excitement prior to the violence?
Data, sampling and analytical procedures
The data consist of judicial files of Dutch juvenile courts. Hence, the maximum
age of the attackers was 18 years at the time the violence was committed. The
files comprise interrogation reports of witnesses, offenders, victims and reports
of the courts and the public prosecutor. Furthermore, reports of the Council
for Child Protection provide additional information on the violent interactions,
including interviews with parents, teachers and the youth themselves.
Based on these files, a textual database was set up, with cases containing
detailed descriptions of the violent interaction. The names reported in this
article are fictitious.
The sampling of case files was based on a series of sections of Dutch penal
law considering violent offenses: public bodily harm (openlijke geweldpleging
tegen personen), battery or bodily harm (mishandeling, considered more
severe than openlijke geweldpleging), grievous bodily harm (zware
mishandeling), (attempt to) manslaughter, (attempt to) murder. Forms of violence
that are linked to economic (e.g. robbing) or ‘sexual’ (e.g. rape) exploitation
were excluded. Complexity was further reduced by restricting the
analyses to violence performed by small groups of about ten attackers at most.
Case files were drawn from archives of four judicial districts. Each archive
provided a list of cases settled by the juvenile court of that district in the years
1995 and 2005. The sampling comprised an interval selection of all the cases
referring to one or more of the above sections of Dutch penal law, taking each
second, third or fourth case on the list, depending on the total number of cases.
In sum, the sample consists of 159 violent interactions.
The aim of the qualitative analyses was to reveal how, in what forms, the
concepts discussed above appeared empirically and these were later coded as
numerical variables. The identification of frenzied attacks was grounded on
two sensitizing concepts: the state of encapsulation of the attackers (absorbed
Frenzied attacks 417
British Journal of Sociology 65(3) © London School of Economics and Political Science 2014
in the action, their sole focus being the destruction of the victims) and the
disproportionateness of the violence (the attackers continue to hurt the
victims even though they do not pose a threat or a challenge to the attackers’
dominance anymore). The following section demonstrates how these two sensitizing
concepts were identified in the data.The cases that met both criteria of
encapsulation and disproportionateness were coded as a binary variable
(0 = no, 1 = yes) that captures the concept of frenzied attacks.
As for weak victims, the qualitative analyses indicated that attackers predominantly
arranged the weakness of the victims by outnumbering them. For
all cases, it was noted how many attackers and how many victims were actually
involved in the violence. The latter number was subtracted from the former to
indicate the extent of outnumbering (ranging from −2 to 4). Considering the
dominance of the attackers’ supportive group, it was coded how many
members of the attackers’ group were present (excluding the ones who were
actually involved in committing the violence), and the same was done for the
victims’ group. To indicate the extent to which the attackers’ supportive group
outnumbered that of the victims, the latter number was subtracted from the
former (ranging from −6 to 9). Shifts in the emotional balances were identified
as the moments at which victims who had been trying to resist or counter
attack their opponent, gave up fighting, became paralysed or started to
apologize. It emerged from the qualitative analyses that a crucial tipping point
in this respect was when victims hit the ground. Thus, for all cases it was coded
whether the victim hit the ground or not (0 = no; 1 = yes).
The emergence of solidarity excitement in the attackers’ group (consisting
of the attackers themselves as well as their supportive group) was identified by
using Hochstetler’s (2001) conceptualization of action alignment as a sensitizing
concept for how groups try to build up a shared focus of attention (this will
be demonstrated below). In addition, expressions were coded that pointed to
a common emotional mood, for instance when group members shared feelings
of joy, anger, humiliation or anxiety. For each case it was coded whether both
a shared emotional mood and a shared focus of attention appeared (0 = no;
1 = yes). Furthermore, the presence of relatives among the attackers’ group
was coded as a binary variable (0 = no relatives, 1 = relatives present). The
degree of alcohol intoxication of the attackers was coded as a scale, consisting
of the following items: 0 = attackers had not used alcohol prior to the attack;
1 = attackers had used some alcohol prior to the attack (maximum 5 glasses of
alcohol or the witnesses, attackers or police officers reported that the alcohol
intake had no tangible impact on the attacker’s behaviour); 2 = attackers were
visibly under the influence of alcohol (6–15 glasses of alcohol or it was
reported they were under the influence of alcohol); 3 = attackers were heavily
under influence of alcohol (over 15 glasses of alcohol or it was reported they
were ‘very’, ‘completely’, or ‘heavily’ drunk or intoxicated, the alcohol inhibiting
normal speech, gait or other movements, or the attackers displayed a
418 Don Weenink
© London School of Economics and Political Science 2014 British Journal of Sociology 65(3)
strong inclination to sleep, directly after the attack). Finally, a control variable
was constructed that distinguishes male-on-male violence (1) from mixed
gender or female violence (0). In line with the argument of this article that
frenzied attacks have distinct situational properties that separates them from
other forms of violence, logistic regression analyses were used to assess the
impact of the above indicators on the likelihood for violence to turn into a
frenzied attack. The ordinal variables included in the regression analyses were
standardized given their varying scales of measurement. Inspection of VIF
values – the highest was 1.585 – indicated that problems of multicollinearity
were negligible.
Frenzied attacks in youth violence: the rage takes over
This section identifies how frenzied attacks appear in youth violence, based on
the notions of encapsulation and the disproportionateness of the violence. In
the first situation, a father and a son had been drinking beer for some time
after a football match. When the victim, Harold, bumped into the father by
accident, the father became angry and shouted that Harold should get out of
the way. The son entered the scene and then both started punching Harold.
Witness 1: They gave Harold a tremendously hard blow in his face. The
father hit him first and then the son joined him. Bystanders tried to pull the
father away, but Junior just carried on. Harold [fell] down on the ground but
Junior kicked him all the time in the face. It was terrible to see, it was bestial,
it was not normal how Junior kicked Harold. He kept on kicking him. I have
no words for how he kicked him. It would not have surprised me if he would
have kicked him to death. (case 57)
Consider first the disproportionateness of the violence. Witnesses frequently
report they were shocked at the view of attackers going berserk, seemingly
unable to stop (‘he kept on kicking him’, ‘I have no words for how he kicked
him’). They find it particularly disturbing and frightening that the attackers
engage in ferocious violence while the victim is defenceless and vulnerable.
Thus, the witness indicates that even though Harold had already hit the
ground, Junior ‘kicked him all the time’. However, it will be explained below
that it is exactly the vulnerability of the victim that unleashes the fury of the
attackers. What frightens witnesses as well is that the attackers are, in the
words of another witness in this case, ‘out of touch’: they have entered a state
of encapsulation. For instance, in the fragment above Junior was not concerned
about the efforts of the bystanders to separate him and his father from the
victim. In fact, the bystanders had difficulty to stop the attackers, the son in
Frenzied attacks 419
British Journal of Sociology 65(3) © London School of Economics and Political Science 2014
In the following situation, Kevin and Marc were cycling back to their homes
after a night in which they had been drinking large quantities of beer. They
encountered a boy coming from the opposite direction who commented assertively
on their wobbly cycling, as they provocatively made it hard for him to
pass them.After an initial phase of mutual scolding, calling names and shoving,
the victim stumbled and fell down to the ground. At that point, Kevin and
Marc entered a state of frenzy. The witnesses reported:
Witness 1 (acquaintance of the attackers): Kevin and Marc were shouting:
‘We break you, we cut you into pieces’. I ran to Kevin and asked him what
he was doing. [. . .] I tried to get between them, but I could not reach them,
they went completely berserk. I have seen some brawls, but not like this:
they could not stop with it.
Witness 2 (acquaintance of the attackers): It was disgusting. I was scared to
death. The boy was bleeding all over. It was so bad, I couldn’t bear it, it was
so dirty to look at. Kevin and Marc seemed insane; they kept on shouting ‘we
will kill you, we will finish you’, hitting and beating him, everywhere they
could. It seemed like it never stopped. (case 54)
The two features of frenzied attacks are clearly discernible in this fragment.
First, the witnesses point to the attacker’s disproportionate violence while the
victim was already down and unconscious (‘they could not stop with it’; ‘it
seemed like it never stopped’) and their encapsulated state (‘they went completely
berserk’, ‘I could not reach them’). In this situation, the attempts to
intervene failed and the violence stopped only when the attackers became too
tired. The fragment emphasizes the feelings of horror and fear that witnesses
often experience when they are confronted with a frenzied attack (‘it was
disgusting’, ‘dirty to look at’, ‘I was scared to death’). Typically, they use words
like ‘bestial’, or ‘insane’ to describe the state attackers are in. The insanity or
bestiality not only concerns the berserk violent action, but also the attackers’
imperviousness to communicate. Normal attuned interaction with them is
impossible as they seem obsessed, unable to pay attention to anything else
than their furious rage. Witnesses report that attackers not just wildly hit or
beat but also shouted, cursed and roared at the victim – even in cases where the
victims were already down and unconscious.These exclamations may proclaim
the nearing destruction of the victim, as in the fragment above:‘we cut you into
pieces’, ‘we will finish you’, or they simply indicate what the attackers are
doing:‘we kick him, we kick him’.Attackers themselves emphasize the feature
of encapsulation to describe what happened to them in a frenzied attack and
they tend to stress the compulsive nature of their actions (‘I could not stop’).
In the following fragment, attacker Nasir describes what happened to him
while he was kicking victim Edwin. Before this violent interaction, the two
boys had been humiliating and threatening one another for some time. This
had culminated in Edwin giving Nasir a head butt.A week later, attacker Nasir
420 Don Weenink
© London School of Economics and Political Science 2014 British Journal of Sociology 65(3)
was with his friends and saw that Edwin was alone, and he took his chances.
After chasing Edwin, he pulled him off his bike.
Nasir: Edwin fell down on the ground. When he was down, I started to kick
him severely. I kicked him everywhere I could. I kicked his ribs and head. I
was wearing heavy shoes with metal toe caps. I don’t know how often I
kicked him. But I kicked him many times. I heard Edwin moaning because
of the pain but I just went on with it. I was out of this world at that moment.
I was so angry at him. My friends were there, but they had nothing to do with
it. One of them said: ‘What are you doing, stop it’. But I couldn’t. (case 128)
During the beating up of Edwin, Nasir felt ‘out of this world’, indicating his
state of encapsulation.While he noticed the pain Edwin suffered and heard his
friend telling him to stop, he could not respond to these appeals. In a few cases,
the state of encapsulation is interrupted. In these short moments, attackers
suddenly understand what they are doing. These cases are of interest as they
illuminate what is not present in most frenzied attacks: a self that monitors the
attacker’s actions. Take for instance the following situation. The attackers had
already delivered some blows to the victim and now one of them (Linda) sat
on the victim, so that Sharleena could cut off her hair. Sharleena recalls what
happened to her at that moment:
Sharleena: So I took the scissors and cut off that girls’ hair.That girl couldn’t
do anything because Linda sat on top of her. I was going completely berserk.
Suddenly I realized I had her hair in my hands and that many people were
standing around me. But then I felt I had not duffed her up enough and I
wanted to beat her up again, but this time really good. (case 73)
The attacker shortly ‘realized’ that she actually had cut the hair of the victim
and that a crowd was watching her. Apparently, for Sharleena the situation
turned into a reality at the moment she could see herself busy. Nevertheless,
she soon felt an urge to launch another fierce attack. Also note that she was
going berserk after the ‘girl couldn’t do anything because Linda sat on top of
her’: this points to the importance of emotional asymmetry for the unleashing
of violent rage, an issue that will be taken up in the following section. In
another situation (case 142), a boy who tried to strangle a girl later reported
that ‘It was only when I heard her screaming that I saw my hand over her neck.
It was a compulsion which I could not resist’. These moments, in which attackers
understand what they are doing are, in Meadian terms, re-appearances of
‘Me’ that suddenly allows them to see what ‘I’ is doing. But these moments are
of short duration only and it may thus be concluded that the monitoring self of
the attacker is temporarily overruled by the rage in frenzied attacks. Thus, the
point is not so much that the violence is out of control. Instead, the violent rage
controls the attackers.
Frenzied attacks 421
British Journal of Sociology 65(3) © London School of Economics and Political Science 2014
Gaining emotional dominance: weak victims and supportive groups
Consider now how weak victims and supportive groups shift the emotional
balance toward the dominance of the attackers. With regard to weak victims, a
tipping-point can be discerned that demarcates the leap to a state of encapsulation
and disproportionate violence. This is the moment when victims hit the
ground. In the following fragment, two boys take revenge on a schoolmate,
Pieter, whom they accused of telling filthy stories behind their backs. After
chasing the victim by bike, the two attackers made him stop and started to
punch him.
Pieter: David came to me, he didn’t say anything and I felt [him] hit me hard.
I couldn’t fight back because they were with two. I fell down on the ground
and then the kicking started. Even though I was on the ground, they just
continued the beating. They started to kick me fiercely. I could not move, I
could not resist them. I felt pain everywhere and I thought they were going
to kick me to death. (case 44)
While the victim indicates that the two attackers started to kick him fiercely,
‘even though’ he was laying on the ground, the general pattern is that the
violence gets worse right after the moment the victim is down on the ground.
In this situation, the victim had shown a sign of weakness already before he
went down to the ground, as he indicated he ‘couldn’t fight back’. By giving up
resistance, the victim widened the emotional asymmetry, that in turn encouraged
the attackers to intensify their violence. Later, his state of paralysis (‘I
could not move, I could not resist them’) further increased the emotional
asymmetry. Indeed, the accounts of some attackers suggest that their anger
flares up when victims hit the ground. Thus, one of the attackers who had
beaten up Pieter reported:
David: Carlo joined me and he grabbed Pieter’s throat from behind and
smashed him to the ground. When he wanted to get up, I kicked him several
times. Carlo also kicked him, with a turning kick, that is a kind of karate
technique. Pieter then stumbled into a parked car and hit the ground again.
At that moment, I went really, really mad and started to kick him wildly.
(case 44)
The first phase of the physical violence started with David punching Pieter.
Then Carlo joined David, grabbed Pieter’s throat and ‘smashed’ him to the
ground. At that point, the victim was not completely defeated yet, as he tried
to get up. To prevent him from running away or taking a less vulnerable
position, the attackers started to kick Pieter. After the turning kick, at the
moment the victim stumbled into a parked car and could not stand on his feet,
422 Don Weenink
© London School of Economics and Political Science 2014 British Journal of Sociology 65(3)
his defeat was clear. His apparent weakness then set the rage in motion (‘I
went really, really mad’).Victims hitting the ground mark a decisive shift in the
emotional balance: now it is clear for both parties who will be dominating and
who will be dominated.Attackers who know this, purposively try to get victims
to the ground for this reason, and because it attracts co-attackers who seize the
opportunity for total dominance:
Robin: I saw that that boy was not yet down on the ground. I walked up to
him and gave him a punch with my elbow in his face. I did so because we
wanted to raise hell with the group. When he was down, I saw that Gradus
began to beat him up completely. This was so hard and fast, I was impressed
by it. (case 138)
When they are down on the ground, victims often confirm or increase the
emotional asymmetry by giving up fighting, by taking a foetal position, paralysing,
apologizing or begging for mercy. However, giving up fighting or
apologizing does not stop the attackers. Instead, it fires up the frenzy: when
resistance is over and complete dominance is within reach, the victims’ submissiveness
renders the attackers furious. The fragments discussed above
also demonstrate another form of arranging dominance: groups of youth
often work together to attack lonely victims or smaller groups. The outnumbering
is twofold. First, it concerns the number of group members who actually
commit the violence versus the number of victims who are actually
attacked, which makes for weak, outnumbered victims. Second, it comprises
the size of the supportive group of the attackers’ group versus that of
victims’ group, which makes for the dominance of the supportive group of
the attackers. Apart from their size, supportive groups may encourage the
attackers by focusing the attention on the violence, by just watching, by
making a lot of noise, or by cheering. For instance, they indicate that a
humiliating act should be revenged (‘are you accepting that?’). Other supportive
groups yell and shout to provoke more action (‘we want more, we
want more!’, ‘kick him, kick him!’, ‘let her crawl, let her crawl!’). In a few
cases, supportive groups suggested ways to degrade and humiliate the
victims: one victim had to kneel down, another had her hair burned, and in
some cases, the attackers were ask to rob the victims of their bikes or glasses,
which the attackers subsequently offered ‘for sale’.
Victims hitting the ground, the presence of weak, outnumbered victims and
the dominance of the attackers’ supportive group are expected to shift the
emotional balance toward the dominance of the attackers.The extent to which
these situational features impact the chance that violence turns into a frenzied
attack will be evaluated statistically below. But first, the appearances of solidarity
excitement among the attackers’ group (including both the attackers
and their supportive groups) will be discussed.
Frenzied attacks 423
British Journal of Sociology 65(3) © London School of Economics and Political Science 2014
Forms of solidarity excitement
The analyses presented here consider how groups develop a shared focus of
attention on violence and how they display a shared emotional mood. In the
following fragment, taken from a voluminous case file on a group of youth who
had committed a series of violent attacks during several consecutive nights,
two group members explain how the violence offered a way to experience fun
and feelings of group membership.
Micha: I was enjoying it at that time. I was acting tough for my friends.After
we had beaten up the guy we went back and had a laugh. We also boasted
about the fight.We told our friends what we had done.At that moment I was
proud of it.
Patrick: We went for the thrill of it and wanted to be tough, that’s why we
gave him a thrashing. We told our friends about this cool story. After that
fight we agreed to beat up someone else. I don’t know who started it, but we
used words like ‘we feel like bashing’. We meant to say that we would beat
up someone without a reason. We all knew, that someone would be beaten
up at the slightest occasion. We would just look at each other and go for it.
I have no idea how this can happen. I think I wanted to act tough. Like I was
part of the group. (case 75)
Several cues point to the presence of solidarity excitement, preceding, during
and after the violence. Consider first the strong focus of attention in the group:
they only needed to look at one another to know what would happen. Apparently,
‘target convergence’ (Hochstetler 2001: 750–1) was easy for this group.
One typical way of target convergence for this and other groups as well is to
yell and scold at passers-by. Another form occurs when group members indicate
that they had been hit, threatened or scolded at by the potential victim.
Most often, this talk is exaggerated or false, and the group knows this. What is
actually being communicated is that a suitable victim is found and that the plan
is to start a fight. Patrick’s reputation (‘whenever we fight, Patrick is there’)
increases the group’s awareness that the attack will be a likely course of action.
This is what Hochstetler (2001: 751–2) calls ‘establishing identity’: the recognition
of group members’ past behaviour is projected upon the upcoming
action. The fragment also indicates the group’s shared emotional mood. First,
there was a joyous spirit (‘we had a laugh’) and the group members apparently
sensed that they shared the same mood, as they used the plural in ‘we feel like
bashing’. Second, both attackers report they were acting tough for their
friends. Micha also indicates that he was proud of beating up the victim and
Patrick says he was part of the group. This means that the parts they played in
the violent action confirm their status as worthy members of the group. Finally,
both attackers indicate that the bashing was talked about later with their
friends. In so doing, the violence has turned into a symbol, providing a way to
424 Don Weenink
© London School of Economics and Political Science 2014 British Journal of Sociology 65(3)
revive the emotional mood again (Collins 2004: 98–9). Another example that
shows how solidarity excitement emerges prior to the attack is provided by the
following fragment. Now the collective mood is one of anger and humiliation.
After some minor confrontations between Turkish and Moluccan boys in the
night life of A. [name of town], the Moluccan boys wanted revenge when one
of them, Samuel, was slapped in the face by a Turkish boy, Kemal. The group
went to search Kemal, but when they could not find him, they readily agreed
on another Turkish victim. Here again, target convergence was easy, given the
already strong alignment of the group to commit violence. This is how the
group developed emotional attunement:
Samuel: When I got home I told the story [that he was slapped in the face by
a Turkish boy] to my nephews. Then Jonathan called. He asked me if I was
beaten by a Turk. He was very angry. Jonathan said:‘Let’s go to A. and show
me that Turk’. He also said that we shouldn’t talk to that Turk, but beat him
up right away. He said: ‘no talking, just trashing’. When I went to the bus
stop, Jonathan was there, and Nahesjon. Later came Ezra and Xander. I
heard from the older boys who joined us in A. that it was one Kemal who
had hit me. They said they were now forced to beat him up, as a Turk had
threatened and hit me. In the bus to A. we spoke about nothing else than
that I was beaten up by that Turkish boy. (case 51)
On their way to revenge, the group established a common mood of anger and
humiliation, as they kept on ruminating about the Turkish boy who had hit
Samuel. Furthermore, the older boys intensified this mood by stating they felt
obliged to pay back the insult. Note also that Jonathan, who was considered as
someone who easily went berserk in confrontations, took the lead in raising the
campaign to vengeance. In Hochstetlers’ terms, his reputation was part of
establishing the violent identity of the group. Related to this, the fragment
above also points to the presence of relatives among the attackers: not only
nephews but also brothers (Ezra and Xander) were joining the fight. In other
cases, fathers and sons or brothers explicitly put kinship ties at stake. For
instance they challenge the opponent to ‘not ever touch my father’ or they
refer to a shared reputation as fighters: ‘show me you are a real Boender’
(referring to the nick name of the family). The degree to which solidarity
excitement and the presence of relatives in the attackers’ group impact the
chance that frenzied attacks occur will be considered below.
Frenzied attacks versus controlled forms of violence
The results of the above analyses indicate that frenzied attacks are qualitatively
different from other forms of youth violence. Moreover, the analyses
suggest that there are specific situational properties that increase the
Frenzied attacks 425
British Journal of Sociology 65(3) © London School of Economics and Political Science 2014
likelihood for violence to turn into frenzied attacks. These findings will be
assessed statistically in this section.Table I compares the situational properties
of frenzied attacks versus controlled forms of violence.
In nearly two-thirds of all frenzied attacks, the victims had hit the ground,
while this is only slightly over 15 per cent in forms of controlled violence.
Concerning the number of attackers, it turns out that in frenzied attacks (2.43),
there are more attackers than in controlled forms of violence (2.01), a marginally
significant difference only. The number of victims is about the same in
the two forms of violence. While weak, outnumbered victims feature in both
forms of violence, they are more likely to be found in frenzied attacks; there
are on average 1.21 more attackers than victims in the latter form of violence,
compared to 0.75 more attackers in controlled forms of violence, a significant
difference. While the average size of the attackers’ supportive groups does not
differ significantly between the two forms of violence, that of the victims’
supportive groups is only 0.43 in frenzied attacks and 1.35 in controlled forms
of violence. This finding points to the tendency to single out lonely victims in
frenzied attacks.As a result, the dominance of the attackers’ supportive groups
differs significantly between the two forms of violence: in frenzied attacks, the
supportive groups of the attackers includes on average two more people, while
in controlled forms of violence the difference is only 0.36 on average. Note that
the standard deviations are large, indicating substantially dispersed variation
within these categories. Both solidarity excitement prior to the violence and
the presence of relatives among the attackers’ group appear significantly more
often in frenzied attacks (respectively 42.9 per cent versus 18.3 per
Table I: Descriptive statistics of the situational properties of frenzied attacks (N = 28) versus controlled
forms of violence (N = 131), frequencies (percentages) and means (standard deviations)
Significance of
Victims hit ground 18 (64.3%) 20 (15.3%) P < 0.001
Number of attackers (a1) 2.43 (1.14) 2.01 (1.06) P = 0.062
Number of victims (b1) 1.21 (0.42) 1.25 (0.63) P = 0.750
Degree of weak, outnumbered
victims (a1 minus b1)
1.21 (1.03) 0.75 (1.05) P = 0.035
Size of attackers’ supportive group (a2) 2.43 (2.46) 1.71 (2.02) P = 0.102
Size of victims’ supportive group (b2) 0.43 (0.79) 1.35 (1.53) P = 0.000
Dominance of attacker’s supportive
group (a2 minus b2)
2.00 (2.60) 0.36 (2.36) P = 0.001
Solidarity excitement prior to the attack 12 (42.9%) 24 (18.3%) P = 0.007
Relatives in attacker’s group 10 (35.7%) 15 (11.5%) P = 0.003
Alcohol intoxication attackers 1.15 (1.43) 0.49 (0.88) P = 0.028
Male-on-male 25 (89.3%) 107 (81.7%) P = 0.251
To determine significance of differences, one sided Chi-square tests were used for dichotomous
variables and T-tests for ordinal variables, using Levene’s F-test to determine equality of variances.
Due to missing values, the analysis of alcohol intoxication is based on 150 cases, including 27
frenzied attacks.
426 Don Weenink
© London School of Economics and Political Science 2014 British Journal of Sociology 65(3)
cent and 35.7 per cent versus 11.5 per cent). Also, attackers had consumed
more alcohol in frenzied attacks. Overall, however this indicator scores rather
low, a score of one means that only five glasses of alcohol maximum were
consumed, or it was reported that the alcohol did not have tangible effects.
While the fragments above often suggested a considerable degree of intoxication,
a great deal of violence in the sample is not so much related to heavy
drinking apparently. Finally, the proportion of male-to-male violence does not
differ significantly between the two forms of violence.
Table II displays the relative impact of the situational properties on the odds
that violence turns into frenzied attacks. Due to missing information on the
alcohol intoxication of the attackers, nine cases were dropped from the
sample.4 The first model shows that both the alcohol intoxication of the attackers
(odds ratio 1.544) and the dominance of the attackers’ supportive groups
(odds ratio 1.857) significantly increase the likelihood that a frenzied attack
occurs. However, the presence of weak, outnumbered victims does not result
into significant effects. In model 2, both the solidarity excitement prior to the
violence (odds ratio 2.888) and the presence of relatives in the attackers’ group
(odds ratio 6.808) result in significant and substantial effects. Contrary to the
expectation, including solidarity excitement does not alter the effect of alcohol
intoxication substantially.To analyse whether it makes a difference if solidarity
excitement is fuelled by alcohol or not, model 3 includes an interaction term
that combines both indicators. However, it does not result in a significant
effect, nor does it contribute to the model’s explanatory power (χ2 = 1.620,
df = 1, p = 0.203). Contrary to what was expected, alcohol intoxication and
solidarity excitement do not appear as two interrelated ingredients of a quest
for violent excitement. Finally, model 4 supports the idea that victims hitting
the ground are a decisive tipping point, as it yields a substantial and highly
significant effect (odds ratio 10.631). Including this indicator reduces the
effects of alcohol intoxication and solidarity excitement and turns them nonsignificant.
Furthermore, model 4 shows that the impact of the presence of
relatives in the attackers’ group slightly increases when the moment victims
hitting the ground is taken into account.
Although these results should be interpreted with some prudence given the
size of the sample and the distribution of the binary variables (see Table I), the
statistical analyses support the idea that frenzied attacks have distinct situational
properties that sets them apart from controlled forms of violence.
Moreover, they suggest that the numerical dominance of the supportive group
of the attackers is of greater importance than the presence of weak, outnumbered
victims. This result remains throughout the analyses. While solidarity
excitement and alcohol intoxication result into significant effects, their impact
is reduced when the model includes whether victims fall down to the ground.
However, as the qualitative analyses suggested, groups may, perhaps driven by
the influence of alcohol and solidarity excitement, try to attain situational
Frenzied attacks 427
British Journal of Sociology 65(3) © London School of Economics and Political Science 2014
Table II: Logistic regression analysis of the likelihood for youth violence to turn into a frenzied attack (N = 150)
Model 1: alcohol intoxication+ outnumbering +
supportive groups
Model 2: model 1 + solidarity
excitement + relatives
B Standard error Odds ratio B Standard error Odds ratio
Constant −2.158 0.637 0.116** −2.878 0.718 0.056**
Male-to-male 0.566 0.683 1.761 0.608 0.720 1.836
Alcohol intoxication attackers 0.435 0.208 1.544* 0.455 0.221 1.576*
Degree of weak, outnumbered victims 0.279 0.227 1.322 −0.016 0.264 0.984
Dominance of attackers’ supportive group 0.619 0.229 1.857** 0.598 0.244 1.818*
Solidarity excitement prior to the attack 1.061 0.548 2.888∼
Relatives in attacker’s group 1.918 0.594 6.808**
Solidarity excitement x alcohol intoxication
Victims hit the ground
Nagelkerke pseudo R2 0.195 0.311
Model 3: model 2 + interaction solidarity
excitement x alcohol
Model 4: model 2 + victims
hit the ground
B Standard error Odds ratio B Standard error Odds ratio
Constant −2.834 0.719 0.059** −3.968 0.866 0.019**
Male-to-male 0.585 0.722 1.794 1.080 0.794 2.944
Alcohol intoxication attackers 0.240 0.287 1.271 0.262 0.252 1.300
Degree of weak, outnumbered victims −0.030 0.264 0.971 −0.372 0.313 0.690
Dominance of attackers’ supportive group 0.552 0.248 1.737* 0.646 0.263 1.907*
Solidarity excitement prior to the attack 0.896 0.589 2.451 0.591 0.613 1.805
Relatives in attacker’s group 1.954 0.590 7.053** 2.034 0.672 7.642**
Solidarity excitement x alcohol intoxication 0.599 0.484 1.820
Victims hit the ground 2.364 0.619 10.631**
Nagelkerke pseudo R2 0.325 0.443
∼ P = 0.053; * P < 0.05; ** P < 0.01.
428 Don Weenink
© London School of Economics and Political Science 2014 British Journal of Sociology 65(3)
dominance exactly by getting the victims to the ground. Finally, the presence of
relatives in the attackers’ group has marked effects. It may very well be that for
the fathers, sons, brothers and nephews in this sample, violent situations are
tests of kinship loyalty.
The most horrendous forms of youth violence are the result of a social figuration
that offers opportunities for total, uncontested dominance. When this
happens, the rage takes over; the attackers become encapsulated and start to
use disproportionate violence. In symbolic interactionist terms, the attackers’
capacities for self-monitoring are temporarily shut off, as ‘I’ is no longer
attended to by ‘Me’. Three features make up the social figuration in which
these emotional dynamics take place. First, the moment victims hit the ground
forms a tipping point. Not only are victims less threatening and more vulnerable
in such position, often they also become passive and paralysed, start to
apologize and beg for mercy, thus contributing to the emotional asymmetry.
The second feature concerns the presence of supportive groups who do not
engage in physical violence themselves. Earlier studies found that the presence
of third parties in general (Felson 1982; Felson and Steadman 1983) and group
ties in particular (Phillips and Cooney 2005) increase the intensity of conflicts.
The findings reported here are in line with these earlier results as they show
that the presence of relatives among the attackers’ group strongly contributes
to the likelihood for violence to turn extreme. This article also extends these
prior findings, as it shows the importance of the numerical dominance of the
attackers’ supportive group vis-à-vis those of the victims’. The third feature of
the social figuration that makes frenzied attacks a more likely outcome concerns
the solidarity excitement among the attackers’ group.While prior studies
of youth street violence and gangs often point to solidarity as a core cultural
value, Gould (1999: 376) concluded that solidarity should be evidenced in
action instead of relying on the ‘rhetoric of kinship’ as it is in real confrontations
that group members have to live up to their claims of solidarity. Combining
Collins’s (2004) and Hochstetler’s (2001) insights, this article aimed to
open up solidarity by analysing it as situationally emergent group feelings that
focus on violence as a future line of action.
This study offers empirical support for the central ideas on which Collins’s
(2008) micro-sociology of violence is based. This is, first, the notion that
extreme violence is different from other forms of violence because of the
specific emotional state the attackers are in. Second, the idea that this emotional
state is the result of a social figuration in which the emotional balances
are shifting toward the total dominance of the attackers. Third, the notion that
weak victims and supportive groups play a crucial role in the shifting of
emotional balances.
Frenzied attacks 429
British Journal of Sociology 65(3) © London School of Economics and Political Science 2014
Finally, this study raises a series of new questions. First, the finding that the
numerical dominance of supportive groups affects the course of violent events
offers a starting point for more elaborated analyses of the role of these groups.
Such work should also follow up on the earlier work on third parties and
partisans mentioned above (see the suggestions made by Phillips and Cooney
2005). In particular, the finding that the dominance of the attackers’ supportive
group is of greater importance than the outnumbering of victims by attackers
themselves needs further unpacking. Close-up analyses of how supportive
groups focus the attention on the violence, how they turn it into a stage and
how they intervene in the interaction will provide useful insights in this
Second, apart from the distinct focus on the emotional dynamics, the
approach followed here offers an alternative view on violence in another
respect. Often, it is found that violence follows a strategic or moral logic.
Thus, violence as retaliation (Jacobs and Wright 2006), as a way to maintain
reputation or respect (Anderson 1999; Decker and Van Winkle 1996; Sanders
1994), or to defend honour (Cooney 1998: chapter 5; Horowitz 1983) all
follow a survival logic of (pre-emptive) protection, to deter future attacks in
unsafe environments. Alternatively, the explanation of violence as a form of
social control to settle grievances, to even wrongdoings or to manage conflict
(Black 1983) follows a moral logic. Instead, this article departs from the idea
that violence follows the logic of dominance. Emotional asymmetry, rather
than survival strategy or morality is supposed to rule the interaction. While
the focus of this article was not on the meanings of violence, the qualitative
analyses do show that at least in some instances, victims are picked opportunistically,
and that attackers can do so without or with fake moral justifi-
cations (e.g. false accusations) that serve to focus the attention of the group
toward violent action. These forms of violence are neither moralistic nor
strategic, and they do not seem to involve conflict in the sense that the opponents
have antagonistic interests either. In relatively pacified environments
where formal law or other adult authority is generally accepted as a means to
manage conflicts, youth violence may not be as moralistic or strategic as it is
often assumed. Under these conditions, youth may be much more likely to
turn to authorities and seek ‘legal partisanship’ (Black 1998: 139–41) for protection
or conflict management, as compared to their counterparts who form
gangs in unsafe environments (Decker and Van Winkle 1996). The violence
that remains may be of predatory nature more often. Thus, one question
for future research is whether the solidarity excitement that is associated with
extreme violence, is more often joyous rather than angry. Another question is
whether (extreme) youth violence under these conditions is more likely to
be part of a quest for excitement (Elias and Dunning 1986), to experience
intense emotions and bodily sensations, rather than as part of social
430 Don Weenink
© London School of Economics and Political Science 2014 British Journal of Sociology 65(3)
Third, subsequent work could try to answer questions that relate the meanings
of violence to the emotional dynamics. For instance, does it make a
difference when victims hit the ground if the violence is a way to express
respect or alternatively, if it is a punishment? And how does the meaning of
violence relate to the development of solidarity excitement? Felson (2009), in
his review of Collins’s theory, argues that violence should be explained as
instrumental behaviour, driven by (emotionally) weighted costs and rewards.
However, perspectives that regard violence as intentional, instrumental and
meaningful behaviour do not necessarily exclude approaches that focus on the
emotional dynamics of violent situations. So, probably the greatest advantage
of Collins’s theory is that social-scientific researchers of violence now have a
coherent conceptual perspective that allows them to see how violence as
meaningful interaction is structured by situational emotional dynamics.
(Date accepted: October 2013)
1. My thanks to Randall Collins, Ali de
Regt and the anonymous reviewers for their
valuable comments on earlier versions of
this article. I also thank the Netherlands’
Organization for Scientific Research for
supporting this research (grant number 016.
2. The term solidarity excitement is used
here instead of Durkheim’s (2001 [1912])
‘collective effervescence’. Although describing
a similar emotional process of group
solidarity, the latter term originally denoted
the ecstatic high moments in which groups
forge religious, moral or other mental
images and representations (Durkheim 2001
[1912]: 158, 164) whereas solidarity excitement
may not produce such lasting results
(see also Collins 2004: chapter 2, on collective
effervescence and the creation of group
3. It can be argued that alcohol consumption
makes attackers slow and clumsy, so
that the most ferocious attacks are probably
not committed by intoxicated attackers (see
Collins 2008: 268). However, the argument is
here that the effect of alcohol consumption
is part of a broader social process that builds
up to solidarity excitement. On the other
hand, alcohol intoxication might make for
weak victims (Felson and Steadman 1983:
65, 67, 72). Additional analyses in which the
victims’ degree of alcohol intoxication was
included (measured in the same way as the
attackers’ intoxication) did not result in a
significant effect of this variable.
4. Additional analyses in which these nine
missing values were replaced with the mean
yielded similar results and explanatory
Adang, O. 1998 Hooligans, Autonomen,
Agenten: Geweld en Politie-optreden in
Relsituaties, Alphen aan de Rijn:
Anderson, E. 1999 Code of the Street.
Decency, Violence and the Moral Life of the
Inner City, New York: W.W. Norton &
Frenzied attacks 431
British Journal of Sociology 65(3) © London School of Economics and Political Science 2014
Black, D. 1983 ‘Crime as Social Control’,
American Sociological Review 48 (1): 34–45.
Black, D. 1998 The Social Structure of Right
and Wrong, revised edition, San Diego: Academic
Collins, R. 2004 Interacton Ritual Chains,
Princeton: Princeton University Press.
Collins, R. 2008 Violence: A MicroSociological
Theory, Princeton: Princeton
University Press.
Collins, R. 2009 ‘The Micro-Sociology of
Violence’, British Journal of Sociology
60(3): 566–76.
Cooney, M. 1998 Warriors and Peacemakers.
How Third Parties Shape Violence. New
York: New York University Press.
Cooney, M. 2009 ‘The Scientific Significance
of Collins’s Violence’, The British Journal of
Sociology 60(3): 586–94.
Decker, S. H. and Van Winkle, B. 1996 Life
in the Gang. Family, Friends and Violence,
Cambridge: Cambridge University Press.
Dunning, E., Murphy, P. and Williams, J.
1988 The Roots of Football Hooliganism.An
Historical and Sociological Study, London
and New York: Routledge & Kegan Paul.
Durkheim, E. 2001 [1912] The Elementary
Forms of Religious Life, Oxford: Oxford
University Press.
Elias, N. and Dunning, E. 1986 Quest for
Excitement. Sport and Leisure in the Civilizing
Process, Oxford: Blackwell.
Felson, R. B. 1982 ‘Impression Management
and the Escalation of Aggression and Violence’,
Social Psychology Quarterly 45(4):
Felson, R. B. 2009 ‘Is Violence Natural,
Unnatural, or Rational?’, The British
Journal of Sociology 60(3): 577–85.
Felson, R. B., Ribner, S. A. and Siegel, M. S.
1984 ‘Age and the Effect of Third Parties
During Criminal Violence’, Sociology and
Social Research 68(4): 452–62.
Felson, R. B. and Steadman, H. J. 1983 ‘Situational
Factors in Disputes Leading to Criminal
Violence’, Criminology 21(1): 59–74.
Gould, R. V. 1999 ‘Collective Violence and
Group Solidarity: Evidence from a Feuding
Society’, American Sociological Review
64(3): 356–80.
Graham, K., Bernards, S., Osgood, D.W. and
Wells, S. 2006 ‘Bad Nights or Bad Bars? Multilevel
Analysis of Environmental Predictors
of Aggression in Late-night Large-capacity
Bars and Clubs’,Addiction 101(11): 1569–80.
Graham, K. and Wells, S. 2001 ‘Aggression
Among Young Adults in the Social Context
of the Bar’, Addiction Research & Theory
9(3): 193–219.
Graham, K. and Wells, S. 2003 ‘ “Somebody’s
Gonna Get Their Head Kicked in
Tonight!” Aggression among Young Males
in Bars – A Question of Values?’, British
Journal of Criminology 43(3): 546–66.
Hochstetler, A. 2001 ‘Opportunities and
Decisions: Interactional Dynamics in
Robbery and Burglary Groups’, Criminology
39(3): 737–63.
Horowitz, R. 1983 Honor and the American
Dream, New Brunswick, N.J.: Rutgers University
Jackson-Jacobs, C. 2004 ‘Taking a Beating:
The Narrative Gratifications of Fighting as
an Underdog’ in J. Ferell, K. Hayward, W.
Morrison and M. Presdee (eds) Cultural
Criminology Unleashed, London: The
GlassHouse Press.
Jacobs, B.A. and Wright, R. 2006 Street
Justice. Retaliation in the Criminal Underworld,
Carmbridge: Cambrdige University
Jankowski, M. S. 1991 Islands in the Street:
Gangs and American Society, Berkely: University
of California Press.
Katz, J. 1988 Seductions of Crime.Moral and
Sensual Attractions in Doing Evil, New York:
Basic Books.
Keiser, L.R. 1969 The Vice Lords, New York:
Holt, Rinehart and Winston.
Lister, S., Hobbs, D., Hall, S. and Winlow, S.
2000 ‘Violence in the Night-Time Economy;
Bouncers: The Reporting, Recording and
Prosecution of Assaults’, Policting and
Society 10(4): 383–402.
Luckenbill, D. F. 1977 ‘Criminal Homicide as
Situated Transaction’, Social Problems
25(2): 176–86.
Parker, R. N. and Auerhahn, K. 1998
‘Alcohol, Drugs and Violence’, Annual
Review of Sociology 24: 291–311.
432 Don Weenink
© London School of Economics and Political Science 2014 British Journal of Sociology 65(3)
Phillips, S. and Cooney, M. 2005 ‘Aiding
Peace, Abetting Violence: Third Parties and
the Management of Conflict’, American
Sociological Review 70(2): 334–54.
Sanders, W. B. 1994 Gangbangs and
Drive-Bys: Grounded Culture and Juvenile
Gang Violence, New York: Aldine de
Spaaij, R. 2008 ‘Men Like Us, Boys Like
Them: Violence, Masculinity, and Collective
Identity in Football Hooliganism’, Journal of
Sport & Social Issues 32(4): 369–92.
Tomsen, S. 1997 ‘A “Top Night”: Social
Protest, Masculinity and the Culture of
Drinking Violence’, British Journal of Criminology
37(1): 90–102.
Trickett, L.F. 2011 ‘ “Fears of the Fearless” ’,
International Journal of Law, Crime and
Justice 39(4): 280–302.
Wilkinson, D. L. and Fagan, J. 2001 ‘A
Theory of Violent Events’ in R. F. Meier, L.
W. Kennedy and V. F. Sacco (eds) The
Process and Structure of Crime, New
Brunswick, NJ: Transaction.
Frenzied attacks 433
British Journal of Sociology 65(3) © London School of Economics and Political Science 2014
Reading Note – Current Issues II
Reading Group:

  • Provide a concise summary (approx. 250 words) of the main ideas of the text

Abstract: …

  • Choose one core quotes that represents the central argument of the text. Briefly discuss why (150 to 200 words).

Quote: ….

  • Formulate two questions you raise after reading the text. These may be questions regarding the content of the text, or critical considerations you wish to discuss

Question: …
Link / Application

  • Connect what you have read in the text with something you picked up in your other courses OR apply it to a contemporary issue (Trump and Russia, migration, etc.)

Link / Application: …

Place this order or similar order and get an amazing discount. USE Discount code “GET20” for 20% discount

Posted in Uncategorized