Show Summary Details

Page of

Printed from Oxford Research Encyclopedias, Communication. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 04 March 2021

Collective Protest, Rioting, and Aggressionfree

  • Stephen ReicherStephen ReicherSchool of Psychology and Neuroscience, University of St. Andrews


In understanding crowd psychology and its explanation of conflict and violence, there are different theoretical approaches that turn on different understandings of communication processes. There are three models of communication in the crowd worth reviewing: classic, normative, and dynamic. Classic models suggest that crowd members are influenced by an idea of emotion presented to them. Normative models suggest that influence is constrained by what is seen as consonant with group norms. And, finally, dynamic models examine how that which becomes normative in the group depends upon intergroup relations. The last of these approaches can explain the patterned, socially meaningful and yet changing nature of crowd action. Crowd action, itself, is a form of communication because it serves to shape the social understandings of participants as well as the social understandings of those beyond the crowd. It is argued that the nature and centrality of crowds contribute to the understanding and creating of social relations in society.


Issues of group communication are central to the development of crowd psychology. Indeed, the very birth of this area can be said to arise out of a crisis of communication.

The impetus came from the process of industrialization in Europe during the 19th century—a process that fundamentally changed relations between different layers of society. The face-to-face contact of village life gave way to the anonymity of the cities. The intimacy of master and craftsman was obliterated by a separation between factory owners and toiling masses. In the new mass society, different classes were separated physically as well as ideologically, living in different parts of town, rarely venturing into the territory of the other (Steadman Jones, 1983).

Through this radical separation, this almost total loss of communication, the population became unknown to the elite. This provided a space for fears and fantasies to grow. What dangerous ideas and dangerous practices would incubate and spread? Which unscrupulous agitators would be let loose to turn the masses against their masters? Freed from the controlling gaze of authority, how could the submission of the majority to the minority be assured and how could the social order be preserved (Giner, 1976)?

If the mass was a potential threat to all that the elite held dear, the crowd was that potential made actual. It was the moment when the masses would actually rise up, run amok, and destroy the social order. As a result the crowd became a dense and over-determined symbol of mayhem and destruction. Just as independent women threatened the micro-social order of the patriarchal family, so the feminine and emotional crowd threatened the rational macro-social order of the nation. Just as mother gin and beer, wine, and spirits threatened the discipline of the home and the workplace, so drunken and uncontrolled crowd threatened the discipline of the state (Barrows, 1981). The crowd became a figure today occupied by the terrorist, a figure of pure unmitigated negativity.

Given this background, it is understandable that crowd psychology developed first in those countries that felt most vulnerable to collective challenges—France and, to a lesser extent, Italy. Following defeat in the Franco-Prussian war, in 1871 France saw the temporary victory of the Paris Commune, dubbed the first socialist society in history (Merriman, 2016). Here, the threat of the masses, and of the masses in action, had passed from the realm of fantasy to reality. The elite had seen the future and they didn’t like it. What is more, the French Third Republic, which grew on the ashes of the Commune, was weak and unstable, beset on all sides by collective agitation whether it be clericalist, populist, socialist, or syndicalist (Nye, 1975).

With all this going on, crowd psychology was not an arcane subject confined to the pages of obscure academic journals. It was debated in the society magazines of the day. Moreover, the discussion was eminently practical rather than theoretical. The first debate in crowd psychology concerned how best to repress crowds. It pitted the Italian Scipio Sighele against the Frenchman Gabriele Tarde and concerned how best to determine criminal responsibility. Is it that everyone in the crowd knows what they are doing, is a bad person, and deserves sanction, or is it that most have lost control of themselves, are manipulated by demagogues, and, hence, only these latter should be subject to the might of the law (van Ginneken, 1992). To this day, the debate has not been fully resolved.

For now, the point to be drawn from all this is that the overwhelming focus of crowd psychology was on crowds as danger. That remains true to this day. Interest in crowds, reporting of crowds, and funding of crowd research spikes after violent episodes (e.g., Reicher & Stott, 2011; U.S. Riot Commission, 1968). Crowds become a topic of concern only when they are dangerous to the status quo. This leads to a highly distorted understanding of crowd phenomena and hence to highly distorted models of crowd process.

The birth site of crowd psychology, fin de siècle France certainly was a contentious place. Barrows (1981) reports the quadrupling of strikes from the 1870s to the 1890s with at least 1,894 incidents between 1890 and 1894—more than one a day. But in only a tenth of these was there any collective demonstration, in only 3.6% was there any violence, and in only one was anyone killed in the crowd. This was the murder of a particularly hated official during a miner’s strike in 1886—the so-called “defenestration of Decazeville.”

Yet, as Barrows goes on to argue, it was this single event: “that was to obsess Le Bon, Tarde, Sighele, Fournial and other writers on the crowd” (1981, p. 21). The exception thereby became the rule. Explanation focused on accounting for, and indeed naturalizing, such exceptional violence. Crowd psychology elided crowds with riots with aggression. The one necessarily entailed the others. The problem with this approach is not simply that it allows the outsider’s obsession with danger to obscure all other aspects of crowd phenomena. It also impedes an understanding of danger and violence itself. For if one takes violence for granted one cannot then explain when it does and when it doesn’t occur. Why are some events violent and others not? What is it that leads to the emergence of violence within a given event? Once violence does occur, what are its targets and what forms does it take? And what are people seeking to do when they act violently?

Paradoxically, then, the key to understanding rioting and crowd aggression is to make it less of a focus. It is to become a little more perplexed and surprised when crowds are violent (rather than tutting with a resigned “crowds will be crowds”). It may be true that there are many events in which some crowd members advocate confrontation, but it is rare that their voices gain traction. So what has to happen for them to command an audience? Such a shift, it can be argued, involves a fundamentally different approach to communicative processes. The argument proceeds in two parts. Different approaches to communicative processes within crowds will be examined, along with how they are used to explain crowd behavior in general and crowd violence in particular. How crowd behavior, including crowd violence, can itself be understood as an act of communication will be discussed.

Communication Processes and the Action of Crowds

The Contagion Approach: Le Bon’s Classic Crowd Psychology

Of all the early writers on the crowd, one name stands out and endures—that of Gustave Le Bon. His 1895 book The Crowd has been described as the most important psychology text of all time, not simply because of the way it influenced understandings of the crowd but also because of the way it shaped the mass politics of the 20th and 21st century. Both Mussolini and Goebbels acknowledged their debt to Le Bon, and Hitler’s Mein Kampf relies heavily on many of his precepts (Moscovici, 1981).

The bases of Le Bon’s influence speak eloquently to the politicized character of the area, for he did not triumph because of his conceptual originality. If anything, the opposite is true. Le Bon appropriated the ideas of his contemporaries without acknowledgment. He systematized them and expressed them with clarity and journalistic flair (Nye, 1975). For his intended audience was not the academy but rather the public and politicians—especially the latter. To that end, his main contribution was pragmatic: how could the understanding of crowd psychology be used to defuse its threat and defend the status quo? Nonetheless, the very lack of original ideas makes Le Bon perfect for our purposes, for his book is emblematic of the main intellectual current of his time, and ensures that this current still flows in the present.

For these reasons, the broad outlines of Le Bon’s position will already be familiar. He essentially argues that, in the crowd, the thin veneer of individualized and civilized life is stripped away and that the primitive inside all of us is released. And because this primitive substrate is irrational and emotional, enthusiastic and heroic, extreme and aggressive, so crowd behavior expresses all these characteristics, irrespective of the individual character of those in the crowd. In Le Bon’s own words, the crowd member “is a barbarian—that is a creature acting by instinct. He possesses the spontaneity, the violence, the ferocity, and also the enthusiasm and heroism of primitive beings” (1947, p. 32).

More technically, Le Bon proposes that the emergence of this violent beast depends upon three processes. The first is submergence, whereby individuals become an anonymous member of the mass, lose their conscious personality, and gain a sense of great power from being assimilated into a larger whole. The second is contagion, whereby crowd members duplicate any passing idea or emotion. The third is suggestion, whereby members revert to the primitive substrate, or racial unconscious (Le Bon was also one of the major “race” theorists of the late 19th century), which is the major source of the ideas and emotions that become contagious.

It is the second of these concepts, contagion, that is of particular interest in the context of the present argument. It is, in effect, a communicative principle that preempts all analysis of communication. The idea is that, having lost our conscious personality and the associated standards that normally allow us to evaluate and filter messages, we are helpless to resist whatever is put before us. Sometimes the analogy is to a hypnotic state, more recently it tends to be more oriented to the disease process. Slutkin, for instance, argued that the spread of violence in the 2011 United Kingdom riots was due to contagion; he compares contagious violence to other contagious health problems such as HIV/AIDS, influenza, and Ebola, and he asserts that “To say violence is a contagious brain process is not a metaphor, but based on scientific evidence” (2017, p. 1).

If this is true, and if mere exposure is sufficient to generate influence, then there is no need to consider the nature of the source, the nature of the target, or the nature of the message. The contagion concept shortcuts all of these and leaves us with nothing to explain about communication beyond the mere fact that a stimulus has been presented. In effect, anything goes, however extreme, however aggressive, however violent.

The problem for this account is that even the most cursory assessment of crowd events shows that anything does not go, particularly when it comes to violence. Violent suggestions and violent acts do not necessarily spread unchecked throughout the crowd. First, as we have already noted, very few crowds are conflictual even when some within them advocate conflict. Second, where there is violence, not everyone joins in. As Milgram and Toch (1969) observe, how come the riot police are not swept along by the urgings of confrontational demagogues? Third, even those who are violent do not act in a random or unconstrained manner. This point is beautifully illustrated by Thompson’s analysis of the wave of English food riots in the late 18th and early 19th century.

One might think that food riots are rather simple events. People get hungry, people gather, they break down the door of the granary, seize the contents, and flee to consume their ill-gotten goods. But, as Thompson documents in meticulous detail, the reality is somewhat different. Riots didn’t generally occur at the moment of greatest hunger but when stocks were replenishing. They occurred around specific events such as grain being transported out of town. The grain was not simply seized but sold at a popular price with the proceeds, and often even the grain sacks, being returned to the merchants. In short, these riots are not an inchoate explosion. Rather they have definite limits and pattern.

Thompson explicates this pattern. The riots he studied occurred at a period of transition from a feudal to a capitalist system. The former was based on the locality. The peasants may have had duties but they also had rights within a fiefdom. Notably, if there was food it was to be distributed and consumed locally. The latter system was based on the market, and commodities were to go to that market that would yield the highest price. These two conceptions of the social (what Thompson calls “moral economies”) clashed precisely at the moment when grain was transported from the locality, and the subsequent process of distribution mirrored the injunctions of a peasant moral economy. Thus, it was not just that riots were patterned but that this pattern was socially meaningful. The actions of crowd members reflected broader cultural systems of belief.

This observation is not true just of food riots. It emerges time and again when we go beyond the headlines and look in detail at what happens in crowd events. Reddy (1977), for instance, studies French labor riots in Rouen over more than a century and shows how the targets of these crowds “glitter in the eye of history” as a reflection of the laborers’ conception of their social world. More generally, it can be concluded that the patterning of all crowd action reflects participants’ shared understanding of society.

The key question is how? How do shared cultural systems of understanding come to shape what members of a crowd do and don’t do? What are the mediating psychological processes that lead to people joining in with actions that are consonant with such systems and eschewing those that are not? Certainly, the contagion concept, and the wider crowd psychology of which it is a part, serves to deny rather than address these questions.

The Normative Approach: Emergent Norm Theory and the Social Identity Model

Le Bon did not go unchallenged. Indeed, his influence on social psychology as a whole probably has more to do with Floyd Allport’s reaction against him than with his direct influence. In his foundational 1924 text, Social Psychology, Allport rejects the notion of a racial unconscious as an empty abstraction. Instead, he insists that the explanation of social behavior is to be found in the characteristics of the individuals involved. If groups or crowds are violent that is a reflection of the coming together of violent people—hence the term convergence theories to describe such an approach (Turner & Killian, 1972).

Despite the fierce polemics between the contagion and convergence approaches it is arguable that they share more in common than that which divides them. Both characterize crowds as pathological but merely dispute the location of such pathology—the one seeing it as a characteristic of crowd process, the other as a characteristic of crowd members. Both ignore the significance of source, target, and content when addressing why violent messages gain influence—the one by positing a model of unconstrained influence, the other by the even simpler expedient of denying that any communicative or influence process is involved. For convergence theorists, violence comes from within and others play little if any role. As a result, both approaches suggest that crowd action is indiscriminate. They cannot explain the patterns of crowd action, still less why these should be socially meaningful.

It was only from the late 1950s that this began to change. One critical factor in all this was the changing demography of the academy. Those who studied crowds were no longer gentleman scholars who gazed on events as horrified outsiders. They were as likely to be part of the civil rights protests or the Vietnam War protests. It was apparent to them that crowd events were not inchoate explosions. They had direct experience of the limits to crowd actions—as Fogelson (1971) observed of the urban riots of the 1960s, these might have been some of the most violent and extreme events in domestic American history for a century, but nonetheless they were principally characterized by restraint and selectivity.

The first sustained attempt to explain this restraint and selectivity was Turner and Killian’s Emergent Norm Theory (ENT), elaborated in their book Collective Behavior, which was first published in 1957 with subsequent editions in 1972, 1987, and 1993. As the name of the theory suggests, the argument is that crowd action is governed by norms that do not exist in advance but that are elaborated within the event itself. At the start, then, crowd events are generally heterogeneous. They consist of diverse individuals with different views doing different things. Individuals mill around trying to make sense of the situation and how to behave. They hear the voices of prominent individuals (“keynoters”) telling them what to do. Over time some of these voices predominate, they define norms that come to shape what is appropriate, what people will do and what they won’t. In this way patterns of action emerge in the crowd.

This is a much richer and more nuanced account than what went before. It paints a compelling account of the micro-interactions and the sense-making process in crowds, one that occurs over time and happens in dialogue and discussion between people. There is much to recommend it. At the same time it suffers from a number of limitations.

To start with, there is an issue of temporality: if norms emerge only after an extended process of milling and keynoting, the implication is that crowd action becomes more ordered and patterned over time as norms transform initial heterogeneity into homogeneity. Moreover, it implies a certain inflexibility whereby, if something unexpected happens, it will take some time for the crowd to settle on a normative response and respond in a unified manner. Yet crowds are characteristically novel and fast-moving events, where one has to react quickly to the actions of the other and where norms emerge quickly to guide these reactions (Wright, 1978).

But there is a larger problem. That is, ENT may describe a process for the emergence of crowd norms, but it fails to specify what precisely becomes normative. As crowd members are exposed to multiple keynoters, who do they listen to and why? What determines which voice prevails and what norms emerge? Turner and Killian are strikingly vague on this point. They point to a variety of factors. Some, such as latent support for a given position, verge on tautology (we endorse a position because we endorse that position). Others focus on the personal characteristics of the speaker (their status, the clarity and brevity of their words). But these suggestions are neither developed nor empirically warranted. As a result, there is no clear and consistent explanation of the content of emergent norms. Certainly, there is no account that can link specific crowd norms to wider systems of meaning. In sum, ENT may account for the patterned nature of crowd action, but it cannot account for the socially meaningful nature of that pattern.

This issue—the linking of local patterns of crowd action to wider systems of meaning—was the basis for developing a social identity model of crowds (Reicher, 1984, 1987). Like Emergent Norm Theory, this also suggests that crowds are norm governed and that local norms are developed within crowd events. It differs, however, in its use of a model of norm formation rooted in the social identity approach in social psychology (Tajfel & Turner, 1979; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987).

The starting point of the social identity approach lies in rejecting a notion of the self that lies at the basis of both Le Bonian and Allportian approaches, and most of psychology besides. This is the idea that selfhood is a singular construct relating to what makes an individual unique compared to other individuals. For Le Bon, that self is the basis of reasoned thought. As it is lost through submergence in the crowd, so crowd members lose their reason. For Allport, that self is maintained or even accentuated in collective settings such that unreasoned action reflects the unreasonable character of crowd members.

In contrast, social identity theorists view selfhood as a system that exists as different levels of abstraction. I can certainly think of myself in terms of what makes me, as an individual, different from other individuals (personal identity), but I can also think of myself as a member of a social category (say as an American, a Catholic, a woman) and in terms of what makes my group different from other groups (social identity). When social identities are salient we then see ourselves and others in terms of the groups we belong to, we act in terms of collective beliefs and standards, and our interests become the good of the group.

On this basis, the social identity model of crowds (SIM) proposes that psychological crowd membership is not based on a loss of identity but on a shift from individual to social identity. Correspondingly, crowd membership does not entail a loss of standards, a loss of reason, a loss of morality but rather a shift from individual to collective standards. However, insofar as crowd events are typically novel or ambiguous it is necessary to elaborate a situationally appropriate set of norms that are consonant with the general tenet of group identity. That is, the question for group members is not simply “what do we do here and now?” but rather “what do we do, as (say) Americans (or Catholics or women), here and now?”

There may be some space for debate as to what, precisely goes. But equally, certain things most definitely do not go. For instance, as a crowd of pacifists confronted by hostile opponents, one might stage a sit-down protest, one might chant or sing songs of resistance, but one would not throw stones. Or, equally, as a member of a sectarian Protestant crowd one might assault Catholic targets in a variety of ways from insult to annihilation, but one is unlikely to attack non-Catholic targets (cf. Davis, 1973). The SIM thereby explains how crowd action is simultaneously volatile but occurs within clear limits and also how the patterns and limits of crowd action reflect broad systems of understandings. In brief, crowd members act as category members and hence their behavior is constrained by the contours of categorical beliefs.

As for exactly how crowds elaborate a situationally appropriate set of norms, SIM draws on a distinction between the deductive and inductive aspects of categorization (cf. Turner, 1982). Deduction is a deliberative process whereby group members infer appropriate actions by discussing how different options relate to the defining beliefs of the group. Induction is a matter of inferring group norms from the behavior of typical group members and, it is argued, is more appropriate to crowd conditions precisely because of their rapid changeability and the impracticality of sitting down to debate “what we believe in.”

Thus, for instance, when the police raided St. Pauls (a black neighborhood in the western English city of Bristol) and locals stood around for a while, uncertain of how to respond, the act of one elderly resident who threw a stone at a parked police car resulted in a hail of stones toward the police and the start of a prolonged confrontation (Reicher, 1984). This inductive process speaks to the key dimensions of communicative processes. In terms of source, it suggests that influence agents need to be seen as in-group members; in terms of targets it suggests that only those who identify as members of the same category will be influenced; in terms of content, it suggests that only those messages seen as consonant with in-group identity will be influential. Thus when stones were thrown at shops owned by locals or else at a bus, others did not join in and, if anything, they purposefully suppressed such acts (Reicher, 1984).

While this focus on the inductive aspect of categorization does help make sense of the fast-moving and volatile events within a riot, and of how these remain within intelligible limits, it is arguable that this presents a somewhat one-sided picture of crowd phenomena. For while crowds may have moments of rapid development—and while it is necessary to be able to explain norm formation within these moments—they also tend to have long periods of relative stasis. For instance, prior to the initiation of conflict in St. Pauls, there were several hours during which the police raided a cafe and removed unlicensed alcohol while residents hung around opposite and discussed what was going on. In this period, there was space for deliberation and for deductive processes to occur.

What is more, Turner and Killian’s concepts of milling and keynoting are useful in describing how deduction took place. That is, there may be value in integrating the attention to the meaning-making process that Turner and Killian bring from symbolic interactionism with the account of norm formation brought from the social identity approach. This can facilitate understanding the importance of both micro-interactions in the crowd and their relation to macro-social categorical processes. Or, to be more concrete, the success of keynoters can be understood in terms of their ability to present themselves as of the in-group, as speaking to in-group members, and as interpreting in-group identity in context (cf. Haslam, Reicher & Platow, 2011).

There may be much mileage in such an integration, but it cannot be the whole story. This is because an account of crowd action as socially patterned is itself only part of the story. Crowds do not only reflect social categories and social understandings. Critically, they also serve to change the organization and the self-understanding of individuals and of societies—that, after all, is why they generate such fear and such interest. What is more, such change is not only a consequence of crowd events. There are many occasions where one can see dramatic changes in the course of a single event.

In his account of the Polish uprising of 1956, Machcevicz (2009) provides a compelling example of such changes. The movement for change was initiated by a mass demonstration in Poznan on June 28, which developed into a series of street battles, eventually suppressed by the military. At the start, participants primarily defined themselves as workers and responded to economic demands: “We want bread!”; “Down with the exploitation of labor.” In time this became intermingled with political, national, and religious categories and demands: “Down with Bolshevism”; “Down with the Russians”; “God in schools” (that is, for religious instructions to be re-introduced into the education system). As demonstrators clashed with the authorities—and, more particularly, in the context of prolonged conflicts with the State Security forces (secret police)—national and religious elements began to predominate (the two being intertwined insofar as Poland was distinguished from Russia in terms of its Catholicism). Rumors flowed that those shooting on the crowd from Security Office windows were Soviets dressed in Polish uniforms. Crowd members called on others as fellow Poles to join in the fighting. As Machcevicz puts it: “in the course of the battle at the Security Office, ‘the people’ that had formed during the march and rally in Stalin Square was transformed into ‘the nation’” (2009, p. 114).

It is important to note that this transformation occurred even though the composition of the crowd stayed largely the same and also that there was some intermingling of all the slogans from the start. In other words, both crowd and keynoters stayed constant. So what else could have changed? The answer appears within the account of Poznan. It is something so obvious that it is easy to forget, and indeed one of the most remarkable things about the various crowd models described so far is that they all overlook this aspect of crowds. That is, they direct their attention exclusively to what is going on in the crowd, as if crowds act in isolation. But they don’t. Crowds exist in interaction with other groups, with the police or militia or army. Crowd events are inter-group encounters, and the development both between and within the various groups can never be understood by reference to just one of them. So, to understand the patterned, socially meaningful, and dynamic character of crowd action, an inter-group dimension must be added to crowd psychology.

The Dynamic Approach: The Elaborated Social Identity Model

The Elaborated Social Identity Model of crowds (ESIM) was developed specifically to address changes in the level of empowerment and of conflict in crowds. It was a response to an empirical pattern of events repeatedly observed in very different types of events, from student demonstrations (Reicher, 1996) to environmental protests (Drury & Reicher, 1999, 2000; Drury, Reicher, & Stott, 2003) to football events (Stott, Hutchison & Drury, 2001; Stott & Reicher, 1998):


In each case the crowd is initially heterogeneous, involving different groupings and different influence agents advocating different types of action. Some (characteristically, the great majority) are conciliatory and have no intention to act conflictually; others (characteristically a minority) are confrontational and come prepared for conflict with out-groups (typically the police).


The out-group (police), seeing people advocating violence within the crowd, both perceives and treats the crowd as a whole as dangerous (guided, in part at least, by traditional contagion models of crowd process). This can involve different tactics: containment, dispersal, use of tear gas, etc. But all these tactics are undifferentiated, they affect everyone equally, and they involve denying crowd members their everyday rights.


In the face of such common treatment, crowd members become more homogenous; they cohere in opposition to the police and around those advocating confrontation with the police. What is more, such cohesion gives rise to an enhanced sense of empowerment and hence a willingness to act in opposition to the police.


The reaction of the crowd confirms the initial police perception and intensifies their undifferentiated and repressive response, which in turn consolidates collective opposition among crowd members and thus creates a spiral process of conflict escalation.

Once again, the key theoretical move in explaining these patterns lies in clarifying the concept of identity. For ESIM, a social identity should be understood as a representation of one’s position in a set of social relations (Drury & Reicher, 2009). Thus, in different contexts, as social relations shift, so will social identity. Along with this, there is a shift in who one sees as in-group and out-group and hence who one will or will not be influenced by. Moreover, if one acts on the basis of one understanding of one’s social position but, as a result, that social position is changed so there will be changes in identity and action within an event. This is precisely what happens in crowd events under conditions where there is an asymmetry between the self-understandings of crowd members and the understandings of an out-group such as the police (e.g., I see myself as a law-abiding citizen expressing my democratic right to dissent; they see me as a dangerous extremist), where the out-group has the power to impose its understanding in action upon crowd members (e.g., the police corral all crowd members and stop them marching to their destination), and, finding themselves positioned in a new way, people redefine who they are (e.g., being positioned as “the opposition” I begin to see myself as an oppositionalist along with many others I previously rejected as radicals).

However there is another key element to stress that has become more explicit in recent elaborations of ESIM (e.g., Reicher, 2017). That is, lived experience does not automatically produce self-understanding (and changes in the structure of lived experience do not automatically produce changes in self-understanding). Rather, experience needs to be theorized, given meaning, and translated into models of self and other. The individual does not undertake this translation process in isolation.

It happens through a process of communication, in dialogue with others, and in interaction with those seeking to influence crowd action. That is, categorization derives from an interaction between the nature of lived experience and the availability in context of sense-making resources to render that experience intelligible.

From this perspective, ESIM can be reframed as a process of communication. As Emergent Norm Theory realizes (along with other group theories, e.g., Mugny, 1982; Subasic, Reynolds, & Turner, 2008), there are always multiple sources of influence seeking to gain sway over the broader population of the crowd. The balance of influence between these different sources within the group is dependent upon the intergroup context. Thus, in a situation of two groups (1 and 2), the way group 2 treats group 1 will determine the balance of influence within—and hence the response of—group 1. This response will in turn shape the balance of influence in group 2 and its further response (see Figure 1, taken from Reicher, 2017).

Figure 1. A dynamic model of group process.

For instance, under conditions where the authorities treat the crowd as a whole as dangerous and deny crowd members their rights, then those sources who represent the world in terms of an opposition with authority and who advocate confrontation with authority are likely to become more influential even if, prior to such treatment, such radical voices had been shunned. In exactly this way, when the security forces in Poznan, seen as controlled by the Russians, responded to demonstrations by indiscriminate gunfire, slogans that articulated a division between the Polish people and the Soviets were more widely chanted and replaced other slogans that were structured around domestic economic demands. This anti-Sovietism in turn strengthened the hand of those more repressive elements in the security apparatus.

To use a somewhat different example, when the Poles demonstrated again in October 1956, and in as large numbers as in Hungary, the reason why Poland (unlike Hungary) did not see an insurrection is that the Polish regime replaced the old Communist Party leadership with Gomulka—believed to be an authentic national voice. As a result the insurrectionary demands to overturn the authorities as a national out-group ceded to reformist demands upon the authorities as an in-group to meet the needs of its workers. By contrast, in Hungary, the initial response was repression and the turn to conciliation came too late and was interpreted as weakness (Machcevicz, 2009).

The more general point is that, while ESIM seeks to explain the conditions under which conflict escalates it certainly does not propose that escalation is inevitable. The conditions under which that occurs (asymmetries of understanding, out-group power, generalized repression) are relatively rare. What is more, under conditions where crowd members perceive themselves as in an antagonistic relationship to authority but authorities do not treat them as other, ESIM is equally capable of explaining how conciliatory voices in the crowd gain influence and conflict de-escalates.

Stott et al. (2001) provide perhaps the best worked example of this. They follow both English and Scottish fans at the football World Cup in France in 1998. Domestically, these fans are equally violent, but in the World Cup context things work out very differently. The English fans are seen as dangerous and treated as dangerous by locals, rival fans, and the police alike. This leads to a process whereby even the moderate fans cohere around more violent elements and join in fights they construe as self-defense. The Scottish fans are seen as boisterous but good natured. They are welcomed and well-treated by all, including the police. In this context not only do violent elements lose influence, but if any individuals act confrontationally, other fans actively surround and stop them. That is, in a context of out-group support, crowd members actively self-police in order to maintain their positive collective reputation. Overall, the contrast in the trajectory of these two groups of fans is eloquent testimony to the role of intergroup dynamics in framing intra-group influence processes.

Crowd Event as Acts of Communication

Thus far, the focus has been on the way that communication processes govern how people within the crowd are influenced and what they do. The crowd event itself, however, is a communicative act aimed not only at crowd members themselves but also at those beyond the crowd. Analysis thus turns from crowds in themselves to the wider social impact of crowds.

Communicating to Those in the Crowd

If, as has been argued, social identities are models of self in social relations, then how does one mobilize people in support of a vision of the social world that does not yet exist in practice? How do would-be nationalist leaders persuade people to define themselves in terms of nationhood and work toward a new nation-state? Or else, how does one embed a new understanding of who and what a group is, of how we relate to each other and how we stand in relation to other groups? One answer is through the use of collective events and by realizing within a social movement the vision that one seeks to bring about in the wider world.

Monica Ozouf (1988) provides a highly influential example of this. Her interest lies in the ways that revolutionaries sought to transform French society after 1789, how they sought to replace the old hierarchical society with a new horizontal society encapsulated in the slogan “liberty, fraternity, equality.” She analyses the pivotal role of festivals in this enterprise. Traditional events in which people were separated into the different “estates” that ordered France were replaced by a new festival calendar and by events that, if they ordered people at all, did so by permeable categories like age, which we all pass through.

This work helped generate a whole literature on the role of commemorations, celebrations, parades, and anniversaries in providing societies with a model of themselves—what they could be and what they should be (e.g., Gillis, 1996; Spillman, 1997). Or, as Gelvin (1998) concludes from his own study of how different visions of Syrian society in the aftermath of World War 1 were embodied in differently organized demonstrations, what these crowd events seek to accomplish is the realization that “the world as lived and the world as imagined, fused under the agency of a single set of symbolic forms, turn out to be the same world” (1998, p. 226).

One of the most notorious instances of this process comes from the Nuremburg rallies as mythologized in Leni Riefenstahl’s film Triumph of the Will (see Spotts, 2002). In the opening sequence, Hitler’s plane descends through the clouds, casting the shape of a cross on the attendant masses. He emerges through then to stand in front of and above them. Everything about the arena is rigidly geometrical. People are formed into homogenous blocks. The leader is placed alone, in dead center, in front of and above the mass, alone in their sight line. The core Nazi principles of order, hierarchy, and above all of messianic leadership (the fuhrerprinzip) shape the performance. Even the building materials of the arena serve to underpin Hitler’s vision of German identity: they are granite, German oak, tokens of hardness and (supposedly) of indestructibility.

Or again, take a more contemporary example: Donald Trump’s 2016 election campaign was often dismissed as a chaotic and laughable affair (Taibbi, 2017); however, this underestimates an underlying clarity and coherence. On the one hand, Trump articulated a vision of America as a nation laid low by the combination of attacks by an enemy without (Mexicans and Muslims) and betrayal by an enemy within (the media, Wall Street, and the Washington establishment). Trump presented himself as the American savior who would clean out the Washington swamp and shut out the migrants (Reicher & Haslam, 2017). On the other hand, Trump rallies can be seen as theatrical performances that dramatized this vision.

Gwynne Guilford (2016) describes how, an hour or so before the candidate came on stage, security guards would fan out, backs to the stage, face to the audience, looking for dissenters or troublemakers. Their mere presence attested to the possible presence of the external enemy. If this enemy was identified (either by protesting or even by not being enthusiastic enough) crowd members were instructed not to intervene but to point at the source of danger and to chant out loud “Trump! Trump! Trump!”—simultaneously broadcasting the threat to the entire crowd and invoking the man who would deal with that threat.

When Trump did come on stage he would involve and encourage the crowd in the often brutal expulsion of protesters. He would also encourage them in goading and insulting the media, penned in behind him. And the media could do nothing but sit and take it. Taibbi describes how: “we in the press, obediently clustered inside our protective rope line . . . would sit looking guilty, like the pampered, narrow-shouldered, overgroomed hypocrites we are, while Trump blasted us as the embodiment of the class that had left regular America behind” (2017, p. xxiii).

Altogether, then, the rallies were morality plays in which good America (embodied and orchestrated by Trump) turns the tables and, for once, vanquishes its various evil enemies. Taibbi again: “we [the media] were part of his [Trump’s] act. And his triumph over us was a major factor in convincing ordinary people that he could deliver on his rebellious rhetoric” (2017, p. xxi). It is important to unpick the various elements in this “triumph.” First, Trump’s version of and vision for America is given credibility. Second, crowd members themselves are empowered and given agency through Trump. Third, and relatedly, they can use their agency to enact their vision of the world and how it should be organized: for once they are not dictated to by their foes but can dictate to them.

This enactment of the group vision is termed collective self-realization, or CSR (Reicher & Haslam, 2006), and there is evidence that the accomplishment of CSR in a crowd event leads to an enhanced commitment to the group identity and group practices beyond the crowd (Hopkins, Reicher, Tewari, Narayanan & Stevenson, 2016). For those who participate in them, crowds both communicate and validate the viability of a certain version of identity. In a phrase, they embed identity by making it real.

Communicating beyond the Crowd

Even if the outcomes described were true of every single person who was part of a crowd, crowds would still be of limited social significance—certainly insufficient to win an election or to change a society. The question therefore arises of what crowds communicate to those who are not members and whether this has any effect.

This raises the more general question of how people come to understand their social position, and how the groups to which they belong stand in society. Personal experience is obviously important, but it is hard to tell if experiences of exclusion or hostility derive from idiosyncratic or collective factors. Is one denied a job or housing because one is black or working class or foreign or simply because one performed badly in the interview? These difficulties are exacerbated by the fact that most of the significant categories that define us are (as Anderson [1983] says of nationhood) “imagined communities.” They are too large ever to assemble together; we can never see what happens to category members as a whole; therefore, we can only ever imagine that we are all together, experiencing the same events in the same ways. Perhaps, but it is arguable that a crowd can be regarded as the imagined community made manifest (Reicher, 2017). How a crowd of black people is treated in Ferguson, Missouri, for instance, illustrates the position of black people in American society or even black people across the world (see Ndlovu, 2014).

If this points to potential scope of crowds as acts of communication and vehicles of influence, it also points to a critical moderating factor. That is, whether people see a crowd event as self-relevant and what they read from it will depend upon how the participants are categorized. Whether Ferguson speaks to black people across the world, for instance, depends upon whether the Ferguson demonstrators are themselves defined as black people as opposed to specifically American blacks or even as troublemakers. This makes the struggle over the categorical representation of crowd events an issue of broad social as well as psychological significance.

This struggle over representation is often encapsulated in the question of whether the crowd should be regarded as the people or the mob. If the former, then the crowd deserves the support of the entire population, its demands should be heeded by government and indeed any government is legitimated by association with the crowd. If the latter, then the crowd is a sign of degeneration, it deserves condemnation and repression and anyone associated with it is rendered illegitimate. Or, to cite Jonsson: “from the French revolution onward the treasured will of the people has gone hand in hand with the wicked rule of the mob” (2008, p. 7).

Jonsson’s argument is that, over time there was a shift. Shortly after it happened, the participation of crowds in the overthrow of the old order was represented as the people entering the political process and established the new regime as an authentic national voice. It is encapsulated by Jacques-Louis David’s famous drawing of “The Tennis Court Oath”: as the deputies of the Third Estate make their vow to break with the old regime (a critical moment in the early development of the revolution), an audience of old and young, men and women, soldiers and civilians enters the scene, crowding in at the windows. A century later, however, crowds are represented in very different ways, as animalistic and dehumanized, frightening, a threat to rather than the foundation of the political process.

There is no need to look to history or to different events to see these different representations and their importance. Thus, the crowds that assembled in Kiev’s Maidan Nezalezhnosti (Independence Square) in November 2013 led to the fall of the old pro-Russian regime of Victor Yanukovych and ushered in a new pro-Western regime. The legitimacy of the so-called Euromaidan revolution turned on the nature of those assembled in the square. Were they, as the demonstrators styled themselves in word and symbol, the Ukrainian people (with the right to reject and install leaders)? For instance, on November 21, as the protests started, the Ukrainian oppositionist Yriy Lutsenko, asserted: “people did not come to the politicians today, but the politicians came to the people” (Shevda & Park, 2016, p. 86). What is more Shevda and Park report that the crowd displayed only national emblems and that sectional party political banners or flags were entirely absent. Or were the demonstrators, as supporters of the old regime and opponents of the new styled them, a faction of extreme right wing nationalists (whose actions were therefore a coup against the legitimate government)? They were styled as Banderites after Stephan Bandera, a Ukrainian nationalist leader who initially fought with the Nazis during World War II before being arrested by them. As Vladimir Putin put it in denouncing the post-Maidan government in Kiev and justifying the annexation of Crimea, Crimea: “will never be Banderite” (Esch, 2015, p. 7).

However, perhaps the most powerful example of this labor of representation comes from the Egyptian uprising of 2011 and the crowds of Tahrir Square. Once again the fate of this uprising (at least in the short term) rested on the representation of the participants. The Mubarak regime systematically attempted to represent them as either alien (foreigners and their dupes), as pathological (hoodlums), or as a privileged and unrepresentative fraction of the population (urbanites). By contrast, the demonstrators waved Egyptian flags, they protested their national allegiance and their national status. According to the April 6th Youth Movement (one of the organizations at the core of events): “nothing brings us together except our love for this country and our desire to reform it”; or again, the protesters “did not belong to any particular group” but rather “this is the revolution of all the people” (Alexander, 2011, p. 8).

One corollary was that “abuse of any individual is against the entire nation” (Alexander, 2011, p. 11). This claim, again by the April 6th Youth Movement, was of more than symbolic value. It shaped the attitude of the military, pledged to defend and protect the people. Hence the question of whether the military intervened to protect or to disperse the protesters was determined by whether the protesters were the people or enemies of the people. In the event, the victory of crowd members in the representational struggle led to instances where the army intervened between the crowd and armed attackers—if not actively repressing those attackers, at least preventing them from harming the crowd.

Ketchley (2014) describes one such incident: “Brandishing his pistol, Boules [an army captain] repeatedly fires in the air. As the column continues to advance, more soldiers appear, heading to collect their weapons from the hatch of the APC [Armored Personnel Carrier]. As one soldier kneels to load his rifle, protestors begin to chant, “The people and the army are one hand.” The chant quickly spreads. The footage ends with protestors jumping up and down, the column of advancing thugs halted, protestors and soldiers buzzing with the emotional energy of averting attack. A weeping Boules is then embraced by protestors after mounting the APC to be cheered by the crowd and proclaimed with further chants of, “The people and the army are one hand” (p. 174).

Clearly, it matters how the crowd is represented. This can (literally) be a question of life or death for those in the crowd. More widely, not only do such representations communicate the state of society, they also determine the fate of society.


The argument has now come full circle. It started with a critique of classic models of communication in the crowd—the assumption that the people stop being the people and revert to the condition of a primitive horde with the result that anything goes, however violent—the argument finished by showing how such views effectively undermine the political attempts of crowd members to communicate their discontent with the existing order. There is a clear coherence here: if the reactionary potential of traditional crowd psychology lies in representing people as “the mob” so crowds can only recover their political potential (within a liberal-democratic society at least) by transforming themselves back into “the people.”

As it developed, the argument addressed both communication processes in the crowd and the crowd as a process of communication have been addressed. In both cases, a challenge was raised to those who dismiss crowd phenomena as inchoate and meaningless—a compelling and exotic but ultimately peripheral phenomenon of little interest to those concerned with the study of everyday society.

Rather, it was suggested that the study of crowds is of interest to the student of society (and of communication in society) in at least three ways. First, communication processes in the crowd, and hence crowd action, are framed by both the macro-social context (broad ideological and structural aspects of society) and the micro-social context (immediate intra-group and intergroup relations). Second, crowds provide a particularly productive site in which to investigate and develop theory about social communicative processes—this precisely because of the instability and dynamism of crowd contexts, which render visible the importance of phenomena often taken for granted. More concretely, the dynamic model sketched in Figure 1 has far wider relevance than simply as a way of explaining crowd phenomena. Third, crowds don’t just produce theories of society, but produce society itself.

Above all, the fascination of crowds lies in the way that they transcend the many dualisms that plague social scientific research: the individual vs. the social (it has been argued that individual agency comes about through becoming a collective subject); reason vs. emotion (it has been argued that people are passionate in crowds because of the opportunity to enact their worldview); and, above all, social determination vs. social change.

There is an intellectual vandalism to those who dismiss crowds as mindless and crowd events as a conflagration or explosion. By so doing, and by ignoring the patterns of crowd action, they trash the best means we have of accessing the social understandings of those who characteristically have no voice in society (it was Martin Luther King who memorably asserted that riots are the voice of the unheard). We need to learn to read crowd events with as much care as we would read a fragile text from a lost civilization (cf. Smith, 1980). In expressing their worldview, crowd members are simultaneously seeking to communicate who they are, to make claims on the solidarity of others, and to create the social forces that can reshape their world.

Further Reading

  • Reicher, S. D. (2001). Crowds and social movements. In M. Hogg & S. Tindale (Eds.), Blackwell handbook of social psychology: Group processes (pp. 182–208). Oxford: Blackwell.


  • Alexander, J. C. (2011). Performative revolution in Egypt. London: Bloomsbury.
  • Anderson, B. (1983). Imagined communities. London: Verso.
  • Barrows, S. (1981). Distorting mirrors. New Haven, CT: Yale University Press.
  • Davis, N. Z. (1973). The rites of violence: Religious riot in sixteenth century France. Past & Present, 59, 51–91.
  • Drury, J., & Reicher, S. D. (1999). The intergroup dynamics of collective empowerment: Substantiating the social identity model of crowd behavior. Group Processes and Intergroup Relations, 2, 381–402.
  • Drury, J., & Reicher, S. D. (2000). Collective action and social change: The emergence of new social identities. British Journal of Social Psychology, 39, 579–604.
  • Drury, J., & Reicher, S. D. (2009). Collective psychological empowerment as a model of social change: Researching crowds and power. Journal of Social Issues, 65, 707–726.
  • Drury, J., Reicher, S. D., & Stott, C. (2003). Transforming the boundaries of collective identity: From the “local” anti-road campaign to “global” resistance. Social Movement Studies, 2, 191–212.
  • Esch, C. (2015). “Banderites” vs. “New Russia”: The battle field of history in the Ukraine conflict. Unpublished paper. Reuters Institute for the Study of Journalism, University of Oxford.
  • Fogelson, R. M. (1971). Violence as protest: A study of riots and ghettos. New York: Doubleday.
  • Gelvin, J. L. (1998). Divided loyalties: Nationalism and mass politics in Syria at the close of empire. Berkeley: University of California Press.
  • Gillis, J. R. (1996). Commemorations: The politics of national identity. Princeton, NJ: Princeton University Press.
  • Giner, S. (1976). Mass society. London: Martin Robertson.
  • Haslam, S. A., Reicher, S. D., & Platow, M. (2011). The new psychology of leadership. London: Psychology Press.
  • Hopkins, N. P., Reicher, S. D., Tewari, S., Srinivasan, N., & Stevenson, C. (2016). How collective participation impacts social identity: A longitudinal study from India. Political Psychology, 37, 309–325.
  • Jonsson, S. (2008). A brief history of the masses: Three revolutions. New York: Columbia University Press.
  • Ketchley, N. (2014). “The army and the people are one hand!” Fraternization and the 25th January Egyptian revolution. Contemporary Studies in Society and History, 56, 155–186.
  • Le Bon, G. (1947). The crowd: A study of the popular mind. London: Ernest Benn.
  • Machcevicz, P. (2009). Rebellious Satellite: Poland 1956. Stanford, CA: Stanford University Press.
  • Merriman, J. M. (2016). Massacre: The life and death of the Paris Commune. New Haven, CT: Yale University Press.
  • Milgram, S., & Toch, H. (1969). Collective behavior: Crowds and social movements In G. Lindzey & E. Aronson (Eds.), The handbook of social psychology (Vol. 4, pp. 507–611). Reading, MA: Addison Wesley.
  • Moscovici, S. (1981). The age of the crowd. Cambridge, U.K.: Cambridge University Press.
  • Mugny, G. (1982). The power of minorities. London: Academic Press.
  • Nye, R. A. (1975). The origins of crowd psychology: Gustave Le Bon and the crisis of mass democracy in the Third Republic (Vol. 2). London: SAGE.
  • Ozouf, M. (1988). Festivals and the French Revolution. Cambridge, MA: Harvard University Press.
  • Reddy, W. M. (1977). The textile trade and the language of the crowd at Rouen 1752–1871. Past & Present, 74, 62–89.
  • Reicher, S. D. (1984). The St. Pauls’ riot: An explanation of the limits of crowd action in terms of a social identity model. European Journal of Social Psychology, 14, 1–21.
  • Reicher, S. D. (1987). Crowd behaviour as social action. In J. C. Turner, M. A. Hogg, P. J. Oakes, S. D. Reicher, & M. Wetherell (Eds.), Rediscovering the Social Group (pp. 171–202). Oxford: Blackwell.
  • Reicher, S. D. (1996). The Battle of Westminster: Developing the social identity model of crowd behaviour in order to deal with the initiation and development of collective conflict. European Journal of Social Psychology, 26, 115–134.
  • Reicher, S. D. (2017). La beaute est dans la rue: Four reasons (or perhaps five) to study crowds. Group Processes and Intergroup Relations.
  • Reicher, S. D., & Haslam, S. A. (2006). Tyranny revisited: Groups, psychological well-being and the health of societies. Psychologist, 19, 146–150.
  • Reicher S. D., & Haslam, S. A. (2017). The politics of hope: Donald Trump as an entrepreneur of identity. In M. Fizduff (Ed.), Why irrational politics appeals: Understanding the appeal of Donald Trump (pp. 25–40). Santa Barbara, CA: Praeger.
  • Reicher, S. D., & Stott, C. (2011). Mad mobs and Englishmen? Myths and realities of the 2011 riots. London: Constable & Robinson.
  • Shevda, Y., & Park, J. H. (2016). Ukraine’s revolution of dignity: The dynamics of Euromaidan. Journal of Eurasian Studies, 7, 85–91.
  • Slutkin, G. (2017). Reducing violence as the next great public health achievement. Nature Human Behaviour, 1, 0025.
  • Smith, D. (1980). Tonypandy 1910: Definitions of community. Past & Present, 87, 158–184.
  • Spillman, L. P. (1997). Nation and commemoration. Cambridge, U.K.: Cambridge University Press.
  • Spotts, F. (2002). Hitler and the power of aesthetics. London: Hutchinson.
  • Steadman Jones, G. (1983). Languages of Class. Cambridge, U.K.: Cambridge University Press.
  • Stott, C., Hutchison, P., & Drury, J. (2001). “Hooligans” abroad? Inter-group dynamics, social identity and participation in collective “disorder” at the 1998 World Cup Finals. British Journal of Social Psychology, 40, 359–384.
  • Stott, C., & Reicher, S. D. (1998). How conflict escalates: The inter-group dynamics of collective football crowd “violence.” Sociology, 32, 353–377.
  • Subašić, E., Reynolds, K. J., & Turner, J. C. (2008). The political solidarity model of social change: Dynamics of self-categorization in intergroup power relations. Personality and Social Psychology Review, 12, 330–352.
  • Taibbi, M. (2017). Insane clown president. London: WH Allen.
  • Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–47). Monterey, CA: Brooks Cole.
  • Turner, J. C. (1982). Towards a cognitive redefinition of the social group. In H. Tajfel (Ed.), Social Identity and Intergroup Relations (pp. 15–40). Cambridge, U.K.: Cambridge University Press.
  • Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. (1987). Rediscovering the Social Group. Oxford: Blackwell.
  • Turner, R. H., & Killian, L. M. (1972). Collective behavior. Englewood Cliffs, NJ: Prentice Hall.
  • U.S. Riot Commission. (1968). Report of the National Advisory Commission on Civil Disorders. New York: Bantam.
  • Van Ginneken, J. (1992). Crowds, psychology, and politics, 1871–1899. Cambridge, U.K.: Cambridge University Press.
  • Wright, S. (1978). Crowds and Riots. London: SAGE.