The New Psychology of Leadership: Identity, Inﬂuence, and Power

3
Foundations for the new psychology of leadership

Social identity and self-categorization

In the previous two chapters, we reviewed prevailing approaches to leadership both in and beyond the academic world. We argued that this work has taken great strides forward in recent years. In particular, there is increasing recognition that:

1 Whether or not leadership is successful depends on context.

2 Leadership is not a quality of leaders alone but rather of the relationship between leaders and followers.

3 Leadership is not just about existing social realities but also about the transformation of social reality.

However, we also identified unresolved issues and residual areas of neglect. Critically, we suggested that leadership is not just a relationship between leaders and followers. It is a relationship between leaders and followers within a social group. As a result, to be effective, leaders and followers need to be bound together by both being part of a common “we.” Moreover, leaders gain their status and their influence over others by being able to represent what this “we-ness” consists of, and they are also constrained in what they can do by the meaning of this “we-ness.” For this reason leadership can never be properly understood simply through an appreciation and analysis of individual qualities. Rather, it is irrevocably bound up with group processes. If we want to understand the nature of leadership—what makes it possible, what makes it effective, and what are its limits—we need to understand something about group processes in general.

The purpose of this chapter is to provide such an understanding. Our perspective is derived from what has, over the last quarter century or so, become the dominant approach to the study of groups in social psychology. This approach derives from the pioneering work of two European researchers: Henri Tajfel and John Turner. At its heart is one core concept: social identity (Tajfel & Turner, 1979; Turner et al., 1987; for a recent overview see Reicher, Spears, & Haslam, 2010; for extensive background and core readings see Postmes & Branscombe, 2010).

Social identity refers to individuals’ sense of internalized group membership. It is a sense of self associated with an awareness that one belongs to a particular social group and that this group membership is important and meaningful (Tajfel, 1972). So, for example, it is social identity that underpins people’s sense that they are part of a particular nation, a particular organization, a particular club, and so on. And it is social identity that allows people to refer to themselves and other members of such groups as “us” (e.g., “us Australians,” “us Ford employees,” “us Lakers fans,” “us Latina women”). However, what is most important for present purposes is the fact that it is social identity that allows people both to lead and to be led. As we will see, this is because social identity—a shared sense of “us”—is central to the social influence that lies at the heart of effective leadership.

The primary goal of this chapter, then, is to lay down the foundations for a new psychology of leadership that is grounded in an understanding of group psychology. In this it is something of a departure from the previous chapters. For the work that we will discuss is not directly or exclusively concerned with issues of leadership. Instead, we develop a series of ideas that are necessary in order to understand effective leadership: we consider the roots of group behavior in shared self-categorization; we examine the relationship between self-categorization, social influence, and social power; we address the ways that psychologically meaningful social categories (including all group memberships) are defined and how they relate to social reality. Along the way, we will point out the relevance of these various aspects of group process for understanding various facets of leadership. But the task of demonstrating these relationships will be left to later chapters.

Another way of saying this is that this chapter is all about agenda-setting. By the time we are done here we will have defined and justified the component elements of a new social psychology of leadership. The remaining chapters will then be devoted to putting flesh onto these bones.

Social identity and group behavior

Towards a group level of analysis

We can set about the process of demonstrating why social identity is so important to leadership by first asking what it is that turns any collection of individuals into a social group. Why do people join groups? And what keeps them there? These are key questions in social and organizational psychology and ones that most researchers have answered in individualistic terms. In particular, along the lines of the exchange approaches that we discussed in Chapter 2 (e.g., equity theory; Adams, 1965), researchers have argued that individuals become group members when they perceive that it is in their personal interests to do so. This suggests that people join groups when they find other group members attractive and, in particular, when they consider the benefits of joining to outweigh the potential costs. Illustrative of this point of view, in the sixth edition of their influential text Groups: Theory and Experience, Rodney Napier and Matti Gershenfeld discuss the question of “Why people join groups” and conclude:

There seem to be three major reasons why people join groups:

1. They like the task or activity of the group…

2. They like the people in the group…

3. Although the group does not satisfy the person’s needs directly, it is a means of satisfying his or her needs.

(1999, pp. 53–75, original emphasis)

The central idea here, then, is that groups are comprised of individuals who become interdependent for essentially instrumental reasons: to satisfy their personal interests and their mutual needs (e.g., Rabbie, 1991).

Psychologically speaking, an important feature of this analysis is that it renders the group itself analytically superfluous. Groups are understood as the constellation and aggregation of personal motivations that bind individuals to them, and thus are seen as nothing more than the sum of their individual parts. To the extent that groups no longer meet their members’ personal needs they should simply disband and disintegrate. Indeed, not only is there nothing psychologically “special” about groups, but also, strictly speaking, there is no such thing as “group process” at all (Turner & Haslam, 2001, p. 31).

Atfirst blush, this model might seem quite plausible, and it certainly speaks to the way that people often think and talk about groups. Nevertheless, it has a number of major empirical and theoretical problems. The first of these is that people’s decisions to join or leave groups are not well predicted by the degree to which those groups satisfy their personal needs. If this were the case, why would people continue to support losing football teams and why would they make a point of being a “die-hard” fan who sticks with their team “through thick and thin”? The validity of this point is confirmed in work by Daniel Wann and Nyla Branscombe that looked at support for baseball and basketball teams in the United States. These researchers found that for fans who lived close to their team’s home base, support bore no relationship to the team’s success or failure (Wann & Branscombe, 1990). For these fans, there was a sense that withdrawing support from their team simply wasn’t an option. Why? Because it was their team.

Probing the logic of established theory still more, we might go on to ask questions about the really big decisions that people sometimes make in life. Ponder, for example, why people join the army and go to war. Why would they do this if they were making rational decisions about what was in their personal best interests? Bearing in mind the probability of death or serious injury (e.g., see Nicholson, 2001), if the question that recruits asked themselves before (or after) signing up for a dangerous campaign was “What’s in this for me?”, the answer would surely have to be “not very much.”

Of course, one might try to explain such decisions by seeing them as a product of peer-group pressure to join up, or as a consequence of individuals’ desire to avoid the stigma associated with failing to volunteer. Yet the personal accounts of soldiers themselves suggest that these are typically not the key factors (e.g., see Lewis, 2003). Instead, along the lines of Burns’s reflections on the impact of transformational leadership (see Chapter 2), it appears that people sign up because they want to and because at a group level they believe it is the right thing to do. As powerful testimony to this point, consider the lines of the World War I poet Edward Thomas, in which he reflected on his own reasons for enlisting:

This is no case of petty right or wrong

That politicians or philosophers

Can judge. I hate not Germans, nor grow hot

With love of Englishmen, to please newspapers.

Beside my hate for one fat patriot

My hatred of the Kaiser is love true…

But with the best and meanest Englishmen

I am one in crying, God save England, lest

We lose what never slaves and cattle blessed.

The ages made her that made us from dust:

She is all we know and live by, and we trust

She is good and must endure, loving her so:

And as we love ourselves we hate her foe.

(Thomas, 1916/1964, p. 57)

Thomas’s poem—written shortly before he himself was killed in action— makes it very clear that his decision to join the army and fight was not an “I thing” but a “we thing.” Support for this same point is provided by witness accounts of Romanian youths who, during the revolution of 1989, bared their chests to Ceausescu’s notorious Securitate police and dared them to fire as they had so many times before. For these demonstrators, a free nation was more important than personal survival. In more controlled experimental contexts too, there is an abundance of evidence that people’s gravitation towards groups, and, when in them, the things they do with and for other group members, are driven not by personal attraction and interest, but rather by their group-level ties (e.g., see Hogg, 1992; Turner, 1984).

The most powerful evidence of this form is derived from a series of classic studies that Henri Tajfel conducted in the 1970s that laid the groundwork for subsequent social identity theorizing (Tajfel, 1970; Tajfel, Flament, Billig, & Bundy, 1971). As George Akerlof notes in the Foreword to this book, these minimal group studies involved assigning participants to groups on the basis of fairly trivial criteria such as their estimation of the number of dots on a screen, or their preference for the abstract painters Klee and Kandinsky. After this, the participants had to award points (each signifying a small amount of money) to an anonymous member of the group that they themselves were in (their in-group) and to a member of the other group (the outgroup). In fact assignment to groups was random, but the key feature of the procedure was that it excluded a range of factors that had previously been considered to play an essential role in driving group behavior—factors such as a history of cooperation or conflict, personal liking or animosity, and interdependence. Individual self-interest and personal economic gain were also ruled out because the participants never assigned points to themselves.

The robust finding that emerged from these studies was that even these most minimal of conditions were sufficient to encourage group behavior. In particular, in the initial experiments, participants tended to award more points to a person from their in-group than to someone from the out-group. In later variants, participants also reported feeling more similar to in-group members than to out-group members, as well as liking in-group members more, perceiving them to be more trustworthy, and wanting to interact with them more (e.g., Billig & Tajfel, 1973; Doise et al., 1972; Platow, Haslam, Foddy, & Grace, 2003; Platow, McClintock, & Liebrand, 1990). In this way, as Turner (1982) noted in his paper “Towards a cognitive redefinition of the social group,” assignment to these “minimal groups” produced all the symptoms of psychological group formation (in-group favoritism, altruism, liking, trust, etc.) even though the factors that individualistic theories suggest are the basis of such behavior (interdependence, attraction, similarity, etc.) were all absent. Indeed, factors like attraction, similarity, and trust seemed to be an outcome, not a cause, of group formation.

This analysis, and the findings on which it is based, make it clear why it is generally sub-optimal for leaders to entreat people to engage in group behavior on the grounds that this will advance their personal interests. Personal interest is typically not what encourages people to support football teams, to pursue organizational goals, or to join armies (although, in time, they may certainly come to see such things as personally rewarding). Moreover, they won’t necessarily do these things more or better if lured with promises of greater personal reward (Tyler & Blader, 2000).

So what does determine these things then? It was this question that led Tajfel and Turner to set about clarifying the importance of social identity for group behavior. A starting point for their analysis was the observation that in the minimal group studies and in research with more complex enduring groups, social identity made a distinct psychological contribution to “creat[ing] and defin[ing] the individual’s place in society” (1979, pp. 40–41). Like Gestalt theorists who had previously concluded that the group was more than the sum of its parts (e.g., Asch, 1952), they therefore argued that group life was characterized by something more than the mere aggregation of individual members’ individual psychologies. Groups have higher-order emergent properties and these transform the individual, while at the same time allowing individuals to engage in group processes that are capable of transforming the world (Tajfel, 1979; Turner & Oakes, 1986).

Social identity theory

In the course of developing the previous point, social identity theorizing went through two distinct phases. In the initial phase, Tajfel and Turner (1979) sought to provide a fuller explanation of the findings from the minimal group studies—and of the roots of intergroup antagonism—by formulating their social identity theory. This suggests that, when people are not only assigned to a group but also take on that group membership as the basis for their own subjective self-definition, then (1) they seek to determine the meaning and standing of the group by making social comparisons between their in-group and relevant out-groups (e.g., so that they understand what it means to be Scottish by comparing Scots to the English) and (2) they seek to define their group favorably by differentiating it positively from out-groups along the dimensions that they value (e.g., seeing the Scots as more communal and more friendly than the English). In other words, people want their own group to be better than rival groups, but what “better” actually means in practice will depend on the values and priorities of the group in question.

This psychological quest for positive distinctiveness is not the end of social identity theory as commentators sometimes assume. Rather, it is the starting point. For Tajfel and Turner recognized that we live in an unequal world where certain groups are defined negatively—for example, black people in a racist society, women in a sexist society. The key question, then, is how the process of seeking positive distinctiveness plays out in different social contexts. More precisely, the theory’s concern is with the question of when, and under what circumstances, those who are members of negatively defined groups will define themselves in terms of that group membership and act collectively to challenge their disadvantage.

Both according to the theory, and as shown by a substantial body of empirical research (e.g., see Ellemers, 1993; Ellemers, Spears, & Doosje, 1999; Haslam, 2001; Reicher & Haslam, 2006b), two sets of factors are critical. The first has to do with whether it is possible for an individual to succeed in society despite their group membership. For instance, can a woman get to the top of an organization or is she held back by a glass ceiling (Barreto, Ryan, & Schmitt, 2009)? In the language of the theory, are the boundaries between groups perceived to be permeable or impermeable? Those who perceive them to be permeable will tend to adopt strategies of individual mobility in which advancement is an individualistic enterprise. They will stress their personal qualities and downplay their group membership—perhaps even explicitly distancing themselves from (and denigrating) the group. “Other women may be emotional and soft,” they might say, “but I am tough and rational, which is why I deserve a seat in the boardroom.” However, a woman who perceives group boundaries to be impermeable will be more likely to identify with other women and to rely on collective strategies to improve her (and their) lot. The nature of those strategies depends on a further set of factors.

When relationships between groups are seen as secure—that is, if they are seen as legitimate, or if it is impossible to conceive of any cognitive alternatives to the status quo—then group members are likely to adopt strategies that aim to reshape the situation without confronting the dominant out-group directly. This might involve trying to redefine the meaning of qualities associated with one’s group. Women could stress the importance of emotions and the dangers of cold logic, for instance. Or else they could claim qualities previously denied: women who keep the family going and are able to do many tasks at once are really the stronger gender. These are strategies of social creativity. However, when relationships between groups are seen both as illegitimate and as possible to change (i.e., when they are insecure), then subordinate group members are more likely to challenge the dominant group directly. Under these conditions they will be more willing to reject the dominant group’s authority and to try to undermine it. This, then, involves strategies of social competition.

In terms of our present focus on leadership, social identity theory makes four contributions that are essential for the analysis we want to develop. The first is to expound the central concept of social identity—the notion that our sense of self can be derived from our group membership and the meanings associated with that group membership. The second is to recognize that different forms of intergroup behavior stem from the definition of the norms and values associated with this social identity. The third is to establish that, when social identities are operative (or “salient”), what counts for an individual is the fate and the standing of the group as a whole, not his or her fate as an individual. The fourth is to observe that the nature of groups and of group processes is always bound up with social context. In particular, if the meaning of who we are depends on comparisons with “them,” then our own social identities will shift as a function of who we are comparing ourselves to in any given context.

Social identity theory introduces us to these concepts, but they are more fully developed, made more explicit, and given wider application to group processes in general (including processes within groups as well as processes between groups) in a second phase of social identity theorizing. This phase centers around the development of self-categorization theory. This theory will be the focus of our interest in the remainder of this chapter because it provides us with the essential conceptual tools for crafting a new psychology of leadership.

Self-categorization theory

Returning to the minimal group studies, for John Turner (1982) the most important implication of their findings was that they suggested that the mere act of individuals categorizing themselves as group members (i.e., defining themselves in terms of a given social identity) was sufficient to produce group behavior. What these studies showed very powerfully was that it was not independence, economic exchange, or attraction that led to group behavior but the cognitive process of defining oneself in terms of group membership. We only act as group members because, and to the extent that, we are able to think about ourselves as “we” and not just “I.” As Turner famously put it: “Social identity is the cognitive mechanism which makes group behavior possible” (1982, p. 21; emphasis added).

In over 10,000 papers that deal with issues of social identity, the idea that is captured in the previous sentence is probably the single most important. Not least, this is because in the 1980s and 1990s it was this insight that led Turner to develop self-categorization theory in collaboration with colleagues at the University of Bristol (Turner et al., 1987) and at the Australian National University (Turner, Oakes, Haslam, & McGarty, 1994).

As well as recognizing social identity as the basis for group behavior, self-categorization theory also specifies a psychological process that underpins the transition from behavior that is informed by a person’s sense of his or her own individuality (what Turner (1982) referred to as personal identity) to that which is informed by social identity. In order to convey the idea that the self is no longer seen in personal terms (as “I”), Turner called this process depersonalization. Note that this is not intended to be a pejorative term. Depersonalization is simply a process of self-stereotyping through which the self comes to be seen in terms of a category membership that is shared with other in-group members. This idea is represented schematically in Figure 3.1.

The depersonalization process leads people to perceive and respond to themselves and others, not as unique persons, but as psychological representatives of the group to which they belong (not necessarily functional representatives, as some people in formal leadership positions may be). To illustrate this point, imagine a situation in which you are playing in a game of football between your team (the blues) and another (the reds). In this situation, would you see yourself and the other players on the two teams simply as unique individuals (Sam, Charlie, George, etc.)? Would you want to? The answer to both questions is probably “No.” Instead, you would see all the players (including yourself) as representatives either of your team or of the

Figure 3.1 The process of depersonalization underpinning the transition from thinking about the self in terms of personal identity (as “I”) to thinking about the self in terms of social identity (as “we”).

Note: Thinking about the self in depersonalized terms involves a process of self-stereotyping through which the self and other in-group members come to be seen as members of the same social category and hence as more similar to each other. Depersonalization (self-definition in terms of a shared social identity) thus reflects a higher-level, more inclusive, and more abstract level of self-categorization.

opposition. As a result, among other things, you would try to pass the ball to another blue player but not to a red one, and you would expect to receive a pass from a blue player but not a red one. Indeed, were you to perceive yourself and the other players as individuals (so that you saw yourself as equally different from all other players, and all players as equally different from you and from each other), it would be highly dysfunctional in this context and it would interfere with your ability to have a meaningful game of football.

But depersonalization isn’t simply about how we respond to others. It is also about how we view and respond to the world in general. Through depersonalization the group becomes the measure of all things to us. The values and norms that guide our behavior are those values and norms associated with the group with which we currently identify—and accordingly they vary from group to group. This means, for example, that if on a Sunday a person goes from a church service in the morning to a football game in the afternoon, then the values that shape his or her behavior are likely to be very different in the two contexts. In church the person may be (and want to be) meek and mild; at the game he or she may be (and want to be) rowdy and raucous.

Equally, when our sense of self is depersonalized, then the interests that concern us are those of the collective. As a result, there are times when we might even be prepared to sacrifice our individual selves in order to advance the common good—as we see when people give up their lives in wars and revolutions. As we noted above, these acts seem senseless and irrational from an individualistic perspective. However, it makes perfect (if tragic) sense to sacrifice the personal self when what really matters to a person is the standing of the collective self.

Another way of making this point is to recognize that depersonalization not only redefines the self but also redefines all self-related terms, so that these relate to “we” not “I.” Note too that depersonalization does not involve a loss of self, or an immersion of self in some amorphous collective. Instead, it involves a redefinition of self. The depersonalized self is just as psychologically (and morally) valid and meaningful as the personalized self. The depersonalized self continues to behave, feel, and think. But now what determines self-esteem is the standing of my group. Likewise, the self of self-efficacy now relates to my group’s capacity to achieve its goals. And, as we have just argued, perhaps most profoundly, the self of self-interest now becomes a matter of my group getting the things that the group values. Self-interest therefore cannot be defined in advance. For some groups it might mean more money and more material possessions. For others, however, it might mean more respect, more love—and this might even involve seeking less in the way of material goods (Sonnenberg, 2003).

Depersonalization matters for social behavior because if people weren’t able to act on the basis of social identity they would have no basis for being able to coordinate their behavior with others, for knowing who is on their side and who isn’t, or for knowing (both implicitly and explicitly) what goals they are aiming for. Without this they wouldn’t be able to play a game of football or to engage in any other form of meaningful group behavior. As we will clarify further below, this point is absolutely crucial for the analysis of leadership. For if self-categorization as a group member is a necessary basis for social collaboration, then it is equally necessary as a basis for someone to guide and shape that collaboration.

This is a point that John Adair conveys in haiku form in the third edition of his best-selling text Not Bosses but Leaders:

Leadership means…

The understanding and

Sharing of a common purpose

—Without that there can be

No effective leadership.

(Adair, 2003, p. 97)

Even more starkly, we can assert that without a shared sense of “us,” neither leadership nor followership is possible. Indeed, this is the foundational premise of our new psychology of leadership.

Social stereotyping and social influence

Social identification and depersonalization make leadership possible, but the way in which they bear on leadership can be spelled out further by distinguishing between two aspects of the depersonalization process. The first is the idea that people self-stereotype. The second is the idea that they share a self-stereotype with other members of the same category (i.e., in-group members). In combination, these two elements produce an explicit model of social influence and, as we have argued previously, leadership is intimately bound up with the ability to exert influence.

Self-stereotyping means that those who define themselves in terms of a particular social identity (e.g., seeing themselves as “a Conservative”), both (1) seek to discover the meaning associated with the category (e.g., “Conservatives value tradition and respect for authority”) and (2) strive to conform to these elements (so that, as a Conservative, “I value traditions and treat authorities with respect”).¹ Those who identify as group members therefore need information from others about the meanings associated with the group (what it means to be a Conservative), and about the implications of those meanings for situated practice (what, as a “good” Conservative, I am meant to do in the here and now).

Such reliance on others is not an exception, to be employed only under special circumstances. It is a necessity. For people cannot simply look at the world dispassionately and “know” whether it is right or wrong to hold a particular view or to perform a particular action. As the essayist William Hazlitt observed: “to know the value of our thoughts, we must try their effects on other minds” (1826, p. 133). Social reality testing is therefore necessary in order to turn our contingent beliefs (e.g., “Ithink global warming may be a serious problem”) into social facts (global warming is a serious problem) that are a basis for relevant social action (e.g., reducing carbon emissions). But, of course, we can’t rely on just anyone to confirm our understanding of the world. So who can we trust to tell us about the way things are, about what counts, and about what we should be doing? Who is in a position to tell us about group values and group action? The obvious answer is our fellow group members.

We expect those who are unambiguously group members to know something about the values and priorities of the group. And because we share the same social identity (and hence the same values and priorities), then we also expect to agree with these fellow group members on issues that are relevant to that identity. More than that, we actively strive to reach agreement on these issues. So, on the one hand, a Democrat might expect and search for agreement with other Democrats on an issue like healthcare provision, but he or she would not expect to agree with a Republican (unless, perhaps, through changes in social context, they both came to define themselves as “American”). Indeed, Democrats might expect and feel validated by disagreement with members of a political out-group. On the other hand, while two Democrats might expect to agree on healthcare, they would neither necessarily expect nor seek agreement on what team to support in the Super-bowl. In the relevant (in this example, political) domain, then, this mutual search for agreement will provide people with a relatively common perspective on the world. In addition, it will motivate them to coordinate that perspective further through processes of communication, persuasion, negotiation, and argument.

In this way, people’s motivation to reach consensus, and their ability to do so, is structured by processes of self-categorization. To test this claim, the first author and colleagues at the Australian National University conducted a program of experimental research to see whether changing participants’ self-categorizations would lead to changes in group consensus. In one such study some of the participants were first asked to think about themselves as individuals (a manipulation designed to make personal identity accessible and salient), while others were encouraged to think about themselves as Australians (making this social identity accessible and salient; Haslam, Oakes, Reynolds, & Turner, 1999). After this, the participants performed a group task in which they had to discuss what it meant to be Australian and then write their views down individually. As predicted, when the participants thought about themselves as individuals, the level of consensus in their views was comparatively low both before and after interaction. However, when participants thought of themselves as Australians, levels of consensus were generally much higher and were especially high after actual interaction with fellow group members. In this condition participants also tended to represent Australians much more positively—a pattern consistent with the motivation to achieve positive distinctiveness for one’s group when it is the basis for self-definition (as predicted by social identity theory; Tajfel & Turner, 1979). Moreover, all the studies in this program demonstrated that it was only when participants were encouraged to think about themselves as “us Australians” that they could agree about what being Australian meant, and could agree that what it meant was good.

What such research demonstrates is the capacity for self-categorization in terms of a shared social identity to organize social perception and social interaction in such a way that people’s idiosyncratic views are transformed into consensual beliefs. Indeed, what we see is that social identity theorizing provides a social psychological analysis of the transformational processes described by Burns (1978) that we discussed in the previous chapter. For it is through social identity-based processes of influence and consensualization that low-level individual inputs are fashioned into higher-order group products. These have emergent higher-order-properties that ensure that the group whole is qualitatively different from (“more than”) the sum of its individual parts. Furthermore, when combined with motivations to achieve positive distinctiveness, we can also see that these processes have the capacity to energize group members in the service of a common purpose by offering them both a sense of collective self-belief and a sense of a collective to believe in.

This is a point that is exemplified by Martin Luther King’s famous “I have a Dream” speech in which he called on his fellow Americans to stop seeing themselves and each other in terms of opposed lower-level identities as Blacks and Whites and to unite instead around the common identity proclaimed in the American Declaration of Independence and enabled through the American Constitution. It was by forging this shared American identity, King asserted, that “jangling discords” could be transformed into a “beautiful symphony of brotherhood,” and through this recategorization that they could collectively garner “the faith to hew out of the mountain of despair a stone of hope” (MacArthur, 1996, pp. 487–491). Likewise, two years earlier in his inaugural address as President, John F. Kennedy had asked his audience:

Can we forge against these enemies [tyranny, poverty, disease and war] a grand and global alliance, North and South, East and West, that can ensure a more fruitful life for all mankind? Will you join in that historic effort?

(cited in MacArthur, 1996, pp. 483–487)

Both these speeches hinge on a key point that the two leaders recognize implicitly: that transformation of the world goes hand-in-hand with transformation of identity. It is the forging of new forms of shared social identity that motivates the collective forging of new worlds.

Social identity and social cohesion

Thus far, we have examined just one aspect of the way in which shared social identity transforms the relations between people—by leading group members to seek agreement and to create consensus. This, however, is just one aspect of a general process whereby group membership transforms the relations between people by making them more intimate and mutually supportive.

As we saw in the previous chapter, psychologists have put a lot of effort into trying to understand what it is that binds people to each other and that leads them to help each other out. Along lines discussed previously, traditional individualistic approaches have sought answers in the dynamics of interpersonal attraction and exchange (for discussions see Hogg, 1992; Turner, 1982). According to such analyses we stick to groups only because, and to the extent that, we find interaction with them and their members attractive and satisfying; we help other group members only because, and to the extent that, they help us. As Napier and Gershenfeld succinctly put it, “a cohesive group is one that members find meets their needs” (1999, p. 144).

But, as we also saw previously, this approach doesn’t do a very good job of accounting for the evidence. To illustrate why this is the case, we can reflect again on our example of a fictitious football game between “reds” and “blues.” Is it personal attraction that makes the red players cohere and help each other out? To help answer this question, imagine a “thought experiment” in which children from the same class are divided randomly into red and blue football teams. When they start playing, what would determine patterns of group cohesion and cooperation during the game: whether individuals are friends in their class or whether they are playing on the same team? The answer of course is that cohesion and cooperation would be determined and predicted by team membership, not by any prior history of personal friendship or liking. Indeed, what we would expect to see here is that cohesiveness and cooperation would be emergent products of the teams that had been created. So the more meaningful the teams became for their members, the more they would cohere and the more they would cooperate.

The point of this example is that it is a sense of social identity, not personal attraction, that makes individuals work together within a group. Football players on the same team help each other out on the playing field, not because they are personal friends, but because their behavior is informed by a shared sense of group membership. Likewise, if they foul a player on the opposing team they can legitimately claim this is “nothing personal.”

Empirical support for these arguments is provided by a number of experiments that have pitted personal attraction and social identity against each other to see which of them is the better predictor of group cohesion and cooperation (for reviews see Hogg, 1987, 1992). In particular, support emerges from variants of Tajfel’s minimal group studies that were conducted by Mike Hogg and John Turner. In one of the most instructive of these, research participants were all given numbers (e.g., 32) and then told that the numbers assigned to other people (e.g., numbers in the 30s or 40s) indicated (1) whether they were people that the participants liked or disliked and (2) whether they were in the participants’ in-group or out-group (Hogg & Turner, 1985). In some situations, participants were told that assignment to groups was meaningful (based on patterns of liking) and in others they were told it was random.

The most interesting situations in the study were those in which participants had to distribute points between a person they liked but who had been randomly assigned to an out-group and a person they disliked but who had been randomly assigned to an in-group. If group processes are determined by personal liking and attraction, then people should obviously prefer to give points to someone who they like but who is in a meaningless out-group than to someone they dislike but who is in a meaningless in-group. But this wasn’t what happened. Instead, there was evidence of the opposite pattern: participants gave more points to the in-group member that they personally disliked than to the out-group member that they personally liked.

What such results suggest is that attraction to group members is as much an outcome of shared group membership as an input. Moreover, it is the depersonalized attraction that flows from a shared sense of social identity that is the critical determinant of group cohesion and cooperation. This is a point that Shakespeare (1599/2002) eloquently articulated in Henry V’s address to English troops preparing for battle at Agincourt:

We would not die in that man’s company

That fears his fellowship to die with us…

We few, we happy few, we band of brothers;

For he today that sheds his blood with me

Shall be my brother; be he ne’er so vile.

(Henry V, Act IV, Scene iii)

What this speech captures particularly powerfully is the point that it is a sense of group-based fellowship that binds the English troops together: however vile they may be as individuals, what matters is that they are brothers in arms.

As brothers in arms, then (or rather, as common group members), people tend not only to agree with each other but also to bond with each other. A wealth of evidence thus indicates that those who perceive themselves to share group membership in a given context are more likely to trust and respect each other, to help each other, and even to seek greater physical proximity to each other (for summaries see, e.g., Haslam, 2001; Haslam, Jetten, Postmes, & Haslam, 2009; Reicher & Haslam, 2010). To provide just one example of this, in a study by Mark Levine and other British colleagues (including the second author), supporters of Manchester United football club were encouraged to think of themselves in terms of their club identity (Levine, Prosser, Evans, & Reicher, 2005). They were then asked to go to another building and, as they went, they saw a man (actually, an actor) run along, fall over and clutch his leg in pain. The man in question was wearing either a Manchester United shirt, a Liverpool shirt (Liverpool are Manchester United’s great rivals), or a plain red t-shirt. When the actor wore the Manchester shirt, the participants helped him almost every time. But they hardly ever helped him if he was wearing another shirt. Next, the study was repeated with the sole difference that, at the start, the Manchester United fans were asked to think of themselves in terms of a more inclusive “football fan” identity. This time, they helped the man both when he was wearing a Manchester and a Liverpool shirt, but not when he had the plain t-shirt.

This study serves to reinforce two important points. First, we help in-group members. However, second, the way that we define our group membership can vary—specifically, it can be wider and more inclusive or narrower and more exclusive. And the more inclusive the definition of the in-group, the greater the number of those to whom we provide help (for similar findings see also Platow et al., 1999).

Pulling all these various strands of research together, we can see that the cumulative effect of shared social identity is to transform a disparate collection of people into a coherent social force. Mutual social influence leads people to agree on what is important and to strive for the same goals. As a result, their efforts, rather than pulling in different directions and canceling each other out, become aligned and additive. What is more, people then are able to coordinate their activities and to support each other. In addition, they can expect support from each other and this gives them the confidence to act in the knowledge that others are behind them and will back them up (see Figure 3.2). In sum, shared social identity is the basis of collective social power.

Social identity and collective power

To invoke the concept of power is to raise our analysis—and its social, if not societal relevance—to a whole new level. For, as Turner (2005) asserts, power is central to human affairs in general. The philosopher Bertrand Russell put it even more forcibly: “the fundamental concept in social science is Power, in the same sense in which Energy is the fundamental concept in physics” (Russell, 1938/2004, p. 4; see also Simon & Oakes, 2006). Why is this the case? The short answer is that those who have power are in a position to remake the social world and not just act in a world made by others.

We saw in the previous chapter that traditional approaches to power (especially as it relates to leadership) focus on social relationships at the individual level. By now, this should come as no surprise. According to this view, those who have power are those who have something to give or take (whether that is information, resources, or security) and those who are powerless are those who are dependent on the resources provided by others. Now, of course, it is certainly true that the control of resources allows one to control behavior. One can indeed get people to do one’s bidding either by promising them riches or by threatening them with a big stick. However, there are very clear limits to this understanding of how power operates. This was recognized by the military and political leader Charles de Gaulle in a series of lectures that he gave to the French War College in 1932. As

Figure 3.2 The role of shared social identity in transforming a collection of disparate individuals into a coherent social force.

Note: In the absence of shared social identity, individuals act idiosyncratically in terms of their personal identities (as “me”; panel 1). The emergence of a sense of shared social identity leads them to see themselves (to self-stereotype) in terms of a common group membership (panel 2). This is then a basis for coordinated, coherent, and concerted social action as individuals work collaboratively (as “us”; panel 3) to achieve identity-related goals.

summarized by David Gergen in his book Eyewitness to Power, the key message of these was that:

Authority … is more than the formal power that comes from holding office or rank; it is the informal power that comes from the respect and deference of others and thus can be infinitely greater in impact.

(Gergen, 2000, p. 65)

Generally speaking, then, the use of reward and punishment to shape the acts of others is rather ineffective and expensive. It is ineffective because while you might be able to use incentives to get people to do your bidding, this is unlikely to persuade them that what they are doing is right. In other words, as some classic studies have shown, use of incentives might lead people to comply, but it won’t lead them to be converted (Lippitt & White, 1953). What is more, as Turner (2005) notes, there is evidence that forcing people into public compliance may actually increase private rejection and hence make people more inclined to do the very opposite of what the influence agent wants—a phenomenon that the University of Kansas professor Jack Brehm termed reactance (Brehm, 1966; Brehm & Brehm, 1981). It is this that renders the process expensive. First, this is because influence agents need to keep people under continuous surveillance in order to ensure that they continue to comply (Reicher & Levine, 1994). Second, this is because the agents need to expend more and more resources to impose their will on an ever more alienated populace (Ellemers, de Gilder, & Haslam, 2004; Tyler & Blader, 2000). One can rule on such a basis, but only for so long.

At this point, it is useful to reintroduce Turner’s (2005) distinction between “power over” (what we have just been describing) and “power through” (see Figure 3.3). Whereas the former is a matter of telling people what you want them to do and using the resources at one’s disposal in order to give them incentives to get them to do it, the latter is a matter of harnessing what people want to do themselves and using that as the motor for action. Or, to use the language employed by Simon and Oakes (2006), “power over” involves imposing an external agency on the group while “power through” involves recruiting the agency of the group.

But how does one gain the latter form of power? Simply put, power through the group is gained by articulating the nature of group identity and its implications for action in context. This is because, as we have already argued, what group members essentially want is to advance the norms and values associated with their social identity. Yet, as we have also argued, people who are recognized as in-group members are in a privileged position when it comes to defining who “we” are and what we should do. Accordingly, those who are in-group members are in a better position to achieve power through the group.

There are three forms of evidence that support this contention. First, research demonstrates that the exercise of power by out-group members is

Figure 3.3 The difference between “power over” and “power through” (after Turner, 2005).

Note: The notion of “power over” (top) accords with traditional models (discussed in Chapter 2) in which leaders gain power as a result of the resources at their disposal and their resultant ability to control others by dispensing reward and punishment. In contrast, “power through” (bottom) derives from leaders’ capacity to articulate a shared group identity that leads other group members to internalize their vision and take it forward as their own. The power associated with control of resources is collectively self-depleting and shrinks as it is used, but the power associated with control of identity is collectively self-replenishing and grows as it is used. As well as this, the effect of “power over” is to reduce a sense of shared identity between leaders and followers, while the effect of “power through” is to build resources.

experienced more negatively than the exercise of power by in-group members: it is seen as more illegitimate, more punitive, more unreasonable, and more pathological (Bruins, Ellemers, & de Gilder, 1999; Haslam, 2001). This is especially true when the decisions that are made have negative consequences for those involved.

Second, as work by Eric Dépret and Susan Fiske at the Universities of Grenoble and Massachusetts has shown, there is evidence that we personally experience the exercise of power by others towards ourselves in very different ways as a function of whether those others are in-group or out-group members (Dépret, 1995; Fiske & Dépret, 1996). When one is subjected to power wielded by someone from another group, the experience is typically negative. It is seen as an imposition, it feels like meddlesome interference, and it detracts from one’s sense of being in control. But exposure to the power of an in-group member is much more positive and can even be uplifting. Most particularly, it does not detract from the sense that one is in control of one’s own fate.

Third, this divergence in the way that we experience power when it is wielded by in-group and out-group members extends to our behavior in response to the use of power (Subašić, Reynolds, Turner, Veenstra, & Haslam, in press). As a program of elegant experimental studies by Naomi Ellemers and her colleagues at the University of Leiden has shown, the more that in-group members exert their power, the more willing we are to collaborate with them (Ellemers, van Rijswijk, Roefs, & Simons, 1997; Ellemers, van Rijswijk, Bruins, & de Gilder, 1998). By contrast, the more out-group members exert power, the less we want to work with them.

Out-group members, then, generally wield power over us (Reynolds & Platow, 2003). This typically invokes resistance and it therefore consumes both energy and resources. But in-group members have much greater potential to exert power through the group. This invokes our enthusiasm and therefore creates both energy and resources. To use two physical analogies, power exercised over the group is like the petrol in a car engine: the more it is exercised, the more it is spent. However, power exercised through the group is more akin to something organic like the leg muscles we use to propel ourselves. The more they are employed, the stronger they become and the more they empower us for the future.

Once again, we can invoke Shakespeare (1623/1990) to communicate these ideas more poetically. Specifically, we see in Macbeth the tragic decline of a ruler whose betrayal of his group takes him from loyal and trusted son to despised and rejected tyrant. At Macbeth’s nadir, Angus observes:

Those he commands move only in command,

Nothing in love: now does he feel his title

Hang loose about him, like a giant’s robe

Upon a dwarfish thief.

(Macbeth, Act V, Scene ii)

And Macbeth himself realizes the consequences of his betrayal:

My way of life

Is fallen into the sear, the yellow leaf;

And that which should accompany old age,

As honour, love, obedience, troops of friends,

I must not look to have; but, in their stead,

Curses, not loud but deep.

(Macbeth, Act V, Scene v)

These, then, are the stakes. The difference between being regarded as an in-group representative and being seen as pursuing in-group goals versus being regarded as an out-group member and as opposing—even betraying— group goals is the difference between having the active, united, and aligned support of the entire membership in bringing one’s projects to fruition versus having (at best) their deep curses. It is the difference between wielding a world-making power and having that power wielded against oneself. As a clear corollary, the first priority of those who want to be effective in shaping their social world—that is, those who would be leaders—is to be seen both as being of the group and as speaking for the group. The first rule of effective leadership, then, is that leaders need to be seen as one of us.

Defining social identities

One way of summarizing the foregoing argument is to say that social identity matters. It matters for individuals because it tells us who we are, how we relate to others, who we can and cannot rely on, what is important in the world, and how we should act within it. These are all things that none of us—not even the most rugged individualist—can do without (Jetten, Postmes, & McAuliffe, 2002). Not least, this is because social identity also allows individuals to be effective in the world—as agents of a group that shapes the world rather than just as subjects who are shaped by the world. But social identities also matter for society. For they create the collectivities that serve to sustain or else challenge the status quo. They are the motors of both social stability and social change.

All this lays great significance on the question of just how categories are defined. How do we come to see ourselves and the world in terms of certain categories rather than others? How do we come to ascribe particular meanings to our group membership? As a consequence, how do particular people come to be seen as more or less representative of the in-group or of the outgroup and their proposals as embodying or else betraying group values and goals? These questions become central for a psychology that can help us understand society. They are certainly at the crux of effective leadership. They are therefore the questions with which we conclude our sketch of the processes that link social identity and group behavior.

Social identity and social reality: 1. From context to categories

For self-categorization theorists, social categories are defined in relation to social reality. The claim is simple, perhaps deceptively simple, and has three key implications. First of all, the general tendency in psychology has been to argue that social categories serve to distort social reality because they are erroneous simplifications that merely allow our limited mental apparatus to cope with the vast complexity of the world (for a review and critique, see Oakes, Haslam, & Turner, 1994). Hence it is a radical and important claim to say that categories represent, rather than misrepresent, reality.

Second, social reality itself is complex. The way we are positioned in relation to others varies constantly in our world—from place to place and from moment to moment. Soldiers killing each other in the trenches one day can be playing football together the next. As Bertolucci’s great film 1900 shows us, close families can be torn apart when members take different sides in a civil war. So, in talking of “social reality” we are required to perform a close analysis of the social relations that exist in a particular time and place, and to recognize that these are highly fluid. That is, we must not think of social reality as something that is static or generic, but as something that is continually changing and evolving, and as something that is constantly renegotiated.

Third, when we say that categories are definedin relation to social reality, it is important to understand that this relationship is, at the very least, bidirectional. That is, categories do not just reflect the existing organization of social reality in context. Categories are also used to invoke a vision of how social reality should be organized, and to mobilize people to realize that vision. To the extent that they are successful, we can say that not only does social reality create categories but also categories create social reality. An obvious example of this is the case of national categories. We use these categories all the time because we live in a world of nations where many different activities (from sport to government to economic activity) are organized along national lines. But, on the other hand, it was the dreams of nationalists— people like Atatürk, Garibaldi, and Jinnah—that inspired the broader mass of people to bring those nations into being (Reicher & Hopkins, 2001). In this section, then, we will address the former path, from contextual realities to social categories, while in the next we will consider the reverse path from categories (through mass mobilization) to social reality.

The theoretical and empirical work of self-categorization theorists suggests that the nature of the categories that we employ to define ourselves and our social world depends on two factors: (1) the fit of a particular categorization with the organization of social reality and (2) the readiness of people to employ particular categories (Oakes et al., 1994; Turner, 1985). This means that a person is more likely to define him or herself as a member of a particular group if this self-categorization maps on to what he or she sees and understands about the patterns of similarity and difference between people in that group and in other salient groups, and if that group has some prior meaning for them. For example, people are more likely to define themselves as Canadian if they see Canadians as meaningfully different from Americans and if they are also patriotic Canadians. Importantly, the very premise of the category is related to these contextual features. This means, for example, that women are more likely to define themselves along the lines of the apparent dual categorization of Latina women if they see Latina women as meaningfully different from an out-group (e.g., Anglo males). It is worth unpacking these ideas more carefully, though, because we will need to draw on them extensively in the chapters that follow.

The principles of fit

A given self-categorization is fitting if it appears to be a sensible way of organizing and making sense of the social world that a person confronts. Fit has two components: comparative and normative. Comparative fit is all about thedistribution of what people say and do, and the extent to which they form distinct clusters that are separate from others. More technically, it suggests that a person will define him or herself in terms of a particular self-category to the extent that the perceived differences between members of that category are small relative to the perceived differences between members of that category and other categories that are salient in a particular context. This is termed the principle of meta-contrast.

Normative fit is all about the content of what people are saying and doing, and the extent to which this meshes with our expectations about what members of a given group should say and do. It suggests that in order to represent sets of people as members of distinct categories, the differences between those sets must not only appear to be larger than the differences within them (comparative fit), but the nature of these differences must also be consistent with the perceiver’s expectations about the categories. If these content-related expectations are violated, then the social categorization will not be invoked.

A critical point is that fit depends on the context and the dimension along which people are judged. As an example of how comparative fit works, a Democrat surrounded by Republicans and by other Democrats would tend to define herself in terms of party allegiance during an election debate because all the Democrats are likely to be adopting a relatively similar position that is very different from the position adopted by the Republicans. But Democrat identity would be less salient at a football game, say. Here it is more likely that patterns of inter-individual differences would make salient group memberships associated with support for different football teams (e.g., “Giants” and “Mets”). As an example of how normative fit works, our Democrat watching the political debate will also be unlikely to classify participants as Democrats and Republicans (or to define herself, and act, as a Democrat) if the members of these two groups are seen to differ from each other in ways that are unexpected—perhaps if the Republicans are arguing for higher taxes and the Democrats are arguing for less spending on welfare.

The principle of comparative fit can also be used to explain how the meaning of category membership is defined—or rather, what position best characterizes the group. This is known as the category prototype. It is the position within the group that simultaneously minimizes intra-category differences and maximizes inter-category differences. Or, in less technical terms, it is the position that best epitomizes both what we have in common and what makes “us” different from “them.”

Again, because the prototype is rooted in comparisons within and between groups, it shifts as a function of who exactly we are comparing ourselves with. This point is illustrated by the research of Nick Hopkins, Martin Regan, and Jackie Abell (1997) into the meaning of “Scottishness.” This research found that when Scots compare themselves to Greeks they consider themselves to be distinctly hard-working, but that when they compare themselves to the English they mark themselves out by their friendliness. This is a phenomenon that, in one form or another, all of us will have experienced. The stance of liberals shifts as a function of whether they are debating with socialists or conservatives. The way we see our team depends on whether we are playing rivals marked by their skill or by their endeavor. And yet, familiar as this phenomenon may be, it cuts against one of the most well-worn assertions within and beyond psychology: that the characteristics ascribed to groups are rigid stereotypes that are resistant to evidence and to change. Although widespread, this view is incorrect. For if one takes care to observe across settings (or to conduct experiments that systematically vary the context of group description) one finds that the ways in which we characterize in-groups is a fluid and sensitive function of comparative context.

One very important implication of the comparative fit principle is that self-categories will become more inclusive and will be defined at a higher level of abstraction as a person’s frame of reference is extended to include a range of very different stimuli. This point is represented schematically in Figure 3.4. Here we see that (other things being equal) a female employee, Beth, who compares herself only with another female employee, Amy, will tend to categorize herself in terms of personal identity and accentuate the difference between herself and Amy. However, as the context is extended to include different others—for example, men—Beth is more likely to categorize both herself and Amy in terms of a higher-level social identity, as “us women” who now appear less different from each other, in comparison to “those men.” To test exactly this idea, Hogg and Turner (1987) conducted a study in which they organized participants either into same-sex pairs (i.e., male–male

Figure 3.4 Variation in self-categorization as a function of comparative context.

Note: In Context 1 Amy and Beth self-categorize in terms of lower-level personal identities that accentuate their differences from each other. However, in Context 2 the comparative context is extended to include more different others (here men), and Amy and Beth are now more likely to define themselves in terms of a higher-level social identity (i.e., as members of the same social self-category) and hence appear more similar to each other.

The important theoretical point here is that as comparative context is extended, people tend to self-categorize at a more inclusive, higher level of abstraction. So, if Amy was from a Northern State in the USA, and Beth from a Southern State, they would be more likely to self-categorize as Americans (rather than as Southerner and Northerner) in a context that included people from a different country (e.g., Canada).

or female–female; Context 1 in Figure 3.4) or into four-person groups comprising two males and two females (Context 2). As predicted on the basis of self-categorization theory, participants were more likely to define themselves in gender-based terms and to accentuate their similarity with other members of the same sex when men and women were present rather than just another person of their own sex.

This claim that self-categorization varies predictably as a function of comparative context has been confirmed in a large number of other studies too (in particular, see Gaertner, Mann, Murrell, & Dovidio, 1989; Haslam & Turner, 1992, 1995). This work serves to underline two important points. The first is that there are no inherent, stable differences between representations labeled “in-group” and “out-group.” The second is that there is no predefined, universal identity in terms of which a person will define him- or herself (and others). Indeed, the very same people can be defined as “in-group” or “outgroup” in different contexts. The party member who is seen as an out-group opponent in the narrow context of the initial campaign to select an election candidate may be redefined as an ally in the broader context of the election itself. In the United States this pattern is typically observed every 4 years once the political focus moves from the Primaries to the Presidential election proper. Thus in 2008 Hilary Clinton and Barack Obama categorized themselves into opposing camps during the Democratic Primaries, but subsequently united as Democrats to fight the Republicans for the right to occupy the White House.

Based on these two principles of fit, we are now in a position to elaborate on what, above, we characterized as “the first rule of leadership”: that a leader must be seen as “one of us.” To be seen as such, an individual must not simply be a member of a currently salient social category, but must also exemplify what makes “us” different from the relevant “them.” Leadership potential thus becomes a function of one’s in-group prototypicality relative to other aspirants (Turner, 1991).

However, leadership is not just a matter of being. It is also a matter of doing. Leadership depends on acting for the group. It involves expressing and advancing the norms and values of the group. Indeed, this message has been implied at various points in our discussion throughout this chapter. If group members are motivated to enhance the relative standing of their group, if group members are concerned with the fate of the group, if their sense of “self-interest” is centered on the group interest, then the leader who represents them (in all senses of the word) will be one who promotes the good of the group. Of course, what we see as the collective good depends on what we believe in and care about. It is expressed through the norms and values of the in-group. All these points converge to define another general lesson for those who seek sway over others. They must not only be seen to be of the group but also for the group. The second rule of effective leadership, then, is that leaders need to be in-group champions. Whatever they do, they must be seen to do it in our collective interest.

The principle of perceiver readiness

Our discussion of comparative and normative fit could be taken as suggesting that people mechanically process information about their social world in a dispassionate, uninvolved way in order to decide how they should define themselves and others. However, this is not the case. For as well as being determined by features of the social world that a person confronts in a particular context (i.e., fit), categorization also depends on a person’s readiness to use a particular set of categories (the principle of perceiver readiness or accessibility; Oakes et al., 1994). This means that self-categorization also reflects people’s prior expectations, goals, and theories—many of which derive from their pre-existing group memberships and previous group encounters. In the above examples, then, Amy and Beth would be more likely to categorize themselves as “us women” if this had been a meaningful self-categorization in the past (e.g., if they were feminists or belonged to a women’s network). Similarly, Barack Obama and Hilary Clinton self-categorized as Democrats because this was a pre-existing political organization of which both had been members for some time. In this way, people organize and construe the world in ways that reflect their social histories and this also lends stability and predictability to their experience.

Social identification—the extent to which a particular group membership is valued and self-involving and contributes to an enduring sense of self—is therefore one particularly important factor that affects people’s readiness to use a given social category in order to define themselves. Among other things, when a person (e.g., Amy) identifies strongly with a given group, she may more readily interpret the world, and her own place within it, in a manner consistent with that group’s values, ideology, and culture. The more Amy identifies with other women, the more likely she is to define herself as a woman; the more that she identifies with the Democratic Party and its members, the more likely she is to define herself as a Democrat.

On the basis of this analysis, we can therefore see that social identity salience is determined by the interaction between people’s present context (the meaningfulness of particular groups in the present) and their prior experience (the meaningfulness of particular groups in the past). That is, category definitions may be constrained by existing realities, but they are not rigidly determined by them. People do have autonomy. They do actively construct the world. This becomes even more apparent when one adds an obvious dimension to the analysis that has been missing so far. Categorization is not only about the past (prior experience) or the present (existing social organization), it is also about the future. We saw this in the case of nations and nationalism. To make the point more generally, categories are as much about saying how things should be as about how they are. This is particularly relevant for leadership, since the tasks of leaders are always future-oriented, whether this is a matter of preserving existing social arrangements or of transforming them. Added to this, categories are not just about envisaging the future. They are also, as we have stressed, tools for making the future. They are world-making things. This takes us to the second aspect of the category–reality relationship: how social categories make social reality.

Social identity and social reality: 2. From categories to context

The general process that leads from definitions of social identity to social reality has already been outlined. Category definitions serve to unite and shape the actions of those who are category members. They motivate and mobilize people as a social force to transform the social world so that it comes into line with their norms and values. This much mostly restates what has gone before. In order to move on, we need to consider first how different elements of category construction relate to different aspects of collective mobilization, and then we need to ask what leads identity definitions to be accepted by their intended audience.

The relationship between category construction and group mobilization can be unpacked by examining the various dimensions along which category definitions impact on collective action. This is precisely the issue that is addressed through the analysis of self-stereotyping. To recap, first people identify with a group. Second, they seek out the meanings associated with its social identity, notably from those who are seen to embody this identity. Third, they seek to act on the basis of these meanings. These ideas can be restated in terms of questions and answers related to three core dimensions of collective action:

Q1:	Who will act together? A1: Those who identify as members of a common social category.
Q2:	What will they do together? A2: Act in terms of shared group norms and values.
Q3:	Who will be able to guide them? A3: Those who embody what makes the group a distinct and meaningful entity.

To flesh these points out, we can reframe this argument in terms of three principles of collective mobilization:

Principle 1: Category boundaries define the size of the mobilization. The wider that category boundaries are defined, the greater the extent of the mobilization and the greater its potential power to shape the social world. For example, an appeal to people on the grounds of social class can only mobilize minority sections of the population, whereas an appeal to nationhood can recruit the great majority. This is one reason why political parties across the spectrum (and not only those who are nationalists) regularly “fly the flag” and root their appeals in nationhood. However, this only applies because the forms of action for which these parties are seeking to mobilize people (i.e., electoral support) depend on winning the widest proportion of the population as a whole. If, say, one were focused on mobilizing trades unionists to effect change through strike action, then it would make perfect sense to use more narrowly defined class categories—because they encompass all those required to make the strike succeed. In general terms, therefore, effective shaping of social reality depends on the deployment of categories that include the widest possible proportion of those required to accomplish the intended action.

Principle 2: Category content defines the direction of the mobilization. The ways in which the norms and values of the group are defined determines what sort of actions are seen as appropriate. As a corollary, any particular form of action will be more liable to gain collective support to the extent that it is seen to reflect the norms and values of the group. So, while it may be true that all electoral politicians employ the same national categories because they are appealing to the same extended audience, they ascribe different content to the categories because they are trying to mobilize them to different ends. In the Scottish case, for example, leftists tend to see Scots as a caring and communal people, while conservatives stress their entrepreneurial spirit (Reicher & Hopkins, 2001). In general terms, then, effective action depends on defining both the action and the content of the category in a way that allows the action to be seen as an embodiment of the category.

Principle 3: Category prototypes define who can influence the mobilization. Those who seek collective influence need to define themselves (their biographies, their character, their actions), the context, and hence the group in ways that make them appear to be the embodiment of group identity. This can be achieved in many ways. As we shall see in later chapters, it may involve focusing on something as small as the hat on one’s head, or something as large as one’s entire ancestry. Nothing is too trivial to include in this process of self-construction. The general message, though, is that effective influence over collective action depends on an effective fusion of personal and group narratives so that the would-be leader becomes the embodiment of the group.

But where do these constructions of boundaries, content, and prototypes come from? By now, we have already begun to answer that question. They come from leaders themselves. Indeed, precisely because social category definitions constitute such a powerful social force, then anyone who is interested in shaping the world—political actors, social movement activists, and so on— needs to be interested in defining categories. Our third rule of effective leadership, then, is that leaders need to be skilled entrepreneurs of identity. Their craft lies in telling us who we are and in representing their ideas as the embodiment of who we are and what we want to be. If they succeed, our energy becomes their tool and our efforts constitute their power (see Reicher, Haslam, & Hopkins, 2005; Reicher & Hopkins, 2001).

Through the operation of the above principles we can see how category definitions shape the collective mobilizations that in turn shape social reality. Moving on, we now need to address the question of when and why people accept particular definitions and act on them. In part this will come down to the plausibility of the construction itself—the extent to which a specific version of identity incorporates well-worn understandings of who and what we are: the events that every child learns at school, the historical figures and cultural icons who appear on our landscape in statues, place names, even postage stamps. In this way, understandings of the past can make for a compelling vision of what the future should be—a vision compelling enough to move people to action.

Yet however compelling a vision, and whatever its ability to mobilize in the short term, vision alone is not enough to sustain our understanding of social identity in the longer term. Vision is only useful if it allows us to see and then create a better future. Accordingly, if collective mobilization fails to translate a definition of identity into experienced reality, then that definition will fall by the wayside. By contrast, where mobilization does succeed in creating realities that reflect a given definition of identity, then that definition will gain in support.

The same goes for those who offer these various definitions. Those leaders who turn visions of society into social realities will succeed over those whose visions remain limited to the imagination. Our fourth rule of effective leadership, then, is that leaders need to be embedders of identity. They must not only tell us who we are, they must also make a world in which our sense of who we are can be made to matter.

Social identity and social reality: 3. A process of historical interaction

We are now at a point where we can bring together the two sides of the identity–reality relationship that we have been discussing in the previous two sections. One way of encapsulating these two sides is to say that categories are about both being (reflecting existing reality) and becoming (creating future realities; see Reicher et al., 2010). However, it is important to understand that the relationship between these things is dynamic, in the sense that each plays a role in determining the other. Thus who is able to represent us depends on the (comparative and normative) definition of our group, but the meaning of our group is also structured by those who represent us. In this way, as Figure 3.5 suggests, the relationship between social reality and social categories is a continuously evolving cycle. Here, which element one sees as primary depends on where and when one enters into the cycle. But in fact no single element has any ultimate priority. Reality feeds into categories, which feed into collective action, which feeds back into reality. There is no natural starting point or finishing point to this process and hence no element predominates over the others. Every element is essential to the overall process.

Appreciation of this cycle has important implications when we attempt to resolve some of the core dilemmas that we identified in our review of the classic and contemporary leadership literatures in the previous two chapters. Notably, we see that it is possible to acknowledge the creativity and agency of leaders without thereby denying agency to followers. And, at the same time,

Figure 3.5 The ongoing and dynamic relationship between social reality, prototypicality, and leadership.

Note: The clear arrows relate to the reactive process whereby social context determines the group prototype and the prototypicality of a given leader, which then affects his or her capacity to display leadership; the shaded arrows relate to the proactive process whereby leaders initiate collective action that redefines the social context so as to change the group prototype and his or her prototypicality.

we see that it is possible to recognize situational and other constraints on leadership without turning leaders into mere ciphers.

As we noted at the start of this chapter, leadership thus involves a relationship between leaders and followers in a group, and it is this collective framing of leader–follower relations that allows the analytical problems of previous research to be overcome. More specifically, what we see is that leaders gain their effectiveness through their ability to represent and advance the social identity of the group. On the one hand this acts as a constraint on them. Leaders cannot say anything or get followers to do anything. They are reliant on their ability to persuade followers of their prototypicality and normativity, and this in turn depends on features of social context. But on the other hand, it is social identity that enables leaders to energize people with their vision, and to recruit the agency of followers in order to transform both their self-understanding and the world they inhabit. Leaders, followers, and situations are not static entities that exist independently of each other, but elements that interact to shape each other—and it is through this interaction that the power of leadership is unleashed.

Conclusion: Setting the agenda for a new psychology of leadership

In this chapter, we have outlined the social identity approach to group processes and used it to derive one framing condition and four rules for effective leadership. The framing principle is simply that the exercise of leadership, in the sense of influence over a collectivity, depends on the existence of shared identity among those who constitute that collectivity. This point was brought forcibly home in some work that two of us conducted a number of years ago that revisited the paradigm of the Stanford Prison Experiment (SPE) that we discussed in Chapter 2.

As in the SPE, the study involved randomly dividing ordinary people into Prisoners and Guards within a simulated prison environment (see Figure 3.6). Unlike the SPE, what we found was that, as the study progressed, the Prisoners identified more and more with their group while the Guards identified less and less. The reasons for this difference need not detain us here (for a full account, see Reicher & Haslam, 2006b). What concerns us for the present are the consequences of this asymmetry.

What we found was that as the Prisoners developed a sense of shared social identity they became more effective in coordinating their actions as a group— notably in challenging the authority of the Guards. At the same time they also became more willing and more able to choose a leader to represent them. Indeed, this emergent leadership was apparent not only to the Prisoners themselves but also to the Guards and—as the data in Table 3.1 indicate—to independent observers (Haslam & Reicher, 2007a).

Table 3.1 Observers’ perceptions of leadership-related processes in the BBC Prison Study (data from Haslam & Reicher, 2007a)

In total contrast, as shared identity among the Guards declined, so they became less and less able to work together. Among other things, this meant that they couldn’t establish a basic system of work shifts because they were afraid that other Guards would do things they disapproved of in their rest periods. This meant that everyone worked all the time, not to help each other but to hinder each other. And the harder that everyone worked, the more exhausted they became and the less they achieved as a group. Their efforts cancelled each other out. Leadership in such a context was quite impossible. No one trusted anyone else, no one would accept anyone else to represent them. Again, this was apparent to both Prisoners and Guards as well as to onlookers. Moreover, the net result of these opposing dynamics was that the Prisoners’ resistance triumphed and the Guards’ regime was overthrown. The study was thus a perfect illustration of our contention that leadership and social identity go hand in hand and that no leader can represent us when there is no “us” to represent.

We have argued, however, that where people do have a shared sense of social identity (a sense of “us”), there are four key rules to effective leadership. Because these provide the structure for the chapters that follow, these are worth reiterating.

Rule 1: Leaders need to be in-group prototypes. The more representative an individual is seen to be of a given social identity—the more he or she is clearly “one of us”—the more influential he or she will be within the group and the more willing other group members will be to follow his or her direction. This is a point that we will expand on in the next chapter.

Rule 2: Leaders need to be in-group champions. In order to take followers with them, leaders must be seen to be working for the group—to be “doing it for us”—rather than to be “in it” for themselves or for another group (“them”). In other words, leaders must advance the collective interest as group members see it. This, together with the allied issue of how leaders promote justice and fairness, will be the focus of Chapter 5.

Rule 3: Leaders need to be entrepreneurs of identity. Leaders don’t just wait around until they and their policies come to be recognized as prototypical of the group. Rather, they work hard to construct identity in order to ensure that they and their policies are influential. The various elements of this identity entrepreneurship are explored in Chapter 6.

Rule 4: Leaders need to be embedders of identity. It isn’t enough for a leader simply to construct a plausible version of identity. As well as this, the sense of who we are and how we believe the world should be organized that is associated with a particular sense of social identity needs to be translated into social reality. The importance of this embedding process and the various levels on which it can be achieved is explored in Chapter 7.

In our final chapter, Chapter 8, we bring our analysis to a close by drawing these various points together and reflecting on three key challenges for a new psychology of leadership. Here we argue that in order for the social identity analysis to move forward, we need (1) to recognize and overcome existing prejudices in thinking about the topic; (2) to clarify how social identity principles can be translated into practice; and (3) to tackle the question of how different models for defining identity serve to sustain different forms of political structure. In this way, we conclude by showing how our approach provides new and important insights into the perennial questions that arise at the “sharp end” of leadership as it is taught and practiced around the world.

3Foundations for the new psychology of leadership