Patterns of Power
In its most general sense, power is the production of causal effects. It is ‘the bringing about of consequences’ (Lukes 1978: 634; Lukes 1986). The power of a river, for example, is manifest in its causal effects: it erodes a bed, transports rock material from one place to another, and produces a delta or a flood plain. Similarly, the power of electricity is manifest in the illumination of light bulbs, the heating of cooker elements, and the operation of underground railways. This idea of power as causal power
is also integral to the very idea of human agency: to be an agent is to exercise causal powers that produce specific effects in the world. These human powers comprise the ‘transformative capacity’ possessed by human agents (Giddens 1976: 110; Giddens 1982). To act is to have causal powers, and these powers constitute the ‘potency’ that defines an organism as a human agent. Power is ‘an actor’s general ability to produce successful performances’ (Wrong 1979: 1).1
To talk or to write about social power involves a move beyond this basic causal vocabulary. Social power is a form of causation that has its effects in and through social relations (Isaac 1992; see also Isaac 1987). In its strongest sense, it is an agent’s intentional use of causal powers to affect the conduct of other participants in the social relations that connect them together. In this book, my concern is with social power in this sense and, unless anything is stated to the contrary, the word ‘power’ will be used exclusively in its social sense.
At its simplest, power is a social relation between two agents, who may usefully be called the ‘principal’ and the ‘subaltern’.2
A principal is the paramount agent in a power relationship, while a subaltern is the subordinate agent. The principal has or exercises power, while the subaltern is affected by this power. Concretely, of course, such relations are rarely so one-sided as this implies. A principal in one relationship may be a subaltern in another, and subalterns often exercise countervailing power to that of their principal. Analytically, however, the dynamics of power relations can initially be understood in terms of this relatively simple relation of principal to subaltern.
The intentions or interests of principals have been central to many discussions of power. To qualify as a social power relation there must be more than simply a causal influence between agents. It is for this reason that Wrong (1979) holds that it is a form of causal influence that involves the production of intended effects. An exercise of power, he argues, typically involves an intentional intervention in a chain of causal effects. An accidental or incidental effect of an agent’s actions cannot be regarded as an exercise of power unless it is a foreseen consequence of these actions (1979: 4). A power relation, then, involves the intention to produce a particular effect or the desire to see a particular effect occurring. Power is an intended or desired causal effect; it is an effect that realises a purpose (Beetham 1991: 43). A power relation cannot, therefore, be identified unless there is some reference to the intentions and interests of the actors involved and, especially, to those of the principal (Wartenberg 1990: 65). An intention or a desire rests upon a felt or perceived interest that the principal believes will be furthered if he or she brings about particular kinds of causal effects in the field of social relations.
As agents, both principals and subalterns are, in crucial respects, free: they have a degree of autonomy in shaping their actions, which are never completely determined by external factors. This is not to say that individuals must be seen, in classical liberal terms, as sovereign individuals making perfectly informed and unconstrained rational choices on the basis of their pure ‘free will’. It is, rather, to recognise that agents always have the ability to choose among alternative courses of action, however constrained these choices may be. Lukes (1974) has stressed that the most important implication of this is that social power has to be seen in relation to the possible resistance that others can offer to it. The subaltern must be thought of as being able to act otherwise than in conformity with the wishes of the principal, as having the capacity to resist. In Foucault’s words, ‘Power is exercised only over free subjects, and only in so far as they are free’ (Foucault 1982: 229). The power of a principal consists in the ability to freely pursue intentions and interests; the power of a subaltern consists in their freedom to resist (Benton 1981: 296). Social power, in its most general sense, then, involves the socially significant affecting of one agent by another in the face of possible resistance.
The exercise of power and the possibility of resistance to it establish a dialectic of control and autonomy, a balance of power that limits the actions of the participants in their interplay with each other. In power relations, then, ‘individual or collective subjects … are faced with a field of possibilities in which several ways of behaving, several reactions and diverse comportments may be realized’ (Foucault 1982: 229). Acts of power occur when principals are able to restrict the choices that subalterns are able to make: the greater this restriction (the more limited the range of choices available to subalterns), the greater is the power of the principal (Wartenberg 1990: 85). As Lukes has put it:
To use the vocabulary of power … is to speak of human agents separately or together, in groups or organisations, through action or inaction, significantly affecting the thoughts or actions of others. In speaking thus, one assumes that although agents operate within structurally determined limits, they none the less have a relative autonomy and could have acted differently. (Lukes 1977: 6–7)
Power relations involve the possibility of conflict because of this choice among alternatives, but resistance is not always expressed in overt conflict or struggle. Consider, for example, a factory manager who orders an employee, on the threat of dismissal, not to smoke at work. If the worker has no intention of smoking – and is, indeed, a non-smoker – then there will be no conflict and the manager has clearly not had to exercise any actual power to prevent the worker from smoking. The manager does, however, still hold the power of dismissal, which is ready and waiting to be exercised should this or any other worker choose to smoke at work. Power relations involve the possibility of conflict, but only the exercise of power need involve actual conflict, however minimal.
These considerations show how important it is to distinguish between exercising power and holding power (Dahl 1968). At its fullest, a power relation involves the deliberate, intentional intervention of a principal in the course of interaction so as to produce a specific and particular effect on a subaltern. Such an exercise of power comes closest to the everyday understanding of social power. An agent who has this capacity to affect others may, however, be able to achieve this without actually having to do anything at all. This occurs when others anticipate their intentions and their likely actions and act in relation to these. Such ‘anticipated reactions’ (Friedrich 1937) are apparent when agents act in a certain way because they believe that, if they do not, they will be affected in some socially significant way by another who has the capacity and the intention to do so.
Action on the basis of an anticipated reaction is an effect of a principal’s power, even though he or she does nothing directly to make this power effective. Indeed, anticipated reactions may even increase a person’s power. The leaders of a political party, for example, may believe that a business leader is wealthy enough to grant or withhold financial favours, and so may formulate policies that accord with his or her wishes. They may, however, misunderstand the true extent of the person’s wealth, his or her actual financial power being amplified by the mistaken beliefs of the party leadership. Such reputational power should not be overstated, but neither should it be ignored:
Power can be effected, then, without being exercised. This conclusion is central to the argument that power is, at root, a capacity. To have a capacity is to be in a position to do something (Morriss 1987: 81), and any capacity may remain latent without thereby ceasing to be a capacity. As Haugaard (1997) has succinctly put it, a Ferrari racing car has the power to travel at 120 miles per hour, even when it is parked in a garage with its engine switched off. Any disposition can persist without being exercised. Someone may, for example, know how to ride a bike even though they are not currently cycling. Their knowledge does not suddenly come into existence when they get on a bike and disappear again when they dismount. A principal may, therefore, have a capacity to act in some way without actually doing so. To have power is to have an enduring capacity or disposition to do something, regardless of whether this capacity is actually being exercised.
It could be suggested that an unexercised capacity might as well not exist, as it might not seem to make sense to describe someone as ‘powerful’ if they never do anything with their supposed power. However, a powerful person who does not exercise their power is like a miser who hoards a fortune but lives as a pauper. The miser retains the capacity to spend and could escape his or her poverty in an instant. Similarly, the actor with the potential to exercise power can, at any moment, choose to realise this potential by affecting the actions of others. Power – like knowledge and money – can be held in readiness for use whenever it is needed. The anticipation of its use, furthermore, means that power can have significant social consequences even when there is no explicit and overt intervention by the principal.
Mainstream and Second Stream
This core idea of power has been developed in two broad directions, forming two streams of power research (Ball 1975; Ball 1976; Clegg 1989). The mainstream tradition has been principally concerned with the episodically exercised power that one agent has over another. The second stream of power research focuses on the dispositional capacity to do something. It is the ability that actors have to facilitate certain things that lies at the centre of attention. Mainstream views concentrate on what in French is called ‘pouvoir
’ while the second stream has concentrated on ‘puissance
The mainstream view of power takes the sovereign power of a state as its exemplar (Macpherson 1962; Abercrombie et al. 1986). The classic statement of this is in Weber’s analysis of the structuring of authority and administration in modern and pre-modern states (Weber 1914). While later work on sovereign power has continued to focus on states and the political power of individuals and groups in relation to states, it has also followed Weber’s recognition that power exists in other sovereign organisations, such as businesses and churches. Economic power, for example, has been studied in national and multinational enterprises and in the actions of the individuals and groups involved in their ownership and control, and in similar ‘stakeholder’ relations. A key area of research has been the relationship between economic power and political power, as explored in elitist and Marxist theories of ruling classes and power elites (Mosca 1896; Mills 1956; Miliband 1969).
Weber saw power as manifested in the chances that an actor’s will can be imposed on the other participants in a social relationship, even against their resistance (Weber 1914: 942). According to this point of view, actors seek to make others do what they would otherwise not do, and they resist the attempts of others to make them act in ways contrary to their own preferences. In this ‘constant sum’ or ‘zero sum’ view, power relations are seen as asymmetrical, hierarchical relations of super- and sub-ordination in which one agent can gain only at the expense of another. They must be seen in terms of the conflicting interests and goals of the participants and the abilities of some to secure the compliance of others. There is a given distribution of power within any society, and some agents have more of this power than others. Struggles over the distribution of power will always involve both winners and losers.
This view of power was forged into a formal model by Lasswell and Kaplan (1950) and was given mathematical form by Simon (1953), Dahl (1957), and Polsby (1960). These writers, however, limited their attention to the behavioural and intentional aspects of the actual exercise of power. They saw power as the exercise of causal influence within the decision-making processes of sovereign organisations. Powerful actors are those who make decisions or who participate in the decision-making apparatuses of sovereign organisations. Dahl, for example, saw a principal having power over subalterns because he or she is able to make decisions to which subalterns conform.
This approach has largely been developed through a reliance on an individualistic and rationalistic view of action that stresses the autonomy and rationality of agents as they choose from among alternative courses of action. The paradigm example of such action is Weber’s type of instrumentally rational action (‘zweckrationalität’). In this framework, individuals have preferences, appetites, desires, or interests, and they pursue their own interests at the expense of those of others. Each agent is a maximiser, or satisficer, of advantages. Drawing on the rational-choice theories of market behaviour produced by economists, power relations in and around sovereign states have been investigated as if they formed a ‘political market’ (Downs 1957; Buchanan and Tullock 1962).
Thus, Dowding (1996) has argued that power should be seen as the capacity of one agent to deliberately change – in line with his or her own interests – the ‘incentive structure’ of costs and benefits faced by another agent. In a similar vein, Wartenberg (1990: 85) holds that an agent becomes a principal in a power relation if, and only if, that agent can strategically constrain the action alternatives available to a subaltern. The constrained alternatives form an integral element in the subaltern’s strategic calculations about future courses of action, and their consideration of the rewards and costs attached to particular alternatives may lead subalterns to act contrary to certain of their own interests. What is important, Dowding argues, is that an altered incentive structure allows a principal to achieve desired outcomes by means of the actions of others.
This purely individualistic and rational-choice version of the mainstream view is more limited than Weber’s own ideas on power, and this led writers such as Wrong (1967–8) and Bachrach and Baratz (1962; 1963) to emphasise a whole second face to the exercise of power. The first face of power, studied by Dahl and his followers, comprises the most obvious and overt processes of formal decision-making. The second face of power, on the other hand, comprises the hidden, behind-the-scenes processes of agenda setting that Bachrach and Baratz termed ‘nondecision-making’. For Bachrach and Baratz, a principal has power over a subaltern to the extent that he or she can prevent the subaltern from doing something that they would otherwise do or that they would like to see happen. This can be achieved, for example, by preventing an issue from coming to the point of decision, thereby excluding the subaltern from any effective say about it.
Lukes’s (1974) important critique of power studies was mainly concerned with the problems that he identified in this mainstream of power research. While he recognised the validity of distinguishing between the two facets of power – though he rather misleadingly described them as two ‘dimensions’ of power – he argued that it was also necessary to add a third facet to the analysis. This aspect of power took more seriously the importance of the ‘real interests’ of which actors may normally be unaware. From this point of view, Lukes argues, the power of a principal can be manifest in the ability to make a subaltern believe that their interest lies in doing something that is, in fact, harmful to them or contrary to their deeper interests.
This argument has generated much critical discussion about the nature of interests, but Lukes (1977) and some other contributors to this discussion (for example, Giddens 1982) have extended the argument to raise a matter that points beyond the bounds of the mainstream approach. In addition to the need to incorporate real interests and ‘false consciousness’ into the model of power, they pointed to the need to take social structure more seriously. Power is not limited to the ‘discrete intervention by a social agent in the life of another social agent’ (Wartenberg 1990: 72), but may also involve the existence of enduring structured constraints over actions. Lukes holds that this is most clearly apparent in what he calls the ‘facilitative power’ that may be held by classes and other collective actors. Though there are problems in Lukes’s distinction between structural constraint and forms of structural determination that do not involve power (Layder 1985), he made the duality of structure and agency central to discussions of power.
In raising these issues, Lukes and Giddens were echoing ideas that had emerged as central themes in the second stream of power research. This second stream of research has not been so tightly defined as the mainstream, and it has no equivalent founding statement to that of Weber. It has, nevertheless, been an important source of critical commentary on that mainstream. The second stream begins from the same core idea of power, but it takes this in a different direction. Its focus is not on specific organisations of power, but on strategies and techniques of power. It sees power as diffused throughout a society, rather than being confined to sovereign organisations. According to this view, power is the collective property of whole systems of cooperating actors, of the fields of social relations within which particular actors are located. At the same time, it stresses not the repressive aspects of power but the facilitative or ‘productive’ aspects. Of particular importance are the communal mechanisms that result from the cultural, ideological, or discursive formations through which consensus is constituted. This is a ‘variable sum’ or ‘nonzero sum’ view of power: all can gain from the use of power, and there need be no losers.
A key figure in the development of this second stream is Gramsci (1926–37), whose concept of hegemony highlighted a mechanism of power through which a dominant class can secure the consent of subaltern classes without the need for any direct use of coercion or repression. Through the cultural formation of individuals in schools, churches, factories, and other agencies of socialisation, a dominant class can secure a more stable position for itself than it could possibly enjoy simply through exercising the repressive powers of a state. Althusser (1971) employed this idea, arguing that the ‘repressive apparatuses’ of a state work alongside its ‘ideological apparatuses’ to sustain social control. It is through ideology, he argued, that individuals are ‘interpellated’ – called out – as subjects with the specific characteristics and desires that commit them to the very actions that are required of them by their class position.
Working from a different theoretical basis, Arendt (1959) also stressed the collective capacities that are inherent in political communities. Power relations, she argued, are formed through communicative actions in discursive communities. People communicate with each other through their speech acts, and the shared symbols that they use allow them to co-ordinate their actions and so to act in concert. Power comes into existence wherever the members of a group are forged together through such bonds of solidarity and organise themselves for collective action. Such a group acquires an identity and purpose and enables or ‘empowers’ its constituent individuals to act in the name of, or on behalf of, the community as a whole (see also Lindblom 1977).
Habermas (1981a; 1981b) shares this view and adds that it is the discursive structures of the socio-cultural life-world that are the bases of such power. Habermas draws on the ideas of Parsons (1963), as well as Arendt, as it was Parsons who saw power as resting on a framework of communal trust and shared values within a ‘societal community’.4
According to Parsons, power is rooted in the shared values that define the goals and purposes of a community. Societal communities are seen as organised around those values in which individuals have trust or confidence, and that define positions of leadership whose occupants are endowed with the legitimate right to issue commands and to make policy in relation to the values and purposes that the members of the community hold in common. Parsons further argues that the diffused character of power makes it a circulating medium analogous to money. It is not confined to sovereign organisations but is something that all individuals can hold, in varying degrees, and can use or exchange in their actions.
Parsons has been criticised for overemphasising value consensus and for implying that societies are generally characterised by the perfect socialisation of their members into this consensus (Wrong 1961). In order to avoid this tendency in Parsons’ work, Barnes (1988) has proposed an approach to power that opens up this argument and takes it in a more acceptable direction. For Barnes, the basis of social order is to be found in shared cognitive meanings – not shared values – and power, therefore, has to be related to the symbolic orders of meaning that underpin particular bodies of knowledge. Power is, he argues, a capacity for action that someone has by virtue of the social distribution of knowledge: an individual’s power is their portion of the collective power of the community as a whole, the community whose knowledge they bear and share (1988: 57). It is particularly closely associated, he argues, with those communal structures of meaning that Weber saw as associated with social status and the social estimation of honour (1988: 144).
The most influential statement of this second-stream view of power in recent years has been that of Foucault (1975; 1976), who argued that analysis of the repressive powers of command within states and other sovereign organisations provides only a part of the full picture.5
Power exists throughout the social sphere that surrounds and penetrates the public, political sphere of sovereign power. What Foucault called ‘discursive formations’ operate through mechanisms of socialisation and ‘seduction’ – to use a term from Baudrillard (1981) – that bring about the cultural formation of individual subjects. They bring particular kinds of mental orientation and routinised actions into being. Where Arendt and Parsons saw discursively formed power in a positive way, as a form of collective empowerment, Foucault stressed its negative face. For Foucault, it remains a source of social control, of ‘discipline’. Discourse constitutes people as subjects who are authorised (as experts) to discipline others, but the most effective and pervasive forms of power occur where people learn to exercise self-discipline. Foucault studied, in particular, the asylums, prisons, scho...