In the last thirty years Scandinavian political science has been preoccupied with studies of power and democracy. It started with a Norwegian study in the 1970s, followed by a Swedish study in the 1980s. Here at the start of the twenty-first century we are witnessing another round of Swedish and Norwegian studies of power, and a new Danish study is well under way.

In this brief review of the Scandinavian studies of power we will argue that it is possible to detect three basic shifts: (1) a theoretical shift from an underlying concept of sovereign power towards an increasing emphasis on discursive power; (2) a methodological shift from a descending analysis of power towards an ascending analysis of power; and (3) a substantive shift from an emphasis on power towards an increasing concern for democracy; Together, the three shifts indicate a break with the positivist orientation of Scandinavian political science and reflect the growing interest amongst political scientists in various forms of social constructivism.

As the latter shift is predicated on the former, we will first explain what is implied by these two interconnected shifts, and then show why the question of democracy necessarily gains importance as a result of these shifts. We will then discuss the five Scandinavian studies of power and democracy one by one in order to show the gradual reformulation of the research agenda.

2 From sovereign power to discursive power

We borrow the concepts of sovereign and discursive power from the French philosopher Michel Foucault. The concept of sovereign power tends to view power as a causal effect from one atomistic agent to another. Steven Lukes (1974) develops this causal, agency-based notion of power into a three-dimensional view of power. Lukes flirts with the idea of seeing power as the causal effect of societal structures but finally rejects the idea. However, actor-based and structure-based concepts of power are familiar in the sense that they both tend to view power as a causal effect of something that is outside and prior to power.

The concept of discursive power calls this causality into question. It views both actors and structures as constituted in and by various power strategies. Power is nominally defined as the name of a complex strategic situation (Foucault 1990; p. 93). Power is a criss-crossing field of discursive strategies. The antagonistic power strategies construct a discursive system of meaning (concepts, cognitions, rationalities, norms, values, etc.) through acts of exclusion. The discursive system is structured in a specific way and establishes specific forms of subjectivity. Hence structure and agency are an effect of power strategies rather than their pre-given point of departure.

The sovereign and the discursive power concepts can be compared on a number of key points, as is illustrated in figure (see Foucault 1990; pp. 92–8). Sovereign power can be possessed by social and political actors. It is a certain strength or resource that can be distributed amongst different agencies. In this perspective, power is often linked to authority, conceived as the right to deploy one’s strength or resources to achieve something. By contrast, the discursive conception of power, views power as only existing in the concrete exercise of power. However, power is not the willful deployment of an individual capacity, but a set of discursive actions that act upon other actions. Power is a political force that determines the identities of the subjects and shapes their capacities and their room for manoeuvre. Power is not a matter of possessing the money, status or right to act, but rather a matter of creating and changing the premises for social action.

Sovereign power has a centre, a residue from absolute monarchy where power emanated from the body of the sovereign. But the idea of an ultimate centre of power has lived on in the form of the Marxist emphasis on the significance of economic power and of the liberal democratic conception of parliament as the incarnation of popular sovereignty. In contrast to these conceptions, the discursive conception of power claims that power comes from everywhere. It is neither anchored in the economic or the political level of society but is present in all social relations as the political force that constitutes and subverts social identity.

Sovereign power is repressive. It draws in its wake a web of restrictions and repressions, and it constitutes a guarantee that human needs and desires are curbed. Power operates through prohibitions, injunctions and taboos. It delimits boundaries and the law is its key medium. By contrast, the discursive conception of power views power as primarily productive. Power produces social identities but since it always places social identities in specific relations of superordination/subordination, it is also repressive. It is exercised by means of institutionalized technologies that discipline the body and normalize the mind.

As figure shows, the two concepts of power have different methodological implications (Foucault 1986). The sovereign analysis of power focuses on the legitimacy of power, and the legal basis of power is viewed as an important source of legitimacy. By asking the question ‘Who has power?’ it is possible to reveal different forms of illegitimate power. Individuals with various amounts of resources are the originators of power. The analysis of who does and who does not have legitimate power, begins at the top of the central political institutions. Thereafter, the analysis moves in a descending direction in an attempt to reveal illegitimate forms of power in the civil service, interest organizations, the private sector, etc.

A discursive analysis of power does not focus on the legitimate exercise of power. It focuses on the political strategies that shape the identity of the actors, their capacity to act and their social context. Political strategies does not refer to calculated acts controlled by rational considerations as to means and ends but to antagonistic struggles about the construction of meaning. In the analysis of discursive power strategies, the key issue is how power is exercised. It asks the questions: what typologies, metaphors, narratives, etc. are employed in the discursive construction of the social and political actors and their political strategies? How are the different strategies institutionalized in and through local power technologies? What kinds of hierarchies of knowledge are sustained by these power technologies, and how do they affect the actors’ ability to think, speak and act in certain ways?

The subject is not the point of departure but rather the fulcrum of discursive power strategies. Discursive power strategies are an expression of ‘intentional non-subjectivity’. On the one hand, all discursive strategies bear the intention of creating a specific discursive order – though this order exists neither as a blueprint nor as a utopia. On the other hand, discursive strategies cannot be invented, controlled or mastered by any single subjectivity, be it a political party, a company or a state. On the contrary, subjectivity is something that is created as part of a discursive strategy.

Since discursive power strategies work through local institutional apparatuses that link power and knowledge it is obvious that power analyses cannot start from the top and thereafter proceed downwards. Power is organized and exercised in locally based institutions and hence the analysis of power must take its point of departure in an ascending direction (Foucault 1986, p. 99). He argues that global power strategies are conditioned by local strategies that are the building blocks of the global strategies. However, at the same time, local power strategies are conditioned by global strategies that strengthen them and provide them with direction (Foucault 1990, pp. 99–100).

Foucault was not very fond of liberal democracy and did not draw any consequences from his power analytics to the study of power in liberal democracies. Our claim is that the concept of sovereign power and the related top-down perspective on power tends to privilege the study of power and marginalize the study of democracy. For if power is viewed as the sovereign’s right to repress other actors by legal means, then the study of democracy can be readily reduced to the question of the distribution of legally exercised power. Democracy prevails as long as the holder of sovereign power, who embodies the will of the free and equal citizens, respects the law and is not constrained by other actors. Whether these fundamental democratic conditions are fulfilled can be assessed through judicial review and a study of the distribution of power.

Now, as soon as we start to see power as a discursive power to be studied in a bottom-up perspective, democracy comes to the fore. The study of democracy can no longer be reduced to a study of the distribution of power, because power is now a question of the strategic interaction of a multitude of contingently constructed agencies. When power ceases to be a question of the legitimate exercise of sovereign power, democracy becomes a question of the regulation of the antagonistic power struggles through the construction of rules for access, procedures for negotiation, norms about mutual tolerance, and values such as freedom and equality. In other words, when power is seen as something that comes from everywhere, democracy becomes a mode of governance to be studied in its own right. Furthermore, it is to be evaluated in terms of its ability to regulate the multiple power strategies in a way which makes power an empty space. It must keep the locus of power open for a multitude of voices and ideas and prevent it from being appropriated by a totalizing political force.

3 Five Scandinavian studies of power and democracy

In 1972 the Norwegian government asked a group of researchers to examine how power was distributed. Two studies were made in Sweden in 1985 and again in 1997. The two final studies – a Danish and another Norwegian were started in 1998 – are not yet completed. In what follows we briefly present the five studies while emphasizing their changing focus.

The first Norwegian study of power and democracy

The Norwegian government inspires the need for at study of power and democracy with reference to a possible discrepancy between the formal and the actual distribution of power in the country. The government is especially interested in studies which examine the illegitimate use of power by strong financial groups, huge organizations, the administrative apparatus, mass media and international corporations. In other words, the government wants to envisage possible threats against the sovereignty of the democratically appointed institutions. Accordingly, it shows little interest in the question of democracy understood as the process and institutions through which sovereign power is constituted and legitimated. Neither does the government require an ascending power analysis. It wants a macro- and actor-oriented study that places the focus of attention on the relationship between the dominating power holders in Norway. The aim is to gather information about the nature of the ‘conditions that give individual actors, groups or institutions power’ (NSPD 1982, p. 1) and ‘clarification about the relations of power between the most important groups’ (1982, p. 1).

The appointed group of researchers adopted the perspective of the government. Power was defined along the lines of a sovereign concept of power as: ‘decision making power, A’s power over B, the ability to act collectively, and power based on technological dependencies’ (1982, p. 4). A descending power analysis focused on the power exercised by the most important groups in society defined as public administrators, interest organizations, firms and mass media.

The main result of the first Norwegian study of power and democracy is that the political system is dis-empowered as a result of: (1) an increasing institutionalization of the relationship between public and private actors that has lead to a sectorialization and segmentation of the public sector; (2) the increasing political influence of organized groups and private finance; and (3) a general decline in the ability of the political system to govern. The researchers conclude that current threats to sovereign power are not only caused by an increase in the power of illegitimate power holders, but also by institutional limits to governance as such (1982, p. 4). What is really at stake is the ability to perform societal governance in Norway.

The first Swedish study of power and democracy

The first Swedish study resembles the first Norwegian study. The main concern of the Swedish government is that economically powerful actors gain an unacceptable amount of influence in political life. Accordingly, the question of democracy is again reduced to an investigation of the threats facing the democratic sovereignty of elected politicians. What is called for is a descending power analysis focusing on: ‘how power resources and influence are distributed in different parts of Swedish society such as business life, the public sector, organisations and mass media’ (SSPD 1990, p. 11).

In contrast to the Norwegian study, the appointed research group is critical of the ‘society-oriented’ and élitist perspectives on power and democracy presented by the government. The concept of democracy sustains the system because of its realistic approach, and it ignores a more ‘individual-oriented’ concept of democracy. Accordingly, the government fails to recognize that a central aspect of the democratization process in Sweden has been the attempt to establish a balance between the two perceptions of democracy. The researchers are in favour of a more complex, normative and change-oriented concept of democracy.

The researchers argue that power traditionally has been about ‘influencing how?’ and ‘influencing who?’ (1990, p. 44). They stress the need for a broader concept of power that makes room for the analysis of mind-controlling power and discursive power (1990, p. 44).

As a consequence of the reformulation of the concepts of power and democracy the question of democracy now becomes a central part of the analysis. A study of sovereign power is supplemented with an analysis of the ability of individual citizens to obtain influence over their own lives and society as a whole. This intention to focus on citizenship and discursive power in everyday life calls for the introduction of elements of an ascending analysis of power as a supplement to the traditional – and still dominant – descending analysis of power.

However, there seems to be some difference between intention and result. Most of the actual research is based on a sovereign concept of power, a society-oriented concept of democracy and a descending power analysis. Therborn, one of its critics, claims that the study overlooks the conditions of possibility of power which is connected to the modern conception of power – power defined as the ability to control and govern society (Therborn 1992, pp. 71, 75). He calls for a historical analysis of the influence of institutions on the nature of power, its consequences, and a process-oriented analysis of concrete policy processes.

The second Swedish study of power and democracy

In 1997 the Swedish government initiated a second study of power and democracy (SG 1997). This study is different from the former studies in several ways. First, it is less ambitious in scope. The main objective is not to produce new research but to present an overview of existing research on power and democracy in Sweden. Second, the study is not led by researchers but by a parliamentary committee. Finally, the government shows more interest in the study of democracy than in the study of power. Hence, the parliamentary committee finds it ‘remarkable that the term power is not mentioned once in the government directive’ (SSPD 2000, p. 35).

The parliamentary committee was asked to shed light on the negative and positive effects of some major societal changes on the functioning of Swedish democracy: the internationalization of the economy; Sweden’s membership of the EU; the spread of information technology; changes in the public sector; and changes in the way citizens engage in the political decision-making process through their participation in elections, the mass media and social movements. In contrast to the first Swedish study these changes are not seen as threats which should be remedied in order to reinforce parliamentary democracy. They are seen as conditions of possibility which Swedish democracy must adjust to. Hence, the former call for a re-authorization of the parliamentary chain of governance is replaced by a change-oriented wish to develop new democratic institutions which are capable of functioning in a world dominated by the New Economy and post modern forms of citizenship. Because of this change of perspective, an ascending power analysis is introduced which focuses on the conditions for developing new democratic institutions.

The parliamentary committee concluded the study in the year 2000 with a final report entitled: ‘The persistent democracy’ (2000). Unlike the first Swedish study of power and democracy, the intentions of focusing on democracy have been realized in practice. Actually, the large majority of publications on which the report is based are concerned with the functioning of democratic institutions and the relationship between citizens and the political system (Amna 1999a, 1999b). Furthermore, a descending power analysis is supplemented by an ascending one that deals with the functioning of democratic institutions in relation to the everyday lives of the Swedish people (Montin 1998; Amna 1999b; Amna 1999c). The central role this analytical approach has obtained, is not only caused by the change-oriented perspective on Swedish democracy. It is also caused by the participatory and deliberative conception of democracy to which the parliamentary committee attests (SSPD 2000, p. 23). In many ways this second Swedish study of power and democracy could be seen as some kind of follow-up on the first Swedish study, dealing with the questions that the first study failed to tackle. This follow-up moves the Nordic studies of power and democracy closer to a discursive and ascending power analysis. It does not take the sovereign power of the Swedish Parliament for granted. The committee investigates the democratic distribution of influence between a multitude of levels and actors. Democracy is conceived and studied in terms of a ‘flexible multi-level governance’ and a ‘negotiated sovereignty’ (2000, p. 112). With regard to Therborn’s critique of the first study this second study gives more room for historical analyses. However, it does not study the discursive conditions of possibilities for the exercise of power and does not undertake any process-oriented studies of policy processes.

The second Norwegian study of power and democracy

In 1998 the Norwegian Parliament appointed a research group to do a second study of power and democracy within a period of five years. The researchers were asked to explore ‘the conditions for the Norwegian democracy and changes within them’ (NP 1997). Democracy was, in this context, defined as representative democracy: ‘The study should focus on how representative democracy and its conditions of being are being challenged and affected’ (NP 1997). The challenges to representative democracy are taken to be internationalization, new technology, public opinion making, environmental problems, multi-culturalism, knowledge society, decentralization and deregulation. These challenges, the politicians argue, make necessary a new study of power and democracy.

The research group agrees. They claim that the nation state is challenged both from the outside and from the inside. The former are caused by increased internationalization and globalization which delimit the realm of state action while the latter is caused by the growing autonomy of subsystems and the emergence of new and more individualistic values in the population. In order to shed light on the consequences of these changes the research group chooses to investigate the relationships and balances of power within and between four societal spheres. Hence, the overall focus is on the role of ‘the state, civil society, the market and community arenas in a nation state under pressure’ (Østerud et al 1999, p. 143).

In many ways the second Norwegian study maintains a sovereign perspective on power. Regardless of the societal changes, representative democracy is maintained as the key reference point for considerations about the future of democracy. We find few of the change-oriented considerations, which dominated the second Swedish study, regarding an adjustment of the concept and institutions of liberal democracy to current societal changes. The study of democracy is more or less reduced to a question of finding new ways to legitimize state power: ‘The conditions under which power and exercise of power is accepted as legitimate establishes much of the connection between power and democracy in this study’ (1999, p. 140).

However, despite the sovereign approach to the study of democracy the research group introduces some elements of a discursive concept of power. Power is perceived as being everywhere, and although an actor-oriented approach dominates, it is emphasized that power also operates through the shaping of symbols and through the construction of meaning and identities. Hence, in an otherwise critical article entitled ‘A new study of power and democracy in old tracks?’, a commentator complements the research group for its interest in ‘identity formation, the power of definitions and knowledge and the hegemonic power connected to the influencing of ideas and images of how things are’ (Gullestad 1999, p. 18). Finally, power is not only regarded as being repressive, but as a platform for empowering ordinary citizens in a democratic society.

This kind of research calls for an ascending analysis. Accordingly, the research group stresses the need for historical studies, process-oriented studies of actual policy making, and studies of relations of power in everyday life.

To sum up, even though this second Norwegian study is still dominated by a sovereign concept of power and a descending power analysis it introduces elements of a discursive concept of power and an ascending power analysis. Furthermore, it is much more oriented towards the study of democracy than the first Norwegian study.

The Danish study of power and democracy

The second Norwegian study was initiated at the same time as a Danish study of power and democracy. Also this study was set to last five years. The study is strongly influenced by the first Norwegian and Swedish studies which, according to the parliamentary committee, have ‘contributed to destroying the myth that the parliament is the only powerful body within the political system’ (DPC 1997, p. 5). Political decision making has been displaced upwards to the EU, downwards to local institutions and user boards, and outwards to increasingly independent semi-public institutions. However, this development has not affected the expectations of the politicians that they are sovereign decision makers on almost any issue concerning the governing of society. The recognition of this problem does not lead the parliamentary committee to ask for a study that aims to reinstate the parliamentary chain of government. Instead, they ask for a redefinition of democratic institutions and the role of politicians in processes of societal governance (1997, p. 11).

This approach calls for a combination of a descending and an ascending power analysis. The former is used to focus on formal and informal distribution of power and influence amongst the actors in the established political system. The latter serves to investigate: the influence which public institutions, organizations, associations and movements grant their users and members; the distribution of influence in actual decision-making processes; and the ability of ordinary citizens to influence their everyday lives.

The appointed group of researchers followed the propositions made by the parliamentary committee. They argue that the analysis of ‘governance in Denmark must deal with both individuals, political institutions and the underlying transformations in society’ (DSPD 1998, p. 1). They opt for a descending power analysis of the functioning of political institutions and the abilities of political élites to govern. At the same time, they recommend an ascending analysis of power in order to deal with the influence of ordinary citizens and the study of concrete policy processes.

However, the researchers go even further in the direction of a discursive power analysis. They explicitly urge Danish researchers to study: (1) the construction of social problems and the formation of structures of meaning; (2) the role of symbols and path dependencies in policy formation processes; and (3) the genealogy of governing ambitions and the historical development of the notions of democracy and power (1998, pp. 9–10, 13). A number of such research proposals are currently being carried out (DSPD 1999). Accordingly, the Danish study, although it is dominated by a sovereign concept of power and democracy, leaves considerable room for discursive power analyses.

Concluding remarks

To sum up, there seems to be three general trends in the Scandinavian studies of power and democracy. First, the focus gradually changes from a focus on power to a focus on democracy. Second, a discursive concept of power is slowly but steadily making its way into the studies. Third, we witness an increased interest in ascending analyses of power. How far the development towards a discursive concept of power and an ascending power analysis will go in the future, depends on the insights that this approach is able to produce. While the outcome of the second Swedish study gives reasons for optimism, the discursive approach to the study of power and democracy stands its real trial in the ongoing Norwegian and Danish studies. When they are concluded in 2003 we will be much wiser as to the fruitfulness of the new approach in the study of power and democracy.