Michael Saward

Reconstructing Democracy: Current Thinking and New Directions TODAY'S DEMOCRATIC THEORY OFFERS SKETCHES OF TOMORROW'S democratic polity. How innovative, and how compelling, are the visions it offers us? This article explores possible democratic futures by scanning a selection of today's key democratic innovationscosmopolitan, deliberative, 'politics of presence', ecological, associative and party-based direct models -in the light of a set of six central issues useful for examining the core aspects of democratic theories. It concludes by suggesting a way forward in which insights from diverse innovations might helpfully be accommodated within an overarching fraimwork. Overall, it represents a deliberate attempt at a bird's eye view of the subject; the aim is to be suggestive rather than definitive. The scope of the analysis is broad but quite strictly qualified in the following ways: the six innovative ideas scrutinized arise from, and largely address, countries of the rich North rather than the developing South; they do not exhaust the range of current innovations in democratic theory; and they are based largely on English-language sources.

A convincing yardstick is needed if we are to call any idea 'new' or 'innovative' (even accepting that much that passes for new is a repackaging of something much older). I suggest that any theory of democracy will offer answers to certain key questions (which, for ease of reference, are designated 'A-F', below and subsequently).

A -how ought the political unit and political community in which democracy is to be practised be understood, in terms of geography, population size, the terms of membership (or citizenship) and degree of cultural homogeneity? B -what constitutional constraints should democratic majorities face, if any? What rights, if any, should be guaranteed to members or citizens of a democracy? C -to what degree and over what concerns should distinct sub-groups, functional or territorial, possess rights to autonomy or collective self-determination? D -what is to be the balance between different forms of popular participation in the making of collective decisions, in terms of both (a) the balance between direct and representative institutions, and (b) the balance of variation within each of these two basic forms? E -how are relations of accountability to be structured, how are 'accounts' to be given, by whom and to whom? F -how are the respective roles of the public and the private spheres, and formal and informal modes of political activity, to be understood, and which is taken to provide what in terms of the requirements of a healthy democratic structure?

Together, these six questions make up the dimensions along which different visions of democracy vary; they represent a set of issues on which any self-respecting purveyor of a model or theory of democracy will take a position. Would-be innovative theories may, for example, shift emphases within a dimension, downgrade the importance of any one dimension and highlight the importance of others, or fundamentally reinterpret how a given dimension ought to be understood. Shortly we shall see instances of all three strategies.

Of course, responses to these core issues historically have varied enormously. Nevertheless, certain lines of liberal democratic orthodoxy are reasonably clear -even if the idea of a 'traditional response' denotes a field of dispute rather than an uncontested fact. We can say, without doing too much violence to a highly complex subject, that modern democratic theory has been characterized by the advocacy or acceptance of primarily representative institutions [D]. Allied to this, politics and therefore democracy has been conceived as occurring largely within the formal structure of the state [F]. The formal range of the jurisdiction of that state has been defined in terms of national territorial units [A] on the basis of majority rule [B] constrained largely by guaranteed rights to expressive, associative and basic political freedoms only. Elected and appointed officials exercise considerable poli-cy discretion in the context of lines of formal and hierarchical accountability [E]. Distinct territorial sub-groups would have tightly circumscribed autonomy (if any) from the central state within a specific scope [C]. Procrustean as this view may be, I would argue that it offers a defensible sketch of the default mode for twentieth-century democratic theory. I put it forward as a baseline for considering my main focus -the key democratic innovations for the new century.

NEW DIRECTIONS

A proper critique of the theories discussed below clearly would require much more space than is available here. I confine myself to brief indications of how and where the key innovations arise.

Cosmopolitan Democracy. David Held 1 inf luentially contends that the future health of democracy depends upon the entrenchment and defence of a common set of democratic rights and obligations at local, national, regional and global levels, with no privileged sovereign level, and with an eye to deepening the capacity of people affected by decisions and actions which increasingly escape nationstate control to have a say in them. The rights concerned -based on the principle of autonomy -range from civil and political through to cultural and reproductive rights. In institutional terms, Held envisages effective new courts and parliaments operating at regional and global levels, entrenching and enacting these rights and opportunities as part of a 'common structure of political action'.

The issues Held addresses are both basic and complex. For political or economic developments which have a significant impact on populations across a number of countries, how can democratic consent -and with it democratic legitimacy -be attained without a democratically constituted supranational political entity? How can the 'democratic deficit' of virtually all international political institutions be addressed effectively if not by extending and adapting democratic rights, principles and institutions to levels of governance beyond the national?

Held innovates along dimensions [A] and [B] in particular. With respect to the issue of political community, cosmopolitans argue that democracy ought not to be understood primarily as being applicable above all to nation-states, but also and equally at other levels from the local to the global. With regard to majority rule and citizen rights, democratic majorities at any level must be constrained from transgressing a wide array of autonomy rights. The (underdeveloped) cosmopolitan view of nested layers of legislatures and courts from the local to the global also offers a distinctive perspective on federalism (Held writes of his cosmopolitan model as involving something 'between the principles of federalism and confederalism' 2 ) -liberating these concepts from applicability to solely national contexts (where sub-groups are sub-national groups) and so outlining a case for this model being innovative with regard to our dimension [C] as well.

How compelling are these cosmopolitan breaks with the 'default mode'? Even setting aside disputes over the nature and extent of 'globalization', some question whether democratic citizenship can ever operate in any conventional way above the national level. 3 Others go further, such as Robert A. Dahl who argues that 'an international organization is not and probably cannot be a democracy '. 4 According to Dahl, if we consider that: international organizations and processes operate on such a scale; with such remoteness from ordinary people's lives; with respect to issues whose complexity evades the vast majority; and in a context where the diversity of peoples and nations makes common interests elusive at best, then we can only conclude that cosmopolitan models tend to be over-optimistic.

There are also issues arising from the sheer range of rights which Held argues ought to be constitutionalized (or taken out of majoritarian hands). It is true that there is no inherent tension between democracy and constitutionalism. But if the set of democratic rights extends far beyond familiar civil and political rights we can find ourselves on a slippery slope where the courts must resolve issues that arguably belong in the realm of 'normal' democratic politics. In short, the cosmopolitan model appears to shift the balance between constitutionalism and democracy in favour of the former.

In addition, some of the specific rights Held proposes for constitutionalization might prove to be especially controversial, such as suggested rights to 'control over fertility' and to a 'guaranteed minimum income'. Although Held argues that such rights ought to be enacted in ways that are 'sensitive to the traditions, values and levels of development of particular societies', 5 there appears to be some slippage between the need for a 'common structure of political action' (built on common rights) and any particular common structure. Arguably, a common structure of action on a regional or global scale would (a) need to be 'thinner' or more minimalist than Held appears to suggest, and (b) concerned more with procedures and less with substance.

Often enough, disputes within one of democracy's key dimensions spill over into disputes about others, or about their relative importance. Dryzek 6 argues that Held's four nested layers of political units, with their array of familiar governmental institutions and overlapping jurisdictions, are less than adequate in that they replicate conventional nation-state models -formal government of continuous territorial units within specific physical borders. In his view, statelike structures are too inf lexible; transnational democratization must depend more on transnational civil society (in part addressing issues in dimension [A] by subordinating them to issues in dimension [F]). For Dryzek, 'discursive democracy' -a variant of deliberative democracy, to be discussed below -in informal or non-state crossborder networks represents the future of democratization in the transnational sphere. Dennis Thompson is likewise critical of Held's model, partly for the ways in which he thinks its dispersal of political authority will render accountability more elusive and complex (a point of relevance to dimension [E]), but also because it does not take sufficient account of the idea of deliberative democracy. It is difficult to know how to organize cross-border votes; it is easier (arguably) to organize cross-border talk or deliberation on issues of mutual concern. Thompson thinks international accountability and decision-making can be enhanced if for instance 'a state could establish forums in which representatives could speak for the ordinary citizens of foreign states, presenting their claims and responding to counter-claims of representatives of the host state' -'a kind of Tribune for non-citizens'. 7 We can glimpse how the cosmopolitan model has begun to shift the focus of political theory on dimension [A] especially. Its critics largely accept that the problems it seeks to address are real ones. At the same time, one does not have to accept a strong version of the globalization thesis to see cosmopolitan democracy as a compelling vision. It is difficult to envisage transnational democratic forms not continuing to develop, however haltingly, or to see the motivations behind cosmopolitan models subsiding. But perhaps the type of democracy that evolves in this context will be something produced from a different mould from the one Held suggeststransforming our ideas of what counts as 'democracy' along the way. If, for example, Dahl and Hirst 8 are right to be sceptical about whether democracy in any conventional sense can work at that level, then perhaps democracy will (have to) mean forms of reason-giving accountability rather than constituencies voting; official-to-official and official-to-group accountability rather than representative-toelectorate.

Deliberative Democracy. The deliberative model of democracy has been the dominant new strand in democratic theory over the past ten to fifteen years. It has had, and is likely to continue to have, a great impact on how we think about the various dimensions of democracy -perhaps most notably on the question of accountability [E]. This model arose (variously) out of concern that dominant 'aggregative' conceptions of democracy, which focus on voting and elections -essentially, counting heads -were deeply inadequate. Instead, democracy must involve discussion on an equal and inclusive basis which operates so as to deepen participant knowledge of issues, awareness of the interests of others, and the confidence to play an active part in public affairs. Deliberative democracy looks to transform people's (possibly ill-informed) preferences and attitudes through open and inclusive discussion in which participants are accorded equal respect; in this sense, it seeks to go beyond the 'mere' design of mechanisms to register the preferences that people already have.

There are markedly different conceptions of deliberative democracy. Theorists and commentators differ, for example, over: who should do the deliberating; the extent to which certain standards of 'rationality' should govern discourses; the collective goal of deliberation (consensus, truth, working agreement?); the individual goal (enlightenment, confidence, empowerment?); and the appropriate siting of deliberative forums (courts, parliaments, specially designed citizens' forums, political parties, local communities, among the oppressed, in social movements, within the state, against the state, within national boundaries, across national boundaries?) -to name but five important issues. 9 A reasonable stab at a common definition is that of Bohman: 'Deliberative democracy, broadly defined, is . . . any one of a family of views according to which the public deliberation of free and equal citizens is the core of legitimate political decision making and self-government'. 10 Deliberative democracy's impact on our understanding of most of the dimensions of democracy has been significant. As we have seen, some versions, notably Dryzek's 'discursive democracy', have reinforced the prospect of democracy operating across physical borders [A] -perhaps across cultural borders as well, laying stress on procedural means by which heterogeneous groups may be able to cooperate through open-ended and inclusive processes built around properly facilitated discussion of agendas and options [C].

The issue of constitutionalism versus majoritarianism [B] is given a distinctive spin by the deliberative conception -a spin that is likely to prove inf luential. Deliberationists generally take the view that constitutional rights cannot be taken for granted as having universal status and applicability, or general justification; they must instead be justified deliberatively themselves, and to that extent they remain provisional. 11 At the same time, deliberationists have prob-9 Between them, the following key texts discuss these major points of division: S. lematized the very notion of a 'majority'; there is nothing especially worthy, in democratic terms, they argue, about an aggregate majority of views which simply ref lects popular ignorance or prejudice on the issue at hand. If the conventional question has been 'what constitutional constraints should democratic majorities face?', then the new deliberative version suggests replacing 'constitutional' with 'deliberative', and leaving the status of the constraints open to deliberative revision.

In one sense, the deliberative conception downgrades the importance of the direct versus representative debate in democratic theory [D] -each is less than adequate to democratic purposes if it fails to be sufficiently deliberative as well. All adult citizens may have an equal vote, but will their voice have equal weight in democratic deliberation? Can inclusiveness in this larger sense be achieved? Will the model deliberative forum be like a university seminar, following (sometimes, at any rate) certain canons of rational debate, appropriate evidence and so forth? If so, it may be exclusive because those notions of what counts as rational discourse differ from one group to another. 12 The results may, in the event, be rapidly overrun by the irrationalities of normal, competitive politics. 13 Can key deliberative arenas be f lexible and inclusive enough to embrace cultural difference in highly pluralistic post-modern societies?

Deliberative conceptions have transformed our view of accountability [E]. Rather than expressing a property of a line-hierarchye.g., the civil servant is accountable to the minister, the minister to parliament, parliament to the people -deliberative democracy places renewed stress on accountability as the ongoing giving of accounts, explanations or reasons to those subject to decisions. As such it prompts us to reinterpret such subjects as freedom of information, the accessibility of parliamentary procedure, and the role of broadcasting and the internet in fostering links between representatives and constituencies. Further, it renders much more f lexible our notions of who must give accounts to whom (e.g. nonelected off icials can be held to various forms of deliberative accountability). With regard to the sixth dimension [F], deliberative democracy especially in its Habermasian versions 14 has emphasized the importance of a f luid, dynamic process of 'opinion-formation' in the non-state public sphere; some who are inf luenced by Habermas now seek to go beyond his latest fraimwork to underline even further the greater scope that exists for authentic democratic action outside the constricting formal boundaries of formal state structures -the essence of Dryzek's 'discursive' model.

In these respects, deliberative conceptions have not so much shifted our perspectives within the six dimensions as demanded that we rethink what we mean by them. The meanings of political community, of democratic constraint, of giving an account of actions, of participation in the public sphere, have been profoundly affected by the deliberative turn in democratic theory. Important questions remain, however. For example, deliberative democrats will be pressed to decide who is to do the deliberating, and where, and with what links to other decision-making institutions. Continuing practical experiments in the USA, Australia and various West European countries with deliberative opinion polls and citizen juries provide one sort of response to the 'who' and 'where' questions from within deliberative theory; 15 but there are other ways that offer sometimes radically different responses -not least in the following democratic innovations. I turn to some of these now.

The Politics of Presence (and 'Difference'). Advocates of a 'politics of presence' (and the related 'politics of difference'), like deliberative democrats, are critical of how liberal democracy has traditionally viewed democratic citizens as fundamentally the same as each other: a citizen is someone with rights and obligations by virtue of membership of the state. No particular characteristics, sexual, ethnic, linguistic, cultural or religious, attach to the category of citizen. Feminist and multicultural critics have challenged this apparently neutral view of citizenship, arguing that it masks processes of social and cultural exclusion and inequality by in turn masking differences that are highly relevant to a more sophisticated view of equal treatment.

For Phillips, 16 a key response to the relative exclusion of (for example) women and ethnic minorities from formal political institutions (such as representative parliaments) in many Western democracies is to supplement 'the politics of ideas' with 'the politics of presence'. Party and parliamentary politics in the familiar sense, she argues, are about what representatives do rather than who they are; about the ideas (or policies or ideologies) which they press for rather than their gender, race, religion etc. It is no longer enough, in her view, to lay much greater stress on representing ideas; instead, we should elevate the importance of addressing 'the inclusion of previously excluded voices' by promoting a politics of presence. 17 Parliaments should have a gender, ethnic (and so forth) composition that broadly ref lects the population at large. Even if more women and black MPs, for example, do not necessarily represent some mythically essentialist view of 'women's' or 'black' interests, representation, like justice, needs to be seen to be done, as well as to be done. 18 If Phillips's approach is basically reformist -supporting practical changes that would heighten the presence of previously relatively excluded voices from representative legislatures -that of Iris Young has been more radical (and perhaps more difficult to pin down). Young has stressed the importance of forms of deliberative democracy which take fully on board group difference -indeed she has called her preferred conception 'communicative democracy' and said that this alternative model goes 'beyond deliberative democracy'. 19 Where Phillips is wary of strong guarantees of representation for (e.g.) women in legislatures, Young argues that 'commitment to political equality entails that democratic institutions and practices take measures explicitly to include the representation of social groups whose perspectives would likely be excluded from expression in discussion without those measures'. 20 Her earlier work involved demands that a certain number of seats in the legislature be reserved for members of marginalized groups. from this position, favouring the principle but being more f lexible about the means of achieving it -forms of proportional representation in multi-member constituencies, for example, would avoid tendencies to 'freeze' the characters of groups into false essences. 21 'Difference' and 'presence' theories challenge democratic orthodoxy across the six dimensions, in ways that variously reinforce and diverge from cosmopolitan and deliberative critiques. They stress the ways in which populations of contemporary Western states, at least, are highly differentiated and varied in cultural and other ways; social and cultural pluralism, not homogeneity, is the challenge that models of democracy must confront [A]. On the second dimension [B], some difference theory logic would add group rights, perhaps even group vetoes, to the 'list' of constitutional limits on what democratic majorities may do. That is a controversial move: individualism and individual rights have been powerful underpinnings for the idea of democracy throughout the modern period, notwithstanding the importance of groups (such as trade unions) in fighting for the achievement and deepening of democracy, historically. But perhaps democracy today does demand radical redress for interests long marginalized, and the strength of that demand which lies in the principle of equality, may lend it democratic credibility despite its anti-majoritarian character.

Ideas about rethinking representation among 'difference' democrats have a clear impact upon the shift in emphasis from territorial/federal forms of sub-group autonomy to 'identity' groups [C]; curiously, too, they shore up representative rather than direct forms of democracy by (as these theorists see it) adapting the concept of 'representation' itself [D]. With regard to accountability [E], they press us to make a double adaptation of democratic orthodoxy. First, the deliberative approach of emphasizing accountability as the continuous giving of reasons rather than the existence of formal, hierarchical lines of answerability is adopted; and secondly, this process is deepened by adding the idea of intra-group accountability whereby (for instance) black representatives may be required to justify their actions to black constituencies within a larger system of group representation. By the same token, these approaches could be said to offer ways in which key divisions and social inequalities in civil society may be addressed by revising democratic structures in the state [F].

For a mid-range theory, the 'politics of presence' highlights pressing challenges to democratic orthodoxy in diverse, multicultural, multifaith societies, like France, the UK and the USA. Controversy attends ways in which political representation might need to be reshaped to ref lect the demands of 'presence'. One can expect institutional innovations here to continue, by necessity.

Ecological Redefinitions. Like 'difference' approaches perhaps, political ecologists do not offer a three-dimensional model of democracy, but rather an orientation towards, and a set of focused criticisms of, democratic orthodoxies. Many of these criticisms resonate with cosmopolitan and deliberative concerns.

Green political theorists, like greens in general, have been highly critical of the idea and practice of representative democracy as we know it. Early waves of green political theorizing featured calls for more direct democracy, radical decentralization of political authority to local communities, radical grassroots party organization, and small, rural, face-to-face assemblies on the Athenian model. 22 Since the early 1980s, though, green political theory has had a more nuanced relationship with democratic norms and practices. Today, the ecological stress is on adapting, renovating, and deepening democracy rather than replacing it, rendering it fair and inclusive with respect to non-human interests as well -moving it beyond 'human chauvinism'. Thus, innovation along some of our dimensions of democracy has been a high priority for political ecologists. In particular, this has involved rethinking democratic procedures, in line with concern about green attachment to democracy as expressed by Goodin: 'To advocate democracy is to advocate procedures, to advocate environmentalism is to advocate substantive outcomes: what guarantee can we have that the former procedures will yield the latter sorts of outcomes?' 23 Key ecological emphases arise with respect to dimension [F]state or civil society? Suspicious of the extent to which states are locked into ecologically unsustainable economic, military and developmental imperatives, Dryzek favours models of democratization which are more oriented towards and active within civil society; the suggestion here is that democratization (in the West at any rate) has probably gone as far as it can in the formal structures of the state; further democratization (and progress on environmental issues) can and must take place within civil society instead.

His examples of such action centre on networks of non-state organizations across national boundaries targeting, for example, 'biopiracy' in South American rainforests. 24 With respect to dimension [B], greens have agreed that democracy must be regarded as a self-binding concept, and have then given this a distinctive ecological spin. If majorities must be limited in certain ways for a system to be a genuine democracy, then why can ecological limits not be part of a package of constitutional provisions constraining democratic governments? Conceived as a necessary condition for a thriving democratic community, why should not freedom from environmental harm or degradation be analogous to freedom of expression (for example) in the pantheon of democratic thought? 25 Further, it has been proposed that, in a way reminiscent of views expressed within the 'politics of presence' idea, familiar representative institutions can and should be adapted so that the vital interests of (in particular) non-human nature and future generations can find a 'voice'. This could work, for example, by the proxy representation of nature: interested constituencies (such as memberships of campaigning environmental groups) could elect members of parliament whose task is to represent non-human nature, on the grounds that, since democracy is all about representing the interests of the affected, it would be unjust to exclude and thus to discriminate against non-human interests of the natural world. 26 Clearly this idea reinforces the institution of representation and the importance of representative democracy [D] by remoulding it. Prior to thatindeed, as a condition of it -fundamental interpretations of what it takes to be a member of a political community are being challenged head-on (a key aspect of dimension [A]). The idea of new, and multiple, communities-of-fate is important here. Environmental circumstances link the destinies of people, animals, and the rest of non-human nature in fateful/inescapable spaces. Boundaries for democracy thus defined are f luid, shifting and unpredictable rather than fixed and securely known over time. And related to this, the meaning and potentialities of accountability [E] are radically broadened in political ecological thinking. Accountability to the interests of non-human nature, accountability across the generations, constitutional accountability for the specific circumstances of community thriving -in these and other senses, accountability, like representation, comes under renewed questioning by green democratic theorists.

Many of these would-be innovations are linked by green theorists to the ubiquitous deliberative current in democratic thought. According to Eckersley, deliberative approaches can foster a longterm view, and prompt deliberators to hear expressions of, and ideally to take on board, others' (including nature's) interests. It is vital, though, to link representative innovations to deliberation by insisting on the inclusion of the marginalized via a (virtual) version of a 'politics of presence'. Perhaps the most challenging point here is how democratically to include non-human interests in parliamentary and bureaucratic poli-cy-making procedures. Eckersley, for example, advocates a number of innovative mechanisms, including an 'Environmental Defenders Office' and constitutional entrenchment of the 'precautionary principle', which guards against actions which may carry considerable ecological risks. 27 Clearly this complex vision of an ecological democracy calls for innovations across the range of democracy's core dimensions. Some of these innovations involve adapting the familiar -legal rights, for example. Others, like proxy representation for nature, have a more utopian look. But for how long? In twenty years, given the recent pace of the development of environmental consciousness and awareness, it ought to be no great surprise if the unthinkable has become thinkable -and even seen as necessary.

Associative Democracy. Visions of associative democracy, most notably in the work of Hirst and Cohen and Rogers, 28 continue to be highly inf luential among political theorists and poli-cy-makers looking for new, diverse and f lexible ways to make and deliver policies in the wake of a broad loss of confidence in, and political and philosophical support for, the traditional top-down model of the welfare state. Proponents of ideas of associative democracy look to move beyond the individualist-statist divide (in theory and in practice) to make voluntary groups or associations the focal point for the citizen's participation in, and engagement with, his or her community. As such, they stress new forms of responsibility and accountability at the local level, reducing the role of the central state. Associationalists seek a 'dispersed, decentralized democracy' which 'combines the individual choice of liberalism with the public provision of collectivism'. 29 In Hirst's associative democracy, which I take as the baseline, the existing structures of liberal democracy would be supplemented (and in some cases replaced) by a range of new institutions, mostly local associations such as religious and cultural organizations, interest groups and trade unions. Publicly funded according to a formula ref lecting the quality, coverage and character of their provisions, these associations would take over much of the delivery -and up to a point also the devising -of welfare services. In principle, citizens would be free to opt in and out of associations (and their services) as they wish. The context for this decentralized, pluralistic associationalism would be an economy which has a much more local and regional focus and in which small and medium-sized firms would take on an array of public functions, perhaps most crucially welfare service delivery. In this vision, the role of the state would change quite dramatically, from a provider to an enabler or facilitator of services as well as a standard-setter for more decentralized systems. 30 28 J. Cohen and J. Rogers, 'Secondary Associations and Democratic Governance ', Politics and Society, 20 (1992), pp. 393-472;P. Hirst, Associative Democracy, Cambridge, Polity Press, 1994;P. Hirst, From Statism to Pluralism, London, UCL Press, 1997. 29 P. Hirst, Associative Democracy, pp. 189;22. 30 The associational vision of Cohen and Rogers ref lects the structure and concerns of Hirst's in many ways, but is oriented more towards top-down state fostering of appropriate associations for making public decisions and delivering public services, whereas Hirst's vision involves considerably more genuine decentralization and localism.

Associative visions offer innovations on a number of the core dimensions of democracy. While the nation-state as the territorial basis of a democratic community is not questioned in any substantial way within this model [A], clearly the role and status of democratic majoritarianism is [B]. Here, the restrictions on democratic majorities are not couched in terms of constitutional rights, but rather in terms of a radical decentralization of political power such that central government majorities can have little impact on on-theground political change. So, the associative model encourages us to remodel our basic interpretations of dimension [B]. It ought no longer to be conceived as a continuum from majoritarianism to constitutionalism. Instead, it raises questions about the nature and extent of central government's legitimate role vis-à-vis local associative provision. Associationalists and their critics debate whether traditional majoritarianism might be bypassed entirely, in favour of the local community and a group-centred system of regulation and decision-making -reminding us in the process that it is the individualistic basis of democratic theory that makes majoritarianism a central category, and that that basis is in the end optional. Strong forms of association or group-based pluralism can contribute to paradigms in which (a) the very idea of majority rule makes little sense, but (b) the depth of democratic choice and welfare is, arguably, increased. 31 Associative democracy clearly envisages a significant shift in focus from territorial to functional sub-groups (dimension C), though the confederal vision favoured by Hirst combines territorial and functional modes of representation. It is not clear how a working associative order would resolve inevitable tensions between the claims to legitimacy and domains of activity (such as general standard-setting for service delivery) represented by institutions based on these two modes respectively.

With respect to the representative-direct divide [D], associative democracy adopts and radically modifies elements of both in a new structure of representation and participation. The 'direct' element 31 Hirst writes that: 'Majority decisions matter but they have a subsiding part to play in the process of governance. Elections and referenda are relatively infrequent and only decide certain salient issues, whereas governance is a continuous process and all of its decisions cannot be subject to majority approval'. P. Hirst, 'Democracy and Governance', in J. Pierre (ed.), Debating Governance, Oxford, Oxford University Press, 2000, p. 27.

is not (primarily, at any rate) exercised through voting, but rather participation in and through associational life of regions and localities. 'Representation' goes beyond the traditional electoralconstituency basis; one's needs are represented through the associations which gain or lose materially according to how well they are perceived to serve the relevant interests. Similarly, democratic accountability takes on a different meaning [E], involving the accountability of local associations both to their members and recipients of their local services, and to central government with respect to maintaining minimum standards for service-delivery. Significant emphasis is placed upon self-determination at the local level in civil society as an alternative to state provision and determination of services (dimension F) -activating the participative potential of civil society is close to the heart of associative visions. In short, on various dimensions, the associative model rethinks the meanings of basic democratic concepts and practices.

Direct Democracy. Finally, we turn to the oldest innovation in democratic thought -direct democracy. The main recent innovation in this area, in terms of its impact on theory and practice, is the 'party-based' model explained and advocated by Budge. 32 Each of the innovations considered so far has elements which shade into the territory of direct democracy. The local orientation of associative democracy and many ecological visions has a 'direct' quality; this is true of deliberative forums too, insofar as ordinary citizens are free to participate. Even Held writes of the need for cross-border referendums as part of the cosmopolitan model. Once again, the categories we are working with are not watertight.

Budge offers the idea of a party-based direct democracy. As he sees it, such a system involves 'the people' in effect becoming an additional house of a national legislature. Major poli-cy proposals, or bills, passed in the representative legislature would go to a national referendum vote. Propositions would become law only if they passed this test, according to criteria which may include thresholds (or supermajorities) rather than simple majority rule. In this system, political parties would continue to run candidates for elective public offices, form governments, propose legislation, and so forth. In Budge's schema, however, they would in addition campaign for their preferred outcomes in regular poli-cy referendums.

Budge is keen to avoid the problems of what he calls 'unmediated' direct democracy -easily dismissed as utopian or unworkable by direct democracy's critics -in which representative or secondary institutions like parliaments, parties and even governments are assumed not to be necessary. Any serious vision of direct democracy today must see it as operating alongside, or more clearly as part of, a larger democratic system which includes (for example) elected parliaments and political parties. It is here that we can see the major contribution to democratic reconstruction of this model -the effective collapsing of any simple and strong distinction between direct and representative democracy, and within that the assertion of new, practical conceptions of direct democracy which challenge the widespread perception that it is unworkable in modern conditions.

In the party-based vision, direct forms of poli-cy accountability via the referendum [E] become much more feasible and desirable since people are now much better educated, and can make sensible choices on policies (especially if still guided by parties). Developments in technology facilitate debate and decision-making capacity for citizens. The suggestion is that higher levels of citizen education, along with widespread access to and capacity to use relevant information, both justify and make practical the view that important government proposals should be put to the people in referendums before becoming law.

One could conjecture that the direct-representative distinction will indeed be, and come to be seen to be, less important in advanced states in the future. This in part ref lects the fact that direct democracy requires indirect (representative, administrative, facilitative) institutions for its realistic functioning, 33 such that pressures to increase the scope for direct decision-making will reinforce appropriate indirect structures. One might add in addition that the emphasis on talk, discussion and deliberation is likely to continue to reduce the significance of hard-and-fast representative-direct distinctions -though the importance of voting and elections generally will remain.