Democracy without Elections

Brian Martin

Go to

Brian Martin's publications on demarchy and democracy

Brian Martin's publications

Brian Martin's website

Published in Social Anarchism, Number 21, 1995-96, pp. 18-51.

Reprinted in Finnish in Timo Ahonen et al. (eds.), Väärin Ajateltua: Anarkistisia Puheenvuoroja Herruudettomasta Yhteiskunnasta (Kampus Kustannus, 2001), pp. 113-144. A truncated version appeared in Bulletin of Anarchist Research, No. 25, pp. 8-20 (Autumn 1991). An earlier, greatly abridged version appeared in Social Alternatives, Vol. 8, No. 4, January 1990, pp. 13-18. Reprinted in Russian in Black Line, supplement, 1993, pp. 7-27. Revised version in Howard Ehrlich (ed.), Reinventing Anarchy Again (Edinburgh: AK Press, 1996), pp. 123-136.

Audio file of this article


One: The case against elections

Two: Alternatives to Elections

Three: Demarchy


For many a jaded radical, the greens are the most exciting political development for ages. The green movements claim to bring together members of the most dynamic social movements, including the peace, environmental and feminist movements, combining their insights and numbers. This is something that many activists have long sought.

Beyond this, the rapidly achieved electoral success of green parties has really captured the imagination. The German Greens have been the centre of attention for a decade precisely because of their election to parliament. A number of other green parties have been electorally successful too.

But wait a moment. Before getting too carried away, isn't it worth asking whether elections are an appropriate way forward? After all, electoral politics is the standard, traditional approach, which has led to those traditional parties which have so frustrated many a radical. Isn't there a danger that participation in the electoral process remains a trap, a bottomless pit for political energy which will pacify activists and masses alike?

My aim here is to take a critical look at elections and their alternatives. I start in Part One with a summary of the case against elections. Much of this will be familiar to anarchists, but it may be useful in bringing together the arguments and perhaps raising one or two unfamiliar ones.

If elections have limitations, then what are the alternatives? This is a harder question. In Part Two I look at some of the methods favoured by those supporting 'participatory democracy,' namely actual rule by the people rather than through elected representatives. These participatory methods, naturally enough, have both strengths and weaknesses. One of their key weaknesses is that it is hard for them to deal with decision making involving large numbers of people without succumbing to some of the same problems as representative systems.

Finally, in Part Three I present the idea of demarchy, a participatory system based on random selection. This is, I believe, a most promising alternative. It is little known, but in recent years there have been important theoretical developments and practical experiences.


One: The case against elections

The idea of elections as the ultimate democratic device is a deep-seated one in the West. It is hard to escape it. Children are taught all about elections in school, and may vote for student councils or club officers. Then all around us, especially through the mass media, attention is given to politicians and, periodically, to the elections which put them in power. Indeed, the main connection which most people have with their rulers is the ballot box. It is no wonder that electoral politics is sanctified. If a country has no elections, or only sham elections, this is taken as a sign of failure [1].

Elections in practice have served well to maintain dominant power structures such as private property, the military, male domination and economic inequality. None of these has been seriously threatened through voting. It is from the point of view of radical critics that elections are most limiting.

The theory of representative democracy and popular sovereignty is based on some hidden, convenient fictions [2]. Here I'll concentrate on the practical shortcomings of electoral systems, though it would be possible to relate these to theoretical assumptions.

Voting doesn't work

At the simplest level, voting simply doesn't work very well to promote serious challenges to prevailing power systems. The basic problem is quite simple. An elected representative is not tied is any substantial way to particular policies, whatever the preferences of the electorate. Influence on the politician is greatest at the time of election. Once elected, the representative is released from popular control (recall is virtually impossible before the next election) but continues to be exposed to powerful pressure groups, especially corporations, state bureaucracies and political party power brokers.

We all know examples of politicians who have 'sold out,' relinquishing their claimed ideals and breaking their solemn promises. Ironically, this is just as true for right-wingers as for left-wingers. The radical right was very disillusioned by Ronald Reagan.

Usually the sell-outs are attributed to failures of personalities, but this is both unfair and misleading. Politicians are morally little different from anyone else. The expectations and pressures on them are much greater. Positions of great power both attract the most ambitious and ruthless people and bring out the worst features of those who obtain them [3]. It is not the individuals who should be blamed, but the system in which they operate.

In principle, elections should work all right for moderately small electorates and political systems, where accountability can be maintained through regular contact. Elections can be much better justified in New England town meetings than in national parliaments making decisions affecting millions of people. In these large systems, a whole new set of reinforcing mechanisms has developed: political party machines, mass advertising, government manipulation of the news, pork barrelling (government projects in local areas), and bipartisan politics. The party machines choose the candidates, canvass voters and impose platforms. Mass advertising treats candidates like soap powders, emphasising personality over policies [4]. Government manipulation of the news includes a variety of techniques by which the mass media become dependent on government suppliers and shapers of information. Government largess in selected regions is a standard technique to attract (or threaten) voters. Finally, bipartisan politics, namely the adoption of identical or near-identical policies by allegedly competing parties, reduces the range of issues which are subject to political debate. In essence, voters are given the choice between tweedledee and tweedledum, and then bombarded with a variety of techniques to sway them towards one or the other.

This is a depressing picture, but hope springs eternal from the voter's pen. Some maintain the faith that a mainstream party may be reformed or radicalised. Others look towards new parties. When a new party such as the greens shows principles and growth, it is hard to be completely cynical.

Nevertheless, all the historical evidence suggests that parties are more a drag than an impetus to radical change. One obvious problem is that parties can be voted out. All the policy changes they brought in can simply be reversed later.

More important, though, is the pacifying influence of the radical party itself. On a number of occasions, radical parties have been elected to power as a result of popular upsurges. Time after time, the 'radical' parties have become chains to hold back the process of radical change. Ralph Miliband gives several examples where labour or socialist parties, elected in periods of social turbulence, acted to reassure the dominant capitalist class and subdue popular action [5]. The Popular Front, elected in France in 1936, made its first task the ending of strikes and occupations and generally dampening popular militancy, which was the Front's strongest ally in bringing about change. The Labour government elected in Britain in 1945 made as few reforms as possible, leaving basic social structures untouched. By contrast, the US New Deal Democratic administration which took office in 1933 did undertake structural changes -- in order to restore and strengthen capitalism. Miliband in these examples writes from the Marxist perspective in which the state is the servant of capitalism. His insights about the reluctance of 'reforming' political leadership of the state to challenge the economic foundation of society applies even more strongly to the unwillingness of this leadership to challenge state power itself.

The experiences of Eurosocialist parties elected to power in France, Greece and Spain in the 1980s have followed the same pattern. In all major areas -- the economy, the structure of state power, and foreign policy -- the Eurosocialist governments have retreated from their initial goals and become much more like traditional ruling parties [6].

Voting disempowers the grassroots

If voting simply didn't work to bring about changes at the top, that would not be a conclusive argument. After all, change in society doesn't just come about through laws and policies. As feminists and others say, 'the personal is political,' and that means just about everything. There are plenty of opportunities for action outside the electoral system.

It is here that voting makes a more serious inroad into radical social action: it is a diversion from grassroots action. The aim of electoral politics is to elect someone who then can take action. This means that instead of taking direct action against injustice, the action becomes indirect: get the politicians to do something.

On more than one occasion, I've seen a solid grassroots campaign undermined by an election. One example is the 1977 Australian federal election in the midst of a powerful campaign against uranium mining. Another is the 1983 Australian federal election at a crucial point in the campaign against the flooding of the Franklin River in Tasmania [7].

At the simplest level, energy put into electioneering is energy not put into direct action. Some activists feel resentful that their campaign is hijacked by election priorities. This can be compensated, to some extent, by the heightened interest in the issues during an election campaign. The more serious problem is the loss of energy that usually occurs after the election.

In the December 1977 election, the pro-uranium Liberal Party was re-elected. This was very demoralising for the anti-uranium movement which had looked with hope for a victory by the Australian Labor Party with its new anti-uranium platform. Yet in retrospect the movement was having considerable success even under the Liberals. A stepped-up campaign should have been called for. But this was hard to achieve. Many anti-uranium activists, notably those who were Labor Party members, had participated in 1977 because of the upcoming election. After Labor's defeat, many of them dropped out of the movement, leaving those remaining feeling less than encouraged. The election campaign was a diversion from long-term strategy against uranium. It raised activity temporarily, to be followed by a more persistent decline.

Another problem is the centralisation of power in social movements which is encouraged by the desire to influence politicians [8]. The campaign against the flooding of the Franklin River in Tasmania illustrates this. A long and well-orchestrated campaign by a variety of means culminated in December 1982 in a 'blockade' against construction work on the dam, using classic nonviolent action techniques. During the blockade, a national election was called for March 1983. The leaders of the Tasmanian Wilderness Society and the Australian Conservation Foundation negotiated with the leaders of the Australian Labor Party, and made a deal. The conservationists would support Labor for the House of Representatives; Labor, if elected, would act to stop the dam.

My main point here is the undemocratic fashion in which deals are made in the political system. It is also interesting to note the aftermath of these negotiations. The blockade was downgraded; Labor, with support of the mainstream environmental organisations, won a close election. The new Labor government did not use its financial power directly against the dam, but rather just supported a legal action in the High Court to use federal power (in relation to a World Heritage listing) to stop a state project. This case won by one vote in the High Court. During all this time, the environmentalists were disempowered, waiting for powerful politicians and judges to decide the fate of the river. The aftermath was a powerful backlash in Tasmania, using the rhetoric of state's rights, against environmentalists.

Incidentally, the 1983 Labor government decided to renege on several remaining planks of its anti-uranium platform. The anti-uranium movement re-emerged on Labor's election, and sank again after this 'betrayal.'

It should be a truism that elections empower the politicians and not the voters. Yet many social movements continually are drawn into electoral politics. There are several reasons for this. One is the involvement of party members in social movements. Another is the aspirations for power and influence by leaders in movements. Having the ear of a government minister is a heady sensation for many; getting elected to parliament oneself is even more of an ego boost. What is forgotten in all this 'politics of influence' is the effect on ordinary activists.

The disempowering effect of elections works not only on activists but also on others. The ways in which elections serve the interests of state power have been admirably explained by Benjamin Ginsberg [9]. Ginsberg's basic thesis is that elections historically have enlarged the number of people who participate in 'politics,' but by turning this involvement into a routine activity (voting), elections have reduced the risk of more radical direct action.

The expansion of suffrage is typically presented as a triumph of downtrodden groups against privilege. Workers gained the vote in the face of opposition by the propertied class; women gained it in the face of male-dominated governments and electorates. Ginsberg challenges this picture. He argues that the suffrage in many countries was expanded in times when there was little social pressure for it.

Why should this be? Basically, voting serves to legitimate government. To bolster its legitimacy, if required, suffrage can be expanded. This is important when mass support is crucial, for example during wartime. It can be seen in other areas as well. Worker representatives on corporate boards of management serve to coopt dissent; so do student representatives on university councils.

Ginsberg shows that elections operate to bring mass political activity into a manageable form: election campaigns and voting. People learn that they can participate: they are not totally excluded. They also learn the limits of participation. Voting occurs only occasionally, at times fixed by governments. Voting serves only to select leaders, not to directly decide policy. Finally, voting doesn't take passion into account: the vote of the indifferent or ill-informed voter counts just the same as that of the concerned and knowledgeable voter. Voting thus serves to tame political participation, making it a routine process that avoids mass uprisings. The expansion of suffrage helps to reduce the chance that a revolt by an oppressed or excluded group will be seen as justified; with the vote, it is easy for others to claim that they should have used 'orthodox channels.'

Voting reinforces state power

Ginsberg's most important point is that elections give citizens the impression that the government does (or can) serve the people. The founding of the modern state a few centuries ago was met with great resistance: people would refuse to pay taxes, to be conscripted or to obey laws passed by national governments. The introduction of voting and the expanded suffrage have greatly aided the expansion of state power. Rather than seeing the system as one of ruler and ruled, people see at least the possibility of using state power to serve themselves. As electoral participation has increased, the degree of resistance to taxation, military service, and the immense variety of laws regulating behaviour, has been greatly attenuated.

The irony in all this, as pointed out by Ginsberg, is that the expansion of state power, legitimated by voting, has now outgrown any control by the participation which made it possible. States are now so large and complex that any expectation of popular control seems remote. Yet, as he comments, the "idea that electoral participation means popular control of government is so deeply implanted in the psyches of most Americans that even the most overtly skeptical cannot fully free themselves from it" [10]. Needless to say, this statement applies to many countries besides the United States.

Using Ginsberg's perspective, the initial government-sponsored introduction of some competition into elections in the Soviet Union and eastern Europe takes on a new meaning. If the economic restructuring seen as necessary by Communist Party leaders was to have any chance of success, then there had to be greater support for the government. What better way than by introducing some choice into voting? Increased government legitimacy, and hence increased real power for the government, was the aim.

Change in Eastern Europe has gone far past that planned by governments, of course. Still, it is revealing that a key demand of reformers has been to introduce multi-party elections. What is sought is a change in the running of government, not in the basic mechanisms of governance.

Although expanding the franchise does help legitimate government, it certainly does not close off political struggle. The introduction of voting and the expansion of suffrage may institutionalise political activity, but they do after all allow the activity. Elections may reduce the chance of radical challenges to the status quo, but that chance does exist. Electoral politics legitimates government to the extent that governments are to some extent dependent on the will of the people -- however routinised and institutionalised the expression of the people's will may be. Because elections provide a channel for radical change, even though a very constrained channel, the hope of radicals is maintained and their reliance on elections is encouraged.

Ginsberg's analysis leads to the third major limitation of electoral politics: it relies on the state and reinforces state power. Of course, this is simply another facet of the two previous objections, namely that elections don't work to bring about radical change (because the state machinery is designed for other interests) and that elections disempower the grassroots (because energy is channelled into the state).

The basis of an anarchist critique of voting is that voting participates in the legitimation of the state [11]. If the state is part of the problem -- namely being a prime factor in war, genocide, repression, economic inequality, male domination and environmental destruction -- then it is foolish to expect that the problems can be overcome by electing a few new nominal leaders of the state.

It is possible to paint a more sophisticated picture of the state, in which there are continual struggles inside and outside the state apparatuses to shape policies and to serve and empower different groups of people. In this picture, it is worth struggling within the state, for example for welfare measures for the poor or against aggressive military policies [12]. There are few who would object to this. But even with this more sophisticated picture, the fundamental critique of the state can still apply.

The basic point concerns whether the organisational structure of the state is neutral or not. If the structure of the state is assumed to be neutral, then the exercise of state power can be seen as the playing out of various power struggles, such as capitalist power versus workers' power or male power versus female power. If the structure of the state is neutral, then the state can be seen as a site for class struggle, gender struggle, etc. This is typical perspective adopted by Marxists, some feminists and most liberals. It is quite an improvement from the picture of the state as a complete tool of the capitalist class. But it does not question the basic assumption of the neutrality of the state structure, which as a consequence can be captured one way or another, either by the simplistic image of taking state power or by the more sophisticated image of working in and against the state.

The basic anarchist insight is that the structure of the state, as a centralised administrative apparatus, is inherently flawed from the point of view of human freedom and equality. Even though the state can be used occasionally for valuable ends, as a means the state is flawed and impossible to reform. The nonreformable aspects of the state include, centrally, its monopoly over 'legitimate' violence and its consequent power to coerce for the purpose of war, internal control, taxation and the protection of property and bureaucratic privilege.

The problem with voting is that the basic premises of the state are never considered open for debate, much less challenge. The state's monopoly over the use of violence for war is never at issue. Neither is the state's use of violence against revolt from within. The state's right to extract economic resources from the population is never questioned. Neither is the state's guarantee of either private property (under capitalism) or bureaucratic prerogative (under state socialism) -- or both.

Voting can lead to changes in policies. That is fine and good. But the policies are developed and executed within the state framework, which is a basic constraint. Voting legitimates the state framework.

One response to the limitations of electoral politics is to campaign against voting and elections. This is useful in raising awareness of the limitations of electing one's rulers. But such a critique needs to be supplemented by the promotion of alternatives to the state. That is a harder task. After all, there's no use in criticising electoral methods if there isn't anything better.


Two: Alternatives to Elections

What participatory alternatives are there to the state and electoral politics? This is a topic on which there is a large literature, especially by anarchists [13]. So I can do no more than highlight some of the relevant answers and experiences. I will emphasise some of the limitations of the standard responses to this problem, since it is essential to be as critical of alternatives as of the existing system.


One set of alternatives is based on direct mass involvement in policy-making through voting, using mechanisms including petition, recall, initiative and referendum. In short, instead of electing politicians who then make policy decisions, these decisions are made directly by the public.

Referendums have been used widely in the United States, often to the consternation of powerful groups. The fluoridation of public water supplies as a measure to reduce tooth decay has resulted in hundreds of referendums, for example. The more frequent result has been against fluoridation, much to the consternation of proponents, who as a result have counselled against referendums and tried for implementation directly by governments.

In practice, referendums have been only supplements to a policy process based on elected representatives. But it is possible to conceive of a vast expansion of the use of referendums, especially by use of computer technology [14]. Some exponents propose a future in which each household television system is hooked up with equipment for direct electronic voting. The case for and against a referendum proposal would be broadcast, followed by a mass vote. What could be more democratic?

Unfortunately there are some serious flaws in such proposals. These go deeper than the problems of media manipulation, involvement by big-spending vested interests, and the worries by experts and elites that the public will be irresponsible in direct voting.

A major problem is the setting of the agenda for the referendum. Who decides the questions? Who decides what material is broadcast for and against a particular question? Who decides the wider context of voting?

The fundamental issue concerning setting of the agenda is not simply bias. It is a question of participation. Participation in decision-making means not just voting on predesigned questions, but participation in the formulation of which questions are put to a vote. This is something which is not easy to organise when a million people are involved, even with the latest electronics. It is a basic limitation of referendums.

The key to this limitation of referendums is the presentation of a single choice to a large number of voters. Even when some citizens are involved in developing the question, as in the cases of referendums based on the process of citizen initiative, most people have no chance to be involved in more than a yes-no capacity. The opportunity to recast the question in the light of discussion is not available.

Another problem for referendums is a very old one, fundamental to voting itself. Simply put, rule by the majority often means oppression of the minority. This problem is more clear-cut in direct voting systems, but also appears in representative systems.

Historically, the referendum approach assumes the existence of a bureaucratic apparatus for implementing the decisions made. Referendums don't implement themselves, certainly. Who does? The state. Referendums, in practice, are a way of increasing participation within the parameters of centralised administration. This latter problem is not intrinsic to the referendum as a method. The challenge is to recast the referendum as part of a more participatory political process.


Consensus decision-making has become widely used in a number of social movements in the past couple of decades, especially in portions of the anti-nuclear power movement. In general parlance, 'consensus' means gaining general agreement, but within social movements it has been given a more precise, operational meaning.

The basic aim is for a group of generally like-minded people to reach a common decision without greatly alienating anyone. This might be a collective working on a newspaper or a group planning a direct action against a military facility. Voting is avoided for several reasons. Those who lose a close or bitterly contested vote often fail to support the majority position, and sometimes even end up leaving the group. Lots of energy is wasted in lobbying and building factions for the purposes of winning votes rather than developing the best campaign. Finally, innovative proposals are often ignored because they seem to stand no chance in a vote.

The basic procedure in consensus decision-making is that various options are canvassed and discussed. If everyone seems to be agreeing, then a test is made for consensus. If no one disagrees, consensus has been reached. If anyone disagrees, they are encouraged to spell out their objections. Consensus is blocked if there is strong disagreement by even one person (or, in modified consensus, by a specified small fraction).

If consensus is blocked, then the group seeks ways to reach agreement. The arguments can be reexamined; new proposals can be raised and discussed; the decision can be postponed until a later time. For example, the group may break itself into a number of small groups which readdress the issue, seeking a resolution.

In many cases, the procedure works remarkably well. Those with divergent views generally see that they are taken seriously, and this builds the cohesion of the group. Sometimes a minority view eventually becomes the consensus view: there is no quick vote to overwhelm it. Most encouraging of all, sometimes brilliant new solutions are developed in the efforts to reach consensus.

That consensus methods often work well should come as no surprise, since they have long been used in an unacknowledged manner in all sorts of situations. For many organisations, official votes are ritualistic only. A vote is seldom taken unless it is obvious beforehand that everyone agrees, or at least that no one strongly disagrees.

The practice of consensus decision-making formalises the process. This is most important as the group gets larger. For large groups there are various methods involving subgroups and delegates which ensure that the basic consensus approach is followed.

An important difference between consensus and normal 'meeting procedure' is the role of leadership. The conventional method has a formal leader (the chair) and a set of formal rules for setting the agenda, speaking, making motions, voting, etc. -- the familiar Roberts rules of order. The consensus approach has no formal leader but instead 'facilitators' who are supposed to help the group do what it wants to. The facilitators are crucial to the success of consensus: they are supposed to test for consensus, encourage less articulate group members to participate, offer suggestions for procedure, summarise views expressed, etc. The ideal is when every group member helps in facilitation, so there is no obvious leader at all.

Consensus, then, is a method of decision-making without voting that aims for participation, group cohesion, and openness to new ideas. Combined with other group skills for social analysis, examining group dynamics, developing strategies and evaluation, consensus can be powerful indeed [15].

Yet anyone who has participated in consensus decision-making should be aware that the practice is often far short of the theory. Sometimes powerful personalities dominate the process; less confident people are afraid to express their views. Because objections normally have to be voiced face-to-face, the protection of anonymity in the secret ballot is lost. Meetings can be interminable, and those who cannot devote the required time to them are effectively disenfranchised. The biggest problem for consensus, though, is irreconcilable conflict of interest.

The best treatment of this problem is Beyond Adversary Democracy by Jane Mansbridge [16]. Mansbridge distinguishes between two types of democracy. What she calls adversary democracy is the familiar electoral approach. It is based on the assumption of conflicting interests, majority rule, secret ballot and equal protection of interests. What she calls unitary democracy is like friendship. It is based on a high degree of common interest, consensus-like methods, face-to-face decision-making and a rough equality of mutual respect.

Mansbridge closely analyses two cases in detail: a New England town meeting which formally uses voting but in practice often seeks consensus, and a work collective which uses formal consensus methods.

Mansbridge points out that the standard approach is to assume conflicting interests and to use adversarial methods, but that unitary interests are much more common than generally realised. Hence seeking unity, rather than assuming conflict, is often preferable. Her most important point though, for my purposes here, is that consensus has a complementary weakness: it can't handle deep-seated conflict.

Much of such conflict is based in inequality of power. To imagine employers and workers in a typical enterprise trying to reach consensus is difficult. They don't have common interests or, very often, equality of respect. In a self-managed enterprise, by contrast, there are no separate employers and consensus becomes more feasible.

Other types of conflict are just as difficult to deal with. Imagine a group of anarchists, Marxists and liberals (with a few conservatives tossed in for good measure) trying to reach consensus on a campaign for reducing crime. Even with the best will in the world, the different perspectives on the world are likely to undermine attempts at consensus on more than the most superficial level.

The larger the group, the more likely there are to be fundamental conflicts of interests. Consensus is most likely to work in small self-selected groups. But as a democratic alternative to elections it has severe limitations dealing with large groups. The problems of consensus are also the problems of self-management in large groups [17].

Small size

One solution to this dilemma is to keep group sizes small. Rather than centralisation of power, decentralisation is the aim. There is no intrinsic reason why education, health, investment and many other functions have to be administered at the level of many millions of people rather than, say, thousands or tens of thousands. Many of the most participatory polities, from ancient Greece to today, have been relatively small. Conversely, many of the ills of electoral politics seem associated with the enormous population in many countries [18].

Small size reduces the severity of many of the problems of decision-making. Even voting is not so limiting when the number of voters is so small that everyone is potentially known to everyone else. The use of consensus can be maximised.

Furthermore, small size opens the possibility of a plurality of political systems. Frances Kendall and Leon Louw propose a Swiss-like federation of autonomous political entities, each of which can choose its own political and economic system [19]. With Kendall and Louw's system, the difficulties of trying new methods, and the costs of failures, are greatly reduced.

Small size may make governance easier, but there will still be some large-scale problems requiring solution. Global pollution and local disasters, for example, call for more than local solutions. How are decisions to be made about such issues?

More fundamentally, small size by itself doesn't solve the issue of how decisions are made. There can still be deep conflicts of interests which make consensus inappropriate, and there can still be problems of domination resulting from electoral methods.

Finally, in all but the very tiniest groups, the basic problem of limits to participation remains. Not everyone has time to become fully knowledgeable about every issue. Consensus assumes that everyone can and should participate in decisions; if substantial numbers drop out, it becomes rule by the energetic, or by those who have nothing better to do. Representative democracy, by contrast, puts elected representatives in the key decision-making roles; the participation of everyone else is restricted to campaigning, voting and lobbying. In both cases participation is very unequal, not by choice but by the structure of the decision-making system.

Delegates and federations

A favourite anarchist solution to the problem of coordination and participation is delegates and federations. A delegate differs from a representative in that the delegate is more closely tied to the electorate: the delegate can be recalled at any time, especially when not following the dictates of the electors. Federations are a way of combining self-governing entities. The member bodies in the federation retain the major decision-making power over their own affairs. The members come together to decide issues affecting all of them. In a 'weak federation,' the centre has only advisory functions; in a 'strong federation,' the centre has considerable executive power in specified areas. By having several tiers in the federation, full participation can be ensured at the bottom level and consultation and some decision making occurs at the highest levels.

Delegates and federations sound like an alternative to conventional electoral systems, but there are strong similarities. Delegates are normally elected, and this leads to the familiar problems of representation. Certain individuals dominate. Participation in decision-making is unequal, with the delegates being heavily involved and others not. To the degree that decisions are actually made at higher levels, there is great potential for development of factions, vote trading and manipulation of the electorate.

This is where the delegate system is supposed to be different: if the delegates start to serve themselves rather than those they represent, they can be recalled. But in practice this is hard to achieve. Delegates tend to 'harden' into formal representatives. Those chosen as delegates are likely to have much more experience and knowledge than the ordinary person. Once chosen, the delegates gain even more experience and knowledge, which can be presented as of high value to the electors. In other words, recalling the delegate will be at the cost of losing an experienced and influential person.

These problems have surfaced in the German Green Party. Although formally elected as representatives, the party sought to treat those elected as delegates, setting strict limits on the length of time in parliament. This was resisted by some of those elected, who were able to build support due to their wide appeal. Furthermore, from a pragmatic point of view (which is often hard to resist), those who had served in parliament had the experience and public profile to better promote the green cause. Thus the delegate approach came under great stress even though the green politicians had little real power. In a situation when the delegates are truly making decisions, the stresses will be much greater.

The fundamental problem with the delegate system, then, is unequal participation. Not everyone can be involved in every issue. With delegates, the problem is resolved by having the delegates involved much more in decision-making, at the expense of others. This unequal participation then reproduces and entrenches itself. The more layers there are to the federation, the more serious this problem will be. Federations, as well, are not a magical solution to the problem of coordination in a self-managing society.

In this brief survey of some of the more well-known participative alternatives to elections, I've focussed on their limitations. But these and other methods do have many strengths, and are worth promoting as additions or alternatives to the present system. Consensus has been developed enormously over the past couple of decades as a practical decision-making method. The potential of decentralisation is undoubtedly great. Indeed, the greatest successes of consensus have been in small groups. As well as the idea of federations, there is also much attention to networks, which do not assume any set of levels for decision making.

Rather than dismissing these possibilities, my aim is to point out some of the problems that confront them. The most serious difficulty is how to ensure participation in a wide range of issues that affect any person. How can the (self-managed) activities of large numbers of people be coordinated without vesting excessive power in a small group of people?

I now turn to 'demarchy,' which is one answer to some of these problems. It is by no means the only or final answer. But it is an approach that holds potential and, in my opinion, is worth much investigation and experimentation.

John Zube advocates 'panarchy,' the peaceful coexistence of a diversity of methods for voluntary association [20]. In this spirit, demarchy can be considered as one candidate for organising society in a participative fashion.


Three: Demarchy

The most eloquent account of demarchy is given by John Burnheim in his book Is Democracy Possible?[21]. Burnheim begins by analysing the state and bureaucracy, and concludes that they are central obstacles to the achievement of true democracy. He includes electoral politics as part of the problem. Since the word 'democracy' is so tainted by association with representative government, Burnheim coined the word 'demarchy' to refer to his alternative.

Demarchy is based on random selection of individuals to serve in decision-making groups which deal with particular functions or services, such as roads or education. Forget the state and forget bureaucracies. In a full-fledged demarchy, all this is replaced by a network of groups whose members are randomly selected, each of which deals with a particular function in a particular area.

For example, in a population of 10,000 to 100,000, there might be groups dealing with transport, health, agriculture, industry, education, garbage, housing, art and so forth, or particular aspects of such functions such as rail transport. Each group would be chosen randomly from all those who volunteer to be on it. The groups could be perhaps 10 or 20 people, large enough to obtain a variety of views but small enough for face-to-face discussion. The groups themselves could use consensus, modified consensus, voting or some other procedure to reach decisions. They could call for submissions, testimony, surveys and any other information they wished to obtain.

Before going further, it is worth looking more closely at random selection (also called sortition). This was used in ancient Athens as a democratic selection device, but has been little used since. Of course, Athenian democracy was limited, excluding women and slaves. Nevertheless, there are many things that can be learned from it [22].

One of the values of random selection was to increase participation and prevent the formation of factions. When the assembly met, the chairman was selected by lot at the beginning of the meeting. In this way there was no opportunity for pre-assembly plotting to push towards particular outcomes by putting pressure on the chairman.

The Athenians used voting too, for example in choosing military leaders. In fact, they used a variety of democratic devices, each chosen for particular purposes. Writers on liberal democracy today draw on the Athenian experience selectively. They use it to justify representation, but ignore or dismiss the use of sortition. Indeed, democracy is often defined today as representative democracy.

The major use of random selection for important decision making today is the jury, which itself prospers in only the United States and a few other countries. 'Ordinary people' are randomly chosen to decide on the fate of their fellow humans. The jury is embedded in a political framework which constrains its potential: the framework of laws which is biased towards the interests of the privileged; the selective enforcement of law; manipulations by lawyers, judges and media. Considering these obstacles, the jury performs remarkably well.

Many governments have dispensed with juries, arguing that professional judges are more suitable. The argument that juries are less capable of dealing with complex technical issues is a vexed one. Arguably, a jury of a dozen people is likely to contain one or two people more technically competent than the average judge.

It is certainly the case that juries are hated by repressive governments. Judges can by pressured more easily by governments than can juries.

From a decision-making point of view, the great advantage of the jury is its capacity for testing opinions. In terms of participation in decision-making, the jury is a form of policy making, though this is greatly discouraged by most judges [23].

Considering that most jury members are given no training in critical assessment of evidence and formation of conclusions, in consensus decision-making techniques or in the role of the jury itself, the decision-making record of juries is remarkably sound [24]. Rather than attacking the failures of the jury system, it would be more appropriate to develop ways of making it function better.

Returning now to Burnheim's picture of demarchy, how does it handle the basic problems of participation? Because there are no elections and no representatives, the problems of unequal formal power, disempowerment of electors, regulation of participation and so forth do not apply -- at least not in the usual way. Formal participation occurs instead through random selection onto 'functional groups,' namely groups dealing with particular limited areas. Random selection for each group is made only from those who volunteer, just as politicians must volunteer. The difference is the method of selection: random selection rather than election.

Few people would volunteer for every possible group. Most are likely to have special interests, such as postal services, art, manufacture of building materials and services for the disabled. They could volunteer to serve on the relevant groups, and also make submissions to the groups, comment on policies and in other ways organise to promote their favoured policies.

Demarchy solves the problem of participation in a neat fashion. Recognising that it is impossible for everyone to participate on every issue in an informed fashion, it avoids anything resembling a governing body which makes far-reaching decisions on a range of issues. Instead, the functional groups have a limited domain. The people who care most about a particular issue can seek to have an influence over policy in that area. They can leave other issues to other groups and the people most concerned about them. This is basically a process of decentralisation of decision-making by topic or function rather than by geography or numbers.

Leaving decision-making to those who care most about a topic has its dangers, of course: self-interested cliques can obtain power and exclude others. That is what happens normally in all sorts of organisations, from governments and corporations to social movement groups. Demarchy handles this problem through the requirement of random selection. No one can be guaranteed a formal decision-making role. Furthermore, the terms of service are strictly limited, so no permanent executive or clique can develop.

Another problem then looms. Won't there be biases in the groups selected, because only certain sorts of people will volunteer? Won't most of the groups, for example, be dominated by white middle-aged men? This poses no problem, given a suitable adaptation of how the random selection is carried out. Suppose, for example, that 80 men and 20 women volunteer for a group of 10, for which it is desired to have an equal number of men and women. The method is simply to select 5 men randomly from the 80 male volunteers and 5 women from the 20 female volunteers. In this way, the sex balance in the group can equal that in the overall population even with different rates of volunteering. The same principle can be applied to characteristics such as ethnic origin, social class, age, occupation and religion.

This may sound logical enough, but who is to make the decisions about what groups are represented in what ways? After all, if a group decides on its own criteria for selection, this is open to abuse. Burnheim's solution to this is what he calls second-order groups. These are groups which act analogously to a judicial system for the operation of demarchy. The second-order groups deal only with procedural issues, such as what (first-order) groups should exist, how the random selection should be carried out, and any other disputed point.

Obviously, members of the second-order groups should have had experience in the first-order groups. How should they be selected? Burnheim suggests that first-order groups should select from among their members those most suitable for second-order groups. Bob James argued to me that this really goes against the guiding principle of demarchy, which is random selection of interested people rather than selection on the basis of performance or popularity. He suggests that second-order group members be chosen randomly from first-order group members. My guess is that the differences between these two procedures would not be so great. Even with a random selection, it is likely that members seen to be performing well would be strongly encouraged by their colleagues to stand for second-order group membership, which would probably not be all that sought after anyway. Finally, the limited term of office on the second-order groups will prevent entrenchment of power.

Several features distinguish demarchy from representative democracy, including random selection, functional groups, limited tenure of office, and elimination of the state and bureaucracy. Some of these could serve as reforms to representative democracy, but there is also a coherency in the entire package.

For example, a limited term of office, say two years, would help prevent entrenchment of power in representative systems. Why should demarchy be better able to sustain such a requirement for turnover of members of decision-making groups? One difference lies in the legitimacy attached to the selection principle. Representatives justify their position in terms of repeated majority preference for their personal selection. Randomly selected individuals have no special legitimation except the random process itself. The legitimacy of random selection lies in regular replacement rather than popular mandate or acquired experience, and this type of legitimacy more easily allows challenges should those in office attempt to extend their term. A similar difference can be seen in the often lengthy tenure of judges, whether appointed or elected, compared to jurors.

There is not the space here to go into many of the issues raised by the concept of demarchy. Suffice it to say that there are many unanswered questions and many areas where further elaboration is required. I'll mention only a few here.

First, implementation of decisions. Burnheim has rejected the state and bureaucracy, so there won't be any permanent staff to carry out decisions made by the demarchic groups. Burnheim says that the groups will carry out the decisions themselves. That sounds fine in theory, but what will it mean in practice?

Second, how will decisions be enforced? Remember, there is no state and hence no military. Essentially, decisions will be effective if people abide by them, and this depends on the overall legitimacy of the system. Actually, this isn't too different from many aspects of present society. Most people accept the need to act in a sensible manner towards babies, public parks and (for that matter) private property, even when the possibility of legal sanctions and apprehension by the police is remote. Force plays only a limited role in the routine operation of society. In a more participatory society, force could play an even more limited role. The corollary of this is that unpopular decisions by demarchic groups would simply lapse through non-observance. The groups would have to take into account the willingness of the population to accept their decisions.

Without the state, there would be no military. How would a community defend itself against external aggression? One possibility is arming the people [25]. However, the most participatory alternative to military defence is social defence, based on popular nonviolent resistance [26]. Demarchy and social defence have many compatible features [27].

Third, a big unanswered question is the nature of the economic system associated with demarchy. In principle, a range of systems are compatible with demarchic decision-making. A group could make a contract for recycling services either with a privately owned company or with a self-managed collective. Demarchy, though, is not compatible with bureaucratically organised economic systems, either socialist or capitalist.

Burnheim argues for extension of the principles of demarchy to economics. For example, there would be demarchic groups to make decisions about particular areas of land. Rents could be charged for uses of the land, and the rents would take the place of taxation, since there is no state to collect taxes. This is an adaptation of Henry George's ideas. The random selection for groups making decisions over portions of land would prevent vested interests from gaining a stranglehold over the political-economic process. Burnheim would also extend this idea to control over labour and money as well as land. These ideas are in a very preliminary form.

One other important problem is the basic one of participation. What if people don't volunteer? What if certain groups don't produce enough volunteers for their quota? In some cases this would be a sign of success. If the way things are operating is acceptable to most people, then there would be no urgency about becoming a member of a decision-making group. By contrast, in controversial areas participation is not likely to be a problem. If topics such as abortion or genetic engineering generated passionate debate, then concerned individuals and groups would find it fruitful to educate as many people as possible about the issues and encourage them to stand for random selection. Indeed, any unpopular decision could generate a mobilisation of people to stand for selection. Furthermore, the people mobilised would have to span a range of categories: men and women, young and old, etc. As a result, participation and informed comment would be highest in the areas of most concern. In other areas, most people would be happy to let others look after matters.

It would be easy to carry on at length about the hypothetical features of demarchy. But what's the point if it's all just a vision?

Burnheim has given the most eloquent expression of demarchy, but he is far from the only person with these sorts of ideas. Random selection, after all, has been around at least since the ancient Greeks, and it should not be surprising that advocates emerge now and again.

Burnheim's vision is a very decentralised and participatory use of random selection. By contrast, others have advocates random selection for the US Congress, for example replacing the elected House of Representatives by a randomly selected 'Representative House' [28]. These proposals have many merits but leave intact the power of the state.

Of special interest are those who have tried out random selection in practice. One such person is Ned Crosby, a political scientist from Minnesota in the United States. In the 1970s, Crosby developed his own idea of a political alternative involving random selection, with a much more centralised system than Burnheim. But failing to find a publisher for his book, he decided to work on practical implementation.

Crosby set up an organisation which is now called the Jefferson Center for New Democratic Processes. It has devoted most of its energy towards practical experiments in random selection for policy-making [29].

One project concerned the question of whether to introduce school-based clinics to deal with teenage pregnancies and sexually transmitted diseases, a very contentious issue in the state [30]. The Jefferson Center convened a number of groups of randomly selected people which they call policy juries. A 12-person policy jury was organised for each of the eight congressional districts in the state of Minnesota. Using the telephone directory, 100 people in each district were selected randomly and contacted and surveyed about their basic views concerning school-based clinics. Then a jury pool was set up from those contacted by ensuring that the demographic characteristics (ethnic origin, sex, social class) of those in the pool matched those in the overall population. Then people from each pool were selected randomly and invited to be policy jurors, until 12 jurors were obtained. In this process, it was ensured that the preliminary views about school-based clinics of the jurors matched the percentages found in the overall survey. Thus, the resulting policy juries very nearly matched the overall population both in demographic characteristics and in preliminary views on school-based clinics.

The policy jury in each district held 'hearings': they listened to various experts, heard testimony from partisans on each side of the issue, and discussed the issues among themselves. At the end of four days of deliberation, each policy jury took a vote concerning various policy alternatives. As well, each jury gave reasons for its views, made additional policy recommendations and evaluated the experience of the policy jury itself.

In addition to the eight district-based policy juries, a state-wide policy jury of 24 people was set up with three members from each district jury. This state-wide jury went through a similar process. The recommendations of the policy juries were made available to Minnesota state legislators, and also widely publicised in the media. Through all this, the Jefferson Center provided the essential support for the process. It carried out the surveys, the random selection, convening of the juries, arranging for expert witnesses, coordination of the jury deliberations and writing up and publicising of the recommendations. To carry weight, it was essential that the Jefferson Center be perceived as committed to a fair process and not to any partisan view on the issue being discussed.

In the above description, I've given only the basic outline of the process. There are many more details for those interested [31]. The basic question to be asked is "how well did it work?" In terms of democratic processes, the answer must be "remarkably well." The key test here is the response of the jurors themselves. They quickly became very committed to the process, taking it extremely seriously. They demonstrated a good grasp of the issues and made 'sensible' recommendations. They also evaluated the process very positively.

From the outside, the policy juries were also well received. They were given favourable reviews by the media and taken seriously by politicians, who recognised the grassroots origin of the views expressed. The Jefferson Center has had similar experiences and success with policy juries on topics such as pollution of water supplies by agricultural chemicals.

The policy juries are not the equivalent of the decision-making groups in demarchy. Policy juries have no formal power, which remains with elected representatives. The policy juries can only influence policy on the basis of the persuasiveness of their views and the process which led to them. But then, in one sense, this is not so entirely different from demarchic bodies, which would gain most of their power from community acceptance.

There are several lessons for promoting demarchy from the Jefferson Center projects. First, random selection can be seen as a legitimate basis for a process leading to policy recommendations. Second, participants become strongly involved in the decision-making process; policy juries are practical experiences in participation which may whet the appetite for more. Third, extensive and careful planning is essential to the success of policy juries. It should be remembered that enormous preparation and energy is put into making elections 'work' in legitimating a certain policy process. To be fairly judged, the same preparation and energy must be devoted to demarchic alternatives.

Finally, policy juries represent a practical intermediate stage for advocates of demarchy. Crosby sees random selection as a means for reforming and revitalising democracy in the United States, making government truly responsive to the will of the people. Demarchy, as presented here, is a more fundamental restructuring of society, eliminating the state altogether. This difference in goals need not cause any special problem. After all, there is a great need for practical steps which are valuable in themselves but also the basis for more fundamental change. Cooperatives can be an experience and a step towards an economy based on production for use rather than profit or control. Similarly, policy juries can be an experience and a step towards demarchy.

Quite independently of the Jefferson Center, similar projects were being undertaken in West Germany beginning in the 1970s, led by Peter Dienel at the University of Wuppertal [32]. The groups of randomly selected citizens brought together for these projects are called 'planning cells.' The cells have dealt with issues such as energy policy, town planning and information technology. The cell members are typically brought together for four days of talks, discussions and evaluations, and are compensated for wages foregone.

Planning cells have many similarities to policy juries. Here I'll just mention a few highlights, focussing especially on differences. First, the planning cells have usually been given wider briefs. Rather than focus on particular policy options on a well defined issue such as school-based clinics, a broader range of scenarios is dealt with. For example, in looking at energy policy, several options were canvassed, ranging from a heavily nuclear future to a soft energy path based on energy efficiency and renewable energy technologies. These widely divergent futures are part of conventional political debate, to be sure. But seldom are they confronted in direct fashion in the normal course of policy-making, which deals for the most part with the issues in terms of particular urgent decisions on particular projects -- a waste disposal site, a regulatory decision, a funding decision. The planning cells are able to deal with broad social issues and take a long term view, certainly far longer than the typical politician concerned about the next election.

Second, the planning cells make more use of small group techniques. Much of the time of members of a 25-person planning cell is spent in groups of 5, discussing the issues. Whereas the policy jury seems to be modelled on an actual jury, hearing testimony and discussing the issue in the full group, the planning cells are somewhat more oriented to mutual support and building up the participation and understanding of the cell members.

But these differences are minor compared to the major similarities: random selection of group members to deal with policy issues. The striking result is that most of the randomly selected volunteers quickly become quite knowledgeable about the subject matter and committed to the decision-making process.


Towards demarchy

Between the few experiments with policy juries and planning cells and Burnheim's vision of demarchy is an enormous gulf. What strategy should be used to move towards demarchy?

Burnheim has some ideas. He thinks that as various government bodies become discredited, they may be willing to switch to demarchic management in order to maintain community legitimacy. For example, a health service might be wracked by disputes over salaries, conflict over provision of high technology medicine or community support services, severe budgetary crises and claims of mismanagement and corruption. This wouldn't be unusual. In this crisis situation, management by randomly selected groups might be seen by state managers as a way of resolving or offloading conflict and relegitimating the health service.

These and other similar scenarios may sound plausible, but they really provide little guidance for action. After all, there are plenty of unpopular, discredited and corrupt institutions in society, but this has seldom led to significant changes in the method of social decision-making. More specifically, how should demarchy be promoted in these situations? By lobbying state managers? By raising the idea among the general population? One thing is clear. The idea of demarchy must become much more well known before there is the slightest chance of implementation.

The experimentation with policy juries and planning cells is vital in gaining experience and spreading the idea of participation through random selection. The limitation of these approaches is that they are not linked to major social groups which would be able to mobilise people to work for the alternative.

Amongst the 'major social groups' in society, quite a number are likely to be hostile to demarchy. This includes most of the powerful groups, such as governments, corporate managements, trade union leaders, political parties, militaries, professions, etc. Genuine popular participation, after all, threatens the prerogatives of elites.

In my opinion, the most promising source of support is social movements: peace activists, feminists, environmentalists, etc. Groups such as these have an interest in wider participation, which is more likely to promote their goals than the present power elites. Social movement groups can try to put demarchy on the agenda by the use of study groups, lobbying, leafletting and grassroots organising.

Demarchy, though, should not be seen only as a policy issue, as a measure to be implemented in the community as a result of grassroots pressure. Demarchy can also be used by social movements as a means. In other words, they can use it for their own decision making.

This may not sound like much of a difficulty. After all, many social action groups already use consensus either formally or de facto. Also, the system of delegates is quite common. It would not seem a great shift to use random selection for decision-making at scales where direct consensus becomes difficult to manage.

Unfortunately, matters in many social movements are hardly this ordered. In many cases, formal bureaucratic systems have developed, especially in the large national organisations, and there are quite a number of experienced and sometimes charismatic individuals in powerful positions. These individuals are possibly as unlikely as any politician to support conversion to a different system of decision-making. (This itself is probably as good a recommendation for random selection as could be obtained. Any proposal that threatens elites in alternative as well as mainstream organisations must have something going for it.)

Nevertheless, social movements must be one of the more promising places to promote demarchy. If they can actually begin to try out the methods, they can become much more effective advocates. Furthermore, the full vision of demarchy, without the state or bureaucracy, stands a better chance within nonbureaucratised social movements than amidst the ruins of bungled government enterprises.

One of the most promising areas for promoting demarchy is in industry [33]. Workers are confronted by powerful hierarchical systems on every side: corporate management, governments and trade union bosses. There is plenty of experience in cooperative decision-making at the shop floor level; difficulties arise at higher levels of decision-making. It is here that random selection presents itself as a real alternative. Works councils, composed of both workers and managers selected randomly to serve a short period, provide a basis for communication and coordination. This approach overcomes the defects of all forms of representation. Workers' representatives on boards of management have served to coopt workers, while representatives in the form of trade union delegates have often become separated from the shop floor. Demarchic groups provide a way to maintain shop floor involvement in large enterprises.

The key point here is that demarchy should not be treated as a policy alternative, to be implemented from the top, but rather as a method of action itself. The ends should be incorporated in the means. It is quite appropriate that groups promoting demarchy use its techniques.

Needless to say, the future of demarchy cannot be mapped out. It is stimulating to speculate about solutions to anticipated problems; Burnheim's general formulations are immensely valuable in providing a vision. But as democracy by lot is tested, promoted, tried out, enjoys successes and suffers failures, it will be revised and refined. That is to be expected.

The message is that the process of developing and trying out alternatives is essential for all those seeking a more participative society. True enough, some worthy reforms can be achieved through the old channels of electoral politics, but that is no excuse for neglecting the task of investigating new structures. Demarchy is one such alternative, and deserves attention.

Demarchy is unlikely to be the final word in participative politics. No doubt it has flaws. But it is certain that present electoral methods provide no final solution.

Electoral methods -- that brings me back to the greens. They may be one of the most exciting political developments in decades, but in entering electoral politics they may have limited their potential for bringing about radical change. Ironically, it is the popular, charismatic green politicians who provide least threat to established power structures. Their electoral success will ensure continuing reliance on the old system of politics.



Valuable comments on a draft were received from Bob James and Ralph Summy.



1 These points are well made by Benjamin Ginsberg, The Consequences of Consent: Elections, Citizen Control and Popular Acquiescence (Reading, MA: Addison-Wesley, 1982).

2 Thomas S. Martin, "Unhinging all government: the defects of political representation," Our Generation, Vol. 20, No. 1, Fall 1988, pp. 1-21; Edmund S. Morgan, Inventing the People: The Rise of Popular Sovereignty in England and America (New York: Norton, 1988).

3 Alex Comfort, Authority and Deliquency in the Modern State: A Criminological Approach to the Problem of Power (London: Routledge and Kegan Paul, 1950); David Kipnis, Technology and Power (New York: Springer Verlag, 1990); Pitirim A. Sorokin and Walter A. Lunden, Power and Morality: Who Shall Guard the Guardians? (Boston: Porter Sargent, 1959).

4 Andrew Wernick, Promotional Culture: Advertising, Ideology and Symbolic Expression (London: Sage, 1991), chapter 6.

5 Ralph Miliband, The State in Capitalist Society (London: Weidenfeld and Nicolson, 1969).

6 Carl Boggs, Social Movements and Political Power: Emerging Forms of Radicalism in the West (Philadelphia: Temple University Press, 1986).

7 Brian Martin,"Environmentalism and electoralism," Ecologist, Vol. 14, No. 3, 1984, pp. 110-118.

8 See Timothy Doyle, "Environmental movement power brokers," Philosophy and Social Action, Vol. 16, No. 3, July-September 1990, pp. 37-52.

9 Ginsberg, The Consequences of Consent, op. cit. See also Benjamin Ginsberg, The Captive Public: How Mass Opinion Promotes State Power (New York: Basic Books, 1986).

11 See for example George H. Smith, "The ethics of voting" The Voluntaryist, Vol. 1, No. 1, October 1982, pp. 1, 3-5; Vol. 1, No. 2, December 1982, pp. 3-6; Vol. 1, No. 4, April 1983, pp. 3-7. The classical anarchists are more polemical or didactic. See P. Kropotkin, "Representative government," The Commonweal, Vol. 7, 1892, weekly instalments, 7 May - 9 July; Errico Malatesta, Vote: What For? (London: Freedom Press, 1942).

12 London Edinburgh Weekend Return Group, In and Against the State (London: Pluto, 1980).

13 A mainstream treatment is Benjamin R. Barber, Strong Democracy: Participatory Politics for a New Age (Berkeley: University of California Press, 1984).

14 F. Christopher Arterton, Teledemocracy: Can Technology Protect Democracy? (Newbury Park, CA: Sage, 1987); Christa Daryl Slaton, Televote: Expanding Citizen Participation in the Quantum Age (New York: Praeger, 1992).

15 Michael Avery, Brian Auvine, Barbara Streibel and Lonnie Weiss, Building United Judgment: A Handbook for Consensus Decision Making (Madison, WI: Center for Conflict Resolution, 1981); Virginia Coover, Ellen Deacon, Charles Esser and Christopher Moore, Resource Manual for a Living Revolution (Philadelphia: New Society Publishers, 1981).

16 Jane J. Mansbridge, Beyond Adversary Democracy (New York: Basic Books, 1980).

17 See for example Murray Bookchin, "What is communalism? The democratic dimension of anarchism," Green Perspectives, No. 31, October 1994, pp. 1-6; Charles Landry, David Morley, Russell Southwood and Patrick Wright, What a Way to Run a Railroad: An Analysis of Radical Failure (London: Comedia, 1985); Howard Ryan, Blocking Progress: Consensus Decision Making in the Anti-nuclear Movement (Berkeley: Overthrow Cluster, Livermore Action Group, 1985).

18 The case is given by Leopold Kohr, The Breakdown of Nations (London: Routledge and Kegan Paul, 1957) and Kirkpatrick Sale, Human Scale (New York: Coward, McCann and Geoghegan, 1980).

19 Frances Kendall and Leon Louw, After Apartheid: The Solution for South Africa (San Francisco: ICS Press, 1987).

20 Publications available from John Zube, 7 Oxley Street, Berrima NSW 2577, Australia

21 John Burnheim, Is Democracy Possible? The Alternative to Electoral Politics (London: Polity Press, 1985). For a short outline, see John Burnheim, "Democracy by statistical representation," Social Alternatives, Vol. 8, No. 4, January 1990, pp. 25-28. On decisions on a world scale, see John Burnheim, "Democracy, nation states and the world system," in David Held and Christopher Pollitt (eds.), New Forms of Democracy (London: Sage, 1986), pp. 218-239.

22 Richard G. Mulgan, "Lot as a democratic device of selection," Review of Politics Vol. 46, No. 4, October 1984, pp. 539-560.

23 Gary J. Jacobsohn, "Citizen participation in policy-making: the role of the jury," Journal of Politics, Vol. 39, No. 1, February 1977, pp. 73-96.

24 Valerie P. Hans and Neil Vidmar, Judging the Jury (New York: Plenum, 1986); Harry Kalven, Jr. and Hans Zeisel, The American Jury (Chicago: University of Chicago Press, 1971); Rita James Simon (ed.), The Jury System in America: A Critical Overview (Beverly Hills, CA: Sage, 1975).

25 Adam Roberts, Nations in Arms: The Theory and Practice of Territorial Defence (London: Chatto and Windus, 1976).

26 Anders Boserup and Andrew Mack, War Without Weapons: Non-violence in National Defence (London: Frances Pinter, 1974); Adam Roberts (ed.), The Strategy of Civilian Defence: Non-violent Resistance to Aggression (London: Faber and Faber, 1967); Gene Sharp, Making Europe Unconquerable: The Potential of Civilian-based Deterrence and Defense (Cambridge, MA: Ballinger, 1985).

27 Brian Martin, Social Defence, Social Change (London: Freedom Press, 1993), chaper 13.

28 Ernest Callenbach and Michael Phillips, A Citizen Legislature (Berkeley: Banyan Tree Books, 1985). See Slaton, op. cit., p. 98, for references to the work of a number of advocates of random selection, including Theodore L. Becker, Robert A. Dahl, James A. Dator, Dennis C. Mueller et al., and Alvin Toffler.

29 Ned Crosby, Janet M. Kelly and Paul Schaefer, "Citizen panels: a new approach to citizen participation," Public Administration Review, Vol. 46, March-April 1986, pp. 170-178; Ned Crosby, "The peace movement and new democratic processes," Social Alternatives, Vol. 8, No. 4, January 1990, pp. 33-37.

30 Jefferson Center, Final Report: Policy Jury on School-based Clinics (Minneapolis: Jefferson Center, 1988).

31 Jefferson Center, 903 Plymouth Building, 12 S. 6th Street, Minneapolis MN 55402, USA. [New address, from April 1997: Suite 405, 3100 West Lake Street, Minneapolis MN 55416-4599, USA, phone +1-612-926 3292]

32 P. C. Dienel, Die Planungszelle: Eine Alternative zur Establishment-Demokratie (Opladen: Westdeutscher Verlag, 1978; second edition, 1988). See, in English, Peter C. Dienel, "Contributing to social decision methodology: citizen reports on technological projects," in Charles Vlek and George Cvetkovich (eds.), Social Decision Methodology for Technological Projects (Dordrecht: Kluwer Academic Publishers, 1989), pp. 133-151; Detlef Garbe, "Planning cell and citizen report: a report on German experiences with new participation instruments," European Journal of Political Research, Vol. 14, 1986, pp. 221-236; O. Renn et al., "An empirical investigation of citizens' preferences among four energy scenarios," Technological Forecasting and Social Change, Vol. 26, 1984, pp. 11-46.

33 Crucial work on this has been done by Fred Emery. See F. E. Emery, "Adaptive systems for our future governance," in F. E. Emery (ed.), Systems Thinking, Vol. 2 (Harmondsworth: Penguin, 1981); F. E. Emery, Toward Real Democracy and Toward Real Democracy: Further Problems (Toronto: Ontario Ministry of Labour, 1989); Merrelyn Emery (ed.), Participative Design for Participative Democracy (Canberra: Centre for Continuing Education, Australian National University, 1989).