In developing my ideas about the relations of polyarchy, pluralism, and scale, I have incurred special debts to a number of close collaborators, but to none do I owe more than to Stein Rokkan. I first met Stein Rokkan at an international meeting in Rome in 1959 or thereabouts. Like virtually everyone who knew him I was at once impressed by range and depth of his knowledge, and remain so to the present day. As acquaintance ripened into a deeply treasured friendship I also came to appreciate, as once again every one who knew him must also have appreciated, his exceptional human qualities. He was a rare scholar, a rare: human being, and a rare friend, whose presence immeasurably enriched our lives and whose loss has left all who knew him, and many others who did not, immeasurably poorer.

Within a few years after Rokkan and I first met, I undertook to edit a collaborative work on political oppositions in western Europe, to which Rokkan contributed not only his extraordinary knowledge, his seemingly unbounded energies, and his unflappable good humor but also an essay on corporate pluralism that was I believe the first scholarly treatment of this important and now widely discussed subject and truly was of seminal influence. During the course of that collaboration,, Rokkan took the leadership in a project on the smaller European democracies, a dozen countries whose experience had until then been largely neglected in political science. Although the volumes we projected on each of the smaller European democracies did not materialize, it is not too much to say, I think, that the project, and Rokkan's unflagging efforts in others, contributed notably to the rapid growth in scholarly work on the political systems of the smaller European countries.¹ It was during this collaboration that, on Rokkan's strong encouragement, Edward Tufte and I undertook to explore briefly the topic of size and democracy.2

* The Rokkan Memorial Lecture, Bergen, May 16, 1984.

Consequences of the Historic Shift in Scale

As we pointed out, the shift in the practical and theoretical locus of democracy from the small city-state to the large and even giant nation state brought with it important consequences, both practical and theoretical, although theory failed to keep pace with practice. By the end of the eighteenth century the city-state, which for over two millenia had been looked on as the natural and even exclusive setting for a democratic order, a view still maintained by Rousseau in the Social Contract (1762), had become almost everywhere so subordinate to the nation-state that democratic efforts, ideas, and ideology inevitably shifted their focus to the problem of democratizing the government of the nation-state. The consequences of the shift in focus, however, were not clearly foreseen. I want to mention seven important consequences.

1. Representation. Because of the practical impossibility of having an assembly consisting of all citizens, or even a significant proportion of them, representation, which was anathema to Rousseau in the Social Contract³ became an unavoidable consequence of the enlarged scale of the political system.

2. Unlimited Extension. Once representation was accepted as a solution, the barriers to the size of a democratic unit set by the limits of an assembly in a democratic city state were eliminated and representative democracy could be extended virtually without limit.

3. Limits on Participation. As a direct consequence of increasing size, however, some forms of political participation necessarily became more limited. Just as a substantial proportion of citizens in a nation state cannot possibly discuss political matters directly with one another, so too only a comparatively small percentage can possibly engage in discussions with their representatives. Even if spatial barriers to communication can in principle be eliminated by electronic means, the limits set by time are inexorable. You can easily see how drastic these limits are by a simple arithmetic exercise. You need only to multiply the number of messages a higly participatory process could reasonably be expected to produce, by the average time you assume a meaningful political message requires.

4. Diversity. Although the relation between scale and diversity is less clearcut, as a political unit increases in size, its inhabitants will tend to exhibit greater diversity, in ways relevant to political life: local and regional, ethnic, racial, religious, ideological, occupational, and so on. The relatively homogeneous population of citizens united by common attachments to city, language, race, history, myth, and religion that was so conspicuous a part of the classical, city-state vision of democracy now becomes for all practical purposes impossible.

5. Conflict. As a consequence, political cleavages are multiplied, political

conflict is an inevitable aspect of political life, and political thought and practices tend to accept conflict as a normal and not aberrant feature of politics.. In contradiction to the classical vision, in which a relatively homogeneous body of citizens could be expected to share essentially the same beliefs about the common good, and to act on those beliefs, the notion of the common good is stretched much more thinly in order to encompass the heterogeneous attachments, loyalties, and beliefs formed among a body of diverse citizens with a multiplicity of cleavages and conflicts. A striking symbol of the change is James Madison, who at the Arnerican constitutional convention in 1787, and later in his defense of it in The Federalist, met head-on the historical view, which was still reflected in anti-Federalist objections to the absurdity and iniquity of the attempt to form a democratic republic on such a grotesque scale as the United States. In a brilliant polemic, Madison contended that because conflicts of interest were in the nature of man and society, and the expression of these conflicts could not be suppressed without suppressing freedom, the best cure for the mischiefs of faction was to enlarge the scale. As he of course intended, it followed that, contrary to the traditional view, a positive advantage of republican government in the nation state was that political conflicts would be much less likely to produce acute civil strife than in the tighter compass of the city-state.

The sixth and seventh consequences of the shift in the locus of democracy
from the city state to the nation state, from small-scale to large scale democracy,
were polyarchy and organizational pluralism, to which I now turn.

Polyarchy

Origin of the Term

Since the term polyarchy does not appear to have a standard meaning, and I myself have doubtless contributed to the confusion of usage, let me say something about its origins. To the best of my knowledge the word was first introduced in modern political science by Lindblom and me in Politics,, Economics, and Welfare in 1953, where we referred to it as a 'process'.4

Considering it as a process was in keeping with the theoretical orientation of the book, the subtitle of which was: 'Planning and Politico-Economic Systems Resolved into Basic Social Processes'. In Part IV we described 'Four Central Sociopolitical Processes': the price system, or control of and by leaders; hierarchy, or control by leaders; polyarchy, or control of leaders; and bargaining, or control among leaders. 'In some societies', we asserted

the democratic goal is still roughly and crudely approximated, in the sense that non-leaders
exercise a relatively high degree of control over leaders. The constellation of social processes
that makes this possible we call polyarchy.

What we were searching for was a distinction between two sometimes confusing usages of the terms 'democracy': one to describe a goal or ideal, an end perhaps never achieved and possibly not even fully achievable in actuality, and the other to describe the distinguishing features of the actual political systems commonly called 'democratic' or 'democracies' in the modern world. According to the authoritative Oxford English Dictionary, for which the portions on the letter p had been completed in 1909, polyarchy is 'the government of a state or city by many: contrasted with monarchy'. The word had fallen into disuse, and seemed to us nicely suited to our needs.⁵

We also set out some six 'criteria' that were meant to give 'operational
significance' to the expression 'a high degree of control'. The first, for example,
was that

Most adults in the organization have the opportunity to vote in elections with no significant
rewards and penalties directly attached either to the act of voting or to the choice among
candidates.

Although the six 'criteria', as we called them there, have altered somewhat in my own writing, the later 'criteria', now seven, are little more than a refinement of the earlier ones. However, later I came to believe that to think of polyarchy as a process is less useful than considering it as a set of institutions.^

Five Interpretations

Like democracy, polyarchy can be viewed from several different perspectives.

As a type of regime

To begin with, polyarchy can be seen as simply a distinctive kind of regime for governing the modern state — a regime with characteristics that distinguish it pretty sharply from all regimes prior to the nineteenth century, and also from most regimes existing among the nation-states of the world today.

Its distinctiveness arises from the combination of two general features: its relatively high tolerance for oppositions — those who oppose the conduct of the government — and the relatively widespread opportunities for participating in influencing the conduct of the government, including removal of incumbent governing officials by peaceful means. More specifically, polyarchies can be distinguished from other regimes by the presence in a realistic sense of seven institutions: a widespread and nowadays nearly universal suffrage; suffrage coextensive with the right to run for public office; fairly conducted elections accompanied by negligible or no coercion; extensive protection of free expression, including criticism of the government, the regime, society, the dominant ideology, and so on; the existence of alternative and often competing sources ofinformation and persuasion not under the control of the government;

a high degree of freedom to form relatively autonomous organizations of great
variety, including, most crucially, opposition political parties; and relatively high
responsiveness of the government to voters and election outcomes.

It is this set of institutions, taken together, that make a polyarchal regime so markedly different from the centralized or feudal monarchies, the Roman Empire, the historical regimes of China and Japan, and even the regimes of the democratic city states of classical Greece, the Roman Republic, or the Italian city-state republics. In the contemporary world, countries in which these seven institutions are, in a realistic sense, present to a high degree, number around 40 out of approximately 160 nominally independent countries.

As a product of democratizing nation-states

Polyarchy can also be understood historically, or developmental^, as a set of institutions that evolved in large part, though not exclusively, as a product of efforts to democratize and liberalize the political institutions of nation states. In this perspective, polyarchy is a unique, historically conditioned set of modern institutions — in particular the complex of institutions I have just mentioned — resulting primarily from attempts since the eighteenth century to adapt democratic ideas and practices to the large scale of the modern nation state. This historically unique complex of political institutions has tended to acquire the name 'democracy', and its institutions have largely superseded the distinctive political institutions of the earlier democratic or republican city-states. In democratic Athens, for example, the citizen assembly was of primary importance whereas organized political parties were unknown, as were many of the other autonomous interest organizations common in polyarchies. I suspect that an Athenian democrat would be totally bewildered by the political institutions of polyarchy and would reject the notion that they were entitled to be called democratic.

As necessary to the democratic process

Thirdly, polyarchy can be understood as a set of political institutions necessary in order to provide a satisfactory approximation to the democratic process when the objective is to apply that process on a large scale, for example, on the scale of the nation state. Viewed in this perspective, our democratic predecessors were not fools: they knew what they were doing in insisting on the suffrage, the right to run for public office, free and fair elections, the right to form political parties, the responsibility of the executive to parliament or the electorate, and so on.

To say that polyarchal institutions are necessary to the democratic process
on a large scale is not to saiy, of course, that they are also sufficient, and I
suppose that few of us believe them to be so.

As a system of control by competition

Fourthly, polyarchy can be understood as a system of political control in which, as a consequence of the set of institutions mentioned earlier, the highest officials in the government of the state face the prospect of their own displacement by means of popular elections, and hence tend to have strong incentives to modify their conduct in such a way as to win elections in political competition with other candidates, parties, and groups. From this perspective, which is of course very close to that of Schumpeter, the most distinctive feature of polyarchy is the open competition among political elites for office. This competition helps to create a certain measure of mutual influence between elites and masses, rather than the unilateral dominance by elites that Michels' iron law of oligarchy would lead one to expect.

As a system of rights

Finally, polyarchy can be interpreted as a system of rights in which certain rights are institutionally guaranteed and protected. Each of the seven institutions of polyarchy prescribes certain rights that are necessary to the existence and functioning of the institution itself. This is self-evidently so with suffrage or freedom of expression. To institutionalize free speech, for example, citizens must possess a legally enforceable claim, an entitlement, a legal right to speak freely on political matters, and it must be an obligation of officials of the state to uphold that claim, if need be by punishing violators. It is obvious, too, that in order for the institution to exist, the right cannot be merely abstract or theoretical, like most of the political rights in the Soviet Constitution. The right must be actually enforceable in courts of law. Although more complex, the other institutions must also necessarily generate a body of enforceable rights; lacking these actually enforceable rights the institution cannot be said to exist in realistic sense.

To one who believes that polyarchy is desirable, the political rights it entails might be valued simply because they are necessary to the institutions of polyarchy. But one might also value a right entailed in polyarchy because that right is thought to be good in itself, or instrumentally necessary to freedom and equality, or to the democratic process. For example, whereas a right to organize a political party to oppose the government might be valued primarily as necessary to polyarchy, a right to freedom of expression might be valued as also good in itself, or necessary to personal freedom.

No doubt polyarchy can be interpreted in still other ways. A Marxist, for example, might interpret it simply as 'bourgeois democracy'. But the point I wish to emphasize here is that the five ways of thinking about polyarchy that I have just described are not inconsistent with one another. On the contrary, they complement one another. They simply emphasize different aspects or consequences of the institutions that serve to distinguish polyarchal from nonpolyarchal regimes.

Pluralism

Origins

Like polyarchy, pluralism., at least as we tend to use the term nowadays, is something of a neologism in political science. It is interesting to take note of the definitions supplied by the Oxford English Dictionary. The section of Volume VII containing the word pluralism was originally published in 1907. Pluralism, we learn from that magisterial source, means

The character of being plural; the condition or fact of being a pluralist. La. Eccl. The system
or practice of more than one benefice being held at the same time by one person, b. The
holding of two or more offices of any kind at one time. ... 2. Philos. A theory or system
of thought which recognizes more than one ultimate principle: opposed to MONISM.

A pluralist, then, is

1. Eccl. One who holds two or more benefices at the same time. ... In extended use, One who
combines two or more offices, professions, or conditions. ... 2. (Philos.) One who holds
the theory of pluralism.

The eccelesiastical usage was old, the earliest reference being 1362. The dispute over the propriety of pluralism in the Church of England had flared up again in the nineteenth century. I recall encountering the term, and implications of that dispute, in one of Trollope's novels, I think Barchester Towers.

What is interesting about all this is the absence of any reference to the meaning, or meanings, attributed to the term by contemporary political scientists and in the last decade increasingly by journalists, politicians, and ideologues. The explanation is simple: The definitions in the O.E.D. were completed about a decade before the appearance of Laski's attack on state sovereignty, in which he explicitly posited pluralism as an alternative to the dominant monistic view.₇ Laski's attack on monism, and his preference for a pluralist interpretation of the state, was not, of course, entirely original. It had antecedents in the works of the French jurist, Leon Duguit, with whose writings Laski was familiar, as well as those by Laski's fellow countrymen, J. N. Figgis and F. W. Maitland, and the still earlier writings of the German jurist Otto Gierke. 'Pluralist' ideas about state and society were also advanced by Guild Socialists. While the idea of Guild Socialism had been formulated by A.J. Penty as early as 1906, it was J.H. Hobson and particularly G.D.H. Cole, writing about the same time as Laski, who perhaps had the greatest impact. Pluralist notions about the state and society were quite prominent in the early 19205, not only in Britain but also in the United States, where several well-known American political scientists analyzed them in considerable detail in the pages of their professional journals.9 After a decade of attention, however, interest in pluralism rapidly waned on both sides of the Atlantic.

Nonetheless, the term, and the essential ideas behind it, maintained a certain currency.lo It was readily available, for example, when Lindblom and I were writing Politics, Economics, and Welfare from 1950 to 1952. We were familiar with the British and European ideas about pluralism as well as the term itself,¹¹ and made use of the word in order to state what we thought to be a requirement of polyarchy:l2

Polyarchy requires a considerable degree of social pluralism — that is, a diversity of social
organization with a large measure of autonomy with respect to one another.

Later, however, the concept took on a life of its own. 'Pluralist theory' came to designate a strange melange of ideas. In fact, a good deal of the 'theory' consisted of interpretations by hostile critics who sometimes constructed a compound of straw men and inferences from the work of assorted writers who by no means held the same views. Frequently the result was a 'theory' that probably no competent political theorist — pluralist or not — would find plausible.l3

Nevertheless, a concept along the lines set out in Politics, Economics, and Welfare retains great utility. Whatever word we may prefer, some such concept appears to be necessary in order to describe countries governed by polyarchal regimes, and thus to grasp one of the important consequences of the change in scale from city-state to nation-state.

The Change in Perspective

Just as the development of polyarchy represented a new way of thinking about democratic institutions, so too the gradual acceptance of pluralism as an inherent, inescapable, and even desirable aspect of democracy meant a radical break with classical democratic ideas. Just as the dominant assumption for nearly two millenia had been that the proper location for a democracy was a small and compact unit like the ideal city-state, so too the prevalent view had been that the citizen body should be fairly homogeneous — in race, ethnicity, religion, language, status, wealth, and knowledge. It was understood, of course, that some functional specialization was necessary. But the notion that citizens might worship different gods, or speak different languages, or retain diverse ethnic affiliations, or differ markedly in any other ways that would tend to generate a diversity of conflicting interests, was anathema. In further pursuit of the ideal of a common good, and therefore to avoid the diversities that would tend to prevent citizens from perceiving or having common interests, in the dominant view and in the practices of city-state republics and democracies, little sympathy and much hostility existed for any notion that subgroups of citizens might properly pursue their special interests in politics by organizing themselves into an explicitly political association. Of course, as Aristotle

recognized and, centuries later, was reaffirmed in the conception of the vivere civile advanced by the Italian civic humanists, citizens would be members of various associations with special purposes, such as families, or economic organizations like the guilds. But the particular purposes of these associations need not conflict with the purposes of others or the common good; in fact, by fulfilling their purposes they would contribute to the common g00d.14 In this view, however, the only proper association in political life was the city itself. To be sure, actual life did not always correspond to the ideal. In practice, factions were often rampant and destructive, particularly in the Italian citystates. But the idea that citizens might properly organize themselves, for example., in the kind of competitive and conflict association we call political parties was completely alien.

The Legitimation of Organizational Pluralism

With the change of scale that accompanied the shift of locus from city-state to nation-state, organizational pluralism became not merely inevitable but also gained legitimacy, not only in social and economic life but also in political life. The change is dramatically marked by the difference between Rousseau and Tocqueville. Rousseau, here following the older tradition, finds associations more or less inevitable, but troublesome and even dangerous.ls In a remarkable passage in Political Economy, he wrote:

All political societies are composed of other, smaller societies of different types, each of which has its interests and maxims. But these societies that everyone perceives, because they have an external, authorized from, are not the only ones that really exist in the State. All the private individuals united by a common interest constitute as many others, permanent or temporary, whose force is no less real for being less apparent, and whose various relationships, well observed, are the true knowledge of mores. It is all these tacit or formal associations which modify in so many ways the appearance of the public will by the influence of their own. The will of these particular societies always has two relations: for the members of the association it is a general will: for the large society, it is a private will, which is very often found to be upright in the first respect and vicious in the latter. ... A given deliberation can be advantageous to the small community and pernicious to the large one.

Seventy years later, Tocqueville, who knew Rousseau's views well, took a contrary position. Although he was not unmindful of the dangers of associations ¹⁶ contemplating democracy on the scale of the United States, whose seeming vastness even then dwarfed the Geneva that so much influenced Rousseau's thought, and greatly preoccupied with the danger of majority tyranny that he believed was inherent in a situation of equality like that of the Americans, Tocqueville concluded that

At the present time, the liberty of association is become a necessary guarantee against the
tyranny of the majority. ... There are no countries in which associations are more needed,
to prevent the despotism of faction or the arbitrary power of a prince, than those which are
democratically constituted.1717

Several years later in the second volume of Democracy in the United States,
Tocqueville returned to the theme of associations, extending now his purview
to include civil as well as political associations.

If men are to remain civilized, or to become so, that art of associating together must grow
and improve, in the same ratio in which the equality of conditions is increased.18

Pluralism and Polyarchy

The advantage of another century and a half of experience helps us to understand more fully the relationship between pluralism and democracy. A monistic view like Rousseau's in the Social Contract arguably might be applicable to democracy on the smaller scale of a city state with a predominantly commercial or agricultural economy. In a small and intimate democratic association the presence of other associations that compete for loyalty and support, weaken cohesion and consensus, and stimulate conflict may be less desirable, or as Rousseau contended, downright undesirable, and to be avoided so far as possible. In these conditions, possibly the adversarial institutions and practices of large scale democracy may be unnecessary and undesirable.¹⁹ However that may be, whenever an effort has been made to realize the democratic idea on the scale of the nation state and the institutions of polyarchy have come to exist, relatively independent associations and organizations of considerable variety and number have also developed. Following Tocqueville we might describe them as both political and civil, but the distinction is far from sharp since as we all know a civil association may also play a political role.

It is not of course an historical accident that polyarchy and organizational pluralism everywhere coexist. For while organizational pluralism may not be a sufficient condition for polyarchy, the institutions of polyarchy are everywhere sufficient to insure that organizations and associations of considerable independence, variety, and number will play an important role in the political life of a country. The advantages of organized cooperation make organizations desirable. Indeed, the existence of relatively autonomous political organizations is necessary to the practice of democracy on a large scale. Finally, the rights necessary to the existence of polyarchy make independent organizations legally possible. That they are desirable and possible makes them inevitable. It takes only the first feeble flickerings of freedom for independent organizations to spring to life when the controls of an authoritarian regime are relaxed: witness Italy, Austria, Germany, and Japan after their regimes collapsed at the end of World War 11, Czechoslovakia in 1968, Poland during the rise of Solidarity, Argentina after the Falklands. Independent organizations can be suppressed only by suppressing the institutions of polyarchy. Just as it is no accident that pluralism and polyarchy go together, so it is not accidental that

among the first acts following a seizure of power by authoritarian leaders is
the suppression of autonomous political organizations: witness Chile and
Uruguay in 1973.

Yet this strict connection, the relation between polyarchy and pluralism is not simple, for at least two reasons. First, it would be a profound mistake to assume that organizational life tends to be alike in all democratic countries. Organizational pluralism is an important feature of political life in both Norway and the United States, for example; but the particular constellation of organizations in Norway is significantly different from that of the United States, and the consequences for political life are different. One has only to look at party systems and trade unions in the two countries to see how different they are, and how different in some ways are their consequences. As we are all aware, party systems vary enormously even among European countries. Let me add in passing that no one contributed more than Stein Rokkan to our knowledge of these differences and our understanding of how they came about.

These differences in the morphology of organizational life bear on the second factor that complicates the relation between pluralism and polyarchy. If pluralism is necessary, inevitable, and desirable in a system of polyarchy, it may also have undesirable consequences. For example, if some interests have ready access to organizations and their resources, and others do not, then such a pattern may help to maintain inequalities among citizens, and some of these inequalities might be unjust. Or consider Rousseau's concern. Associations may do more than simply protect or advance the interests of their members. They may also sharpen and exaggerate particular aspects of a citizen's interests as against other, perhaps broader attachments, loyalties, concerns, and interests, and in this way help to produce and maintain a deformed civic consciousness. When organizational pluralism helps to produce the consequences I have just mentioned, it may also distort the public agenda by focussing the political process on alternatives that promise visible short-run benefits to a small minority of well organized citizens rather than on alternatives that would provide significant long-run benefits to to a larger number of unorganized citizens.

Finally, relatively autonomous organizations — or, more commonly, coalitions of organizations — may take on what are essentially public functions. However distasteful this must be to an advocate of monistic democracy like Rousseau, in itself it need not be particularly alarming. But it does alert us to some possibilities that must give concern even to one who accepts the inherent relation of organizational pluralism to large scale democracy. One such possibility is that control over some important public matters will be transferred to organizations that are not themselves controllable, as a practical matter, by the demos and its representatives in parliament and the executive.

This transfer of control then becomes much more than mere delegation of power by the demos to an organization that it controls, as at least formally is the case when decisions are delegated by parliament to a bureaucratic agency of government. If in practice the demos cannot exercise adequate control, then what occurs is not delegation but alienation of authority.

That something of this kind may indeed have taken place in many countries accounts at least in part for the numerous attempts in recent years to understand the empirical and normative aspects of what has been variously called corporatism, democratic corporatism, corporate pluralism, and so on. The seminal work, published nearly twenty years ago, was an essay by Rokkan on Norway, the thesis of which may be summed up in his own words: 'Numerical Democracy and Corporate Pluralism: Votes Count but Resources Decide'. The Cabinet, he wrote, 'stands at the top of the electoral hierarchy but it is only one of four corporate units at the bargaining table'. The other three he had in mind were, of course, labor, business, and farmers. He went on to say:

The Cabinet has increasingly had to take on the role of mediator between the conflicting interests in the national community. At least in matters of internal policy it can rarely if ever force through decisions solely on the basic of its electoral power but has to temper its policies in complex consultations and bargains with the major interest organizations.2o

That systems combining numerical democracy with corporate pluralism have great advantages seems to me undeniable; but it is also undeniable that they raise perplexing problems for democratic theory and institutions. It is not difficult to justify corporate pluralism, at least of the kind Rokkan described for Norway, on pragmatic and utilitarian grounds. Yet to the extent that it allows control over crucial public matters to be alienated, it would seem to violate democratic criteria. It may be that our understanding of democracy will have to adapt somehow to practice, but at present no satisfactory reformulation seems to have been found that would give democratic legitimacy to corporatism.

Yet corporatism in the Scandinavian manner is only one manifestation of the problem. The United States, for example, lacks the rather centralized national organizations of unions, business, and farmers that make the structure of democratic corporatism possible in Sweden, Norway, the Netherlands, and Austria. There the problem appears in a very different form, in, for example, the famous 'iron triangles' of Congressional committees, bureaucracies, and interest organizations that exercise great influence over policy-making. The national bargaining that takes place in corporate systems has more dramatically decisive consequences but the steady working of the iron triangles may be no less decisive in the long run — and more concealed from public view.

Conclusion

The shift in the scale of democracy from city-state to nation-state has, then, led inexorably away from monistic toward pluralist democracy. The transformation of practices and institutions brought about by the enormous change of scale is dramatic and far-reaching. Democratic theory, too, has altered. The Social Contract required a state and a society on a scale that inexorable changes had made impossible. In that sense, it was a reactionary vision, which is why it has remained so revolutionary in its implications. Writing only seventy years later Tocqueville has already accepted the idea that the large nation state, not the small city state, is the appropriate locus for democracy in the modern world, and another generation later John Stuart Mill takes it utterly for granted.

Yet democratic theory, originally formulated as a vision of a small scale system of politics and society, is by no means wholly at ease with large scale democracy. The monistic bias of the classical democratic vision clashes with the pluralistic actuality of large-scale democracy, and theory, straddling both, often seems inadequate both descriptively and normatively.

In order to take the inherent connection between polyarchy and pluralism more fully into account, it might be useful to suggest yet a sixth interpretation of polyarchy to add to the five I described earlier. In this perspective, polyarchy is a kind of regime for governing nation-states in which power and authority over public matters are distributed among a plurality of organizations and associations that are relatively autonomous in relation to one another and in many cases in relation to the government of the state as well. These relatively autonomous units include rot only organizations that are, legally and sometimes constitutionally, components of the government of the state but also organizations that legally are — to use a term that in this connection may often seem singularly inapt — 'private': that is legally, and to an important extent realistically , they are independent, or mainly independent, of the state.

Among other ways, polyarchy is distinguished from classical monistic democracy by the salience, power, and legitimacy of autonomous organizations in political life and decisions on public matters. Polyarchy is also distinguished from authoritarian regimes in two ways: (1) By the institutions of polyarchy which, by definition, no authoritarian regime completely possesses, and which provide much more scope to the democratic process than any authoritarian regime can provide, lacking as it does one or more of the crucial institutions necessary to, if not sufficient for, large-scale democracy; (2) By the scope of organizational pluralism, which sharply distinguishes polyarchy at one extreme from monistic authoritarian regimes — that is, totalitarian systems — and at the other extreme from authoritarian regimes of limited pluralism, to use Juan Linz's expression, where a plurality of relatively autonomous political parties, for example, does not exist.²¹

Compared against the ideal of monistic democracy — the dominant perspective from classical Athens to Rousseau — the power and authority of the organizational subsystems — subgovernments, they have sometimes been called — is nothing less than a shocking outrage to the democratic vision. Paradoxically, however — or perhaps not so paradoxically after all — in our world the theory and practice of monism is best exhibited in authoritarian regimes. If the most relevant and likely alternative to polyarchy in the modem world is not city-state democracy but an authoritarian regime, then even from a democratic perspective the untidy systems of polyarchy and pluralism begin to look much more charming. For if by comparison with the ideal monistic democracy the subsystems often are outrageously powerful, by comparison with the monism or limited pluralism of authoritarian regimes what is striking is the limits of their power. They are limited by the existence and legitimacy of the institutions of polyarchy, and by the existence and legitimacy of the system of rights necessary to those institutions.

That the power and authority of organizations are limited in this way is surely one reason why these relatively powerful organizations are permitted to retain as much autonomy as they do. A second reason, at least for some of them, is that they are, as we have seen, necessary to large-scale democracy. In addition, however, the claims made by the early legal pluralists, from Gierke to Laski and Cole, are today widely accepted: relatively autonomous associations are necessary to a good life and a decent sociopolitical order, and within limits which cannot be carefully specified their existence is as justifiable, morally and practically, as that of the state itself. Furthermore, the complex system of making decisions about public matters in which they participate is often seen to be advantageous on purely utilitarian grounds, in comparison, at least, with any alternative that appears to be available. Yet there is also a final reason: it is unlikely that any government in a polyarchy could effectively enforce a prohibition against either the autonomy of many important organizations or their authoritative participation in decision-making.

Nonetheless, I do not believe we have yet found an altogether satisfactory way of resolving the tensions, which exist both in theory and practice, between pluralism and democracy. The anomaly of democratic pluralism that Stein Rokkan remarked on with such insight nearly two decades ago still remains with us: votes count, but often organizational resources decide.

---

NOTES

1. The project did result in the distinguished work of Basil Chubb's, The Government and Politics of Ireland, Stanford: Stanford University Press, 1970.

2. Size and Democracy, Stanford: Stanford University Press, 1973.

3. However, Richard Fralin has shown, conclusively I think, that Rousseau's hostility to representation in the Social Contract was contrary to views he expressed elsewhere, both before and after that work. Rousseau and Representation, New York: Columbia University Press, 1978.

---

3. However, Richard Fralin has shown, conclusively I think, that Rousseau's hostility to representation in the Social Contract was contrary to views he expressed elsewhere, both before and after that work. Rousseau and Representation, New York: Columbia University Press, 1978.

4. Although we were unaware of it, Ernest Barker had used the term 'polyarchism' in 1913 or 1914 in an essay later published in Church, State, and Study (London: 1930). 1 owe this discovery to Susan Gross Solomon, ' "Pluralism" in Political science: The Odyssey of a Concept', in the volume edited by her, Pluralism in the Soviet Union, London: Macmillan, 1983, p. 15 and footnote 51. However, she in turn attributes the discovery to Claude Burtenshaw, 'The Political Theory of Pluralist Democracy', Western Political Quarterly, 21 (December 1968) 586.

5. Along with the adjective polyarchal or (rare, and avoided by Lindblom and me, then and later), polyarchical. The term polyarchist was also available for 'one who advocates or believes in polyarchy' — such as Lindblom or myself. The first usage listed is 1609: 'The Bees abhor as well Polyarchy as Anarchy', and the last, in 1890 by J.H. Stirling in the Gifford Lectures, is 'Polyarchy is anarchy'. Much later, Arendt Lijphart called my attention to its use by Althusius, who treated it as one type of 'supreme magistrate': 'a polyarchic supreme magistrate is one in which those who are furnished by the subjects with equal or the same supreme imperium rule and administer the rights of sovereignty. That is to say, the succession of administration is communicated among a number of persons. ... This polyarchic magistracy is either aristocratic or democratic'. The Politics of Johannes Althusius, F.S. Carney, trans., Boston: Beacon Press, 1964, p. 200.

6. In A Preface to Democratic Theory (1955), where polyarchy is again interpreted as an actuality, as distinguished from the ideal of democracy, I defined it 'loosely as a system in which (eight) conditions exist to a relatively high degree'. The 'conditions' were essentially the 'criteria' that Lindblom and I had specified earlier, but transmogrified into the ideal or limiting conditions of democracy. My sketch of an attempt 'to measure polyarchy quantitatively', assuming that each of the eight conditions could be scaled, then led to a definition of polyarchies as 'organizations in which all eight conditions are scaled at values equal to or greater than 0.5', further distinguishing 'egalitarian polyarchies' as 'polyarchies in which all eight conditions are scaled at values equal to or greater than 0.75'. (pp. 84-7). Some readers, not surprisingly, misunderstood my intentions and assumed that in my view polyarchy is hardly distinguishable from perfected democracy. In later work I wholly abandoned the heroic, but confusing, achievement of pseudo-quantitative scholasticism represented in those two or three pages of a chapter appendix.

7. In The Problem of Sovereignty, New Haven: Yale University Press, 1917, and Authority in the Modern State, New Haven: Yale University Press, 1919.

8. He and Frida Laski translated Duguit's Transformation du Droit Public (1913), which was published in 1919 as Law in the Modern State.

9. For example, Francis W. Coker, 'The Technique of the Pluralistic State', 186-213; George H. Sabine, 'Pluralism: A Point of View', The American Political Science Review, 17 (February 1923) 34-50; W.Y. JElliot, 'Sovereign State or Sovereign Group', American Political Science Review 19 (August 1925) 475-499.

10. F. W. Coker's widely read text, Recent Political Thought, published in 1934, contained a chapter on 'The Pluralists' Attack on State Sovereignty'. He published an essay on the same subject the following year in a volume edited by Charles Merriam and Harry F. Barnes, Political Theory, Recent Times.

11. In 1936-7 I had taken F. W. Coker's year-long seminar at Yale on Recent Political Thought in which, though somewhat critical, he gave considerable attention to the subject of pluralism. Because I found the ideas attractive, I read, or at least dipped into, the works of the- writers mentioned above. By coincidence, as an undergraduate at the University of Washington, I had read Duguit's Law in the Modern State in a course on jurisprudence. Although my copy, which I still possess, is well marked, I did not grasp much of what Duguit was up to until I read Laski a year or two later.

---

12. Politics, Economics, and Welfare, p. 302. The sentence is italicized in the original.

13. Since one of the sources of the 'theory' was Who Governs?, probably I bear some responsibility for the confusion. I described New Haven as having changed over two centuries from oligarchy to pluralism, a judgment I believe to be correct. In several places I also referred to New Haven as a 'pluralist democracy'. But I did not attempt to define these concepts with any rigor, nor to distinguish sharply between the generic features of 'pluralism' and the various aspects of New Haven political life that it might or might not share with other 'pluralist' systems. Thus some readers were tempted to over-interpret the meaning of pluralism. Pace some interpretations, the book was not written to advance a general 'pluralist theory of polities'; in fact, 'pluralism' and 'pluralist democracy' are not included in the index. In hindsight, it might have been better to have set out a more explicit theory. But perhaps not.

14. J.G.A. Pocock, The Machiavellian Moment. Princeton: Princeton University Press, 1975, pp. 66-80.

15. On the Social Contract with Geneva Manuscript and Political Economy, ed. by Roger D. Masters, New York: St. Martin's, 1978, pp. 212-13. Italics added.

16. In Vol. 2he repeated the warning of the first, ' "that the unrestrained liberty of association for political purposes is the last degree of liberty which a people is fit for. If it does not throw them into anarchy, it perpetually brings them, as it were, to the verge of it!' Thus I do not think that a nation is always at liberty to invest its citizens with an absolute right of political association for political purposes; and I doubt whether, in any country or in any age, it be wise to set no limits to freedom of association'. Vol. 2, p. 143. The quotation cited from Vol. 1 is at p. 222; there, however, the translator (Henry Reeve) rendered the passage slightly differently.

17. Alexis de Tocqueville, Democracy in America, Volume 1. New York: Schocken Books, 1961, pp. 220-221.

18. Ibid., Volume 2, p. 133.

19. Cf. Jane Mansbridge, Beyond Adversarial Democracy.

20. 'Norway: Numerical Democracy and Corporate Pluralism', in Dahl, Political Oppositions in Western Europe, New Haven: Yale University Press, 1966, pp. 105, 107.

21. Juan Linz, 'Totalitarian and Authoritarian Regimes', in Fred I. Greenstein and Nelson W. Polsby, Handbook of Political Science, Vol. 3, Macropolitical Theory, Reading, MA., Addison-Wesley, 1975, pp. 175-411.