Joshua Cohen and Archon Fung, MIT

Forthcoming in Constitution, Democracy, and State Power: The Institutions of Justice, 4 volumes, ed. Joshua Cohen and Archon Fung (Cheltenham: Edward Elgar, 1996).


A just society guarantees its members rights to basic civil liberties; it protects the political liberties associated with democratic governance (rights of suffrage, association, and political speech), while ensuring state accountability and responsiveness to citizens; it distributes fairly the resources required for exercising civil and political liberties; and it promotes the general welfare. Despite broad agreement on these abstract requirements, the conditions that foster justice, thus understood, are a matter of long-standing controversy in political theory. That controversy is fueled by competing interpretations of fair distribution, general welfare, and the other norms of justice. But even when we set those differences aside, much disagreement remains. Candidates for favorable conditions range from a supportive political culture, an absence of sharp ethno-religious cleavages, decent material circumstances, and the absence of powerful enemies, to the work of political artifice -- in particular the work of constructing and preserving suitable constitutional and democratic institutions. Although such institutions seem clearly to make a difference to justice, considerable dispute surrounds the details of that difference: to what extent and in what ways do institutions contribute; and how much does their contribution depend on the inner logic of institutions, as distinct from a favorable economic, cultural, historical context.[2] Moreover, the interest of these disputes is not purely intellectual. In a period of constitution-making and constitutional revision, their resolution has large practical implications.

To address these controversies about the role of basic political institutions in advancing the requirements of justice, we have collected more than 50 articles on such institutions as the rule of law, judicial review, federalism, separation of powers, freedom of speech, elections and parties, direct democracy, organized social groups, and administrative agencies. The articles address three broad issues:

1. How, as a theoretical matter, are particular institutions supposed to meet the requirements of a just political order: the protection of basic liberties, accountability and responsiveness, a fair distribution of resources, and the general welfare?

2. How well, as a practical matter, do those institutions serve the goals that have classically been assigned to them? For example, does judicial review succeed in protecting individual liberties, or elections in securing government responsiveness? And in what ways are these contributions (or failings) sensitive to other aspects of the material, cultural, social, or institutional context?

3. What reforms might increase the capacity of these institutions to promote fundamental political values?

The focus on institutions aligns these volumes more closely with the tradition of political theory from Hobbes to Mill than with contemporary work in the field.[3] Much contemporary political philosophy is devoted to stating, clarifying, and justifying political norms, reflecting on the possibility of such justification, or examining the political morality of public policy rather than to exploring the institutional implications of political principles.[4] Perhaps the setting of constitutional democracy is simply assumed, and political philosophy is offered as a contribution to public discussion among citizens in that setting; or perhaps the study of institutions of justice is simply assigned to another discipline, more empirical than political philosophy. Whatever the explanation, current discussion departs from classical constitutional and democratic theory. Hobbes's Leviathan, Locke's Second Treatise, Montesquieu's Esprit des Lois, Rousseau's Contrat Social, the Federalist Papers, Kant's Rechtslehre, Bentham's many constitutional codes, Hegel's Philosophie des Rechts, and Mill's Representative Government all address institutional implications as much as political principles. The assumption throughout is that these principles must be realizable in a stable social order, that their realization depends, inter alia, on the organization of public institutions, and that, for this reason, political theory must incorporate debatable empirical claims about the workings of institutions, not only normative principles.

In his defense of constitutional government, for example, Locke rejected Hobbes's empirical assertion that constitutional limits on government would generate civil conflict, as well as his theory of natural law.[5] Similarly, in rejecting representative government in favor of democratic assemblies, Rousseau combined a strong defense of the value of political autonomy with the empirical hypothesis that only a vigilant citizenry could prevent government from appropriating political power. Our aim in these volumes has been to collect contemporary assessments of the validity of such large empirical claims -- articles that address the actual or likely success of basic institutions in achieving their constitutional and democratic objectives.

The rest of this introduction provides a guide to the collection. We first sketch the broad, substantive themes about constitutional and democratic government that recur throughout the volumes; then we state the criteria for selecting articles and some of the obstacles we faced in finding suitable materials; finally, we introduce the ten sections, indicating how the articles in each connect to our larger themes.


In his Leviathan, Hobbes argues that absolute political authority is required to keep peace: limits on sovereign authority invite political instability, social dissolution, ultimately the collapse of ordered cooperation into a war of all against all.[6] Hobbes makes a forceful case for political authority, but lacks a compelling account of why an unlimited sovereign will not itself provoke conflict. And that deficiency leaves us uncertain about why domestic peace is more likely under absolute sovereignty than under a limited government designed to respect citizens' lives and liberties. John Locke provides the classical statement of the difficulty:

To ask how you may be guarded from harm, or injury, on that side where the strongest hand is to do it, is presently the voice of faction and rebellion: as if when men quitting the state of nature entered into society, they agreed that all but one should be under the restraint of the laws, but that he should retain all the liberty of the state of nature, increased with power, and made licentious by impunity. This is to think, that men are so foolish, that they take care to avoid what mischiefs may be done them by pole-cats, or foxes; but are content, nay, think it safety, to be devoured by lions.[7]

Locke's rejection of the leviathan state as an institutional guarantor of social peace -- and of the individual liberty and personal security that depend on peace -- points to an important strand in modern conceptions of just institutions. Guided by Locke's claim that the unification of legislative and executive power in the same hands is the essence of despotic government and the enemy of political society,[8] and Montesquieu's warning that "every man invested with power is apt to abuse it, and to carry his authority as far as it will go,"[9] this strand argues that political institutions must address the threat that state power itself poses to the liberty and security of individuals, not simply reduce the dangers of conflict among those individuals themselves. It proposes to meet this threat by confining state power, and endorses constitutional government as its principal institutional instrument. Constitutional government is to be limited through a combination of public rules delineating the boundaries of legitimate state authority (rule of law), devices for fracturing state power to keep it from transgressing those boundaries (federalism, separation of powers), and arrangements for monitoring government conduct for possible transgressions (judicial review, legislative supremacy).

An alternative response to Hobbesian absolutism begins by emphasizing that restraints on public power are of uncertain benefit when it comes to meeting the demands of justice. Such restraints may, for example, cripple the state's capacity to advance the general welfare, or to correct conditions of economic and social subordination, or to remove substantial "private" obstacles to a fair distribution of resources. For example, the general welfare may depend on the provision of public goods -- education, infrastructure, public health, national security. And such provision may require an affirmative state with the capacity to resist the supplications of particular interests, articulate a view of the general welfare, pursue sustained projects in the name of that view, and raise the revenues required by such projects. Furthermore, assuming a market economy, affirmative public power may be necessary to ensure fair access to resources for citizens whose fate would otherwise be sealed by accidents of natural or social fortune. Moreover, justice requires fair political equality of citizens, and a more affirmative state may be required, as well, to guarantee all citizens the rights and powers they need to participate as equals.

According to this second line of thought, then, the problem of institutional design is not to fracture public power, but to establish a structure which ensures its proper direction -- to ensure, as Rousseau argued, that the state is guided by the general will and reason of citizens as a body, not by the particular desires and judgments of a Hobbesian sovereign. More concretely put, the institutional task is to enable citizens to participate as free and equal in defining common political objectives, to ensure that the state has the capacity to pursue those objectives, and to provide citizens with the powers they need to make certain that the state properly exercises that capacity.

Our sketch of these two perspectives no doubt exaggerates their differences. Affirmative states are not unlimited, and familiar constitutional devices play an important role in keeping them within bounds;[10] correspondingly, democratic government can be understood in part as a device for limiting governmental power and protecting civil liberties, rather than as an arrangement that helps to form a general will and ensure that government is guided by it.11 Still, the contrast illuminates an historically important distinction between views about just institutions that emphasize the use of constitutional instruments to confine the exercise of power, and views that stress the importance of democratic will-formation and state capacity.

This division runs through the four volumes. Part 1 (sections 1-3) collects important articles which describe and assess the main devices of constitutional design: the rule of law, judicial review and the protection of basic rights, and the vertical and horizontal "deconcentration" of power through federalism and the separation of powers. The articles describe the roles conventionally assigned to these devices -- limiting political power, or establishing conditions of democratic authorization -- and evaluate their success in achieving those aims.

Part 2 (sections 4-7) examines several institutions associated with democratic governance: protections of free speech, regular elections and competing parties, direct (rather than representative) democracy, and secondary associations. Here, too, the articles explore the capacity of such arrangements to secure justice by preventing the arbitrary exercise of power, or by forming individual preferences and opinions into public judgments and then enacting those judgements as public policy.

Part 3 (sections 8 and 9) shows how these concerns inform two debates about the American scheme of constitutional democracy. The articles explore problems of ensuring suitable forms of constitutional and democratic governance that arise with the expansion of administrative power in the modern welfare state and the expansion and extensive deployment of US military power since the end of World War II.

Part 4 (section 10), finally, underscores the limits of the earlier debates by examining institutions of justice that lie "beyond the state." These articles discuss such institutions in light of contests in the twentieth century over gender and family, the organization of work, and the expanding power of corporate capitalism. How suited are eighteenth century political forms to responding to those contests? What contribution to justice might be made by more egalitarian families, and by democracy in firms, secondary associations, and international institutions?


These volumes bring together articles assessing the role of constitutional and democratic institutions in protecting civil and political liberties, ensuring fair equality, and promoting the general welfare. Because of our concern with institutions of justice, we have not, as a general matter, included philosophical essays that explore the nature of political justification, or defend a particular set of individual rights, or explore the legitimate scope of state action. Instead, we concentrate on articles that propose institutional strategies for protecting individual rights, or keeping the state within its legitimate boundaries, or securing the accountability of government to the governed -- or that discuss the failure of conventional institutions to achieve these ends.

That, at any rate, is the aim, though we have been forced to depart from it in several places. The departures reflect the history of political science as a discipline. In the 1950s much political science got out of the business of studying institutions, and, more particularly, evaluating their success in achieving a just society. Responsibility for the troubles may lie with the anti-institutional thrust of the 1950s behavioral revolution in political science, or perhaps with a strange understanding of the requirement of value-freedom according to which ethical-political values cannot play a role in explicitly defining research questions.[12] Whatever the explanation, we were less well-served by the political science literature than we had hoped. To be sure, legal scholarship (particularly in constitutional law) never lost its interest in normative institutional argument. But that scholarship is often decidedly non-empirical in its institutional evaluations, and concerned more with developing doctrine than assessing the empirical plausibility of theories. So in several cases we were compelled to depart from our central idea.

Take the case of judicial review: since Robert Dahl's 1957 article on the U.S. Supreme Court's limited success as a guardian of minority rights,[13] the literature has offered virtually no serious empirical evaluation of judicial review as a device for ensuring rights. Countless articles and books state the rationale for judicial review and aim to reconcile judicial protection of rights with democratic government. And we decided to include several such defenses, including important articles by John Hart Ely and Ronald Dworkin. But very few articles fit more perfectly with our larger aims.

Things may now be changing. Over the past 15 years, a "new institutionalism" has emerged, returning institutions to the center of political science.[14] But it has not yet produced a rich literature explicitly connecting its own investigations to the large institutional claims of classical theories of constitutional and democratic government. As Adam Przeworski observes, we still lack "sufficiently reliable empirical knowledge to answer questions about institutional design," particularly when we aim to provide answers appropriate to "specific historical conditions."[15] Considering the extensive institutional advice that lawyers and political scientists have recently been giving to newly democratizing countries, this observation may come as a surprise. But having surveyed the literature, we suspect that the advice largely reflects constitutional traditions in advisors' home countries. The literature itself does not settle much about institutional design.

Within the constraints just noted, several criteria guided our selection of articles.

1. The bulk of these essays focus upon political institutions broadly conceived. For better or worse, the guiding assumption in the field is that such institutions simultaneously represent the greatest threat and the best hope for (social) justice. So we have focused the collection on work about limiting the dangers and harnessing the potential of democratic states. But justice will not issue from political institutions alone. Institutions of family, workplace, and economy, and global arrangements between and among states and societies present important obstacles and opportunities to the advance of social justice. The essays in section 10 address some these arenas and challenge our discipline's assumption about the importance of political institutions.

2. We have included some articles because they strike us as especially interesting, others because they present an unconventional outlook, addressing aspects of a problem that merit closer attention. At the same time, we have sought to achieve a certain coherence within each of the ten sections. For each subject, we included classical, research-agenda-defining articles that state the traditional reasons for thinking that a particular institution will advance justice, as well as work that criticizes those views analytically or empirically. Each section, then, may be read as a debate between and among competing perspectives, and, in the next part of the Introduction, we sketch the terms of those debates.

3. An Introduction to 2,000 pages, spread over four volumes, is hardly the place to emphasize limits. Still, a quick glance at the table of contents reveals a geographic concentration imposed by both the bounds of our expertise and the requirement -- imposed by the design of the volumes -- to find articles in English. The majority of the essays are written by American authors, and many focus on the United States. Moreover, the comparative essays deal almost exclusively with Western industrialized nations (the articles by Linz, Riker, and Lijphart in section 3 are exceptions). Of course the issues addressed in these volumes arise throughout the world.[16] But precisely because cases that fall outside our competence are so relevant and important, we have left the reexamination of this work to those more capable.

4. Finally, we have included only articles that originally appeared in journals (and therefore not as chapters in books, or contributions to edited collections). This constraint came from the definition of the Political Thought series in which these volumes appear. The series aims to make important journal articles more widely available by collecting them in thematically organized volumes. We decided early on to treat this constraint as absolute, even when -- as, for example, in the literature on federalism -- much of the best discussion appears in books.


The Rule of Law

The rule of law is a basic hedge against the arbitrary exercise of state power. John Locke, for example, thought that whether government is democratic, aristocratic, or monarchic, "the ruling power ought to govern by declared and received laws, and not by extemporaneous dictates and undetermined resolutions. For then [i.e., in the absence of the rule of law] mankind will be in a far worse condition than in the state of nature. . . ."[17] And Jean-Jacques Rousseau counted among the "marvels of law" that people "obey and no one commands, that they [people] serve and have no master, and all are the freer . . . because under what appears as subjugation, no one loses any of his freedom except what would harm the freedom of another."[18]

The core idea of the rule of law -- a government of laws, not men[19] -- is that government ought both to guide and limit the exercise of its coercive power by reference to norms that are general, public, prospective, and reasonably stable.20 Government ought consistently and impartially to enforce such norms, and it ought to use its power only to enforce them, no matter how attractive or honorable the ends it aims to advance.[21] Thus restricting the government's use of coercive power will, according to the standard argument, advance values of personal security, liberty, equality, and community. The requirement of generality promotes personal security by limiting highly focused, discretionary exercises of power, thus protecting people from abuse. Generality and stability serve the value of liberty in part by requiring that individuals not be subordinate to (subject to the discretion of) other individuals; and publicity and prospectivity serve liberty by requiring that people be adequately informed about the limits of permissible behavior. Moreover, all these conditions make the exercise of power more predictable, enabling people to plan and to regulate their own conduct. Restrictions serve the value of equality, too, because generality of rules (assuming consistent enforcement) grounds a minimal expectation of equal treatment for equal cases, regardless of the particular people involved. And generality serves the value of community by establishing a common framework of norms for all.[22]

That conventional story has often been criticized. Some critics argue that legal categories are always objectionably over-inclusive, that their generality blinds us to individual cases in their full particularity, and that attention to particularity is required for a reasonable resolution of complex cases.[23] Others claim that the legal form presupposes market relations and associated conflicts of individual interest, and can be expected to disappear in favor of more technical modes of problem-solving as those relations are transformed.[24] In his contribution on the rule of law, Hayek, by contrast, embraces the conventional account in the strongest terms. He claims that the rule of law is the principal institutional requirement for fulfilling the classical liberal commitment to liberty.[25] Indeed, we guarantee the protection of individual liberty when we limit the coercive powers of government by reference to a stable scheme of general abstract rules, requiring the lawmaker "to prove his belief in the justice of his pronouncements by committing himself to their universal application to an unknown number of future instances and renouncing the power of modifying their application to particular cases" (99).[26] When legislation is thus restricted, it guarantees a private sphere in which "a free person can count on using a known domain of resources" (99).[27]

Hayek's view is based on a sharp distinction between legislation and governance. The purpose of legislation is to provide a system of general rules fixing boundaries of right that constrain coercion; governance, by contrast, is a matter of pursuing goals deemed socially desirable: democratic governments, for example, are "dedicated to the realization of a particular system of concrete ends" (102). Such dedication is unexceptionable. But liberal democracies open individual freedom to public invasion by conflating the goal-pursuing function of government with the generality-requiring, constraint-setting, freedom-protecting role of legislation. This conflation arises from mistakenly identifying law by reference to its pedigree -- in particular, its source in majority will -- rather than its intrinsic characteristics (in particular, generality); and that mistake eliminates any restriction on the scope and content of law. Legislatures, for example, sometimes pursue policy goals by enacting highly specific regulations, or by enacting vaguely worded statutes and handing over to administrative agencies operating without clear rules the power to regulate individual conduct and property pursuant to those statutes.[28] Although those regulations and statutes have democratic pedigree and are therefore conventionally understood to be genuine laws, their specificity and vagueness deprive them of that status. But when they are treated as laws, and relied on to authorize the exercise of coercive power, citizens are subjected to power unconstrained by rule. To secure the "essence of a free society," and rescue us from the road to serfdom, governments must separate lawmaking and governance, and cabin government discretion by keeping the democratic pursuit of social ends within the general rules prescribed by the strict rule of law.

Joseph Raz broadly agrees with Hayek's conception of what the rule of law is, but accuses him of "exaggerat[ing] its importance" (226). For Raz, the central aim of the rule of law is to establish standards by reference to which people can guide their own conduct. This abstract conception yields various formal conditions on the rule of law: prospectivity, stability, clarity, and generality. When these requirements are met, the result is a more predictable environment, and predictability is good for freedom of conduct, which is an important political value. So government ought, all else equal, to limits its coercive power to enforcing rules that meet these conditions. But such limitation does not guarantee protections from torture, nor does it assure political freedom, understood to encompass freedoms of speech, association, and conscience. In short, the rule of law is not a sufficient guarantee of freedom. Moreover, the rule of law is only one specific virtue among others: it is not the rule of good law, democratic law, or just law. For this reason, "Conflict between the rule of law and other values is just what is to be expected" (228). And when such conflicts arise, there may be good and sufficient for letting the rule of law give way.

Raz's more measured enthusiasm for the rule of law echoes H. L. A. Hart's "Positivism and the Separation Between Law and Morals." With the classical legal positivists (Jeremy Bentham, John Austin), Hart embraces the thesis that law is a matter of social fact. Although he disagrees with Bentham and Austin about what the relevant social facts are,[29] he rejects content-based tests of legal validity, accepts the separation of law and morality, and criticizes natural law theories for denying that separation.[30] According to Hart, moral rightness is neither necessary nor sufficient for legal validity -- indeed moral rightness is no part of the ultimate test of legal validity set out in a community's "rule of recognition." To be sure, the rule of law is an important good. Laws must be general, and legal generality connects to a basic precept of justice: that like cases be treated alike. And given such basic facts of human life as our common physical vulnerability, all legal systems will overlap with morality in condemning theft, murder, and violence. But these points of overlap permit substantial conflict between legal and moral obligations: individual laws may be evil, and legal orders "hideously oppressive." Moreover, maintaining a strong distinction between legal and moral requirements allows us to steer clear of two important moral errors. With Bentham, Hart rejects the anarchist's contention that morally repugnant laws are no laws at all and impose no legal obligations. But distinguishing law and morality also enables Hart to reject appeals to legal obligation as excuses for immoral conduct. The separation of law and morals enables us to condemn evil laws as such, while still giving law, as law, its due.

Fuller responds by arguing that this fundamental distinction between law and morality is intellectually and morally indefensible, and defending a qualified natural law view.[31] Fuller agrees with Hart that laws are rules, and that particular rules come as elements in systems of rules.[32] But he argues that Hart fails to appreciate the "internal morality" of a legal system -- the fact that, by consistently applying coercion in accordance with rules, a legal order acknowledges, in practice, the importance of values of fairness and justice. Thus, contrast two cases. In the first, citizens can be arrested and detained on the command of authorities. Here, the authorities violate requirements of procedural fairness, and do so precisely because they fail to treat those citizens in accordance with a system of rules: in short, there is no rule of law, and the absence of such rule is a source of substantial evil. In the second, rules are effectively in play. But officials must, then, be following such precepts of justice as "treat like cases alike," "only enforce sanctions announced in advance," and "give fair notice:" following these precepts is simply part of what it is to coordinate social conduct through rules.

Apart from noting the law's intrinsic morality, Fuller thinks that a frank acknowledgment of that morality is likely to have positive practical consequences for legislation and adjudication. But however desirable those effects are, it remains true that the rule of law does not itself suffice to ensure justice, and that the law's intrinsic morality appears to be compatible with -- in Hart's phrase -- a "hideously oppressive" legal order.

Franz Neumann and Catharine MacKinnon shift the focus, both methodologically and substantively. Methodologically, their aim is not to do the work of analytical jurisprudence by explicating the concept of law and legal order, but to assess the practical implications of the rule of law in settings of unequal power. Substantively, they argue that the rule of law, operating in such settings, enlists the force of state to reinforce a status quo of social inequity and domination.

Like Hayek, Neumann describes large historical changes in the character of law with the shift from a classical-liberal to an interventionist state. Neumann traces that evolution to the transformation of capitalism from a liberal-competitive to a monopoly phase, rather than to a confusion of legislation and governance. In its classical liberal form, law was described by reasonably determinate rules, forming a coherent system, satisfying requirements of generality and prospectivity. With Marx, Neumann emphasizes that this rule of law stabilized capitalism by "obscuring the domination of the bourgeoisie; with Weber, he emphasizes its role in "rendering the economic system calculable" (42). But the generality of rules and regularity of application required by competitive capitalism also served an "ethical" role: by protecting citizens from arbitrary state action, it ensured "a minimum of liberty and equality" (42). Monopoly capitalism transforms the character of law. Formulated in vague terms, lacking determinate content,[33] regulations leave vast areas of discretion to administrative agencies, powerful private actors, and judges, whose translation of law into results in specific cases requires resort to "general principles" of natural or common law. Because the content of law is so indeterminate, the resort to such principles is easily turned to the advantage of "the power position of the monopolies," whose power enables them "to manage without the formal rationality of the law" (58).

Though Neumann endorses the rule of law, then, his enthusiasm is qualified by his attention to its two-sided character -- though it obscures domination, it also protects liberty. The practical implications of this qualified endorsement are uncertain, however. Neumann believes that liberty requires the rule of law, that the rule of law can coexist with competitive capitalism, and that it is incompatible with monopoly capitalism. But he does not explain whether a successor to monopoly capitalism that does not rest on domination is possible, or whether a liberty-protecting rule of law is compatible with such a successor.

MacKinnon, too, analyzes the rule of law under conditions of unequal power. But she is concerned with sexual hierarchy, not principally with economic power. Her central claim is that the liberal rule of law is substantively and formally male: "The law sees and treats women the way men see and treat women" (644). Substantively, legal rules reflect male outlooks and advance male interests; for example, in a world of sexual hierarchy, the abstract idea of a private sphere off-limits to public power implies in practice that the state will "keep some men out of the bedrooms of other men," but not interfere with the sexual subjection of women within families. Formally, the rule of law reproduces the social subordination of women through its conceptions of neutrality, fairness, and objectivity. Men have power, and abstract rule-of-law values are interpreted and given determinate content against the background of that power, which is taken for granted. To be neutral, objective, fair means not to take the side of any particular group. But in a society defined by sexual hierarchy, the contest is unequal; so if neutrality, objectivity, and fairness mean not taking sides, then in practice they mean acceding to male domination.

Transforming the rule of law from a tool of sexual subordination into an instrument for gender equality, then, requires dislodging the male point of view and substituting a feminist jurisprudence that directly challenges gendered social power, rather than simply assuming it as a fixed background. Such jurisprudence will of course be criticized for taking sides, for a partisanship that rejects the high ideals of neutrality, objectivity, and fairness associated with liberal legalism in the name of the interests of a particular group. But that is the inevitable consequence of the role of male power in giving content to those apparently formal norms.

MacKinnon's critique of the rule of law seems more fully rejectionist than Neumann's. She does not discuss what he calls the "ethical" role of the rule of law (in protecting liberty or equality), and suggests that notions central to the rule of law -- objectivity, neutrality, and fairness -- are essentially tied to practices of male domination, that a feminist jurisprudence has no need for them, and that they would have no role in a society without sexual hierarchy. To say this is not simply to note, with Raz, that the rule of law is insufficient to protect freedom, and that other values may override it; it is to condemn the rule of law as such. But MacKinnon does not explicitly state these strong claims; she is, as a general matter, silent on general jurisprudential questions. Her criticisms of the form and substance of the rule of law may, therefore, be read as more contextual -- as showing that, in settings of sexual inequality, the rule of law operates to obscure and justify subordination -- rather than as implying a more sweeping rejection.

Judicial Review

Judicial review adds institutional force to the notion that laws rather than men should ultimately rule. In democratic systems with judicial review, ordinary legislative processes operating on majoritarian norms are responsible for advancing the general welfare; courts with powers of review are responsible for enforcing the constitutional framework itself, declaring legislation void should it violate basic law, as expressed in constitutional rules and principles. Judicial review is particularly important when the constitution includes a scheme of individual rights. Because a majority indifferent to minority rights may prevail in the legislature, judicial review provides an essential institutional backstop which enables minorities to vindicate their (constitutional) rights and compels winners to treat losers with respect.

This sketch of the rationale for judicial review prompts two questions. First, are individual rights particularly well-protected in systems with judicial review? Does the U.S. system, for example, with a written constitution, bill of rights, and judicial review protect individual liberty or ensure fair treatment more effectively than a system with no written constitution, or with a written constitution containing exclusively structural provisions, or with a written constitution and bill of rights whose enforcement are left in the hands of citizens and the legislature?[34]

Second, what limits the power of the judges themselves? This latter question has particular force when constitutional provisions protecting rights are framed in abstract terms -- such as the free speech and equal protection clauses in the first and fourteenth amendments to the US constitution. These clauses are stated in open-ended language: "Congress shall make no law abridging the freedom of speech;" "Nor shall any state deny to any person within its jurisdiction the equal protection of the laws." The language requires interpretation, the interpretations are always controversial, and the interpretive controversies often closely track broader ideological-political controversies. In the face of such disagreement, judicial interpretation may appear simply to usurp popular sovereignty, with judges subverting popular rule by turning their political preferences into higher law (even as they claim to be finding the right legal answer to a disputed question).[35]

Ely, Dworkin, and Riker and Weingast focus on this second question, and offer three competing rationales for judicial review and corresponding prescriptions for deciding controversial cases. Although we may be inclined to think of judicial review as limiting democracy, Ely's "representation-reinforcing" theory argues that judicial review helps to ensure fair, democratic procedures of collective choice.[36] According to Ely, the U.S. constitution is principally a procedural document; it does not mandate political goals, but helps establish a democratic process enabling citizens themselves to decide on the substantive goals that will guide public policy. In particular, it calls for a process in which all are able to participate, discussion is open, every vote is counted, and all interests are taken into consideration. To ensure such a process, the constitution establishes certain rights: rights of political speech[37] and association, for example, are required for an open process; rights not to be a target of hostility and stereotyping are required if all interests are to be taken into consideration. Thus, majority rule with government regulations of political speech is not a democratic process in the relevant sense; nor is majority rule in a racially divided society, unless there are constraints on the use of racial classifications in legislation.

But a democratic process is not certain to preserve the rights required for its own integrity. A majority might decide to freeze a group out of the process, either through formal exclusion (denial of voting rights, limits on free speech), or by systematically discounting their interests (as when legislation is rooted in racial prejudice or sexual stereotyping). The role of judicial review, then, is to protect those rights; the Supreme Court is to strike down legislation that formally disenfranchises, or that arises from processes that effectively exclude the interests of some citizens. Those protections are not constraints on democratic process imposed by courts in the name of higher laws, transcendent values, or natural rights; instead they are required if the political process is to be democratic. Correspondingly, when judges protect those rights and interpret the constitution's open-ended clauses in light of the ideal of democracy, they serve as guardians of popular sovereignty, not as authorities superior to the people. By interpreting the constitution as a design of democracy, and assigning to judicial review the task of upholding that constitution, Ely finds a role for judicial review that is limited, essential, and compatible with a commitment to democratic government -- indeed aimed at ensuring such government, rather than at tying its hands by reference to independent and prior claims.

Dworkin agrees that judicial review is necessary to uphold the ideal of democracy. But, rejecting Ely's exclusively procedural interpretation of the democratic ideal, he concludes that judges ought to protect a more expansive set of rights, some of which are not justified by their connection to a fair political process.[38] According to Dworkin, the abstract concept of democracy is too thin to guide judgment in controversial cases; so judges who wish, as Ely urges them, to safeguard democracy must develop a more specific interpretation of democracy that justifies settled aspects of democratic practice and offers guidance about unsettled aspects. To provide such justification, however, the interpretation of democratic process must be set within a system of moral-political principles, stating what is required for the law to treat people with equal concern and respect. Furthermore it is unlikely that the implications of that system of principles will be confined to rights associated with fair procedures of collective choice. Judges who understand their responsibility to protect democracy will aim to ensure the protections of rights that flow from those principles without worrying whether the rights fall into more procedural (by the people) or more substantive (for the people) precincts of the democratic ideal. More broadly, if the judiciary is to protect democracy, it cannot avoid addressing issues of political morality, and in so doing it serves as a forum of principle in which "the most fundamental issues of political morality will finally be set out and debated as issues of principle and not simply issues of political power."

Riker and Weingast use the insights of social choice theory to argue for a more skeptical assessment of normal democratic processes and firmer judicial constraints in the area of economic regulation. The need for such constraint arises, they claim, from a very general defect of democratic methods of collective choice: given only slight diversity of individual preferences, the results of majoritarian decision-making are highly unstable, and difficult to interpret as expressions of a coherent popular will.[39] Because of that instability, outcomes of democratic process will reflect the circumstances of voting -- for example, the sequence in which choices are considered -- as much as the preferences and opinions of voters. And because majority rule is so fickle, economic rights lack even minimal protection in normal politics. Though Riker and Weingast do not propose a theory of rights, they conclude from their pessimistic account of the rationality of majority rule that judges should read constitutional protections of liberty more strictly, and that judicial review should more firmly tie the hands of government, thus providing a more secure foundation to rights in general, and economic rights in particular.[40]

Despite their differences, then, these views each assign a central role to judicial review in protecting individual rights. Moreover, while Riker and Weingast see judicial review as a way to limit the power of government to regulate the private economy, Ely and Dworkin suggest ways that judicial review might instead be seen as a constitutive institution of democracy, thus addressing the second question noted earlier. But none of these views of judicial review speak to the first question: how does judicial review perform in practice? Does it protect rights more successfully than alternative arrangements?

Robert Dahl and Alec Stone address this question, but arrive at different conclusions. Dahl concludes from an historical examination of the U.S. Supreme Court that, contrary to common expectations, the Court does not generally overturn acts of Congress; and when it does, it typically works to protect privilege, rather than to defend vulnerable minorities from procedural or substantive exclusion.[41] These features of judicial review may be surprising if we focus exclusively on the Court's institutional independence, the formal autonomy of legal reasoning from political calculation, and jurisprudential theories of judicial review. But Dahl proposes a natural explanation. Justices and legislators are, after all, drawn from roughly the same elite classes, and U.S. Supreme Court appointments are politically controlled -- justices are nominated by the President, and subject to approval by the Senate. For these sociological and political reasons, the decisions of Court and Congress are likely to be broadly congruent, though lifetime appointments to the Supreme Court will cause some lag. Although the Court does not defend minority rights, Dahl does assert that it helps to ensure basic democratic process and gives the stamp of constitutional legitimacy to congressional decisions.

In arguing that the Supreme Court has played a limited role, Dahl looks only to cases in which the Supreme Court has overturned congressional decisions: the infrequency of such cases demonstrates those limits. He fails to consider how the institution of judicial review and the associated possibility of such overturning might shape the expectations and, therefore, the behavior of legislatures. Perhaps legislative choices reflect their beliefs about judicial decisions. If they do, then legislators would avoid passing legislation that they expect to be overturned, and the effects of judicial power would not be revealed through its exercise.

Emphasizing such legislative anticipation, Alec Stone's examination of judicial review by the French Constitutional Council arrives at a view very different from Dahl's. Constitutional review by the Council takes place immediately after Parliament adopts a bill, without any "case or controversy" requirement as in the U.S. case, and exclusively on petition from an elected official. According to Stone, this scheme alters the structure of legislative decision-making, effectively "juridicising" it. Constitutional review by the Council forces contending legislative groups to consider the possibility that opponents defeated in the legislative process will carry their challenge forward before the Courts. This threat ensures legislative attention to civil rights and democratic process. It also changes legislation from a majoritarian to a consensus process, because the legislature formulates laws which can gain super-majoritarian support, thus avoiding constitutional challenge. In Stone's view, this de facto consensus requirement frustrates the pace of social reform by limiting bold legislative initiatives.

Federalism and the Separation of Powers

Constitutional theorists and architects who favor formal divisions of state power, whether vertical (between national and local authorities) or horizontal (a separation of legislative, executive, and judicial powers), sometimes make their case by appealing to the virtues of fragmentation. By multiplying centers of power, federalism and separation of powers increase the costs of coordination among officials; moreover, they encourage officials to identify their ambitions with their institutional roles, thus pitting ambition against ambition. As the costs of coordination increase, and power serves as a limit to power, the likelihood decreases of threats to liberty from an illegitimate capture of political authority. Thus, premising a general need to check power with power, Montesquieu claims that "When the legislative and executive powers are united in the same person, or in the same body of magistrates, there can be no liberty," and that "Again, there is no liberty, if the judiciary power be not separated from the legislative and executive."[42]

An alternative case for vertical and horizontal division downplays the virtues of fragmentation, and emphasizes instead that the separation of powers and federalism help to institutionalize a more reasonable form of democratic politics than simple majoritarianism. For example, because the separation of powers requires that policy issue from the collaboration of different branches with separate competencies, it encourages more deliberative forms of decision-making. Moreover, the separation of legislative and executive power makes it harder for legislators to exempt themselves (or the factions they represent) from laws they impose on others; instead "they are themselves subject to the laws they have made; which is a new and near tie upon them, to take care, that they make them for the publick good."[43]

Similarly, the multiplication of units of government characteristic of federal schemes is sometimes said to improve the state's capacity to advance the public good.[44] Because each unit governs a smaller territory and population, policies can be tailored to regionally diverse preferences. Furthermore, a government operating closer to its subjects is likely to have greater knowledge of public problems and improved means for assessing policy solutions. A federal system also encourages governmental efficiency by creating an environment of competition between units; it allows policy innovation and learning by permitting units to pursue separate experiments, and subsequently discuss the results of their distinct efforts; and, by expanding the power of local government, deepens citizenship and accountability by creating greater opportunities for political participation. Finally, when national cooperation is required, federalism -- particularly in a religiously or ethnically divided society -- encourages more a consensual style of decision-making, inasmuch as the approval of separate units -- and not simply the majority of a national aggregate -- must be secured before proceeding.

Michael McConnell summarizes these distinct lines of defense of federalism. He distinguishes its contributions to the public good, the protection of rights, and popular government, traces each strand of argument to the Federalist Papers, and argues that the considerations advanced in support of the "Founders' Design" of federalism retain undiminished validity for current constitutional debates in the United States. McConnell does not, however, examine whether the practice of federalism supports the theory.

Riker does, and finds the conventional case for the virtues of a federal constitution in a divided society wanting. The quality of government does depend, Riker emphasizes, on effective decentralization -- regional autonomy and diversity. But he argues the standard recommendation of federalism to accommodate diversity and ensure political stability in divided societies reflects a conflation of effective with formal-institutional decentralization. Controlling for political culture, unitary constitutions are as good as federal ones at permitting local autonomy. What matters for virtuous political decentralization is not constitutional specification, but economic decentralization, responsive local officials, and strong local party organizations. If Riker is right, then political architects in divided societies should worry less about deciding between federal and unitary constitutional forms, and more about ensuring the strength of those "real forces" that determine local autonomy and diversity.

But whether effective loca autonomy is secured through constitutional mechanisms or other forces, its effect outcomes remains a major concern. According to Peterson and Rom and to Scharpf, it does make such a difference, and for the worse. Whereas the defense of federalism points to competition among units as a source of individual security, government efficiency, and policy innovation, Peterson and Rom argue that federalism-induced policy competition in U.S. welfare policy has created a "race to the bottom," with individual states seeking to attract investors by reducing levels of welfare provision.[45] Unitary national policies would eliminate the incentive for states to displace the costs of welfare policy onto other states; the result would be higher welfare floors and policies better matched to public preferences.[46]

Drawing on experience in Germany and the European Community, Scharpf, too, criticizes federalism, though, unlike Peterson and Rom, he is concerned with federalism as a structure of collective decision-making, not of decentralized policy-formation. Three conditions define his target: there is a decision-making body for an entire federation, composed of representatives of units of government; that body operates under a unanimity rule; and, because the parties do not agree on distributive norms, decision-making takes the form of bargaining rather than collective problem-solving, with decision-makers all pressing for the largest possible advantage for their own units. Together, unanimity and bargaining imply that programs remain in effect so long as a single member benefits. Policies are, therefore, likely to continue long after they have become inefficient. Instead of an attractively consensual form of decision-making, then, federal structures erect a "joint-decision trap," which freezes collective choice into sub-optimal outcomes. According to Scharpf, this structural analysis explains why European integration has resulted in "frustration without disintegration and resilience without progress." In the end, however, Scharpf's objections to federalism depend not simply on the structure of collective decision-making but also on the dominance of bargaining style within it.[47] The practical implications depend, therefore, on the availability of joint problem-solving as an alternative to bargaining.

Wilson and Linz evaluate the second great device of constitutional division: the separation of powers. Wilson defends the American scheme of vertical separation against critics who condemn it for hampering affirmative government policy. Relying on the same facts, Wilson comes to the opposite conclusion. He argues that vertical separation fragments power, and that the resulting restraints on affirmative government conduct simply underscore the success of the founder's design, which sought to protect liberty against government encroachment and slow the pace of constitutional change. Turning to recently democratizing countries, Juan Linz cautions against copying presidentialist, U.S.-style, separation of powers. Separate elections of a single executive increase the stakes of political conflict and reduce incentives to compromise, and a fixed term of presidential office reduces political flexibility. Countries that have only recently democratized, in which government operates against a background of deep social division and instability, may be better served by a parliamentary form of government.[48]

Using examples from Southern Europe to draw lessons for democratizing, socially-divided, Latin American polities, Arendt Lijphart explores both elements of deconcentration. He emphasizes the virtues of federalism, and of the separation of powers, though he argues that in presidential systems the dominance of a single executive commonly leads to executive dominance of the state. Lijphart's case draws on a general distinction between majoritarian (Westminster) and consensus forms of democracy.[49] The majoritarian model includes unitary national government and a fusion of executive and legislative power. Although this model is attractive in relatively homogeneous societies, it is ill-suited to socially divided systems because its winner-take-all character increases political conflict by heightening its stakes. Such systems ought to reject majoritarianism in favor of more consensus-promoting structures for collective decision-making -- for example, federalism, which requires agreement among separate units for collective action, and a separation of powers freed from Presidentialist distortions.

Freedom of Expression as an Instrument of Democracy

Two competing pulls define debate about the role of free speech and its permissible regulation in a democratic society. On the one hand, speech can be directly harmful, or provoke harmful conduct. So there is some temptation to think it ought to be subject to the same regulation as other causes of harm: regulated just in case the injuries outweigh the benefits. On the other hand, free expression appears to play an especially fundamental role in a democracy -- in checking tendencies to government misconduct,[50] facilitating the open debate that is a precondition for rational collective choice and well-grounded confidence in our own individual judgments,[51] respecting the sovereignty of citizens (rather than government) in a democracy, and/or treating citizens as free, responsible agents, capable of arriving at independent judgments about practical affairs. And for these reasons, we ought perhaps to treat free expression as a constitutive element of a democratic constitution, thus rejecting a balancing of costs and benefits in favor of a more absolutist view of free speech. In his dissenting opinion in Abrams v. US, Justice Oliver Wendell Holmes powerfully states this tension:

Persecution for the expression of opinions seems to me perfectly logical. If you have no doubt of your premises or your power and want a certain result with all your heart you naturally express your wishes in law and sweep away all opposition. To allow opposition by speech seems to indicate that you think the speech impotent, as when a man says he has squared the circle, or that you do not care wholeheartedly for the result, or that you doubt either your power of your premises. But when men have realized that time has upset many fighting faiths, they may come to believe even more than they believe the very foundations of their own conduct that the ultimate good desired is better reached by free trade in ideas -- that the best test of truth is the power of the thought to get itself accepted in the competition of the market, and that truth is the only ground upon which their wishes safely can be carried out.[52]

Holmes's marketplace-of-ideas argument provides one rationale for the free-speech "absolutism" that has played a leading role in American constitutional law since the 1960s. Alexander Meiklejohn's "First Amendment is an Absolute" strengthens the case for absolutism by shifting the foundation of a strong free-speech guarantee from the market to the fundamental ideal of democratic self-government that informs the U.S. constitution.[53] With James Madison, he claims that the American constitution is a design for citizen self-government, standing for popular sovereignty, not governmental sovereignty. Because the exercise of popular sovereignty requires free and open discussion among citizens, the government mocks the constitution's defining principle when it interferes with such discussion, for that interference treats citizens as, in effect, subjects of government rather than its masters.

Meiklejohn's account of free expression suggests a more general point about the role of rights in a democracy. We commonly think of rights as devices for protecting individuals from government decisions, supported by a majority of citizens. And the first amendment takes the form of a ban of congressional abridgments of free speech. But, according to Meiklejohn, the role of the free-speech guarantee is not to protect subjects from government tyranny, but to empower citizens to control government by providing immunity from sanction when they act in their capacity as rulers.[54] Because it links the free-speech guarantee to popular self-government, the Meiklejohnian defense assigns to political speech a position of special privilege.[55] But he construes "political speech" broadly, to include philosophy and science, literature, and discussion of public issues generally. Because inquiry in any of these fields may bear on public affairs, the government should be forbidden from restricting any of it.

Like Meiklejohn, Scanlon defends stringent restrictions on the legitimate authority of government to regulate speech. But the foundations and implications of his defense are different: the foundations do not lie in the decision to adopt a democratic form of government, and the implications are not confined to political speech, however expansively understood. Instead, Scanlon's strong free speech guarantee is one element in an account of political legitimacy, part of an answer to the question: "How could citizens recognize a right of governments to command them while still regarding themselves as equal, autonomous, rational agents?" (225) Government can be acceptable to citizens, thus regarded, only if it permits them to arrive at their own judgments about what to believe. That constraint implies, in turn, that autonomous citizens cannot grant to the state the right to decide what they may hear. Scanlon's view does not emphasize the autonomy of individuals as speakers with interests in self-expression, but the insult to listeners -- as free, equal, autonomous agents -- when government regulations of speech prevent them from hearing views that may influence their conduct by influencing their thought. According to this "listener autonomy theory," then, government is barred from preventing harms "by controlling people's sources of information to ensure that they will maintain certain beliefs" (222).[56] Scanlon of course acknowledges that speech can lead to harm by persuading people to act in ways that injure others; but a condition of political legitimacy is that government must find speech-protective strategies for preventing such injury.

The essays by Riesman, MacKinnon, and Kretzmer reject absolutism. They draw attention to the harms of unregulated speech -- speech can reinforce social inequality and, under conditions of conflict, jeopardize democracy itself -- and criticize the effectiveness of speech-protective strategies for preventing those harms.

Riesman's essay is the rare combination of legal-doctrinal argument and comparative political analysis: he explores the role of libel and defamation law in the struggles between fascism and democracy in pre-World War II Germany, England, France, and the United States. More substantively, he emphasizes the potentially destructive effects of speech, and sharply rejects the speech-is-special outlook in favor of treating speech on a par with other potential causes of harm.

According to Riesman, the Nazis first weakened their domestic opponents and discredited a democratic regime by exploiting loopholes in defamation law, then used that same law to repress their critics. By contrast, England's strict libel law, while not subject to fascist manipulation, repressed legitimate political criticism and debate. To avoid these dangers, Riesman proposes that U.S. courts should approach the law of libel and defamation with less attention to bright-line rules and abstract, free-speech principles, and more to the situational dangers of fascism and the need to protect democracy.[57] The result would not be a new scheme of legal rules, but an administrative approach to speech regulation (1314) which is profoundly at odds with the Meiklejohn-Scanlon view that speech deserves special treatment. Instead of formulating general principles and applying them irrespective of the power, purposes, and viewpoints of the litigants, courts would make contextual, fact-sensitive libel judgments, distinguishing "verbally similar attacks on the basis of the social context and the social purpose of the attacker and the litigant" (1318); and they would tailor damage awards to particular cases. More broadly, courts would understand better the need to "guide the direction of social change by deciding what groups are to be free to criticize and what groups are to be curbed" (1310). And they would make decisions that "protect those weaker groups and weaker critics who cannot rely on wealth or power over public opinion as their safeguard" (1310-11).

Like Riesman, Catharine MacKinnon attacks free-speech absolutism.[58] Acknowledging that we often have compelling reasons for protecting speech despite its harmful consequences, she argues that those reasons do not extend to pornography. Her case does not explicitly rely on the conventional reasons of immorality, offensiveness, or sheer vulgarity for regulating sexually explicit speech. She argues instead that pornography harms women by reproducing their sexual subordination, and that this harm is sufficiently great to justify regulation of pornography, irrespective of any political or artistic value that it might possess.

Pornography works its harm by making sexism sexy: it enlists our sexuality -- as idea, identity, desire, and practice -- in support of subordination, making it seem right, look natural, and feel good. Responding to the insistence that these harms of pornography be met with "more speech," MacKinnon argues that pornography defeats such efforts by silencing women. By producing a psycho-cultural setting in which we experience sexism as irresistible, and by silencing opposition, it closes off all avenues of exit from subordination, except the avenue of regulating pornography itself. MacKinnon defends such a regulation, drafted by Andrea Dworkin and herself. Defining pornography as the graphic, sexually explicit subordination of women that meets one of a series of tests, it would permit women to bring civil actions for the harms resulting from pornography.[59]

David Kretzmer focuses on racial hate speech, and proposes a more tentative defense of restrictions on expression.[60] He accepts that the considerations of democracy and autonomy advanced by Meiklejohn and Scanlon often -- perhaps normally -- provide compelling reasons for protecting speech. He notes, however, that the United States is the only nation whose constitutional doctrine forbids restriction of racist speech. And he argues that the particulars of racist speech -- its injuries to dignity and the relatively easy spread of racial prejudice through mass communications -- suffice to rebut the presumption against regulation. Indeed, because of those specific dangers, democracy and autonomy themselves are arguably enhanced by restrictions on racist hate speech. Whether such restrictions are wise legislation is another matter, though Kretzmer argues that on balance they are.

Owen Fiss, too, rejects absolutism. But he is less concerned with regulating particular categories of harmful speech than with defending a positive role for state action in advancing free speech values. More particularly, Fiss draws our attention to conflicts between democracy and autonomy arguments for free speech, urges that democracy is the basic free-speech value, and argues that affirmative state action is sometimes required to secure that value -- for example, by subsidizing speech that would not otherwise be heard, and regulating the speech of some to prevent the silencing of others.[61]

In the eighteenth century, the protection of individual autonomy from state repression was arguably the best formula for promoting the robust public discussion essential to a well-functioning democracy. Thus the first amendment to the U.S. constitution could promote democratic deliberation by barring Congress from interfering with individual speech. But changes in communications technology and increased inequality in the distribution of social resources invalidate this coincidence of autonomy-protective and democracy-promoting measures. State suppression of dissident speech is no longer the principal danger to free expression; instead, individual autonomy itself -- for example, campaign contributions by the wealthy -- may threaten the "uninhibited, wide-open, and robust" public debate required for a well-functioning democracy. Because of concentrated wealth and control of major media, a "free-market" in speech cannot be counted on to produce high-quality public debate or fair access for all citizens to the public arena.

Promoting democratic deliberation, then, may require the state itself to ensure fair access and robust debate, perhaps by restricting individual autonomy. To ensure the fair political equality required in a well-ordered democracy, for example, government may need to limit campaign contributions and expenditures, restricting the autonomy of wealthier citizens; to ensure that the speech of women is not silenced by pornography (as MacKinnon argues), government may need to restrict the autonomy of private producers and consumers of pornography. Emphasizing (with Meiklejohnian) the role of free speech in democratic self-government, then, Fiss urges that we shift from seeing the state as the principal threat to public discussion, and free speech values as providing an especially compelling case for fracturing state power, to seeing the state as perhaps the ultimate guarantor of democratic debate.

Voting, Parties, and Popular Control of Government

Universal suffrage, regular elections, and competing political parties lie at the heart of democratic government in a mass democracy. Combining political participation with political contestation,[62] these institutions routinely accomplish a peaceful transfer of power under conditions that suggest government by and for the people, not simply of them.[63]

Although a system of regular elections enjoys widespread support, advocates offer diverse and sometimes competing justifications for it. According to an ambitious interpretation, party competition provides the political infrastructure required for popular self-government in a mass democracy. When strong political parties compete for electoral support by articulating distinct, coherent conceptions of the common good, citizens can participate meaningfully in public debate about the merits of these conceptions, arrive at reasonable collective judgments about the substance of public policy, and hold government accountable when it fails to enact their choices.[64] According to a variant of this ambitious theory, social and economic context are decisive in determining the democratic implications of institutions of contestation and participation. Some argue, for example, that the ideal of substantive self-government can be realized through universal suffrage, elections, and parties, but only if the constraints imposed by the private control of investment are removed.[65] An alternative view expects less from election. It holds that electoral competition is unlikely to yield coherent collective judgments, either because such judgments do not in general exist, or because party competition is too blunt an instrument to reveal them. Still, it recommends regular competitive elections as supplements to the classical constitutional repertoire of tools to limit public power, and defend individual liberty against government abuse.[66]

Downs, Fiorina and Noll, and Lindblom all present grounds for skepticism about the more ambitious interpretation of electoral democracy, though their doubts have different sources and implications. Downs, who assumes a population of rational (utility-maximizing) voters, argues from the very general fact that information is costly.[67] Voters will, therefore, collect information only when the benefits, discounted by the improbability of influencing the outcome, outweigh the costs; and that implies that they will choose to remain uninformed about most issues. Interested parties will, in turn, exploit this rational ignorance by supplying voters with information biased in the supplier's favor (tobacco companies will support and distribute studies on the health benefits of smoking, domestic sugar producers on the medical virtues of bad teeth and the economic virtues of agricultural trade barriers). The public policies that result from this process will favor subsidizers of information, who will typically be producers with concentrated interests in particular areas of policy, rather than consumers, whose interests range over a wide range of goods.

Fiorina and Noll focus their skepticism on the contribution of elections to popular control of public policy in the context of modern welfare states. Such states create opportunities for elected representatives to deliver discrete benefits to their constituents -- through pork-barrel projects and constituency casework services. Voters correspondingly have incentives to choose candidates on the basis of their capacity to deliver such benefits, not of their policy positions. Once elected, then, representatives are able to insulate themselves by investing in service delivery, and in the bureaucracies that provide the benefits they deliver to constituents. As a result of this perverse dynamic of incumbency advantage, democratic states exhibit three negative developmental tendencies: they become bigger, more bureaucratic, and more "distributive" (as opposed to regulatory or redistributive). These results do not, however, reflect citizens' collective preference for destructive policies, but a structure of political decision-making that aggregates the locally rational choices of voters, representatives, and bureaucrats into collectively irrational results.

Whereas Downs traces the limits of popular control of policy to information costs, and Fiorina and Noll to the dynamics of electoral support within the political framework of a modern welfare state, Lindblom argues that those limits arise from the market.[68] Because markets are decentralized systems of coordination, the cooperation of economic agents must be elicited through incentives, not commands. The need to provide such incentives privileges the interests of powerful economic agents -- in the case of private ownership economies, that means business interests. If public policy displeases owners of capital, they will choose not to invest or hire workers; but low investment and unemployment have potentially disastrous consequences; anticipating those consequences, voters and officials therefore defer to business interests, and treat the will of business as the general will. Thus markets operate as "automatic punishing recoil" mechanisms that assign effective veto power over public policy to business, restricting the capacity of electoral institutions to serve as a framework of substantive self-government.

"Toward a More Responsible Two-Party System" brings political parties into the picture of electoral control, and argues that they play an essential role in achieving a more ambitious form of popular self-government. Written by a special committee of the American Political Science Association in 1950, and set against the background of the New Deal nationalization of American politics, it proposes a reconstruction of the American party system to transform the two parties from "loose associations of state and local organizations, with very little national machinery and very little national cohesion" (v) into national organizations capable of structuring national policy debates around party platforms that present alternative, coherent conceptions of the public interest. These organizations, while accountable to members on the inside, would be strong enough to resist the external pressures of wealth and interest-groups; because of their programmatic character, they would provide voters with targets for blame and discipline in the face of failures of public policy, thus improving the accountability of government to the governed; their legislative role would improve the capacity of national government to handle expanding responsibilities; and, by organizing coherent choices between alternative courses of action, they would establish a framework of "popular self-government" on a national scale.[69] Absent such parties, an expanding national government would be executive-centered, interest-group-dominated, and populated by extremist parties. Citizens would lack an organizational counterweight to the "natural" advantages of wealth, with predictable consequences for social and economic inequality.

Budge and Hofferbert find suggestive empirical support for the thesis that strong parties contribute to democratic governance. Examining British data from 1948-1985, they argue that party competition allows citizens to control public policy.[70] Analysis of party platforms and government budgets shows that parties, once elected to power, reliably translate their campaign promises into public policy. At least in the context of parliamentary government, then, there is some truth to the party mandate theory: that the choice voters make between parties is also a choice between policy alternatives, and that the winner delivers the program it was elected to pursue.

Anthony King is more skeptical about the importance of political parties in ensuring government accountability or responsiveness. King identifies common functions attributed to parties: structuring votes, mobilizing citizens, recruiting leaders, organizing government, formulating policy, and aggregating interests. He argues that these functions are often imprecisely specified, that political parties are only one among a number of organizations that perform them, and that parties are by no means the most significant such organizations. Although King's article aims more at describing a research program than at establishing definitive conclusions, he speculates that the results of research will reveal that political scientists commonly confuse an idealized image of strong parties in a well-functioning democracy with the role that real parties play in real democracies.[71]

Przeworski, too, focuses on parties, and like King emphasizes that parties offer only limited help to citizens in controlling government. But rather than tracing those limits to defects in party organization and strategy, Przeworski, like Lindblom, locates their roots in the basic structure of capitalist democracy.[72] He offers an historical account of how the structural constraints of capitalist democracy caused a transformation of socialist into social democratic parties. In the nineteenth century, revolutionary worker's movements chose parliamentary democracy to advance their goals. Given the material uncertainty of workers resulting from their lack of control over their product, together with the possibility of material gains through the state, no proletarian party could survive under democratic conditions unless it made that choice. Once having made it, socialist parties faced an electoral dilemma: because workers were everywhere a minority, their parties could either preserve their class identity while remaining in a permanent minority status, or redefine their identity in universalistic terms in the hope of winning an electoral majority. Desiring victory, they chose redefinition. Because of that redefinition, workers were no longer organized as a class, and the political relevance of class identity faded. Keynesianism made the reform path even more compelling by offering socialist parties a program of regulation and distribution that favored their constituents' class interests, without requiring a transformation of capitalism itself. Unable to change the structure of capitalist democracy, socialist parties pursued a strategy that revealed more about the structural constraints on their conduct than the aspirations of their supporters.

Lani Guinier suggests a different social constraint on the democratic character of electoral institutions. She considers the performance of majoritarian, winner-take-all electoral institutions in a racially-divided society characterized by racially polarized voting, and a relatively small and geographically dispersed minority.[73] Given this background, majority rule produces racially-exclusionary results. In the United States, the principal strategy for addressing this inequity has been to correct violations of the Voting Rights Act by establishing majority-black electoral districts.[74] While this strategy of "race-conscious districting" has the virtue of ensuring Black elected officials, it also frustrates central democratic values: it reduces political competition, discourages citizen participation, stifles debate, and limits opportunities for political bargaining and deliberation.

As an alternative to race-conscious districting, Guinier proposes a variant of proportional representation. She proposes a system of cumulative voting in which each voter may cast a number of votes equal to the number of seats (if there are three seats on the City Council, each voter gets three votes), and can distribute those votes as she pleases.[75] For example, a voter can give all three votes to a single candidate. Under such a system, voters can form themselves into more voluntary constituencies rather than being assigned by the state to racially-defined constituencies (if Blacks wish to vote as a bloc, they can by simply putting most of their votes on a single Black candidate); electoral results reflect intensities of support and not simply numbers; and electoral competition increases because incumbents are easier to challenge. Guinier is, then, hopeful about the role of electoral institutions in forming and expressing a reasonable collective will, but only if electoral arrangements are suitably adjusted to the social context of their operation.

Direct Democracy

At least since Rousseau, one strand of modern democratic theory has favored direct, participatory democracy as a way to remedy the deficiencies of conventional representative institutions.[76] For historical antecedents, proponents of direct democracy point to ancient Athens, which was ruled by a citizen assembly and in which citizen officials rotated through public offices. Analytically, they argue that direct democracy has four principal attractions: it gives manifest institutional life to an ideal of political self-government in which citizens are sovereign and subject, authors of the laws with which they are expected to comply;[77] it promises a higher level of civic engagement and encourages a deeper appreciation of the intrinsic importance of public life ("public affairs dominate private ones in the minds of the citizens"[78]); it encourages greater accountability and responsiveness of government to citizens, who directly supervise government conduct; and it enhances political equality by effacing the distinction between a political class and the subjects over whom they rule: "The instant the people is legitimately assembled as a sovereign body, all jurisdiction of the government ceases, the executive power is suspended, and the person of the humblest citizen is as sacred and inviolable as that of the first magistrate."[79] Acknowledging the difficulties of participatory institutions in a mass democracy, defenders of such arrangements nevertheless argue that the benefits of participation offer a way out of the political alienation experienced by mass publics in representative democracies.[80]

Critics of direct democracy urge that its deficiencies outweigh these alleged benefits.[81] On the historical side, they point to the exclusivity of the Athenian system, in which citizenship rights were not extended to slaves, women, or residents without two Athenian parents. Add to this narrowness the sheer differences in political scale between Athens, with 40,000 citizens, and any modern state, and it should be apparent that direct democracy is not suited to contemporary conditions.[82] Furthermore, critics challenge the analytical case for direct democracy. Although the majority may author the laws, others must simply obey laws authored by others, thus limiting the achievement of autonomy;[83] civic engagement is simply one value, and political life ought not to be organized around its presumed supremacy;[84] and direct democracy may purchase an increase in accountability and responsiveness at a severe cost in competence. Finally, referenda and initiatives -- the principal modern alternatives to participatory assemblies as forms of direct democracy[85] -- allow people to register their opinions directly, but, by requiring a yes/no vote, undermine the political deliberation required for eliciting reasoned popular judgments and sensible political results.[86] As this reference to limits of "reasoned popular judgments" indicates, some critics of direct democracy reject it not for being overly democratic and insufficiently attentive to the need to limit the state's power, but as a poor strategy for institutionalizing rule by the people. Such rule requires a framework of public discussion, not simply the episodes for registering preferences and opinions that are ensured by referenda and initiatives.

The essays in the sixth section debate these issues in light of recent and historical experience with less mediated forms of democracy such as participatory forums, referenda, recalls, and mandates.

Assume, then, that the direct experience of self-government is of considerable importance, for all the reasons advanced by defenders of direct democracy. Suppose, too, as the critics say, that highly participatory forms of democracy cannot succeed on the scale of the nation-state -- that large-scale requires political representation, rather than the direct discussion between citizens and political leaders that is essential to the classical democratic experience. Is there a way to accommodate both observations? Dahl offers a qualified affirmative in his case for the medium-sized city as a fundamental unit of democratic participation. His support is qualified in that he does not look to eliminate the nation-state: some large problems cannot be addressed on a smaller scale, and, given the scale, representation is mandatory. Still, faced with widespread disillusionment about national-level democratic institutions, Dahl argues that the democratic city is the optimum scale for realizing the values of direct participation; the alternatives -- committees, workplaces, villages -- are too small, too narrowly focused, too confining. Cities are small enough to allow widespread communication between citizens and officials, and to socialize citizens into civic virtues, yet large enough that their policies make a difference to important aspects of our lived environment: education, housing, public safety, health, and social equality.

Poulantzas and Bobbio, both critics of capitalist democracy and Soviet authoritarianism, evaluate direct democracy within a democratic socialist framework. Although Poulantzas criticizes Lenin's proposal for an entirely rank-and-file, soviet democracy -- an "exclusive reliance on council democracy and complete elimination of representative democracy" (253) -- he argues that direct democracy should play an important role with a national framework of representative democracy. In criticizing a pure council democracy, he agrees with Rosa Luxemburg,[87] who argued that, in the absence of democratic institutions at the national level, democracy based on workers' councils would degenerate into authoritarian rule. But Poulantzas is hostile to the "state-worship" associated with Stalinism and social-democracy, and so proposes a strategy that would extend democratic political rights at the national level, reinvigorate local institutions of participation and self-management (factory committees, citizen committees), and recognize the interdependence of such extension and reinvigoration. In the end, Poulantzas is less specific than Dahl about the appropriate sites of direct democracy. But that is in part because he is concerned less about finding the optimal participatory unit, and more about simply ensuring a substantial role for the practice.

Bobbio considers several alternatives to representation -- referenda, assemblies, and mandates -- each of which has been proposed as a device to give content to the socialist ideal of direct democracy. He is critical of each. Agreeing that referenda are sometimes desirable, he highlights the implausibility of using them to decide the vast range of issues now addressed by parliaments. Moreover, assembly democracy is incompatible with the size of modern states; and revocable mandates are either undesirable -- if the right to revoke is vested in a sectional interest -- or acceptable -- if that right is vested in a mediating institution of precisely the kind rejected by advocates of direct democracy. Rather than replacing representative with direct democratic arrangements, then, Bobbio suggests that many socialist criticisms of representative democracy could be answered by extending the authority of representative governments to economic matters.

Much of the debate about direct democratic arrangements under modern conditions is speculative -- like the discussions by Dahl, Poulantzas, and Bobbio. Gordon Smith proposes a scheme for a more systematic, empirical study of the political importance of referenda. Emphasizing widespread disagreement about their virtues, he distinguishes two particular dimensions of disagreement: whether referenda are typically initiated by political elites (are they controlled or uncontrolled?) and whether referenda reinforce the status quo or bring change (are they pro- or anti-hegemonic?). Although few referenda are both uncontrolled and pro-hegemonic, Smith finds examples that fit the other three categories. The diversity of functions served by these different types of referenda make it difficult to assess the desirability of the institution as a form of collective choice, without reference to contextual details.

The essays by Bell and Lowenstein do not rely on Smith's specific analytics. Nevertheless, by examining the implications of direct democracy under conditions of racial and economic inequality, they highlight his more general point about the importance of context in assessing the virtues of direct democracy. Arguments for direct democracy often assume a relatively homogenous polity -- for example, the enfranchised citizens of ancient Athens. Derrick Bell argues that referenda operating in the American context of racial division reinforce subordination: "the more direct democracy is, the more threatening it is" (1). The conclusion turns on the role of deliberation. Representative democracy encourages public deliberation in representative assemblies, which tends in turn to filter out racially-biased views and reduce their impact on legislation; direct legislation, by contrast, does not require citizens to justify their views publicly, nor does it issue from an institution in which reasons, whether sincere or not, are stated in support of views. As a result, "appeals to prejudice, oversimplification of the issues, and exploitation of legitimate concerns" commonly dominate direct-democratic processes. In a social context in which racial prejudice is common, the absence of filters on racist preferences in the direct-democratic setting means that such preferences are more easily translated into law.

Where Bell focuses on unreasonable outcomes resulting from racial prejudice, Lowenstein raises troubles about unfair processes resulting from economic inequality. He rejects the aprioristic argument that direct democratic procedures of collective choice are less susceptible to influence from campaign spending than candidate elections, and tests their relative dependence through an examination of ballot propositions in California. The evidence indicates that differential spending does make a difference to outcomes on ballot propositions when opponents outspend proponents: he finds a substantial range of cases in which measures would likely have passed, had it not been for one-sided spending by the opposition. To correct this unfairness, Lowenstein suggests a combination of public financing and limits on private contributions similar to that proposed by advocates of campaign finance reform in the context of candidate elections.[88]

Group Politics and Democratic Governance

Rights of association protect a fundamental democratic liberty: to join with like-minded others in pursuit of shared interests or common aspirations. But the constructive contributions of secondary associations -- the groups intermediate between family and state that result from the exercise of those rights -- are a matter of deep disagreement.

Skeptics focus on the problem of "faction." They argue that group formation threatens to harden social inequality and undermine civic identity, replacing it with fragmented and conflicting group sensibilities. To the extent that a political system permits the formation and participation of associations, some citizens -- for example, those with greater wealth -- will find it easier to form organized groups. Groups can then turn this organizational strength into political power, amplifying their initial advantage, and transforming the pluralist bazaar into a self-reinforcing system of social and political privilege. Group particularism only makes things worse. Only some citizens are represented in groups; leaders are (at best) accountable only to their members; and organization encourages individuals to focus on what they share as members rather than on their common interests and concerns as citizens. As a result, particular groups are prepared to promote policies that impose net costs on society, so long as the policies deliver net gains to their own members; and group bargaining and competition replaces civic deliberation about the common good.

One response is that such problems of faction underscore the importance of sharply-defined constitutional limits on government. According to Hayek, the "domination of government by coalitions of organized interests" -- by "an enormous and exceedingly wasteful apparatus of para-government . . . has arisen only in response to (or partly as defense against being disadvantaged in) the increasing necessity of an all-mighty majority government maintaining its majority by buying the support of particular small groups." Such factional domination is "the inescapable result of a system in which government has unlimited powers to take whatever measures are required to satisfy the wishes of those on whose support it relies."[89] Skeptical about the efficacy of such constitutional solutions, but also troubled by the pathologies of group politics, Rousseau concluded that there ought to be either sufficiently many groups to nullify the effects of each -- a perfectly competitive pluralist bazaar -- or "no partial society in the state." For "when intrigues and partial associations come into being at the expense of the large association, the will of each of these associations becomes general in relation to its members and particular in relation to the state."[90] Madison, too, saw the dangers of faction but thought it futile to attempt to eliminate these groups or even to "adjust these clashing interests, and render them all subservient to the public good." His prescription for the "mischief of faction" was, in effect, to insulate the state from undue pressure through a combination of representative government and a large population. Minority factions would be controlled by the requirement of winning majority support, and majority factions by creating a system that includes a large variety of parties and interests, thus making it "less probable that a majority of the whole will have a common motive to invade the rights of other citizens; or if such a common motive exists . . . more difficult for all who feel it to discover their own strength, and to act in unison with each other."[91]

Tocqueville's observations of associations in the United States suggest an alternative, more optimistic line of thought about associations: that the "civilizing" effects of widespread participation could outweigh the dangers of faction. "It is through political association," he argued, "that Americans of every station, outlook, and age day by day acquire a general taste for association. . . . Through them large numbers see, speak, listen, and stimulate each other to carry out all sorts of undertakings in common. . . . In this way, by the enjoyment of a dangerous liberty, the Americans learn the art of rendering the dangers of freedom less formidable."[92] More generally, this second line of argument emphasizes the constructive role associations can play in promoting the general welfare and achieving fair equality in a democracy. They provide information to policy-makers regarding member preferences, the impact of proposed legislation, or the implementation of existing law; they provide representation to their members, and equalize representation across different citizens; they can provide a means of social ordering alternative to markets or public hierarchies, one that can promote trust and coordination and so permit society to achieve important benefits of cooperation among member citizens; and they can function as "schools of democracy" by developing the confidence and democratic disposition of members. The task is to capture these large benefits, while avoiding the destructive tendencies to faction.

The articles by Dahl and Schlozmann explore the dimensions of the problem of faction. Dahl sketches the roots, implications, and possible remedies for the troubles created by the modern phenomenon of organizational pluralism. He traces such pluralism to the growth in territorial scale associated with modern states. The change from ancient city-state to modern nation-state necessitated a dramatic shift in the institutions of political equality from classical "monistic" democracy to "polyarchy" -- a system that disperses control over political leaders, and ensures to citizens the rights and capacities required for exercising that control. The increased scale associated with polyarchy, and the liberty of association characteristic of it, imply, in turn, that polyarchy will be characterized by social pluralism -- social diversity and conflict, multiple allegiances of citizens, and a variety of organizations independent of the state.[93] Associative pluralism is, then, inevitable in democratic regimes; it is also desirable inasmuch as associations are one of the instruments through which citizens control government. These benefits notwithstanding, pluralism threatens to stabilize inequality, deform civic consciousness, distort the political agenda, and vest final control of the agenda in particular groups. Thus any democracy will need to develop political strategies for addressing these threats. But while the fundamental "dilemma of pluralist democracy" -- that independent associations are both essential and threatening to democracy -- can be more or less successfully addressed, it is part of the terrain of democracy itself.[94]

Kay Schlozmann pursues one of the troubles created by organizational pluralism. She asks whether representation through interest organizations (political pressure groups) equalizes political influence or simply reflects economic and social inequalities. Her answer, based on a study of 7,000 registered Washington lobbying organizations, is critical of the achievements of pluralism as a device for equalizing political representation. She finds strong support for E.E. Schattschneider's contention that "The flaw in the pluralist heaven is that the heavenly chorus sings with a strong upper-class accent. Probably about 90 percent of the people cannot get into the pressure system."[95] More precisely, business and professional interests constitute fully 88% of all Washington pressure groups. Despite the proliferation over the past 30 years of organizations devoted to representing historically underrepresented interests, economically advantaged citizens continue to receive -- as advocates of strong parties have always argued -- better representation in the pressure system than any other group.

Laski, Schmitter, and Cohen and Rogers discuss different strategies for capturing the benefits of group organization through deliberate acts of political construction. Departing from Tocqueville's proposition that a healthy democracy requires the cultivation of habits of participation, Harold Laski argues that geography is increasingly less relevant to the definition of interests and therefore to the appropriate sites of participation in modern societies. So he urges the replacement of territorial representation with a system of group representation based on industrial function.[96] Under received democratic institutions, "the average elector, except in times of crisis or abnormal excitement, is but partially interested in the political process," because so much energy is consumed at making a living, while political representation is removed from the workplace. Three major failings stem from reduced working class participation: workers suffer from a government that principally serves the interests of wealthier citizens; the quality of administration and government suffers because it lacks the vital energy of citizens' creativity and instead leads to "a curious intellectual inertia which suggests too rigid stratification;" and citizens themselves are deprived of "the credit of their humanity" because "the offices of men do not make their souls erect and their minds intelligent." A form of representation based on industrial "function," and the creation of an "economic sovereignty" that coordinates separate economic units, would address these failings by creating arenas of democratic participation more organically connected to the lives of ordinary citizens.

Neo-corporatism is a second response to the shortcomings of informal pluralism, and, like Laski's industrial democracy, it would also establish more official group representation. Here, Phillipe Schmitter assesses its democratic credentials.[97] Neocorporatism is a form of "intermediation between interest associations and interlocutors" (897), in which hierarchically organized "peak" associations -- typically organizations representing the interests of labor and capital -- are recognized and licensed as monopoly representatives of certain interests.[98] The resulting arrangement insulates some arenas of public decision-making -- classically, wage-bargaining and macro-economic policy -- from electoral mechanisms. But while neocorporatism arguably contributes to effective problem-solving, it also appears to conflict with basic democratic norms. Precisely because it formally privileges certain interests and organizations -- deliberately elevating some social interests to the exclusion of others -- it opposes the equal representation and widespread citizen participation that define political democracy. Furthermore, these problems cannot be addressed by simply extending the corporatist form to conventionally underrepresented interests. Such inclusion would require extensive state intervention, would be rejected by groups whose identity is incompatible with official representation, and would ultimately undermine the very conditions "of small-group interaction, specialized competence, reciprocal trust, and propensity for compromise" (918) that make the neocorporatist formula successful.

Agreeing with Dahl that secondary associations inevitably and reasonably play a role in the politics of mass democracies, with Dahl and Schlozmann that existing forms of pluralism create problems of inequality and particularism, and with Schmitter about the deficiencies of corporatism as a form of democratic governance, Cohen and Rogers suggest a way to encourage a constructive role for secondary associations in improving democratic process and performance.[99] They describe an associative democratic strategy for advancing norms of popular sovereignty, political equality, distributive equity, civic consciousness, economic performance, and state competence. The associative model has secondary associations playing a large role in ensuring a fair representation of interests and participating in problem-solving, governance arrangements that provide alternatives to command-control regulation.

Such favorable associational conditions do not arise naturally; problems of faction are serious, and associative democracy must be aimed at to be achieved. So Cohen and Rogers suggest a deliberate use of public powers that would adjust the "terms, conditions, and public status of groups" (395) with the aim of creating a more favorable associative-democratic setting. Where manifest inequalities in political representation exist, they recommend promoting the organized representation of presently excluded interests; where group particularism undermines democratic deliberation or popular sovereignty, they recommend efforts to encourage the organized to be more other-regarding in their actions; and where associations have greater competence than public authorities for achieving democratic ends, or where their participation could more effectively promote political values, they recommend a more formal governance role for groups. Although it is no prescription for a minimal state, associative democracy would shift the balance of affirmative state conduct from direct regulation to the construction of deliberative arenas in which associations themselves address issues of public concern.[100]

Constitution, Democracy and Administrative Power: The U.S. Case

Modern welfare states present a serious challenge to classical understandings of constitutional government and democracy. Such states aim to correct market imperfections and provide social welfare. The complexity of those tasks leads legislatures to pursue them by dolloping out discretionary authority to administrative agencies. But, according to critics, the power granted to agencies undermines important rule-of-law values and democratic protections. For example, agencies interfere with the exercise of individual rights by intervening in essentially private matters; they use highly discretionary modes of decision-making, thus depriving individuals of the procedural protections associated with the rule of law; they violate the separation of powers by simultaneously exercising legislative, executive, and judicial authority; and, with administrators operating under vague delegations shielding their conduct from legislative oversight, they frustrate the popular accountability required for democratic legitimacy.

Theodore Lowi's The End of Liberalism provides the classic statement of these concerns about constitutional and democratic governance in the context of the American welfare state.[101] Writing in 1969, Lowi described a broad liberal-conservative consensus on a public philosophy of "interest-group liberalism" which provides the framework for the affirmative American state. This brand of liberalism reduces political process to interest-group bargaining, the role of government to facilitating deals and enforcing their results, and the common good to the bargains struck by organized interests. Interest-group liberalism "feudalizes" the American state by encouraging groups to compete to appropriate the discretionary power granted to administrative agencies. The public power captured by groups operates free from the constraints of democratic accountability and the rule of law, thus rendering effective governance impossible. Articles in this section explore the problems crystallized in Lowi's work: how might the troubles created by administrative discretion be remedied, and are they as severe as Lowi suggests?

Carl Friedrich sets the stage with his defense of agency independence. Writing in the wake of the New Deal expansion of the federal bureaucracy, he urges such independence as an alternative to a "transmission belt" model of administration which would require agencies simply to implement legislative commands. That latter model is defeated by the complexity of the tasks facing the modern state: the directives that legislatures issue to their bureaucratic agents are inevitably incomplete and leave a large residue of discretionary power. Furthermore, the transmission belt model of the state presumes a sharp conceptual distinction between policy formation and execution which does not fit political reality: "Public policy, to put it flatly, is a continuous process, the formation of which is inseparable from its execution. Public policy is being formed as it is being executed, and it is likewise being executed as it is being formed" (6). The conception of administrative responsibility as obedience to legislative commands cannot withstand these two objections, so Friedrich urges that it be replaced by an expectation of official responsiveness to a dual standard of technical knowledge and informed popular sentiment.

But when this dual standard fails -- when technical knowledge is contested and popular sentiment is divided and lacks determinate institutional expression -- administrative discretion can no longer claim legitimacy by reference either to expertise or democratic authorization. One response to this crisis within administrative law was to increase the representation of interests in agency decision-making. Richard Stewart's "Reformation of American Administrative Law" explores this interest-representation strategy. Faced with an "enormous expansion of governmental activity, not only in the regulation of private activity, but also in the provision of goods, services, and advantageous opportunities" (1811), an inability of legislatures to control administrative discretion, and citizen demands of agency accountability and responsiveness,[102] courts used a variety of devices to ensure citizens themselves greater access to administrative decision-making. The idea was to invite pluralism into administration, thus restoring legitimacy by turning agencies into a second tier of quasi-legislatures, whose decisions would reflect fair bargaining among represented interests.

Stewart criticizes this strategy on four grounds: it increases costs of decision-making; it creates troubles for the rule of law by generating decisions for particular cases that do not easily generalize into legal rules; it encourages excessive legal intervention; and it fails to address problems of bias and private capture of agency power, in part because it inherits pluralism's inability to make principled decisions about which groups are to be represented. Stewart speculates that the interest-representation model of administrative law, which assumes circumscribed public power and private autonomy, might ultimately give way to "a new and comprehensive theory of government and law that would successfully reconcile our traditional ideas of formal justice, individual autonomy, and responsible mechanisms for collective choice with the contemporary realities of decentralized, uncoordinated, discretionary exercises of governmental authority and substantial disparities in the cohesiveness and political power of private interests" (1807). However, he acknowledges that such a model -- whose content is not further specified -- is unlikely to emerge soon, and pins near-term hopes on a blend of strategies, including more discrete uses of interest-representation, limits of agency authority, judicial review, and increased legislative oversight.

One response to these troubles about delegation -- suggested in the Cohen-Rogers account of associative democracy (see pp. 000-000) -- would assign a larger role to associations themselves, while transforming them from pressure organizations and bargaining partners into problem-solvers.[103] Another, which Friedrich and Stewart reject in their quests for administrative accountability, would revoke broad delegations and renew political control of agencies by imposing demands of clarity, stability, and generality on legislation. Lowi himself endorses this rule-of-law prescription, which he calls "juridical democracy,"[104] and argues that it is necessary to counter the enormous constitutional risks of broad delegation. He emphasizes that broad agency discretion undermines control of government by the governed, and replaces it with a system of patron-client relations, the "moral equivalent" of serfdom.[105] More particularly, such discretion encourages an overweening executive, weak drafting of laws, excessive deference of courts to agencies, and over-regulation. But Lowi rejects the Hayekian claim that these concerns mandate a return to minimal government. Believing that broad delegations are simply one particular strategy for achieving more ambitious goals -- a contingent product of political history rather than the inevitable result of regulatory complexity -- Lowi endorses juridical democracy as a way to combine constitutional and democratic norms with an affirmative state. Though he doubts that there is sufficient political will to achieve this combination, he urges a renewal of rule of law values, with Congress supplying and codifying clear rules, and Courts enforcing requirements of clarity by selectively invalidating overbroad or excessively vague delegations.

Cass Sunstein carries forward this ideal of juridical democracy by sketching a constitutionalism suited to a post-New Deal affirmative state.[106] He distinguishes the intellectual motivation of the New Deal project from its institutional expression, and proposes that we embrace the former but reject the later. Intellectually, that project accepted the legal realist critique of common law rights of property and contract as pre-political baselines for assessing state action, and, correlatively, rejected the idea that to protect those rights was simply to act as the neutral, impartial enforcer of a natural ordering.[107] Recognizing the status quo itself as a political ordering that could not be justified by appeals to fairness or efficiency, New Dealers insisted on the need for deliberate state action to address market failures and achieve a more just distribution of resources. Sunstein's constitutionalism affirms this intellectual component of the affirmative state and extends it from the economy to understandings of rights in areas of race and gender.

On the institutional side, the New Deal largely rejected federalism and the separation of powers as holdovers from an era of limited government and formalist obstacles to an affirmative state. Put more constructively, they nationalized government and created a system of administrative agencies operating with considerable discretionary power. Sunstein is critical of both centralization and agency independence, urging in particular that the New Deal reform of agencies was "largely a mistake" (452). We now live with the consequences of that mistake: unresponsive agencies vulnerable to capture and an eclipse of the Madisonian ideals of participation and deliberation. To correct it, Sunstein's positive program would reestablish legislative, judicial, and executive oversight of agencies: Congress would set the ends for policy, courts would look closely at the beneficiaries of agency rule-making, and the executive would ensure their budgetary soundness. Restoring federalism, separation of powers, and the rule of law would preserve a New Deal/social-democratic vision of positive government action, while insulating the state from group politics, and subjecting agencies to democratic control by more sharply defining their tasks and shining a brighter light on their operations.[108]

In "Congressional Oversight Overlooked: Police Patrols versus Fire Alarms," McCubbins and Schwartz reject Lowi's description of the contemporary relationship between congress and agencies, and, by implication, oppose reform proposals that assume that unbounded administrative discretion is a large problem. Although they agree that Congress does not perform its oversight function through the standard "police patrol" machinery of centralized monitoring, they criticize conventional views for confounding the failure of this particular mechanism with a failure of legislative oversight generally. According to McCubbins and Schwartz, Congress has developed effective alternative oversight mechanisms that rely upon feedback from citizens and organized interest groups -- those mechanisms operate as fire alarms rather than police patrols. These "rules, procedures, and informal practices," which facilitate citizen complaints about administrative misbehavior, are effective because they insure that the worst abuses will come to light and allow initially-vague legislative goals to be elaborated in light of citizens' grievances. If McCubbins and Schwartz are right, then dysfunctions of the affirmative state should not be blamed on agency power, but perhaps on general deficiencies of democratic collective choice, or on problems with any form of state intervention in markets, or on the limits of the command-and-control regulatory style associated with fire-alarm enforcement.

Constitutionalism, Democracy, War

One of the great challenges to constitutional democracy, with its ideal of government by the consent of the governed, is to forge suitable institutions for controlling decisions about national security and war. The war-making tendencies of rulers operating independently of constitutional and democratic constraint provided an historically important motivation for imposing such constraints.[109] And those constraints appear to have had at least some benefit: since 1800 established constitutional democracies that are also major powers have not gone to war against one another.[110] But democratic states still go to war, and that decision is still a state's most extreme action in the community of states. Moreover, no other policy imposes so heavy a burden on citizens or is so consequential for domestic institutions. Because of this special gravity, the case for limiting the war-making power, and subjecting it to the will of the people, seems particularly strong.

Since Hobbes, however, defenders of a strong executive have been quick to remind us that wars are emergencies, threatening to the survival of citizens and peoples. They may require political authorities to act in secret, and with a decisiveness forbidden by the slow-grinding constitutional and democratic machinery devised for normal conditions. To be sure, emergency measures may threaten constitutional democracy, but governments unwilling to take that risk put the state itself in jeopardy.[111] Fretting over procedural formalities, they endanger the lives of citizens, thus failing to discharge their elementary political responsibility. Abraham Lincoln wondered in 1861 whether a republic must "of necessity, be too strong for the liberties of its own people, to too weak to maintain its own existence."[112]

Since the late 1940s, these general concerns have become especially pointed in the United States. The world's largest military power, the United States also has longstanding traditions of popular and constitutional suspicion of centralized executive power. In the postwar period, those suspicions have repeatedly been provoked by an imperial presidency, national security ideology, and routine military interventions (justified during the Cold War by reference to the Soviet Threat, justified since by a host of considerations ranging from the health of the domestic economy, to international drug trafficking, to the requirements of a New World Order). The three articles in this section probe America's troubled attempts to reconcile its military power with the requirements of constitutional, democratic government.

Gerhard Casper's "Constitutional Constraints on the Conduct of Foreign Policy" addresses the problem of executive power in light of the near-constitutional crisis occasioned by the US invasion of Indochina, and the subsequent conduct of the Vietnam War. Casper emphasizes two central elements of American constitutional doctrine and practice whose combined effect has been to obstruct constitutional constraints on Presidential conduct of foreign policy. On the one hand, an abiding feature of U.S. constitutional practice is an almost exclusive reliance on courts as sources of constitutional interpretation and limits. On the other, the Supreme Court has judged certain areas of government -- foreign and security policy in particular -- to present essentially "political questions" that lie beyond the reach of judicial competence. The result is an "absence in practice of constitutional constraints and the presence of few political constraints" on presidential decisions about foreign and defense policy (463). Because foreign policy requires proper coordination between executive and legislative powers, checks and balances are essential. Casper argues that they can be ensured through a non-judicial brand of constitutional interpretation, with the legislative branch taking the responsibility to define the roles of separate constitutional powers through quasi-constitutional "framework statutes."[113]

Jules Lobel is considerably less sanguine about the prospects for controlling executive war powers through such institutional innovation. According to Lobel, liberal constitutionalism depends on a bright-line distinction between constitutional order and emergency power. While recognizing that extreme emergencies might require resolute executive action, the liberal solution judges it best not to make constitutional provision for such action by including an emergency powers clause within the constitution itself. Constitutionalizing emergency powers reduces the costs of, and thus threatens to multiply, declarations of emergency; by contrast, rejecting an emergency powers clause preserves a sharp distinction between constitutional and emergency authority, thus encouraging the executive to declare the need for crisis powers only when faced with genuinely extraordinary and grave threats to the nation. The substantive precondition of this institutional strategy, however, is infrequent military entanglements. As the frequency of military conflict increases, states become less able to treat military emergencies as constitutionally exceptional.

This liberal framework served the United States well through nineteenth century and the first part of the twentieth. Since the Second World War, however, the regular deployment of U.S. forces[114] has transformed practice and doctrine: it has in effect generated a state of permanent emergency, a constitutional doctrine of substantial, inherently-exclusive executive power in foreign affairs,[115] and the lack of accountability noted by Casper. But this expansion of executive power has been driven "less by legislative failure or by executive arrogance, than by the transformation of the eighteenth century world and the rise of American power in global affairs" (1386). For this reason, reestablishing domestic liberalism requires much more than institutional reform; it demands a substantive change of policy in which the United States embraces a less aggressive posture toward other states.

Like Lobel, Elaine Scarry argues that a suitable constitutional framework for U.S. foreign and military policy requires a change in the substance of that policy: in particular, policy allowing first-use of nuclear weapons. But she goes further than Lobel in rejecting the view that wars present constitutional emergencies in which requirements of citizen consent are suspended in favor of extra-constitutional executive directives. More immediately, she argues that the U.S. constitution rejects that premise, and instead makes the legitimate use of military force contingent on active citizen consent. That dependence is manifest in Article I, section 8, which invests in Congress the power to initiate war through declaration. But it is also ingredient in the second amendment guarantee of the right to bear arms. Not simply a protection for hunters, this guarantee disperses the practical right to authorize war by mandating a broad distribution of control over the instruments of war. As Meiklejohn and Fiss argue that a constitutional free speech guarantee helps to constitute democracy by empowering citizens to govern the government, and the ambitious theory of electoral control argues that strong parties help to constitute democracy by establishing a coherent framework of popular electoral choice, Scarry interprets the right to bear arms as constituting democracy by empowering citizens to regulate the conduct of military force. With a just distribution of means of coercion, the state can only make war by eliciting agreement from citizens to place the material conditions of war -- their own bodies and weapons -- at the state's disposal.

Modern weapons, especially nuclear devices, threaten the second amendment's requirement of ongoing citizen consent to war by radically concentrating control of the means of violence. U.S. nuclear policy, which empowers the President to initiate the use of atomic weapons, is especially troubling in its extreme destructive potential and concentration of authority. By vesting such an extraordinary power in the executive, current policy undermines the core of the constitution as a design of consensual, democratic governance. While affirming the importance of institutional design to democratic success, then, Scarry reminds us, too, that even well-designed institutions may not be able to withstand the pressure of indecent policy.

Beyond the State

The articles in the final section explore relationships of authority and power outside the state. An implicit theme running through these articles is that the institutions of justice extend beyond the state, to include corporations, families, associations, and international organizations. Because these sites lie outside the traditional public sphere, many have not considered them appropriate subjects of constitutional and democratic concern. But they exert substantial influence on the achievement of justice, in part because they can, in Dahl's phrase, pose "impediments to democracy." The contributions in this section therefore underscore the importance of extending democratic theory and practice beyond the state.

Carole Pateman juxtaposes a consensual theory of legitimate political authority with the institution of the family and the social position of women. She argues that the requirement of consent to political authority is best understood as one part element of a "broader voluntarist theory of society" that presents individuals as free and equal, and holds "that relationships of authority and obligation must be grounded in the voluntary acts or commitments of individuals," understood as free and equal (72).[116] Such voluntarism appears to have radical social and political implications: if obligations, quite generally, require consent, and genuine consent depends on background conditions of freedom and equality, then much of social life needs reordering. But liberal political theory avoids those implications by arbitrarily circumscribing the individuals whom it represents as free and equal, and arbitrarily limiting the application of the voluntarist principle to more formal, political arenas.[117] Liberal conceptions of marriage and of the social position of women illustrate the arbitrary exclusion of women from the voluntarist principle. The presumption that women are not "free and equal individuals" capable of consent is revealed in classical liberal defenses of the patriarchal family and contemporary rape and marriage laws -- in the latter case, by treating men's judgments of reasonable conduct as authoritative. A reformed consent theory would take seriously that women have such capacity, and, therefore, would embrace the reconstruction of the family as one step in ensuring the circumstances of freedom and equality required for genuine consent.

The next two essays explore the role of power in social relations that are commonly considered voluntary and thus coercion-free, and argue for formally democratizing those relations. Grant McConnell focuses on the exercise of power within private associations -- for example, unions, professional associations, and trade groups. As the articles in section 7 indicate, such associations play an essential role in organizing democratic society and are sometimes vested with public powers in settings of licensing, rule-making, and policy formation. But their internal governance is commonly undemocratic: executive-dominated, unconstrained by member rights, and lacking in an institutionalized form of leadership competition. Because these associations affect the welfare of their members and commonly exercise coercive powers over them, leaders should be made accountable to members. McConnell suggests that such traditional constitutional mechanisms as rights and representation, usually reserved for formal public institutions, could help to ensure accountability. Establishing those mechanisms requires, however, that we first break from false assumptions about the homogeneity of group membership and the majoritarian conceptions of organizational democracy that follow from those assumptions.

Radical democratic thought has consistently rejected the idea that a market economy is an arena of private ordering unsuited to democratic norms, and argued instead for an extension of democratic norms from the state to firms. That argument depends on two key premises: that arrangements of democratic accountability are appropriate wherever there are stable asymmetries of power, and that such asymmetries characterize employer-employee relations in market economies. In "Power and Wealth in a Competitive Economy," Bowles and Gintis defend the second premise by arguing against the claim that competitive markets are power-free arenas.

In the Walrasian model of a competitive economy, markets clear and workers can move costlessly between positions; to be sure, equilibrium wages may be low and the distribution of income grotesque, but workers' exit options deprive employers of any power over them. That model depends, however, on the implausible assumption that employers can costlessly monitor employee performance. In the presence of monitoring costs, profit-maximizing firms will increase wages; higher wages mean a greater cost of job loss, and that creates incentives for workers to exert greater effort to avoid losing their jobs. Such "efficiency wages," however, generate unemployment. And the combination of unemployment and higher-than-market-clearing wages gives employers power: they can impose the costly sanction of firing if workers do not work hard enough. Because the ideal competitive market itself is an arena of power, standard arguments for democratic accountability in public institutions also apply to firms operating within it, whatever their juridical status.[118]

Robert Dahl's "On Removing Certain Impediments to Democracy" presents a broader program of democratic reform. He describes how U.S. commitments to corporate capitalism, social welfare, and international power clash with an ideal of procedural democracy.[119] Corporate capitalism poses obstacles to procedural democracy by generating substantial social and economic inequality, while ensuring that the world of work is governed hierarchically rather than democratically. The creation of a national welfare state during the New Deal, although intended to reduce the dependence of citizens on the vicissitudes of the market, further enhanced presidential power and brought forth expansive state bureaucracies. Finally, superpower status extends the scope of the imperial presidency, and requires a third organizational hierarchy, the military establishment. Dahl argues, then, that we face a choice between these three commitments and a commitment to democracy, that procedural democracy should take priority over the others, and that its realization requires the reorganization of political and economic institutions to promote equality and openness.

We end by stepping entirely outside the domestic system. In the concluding article, Robert Johansen explores the relationship between international security and democracy. Inspired by the possibilities created at the end of the Cold War, but disheartened at the lack of international movement to exploit this enormous opportunity, Johansen proposes a democratic conception of security to replace the competitive nation-state system. His scheme focuses on individuals rather than states, emphasizes the need for mutual respect rather than narrow self-interest, and broadens security concerns to encompass environmental and economic, as well as military, issues. Using this reconstructed notion, Johansen argues that there is a false dichotomy between security and other elements of social welfare. Security, properly understood, is best advanced through policies that foster reciprocity, equity, sustainable economic development, and demilitarization of international dispute resolution. Most importantly, security depends on a project of democratization that would strengthen domestic democratic institutions, increase accountability across borders, and democratize international institutions themselves. Johansen does not explain how the power vested in those new institutions is to be limited, made accountable, or directed to its proper ends, but that task is perhaps best left to a future, cosmopolitan political theory.