Deliberation, Social Choice and Absolutist Democracy
Social choice theory and theories of deliberative discourse have ...
9 downloads
1248 Views
1007KB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Deliberation, Social Choice and Absolutist Democracy
Social choice theory and theories of deliberative discourse have strongly influenced the way political scientists understand the dynamics of democratic politics and decision-making. This new book addresses the dispute between these competing schools of thought. Deliberative democrats and social choice theorists offer the two dominant conceptions of participation in contemporary democratic theory. The former hold that through the democratic process we can arrive at consensus, rational outcomes and even principles of justice, while the latter suggest that fair and equal participation is more likely to lead to instability and irrational outcomes. With an in-depth examination of social choice theory and deliberative democracy, David van Mill: • • •
Presents two case studies on the American Continental Congress 1774–1789. Provides an assessment of the types of institutions that will promote democracy and create stable outcomes with minimum sacrifice of the freedom and equality of participants. Defends a more radical idea of absolutist democracy, gleaned from the writings of Hobbes, against the claims made in favour of limited constitutional government.
This book will be of interest to students and researchers of political theory, particularly those with an interest in democratic theory. David van Mill is Senior Lecturer in the Department of Political Science in the University of Western Australia.
Routledge innovations in political theory
1 A Radical Green Political Theory Alan Carter 2 Rational Woman A feminist critique of dualism Raia Prokhovnik 3 Rethinking State Theory Mark J. Smith 4 Gramsci and Contemporary Politics Beyond pessimism of the intellect Anne Showstack Sassoon 5 Post-Ecologist Politics Social theory and the abdication of the ecologist paradigm Ingolfur Blühdorn 6 Ecological Relations Susan Board 7 The Political Theory of Global Citizenship April Carter 8 Democracy and National Pluralism Edited by Ferran Requejo 9 Civil Society and Democratic Theory Alternative voices Gideon Baker 10 Ethics and Politics in Contemporary Theory Between critical theory and post-marxism Mark Devenney 11 Citizenship and Identity Towards a new republic John Schwarzmantel
12 Multiculturalism, Identity and Rights Edited by Bruce Haddock and Peter Sutch 13 Political Theory of Global Justice A cosmopolitan case for the World State Luis Cabrera 14 Democracy, Nationalism and Multiculturalism Edited by Ramón Maiz and Ferrán Requejo 15 Political Reconciliation Andrew Schaap 16 National Cultural Autonomy and Its Contemporary Critics Edited by Ephraim Nimni 17 Power and Politics in Poststructuralist Thought New theories of the political Saul Newman 18 Capabilities Equality Basic issues and problems Edited by Alexander Kaufman 19 Morality and Nationalism Catherine Frost 20 Principles and Political Order The challenge of diversity Edited by Bruce Haddock, Peri Roberts and Peter Sutch 21 European Integration and the Nationalities Question Edited by John McGarry and Michael Keating 22 Deliberation, Social Choice and Absolutist Democracy David van Mill
Deliberation, Social Choice and Absolutist Democracy
David van Mill
First published 2006 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Simultaneously published in the USA and Canada by Routledge 270 Madison Ave, New York, NY 10016 Routledge is an imprint of the Taylor & Francis Group This edition published in the Taylor & Francis e-Library, 2007. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.”
© 2006 David van Mill All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data van Mill, David. Deliberation, social choice and absolutist democracy / David van Mill. p. cm. — (Routledge innovations in political theory ; 22) “Simultaneously published in the USA and Canada by Routledge.” Includes bibliographical references and index. ISBN 0–415–39092–3 (hardback : alk. paper) 1. Democracy. 2. Social choice. 3. Decision making. I. Title. II. Series. JC423.V345 2006 321.8—dc22 2006002352
ISBN 0-203-08640-6 Master e-book ISBN
ISBN10: 0–415–39092–3 (hbk) ISBN10: 0–203–08640–6 (ebk) ISBN13: 978–0–415–39092–7 (hbk) ISBN13: 978–0–203–08640–7 (ebk)
For Justine
Contents
Acknowledgements
ix
1 Introduction
1
Overview of the chapters 2
2 A comparison of social choice and deliberative theories of democracy
8
Introduction 8 Discourse and democracy 9 Social choice theory 18 Criticisms of deliberative democracy 24 Criticisms of social choice theory 33 Conclusion 35
3 Circumventing Arrow’s theorem Introduction 37 What hope for democracy? 38 Can deliberative democracy and social choice theory be made compatible? 46 Abandoning democratic purity 49 Limiting participation 61 Conclusion 74
37
viii Contents 4
An empirical test of social choice and deliberative theories of democracy
76
Introduction 76 A broad description of the institution 81 Case studies: Administration and governance of the west, and the location of the capital 86 Conclusion 97
5
Hobbesian sovereignty and the spectre of tyranny
99
Introduction 99 Hobbesian sovereignty 100 Hobbes’s intentions 105 Success? 127 Conclusion: Hobbesian absolutism revisited 136
6
Democracy versus constitutionalism
140
Introduction 140 The fear of the majority 141 Free speech 151 Back to constitutions 162 The British Constitution 168 Conclusion: Limits on parliamentary supremacy 171
Bibliography Index
174 183
Acknowledgements
I would like to thank the Journal of Politics for allowing me to reprint the article ‘The Possibility of Rational Outcomes from Democratic Discourse and Procedures’ and the accompanying ‘Reply’ that originally appeared in Vol. 58, No. 3. I would also like to thank the Australian Journal of Political Science for permission to publish sections from ‘Civil Liberty in Hobbes’s Commonwealth’ that first appeared in Vol. 37, No. 1 and for permission to publish ‘An Empirical Test of Social Choice Theories of Disequilibrium’ that first appeared in Vol. 37, No. 2.
1
Introduction
There are two main goals of this book. The first is to address the dispute between social choice theorists and deliberative democrats, and the second is to defend a more radical idea of absolutist democracy (gleaned from the writings of Hobbes) against the claims made in favour of limited constitutional government. Deliberative democrats claim that democratic procedures (of a certain discursive kind) can be defended because they produce outcomes that are consensual and legitimate. Social choice theorists tell us is that it is difficult to justify democracy from a purely procedural point of view because the results are likely to be arbitrary or coercive. The first three chapters of the book examine the merits of these claims. I argue that social choice theorists present a more accurate picture of democracy given certain assumptions about the workings of ‘institution-free’ (IF) settings, but that deliberation is useful in more structured democratic environments. I argue that procedural defences of democracy need to focus on how to attain stability at the minimum cost to freedom and equality. I also recommend that support for democracy needs to focus on non-procedural arguments, and in the final chapter I claim that democracy can be defended because it provides better protection against tyranny than other types of political systems. Social choice theory poses difficulties for liberal democracy, even though most of its supporters are indeed liberals. It does so in two respects. First, as already noted, it shows that procedural defences of democracy are weak. Second, it suggests that we are faced with a choice between anarchy and absolutism, and most liberal democrats want to avoid both extremes. In Chapter 5 I claim that absolutism of a particular type found in the writings of Hobbes is nothing to be feared and should be embraced by supporters of democracy.
2 Introduction
Overview of the chapters Chapter 2: A comparison of social choice and deliberative theories of democracy This chapter focuses on the possibility of reaching rational outcomes from democratic procedures. I examine two dominant and competing conceptions of participation in contemporary democratic theory. These are (a) theories of democratic discourse, which find their strongest expression in the work of Habermas, Cohen, Dryzek and others, and (b) disequilibrium theories of social choice supported by Arrow, Riker, Black and McKelvey, to name just a few. Theories of democratic discourse tell us that through the democratic process we can arrive at consensus, rational outcomes and even principles of justice. They also importantly suggest that preferences can be changed through the process of dialogue. Social choice theorists claim that fair and equal participation is more likely to lead to instability and irrational outcomes. Hence, the arguments seem to be mutually exclusive, and with some exceptions, theorists who work in one field have tended to ignore the conclusions reached in the other. Despite the fact that these arguments seem to head off in different directions, I demonstrate that both theories present almost identical assumptions concerning the requirements for a fair democratic procedure; both demand equal access to debate, the absence of a powerful agenda setter, unrestrained access to raise and object to amendments, and so on. This chapter details the common assumptions that underlie the competing arguments and provides an overview of the predicted outcomes of democratic participation. At heart, the difference between the theories is that supporters of discourse argue that the more democratic the system the better, while social choice theorists suggest that more democracy can often lead to unstable and arbitrary results. I examine some of the criticisms that have been levelled at each theory and conclude that at the level of abstract theory Arrow gets the better of the argument, but that this might not translate into a more workable theory for real world democratic institutions.
Chapter 3: Circumventing Arrow’s theorem The conclusion I draw at the end of Chapter 2 fits with a line of argument that has led to a pessimistic outlook for democracy, and political scientists such as William Riker and Russell Hardin have concluded that democracy rests on a bed of sand. This conclusion is particularly
Introduction 3 damaging to democracy if we agree that the ‘reasonable’ conditions imposed on procedures by Arrow and Habermas are acceptable. In this chapter I argue that we should abandon some of these conditions. For example, the universal domain condition which suggests that all preference orderings are allowed into the democratic arena is an unreasonable condition to place on a social welfare function. Some preference orderings, for example those of the Ku Klux Klan and similar groups, might be disallowed prior to any deliberation because they undermine the democratic requirement that all citizens are to be treated with equal respect. I conclude that attempts to produce a ‘power-free’ arena for decision-making that treats all preferences with equal respect cannot succeed and that a workable system of democracy has to strike a balance between freedom, equality and coercion. If the findings of this chapter are accurate, serious attention has to be paid to the design of democratic institutions; otherwise it is likely that too much democracy will lead to volatility and too little will snuff out the freedom for which the democratic process is praised. I do not, therefore, share the pessimism of Riker. Although I have utilized the arguments of social choice theory to highlight some of the weaknesses of deliberative models of democracy, this is not meant to suggest that democracy cannot be defended and in the rest of the book I provide a case in its favour. I do suggest, however, that theories of deliberation and social choice are ultimately incompatible. I recommend that we need to reign in some of our expectations about what democracy can deliver from a procedural point of view and, in particular, I argue that it cannot provide the sort of authenticity that is claimed for it by many supporters of deliberative democracy. The limitations that need to be placed on democratic participation in order to avoid instability do mean that outcomes of democratic procedures lack full legitimacy. I conclude by arguing that the best grounds for defending democracy are non-procedural. Chapter 4: An empirical test of social choice and deliberative theories of democracy This chapter presents an empirical test of ‘institution-free’ theories of democratic procedures, exemplified by Habermas, Kenneth Arrow and William Riker. Social choice theorists claim that instability is the likely result of a pure majority rule institution, whilst deliberative democrats point to the normative benefits that accrue from free and equal participation. The conclusions reached by discursive and disequilibrium theorists seem to be at odds with one another and
4 Introduction it would be useful to test the claims empirically. Despite a surge of empirical research on the effects of discourse, it has been difficult to find a political institution that fits with the procedural requirements of deliberative democrats and social choice scholars. The problem is to find examples of collective decision-making procedures that come close to resembling the conditions of freedom and equality stipulated by both sets of theorists. When we look at contemporary institutions we see that most empirical work is aimed at understanding how complex institutional structures shape the collective decision-making process. The arguments of deliberative democrats and social choice theorists are often treated as abstract hypothesis; for example, the requirement for an open agenda process is usually regarded as a highly specialized and arbitrary concept with little relevance to actual political behaviour. Many people working in democratic theory suggest that institutions organized according to principles of pure majority rule cannot exist. We need only accept this argument, however, if we also accept that institutions always resemble something like modern legislative bodies and if we assume that procedural rules always constrain actors. The political institution that comes closest to meeting the requirements of pure majority rule no longer exists. When it was in operation, it was organized according to a few basic democratic principles that were designed to allow the free and equal participation of all members of the institution. Its structure was simple enough to allow us to examine it as a whole, and because it operated in the world of ‘real politics’, we do not have to concern ourselves with many of the counter-intuitive assumptions used by social choice theorists, such as gate-keeping powers, ex-post vetoes, perfect information, dimension by dimension decision-making and calculations of the relative ‘sincerity’ of voting behaviour. The institution to which I refer is a real legislative body rather than a stylized theoretical creation. Hence, we can engage in the crucial enterprise of testing theories against the backdrop of firm empirical evidence and it allows us to move a step beyond the usual approach of theorizing from a ‘stylized’ institutional setting. The legislative body to which I refer is the American Continental Congress 1774–1789. In this chapter I will show that the Congress was an open, egalitarian, and democratic legislature and fully reflected the preferences of its members. It is an excellent example of what happens when institutional norms and procedures are weak, and when power is widely and evenly dispersed. I provide a detailed description
Introduction 5 of how the Congress dealt with two of the most pressing issues of the day (the Northwest Ordinance and the location of the capital) and conclude that free and equal democratic procedures can lead to unstable outcomes.
Chapter 5: Hobbesian sovereignty and the spectre of tyranny The more general conclusion I draw from the previous three chapters is that Arrow is correct to claim that we can have free institutions and the potential for instability and disorder, or the buck has to stop somewhere, in which case we get stability but at the ‘cost’ of absolutism. An important conclusion of social choice theory is that we are left with a choice between absolutism and chaos. This conclusion has often been bypassed by Arrow’s supporters because it is usually assumed that social choice theory is within the rubric of liberalism, which goes to great lengths to circumvent absolutism. Unfortunately, if we do away with a final decision-maker, we end up with instability once again. We seem to be between a rock and a hard place. Hobbes came to the same conclusion about chaos and absolutism 300 years earlier and it is not as unusual as it first seems to turn to him for inspiration on matters of democracy. Arrow and Hobbes are right about the need for a final arbiter to guarantee outcomes, and the latter bit the bullet and followed this argument to its logical conclusion. To give full support to social choice theory might, therefore, require rehabilitating the idea of absolute sovereignty found in the work of Hobbes. In this chapter, I spell out the benefits for contemporary democratic theory if we take Hobbes’s theory of sovereignty seriously. It has been a common complaint about Hobbes that he provides a theory of sovereignty more in line with a totalitarian brand of politics. A common interpretation of his political philosophy is that he proposes a form of absolutism in order to tame savage and instrumentally self-interested individuals who will, to use Madison’s language, vex and oppress one another given the chance. I argue that this interpretation does not fit well with Hobbes’s motives, which were to provide logical and clear definitions of political concepts, and to produce a political system that protected the liberty of the individual. His definition of sovereignty stems from his civil science and not from an intention to terrorize people. Commentators who suggest that Hobbes displayed an unrepentant willingness to incarcerate the natural individual face a severe textual problem because Hobbes repeatedly states that his goal is to find the logically necessary truths
6 Introduction about politics. The point he wishes to make is that absolutism occurs wherever political rule is properly exercised. It is a necessary part of politics that there has to be an end point to political deliberations if we want to guarantee stable outcomes. Hobbes’s definition of sovereignty does not rule out a democratic form of politics. He states that: ‘the sovereign is either in one man or in an assembly of more than one, and into that assembly either every man hath right to enter or not every one (but certain men distinguished from the rest), it is manifest there can be but three kinds of commonwealth’ (1968, 239). The three kinds are monarchy, aristocracy and democracy. Hobbes, therefore, does not reject democracy as a legitimate form of sovereign power. He thinks it is dangerous and inefficient, but it is not logically ruled out by his system of thought: ‘the difference between these three kinds of commonwealth consisteth not in the difference of power, but in the difference of convenience or aptitude to produce the peace and security of the people’ (1968, 241). In fact, Hobbes claims that he has proven all his claims in Leviathan except his arguments in favour of monarchy, which he says are not truths but only his opinion. Chapter 6: Democracy versus constitutionalism In the concluding chapter of the book, I argue that a Hobbesian brand of absolutism is more compatible with democratic politics than liberal constitutionalism. This is not a popular idea and liberals have traditionally been very wary of the power of the majority and have tried to limit it in various ways through constitutions, bills of rights, separation of powers, checks and balances and federalism. I argue in favour of democracy against liberalism and suggest that constitutionalism perverts democratic practices and dilutes the essence of democracy, which in its most basic form is rule by the people. Constitutionalism replaces popular sovereignty with a form of guardianship that takes decision-making out of the hands of the majority and places it in the hands of a supposedly more tempered and wise elite minority. The major justification for hijacking political power in this way has a long history and can be traced all the way back to Plato. It is the claim that unchecked political power in the hands of the masses will lead to the tyranny of the majority. I argue that this fear is ungrounded and that elites have traditionally been the ones responsible for tyranny. I claim that citizens are better protected by a political culture that is supportive of democratic freedom and equality rather than by legal documents and conclude that the logic and practice of democracy
Introduction 7 protects rather than endangers the freedom of the individual. I utilize the British political system to demonstrate how a Hobbesian form of parliamentary sovereignty is not tyrannical and to show that there can be checks on the executive that are consistent with the idea of absolutism. I also argue that constitutions are incapable of providing useful guidelines for complex matters of public policy. I examine the issue of free speech and argue that constitutional statements such as, ‘Congress shall make no law abridging freedom of speech’, are of little use when deciding the appropriate limits to freedom of expression.
2
A comparison of social choice and deliberative theories of democracy
Introduction This chapter focuses on the debate about how much rationality we can expect from the outcomes of democratic procedures. I will examine two contemporary arguments within democratic theory that posit competing and opposing conceptions of participation. These are (a) theories of democratic discourse and (b) disequilibrium theories of social choice. Theories of discourse, presented to us by Habermas (1984), Cohen (1989), Dryzek (1990, 2000), Benhabib (1994) and others, tell us that through the democratic process we can arrive at consensus, rational outcomes and even principles of justice. In order for the results of such a procedure to be fair we need an unrestrained discourse setting in which each participant has equal access to debate, and equal opportunity to raise issues, voice objections, and enter new alternatives into the discourse. These procedural requirements make sure that the rules of the setting do not favour any particular participant. Discourse is directed at agreement concerning reasonable speech acts that, because they are agreed upon and are rational, gain legitimacy. As an alternative view of democratic practice, we have social choice arguments of an ‘institution-free’ type, described in the writings of Black (1958), Arrow (1963), McKelvey (1976) and Riker (1982). Such theorists also posit certain procedural settings for democratic fairness and argue that in a fair majority-rule setting where there are three or more participants and three or more alternatives, the forthcoming results can potentially end up anywhere in the policy space. When this is combined with McKelvey’s insight that even a slight change in preference orientation can cause a complete breakdown in equilibrium, the conclusion of social choice theory is that the outcomes of majority rule are inherently irrational and unstable.
A comparison 9 An interesting feature of both theories is that they present almost identical assumptions concerning the requirements for a fair procedure; they both demand equal access to debate, the absence of a powerful agenda setter, unrestrained opportunities to raise and object to amendments, unrestricted domain regarding preferences and so on. Despite these similarities, one argument suggests that the procedure of democratic discourse will yield morally legitimate outcomes, while the other suggests that it has a tendency to produce chaos and disequilibrium. At heart, the difference between the theories is that supporters of discourse argue that the more democratic the system the better, while social choice theorists suggest that more democracy leads to instability and arbitrary results. Both cannot be correct, and a resolution to the debate would seem crucial, considering the current support afforded both discourse theory and social choice theory in the social sciences. A crucial point to be made at this juncture is that both theories are very abstract and stand a long distance from everyday politics. Riker provided some examples of cycling from real world legislatures, but Mackie (2003) has recently undermined the validity of this research. Given that Riker’s research was conducted on institutions that in no way conform to the requirements of fairness just discussed, it is not surprising that we still know very little about whether instability does reign in pure majority-rule settings.
Discourse and democracy John Dryzek claims that ‘[t]he essence of democracy itself is now widely taken to be deliberation, as opposed to voting, interest aggregation, constitutional rights, or even self-government’ (2000, 1). I am not sure that all scholars or lay people would agree with this assessment, but it is certainly the case that the ‘deliberative turn’ in democratic theory has been popular and influential. In this section of the chapter, I will focus on the arguments of Habermas and others concerning the benefits that accrue from free and equal dialogue. It is true that Habermas’s position has changed somewhat over the years and his concept of democratic participation has taken a more liberal bent, but certain key elements have remained constant in his writings and in those of other discourse advocates. There are three main points that discourse theorists such as Habermas, Benhabib, Cohen, Dryzek and Chambers (1996) still share. First, the procedure has to allow free and equal access for all participants. Second, each theorist expects that interests can and will be transformed through debate. This notion
10 A comparison of discursive rationality, as opposed to instrumental rationality, is considered crucial by supporters of discourse because it is argued that we can change people’s preferences through persuasion so that consensus, or something close to it, is achievable. The claim, therefore, is not only that preferences can be transformed, but that they converge. Finally, they also claim that the most persuasive source of legitimacy is reasoned deliberation. As Habermas says, we would achieve legitimate consensus ‘if only the argumentation could be conducted openly enough and continued long enough’ (1984, 42). Benhabib summarizes the above points when she says, ‘legitimacy results from processes of collective deliberation conducted rationally and fairly among free and equal individuals’ (1994, 31). If these three aspects of discourse hold, we can hope for outcomes that are rational, fair, and hence legitimate. The viability of such claims is precisely what is questioned by the social choice theorists examined in the next section of this chapter. To get to the root of modern society, Habermas advocates a ‘paradigm of understanding’ that looks closely at intersubjective relationships found in communication. According to Habermas, there must be a model of the subject at the core of the social sciences dealing with human interaction. Hence, at the centre of all social science there is a subjective premise of what it is to be human, and for Habermas the premise is that humans are rational. In particular, he is concerned that we are able to give good reasons for an action or statement. He claims that there are three domains of the rational – the pragmatic (instrumental, technical), the ethical (clinical; rationality in relation to one’s life-plan), and the moral (other-regarding, substantive) – and practical reason is utilized differently in each circumstance: the illocutionary meaning of “must” or “ought” changes with the practical relation and the kind of decision impending but also the concept of the will that is supposed to be open to determination by rationally grounded imperatives in each instance. The “ought” of pragmatic recommendations relativized to subjective ends and values is tailored to the arbitrary choice of a subject the faculty of rational choice does not extend to the interests and value orientations themselves but presupposes them as given. The “ought” of clinical advice relativized to the telos of the good life is addressed to the striving of self-realization [t]he categorical “ought” of moral injunctions, finally, is directed to the free will emphatically construed, of a person who acts in accordance with self-given
A comparison 11 laws Only a will that is guided by moral insight, and hence is completely rational, can be called autonomous. (1993, 9–10) The pragmatic/instrumental view suggests that rationality is simply about finding the best means to a particular end, whereas Habermas wants to claim that it can tell us how best to pursue our ends and what ends we should pursue. He does not deny that the rational choice view of instrumental rationality can be applied to many aspects of the human condition, but he does reject the claim that it captures everything to be said about rational behaviour, particularly regarding questions about how one ought to behave. The instrumental view does make ‘ought’ recommendations to agents, but these are only of the self-regarding kind, i.e. that one ought not to do an action if it will prevent one from attaining a desired goal. This is not a normative usage of the word ‘ought’ because the recommendation is offered regardless of the morality of the act or of the preference. For Habermas, the ‘ought’ can refer to actions that are primarily selfregarding or to actions that have a moral other-regarding focus and it is this latter aspect of rationality that is particularly important to him because he is concerned that the instrumental view banishes morality from the definition of rationality. When we engage in democratic deliberation, we have to accept that there are rules and conditions for acceptable behaviour, and if we break these rules we are acting as if the rules are a necessary condition for participation but they nevertheless do not apply to ourselves. As participants we would be engaged in a logical contradiction, which means that we are not behaving rationally. In particular, he suggests that some goals are irrational because they treat others only as a means for one’s purposes rather than as ends in themselves. This involves a performative contradiction because one is acting towards others in a fashion that would be unsatisfactory if the roles were reversed. Habermas’s suggestion is that it is rational to be morally reasonable. The argument that there is an uneasy tension between rationality and morality goes back a long way in the history of philosophy, but Habermas is part of the Kantian tradition in which the universalistic conception of rationality insists that what makes it rational to satisfy an interest does not depend on whose interest it is. Thus the rational person seeks to satisfy all interests If I have a direct interest in your welfare then on either conception I have reason to promote your welfare. But your interest in your welfare
12 A comparison affords me such reason only given the universalistic conception. (Gauthier 1969, 7) The measure of rationality is whether one can give good reasons, regardless of self-interest, for one’s actions. Rationality rests upon agreements that have been arrived at through dialogue, and the individual who is incapable of engaging with others ‘by appeal to standards of evaluation is not behaving rationally’ (Habermas 1984, 17). Simone Chambers suggests that At the center of Habermas’s moral theory is a strong cognitivist claim. Moral statements are open to rational evaluation and are not mere statements of preference. Habermas maintains that what we mean when we say something is rational is that it could be defended with reasons. On this view, rationality is embedded in essentially public practices of communication. A statement or an action is rational to the extent that it could be explained to others. (1996, 90) If Chambers is correct, it means that Habermas has a two-tiered concept of rationality. The first tier encompasses the norms and values of a society and hence is ‘universally’ valid only within that specific context. The second tier relates to the necessary procedural conditions that have to exist for dialogue to take place, regardless of social context. If an agent accepts that such conditions are a necessary part of discursive interaction, the agent acts irrationally when he or she contravenes them. It is in this sense that Habermas claims it is irrational to be unreasonable. It is now clear that moral constraints (because they are a part of the definition) do not pose a threat to his theory of rationality. Chambers concludes, For the Kantian, rationality is associated with reasoning in a world populated by other rational agents who are recognized as being worthy of the same consideration I am worthy of. Practical reason has an element of reciprocity built into it. Morality finds its underpinnings in a conception of practical reason which links the two (morality and reason) prior to the agreement. In the case of Scanlon (and Habermas), this Kantian idea is articulated in a notion of accountability or justification. (1996, 24) Chambers also suggests that we distinguish between rational-choice theory, which she traces back to Hobbes, and rational-agreement
A comparison 13 theory, which finds its roots in the philosophy of Kant: ‘Choice versus agreement adequately captures the individualistic versus intersubjective starting points of these two traditions’ (1996, 25). The problem that the heirs of Kant have to face is how to defend this more encompassing definition of rationality; as Gauthier notes, ‘this assumption, of the impersonality or impartiality of reason, demands defense’ (1969, 8). Habermas thinks that we all have the potential to use sentences and issue statements that allow us to reach mutual understandings concerning reasons for action. The validity of a sentence is open to examination and is judged rational or not in terms of its correspondence to certain specifications. As Stephen White lists them, the signs of communicative competency rest on mastery of the rules of discourse, an understanding of linguistic rules that help foster understanding, a grasp of the rules of grammar that allow for understandable sentences, and the ability to take part in complex forms of interaction. These are bracketed by Habermas into a general theory of speech actions. Habermas argues that rational individuals are capable of these discursive demands and as autonomous beings will be able to critically assess their own beliefs, give justifications for them, overcome old identities, view themselves as separate from roles they play in society and make judgements based on rational principles. Hence, Habermas’s social science is based on the liberal premise of the rational, autonomous self. Additionally, there has to be some link between the individual and the collective because we all have to be intersubjectively recognized by others while at the same time retaining our individuality. Actors can use all of their intersubjective life-world experience to coordinate action to help them reach understanding, i.e. rational agreement. They can, therefore, judge the rationality of speech and action in terms of the truth, authenticity and normative legitimacy of other agent’s actions and statements. Rational outcomes crucially depend upon whether we can create consensus through competent persons communicating with one another. Language is the basis for autonomy and is the foundation for moral agreement. As Habermas puts it, ‘the concept of truth combines the objectivity of experience with a claim to the intersubjective validity of a corresponding descriptive statement, the idea of an idealized consensus’ (1987, 72). In order for this to happen, the speaker is required to be honest and the listener has obligations to test rationally the speaker’s claims if the listener should disagree with the statement. Most statements are not put to a validity test because they are accepted by the listener. But once
14 A comparison they are challenged, the actors must attempt to come to some form of rational consensus about the truth value of the statement. It is the ideal speech situation (ISS) that provides the opportunity to arrive at such consensus. Habermas argues that the ISS allows normative claims to be treated like truth statements through a universal procedure that treats all participants as free and equal in the discourse. The test ‘is whether or not a proposed norm is acceptable in an actual argumentation to all who are potentially affected by the norm, and acceptable is taken to mean that the norm satisfies the interests of each participant in the argument’ (1987, 49, my emphasis). Validity, therefore, springs from consensus according to the intersubjective recognition of statements. The crucial point to draw from the above is the assumption that discourse leads to resolution and agreement. As White says, we have a speech situation ‘in which a conflict would be resolved solely by the force of the better argument’ (1988, 55). In fact, Habermas seems to suggest that the only way to tell if a decision is rational is if it is a consensual one, arrived at through pure, unadulterated discourse: ‘Only those norms may claim to be valid that could meet with the consent of all affected in their role as participants in the practical discourse For a norm to be valid, the consequences and side effects of its general observance for the satisfaction of each person’s particular interests must be acceptable to all’ (1990, 197). As Rescher says, ‘Habermas sees the impetus to consensus as the constitutive core of rationality itself’ (1993, 26). As we will see shortly, social choice theorists question strong claims such as these and also place into doubt the weaker claim that rational outcomes do come from free, equal and hence non-coercive dialogue. The ISS procedure is universal because anyone who enters into the argumentation already presupposes that universality is valid and hence that those participating should be treated equally. As already noted in the discussion on rationality, what Habermas means is that the discourse structure itself implies mutual intersubjectivity and to enter the debate is to be immediately morally obligated to a few minimal, universal values that make dialogue possible: As a procedure, practical discourse insures that all concerned in principle take part, freely and equally, in a cooperative search for truth, where nothing coerces anyone except the force of the better argument. Practical discourse is an exacting form of argumentative decision making it is a warrant of the rightness (or fairness) of any conceivable normative agreement that is reached under these conditions. (Habermas 1990, 198)
A comparison 15 There are certain rules that have to be followed to make the ISS fair. White (1988, 56) lists them as follows: 1 2 3 4 5 6
Each subject is allowed to participate in discourse. Each is allowed to call into question any proposal. Each is allowed to introduce any proposal. Each is allowed to express attitudes, wishes and needs. No speaker is to be hindered by compulsion. Universality in agreement concerning outcomes.
The goal is to produce a fair system in which deception and ideology do not influence decisions. Each participant is an equally autonomous actor in the discourse. Cohen demands a comparable institutional framework when he requires that each can put issues on the agenda, propose solutions, and offer reasons in support of or in criticism of proposals and each has an equal voice in the decision. The participants are substantively equal in that the existing distribution of power and resources does not shape their chances to contribute to deliberation (1989, 23). From this procedure, deliberative democracy ‘aims to arrive at a rationally motivated consensus’ (1989, 23). Benhabib and Dryzek propose similar procedural requirements: ‘There are no prima facie rules limiting the agenda of the conversation’ (Benhabib 1994, 31) and ‘there should be no hierarchy of formal rules A decision rule of consensus should obtain all features should be redeemable within the discursive design itself. Participants should be free to reflectively and discursively override any or all of them’ (Dryzek 1990, 43). Dryzek (2000) prefers to talk of discursive rather than deliberative democracy because the latter does not necessarily require personal interaction; as Rousseau noted and advocated, one can deliberate alone. In contrast, the discursive process is social and intersubjective. ‘Deliberation’, when it refers to interaction, suggests a calm and gentle argument which is often far removed from how we talk to one another. It is this form of deliberation that was seen by Foucault and his supporters as a constraining form of interaction. Dryzek notes that the terms (‘deliberation’ and ‘discourse’) are now used interchangeably and we would not be able to separate them, so he suggests instead that we think of discourse as an inclusive category that encompasses the more critical deliberative element.
16 A comparison I interpret this brand of democracy to be a ‘pure’ form in which power relations are purged and rational outcomes are supposed to rest only on mutual and consensual agreement that springs from a fair procedure. Such ‘agreement is regarded as valid not merely “for us” (the actual participants) but as “objectively” valid for all rational subjects (as potential participants). In this sense, discourse is, as Habermas puts it, “the condition for the unconditioned” ’ (McCarthy 1978, 292). It is crucial to note the strong claim being presented here concerning the truth value of decisions arrived at through consensus: ‘His thesis is that the testing of these claims in argumentative discourse warrants attaching to those that emerge unscathed the honorifics: “true,” “objective,” “valid,” and so forth’ (McCarthy 1978, 294). Benhabib makes a similar claim when she argues that such decisions ‘have a presumptive claim to being rational’ (1994, 33) and Cohen suggests that ‘ideal deliberation aims to arrive at a rationally motivated consensus’ (1989, 23). Dryzek and List quote Elster, who suggests that under deliberation ‘there would not be any need for an aggregating mechanism, since a rational discussion would tend to produce unanimous preferences’ (2003, 6). As they correctly note, there is a very large assumption being made here, namely that we will all come to agree if we simply talk for long enough. Alford (1996) has suggested that Habermas might not be the proceduralist that the above interpretation suggests. He places Habermas in the classical republican tradition of democratic theory that is concerned not so much with procedural rationality but with finding the best form of life for human beings through dialogue. It is not clear what he means by ‘republican tradition’ but he seems to be suggesting that Habermas is closer to Aristotle than to contemporary procedural theories of democracy. Habermas, however, has stated on many occasions that discourse is not aimed at the good but is directed towards discovering rational principles. He does not claim to have a teleological view of the good life as can be seen in his reply to Albrecht Wellmer, who criticized him for holding such a view: We can specify only certain formal conditions of a rational life – such as a universalistic moral consciousness, a universalistic law, a collective identity that has become reflective and so forth. But insofar as we are dealing with the possibility of a rational identity, there is no ideal limit value describable in terms of formal structures. There exists rather only the success or failure of the efforts to achieve a form of life in which the unconstrained identity
A comparison 17 of individuals, along with unconstrained reciprocity among individuals, becomes an experienceable reality. (1982, 26) Nor is Habermas in the Rousseauean tradition: ‘According to Mead, no individuation is possible without socialization, and no socialization is possible without individualization. For this reason, moreover, a social theory that captures this insight in a linguistic pragmatics must also break with the sort of Rousseauism that Henrich attributes to me’ (1993, 26). Neither Aristotle nor Rousseau championed dialogue as a means of discovering true principles; both favoured rational contemplation rather than democratic discourse. If one wishes to place Habermas within a philosophical tradition, it is more fruitful to look back to Marx rather than Aristotle or Rousseau. Habermas, however, has taken a steady trajectory away from Marx and can now be safely placed close to or within the liberal camp despite his own claim that he is the last Marxist. This shift to a liberal proceduralism, however, also brings with it some problems as outlined in the next section of the chapter. This is not to suggest that Habermas is a proceduralist without a substantive goal in mind; the point I am making is that procedure plays a large part in his theory of democratic discourse, and this is important because the theory of social choice, which will be discussed in the next section, suggests that the procedural aspect of his theory cannot get him to his substantive goals. One reason that Alford thinks Habermas is not a proceduralist is that he does not demand that preferences, after dialogue, have to be revealed through the voting procedure. Discourse is not about the expression of preferences he tells us; rather ‘it is a procedure to form them’ (1996, 754), (apparently Habermas is now a proceduralist). But what good is the formation of preferences if we do not have a procedure for expressing them? If the ISS is not a procedure then what is it? Habermas certainly gives us a set of formal rules meant to govern the way we talk to each other and the way we express collective decisions. If it is not in some sense a procedure, then it seems useless as a model of discourse. It is important to remember that legitimacy comes not only from discourse but from the results of such discourse. Alford at times suggests that Habermas presents us with an ideal that tells us what we would agree to if we all truly knew our minds, if we all knew our real interests, and if power relationships were negated. What he is getting at with the first two claims is that genuine consensus does not occur when we suffer from a form of false consciousness and he blames critics for ignoring this potentiality: ‘The possibility
18 A comparison that what people prefer may not be what they want finds no place in their models’ (1996, 756). I can only say ‘thank goodness’ for that. Telling people what their true preferences are and informing them that their current preferences stand in opposition to their true wants is incompatible with democratic theory. I can claim Habermas’s support on this issue. He is very careful to avoid any suggestion that participants in discourse suffer from false consciousness. He is willing to argue that preferences change through discourse and even that we can recognize that preferences after dialogue are more sophisticated, but he does not say that preferences we hold prior to discourse are held because of false consciousness. Habermas’s demand for rational decisions does mean, however, that justification is not based purely on discursive procedures, although the procedure is, of course, important, and the closer the discursive arena resembles the ideal the better. Justification also has to be results based. Results have to be rational as well as consensual for the system to be legitimate. Hence, even assuming that we can set up the correct conditions for discourse we are still a long way from justifying this method in terms of procedural outcomes. If it can be shown that uncoerced discourse cannot provide for rational outcomes that represent the wishes and judgements of the participants, some of the moral benefits ascribed to participation will have to be abandoned. The social choice argument presented in the following section poses a serious threat to such claims and suggests that the closer the institutions of decision-making resemble the ISS, the more likely they are to produce arbitrary results. Social choice theorists present an analysis of democratic procedures that, they argue, demonstrates that we can never know whether the outcomes of democratic decision-making are rational, or the preferred outcome of a majority.
Social choice theory Social choice theorists analyse politics in terms of subsets of institutions that aggregate individual preferences into social choices. According to Amartya Sen, the central question of social choice theory is, ‘how can it be possible to arrive at cogent aggregative judgments about the society (for example, about “social welfare,” or “the public interest”, or “aggregate poverty”, given the diversity of preferences, concerns, and predicaments of the different individuals within the society? How can we find any rational basis for making such aggregate judgments?’ (2002, 66). Historically, the two most important figures in this school of thought are Condorcet and Arrow. As Russell Hardin
A comparison 19 (1993) and others have noted, Condorcet’s problem of instability differs somewhat from Arrow’s. The former is concerned with voting on specific issues whereas Arrow is concerned with choosing a total world where everything is settled. Arrow wants us to make a single social choice of a total world and then no more choices are left to make. Hence, it is a situation that is not immediately practical to our everyday lives, where we never make such a choice. Arrow thought of the issue in this way because he did not want to assume anything other than purely ordinal preferences for individuals and for the final collective choice. Arrow thinks we can only make ordinal choices over whole states of affairs because in real life there is never a situation when I would prefer ‘a’ over ‘b’ regardless of all other factors; in order to make certain of absolute ordinal choices he assumed that we are choosing over whole states of affairs. One of the strengths of Arrow’s argument is that it does not rely on some form of anti-democratic statement about the incapacity of individuals to form coherent preference orderings because they are fickle, ignorant or stupid. He assumes that they do have coherent and rational preference orderings and then sets about demonstrating that outcomes can still be irrational. Amartya Sen (1970) gives one of the best descriptions of Arrow’s position. He notes that in 1938, Bergson came up with the idea of thinking of orderings over all possible alternative states. This as yet says nothing about how such collective decisions are to be made. This is why Arrow is important; he asks that the collective choice rule meets four criteria of reasonableness. His ‘General Possibility Theorem’ proved that such ‘mild’ conditions rule out every possible social welfare function (SWF). The four conditions are 1
2 3
4
The SWF must be wide enough to include any logically possible set of individual orderings. Sen calls this the condition of unrestricted domain or ‘condition U’ for short. It must be Pareto optimal in weak form, i.e. if everyone prefers ‘a’ to ‘b’ then society should also prefer ‘a’ to ‘b’. The social choice over alternatives depends only on the orderings of individuals over those alternatives and not on a ranking irrelevant to the choice. Sen calls this ‘condition I’ and Arrow refers to it as the independence of irrelevant alternatives (IIA). The SWF should not be dictatorial (‘condition D’).
Arrow proved that there is no SWF that can meet all four of these conditions and ‘together they seem to produce a monster that gobbles up all the little SWF’s in the world’ (Sen 1970, 38). Unfortunately, as
20 A comparison Hardin notes, it could be the case that almost everyone prefers ‘a’ over ‘b’ and ‘b’ over ‘c’ and ‘c’ over ‘a’, and we still end up in a cycle. Hence, the theory does not say that we cannot make democratic social choices, it simply states that even in the case of unanimity, there is no one rule for the social ordering of all alternate states of affairs. This might not be too troubling from an empirical point of view because it might be the case that we will often prefer ‘a’ over ‘b’ and ‘b’ over ‘c’ and ‘a’ over ‘c’, in which case the result is stable and not a cycle. On the other hand, we might not. As Hardin notes, even if Arrow’s total choice situation is not appealing, this does not show us a positive argument for a social choice procedure that provides outcomes that are also rational and fair. Once we leave the rarefied atmosphere of Arrow, we are no better off because Condorcet highlights the fact that we do not have a democratic procedure that is going to meet the requirement of being rational, free and equal and we still face the same type of problem Arrow identified, but this time in real working democracies. The crucial aspect of democracy for contemporary social choice theorists is popular participation in government through voting. Only through such participation is freedom maintained, argues William Riker, because the major goal of political participation is preventing representatives from acting in a tyrannical manner (1982). Hence we are free in a ‘negative’ sense to the extent that we prevent our representatives from abusing power. Such participation, as well as protecting our liberty, also provides for equality because everyone’s vote counts the same. In its ideal form, democracy promotes citizens who are free and equal politically, and it is, therefore, a just means of making collective decisions. Riker’s argument presents a rather ‘thin’ conception of the benefits of participation. He does not suggest that we can reach the mantle of autonomous beings through participation, or that voting unveils a form of ‘general will’. He values democracy almost solely because it protects us against political elites in a manner that other systems cannot guarantee. Riker argues that the democratic process in its ideal form blends together ends and means. That is to say, the goal for strong democrats, or populists as Riker calls them, is self-determination and self-realization, and these are gained through participation in democratic politics. In other words, democratic ends can only be achieved through democratic means. If these claims cannot be realized, ‘democracy is meaningless’ (1982, 3). A procedural defence of democracy fails, therefore, if it can be shown that the means do not promote the ends. These arguments presented by social choice theorists stand in opposition to the claims made by supporters of democratic discourse. Riker’s
A comparison 21 major point is that democratic participation cannot show us the will of the people. This conclusion stems from the argument that every participatory system that demands procedural fairness will produce outcomes that are potentially arbitrary and, therefore, meaningless as expressions of the ‘truth for us’. By fairness he means monotonicity, neutrality among options, simple majority decision, no agenda setter, equal access to the agenda, equal opportunity to debate, equal opportunity to introduce proposals, and each vote counting equally. In other words, the requirements for pure majority rule as presented by social choice theorists are almost identical to the ISS procedural demands of Habermas, who does not want decisions made by a dictator, and who supports condition U and weak Pareto optimality (it is difficult to know what he would think of condition I). The gulf that supposedly exists between social choice theories of democracy and Habermas’s discourse setting does not exist on a procedural level. The gap shows itself when the discussion turns to the likely outcomes from this sort of procedure. In such ‘ideal’ democratic conditions we still have a diversity of goods, variations in levels of information, and different epistemologies, classes, genders, races and personality types. We can, therefore, expect differences in knowledge, behaviour and preferences. The question to be asked is what type of results can we expect from this mixed bag? Habermas claims that the results will be consensual and hence hold more legitimacy than any other form of decision-making. Riker argues that the results will often be meaningless. More specifically, he does not deny that consensus is possible, but he does deny that we can necessarily attach moral worth or the label ‘rational’ to democratic agreements. Hobbes’s arguments about the state of nature are similar to the claims of Riker. Hobbes argued that a condition with no rules is chaos. Riker suggests the same is true for democracy; what Habermas is doing is offering the democratic equivalent to the state of nature, and Riker claims that we can expect the same sort of chaos to ensue. Riker arrives at his conclusions by embracing Condorcet’s insight about the paradox of voting. In any given policy space where there are three or more participants and three or more alternatives, the result of a fair majority rule procedure can potentially end up anywhere in the policy space. This happens because a voter may have an ordinal preference among policies ‘a’, ‘b’ and ‘c’, but when they are paired off in competition, the different pairings of preferences and the different order in which they are presented can lead to many outcomes. It is assumed that voters can order preferences in any manner they wish. If they all have the same preferences, the result will be single peaked and
22 A comparison stable with a clear Condorcet winner. If preferences differ, however, which seems highly likely, cycling can occur under pure majority rule. At the risk of boring readers with another formulation of the problem, consider the following example with three options open to the participants:
Individual
First preference Second preference Third preference
X
Y
Z
c a b
b c a
a b c
It is quite possible to have majority opinions on all three of these options even though they are mutually incompatible when taken together. Given the options and preferences above, there is no one logical or rational outcome. Each preference ordering wins when paired off against one of the other alternatives. Hence, any decision will have to be arrived at by some form of rule that stops the debate and arbitrarily decides the issue in favour of one option over another. If option ‘a’ beats ‘b’ and then ‘c’ beats ‘a’ we will often assume that ‘c’ is the winner but we have not tested whether ‘c’ would beat ‘b’, so we could have a potential cycle without knowing it. Riker suggests that an agenda setter who knows this can manipulate results to his or her advantage. This argument was first presented by Condorcet and is now an axiomatic principle of social choice democratic decision-making theories. When this is combined with Richard McKelvey’s finding (1976) that even a slight change in preference orientation can cause a complete breakdown in equilibrium, the conclusion of social choice theory is that majority rule is inherently unstable. There are many possible outcomes of majority rule; a fair decision-making process can produce inconsistent results, and hence cannot show in any certain fashion the popular will. The result may, of course, actually represent such a will, but given the dynamics of a fair democratic procedure we can never know if this actually is the case. Decisions have no legitimacy other than that they came about through a just procedure. But as this procedure is arbitrary in the sense that its outcomes do not necessarily bear any firm relationship to actual preferences, the type of procedural defense offered by Habermas is deemed inadequate.
A comparison 23 Coleman and Ferejohn summarize the dilemma thus: ‘The problem social choice raises for democratic theory is that any democratic voting procedure that is fair in the appropriate sense will be normatively defensible but meaningless, that is its outcomes will be arbitrary. Only voting that is meaningful and fair can be justified. Unfortunately, no voting procedure can be both’ (1986, 11). The production of rational outcomes requires some form of concentration of power: ‘a dictator, an oligarchy, or a colloquium’ (Riker 1982, 132). Some form of transitivity is required within the political system to prevent tyrannical control of the agenda. But then we arrive at instability again, and the problem of assigning meaning to decisions. A defensible form of democracy requires some potential avenues for instability, and social choices will have to be unordered and inconsistent to a certain extent. The above arguments, if correct, imply that the strong claims made by deliberative democrats must be severely tempered because of the logical limits of the process. We have either procedural restrictions or ‘anything goes’. The problem with theories of democratic discourse from a social choice perspective is that the procedure of the ideal speech situation sets up a condition in which all participants can enter alternatives, which creates huge problems for building any sort of consensus. The policy space can, in theory, be infinitely stretched. And yet, the goal of discourse for Habermas is precisely to foster consensus: ‘he has, in recent years most emphatically and cogently envisioned a central position for consensus in human praxis – and has, indeed, made the quest for consensus a foundation-stone of his philosophy’ (Rescher 1993, 25). From the social choice point of view, the suggestion is that we are much more likely to find a lack of consensus than any form of universal agreement. The deliberative democrat can respond that we will all come to agree if we talk long enough, but the social choice theorist can, with equal justification, reply that dialogue will often move people further apart in the policy space and create intransigence and deadlock. The theoretical arguments we have from Black, Arrow, McKelvey, Riker, etc. imply that the outcomes of an ideal type of discourse setting have the potential to be irrational. In fact, for Riker, it becomes difficult to make sense of what can be meant by the term ‘popular will’ and he would no doubt suggest that it does not seem unreasonable to demand from deliberative democrats some compelling evidence that equilibrium and an open democratic forum can sit comfortably together. It is essential for deliberative democrats that outcomes at least make sense, otherwise the notion that we are living according to the dictates of our own will is undermined; we cannot embrace the
24 A comparison idea that freedom and irrationality are compatible. It is important to remember that democratic discourse is not justified if it only allows us to agree to differ; it also has to be demonstrated that discourse will lead to a rational and consensual outcome.
Criticisms of deliberative democracy One problem identified with deliberative democracy is that there seems to be an inbuilt tension in the theory between the twin demands of pluralism and consensus. All supporters of deliberation lean towards greater diversity and inclusion in the decision-making process. This is to be expected of a theory that requires free and equal participation of all those who are affected by the final result. In tension with this aspiration for pluralism, however, is the desire for unanimity because it seems reasonable to suppose that greater pluralism is unlikely to lead us to the (supposedly) desirable goal of consensus. To suggest that expanding the policy arena of competing preferences is compatible with boiling outcomes down to a unifying consensus seems at face value to be unrealistic. I am not talking here of legislative bodies that currently exist which have all sorts of mechanisms for limiting pluralism. I am talking of an environment like the ISS that allows for unlimited access. Arrow’s condition of universal domain is, therefore, similar to Habermas’s claim that all points of view must be heard in deliberation. They are both requirements for extensive pluralism; the difference is that Arrow claims that leaving the domain unlimited makes democracy unfeasible. If we do not get consensus then we have to vote on the issue. Even if we allow for the fact that preferences are not fixed and can change through the process of dialogue, we still cannot escape the potential difficulties of voting highlighted by Arrow and Riker. The problem of deciding whether preferences are fixed or malleable, a problem that has been a divisive issue between the two theories, is actually a false problem once we allow for the reasonable assumption that preferences can change, but not necessarily converge to a single point of agreement. The claims for deliberative democracy are also threatened by the arguments of Mancus Olson (1965). If one’s participation will not affect the result of collective decision-making then it becomes very questionable whether it is rational to expend all the additional cost of deliberative participation. Given the costs associated with gathering information, and the fact that it is infinitesimally likely that one’s participation will make a difference, it is not at all surprising
A comparison 25 that voters are uninformed – the thing that needs explaining is why they vote at all. But given this lack of knowledge it seems to be stretching things to say that elections in any way accurately reflect the preferences of voters – they often do not have enough information to form an opinion on what is being offered. It seems difficult to support some of the claims for deliberative democracy in the face of such arguments because the opportunity costs go up drastically with this type of participation without seeming to provide any comfort in the way of making a difference. It is sometimes claimed that citizens recognize that their vote will not make a difference, but they still participate in order to express a moral point of view. But if voting does not make any difference, it does not matter if a person votes for moral, economical or political reasons. That person’s vote will still be an infinitesimal part of a large total number of votes. Hardin (1993) suggests that if we were to live in a strong communitarian society where there is large scale agreement on all issues, the results of voting might be stable and democracy might not fall victim to the sorts of problems identified by social choice. But, he says, we do not live in such a society and it is not one that we can simply adopt. Hence, the problems of aggregation will not go away for us, even if they might not exist for some communitarian ideal. Even this claim by Hardin seems too optimistic given that he also says, ‘we can face this quandary [cycles] no matter what our majority must be’ (1993, 161). Only if we had consensus over all possible states of affairs would we be assured of stable majority rule. But, of course, if we were to live in such a society we would not need a decision-making procedure in the first place. Albert Weale makes the point that ‘all [democratic] forms of government make processes of deliberation central to the way in which decision-making is conceived’ (1999, 37). Hence, the distinctions made by deliberative democrats that contrast their own form of democracy with that of competitive party models are overblown: the model of deliberative democracy rests on a contrast with an implausible, if widely held, view of democratic decision-making in which choices are made as a result of competing pressures. The institutional theories of all the variants of democracy that I have been discussing stress the importance of deliberation, including liberal constitutionalism. (1999, 37) This point is brought home by the fact that Cass Sunstein (1997) includes American constitutionalism within the deliberative camp. If
26 A comparison he is correct about this, what is the distinction that is supposed to be so crucial between liberal and deliberative democracy? John Rawls, the great defender of liberal democratic procedures, also classes himself as a deliberative democrat despite the fact that his principles of justice are arrived at through philosophical reflection rather than interpersonal deliberation. It would seem that the concept of deliberation is part of the logic of most theories of democracy with the possible exception of Rousseau. Weale’s argument suggests that deliberative democrats are not really saying anything very novel; or if they are, it is not because they find deliberation to be a necessary part of democracy, but because of the grand claims they make about the effects of deliberation. Habermas has now moved away from some of his earlier statements about the benefits of deliberative democracy, and in some respects is offering a variant of liberal constitutionalism, a position we might want to describe as ‘Habermas Lite’. One reason Habermas and his supporters do not bring too much attention to voting procedures might be because we now have to admit that the minority is not ruling itself. The minority is not as free as the majority because it is compelled to obey a decision it did not choose. Assuming that one will be part of the majority some of the time, the best we can say is that we will sometimes rule ourselves. Permanent minorities pose a significant problem for any version of democratic theory because they are always compelled to obey. As Jane Mansbridge (1996) has reminded deliberative democrats, coercion plays a part in all legitimate democracies because those on the losing side are forced to accept the result. Her point is not only that coercion is required, but that it is valuable – we cannot have a stable political system without it. Hobbes makes similar points in his usual blunt manner when he notes that all forms of political rule require the use of force. The task, therefore, is not to try and escape power relations, because this is impossible; the task is to try and make those power relationships legitimate. This has led some critics to suggest that Habermas’s recommendation that we should seek ‘coercion-free’ communication as an ideal misconstrues the very nature of politics. Chantel Mouffe, for example, claims that the tension between liberty (liberalism) and equality (democracy) cannot be reconciled. Consequently, ‘the “consensus model” of democracy which informs theories of “deliberative democracy” is unable to grasp the dynamic of modern democratic politics which lies in the confrontation between the two components of the liberal-democratic articulation’ (2000, 8). Rawls (who also tries to achieve consensus) and Habermas have not succeeded in their stated task of reconciling
A comparison 27 freedom and equality because, she claims, they are incompatible: ‘this ultimate irreconcilability need not be visualised on the mode of a contradiction but as the locus of a paradox’ (2000, 9). If theorists attempt to keep liberalism and democracy together, they have to realize that perfect liberty and equality are no longer attainable. Instead, ‘pluralist democratic politics consists in pragmatic, precarious and necessarily unstable forms of negotiating its constitutive paradox’ (2000, 11). Because Rawls and Habermas fail to recognize this they attempt to ‘erase the very place of the adversary’ (2000, 14). Carl Schmitt thought that there was no place for pluralism within a democratic state and he argued that it could only exist among states. Within the state, pluralism undermines loyalty and the state becomes only a clearing house for conflict. Mouffe disagrees and champions pluralism. Her argument is that instead of trying to avoid conflict by banning it, as Schmitt suggests, or attempting to negate it through consensus as do Rawls and Habermas, we should leave the public arena as a space forever open to contestation. She notes that [f]or the Habermasian, the process of deliberation is guaranteed to have reasonable outcomes to the extent that it realizes the condition of the ‘ideal discourse’: the more equal and impartial, the more open the process is, and the less the participants are coerced and ready to be guided by the force of the better argument, the more likely truly generalizable interests will be accepted by all those relevantly affected. (2000, 88) This is bound to fail because deliberative democrats ‘are unable to recognize that bringing a deliberation to a close always results from a decision which excludes other possibilities’ (2000, 105). Mouffe agrees with the basic point made by Mansfield when she tells us that politics entails ‘an element of force and violence that can never be eliminated and cannot be adequately apprehended through the sole language of ethics or morality. We need a reflection of the political proper’ (2000, 130). Mouffe addresses other complaints against Rawls and Habermas. Rawls tries to maintain a clear separation between private and public in which the latter is independent of any concept of the good. It is an area of overlapping consensus and hence is not a realm of pluralism because there is a shared concept of justice. Mouffe says that this attempt to purge the public realm of conceptions of the good fails and Rawls’s descriptions of the reasonable and of the person are not
28 A comparison neutral. Habermas thinks he sidesteps these problems because his ISS is a strictly neutral procedure that leaves most things open to rational debate; for example, there is no preconceived notion of the person. Mouffe suggests that his procedure is not as neutral as he thinks and is infused with liberal principles. Neither philosopher can separate the procedural from the substantial, or the public from the private, and both are engaged in a futile attempt to find a way of insulating the public realm from the pressures and tensions of pluralism, and the paradox that exists between liberalism and democracy. My point would be that even if they could succeed in their task, the result of such procedures, if they really are free, fair and equal, has the capacity to be unstable and arbitrary. Mouffe thinks that ‘[s]uch a tension, though ineradicable, can be negotiated’ (2000, 93). She accepts pluralism as a fact, as does Rawls, but it is also constitutive of the very concept of modern democracy, meaning that it is something that should be seen in positive terms and embraced rather than viewed as a problem to be solved. While Rawls and Habermas recognize pluralism as a fact of life, they do not take seriously the pluralism of politics. Their vision is devoid of politics because they are advocating a form of pluralism without agonism. Rawls’s political structure becomes utopian because conflict over political and economic interests is removed and all competing ideas of the good are relegated to the private realm: ‘politics is reduced to a mere activity of allocating among competing interests susceptible to a rational solution’ (2000, 30). Rawls has already sorted out the distribution of primary goods, rights and resources from behind the veil of ignorance. There is no longer a place for democratic struggle in Rawls’s society except to decide who controls distribution. Consequently, the quest for consensus is the real threat to democracy because it misses the true nature of modern democracy which is pluralistic. She sees Rawls’s call for a consensus around reasonable pluralism as a way of sneaking a political position into a supposedly politically neutral claim. Rawls is really saying that those who are not liberals are not reasonable. She agrees with Rawls that people who refuse principles of liberalism should be excluded: ‘I have no quarrel with him on this issue. But this is the expression of an eminently political decision, not of a moral requirement’ (2000, 25). A similar criticism can be levelled at Habermas’s desire for consensus. Consensus means that there is no disagreement; no disagreement means we no longer need politics. Habermas does not doubt that there will be obstacles to consensus. For example, we might not be able to put aside our interests in the way he hopes; but these obstacles are not ones that are
A comparison 29 inscribed in the logic of the democratic procedure. Mouffe suggests that drawing boundaries between ‘us’ and ‘them’ is woven into the very fabric of democracy and hence consensus is neither achievable nor desirable. Instead, we need to ‘relinquish the very idea that there could be such a thing as a “rational” political consensus; namely, one that would not be based on any form of exclusion’ (2000, 32). I think she is correct to make this claim, and this is why I also think that the recent attempt to reconcile deliberative democracy and social choice theory is not going to work; at its core, social choice theory recognizes it is highly unlikely that there will be stable decisions without coercion. The drive for consensus that still animates Habermas and deliberative democrats is an illusionary dream that banishes the core of politics to the periphery, and this is something that social choice does not share. In this sense, social choice theorists are more attuned to real world power structures than discourse theorists and I agree with Ian Shapiro (2003), who argues that deliberative democrats are insufficiently attentive to the very unequal power relationships that exist in modern democracies. As Shapiro also notes, democracy seems to keep the minority playing the game rather than abandoning the process. One problem with Habermas’s idea of consensus is that the nearer we get to it, the less likely it is for those in the minority to ever envisage winning. This might lead to other avenues for political expression, including violent ones. One of the primary aims of deliberative democracy is to transform rather than aggregate preferences, which means that the common good is manufactured rather than discovered. Shapiro makes the reasonable request that democrats of this mould need to demonstrate how this would work with hard cases such as the abortion debate. Such a model only works ‘for those fundamentalists who also count themselves fallibilist democrats’ (2003, 26). There is also the danger that individuals and groups who hold out against the quest for consensus are easily identified, blamed and victimized. Fundamentalist politics is more likely to thrive in places with tight and shared identities than in places where there is a wide variety of viewpoints. In some respects, Mouffe is claiming that Habermas and Rawls are after a win-win situation in which the procedure is set up in such a way that there can be no losers in politics. Her complaint is a powerful one and suggests nothing less than that Habermas and Rawls have misunderstood the nature of the task at hand. They have misunderstood that ‘the political is not reducible to a rational moral calculus and always requires decisions’ (2000, 140). She also notes that some post-moderns also think of politics as a no-loss situation
30 A comparison in which they advocate an endless conversation with the ‘other’. As she points out, this also ignores the antagonistic dimension of the political and is equally blind to contemporary power relations. Such theorists tend to forget that there has to be a moment of decision in democracy. When we have this moment of decision, as we must, we reach a conclusion through the use of coercion, which brings to an end, at least temporarily, the possibility of other outcomes. In this respect, Mouffe is echoing the claims of Hobbes, Arrow and social choice theory. Like Mouffe, they ‘reject the very possibility of a nonexclusive public sphere of rational argument where a non-coercive consensus could be attained’ (2000, 33). Somewhat surprisingly, poststructuralists and social choice theorists have some things in common. A more down-to-earth criticism of deliberative democrats that also suggests they do not fully recognize the nature of politics comes from Robert Dahl and Edward Tufte, who argue that the ideal of deliberation is simply not logistically possible. Their arguments about the limitations of size on deliberation are very compelling and it is worth presenting the basic claim once again. One of the key conclusions Dahl and Tufte (1973) draw is that the larger the number of participants, the less we can accommodate sequential participation: if each citizen speaks once on the main issue, and his speech is then followed by a rejoinder from every other citizen, the number of speeches by each participant must increase as the square of the number of participants critics will point out that if the average time required for a speech were 15 minutes, if the bare quorum of 6,000 citizens were present, and if sessions were held 10 hours a day for 300 years, then [it] would require the Assembly to meet for 3,000 years to reach a single decision. Our axiomatics have enabled us, at the very least, to push an example or two to the point of utter absurdity. The essential point, however, is not at all absurd, even though it is painfully obvious: the number of people who can participate directly in a decision by speaking so as to be heard by all the other direct participants is extremely small. (1973, 70) Given these logistical problems it seems difficult to picture the sort of participation favoured by Benhabib: (1) participation in such deliberation is governed by the norms of equality and symmetry; all have the same chances to initiate speech acts, to question, to interrogate, and to open debate; (2) all
A comparison 31 have the right to question the assigned topics of the conversation; and (3) all have the right to initiate reflexive arguments about the very rules of the discourse procedure and the way in which they are applied and carried out. There are no prima facie rules limiting the agenda of the conversation, or the identity of the participants, as long as any excluded person or group can justifiably show that they are relevantly affected by the proposed norm under question. (1996, 86–7) Who do deliberative democrats have in mind when they talk of all those affected? People in Australia are affected by the result of American elections; should they get a say in the outcome? Put simply, on practical grounds it would be impossible for all affected persons to participate in the manner suggested by Benhabib. We can, of course, still have simultaneous action, but this requires participation through voting rather than through talking. Voting gets around the time restrictions that hamper deliberation and deliberative democrats, who provide us with a somewhat impractical model of democracy, are too quick to write it off as a second-rate democratic activity; it is, and always will be, the primary means of democratic participation. ‘Participation’ through reading, watching the news and thinking for oneself is also immune to sequential time limitations. A final related word of caution regarding the unrealistic expectations of deliberative democrats comes from Anne Phillips. She notes that feminism has often been attracted to a radical notion of small-scale participatory politics but cautions that this type of activity, where women are constantly in committee meetings, is not going to make it any easier for them to succeed in the world. The more demanding the time requirements of democracy, the more likely it is that it will be dominated by a relatively small number of people. Iris Marion Young (1996) finds the idea of deliberation broadly attractive, but suggests that it fails on two counts. The first is that it is too narrow and does not allow modes of expression other than critical argument (although John Dryzek (2000) has since attempted to address this issue). Young claims that this creates a bias against certain people and groups. Deliberative democrats tend to assume that factoring out political and economic power makes for a level-playing field, but Young claims that this still leaves some forms of communication prioritized over others, particularly along lines of class, race and gender. Deliberative democracy, in other words, favours white middle-class men. She says white men talk in more deliberative ways while women and minorities value more the ‘expression of emotion,
32 A comparison the use of figurative language, modulation in tone of voice, and wide gesture’ (1996, 124). The second failure is that we must begin from shared understandings and take the collective good as our goal if we are to reach common ground. Young echoes the criticisms offered by Mouffe that deliberative democrats tend to assume we can tap into an underlying consensus if we simply talk for long enough, and difference is seen as something to be transcended and overcome. Again, this favours some over others. Those persons and groups closest to the symbolic definers of society will be privileged by this approach. When we look to find commonality, we are not using dialogue to transform our preferences and values; we are looking to reinforce them. Dialogue should show us the partiality of our own positions, that we differ from others, and consequently that there is no consensus to be had other than accepting the best procedure for arriving at collective decisions. She offers ‘communicative democracy’ as an alternative in which difference is seen as a resource rather than a barrier for arriving at understanding. She also wants to include such things as greeting, rhetoric and storytelling as part of democratic communication. One might agree with her critique of deliberative democracy but not find her alternative any more compelling. Storytelling and rhetoric might perhaps lead to understanding, but it could also muddy the waters and the problems identified by social choice still remain once we try to reach a decision. Carol Gould (1996) makes the same point about Habermas’s reluctance to embrace pluralism and claims that he advocates discourse as a means of finding common ground rather than articulating diversity. She also claims that those who are inarticulate because of culture, fear, habit, inability, etc. are at a disadvantage. One could also add that those who are aggressive, argumentative, will not shut up, and have a great deal of stamina are unjustifiably privileged. When this is tied to Habermas’s rejection of strategic instrumental action, the consequence is that a lot of things get left out of deliberation. She prefers a public life in which there is more rather than less room for a diversity of views. Not only does Habermas drive at agreement, he aims at consensus among all who are affected by the decision. Gould claims that this has a chilling effect on those who might want to be dissidents and is likely to produce a culture of conformity. Gould also criticizes Habermas for ignoring decision-making in favour of discourse: ‘We may say that while decision without deliberation is blind, deliberation without decision is empty’ (1996, 176). Socrates might be a good example of why deliberation is not necessarily a good thing. His use of Elenchos demonstrates why discussions
A comparison 33 often do not lead to the conclusions we might hope for. In the process of discourse, Socrates usually asks a question, such as ‘what is x?’ Someone answers, often by providing a particular example, and Socrates then provides a counterexample of something we think of as ‘x’ but that differs in some crucial way from the example offered. The discussions that take place in Plato’s dialogues often end in confusion and a mutual scratching of the head, and by the end of the debate Socrates has often upset his interlocutors even though his sole aim was to arrive at truth. This is particularly clear in Book One of The Republic and it is not beyond the bounds of possibility that Plato is deliberately highlighting some of the problems that stem from unrestrained discourse. Some of these problems are as follows: 1 2 3 4
5
Discourse does not work against logical consistency, even if the person is starting from the wrong premise. Discourse often leaves us with no firm answers to important questions. Discourse often leads to antagonism. Discourse depends on the intellectual and moral qualities of the interlocutors. Deficiencies in either one will often lead the discussion away from the truth. The method requires that people are willing to change their beliefs in response to compelling arguments.
It is perhaps not too surprising that after Book One, the tenor of The Republic changes and it is left to Socrates to explain the requirements of justice to a willing and passive audience.
Criticisms of social choice theory Social choice theory provokes strong reactions from its critics. Perhaps, the most serious objection by discourse theorists concerns the nature, and more particularly, the creation of preferences. The social choice theorist assumes that preferences are fixed before the participant enters the decision-making arena, whereas the proponent of discourse focuses on the creation and alteration of preferences through the decision-making process itself. Discourse can change preferences in such a way that the end result of dialogue is consensus, even though the participants may have entered the discourse with very different preferences. Hence, it is argued that the sorts of problems discussed by Riker et al. do not apply at all to the ISS.
34 A comparison It is a valid criticism of social choice theory that it does not pay attention to the formation of preferences. Many critics of liberal democracy rightly suggest that there is more to the democratic process than simply aggregating preferences. Democrats should also be concerned with how preferences are formed. Votes tell us about people’s judgements but they do not tell us how or why these judgements came about. The assumption of deliberative democrats is that we are not solely instrumentally self-interested, that we will reformulate the issues from a moral point of view, and that we can engage in meaningful interpersonal comparisons. Through this process, our preferences change and converge. It is also suggested that social choice theorists have a very emaciated concept of democracy. They view it as a battleground for competing interests and pay little attention to the collective aspects of the democratic process that produce a sense of belonging and shared identity within a community. Focusing on voting rather than participation in a broader sense misses the fundamental character of a democratic polity. William Riker, for example, defends democracy on the very minimal grounds that it allows us to throw out elites. Democracy demands only that we have a popular veto that we can impose upon elected officials if they are threatening to become tyrannical. There is no sense of a collective good beyond securing peace and safety. One can question whether even such a limited demand from democracy holds up to Riker’s own arguments. He obviously thinks so, although he never justifies his position other than through his own procedural arguments. It seems to be the case, however, that even this thin requirement cannot be met if we accept Arrow’s conditions. If the outcomes of the democratic process are ambiguous, then the outcomes of the minimal brand of democracy supported by Riker must suffer the same fate as those arrived at through discourse. Elected officials thrown out of office by voting procedure ‘a’ might have held on to their position under procedure ‘b’. If no procedural method of voting can adequately produce unambiguous results, collective decisions are as troublesome for Riker as they are for Habermas. It would seem that no aggregative method can be defended on procedural grounds no matter which side of the debate one is positioned and Riker undermines his own defence of democracy. Many critics have pointed out that the amount of randomness and instability that is predicted by Arrow does not materialize in real world legislative bodies. The claim is that Arrow’s findings are interesting from a logical point of view but they are not troubling because the defining feature of institutions is stability rather than chaos, and the
A comparison 35 claim that cycling is a feature of legislative bodies has faced withering criticism. Mackie goes to great lengths to show that the instability identified by Arrow simply does not exist and he devotes hundreds of pages of his book to demolishing Riker’s studies that purported to demonstrate cycles. Mackie also makes a very impressive attempt to undermine the claim that outcomes vary drastically depending on which voting procedure is used. These criticisms of social choice theory are powerful and I will return to them in much greater detail in the next chapter.
Conclusion I made two suggestions in this chapter. The first is that social choice theory and theories of discourse are very similar on a procedural level. The second is that, once this is recognized, discourse theorists should be seriously concerned about the arguments of social choice and vice versa. If the claims of social choice theory are correct there is no way of deriving the legitimate consensus demanded of discursive decisionmaking and the arguments are fatal to the claim that discourse can provide what everyone wants, that it produces outcomes that reflect the preferences/beliefs of everyone, and that through discourse we can discover principles of justice. Habermas’s claim that we can have unfettered dialogue, greater pluralism and a reasonable chance at rational consensus would turn out to be false, and all that a popular decision can tell us for certain, under Arrow’s conditions, is which policy, or decision, or candidate wins, not which one represents a legitimate consensus; the more pluralistic a society becomes, the less we can know about shared outlooks. If the argument is correct that democracy and pluralism are interdependently linked, then the claims for consensus politics become less realizable the more democratic a society becomes. In reply, deliberative democrats argue that social choice theory is based on an incorrect assessment of how individuals think and behave. Our preferences are not fixed and they can be changed through reasonable and rational discourse. We can eventually arrive at consensus if we treat people with respect and accept the force of the better argument. Individuals are capable of treating other people fairly and they do not always try to maximize their own interests. It is also suggested that the conditions imposed on democracy by Arrow are unreasonable and that evidence from voting procedures in the real world of politics does not show the sorts of instability that is predicted by social choice models.
36 A comparison Which argument is right? The crucial point to keep in mind is that Arrow’s is engaged in an exercise in deduction and he is correct to say that the conclusions he arrives at flow logically from his starting point. In a sense, the argument is already rigged because no other outcome can be gleaned from the ‘mild’ conditions he imposes on democratic decision-making. If we accept these conditions we also have to accept his impossibility results. I have already suggested that Habermas does largely accept Arrow’s conditions, which means that the outcomes he hopes for cannot materialize. From a purely theoretical standpoint, Arrow is in the box seat. However, this does not mean we should abandon the ideas of deliberative democrats because they may be more helpful than Arrow’s in the context of real world political institutions. It might turn out that we get the best outcomes from deliberative procedures once we have left the rarified world of ‘pure’ democracy. As Mackie says, ‘[t]he Arrow theorem is a great piece of work. It illustrates an abstract limit case. It is a logical exercise, it does not describe the real world. The conditions of the theorem, especially IIA (A), are methodological assumptions, with no descriptive or normative force of their own’ (2003, 156). I argue that we have to abandon the claims of Habermas to the extent that his requirements for ideal speech resemble Arrow’s conditions, but in the next chapter I also suggest that we need to abandon Arrow’s conditions. Once we do this, the question remains open as to the most appropriate democratic procedures, including ones that place a high value on deliberation.
3
Circumventing Arrow’s theorem
Introduction The conclusion of the previous chapter is that Arrow gives a more accurate assessment of outcomes than Habermas, given the strict conditions both impose on participation. As a consequence of Arrow’s arguments, Riker claimed that we need to abandon all defences of democracy other than his own very thin ‘liberal’ concept. He supplied empirical examples, largely demolished by Mackie, to support the claim that democracy and rational outcomes are strangers to one another. At the bottom of Riker’s model of democracy is nothing more than an appeal to judicial retrieval; voters have the procedural capacity to produce a decisive result and remove office holders if they so choose. This is just about the ‘thinnest’ defence of democracy one can have. No attempt is made to foster support for democracy on moral grounds because Riker thinks it is not possible to morally defend results that are always potentially arbitrary, meaningless and unintelligible. The sole defence is that democracy is useful for preventing tyranny by promoting ‘negative’ liberty through a veto. Russell Hardin (1993) reaches a similar conclusion and suggests that we can only get stability if we limit choice through coercive means that are anti-democratic. Both theorists suggest that we have to abandon the attempt to defend democracy from a procedural point of view. Riker concludes that we should place our faith in market mechanisms rather than political institutions, given the inherent instability of the latter. But as Mackie rightly points out, the arguments of Arrow also apply to the marketplace. Riker is wrong to think that the market avoids the issues raised by social choice; most social choice theorists, however, recognize the general character of the theory and how it applies to social choices made in politics and in the market. Arrow himself states,
38 Circumventing Arrow’s theorem In ideal dictatorship there is but one will involved in the choice [and] no conflict of individual wills is involved. The methods of voting and the market, on the other hand, are methods of amalgamating the tastes of many individuals in the making of social choices. (1963, 2) And when Arrow is specifically discussing voting, he says that no method can remove the paradox of voting, and ‘[s]imilarly, the market mechanism does not create a rational social choice’ (1963, 59). The rather pessimistic conclusion that many have drawn from such comments is that democratic outcomes lack legitimacy.
What hope for democracy? Do Arrow’s arguments mean that democracy is doomed to struggle on without a firm theoretical foundation? Does it mean that there are no good justifications for the democratic process? I do not think we have to be overly pessimistic in our answers to these questions; I have been using social choice theory to undermine a particular view of democracy, not democracy per se. I too want to defend democracy against some of the claims made by Riker. Arrow himself suggested that social choice theory offers a particular and limited criticism of democratic procedures and I would agree with this assessment. I think we can utilize social choice to critique Habermas and some of his supporters, but still maintain a general support for democracy. However, such support does not have to be on procedural grounds and it is important to recognize that the theorem does shed light on the fact that stability does come at a cost. We have to pay close attention to the conditions that Arrow places on democracy if we want to give a compelling answer to the above questions. The problem of defending democracy in the face of Arrow’s findings should not make us as agitated as Mackie, who says that to ‘give up democracy, for example, in order to avoid violation of Arrow’s condition of the independence of irrelevant alternatives seems to me to be an absurd bargain’ (2003, 389). Not many people would suggest such a bargain. Arrow has attempted to address the concept of democracy in the light of his findings, and Sen notes that the pessimistic conclusions that one can draw from Arrow’s results should not overshadow his ‘immensely important constructive program of developing a systematic social choice theory that could actually work’ (2002, 70). Mackie might also not like Riker’s views or agree with his definition of democracy, but one cannot deny that Riker supported a form of
Circumventing Arrow’s theorem 39 constitutional democracy, even if the support comes at the loss of logical consistency. In fact, Riker seemed to think that in many respects the American system was highly commendable. It is false, therefore, to argue that social choice theorists are necessarily anti-democrats. As noted in Chapter 2, some commentators have suggested that while Arrow’s theorem is safe from refutation on a logical level, the amount of randomness in outcomes is much less than we might expect (Shepsle and Weingast 1981a; Weale 1999; Mackie 2003). This argument suggests that Arrow’s findings can be ignored because they do not map neatly onto the empirical world. For example, the spread of preferences may be limited to a fairly narrow part of the policy space so that even if we do get movement it will be narrowly confined. Mackie does grant that instability is possible, but he asks whether it is probable, and answers that as long as ‘voter preferences are sufficiently similar, a variety of voting systems lead to similar choices’ (2003, 50). There are several responses to these types of arguments. The first is that Mackie, like Arrow, is simply stating a tautology; if voter’s preferences are similar then they are obviously going to agree on the same outcomes. Arrow himself granted as much but recognizes that this does not solve the problem. We noted his requirement for nondictatorship in the Chapter 2, but it is worth reminding ourselves of his argument in a slightly different form: If we admit meaning to interpersonal comparisons of utility, then presumably we could order social states according to the sum of the utilities of individuals under each Even in this case we have a choice of different mathematical forms of the social utility function in terms of individual utilities; thus, the social utility might be the sum of the individual utilities or their product or the product of their logarithms or the sum of their products taken two at a time. So, as Professor Bergson has pointed out, there are value judgments implicit even at this level. (1963, 4) Arrow concludes, Even in the case where it is possible to construct a procedure showing how to aggregate individual tastes into a consistent social preference pattern, there still remains the problem of devising rules of the game so that individuals will actually express their true tastes even when they are acting rationally. (1963, 7)
40 Circumventing Arrow’s theorem An additional problem is that lots of preferences never see the light of day in contemporary democracies. If the median voter model is at all accurate regarding the spread of preferences across the left/right continuum, it is not surprising that politics is dominated by similar preference structures. But this is at the cost of those on the margins who are not represented. This is not necessarily a bad thing because democracy is supposed to represent the majority, but we should also accept that democracy is very much about creating winners and losers. Mackie suggests that cycles are not much of a problem because ‘people have similar preference orders because they live in the same world and they have similar interests in that world’ (2003, 98). The example he uses to support this claim is that people prefer torturing kittens over engaging in suicidal nuclear war. He could also have suggested that we all agree that eating ice cream is better than watching children being tortured. It is very likely true that preferences do converge in these instances but when we look to real political dilemmas concerning the distribution of scarce resources, such as health care, work safety, education, defense, etc., we are going to find large disagreements. There are also many policy issues, ranging from abortion to the right to die, where there are widely divergent opinions. In another example, Mackie asks us to imagine that there are three alternatives: (a) less tax collection, (b) more parks and (c) public safety. A cycle could take place among these alternatives but he claims it would be busted through deliberation, which would show that the cycle was the result of incoherent preferences because it is not coherent to vote for less taxes and more public safety. It is not, in fact, incoherent to support fewer taxes and more expenditure on safety because one might simply want the revenue raised by taxes to be lowered and redistributed in a manner that promotes public safety. But even assuming incoherent preferences, a better explanation would be that deliberation has not busted a cycle by changing preferences, it has simply made the preferences more coherent. Once the preferences are coherent, it does not mean that consensus will emerge because there could still be a fundamental divide between those who want to see a reduction in public spending and those who want to see an increase in expenditure on parks and public safety. Mackie argues that we can get close to consensus because we all live in a shared world and have similar preferences. If this is the case we do not need discourse because we already want roughly the same thing. And why is it so important that preferences are transformed by debate? It would seem that changing people’s preferences through deliberation would be counter productive given that we start from a shared perspective.
Circumventing Arrow’s theorem 41 He cannot have it both ways. If the task of deliberation is to transform preferences, we have to accept that they differ significantly prior to discourse, which means that the world we inhabit is not as homogenous as he would have us believe. Mackie also suggests that cycles are not a problem as long as people are motivated by considerations of fairness, which again supposedly leads to a convergence of preferences. This assumes that we all agree on what is fair, which is not the case. My preference would be to outlaw inheritance and private schools in an attempt to create greater fairness and equality, but I find little support for such views from the undergraduates in the classroom. Even if it is true that we are motivated more by fairness than selfishness, it is unrealistic to make the further assumption that this will lead to consensus through dialogue. It seems to be more realistic to assume that preferences will not all change to one focal point of agreement simply because we talk about the issue for long enough. Note that it is not enough for the supporter of discourse to demonstrate that preferences can change through dialogue; this seems like a sensible assumption. What also has to be demonstrated is that such preferences will all converge upon a single answer. This may be possible when participants are few and already have similar preferences before they enter into dialogue; workplace democracy may be an example of this. It becomes increasingly unrealistic when we expand the scope of discourse to include questions that are inherently pluralistic in nature, such as larger questions of ethics and justice or complicated policy issues. It is worth pointing out that there is always the possibility of introducing new alternatives into a fair procedure and hence broadening the policy space boundaries, especially if there is no germaneness rule. It is only when we start to introduce rules limiting access that we can also begin to limit arbitrary shifts in outcomes. Even if there is agreement among legislators, this does not demonstrate that consensus also exists; agreement often comes about through compromise and bargaining. The inevitable discrepancy between voter’s choices and the eventual policy outcome of legislatures makes it very difficult to judge if the wishes of the people are translated into law. The general claim made by Arrow’s critics is that we do find equilibrium in actual democratic institutions. The criticism largely misunderstands the force of Arrow’s argument. It is undeniable that most results seem to be stable in contemporary politics, but this is completely unsurprising given that all modern legislative bodies violate Arrow’s conditions. A major branch of social choice theory claims that stability can only occur if some form of order is imposed upon the system.
42 Circumventing Arrow’s theorem This order, or structurally induced equilibrium (SIE) as it is often called, comes from institutional rules and constraints placed upon participants that violate either condition U or D or both. In other words, equilibrium can be guaranteed but only through coercion of one form or another. Consequently, there is rationality in the form of forced stability, but this comes at a cost because the social welfare function has been weighted in favour of some participants over others. Riker was obsessed with outcomes and devoted his professional life to the issue. Mackie is similarly obsessed with Riker and he has used up a significant amount of time and energy demonstrating that his empirical results are mistaken. Both theorists are looking at democracy as it is practised within liberal democratic institutions that do not possess the properties of an IF setting and consequently it is not clear how useful their findings are for judging the value of Arrow’s claims. What Mackie has ably demonstrated is that it is not very fruitful to search for voting cycles in institutions that are not pure IF settings because one of the defining features of such institutions is that they are set up to provide stability. It is puzzling why Riker devoted so much time to this task, and why he would leave himself exposed to the harsh criticism aimed at him by Mackie and others. There is no point looking for apples in an orange grove. Perhaps, Mackie could also have put his resources to better use because the major conclusion we can draw from his exhaustive assault on Riker is that the latter was better at formal theory than he was at empirical research. A more apt title for Mackie’s book might have been Riker Attacked rather than Democracy Defended and the assault seems to be fuelled by more than intellectual curiosity given that he basically accuses Riker of plagiarism. The book also has dire warnings about the threat to humanity that is posed by social choice theory. Sen suggests that the motivation behind social choice is to avoid instability and arbitrariness in social choices, particularly within the realm of democratic decision-making, while paying ‘adequate attention to the preferences and interests of all its members’ (2002, 68). Note the words ‘all its members’. This seems like a laudable task and it is unclear why it stirs such animosity from Mackie. It is also difficult to understand the antagonism when we remember that the task of social choice theory is ‘to describe social welfare and not some sort of illfare’ (Arrow, 1963, 25). Mackie undermines some of his own claims about the virtues of democracy at the same time as he crushes the empirical results of Riker. One lesson to be gleaned from his in-depth case studies is that much of democratic politics is mired in duplicitous and self-aggrandizing behaviour.
Circumventing Arrow’s theorem 43 If we are going to test the claims of IF democratic theorists, we need institutions that fit within the parameters of the argument and in the next chapter, I provide the Continental Congress as one possibility. Even in an institution that did resemble the IF setting described by Arrow, I found examples of instability but was not able to discover a voting cycle. This is probably a good thing because the issue of cycles has been overstated. The discovery of such cycles would not have rendered democracy meaningless, and not finding any does not mean that democratic outcomes are legitimate. A more balanced conclusion is that democracy can be a messy business, and this also should not make us despair. It is only those who think it can deliver us the truth, consensus, and/or a pure and unadulterated general will who will be dissatisfied with this outcome. I conclude that it is not very fruitful to try and dislodge Arrow’s deductive argument through empirical examples of institutions that do not reflect the necessary conditions of his argument. We should remember that Riker does not claim that results will always be unstable; his point is that a fair procedure is always open to the possibility of randomness and distortion. Given this uncertainty, we should be very wary about claiming that the results produced by majority rule actually are manifestations of the popular will. Yet, this seems to be a reasonable requirement of a democratic system for those who make large claims for the benefits of participation. This argument does not have to rest on voting as a means of measuring what people want; any procedure for expressing democratic will formation will suffer the potential consequences identified by social choice theory if the means of will formation and the method of its expression treat all participants as free and equal. Social choice theory is also criticized because of its minimalist, instrumental view of rationality; discourse theory, as already discussed, posits agents who act rationally when they behave as intersubjective discoursers. These competing views of rational behaviour have been very damaging to any form of dialogue between the two camps and have led some practitioners to assume that the theories must be mutually incompatible. I suggest that the different conceptions of rationality are actually irrelevant to the specific topic at hand, i.e. to the assessment of the rationality of outcomes. The procedural mechanics of the ISS and of the IF setting operate as they do independently of the behavioural properties of the participants. It does not matter which theory of rationality is proposed given the procedural structure of a non-coercive discourse setting. Arrow seems to suggest that he is starting from a neutral position in regard to individual motivations: ‘if no prior assumptions are
44 Circumventing Arrow’s theorem made about the nature of individual orderings, there is no method of voting which will remove the paradox of voting neither plurality voting nor any scheme of proportional representation no matter how complicated’ (1963, 59). I think Arrow is correct to suggest that outcomes from a free and equal democratic setting are independent of any particular manifestation of human motivations about preference orderings. Let us assume that rational choice theory has a false view of the self, and also assume that we are completely socially integrated beings; let us assume that the self actually fits a more discursive ideal. Would this invalidate the social choice argument? No: even assuming that individuals fit this description, the results from pure majority rule are potentially the same. A discourse among cooperative, rational altruists is as likely to be disturbed by disequilibrium as one among selfish maximizers, as long as we make the reasonable assumption that participants have competing preferences after dialogue that cannot be reconciled. Habermas clearly does grant this assumption: ‘The political interests and values that stand in conflict with each other without prospects of consensus are in need of a balancing that cannot be achieved through ethical discourses [t]he required balance of competing interests comes about as a compromise between parties that may rely on mutual threats’ (1994, 5). Democratic results can remain unstable even if we have the most sympathetic view of human nature. This is because the procedural logic of Arrow’s argument remains undisturbed regardless of how we view human beings. For the social choice model to work, we do not need to assume instrumental self-interest; we simply have to assume that people have different preferences. The argument that results can theoretically cycle is not based upon a philosophical view of human nature, but is a logical conclusion of the model. The statement that agendas can be manipulated is based on an examination of how agendas actually work and not on a particular view of human psychology. It is also a fact that the structure of institutions does limit the size of the policy space and consequently creates equilibrium through limiting choice. Simply because one may dislike the instrumental view of the self, therefore, is not a good reason to ignore the serious implications of social choice theory. Even if we could create the ISS and populate it with the discursively rational character types demanded by Habermas, we are still not guaranteed rational and consensual outcomes. It is not enough, therefore, to criticize social choice theorists because of a perceived view of rationality; if one wishes to refute them, one has to address the procedural logic of their arguments. And as Mackie has argued, the
Circumventing Arrow’s theorem 45 rational choice view of the self is often mistaken for something it is not. He notes that in rational choice theory the individual is supposed to maximize utility but this does not necessarily entail the idea that the person is selfish. The person is judged rational in terms of how he or she pursues preferences and it is better to think of individuals as purposive rather than selfish. If we are concerned with psychological motivations rather than procedural outcomes, the only way to guarantee consensus would be to make all those who take part in the discourse exactly the same. If each participant has the same knowledge, the same epistemological approach, a desire for the same ends and the same strategy for attaining those ends, then, and probably only then, can we guarantee consensus. Arrow recognized this, and it is worth quoting him at some length on the issue: Suppose that we do not assume in advance the shape of the preferences of any one individual, but we do assume that all individuals have the same preferences for social alternatives. This implies a social-minded attitude and also a homogenous society. If we consider the preferences in question to refer not to expressed preferences but to the preferences which would be expressed if the corruptions of the environment were removed, the assumption of unanimity is the idealist view of political philosophy. In this case, the obvious way of defining the social welfare function is to choose some one individual and then say that the social welfare preference scale shall be the same as his. This satisfies all the conditions except the conditions that the social welfare function not be dictatorial. Under the assumptions of this section, since it makes no difference who is the dictator, the condition of nondictatorship loses its intrinsic desirability. (1963, 74) This could be an interesting way to reformulate Habermas’s argument, because what Arrow is saying is that the ideal speech situation, if it allows us to reach unanimity, would satisfy his conditions (because dictatorship is no longer an issue) and democracy is in the clear. The problem is that unanimity is an unreasonable requirement to build into a social welfare function, and as Arrow notes, ‘may seem obviously contrary to fact’ (1963, 81). In Chapter 2, I noted that many commentators think that such unanimity is neither possible nor desirable and now it seems that if it was possible we would no longer need dialogue anyway. Under such conditions we would not have consensus through discourse but a pre-conditioned outcome that all could agree to in
46 Circumventing Arrow’s theorem isolation from one another. We end up with a Rawlsian hypothetical justification of agreement – but this is precisely what Habermas is trying to get away from: ‘practical discourse transforms what Mead [and Rawls] viewed as individual, privately enacted role taking into a public affair, practiced intersubjectively by all involved’ (1990, 198). The whole point of dialogue for Habermas is the fact that we do disagree. But as soon as we admit this, it becomes much more difficult to demonstrate that there is a link between unfettered dialogue, consensus and truth. Social choice theorists think it is better to assume that we do not all have the same preferences, even after extended dialogue, and because of this to assume also that instability is always possible in a pure democratic setting.
Can deliberative democracy and social choice theory be made compatible? John Dryzek and Christian List (2003) have recently provided a very interesting argument in which they attempt to reconcile social choice and deliberative democracy. I will focus on this paper because it is the best effort so far to marry the two theories together. They claim that (1) if certain conditions are present, meaningful collective choices are possible, (2) the constraints required to bring about these conditions are consistent with deliberative democracy, and (3) deliberation also helps to bring about these conditions. The best type of procedure is when judgements are expressed in a decision, either by voting or group discussion, prior to which deliberation has taken place. This applies to first-order decisions regarding particular outcomes and second-order decisions regarding institutional arrangements. They note that deliberation can affect preferences by providing new information, drawing attention to new arguments, making people justify their preferences to others, and making them talk and listen to each other. Each of these, they claim, can help to overcome problems of collective-decision-making identified by social choice theorists. One way to demonstrate this is to avoid the unwanted results that Arrow identified by relaxing the conditions of social choice theory. They claim that deliberation is a way to do this. The argument is presented as a series of five if/then statements: 1
‘If deliberation induces individuals to reveal their preferences and views truthfully, then strategic manipulation becomes less of a threat to deliberation’
Circumventing Arrow’s theorem 47 2
3
4
5
‘If deliberation induces preference structuration then both cycling and strategic manipulation become less of a threat in deliberation’ ‘If deliberation helps uncover or create the tacit issue-dimensions that ‘cause’ a lack of preference structuration then dimensionspecific aggregation in accordance with all of the conditions of Arrow’s theorem (except condition (u)) becomes possible’ ‘If deliberation can produce agreement on what the set of relevant alternatives is then agenda manipulation becomes less of a threat’ ‘If deliberation can produce agreement on an interpersonally comparable evaluation variable for assessing individual interests and a decision principle for aggregating individual interests into a collective outcome, then a solution to Arrow’s problem that is consistent with all of Arrow’s conditions becomes available’ (2003, 27–8).
It might strike the reader that there are a lot of if s in the above statements. I will examine a couple of them before making a more general comment on the argument. The first claim they make is that ‘Group deliberation induces individuals to reveal their preferences and views truthfully’ (2003, 9). This statement (Hypothesis 1 in the list) immediately rules out, by definition, claims that individuals might act strategically, and this means we do not have to place limits on actors that would undermine the free and equal nature of participation. If individuals do not act strategically, Dryzek and List claim that there might be decision procedures that avoid cycling, such as the single transferable vote. Everything hinges, therefore, on the empirical accuracy of this hypothesis. Dryzek and List claim that deliberation leads to truth-telling because (1) the risks and penalties attached to false disclosure are too great and (2) that truthfulness begets greater cooperation. There are a variety of reasons for doubting these claims. First, if we do expose the strategic actor for a cad, it might have a minimum impact because the actor still has to be treated as a free and equal participant in the process. Otherwise, we are likely to have people barred because of accusations of manipulation, which could become a method of strategic manipulation in itself. Second, as current politics shows us, dishonesty is often accompanied by large-scale support. Third, there might be greater risks for true disclosure than for dishonesty – for example, where deliberation is over the issue of guilt or innocence related to an offence, or where defeat in deliberation leads to a large
48 Circumventing Arrow’s theorem financial loss. Fourth, for the same sorts of reasons, truthfulness might bring all cooperation to an end: ‘Did you like the meal I tirelessly prepared for you?’ is sometimes a risky question to answer truthfully. Fifth, the argument makes the optimistic assumption that those who do not tell the truth will be easily exposed. The argument about truth-telling and cooperation does not assume that there will be repeated interaction, but it is much more persuasive if this is the case. However, as Dahl’s arguments about size indicate, it is very unlikely that a person will speak to all participants once, let alone several times. This suggests that even if Hypothesis 1 is correct for repeated group interaction, it is very unlikely to be correct for anything approaching large-scale democracy. Dryzek and List point to some evidence from prisoner’s dilemma experiments that show cooperation can take place in one-shot situations. However, it is a long stretch to draw conclusions about politics from experiments with very small payoffs. If deliberation does limit strategic action, a better place to measure it would be in real legislative bodies, which, as Weale notes, are already deliberative arenas. The evidence from here suggests that strategic action and dishonesty are alive and well. Next, they claim in Hypothesis 2 that ‘the profile of personal preference orderings after a period of group deliberation will satisfy (or approximate) single-peakedness’ (2003, 15). In other words, ‘[w]e suggest that, if, through deliberation, (i) a particular generalizable interest becomes focal, and (ii) this generalizable interest can be associated with a single dimension, then (a high level of) single-peakedness is a likely consequence’ (2003, 16). This is true but does not greatly worry social choice theorists. For example, they quote Riker, who states that ‘[i]f, by reason of discussion, debate, civic education, and political socialization, voters have a common view of the political dimension (as evidenced by single-peakedness) then a transitive outcome is guaranteed’ (Riker [1982, 128], quoted in Dryzek and List [2003]). Quite correctly, Riker does not suggest that all of politics is unstable, but as he also points out, on many issues such agreement is highly unlikely and when it does occur we never know if this result is the outcome of fair procedures or bargaining and manipulation. A related suggestion from Dryzek and List is that issues should be broken down through discussion into more manageable dimensions where it will be possible to get a single peaked outcome. This again might be true, but how is one to do this given the requirement of deliberative democracy that all those affected by the result are part of the discourse? The example they use concerns whether
Circumventing Arrow’s theorem 49 there should be tariffs on exports and imports. It would be impossible to allow everyone affected by such a result to participate in the discussion. This leads me to my more general criticism. The force of their argument rests on the power of deliberation to circumvent the problems identified by Arrow. Even assuming that the argument is correct, it can only be useful in the very specific and limited circumstances in which democratic deliberation does take place. The authors state at the beginning of the paper that ‘the essence of democratic legitimacy is the capacity of those affected by a collective decision to deliberate in the production of that decision’ (2003, 1). As Dahl has shown, this is a hopelessly unrealistic way to think of democracy in anything other than a meeting of a dozen people. Once the group size expands beyond this, then all of the if statements noted earlier are nullified and democratic politics as it is practised is still prone to the problems identified by Arrow. We will never be able to justify democracy if all those affected by the decision have to participate in the production of that decision. The problem, therefore, is that Dryzek and List are still unable to adequately address the U condition required by Arrow (and Habermas).
Abandoning democratic purity Given that Arrow’s argument is logically sound, anyone who embraces the conditions he places on democracy, such as Habermas, is bound by logic to agree with his conclusions or commit an error of reasoning. How then to defend democracy in the face of Arrow’s results? One of the flaws with theories of social choice and deliberation is that they both paint an unrealizable picture of democratic practices. Arrow and Habermas have a shared vision of a power-free democratic zone in which all are treated as free and equal inviolable sovereign bodies. I think that the real lesson we should learn from this idealized view is that power-free democracy is unworkable. A trade-off in values is necessary if democracy is to be feasible. We cannot have the full democratic cake and eat it as well. The key question, then, is what type of trade-offs should we accept? The task is to look for rules of participation that might violate the U domain requirement, or IIA, but that still give us a viable democratic procedure. Some might claim that this is a second-best defence of democracy given that the best option of free, equal and uncoerced participation seems to have been abandoned. But this is not necessarily the case because the conditions imposed on participation by Arrow and Habermas might
50 Circumventing Arrow’s theorem themselves be subject to criticism on democratic grounds. This is what Arrow says: We have now imposed apparently reasonable conditions on the construction of a social choice function. These conditions are, of course, value judgments and could be called into question; taken together they express the doctrines of citizen’s sovereignty and rationality in a very general form, with citizens being allowed to have a wide range of values. (1963, 31) I suggest that we abandon Arrow’s brand of democratic purity and do as he suggests above, which is to question some of the value judgements he makes. If we want to escape Arrow’s conclusions, we have to demonstrate that the conditions he places on democracy are unnecessary, too stringent, or undesirable. Given the deductive nature of Arrow’s claims, it is best to take a closer look at what he assumes to be the necessary conditions of democratic decisionmaking; we should examine the inputs rather than the outputs to find chinks in the argument, and the best way to do this is to take a close look at the supposedly weak conditions he imposes on social welfare functions. The conditions of universal domain, non-dictatorship, the independence of irrelevant alternatives and weak Pareto efficiency are presented as necessary conditions of a free and fair procedure, but is this really the case? Mackie has criticized condition P, which states that if all voters prefer an option, then society prefers that option. Mackie argues that the condition is undermined by certain anomalies in human rights claims. For example, all members of society might agree with the idea of universal rights but still vote to deny these rights to a subset of the population. This might well happen, but it seems that this would not be a violation of condition P because it is clear (assuming for the moment that we can map preferences onto outcomes) that the preference of voters is to deny rights to some members of the community, and this is expressed in the vote. I think Arrow is correct therefore to maintain that condition P is a necessary condition of democracy. Mackie has also gone to great lengths to challenge the IIA condition. His argument is sophisticated and complex and cannot be repeated here, but the conclusion he draws is that it is acceptable for decision rules to break IIA, which, he suggests, has anyway been much misunderstood. There are two main versions of IIA doing the rounds: IIA(A) (A for Arrow’s version) and IIA(RM) (RM for Radner and Marshak),
Circumventing Arrow’s theorem 51 which is also called contraction consistency. Mackie suggests that most commentators have misunderstood Arrow’s account and have inadvertently been using the RM version. He claims that IIA(A) is obviously not an intuitive condition of democracy as many people claim given that most social choice theorists do not get it right. He concludes that Arrow’s proof has to rest on the (A) version of IIA rather than the (RM) version and that this condition can be broken without violating rationality: My goal is to show that the conditions are not requirements of rationality, are not justified by naked appeal to intuition, and to do so I present examples that illustrate the absurdity of obeying the condition. If I present plausible counterexamples to the conditions, then my argumentative goal is achieved. (2003, 131) The reason that IIA(A) is so crucial is as follows: Arrow’s possibility theorem shows that social ordering, universal domain, Pareto principle, nondictatorship and IIA(A) are inconsistent. Ray (1973) shows that social ordering, universal domain, Pareto principle, nondictatorship and IIA(RM) are consistent. (2003, 130) Mackie begins by using an example borrowed from Arrow to undermine the IIA(RM) version. Take a person who has an ordering over three possible states of the world – she prefers Cold War to Hot War to Disarmament. She rationalizes her priorities by claiming that Cold War is preferred to Hot War because there are few casualties, but Hot War, even though the casualties would be great, is better than capitulating to the enemy. If, however, she is faced with pair-wise decisions, she still prefers Cold War to Hot War, but she now prefers Disarmament to Cold War. She seems to have gone against the requirement for contraction consistency because her choice should not change when one of the alternatives has been removed. This seems to violate the RM version of IIA but Mackie says that it also seems rational for the woman to make the choice. He claims, ‘I have just shown by example that it is possible for a rational person’s preferences over alternatives to vary by the menu of alternatives available’ (2003, 133). It is not clear that Mackie has demonstrated that IIA(RM) can be violated with this example. What has happened with the withdrawal of Hot War as an option is that the parameters of Cold War and Disarmament have now changed and, as Mackie himself notes, Disarmament
52 Circumventing Arrow’s theorem no longer means that one has to capitulate to the enemy. It is not the case therefore that one alternative has simply been removed from the preference structure; what has happened is that the very meaning of Disarmament has now changed and as it is no longer the equivalent of surrender, it is not possible to think of it as the same preference that existed during the possibility of Hot War. So while it is true that her vote does change when one option drops out, which seems to violate IIA(RM), it is not clear that the vote does reflect a change of mind because the preferences now mean altogether different things. It depends on what IIA demands. If it demands that no change whatsoever occur because an alternative drops out, then the example works to support Mackie’s claim. If it demands that no change occur only as long as the options that are left mean the same thing, then it is not clear that Mackie’s argument works. A more plausible scenario would be to imagine a person who prefers ice cream to chocolate to apple pie. IIA(RM) demands that ice cream still be preferred to apple pie should chocolate no longer be an option, and this seems quite reasonable. We would certainly require a good explanation for why apple pie is now preferred to ice cream if we are to continue to think of the person as rational. It seems, therefore, that Mackie has not provided a compelling counterexample and he has not demonstrated the absurdity of the condition. Mackie is more concerned with Arrow’s version of IIA. After his discussion of IIA(RM), he says, ‘[n]ow for the more important chore, to show that violation of Arrow’s IIA(A) may be substantively rational’ (2003, 133). It should be noted that in this attempt to refute IIA(A), the focus changes slightly from the argument related to IIA(RM). It differs because we are now dealing with groups rather than individuals, one of whom drops out of the scenario. He asks us to imagine a reception where the ‘caterer will only provide one beverage, either beer or coffee’ (2003, 133). The organizer inadvertently e-mails possible attendees the previous year’s itinerary and asks them to rank their preferences over beer, coffee, water, tea, milk and pop. Five people from the business school reply, and have the same preference structure of beer > coffee > water > tea > milk > pop. Four people from the law school reply, and they all have the preference structure of coffee > beer > water > tea > milk > pop. Beer prevails under the Condorcet and Borda methods so we have a clear winner. Unfortunately, the people from the law school are dropped and are replaced by four theologians who rank the beverages coffee > water > tea > milk > pop > beer. Beer still wins using the Condorcet method, but the Borda count winner is coffee because the
Circumventing Arrow’s theorem 53 theologians ranked beer last (it should be noted as an aside that this outcome is troubling for Mackie because it highlights something that he has been eager to play down, namely that results can differ according to the voting mechanism that is used). Because the organizer is ‘a political scientist indoctrinated in the Arrow theorem, and a believer in the IIA(A) condition’ (2003, 133), he only counts the preferences for the relevant alternatives of beer and coffee, which means that the theologians end up thirsty and angry. Mackie suggests that this example undermines IIA(A). Again, I am not sure this is the case although I grant that this might be because I have not fully understood the intricacies of IIA(A) or Mackie’s argument. He claims that ‘almost everyone would find the Borda winner normatively superior to the Condorcet winner, in other words, that they would find it intuitively obvious that the IIA(RM) and the IIA(A) should be violated’ (2003, 135). It is interesting to note the normative element that has stepped to the forefront of the argument. The reason that IIA(A) is unacceptable in the above example is because it is unfair. Mackie asks us to imagine a different scenario in which the participants had talked beforehand and the theologians were able to express their dislike for beer; we are invited also to imagine that they are teetotalers who have a right to be served a beverage at the meeting. These are important considerations and they demonstrate that the caterer and organizer are not the most sympathetic characters, but it is not clear that they undermine IIA(A). Mackie does make an intuitively appealing case for choosing coffee over beer in his example. It does seem unfair that the teetotalers have nothing at all to drink. But it also seems reasonable to suggest that a decision on the only two possible alternatives should reflect the preferences over those options, and not preferences over phantom beverages that are unavailable. If the vote is limited to the available beverages, then beer remains the Condorcet and Borda winner and this gives us strong reason to get out the bottle opener. If we are going to base our decision on the possible rather than the actual alternatives, how are we to decide on the extent of the inventory of drinks? Why should we limit it to the beverages listed above? It is a well-known feature of the Borda count that it can produce different results depending on the alternatives on offer. In Mackie’s example, the relative place of coffee and beer in the ordinal rankings can change significantly with the introduction of further beverages. The inclusion of more alcoholic tipples will push coffee down the list of the business school types whilst not necessarily dislodging beer from the most preferred option. And coffee might slide down the
54 Circumventing Arrow’s theorem preference order of the theologians when healthier drinks are added to the menu. Why not also find out whether preferences for beverages are linked to preferences for food? The ranking of red wine might well be affected by what is being served as the main course. Mackie accepts that the Borda count is logically susceptible to manipulation but he claims that this has little practical significance; unfortunately his own example shows how the addition and omission of drinks to his list leads to the conclusion he is seeking. It is not clear, therefore, that IIA(A) is undermined. Mackie’s argument seems stronger when he questions the requirement of IIA(A) that our preferences over alternatives in the available set should not influence our choice between two options within that set. For example, if there are four alternatives ‘a’, ‘b’, ‘c’ and ‘d’ in set ‘S’, Arrow’s condition requires that our choice over ‘a’ and ‘b’ is not influenced by our preferences for ‘c’ and ‘d’. Mackie suggests that a better name for IIA(A) would be the ‘pairwise comparison condition’ (2003, 137) and he makes a persuasive case that a rational choice to spend taxpayers’ dollars on either a college, a stadium, or a museum might be based on more than a simple pairwise comparison between each option. There seem to be good arguments for and against the IIA(A) condition and Mackie has succeeded in his task of providing a plausible counterexample, but he has not demonstrated ‘the absurdity of obeying the condition’. As we have seen, one of the main reasons Mackie is so concerned with IIA(A) is that it rules out the Borda count as a method of vote aggregation: ‘Arrow’s is a strong independence condition. Slight weakenings of it allow the Borda count or the Young-Kemeny rule as possible social welfare functions and further weakenings permit further voting procedures’ (2003, 123). The Borda count can prevent cycles that occur under pairwise comparisons and Mackie asks, ‘[i]f cycles are indeed a threat to the very intelligibility of democracy, then shouldn’t we eagerly seek for voting rules that resolve cycles’ (2003, 121). He makes a strong argument for the Borda count as an aggregative procedure and I share a lot of his enthusiasm for this method of counting votes, but this is a value judgement we both make, not a logical proof that it is better than any other form of aggregation. The value judgement comes from accepting the fact that the Borda count is an ordinal measure in which we list our preferences from the most to the least desirable. Such a mechanism does not allow the intensity of preferences to enter into our vote-counting procedure. The difference in score with the Borda count depends on the number of available alternatives and not on the intensity of feeling about each alternative.
Circumventing Arrow’s theorem 55 Even if we grant that IIA(A) can be breached, it is difficult to see how such a voting mechanism can tell us the general will. If other people hold that preference intensity should be included in the social welfare function, there is not much I can do to persuade them otherwise. Mackie says that ‘in the pure voting exercise all we have are ordinal data, and thus we cannot go beyond the Borda count. But if we have sound cardinal data then why not use it’? (2003, 135). I prefer not to count the intensity of preferences (cardinal data) because it is a very difficult thing to measure. It also has a strong undercurrent of anti-majoritarianism about it because intense minorities can prevail over less intense majorities. The option of counting the intensity of preferences has similarities with the sort of constitutionalism I will be arguing against later in the book. I think it is safe to assume that many liberals would not be as taken with the Borda count as Mackie given their fear of the rule of the majority. No amount of impartial deliberation is going to resolve an issue such as this, and consequently there is always going to be disagreement even on the most fundamental issue of aggregation. One might also ask how far we should relax the IIA(A) condition. This question demands value judgements that will provide many different answers, which will in turn influence which voting procedures are allowed and which are not. So, even if we can legitimately use the Borda count, we still do not have a democratic system that would find unanimous agreement. Colin Bird (2000) has made a very thorough attempt to circumnavigate the whole of Arrow’s conditions. His argument revolves around the concept of self-government, which he takes to mean consciously pursuing one’s own will. Social choice theory denies this can happen on a societal level because we can only arrive at unambiguous results if there is a dictator or if the range of preference possibilities is narrowed in advance. But if either of these things happens we cannot say that we are getting a true reflection of the will of the people. Bird somehow has to address this argument. Next, he identifies three notions of self-government. The first is when institutions are operated and controlled by people drawn from the populace rather than from outsiders. The second is that public decisions reflect and reveal the will of the collective that is ‘authentically their own’ and which is revealed to governors who execute this will. The third is a moralized concept where government acts only on an identifiable conception of citizen’s common good that is oriented to the real good of the whole community, impartially considered. This is a Rousseauian way of thinking of the issue.
56 Circumventing Arrow’s theorem Bird rejects this third sense of the term because, he argues, a society can have a will that is unpalatable. This third sense assumes that if society does not exist in accord with its common good then it is not properly self-governing (Habermas would also say that it is irrational). Self-government 2 is independent of self-government 3 and this is a key point because many problems of self-government are removed once we take the third variety out of the picture. To illustrate the difference between self-governments 2 and 3, Bird gives the example of a society that recognizes putting people in space is not the best use of tax dollars for promoting the common good, but does it anyway. Here the common good and the will of the people do not coincide. Bird argues that the question needing to be answered collectively is not ‘what is the common good?’ but ‘what do we want to do?’ His claim is that even if we cannot know the general will in a Rousseauian sense, we can still know it in the second sense of the term. Having identified the different ways we can think about selfgovernment, the next step in the argument states that societies and individuals structure the will in analogous ways. For either one to be self-governing requires what he calls ‘reflective authenticity’ (RA). One measure of this is provided by the conditions placed on social choice by Arrow, but does this define the only possible standard for RA? Bird suggests not, and claims that Arrow’s conditions do not apply to individual choices. If it is true, therefore, that individuals and societies are self-governing in similar ways, it is possible for society to violate Arrows conditions and still be self-governing. How, then, are individuals self-governing? Bird leans heavily on Frankfurt’s distinction between first-order desires and second-order desires. The latter are the authentic reflection of the will and manifest our true personal identity. These desires allow us to make an ‘all things considered’ judgement that might override our more immediate but less important first-order desires. The crucial point Bird makes is that such claims undermine Arrow’s requirement for non-dictatorship because preferences are given an unequal weighting and the second-order desires act in a dictatorial manner in relation to the first-order desires. The example he provides is of a wife who might have an immediate preference for infidelity, but whose secondorder desire for fidelity triumphs; RA is demonstrated when she acts faithfully. Bird suggests that the main problem with Arrow’s argument is that it does not allow for the privileging of some preferences over others and hence does not seem like a useful model for decisionmaking. In particular, he questions ‘the requirement that the individual will depend exclusively on the profile of currently emergent
Circumventing Arrow’s theorem 57 preferences without any assessment of their character in relation to operative ideals’ (2000, 570). Bird’s argument in turn rests on the validity of his claim about the lexical ordering of preferences. It is not altogether clear how this applies at the collective level, but it seems that, as with the wife, society can have first-order values and second-order values. The latter are the most important and can override the former if they go against the prescriptions of the latter. For example, a society has a second-order value that individuals are treated equally; despite this we might find public policies that discriminate based on skin colour. The argument must be that the second-order desire is the real will of the people and can dictatorially trump the first-order desires, just as fidelity is the real will of the wife. This argument requires two things in order to be persuasive: (1) that the concept of individual self-governance is compelling, and if it is compelling, (2) that the model of self-government applies to individuals and collectives. I argue that the description of self-governance is not compelling. A similar argument to Bird’s is made by Charles Taylor (1985). I will spell the argument out in some detail because it provides the best case for the self-governance thesis. Taylor suggests that self-governing of the form discussed by Frankfurt and Bird is about ‘the exercising of control over one’s life’ (1985, 213). He proposes ‘[a]n exercise concept of freedom [that] requires that we discriminate among motivations If we are free in the exercise of certain capacities, then we are not free, or less free, when these capacities are in some way unfulfilled or blocked’ ( 1985, 215). Taylor suggests that we are not properly self-governing when we act upon a desire we do not identify with: ‘Freedom is important to us because we are purposive beings. But then there must be distinctions in the significance of different kinds of freedom based on the distinction in the significance of different purposes’ (1985, 219). In a similar manner to Frankfurt, Taylor distinguishes between ‘brute’ and ‘import-attributing’ desires, and ‘lower’ and ‘higher’ (firstand second-order) desires. These categories of desires are important because they allow us to determine whether a person is self-governing. It is not necessarily the case that brute desires rob us of freedom; I may have a strong desire for many basic pleasures, golf for example, but if I identify and exercise control over this desire then I am acting in a self-governing manner. However, generally ‘we have to see our emotional life as made up largely of import-attributing desires’ that ‘are of great significance for me, meet important, long lasting needs, and represent a fulfillment of something central to me’ (1985, 224).
58 Circumventing Arrow’s theorem Such desires can deprive us of freedom and self-government if they are mistaken: ‘Now how can we feel that an import-attributing desire is not truly ours? We can do this only if we see it as mistaken, that is, the import or the good it supposedly gives us a sense of is not a genuine import or good’ (1985, 224). Taylor asks, ‘what is it to feel that a desire is not truly mine? Presumably, I feel that I should be better off without it, that I do not lose anything in getting rid of it, I remain quite complete without it [l]osing these desires we lose nothing, because their loss deprives us of no genuine good’ (1985, 224–5). Taylor utilizes the examples of a man whose spiteful behaviour is ruining a relationship he values highly and a woman who will not engage in public debate despite a desire to do so to illustrate his point. These people recognize their hierarchically superior second-order desires but lack self-governance because they do not act upon them. A person can follow a higher, second-order desire and still be unfree because ‘[f]reedom now involves my being able to recognize adequately my most important purposes, and my being able to overcome or at least neutralize my motivational fetters I must be actually exercising self-understanding in order to be truly or fully free’ (1985, 228–9). Rather ominously, Taylor now suggests that the person cannot be the final judge as to whether or not the person is free: ‘the subject himself cannot be the final authority on the question whether he is free; for he cannot be the final authority on the question whether his desires are authentic, whether they do or do not frustrate his purposes’ (1985, 216). Sometimes, Taylor suggests individuals without the correct ordering of desires are unfree, and sometimes he suggests they are less free. The first claim seems to more accurately represent his position on selfgovernment, especially when he claims that immoral desires make us unfree: our ‘fundamental purpose (can be) shot through with confusion and error’ (1985, 227). The lack of freedom in this sense is due to a flawed understanding of what is important in life: ‘the capacities relevant to freedom must involve some self-awareness, selfunderstanding, moral discrimination and self-control, otherwise their exercise could not amount to freedom in the sense of self-direction’ (1985, 215). Even if an individual has an ordinal ranking of preferences the individual can lack self-government if the preferences do not fit within a defensible normative framework. Taylor suggests that full self-governance requires a life-plan that provides moral coherence. A summary of Taylor’s position suggests several ways we can lack self-government:
Circumventing Arrow’s theorem 59 1 2 3 4 5
When external factors prevent the pursuit of the good. When brute desires take precedence over higher desires. When some import-attributing desires are pursued to the detriment of other import-attributing desires. When we are mistaken about the value of an import-attributing desire. When an import-attributing desire is morally flawed.
I have spelled out Taylor’s argument in some detail because it is a more detailed and nuanced formulation of the first-order and second-order values discussed by Bird. Both arguments are unsound because they rest on a flawed understanding of desires and preferences. One problem with identifying self-government with higher and lower desires is that we have to claim a person is not responsible for their actions while in the grip of a lower desire. A man who cheats, lies, steals, etc. is not acting according to a higher desire and hence is not self-governing. This will not come as pleasant news to those who have been cheated, lied to or robbed, and who might want to suggest that the man is responsible for his actions. Bird and Taylor claim that making normative judgements is part of our humanity; to then claim that we might be unfree when making such judgements is unconvincing. Taylor even claims that desires that ‘are of great significance to me, meet important long lasting needs [and] will bring me closer to what I really am’ (1985, 224) can prevent self-government. Flathman (1987) disagrees and suggests that one cannot make a mistake and be unfree. The very idea of something being a mistaken decision suggests that there were alternative choices available. Free action requires the ability to evaluate options, weigh consequences, and make judgements according to the available evidence. It stretches credibility to claim that such decisions are not necessarily a reflection of self-government. The distinction between first- and second-order desires also suggests that there is a link between self-government and morality. Are we to abstain from condemning the racist because he is acting to a firstorder desire and hence not self-governing? If he is not self-governing he is not free and cannot be held responsible. Taylor expands selfgovernment in a way that incorporates the ethical (how we ought to live), and consequently jumbles up issues of freedom with those of moral behaviour. The higher/lower distinction is a useful resource for judging the morality, but not the freedom, of an action. I conclude that the psychology adopted by Bird is not persuasive for the same reasons that Taylor’s argument fails. Bird is critical of the notion of
60 Circumventing Arrow’s theorem self-government 3 because it changes the question from ‘what do we want?’ to ‘what should we do?’ but he reintroduces it when he accepts Frankfurt’s distinction between first- and second-order desires. I am also troubled by the claim that the individual model applies at the societal level. If it is correct, it would mean that voting is bypassed as the method for ascertaining the will of the people because it is more likely to be an expression of first-order preferences. To identify second-order values that express the real will of the people, we will have to examine legal documents and constitutions that are removed from the immediate statements of preferences found in voting. The argument seems to rest on the claim that the will of the people can be known independently of any particular social welfare function. Even if this is true, it would only help us with constitutional matters. How are we to know the will of the people if we are deciding among the huge number of issues that do not signal any particular second-order value? How are we to know the will of the people if there is a clash among second-order values? If there is a tension between the so-called first- and second-order values on a societal level, it is preferable to accept the first as the real preference. Even this assumes we are in a position to identify voting as a proper reflection of the will of the people, which is, of course, denied by social choice theory. My final concern is that the door is left open to a form of guardianship. If there is a distinction to be made between first-order and second-order desires and values, then we have to agree with Taylor that the individual is not the final arbiter regarding his or her best interests. This also holds on a societal level, which necessitates an elite such as legislators or judges to distinguish between our ephemeral and possibly dangerous first-order impulses and our real and authentic second-order preferences and values. I address this issue in the final chapter. Despite these criticisms, I think Bird is heading in the right direction, which is to accept the logical strength of Arrow’s conclusions but to question the necessity and validity of his conditions. I concur, but for different reasons, with his assessment that the U domain requirement is too restrictive. Bird rejects the U condition because, he claims, it is acceptable to remove some first-order preferences from the policy space in order to better reflect authentic second-order preferences. The U requirement has been generally understood to mean, according to Pildes and Anderson, that ‘citizens should be free to prefer any policy option at all and to rank any options in any way they want, meaning that no institution should have power to declare certain choices out of bounds at the start’ (Mackie 2003, 93 quoting Pildes
Circumventing Arrow’s theorem 61 and Anderson). I have argued that the distinction between first- and second-order desires is not a good reason to violate the U condition, but I do agree with Mackie, who also questions the legitimacy of this requirement and is willing to accept that ‘destructive self-seeking preferences should be excluded from public consideration’ (2003, 94). Hence, he is willing to place boundaries on the decision-making space at the cost of freedom and equality of participation. There is obviously much ground for debate regarding the limits to government intervention, but there are very few political philosophers who would deny that the dimension space does not have to be limited at some point, even for self-regarding actions. There is considerably more agreement when it comes to placing limits on other-regarding activity that causes illegitimate harm. It seems, therefore, that most political philosophers would reject condition U because it is not normatively defensible to argue that all preference structures should be admitted. Some alternatives are irrational, others are nasty, irrelevant, too time consuming or costly. Arrow insisted on condition U because he wished for complete information in his social choice world, but such a requirement is normatively unappealing, and it is an impossible requirement to realize in practice. It is not such a bad thing, therefore, to explicitly abandon the requirement for free and equal participation that fuels the ideas of Arrow and Habermas.
Limiting participation The general conclusion of the last section is that, contra Arrow and Habermas, it is necessary from a practical and normative point of view to put limits on the democratic arena. The U condition seems to be normatively unappealing and it is at least open to question whether IIA(A) can be breached without violating certain rationality requirements. Democracy only works when it is infused with power, not when power is removed. This is particularly important for proponents of democratic discourse; rather than defending a procedure without a power structure, they need to discover specific rules that create stability without, at the same time, completely undermining the deliberation that is so necessary for the legitimacy of outcomes. The task is not to create a coercion-free system of participation as Habermas suggests, but to develop a system with justifiable limitations on freedom. Too much participation and we do not get stability; too little and we end up with an overly constrained system. Sen suggests that social choice is a game of brinkmanship in which possibility and impossibility are a short step from one another. This
62 Circumventing Arrow’s theorem brinkmanship requires that we find the best and least cumbersome rules that lie as close as possible to Arrow’s impossibility results. To avoid the undesirable poles of chaos and tyranny ‘[w]e have to get on with the basic task of obtaining workable rules that satisfy reasonable requirements’ (2002, 75). Two of the primary tasks of democratic theory, therefore, are to decide (1) rules for arriving at decisions and (2) rules for deciding what should be allowed in and what should be kept out of the democratic arena. The way forward, if we wish to defend democracy at the procedural level, is to find institutional rules that limit the dimension space but that do so without giving up too much freedom and equality. The question is whether it is possible to do this, or whether we have to abandon procedural defences of democracy. The branch of social choice theory that highlights structurally induced equilibrium might help us answer this question. SIE suggests that through the introduction of a few rules we can solve the problem of cyclical instability. Hence, there is no problem producing equilibrium in actual legislative bodies, but this equilibrium is the result of rules that undermine the strict requirements for equal dialogue found in democratic discourse. As Shepsle argues, stability comes from ‘jurisdictional arrangements and rules of procedure’ (1979, 51). Shepsle links stability to committees with gate-keeping powers in which decisions are broken down according to a division of labor within committees. Once the committee sends its decision to the floor there can be no further amendments to the committee’s proposals. In Shepsle’s model, committees decide on one issue at a time; the decision is then passed to the floor for acceptance or rejection as a whole. This does create stability but the problem is that the procedure is not very fair or democratic because the committee’s decisions may in no way reflect the preferences of the majority. But it does create median voting in a dimension-by-dimension policy space. No result other than equilibrium is possible with this model, but stability comes at severe costs to democratic participation; results coincide with the median position of the committee but can be a long way from that of the floor. More complicated models that incorporate sophisticated voting behaviour and more complex rules tend to show that such features lessen the chances of equilibrium. This means that once we loosen the rules of participation somewhat, the preferences of actors become more influential on outcomes; the problem is that this leads us back to greater instability. It seems that we are never going to have a completely fair and rational system; there is always going to be some
Circumventing Arrow’s theorem 63 trade-off between fairness and efficiency. Shapiro (2003) claims that cyclical instability is not a big problem for democracy because cycles at least give those in the minority the hope that they will someday gain power. This is correct, but this does not get to the deeper philosophical problem with democracy, which is how to justify a decision-making procedure that will always favour some over others. We can never have a system that treats all participants in a fair and equal manner. If there was one, it is a fair bet that it would have come to light by now. Even in an IF body the discussion had to stop at some point and a vote had to be taken. On a different day using a different counting mechanism the decision might well have been different. Hence, even if cycles do distribute power, the fact that we decide issues with one voting procedure rather than another injects an element of unfairness. I think we simply have to accept this, and with it the conclusion that democracy cannot be completely fair. Once we introduce specific rules that place limits on discourse, and promote some desiderata over other desiderata, we also have to abandon claims that outcomes are fully legitimate, or true or represent a consensus. But this is still probably the most productive path to follow because, as demonstrated by Arrow and supported by my case study in the next chapter, we also cannot make such claims for the results of a pure democratic procedure. It remains difficult, therefore, to suggest that democracy approximates the popular will. How do we know it approximates, and by how much? 60 per cent? 75 per cent? Even if voting rules do not produce drastically different results, we still cannot conclude that the result is a reflection of the popular will. This is particularly the case when democracy is practised on a large scale. The relation between democracy and size leads to what Dahl and Tufte call the ‘Chairman’s problem’. They note that [a]s the number of citizens in an assembly increases the problem of the chairman becomes more severe. He cannot possibly give everyone the floor; hence he must pick and choose [t]o avoid confusion, perhaps an agenda must be prepared. If the issues on the agenda are complex, additional information will be needed. To focus debate, a recommended policy may be desirable. Who is to set the agenda, gather the needed information, and make the initial recommendations? How are these citizens chosen?’ (1973, 72) Problems such as these lead Dahl and Tufte to state the seventh axioms in their list: ‘Whenever a democratic system is larger than a working committee, leaders emerge who participate much more fully
64 Circumventing Arrow’s theorem in political decisions than the average citizen’ (1973, 71). Hence, we soon move to conditions that violate the requirement for free and equal participation. Finding the correct decision rules to impose order is not easy. For example, too much information distributed to all members about the other participant’s beliefs is more democratic, but can make it easier to manipulate the agenda. Here we have a rule that is open and democratic and creates stability, but a form of stability that is in the interests of those who can manipulate outcomes. Hence, not even rules that create stability and democracy are always useful. I think Rawls is correct when he suggests that we can only arrive at a consensus if we deprive individuals of a lot of information. Supra-majority rules that demand unanimity, or close to it, can greatly favour the status quo. This is particularly a problem for discourse theorists because it demonstrates that the demand for consensus can be very restrictive and greatly favours the possibility of non-decisions. In fact, discourse can create great stability, but of the wrong kind because it is a stability that results in deadlock if no consensus is reached, which privileges the status quo. There are, therefore, good and bad ways of creating stability; logrolling and vote trading create stability but are again questionable. We also have to recognize that equality can hinder the democratic process because as long as power is equally distributed we might fail to reach a decision. Again, this is not because actors are self-interested, but simply because when equal power is combined with alternative preferences there is no guaranteed way to move from deadlock. The more diversity we have, the less likely we are to have single-peaked distributions and hence the less likelihood there is of consensus. We have to ask ourselves whether at some point it is better to have coercive rules that distribute power unevenly. Although it sounds unpalatable, what is needed is a system of rules that prevent certain preferences from being heard; this may be because they are grossly intolerant to the interests of others, or ludicrously simplistic, or unworkable in practice or simply because they are specifically aimed at preventing agreement. One legitimate limitation could be to disallow preference structures that undermine the equal right to citizenship. We might also want some rules that prevent filibustering or that allow only one issue to be voted on at a time (this proposal was passed recently concerning the initiative process in Colorado). Such rules could also limit proposed options to those supported by a sizable number of people. These rules limit the freedom of participants, but they might be necessary if we want
Circumventing Arrow’s theorem 65 to promote stability and agreement (on a severely constrained choice) over the alternatives of chaos or the imposition of decision by one person. We also need to look at meta-rules that govern the institution. These are crucially important because they determine whether policy outcomes have any resemblance to the preferences of actors before they enter into the dialogue. Focusing on meta-rules helps us to realize that we are never going to get full stability; as the last few years in Eastern Europe have shown, institutions themselves are also open to being changed. Shepsle has argued that institutions are susceptible to ‘inherit’ the instability that we find within the institution itself: ‘If the social preference relation defined by decisive coalitions is cyclical over outcomes, then the cyclicity will be inherited in social preferences over institutions’ (Shepsle 1986, 68). Institutions are themselves nothing more than congealed tastes: ‘the people whose values and tastes are influential live in a world of conventions about language and values themselves. These conventions are in turn condensed into institutions, which are simply rules about behavior, especially about making decisions’ (Riker 1980, 4). Institutions are chosen as policy issues are chosen; the only difference is that such choices tend to be more stable. Shepsle has tackled this issue and argues that one way to protect institutions from inheriting instability is to make it risky, through the imposition of sanctions, for those who attempt such change. This is quite draconian but may be necessary for a democratic system to function at all. Institutions are also stable because actors are loath to meddle with a mechanism that facilitates the cooperation necessary for them to achieve their goals: ‘Institutions, then, look like ex ante agreements about a structure of cooperation’ (Shepsle 1986, 74). Rather than face the uncertainty of instability at the level of the institution itself, participants are more likely to favour maintaining known rules: ‘inheritability is short-circuited by uncertainty’ (Shepsle 1986, 75). This is perhaps why ‘institutions persist in ways that ordinary policies do not’ (Shepsle 1986, 76). I would also propose that another way around the inheritability problem is to make it a rule that a proposal needs, if not to reach, then at least to approach unanimous support in order to change the institution itself. The stability that comes from such a requirement, while often a drawback in everyday policy making, becomes a bonus at the level of meta-rules. When we get stability in a system that does not overwhelmingly favour the status quo, we do so because the system has avenues of access, but also has rules that bring an end to dialogue; hence the system is not
66 Circumventing Arrow’s theorem fully legitimate. A workable theory of democracy has to deal with this muddled compromise. I just argued that a case can be made for a unanimity rule at the meta-level, but any rule that attempts to promote unanimity on all democratic decisions will run into trouble. Habermas’s concept of consensus is similar to the social choice understanding of unanimity where, ‘once one of the proposals has achieved a unanimous majority no other proposal from the Pareto-efficient set can achieve unanimity when placed against it’ (Mueller 1989, 103). Buchanan and Tullock (1962) argued for consensus as the fair basis of collective decisionmaking because it is the only rule that can, with certainty, lead to Pareto-optimal outcomes, but they also accepted non-unanimity because the costs of reaching unanimous decisions are too high. Many social choice theorists have accepted the claim that unanimity (consensus) has the potential to be a fair decision rule if it can be realized. The problem is that Buchanan and Tullock are wrong to suggest that unanimity does guarantee fair results. A unanimity rule can lead to manipulation of outcomes in numerous ways. For example, a minority can hold out for unfair advantages, knowing that the majority will settle for a less than optimal decision rather than no decision at all. Consensus does not necessarily result in equity and can instead perpetuate pork barrel politics. This can even happen when there is agreement on the desirability of the public good under consideration. Consensus or unanimity also does not provide us with a solution to the free-rider problem; as Mueller says, [w]here an individual might, under a voluntary revelation scheme, gamble on the rest of the group providing an acceptable quantity of the public good without his contributing, under the unanimity rule he might gamble on the group’s reducing the size of his contribution rather than risk his continual blocking of the collective outcome. (1989, 51) The major problem with unanimity and hence with Habermas’s demand for consensus is that each individual has an absolute right, which cannot be overridden, to defend his or her interests. This right of non-interference can block the interests of an overwhelming majority and the effect is to vastly favour the status quo; if one participant or one small group abstains then no decision is made. The unanimity rule does not allow for a decision where there is unresolved conflict over preferences, which means that any policy issue that is zero-sum in nature cannot be resolved by consensus. Unanimity is
Circumventing Arrow’s theorem 67 only possible when outcomes are positive-sum. There are many issues where unanimity is not possible because preferences diverge. This can sometimes happen, as Tocqueville suggested, because participants are often driven to the extremes of their own positions in the heat of debate. Hobbes was another theorist who saw the possibility of debate leading to intransigent, aggressive behaviour. This is why he went to such great lengths to clarify the meaning of words. Leviathan, as well as many other things, is a political dictionary that is supposed to clear up the meaning of controversial words and concepts. Even within groups that one would expect to be fairly cohesive, environmentalists and animal rights activists for example, we see large divisions between fundamentalists and pragmatists. What are we to do in this instance? Sen suggests that ‘one answer is to insist on unanimity for a change, and if there is no such unanimity for any proposed change, then to stick to the status quo’ (1970, 25). But, as he also says, ‘[t]his method is one of supreme conservatism’ (1970, 25). It is also the case, as Buchanan and Tullock note, that the requirement for unanimity is akin to the minority rule of one person. This seems difficult to justify in the name of democracy. It is possible to arrive at consensus through discussion and debate and through compromises and trades ‘equivalent to the logrolling process’ (1970, 26), but such outcomes might not be fair: it could not be overemphasized that what compromises people are ready to accept depends much on their own assessments of their relative bargaining power. The fact that all members of a community come to accept a certain social situation does not necessarily mean that it is unanimously preferred to other social alternatives. (1970, 26) Deliberative democrats state as a positive that people’s preferences change and converge on consensus as the result of discourse, but this might not be as unproblematic as we are led to believe. As Arrow says, ‘If individual values can themselves be affected by the method of social choice, it becomes much more difficult to learn what is meant by one method’s being preferable to another’ (1963, 8). One might also ask why consensus should be the point at which discourse stops. It might be only a passing moment and other alternatives could be introduced that will again move people away from agreement. The general difficulty, therefore, with rule-induced equilibrium is that the outcome is biased in unavoidable ways. For instance, if a democratic decision is made between various alternatives, the means
68 Circumventing Arrow’s theorem of calibrating the votes can determine which option wins or loses. It is quite feasible for option ‘a’ to win under a Condorcet form of voting, option ‘b’ to win when plurality is the deciding principle, option ‘c’ when using cardinal utility and option ‘d’ when the winner is decided according to Nash’s calibrations. The preferences of the participants remain unchanged, yet we arrive at many different winners simply because of the rules of the aggregating procedure. Mackie has shown that this is often not a problem within contemporary legislative bodies. But these bodies have already winnowed the vast array of possible preferences down to a small number that are already of a similar nature, and hence he has not demonstrated that different voting methods lead to similar outcomes over all possible states of affairs, as demanded by Arrow. When we start introducing powers of agenda setting, limiting access to debate and agenda formation, etc., we move further away from a purely discursive setting. Any form of decision, whether decided by vote or some other method, distorts results in a way that favours one outcome over another. Weale attempts to negate Arrow’s claims by pointing to SIE; he suggests that many institutions gain their legitimacy because of their established pathways to choice in which outcomes are understood in terms of the institutional limitations that are placed on decision-making. This is what happens with pair-wise majority rule, which satisfies all Arrow’s conditions except transitivity, a condition demanded by Arrow so that options are not favoured or hampered depending on the order in which they come up for decision. Weale makes the point that is later echoed by Mackie, namely that if we ask how frequent in practice is the turbulence associated with majority-rule cycling, the answer appears to be surprisingly little there are strong institutional pressures to prevent a synoptic evaluation of alternative policy combinations. The principle means by which this is accomplished in functioning democracies is by political parties, which reduce the many dimensions of potential competition down to two or three. Moreover, policymaking is in practice highly sectorized. Policy communities are an established feature of different fields of policy, forming intellectual and political reference-groups for different issues. (1999, 145) Weale is right to note that cycles do not seem to occur in such majority-rule settings, but this is precisely because of SIE that determines path dependency, which seems to support the social choice claim that we can have either stability through rules and procedures that
Circumventing Arrow’s theorem 69 limit options, or fairness and the potential for instability. It needs to be stressed that the importance of Arrow’s results is unrelated to how real world institutions actually come about decisions; the importance lies in the findings about decision-making under certain specified conditions. I agree, therefore, with Weale, Mackie and others that instability is not a big issue in contemporary legislative bodies, but this is because such bodies do not incorporate the type of procedures that one would find in an IF setting. The SIE that we find in the world does not support Mackie’s conclusion that ‘the cycling hypothesis must die from lack of evidence’ (2003, 92). There is too little evidence available from IF settings to make so bold a claim and if we are interested in getting a fair result rather than a binding result, Arrow’s results are still of real concern. I disagree with the conclusions they draw from the lack of cycles, partly for reasons already mentioned, but also because the absence of cycles does not mean the presence of stability. Weale grants that he is more concerned with getting a result than finding the purest form of democracy and hence he supports a Condorcet mechanism of pair-wise competition because this still gives us a majority-based outcome. He is willing to sacrifice Arrow’s requirement that the final choice is independent of the pathway to the choice. He is also willing to agree that such decisions do not tell us what the general will is because we are only voting on an issue-byissue basis, which leaves many alternatives never seeing the light of day. Under such a system, we would know that the majority prefers ‘a’ to ‘b’ and Weale is satisfied with this result. The pair-wise method, therefore, is only a comfort to those who do not mind having a result that primarily reflects the procedure of agenda setting itself rather than the free and equal choice of participants. Voting on pasta over pizza for dinner is never going to tell us anything about our preferences for Chinese, Indian or French food. Ultimately, pair-wise comparison might be the best we can do, but if so we should recognize it as a less-than-optimum outcome and reign in our claims for the possibilities of democracy. It is less than optimum because there does not seem to be a clear way of deciding how to restrict the domain to two options. It is worth asking how much instability there would be in real world democratic systems if the domain was not limited by political parties, and we really could vote for anyone and any policy we wished. Is it sensible to think that we could find a rational formulation of the general will under such conditions? And is it really accurate to suggest that because voters choose from among a few candidates and policy platforms of political
70 Circumventing Arrow’s theorem parties, we are getting an accurate reflection of their preferences? The answer to both questions is no. As noted in Chapter 2, one of the major criticisms of social choice is that it places too much emphasis on voting as a means of judging popular preferences and it is suggested that we can discover the collective will by some other procedural method that avoids the problems mentioned above. It is worth noting at this point what I think are the foundations of political democracy within social choice and within more participatory theories. At a minimum, they both require some form of participation and the most fundamental and central aspect is voting. That is to say, voting is the major procedure for facilitating popular choice, whether in small groups or national assemblies. More recent work of Habermas suggests that he thinks decisions arrived at through liberal democratic institutions and voting mechanisms are legitimate: it is not discourse of an ethical type that could grant on its own the democratic genesis of law. Instead, deliberative politics should be conceived as a syndrome that depends on a network of fairly regulated bargaining processes In legislative politics the supply of information and the rational choice of strategies are interwoven with the balancing of interests. (1994, 5–6) Habermas continues by claiming that ‘Informal public opinionformation generates “influence”; influence is transformed into “communicative power” through the channels of political elections’ (1994, 8). Habermas makes it clear that he leans more towards the liberal than the republican view of democracy when he states that [a]ccording to the republican view, the people are the bearers of a sovereignty that in principle cannot be delegated. Liberalism opposes this with the more realistic view that in the constitutional state any authority originating from the people is exercised only ‘by means of elections and voting and by specific legislative, executive, and judicial organs’. (1994, 9) It seems clear that the practice of voting is now an integral part of Habermas’s theory of democracy. It is important to recognize the radical shift that takes place with such statements because once we vote on an issue we have already given up on achieving consensus. As Mackie notes, voting is simply an aggregative mechanism and would
Circumventing Arrow’s theorem 71 not be necessary if we had already reached a decision that reflected the preferences of all involved. Cohen takes a similar line and argues that when, as is often likely, consensus is not attained then ‘deliberating concludes with voting’ (1989, 23). This line of thought opens the door for all of the social choice criticisms. Despite this nod towards majority rule, Cohen also claims that justification ‘rests on terms acceptable to others that all who are governed by collective decisions must find the bases of these decisions acceptable’ (1996, 100). It is not clear what this means. If it means that the procedure and discourse were acceptable then it is not a strong requirement because liberal democracies already provide this. But this also holds the possibility of a very divided polity as we find at the current time in the United States. The problem is that many of the procedural requirements of deliberation are ephemeral and vague and more attention needs to be paid to the complexity of issues, the opportunity costs of participation, avenues and opportunities related to communication, and the incentives to participate, if we are going to accurately assess the benefits of deliberative politics. Central to this more limited defence of democracy is still the demand that the results of voting in some way are a manifestation of the popular will. This is facilitated by free speech, organized platforms, freedom of association, etc., all of which are in themselves part of the democratic participatory process. But they are useful in that they facilitate the formulation of a popular preference that is displayed through the actual procedure of voting. It is crucial to focus on the fact that, for most forms of democratic theory, the only way we can attempt to measure will formation is through the casting of votes. For deliberative democrats, the result of voting after deliberation will supposedly be the final expression of the common interest, otherwise some of the claims for the liberating effect of participation would be lost. We could not argue very convincingly, for example, that we are free because we live according to laws we impose upon ourselves, or that we have arrived at a consensual understanding of justice, if at the same time the procedure by which we arrived at these laws does not in some fashion demonstrate the popular will. Habermas may respond that dialogue is still more important than voting. But as he sees democracy as a method for dealing with conflict, at some point the talking has to end and a decision must be made. This decision is then meant to be an expression of the preferences of the participants; it is an uncovering of a collective understanding. Once again we are left with voting as the crucial final stage in the decision-making process for all forms of proceduralist democracy.
72 Circumventing Arrow’s theorem If one does not concede that voting is the clearest manifestation of democracy, one will have to come up with some other means of socially aggregating collective decisions that escape the paradoxes of voting. In fact, we will probably no longer be able to define democracy in procedural terms, as deliberative democrats wish to do, but instead we will have to define it as a very specific form of social living in which the collective will is expressed in a variety of social interactions. If something as straightforward as voting cannot give us coherence, however, it is difficult to imagine more complex forms of social intercourse providing a clearer view of the popular will. To cast voting aside as irrelevant is to increase the stakes considerably for the strong democrat. The most recent work on democratic discourse does not take this path but relies instead upon liberal institutions and voting practices to arrive at legitimate decisions. I have neither the space nor the competence to draw up a specific list of institutional rules that would allow us to walk the tightrope between freedom and stability. The major point I wish to make is that the rules, whatever they eventually look like, are going to be a compromise that will appeal to the value judgements of some people but not to others. Some people are going to prefer the Condorcet method because of the arguments presented by Weale. Others will favour the Borda count after reading Mackie’s impressive description of its merits even though he says that these arguments ‘are not a plea for the widespread adoption of the Borda count’ (2003, 156). Mackie also notes that a combination of discourse and ‘constraints of communication, ideological coherence and budget’ (2003, 390) might act to collapse dimensions. This echoes the argument of Miller (1992) and it is a good point, but we have to remember that we are still imposing constraints. Mackie suggests that deliberation and aggregation are not an either/or situation because they complement one another. As noted earlier, Weale makes a similar point and I agree with this as long as we are clear about what we mean by deliberation. It is not the kind outlined by deliberative democrats where all involved are allowed to speak, because Dahl has shown that time constraints make this impossible. Even if this type of deliberation does possess all the qualities that its supporters suggest, it is of little use for real world democratic systems because only a small number of people can deliberate on a limited number of central issues. Despite Mackie’s sterling effort to justify democracy from a procedural point of view, he does also confess to the shortcomings of this approach when he says that ‘there is no ideal voting system, in the same sense that there is no ideal dinner, no ideal residence and
Circumventing Arrow’s theorem 73 no ideal holiday, simply because there are always tradeoffs among desiderata’ (2003, 441). Ultimately, the difference between his position and mine might not be that large because we both recognize that ‘in practice democracy is a messy and imperfect business’ (2003, 30). Where I part company with Mackie and Weale is my unwillingness to grant too much legitimacy to the outcomes of the various aggregative methods for making social choices, and I remain convinced that we cannot get too far in our support for democracy from a procedural point of view. Thankfully there are grounds other than procedural ones for defending democracy. J.S. Mill argued that democracy promotes autonomous individuals who are capable of self-determination. Selfrule and political rule reinforce one another and Mill hoped that the final outcome of democracy would be an improvement in the condition of humankind. It is easy to get too carried away about the educative effect of participation, particularly when it is limited to voting once every few years. But if democracy is expanded to include greater participation on a day-to-day level some of Mill’s hopes might be realized. Carole Pateman took this approach and claimed that ‘the theory of participatory democracy stands or falls on the educative function of participation’ (1970, 44). Radcliff and Wingenbach make a persuasive case that democracy can be defended on these sorts of grounds even if it does produce incoherent outcomes: ‘one can believe in the legitimacy of the democratic process while being aware of its shortcomings, for much the same reason that one can believe in the general notion of juries while admitting that they sometimes err’ (2000, 990). Despite my criticisms of deliberative democracy, I support the claim that discourse in general is a good thing and can promote legitimacy, even though I do not think that it leads to consensus, or allows us to escape from coercive power structures. Talking to one another is a valuable exercise in itself. It helps us to gain a better understanding of similarities and differences regarding our goals, ambitions and aspirations. It can shake us from our prejudices and provide us with a different point of view. This might ultimately lead to deeply held antagonisms because, despite the hopes of deliberative democrats, dialogue has incendiary qualities that can fan the flames of disagreement; but an informed prejudice is better than an uninformed one. Discourse and democracy are not necessarily linked, but the empirical evidence supports the claim that freedom of speech, particularly on political matters, is best protected in democracies.
74 Circumventing Arrow’s theorem Democracy also fosters equality. As I argue in Chapter 6, it is a necessary condition of democratic practices that participants are treated with equal respect. This does not mean that everyone has to be treated the same because democracy distributes its rewards primarily to those in the majority and minorities can be coerced into accepting decisions to which they are opposed. Equality means that everyone can take part in the decision-making process, but it does not mean that political rule cannot be imposed on citizens. This, however, should not be confused with arguments about the tyranny of the majority, and as I will also argue in Chapter 6, democracies can be defended because they tyrannize less than any other form of political association. Majority rule is very unlikely to lead to extremism which, by definition, exists on the margins rather than with the masses. I agree, therefore, with Riker, who claims that the best defence of democracy is that it protects us from tyranny. But I think it does so for very different reasons than the ones he articulated. It is important, of course, that political leaders feel that they are held accountable, but the real reason democracies are reasonably safe is that they generate a type of social and political culture that is not amenable to tyranny. I do not have the space to outline these sorts of arguments in great detail and it is not the aim of this book to develop a fully fledged defence of democracy, but I do think that any such defence will have to rest primarily on these sorts of claims rather than procedural ones.
Conclusion In this chapter, I have suggested that we should abandon some of Arrow’s conditions for democracy. I argued that social choice and deliberative theories of democracy are incompatible, and I conclude by suggesting that we should abandon the rarified world of Arrow and Habermas, who place unworkable conditions on the democratic process. In Chapter 4, I intend to put the theoretical assumptions made by deliberative democrats and social choice theorists to a test, and I present an empirical study that will hopefully shed some light on the general issue of democratic stability. As I stated at the beginning of this chapter, there is very little evidence from the real world of politics to support the speculative claims of either camp. Does deliberation help or hinder collective decision-making? Do we get more instability the closer we get to ‘pure’ democracy? Riker attempted to shore up his arguments about cycling with examples from American politics, but without much success. There is also a growing body of work testing
Circumventing Arrow’s theorem 75 the consensus-building properties of deliberation. The results from both camps have failed to decide the issue, primarily because it is very difficult to find an institutional setting that resembles the procedural requirements of the competing theories. A detailed examination of these issues is beyond the scope of this book but in the next chapter I provide a study of an institution that was exceedingly democratic in its basic structure. I will utilize the Continental Congress (1774–1789) to try and better understand the likely outcomes from free and equal democratic institutions.
4
An empirical test of social choice and deliberative theories of democracy
Introduction As was noted in Chapter 3, social choice theory and theories of deliberative discourse have had a large impact on how political scientists understand the dynamics of democratic politics and decisionmaking. Such theorists ask us to imagine a procedure in which all have equal opportunities to raise and object to issues and amendments, where there is no powerful agenda-setting agent, where there is equal access to debate and where all votes count the same. What is likely to result from such a highly democratic setting? We saw that some of the theoretical findings that stem from the social choice research agenda in answer to this question have been disturbing to supporters of the democratic process. On top of the results already mentioned from Arrow, Riker, Condorcet and McKelvey, work by Plott (1967), Davis and Hinich (1968) and Schofield (1978) added further weight to the arguments that inconsistency would run rampant in IF settings. McKelvey (1976) and Plott and Levine (1978) were also able to demonstrate that control of the agenda was crucial for affecting the final outcome of choice mechanisms. Riker (1982) pushed the argument to its limits by claiming that institutions themselves are only ‘congealed preferences’ that are open to the same patterns of instability. One response to these assertions that might give sustenance to deliberative democrats is the claim, obviously true, that voting cycles seem rare to non-existent. Mackie has gone to great lengths to debunk Riker’s examples of cycles and Tullock (1981) asked the important question of why there is so much stability in politics? This led to the second string of social choice theorizing that tries to understand the dynamics of SIE. Shepsle and Weingast claim that ‘the rules employed by legislatures significantly restrict the potential outcomes
An empirical test 77 of the legislative process’ by placing ‘constraints on the completeness of the majority rule relation by restricting social comparisons’ (1981a, 515). This line of work has suggested that institutions have evolved in such a way as to negate the instability that would ensue in a pure majority-rule setting. Hence, the arguments that stem from Arrow’s findings are seen as theoretically interesting but not unduly worrisome because they do not affect the workings of highly intricate, contemporary legislative bodies. SIE theorists do not deny the potential for chaos in a purely democratic setting but they tell us that in real legislatures there are inevitably mechanisms that create stability. A social choice theorist might respond that this line of thought does not offer a helping hand to those with a deliberative bent because the rules and procedures that create stability favour some participants rather than others; stable outcomes are inherently biased in a way that does not treat all participants as free and equal. Even if we allow for stable preferences among participants, we can still arrive at different collective decisions depending on the aggregating procedure (e.g. first past the post, cardinal utility, plurality, or Nash’s calibrations). As already noted, Arrow simply provided a logical deduction that is sound given the parameters he sets. Demonstrating that Riker’s empirical work is flawed does not refute Arrow, because Riker’s examples are very weak, as Mackie has so amply demonstrated, and the institutions he examined do not fit with Arrow’s conditions. But the arguments of social choice and deliberative democracy are open to empirical scrutiny. We should expect more instability the closer we get to an IF setting if the social choice argument is correct, and we should get more consensus if Habermas and his supporters are right. The task is to find an institution satisfying the relevant criteria of no agenda setter, each person participating, weak Pareto efficiency, etc. There is a growing body of empirical research on how deliberation affects preferences and attitudes. Most of this research is small group experimental work that is not applicable to politics on a large scale, which is what I am concerned with in this book. The results of these experiments are mixed, with some showing greater cooperation after discourse (Sulkin and Simon 2001) and much of the social psychology literature sounding a pessimistic note (Mendelberg 2002; Rosenberg 2003). There is some evidence from small study groups suggesting a convergence to unanimity, but Weale questions this by noting the pressures to conform in such situations. The drive to consensus might be placing undue pressure on some people to conform to the wishes of the
78 An empirical test rest and, hence of conceding too much in terms of their own interests. This type of argument stretches back at least to John Stuart Mill who argued that consensus was stultifying because it bred mass conformity and Ian Shapiro (2003) argues in a similar vein that bi-partisan agreement is damaging to democratic vibrancy. He claims that America has very little competitive democracy because of the similarity between the two major parties, something that consensus theorists would presumably see as a good thing. Lucio Baccaro expanded the study of deliberation to factory workers in Italy and found that discourse seemed to produce a different outcome than if participants simply voted on an issue. However, as Steiner et al. point out, his meaning of deliberation is not the same as Habermas’s, and it is difficult to draw firm conclusions simply from the fact that people talk to one another. A similar complaint can be made against Fishkin’s (1997) results. I do not think anyone would deny that talking to one another is a good thing and can help people become more informed, but the Habermasian model demands more than this. Simone Chambers (1999) has conducted research on discourse during the referendum on Quebec in Canada. She found that during the conference stage of the process, discourse was of a high quality and there was wide-scale agreement such that ‘[a]dvocates even of quite extreme positions were willing to sign on to ‘consensual’ documents that reflected the general tenor of the weekend-long debates’ (quoted in Steiner et al. 2004, 48). But in the nitty-gritty of competitive politics ‘the referendum campaign appeared to harden positions. Furthermore, leaders and spokespersons began to talk in zero-sum terms, to the effect that any concession to other interests would be a loss for their side’ (quoted in Steiner et al. 2004, 48). Chambers concluded that the quality of discourse erodes as the decision-making time approaches. Jurg Steiner et al. (2004) have provided an interesting attempt to study the effects of deliberation in large-scale democracies. Their aim ‘is not to defend the Habermasian model but rather to establish in an empirical way the preconditions of a high-quality discourse and its consequences for policy outcomes’ (2004, 5). They note that real world democracies are a long way from the Habermasian discursive ideal but they do accept the argument that the ISS should count as a benchmark by which to judge actual democracies: ‘the ideal type of deliberation must be seen as the end point of a continuum that will never be fully reached’ (2004, 18). The assumption being made here is that the closer we get to the ideal, the better off we will be. Some discourse is good but more is better. The authors of the
An empirical test 79 study test a number of hypothesis such as, ‘[c]onsensus democracies are characterized by a higher level of parliamentary deliberation than competitive ones’ (2004, 82) and ‘[t]he quality of parliamentary deliberation is higher in second chambers than in first chambers’ (2004, 87). They found that favorable institutional antecedents include consensus institutions, the presence of veto players, debate in second chambers, and debate in non-public arenas. Debilitating factors include competitive settings, the absence of veto-players, debate in first chambers, and debate in public arenas. (2004, 135) I agree with the general conclusion arrived at by Steiner et al. ‘that institutions matter for political discourse Changes in the institutional parameters can have a profound effect’ (2004, 114). These results say nothing as yet about the quality of outcomes, but they continue by studying ‘the implications of discourse (for outcomes) in one institutional setting: the German Mediation Committee’ (2004, 138). They test the hypothesis that ‘a genuine consensus or a reasoned compromise, are more likely when there is a high level of parliamentary deliberation’ and conclude that ‘there appears to be a strong relationship between outcomes and discourse quality’ (2004, 150). These findings are very interesting and do suggest that discourse can help to resolve conflict and move people closer to one another on specific issues. The key problem with this research, and with all of the empirical work to date, is that it is measuring discourse and outcomes in institutions that undermine some of the conditions Arrow and Habermas place on decision-making and deliberation, particularly condition U. Modern legislative bodies are full of rules, regulations and norms that create an environment a long way from the ‘ideal’ of an IF setting. Agenda setters decide on what will be discussed, the order in which it will be debated and how long it will be talked about. Studies of these institutions can show that it is better to talk than not to talk, but few would question this. They do not show us that deliberation on the Habermasian end of the continuum is a good thing. If we really want to know what will happen in an IF environment, we need to find and IF institution. Steiner et al. recognize these problems and ‘do not claim that the findings with respect to the German Mediation Committee generalize to other contexts’ (2004, 139). The major empirical problem we face is that examples of pure IF decision-making institutions are extremely rare. If we can unearth
80 An empirical test such institutions, they might go a long way to answering our questions about the likely outcomes of free and equal democratic procedures. When we look at contemporary institutions we see that Shepsle is correct about SIE, and it is not surprising that most empirical work is aimed at understanding how institutional structures shape the collective decision-making process. Consequently, the arguments of IF theorists are usually treated as abstract hypotheses that are largely untestable. In social choice theory, an open agenda process is often regarded as a highly specialized and arbitrary concept with little relevance to actual political behaviour. According to Shepsle, ‘Institutions by their very structure induce an element of stability in policy outcomes that does not emerge in the more atomistic world of pure majority rule’ (1986, 76). The suggestion here is that institutions as we know them and pure majority rule (PMR), of the type discussed by Arrow and Habermas, cannot coexist. We need only accept this argument, however, if we also accept that institutions always resemble something approximating modern legislative bodies and if we assume, as Shepsle does, that rules always constrain actors. Fortunately, history provides us with a legislative body that is simple enough in its structure to allow us to examine it as a whole, without having to remove and ignore essential components of ‘real politics’ and without having to add some of the counter-intuitive assumptions from social choice, such as gate-keeping powers, ex-post vetoes, perfect information, dimension-by-dimension decision-making and calculations of the relative ‘sincerity’ of voting behaviour. In other words, I will examine a simplified institutional arrangement but it is a real legislature rather than a stylized creation. In this manner, we can engage in the crucial enterprise of testing theories against the backdrop of a firm empirical base and move a step beyond the usual approach of theorizing on a purely abstract level. The legislative body to which I refer is the American Continental Congress, 1774–1789. I will show in a later section that the Continental Congress was organized around norms and practices that fit well with the assumptions made by deliberative democrats and social choice advocates. The Continental Congress was an open, egalitarian and democratic legislature and fully reflected the preferences of its members. I intend to use the Congress as an example of what happens when institutional norms and procedures are weak, and where institutional power is dispersed. In essence, I am responding to the still largely unheeded call by Cooper and Brady (1981a,b) and Green and Shapiro (1994) to use history as a laboratory for investigating the role and place of contemporary concepts in political science. As they correctly note, ‘one cannot
An empirical test 81 but be impressed by the paucity of sustainable empirical analysis flowing from rational choice theory’ (1994, 244). In responding to the call, I utilize the early Congress as a simple, though real and historically important, legislative setting within which assertions of modern democratic theory concerning the dynamics of IF procedural politics can be explored. I will attempt to show that there was an earlier world in which some of the theoretical claims of Arrow, Riker and Habermas played out in political practice. From the description of the institutional structure of the Continental Congress and the results that flowed from it, I will draw certain implications regarding the dynamics that underlie the operations of open, democratic legislatures. The focus of my analysis, given the task of highlighting the IF nature of Congress, will be on the powerlessness of the President (as speaker, leader, or agenda setter), the organization, or lack thereof, of committees, and, in particular, the openness of the floor. I conclude the chapter with an assessment of the theoretical claims of IF theorists in the light of the empirical evidence. The evidence suggests that cycles did not occur but that instability was rife.
A broad description of the institution It would require a lot more space than is available to me here to outline the workings of the Congress. What I intend to do is to look briefly at some of the prime stability-inducing institutional mechanisms that have been identified by political scientists. My aim is to demonstrate that such features in the Congress did not interfere with its deeply democratic nature. In particular, I draw on the work of Jillson and Wilson (1994) to flesh out the structural dynamic of the institution, and I recommend this work to anyone interested in a more detailed discussion of the Congress. A theme that animated most of those who attended the Congress was a fear of abuses that could stem from an exuberant political body with too much power (Bowling, 1990). The result was that the delegates arrived in Philadelphia with a wary eye on republic building, which was reinforced by similar feelings of those left back at home in the colonies. To make sure that the predominance of the states was clear, Article 2 of the Confederation provided that each state would retain ‘its sovereignty, freedom, and independence, and every power, jurisdiction, and right, which is not by this confederation expressly delegated’ to the Congress (Journals, 9:908). Members of the Congress also wanted power to be widely diffused within the body itself and took considerable pains to make each
82 An empirical test member the equal of all others. This even extended to the powers allocated to the President of Congress who was elected to be a figurehead of the body rather than as the seat of institutional power. Most of the colonies had sent along their political elite and such men were not about to allow one amongst them to take a dominant role. Out of the 56 men who attended in the Spring of 1774, 12 of them had been the Speaker in their own colonial legislative body. Jillson and Wilson note that ‘delegates deliberately chose to constrain and weaken leadership positions within the Congress – especially the presidency it is important to recognize that the choice was quite deliberate and was confirmed and maintained throughout the life of the institution’ (1994, 71). Randolph was the first person to be elected President and his main function was to preside in a judicious manner over the floor debates; by the end of the Congress, the role of the President was still largely the same. He certainly had none of the agenda-setting powers that Shepsle identifies as mechanisms of equilibrium. Shortly after the opening of the first Congress, a delegate from Connecticut wrote that ‘our president, thought a gentleman of great worth, and one who fills and supports the dignity of his station to universal acceptance, yet cannot urge forward matters to an issue with that dispatch, which he might in a different assembly,’ and President Henry Laurens said of himself that he was ‘a silent auditor and spectator with no will of my own The reasonings in Congress and Conscience, were my oracles. I neither consulted, nor received advice out of doors, directly or indirectly, from any man’ (Burnett, Letters, 3: 221, 227, 272; 4: 147). If we cannot find the source of equilibrium-inducing power in the President it is worth taking a look at the committee system. Here we find that the creation of, and appointment to, committees was a very democratic matter. A committee was appointed whenever a majority of the states voted in favour of a motion to form a committee. If the vote was in favour, then nominations for members of the committee came from the floor and each member of Congress voted secretly for one of the nominees. The ones gaining the most votes were duly elected and the most popular was chosen as the Chair. The dominant form of dealing with business was through the formation of ad hoc committees. Although there were standing committees, these were not very successful; the same can also be said for the appointment of independent boards and the election of single executives who were not members of Congress (Jillson and Wilson 1994, 93). The history of the organizational strategies of members of Congress is too complicated to detail here, but it is safe to say that the bulk of
An empirical test 83 the activity from 1774 to 1789 was carried out by ad hoc committees; by the demise of the institution, 3249 committees had been formed. Shepsle has noted that one way of creating stability is to provide a committee with exclusive control over a policy issue. This never happened in the Continental Congress because jurisdictional boundaries never formed and most committees had a narrow agenda that was issued directly from the floor. The turnover of members also negated the formation of a strong power base for any particular member. The usual length of term was one year, which precluded incentives to specialize in particular policy issues. As President Jay wrote, while the maritime Affairs of the Continent continue under the direction of a Committee they will be exposed to all the Consequences of Want of System Attention and Knowledge. The Marine Committee consists of a Delegate from each State. It fluctuates, new members constantly coming in and old ones going out Very few of the members understand even the state of our naval Affairs or have time or inclination to attend to them The commercial committee is equally useless. (Burnett, Letters, 4: 176) It seems, therefore, that the committee system, like the powers of the President, was not a source of equilibrium; ‘The general environment in the Congress, then, was one in which a plethora of small committees confronted an extensive and varied workload. Most members served on many such committees’ (Jillson and Wilson 1994, 98). If we turn our attention to the rules for the floor, we find that democratic deliberation was encouraged. Each member had full and equal access, legislation could be easily blocked and there was no one to set a specific agenda to be debated: ‘from the outset delegates sought to guarantee open debate in the Committee of the Whole and on the floor to ensure that everyone had full opportunity to shape the resolution of important issues’ (Jillson and Wilson 1994, 134). As Trumbull noted, ‘unanimity being in our view of the last importance, every one must be heard, even on those points or subjects, which are in themselves not of the last importance’ (Burnett, Letters, 1: 70). Members wished to guard against other colonies (states) from dominating affairs and so made sure that agenda setting itself was off the agenda. They did this by limiting the power of the President, doing the same with committee control of issues and by prohibiting the bundling of issues over policy dimensions. This meant that each issue was dealt with as an autonomous whole. Rule 6 demanded that
84 An empirical test if two reports were called for at the same time, the President had to call for a vote from floor members to decide which one would be dealt with first. The reports from the committees were simply taken up in the order in which they were finished. Henry Laurens complained that he had been witness to a Report made by a Committee of the Whole, which had been entered upon the Journal, superseded by a new Resolution even without reference to the Report. A Resolution carried almost Nem Con – entered, and half an hour after reconsidered and expunged. When I add that such irregularity is the work of almost every day, you will not wonder that I wish to be any where but in Congress. (Burnett, Letters, 2: 482; see also 2:488) His disenchantment only increased over time: We are now busily engaged on the Report for an half pay establishment. Long and warm debates for many a day had led us to the threshold of the Report from the Committee of the Whole. We had Entered fairly the Door, by reading the whole for information, the first Clause for debate, and received an amendment which was read by the Chair and the question half put, when we were turned out by a New Motion – debates arose upon the point of order (Burnett, Letters, 3: 170) Thomas Burke from North Carolina noted the nature of the Congress when he stated, ‘the Imperfect Constitution of Congress which cannot reject any Business, addressed to them by way of the despatch through the President, before it has undergone some Consideration’ (Burnett, Letters, 4: 367–8). He continued by observing that the Nature of Congress is a deliberating Executive assembly, to whose proceedings the rules of order Established for deliberating Legislative assemblies will not always apply those circumstances make rules of order in that assembly very arbitrary and uncertain, hence frequent disputes arise thereon and the decisions at length depend upon the Integrity of the Majority. Thus Rules of order cease to be, what they ought, common checks upon excesses (Burnett, Letters, 4: 367–8) Rules governing floor procedures prevented complex pieces of legislation from being voted up or down as a package. Rule 11 from the
An empirical test 85 revision of 1776 stated that ‘If a question in debate contain more parts than one, any member may have the same divided into as many questions as parts’ (Journals, 5: 574). In the formal rules adopted in 1781, Rule 8 still allowed delegates to break down a report ‘if it comprehends different subjects, independent of one another, in the form of distinct acts or resolutions a question shall be taken on each’ (Journals, 28: 477). This denied committees the opportunity to bring forward reports that bundled together different issues in order to build a majority coalition on the floor. As we can see, the First American Congress, the Continental and Confederation Congresses of 1774–1789, was a fascinatingly accurate embodiment of what modern social choice theorists have described as an IF legislative setting. The procedures allowed for open discussion on the floor, no one had agenda-setting powers, all could introduce new amendments at any time, the President did not have power concentrated in his hands, and neither was it found in a few dominant committees. As Jillson and Wilson state, the members chose a president who was expected to preside over debate and maintain order, but not to control or manipulate outcomes; created a committee system that revolved around the rights of the individual colonies and members on the floor; and established a system of floor debate that guaranteed equal rights to all members. (1994, 8) This also fits remarkably well with the sorts of institutional arrangements favoured by deliberative democrats. How then did the First American Congress work and how does our empirical knowledge of its structures, rules and patterns of activity comport with expectations of deliberative democrats and social choice theories? As an example of the sorts of decision-making created by the rules and procedures of the Congress, I present two case studies. Both were chosen because they represent how Congress dealt with difficult issues over a period of time, and because they were two of the most important issues debated during the history of Congress. The first is the Northwest Ordinance, which, along with being important and much debated in Congress, has also gone down as one of its major successes. As Jillson and Wilson note, ‘without exception, historians agree that the Northwest Ordinance was the one shining piece of post-war congressional policy making’ (1994, 276). The second case concerns the very location of Congress, an issue that was of crucial significance to the members themselves.
86 An empirical test
Case studies: Administration and governance of the west, and the location of the capital Case study 1: The Northwest Ordinance of 1784 As soon as Congress agreed to Virginia’s cession of Western lands, the delegates’ attention was turned towards the twin questions of how to divide the land for settlement and how those territories should be governed. Shifting factional alignments across these issues were the norm. Jefferson brought to the floor a committee report on western governance that contained five general provisions. First, it delineated the boundaries of the territory, including both land ceded or to be ceded. Second, it offered provisions for the creation of temporary and permanent governments. Third, it offered a set of five principles upon which those governments would be based, including the exclusion of hereditary titles and the prohibition of slavery. The fourth element detailed the process for the admission of new states, while the final provision defined the boundaries of ten new states. Congress sought to rein in Jefferson’s imagination by recommitting the report to his committee (Journals, 26: 118–20). On 19 April 1784, Jefferson’s suggestion ‘That after the year 1800 of the Christian era, there shall be neither slavery nor unvoluntary servitude in any of the said States ’ (Journals, 26: 247) was to have a short life. Richard Dobbs Spaight (NC) moved to strike out this section. On the question, ‘shall the words moved to be struck out stand?’ sixteen delegates voted yes while only seven voted no (see vote #1128). Nonetheless, the words were struck out because an affirmative vote of seven states was not achieved. The vote almost carried as every delegate representing the seven northern states voted yes. As Jefferson noted, the issue ‘was lost by an individual vote only Jersey would have been for it, but there were but two members, one of whom was sick in his chambers’ (Burnett, Letters, 7: 500). After pushing past the issue of slavery, the delegates focused on a series of questions dealing with how the territories would be monitored as they moved towards full statehood. The relevant paragraph read, ‘That they shall be subject to the government of the United States in Congress assembled, and to the Articles of Confederation in all those cases in which the original States shall be so subject’ (Journals, 26: 249). Sherman (CT), wishing to loosen the reins a bit, moved to strike out ‘to the government of the United States in Congress assembled, and’ so that the territories would be subject only ‘to the Articles of Confederation’ rather than directly to the Congress.
An empirical test 87 On the question, ‘shall the words stand?’ only Maryland voted yes, New York and the Carolinas divided, while all of the New England states, as well as Pennsylvania and Virginia voted no (see vote #1129). Vote #1133 made the further amendment that new states were to be admitted ‘provided that the consent of so many states in Congress is first obtained as may at the time be competent.’ Read’s motion that the original settlers ‘shall be ruled by magistrates to be appointed by the United States in Congress assembled’ (Journals, 26: 259) was rejected with a clear regional division on the vote. It was clear that most of the delegates wished to assure that the states that would rise in the west would be subject to, and eventually members of, the Confederation but they were uncomfortable with direct congressional governance. As soon as the Ordinance of 1784 was passed, delegates turned to consider means for surveying and distributing the vast western territory. Jefferson (VA), this time chairing a five-person committee which included Williamson (NC), Read (SC), Gerry (MA) and Howell (RI), reported to the floor on 30 April 1784. The model proposed for surveying and distributing land was based on the usual southern convention of the rapid distribution of large and often scattered tracts. This original plan called for surveying the land into ten square mile tracts called ‘hundreds’, to be sold complete or in one square mile units. Jefferson’s committee report was rejected almost unanimously on 28 May 1784 (Burnett 1941, 598–604). Settlement and distribution of western lands – 1785 Congress did not return to the issue of how to survey and sell the western lands until 1785. On 3, 4 and 5 May 1785, a number of votes were taken over a series of proposed compromises between the eastern and southern modes of land disposal. The real fight in the Congress was not over the size of the townships to be laid out in the west, but over whether they would be sold in parts or as units. The four options placed before members were the following: 1
2
Option A: Howell’s (RI) language as embedded in the revised committee report provided the status quo against which alternatives were posed (Journals, 28: 300). Option B: Grayson’s (VA) motion to strike Howell’s (option A) language and insert ‘he shall offer for sale every third Township by sections of 640 acres provided that all the fractional parts of Townships shall be sold only by sections’ (Journals, 28: 328).
88 An empirical test 3
4
Option C: Wilson’s (PA) proposal to amend Grayson’s motion (option B) to require every ‘second’ rather than every ‘third’ township and all ‘fractional parts of Townships’ to be sold by sections (Journals, 28: 328). Option C1 : McHenry’s (MD) motion, a refinement of Wilson’s (option C), provided that every ‘second’ township and ‘half’ of the ‘fractional parts of Townships’ be sold by sections (Journals, 28: 335).
Grayson (VA) opened the debate on 3 May by moving to strike out Howell’s (RI) language in the amended committee report that ‘he may sell any Township by sections, provided that he does not offer a second Township for sale by sections, till the whole of the former is sold’ and to replace it with language requiring that ‘he shall offer for sale every third Township by sections of 640 acres ’ (Journals, 27: 328). If the language in the amended committee report held, Grayson knew that complete townships could be sold apace while sales of single sections in a particular township might languish, thereby blocking additional sales by section in other townships. Wilson (PA), not satisfied with the proposal that every ‘third’ township be available for sale by sections, moved to amend Grayson’s proposed amendment by striking ‘third’ and inserting ‘second’. Wilson’s motion to amend failed narrowly, five states to three, though every New England delegate voted no and every delegate from New York to the south voted yes. At the beginning of the debate, therefore, it was clear that the New England delegates wanted to limit the amount of land sold by sections, while the delegates from New York South wished to make half or more of the land available for sale in this manner. Ellery (RI) then called for a vote on the first part of Grayson’s motion – that every third township be made available by sections. Unlike the preceding vote, which was almost strictly regional, this time delegates from New England and the South (with Grayson voting against his own motion) joined to defeat option B. The North Carolina delegation was split and single delegates from South Carolina and Georgia sided with Virginia. The question was lost, as most southern delegates, sensing victory, held out for an alternative nearer to C while the New England delegates voted no because they opposed dispersed sales altogether. Hence, we find that in the first round of voting, A defeated B and B defeated C. The following day, 4 May, James McHenry (MD) presented option C1 to strike the provision of the Ordinance reading, ‘or he may sell any Township by sections, providing he sells them in the order
An empirical test 89 of their number ’ and replace it with language saying that ‘the alternate Townships in each range shall be sold entire The other Townships shall be sold in sections of one mile square’ (Journals, 28: 335). The southern states were clearly intent on pushing for option C and were prepared to fall back incrementally to C1 . King (MA), recognizing that A was no longer tenable and that compromise somewhere between B and C was becoming a strong possibility, then sought to limit the impact of McHenry’s motion by reintroducing Grayson’s motion (option B), replacing ‘one-half’ with ‘one-third’ so that only every third township could be sold by sections. King’s motion for ‘one-third’ lost when only two states, New Hampshire and Massachusetts, gave support, New York divided, and Connecticut, Delaware and Georgia could not vote because only a single member was in attendance. At this stage in the voting procedure, therefore, the northern delegates, having abandoned Howell’s language (option A), fell back to ground originally occupied by Grayson (option B), only to find the Virginian and his southern colleagues had abandoned that ground to join Wilson at McHenry’s compromise position (option C1 ). The question on McHenry’s motion was postponed until the next day at the request of New Hampshire. When McHenry’s option C1 came up the next morning, it passed easily, eight states in favour and Rhode Island divided. This sealed the compromise between eastern and southern modes of land distribution and settlement. What we find in this little sequence of voting is that A beat B and B beat C and that C1 , a close relation to C, beat A. This is not quite a voting cycle because of the slightly different wording of C and C1 , but it is probably the closest empirical example yet of a cycle and it does show the potential for instability in outcomes. What also comes through from this example is that the strict demands of majority rule can often prevent the most popular position from winning. In fact, when Congress changed its voting procedure to demand a supra-majority of 9 states, we find that the status quo is massively favoured on all issues, especially when we factor in the problem of member absenteeism. This is the only rule in the Continental Congress that makes it differ markedly from the requirements set down by Arrow, and it makes the likelihood of cycling more remote precisely because it acts as a check against any change to the existing policy position. Hence, the instability noted above and in the case study to follow is all the more remarkable because a super-majority was required to move from each previous decision. Though desultory skirmishes continued for two more weeks, few important changes were made and the land ordinance was eventually
90 An empirical test passed by a nearly unanimous Congress on 20 May 1785. Prior to the discussion concerning the distribution of lands, Congress once more addressed the question of slavery. On 16 March 1785, Rufus King presented a resolution ‘That there shall be neither Slavery nor involuntary servitude in any of the States, described in the resolve of Congress of the 23 April, 1784 And that this regulation shall be an article of compact between the 13 Original States, and each of the States described in the said resolve of the 23 April, 1784’ (Journals, 28: 165). The vote was split among the same states exactly as with Jefferson’s proposal the previous year, but this time New Jersey had a full compliment and the proposal was resolved in the affirmative. The final outcome, therefore, was as follows: regarding the government of the lands it was originally agreed that there be no anti-slavery clause. This decision was then placed at risk by a proposal to amendment at a later date. The new states were to be subject to the Articles of Confederation and only loosely controlled by Congress. There were to be ten new states, the population of which had to be at least twenty thousand. They were to be parcelled out into blocks of six square miles, and sold alternately as entire blocks, and in sections of one square mile. The arbitrariness of the process to get to this outcome lends support to the theoretical arguments of the theorists of disequilibrium discussed earlier in the paper. However, the key question to be addressed, given that a final outcome did materialize, is, ‘how stable was this decision?’ a decision that Henderson describes as ‘the most enduring legislation yet passed by the Confederation Congress’ (1974, 376). The Northwest Ordinance – 1787 The issue of western lands was addressed again in 1787. Prior to this, however, in the autumn of 1786, Monroe paid a visit to the western territories. He concluded that small and sparsely populated states were unlikely to have interests compatible with the objectives of the federal government, even assuming they would have sufficient population to entitle membership to the confederacy. On his return to Congress, a grand committee was established to examine the merits of the 1784 Ordinance concerning these issues. The report of the committee, discussed in Congress on 7 July, resolved to ask Virginia to alter her cession agreement from the formation of ten new states ‘into distinct republican states, not more than five nor less than three, as the situation of that country and future circumstances may require’ (Journals, 31: 392). The vote on the resolution to propose alteration
An empirical test 91 was almost unanimous with 26 of the 27 members present voting in its favour, clearly demonstrating, as did the proposal to amend the slavery clause, that the 1784 Ordinance would not be afforded much safety from debate and amendment. It was not until 1787, however, that Congress was moved to fully act once more on the question of the western lands. Despite the supposed ‘stickiness’ of the 1784 ordinance, events in 1787 were to take ‘an unusual, even bizarre, turn’ (Henderson 1974, 410). Congress, slow in its implementation of decisive action regarding the governance and distribution of the western lands, was moved to action by a memorandum from the Ohio Company, which offered to substantially reduce the public debt in exchange for a million acre block of land immediately west of the lands then currently under survey. To judge how this Ordinance differed from its predecessor I break it down, as with the 1784 version, into the two areas of distribution and governance. With respect to the first of these questions, there is a great change from the Ordinance passed three years previously; the 1787 decision to sell a million acre bloc of land was a radical departure from all previous proposals. The second difference is the acceptance of the 1786 proposal to limit the number of new states to between three and five, again by a unanimous decision among the states and with the same wording as the 1786 proposal. The third difference is an increase in the population requirement for joining the Union from twenty thousand to sixty thousand. This question did not seem to be divisive in 1784 or 1787, and so the amendment to the original Ordinance seems to be of a particularly ad hoc nature. On the question of governance, the spirit and letter of the 1784 Ordinance was again substantially altered. The southern states favoured a more piecemeal progression towards state autonomy. This included, in particular, the establishment of a governorship and a court which ‘shall adopt and publish in the district, such laws as the original States, criminal and civil, as may be necessary, and best suited to the circumstances of the district’ (Journals, 28: 315). The governor was to be the ‘commander in chief of the militia, appoint and commission all officers in the same, below the rank of General Officers and appoint such magistrates and other civil officers in each County or township, as he shall find necessary for the preservation of the peace’ (Journals, 28: 315). Only when the free male adult population reached 5000 were the states allowed to begin to govern their own affairs. All of these proposals had been discussed the previous year, again highlighting the tenuous condition afforded all congressional legislation.
92 An empirical test Another immensely important feature of the new plan and the most telling regarding institutional instability was the prohibition of slavery from the territories. As discussed above, this had originally been part of the intent of Jefferson’s proposal in 1784 but had been narrowly rejected by ballot. The 1787 Ordinance, with the exception of the clause ‘after the year 1800 of the Christian era’ and the introduction of a proviso concerning fugitive slaves, returned to Jefferson’s original proposal. Article six of the Ordinance, which was added only on the last reading, read ‘There shall be neither Slavery nor involuntary Servitude in the said territory otherwise than in the punishment of crimes’ (Journals, 28: 343). The policy decisions of Congress concerning the slavery issue are extremely enlightening concerning the unstable nature of a PMR body. In the space of three years we see widely fluctuating positions on this most crucial issue. There is, of course, much disagreement in the literature concerning the differences and similarities between the two ordinances. It is not my intention to enter this debate; the task has been to point to the areas of change and as a consequence highlight the fact that in a PMR decision-making body no decision is safe from amendment. This condition of potential instability was the same for every decision made in the Continental Congress, even, to repeat the claim of Henderson, for the Ordinance of 1784, which represented ‘the most enduring legislation yet passed by the Confederation Congress’ (1974, 376). Case study 2: The location of the capital In June of 1783 Congress was forced to meet in emergency session to respond to the threat of mutinous troops in Pennsylvania. The emergency became even more acute when Congress itself was threatened by the close proximity of the troops, and members agreed, in an uncharacteristic display of unity, to adjourn to Princeton. Once ensconced in the safety of their new environment, they turned to what was to become one of the most divisive issues in the history of Congress, namely the debate over the permanent location of the capital. I will look in some detail at this debate and use it to demonstrate how an issue of obvious importance to the delegates could nevertheless become bogged down by debate, amendments, and what seemed like an almost endless series of votes. Members deemed ‘that the question be taken, in which state buildings shall be provided and erected for the residence of Congress; beginning with New Hampshire, and proceeding in the order in which they stand’ (Journals, 25: 649). Not surprisingly each delegate voted for locating Congress in his own state. When delegates
An empirical test 93 started casting further afield than their own state, New Englanders were not willing to support New York and focused upon Trenton as an alternative. Southern delegates favoured locating the capital on the Potomac close to Georgetown once they had resigned themselves to the loss of Congress in their own state. The middle states preferred Philadelphia over the other alternatives, but this site was unsupported by New York, New England and the South. After the first round of voting, therefore, we can see clearly the troubles that lay ahead. When Congress returned to the issue the next day, Gerry (MA) moved ‘that buildings for the use of Congress be erected on the banks of the Delaware, near Trenton, or of Potomac, near George-Town, provided a suitable district can be procured on one of the rivers aforesaid for a federal town’ (Journals, 25: 654). These had been the two most favourable positions the previous day and the hope was that here at least the options could be reduced to two and that eventually one or the other would be selected. Bedford (DE) had different ideas, however, and as was his prerogative in an institution which allowed equal access to the floor, moved to strike out the words ‘near Trenton’ and ‘near George-Town.’ On the vote as to whether these words should remain only five states voted yes with three voting no and three being divided. Hence, because of a lack of a majority, the words were struck out with the bizarre result that the two most favourable alternatives had both been lost. The next few votes for alternatives also failed to gain the required majority, but finally on vote #1039 a majority was reached and it was resolved ‘that buildings for the use of Congress be erected on or near the banks of the Delaware, provided a suitable district can be procured on or near the banks of the said river’ (Journals, 25: 657). On 10 October, delegates addressed the issue of making a temporary move until the permanent site was finalized. Bedford moved ‘that for the more convenient transaction of the business of the United States, and the accommodation of Congress, it is expedient for them to adjourn from their present residence’ (Journals, 25: 666). This motion passed, again on a strict regional division with all the New England states voting against the motion. Williamson (NC) moved in favour of adjourning to Philadelphia until the following June, and then switching to Trenton. This vote failed to achieve a majority, as did Duane’s attempt to strike out Philadelphia and simply adjourn to Trenton. Howell’s (RI) motion to adjourn to Annapolis, and Mercer’s (VA) motion for Williamsburg met similar fates. For the time being, it seemed impossible to find a solution which was satisfactory to the majority. The success in finding a single, permanent home was also
94 An empirical test to be short-lived. Although members of Congress had voted in favour of housing Congress near the falls of the Delaware, the mood among delegates was such that this result was unlikely to last very long. In the Continental Congress, every majority decision was always vulnerable to future amendment or rejection. This was to prove to be the case on this occasion, and on 17 October Gerry (MA) put forward a motion which read that ‘the resolutions of Congress of the 7th instant, to erect buildings for their use at or near the falls of the Delaware, are not satisfactory to a respectable part of the United States And whereas Congress have no prospect of a general assent to any one place for their residence there is every reason to expect that providing buildings for the alternate residence of Congress in two places, will be productive of the most salutary effects’ (Journals, 25: 697). It was moved, therefore, to override the vote of the 7th in order to adopt two sites to house Congress, one on the Delaware near Trenton and the other on the Potomac near Georgetown. The two sites which had previously been lost individually and then been struck from the motion designed to pose a choice between the two were now back on the table again, precisely because members of Congress realized they were not going to reach agreement on a single site. On 20 and 21 October, through a series of votes Congress acceded to this unlikely compromise and also finally agreed to temporarily divide its time between Trenton and Annapolis until the permanent sites had been built. Hence, it was resolved that as well as building a site at Trenton, ‘buildings be likewise erected for the use of Congress, at or near the lower falls of Potomac or George-town,’ with the hope that this would have ‘the most salutary effects, by securing the mutual confidence and affections of the states’ (Journals, 25: 707). Given the IF nature of Congress, it had proven impossible to build a winning majority coalition around one site. Members of Congress were unhappy with the solution for the temporary location and continued to address the issue. In April 1784, delegates of Rhode Island tried unsuccessfully to move Congress to Newport. Montgomery (PA) then suggested Philadelphia, which only gained the vote of Pennsylvania. Jefferson (VA) suggested Alexandria and this also only gained a single vote in its favour. Congress eventually voted once more for Trenton and Annapolis, but this series of votes again demonstrates how any decision could be repeatedly challenged. In fact, the delegates for Rhode Island tried for Newport and failed once again on the 26th of April. Hardy (VA) argued that the attempts by delegates of Rhode Island suggested that ‘it is evidently
An empirical test 95 the sense of that state, that the resolution for two federal towns should not be carried into effect’ (Journals, 26: 293–4). The compromise agreement over the permanent location was also showing signs of strain. Only one state voted yes on McHenry’s motion that ‘as soon as a proper place shall be reported by the committee on Potomac as on Delaware, Congress will, forthwith, proceed to erect such public buildings for their sessions’ (Journals, 26: 292). Hardy continued to express his discontent by stating that this decision ‘forbidding the committee appointed to view the grounds until the further order of Congress amounts to a virtual repeal of the aforesaid act’ (Journals, 26: 294). The temporary residency decision came under threat once again in December 1784 when a motion was made to repeal ‘the alternate temporary residence of Congress at Trenton and Annapolis’ (Journals, 27: 697). This narrowly lost, but may have won if the full complement of states had been present, requiring only another two votes for victory. The survival of the joint temporary residences was short-lived, however, for on 23 December a motion to continue meeting at Trenton gained only two votes. Hardy (VA) moved for Philadelphia as an alternative and was defeated, Ellery (RI) proposed Newport and lost, and finally Howell (RI) suggested New York, which gained the required majority. Hence, the joint temporary location compromise finally gave way under the weight of the constant attempts at amendment and New York was to remain the location of Congress until its demise in 1789. At this point, the permanent residency problem was again addressed, and it was resolved that it would be ‘inexpedient for Congress at this time to erect public buildings for their accommodation at more than one place’ (Journals, 27: 696). An ordinance was drawn up to appoint three commissioners to examine land on the banks of the Delaware, and it was agreed that a sum of $100,000 be set aside for the federal buildings. Not surprisingly, those who wished for the permanent residence to be in Georgetown moved ‘to strike out the words on the banks of either side of the Delaware and in lieu therof to insert, at Georgetown on the Potomac.’ This, however, failed to gain majority support and the original words stood. It was also passed that Congress would continue to meet in New York until the buildings were completed. The new ordinance, therefore, bore no resemblance to the original compromise decisions concerning the permanent and temporary locations for Congress. It now read that the permanent site of Congress would be on the banks of the Delaware and until that time Congress would meet permanently in New York.
96 An empirical test The final series of votes on the issue of the permanent site of Congress began on 28 July 1788. The ratification process of the Constitution being under way, the relevant paragraph for putting Congress into effect dealt in part with where the Congress should be located. It read ‘that the first Wednesday in March next be the time and [blank] the place for commencing proceedings under the said constitution’ (Journals, 34: 359). A motion was made to ‘fill the blank with ‘Philadelphia’ (Journals, 34: 359), which lost on a 6-6 vote. Two days later, it was suggested that the blank be filled with the words ‘and at such a place as shall hereafter be appointed by Congress’ (Journals, 34: 367). This also failed to gain a majority. On the 4th of August, Lancaster lost but Baltimore gained a 7-6 victory, and this decision was upheld in a later vote that same day. For the time being, therefore, the paragraph read ‘that the first Wednesday in March next be the time and the town of Baltimore in the state of Maryland the place for commencing proceedings under the said constitution’ (Journals, 34: 395). This decision immediately came under attack when Tucker (SC) moved that Baltimore be struck out and New York put in its place. This motion itself was challenged when Williamson (NC) proposed an amendment to the amendment by removing the words New York and putting in their place ‘the seat of the Congress ought to be in some place to the southward of New York’. This amendment narrowly failed 6-7, but immediately Carrington (VA) moved to strike the words ‘New York’ and in their place insert ‘Philadelphia’, a motion which also lost. Finally the original motion for New York was voted on and was successful by a 7-6 vote, thus ousting Baltimore. The dust settled for three weeks, but no decision seemed final on this issue and on 26 August Kearny (DE) moved to strike out ‘New York’ and insert ‘Wilmington’. This motion lost as, on 2 September, did Clark’s motion for the words ‘at the seat of the federal Government’ and Irvine’s (PA) motion for Lancaster. Exasperated, members finally produced a motion two days later that read ‘The organization of the said system of government has met with undue procrastination by which high inconveniences must accrue to the union at large the same unhappy Cause of Delay still exists and there is but little apparent likelihood that such accommodation will result as to gain the assent of the United States in Congress Assembled to any Place for the meeting of the said Government. Therefore resolved that the first Wednesday in March next be the time for commencing Proceedings under the said constitution at Such Place as Congress shall hereafter appoint’ (Journals, 34: 517). It was agreed, therefore, on 13 September, by a 9-0
An empirical test 97 vote to remain at the present site until the question could be addressed by the new government set up under the constitution. Despite five years and 60 votes on the issue, the members of the Continental Congress had been unable to resolve one of the most pressing issues of the day, and had finally to abandon the problem altogether as irresolvable and leave it for the new congress, which, as we know, is not an IF setting.
Conclusion In this chapter, I have made a first stab at an empirical test of two of the most influential theories of democratic institutions to emerge over the last 50 years. Although the workings of the Congress have to be sifted through in greater detail than is possible here, I would suggest that the theoretical importance of the case studies is significant. The two examples from the Continental Congress provide good evidence that pure majority rule and stability can easily part ways. There were many issues, of course, that were rapidly resolved and remained stable over time, a conclusion that would neither surprise nor worry Arrow, who only ever claimed that instability is a possibility in PMR settings. But there were other issues, such as half-pay, peace settlements, currency reform and national debt, to mention a few, that exhibited the same sorts of instability we find with the case studies above. The study of the Continental Congress demonstrates that there can be a lot of arbitrary decisions in the absence of cycles and an open democratic forum by no means guarantees that preferences will converge to points on a single dimension. I am not suggesting that the Congress bore a full resemblance to the stylized creations of social choice or Habermas’s ISS, but I will claim that it is as close as we are going to find in real politics and hence is important for assessing the validity of claims for PMR. Nor am I suggesting that the members of Congress debated in a manner that matches all of Habermas’s requirements for discursive rationality. I do think, however, that they were as likely to approach these ideals as the members of the German Mediation Committee studied by Steiner et al. A balanced conclusion is probably that democracy will always be a somewhat disorderly business and the more democratic the process, the more likely it is that we will have to deal with the problems that come along with free and equal participation. It seems probable that we can have order and hierarchy and equality and instability, but it seems less likely that we can have order and equality and we should probably not get our hopes up too high regarding PMR.
98 An empirical test I conclude this chapter as I began by heralding the importance of measuring contemporary concepts against the contextual yardstick of history, in which attention is paid to both behavioural and institutional features of real legislatures. By examining only theoretical abstractions or highly specialized experiments, we are blinded to the evolution of political institutions. Understanding the foundations of those institutions and how the choices of members contribute to their development remains a primary task for theorists of democracy. I noted at the beginning of the chapter that Steiner et al. have demonstrated that constrained discourse within institutions can lead to more consensual outcomes (although they did not find that they converged on decisions that are more fair and just), and my study has shown that unconstrained discourse can lead to instability. One of the arguments they make is that ‘the ideal type of deliberation must be seen as the end point of a continuum that will never be fully reached’ (2004, 18). My findings suggest that a completely open discourse as idealized by Habermas is a hindrance and needs to be constrained. I tentatively conclude that some deliberation is good, but that one can have too much of a good thing. Given the structural similarities between the theories of social choice and deliberation, I claim that the arguments of Arrow are more persuasive than those of Habermas in relation to decision-making bodies that have ‘institution free’ characteristics. To repeat my claim from an earlier chapter, this argument is very specific. I am utilizing social choice theory to show that if proponents of democracy insist on certain procedural requirements for participation, as I believe Habermas and many of his supporters do, then the arguments of Arrow demonstrate that these requirements lead to an unfeasible brand of democracy. But given the findings of Steiner et al., it also seems likely that deliberation can become a positive tool once we have institutional mechanisms in place that limit participants. A fruitful area of research, therefore, is to examine the type of institutional structures that take freedom and equality seriously, but not so seriously that the institutions themselves become unworkable. A hard-line social choice theorist might complain that this is illegitimate and continue to maintain that the only options open to us are IF settings that are plagued by instability or a dictator who independently decides results. I have not yet addressed Arrow’s non-dictatorship condition, and in Chapter 5 I argue that dictatorship in a loose sense of the term (i.e. that there has to be a body that makes the final decision) is not detrimental to the case for democracy. To make the argument, I will turn to the unusual source of Thomas Hobbes for inspiration.
5
Hobbesian sovereignty and the spectre of tyranny
Introduction The conclusion of the previous chapters supports Arrow’s claim that one can have free institutions and the potential for instability and disorder, or one can have stability if limits are placed on freedom and equality; stable decisions are only guaranteed when the buck has a very definite place to stop. This conclusion is unpalatable for some democrats, Arrow included, because they wish for a power-free democratic arena in which all preference structures can be expressed. I argued that limiting the U condition, which means limiting freedom and equality, is the normatively correct thing to do. Many other democrats, especially those emanating from America, will not be troubled by this argument because they claim that constitutional limits on voters and decision-makers are necessary to safeguard rights. Most democrats of a liberal persuasion wish to limit the power of the majority through mechanisms such as the separation of powers, checks and balances and bills of rights and are hence not philosophically disposed towards the U condition. However, placing limits on participation through anti-majoritarian measures can also have the effect of creating instability because power is dispersed among competing branches of government. If instability occurs when there are differences in preferences among individuals, it will also occur when there are differences in preferences across political institutions. One way of securing more coherent outcomes, therefore, is to abandon the idea that power should be dispersed among several political bodies and embrace instead the Hobbesian idea that power should be absolute and centralized. This is the major policy recommendation of the book. It will not prevent the volatility that comes from voting procedures, but it will at least get rid of the instability that occurs because of the dispersal of power. If social
100 Hobbesian sovereignty and the spectre of tyranny choice theorists really do take Arrow’s argument seriously, they should give up their liberal sensibilities and embrace a brand of democracy that circumvents at least some of the problems identified by Arrow. Arrow argued that the non-dictatorship condition cannot be violated, and I agree with him in the sense that if all society prefers ‘a’ and one person prefers ‘b’, then the social welfare function should choose ‘a’. But if we think of dictatorship in the sense of the citizen body having the final and absolute say on political decisions, then it can be made normatively appealing. In Chapters 5 and 6 I argue that democracy should take precedence over liberal limits on majority rule. Because Hobbes’s brand of absolutism seems such an unlikely source of inspiration for democratic theory, the bulk of this chapter is an attempt to rehabilitate his political philosophy and demonstrate that it is compatible with freedom and democracy. In the Chapter 6 I will address the main concern about absolute power, namely that it is too dangerous, and I will argue that a democratic form of absolutism is not to be feared.
Hobbesian sovereignty Perhaps the most enduring criticism of Hobbes’s political philosophy is that it provides for an absolute sovereign that poses a great threat to individual freedom. It is a criticism that remains a dominant theme in contemporary political thought; Judith Shklar suggests that the ‘convoluted genealogy of liberalism that insists on seeing its origins in a theory of absolutism’ (1989, 24) is incorrect. Susan Moller Okin claims that ‘Hobbes was no liberal in his conclusions, advocating an absolute rather than restrained state’ (1989, 257), and Benjamin Barber argues that ‘Liberals rightly pall at the idea of Hobbes as a liberal predecessor because his fear of anarchy leads him to embrace an authoritarian conception of the state incompatible with limited government’ (1989, 261). Work by William Connolly (1988) David Gauthier (1995) Michael Ridge (1998) and Victoria Kahn (2001) are in the same vein. Richard Boyd argues that Hobbes ‘has few misgivings about the state’s unlimited and undivided power to coerce individuals’ (2001, 402). Charles Tarlton (2001, 2002) has published recent articles in which the attack on Hobbes is driven home with great force. For example, he tells us that Hobbes is derisive, dogmatic, coarse and disparaging. He also uses language such as ‘the sly, petty despot’s sneer’ (2002, 83) to describe Hobbes’s arguments. Tarlton accuses Hobbes of promoting tyranny, despotism, arbitrary rule, indoctrination and the total subordination of the individual: ‘Peace was secured
Hobbesian sovereignty and the spectre of tyranny 101 under Hobbes’s system by means of the complete subordination of the wills of the subjects to the will of the sovereign’ (2002, 82). Are such criticisms valid? As the old saying goes, there is no smoke without fire, and it would be idle to deny that on first reading one can see why many commentators are worried about the nature and the extent of sovereign power in Hobbes’s work. I argue, however, that the above commentators are incorrect to imply that Hobbes’s sovereign is a great threat to liberty. I do not deny that Hobbes provides the sovereign with a considerable range of powers. The majority of these, however, are not troublesome; few would quarrel with his claim that the sovereign should have the capacity to raise taxes, make laws, form a military, and coin money. A warning bell does sound for some commentators, however, when Hobbes discusses other powers concerning issues such as censorship, civil education, free speech, free association and property rights. In these areas, Hobbes is accused of introducing the spectre of tyranny. I will examine these issues later in the chapter and demonstrate that Hobbes’s political philosophy has been seriously misunderstood and that it still has great purchase for contemporary democratic theory. A close reading of Hobbes’s views on sovereignty suggests that his intentions were to provide a logical definition of the office rather than to support a political body that aims to oppress individuals. Once peace is secured, Hobbes allows for a wide expanse of freedom in associational life. The usual interpretation suggests that he proposes a form of absolute sovereignty in order to tame wild, savage and instrumentally self-interested individuals who will, to use Madison’s language, vex and oppress one another if given the chance. Hence, the supposed tyranny we find in his writings is fully intended and justified because of the base material with which he has to work. The first line of defence against this criticism is to note that Hobbes’s views on sovereignty are based on his civil science rather than any desire to terrorize people. Hobbes was concerned with definitions, and his statements on absolute sovereignty are about logical consistency. Tarlton suggests that those who think Hobbes is not presenting a tyrannical form of politics perform a sleight of hand in which Hobbes’s ‘unrepentant willingness to incarcerate the natural individual became only a dreary matter of logical necessity’ (2001, 597). This argument faces a severe textual problem because Hobbes repeatedly states that his goal is to find the logically necessary truths about politics. The point he wishes to make is that absolutism occurs wherever political command is properly exercised. When Hobbes defines sovereignty he separates it into democracy, aristocracy and monarchy. Despite the
102 Hobbesian sovereignty and the spectre of tyranny fact that the sovereign can take on different forms, he says that ‘the difference between these three kinds of commonwealth consisteth not in the difference of power, but in the difference of convenience or aptitude to produce the peace and security of the people’ (1968, 241). It is a necessary part of politics that there has to be an end point to political deliberations: It is therefore manifest, that in every city there is some one man, or council, or court, who by right hath as great a power over each single citizen, as each man hath over himself considered out of that civil state; that is, supreme and absolute, to be limited only by the strength and forces of the city itself, and by nothing else in the world. For if his power were limited, that limitation must necessarily proceed from some greater power. For he that prescribes limits, must have a greater power than he who is confined by them. Now that confining power is either without limit, or is again restrained by some other greater than itself; and so we shall at length arrive to a power, which hath no other limit but that which is the terminus ultimus of the forces of all the citizens together. (1972, 186–7) Hobbes suggests that knowledge is arrived at by making necessary consequences, having first taken the beginning from experience. But men’s reasonings are sometimes right, sometimes wrong; and consequently, that which is concluded and held for a truth, is sometimes truth, sometimes error. Now errors, even about these philosophical points, do sometimes public hurt, and give occasions of great seditions and injuries. It is needful therefore, as oft as any controversy ariseth in these matters contrary to public good and common peace, that there be somebody to judge of the reasoning, that is to say, whether that which is inferred, be rightly inferred or not; that so the controversy may be ended [m]oreover, if a controversy be raised of the accurate and proper signification, that is, the definition of those names or appellations which are commonly used, insomuch as it is needful for the peace of the city, or the distribution of right, to be determined; the determination will belong to the city [f]or example, if a woman bring forth a child of an unwonted shape, and the law forbid to kill a man; the question is, whether the child be a man. It is demanded therefore, what a man is. No man doubts but the city shall judge it. (1972, 344–5)
Hobbesian sovereignty and the spectre of tyranny 103 Men have a hard time accepting this logical fact, according to Hobbes, and they will often tie themselves in knots to escape what he thinks are inevitable conclusions: This same supreme command and absolute power, seems so harsh to the greatest part of men, as they hate the very naming of them [to] avoid this kind of supreme authority, some of them will have a city well enough constituted to which purpose, and also to the repelling of a foreign enemy, they appoint a certain and limited return, with this condition, that if that suffice not, they may call a new convention of estates. Who sees not in a city thus constituted, that the assembly who prescribed those things had an absolute power? [f]or he that by right hath this might given, by punishments to restrain what citizens he pleaseth, hath such a power as a greater cannot possibly be given by any citizens. (1968, 265) It is the pursuit of logical conclusions, rather than a desire to tyrannize, that informs Hobbes’s arguments and he suggests that anyone who rejects his claims has failed to understand the nature of political rule; politics is necessarily absolutist. Tinkering with political power through mechanisms such as the separation of powers simply makes it more difficult to specify which branch of government has the final say, but at some point in the process one branch of government will demonstrate its sovereign power by making a decision that has binding force. The more difficult it becomes to identify the source of political power, the more likely the polity will collapse; sovereign powers are ‘incommunicable, and inseparable [and] a Kingdome divided in it selfe cannot stand’ (1968, 236). These are, again, points about logical consistency, and to those who claim that the sovereign is a greater power than each individual but not as great a power as the people collectively, he says: if by all together, they understand them as one Person (which person the Soveraign bears), then the power of all together, is the same with the Soveraign’s power; and the speech is absurd: which absurdity they see well enough, when the Soveraignty is in an Assembly of the people; but in a Monarch they see it not; and yet the power of the Soveraignty is the same in whomsoever it be placed. (1968, 237) Boyd correctly notes that Hobbes has few misgivings about the state’s unlimited and undivided power; his error is to link this support with
104 Hobbesian sovereignty and the spectre of tyranny a corresponding desire to terrorize individuals. Hobbes is concerned with deductive definitions, not with coercion and compulsion. The question of sovereignty is very important for Hobbes because if we cannot solve our differences politically we are likely to settle them through war. The most crucial task of political science, therefore, is to pinpoint a single source of sovereignty. The best way to do this, as far as Hobbes is concerned, is to have only one seat of political power. A second-rate solution, but one that is necessary if we have multiple areas of power, is to make sure that one of the political bodies is more powerful than the rest. The key point to reinforce is that once a decision is binding, the body that made the decision must be the final arbiter on the issue and hence is, by definition, absolute: ‘that king whose power is limited is not superior to him or them that have the power to limit it; and he that is not superior is not supreme, that is to say not sovereign. The sovereignty therefore was always in that assembly which had the right to limit him; and by consequence the government not monarchy, but either democracy or aristocracy’ (1968, 246–7). Hobbes would suggest that Americans are wrong to think they have escaped from absolutist politics. The buck has to stop somewhere in US politics, just as it does elsewhere, and wherever it stops that is where absolute sovereignty resides. An example that would have pleased Hobbes was the initiative proposal (Amendment II) in Colorado to limit the right of homosexuals to claim discrimination before the courts. It passed by majority rule but was then taken to the state court where it was deemed unconstitutional, and eventually worked its way to the Supreme Court, where it met with the same decision. A Hobbesian analysis of this process would declare that on this issue the Supreme Court was absolute because no other political body could overrule its decision. The amendment worked its way through a variety of political institutions until it came to rest with the final arbiter. At this point, the parties involved have to accept the decision of the court or use extra-political means to try and achieve their goals. There are two problems with the American political system from a Hobbesian point of view. The first is that it is often not so easy to identify exactly where sovereignty does lie. The second is that even when we can identify sovereignty, it is spread across a variety of political institutions. This is precisely why Hobbes preferred a unitary form of government. On the Colorado initiative, the court decided, but on other issues, Congress or the President has a similar decisive power. Hobbes would suggest that politics in the United States does
Hobbesian sovereignty and the spectre of tyranny 105 not escape absolute power it is simply that there are various powers that are absolute in certain areas. And if these powers can also be interfered with, by changing the constitution for example, then absolute sovereignty ultimately resides with the people. Despite the best efforts of its founders, the US Constitution does not circumnavigate the issue of absolute sovereignty. It may often be the case that there is no sovereign to be found, but we always know when one is present because the sovereign is the one that makes the final decision. One need not sympathize with recent attempts to turn Hobbes into the founding figure of rational choice theory to accept that he is concerned with issues of transitivity in power relationships. Imagine that we have three sources of political power: (a) the people, (b) the legislature and (c) the executive. Hobbes wants to make sure that if ‘a’ beats ‘b’ and ‘b’ beats ‘c’, then ‘a’ also beats ‘c’. If this does not happen, we cannot be guaranteed a binding decision, which means that we cannot have political stability. Hobbes is stating what Kenneth Arrow noted 300 years later, namely politics is either absolutist or potentially unstable. His solution to this problem is simple and persuasive: remove the transitivity problem by concentrating power in one institution.
Hobbes’s intentions Liberty Another way to address the criticism of Hobbes’s arguments about the unlimited nature of political power is to examine his definition of natural freedom. He tells us, ‘I conceive of liberty to be rightly defined in this manner: Liberty is the absence of all the impediments to action that are not contained in the nature and intrinsical quality of the agent’ (1968, 38). He continues by claiming that ‘LIBERTY, or FREEDOM, signifieth, properly, the absence of opposition and may be applied no less to irrational, and inanimate creatures, than to rational’ (1968, 261). Hobbes concludes his discussion by telling us that by opposition he means external physical impediments to motion. It is important to note that when Hobbes talks of external impediments he means some material body that physically impedes motion, such as a wall, a door, chains, or the banks of a river: ‘For whatsoever is so tyed, or environed, as it cannot move, but within a certain space, which space is determined by the opposition of some externall body, we say it hath not Liberty to go further (1968, 261–2). This is why Alan Ryan suggests, for Hobbes, ‘Liberty as unimpededness is the only sort of liberty there
106 Hobbesian sovereignty and the spectre of tyranny is’ (1988, 199). In this definition of freedom there is no requirement for rationality, choice or animation, and freedom is compatible with fear and determinism. If an object is in motion, whether it is a stone rolling down a hill, a river flowing through a valley, or a human being walking down the street, it is free until its motion is stopped. There are many problems with this definition of freedom but we need not dwell on them here. The question I am interested in is the impact this understanding of freedom has on political life. If freedom is as Hobbes describes it, there is a huge amount of it in his commonwealth because there are very few physical, external objects that prevent motion. A small number of people will find themselves in jail or prison and some will perhaps be physically impeded in other ways. But by and large most people will not be limited in their movements regardless of the actions of the sovereign. This conclusion suggests that a concept of freedom that only deals with physical externalities is severely flawed. In fact, Hobbes himself cannot hold to this basic definition of freedom and provides an alternative meaning when he turns his attention to civil society. In particular, he claims that laws can limit liberty, and that the end of civil laws ‘is no other but to limit the naturall liberty of particular men, in such a manner, as they might not hurt, but assist one another’ (1968, 314). It is true that such laws are external to the agent, but they are not physical impediments like a brick wall or a locked door. According to Hobbes, laws limit our freedom, not because they physically impede our motion, but because we fear the consequences of disobeying the law. This, however, is an internal constraint, something that should have no impact on freedom given his original definition. In Leviathan he makes it clear that the use of Lawes is not to bind the People from all Voluntary actions; but to direct and keep them in such a motion, as not to hurt themselves by their own impetuous desires, rashnesse, or indiscretion, as hedges are set, not to stop Travellers, but to keep them in the way. And therefore a Law that is not Needfull, having not the true End of a Law, is not good. (1968, 388) He continues, ‘It belongeth therefore to the Office of the Legislator to make the reason Perspicuous, why the Law was made; and the Body of the Law it selfe, as short, but in as proper, and significant termes as may be’ (1968, 389). The very task set for the sovereign by Hobbes is to navigate the treacherous path between too much liberty and too much coercion, which in itself suggests that he takes freedom seriously.
Hobbesian sovereignty and the spectre of tyranny 107 Boyd disagrees and claims that Hobbes is not very concerned with protecting individual liberty: ‘If individuality is really such an absolute value for Hobbes, it is unclear why he should so willingly acknowledge that it must be limited when it conflicts with civil order’ (2001, 402) and Tarlton claims that ‘the utter surrender of liberty and judgment lies at the root of his political theory’ (2002, 61). When we examine Hobbes’s specific arguments, however, we find that he embraces the idea of individual freedom. Members of his commonwealth are granted the ‘Liberty to buy, and sell, and otherwise contract with one another; to choose their own aboad, their own diet, their own trade of life, and institute their children as they themselves think fit; and the like’ (1968, 264) and should be free to get ‘food, ayre, medecine, or any other thing without which he cannot live’ (1968, 269, my emphasis). In a well-governed city, Hobbes suggests ‘that there be infinite cases which are neither commanded nor prohibited, but that every man may either do or not do them as he lists himself. In these, each man is said to enjoy his liberty the more is left undetermined by the laws, the more liberty they enjoy’ (1972, 268, my emphasis). Even John Stuart Mill agrees with these sentiments and states that ‘[b]oth extremes [too much or too little liberty] are faulty’ (1972, 268). Hobbes suggests that ‘[a] good Law is that, which is Needfull, for the Good of the People, and withall Perspicuous’ (1968, 388). When Hobbes talks about civil liberty, he always refers to the rule of law. Hence, it was clear in his own mind that freedom is linked to procedures that provide clear boundaries for action. He thought that arbitrary rule was the opposite of freedom because men could not predict the actions of the government. He questions whether ‘a law may be good, when it is for the benefit of the soveraign; though it be not necessary for the people; but it is not so. For the good of the Soveraign and the people cannot be separated’ (1968, 388). All legislation should be made as clear as possible: ‘It belongeth therefore to the Office of the Legislator to make the reason Perspicuous, why the Law was made; and the Body of the Law it selfe, as short, but in as proper, and significant termes as may be’ (1968, 389). Laws should be few in number: For since men are wont commonly to debate what to do or not to do, by natural reason rather than any knowledge of the laws, where there are more laws than can easily be remembered, and whereby such things are forbidden as reason of itself prohibits not of necessity, they must through ignorance, without the least evil
108 Hobbesian sovereignty and the spectre of tyranny intention, fall within the compass of laws, as gins laid to entrap their harmless liberty; which supreme commanders are bound to preserve for their subjects by the laws of nature. It is a great part of that liberty, which is harmless to civil government and necessary for each subject to live happily, that there be no penalties dreaded but what they may both foresee and look for. (1972, 268–9) These are just a few of the dozens of similar comments made by Hobbes on this issue. Further evidence that Hobbes advocated rather than simply tolerated freedom can be found in his assessment of human nature: ‘there are very few’ he tells us, that are ‘so foolish that had not rather governe themselves’ (1968, 211). The best way to maintain a peaceful commonwealth is to leave people alone unless they are transgressing laws that preserve peace. For reasons concerned with promoting personal choice, self-development and autonomy, Hobbes suggests limited interference in the lives of peaceful citizens: ‘A plain husbandman is more prudent in the affairs of his own house than a privy councillor in the affairs of another man’ (1968, 138). Hobbes makes it explicit in De Cive that liberty is one of the positive goods to develop out of civil association. He suggests that, ‘[t]he benefits of subjects respecting this life only, may be distributed into four kinds. 1. That they be defended against foreign enemies. 2. That peace be preserved at home. 3. That they be enriched, as much as may consist with public security. 4. That they enjoy a harmless liberty’ (i.e. that they are left alone as long as they do not harm others)’ (1972, 260). I think it is worth paying more attention to the distinction that Hobbes makes between natural and civil freedom. The latter fits closely with his description of the state of nature, and the basic point Hobbes wishes to make is that the human condition manifests its most basic and brutal form when liberty is reduced to nothing more than license. Hence, it is true that Hobbes is not a supporter of the view that freedom is an unqualified good. What Hobbes wants instead is a bounded condition of freedom that he calls civil freedom. This condition stands in sharp contrast to the one so famously described in Chapter 13 of Leviathan. Instead of a condition in which arts, society and commerce are absent, we find a description of the opposite. A flourishing society for Hobbes has to include space for the pursuit of science, industry, art, leisure and education; such things are lacking when there is either too much freedom or when there is not enough. Civil freedom is a bounded freedom, and the correct parameters for
Hobbesian sovereignty and the spectre of tyranny 109 this are to be found in the laws of nature that suggest individuals should not harm one another and that everyone should be treated equally before the law; it is the task of the sovereign to turn these natural laws into positive law. As already quoted, the purpose of civil law, ‘is no other but to limit the naturall liberty of particular men, in such a manner, as they might not hurt, but assist one another’ (1968, 315). Legislation sets the boundary of civil freedom, and Hobbes recommends that laws should be few in number and aimed primarily at preventing harm. The upshot is that even though Hobbes gives conflicting arguments about what freedom is, all of his claims suggest that his commonwealth will be a place of considerable liberty. What we find is that Hobbes is using the term civil liberty in much the same way that we do. One of the consequences of this is that it reopens the topic of how much freedom there is in his commonwealth because we can no longer assume that people are free to the extent they are physically unimpeded. It is worth looking at some specific topics to illustrate the nuanced way in which Hobbes approaches the issue of civic freedom and in the next section I examine his ideas on censorship, freedom of association and the ownership of private property. Censorship Hobbes tells us that the governing power has to decide ‘how far, and what men are to be trusted withal, in speaking to multitudes of people, and who shall examine the doctrines of all books before they are published’ (1968, 233). In De Cive he notes that ‘Man is made fit for society not by nature but by education’ (1972, 110). What are we to make of these statements? Do they suggest that his critics are right to worry about his ideas on censorship? The first line of defence for Hobbes is to make clear that such powers are provided, and are to be used, for a specific purpose, namely to preserve peace. The power of censorship is ‘annexed to the sovereignty to be judge of what opinions and doctrines are averse, and what conducing, to peace It belongeth therefore to him that hath the sovereign power to be judge (or constitute all judges) of opinions and doctrines, as a thing necessary to peace, thereby to prevent discord and civil war’ (1968, 233, my emphasis). Hobbes supports the dissemination of ideas and knowledge, and one of his criticisms of the clergy and universities is precisely that they obfuscate instead of educate. He tells us that one of the great benefits of civil society is that it is characterized by activity in the realms of science, letters and arts. The critic may suggest that
110 Hobbesian sovereignty and the spectre of tyranny I am being too lenient and point to passages where Hobbes seems to advocate banning certain texts by Greek and Roman authors: And as to rebellion in particular against Monarchy; one of the most frequent causes of it, is the Reading of the books of Policy, and Histories of the antient Greeks, and Romans; from which, young men, and all others that are unprovided of the Antidote of solid Reason, receiving a strong, and delightfull impression, of the great exploits of warre [n]ot considering the frequent Seditions, and Civill warres, produced by the imperfection of their Polity. From the reading, I say, of such books, men have undertaken to kill their Kings, because the Greek and Latine writers, in their books, and discourses of Policy, make it lawfull and laudable for any man to do so; provided before he do it, he call him Tyrant. (1968, 369) Hobbes is clearly worried about some of this material because it championed regicide, and he thinks it can never be in the interests of peace to kill the very person we have introduced to maintain order. He does not, however, advocate that such texts should be banned, even though they are the ones most likely to lead to chaos. His solution for dealing with such dangerous books is to say, ‘I cannot imagine how anything can be more prejudiciall to a Monarchy than the allowing of such books to be publikely read without present applying such correctives of discreet masters as are fit to take away their Venime; Which I will not doubt to compare to the biting of a mad Dogge’ (1968, 369–70). The correct response to such material is not the heavy hand of censorship but the introduction of challenging works that are the product of ‘discreet masters’. It was no doubt clear in Hobbes’s own mind that one such masterful corrective would be his own body of work. Hobbes suggests in De Cive that education rather than force is the best path to peace: ‘It is therefore the duty of those who have the chief authority, to root those [ideas such as regicide] out of the minds of men, not by commanding, but by teaching; not by the terror of penalties, but by the perspicuity of reasons’ (1972, 262–3). We can see, therefore, that Hobbes’s ideas on education fit well with his thoughts on censorship. To have an environment in which there is enough stability, order and peace for meaningful agency, it is necessary that members of the commonwealth develop a degree of reason and rationality. The best way to do this is not through a threat of legal violence, but by promoting a system of education that teaches us to
Hobbesian sovereignty and the spectre of tyranny 111 overcome, or at least control, our passions. As Flathman notes (1993, 149), the fact that Hobbes thinks that the state needs to teach is symptomatic of the lack of coercive power the sovereign actually has. If it could use force it would not have to worry about education. Another cause for concern is when Hobbes states that ‘the actions of men proceed from their opinions: and in the well governing of opinions, consisteth the well governing of men’s actions, in order to their peace and concord’ (1968, 233). One response to this statement is to accept that Hobbes is guilty as charged by his opponents, who claim that he wants the sovereign to censor and manipulate the thoughts of citizens. William Connolly, for example, suggests that ‘[t]he Hobbesian individual is a domesticated human’ (1988, 28) who is the subject of constant ‘regulation and control’ (1988, 34). Victoria Kahn takes a similar tack: Precisely because manly mimetic desire is synonymous with the state of war, the ideal Hobbesian reader and subject is closer to the female subject of romance or of seventeenth-century domestic manuals, the wife who consents to be bound by her passions to a hierarchical, inequitable, irrevocable marriage contract the ideal Hobbesian subject is the docile, effeminized political subject of an absolute sovereign that leads to appropriate subordination and reverence rather than insubordination and emulation. (2001, 23–4) The mistake here is to jump from the fact that Hobbes is concerned with peace and order, to the conclusion that he wants to make people servile and subordinate. When it comes to the specifics of Hobbes’s thoughts on education, he suggests the sovereign focuses on the enunciation and clarification of the rights and duties of all members of the commonwealth, the instruction that no individual has the same right as the sovereign to claim obedience, and that citizens should be taught to love and respect their own form of government institutions. We are to be taught the sum of the laws of nature, which is, ‘Do not that to another, which thou wouldest not have done to thy selfe’ (1968, 214). Hobbes challenges us to respect other persons and to support the precepts of justice that demand citizens do ‘not deprive their neighbors, by violence, or fraud, of anything which by the Sovereign Authority is theirs’ (1968, 383). Further teachings include that one should obey the law of the land, that no one should injure another person, that everyone should act charitably to all others persons, rich and poor alike, and that all should try to live soberly. Hobbes
112 Hobbesian sovereignty and the spectre of tyranny summarizes all of the above by claiming his argument ‘is reduced to this one commandment of mutual charity: Thou shalt love thy neighbour as thyself ’ (1968, 383). Hobbes thinks that the state has to inculcate certain virtues of character that promote truth telling, respect for others, toleration and the acceptance of reasonable difference. He does not support a platonic form of education in which individuals are shaped to take on a particular form. Instead, he wishes to strike a balance between, on the one hand, leaving people alone to develop and grow, and on the other hand, providing the necessary requirements for peace. The state should be interventionist without at the same time being repressive. In the dedicatory note at the beginning of Leviathan, he presents himself as a moderate who recognizes that liberty and authority are important boundaries: ‘for in a way beset with those that contend on one side for too great Liberty, and on the other side for too much Authority, ‘tis hard to passe between the points of both unwounded’ (1968, 75). It may be a difficult task to pass between these two points, but it is the basic charge Hobbes sets himself, once peace has been secured. The best arenas for such teachings are Oxford and Cambridge. The political elite need to be taught a civic doctrine that fosters peace and promotes equality: if any man would introduce sound doctrine, he must begin from the academies. There the true and truly demonstrated foundations of civil doctrine are to be laid; wherewith young men, being once endued, they may afterward, both in private and public, instruct I therefore conceive it to be the duty of supreme officers, to cause the true elements of civil doctrine to be written, and to command them to be taught in all the colleges of their several dominions. (1972, 263) As he says in Behemoth, ‘The core of rebellion, as you have seen by this, and read of other rebellions, are the Universities; which nevertheless are not to be cast away, but to be better disciplined’ (1990, 58). It may still be claimed that freedom of expression is unduly curtailed by Hobbes’s demand that the sovereign is the judge of good and evil. This amounts to the recommendation that every Member of the commonwealth has to abandon his or her private judgement on such matters. Is this a reasonable requirement? Hobbes is correct to note that no society could survive for very long if its members did not abandon the right of judgement on certain issues, especially those that can cause harm to others. As he says in De Cive, ‘first of all, it is
Hobbesian sovereignty and the spectre of tyranny 113 necessary to peace, that a man be so far forth protected against the violence of others, that he may live securely’ (1972, 176). The task, although he concedes it is impossible, is ‘[t]o make men altogether safe from mutual harms, so that they cannot be hurt or injuriously killed’ (1972, 176). It would be peculiar for Hobbes to make this claim if he thought that it was acceptable for the sovereign to harm people or for individuals to act on their own private judgement concerning good and evil. Hobbes’s argument, therefore, is quite clear and straightforward regarding censorship. He claims that all rational people will agree that a state is necessary for any form of decent human life. All rational people will agree also that the maintenance of this state has to take precedence over discourse that threatens its very existence. This is even in the interest of those engaged in dangerous speech because they are better off in a condition of peace than war. It necessarily follows that if one accepts the initial premise, one also has to accept that the sovereign can limit certain forms of speech that directly threaten the very survival of the state. Such speech will be very rare and does not even stretch to include regicide. Hobbes allows that people can challenge state decisions, make alternative suggestions and criticize the validity of certain laws. But he is saying that when the very survival of the state comes under threat, it is in the interests of everyone that this threat is addressed. Hence, he remarks in De Cive: for since there is no man who grants not to the city the judgement of those things which belong to its peace and defence, and it is manifest that the opinion which I have already recited do relate to its peace; it follows necessarily that the examination of those opinions, whether they be such or not, must be referred to the city; to him who hath the supreme authority. (1972, 180) I think it is worth noting that this is the same argument that Locke makes in A Letter Concerning Toleration and we do not have much difficulty categorizing him as a proponent of freedom. Associational life Richard Boyd (2001) is sympathetic to the type of criticisms made by Tarlton and claims that the primary task of Hobbes’s sovereign is to subdue individuals to maintain order. Boyd adds an intriguing twist to the usual line of criticism and claims that Hobbes promoted individualism to undermine all forms of associational life in order to
114 Hobbesian sovereignty and the spectre of tyranny leave the individual isolated and weak. Once we stand alone in the world we become timid and susceptible to manipulation; modern totalitarian states engage in such tactics, and Boyd suggests that Hobbes wished for something similar. It was Hobbes’s concern about pluralism rather than his desire to promote liberty that ‘contributed to the more familiar individualism at which he ultimately arrived’ (2001, 393). Hence, Hobbes supports individualism not because it was good for people but because it helps to keep them in check. Boyd claims that ‘Hobbes’s reasons for extolling the life of the rational, dissociated individual may escape his interpreters in a liberal age’ (2001, 403). In fact, Boyd goes further and makes the very strong claim that ‘[i]t is difficult to imagine why he would seek to found his Leviathan on anything but the firm bedrock of the isolated, rational, and instrumental individual’ (2001, 405), given his fear of associational life. This isolated individual is a cowed and trembling figure and the whole image presented by Boyd is pretty grim: ‘twentieth-century totalitarianism testifies to the fact that the isolated individual faces the most treacherous life of all’ (2001, 403). Boyd suggests that Hobbes is ‘Anglo-American political thought’s most extreme critic of pluralism’ (2001, 392) and this distrust of associational life is fuelled by the fact that a ‘war of group against group looms in the back of Hobbes’s mind’ (2001, 398). As noted, Boyd suggests that this fear leads to certain illiberal aspects within Hobbes’s political philosophy. For example, ‘[i]f individuality is really such an absolute value for Hobbes, it is unclear why he should so willingly acknowledge that it must be limited when it conflicts with civil order’ (2001, 402). Hobbes’s desire for order ‘leaves precious little room for associational life’ (2001, 403) and the general result of Hobbes’s fears is that he ‘has few misgivings about the state’s unlimited and undivided power to coerce individuals’ (2001, 402). The end result of such thinking is that ‘Hobbes comes down on the side of political order, depreciating the inherent dignity of religion, university life, family life, and private association. Following Hobbes’s logic to its extreme would dictate that we tolerate only the “right” kinds of associations’ (2001, 410). Despite the novel twist to Boyd’s interpretation, the message is a familiar one, namely that Hobbes supports absolutism as the best means of controlling and manipulating the members of the commonwealth. Does Hobbes provide pluralism with ‘one of the least generous treatments it has received in the history of political thought’? (2001, 410). As Boyd notes, when we examine Hobbes’s vocabulary we find that he uses inflammatory language. Hobbes has a tendency to do this with a variety of issues that concern him, whether it is freedom,
Hobbesian sovereignty and the spectre of tyranny 115 equality or the powers of the sovereign. Hence, he is often his own worst enemy when it comes to his ideas being fiercely contested. His rhetoric often sounds frenzied, and this has the effect of masking the formal logic of his arguments. Once we get past his rhetorical flourishes we usually find a more tempered and reasonable discussion of the matter at hand, and this turns out to be the case with his analysis of groups. Hobbes was very distrustful of associations that fomented the civil war in England. It would be idle, therefore, to deny Boyd’s suggestion that Hobbes is concerned about the effects of groups; he saw at first hand the terrors that one group could inflict upon another. But this does not mean that he necessarily distrusted all forms of associational life. Hobbes is surely correct to insist that it is up to the sovereign to decide which groups should be lawful and unlawful. Every political order does this so the heart of his supposed tyrannical regime cannot be his concern that groups should be monitored to some extent by the state. There is nothing necessarily dangerous about a concern for order and it could only be classified as such if Hobbes wanted to seriously curtail associational life within his polity. Is this the case? I argue that it is not. Hobbes is only concerned to deter groups that have an evil intent, pose a great danger to peace, or that are ‘perpetually meddling with Fundamental laws’ (1968, 373–4). As Boyd himself says, Hobbes is concerned with ‘the dangerous consequences of group fanaticism’ (2001, 396, my emphasis). When Boyd identifies the types of groups Hobbes is particularly worried about, he states that ‘foremost on Hobbes’s mind are those sectarians, congregations, parties, militias and other mass movements at the heart of seventeenth-century social unrest’ (2001, 397). There is nothing illiberal about being worried over militias and religious groups that threaten the very existence of the commonwealth; they worry me! It does not follow that all groups are a threat, or that individuals need to be manipulated and coerced. Hobbes actually suggests that groups can petition the sovereign over a wide range of issues and express dissatisfaction with the laws of the commonwealth; it is only when they perpetually meddle with fundamental laws that his concern grows. Even in this case, Hobbes provides a warning, rather than a call to eradicate such groups. I think we can arrive at conclusions other than those reached by Boyd. One of the groups Hobbes was most interested in is the private, unlawful variety: those that unite themselves into one person representative, without any public authority at all, such as are the corporations
116 Hobbesian sovereignty and the spectre of tyranny of beggars, thieves and gypsies and the corporations of men that by authority from any foreign person unite themselves on another’s dominion, for the easier propagation of doctrines, and for making a party against the power of the commonwealth. (1968, 285) Hobbes is clearly insensitive to the needs of the destitute with this statement, but his main concern is with preserving peace, particularly in response to the threat that is posed by agents of foreign powers or standing private armies within a commonwealth. This becomes more apparent in De Cive: I call a faction, a multitude of subjects gathered together either by mutual contracts among themselves, or by the power of some one, without his or their authority who bear the supreme rule. A faction, therefore, is as it were a city in a city: for as by an union of men in the state of nature a city receives its being, so by a new union of subjects there ariseth a faction. According to this definition, a multitude of subjects who have bound themselves simply to obey any foreign prince or subject, or have made any pacts or leagues of mutual defence between themselves against all men, not excepting those who have the supreme power in the city is a faction. Also favour with the vulgar, if it be so great that by it an army may be raised Forasmuch therefore as it is true, that the state of cities among themselves is natural and hostile, those princes who permit factions do as much as if they receive an enemy within their walls: which is contrary to the subject’s safety, and therefore, also against the laws of nature. (1972, 266) The other type of systems he is concerned about is called ‘irregular.’ A concourse of people is an irregular system the lawfulness or unlawfulness whereof dependeth on the occasion and on the number of them that are assembled. If the occasion be lawful and manifest, the coucourse is lawful (as the usual meeting of men at church, or at a public show) in usual numbers It may be lawful for a thousand men to join in a petition to be delivered to a judge or magistrate; yet if a thousand men come to present it, it is a tumultuous assembly, because there needs but one or two for that purpose. But in such cases as these, it is not a set number that makes the assembly
Hobbesian sovereignty and the spectre of tyranny 117 unlawful, but such a number as the present officers are not able to suppress and bring to justice. (1968, 287–8) Again, one can disagree with the specifics of Hobbes’s thoughts on these matters, but I think Boyd exaggerates somewhat when he concludes that Hobbes ‘leaves precious little room for associational life’ (2001, 403) and especially when he suggests that Hobbes shows no respect for ‘religion, university life, family life, and private association’ (2001, 410). In the above quote, Hobbes specifically grants that religious association is perfectly valid, and that large assemblies, even when expressing heartfelt discontent, are lawful. Hobbes is worried primarily about religious fanaticism. Boyd states, Knowing full well that individuals will err, Hobbes nevertheless takes it for granted that so long as these errors are individual and not collective, their socially disruptive effects will be negligible. The ‘Madnesse’ of the religious virtuoso, asectic, prophet is of little danger to anyone but himself. Placed at the head of a militant sect, however, he undermines any ground for civil order. (2001, 407) It is worth asking if anyone would not think that militant sects, headed by religious madmen, are dangerous. We should not conclude that Hobbes is fearful of all groups simply because he has a very rational fear of extreme ones. Boyd seems to accept this in part because he suggests that the ‘natural’ man Hobbes discusses is really ‘the contentious Protestant sectarians of the seventeenth century’ (2001, 407). These are the people Hobbes is worried about, not the average person in civil society. Boyd suggests that ‘Hobbes admittedly grants some discretion for the sovereign to determine if a particular “concourse” endangers civil liberty’ and ‘[u]nder the best of circumstances, this may afford significant room for associational life’ (2001, 394). I suggest that this is partly correct; Hobbes is concerned to limit interference with group activities to protect civil liberty, and this means that under normal rather than the ‘best’ conditions there is plenty of space for a wide range of associations. This is not to deny that ‘Hobbes comes down on the side of political order’, or that ‘[f]ollowing Hobbes’s logic to its extreme would dictate that we tolerate only the “right” kinds of associations’ (2001, 410). Without order we cannot have group life at all because we would have no society. This is not a controversial position to hold, and John Stuart Mill suggested that the most important political issue we face is finding the correct limits to what people can and cannot do: ‘where to
118 Hobbesian sovereignty and the spectre of tyranny place the limit – how to make the fitting adjustment between individual independence and social control. What these rules should be is the principle question of human affairs’ (1978, 63). Boyd is correct to argue that Hobbes’s logic leads to the conclusion that only the right kinds of association can be tolerated, but most political philosophers have reached the same conclusion. Whether this is a good or a bad thing depends on what counts as the ‘right’ kind of group; the groups that Hobbes seems to discriminate against are those that try to destroy the state. Arguments that portray Hobbes as an opponent of associational life are usually based on a claim that he supports a privatized form of individualism. This interpretation is popular because Hobbes wishes to ‘deny man to be born fit for society’ (1972, 110). Statements such as this one lead people to assume that Hobbes thought of humans as naturally isolated individuals who have to be subdued and controlled. Despite the overwhelming prominence of this view, the stronger argument is that he thought the opposite; certainly the logical thrust of his argument was to suggest that we are communal beings. It is better to think of Hobbes’s statement above as referring to his views on education. Hobbes thought that we are born not knowing how to interact with other human beings in a civil society. There is no innate knowledge that makes us naturally social beings. This is perfectly compatible with Hobbes’s claim that civil society is the best environment for us to survive and flourish. In fact, Hobbes thought we can live fulfilling lives only as communal beings. The state of nature is not the natural condition for human beings and is presented by Hobbes as counter factual to our actual social existence. The ‘natural’ state is, in fact, unnatural and that is why we would not want to be a part of it; it undermines everything that makes us human, such as arts, science, letters, companionship and commerce. It is also unnatural in the sense that it does not fit us as biological entities; even the strongest of us will perish in such a condition. It is not surprising, therefore, that Hobbes suggests we would wish to leave it at our earliest convenience and that we would make all possible attempts to ensure we do not end up in such a condition in the future. The message Hobbes is sending is that there can be no existence without community. When he tells us that man is not born fit for society, he means that the raw material has to be shaped and moulded to become sociable. The truth of the matter, he tells us, is ‘to man by nature, or as man, that is, as soon as he is born, solitude is an enemy; for infants have need of others to help them to
Hobbesian sovereignty and the spectre of tyranny 119 live, and those of riper years to help them to live well’ (1972, 110). Our existence depends upon association with others. This must mean that social conventions do not oppose our nature (except perhaps for the vain glorious few, whom he calls madmen, who deny equality) because we could not maintain a community in their absence; we want them, and we cannot do without them. For conventions to have force we need to be disposed to live by them and Hobbes suggests that this happens due to a mixture of self-interest, fear and the threat of punishment. Just as importantly, we also need to hold a favourable disposition to one another and be willing to engage in social cooperation. I propose that rather than advocating brainwashing, Hobbes is attempting to promote a rational and legitimate form of self-regulation. Self-regulation is the crucial piece of the puzzle because Hobbes knew that force alone was not enough to guarantee peace: ‘These rights have to be diligently and truly taught because they cannot be maintained by any civil law or terror of legal punishment’ (1968, 377). As Holmes has noted, ‘Hobbes stressed the self-defeating character of attempts to change people’s minds by brutal means’ (1990, 140). The key is to safeguard peace, and this is intended should be done, not by care applied to individuals, further than their protection from injuries when they shall complain, but by a general providence, contained in public instruction, both of doctrine and example, and in the making and executing of good laws, to which individuals may apply their own cases. (1968, 376) Hobbes’s goal is to show us that we cannot develop the human faculties independently of civil society; indeed his goal is to demonstrate that we cannot even survive without a political community. We are all familiar with Hobbes’s famous description of the state of nature, but not enough attention is paid to the fact that it is supposed to represent the opposite of what a civil society should look like. Seen in this light, we should reverse Hobbes’s famous quote about life being nasty, brutish and short, and think of society as characterized by cultivation of the earth, navigation of the seas, commodious buildings, knowledge of the face of the earth, an account of time, arts, letters, society, and what is best of all, peace of mind, safety, and the life of man, communal, rich, kind, civilized and prolonged. If this is right there is no need for the sovereign to attempt to create isolated monads because Hobbes did not see groups posing the sorts of threat identified by Boyd. I am not suggesting that Hobbes is opposed to individualism
120 Hobbesian sovereignty and the spectre of tyranny (he is favourably disposed to individual liberty once security and order have been secured), but I do claim that the individualism favoured by Hobbes does not come in the form identified by Boyd. I would like to reiterate that I am arguing with Boyd’s, Tarlton’s (and many other commentators’) views about Hobbes’s intentions. The usual claim is that Hobbes consciously wishes to oppress groups and individuals via a coercive sovereign. I suggest that his goals are somewhat different. Property Is property well protected in Hobbes’s commonwealth? One could argue that it is not because he states that property rights depend on the sovereign. He is quite clear about this and suggests, in De Cive, that anyone who disagrees simply does not understand the nature of the problem: The seventh doctrine opposite to government, is this; that each subject hath an absolute dominion over the goods he is in possession of: that is to say, such a propriety as excludes not only the right of all the rest of his fellow subjects to the same goods, but also the majistrate. Which is not true Before the yoke of civil society was undertaken, no man had any proper right; all things were common to all men. Tell me therefore, how gottest thou this propriety but from the majistrate? Thy dominion therefore, and propriety, is just so much as he will, and shall last so long as he pleases But the greatest part of men who profess civil prudence, reason otherwise They who talk thus know not, that what they would have, is already done from the beginning, in the very constitution of government; and therefore speaking as in a dissolute multitude and yet not fashioned government, they destroy the frame. (1972, 250) Hobbes makes it clear that his mind remained unchanged on this matter when he wrote Leviathan, although he expressed himself in a more flamboyant manner: The constitution of Mine, and Thine, and His; that is to say, in one word Propriety; and belongeth in all kinds of Commonwealth to the Soveraign Power. For where there is no commonwealth, there is a perpetuall warre of every man against his neighbour; And therefore every thing is his that getteth it, and keepeth it by force; which is neither Propriety nor Community;
Hobbesian sovereignty and the spectre of tyranny 121 but Uncertainty [t]ake away the Civill Law, and no man knows what is his own, and what another mans. (1968, 296) Hobbes asks us to imagine how we could claim a right to property against an assailant if there was not a sovereign power in place. His answer is that we cannot because we have no claim on something if there is no law, and there cannot be law without a sovereign. Rights for Hobbes are social rights; there is no sense to the claim that one can have a right to property outside a political system. What stems from this is that we can never have an absolute property right in society, because the sovereign can legally prescribe how we utilize our possessions. This, however, is perfectly reasonable; all states place limits on property so that it is utilized in a manner that does not disturb the public peace. Liberal states have also been quick to seize property for the collective good in times of warfare. Even Robert Nozick (1974) denies that persons can do whatever they wish with their property. The result of such inconveniences is more than compensated for, according to Hobbes, because we gain a huge amount of security against the molestation of others: ‘the Propriety which a subject hath in his lands, consisteth in a right to exclude all other subjects from the use of them; and not to exclude their Soveraign, be it an Assembly, or a Monarch’ (1968, 297). Such protection is lost if we prevent the sovereign from intervening: ‘But if the right of the sovereign also be excluded, he cannot perform the office they have put him into; which is to defend them both from foreign enemies, and from the injuries of one another; and consequently there is no longer a commonwealth’ (1968, 367–8). We cannot exclude the sovereign, but we gain the inestimable advantage that we can exclude everyone else. To suggest otherwise is to wish impossibilities; we would have our security against all the world upon right of property, without paying for it; this is impossible. We may as well expect that fish and fowl should boil, roast, and dish themselves, and come to the table, and that grapes should squeeze themselves into our mouths Secondly, there is no nation in the world where he or they that have the sovereignty, do not take what money they please for the defence of those respective nations, when they think it necessary for their safety. (1966, 20) It has never been the case, Hobbes tells us, that Englishmen have had an absolute right to property, and the
122 Hobbesian sovereignty and the spectre of tyranny Magna Charta, that was made by some King more ancient yet that statute was made, not to exempt any man from payments to the public, but for securing every man from such as abused the King’s power by surreptitiously obtaining the King’s warrants, to the oppressing of those against whom he had any suit in law. But it was conducing to the ends of some rebellious spirits in this Parliament, to have it interpreted in the wrong sense. (1990, 37–8) Hobbes certainly thinks that the sovereign should use its powers in a very restrained manner: if the ruler levy such a sum of vast monies from his subjects, as they are not able to maintain themselves and their families, nor conserve their bodily strength and vigor, the disadvantage is as much his as theirs [b]ut if he raise no more than is sufficient for the due administration of his power, that is a benefit equal to himself and his subjects. (1972, 223) It is in the interests of the sovereign, just as much as in the subject, that members of the commonwealth are allowed to prosper and trade in an environment free from the fear of arbitrary interference with property. Capricious acts by government go against the law of nature, which tells everyone, including the sovereign, that the path to peace is to treat all people equitably. Hobbes also suggests that history is on his side in this matter: Nor can the people, nor any man that humours their disobedience, produce any example of a King that ever raised any excessive sums, either by himself or by the consent of his Parliament, but when they had great need thereof; nor can show any reason that might move any of them so to do. (1966, 11) Concerns about the protection of property are not only pertinent to his preferred form of monarchical politics, but are sown as the seeds of any political association. When people worry that the sovereign will squander public resources, he admits that this can happen: I confess this is a grievance, but of the number of those which accompany all kinds of government, but are more tolerable in a monarchy than in a democracy in a democracy, look how many demagogues, that is, how many powerful orators there are
Hobbesian sovereignty and the spectre of tyranny 123 with the people, so many children, kinsmen, friends, and flatterers are to be rewarded. (1972, 226) It is clear that Hobbes prefers monarchy rather than democracy precisely because he thinks that liberty and property are under more of a threat from the latter. The fifth law of nature suggests that the sovereign can distribute certain goods if they are unnecessary to one person but crucial to another. A huge economic imbalance between the rich and poor is a major threat to peace and security because ‘nothing more afflicts the mind of man than poverty’ (1972, 135). Too much wealth stored in the hands of a few powerful persons is just as dangerous; he describes monopolies of this kind as diseases that afflict the body politic. Monopolies, along with the claim that the sovereign cannot have access to personal property, can lead to the extinction of the commonwealth: when the treasure of the commonwealth, flowing out of its due course, is gathered together in too much abundance in one, or a few private men, by monopolies breedeth there an inflammation, accompanied with a fever, and painful stitches whereby there succeedeth at first a cold contraction, and trembling of the limbs; and afterwards a hot, and strong endeavour of the heart, to force a passage for the blood; and before it can do that, contenteth itself with the small refreshments of such things as cool for a time, till (if nature be strong enough), it break at last the contumacy of the parts obstructed, and dissipateth the venom into sweat; or (if nature be too weak) the patient dieth. (1968, 373–4) It is worth noting why Hobbes was so bullish on this matter. He thought that one of the primary reasons for the civil war was the battle that raged between the King and Parliament over raising revenue, and in particular he blamed the latter’s unwillingness to provide finances to support the military. In A Dialogue the character of the philosopher states that when there is a Parliament, if the speaking and leading men should have a design to put down monarchy, as they had in the Parliament which began to sit the third of November, 1640, shall the King, who is to answer to God Almighty for the safety of the people, and to that end is intrusted with the power to levy and dispose of the soldiery, be disabled to perform his office, by virtue of these acts of Parliament which you have cited? If this be reason, it is
124 Hobbesian sovereignty and the spectre of tyranny reason also that the people be abandoned, or left at liberty to kill one another, even to the last man. (1966, 13–14) The philosopher is clearly an unhappy fellow because he continues his complaint by stating that ‘the King cannot make his laws effectual, nor defend his people against their enemies, without a power to levy soldiers’ (1966, 10). As Hobbes caustically notes, the opponents of the King were quick to gain access to the property of Englishmen when they had power. Parliamentarians ‘no sooner took upon them the sovereign power, than they levied money upon the people at their own discretion. Did any of their subjects dispute their power or made they any doubt but to be obeyed in all that they commanded, as a right absolutely due to the sovereign power in whomsoever it resides?, (1966, 17–18). Finally, Hobbes suggests that the sovereign shall assigne in what places, and for what commodities the Subject shall traffique abroad it belongeth to the Common-wealth (that is, to the soveraign only) to approve, or disapprove both of the places, and matter of forraign Traffique’ and ‘to appoint in what manner all kinds of contract between Subjects, (as buying, selling, exchanging, borrowing, lending, letting, and taking to hire,) are to bee made. (1968, 299) This suggests that Hobbes wanted a considerable amount of freedom in trade, but that he also thought it was necessary that the sovereign place certain rules of engagement upon participants. Hence, Hobbes places limits on the accumulation of property out of a concern for liberty, peace and commodious living. We should not conclude, however, that Hobbes is arguing for arbitrary interference with property; instead, he provides an alternative view to the radical and inflammatory claim that men are thought ‘to be so much master of whatsoever he possessed, that it could not be taken from him upon any pretence of common safety without his own consent’ (1990, 4, my emphasis). He provides a variety of reasons why the sovereign has to have access to property, but none of them concern the personal aggrandizement of the sovereign or condone the arbitrary removal of holdings. Instead, he suggests access should be granted for the following reasons: protecting rights, punishing offenders, raising revenue (in particular to provide adequate defense), preventing monopolies, providing protection for the poorest members of society, preventing the use of property that causes harm or nuisance to others,
Hobbesian sovereignty and the spectre of tyranny 125 governing the use of property even when it is not harming others (zoning laws would be an example), placing limits on trade, borrowing money and distributing awards. Hobbes’s position turns out to be similar to that of one of the great defenders of property, John Locke. Both suggest that property cannot be protected without state power. In one sense, Locke is more stringent than Hobbes because he claims that there are limits placed on the use of property in nature. The property right we have in our own bodies does not extend to killing or enslaving ourselves and the property we extract from nature also comes with limits. We cannot make the situation of others worse than before we acquired our property, and hence we have to leave as good and enough for others. The land we take from the common also has to be used in an appropriate manner; we cannot simply put a fence around it and let it rot, nor can we deliberately destroy it. The spoilage proviso does not allow us to let our produce go to waste and holds even when everyone else is well taken care of; the price of creating spoilage is punishment under the laws of nature. Property must be acquired justly and we are not to use our goods to cause harm to others. Locke also suggested that ownership of goods is limited by the demands of charity and he believed that the desperately poor have a claim to our surplus. Finally, we cannot utilize our resources in a manner that negates the obligations we have to our children. Hence, Locke suggests there are moral limits on property, even in the state of nature, and the claims that Locke is providing a theoretical defence of capitalist appropriation are very questionable. Locke is seen to differ from Hobbes when the discussion turns to the ownership of property in civil society: ‘[t]he Supreme power cannot take from any Man any part of his Property without his own consent’ (1988, 360). Hence, the argument goes, property is more secure under a Lockean government than the Hobbesian variety. Things are not quite so simple, however, once we discover what Locke counts as consent. He makes a variety of statements about consent, and the ones that are of interest here include [t]is true, Governments cannot be supported without great Charge’, and ‘tis fit every one who enjoys his share of the Protection, should pay out of his Estate his proportion of the maintenance of it. But still it must be with his own Consent, i.e. the Consent of the majority, giving it either by themselves, or their Representatives chosen by them’ (1988, 362). It is a startling fact that Locke links consent to the rule of the majority nine times in the first five paragraphs of Chapter 8 and it suggests that property rights are more limited than might at first appear. The very fact that one has
126 Hobbesian sovereignty and the spectre of tyranny joined the community means that one has also given up the right to property as it existed prior to society: Men at first, for the most part, contented themselves with what un-assisted Nature offered to their Necessities: and though afterwards, in some parts of the World, (where the Increase of People and Stock, with the Use of Money) had made Land scarce, and so of some Value, the several Communities settled the Bounds of their distinct Territories, and by Laws within themselves, regulated the Properties of the private Men of their Society, and so, by Compact and Agreement, settled the Property which Labour and Industry began. (1988, 299) This is made even more clear when Locke discusses tacit consent, where he suggests that every Man, that hath any Possession, or Enjoyment, of any part of the Dominions of any government, doth thereby give his tacit Consent, and is as far forth obliged to Obedience to the Laws of that Government, during such Enjoyment, as any one under it; whether this his Possessions be of Land, to him or his Heirs for ever, or a Lodging only for a Week; or whether it be barely travelling freely on the Highway; and in Effect, it reaches as far as the very being of any one within the Territories of that Government. (1988, 348) Here, Locke makes it clear that simply owning property is a sign that a person has consented to the actions of the sovereign body, which can then place limits on property for the common good. Hence, his claim that the state cannot take one’s property without consent is no stronger than that of Hobbes. What Locke is really concerned about is the threat that the state will ‘dispose of the Estates of the Subject arbitrarily’ (1988, 361), but of course Hobbes expresses exactly the same sentiment, and I would argue that property is as safe in Hobbes’s commonwealth as it is in Locke’s. The difference between Hobbes and Locke has been somewhat exaggerated. Hobbes is depicted as the supporter of tyrannical absolutism, whereas Locke is seen as the first great defender of limited government. In fact, Locke provides the executive branch of government with extensive powers. He describes the executive as the supreme governmental power that implements the law and exercises federative power, by which Locke means the power to deal with persons
Hobbesian sovereignty and the spectre of tyranny 127 and political bodies outside of the commonwealth. More importantly, Locke provides the executive with prerogative powers that allow it to operate at its own discretion beyond the legal purview of the office. This power is mainly used where the law is silent, but it can also override or oppose established law. Its use is legitimate when used for the public good under conditions that require swift and decisive action. Hobbes states repeatedly that the sovereign should act in accord with known and settled laws whereas Locke tells us that good executives cannot have too much prerogative power. Basically, prerogative is the exercise of power beyond right. A prince who uses prerogative is not far removed from one who uses tyranny because the defining feature of both, according to Locke, is the use of power beyond the boundaries of the law. The difference is that prerogative is used for the collective good and tyranny is used for personal gain; it is left to the people to decide which of these circumstances hold. The difference between the best and the worst prince seems to rest on the perception of the people, and as long as they are happy with the actions of the executive it is permissible for it to act outside the boundaries of the law. When it is decided by the people that tyranny has taken hold it can be opposed by force. It is unclear why this form of government is more legitimate than Hobbes’s sovereign. I hope the above arguments support the claim that Hobbes was a supporter of liberty. However, a second criticism is perhaps most pressing: this suggests that regardless of Hobbes’s intentions, liberty is simply not compatible with absolutism. If this is correct we will have to conclude that Hobbes did not succeed in promoting liberty, even if this was his goal, and my attempt to utilize his argument to defend democracy will have to be abandoned.
Success? Even if it is granted that Hobbes’s intentions are good, it might still be argued that liberty is better protected by institutional mechanisms such as the separation of powers and a bill of rights. It is worth addressing this question here briefly before turning to it more fully in Chapter 6 and one way to do this is to see what Hobbes himself has to say. His main response has already been covered, namely that the separation of powers and written documents do not circumvent the issue of absolutism. He has three other responses, and the first is to suggest that the human condition is better in any political system than in the state of nature. The incommodities of life that will descend upon us
128 Hobbesian sovereignty and the spectre of tyranny if we divide power are ‘scarce sensible, in respect of the miseries, and horrible calamities, that accompany a Civill Warre; or that dissolute condition of masterlesse men, without subjection to Lawes’ (1968, 238). I think it is fair to say that Hobbes exaggerated the consequences of separating power. His second response is to suggest that the sovereign will not harm individuals because it is not in its interests to do so. Immediately following the previous statement, he says, nor considering that the greatest pressure of Soveraign Governours, proceedeth not from any delight, or profit they can expect in the dammage, or weakening of their subjects, in whose vigor, consisteth their own strength and glory (1968, 238) [f]or no king can be rich, nor glorious, nor secure; whose Subjects are either poore, or contemptible, or too weak through want, or dissention. (1968, 242) Individuals have a tendency to note harms done to them and ignore actions for their benefit: For all men are by nature provided with notable multiplying glasses, (that is their Passions and Self-Love) through which, every little payment appeareth a great grievance; but are destitute of those prospective glasses, (namely Morall and Civill Science), to see a farre off the miseries that hang over them. (1968, 239) Hence, the sovereign has to take special care not to antagonize the populace. Third, and most importantly, Hobbes’s strongest reply to his objectors is simply to suggest that no iron-clad guarantees against the abuse of political power can be given, and to think otherwise is again to misunderstand the nature of politics. Hobbes himself thought that monarchy was the safest form of sovereignty. A common complaint is that the ruler hath power not only to appoint what punishments he lists on any transgressions, but that he may also in his wrath and sensuality slaughter his innocent subjects But it is the fault of the ruler, not of the government. For all the acts of Nero are not essential to monarchy; yet subjects are less often undeservedly condemned under one ruler, than under the people in a popular dominion, there may be as many Nero’s as there are orators who soothe the people, and they mutually give way to each other’s appetite. (1972, 226–7)
Hobbesian sovereignty and the spectre of tyranny 129 The crucial point that Hobbes is making is that political power is necessary, but that it is also necessarily dangerous, and we forget this message at our peril. Hobbes himself did not shy away from this disconcerting conclusion, nor did he try and hide the unpalatable truth from his readers. As noted above, alternative systems such as the separation of powers do not solve the problem and simply muddy the waters of political accountability. The key point to keep in mind is that Hobbes thought that absolutism is necessary for the preservation of liberty. The crucial task for the person engaged in science (which for Hobbes is about definitions and deductive reasoning) is to pinpoint a single source of authority; if we cannot do this then liberty is threatened. Note also the democratic potential of Hobbes’s sovereign. The fact that Hobbes repeatedly calls the sovereign a person should not lead us to the conclusion that it can be only one man. Although Tarlton notes that Hobbes’s sovereign can take on different guises, his critique is based heavily on assuming that power resides with one (power-crazed) individual: ‘for him [Hobbes] the minimum power of government had to be absolute and unlimited, at the same time that it was in the hands of a Hobbesian man and despotism, tyranny, dictatorship, oppression and the like may very well result from bestowing such power upon a mere human being’ (2002, 83–4). Hobbes, however, states in Chapter 18 of Leviathan that the power of sovereignty can be given to either a ‘Man, or Assembly of Men’ (1968, 228). He makes the same point throughout his writings: ‘The person to whom this authority of defining punishments is given, can be no other, in any place of the world, but the same person that hath the sovereign power, be it one man or one assembly of men’ (1966, 122). It follows that the sovereign could be a monarchy, an aristocracy or a democracy: ‘the difference between these three kinds of commonwealth consisteth not in the difference of power, but in the difference of convenience or aptitude to produce the peace and security of the people’ (1968, 241). In fact, Hobbes claims that he has proven all his claims in Leviathan except his arguments in favour of monarchy, which rest on opinion rather than on logical necessity. A fully participatory democracy is as justified by Hobbes’s philosophy (i.e. civil science) as hereditary sovereignty. The arguments presented above do not mean that there are no checks on Hobbes’s governing body. What is often forgotten about Hobbes is that there is also a form of separation of functions, if not powers, in his commonwealth. Hobbes saw this as a problem, in fact he saw it as one of the causes of the civil war, but this does not mean
130 Hobbesian sovereignty and the spectre of tyranny that he was opposed to all forms of dispersed authority. He never suggested that the whole activity of governing has to be concentrated in the hands of one institution. What he does say is that the sovereign can be different from the government. For example, when in one Common-wealth there be divers Countries that have their Lawes distinct one from another, or are farre distant in place, the Administration of the Government being committed to divers persons, those Countries where the Soveraign is not resident, but governs by Commission are called Provinces. (1968, 279) A little later he tells us that ‘A PUBLIQUE MINISTER, is he, that by the Sovereign is employed in any affaires, with Authority to represent in that employment, the Person of the Common-wealth’ (1968, 289). His main concern was to ensure that when push comes to shove, we have one political authority that can decide the issue. In practice, the sovereign distributes a wide variety of governing tasks to the police, to judges, to magistrates and to local governments. It is crucial to recognize how much effort Hobbes puts into his discussion of law. This would make no sense if his motive was arbitrary rule. In Chapter 26 of Leviathan, titled ‘Of Civill Law’, he states that he is not referring to the laws of a particular commonwealth, but of commonwealths in a universal sense, ‘my designe being not to shew what is Law here, and there; but what is Law’ (1968, 311–12). This should immediately alert us to the fact that he is interested in the rule of law and not the rule of individual tyrants. Hobbes wanted judges to decide according to the body of the law, not according to what they guessed was the prevailing mood of the sovereign. Judges in Hobbes’s commonwealth are independent and are instructed to seek justice. When a judge is passing sentence, he is to do so according to what he assumes is ‘the reason of his Soveraign, which being alwaies understood to be Equity, he is bound to it by the Law of Nature’ (1968, 319). The civil law is basically the natural law with the weight of political enforcement behind it: Civill, and Naturall Law are not different kinds, but different parts of Law; whereof one part being written, is called Civill, the other unwritten, Naturall And law was brought into the world for nothing else but to limit the naturall liberty of particular men, in such a manner, as they might not hurt, but assist one another. (1968, 315)
Hobbesian sovereignty and the spectre of tyranny 131 Natural law is promulgated through some form of word, document or act; it is made known as the directive of the sovereign, and ‘to be supposed alwaies consonant to Equity and Reason’ (1968, 319). The task of judges in the United Kingdom is the same as in Hobbes’s commonwealth; their job is not to make law but to interpret it as best they can to reflect the wishes of the sovereign. When the sovereign’s will is unclear, Hobbes’s judges are supposed to use reason to come up with the most equitable conclusion because it is always to be assumed that equity is the law of nature, and that the sovereign’s role is to translate natural law into positive law. The sovereign and judges should always keep an eye on the distribution of justice: Because there is no Judge Subordinate nor Soveraign, but may erre in a Judgement of Equity; if afterward in another like case he find it more consonant to Equity to give a contrary Sentence, he is obliged to doe it. No mans error becomes his own Law; nor obliges him to persist in it. (1968, 323) Hobbes follows this with a remarkable statement. He tells us that ‘Neither (for the same reason) becomes it a Law to other Judges, though sworn to follow it’ (1968, 323). What Hobbes seems to be saying here is that judges are not obliged to follow the lead of the sovereign or previous judges when their own natural reason tells them that equity has been distorted, ‘For he that judged it first judged unjustly; and no Injustice can be a pattern of Judgement to succeeding Judges’ (1968, 324). This is an astonishing statement for Hobbes to make if Tarlton and Boyd are right and it seems more accurate to suggest that Hobbes had a great respect for the rule of law. A few pages further Hobbes says that [t]he things that make a good Judge, or good Interpreter of the Lawes, are, first, A right understanding of that principall Law of Nature called Equity; which depending not on the reading of other mens Writings, but on the goodnesse of a mans own natural Reason, and Meditation Secondly, Contempt of unnecessary Riches, and Preferments. Thirdly, To be able in judgement to devest himselfe of all fear, anger, hatred, love and compassion. Fourthly, and lastly, Patience to heare; diligent attention in hearing; and memory to retain, digest and apply what he hath heard. (1968, 328–9)
132 Hobbesian sovereignty and the spectre of tyranny The basic requirement is for a judge to use reason to make impartial decisions. In relation to the laws of nature Hobbes says that Heaven and Earth shall passe; but not one title of the Law of Nature shall passe; for it is the Eternall Law of God. Therefore all the Sentences of precedent Judges that have ever been, cannot all together make a law contrary to natural Equity or discharge the present Judge of the trouble of studying what is Equity from the principles of his own naturall reason. (1968, 324) The workings of power rest on how laws are interpreted and this provides judges in Hobbes’s commonwealth with a lot of leeway. In a Hobbesian commonwealth, there would not be a distinction between constitutional and unconstitutional; it would, instead, be between laws that promoted and those that undermined the laws of nature. These commands include the following: the 2nd law states that each person has to lay down the liberty of hindering other people; the 5th law counsels that every man should strive to accommodate himself to the rest; the 6th law advises that we should pardon those who repent; the 8th law limits certain forms of speech, but only speech that declares hatred for others; the 9th law asserts that everyone should be regarded as naturally equal, which also extends to a form of legal equality; the 10th law holds that no person can claim a right that is not available to everyone else; the 11th law commands that any person who acts as a judge is to treat all equally; the 16th law requires that anyone submitting to an arbitrator should accept the decision; the 17th law directs that no person shall be arbitrator in his own cause; and the 18th law recommends that no one shall be arbitrator who has interests in the outcome. These are the rules that have to be made into positive law by the governing body, and because they are laws of nature they apply as much to members of the sovereign body as to anyone else. Hobbes tells us that these laws are ‘contracted into one easie sum, and that is, Do not that to another, which thou wouldest not have done to thyself’ (1968, 214). These are hardly the rules of tyranny. Such laws make it clear why Hobbes’s sovereign can be charged with iniquity even if it cannot be charged with injustice (because, by definition, the sovereign cannot be charged with injustice). Hobbes makes many statements in Leviathan where he suggests that the safety of the populace is the ultimate goal of the sovereign. He also made this point very clearly in De Cive:
Hobbesian sovereignty and the spectre of tyranny 133 Now all the duties of rulers are contained in this one sentence, the safety of the people is the supreme law [I]t is their duty in all things, as much as possibly they can he, who being placed in authority shall use his power otherwise than to the safety of the people, will act against the reasons of peace, that is to say, against the laws of nature the city was not instituted for its own, but for the subject’s sake he hath fully discharged himself, if he have thoroughly endeavoured by wholesome constitutions to establish the welfare of the most part, and made it as lasting as may be by safety must be understood, not the sole preservation of life in what condition soever, but in order to its happiness to furnish their subjects abundantly, not only with the good things belonging to life, but also with those which advance delectation. (1972, 258–9) Hobbes himself was in no doubt that his form of sovereignty, particularly expressed through monarchy, was the safest. At this point, it is worth noting Hobbes’s distinction between a subject and a slave. The latter may obey a sovereign who comes to power through force, but one is equally at liberty to escape if one can. Hence, political power based solely on force is not enough for stability and has to come, ultimately, from a voluntary agreement. This gives the sovereign the right to obedience and creates an obligation to obey on the part of the subject. One should keep in mind that Hobbes wants the sovereign’s interests to coincide with those of the subjects; in this way, self-interest also works to provide protection against abuse. A brief examination of the frontispiece of Leviathan shows that the sovereign is made up of the people, and hence any harm done to them is a self-inflicted wound: ‘It is a weak Soveraign, that has weak Subjects; and a weak People, whose Soveraign wanteth Power to rule them at his will’ (1968, 388). It is claimed by Tarlton that Hobbes does not provide for legitimate political resistance, or even for the right of appeal. Commentators such as Baumgold (1993) and Burgess (1994) have shown this is not the case. In some respects, Hobbes gives a stronger right of resistance than Locke. Locke suggests that civil disobedience is only justifiable after a long train of abuses. Hobbes argues that our obligation to obey the sovereign is removed as soon as the latter uses force against us: A covenant not to defend myself from force by force is always void. For (as I have showed before) no man can transfer or lay down his right to save himself from death,
134 Hobbesian sovereignty and the spectre of tyranny wounds, and imprisonment and therefore the promise of not resisting force in no covenant transferreth any right nor is obliging notwithstanding that such Criminals have consented to the law, by which they are condemned. (1968, 199) If the sovereign does not provide safety, the people ‘may nevertheless, without injustice, refuse [to obey], as every subject has liberty in all those things, the right whereof cannot by Covenant be transferred’ (1968, 268). This last point, noted by Baumgold and Burgess in particular, is crucial because it directs us to the conclusion that members of the commonwealth do not fully abdicate the right of nature or the capacity to make private judgement. If the sovereign uses force against us we are fully entitled to defend ourselves. He makes the case for this forcefully in De Cive: it is one thing if I say, I give you right to command what you will; another, if I say, I will do whatsoever you command. And the command may be such, as I would rather die than do it. Forasmuch, therefore, as no man can be bound to will being killed, much less is he tied to that which to him is worse than death [t]here are many other cases in which, since the commands are shameful to be done by some and not by others, obedience may by right be performed by these and refused by those. (1972, 183) Hence, our obligation ends not only when the sovereign is trying to kill us, but even when it orders us to do something we find shameful. This is an unusual statement, to say the least, from a man who supposedly endorses a tyrannical brand of politics. It is also an untenable position because it means that the sovereign, which has to use force to implement the law, can be resisted every time it seeks to preform its policing duties. The fact that Hobbes would make such an argument does suggest a healthy respect for individual freedom. Hobbes makes it very clear that one can appeal the decision of a judge, at which point the sovereign can appoint another judge or examine the case itself. People can petition the sovereign in their own interests: if he, whose private interest is to be debated, and judged in the Assembly, make as many friends as he can; in him it is no Injustice; because in this case he is no part of the Assembly. And though
Hobbesian sovereignty and the spectre of tyranny 135 he hire such friends with mony (unlesse there be an expresse Law against it) yet it is not Injustice. (1968, 286) It is also perfectly lawful to gather signatures for petitions that are to be delivered to the sovereign. Hobbes suggests that ‘These properties of just and rationall Judicature considered, I cannot forbeare to observe the excellent constitution of the Courts of Justice, established both for Common and also for Publique Pleas in England’ (1968, 292). By public he means ‘those, where the Complaynant is the Soveraign’ (1968, 292), which suggests the sovereign is to act through legal channels as does everyone else. It is a misrepresentation on Tarlton’s part to suggest that ‘the arbitrator’s judgment could not be analysed, criticized or evaluated’ (2002, 76). Hobbes repeatedly makes the claim that the sovereign is limited in what it can do by religious considerations. There is much debate concerning Hobbes’s true thoughts about religion, but whether he was a believer or not is irrelevant to the argument made here. I simply want to highlight that Hobbes at least utilizes religious claims to demonstrate that his sovereign will not be despotic. In The Questions Concerning LIBERTY, NECESSITY AND CHANCE, he distinguishes between the justice of a law in relation to a particular people and to God. He says that in relation to a polity, what the sovereign decides cannot by unjust but that: in relation to God, if God have by a law forbidden it, the making of such a law is injustice. If therefore, those laws were ordained out of wantonness, or cruelty, or envy, or for the pleasing of a favorite, or out of any other sinister end, as it seems they were, the making of it was an unjust action, of which they were to give account to none but God. I fear the Bishop will think this discourse too subtile; but the judgement is the reader’s. (1966, 17) Moral limitations also apply. One may suggest, and certainly expect, that Hobbes would argue that the sovereign can still dictate the content of the natural law; in fact he does not. In Chapter 26 of Leviathan, he states that the sovereign cannot act contrary to the natural (i.e. moral) law: ‘for whatsoever is not against the Law of Nature, may be made law’ (1968, 333). Hence, ‘Injustice, Ingratitude, Arrogance, Pride, Iniquity, Acception of persons, and the rest, can never be made lawfull. For it can never be that Warre shall preserve life, and Peace destroy it’ (1968, 215). He makes a similar comment in De Cive, where he
136 Hobbesian sovereignty and the spectre of tyranny suggests that even though the sovereign has the right to kill, ‘if they use that right otherwise than right reason requires, they sin against the laws of nature, that is, against God’ (1972, 183). The sovereign’s task is to interpret rather than dictate natural law. Hobbes reinforces this point with the 11th law of nature when he states, ‘if a man be trusted to judge between man and man [the sovereign], it is a precept of the Law of Nature, that he deale Equally between them’ (1968, 212). Hobbes does not suggest that we simply wait to see what the sovereign decides are the laws of nature. Instead he tells us with great precision what they are, and the sovereign is judged according to how well it implements them. The only real difference between the natural and the civil law is that the former are accessed through reason while the latter are made known to everyone ‘by Word, Writing, or other sufficient Sign of the Will’ (1968, 312). The laws of nature provide an objective description of good and bad regardless of the views each individual holds, including the sovereign: The End of this institution is the Peace and Defense of them all For in this consisteth Equity; to which, as being a Precept of the Law of Nature, a Sovereign is as much subject as any of the meanest of his People. All breaches of the Law, are offences against the Common-wealth. (1968, 385)
Conclusion: Hobbesian absolutism revisited The critic may still be unpersuaded and claim that it is still the case that there is nothing concrete preventing the political state rampaging over the interests of the people. I would suggest that this is a problem for every political system because all states have the capacity to cause great harm. Hobbes recognized this in De Cive and suggested that wise princes ‘sometimes forbear the exercise of their right; and prudently remit somewhat of the act, but nothing of their right’ (1972, 181). The point is that political power is necessary but because of this it is also necessarily dangerous. Hobbes thought that the ill-use of such power will usually end in a disruption of the peace and hence is bad for all involved. Hobbes’s enterprise is to allow people to pursue their own form of ‘commodious living’ to the extent they do not harm others. Freedom requires authority, but Hobbes is no authoritarian. If Hobbes is right about these things, we have to demand a certain form of public disposition from persons that want to be members of a society. As noted by
Hobbesian sovereignty and the spectre of tyranny 137 Samantha Frost, this means that we need a policy of exclusion if we are to have any form of politics at all: Hobbes is no theorist of universalism. In his theory there is exclusion, but the lack of political recognition afforded to ‘unintelligible’ subjects is a consequence not of what they are but of their refusal of the ethical posture. When Hobbes articulates the conditions of political subjectivity in this way, he pushes us to take seriously the contemporary theoretical insight that exclusion is a necessary condition and an unavoidable effect of the constitution of identity or of a political community. (2001, 50) Identity has to be publicly shaped and formed to some extent; Hobbes takes this tough bull by the horns and argues for exclusion, not on the grounds of gender, race, class or natural differences along the lines of Aristotle, but only for those who deny the very existence of politics. In Political Liberalism, John Rawls identifies three things that led to contemporary liberalism: the reformation, the rise of the modern state and the development of modern science. All three informed Hobbes’s thinking and it is perhaps not surprising that we can trace modern political thought back to a figure dealing with the direct effects of these three waves of modernity and particularly to the rising historical experience of diversity, disunity and competing world views. Hobbes’s brilliance was to be one of the first to recognize the turbulence that comes from such a shift, and to formulate a politics that could deal with the ensuing problems. He does not try and force one overarching notion of the good life and he accepts that people will disagree violently over conceptions of the good; consequently, we have to make society safe in the face of this pluralism. For Rawls, pluralism is a fact of society that necessarily stems from the human condition. Hobbes also recognized this to be one of the main features of his own time and was well aware that a political philosophy that could not deal with competing conceptions of the good would be incapable of maintaining peace. Given this, the task of the political philosopher is twofold: the first responsibility is to find a political concept of justice that all can accept; the second is to determine how people are to be motivated to act according to this concept. To perform the first task, Hobbes claims we have to wrestle authority away from the church and the universities, and place it in the hands of a rational political entity that will interpret the rules of justice found in the laws of nature. Hobbes is the first great egalitarian. He presents a political concept that is accessible to all,
138 Hobbesian sovereignty and the spectre of tyranny purely through the exercise of reason, and he suggests that all rational people have the capacity to act in a reasonable manner when others do the same. Hobbes artfully produces arguments that will persuade religious, moral and even selfish persons to accept the rule of law. The usual view is that Hobbes performs these tasks by instilling fear into the hearts and minds of those who populate his ideal commonwealth. Contemporary theorists on the other hand expend great effort attempting to defend non-coercive principles that promote a fair system of cooperation, foster reasonable pluralism and create rules that can be agreed upon (and hence are legitimate) by all members of the political community. The gap between modern political theorists and Hobbes, therefore, seems to be huge. I have argued instead that Hobbes places the focus on educating rather than coercing. The preservation of peace can only be achieved through a complex mix of public authority, and civic and moral education. The laws of nature aid us in this enterprise and can be viewed as a public philosophy of reasonableness. Hobbes knew perfectly well that fear alone cannot provide peace. This is not to say that fear is unimportant in his writing because it is the only thing that will keep some people from breaking the law. But what Hobbes has in mind is a healthy type of fear based on an understanding of the legitimate consequences of breaking the law. This is not a cringing type of fear and allows, rather than curtails, freedom for most members of the commonwealth. Hobbes used natural law to set out the concepts necessary for any fair system of cooperation. His mission is ‘not to dispute the laws of any government in special, that is, not to point which are the laws of any country, but to declare what the laws of all countries are’ (1972, 104). Such laws are universal and eternal but they are also rules concerned primarily with people living together in a cooperative manner in civil society. As Rawls suggests, cooperation is about ‘fair terms of cooperation that each participant may reasonably accept’ and which ‘specify an idea of reciprocity’ (1993, 16) that requires each individual to search for the point of view from which thinking of one’s own interests is the same as thinking about the interests of everyone. Hobbes also had something similar in mind. Each cooperating member of society has no more to do in learning the Lawes of Nature, but, when weighing the actions of other men with his own, they seem too heavy, to put them into the other part of the ballance, and his own in their place, that his own passions, and self-love, may adde nothing to the weight; and then there is none of these
Hobbesian sovereignty and the spectre of tyranny 139 Lawes of Nature that will not appear unto him very reasonable. (1968, 215) Judith Shklar rejects Hobbes, but their interest is the same: ‘Liberalism has one overriding aim: to secure the political conditions that are necessary for the exercise of personal freedom’ (1989, 21). Hobbes, therefore, would also surely agree with Shklar that ‘Systematic fear is the condition that makes freedom impossible’ (1989, 29). For both theorists there is no difficulty identifying a summum malum. Tarlton suggests that ‘if authority is necessarily as he described it, then maybe anarchy, disorganization and what Kenneth Burke used to call “the human barnyard” are really no worse’ (2002, 89). Hobbes notes that people often fall into the trap of describing political arguments they do not happen to like as tyrannical; I think Tarlton has done the same. I suggest that Hobbes’s intentions would be familiar to modern philosophers and that he should be allowed his rightful place as a founding figure of contemporary political thought. It has to be granted, however, that despite the many similarities mentioned above, modern liberalism has not taken on board Hobbes’s absolutist arguments. I will take up this issue shortly. What I hope is clear from this chapter is that arguments for absolutism are not by definition, and certainly not by intention, arguments for tyranny and that they might go a long way to solving some of the problems of instability in politics. I suggest that we should be wary of dismissing Hobbes’s views on liberty and the nature of political power, and that the ‘barnyard’ is still a lot better than foraging around in nature.
6
Democracy versus constitutionalism
Introduction In Chapter 5 I claimed that Hobbes’s argument for absolutism is logically coherent and that it better explains the nature of politics than the current liberal democratic alternative. In this chapter, I will extend the argument and claim that Hobbes’s brand of absolutism, when linked to democracy, is a better model for the exercise of political power than liberal constitutionalism. The reason that Hobbes’s argument about the nature of politics is more compelling is that sovereignty always has to reside outside written documents because these documents have to be modified and interpreted; all constitutions can be changed and those who change them must have more power than the document itself. There are mechanisms that make such alterations easy or difficult but all constitutions have provisions to accommodate change, and hence there is always some body that has sovereignty over and above the formal structures and norms that make up a constitution. In the United States three-fourths of states have to be in favour of a change to the constitution whereas the constitutions of Iceland, Israel, New Zealand, Sweden and the United Kingdom can be changed by a simple parliamentary majority. In Denmark, France, Ireland and Italy, a simple majority in a referendum is required, and most other democracies embrace the same principle but require larger majorities. Judicial review is nowhere as powerful as it is in the United Sates, and judges cannot rule a law unconstitutional in Belgium, Finland, Luxembourg, Netherlands, Switzerland, France, Israel, New Zealand and the United Kingdom. The Scandinavian countries and Iceland do allow for judicial review, but use it with extreme caution (Dahl 1998, 193–4). Hobbes is right to suggest that logically the buck has to stop somewhere, and wherever it stops, whether it is a king, parliament or the
Democracy versus constitutionalism 141 people, is where we find sovereignty. Formulating the issue in this way undermines one of the main arguments against absolute sovereignty, namely that power corrupts and absolute power corrupts absolutely. A Hobbesian analysis suggests we should not distract ourselves trying to ascertain whether one political system is more absolute than another; instead we should focus on identifying procedures that make it more or less difficult to exercise sovereignty. If political power resides with one individual it is much easier (but potentially more dangerous) for sovereignty to be exercised than if three-fourths of the people, in three-fourths of the electoral districts are required to agree to a decision. One of the crucial strategies for limiting political abuse, therefore, is not to argue against absolutism, but to argue for greater difficulty in utilizing the power that all sovereigns exercise absolutely. This is what we find in real politics – all constitutions can be changed, it is simply that it is easier to do this in the United Kingdom than it is in the United States, and it is particularly difficult to do so in Belgium. In Hobbes’s system, a democratic sovereign would find it more difficult than a king to exercise power in a tyrannical manner. These arguments can be used to support majority rule; if sovereignty poses a threat to rights, it is better that it rests with the majority because only a minority of people can possibly have their rights invaded; if political power is necessarily absolutist, we have a strong argument in favour of participatory democracy. I suggest that instead of trying to control power, or divide it up or make it safe, the best solution is simply to give it back to the people and trust them, because as Hobbes noted many years ago, if sovereigns are going to abuse power, there is little to be done about it at the constitutional level.
The fear of the majority The claims made above find opposition from supporters of constitutional checks and balances, separation of powers and bills of rights, who assume that institutional mechanisms can protect us from the abuse of political power, primarily because legislators and judges act as a bulwark against the tyranny of the majority. As noted in Chapter 1, this leads Mouffe to suggest that liberal forms of political communities are always in tension with the practice of majority rule. The fear of the majority has a well-documented history in the life and times of liberalism. James Madison went to extensive lengths to try and limit popular rule. The details of his argument are well known and do not need repeating here, but a brief examination of the view
142 Democracy versus constitutionalism of human nature that underpins his version of constitutionalism does not provide much in the way of optimism. Madison describes human beings as intolerant, vindictive, petty, full of animus, driven by self-love and ruled by passion rather than reason. Given the chance, we will ‘vex and oppress’ other individuals and the desire to form and join factions runs deep in our blood. Unfortunately, a faction is ‘a number of citizens who are united and actuated by some common impulse of passion, or of interest, adverse to the rights of other citizens, or to the permanent and aggregate interests of the community’ (1961, 78, my italics). It is part of our nature, according to Madison, to act in a manner that adversely affects the rights of individuals and the interests of the community. These same drives obviously inhabit those who hold office, and government is simply an expression of human nature writ large. It is no surprise, therefore, that Madison designed political institutions to promote conflict as a means of ensuring that ambitious men check and control one other. The result is a political system that is deliberately structured to result in deadlock, strife, infighting and disruption. This is where the fear of the majority leads us because the worst faction of them all is the majority faction, and the political system has to be structured in a manner that will thwart its ambitions. John Stuart Mill took a more parsimonious approach and claimed that he could use one ‘simple’ philosophical principle to control the majority. As Mill argued in On Liberty, a struggle always takes place between the competing demands of liberty and authority, and we cannot have the latter without the former: All that makes existence valuable to anyone depends on the enforcement of restraints upon the actions of other people. Some rules of conduct, therefore, must be imposed – by law in the first place, and by opinion on many things which are not fit subjects for the operation of law. (1978, 5) The ‘one very simple principle’ (1978, 9) is now usually referred to as the Harm Principle, and it states that ‘the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others’ (1978, 9). There is a great deal of debate about what Mill means by harm, but I will take him to mean that an action has to directly and in the first instance invade the rights of a person (Mill himself uses the term rights despite basing the arguments in the book on the principle of utility).
Democracy versus constitutionalism 143 Mill suggests that actions are of a self- or other-regarding nature. Self-regarding actions, whether or not they cause harm, require no intervention. The same applies to other-regarding actions that cause no harm. Other-regarding actions that cause harm are divided into the categories of legitimate and illegitimate harm and public and private harm. The main thrust of Mill’s argument in Chapter 1 of On Liberty is that only other-regarding actions that cause illegitimate harm in the public and private spheres are open to censure. The harm principle will check the majority precisely because it bars all interference with the individual unless that person is harming the rights of another person. Chapter 3 of On Liberty is an incredible assault on social censorship, expressed through the tyranny of the majority. Mill suggests that it produces stunted, pinched, hidebound and withered individuals; it limits originality and genius, snuffs out spontaneity, and leaves the masses with nothing else to do other than imitate their betters in an ape-like manner: ‘everyone lives as under the eye of a hostile and dreaded censorship [i]t does not occur to them to have any inclination except what is customary’ (1978, 58). He continues, the general tendency of things throughout the world is to render mediocrity the ascendant power among mankind at present individuals are lost in the crowd the only power deserving the name is that of masses [i]t does seem, however, that when the opinions of masses of merely average men are everywhere become or becoming the dominant power, the counterpoise and corrective to that tendency would be the more and more pronounced individuality of those who stand on the higher eminences of thought. (1978, 63–4) He says he is not complaining of any of this and is simply describing the low state of the human mind in nineteenth-century England. With these comments and many of a similar ilk, Mill demonstrates his distaste for the apathetic, fickle, tedious, frightening and dangerous majority. He suggests that engaging in eccentric behaviour is one way to hold off the normalizing tendencies of the masses. This is hardly the best solution I have heard for fighting against tyranny, and I doubt that performing a John Cleese–style funny walk is going to keep the mob at bay. Is the fear of majority tyranny as exemplified by Madison and Mill justified? Mill’s arguments are overstated. A political system that tyrannizes can be described as unjust, oppressive, despotic and cruel
144 Democracy versus constitutionalism and the type of social pressure that Mill describes hardly qualifies. I also suggest that Madison’s view of majoritarian politics is flawed and lean on Robert Dahl for support. As long ago as 1956, Dahl noted that constitutional mechanisms are of minor importance for the prevention of tyranny. He suggested that the key to preserving political liberties is a commitment to democratic practices and norms, and that these are, not surprisingly, more likely to flourish through repeated acts of political participation. It is worth asking oneself whether democracy is more likely to thrive in a society where everyone participates frequently, or in one where people only exercise their democratic capacities once every few years. It seems plausible to claim that democracy teaches democracy. Dahl suggests that once one learns to be democratic, one is likely to promote the equal participation of everyone as a valuable norm, and this realization can stem simply from reflecting on one’s own self-interest. Hence, it is rational to be a democrat. Despite my earlier criticisms of Cohen, he is right when he says, ‘The background conception of citizens as equals sets limits on permissible reasons that can figure within the deliberative process’ (1996, 106) (which, of course, requires the rejection of the universal domain requirement), and Cohen thinks that Rawls’s difference principle gets this about right. The upshot is that the democratic process itself rules out treating citizens unequally, and if minorities are abused then the system is, by definition, not democratic. Arguments of this type suggest that democracy and tyranny are logically incompatible. This does not mean, of course, that states that are described as democracies do not abuse people. The problem in such instances is how to stop political systems that are democracies from degenerating into regimes that do not treat citizens with respect. One answer is a constitution but it is unclear why a piece of paper should be able to do so much work. Arneson (1993) criticizes the idea that democracy and rights are necessarily compatible. He gives a protectionist defence of liberal democracy and describes it as a political system in which a constitution asserts certain rights that are protected by non-elected judges. He does accept that some rights are reinforced by political participation. For example, one can argue that the right to free speech is a logically necessary component of democratic politics. But he says that not all fundamental rights, such as the right to abortion, are so clearly linked to democratic participation; whether we do or do not allow abortion is independent of basic democratic rights in a way that free speech is not. Arneson suggests that abortion rights might, in fact, be
Democracy versus constitutionalism 145 upheld against the wishes of the majority in a constitutional system. There are many other rights, according to Arneson, that are crucial to a fulfilling life but that are not logically entailed by democratic participation. One response to this criticism is to argue that abortion can be linked to the right to participate. Raising a child can place huge burdens on a woman that might severely curtail her ability to actively participate in the political arena. A stronger response, however, is to reinforce the claim that the best way to decide whether rights and the principles of democracy are being ignored or promoted by decisions about abortion is through the democratic process. If we do not let the majority decide such controversial issues, we have no option but to place the decision in the hands of an elite minority. My defense of absolutist democratic sovereignty is not neutral and assumes that self-rule is a good. Once this is accepted, the procedure I advocate is a consequence of that good. This is an integral part of the argument that democracy is not in tension with rights. It is a necessary condition of democracy that citizens are self-governing and in order for this to happen, we need the space to engage in self-expression, free speech, free association, etc. Hence, majorities cannot strip these rights from a minority and still claim that the system is democratic. Democracy has limits built into it because of the requirement that people govern themselves, and if we are committed to the democratic process we also have to be committed to the means of self-government, and all the rights this entails. Of course, democracy can break down, but we will always have recourse to strong arguments to show why this is invalid. As Dahl notes, once we reach the point ‘when the democratic process can no longer be sustained in the face of a weak or hostile political culture, it strains credulity to believe that primary political rights will be preserved for long by courts or any other institution’ (1989, 173). This is the crucial point that Hobbes also makes. There are other factors beyond basic political rights that are necessary to make self-government possible, such as literacy, minimum living standards, health care, etc., and any democratic state will need significant welfare provisions to make sure that equal participation is upheld. Such a claim is again based on the logical structure of the argument for democracy. Dahl notes that taking democracy seriously means that nothing is outside the purview of the democratic process. Even if citizens wish to have a bill of rights, it will need to be agreed to by a majority vote and it can be changed by the same mechanism. Democratic bodies should have the capacity to limit
146 Democracy versus constitutionalism themselves to make sure they do not act hastily or ignore due process, but these restrictions have to be self-imposed and reversible. These mechanisms of restraint should not be extended to the requirement for supra-majorities. Majority rule maximizes the number of people exercising self-rule and supermajorities are simply a way of reintroducing minority rule. It is sometimes suggested that supermajorities are more democratic because they require a bigger majority. In practice, requiring a bigger majority simply strengthens the rule of the few and/or the status quo. Offering a democratic defence of rights provides a better justification than simply asserting that rights are universal. If we claim that the right of free speech is natural and inalienable, it becomes difficult to justify any limits to the right. However, if we ground rights within the framework of democracy, we can place limits on these rights in the name of protecting democracy. As I will argue later in the chapter, we can place certain limits on hate speech because it undermines the democratic principle of equal citizenship. Charles Taylor’s (1985) criticism of the atomistic view of rights provides useful support for the arguments put forward in this chapter. The traditional liberal defence of rights is to suggest that individuals are self-sufficient, possess rights prior to society, and form political communities to protect these rights. Such individuals have rights by nature, but they do not have any corresponding obligations which can only arise from consent. The individual, therefore, is chronologically and morally prior to society. This argument is supposed to be neutral regarding conceptions of the good and human nature. Taylor argues that such a position is logically incoherent because any list of rights has to depend on a prior conception of what is good for persons. Why, for example, do lists of rights usually include the right to free speech, assembly and religious worship, but not the right to play scrabble, eat Chinese food, and watch football? The answer is that the first list of rights identifies things that are seen as essential to human flourishing, which in turn requires some prior understanding of what is good for persons. Rights claims only make sense against a background understanding of capacities that ought to command our respect. Such rights can only be promoted and fostered in certain types of society and consequently it only makes sense to talk of the right to free speech within a social context. Traditionally, this has been a democratic context. The practice of democracy means that we have to engage with others and listen to their point of view if we wish to reach agreement with them. This leads to a society in which individuals are less likely to be abused. The empirical evidence supports this and
Democracy versus constitutionalism 147 suggests that freedoms and rights are best protected in democracies. Despite this, constitutionalists expect us to believe that the heavy hand of tyranny will be upon us if we extend the reach of democracy beyond its current practice. Arneson’s argument turns out to be embedded in the usual paranoia about the tyranny of the majority. Discussing Marx’s writings on the Paris Commune, he says that ‘[t]o modern readers educated on Madison and De Tocqueville, Marx’s constitutional thinking might seem naïve and simplistic. Marx altogether ignores the dangers of majority tyranny that are part and parcel of the unbridled majoritarianism he favours’ (1993, 131). While Arneson says that such a reading misses some of the nuance of Marx’s position, he shares with these modern readers the same fear of majority tyranny. His alternative suggestion is guardianship, i.e. that a few wellchosen individuals are capable of making decisions for the majority. J.S. Mill thought that the better educated should be given more influence through the mechanism of plural voting – everyone should have a voice, but not an equal voice. Mill did not find much support for this view, but Arneson argues that plural voting is a close bedfellow of judicial review. They are both brands of platonic guardianship because both suggest a greater say for those with more education and wisdom. He does not shy away from the elitist aspect of these claims and points out that if one is opposed to Mill’s plural voting method for the better educated, then one should also be opposed to judges being the final arbiter: ‘Judicial supremacy is just plural voting by other means’ (1993, 135). We can presumably switch this around and argue that if one is opposed to plural voting then one should also be opposed to judges acting as the final arbiter. Dahl (1989) notes that guardians require two things: a knowledge of the common good and the technical ability or science to attain it. It is assumed that only a small subset of the population has this information. If one wishes to abandon the claim that some individuals have more access to the truth than others, one can still support guardianship by arguing that a small elite should rule because they are more virtuous. This position has its supporters among Platonists and Aristotelians, but it is a difficult case to make given the current widespread distrust of those professionally engaged in politics. The desire for judicial protection against the majority also has its roots in romanticism’s critique of the rational idealism that led to tyranny after the French Revolution. Burke thought that democratic reformers meddled with the natural order of things and undermined the requirements for sociability by letting loose passions that need to
148 Democracy versus constitutionalism be restrained. Burke suggests that it is better to rely on tradition to guide politics and morality: [w]e know that we have made no discoveries; and we think that no discoveries are to be made in morality; nor many in the great principles of government, which were understood long before we were born, altogether as well as they will be after the grave has heaped its mould upon our presumption. (Burke 1968, 182) Society is naturally hierarchically organized and this is incompatible with the levelling impetus of democratic reform. Society is too complex to be structured according to the dictates of rationalists whose egalitarian philosophising can undermine hundreds of years of political practice. Natural rights, metaphysics, pure reason and social engineering are terms to be avoided. Reason has its place, and that place is to play second fiddle to pragmatism, emotion, habit, mystery, chivalry and nobility. Abstract reason should particularly be avoided when it begins to undermine our traditions and institutions. Society cannot be reduced to a self-interested contractual agreement; rather, it is ‘a partnership in all art; a partnership in every virtue, and in all perfection. A partnership not only between those who are living, but those who are living, those who are dead, and those who are to be born’ (1968, 194–5). One could respond to Burke by arguing for a more gradual introduction of democratic equality but this would be deemed inadequate because he thought that democracy would simply release the desire to dominate in a new group of people who would be more dangerous than the older elite trained in ideas of noblesse oblige: [n]othing is more certain, than that our manners, our civilization, and all the good things which are connected with manners, and with civilization, have, in this European world of ours, depended for ages upon two principles I mean the spirit of a gentleman, and the spirit or religion. The nobility and the clergy, the one by profession and the other by patronage, kept learning in existence [h]appy if they had all continued to know their indissoluble union, and their proper place! Along with its natural protectors and guardians, learning will be cast into the mire, and trodden down under the hoofs of a swinish multitude. (1968, 173) Burke was undoubtedly right to see that democracy will have a hard time grafting onto societies in which more traditional views are
Democracy versus constitutionalism 149 dominant. But once it has been grafted on, its fundamental principle of equality should do away with the more traditional view of guardianship that he advocated. Democratic equality is now part of our fabric; we no longer accept Plato’s arguments for philosopher rulers or Burke’s cries for an educated elite to rule, and nor should we accept similar claims for judges. How are we to know who the best elite will be? A problem faces all elite theorists: if the masses are to be trusted with choosing the best leaders (elite theorists are usually but not always liberal democrats of some persuasion) then presumably the masses also have the intellect to listen to elites and decide to accept or ignore their advice without giving them the sorts of political privileges that accompany legislative and judicial bodies. Many theorists dealing with the tyranny of the majority make the same mistake, which is to equate being in a mob with being in a democracy. They are not the same and democracy is more likely to produce individualism than collective despotism precisely because it is founded on, and practised according to, the principles of personal autonomy. One of the major drivers of the fear of majority is the concern that people will be coerced to follow the general will. However, in a pluralist society populated by autonomous individuals there is no desire for, or likelihood of, a general will, and hence there is no impulse to force people to accept it. In contemporary democratic societies, the general will means nothing other than what the majority decides. As Dahl has noted, the majority is often a shifting, temporary and changing alliance that hardly fits with the image of the unchanging homogenous mass described by some anti-democrats. The evidence is pretty clear that democracies do not harm their citizens as much as elitist forms of government. It is also the case that there are more avenues for the free expression of individuality in democracies than in any other form of political organization. Political systems of any persuasion have to fail if they do not pay sufficient attention to civic culture, education and the inculcation of political virtues. Liberals often claim to avoid such tricky issues because they put in place constitutional checks that prevent the majority abusing individuals or minority groups. Kant attempted to persuade us that a race of devils can be controlled if we only put in place the correct political institutions. This, however, is a flimsy defence against the abuse of political power because a constitution is only a piece of paper that reflects a particular philosophy and has no actual coercive powers. This is why the Soviet constitution offered little protection to persons who wanted to leave the country, or express certain opinions, or practise certain religions. It is also why
150 Democracy versus constitutionalism the American version did not protect slaves, or women, or communists, or Japanese Americans at certain points in history. If the people or their political leaders have tyrannical leanings, a constitution will not stop them from engaging in tyrannical practices. Hence, every stable and peaceful society, liberal or otherwise, has to be populated by persons of a certain disposition; without this, no amount of rights and liberties, carefully crafted, written down and formulated, will provide adequate protection. Stephen Holmes (1999) suggests that civil society is about nonpredatory relations among strangers. One of the major problems facing post-communist Europe is how to produce even this minimal requirement when the state is weak. An atrophied state might not be very good at tyrannizing, but it is also not very good at providing stability and order. A strong state is a prerequisite for the exercise of rights and the more rights claims there are (and today we claim them to all sorts of things) the stronger the state has to be in order to adequately enforce these rights. Hence, rights claims are a doubleedged sword: the more we claim, the stronger the state has to be and the more intrusive it becomes. The excessive rights claims that characterize modern liberalism might be more perilous than the supposed fickle majority. Dahl argues that no constitutional design will work in unfavourable conditions and that a variety of constitutions can work when conditions are favourable. This supports the point, pressed in this book, that the crucial ingredient for a safe political system is not constitutional documents but a democratic political culture. Dahl lists the following as the essential conditions for democracy: control of the military and police by politicians, a democratic political culture, i.e. a belief in democracy as a viable form of government, no foreign influence that is hostile to democracy, a market economy and weak sub-cultural pluralism. It is interesting to note that Dahl does not place a constitution, a bill of rights or the separation of powers as essential conditions of democracy. If we are going to have a constitution, Albert Weale (1999) is probably close to the mark when he suggests that it should be used to codify the democratic rules of the game rather than place limits on the democratic process. Such a constitution is useful because it prevents temporary holders of political power from abusing the rights of citizens and from using it for their own advantage. Even if we grant this, a further problem arrises when we consider how constitutions are interpreted. Some scholars suggest that the wording of constitutions is left deliberately vague to allow for a variety of interpretations. Others recommend that we can find an objective
Democracy versus constitutionalism 151 original intent in the document. The latter point of view suggests, for example, that the equal protection clause in the US Constitution should be understood in terms of those who wrote it, not in our own terms. This line of argument leads to the unfortunate consequence that bad decisions made by judges will be reinforced over time. It also suggests that the wording of constitutions is exact and does not require further elucidation; we are tied forever to the ideas of a few people at a particular point in history. This view is not persuasive, as the first amendment of the American Constitution makes clear. Among other things, this amendment states that ‘Congress shall make no law abridging the freedom of speech’. It is difficult to find a more parsimonious statement than this one. It does not say that ‘Congress shall occasionally abridge free speech’ or ‘Congress shall make no more than six laws prohibiting speech’. It states quite clearly and unambiguously that Congress shall make no law abridging freedom of speech – not one. Yet something that is so clearly stated has become a huge battleground for various interpretations. This is just as well because the First Amendment as it is stated is an untenable position to hold on the issue of free speech. In the next section, I will examine this issue in some detail. If we can highlight the complexity of the topic it might become more apparent why it cannot be settled with simplistic constitutional statements.
Free speech Free speech becomes a volatile issue only when it is highly valued because only then do the limitations placed upon it become controversial. The first thing to note in any discussion of freedom of speech is that it does have to be limited. No society has yet existed where speech has not been interfered with to some extent. This should make us wary of those (constitutionalists in particular) who warn of the dangers of the ‘slippery slope’. The slippery slope argument is that we should not place limits on free speech or we will slide down the slope that heads in a one-way direction from liberty to tyranny. Such arguments assume that we can be on or off the slope. In fact, no such choice exists: we are necessarily on the slope whether we like it or not, and the task is always to decide how far up or down we choose to go, not whether we should step off altogether. The second thing to note when discussing freedom of speech is that it can never be fully extinguished. Most forms of free action differ from speech; for example, if the government wishes to stop citizens riding motorbikes it can limit the freedom to do so by making sure that
152 Democracy versus constitutionalism such vehicles are no longer available; current bikes could be destroyed and a ban can be placed on imports. Freedom of speech is a different case. A government cannot make it impossible to say certain things. The only thing it can do is punish people after they have spoken. The threat of punishment can have a chilling effect but it cannot stop those who are determined to speak. Given that we have already discussed John Stuart Mill, who presented one of the first and perhaps still the most famous liberal defence of free speech, I will focus on his claims and use them as a springboard for a more general discussion of free expression. In the footnote at the beginning of Chapter 2 of On Liberty, Mill makes a very bold statement: ‘If the arguments of the present chapter are of any validity, there ought to exist the fullest liberty of professing and discussing, as a matter of ethical conviction, any doctrine, however immoral it may be considered’ (1978, 15). This is an incredibly strong defence of free speech; Mill tells us that any doctrine should be allowed the light of day no matter how immoral it may seem to everyone else. And Mill does mean everyone: ‘If all mankind minus one were of one opinion, and only one person were of the contrary opinion, mankind would be no more justified in silencing that one person than he, if he had the power, would be justified in silencing mankind’ (1978, 16). Liberty of speech should exist with every subject matter, such that we have ‘absolute freedom of opinion and sentiment on all subjects, practical or speculative, scientific, moral or theological’ (1978, 11). Mill claims that we need the fullest liberty of expression to push our arguments to their logical limit, not to the limit of social embarrassment. This is a very powerful argument in favour of freedom of speech and at first glance seems to rival the First Amendment in its scope. But Mill’s argument is more sensible and as I already noted above, he suggests that we need some rules of conduct that regulate the actions and words of members of a political community. The limitation he places on free expression is the previously discussed harm principle. The limits imposed on free speech by this principle appear to be few because it is difficult to support the claim that most speech actually causes illegitimate harm to the rights of others. If we accept for a moment that the harm principle is the appropriate instrument for governing social and political intervention we need to ask what types of speech, if any, cause harm; once we can answer this question we have found the correct limits to free expression. Mill uses the example of speech related to corn dealers; he suggests that it is fine to claim that corn dealers starve the poor if such a view is expressed through
Democracy versus constitutionalism 153 the medium of the printed page, but that it is not permissible to promulgate the same view to an angry mob, ready to explode, that has gathered outside the house of the corn dealer. The difference between the two situations is that the latter is an expression ‘such as to constitute a positive instigation to some mischievous act’ (1978, 53), namely to place the rights, and possibly the life, of the corn dealer in danger. It is important to remember that Mill will not sanction limits to free speech simply because someone is harmed by the statements of others. For example, the corn dealer may suffer severe financial hardship if he is accused through the printed word of starving the poor. Mill distinguishes between legitimate and illegitimate harm, and it is only when speech causes a direct and clear violation of rights that it can be limited. Other examples where the harm principle may apply include libel laws, blackmail, advertising blatant untruths about commercial products, advertising dangerous products to children (e.g. cigarettes), and securing truth in contracts. In most of these cases, it is possible to make an argument that harm has been committed and that rights have been violated. There are other instances when the principle has been invoked but where it is more difficult to demonstrate that harm, in the appropriate sense, has occurred. Perhaps the most obvious example of this is the debate over pornography. As Feinberg notes (1985), most attacks on pornography up to the 1970s were from social conservatives who found such material to be immoral and obscene. (Feinberg states that there is no necessary link between pornography and obscenity; pornography is material that is intended to cause sexual arousal, whereas something is obscene when it causes repugnance, revulsion and shock. Pornography can be, but is not necessarily, obscene.) In recent times, the cause has been joined by some feminists who have avoided the claim that depictions of sexual acts are immoral; instead they have maintained that pornography degrades, harms and endangers the lives of women. This argument, to have force, must distinguish between erotica (as a sub-class of pornography aimed at sexual arousal) and pornography that threatens rights by depicting acts of violent abuse against women. If it can be demonstrated that this latter material significantly increases the risk that men will commit acts of physical violence against women, the harm principle can legitimately be invoked. When pornography involves young children, most people will accept that it should be prohibited because of the harm that is being done to persons under the age of consent. It has proved much more difficult to make the same claim for consenting adults. It is not easy to
154 Democracy versus constitutionalism demonstrate that the actual people who appear in books, magazines, films, videos and on the Internet are being physically harmed, and it is even more difficult to demonstrate this is the case for women as a whole. Very few people would deny that violence against women is an abhorrent feature of our society, but how much of this is caused by violent pornography? One would have to show that a man who would not otherwise have raped or beaten females was caused to do so through exposure to material depicting violence to women. It has proven very difficult to draw such a conclusive causal relationship. If pornographers were exhorting their readers to commit violence and rape, the case for prohibition would be much stronger, but they do not do this, just as films that depict murder do not actively incite the audience to mimic what they see on the screen. Remember that Mill’s formulation of the harm principle suggests only speech that directly harms the rights of others should be banned; finding such material offensive, obscene or outrageous is not sufficient grounds for prohibition. Overall, it seems very difficult to mount a compelling case for banning pornography (except in the case of minors) based on the concept of harm as formulated by Mill. Another difficult case is hate speech. Most European liberal democracies have limitations on hate speech, but it is debatable whether these can be justified by the harm principle. One would have to show that such speech violated rights, directly and in the first instance. A famous example of hate speech was the proposed Nazi march through Skokie, Illinois. In fact, the intention was not to engage in political speech at all, but simply to march through a predominantly Jewish community dressed in storm trooper uniforms and wearing swastikas (although the Illinois Supreme Court interpreted the wearing of swastikas as ‘symbolic political speech’). It is clear that most people, especially those who lived in Skokie, were outraged and offended by the thought of the march (which never eventuated) but were they harmed? There was no plan to cause physical injury and the marchers did not intend to damage property. The main argument against allowing the march, based on the harm principle, was that it would cause harm by inciting opponents of the march to riot. The problem with this line of argument is that it is the harm that could potentially be done to the people speaking that becomes the focal point of the prohibition and not the harm done to those who are the subject of the hate. To ban speech for this reason, i.e. for the good of the speaker, tends to undermine the basic right to free speech in the first place. It is possible to suggest that persons on the wrong end of hate speech are psychologically harmed, but this is more difficult to demonstrate
Democracy versus constitutionalism 155 than harm to a person’s legal rights. It seems, therefore, that if we are to base our defence of speech on the harm principle, we are going to have very few sanctions imposed on the spoken and written word. It is only when we can show direct harm to rights, which will almost always mean when an attack is made against a specific individual or a small group of persons, that it is legitimate to impose a sanction. This is the position staked out by Mill in the first two chapters of On Liberty and it is a good starting point for a discussion of free speech because it is hard to imagine a more liberal position. It becomes very difficult to defend free speech once it can be demonstrated that its practice does actually invade the rights of others. It is, however, too restrictive if we take the First Amendment seriously because it states that Congress shall make no law abridging free speech which would rule out the harm principle. This view is not often expressed because, as already noted, most people think that free speech should be limited if it does cause harm. We have to conclude, therefore, that most Americans do not take the first Amendment seriously. George Kateb (1996), however, is an exception and he has made an interesting argument that runs as follows. If we want to limit speech because of the harm it causes, we will have no alternative other than to ban a lot of political speech. Kateb claims that most political speech is useless, that a lot of it is offensive, and that some of it causes harm because it is deceitful, and because it is aimed at discrediting specific groups. It also undermines democratic citizenship and stirs up nationalism and jingoism, which results in harm to citizens of other countries. Even worse than political discourse, according to Kateb, is religious speech; he claims that a lot of religious speech is hateful, useless, dishonest, and ferments war, bigotry and fundamentalism. It also creates a bad self-image for many people and causes feelings of guilt that can haunt a person throughout his or her life. Pornography or hate speech, he claims, cause nowhere near as much harm as political and religious speech. His conclusion is that the harm principle casts its net too far and we should allow almost unlimited speech. This is an intriguing argument and seems to take the sentiment of the First Amendment seriously, but there is a major problem with the analysis. The harm principle would actually allow religious and political speech for the same reasons that it disallows limits on pornography and hate speech, namely that it is not possible to show that such speech does cause harm. It is very doubtful that Mill would have supported using his arguments about harm to ban political and religious speech. However, if one accepts that Kateb is right and it can be shown that such speech does cause harm in the sense of violating
156 Democracy versus constitutionalism rights, the correct response is surely to start limiting political and religious speech. If Kateb’s argument is sound, he has shown that harm is more extensive than we might have thought; he has not demonstrated that the harm principle is invalid. It is precisely because we want to prevent unjustified harm that the blanket statements put forward in the First Amendment and by Kateb seem so strange and unconvincing. There are two basic responses to the harm principle as a means of limiting speech. The one already discussed is that it is too broad; the other is that it is too narrow and needs to be stretched to allow greater interference with speech acts. One of the most impressive arguments for this position comes from Joel Feinberg, who suggests that the harm principle cannot shoulder all of the work necessary for a principle dealing with all the nuances of free speech. He argues that we also need an offence principle that can act as a guide to public censure. The basic idea is that the harm principle sets the bar too high for legal interference and that we should prohibit some forms of expression because they are very offensive. Causing offense is less serious than harming someone, so the penalties imposed should be less than those for acts that cause harm. As Feinberg notes, however, this has not always been the case and he cites a number of instances in America where penalties for sodomy and consensual incest between adults have ranged from twenty years’ imprisonment to the death penalty. These are victimless crimes and hence the punishment has to have a basis in the supposed offensiveness of the behaviour rather than the harm that is caused. An offence principle is difficult to apply given that many people take offence because of an overly sensitive disposition, or worse, because of unjustified prejudice. Despite the difficulty of applying a standard of this kind, something like the offence principle operates widely in liberal democracies where citizens are penalized for a variety of activities, including speech, that would escape prosecution under the harm principle. Wandering around the local shopping mall naked or engaging in sexual acts in public places are two obvious examples. Feinberg suggests that a variety of factors need to be taken into account when deciding whether speech can be limited by the offence principle and these include the extent, the duration and the social value of the speech, the ease with which it can be avoided, the motives of the speaker, the number of people offended, the intensity of the offense and the general interest of the community at large. How would the offence principle help us deal with the issue of pornography? Given the criteria just mentioned, Feinberg argues that books should never be banned for reasons of offence; if one has freely decided
Democracy versus constitutionalism 157 to read the book for pleasure, the offence principle obviously does not apply, and if one does not want to read it, it is easily avoidable. A similar argument could be applied to pornographic films. The French film Bais-Moi was in essence banned in Australia in 2002 because it contained material that was deemed to be offensive (it was denied a rating which meant that it could not be shown in cinemas). It would seem that the offence principle outlined by Feinberg would not permit such prohibition because it is very easy to avoid being offended by the film. It should also be legal to advertise the film, but some limits could be placed on the content of the advertisement so that sexually explicit material is not placed on billboards in public places (because these are not easily avoidable). At first glance, it might seem strange to have a more stringent speech code for advertisements than for the thing being advertised; the harm principle would not provide the grounds for such a distinction, but it is a logical conclusion of the offence principle. What of pornography that is extremely offensive because of its violent content? In this case, the offence is more profound and simply knowing that such films exist is enough to deeply offend many people. The difficulty here is that bare knowledge, i.e. being offended by merely knowing that something exists or is taking place, is not as serious as being offended by something that one does not like and that one cannot escape. If we allow that films should be banned because some people are offended by them, even when they do not have to view them, logical consistency demands that we allow the possibility of prohibiting many forms of expression. Lots of people find strong attacks on religion, or television shows hosted by religious fundamentalists, deeply offensive. Feinberg argues that even though some forms of pornography are profoundly offensive to a lot of people, they should still be permitted. Hate speech causes profound and personal offence. The discomfort that is caused to those who are the object of such attacks cannot easily be shrugged off. As in the case of violent pornography, the offence that is caused by Nazis marching through Jewish neighbourhoods cannot be eluded simply by staying off the streets because the offence is taken over the bare knowledge that the march is taking place. As we have seen, however, bare knowledge does not seem sufficient grounds for prohibition. If we examine some of the other factors regarding offensive speech mentioned above, Feinberg suggests that the march through Skokie does not do very well: the social value of the speech seems to be marginal, the number of people offended will be large and it is difficult to see how it is in the interests of the community. These reasons, however, also hold for violent pornography.
158 Democracy versus constitutionalism The key difference is in the intensity of the offence; it is particularly acute with hate speech because it is aimed at a specific audience. The motivation of the speakers in the Skokie example seemed to be to incite fear and hatred and to directly insult the members of the community with Nazi symbols. Nor, according to Feinberg, was there any political content to the speech. The distinction between violent pornography and this specific example of hate speech is that a particular group of people were targeted and the message of hate was paraded in such a way that it could not be easily avoided. Feinberg also claims that when fighting words are used to provoke people who are prevented by law from using a fighting response, the offence is profound enough to allow for prohibition. If pornographers engaged in the same behaviour by parading through neighbourhoods where they were likely to meet great resistance and cause profound offence, they too should be prevented from doing so. It is clear, therefore, that the crucial component of the offence principle is whether the offensive material can be avoided. For the argument to be consistent, it must follow that many forms of hate speech should still be allowed if the offence is easily avoidable. Nazis can still meet in private places, or even in public ones that are easily bypassed. Advertisements for such meetings can be edited (because they are less easy to avoid) but should not be banned. Very few liberals take the Millian view that only speech causing direct harm should be prohibited; most support some form of the offense principle. As noted in the Skokie example, some are willing to extend the realm of state interference further and suggest that speech can be limited to prevent harm being done to the speaker. The argument here is that the agent might not have a full grasp of the consequences of the action involved (whether it is a speech act or some other form of behaviour) and hence can be prevented from engaging in the act. Most liberals are wary of such arguments because we are now entering the realm of paternalistic intervention where it is assumed that the state knows better than the individual what is in his or her best interests. Despite this, it is probably safe to say that not many people would support a complete ban on paternalistic intervention. Mill, for example, is an opponent of paternalism generally, but he does believe there are certain instances when intervention is warranted. He suggests that a public official can prevent a person crossing a bridge if he is certain that it will collapse. If, however, there is only a danger that it will collapse the person can be warned but not coerced. The decision here seems to depend on the likelihood of personal injury; the more certain it becomes, the more legitimate the intervention.
Democracy versus constitutionalism 159 Prohibiting freedom of speech on these grounds is very questionable (it was not persuasive in the Skokie case) because it is very rare that speech would produce such a clear danger to the individual. We have examined some of the options that are open to us for limiting free speech. It would be difficult to class oneself as a strong liberal if one were willing to stray a great deal further into the arena of state intervention than already discussed. Liberals like Kateb have many more reasons for supporting free speech than I have mentioned but they tend to be united in opposing paternalistic and moralistic justifications for limiting free expression. They have a strong presumption in favour of individual liberty because, they argue, this is the only way that the autonomy of the individual can be respected. To wish to prohibit speech for reasons other than those already mentioned means that one has to make an argument that it is permissible to limit speech because of its unsavoury content, or as Feinberg puts it, one has to be willing to say that ‘[i]t can be morally legitimate for the state, by means of the criminal law, to prohibit certain types of action that cause neither harm nor offense to any one, on the grounds that such actions constitute or cause evils of other kinds’ (1985, 3). Acts can be evil because they are, amongst other things, dangerous to a traditional way of life, because they are immoral or because they hinder the perfectibility of the human race. However, a strong claim can be made that hate speech should be banned, even if it does not cause harm or unavoidable offence, without having to rely on conservative principles regarding tradition or morality. The reason it should be banned is that it is unacceptable (and anti-democratic) to brand some citizens as inferior to others on the grounds of race or sexual orientation. The same applies to pornography; it should be prevented because it is wrong to portray women as sexual objects who are fair game for violent mistreatment. To make this case, one has to dilute one’s support for freedom of expression in favour of other values, such as equal respect for all persons. Rae Langton (1990), for example, has argued that democracy is undermined when one group uses speech acts of a hateful or pornographic nature to deny equal respect for other groups. Censorship is permissible if it can be demonstrated that such material is inconsistent with the underlying principles of democracy which demand that no group should be treated as inferior to others on the grounds of race or sex. Langton is careful to distinguish between erotica and pornography, and the latter is defined as ‘a graphic sexually explicit subordination of women through pictures or words’ (1990, 346). She concludes that ‘women as a group have rights against the consumers of pornography
160 Democracy versus constitutionalism and thereby have rights that are trumps against the policy of permitting pornography the permissive policy is in conflict with the principle of equal concern and respect, and that women accordingly have rights against it’ (1990, 346). If we are concerned about tyranny, a carte blanche right to say whatever one wants is more dangerous to the rights of minorities than majority rule. Because Langton is not basing her argument on the harm principle, she does not have to show that women are harmed by pornography. For the argument to be convincing, however, she does have to demonstrate that pornography does not treat women with equal concern and respect. One also has to agree that treating people with equal concern is more important than unlicensed freedom of expression. The arguments presented in this book would be compatible with limiting some forms of speech that violate the democratic principle of treating all citizens with equal respect. We began this examination of free speech with John Stuart Mill; let us end with him and assess whether he remains true to his idea of the harm principle and whether his arguments in favour of free speech are persuasive. Mill suggests that we need to distinguish between legal sanction and social disapprobation as a means of limiting speech. As already noted, the latter does not ban speech but it makes it more uncomfortable to utter unpopular statements. Mill does not seem to support the imposition of legal penalties unless they are sanctioned by the harm principle. As one would expect, he is very concerned about the use of social pressure as a means of limiting speech. It is quite a surprise, therefore, to find that Mill also seems to embrace a fairly encompassing offence principle when the sanction does involve social disapprobation: ‘Again, there are many acts which, being directly injurious only to the agents themselves, ought not to be legally interdicted, but which, if done publicly, are a violation of good manners and, coming thus within the category of offenses against others, may rightly be prohibited’ (1978, 97). Similarly, he states that ‘The liberty of the individual must be thus far limited; he must not make himself a nuisance’ (1978, 53). In the latter parts of On Liberty, Mill also suggests that distasteful characters can be held in contempt, that we can avoid such persons (as long as we do not parade it), that we can warn others against the person, and that we can persuade, cajole and remonstrate with those we deem offensive. These actions are legitimate as the free expression of those people who happen to be offended as long as they are done as a spontaneous response to the other person’s faults and not as a form of punishment. But those who exhibit cruelty, malice, envy, insincerity, resentment and crass egoism are open to the
Democracy versus constitutionalism 161 greater sanction of disapprobation as a form of punishment, because these faults are wicked and other-regarding. It may be true that these faults have an impact on others, but it is difficult to see how acting according to malice, envy or resentment necessarily violates the rights of others. The only way that Mill can make such claims is by expanding his argument to include an offence principle and hence by giving up on the harm principle as the only legitimate grounds for interference with behaviour. Overall, Mill’s arguments about ostracism and disapprobation seem to provide little protection for the individual who may have spoken in a non-harmful manner but who has nevertheless offended the sensibilities of the masses. The concern that Mill expressed about the tyranny of the majority seems to dissipate by the end of On Liberty. It is still the case that Mill provides one of the great arguments for individual freedom in the liberal tradition. Why does he value liberty so highly? He defends freedom generally, and free speech in particular, for a variety of reasons; it fosters authenticity, genius, creativity, individuality and human flourishing. More specifically he tells us that silenced opinions may be true, or contain a portion of the truth, and that unchallenged opinions become mere prejudices and dead dogmas that are inherited rather than adopted. These are empirical claims for which Mill provides very little evidence (neither do Mill, Madison and Tocqueville provide much in the way of evidence to support their claims about the tyranny of the majority). How much is truth promoted by free speech? Is it likely that we enhance the cause of truth by allowing hate speech or violent and degrading forms of pornography? What is the relationship between speech and truth? What would the graph look like if one axis was truth and the other was free speech? Do we get one extra unit of truth for every extra unit of free speech? How do we know that ideas degenerate into prejudice if they are not constantly challenged? Is playing the Devil’s advocate useful or tedious? Mill asserts many things in On Liberty but provides little evidence or argument to back them up. In conclusion, we have found that the harm principle provides legitimate reasons for limiting free speech when doing so prevents direct harm to rights. Taken on its own, the principle allows speech acts to be largely unmolested. The offence principle has a wider reach than the harm principle, but it still recommends very limited intervention in the realm of free speech. All forms of speech that are found to be offensive but easily avoidable should go unpunished. This means that all forms of pornography and most forms of hate speech will escape punishment. If this argument is acceptable, it seems only logical that
162 Democracy versus constitutionalism we should extend it to other forms of behaviour. As noted, public nudity is prohibited because it is offensive to some people, but most of us find it at most a bit embarrassing, and it is avoided by a simple turn of the head. The same goes with nudity and coarse language on television. Neither the harm nor the offence principles support criminalizing bigamy or drug use, or the enforcement of seat belts, crash helmets and the like. To extend prohibitions on speech and other actions beyond the harm and offence principle requires an argument for permitting the state to decide what is necessary for the equal treatment of members of the community, and what is desirable for the safety of citizens. It has certainly been the practice of most societies, even liberal-democratic ones, to impose restrictions of this kind and to limit speech because it causes profound but avoidable offence. As we have seen, even Mill seems to back away from the harm principle. Hence, the freedom of expression supported by the harm principle outlined in Chapter 1 of On Liberty and by Feinberg’s offence principle is still a possibility rather than a reality. It is up to the reader to decide if it is a possibility that sounds appealing. I am not persuaded and have suggested that Langton’s argument for limiting speech that undermines democratic equality is worthy of support. I will not provide a detailed argument in support of this position because this brief examination of free speech is simply meant to highlight the point that constitutional statements such as the First Amendment (which is one sentence long) do little to help us comprehend a complicated and nuanced issue of public policy.
Back to constitutions A problem with constitutions is that they often reify very bad arguments; the second amendment to the American Constitution comes to mind at this point. To combat this problem, it is sometimes suggested (as I noted earlier in the chapter) that the wording of constitutions should be left somewhat imprecise in order to foster lively debate. In this case, however, the minority who are interpreting the text seem to be even less answerable for the decisions they make. A small group of judges can override elected officials, or in a more democratic system, the direct wishes of the people. Unfortunately, this turns a small unrepresentative elite into an absolute sovereign. The US Supreme Court is fairly unique in this regard. As Dahl notes, other polyarchies have supreme courts, but they are not as powerful or used in the same way as the American version, and most rely on the democratic process rather than the US style of quasi guardianship. In half of
Democracy versus constitutionalism 163 the polyarchies identified by Dahl, the judiciary cannot strike down legislation as unconstitutional. These countries do not seem to have a worser human rights record than America. As Dahl notes, quasi guardianship might actually make the people less diligent because they assume that elites will do the heavy lifting when it comes to protecting individuals – hence it might have a negative impact on the protection of rights. Certainly, the history of judicial review in the United States does not suggest that Supreme Court judges are the guardians of the downtrodden and abused. A common response to those who argue for judicial interpretation is that constitutions are set in stone and are above manipulation. However, as Ronald Dworkin (1999) has noted, this depends on the false belief that constitutions do not need to be interpreted. To think of them in this light is to make the same mistake as the strict originalist discussed earlier. The truth of the matter, according to Dworkin, is that moral interpretation is at the forefront of constitutional law because constitutions are simply a collection of abstract moral principles. There is no objective interpretation of a constitutional principle that everyone will agree upon and hence it is not possible to ‘keep faith with the text’. This is what Dworkin calls the ‘moral reading’ of the constitution, something he claims that lawyers and judges already utilize in their everyday practice. If judges did not apply the moral reading it would not be possible to classify some of them as liberals and others as conservatives. There would be no fierce battles over nominations of Supreme Court justices if they simply provided an objective interpretation of the constitution. If Dworkin is correct, it seems that even the strict constitutionalism of the United States allows no escape from the human element. How else, Dworkin asks, are we to understand the fourteenth amendment to the Constitution that demands people should be treated equally? There is no agreed upon definition of equality and it has to be interpreted as an abstract moral principle by judges. He refers to judges as writing chapters in a novel that make sense in relation to previous and current chapters. Judges have some wriggle room but they are also constrained by this historical narrative: ‘[o]ur constitution is law, and like all law it is anchored in history, practice and integrity’ (1999, 90). Hence the moral reading, often accused of being revolutionary in practice, simply describes how judges do in fact behave. Dworkin does claim, however, that it would be revolutionary if judges admitted that this is how they work because the moral reading is never openly endorsed as a method of decision-making. Instead, judges claim to be acting in a morally neutral manner.
164 Democracy versus constitutionalism Dworkin claims that judges will not admit to the moral reading because it seems to suggest that the law will often be a reflection of the moral opinions of the judiciary at any particular point in time. More importantly, from the point of view of this book, he suggests that it also seems to undermine the sovereignty that is supposed to lie with the people. The moral reading argument is unpopular because it has elitist and anti-democratic connotations. This seems to be a peculiar explanation because he is suggesting that an institution that is meant to check the majority as one of its functions is criticized because it is anti-democratic. This line of argument arises, Dworkin says, because the concept of democracy is not properly understood in the United States. He suggests that the moral reading is not antidemocratic because he rejects the majoritarian premise which suggests that political outcomes should coincide with the preferences of the majority. Most Americans, he claims, agree that the majority can be checked, and if democracy is about equal concern and respect for each individual it will not necessarily be regrettable if the majority is occasionally thwarted. Rather than being anti-democratic, Dworkin suggests that the moral reading is indispensable to the working of democracy. Democracy, for Dworkin, is about protecting rights rather than majority rule and he comes down on the side of the separation of powers in which judges, rather than the people, make crucial political decisions. A large part of this conclusion rests on the fact that he sees courts as institutions that are more deliberative than legislatures. As has already been pointed out in an earlier chapter, this misrepresents the deliberative nature of many legislative bodies. But a big part of the argument is also that minorities need protecting, and once more the argument about the tyranny of the majority is used to support a political system in which political power is removed from the people and placed in the hands of a supposedly intellectually and morally superior elite. Despite Dworkin’s protestations, the moral reading is indeed elitist and anti-democratic. It would be more appropriate for Dworkin, and liberals generally, to state that their fear of majority rule only allows support for a very limited notion of democracy. Despite this criticism, Dworkin does provide a very persuasive argument about the moral reading of the constitution. This argument, however, better serves the interests of those who are wary of placing too much faith in legal documents. If moral judgements are at the heart of law-making, why should we entrust them to a small unrepresentative minority? Sheldon Wolin (1996) sympathizes with the critique of constitutionalism put forward here and agrees that constitutions undermine
Democracy versus constitutionalism 165 democracy. He defines democracy as a project for realizing the political potential of citizens rather than as a form of government. The trouble with liberal democracies is that the boundaries of the political are drawn by the constitution, which ‘regulates the amount of democratic politics it lets in’ (1996, 34). The liberal democratic state is also based on a model that draws boundaries to keep others out completely. It promotes an inward-looking approach to politics that is increasingly difficult to fit with the world of global capitalism and communication. Wolin also makes the familiar point that liberalism is built on a suspicion of democracy. Madison, Tocqueville and Mill are three obvious examples of this. Such fears, Wolin claims, promote a constitutionalism that removes the political (the act of collective decisionmaking) and replaces it with politics (the battle for resources). The political is removed and is replaced by boundaries, institutions and bureaucracies, leaders, experts, and procedures. He suggests that a return to Aristotle’s idea of knowing how to rule and be ruled is necessary if the political is to be returned to its rightful place in the life of the demos. Democracy, the act of political decision-making by citizens, has almost disappeared, and its presence is now only ‘occasional and fugitive’ (1996, 39). What has been lost is the ‘demos as actor’ (1996, 34). For Wolin, ‘[t]he power of the demos is majority rule; but constitutionalism, especially in its Madisonian version, is designed to strew as many barriers as possible in the way of democratic power’ (1996, 41). Constitutions set specific requirements for liberal democracy, such as a free press, political parties and a market economy: ‘The specifications are so precise that the United States periodically dispatches experts to Central America to determine whether those requirements have been met’ (1996, 42). The model of politics as a continuous activity by elites to gain control of the state apparatus is always going to be at odds with an ideal of democracy, with citizens as actors and politics as periodic activity. Democracy should be seen as something other than government; it is a rebellious moment when the political rather than politics comes to the fore. What Wolin is identifying here is the usual liberal fear of the tyranny of the majority. This is the reason that politics has been displaced by the political; the latter is seen as too unstable and dangerous and the demos cannot be trusted to make decisions. Chantel Mouffe (2000) makes a similar point when she claims that we are dealing with a new form of political society stemming from two independent sources, namely the liberal tradition and the democratic
166 Democracy versus constitutionalism tradition. The liberal point of view has won out and democracy tends to be limited by liberal concerns for rights which have created a ‘democratic deficit’. In Wittgenstein’s terms, both traditions have different and incompatible grammars. Benjamin Barber (1996) seems to be in agreement, judging by his remarks about foundationalism in politics. The crucial aspect of democratic politics, according to Barber, is that we make decisions collectively. When we engage in such activity we are not unduly interested in the question of what is true, but, rather, of what we should do: ‘Foundationalism tends then to undermine democracy, and democracy both requires and entails an immunity to its own foundations if it is to flourish’ (1996, 352). The reason foundationalists are immune to democracy is that they already know the truth about certain issues from the start; they place these truths in constitutions that limit the democratic process. Foundations tell us certain things we can and cannot do, and hence prevent us choosing and willing our own existence. We are bound by a decision taken at a particular time in history that binds us for all time, even if those bonds can be interpreted in different ways by a small elite of judges. Barber suggests that such a system gets things the wrong way round: ‘What is right, or even what a right is, cannot in itself determine political judgment. Rights themselves are both constantly being redefined and reinterpreted and depend for their normative force on the engagement and commitment of an active citizen body’ (1996, 354). Such sentiments seem eminently sensible, as long as we realize we are not dealing with epistemology and that we reject (a) the sort of Habermasian claims that democratic deliberation leads to consensus, and justice, and (b) the sorts of pragmatism that suggest all knowledge and reality rest on the collective decisions of the community. The social choice arguments, examined at the beginning of the book, place such arguments in serious doubt. But as I argued in Chapter 4, we can defend democracy regardless of its procedural outcomes. Democracy is useful simply because it engages us and starts us thinking about what is important. A major criticism of democracy suggests that the masses do not think clearly about politics, are uninformed, and hence will continually make incorrect decisions. Condorcet’s Jury Theorem is interesting in this regard. Condorcet tells us that if we have a large number of voters each of whom is right 51 per cent of the time, the chances of the majority getting it right are 99.97 per cent. In a very interesting argument, Robert Goodin (2003) suggests we think of the argument in terms of tossing a coin. A small number of tosses might be very skewed, but the larger the number, the greater the probability that
Democracy versus constitutionalism 167 heads or tails will win 50 per cent of the time. The same is true with voters or juries, and the larger the number of participants the more likely it is that the 51 per cent who are right will overrule the 49 per cent who are wrong. The jury theorem has been criticized because it relates only to the options of guilty or innocent and relies on the assumption that more people know the answer than do not. Goodin shows that even if these conditions are relaxed, the results are still persuasive. With three or more options and less than 51 per cent being right, the majority winner is still much more likely to be right and as the number of voters grows, this figure converges on 1. Goodin also examines the arguments of Bayes, who made similar claims to Condorcet: ‘Bayes tells us what is the probability that outcome K is right given that it was chosen by the majority, whereas Condorcet tells us the probability that the majority will choose K given that K is right’ (2003, 111). Bayes concurs with Condorcet that it is rational to accept the decision of the majority because it is highly probable that the majority is right. Hence, if one is rational one must revise one’s beliefs and bring them in line with the majority decision. Minorities should cave in to the majority and they are acting irrationally if they do not. This argument supposes that people are voting based on the assumption that one answer is correct and another is wrong, i.e. that a vote has propositional content. This assumption might not be accurate from the instrumentally rational point of view. If one’s vote will not have an impact on the outcome, as rational choice theory tells us it will not, it does not matter for which option one votes. Hence, it is rational for a vote not to have any propositional content and if this is the case one cannot derive the conclusions Bayes hopes for. Of course, people do have reasons for why they vote and they do think their vote counts, so the Bayes argument holds up on empirical grounds. If Condorcet and Bayes are correct then it seems there is a good case for supporting pure majority rule and for abandoning our concerns about the uneducated majority. Democracy is a very good truth tracker – as long as democracy is about facts and close to 50 per cent of the participants are accurately informed. But as Goodin says, it is also about values that are not right or wrong in the way that facts are. Hence, it is still rational for minorities to oppose majorities on normative issues. But at least at the level of factual evaluation, it seems that the arguments of Condorcet and Bayes support the claim that majority rule is superior to other decision-making procedures. And as Goodin notes, those who think there is no difference between facts and
168 Democracy versus constitutionalism norms should be highly supportive of democracy as the way to arrive at normative decisions. The crucial assumption that underlies these conclusions is that large numbers of people hold views that are factually correct. This suggests that democracies need to foster a culture that prizes and promotes political awareness and removes some of the power currently in the hands of judges, representatives and bureaucrats. It is particularly important that democratic societies have an adequate system in place to supply politically relevant information to citizens. None of this suggests that voting can reveal the general will because judgements about specific facts will not reveal a result that reflects what everyone wants.
The British Constitution One way to assess the intent and success of Hobbesian democratic absolutism is to examine the constitution of the country in which Hobbes lived. I could also have used New Zealand as an example of a sovereign body that is absolutist and yet compatible with democracy. Lijphart notes that New Zealand is ‘a virtually perfect example of the Westminster model of democracy’ (1984, 16). Lijphart quotes a famous New Zealand legal scholar, who states that ‘the central principle of the Constitution is that there are no effective legal limitations on what Parliament may enact by the ordinary legislative process’ (1984, 19). Despite this supposedly dangerous system, New Zealand has a better record (though still bad) of protecting its indigenous population than America or Australia and it was the first democratic society to give women the franchise. The British Constitution was seen, from Dicey to Jennings, as being in the Hobbesian mould, that is to say the executive is absolute. According to such scholars the beauty of the system is its absolutist nature, which gives Britain its distinctive and laudable form of politics. Bagehot suggested that the fusion of the legislative and executive power in the British system was the most important key to its success, and as a young man, Woodrow Wilson advocated scrapping the presidential form of politics in favour of a British style parliamentary system. The lack of legal rigidity that informs the American system was seen as one of its primary blessings, and British exceptionalism was perceived as something in which to take pride. The particularly interesting feature of British government is that it has become democratized in practice, with very little change to its formal structure. The executive power of the crown remains unchecked. From appointing judges, generals and political figures, up
Democracy versus constitutionalism 169 to and including the prime minister (and, crucially, dismissing them), to the day-to-day running of the government and the signing of bills into law, the crown is unrestricted. There are no formal constitutional rules for such functions, and the very office itself is not specified. Hence, the executive power in Britain is precisely the type that worries Tarlton, Boyd and other commentators on Hobbes so much. These powers remain with the monarch only nominally and have been democratized so that the executive function is now performed by the prime minister. The head of state and the head of government have parted ways: In the British context, the exercise of executive power was carried out in the name of the monarch and the rules specifying the role of the head of the government and the ministry were not set out in a constitutional document. This meant that little appeared to have changed in formal terms from a system where all aspects of executive power were carried out autocratically by the monarch. (Sharman 2001, 171) What this means is that absolute power has changed hands, but the power has not disappeared or fundamentally changed. As Campbell Sharman says, the upshot is that the executive has been ‘democratised but not really checked’ (2001, 171). The quotes above do not, however, get to the real heart of the sovereign nature of the British government because within Parliament the crown formally exercises veto power over all legislation. Hence, even with the democratization of certain aspects of Parliament, the untrammelled nature of executive power, still nominally with the crown, remains to this day. As Campbell Sharman also observes, ‘The notion of a constitution as an explicit statement of basic rules limiting the mode and sphere of government operation is alien to the British style of government’ (1990, 207). In 1800 the British system of government did still reflect a significant separation of powers, but as Sharman suggests, ‘the parliamentary system has an inherent propensity towards the concentration of power in the hands of the executive branch’ (1990, 205). But as he also crucially notes, this does not exclude Britain from the ranks of democracies. Has liberty been under much greater threat in Britain than in a system with formal checks on power, such as America? I would venture that it has not. This is not to say that the British are more free than the Americans; it is simply to suggest that the level of freedom available to citizens in both countries is somewhat similar and that
170 Democracy versus constitutionalism it has relatively little to do with the structure of government. It has more to do with political culture. As Munro suggests, [t]here is nothing legally to prevent Parliament from denying used car salesmen the right to vote or making the practice of Methodism a criminal offence, but respect for the principles of universal adult suffrage and freedom of religion is such as to render the possibilities unreal. (1987, 6) These things would not happen in Britain because there is a cultural presupposition in favour of political and religious freedom; and if there was not such a disposition, legal documents would not be capable of providing much in the way of protection. A cursory knowledge of US history is enough to show that freedom cannot be guaranteed simply through the attempt to curtail political power by constitutional means. The Constitution did not provide protection for an indigenous population that was systematically eradicated. Neither did it make the bonds of slavery any looser; in fact the unequal worth of slaves was written into the very framework of the document. The Constitution did not provide for the political liberty of women for over 100 years after it was written and did not prevent segregation in the South. It did not even protect a person’s freedom to have a beer at the end of the day for certain periods of history (and still persecutes rather than protects 20-year-olds who want to engage in that activity today). It was under elitist liberal democracy that these things happened. A constitution of a different political system provided the following rights: all accused persons are guaranteed a fair trial and judges are independent and subject only to the law; each person has a right to an education; women have all the same rights as men; maternity leave is guaranteed with full pay and there is a right to the provision of nurseries and kindergartens; there is to be freedom of conscience, religious worship and speech, freedom of the press, and freedom of assembly, mass meetings and demonstrations (including the provision of organized printing resources so that these rights can be exercised); there is the freedom to join trade unions; no one can be placed under arrest without a decision of the court; the inviolability of the home and the right to privacy is guaranteed; and there is to be universal, equal and direct suffrage by secret ballot for all men and women. This sounds like the constitution of one of the Scandinavian countries but is in fact that of the Soviet Union. None of this is to suggest that democratically introduced limitations on power would make the United Kingdom a worse place to live. It is
Democracy versus constitutionalism 171 simply to point out that not having such safeguards in a constitutional format does not equal tyranny, and having them does not ensure safety and liberty. This seems particularly important given the current political climate of homeland security and the war against terror. My point is simply that the British/Hobbesian model, while absolutist in nature, is exercised in a manner compatible with the protection of individual liberty, and to write it off as necessarily tyrannical is to have an unnecessarily narrow view of political possibilities.
Conclusion: Limits on parliamentary supremacy The evidence in the section above suggests that despite the concentration of power and the lack of institutional checks favoured by the American founders, Britain does not seem to have a significantly worser record than the United States when it comes to the protection of liberty, and it has a notably better record than the Soviet Union. One has to conclude that the Hobbesian concept of sovereignty is not synonymous with tyranny. One explanation for this might be that the British Parliament is not quite as unchecked as it would seem. A criticism of early constitutional theorists such as Dicey is that they failed to take into account the nuances of the British political system; once we do, we find that Parliament is limited in a variety of different ways. To begin with, Parliament has always been restricted by laws even if they are of its own making and can be repealed if Parliament so chooses. Lijphart notes that British governments operate according to a number of basic laws, customs, and conventions. Parliament will normally obey these constitutional rules, but it is not formally bound by them. Even the basic laws have no special status and they can be changed by Parliament in the same way as any other laws Parliament is the ultimate, or sovereign, authority there are no formal restrictions on the power of the majority of the House of Commons. (1984, 9) What sovereignty means, and here Hobbes would agree, is that ‘Parliament is the supreme law-making body whose legislative acts override those of the courts or of subordinate legislation’ (Mount 1992, 82). Parliament, however, still has to abide by rules of procedure for as long as it decides not to formally change those laws. Hence, we should keep in mind that arguments for absolute sovereignty do not mean that there cannot also be strict rules for exercising such power. Once a law is made, Parliament is bound by it until such time
172 Democracy versus constitutionalism as it again decides to make a change. Parliament has always taken its procedures very seriously and previous Parliaments do bind future ones to certain forms of behaviour; it is just that such bonds are not permanent. As a matter of practice, such procedures place significant limits on the day to-day activity of the legislative body. What British parliamentarians have in effect done is to check themselves through self-imposed procedures. There are also many informal practices that place limits on behaviour. These are what Dicey called ‘the Conventions of the Constitution’ and are rules that regulate conduct even though they are not enforceable by the courts. Such conventions place customary limits on governments and tend to have varying levels of acceptance, binding force and precision. As Munro suggests, they are best seen as existing on a continuum. The point is that they exist in any political system and influence political behaviour even when they are somewhat ephemeral. They include such things as the Monarch’s capacity to withhold assent when presented with a Bill. The Queen can theoretically appoint or dismiss any minister at her choosing; the very office of prime minister is a convention that began with Walpole and only took its modern shape with the party system of the nineteenth century. To say that Parliament is supreme is to say that there are no legal limits on it, not that there are no political restraints. These restraints extend beyond Westminster. Parliament can pass horrible legislation but this does not mean that it would not meet with large-scale political protest. We should also keep in mind that simply because Parliament can change any law does not mean that there is not a great respect for the rule of law. Britain has a political system that fosters the rule of law rather than the rule of individuals. There is lot of ‘separation of functions’ (a better term than ‘separation of powers’ for Britain) even though the crown is still supreme. The judiciary cannot legislate but the courts make themselves heard when they feel that Parliament is trying to influence sentencing. Finally, we should remember that Britain has many legal documents as part of its ‘unwritten’ constitution, and that the American system is not exhausted by the rules and regulations of its own written constitution. The difference between the two is that some of the aspects seen as most important in the United States have been put into a single document and the idea seems to be that such a grouping provides a much greater defence of liberty. The documents in the British system stretch over centuries and include the Magna Carta, the Instrument of Government that provided Cromwell with the Protectorate, the Humble Petition and Advice that followed shortly after, the Acts of Union, and the Bill of
Democracy versus constitutionalism 173 Rights that set the terms offered to William and Mary. None of these undermines the legal concept of parliamentary sovereignty, but all go some way in providing for tempered and reasonable government. None of the discussion in this chapter is meant to suggest that a more active, participatory and unlimited democracy is without flaws. It might well lead to abuses of individuals on occasions but unless we can show that it would be worse than constitutional liberal democracy it should be the system of preference. The track record of the Supreme Court in America has not been too inspiring when it comes to defending citizens, and Shapiro and others have noted that the court has a long record of violating rights. The absolutism I recommend fits better with the democratic spirit and conveniently avoids some of the intransitivity problems identified earlier. Shapiro recommends that courts should be allowed to delay legislation when the basic rights and interests of citizens seem to be threatened, but when we are dealing with issues of less importance the right to delay simply favours the status quo. Hence, he argues for ‘democracy-reinforcing judicial review – as distinct from the democracy-limiting view’ (2003, 51). I think that a court or some other body without the power to negate popular rule is acceptable. It should serve as a moral watchdog to warn us when we stray from the basic principles of democracy and democratic citizenship. Opposition parties currently do this, but because of their partisan nature, they do not carry the moral authority of the type of institution I am talking about. But such an institution cannot have the sort of power that would let it act as the US Supreme Court did when it decided on the Gore/Bush election. Nor should such an institution be politically appointed. A strong case could be made for a non-elected second chamber, such as the reformulated House of Lords (perhaps chosen by lottery from the general population) as long as its role is clearly defined and its powers are limited to moral censorship. This is perfectly compatible with the absolutism I have advocated in this book. It is fitting to leave the last words to Hobbes, who warns us against the complacency that can arise when we think that political power can be tamed by institutional means: But grant then, that thou hadst given him a power which were not absolute, but so much only as sufficed to defend thee from the injuries of others; which if thou wilt be safe, is necessary for thee to give; are not all the same things to be feared? For he that hath strength enough to protect wants not sufficiency to oppress all. (1972, 181)
Bibliography
Alford, C.F. (1996) ‘Comment’, Journal of Politics 58: 753–6. Arblaster, A. (1987) Democracy, McGraw-Hill: Open University Press. Arneson, R.J. (1993) ‘Democratic Rights at National and Workplace Levels’, in D. Copp, J. Hampton and J.E. Roemer (eds) The Idea of Democracy, Cambridge: Cambridge University Press. Arrow, J. (1963) Social Values and Individual Values, New York: Wiley. ——(1984) Social Choice and Justice: Collected Papers of Kenneth J. Arrow, Oxford: Basil Blackworth. Austen-Smith, D. (1987) ‘Sophisticated Sincerity: Voting over Endogenous Agendas’, American Political Science Review 81: 1323–9. Austen-Smith, D. and Banks, J.S. (1998) ‘Social Choice Theory, Game Theory and Positive Political Theory’, Annual Review of Political Science 1:259–87. Baccaro, L. (2001) ‘Aggregative and Deliberative Decision-Making Procedures: A Comparison of Two Southern Italian Factories’, Politics and Society 29: 243–71. Barber, B. (1984) Strong Democracy: Participatory Politics for a New Age, Berkeley: University of California Press. ——(1989) ‘Liberal Democracy and the Costs of Consent’, in N. Rosenblum (ed.) Liberalism and the Moral Life, Cambridge, MA: Harvard University Press. ——(1996) ‘Foundationalism and Democracy’, in S. Benhabib (ed.) Democracy and Difference: Contesting the Boundaries of the Political, Princeton: Princeton University Press. Barry, B. (1972) ‘Warrender and His Critics’, in M. Cranston and R.S. Peters (eds) Hobbes and Rousseau: A Collection of Critical Essays, New York: Anchor. Baumgold, D. (1988) Hobbes’s Political Theory, Cambridge: Cambridge University Press. ——(1993) ‘Pacifying Politics: Resistance, Violence, and Accountability in Seventeenth-Century Contract Theory’, Political Theory 21: 6–27. Benhabib, S. (1994) ‘Deliberative Rationality and Models of Democratic Legitimacy’, Constellations 1: 26–52.
Bibliography 175 ——(1996) ‘Towards a Deliberative Model of Democratic Legitimacy’, in S. Benhabib (ed.) Democracy and Difference: Contesting the Boundaries of the Political, Princeton: Princeton University Press. Berlin, I. (1969) Four Essays on Liberty, Oxford: Oxford University Press. Bernholz, P. (1975) ‘Logrolling and the Paradox of Voting: Are They Logically Equivalent?’, American Political Science Review 69: 961–2. Bird, C. (2000) ‘The Possibility of Self-Government’, American Political Science Review 94: 563–78. Black, D. (1958) The Theory of Committees and Elections, Cambridge: Cambridge University Press. Blydenburgh, J. (1971) ‘The Closed Rule and the Paradox of Voting’, Journal of Politics 33: 57–71. Bohman, J. (1993) ‘Communication, Ideology, and Democratic Theory’, American Political Science Review 84: 93–109. ——(1996) Public Deliberation, Cambridge, MA: MIT Press. Boonin-Vail, D. (1994) Thomas Hobbes and the Science of Moral Virtue, Cambridge; Cambridge University Press. Bowling, K.R. (1990) ‘The Revolutionary Origin of the Exclusive Jurisdiction of Congress over the District of Columbia’, paper presented at the American Political Science Association Meeting, San Francisco, August 30–September 2. Boyd, R. (2001) ‘Thomas Hobbes and the Perils of Pluralism’, Journal of Politics 63: 392–413. Brennan, G. and Hamlin A. (2000) Democratic Devices and Desires, Cambridge: Cambridge University Press. Brown, K. (1962) ‘Hobbes’s Grounds for Belief in a Deity’, Philosophy 37: 366–44. ——(1965) Hobbes Studies, Oxford: Blackwell. Buchanan, J.M. and Tullock, G. (1962) The Calculus of Consent: Logical Foundations of Consitutional Democracy, Ann Arbor: University of Michigan Press. Budge, I. (1996) The New Challenge of Direct Democracy, Cambridge: Polity Press. Burgess, G. (1994) ‘On Hobbesian Resistance Theory’, Political Studies XLII: 62–83. Burke, E. (1968) Reflections on the Revolution in France, C.C. O’Brien (ed.), London: Penguin. Burnett, E.C. (ed.) (1921–36) Letters of the Members of the Continental Congress, Volumes 1–8, Washington, DC: Carnegie Institution. ——(1941) The Continental Congress: A Definitive History of the Continental Congress from its Inception in 1774 to March 1789, New York: W.W. Norton. Chambers, S. (1995) ‘Discourse and Democratic Practices’, in S. White (ed.) The Cambridge Companion to Habermas, Cambridge: Cambridge University Press.
176 Bibliography ——(1996) Reasonable Democracy: Jurgen Habermas and the Politics of Discourse, Ithaca: Cornell University Press. ——(1999) ‘Talking versus Voting: Legitimacy, Efficiency, and Deliberative Democracy’, unpublished manuscript, University of Colorado. Cohen, J. (1986) ‘An Epistemic Conception of Democracy’, Ethics 97: 26–38. ——(1989) ‘Deliberation and Democratic Legitimacy’, in A. Hamlin and P. Pettit (eds) The Good Polity, London: Basil Blackwell Ltd. ——(1996) ‘Procedure and Substance in Deliberative Democracy’, in S. Benhabib (ed.) Democracy and Difference: Contesting the Boundaries of the Political, Princeton: Princeton University Press. Coleman, J. (1990) Foundations of Social Theory, Cambridge, MA: Harvard University Press. Coleman, J. and Ferejohn, J. (1986) ‘Democracy and Social Choice’, Ethics 97: 6–25. Coles, R. (1992) ‘Communicative Action and Discourse Ethics: Habermas and Foucault’, Polity 25: 71–94. Connolly, W.E. (1988) Political Theory and Modernity, Oxford: Basil Blackwell. Cooper, J. and Brady, D. (1981a) ‘Institutional Context and Leadership Style: The House from Cannon to Rayburn’, American Political Science Review 75: 411–25. ——(1981b) ‘Toward a Diachronic Analysis of Congress’, American Political Science Review 75: 988–1012. Cox, G.W. (1987) ‘Electoral Equilibria under Alternative Voting Institutions’, American Political Science Review 34: 82–108. ——(1997) Making Votes Count, Cambridge: Cambridge University Press. ——(1999) ‘The Empirical Content of Rational Choice Theory: A Reply to Green and Shapiro’, Journal of Theoretical Politics 11: 147–69. Cranston, M. and Peters, R.S. (eds) (1972) Hobbes and Rousseau, New York: Anchor. Cronin, T. (1989) Direct Democracy: The Politics of Initiative, Referendum and Recall, Cambridge: Harvard University Press. Dahl, R.A. (1956) A Preface to Democratic Theory, Chicago: University of Chicago Press. ——(1960) Who Governs?, New Haven, CT: Yale University Press. ——(1971) After the Revolution, New Haven, CT: Yale University Press. ——(1989) Democracy and its Critics, New Haven, CT: Yale University Press. ——(1998) On Democracy, New Haven, CT: Yale University Press. Dahl, R.A. and Tufte, E. (1973) Size and Democracy, Stanford: Stanford University Press. Davis, O. and Hinich, M.J. (1968) ‘On the Power and Importance of the Mean Preference in a Mathematical Model of Democratic Choice’, Public Choice 5: 59–72. Dietz, Mary (ed.) (1990) Thomas Hobbes and Political Theory, Lawrence: University of Kansas Press.
Bibliography 177 Donovan, T. and Bowler, S. (1998) ‘Direct Democracy and Minority Rights: An Extension’, American Journal of Political Science 42: 1020–4. Dryzek, J.S. (1990) Discursive Democracy, New York: Cambridge University Press. ——(2000) Deliberative Democracy and Beyond: Liberals, Critics, Contestations, Oxford: Oxford University Press. Dryzek, J.S. and List, C. (2003) ‘Social Choice Theory and Deliberative Democracy: A Reconciliation’, British Journal of Political Science 33: 1–28. Dworkin, R. (1999) ‘The Moral Reading and the Majoritarian Premise’, in H.H. Koh and R. Slye, (eds) Deliberative Democracy and Human Rights, New Haven: Yale University Press. Ettema, J.S. (1994) ‘Discourse That is Closer to Silence Than to Talk: The Politics and Possibilities of Reporting on Victims of War’, Critical Studies in Mass Communication 11: 1–21. Ewin, R.E. (1991) Virtues and Rights: The Moral Philosophy of Thomas Hobbes, Boulder: Westview Press. Feinberg, J. (1985) Offense to Others: The Moral Limits of the Criminal Law, New York: Oxford University Press. Femia, J. (2001) Against the Masses, Oxford: Oxford University Press. Fishkin, J. (1991) Democracy and Deliberation, New Haven, CT: Yale University Press. ——(1997) The Voice of the People, New Haven, CT: Yale University Press. Flathman, R. (1987) The Philosophy and Politics of Freedom, Chicago: University of Chicago Press. ——(1993) Thomas Hobbes: Skepticism, Individuality and Chastened Politics, Newbury Park: Sage. Frost, S. (2001) ‘Faking it: Hobbes’s Thinking-Bodies and the Ethics of Dissimulation’, Political Theory 29: 30–57. Gauthier, D. (1969) The Logic of Leviathan, Oxford: Oxford University Press. ——(1976) Morals by Agreement, Oxford: Clarendon Press. ——(1995) ‘Public Reason’, Social Philosophy and Policy 12: 19–42. Gert, B. (1965) ‘Hobbes, Mechanism and Egoism’, Philosophical Quarterly 15: 341–49. Goldsmith, M.M. (1966) Hobbes’s Science of Politics, New York: Columbia University Press. Goodin, R. (2003) Reflective Democracy, Oxford: Oxford University Press. Gould, C. (1996) ‘Diversity and Democracy: Representing Differences’, in S. Benhabib (ed.) Democracy and Difference: Contesting the Boundaries of the Political, Princeton: Princeton University Press. Green, D.P. and Shapiro, I. (1994) Pathologies of Rational Choice Theory: A Critique of Applications in Political Science, New Haven, CT: Yale University Press. ——(1995) ‘Reflections on our Critics’, Critical Review 9: 235–76. Habermas, J. (1982) ‘A Reply to My Critics’, in J.B. Thompson and D. Held (eds) Habermas Critical Debates, Cambridge, MA: MIT Press.
178 Bibliography ——(1984) The Theory of Communicative Action, Volume One, Boston: Beacon Press. ——(1987) The Theory of Communicative Action, Volume Two, Boston: Beacon Press. ——(1990) Moral Consciousness and Communicative Action, Cambridge, MA: MIT Press. ——(1993) Justification and Application, Cambridge MA: MIT Press. ——(1994) ‘Three Models of Democracy’, Constellations 1: 1–10. ——(1996) Between Facts and Norms, Cambridge MA: MIT Press. Hamilton, A., Madison, J., and Jay, J. (1961) The Federalist Papers, C. Rossiter (ed.), New York: Penguin. Hardin, R. (1993) ‘Public Choice versus Democracy’, in D. Copp, J. Hampton and J.E. Roemer (eds) The Idea of Democracy, Cambridge: Cambridge University Press. Henderson, H.J. (1974) Party Politics in the Continental Congress, New York: McGraw-Hill. Hirst, P. (1994) Associative Democracy: New Forms of Economic and Social Governance, Cambridge: Polity Press. Hobbes, Thomas (1966) English Works, Vols. 1–10, Sir W. Molesworth (ed.), London: John Bohn. ——(1968) Leviathan, C.B. Macpherson (ed.), London: Penguin Books. ——(1972) De Cive, London: Appleton-Century-Crofts. ——(1972) Man and Citizen, B. Gert (ed.), New York: Doubleday. ——(1990) Behemoth or the Long Parliament, F. Tonnies (ed.), Chicago: University of Chicago Press. —–(1994) The Elements of Law, Natural and Politic, J.C.A. Gaskin (ed.), New York: Oxford University Press. Holmes, S. (1990) ‘Political Psychology in Hobbes’s Behemoth’, in M.G. Dietz (ed.) Thomas Hobbes and Political Theory, Lawrence: University of Kansas Press. ——(1993) ‘Tocqueville and Democracy’, in D. Copp, J. Hampton and J.E. Roemer (eds) The Idea of Democracy, Cambridge: Cambridge University Press. ——(1999) ‘Constitutionalism, Democracy, and State Decay’, in H.H. Koh and R. Slye (eds) Deliberative Democracy and Human Rights, New Haven: Yale University Press. Horster, D. (1992) Habermas: An Introduction, Philadelphia: Pennbridge Books. Ingram, D. (1993) ‘The Limits and Possibilities of Communicative Ethics for Democratic Theory’, Political Theory 21: 294–321. Jillson, C. and Wilson, R.W. (1994) Congressional Dynamics: Structure, Coordination, and Choice in the First American Congress, 1774–1789, Stanford: Stanford University Press. Johnson, J. (1993) ‘Is Talk Really Cheap? Prompting Conversation Between Critical Theory and Rational Choice’, American Political Science Review 87: 74–84.
Bibliography 179 Kahn, V. (2001) ‘Hobbes, Romance, and the Contract of Mimesis’, Political Theory 29: 4–29. Kateb, G. (1996) ‘The Freedom of Worthless and Harmful Speech’, in B. Yack (ed.) Liberalism Without Illusions: Essays on Liberal Theory and the Political Vision of Judith N. Shklar, Chicago: University of Chicago Press. Knight, J. and Johnson, J. (1994) ‘Aggregation and Deliberation: On the Possibility of Democratic Legitimacy’, Political Theory 22: 277–96. Langton, R. (1990) ‘Whose Right? Ronald Dworkin, Women, and Pornographers’, Philosophy and Public Affairs 19: 311–59. Lijphart, A. (1984) Democracies: Patterns of Majoritarian and Consensus Government in Twenty One Countries, New Haven, CT: Yale University Press. Locke, J. (1988) The Two Treatises of Government, P. Laslett (ed.), Cambridge: Cambridge University Press. Lohmann, S. (1995) ‘The Poverty of Shapiro and Green’, Critical Review 9: 127–54. McCarthy, T. (1978) The Critical Theory of Jurgen Habermas, Cambridge, MA: MIT Press. McKelvey, R.D. (1976) ‘Intransitivities in Multidimensional Voting Models and Some Implications for Agenda Control’, Journal of Economic Theory 12: 472–82. Mackie, G. (2003) Democracy Defended, Cambridge: Cambridge University Press. Mannin, B. (1987) ‘On Legitimacy and Political Deliberation’, Political Theory 15: 338–68. Mansbridge, J. (1996) ‘Using Power/Fighting Power: The Polity’, in S. Benhabib (ed.) Democracy and Difference: Contesting the Boundaries of the Political, Princeton: Princeton University Press. Mendelberg, T. (2002) ‘The Deliberative Citizen: Theory and Evidence’, in M. Delli Carpinin, L. Huddy, and R. Shapiro (eds) Political Decision Making, Deliberation and Participation, Amsterdam: Elsevier. Mill, J.S. (1978) On Liberty, Indianapolis: Hackett. Miller, D. (1992) ‘Deliberative Democracy and Social Choice’, Political Studies 40: 54–67. Mouffe, C. (1996) ‘Democracy, Power, and the “Political” ’, in S. Benhabib (ed.) Democracy and Difference: Contesting the Boundaries of the Political, Princeton: Princeton University Press. ——(2000) The Democractic Paradox, London: Verso. Mount, F. (1992) The British Constitution Now: Recovery or Decline? London: Heinemann. Mueller, D.C. (1989) Public Choice 11, Cambridge: Cambridge University Press. Munro, C.R. (1987) Studies in Constitutional Law, London: Butterworth & Co. Nozick, R. (1974) Anarchy, State and Utopia, New York: Basic Books.
180 Bibliography Okin, S. Moller (1989) ‘Humanist Liberalism’, in N. Rosenblum (ed.) Liberalism and the Moral Life, Cambridge: Harvard University Press. Olson, M. Jr. (1965) The Logic of Collective Action, Cambridge MA: Harvard University Press. ——(1982) The Rise and Decline of Nations, New Haven: Yale University Press. Outhwaite, W. (1994) Habermas: A Critical Introduction, Stanford: Stanford University Press. Pateman, C. (1970) Participation and Democratic Theory, Cambridge: Cambridge University Press. Phillips, A. (1991) Engendering Democracy, Cambridge: Polity Press. ——(1993) Democracy and Difference, Cambridge: Polity Press. ——(1994) ‘Dealing with Difference: A Politics of Ideas or a Politics of Presence?’ Constellations 1: 74–91. Plato (1968) Republic, trans. Allan Bloom, Basic Books: Harper Collins. Plott, C.R. (1967) ‘A Notion of Equilibrium and its Possibility Under Majority Rule’, American Economic Review 57: 787–806. Plott, C.R. and Levine, M.E. (1978) ‘A Model of Agenda Influence on Committee Decisions’, American Economic Review 68: 146–60. Radcliff, B. and Wingenbach, E. (2000) ‘Preference Aggregation, Functional Pathologies, and Democracy: A Social Choice Defence of Participatory Democracy’, Journal of Politics 62: 977–98. Rawls, J. (1993) Political Liberalism, New York: Columbia University Press. Rescher, N. (1993) Pluralism, Oxford: Oxford University Press. Ridge, M. (1998) ‘Hobbesian Public Reason’, Ethics 108: 538–68. Riker, W.H. (1963) The Theory of Political Coalitions, New Haven CT: Yale University Press. ——(1980) ‘Implications from the Disequilibrium of Majority Rule for the Study of Institutions’, American Political Science Review 74: 432–46. ——(1982) Liberalism Against Populism: A Confrontation Between the Theory of Democracy and the Theory of Social Choice, Prospect Heights: Waveland Press. ——(1984) ‘The Heresthetics of Constitution Making: The Presidency in 1787, with Comments on Determinism and Rational Choice’, American Political Science Review 78: 1–16. ——(1986) The Art of Political Manipulation, New Haven CT: Yale University Press. Rosenberg, S. ‘Reason, Communicative Competence and Democratic Deliberation: Do Citizens Have the Capacities to Effectively Participate in Deliberative Decision-Making?’ Paper presented at the American Political Science Association, Philadelphia, August 2003. Ryan, A. (1983) ‘Hobbes, Toleration, and the Inner Life’, in D. Miller and L. Seidentop (eds) The Nature of Political Theory, Oxford: Clarendon Press. ——(1988) ‘Hobbes and Individualism’, in G.A.J. Rogers and A. Ryan (eds) Perspectives on Thomas Hobbes, Oxford: Clarendon Press. Sanders, L. (1997) ‘Against Deliberation’, Political Theory 25: 347–76.
Bibliography 181 Schlosberg, D. (1995) ‘Communicative Action in Practice: Intersubjectivity and New Social Movements’, Political Studies XL111: 291–311. Schofield, N. (1978) ‘Instability of Simple Dynamic Games’, Review of Economic Studies 45: 575–94. ——(1995) ‘Rational Political Economy’, Critical Review 9: 189–211. Sen, A. (1970) Collective Choice and Social Welfare, San Francisco: HoldenDay. ——(1982) Choice, Welfare and Measurement, Cambridge MA: Harvard University Press. ——(1999) ‘The Possibility of Social Choice’, American Economic Review 89: 349–78. ——(2002) Rationality and Freedom, Cambridge: Harvard University Press. Shapiro, I. (2003) The State of Democratic Theory, Princeton: Princeton University Press. Sharman, C. (1990) ‘Parliamentary Federations and Limited Government’, Journal of Theoretical Politics 2: 205–30. ——(2001) ‘Over Powered and Under Legitimized: Redesigning the Australian Head of State’, in Irving et al. (eds) Trusting the People: An Elected President for an Australian Republic, Perth: Optima Press (for Andrew Murray), 164–81. Shepsle, K.A. (1979) ‘Institutional, Arrangements and Equilibrium in Multidimensional Voting Models’, American Journal of Political Science 23: 27–59. ——(1986) ‘Institutional Equilibrium and Equilibrium Institutions’, in H. Weisberg (ed.) Political Science: The Science of Politics, New York: Agatha Press. ——(1995) ‘Statistical Political Philosophy’, Critical Review 9. Shepsle, K.A. and Weingast, B. (1981a) ‘Structure-Induced Equilibrium and Legislative Choice’, Public Choice 37: 503–19. ——(1981b) ‘Political Preferences for the Pork Barrel: A Generalization’, American Journal of Political Science 25: 66–111. ——(1984) ‘Political Solutions to Market Problems’, American Political Science Review 78: 417–34. ——(1987) ‘The Institutional Foundations of Committee Power’, American Political Science Review 81: 85–104. Shklar, J. (1989) ‘The Liberalism of Fear’, in N. Rosenblum (ed.) Liberalism and the Moral Life, Cambridge, MA: Harvard University Press. Skinner, Q. (1996) Reason and Rhetoric in the Philosophy of Hobbes, Cambridge: Cambridge University Press. ——(1998) Liberty before Liberalism, Cambridge: Cambridge University Press. Slomp, G. (1994) ‘Hobbes and the Equality of Women’, Political Studies XLII: 441–52. Spitz, E. (1984) Majority Rule, New Jersey: Chatham House. Sulkin, T. and Simon, A. (2001) ‘Habermas in the Lab: A Study of Deliberation in an Experimental Setting’, Political Psychology 22: 809–26.
182 Bibliography Steiner, J., Bachtiger, A., Sporndli, M. and Steenergen, M. (2004) Deliberative Politics in Action: Analysing Parliamentary Discourse, Cambridge: Cambridge University Press. Sunstein, C. (1996) Free Markets and Social Justice, New York: Oxford University Press. ——(1997) Legal Reasoning and Political Conflict, New York: Oxford University Press. Tarlton, C. (2001) ‘The Despotical Doctrine of Hobbes, Part 1: The Liberalization of Leviathan’, History of Political Thought XXII: 587–618. ——(2002) ‘The Despotical Doctrine of Hobbes, Part II: Aspects of the Textual Substructure of Tyranny in Leviathan’, History of Political Thought XXIII: 61–89. Taylor, C. (1985) Philosophy and the Human Sciences: Philosophical Papers 2, Cambridge: Cambridge University Press. Tullock, G. (1981) ‘Why So Much Stability?’, Public Choice 37: 189–202. Watkins, J.W.N. (1965) ‘Philosophy and Politics in Hobbes’, in K. Brown (ed.) Hobbes Studies, Oxford: Blackwell. ——(1973) Hobbes’s System of Ideas, London: Hutchinson & Co. Weale, A. (1999) Democracy, London: Macmillan. White, S. (1988) The Recent Work of Jurgen Habermas: Reason, Justice, and Modernity, Cambridge: Cambridge University Press. Wolin, S.S. (1996) ‘Fugitive Democracy’, in S. Benhabib (ed.) Democracy and Difference: Contesting the Boundaries of the Political, Princeton: Princeton University Press. Worthington C.F. (ed.) (1904) Journals of the Continental Congress 1774–1789, Washington DC: US Government Printing Office. Young, I.M. (1996) ‘Communication and the Other: Beyond Deliberative Democracy’, in S. Benhabib (ed.) Democracy and Difference: Contesting the Boundaries of the Political, Princeton: Princeton University Press.
Index
absolutism 1, 6, 99, 100, 126; and associational life 113–20; and censorship 109–13; defence of 145; Hobbesian notions 100–39; and liberty 105–9; as logically coherent 140; and property 120–7; and sovereignty 100–5, 122–4, 128, 129–30, 132–6; and US politics 104, see also Hobbes, Thomas agenda setting 9, 76, 79, 83 Alford, C.F. 16 Aristotle 16, 137 Arneson, R.J. 144–5, 147 Arrow, Kenneth 8, 18, 23, 30, 37, 46, 49; challenges to IIA condition 50–5; and collective choice 19–20; conditions imposed by 24, 35–6, 38, 49, 60–1, 77; criticisms of 41–2; General Possibility Theorem 19, 36; and ideal speech situation 45; and individual motivations 43–4; and instability 34–5; and irrational outcomes 19–20; and non-dictatorship 39; persuasiveness of arguments 98; pessimistic conclusions drawn from 38; and social choice theory 38; and social welfare function 19; and transitivity 68;
and values 67; and violation of condition P 50 associational life, and factions 116; and group fanaticism 115; and individualism 113–14, 118, 119–20; and irregular systems 116–17; and lawfulness of groups 115; limitations on 117–18; and need for community 118–19; and preservation of peace/order 115–16 Baccaro, Lucio 78 Bais-Moi (film) 157 Barber, Benjamin 100, 166 Baumgold, D. 133–4 Bayes, Thomas 167 Benhabib, S. 8, 30–1 Bergson, Henri 19 Bird, Colin 55–7, 59–60 Black, D. 8, 23 Borda, J.C. de 52, 53, 54–5 Bowling, K.R. 81 Boyd, Richard 100, 103, 107, 113–14, 115, 117 Buchanan, J.M. and Tullock, G. 66–7 Burgess, G. 133–4 Burke, Edmund 147–9 Burke, Kenneth 139 Burnett, E.C. 82–4
184 Index censorship, and banning of texts 110; and education 109–11, 112–13; and free speech 159–60; and mutual charity 112; power of 109; and servility/subordination 111; and sovereignty 109; and the state 112, 113; and tyranny of the majority 143 Chairman’s problem 63 Chambers, Simone 12, 78 Cohen, J. 8, 15, 71, 144 Coleman, J. and Ferejohn, J. 23 committees 62–4; formation of 82; and as source of equilibrium 83 Condorcet, M.J.A.N.C., marquis de 18, 19, 20, 21–2, 52, 53, 69, 167 Connolly, William 100, 111 consensus 44; coercive 30; and culture of conformity 32; and democracy 10, 14, 16; and discourse 33, 41; equitable 66, 67; existence of 41; and fallacy of stable state of affairs 25; and free-rider problem 66; Habermasian 66; loss of 70–1; necessity for 18; and pluralism 27; possibility of 21, 23; as pre-conditioned outcome 45–6; quality of 79; quest for 28–30; and rationality 14, 16; and sameness of participants 45; as stultifying 78; and truth value of decisions 16; and validity 14 constitutions 6–7, 144; American 150–1; British 168–71; Conventions of 172; and fear of majority rule 164; judicial 162–4; and liberalism 165; and reification of bad arguments 162; and undermining of democracy 164–5 Continental Congress 4–5, 43, 75, 80–1; administration/governance 86–92; broad description of
81–5; location of capital 92–7; Northwest Ordinance (1784) 86–7; Northwest Ordinance (1787) 90–2; settlement/ distribution of western lands (1785) 87–90 Cooper, J. and Brady, D. 80 Dahl, Robert 49, 72, 140, 144, 145, 147, 150, 162–3; and Tufte, Edward 30, 63–4 Davis, O. and Hinich, M.J. 76 decision-making 74, 77, 98; arbitrary 97; collective 19, 24, 46, 55; by committee 62; conditions of 50–5, 79; and deliberation 27; first-order 46; institutional limitations 68; and participation 24; popular 35; and preferences 33, 71; and rules 64, 66, 85; shaping of 80; and unanimity 66–7 deliberative democracy 1, 2; and Arrow’s theorem 36; and coercion 26; compatibility with social choice 46–9; and consensus 35; criticisms of 24–33; and decision-making 25; and deliberation/truth-telling 47–8, 49; and finding commonality 32; and the good 27–8; and if/then statements 46–7; and liberal democracy 26; and liberty/equality tension 26; and misunderstanding of the political 29–30; narrowness of 31–2; and (non)-effect of participation 24–5, 30–1; novelty of views 26; pluralism/consensus tension 24, 27–9, 32; and political antagonism 27, 28, 30; and preferences 29, 67; and public/private separation 27; and simultaneous action 31; and single-peakedness 48–9; and
Index 185 storytelling/rhetoric 32; and unrestrained discourse 33; and voting procedures 24–5, 26, 71 democracy, and being in a mob 149; break down in 145; and coercion 26, 30; compatibility with rights 144–5; and consensus 10, 14, 16; criticism of 166–7; and decision rules 64–5; defence/justification of 1, 3, 72–3, 145–6; defined 165; and deliberation 9, 72; and discourse 9–18; discursive/deliberative difference 15–16; as disorderly 43, 97; and duplicitous/ self-aggrandizing behaviour 42; and equality 64, 74; equality in 148–9; and false consciousness 17–18; and fear of tyranny 165; flaws in 173; and foundationalists 166; and free/equal dialogue 9; ideal conditions 21; and judiciary 173; and meta-rules 65; optimism/support for 38–46; and political awareness 168; and power 29; procedural requirements 14–15, 17; pure form 16; and rationality 10–13, 16–18; reform of 147–8; and removal of power 168; and rights 144–5, 146–7, 150; and rules/conditions of behaviour 11; and size 63–4; and treatment of citizens 144; as truth tracker 167; and truth value of statements 13–14, 16; undermining of 164–5; and voting 20, 21 democratic purity, abandoning 49–61; and challenges to Arrow’s IIA condition 50–5; and participatory conditions 49–50; and self-government 55–60; trade-offs 49; and U domain condition 49, 60–1
Dicey, A.V. 172 dictatorship 98, 100 discourse, common assumptions 2, 4, 8; and communicative competency 13; and conflict resolution 79; and deliberation 15–16; and democracy 9–18, 73; erosion of 78; and fair/equal participation 9; and the good life 16–17; and ideal speech situation 14–15; and justification 18; and pragmatic, ethical, moral rationality 10–13; and preferences 10, 17–18, 33; and reasoned deliberation 10; and resolution/agreement 14; rules concerning 15; rules limiting 63; similarities/differences with social choice 9, 35; and transformation through debate 9–10; and validity of statements 13–14, 16 Dryzek, John 8, 31; and List, Christian 16, 46–9 Dworkin, Ronald 163–4 Feinberg, J. 153, 156–7 Fishkin, J. 78 Flathman, R. 111 Foucault, Michel 15 free speech 7; and censorship 159–60; defence of 152; and erotica/pornography 153–4, 156–7, 159–60, 161; extinguishing of 151–2; feminist view 159–60; and harm principle 152–3, 155–8, 161, 162; and hate speech 154–5, 157–8, 159; limitations on 152–9; Millian 152–3, 160–1; and offence principle 156–8, 160, 161–2; support for 159; and truth 161; volatility of 151 freedom see liberty Frost, Samantha 137
186 Index Gauthier, David 100 General Possibility Theorem 19 Goodin, Robert 166–7 Gould, Carol 32 Green, D.P. and Shapiro, I. 80 Habermas, Jürgen 8, 34, 44, 45; and consensus 14, 16, 26–7, 28, 66, 70; and decision-making 32, 70, 79; and dialogue 46; and false consciousness 17–18; and free/equal dialogue 9–18; general theory of speech actions 13–14; and ideal speech situations (ISS) 14–15, 28, 36; and influence 70; and intersubjectivity 13; and listening to all point of view 24; “ought” recommendations 10–11; and paradigm of understanding 10; and pluralism 28, 32; as proceduralist 16–17; and rationality 10–13, 18; and rationality/morality tension 11–12; and voting procedures 26, 70, 71 Hardin, Russell 18–19, 20, 25, 37 Hobbes, Thomas 1, 21, 30, 67, 98; and absolutism 99, 136–9, 140; and American political system 104–5; and associational life 113–20; and censorship 109–13; and complacency 173; criticisms of 100–1; and desire for order 114, 115–17; and education 109–12, 138; and exclusion 137; and factions 116; and fear 138; and group fanaticism 115–16; and human nature 108; and individualism 113–14, 117, 119–20; intentions 105–27; and irregular systems 116–17; and knowledge 102; and laws 107–8, 111, 130–2, 136, 138–9; and liberty 105–9; and necessity of a state 113; and political deliberation 102; and political
power 129, 139; and property 120–7; and reason 138; and religion 135–6; and respect for others 111; rhetorical flourishes 115; and right of resistance 133; and safety of the populace 132–3; and self-regulation 119; and sovereignty 100–5, 122–4, 128, 129–30, 132–6; and state of nature 118–19; and state power 103–4; and subject/slave distinction 133; success of ideas 127–36; and tyranny 101, 103, 104, 130, 134, 139 Holmes, Stephen 150 ideal speech situation (ISS) 14–15, 28, 36, 43, 44, 45, 78 instability 3–4, 9, 28, 99; avenues for 23; and chaos 21, 34; Condorcet–Arrow difference 19; and creation of 64; cyclical 63; and equilibrium 67–8; and fair procedure 43, 48; and imposition of order 41–2; and institutions 65–6, 69; and preferences 44, 46; protection from 65–6; in the real world 34–5; and rules 64–5, 67–9, 72, 76–7; and structure of institutions 80 institution-free (IF) 1, 3, 8, 42, 43, 76, 77, 79, 79–80, 81, 98, see also Continental Congress Jillson, C. and Wilson, R.W. 82–3, 85
81,
Kahn, Victoria 100, 111 Kant, Immanuel 13, 149 Kateb, George 155–6, 159 Langton, Rae 159–60 laws, civil/natural 130–2, 136; Hobbesian 107–8, 111, 130–2, 135, 138–9; interpretation of 131–2; and liberty 106, 107–8
Index 187 liberty 57–9; and coercion 106; definition 105–6; individual 107; and laws 106, 107–8; Millian 160–1; natural/civil distinction 108–9; as physical 105–6; as positive good 108; protection of 171; UK versus US 169–70 Lijphart, A. 168, 171 Locke, John 113, 125–7 McCarthy, T. 16 McKelvey, Richard 8, 22, 23 Mackie, G. 9, 35, 37, 39, 68; and alternative preferences 38, 40; challenges to IIA condition 50–5; and consensus 40; and democracy 38, 42; and fairness 41; and (in)stability 39; and messiness of democracy 73; procedural view 72; and rational choice of self 44–5; and voting cycles 41, 42, 76 Madison, James 141–2, 143–4, 147, 161 majority rule, and abuse of individuals/minority groups 149; and argument for democracy 145–6; and coercion 149; and constitutional design 144, 150–1; and defence of rights 146–7; and democracy/rights compatibility 144–5, 150; and democratic equality 148–9; fear of 141–51, 164; and guardianship 147; and the harm principle 142–3; and judicial protection 147–8; and limiting of popular rule 141–2; as overstated 143–4; and prevention of tyranny 143–4; and self-government 145–6; and strength/weakness of state 150; superiority of 167; tyrannical 143–4, 147, 149–50 Mansbridge, Jane 26, 27
Marx, Karl 147 Mead, G.H. 17, 46 Mendelberg, T. 77 Mill, John Stuart 73, 78, 107, 117; and control of the majority 142; and free speech 152–5, 158, 160–2; Harm Principle 142–4, 152–3, 162; and plural voting 147 Miller, D. 72 Mouffe, Chantal 26, 27, 28–30, 141, 165–6 Mueller, D.C. 66 Munro, C.R. 170, 172 Nozick, Robert
121
Olson, Mancus
24–5
participation 2, 20; and acceptability of a norm 14; benefits of 20; conditions 49–50; educative effect of 73; effect of 24–5, 30–1; and equality/symmetry 9, 30–1; and feminism 31; and (in)stability 62–3, 64, 65; limiting 61–74, 99; opportunity costs 24–5, 71; sequential 30; and voting 20–3 Pateman, Carole 73 Phillips, Anne 31 Pildes and Anderson 60–1 Plato 33 Plott, C.R. 76 preferences, accuracy of 70; affect of deliberation on 77, 78; and assumption of unanimity 45; change/converge 40, 67–8; coherence of 40; exclusions 61; fixed/malleable 24; formation of 33–4; and human motivation 44; informing people of 17–18; revealing of 17; similarity in 39–40, 41; spread of 39; transformation/convergence of 10, 29, 40–1; unrestricted domain regarding 9
188 Index property 120–7; civil ownership 125–6; limits on accumulation of 124–5; rights/protection of 120–2; and sovereignty 121–5 public–private sphere 27, 46 pure majority rule (PMR) 4, 77, 80, 97, see also Continental Congress Radcliff, B. and Wingenbach, E. 73 Radner and Marshak 50–3 rational choice theory 45, 81 rationality 44; and agreement through dialogue 12; competing views 43; discursive/ instrumental 10–11; Habermasian 9–18; Hobbesian 138; and intersubjectivity 13–14; Kantian 12; and rational-choice/rational-agreement distinction 12–13; rules/conditions of behaviour 11; and speech actions 13; tension with morality 11–12; three domains 10–11; two-tiered concept 12; universal conception 11–12 Rawls, John 46; and consensus 26–7, 28, 64; and cooperation 138–9; as deliberative democrat 26; and distribution of primary goods, rights, resources 28; and liberalism 137; and pluralism 137; and public/private separation 27; and reasonable/person neutrality 27–8 Rescher, N. 23 Ridge, M. 100 Riker, William 8, 9, 23, 24, 33, 34, 37, 65, 77; accused of plagiarism 42; an preferences 76; and benefits of participation 20; and commendability of American system 39; and (in)stability of politics 48; obsession with outcomes 42;
and paradox of voting 21; and randomness/distortion 43; support for democracy 38–9; and tyranny 74; and will of the people 21 Rosenberg, S. 77 Rousseau, Jean-Jacques 26, 55 rules, of behaviour 11; and decision-making 64, 66, 85; and discourse 15, 63; and instability 64–5, 67–9, 72, 76–7; meta-rules 65 Ryan, Alan 105 Schmitt, Carl 27 Schofield, N. 76 self-government 55–60, 145–6; defined 55, 57; and desires 57–60; and freedom 57–9; individual 56–7; lack of 58–9; notions of 55–6; and reflective authenticity 56; society 56, 57 Sen, Amartya 18, 19, 61–2, 67 Shapiro, Ian 29, 63, 78, 173 Sharman, Campbell 169 Shepsle, K.A. 62, 65; and Weingast, B. 39, 76–7 Shklar, Judith 100, 139 SIE 80 social choice 1, 2, 30; and aggregate judgments 18; and Arrow’s theorem 19–20; and collective choice 19–20; common assumptions 2, 8; compatibility with deliberative democracy 46–9; criticisms of 33–5, 43, 44; and deliberation/truth-telling 47–8, 49; emaciated concept of democracy 34; and formation of preferences 33–4; as game of brinkmanship 61–2; if/then statements 46–7; and irrationality 23–4; and popular participation through voting 20–3; and pure majority rule 21;
Index 189 and randomness/instability 19, 22, 34–5; similarities/differences with discourse theories 9, 35; and single-peakedness 48–9; and stability 41–2, 43, 46; as threat to humanity 42; and voting 20, 21–2, 70 social welfare function (SWF) 19 Socrates 32–3 sovereignty 5–6, 141; and absolutism 100–5, 122–4, 128, 129–30, 132–6; and abuse of power 128–9; and censorship 109; democratic potential of 129; and dispersal of authority 129–30; Hobbesian 100–5, 122–4, 128, 129–30, 132–6; and the individual 128; kinds of 101–2; and liberty/coercion balance 106; and limiting of power 141; moral limitations 135; obligation to 134; and Parliament 171–2; petitioning of 135; and political power 103, 104–5; prerogative powers 126–7; and property 121–5; and religion 135–6; and safety of the populace 132–3, 134; and squandering of public resources 122–3; as threat to liberty 101; and use of force 134–5 stability see instability Steiner, J. 78–9; et al 78, 79, 97, 98 structurally induced equilibrium (SIE) 42, 61, 68–9 Sulkin and Simon 77 Sunstein, Cass 25–6
Tarlton, Charles 100, 107, 139, 169 Taylor, Charles 57–9, 146 Tocqueville, Alexis de 147, 161 transitivity 23, 68 Tullock, G. 76 tyranny 23, 74, 101, 103, 104, 126, 127, 130, 134, 139, 141, 143–4, 147, 149, 165 unanimity 45, 66–7, 77 universal domain 3, 24, 49, 50, 60–1, 79, 144 voting 68, 99; cycles of 41, 42, 43, 76; and information 25; and jury theorem 167; and loss of consensus 70–1; as manifestation of democracy 72; median 62; no ideal system 72–3; and pair-wise comparison 69–70; paradox of 38, 44; participation in 20; and policy outcomes 41; and popular choice/will 70, 71; tossing a coin analogy 166–7; and voter preference 39–40; and will formation 43 Weale, Albert 25, 26, 39, 68–9, 77–8, 150 Wellmer, Albrecht 16–17 White, S. 14, 15 Wittgenstein, Ludwig 166 Wolin, Sheldon 164–5 Young-Kemeny rule 54 Young, Iris Marion 31–2