The Rule of Law and the Separation of Powers [1 ed.] 0754624633, 2004057431, 9780754624639

Table of contents :
Cover
Half Title
Title Page
Copyright Page
Contents
Acknowledgements
Preface to the Second Series
Introduction
Defining the Rule of Law
The Rule of Law and the Circumstances of Politics
Arbitrary Rule and Rule by Law
The Rule of Law and Judicial Discretion
Judging on Principle
The Separation and Balance of Power: A Republican View
Democratizing the Judiciary
Conclusion
Notes
PART I: DEFINING THE RULE OF LAW
1 ON THE MORAL STATUS OF THE RULE OF LAW
I. REASONS RATHER THAN PROBABILITIES
II. SIMMONDS AND THE RULE OF LAW
A. Simmonds's First Major Retort: The First Strand
B. Simmonds's First Major Retort: The Second Strand
C. A Further Retort
2 RECONSIDERING THE RULE OF LAW
I. INTRODUCTION
II. THE RULE OF LAW AS A CONTESTED CONCEPT
A. The Instrumental Conception of the Rule of Law, or "How to do Things with Rules"
B. The Substantive Conception of the Rule of Law, or "How to Foster Liberty and Constrain Leviathan"
C. A Contested Concept
III. TRADITIONAL FORMALISM AND THE RULE OF LAW
A. Senses of Traditional Formalism
1. Formal connection between rule and particulars
2. Formal connection between foundations and rules
3. Formal connection between words and things
4. Formal realizability and the formalist conception of rules
B. Traditional Formalism and the Rule of Law
IV. THE WITTGENSTEINIAN CONCEPTION OF RULES AND ITS CONSEQUENCES (IF ANY)
A. The Wittgensteinian Social Conception of Rules
B. Consequences (if any) for the Model of Rules and the Rule of Law
l. Does Our Conception of Rules Matter?
2. Compromise Positions on the Model of Rules
3. On the connection between philosophy and normative social activity: the repercussions of reconceiving rules
V. TOWARD REINTERPRETATION OF THE RULE OF LAW
A. Liberal Legalism and Non-Formalist Natural Law
B. The Rule of Law and Interpretive Community: Toward a Pragmatic View
VI. CONCLUSION
3 THE RULE OF LAW AND ITS VIRTUE *
1. THE BASIC IDEA
2. SOME PRINCIPLES
3. THE VALUE OF THE RULE OF LAW
4. THE RULE OF LAW AND ITS ESSENCE
5. SOME PITFALLS
4 Formal and Substantive Conceptions of the Rule of Law: An Analytical Framework
1. The formal conception of the rule of law
(a) joseph Raz
(b) Dicey
(c) Unger
2. The substantive conception of the rule of law
(a) Dworkin
(b) Sir john Laws
(c) Allan
3. A middle way?
(a) Raz
(b) Jowell
4. Conclusion
5 IS THE RULE OF LAW AN ESSENTIALLY CONTESTED CONCEPT (IN FLORIDA)?
I. THE FLORIDA DEBACLE
II. ANALYTIC CONFUSION
III. A HISTORY OF CONTESTATION
IV. RULE-OF-LAW ISSUES IN FLORIDA
V. WHAT DOES "ESSENTIALLY CONTESTED" MEAN?
VI. THE RULE OF LAW AS AN ESSENTIALLY CONTESTED CONCEPT
VII. FINALLY, THE STREETS OF FLORIDA
PART II: THE RULE OF LAW AND JUDICIAL DISCRETION
6 Freedom and the Rule of Law
Preventing Arbitrary Coercion
Unavoidable Duties
Adapting British Traditions Ahroad
Independent Judges and State Policy
Dicey on Administrative Law
A Political Argument
7 JUDICIAL DISCRETION
8 THE THIRD THEORY OF LAW
9 INCOMPLETELY THEORIZED AGREEMENTS
INTRODUCTION
I. AGREEMENTS WITHOUT THEORY
A. In General
B. How People Converge
C. Rules and Analogies
II. JUSTIFICATIONS AND INSTITUTIONS
A. The Case for Incomplete Theorization
B. Judges, Theory, and the Rule of Law
C. Full Particularity?
Ill. HERCULES AND THEORY
A. An Ambitious Alternative
B. Conceptual Ascent?
C. Legitimacy
IV. INCOMPLETELY THEORIZED AGREEMENTS OVER TIME
A. Change
B. Disagreement
C. Principle, Politics, Law
10 Stability and Change in Judicial Decision-Making: INCREMENTALISM OR STARE DECISIS?
Politicallurisprwlence
Incrementalism
Incrementalism and Judicial Decision-Making
PART III: THE SEPARATION OF POWERS
11 The Political Form of the Constitution: the Separation of Powers, Rights and Representative Democracy
The Separation of Powers and Rights: a Preliminary Analysis
From Mixed Government to the Balanced Constitution
The Separation of Powers and the Modern Republic
Conclusion
12 Separation of Powers and Constitutional Government
The values served by the separation of powers
The doctrine under attack
The doctrine defended
The partial separation theory
Implications of the separation of powers
The separation of powers in the United Kingdom
The separation of powers and constitutional reform
13 ON SPEAKING SOFTLY AND CARRYING BIG STICKS: NEGLECTED DIMENSIONS OF A REPUBLICATION SEPARATION OF POWERS
I: The Concentration of Private Power
A. TOWARDS THE 'PURE DOCTRINE' OF THE SEPARATION OF POWERS
B. BEYOND THE PURE DOCTRINE
II: The Concept of the Separation of Powers
A. ON METHOD
B. FROM SPARTA TO MADISONIAN SEPARATIONS OF PRIVATE POWERS
III: Why Big Sticks Rebound
A.DETERRENCE FAILURE
B. EMOTION AND DEFIANCE
C. COGNITIONS OF STIGMA AND PROCEDURAL INJUSTICE
D. REACTANCE
E. RATIONAL COUNTER-DETERRENCE
F. THE DETERRENCE TRAP
IV: The Strength of Weak Sanctions
A. OVERDETERMINATION AND SOFT TARGETS
B. THE SOFT TARGET AT SOLOMONS
C. THE ACCOUNTABILITY MODEL
D. GENERALIZED RESTORATIVE JUSTICE AND SOFT TARGETS
V: Separating Enforcement Targeting from the Actor who Benefits from the Abuse
A. TARGETING GATEKEEPERS
C. THE PLETHORA OF THIRD PAR1YTARGETS
D. BUILDING A THOUSAND GATES TO THE POWER OF CORRUPT OFFICIALS
VI: Separating Powers Within and Between the Public and Private Sectors
A. SUMMARY SO FAR
B. PLURAL PRIVATE SEPARATIONS; PLURAL PUBLIC SEPARATIONS
VII: How to Separate Powers in the Private Sector
A. POWERS OF SHAREHOLDERS, DIRECTORS, MANAGERS
B. AUDIT AND AUDIT OF AUDIT
C. ENTERING THE CIRCLE WITH OUTSIDE POWER
E. THE STRATEGIES
VIII: Separation of Powers and Efficiency
A. TWO CONCERNS
B. OVERCOMPLIANCE
C. FISCAL IRRESPONSIBILI1Y
14 A REVISIONIST VIEW OF THE SEPARATION OF POWERS
I. Introduction
II. The Monopoly Analogue
III. The Model
(a) The Legislature
(b) The Executive
(c) The Game between Citizens, Parties and the Executive
IV. Extensions and Generalizations
V. The Separation of Powers
VI. Concluding Remarks
REFERENCES
15 Institutionalizing the Public Interest: The Defense of Deadlock and Beyond
THE FEDERALIST DEFENSE OF DEADLOCK
Making a Constitution and Defending It
Defining the Public Interest
Determining and Defending a Separation of Powers
U.S. Constitutionalism Revisited
ALTERNATIVE POLITICAL STRUCTURES
Interest-Based Partitions
Responsible Party Government
Legislative Unanimity
Direct Democracy
ALTERNATIVE CONCEPTIONS OF THE PUBLIC INTEREST
Reconceptualizations
Implementation
CONCLUSION
REFERENCES
PART IV: PARLIAMENTARISM AND FEDERALISM
16 THE NEW SEPARATION OF POWERS
I. DEMOCRATIC LEGITIMACY
A. Against Presidentialism
B. Constrained Parliamentarianism
C. The One-and-a-Half House Solution
II. FUNCTIONAL SPECIALIZATION
A. The Intellectual Challenge
B. Two Modest Proposals
C. Warring Separationisms
D. From Theory to Practice
E. Separationism and the Rule of Law
III. FUNDAMENTAL RIGHTS
A. The Democracy Branch
B. Safeguarding Fundamental Rights
IV. THE SHAPE OF THE NEW SEPARATIONISM
17 Constitutionalism and the Many Faces of Federalism
I. THE DIVISION OF POWERS BETWEEN THE CENTRAL GOVERNMENT AND THE COMPONENT ENTITIES IN A MULTI-SOVEREIGN CONSTITUTIONAL ORDER
A. Integrative Federalism
A.1. The European Community
(b) The division of powers in practice - ERASMUS as a case study
(c) The absence of a constitutionally protected nucleus of sovereignty for the Member States
(d) The relationship between Community and Member State powers
(e) Executive federalism within the European Community
A.2. Switzerland
B. Devolutionary Federalism
B.l. Belgium
B.2. Canada
B.3. Spain
II. THE JUDICIAL UMPIRING OF INTEGRATIVE AND DEVOLUTIONARY FEDERALISM
(a) The mechanisms of the judicial umpiring of the federal system
(b) The relationship between the judiciary and the legislature as to the umpiring of the federal system
Name Index

Citation preview

The Rule of Law and the Separation of Powers

The International Library of Essays in Law and Legal Theory Second Series Series Editor: Tom D. Campbell Titles in the Series: Freedom of Speech, Volumes I and II

Human Rights

Larry Alexander

Robert McCorquodale

Fernando Atria and D. Neil MacCormick

Law and Legal Interpretation

Anti-Discrimination Law Christopher McCrudden

Privacy Eric Barendt

Medical Law and Ethics Sheila McLean

Comparative Legal Cultures

John Bell

Constitutionalism and Democracy

Mediation

Carrie Menkel-Meadow

Environmental Law

Richard Bellamy

Peter S. Menell

The Rule of Law and the Separation of Powers Richard Bellamy

Criminal Law Thomas Morawetz

Contract Law, Volumes I and II Brian Bix

Thomas Morawetz

Corporate Law William W. Bratton Law and Democracy Tom Campbell and Adrienne Stone Legal Positivism

Tom D. Campbell

Administrative Law Peter Cane

Law and Language

Law and Anthropology Martha Mundy Gender and Justice Ngaire Najjine Seminal Issues in Mental Health Law

Jill Peay

Law and Economics Eric A. Posner Japanese Law

International Trade Law Ronald A. Cass and MichaelS. Knoll

J Mark Ramseyer

Sociological Perspectives on Law, Volumes I and II Roger Cotterrell

Declan Roche

Intellectual Property

Peter Drahos

Cyberlaw, Volumes I and II

Brian Fitzgerald

Family, State and Law, Volumes I and II Michael D. Freeman Natural Law

Robert P George

The Creation and Interpretation of Commercial Law Clayton P. Gillette Competition Law

Restorative Justice Constitutional Theory Wojciech Sadurski

Justice Wojciech Sadurski Regulation Colin Scott War Crimes Law, Volumes I and II Gerry Simpson Restitution

Lionel D. Smith

Freedom of Information Robert G. Vaughn Relocating Sovereignty

Rosa Greaves

Neil Walker

Law and Morality Kenneth Einar Himma and Brian Bi.x

Ernest J Weinrib

Chinese Law and Legal Theory Perry Keller

Rights Robin West

Constitutional Law

Ian D. Loveland

Tort Law

Welfare Law

Lucy A. Williams

The Rule of Law and the Separation of Powers

Edited by

Richard Bellamy University of Essex, UK

~ ~ ~~o~~~;n~~:up LONDON AND NEW YORK

First published 2005 by Ashgate Publishing Published 2016 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN 711 Third Avenue, New York, NY 10017, USA

Routledge is an imprint of the Taylor & Francis Group, an inform a business Copyright © 2005 Richard Bellamy. For copyright of individual articles please refer to the Acknowledgements. All rights reserved. No part of this book may be reprinted or reproduced or utilised 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. Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing in Publication Data The rule of law and the separation of powers. - (The international library of essays in law and legal theory. Second series) 1. Rule of law 2. Separation of powers I. Bellamy, Richard (Richard Paul) 340.1' I Library of Congress Cataloging-in-Publication Data The rule of law and the separation of powers I edited by Richard Bellamy. p. em.- (The International library of essays in law and legal theory (2nd series)) Includes bibliographical references. ISBN 0-7546-2463-3 (alk. paper) 1. Separation of powers. 2. Rule of law. I. Bellamy, Richard (Richard Paul) II. International library of essays in law and legal theory. Second series

K3173.R85 2004 340' .11--dc22 2004057431 Transfered to Digital Printing in 2011 ISBN 9780754624639 (hbk)

Contents Acknowledgements Series Preface Introduction: The Rule of Law as the Rule of Persons PART I

DEFINING THE RULE OF LAW

Matthew H. Kramer (2004), 'On the Moral Status of the Rule of Law', Cambridge Law Journal, 63, pp. 65-97. 2 Margaret Jane Radin (1989), 'Reconsidering the Rule of Law', Boston University Law Review, 69, pp. 781-819. 3 Joseph Raz (1977), 'The Rule of Law and its Virtue', Law Quarterly Review, 93, pp. 195-211. 4 Paul Craig (1997), 'Formal and Substantive Conceptions of the Rule of Law: An Analytical Framework', Public Law, pp. 467-87. 5 Jeremy Waldron (2002), 'Is the Rule of Law an Essentially Contested Concept (in Florida)?', Law and Philosophy, 21, pp. 137-64.

PART II

6 7 8 9 10

11

3 37 77 95 117

THE RULE OF LAW AND JUDICIAL DISCRETION

F.A. Hayek (1956), 'Freedom and the Rule of Law', Listener, pp. 989-90 13 December 1956, and pp. 1067-68 27 December 1956. Ronald Dworkin (1963), 'Judicial Discretion' ,Journal of Philosophy, 60, pp. 624-38. J.L. Mackie (1977), 'The Third Theory of Law', Philosophy and Public Affairs, 7, pp. 3-16. Cass R. Sunstein (1995), 'Incompletely Theorized Agreements', Harvard Law Review, 108, pp. 1733-72. Martin Shapiro (1965), 'Stability and Change in Judicial Decision-Making: Incrementalism or Stare Decisis?', Law in Transition Quarterly, 2, pp. 134-57.

PART III

vii ix xi

147 157 173 187 227

THE SEPARATION OF POWERS

Richard Bellamy (1996), 'The Political Form of the Constitution: The Separation of Powers, Rights and Representative Democracy', Political Studies, 44, pp. 436-56.

253

vi

The Rule of Law and the Separation of Powers

12

Eric Barendt (1995), 'Separation of Powers and Constitutional Government', Public Law, pp. 599-619. 13 John Braithwaite (1997), 'On Speaking Softly and Carrying Big Sticks: Neglected Dimensions of a Republication Separation of Powers', University of Toronto Law Journal, 47, pp. 305-61. 14 Geoffrey Brennan and Alan Hamlin (1994 ), 'A Revisionist View of the Separation of Powers', Journal of Theoretical Politics, 6, pp. 345-68. 15 Robert E. Goodin (1996), 'Institutionalizing the Public Interest: The Defense of Deadlock and Beyond', American Political Science Review, 90, pp. 331-43. PART IV

16 17

297 355 379

PARLIAMENTARISM AND FEDERALISM

Bruce Ackennan (2000), 'The New Separation of Powers', Harvard Law Review, 113,pp.634-729. Koen Lenae1ts (1990), 'Constitutionalism and the Many Faces of Federalism', American Journal of Comparative Law, 38, pp. 205-63.

Name Index

275

395 491

551

Acknowledgements I am grateful to Dario Castiglione for his advice on the selection of the articles and to Louise Hawkridge for her efficiency in tracking them down. Dario also offered invaluable comments on my introduction, as did members ofthe Glasgow Law Department seminar, especially Adam Tomkins, and my colleagues in the Essex Political Theory Reading Group, Jason Glynos, David Howarth, Aletta Norval, Albert Weale and most particularly Sheldon Leader. The editor and publishers wish to thank the following for permission to use copyright material. American Journal of Comparative Law for the essay: Koen Lenae1ts ( 1990), 'Constitutionalism and the Many Faces of Federalism', American Journal of Comparative Law, 38, pp. 205-63. Copyright © 1990 American Association for the Comparative Study of Law, Inc. Blackwell Publishing Limited for the essays: Richard Bellamy (1996), 'The Political Form of the Constitution: The Separation of Powers, Rights and Representative Democracy', Political Studies, 44, pp. 436-56; J.L. Mackie (1977), 'The Third Theory of Law', Philosophy and Public Affairs, 7, pp. 3-16. Cambridge University Press for the essay: Robert E. Goodin (1996), 'Institutionalizing the Public Interest: The Defense of Deadlock and Beyond', American Political Science Review, 90, pp. 331--43. Copyright© 1996 American Political Science Association. Harvard Law Review for the essays: Cass R. Sunstein (1995), 'Incompletely Theorized Agreements', Harvard Law Review, 108, pp. 1733-72; Bruce Ackerman (2000), 'The New Separation of Powers', Harvard Law Review, 113, pp. 634--729. Journal of Philosophy for the essay: Ronald Dworkin (1963), 'Judicial Discretion', Journal of Philosophy, 60, pp. 624-38. Kluwer Academic Publishers for the essay: Jeremy Waldron (2002), 'Is the Rule of Law an Essentially Contested Concept (in Florida)?', Law and Philosophy, 21, pp. 137-64. Copyright © 2002 Kluwer Academic Publishers. Matthew H. Kramer (2004 ), 'On the Moral Status of the Rule of Law', Cambridge Law Journal, 63, pp. 65-97. Copyright © 2004 Matthew H. Kramer. Listener for the essay: F.A. Hayek (1956), 'Freedom and the Rule of Law', Listener, pp. 98990 13 December 1956, and pp. 1067-68 27 December 1956.

viii

The Rule of Law and the Separation of Powers

Margaret Jane Radin (1989), 'Reconsidering the Rule of Law', Boston University Law Review,

69, pp. 781-819. Copyright© 1989 Margaret Jane Radin.

Sage Publications for the essay: Geoffrey Brennan and Alan Hamlin (1994), 'A Revisionist View of the Separation of Powers', Journal of Theoretical Politics, 6, pp. 345-68. Sweet and Maxwell for the essays: Joseph Raz (1977), 'The Rule of Law and its Virtue', Law Quarterly Review, 93, pp. 195-211; Paul Craig (1997), 'Formal and Substantive Conceptions of the Rule of Law: An Analytical Framework', Public Law, pp. 467-87; Eric Barendt (1995), 'Separation of Powers and Constitutional Government', Public Law, pp. 599-619. University of Toronto Press Incorporated for the essay: John Braithwaite (1997), 'On Speaking Softly and Canying Big Sticks: Neglected Dimensions of a Republication Separation of Powers', University of Toronto Law Journal, 47, pp. 305-61. Every effort has been made to trace all the copyright holders, but if any have been inadvertently overlooked the publishers will be pleased to make the necessary anangement at the first oppmtunity.

Preface to the Second Series The first series of the International Library of Essays in Law and Legal Theory has established itself as a major research resource with fifty-eight volumes of the most significant theoretical essays in contemporary legal studies. Each volume contains essays of central theoretical importance in its subject area and the series as a whole makes available an extensive range of valuable material of considerable interest to those involved in research, teaching and the study oflaw. The rapid growth of theoretically interesting scholarly work in law has created a demand for a second series which includes more recent publications of note and earlier essays to which renewed attention is being given. It also affords the opportunity to extend the areas of law covered in the first series. The new series follows the successful pattern of reproducing entire essays with the original page numbers as an aid to comprehensive research and accurate referencing. Editors have selected not only the most influential essays but also those which they consider to be of greatest continuing importance. The objective of the second series is to enlarge the scope of the library, include significant recent work and reflect a variety of editorial perspectives. Each volume is edited by an expert in the specific area who makes the selection on the basis of the quality, influence and significance of the essays, taking care to include essays which are not readily available. Each volume contains a substantial introduction explaining the context and significance of the essays selected. I am most grateful for the care which volume editors have taken in carrying out the complex task of selecting and presenting essays which meet the exacting criteria set for the senes. TOM CAMPBELL Series Editor Centre for Applied Philosophy and Public Ethics Charles Sturt University

Introduction When judges criticize legislatures and governments they typically do so in the name of the Rule of Law. Unfortunately, politicians do the same when they attack judicial activism. For the first, the Rule of Law generally refers to the judiciary's authority to determine the state of the law in a given case, to pass sentence and to review legislation for its compatibility with prevailing legal norms, including constitutional rights. Any attempt by politicians to contest or interfere with these powers is condemned as undermining the integrity of the legal system. 1 For the second, however, the Rule of Law refers to the right of a legally authorized government to pass laws according to the due formalities and to have them obeyed. Judicial moves to reinterpret or question duly passed laws usurp the government's right to legislate. On this view, the prime danger to consistency in the legal process comes from courts aspiring to make the laws rather than simply applying them. In the process, people's ability to influence and contest legislation via the ballot box becomes undermined. 2 Given the self-serving and apparently incompatible character of these two positions, it is perhaps small wonder that some theorists have regarded the concept as at best 'essentially contested' (Waldron, this volume, Chapter 5) and, at worst, no more than a 'meaningless ... self-congratulatory rhetorical device' (Shldar, 1987, p. 1). As the two accounts outlined above indicate, different views of the respective powers of the legislature and the judiciary and the nature of their separation lie at the heart of debates regarding the Rule of Law. Any theory has to somehow reconcile their rival claims and show whether there is any meaningful and valuable sense in which the law, rather than judges or politicians, might be said to rule. This Introduction sketches the main lines of argument found in the essays reproduced in this volume and proposes a political view of the Rule of Law as arising from a balance of power within, as well as between, the legislature and courts. The first section of this Introduction explores the nature of law and its relationship to political power. I argue that the Rule of Law has to be distinguished from both the Rule of Good Law and rule by law. Because notions of the good and just are subject to political disagreement, the Rule of Law cannot be associated with some notion of the good or just that lies outside the circumstances of politics and is not the product of any political or legal authority. However, merely to rule by law proves insufficient to curtail the arbitrary abuse of power. Consequently, both court-centred and legislature-centred approaches must confront what I dub the Hobbesian challenge: namely, how can they avoid empowering either courts or legislatures as a sovereign power, able to act in an arbitrary way? The former seek to do so by offering what might be called 'a third theory of law', lying midway between natural law and legal positivism. 3 They have to show how the very nature of law implies certain norms that can guide in all cases, directing judicial discretion so that law can indeed rule and legal forms bind those governments committed to employing them. I explore two attempts at such a theory in the second section. The latter have to indicate how the political process itself might lead legislators not to rule in their own or their supporters' interests alone. They must also explain how and why judges and other officials might follow their lead. I defend a version of this position in the third section. The legal, court-centred view attempts to distinguish law strictu sensu from legislation. The political, legislature-centred

xii

The Rule of Law and the Separation of Powers

view suggests that no such distinction can be made. Judges and politicians think and act in similar ways. The key is to provide each side with the motivation to treat all as equals under the law when making their decisions. This approach tackles the problem of arbitrariness at its source by so dividing power that it cannot be employed in a sovereign, and hence potentially arbitrary, manner. In this regard, many of the attributes associated with the Rule of Law tum out to depend on the background social and political conditions fostering democratic rule more generally. Indeed, in many crucial respects the Rule of Law is simply rule by democracy.

Defining the Rule of Law This section of the Introduction sets the scene for the essays collected in Part I of the book. Matthew Kramer's essay (Chapter 1) revisits the classic debate between legal positivists and theorists of natural law (and their secular counterparts) regarding the morality of law. The essays by Margaret Radin, Joseph Raz and Paul Craig (Chapters 2, 3 and 4) all largely follow Kramer in developing a largely formal and limited notion of legality as rule by law, rather than the rule of law. In Chapter 5 Jeremy Waldron goes even further, suggesting that even this formal notion is contested, placing the rule of law within the circumstances of politics - a view elaborated below.

The Rule of Law and the Circumstances of Politics If we were ruled by philosopher kings, who simply discovered the law through their devotion

to truth and justice and could be counted on to apply it with an angelic rectitude and a divine omniscience, the Rule of Law would be unproblematic. The Law's agents, whether we called them monarchs, legislators or judges, would be free from the uncertainties, as well as the biases, that motivate political debate. They would be the mere mouthpieces of a superior wisdom that infallibly acted to promote just solutions that harmonized the individual's interest with that of the public. Clearly this is a fantasy, yet many versions of the Rule of Law have a tendency to embrace it. The danger lies with potentially bad political rulers. Enshrine good laws in the constitution and entrust a special caste of legal guardians to oversee them, and the rule of persons can be subordinated to the Rule of Law. Unfmtunately, this solution simply begs the very questions that lie at the heatt of the problem. As Joseph Raz (Chapter 3) has noted, many accounts of the Rule of Law use the term as a catch-all slogan for every desirable policy one might wish to see enacted. He cites, as an instance of this approach, the International Congress of Jurists' equation of the Rule of Law with the creation and maintenance of 'the conditions which will uphold the dignity of man as an individual'- a requirement that includes 'not only recognition of his civil and political rights but also the establishment of the social, economic, educational and cultural traditions which m·e essential to the full development of his personality' (p. 77). However, people reasonably disagree over the nature of the right and the good. Although most people will find the general aim of the ICJ s' declaration unexceptionable, many will differ over what human dignity consists in and requires. Likewise, they will dispute the vm·ious ways in which such aims of making all equal under the law might be achieved- including what formal and substantive procedures and entitlements would need to be in place. These differences also lie at the heart of the divergent

The Rule of Law and the Separation of Powers

xiii

views of the Rule of Law outlined earlier: whereas some see human dignity and equality being best promoted through democracy, others will regard it as best protected via the courts and bills of rights (see Waldron, this volume, pp. 120-28). In other words, the Rule of Law lies within the circumstances of politics; it represents a necessary collective agreement in an area where there is deep disagreement over what its character might be and where no demonstrably best solution is available (Waldron, 1999, pp. 107-18; Weale, 1999, pp. 8-13). Amore fonnal approach attempts to overcome these difficulties. 4 This focuses on the benefits, values and constraints inherent in the very existence of legal forms and procedures. After all, people often obey laws with which they disagree simply out of respect for the advantages of living in a law-governed environment. Law facilitates social interaction and helps curb the abuse of power. It can even provide regular procedures for contesting and changing laws and decisions which they dislike. None of these considerations needs to involve substantive agreement with the goodness of the law. Yet, as we shall see in the next part of this section, legality per se can only play a limited role in protecting against oppression and domination. To achieve this end, we may need good laws. Moreover, it is not clear that even good laws can rule in and of themselves- covering all possible eventualities and infallibly guiding those entrusted to uphold it towards the right answer. Does this return us to the original dilemma of relying on good men to give us good laws? This question brings us to what can be termed the Hobbes challenge. Stated crudely, Hobbes argued that, in circumstances of conflicting interests and deep disagreements about values and judgements, laws would only be equitably and coherently drafted and applied by all individuals being equally in awe of a sovereign who was outside the law and whose power was indivisible. For a start, the meaning oflaws is rarely clear, so that 'all Laws, written and unwritten, have need oflnterpretation' (Hobbes, 1991, p. 169), and these interpretations are usually controversial. Even when the meaning of laws is clear, their bearing on patticular situations often is not producing another source of controversy. Indeed, even if laws could be defined absolutely clearly in terms of both their meaning and application, self-love, pattiality and passion can lead people to employ them in self-serving ways and hence into conflict with each other. Hobbes believed that these difficulties arise as much with the hypothetical imperatives of the Law of Nature as with human laws. Therefore, laws or rules do not themselves provide the basis for social cooperation. Rather, a peaceful society results from having a political authority vested with the power to formulate, interpret and apply the laws and, crucially, to oveiTule rival views of their bearing in any given case. The claim that laws could be set above the sovereign person or persons empowered to enact and implement them was incoherent for Hobbes. It could only mean to set up another power able to judge and enforce these laws 'which is to make a new Sovereign' leading to an infinite regress with the need of 'for the same reason a third, to punish the second; and so continually without end, to the Confusion, and Dissolution of the Commonwealth' (Hobbes, 1991, p. 169). On this view, law is not only subject to political contestation but intrinsically political by virtue of being the creation of political authority. Of course, as H. L.A. Hart observed, the sovereign's rule is recognized in patt because it has legal foundations. In his terminology, we need 'secondary' rules of recognition to identify what counts as law as well as the 'primary' rules or laws that m·e created within a recognized legal system (Hart, 1994, pp. 94-95). Among these secondary rules are those identifying the sovereign as authorized to decide what the law is. Such rules do place a legal constraint on personal rule, but not necessarily a very strong one. It is not just that tyrants have a habit of securing legal

xiv

The Rule of Law and the Separation of Powers

legitimacy for their rule after, rather than before, seizing power. Political rulers of all stripes habitually claim that the Rule of Law simply requires obedience to their commands as the legally recognized authorities. Yet if government is, by definition, the agency authorized by law to issue laws, then, on this interpretation, the Rule of Law barely constrains the rule of persons. To say that all acts of government must have a foundation in law becomes almost a tautology. If law is merely (and only) what the lawful government decrees, then anything the lawful government decrees is authorized by law and what is not authorized or so decreed is illegal and so cannot be an action of the government (see Raz, this volume, pp. 78-79). Naturally, it would be wrong to dismiss this view completely. After all, British courts have frequently done a valuable job in restraining ministers and officials from acting beyond their remit. A government (or other body) that has to act within and according to its legally defined powers can be held to account in ways in which bodies or people that are not so constrained cannot. The rule of the socially influential and wealthy, of the mafia or of the mob, reveal in their different ways the disadvantages of lawless compared to lawful government. However, here the Hobbes challenge re-emerges since some body of people, be they judges or other politicians, have to decide whether the government has breached the rules. And although they, too, will be legally constituted, at some point some person or persons possesses the competence to decide questions regarding its own competence. 5 Meanwhile, an underlying set of problems remains: namely, why should the powerful have accepted to be rule-bound in the first place, unless it had somehow become in their interest to do so, and how can the legal forms of legislation and judging ensure that the law prevails rather than the will of those authorized to make decisions? Therefore, the central dilemma posed by the Rule of Law is how to enjoy the benefits of legality while overcoming the Hobbes challenge. The law cannot rule without legislators and the judiciary. Can the rule of persons be modified by the Rule of Law if they are necessarily the instruments through which it governs? Legal constitutionalists claim that, by ruling through law, those who make and enforce it are somehow led to be ruled by it in their tum. Fidelity to law is an intrinsic aspect of a coherent legal system, and, for that to obtain, the laws and judging must possess certain desirable qualities (Fuller, 1969, ch. 2). I shall dispute this view, arguing that it ends up perilously close to the fantasy of a system of good laws under the jurisdiction of beneficent philosophical kings. By contrast, I shall contend that a political constitutionalism tackles the Hobbes challenge at its source- the problem of the sovereign ruler. Turning rule by law into the Rule of Law is largely a matter of the process of legislation rather than a matter of the form or nature of law per se. It is the political system that de-sovereigntizes sovereign power. Of course, the issue of how judges and others interpret and apply the law remains. Here, too, the solution largely lies in their place within the political system and the incentives it offers. However, the outcome is a less heroic view of judging than many jurists suppose- one much more sensitive to the disagreements of ordinary citizens and the political mechanisms which they employ to resolve them. Arbitrary Rule and Rule by Law No theorist of the Rule of Law believes that rule by persons can be avoided in all respectsmerely that certain types of personal rule can be placed outside, or in certain respects are simply incompatible with, any workable system of law. Similarly, no critic of the Rule of Law

The Rule of Law and the Separation of Powers

XV

denies that there are many advantages to rule by law - merely that these in themselves pose only mild constraints on personal rule, not least because they are inseparable from it. Therefore, to assess the respective merits of political and legal constitutionalism, we need to identify the specific dangers posed by the rule of persons, on the one hand, and the attributes, benefits and potential weaknesses and lacunas of rule by law, on the other. The chief danger associated with the rule of persons is the possibility of arbitrary rule (see Raz, this volume, pp. 84--85). Arbitrariness, in this context, involves the capacity of rulers to coerce, obstruct, manipulate or otherwise impact upon the ruled without consulting their diverse views or interests and showing them equal concern and respect. As a result, not all are equal before the law either formally or substantively. Such arbitrariness manifests itself in three main ways. First, arbitrary rule can involve wilful or capricious government. In which case, there will be no consistency or coherence to policy, so that people will not know where they stand. For example, an act that had hithe1to appeared unexceptional might suddenly, and for no apparent reason, attract a severe penalty simply because the ruler has taken a dislike to it or become unusually attentive. Indeed, arbitrariness in this sense can result from unconstrained judgement more generally. Even a good decision to refonn a poor or unfair practice, say, can be the cause of injustice in its tum if it can be acted on at will and without any warning or consultation in ways that unsettle established expectations. Second, arbitrariness can stem from oppression, as when people's legitimate expectations and needs are ovenidden or ignored out of pure maliciousness or for the sake of another's self-interested ends. Finally, the presence of a ruler capable of acting arbitrarily in a systematic way entails domination. 6 The capacity for arbitrary interference is never absolute. No agent can always choose when, where, how and who to interfere with. However, so long as there is the potential for regular, if not total, interference we will be dominated by the ruler. 7 The ruled become permanently in awe of the power of their rulers, attempting to second-guess their next move - either to escape their wrath or win their favour. Although linked, one kind of arbitrariness can be present without necessarily involving the others. For example, an enlightened despot may not act capriciously or oppress people's interests but, by virtue of his comparatively unchecked power, will still exert domination over them. Likewise, bureaucratic machines may not be wilful yet nonetheless be both oppressive and dominating. Indeed, avoiding this situation may require an occasional wilful disregard of the rules on the part of an official. Yet that ability places him in a potentially dominating role in his tum. As we shall see, this tendency for the solution to one form of arbitrariness to generate a different form proves to be a general problem. The effect of all three kinds of arbitrary rule, whether singly or in combination, is to reduce the ruled to slaves of their ruler by removing their ability to act as free and equal citizens. Individuals lose the capacity to act autonomously and plan ahead on terms which they can accept as fair for all. Instead, they have the anxiety of being subject to unpredictable, possibly malicious and usually prejudiced and unfairly discriminatory interference by the public authorities and their supporters and/or the inadve1tent, yet preventable, or deliberately malign actions of others. How far does law per se mitigate arbitrariness? Law can be broadly defined as a set of authoritative coordinating rules, regulations, directives and constraints that are backed by punitive sanctions and, less commonly, by preventive restraints (Finnis, 1980, p. 226). When designating what is required to rule by law, most commentators offer elaborations on, or synthetic summaries of, the eight desiderata listed by Lon Fuller as expressing the 'internal morality of law'. R These

xvi

The Rule of Law and the Separation of Powers

conditions consist of ce1tain largely formal characteristics that are used to distinguish a system of laws from either a number of ad hoc directives or a collection of customary practices. Some of these qualities relate to the fonn of law, others to the form of the legal system. With regard to the first, standard criteria include that laws should be general and apply equally to all (though, as we shall see, they may need to take into account relevant differences), be prospective (only invoking retroactivity as a curative measure), public (albeit often through publicly funded experts), clear (avoiding vague tenninology open to wide discretionary interpretation) and relatively stable (but not so as to ossify). So far as the second is concemed, laws should be consistent with each other, feasible and congment with official action. Most importantly, the promulgation, execution and ability to contest the laws should be guided by general, equal, public, clear and relatively stable mles. There are numerous familiar legal norms and practices that exemplify these formal qualities of the law, on the one hand, and the legal system, on the other. For example, those associated with the first are reflected in the notion that there can be no crime without a law, while those linked to the second underpin the view that the court system should be free from extraneous pressures and independent of all authority but the law, and operate according to settled and open procedures that are the same for all citizens. These features give law the regularity needed for it to be a system in any moderately well functioning sense. Without at least some of these elements, neither those subject to the laws nor those appointed to implement them will be able to follow and administer the law respectively in any sort of systematic way. To the degree that mle is by laws which satisfy these criteria, then mle by law will generate a weak form of the Rule of Law. At least for much of the time, people will comply with the law not out of deference to (or fear of) the persons who have made or are upholding it, nor necessarily because they agree with it, but simply because it is the law. Law so understood will be necessary to some degree in societies of any complexity. It satisfies the need for common decisions that can coordinate the collective life of the community in conditions of reasonable disagreement over the nature of the good society. There are many aspects of social life where having almost any mle outweighs the perils or inconvenience of having none. Those with sharp elbows may not like the mle of 'first come first served' in shops, but most people benefit from the anangement and have a less stressful time as a result. Continuous white lines may annoy those drivers who claim to 'know' the road or to possess super-fast reactions, but the rest of us have an interest in such people being severely detened from acting on this belief and in obeying the law. Even in areas which are not collective action problems and involve more contentious moral issues, the need for a common mle so that we know where we stand can lead us to accept laws we dislike - although, as I noted earlier, the place of that law within a more general system of legality (usually including mechanisms for changing or challenging it) also influences our acceptance. A regular legal system generates a degree of predictability in social life by stabilizing, regulating, securing and even constituting numerous valuable relationships. Most obviously, the presence of law reduces certain risks by protecting us from various potentially harmful, yet unforeseeable, acts by others. Criminal law achieves this reduction by attempting to deter specific wilfully harmful acts. However, by coordinating activity, laws can also help prevent those harms that would otherwise arise simply from operating in an umegulated environment. For example, a traffic regulation directing that we all drive on one side of the road rather than the other undoubtedly reduces accidents by virtue of authoritatively deciding the issue and removing unce1tainty. Indeed, without law, many social relationships and interactions would

The Rule of Law and the Separation of Powers

xvii

be insecure, prone to collapse, liable to develop in erratic and unpredictable ways, or simply not exist. Thus, marriage law promotes, and to some extent creates, a certain kind of social bond, shaping its development and shoring it up against those social forces that might erode or destabilize it, as well as partly regulating behaviour within it. Partners are given a clear understanding about what happens should they separate, one of them dies unexpected! y and so on. People are secured against various potential pitfalls of joint property ownership, merging their incomes, starting a family and the like, and so acquire the confidence to enter into such relations in the first place. Likewise, much commercial law not only protects us against fraud and particular smts of economic loss, it actually facilitates trading and trust between buyers and sellers by informing each party as to the terms of their contract and holding them to it. One does not have to know the prevailing social customs of a given locality or learn and pander to the whims of the individuals with whom you are dealing in order to know one's rights and obligations. In a world where much human interaction is with numerous strangers and rather faceless yet powerful institutions -from state bureaucracies to big corporations -these are real gains. Indeed, as the case of marriage shows, even in the most intimate relationships a clear understanding of one's basic entitlements and duties, should things come unstuck, can be important. Plainly, a ce1tain complementarity obtains between the promotion of stability and predictability through law and the reduction of arbitrariness. To the extent that officials and other powerholders follow prearranged rules, arbitrariness in all three senses is diminished because these rules will restrict their room for manoeuvre. Keeping a government by and large within legal bounds protects against the 'lawless authority' that Dicey famously associated with the Bastille where a figure such as Voltaire could be imprisoned solely at royal whim (Dicey, 1959, pp. 189-91 ). Indeed, even if rulers (or our fellow citizens) possess good judgement on a given matter that is superior to the prevailing law, we can still prefer that they abide by the comparatively imperfect rule and only improve upon it by following established procedures. Forcing policy changes to be made by a due process gives people warning of any new regulations and time to adapt to them. It also guards against those occasions when even good and wellintentioned rulers make mistakes. To the extent that wilfulness promotes oppression and domination, constraining this form of arbitrary rule will also reduce the others. Even under a tyrant, one may be able to avoid falling foul of the law, while a regard for legality, however superficial, will make the persecution of a particular individual more cumbersome if the due formalities are to be observed. Law also embodies a certain rough and ready form of equality. If we are to know how to react to the law so that it can guide our actions and interactions with unknown others, then we have to assume that it will have the same implications for everyone to whom it applies. Nevertheless, how deeply equality before the law reaches is a moot point. The question of whether law is just an instrument that can be employed for any purpose, or whether oppressive and exploitative regimes are incompatible with law, has spawned a huge literature. 9 Clearly, even a tyranny within a modem society will need laws to regulate many of the activities that are incidental to its purposes, such as traffic. But some theorists believe that tyrants seek to act in ways that prove antipathetic to a working legal system- they want discretion to do whatever they want, to be able to show pmtiality towards their suppmters and to persecute whomever they dislike. All these activities m·e said to involve not placing themselves on a par with their fellow citizens as equals under the law. As a result, they end up unde1mining law because it

xviii

The Rule of Law and the Separation of Powers

ceases to offer a reliable guide to action. Why should I even stop at red traffic lights if many of my fellow drivers can flout it with impunity? If waiting for a green light gives me no greater assurance that I won't be crashed into and blamed for the incident than going through the red, I might as well rely on my own judgement and go whenever I think the way is clear. It cannot be doubted that tyrants typically use the law for largely cosmetic purposes when pursuing their evil ends. Nor do they usually see all, if any, of their fellows as equals. Nevertheless, the general benefits of legality as a system of coordination and legitimation may make them wary of departing from its forms too obviously for fear of so undermining respect for law that people only act when ordered to do so by armed officials. For rule by coercion-backed commands alone is highly inefficient and makes much of social life difficult to sustain. In other words, even tyrants will want traffic laws to be obeyed. They will also wish to provide citizens with regular incentives to comply with some of their more iniquitous policies and to coordinate the actions of the officials responsible for their implementation. Of course, tyrannical regimes standardly possess an army of secret police, loyal only to the ruler and operating outside the rules governing the ordinary forces of law and order. However, they would cease to be 'regimes' if they relied on these agencies alone to enforce their will (and even these groups usually operate to a degree within the (admittedly despicable) laws of the land). 10 Moreover, it is not obvious that operating outside the law, or according to different rules, in some areas, necessarily precipitates a general decline of law-abidingness throughout the legal system. After all, no legal system is perfect. Part of the problem is that some version of the tyrant's use of discretion, discrimination and use of special laws seems unavoidable if the law is not to be arbitrary in its tum. Take the simplest expression of the notion of equality before the law- the idea, that held sway for a long time in British law, that there should be 'one law for all' (Dicey, 1959, p. 193; Waldron, 1990, pp. 40-41 ). This notion seems to neatly encapsulate the ideal of the Rule of Law. If law covers all and is the same in both content and application for all, then surely none is above it and law rules. Bias and self-interest should be ruled out because a tyrant who genuinely believed in law would have to apply it as severely to himself and his followers as those he wishes to persecute. Unfortunately, matters are not so neat. Let's start with discretion. The almost mechanical application of rules, however well fonnulated, will not always be a good thing. An officious insistence on the letter of the law is often itself arbitrary and productive of injustices, even when the rule itself is well motivated and just. As anyone who has ever had to deal with a bureaucracy knows full well, rule-bound procedures that appear excessive, or in some way inappropriate to the task at hand, exert a stultifying domination over administrators and their clients alike. The domination of arbitrary impersonal rules may be less degrading than that of an arbitrary ruler but they can prove as irksome, if not more so through being all-pervasive. Indeed, turning an unjust or plain inept policy into law may make its enforcement all the more regular. If the consistent application of unjust laws means that one can learn how to get around or evade them (Rawls, 1971, p. 59), consistency also prevents sympathetic judges and officials quietly bending the rules to avoid injustices (Waldron, 1989, p. 82). Within an evil regime, obedience to the law may allow officials to excuse themselves of any moral responsibility for the acts they commit. Moreover, it is not just bad rules that are the problem. There can also be an absence of rules - after all, it is hard to anticipate all eventualities. In many cases, the discretionary exercise of good judgement will be preferable to no decision or the attempt to adapt an unsuitable existing rule.

The Rule of Law and the Separation of Powers

xix

The universality requirement is sometimes thought to rule out discrimination. Certainly it would be hard, if still possible, to formulate a general law, applying equally to all, that is aimed solely at a particular individual. It is fairly easy, though, to construct such laws so that they discriminate against groups. Nazi race laws and South African apartheid laws had this fonn, applying, as they did, to all Jews and non-whites. True, if the laws apply also to the rulers, they may make them less harsh or allow for exemptions that benefit others beside themselves but it all depends how ideologically pure and committed they are - after all, prominent Italian Jewish fascists were not spared the rigour of Mussolini's late conversion to Hitler's Aryan ideal. Meanwhile, although racist and similar laws are rightly condemned, it is unclear whether all laws singling out groups as different should be criticized. Surely the fact that all cases are not alike may often require laws that do differentiate between groups in order to avoid discrimination. Rules regarding matemity leave, for example, have been generally placed in this category. What about the special laws goveming party members and the political police? Or those laws that place the leader outside the law? Once again, these exceptions to 'one rule for all' do not appear unjustified in themselves. Officials often undertake tasks which one would not wish ordinary citizens to be allowed to do, such as being able to detain and intenogate those suspected of criminal activities. Don't we need special laws to regulate such special powers, granting dispensations from ordinary legal constraints but setting possibly higher standards according to which such powers must be exercised? Indeed, having literally the same laws for everyone in all situations seems absurd. Most laws are framed in a conditional form so as only to apply to all those engaged in a given activity in certain circumstances, and not to everyone all the time. Health and safety standards in private houses are different to those in hotels, adults and children are treated differently and so on. As to there being some powers literally outside the law, executives are standardly accorded exceptional powers in an emergency that allow them to go beyond - or even suspend - the law in order to preserve it. We seem to be faced with a paradox. The capacity of law to avoid the arbitrariness of the rule of persons seems to rest on its being a rule-bound system that makes everyone equal before the law. Yet, without discretion -laws that treat different people differently and special laws that place some under different laws to others -it risks becoming arbitrary in its tum. Law has to be flexible enough to make morally or pragmatically justified distinctions and discriminations, without becoming so flexible that even a tyrant can use it. There also has to be some scope for discretion in the event that no law appears (at least prima facie) to apply. The key is whether the form or inner morality of law can guide the legitimate use of such discretionary and differential treatment in ways that preserve the consistency and integrity of law as a system. If so, then law may be said to rule. However, underlying the notion of the Rule of Law is arguably a set of values that goes deeper than anything law in itself can provide. It is an inherent property of rules that they treat people in the same way, and, as we saw, this feature can yield many desirable benefits. But there is more to equal treatment than treating everyone the same. We need to see other people as equals. There are elements of impartiality and of reciprocity involved here. On the one hand, we should endeavour to avoid prejudice or self-interest in our judgement of others. On the other hand, we should only engage in collective enterprises which all can accept and undertake to play our part in sustaining them. Both these notions seem bound up with the ideal of equality before the law: we want the law to be not only blind and free from partiality but also in the

XX

The Rule of Law and the Separation of Powers

public interest. Each of these qualities gives us faith in, and promotes fidelity to, the law. However, the second proves to be especially important. Procedural fairness alone will not generate a responsibility to abide by and uphold the law. We must see the law as in some sense ours - a feeling that flows in large part from the law being a public good, depending on and making possible mutually beneficial cooperation. Yet, seeing ourselves as on a par with our fellows and bound in reciprocal ties to them is a fundamentally political ideal, linked to democratic citizenship. This ideal literally has the rule of a tyrant as its antithesis. Indeed, it points to why authoritarianism would also be incompatible with the Rule of Law. Theoretically, an authoritarian regime could be a benevolent, rather than a tyrannical, dictatorship although, obviously, there are prudential reasons for doubting that this would ever be so in practice. But if we see the Rule of Law as subject to the circumstances of politics, so that there is no agreement on which cases are different and how they should be treated differently, then the only way to treat people as equals under the law will be to allow them to decide this issue. In this case, authoritarianism proves incompatible with the Rule of Law in theory as well. Put another way, authoritarianism will always be arbitrary by virtue of being dominating (Pettit, 1999, pp. 56-57). No privileged evaluative viewpoint, be it of a judge or some other third-party arbitrator, can decide whether government action serves public, rather than sectional, interests and so shows equal regard for all. The only available heuristic is a political process that allows people to speak for themselves and to contest the proposals of others. In other words, only democracy is capable of promoting laws that are fairly framed and applied. Before exploring this thesis further, we need to investigate more fully the problems of legalistic conceptions of the Rule of Law.

The Rule of Law and Judicial Discretion This section prepares the way for a consideration of the essays by and on F.A. Hayek and Ronald Dworkin in Part II of the volume. These thinkers have offered two of the most important defences of the legal view of the Rule of Law in recent times. Superficially, their theories could not be more different. Hayek adopted a largely formal and rule-based approach of a kind that Dworkin sees as ultimately vacuous. By contrast, Dworkin's principled approach appeals to a sense of moral community in a way that Hayek regarded as pernicious. Although both see the Rule of Law as a bulwark of individual freedom, Hayek believed that it defends the market against redistributive measures whereas Dworkin considers the equitable distribution of resources as a condition of, rather than a constraint on, liberty. However, there are also certain similarities between their theories. Both distinguish law proper from ordinary legislation, which they perceive as tainted through being particularistic and goal-oriented. In their view, the former prioritizes the individual and the latter prioritizes the collective, and a prime function of the Rule of Law is to ensure that the one is never sacrificed to the other. They also claim that law never runs out, so that judges can also be guided by it. Each of these points will be disputed below. Going by the Rule Book

Hayek conceived the Rule of Law as being diametrically opposed to the arbitrary rules of a Hobbesian sovereign, a form of governing he associated with totalitarian regimes of both the

The Rule of Law and the Separation of Powers

x.xi

right and left. He counterposed 'law' as a system of generalmles to 'legislation' seen as the commands of the sovereign mler- including a democratically elected sovereign parliament (Hayek, 1973, ch. 2). Legislation was appropriate to organizations, such as an army, that had clear and determinate goals. However, Hayek also contended that society at large has no such common purposes, just the very diverse and evolving ends of the individuals who compose it. Consequently, to govern all social life as if it were an organization is an impossible task that leads all who attempt it towards tyranny. Instead, a well-ordered society needs clear and general mles designed to provide everyone with the maximum possible freedom to act and think without infringing on the freedom of others to do likewise. As he put it, the law should operate like a Highway Code: it should provide the mles of the road needed for people to drive about with a reduced risk of accidents, not a set of orders directing people where to go, when and how (Hayek, 1994, pp. 55-56). According to Hayek, the Rule of Law consisted in the primacy of such mles over legislation. If legislation was subordinated to law, citizens would no longer be made to serve the momentary purposes of government policy but would be free to pursue their own ends on equal terms with others and in ways that avoided mutual interference. Hayek developed this thesis most clearly with regard to the economy. To mn an entirely planned economy, the state would have not only to know what people currently want and need and the best way to satisfy these requirements, but also to second-guess a whole range of unpredictable events, including the conceptual innovations, unde1taken by individuals, that alter supply and demand in the future. Since no human agency can do this, planning inevitably proves inefficient and, to the extent that it necessarily constrains human choice and experimentation by forcing all to conform to the plan, coercive as well. He concluded, social justice, interpreted as an ideal distribution of goods that accords with every individual's needs and deserts, is a mirage (Hayek, 1976, chs 7 and 9). Unfortunately, democracy encourages the pursuit of this very idea since it offers the perfect legitimation for organized groups seeking various benefits from the state (Hayek, 1973, p. 2). As a result, state legislation was gradually encompassing all spheres of life and moving Western democracies along 'the road to serfdom' (Hayek, 1994, chs 5 and 6). Upholding the Rule of Law would curb this shift. In Hayek's view, mles arise not from human agency but spontaneously, through individuals trying to adapt to each other and their environment. Rational appraisal involves weeding out poor conventions through trial and enor, not assessing their fit within a supposed rational system oflaw (Hayek, 1973, ch. 1). To continue the Highway Code analogy, it is custom rather than a priori reasoning that determines whether one drives on the right or left, and experiment that tends to fix the most appropriate speed limits. Nevertheless, laws had to have certain fonnal features to avoid becoming instmments of arbitrary power. They had to be fixed, general, abstract, universal and, above all, independent of any given policy goal. Hayek believed that these criteria are sufficient to distinguish law from legislation (ibid., p. l ). Their very generality and abstractness make it impossible to know how they might affect particular people. As a result, law-makers have an incentive to ensure legalmles and procedures are as fair as possible. Moreover, he maintained that mles of this nature cannot be used to pursue a particular purpose. Laws so conceived only provide a framework for human interaction that prevents others, especially governments, seeking to employ us as a means to their ends. I noted in the previous section how a working system of law needs to possess some formal mle-like qualities and that these reduce certain kinds of ad hoc and discretionary decisionmaking. However, Hayek's account of the Rule of Law is a radical version of this thesis. He

xxii

The Rule of Law and the Separation of Powers

maintains that we can literally rule through formal, general rules and, indeed, that this is the only way for a complex society to be govemed without degenerating into an inefficient tyranny. How true is this argument? Hayek certainly underscores the ways in which general rules can overcome many of the informational, psychological and political costs of substantive, case-bycase decision-making. 11 Obtaining all the relevant infonnation imposes a massive administrative burden and may not be possible in any case. It also places a huge responsibility on officials and citizens, obliging them to constantly exercise their judgement. Getting substantive political agreement on a given matter is often much harder than agreeing on a general rule. In the former case, people's different concems and interests are likely to play a much greater role. Consequent!y, decision-making takes much longer and the incentives for rent-seeking behaviour increase as each group seeks some pay-off for giving its assent. The result can be incoherent and expensive legislation. By contrast, people can often agree on the reasonableness of a rule without entering into debates over its ultimate rightness and wrongness. Indeed, a procedural or formal rule, such as 'first come first served', can be a fair way of resolving an issue without judging the merits of each case. A rule in an uncontentious area may even be extendable as a solution to resolve contentious cases in other areas. Even if agreeing on the rule does have informational, psychological and political costs of its own, it may produce savings overall by deciding many cases at one go. Rules also guard against myopia, carelessness, bias or weakness of will. The burden of making judgements is lifted from citizens and officials alike. Moreover, both are enabled as well as constrained. For example, they can be emboldened to confront the rich and powerful by being able simply to insist that the rule be followed rather than having to challenge their actions on the basis of a personal judgement. In similar ways, by increasing visibility, rules enhance accountability. Citizens and officials alike can hold each other to the rules. The impersonality of rules can also reduce the feeling of humiliation felt by those to whom they are applied. They need not feel as though they are the hapless victims or the fawning beneficiaries of an individual's discretion. All these features, along with the others we explored earlier, diminish the scope for wilfulness, oppression and domination. However, we also saw that rules have various drawbacks. In many respects, Hayek's formalist approach highlights these shortcomings rather than overcoming them. Two of his criteria for law prove particularly problematic- the insistence that there be no reference to specific people or policy outcomes, and his desire to eliminate discretion, so that the law always guides what judges and officials should do. Both restrictions are crucial to his attempt to distinguish law from legislation, yet neither appears tenable. Formal properties alone do not define the appropriateness of a rule. Their generality and blindness to particulars often prove a vice rather than a virtue (Sunstein, 1996, ch. 5). As I noted, rules that are reasonable in some contexts may be inappropriate in others - health and safety rules suitable for large establishments may be excessive for small concems, for example. Circumstances can also change in unforeseen ways that make a nonsense of prevailing general rules- it would be absurd to apply regulations imposing catalytic converters on 'all cars' to electric vehicles that were unimagined at the time the law was passed. Rules can also entrench certain biases and injustices. I remarked how racial laws can satisfy the criteria of generality and universality. But I observed, too, how failing to discriminate can be problematic. In general, formal equality of treatment promotes, rather than reduces, substantive inequalities. That can often be a legitimate goal. To discover who can run the fastest in given conditions, the rules of the race must be the same for all. However, equity often requires ensuring that the rules do not

The Rule of Law and the Separation of Powers

xxiii

treat differently situated individuals in inappropriately similar ways. By treating all in the same way rules can harm those who ought to be treated differently or, worse, legitimate practices that exploit those differences. Thus, feminists have criticized equal opportunities law for including a covert (and sometimes ove1t) male comparator test that overlooks relevant differences such as pregnancy or the structural factors that have relegated women to low-paid, casual employment (Mackinnon, 1989, p. 36). Instead of rethinking work and gender roles in ways that are compatible with shared childcare duties, family responsibilities become a disqualification for certain jobs. As a result of these difficulties, it seems impossible for laws not to deal either with policy outcomes or persons. Hayek is no doubt correct to observe that legislators rarely set out to frame laws in fonnal tenns per se. But that is because it would be nonsensical to do so. Legislators devise laws with pmticular ends in mind and then give them a certain form in order to obtain the lawlike characteristics which enable these purposes to be achieved. Although this means that laws can be turned to immoral purposes, it also enables them to avoid the pitfalls outlined above. Thus, traffic regulations have elements of generality to secure fairness and the benefits of cooperation but m·e also determined in pmt by notions of expediency and vm·ious substantive purposes. Of course, Hayek did not dispute that legislation was required in certain m·eas and presumably thought that, in these cases, judges should be guided by the policies laid out by governments. However, he did believe that, in deciding cases where the law was unclem· or when detennining whether legislation was appropriate or not, only lawlike considerations should prevail. Yet it is not clem· whether even this more limited position is feasible. Once we concede that there can be occasions when a rule such as 'nobody can drive faster than 30 miles per hour in a built-up m·ea' might admit of exceptions- as when an ambulance, or possibly a private em·, is rushing an emergency patient to hospital - then it becomes difficult not to assess the rule in the light of its justification. If that occurs, the rule loses its rule-like chm·acter and becomes a piece of advice or a weighty consideration that has to be judged in terms of its relevance for the individuals in the case at hand and the likely effect of making a precedent of the exception or given reading of the rule. After all, notions of 'reasonableness' and 'proportionality' are standm·dly invoked in these sorts of cases and usually involve attention to pmticulars and some calculation of social hm·ms and benefits. In other words, the judge begins to act in ways that Hayek deems as appropriate to the legislator. This brings us to Hayek's belief that rules could ever eliminate judicial discretion. The problem is that rules almost never operate in the manner he proposes- as mid-level generalizations that settle all cases in advance. A degree of vagueness that requires the exercise of judgement on those who follow or apply them is built into their very nature. Wittgenstein (1953, pp. 87, 185, 219) famously noted how easy it is to misread a rule because a rule builds on fm· more assumptions than can ever be expressed in the rule itself. Is a rule about the size of 'trunks' referring to trees, a person's body, the luggage compartment of a car, male swimming shmts or an elephant's nose? Only context and certain shm·ed understandings will enable us to know. In these cases, most people probably could fix on a detenninate meaning without too much difficulty. However, it is not difficult to envision harder cases. For example, does an 'electric' bike come under the rules relating to pedal cycles or mopeds - should the rider wear a motorcycle crash helmet, say? Or take the references to 'equality' or 'freedom of speech' in bills of rights. Interpreting these terms raises substantive disagreements stemming from contrasting moral, political and epistemological positions. Sometimes the assumptions that come to guide judicial understanding

xxiv

The Rule of Law and the Separation of Powers

of these sorts of terms end up being of a highly technical character, making it tricky for citizens to remain law-abiding when these technical definitions depart considerably from lay understandings. But, often, there simply is no totally settled view -making rule-following even harder and official discretion all the greater. Its doubtful whether tightening up the rules or providing a glossary of definitions, as law books often attempt to do, will ever remove all ambiguity. Rules will always seem either too natTow or too broad when faced with the unanticipated case, leading those charged with enforcing them to consider whether or not it is sufficiently different in relevant ways for them not to apply. Indeed, some vagueness in the rules may even be desirable at times. I noted em·lier how sticklers for the rules can sometimes seem overly rigid and officious. In fact, strict rules that seem insufficiently discriminating can lead to people suneptitiously bending or selectively ignoring them, as occasionally occurred when British juries refused to convict certain classes of murderers in cases where the death penalty was compulsory. Although such deception may be watTanted, transpm·ency of decision-making can be lost in the process. Rules that are too fixed and predictable can also aid evasion. Exploitable loopholes will always exist, and people can lem·n to play the rules- as tax accountants m·e paid to do. In these sorts of cases, it might be better to have a flexible standard rather than a fixed rule. In sum, the model of a system of general rules seems unable to deliver the goal-independent, impartial and fixed set of prospective judgements that Hayek seeks from it. Many of the difficulties become apparent if we tum to Hayek's two main claims with regm·d to the Rule of Law- that rules will operate against redistribution and economic planning while protecting the freedom of individuals from interference by others. He assumes that the formal prope1ties of rules will exclude the pursuit of patticular goals. However, there are no patticular purposes that cannot be framed within some general description that applies to them alone or be derived from some suitably designed general rule. 'All persons eaming above £100000 pay supertax' is a general, equal rule. Likewise, stable prices and plans announced well in advance may also be consistent with a purely formal view of the Rule of Law. Far from being incompatible with the pursuit of social purposes, rules help promote them by moving officials and citizens consistently towards that objective without the need for continuous commands. The point in such cases is whether these policies m·e socially or economically sound: often they are (Raz, this volume, pp. 91-92). For example, Hayek suggests that 'measures designed to control the access to different trades and occupations, the terms of sale, and the amounts to be produced and sold' require reference to particulm· goals and so 'involve arbitrm·y discrimination against persons' thereby contravening the Rule of Law as he conceives it (Hayek, 1960, pp. 227-28). Yet he immediately backtracks, admitting that it is sensible to ensure that doctors are suitably qualified before being allowed to practise, that pilots pass eye tests, and even that sellers of firearms and poisons should be 'persons satisfying certain intellectual and moral qualities (ibid., p. 228). In fact, most govemment regulations - including much social legislation- arise not as patt of a rational plan but as responses to particulm· problems, being progressively modified through the trial-and-error mechanism of which Hayek approves (McCormick, 1989). These problems prove even more acute in the case of liberty. Hayek's central contention is that 'when we obey laws, in the sense of general abstract rules laid down irrespective of their application to us, we are not subject to another's will and are therefore free' (1960, pp. 15354). At times, he appears to suggest that, so long as everyone is similm·ly affected by a rule, it is not aimed at anyone personally and infraction is avoidable, then no coercion is involved.

The Rule of Law and the Separation of Powers

XXV

Law-givers do not know the patticular cases where their rules will be applied and the judge is simply applying that law. Consequently, law is like a natural obstacle. But this leads to absurdities. As Hamowy (1987) has observed, by these criteria a gangster-ridden neighbourhood- being like a plague-infested swatnp, neither aimed at me personally nor unavoidable- represents no limit on my freedom. This is a serious problem given that his purely formal criteria not only offer no guidance as to which rules should apply to what sorts of activity, but also are consistent with all kinds of hidden or ove1t biases that can discriminate against particulm· groups. Hayek pattially acknowledged this difficulty in accepting that it can be misguided to apply the same rules to everyone in all circumstances. As I have noted, law frequently discriminates on grounds of age or sex, for example. The key is to discover when such discrimination is reasonable or not. His response to this problem is extremely suggestive: 'Such distinctions will not be m·bitrary', he wrote, 'will not subject one group to the will of others, if they are equally recognised as justified by those inside and those outside the group' (Hayek, 1960, p. 54, emphasis added). However, this criterion makes the test of the Rule of Law not its formal qualities but its capacity to evince reciprocity and hence obtain mutual assent from citizens. Equality before the law involves the content of legal rules taking everyone into account and giving equal weight to different points of view. Put another way, Hayek appears to be suggesting that law must reflect the general, rather than any particulm·, will. Yet, this m·gument pushes us towards the political conception of constitutionalism, since such mutual acceptance is only likely from a political process in which all must actually consult the interest of others. Before exploring that possibility, though, we need to see whether an alternative view of legal integrity to the Hayekian rule book might be capable of leading law to rule. Judging on Principle

Among contemporary jurists, Ronald Dworkin has probably mounted the boldest defence of the judiciat-y's role in upholding the moral integrity of the law. However, he focuses on principles, which he largely identifies with constitutional rights, rather than general, abstract rules. He contends that law's principles can be consistently and prospectively applied to every possible case, securing individual rights in the process. As a result, he believes that law's empire can be legitimately and coherently extended to cover all persons so as to render everyone, including the courts, equally subject to it. Although Dworkin shares Hayek's view of legislation as policy-oriented and designed to promote the general welfare, he grants it a broader scope and sees democracy as the fairest mechanism for aggregating individual interests in ways likely to promote the public good. Nevertheless, he too believes that democracy needs to be supplemented and circumscribed by law. Legislation suffers from the sorts of problems I identified with rule by law. Somewhat contentiously, Dworkin maintains that legislative rules tend to be quite specific, applying to any given instance in an 'all or nothing' manner that lacks the possibility of 'weighing' different considerations ( 1978, pp. 24, 26). As a result, a rule's bearing in an unforeseen case is not just hm·d to fathom but literally non-existent- there simply is no relevant rule to apply. Even if this contention goes too fm·, legal positivists would agree that legislation frequently contains 'gaps', albeit often arising from its deliberate vagueness rather than any overspecificity. These 'gaps' present judges and officials with 'hm·d cases'. The legal positivist argues that, in this instance, law runs out so that they must use their discretion to decide the matter, although this is usually

xxvi

The Rule of Law and the Separation of Powers

guided in various ways by prevailing legal norms and parallel cases from which analogies might be drawn. By contrast, Dworkin believes a 'right answer' may be discerned by consulting the moral principles that underpin the legal system (ibid., p. 81). These principles consist of rights rather than policy goals (ibid., pp. 22-23, p. 82). As such, they constrain ordinary legislation by protecting individual interests from being overlooked or curtailed for the benefit of the collective welfare. Moreover, by not being tied to the needs of the moment, they allow a consistent and prospective application of the law. How does this work? In the most developed version of his thesis, Dworkin argues that judges must interpret the law so that it both achieves a 'fit' with generally accepted aspects of legal doctrine and has the most justificatory 'appeal' from a moral point of view. In a pre-interpretative phase, judges collect the available legal materials. Then, within the constraints of 'fit', they seek to interpret the law in a way that offers the morally 'best constructive interpretation of the political structure and legal doctrine of their community' by reference to a 'coherent set of principles about people's rights and duties' (Dworkin, 1986b, p. 225). Finally, in the postinterpretative phase they develop the law by adjudicating upon its bearing in patticular cases in the light of their interpretation. Dworkin believes that this procedure constrains judicial discretion by promoting consistency of judgement and removing 'gaps' -even in hm·d cases. For in every case the judge is simply 'discovering' the existing entitlements ofthe different patties. Although judges necessarily employ a 'weak' form of discretion, given that there is no clear-cut decision procedure, they do not have discretion in the 'strong' sense of being able to decide a case as they wish. Such discretionm·y judgement is ruled out not only by the need for 'fit' with settled and relatively uncontroversial aspects of the law, but also by his contention that the rights and principles that best justify it will supply, theoretically at least, determinate answers that it is the duty of the judge to uncover. Nevertheless, these justificatory moral values m·e not simply those which a legal positivist would recognize as legally valid because they have been explicitly incorporated into the law, as is the case with bills of rights, or expressly refened to by it, as m·ises with notions of fairness in contracts. In giving the 'morally best' justification of the law, judges must draw on their own moral views and not just take account of conventional morality. It is not simply that settled social rules may not exist in all m·eas and that bills ofrights can be compatible with more than one theory of morality. Dworkin also believes that the law's integrity rests on its reflecting certain objective moral considerations. For Dworkin, therefore, law rules by virtue of its 'fit' with 'a coherent set of principles about justice and fairness and procedural due process' (1986b, p. 243). Indeed, he believes that by giving law a degree of moral and practical integrity, these principles promote faith in the legal system as a whole, generating an obligation to abide by it on the patt of citizens and officials alike. Both these claims are disputable. Law is too complex and capable of being given too many conflicting moral readings for any such principled 'fit' to appear anything but arbitrary. In this case, fidelity to law is more likely to m·ise from allowing the expression of p1incipled disagreement through the legislature than through promoting a non-existent consensus via the courts. First, Dworkin's holistic approach to the law can often be arbitrm·y, encouraging the application of principles and considerations in settled parts of the law that m·e remote from, and inappropriate to, the case at hand. Dworkin concedes that constructing a coherent picture of law, let alone providing it with a principled justification, requires an almost Herculean effort beyond the capacity of ordinary judges. The law contains many branches, and he accepts that the time and intellectual effort involved makes it unlikely that any real judge could hope to construct a

The Rule of Law and the Separation of Powers

xxvii

comprehensive picture of the law in each area and an overarching general theory to unite them all. Instead, real judges necessarily operate with an understanding of the law that is partial. However, he believes that for all their limitations, they still strive for as coherently principled a view as they can ( l986b, p. 265). But this may not be available. Different parts of the law give different weights to different goods, values and types of moral claim. Civil and criminal cases tend to operate with different standards of evidence and notions of responsibility, for example. Even similar areas of the law may involve quite different criteria. Principles and considerations relevant to cases of sex discrimination, say, may be quite inappropriate to issues of race. Dworkin's holistic approach may force a coherence that does not and should not obtain, by treating very different areas as if they were the same. He partly conceals these difficulties by arguing at the level of very abstract principle. But when one descends from considering equality as such to equality in particular cases and areas, dissimilarities as well as similarities soon emerge. Moreover, it may be that the moral considerations involved are of quite diverse and incommensurable kinds. In these circumstances, to appeal from rights in one part of the law and apply them to quite another could be incoherent (see Sunstein, 1996, pp. 48-50). Second, and related to this last point, decisions based on rights need not yield a right answer. There may be two or more answers that could be seen as being, in very different ways, equally right. Of course, Dworkin does accept that rights can clash and may need to be balanced. But he believes that an appropriate balance always exists, just as he contends that a given party must always have a right to win (Dworkin, 1978, pp. 286--87). Again, this argument assumes a degree of commensurability that may not be available (see Mackie, this volume, p. 179). For example, take the balance between privacy and freedom of speech. People holding different normative theories and priorities are likely to disagree not only about their relative weightboth with regard to each other and to other important values - but also over the nature of the two concepts. Thus, some may view the private sphere as one's immediate family circle whereas others may include close friends, business associates (who may also involve family members and friends) or even everything that happens behind closed doors in a privately owned establishment and so on. The same holds for 'speech', as disputes about the first amendment to the US constitution testify. Consequently, for some, no clash between the two will have occurred whereas, for others, it has. Because there is no commensurable scale on which privacy and speech can be weighed, and because each is in any case defined and applied in ways that are subject to the 'differing empirical and normative appraisals', then, even if ideally there is an 'objective' balance between the two, any view of what that is will always be subject to reasonable disagreement. Third, even though legal reasoning refers to rights, it does not follow that there should be no reference to interests. In this respect, Dworkin's division between 'policy' and 'principle' proves to be as contentious as Hayek's similar attempt to separate rules from any reference to specific ends. Rights certainly deserve to be regarded as weighty interests that any view of the public interest should take into account. However, even Dworkin accepts that not all rights are absolute. They often derive their rationale not from their benefit to the right-holder per se but from the public goods they promote. As Raz has noted, only journalists and politicians draw a direct benefit from freedom of speech. By and large, ordinary citizens value it for the diffuse public benefit that derives from living under a political system that is open to scrutiny in the way free speech allows (Raz, 1994, pp. 52-54). This broader context often comes into play when fixing the boundaries of rights, with courts being less inclined to give any guarantee an expansive

xxviii

The Rule of Law and the Separation of Powers

reading if that would not be conducive to the general welfare. Thus, we restrict speech in areas such as official secrets because it seems to undercut the public purposes that speech serves. Likewise, Greenawalt reports the US Supreme Court as rejecting an official's claim not to produce an organizational record on the grounds that it might incriminate him because such records offer the main source of evidence of wrongdoing and 'were the cloak of the privilege to be thrown around these impersonal records and documents, effective enforcement of many federal and state laws would be impossible'. 12 Similar!y, when rights clash, interests and considerations of the general welfare may help detennine which right is stronger in the particular circumstances. When comparing, say, the rights of a property developer to build an airpmt on land legally purchased for the purpose with the rights of nearby home owners to enjoy their houses and gardens without suffering damaging noise pollution, then at least some view of where the public, as well as private, interests lie in this case seems a necessary part of any decision. Little would ever be done if collective benefits could never outweigh individual rights. Yet if this is true, we appear to have a problem. Not only does Dworkin himself concede that weighing up the collective interest is best done by representative assemblies, he also notes that such considerations are likely to alter according to circumstances, making the prospective almost a priori account of law he seeks impossible. Finally, Dworkin suggests that what the law is depends in certain instances on what is morally best. He believes that this moral dimension gives a certain objectivity to judicial decisions. Certainly, moral judgements typically distinguish themselves from mere custom or convention by claiming to be objective. The difficulty lies in sustaining such claims in the face of the limitations of human reasoning which appear to allow for rival claims to be made. There is absolutely no reason to suppose that the judiciary will concur on the morally best view of the law any more than the rest of the population if reasonable disagreements in this area are possible. Indeed, insisting that judges look for some comprehensive, objective theory of the law is more likely to generate such disagreements than a more modest approach. For a start, the judge or official will be invited to interrogate even law that appears to be settled to see if it 'fits' the preferred morally best view of law as a whole. That will have the potential effect of making all cases 'hard'. Moreover, if - as I have suggested - there are good grounds for assuming that judges will disagree, consistency in the law will also be jeopardized. We have observed that applying rules to cases is far from clear-cut. But the application of abstract principles of justice seems likely to be even more indeterminate. In sum, Dworkin's principled approach fails to rule out judicial or official discretion and, in many cases, actively encourages it. By inviting judges to offer a view of 'good' law rather than law per se, Dworkin tums judges from third-party arbiters into patticipants in many of the disagreements that it is politics' rather than the law's role to resolve. They become advocates like other citizens of a patticular view of how society could best be organized. In the process, Dworkin undennines the distinction between law and legislation. Indeed, rights-based arguments more generally cannot avoid taking the consequences for policy into account and do not always offer a determinate 'right' answer. Fm· from promoting fidelity to law, the very atnbition of Dworkin's approach risks undermining it by making law appem· to be little more than the contentious opinion of a particulm· person. Rather than taking law out of politics, his theory serves to emphasize the degree to which law lies within the circumstances of politics. It remains to be seen whether a more overtly political approach can address the worries that lie behind the desire of legal constitutionalists to exclude its influence on the law.

The Rule of Law and the Separation of Powers

xxix

The Separation of Powers, Parliamentarism and Federalism This section turns to the discussion of the separation of powers and parliarnentarism and federalism represented by the articles in Parts III and IV of this volume and relates these mechanisms to an alternative view of judicial review and the Rule of Law to those of Dworkin and Hayek as put forward by Cass Sunstein and Martin Shapiro in their essays reproduced as Chapters 9 and 10 in Patt II. Both Hayek and Dworkin fem·ed that because legislation was the product of the rule of persons it risked being arbitrary. The bogey figure in each case appears to be the Hobbesian sovereign- potentially as potent a force in the form of a sovereign parliament or people as it is in the guise of a monm·ch or dictator. Yet, ironically they fail to meet the Hobbesian challenge and end up turning judges into sovereigns and potentially m·bitrary legislators in their tum. Two problems emerge from the analysis thus far. First, is it possible to conceive of the process of legislation in what one might call a 'de-sovereigntized' way, so that the legislators make laws that evince equal concern and respect for the ideals and interests of those who are to live under them? Second, is it possible for judges and other officials to follow these rules without substituting their judgement for those of the legislature? Both questions m·e briefly addressed below. In each case, I shall argue that the answer lies in so dividing power that politicians and courts alike m·e moved to make decisions in equitable and consistent ways that respect the reasonable disagreements that constitute the 'circumstances of politics'. This response draws inspiration from the republican idea of a mixed commonwealth. In this schema, the Rule of Law is the product of separating and balancing power so that no single group or person exercises sovereign power over others and each must consult the interests of everyone else.

The Separation and Balance of Power: A Republican View That a political system should constitute 'an empire of laws and not of men' is a prime tenet of the republican tradition (see, for example, HatTington, 1992, pp. 20-21), at least in the neoRoman variant of Cicero, Machiavelli and HatTington recently identified by Quentin Skinner (1998,pp. 44-45) and Philip Pettit (1999,pp. 172-83; also Viroli, 1998, pp. 121-25). Republicans contrast a law-governed polity with the domination of m·bitrary rule. As we saw, the distinguishing features of m·bitrary rule m·e the capacity to exercise discretionm·y power at whim and without consulting the interests of those affected. Republicanism's distinctiveness resides in noting how even an enlightened and benevolent ruler who possesses this power still dominates, even if he does not actually oppress, the ruled, for they m·e subjects of their ruler's will. The only alternative to the domination of personal rule is for the people to be citizens and rule themselves. Pm·adoxically, therefore, the rule of law depends on the democratic self-rule of persons. It can be secured only if all citizens, usually through their elected representatives, can command equal consideration in the making of collective rules, and everyone within the body politic including those authorized to rule - is equally subject to whatever laws they impose upon themselves (Skinner, 1998, p. 74). According to this way of thinking, rulers will only accept restrictions on their power and take account of the interests of others when obliged by necessity or because of the perceived benefits. 13 I noted above how rule by law is no gum·antee of the Rule of Law in the sense of treating all as equals. Even tyrannical rulers can have an interest in making their rule reasonably

XXX

The Rule of Law and the Separation of Powers

predictable in order to create a regime - albeit of an authoritarian kind. Likewise, rich and influential groups can desire to have their privileges secured against both the rulers and the less well-off and powerful. Because rulers need their cooperation and support most of all, and they are best situated to get what they want by extra-legal means, the wealthy and well organized will be the first to be provided with legal protection. Hence the primacy of property over other rights, both chronologically and (certainly in the past and, in many respects, today) substantively. Legality in these cases helps stabilize the position of those concerned and incorporates potential rivals to the ruler's power and authority into the system. However, it also skews the law against the poor and less well placed. If the law is to serve all citizens equally, then these groups must organize themselves in their tum and obtain some leverage on the powers-that-be. To take an example from an area close to traditional republican concerns, it is no accident that the era of mass war, when most of the fit, adult male population had to be conscripted into the armed forces and their place in the workforce filled by women, should have coincided with the advancement of mass democracy through the extension of the franchise- most patticularly to women - and through the creation and deepening of welfm·e and social insurance schemes. Of course, the ideal would be a society without any factions or special interests, where all were committed to the common, rather than their own particulm·, good. However, this scenario is unlikely in a society of any complexity and diversity. Therefore, the best chance for the Rule of Law prevailing is a pluralistic society in which many groups have roughly equal bm·gaining power. Any group will be tempted to employ the law to support their own interests. Consequently, asymmetries in power will lead to laws entrenching the privileges of the powerful at the expense of the weak. To overcome this imbalance, the weak must put themselves in a situation whereby rulers enable them to make and use laws to defend their concerns. Industrialists may get laws regulating ownership, trade, labour and capital in ways that they find beneficial, but they will also have to accept health and safety and employment legislation that favours workers. This picture may seem to sacrifice justice to power. But that view would be wrong; rather, it shows how justice emerges from a balance of power. Every group may seek to employ the law for their own purposes, but when they are compelled to deal with each other they tmn the law into a common asset. Thus, worker and industrialist, in the exatnple given above, each one has to accommodate the interests of the other. The result is that employment law, even if devised with the worker in mind, by being equally available to all will occasionally favour the industrialist. Similm·ly, workers may benefit from freedom of contract and trade laws designed with the industrialist in mind. Indeed, because the laws will be equally binding on all there will be an incentive to render them fair and make reciprocal concessions that recognize the special concerns and situation of each. In other words, laws will be fratned in mutually acceptable ways and bem· consistently and equally on all to whom they apply- whether they m·e common rules for all or special legislation that takes into account the peculiar circumstances of a few. Thus, the key to the republican approach is to create a pluralist democracy in which different interests and values can counterbalance each other. Republicanism concentrates on the form of government, rather than any formal qualities of the law per se, as crucial to ensuring laws are made and applied in ways that show equal concern and respect to all. From this perspective, the Rule of Law arises from a particulm· civic condition- one whereby all citizens enjoy an equal political status and have no dominion over each other. To quote HatTington again, it is only when all are equal in the making of the laws that they will be 'framed by every private man unto no other end (or they may thank themselves) than to protect the liberty of every man'

The Rule of Law and the Separation of Powers

x.xxi

(Harrington, 1992, p. 20; see also Skinner, 1998, pp. 74--77). Republicanism operates on the basis of what Habermas has called the radical democratic 'hunch' that 'private legal subjects cannot come to enjoy equal individual liberties if they do not themselves, in the common exercise of their political autonomy, achieve clarity about justified interests and standards. They themselves must agree on the relevant aspects under which equals should be treated equally and unequals unequally' (Habermas, 1996, p. xlii). Democracy has both a negative and a positive role within this argument. Negatively, it provides a control mechanism. Governments can be held to account for neither taking people's interests seriously nor treating them with equal concern and respect. Positively, it informs people about each other's interests and values, enabling them to give these due consideration and negotiate appropriate compromises when deliberating common rules and policies. This democratic approach avoids a formulaic view that laws must be general, abstract and universalizable to preserve equality and freedom. For example, it enables special rules to be tested for their mutual acceptability by checking that they are regarded as neither discriminatory by those to whom they apply nor as unfair privileges by the rest. To achieve these ends, democracy must have a ce1tain form. Republicans advocate a 'mixed' type of government to derive the Rule of Law from the democratic rule of persons. There are two central mechanisms within this peculiarly republican regime: the 'separation of powers' and the 'balance of power' (Bellamy, this volume, Chapter 11 ). The first reduces the discretionary aspect of the law, preventing it from degenerating into a mere command. The second encourages the law to track the interests of those to whom it applies and gives them a sense of ownership over it. Taken together, they serve to disperse power so that the law becomes more sensitive to the diversity of ideals, interests and situations within the polity. They block the abuse of power while facilitating and legitimating its constructive and more differentiated use. The 'separation of powers' divides the legislative from the executive and judicial functions to prevent any person or group becoming a judge in their own cause. Separating those who formulate the laws from those entrusted with their interpretation, application and enforcement brings all within the law. The legislators are constrained in their ability to decree ad hoc or selfserving laws by the judiciary's role in applying the law to all in an impartial and consistent manner. Meanwhile, the discretionary and interpretative powers held by the executive and judicial branches are checked because exercised under laws they do not make. However, there are two well-known problems with this thesis. The first concerns the conceptual and practical difficulty of separating functions. For example, when judges adjudicate on which rules apply in given cases, they often end up setting precedents that come to constitute new rules. Similarly, officials frequently create rules in the course of implementing a law. Meanwhile, legislators are inevitably concerned with how the laws which they frame will be interpreted and applied to specific cases. Indeed, we have seen how self-serving biases can be built into the most formal and general of rules. Thus, the three functions are interrelated, with each branch of government engaged to some degree in the activities of the other. The second problem arises at this point. For the constraints imposed by what functional separation is possible will be undermined if all branches of government represent similar groups and interests. Having each function run by different people will not necessarily prevent their working for a partial interest if all belong to the same party or class. This problem has been pmticularly acute in systems, such as the British, where the executive controls the legislature and can exert direct and indirect influence over the judicim·y through appointments or other means.

xxxii

The Rule of Law and the Separation of Powers

The notion of the 'balance of power' comes in here. Republicans are not overconcerned with maintaining a strict separation of powers. That each branch possesses elements of all three functions not only resolves some of the impracticalities of purely fonnal accounts of the rule of law, but also allows these branches to act as more effective checks on each other. The crucial factor is that power is dispersed between and within these branches so that each constrains the discretionary power of the other. The purpose of such mutual constraining is to ensure that the law addresses the concerns of those it affects in a fair manner. So the separation of powers needs to be so organized as to produce a balance of power between the various interests and values of individuals and groups within the polity, obliging them to interact with each other in ways that promote equal concern and respect. 14 For example, federalism has been a standard device of this kind. It operates to produce a balance between various sorts of local and national concern, reinforcing the separation of powers by ensuring that there are rival courts, legislatures and executives that are unlikely to be dominated by any one interest. Bicameral legislatures that are elected according to different systems operate in a similar way. The type of interaction advocated by republicans is one which obliges the different parties to 'hear the other side'. Public measures- the res publica- must be publicly justifiable. Such public justification involves more than a lowest-common-denominator test, whereby the only legitimate collective rules relate to goods that de facto are in everyone's rational interest to have publicly provided. In this case, those able to provide for themselves could object to supporting collective arrangements and even standard public goods might be deemed unacceptable. Instead, public justification entails the giving of reasons that can be shared by others. Thus, common rules should not only treat all individuals as moral equals capable of autonomous action, but also be attentive to the variety of circumstances in which they find themselves and the diverse fonns of practical reasoning they adopt. Consequently, legislators must drop purely self-interested and self-referential reasoning and look for fonns of argument that could be accepted by other individuals who are similarly constrained. In other words, there will be an assumption that, in evaluating laws, we strut by taking into account the effects of their general performance for securing the various generic goods that one could expect individuals to value in the different situations they might find themselves. This assumption implies neither that all are similarly situated nor that they value the same goods. On the one hand, it would exclude any m·guments that failed to heed the plight or concerns of others and could not be plausibly shared. Thus, self-serving m·guments by the prosperous that there could never be grounds for mutual aid would be unlikely to pass this test. On the other hand, it merely requires that arguments be made in tenns all could relate to. This requirement is consistent with groups or individuals pointing out either how their peculim· circumstances create special demands which would be felt by others in their place, or requesting that their currently ignored claims be recognized on grounds of fairness by drawing parallels with certain existing entitlements of others. When incommensurable goods and values are in play, it also allows for collective agreements to take the fonn of a compromise involving reciprocal concessions of vm·ious kinds (Bellamy, 1999, ch. 4). These constraints will often force political arguments to manifest many of the formal features of abstractness, generality and equality typically associated with the Rule of Law. However, such formalism need not be applied in the mechanical manner advocated by Hayek and others so as to prevent overtly raising substantive positions or an appreciation of the peculim· requirements of different people's patticular circumstances or experiences. Rather, the law can

The Rule of Law and the Separation of Powers

xxxiii

evolve through the reciprocal recognition of precisely such considerations, with laws and regulations becoming more diverse and specialized as societies grow more differentiated and pluralist. From a republican point of view, this development need not signal a disregard for justice - quite the contrary. The central requirement is not that the law avoids any deviation from generality and abstractness but that any particular and specific provision should be justifiable in a mutually acceptable way and bear equally and consistently on all to whom it applies. It will not be possible to decide a priori what rules are likely to pass this test. Much will depend on the policy and the complexion of those involved. Thus, a call by Muslims for the education system to recognize their religion will play differently in a state that supports religious schools than in one where it does not. In the former, a demand for equal treatment might support the establishment oflslamic schools; in the latter it might only lead to a special provision within the national curriculum, such as the ability to take a state exam on Islam. What is important is that all sides listen to each other and couch the argument in public reasons acceptable to each. Republicans insist that such reciprocal acknowledgement and negotiation rests not on the formality of the law but on the republican character of law-making as a result of an appropriate mix being achieved through a suitable balance of power. At this point, a legitimate question arises as to how one can tell whether the system is appropriately mixed or not. The solution lies in making the constitutive rules of the political system as open to democratic contestation as the regulative rules are. The latter result from ordinary legislation, which provides the legal framework within which citizens, public agencies and officials operate, while the fonner regulate the system whereby these laws are made. Clearly, the constitutive rules will influence the likely content of the regulative rules. Unsurprisingly, therefore, criticism of the injustice of the second are often linked to demands for altering the first to change biases in the legislative process. Workers and women, for example, linked their campaigns against various discriminatory laws with campaigns for suffrage reform. More recently, claims for polyethnic rights by multicultural groups have likewise been connected to demands for self-government and special representation rights (Kymlicka, 1995, pp. 26-33). Clearly, there are good Rule of Law-like reasons for making the constitutive rules harder to change than ordinary legislation. People are made more secure if there is a relatively stable institutional system which, among other things, prevents individual rights being curtailed by the whim of the majority. However, from a republican perspective, the justification of such constitutional rules is ultimately a democratic one. So, there must be reasonably accessible and practical mechanisms for their reform and review, such as referenda, which render them open to public scrutiny and discussion. As Skinner ( 1996, pp. 10-11, 59-60) has shown, republicanism was a key alternative to both Hobbes's view ofthe Rule of Law and those liberal arguments, such as Hayek's and Dworkin's, that - for all their antagonism to the power accorded by Hobbes to the sovereign ruler nevertheless develop aspects of his anti-republican critique. This Hob be sean attack raises three common criticisms that republicans need to counter. First, Hobbes denied any connection between public and private liberty (1991, p. 110). He contended that individuals have no more immunity from the laws in a republic than under a monarchical government. What guarantees individual freedom is not the source of law but its extent. Liberals have developed this thesis to argue that freedom depends on limiting interference and the silence of the law. They argue that the link between liberty and political participation rests on a false account of freedom as selfrealization, in which politics is seen as part of the human good. Such accounts risk identifying

xxxiv

The Rule of Law and the Separation of Powers

serving the state with personal freedom - a position with potentially totalitarian implications (Berlin, 1969, pp. 162-66). Skinner and Pettit have revealed this interpretation to be mistaken. As we have seen, the republican's blunt response to Hobbes is that in a monarchy- however benign- one will always be subject to the will ofthe ruler (Harrington, 1992, pp. 19-20). One can only be sure that the law will track one's private interests by ruling oneself through participating in the public sphere. Thus, private and public freedom are intimately related. However, the reasons for political participation are largely prudential. The motivation is not a strong identification with the state per se, so much as the desire to preserve one's liberty from arbitrary interference by dominating princes. Second, the view of liberty as an absence of interference rather than of domination leads to a view of rights as pre-political, with law an unfortunate, if necessary, intervention with natural liberty, justified only by virtue of constraining the even greater hindrances that might otherwise arise from other people or agencies. Although social liberals construe such potential interferences quite broadly so that this view need not be equated with the minimal state favoured by liberta:tians, the presumption is always against state action. Liberals standm·dly draw the institutional corollary that, since rights are both the basis of and constraint, upon politics, they should be protected by the judicia:t-y. By contrast, we noted how republicans view rights as the products of law. They argue that rights and liberty do not belong to any putative natural condition that the state must attempt to preserve. They m·e a civic achievement, the result of living under a certain sort of populm· political system. Rights are identified, enacted and defended through ordinary and constitutional politics regarding the regulative and constitutive rules of the polity. Moreover, because libe1ty is linked to the absence of domination and m·bitrary (as opposed to all) interference, intervention in the public interest is seen as often promoting the personal liberties of citizens because it creates new opportunities that m·e only available in a society rather than merely preserving the natural liberty of the pre-social state. However, republicanism is not populism, and there are guarantees for minority rights both through the anti-majoritarian bias of the balance of power and the fonnal judicial review allowed by the sepm·ation of powers (Pettit, 1999, pp. 180-83 ). However, although the judicim·y plays a vital role in upholding the consistent application and general coherence of the law, this function does not entail anything like the judicial activism endorsed by those who take a more rights-based view of law. From a republican point of view, such a policy simply produces judicial domination. Finally, Hobbes argues that equitable treatment requires uniform rules, authoritatively defined and interpreted, and that this scheme requires that power be concentrated in the hands of a monm·ch or some other central body. Such concentration is also necessary to ensure that rulers act responsibly. Republicanism rejects this thesis on both prudential and substantive grounds. Prudentially, it would be unwise to trust any single agent or agency to employ their power for other than self-interested purposes unless constrained by the countervailing power of others to do so. Substantively, circumstances are so various and the forms of practical reasoning employed in different activities and by different groups so diverse that there cannot be the sort of essential and universal rules of conduct that Hobbes seeks. Acceptance of common rules cannot be hypothetically derived from pure reason on a priori grounds, but must be negotiated by each participant in the decision-making process 'hem·ing the other side'. The Rule of Law involves the reciprocal giving and responding to the vm·ious reasons of others, not the imposition of a common mode of reasoning on all. This sort of dialogue will only take place when power is divided. As I have m·gued elsewhere (Bella:tny, 1999, ch. 5), the republican notions of dispersing

The Rule of Law and the Separation of Powers

XXXV

and balancing power for the equitable management of social division and conflict also make it highly suited for the governance of complex and pluralist societies. It encourages greater attention to practicalities and increases commitment to obeying and enforcing the rules fairly through promoting civic participation in the setting and monitoring of standards. For the Rule of Law will only be generated if political and legal officials, policies and institutions are responsive to legislation that reflects citizens' evolving experience of the effects of law. Thus, the inevitable delegation of policy-making and interpretation to administrative bodies and officials has to be met by opening up such agencies and agents to democratic accountability. Within republicanism, therefore, the Rule of Law simply is the rule of persons. What is crucial is that it is the equal rule of all persons. But that is the product of a certain form of politics that prevents the exercise of arbitrary power- namely, a system of mixed government involving the separation of powers, a representative legislature, an independent judiciary and accessible courts - that gives wide scope for civic participation and contestation. Moreover, how far the system and its attendant rules allow power to be equitably shared will itself be at issue and a matter of political debate. The result is an evolving practice of political negotiation that embodies many of the forms of the Rule of Law, but tailors decisions to the different values, interests and circumstances of those to whom they are to apply. Democratizing the Judiciary Democratic legislation still has to be interpreted and acted upon by judges and officials. How can we avoid arbitrariness here? Obviously, from a democratic point of view it is desirable that laws are sufficiently detailed and clear to yield a very strict application in most cases. However, an absolute application of even the best drafted laws is virtually impossible. 15 No matter how specific the rules and regulations passed by the legislature may be, these will still give rise to the problems of indeterminacy- the various 'gaps' in law- and the need for discretion that we have explored earlier. 16 Even the plainest language will not eradicate all 'linguistic' ambiguity. Nor can judges avoid interpretative arguments relating to the systemic properties of the lawas when they ponder under which rule a given case might fall. There will also be instances when the mechanical application of the law appears to produce unjust or dysfunctional outcomes. 17 Of course, the legislature can stipulate rules for the interpretation of rules, but these, too, will be incomplete. Nor can appeals to legislative intent overcome the dilemma, for they likewise will be subject to interpretation. Not only is the history of any act usually unclear and the imputing of intent to a collective agent notoriously difficult, but also the self-same linguistic ambiguities, problems of systemic coherence and unforeseen injustices and disutilities will reappear in a new guise as judges seek to ascribe to legislatures an intention to use words in given ways, to see the law as cohering with other laws in a given fashion and so on. Thus, the issue cannot be to outlaw judicial interpretation altogether, but to see if a form of judging exists that is consistent with the circumstances of politics and the primacy of democratic legislation. In fact, the solution among the judiciary turns out to parallel that of the legislature- namely, to institute a balance of power between different courts. This structure helps institutionalize a certain practice of judging that rests on analogy and allows incremental adaptation of the law to unforeseen contexts. Cass Sunstein (1996, chs 2 and 3) has provided a powerful argument for why analogical, case-by-case reasoning offers a better account of how judges can maintain fidelity to law than

xxxvi

The Rule of Law and the Separation of Powers

either Dworkinian judgement on principle or the view of Hayek and others that they should simply abide by the rules. He identifies three advantages to this approach that are germane to our inquiry. First, focusing on the details of the case and looking for parallels with other cases in settled parts of the law in similar areas offers a means for overcoming 'reasonable disagreements' within the circumstances of politics. As we saw, a major problem with the Dworkinian approach is that it courts controversy. By contrast, the search for analogies offers a way of skirting around political conflict. People may have different views of why a previous case was settled correctly or why a given agreed outcome would be correct. By restricting their discussion to points of law, they can avoid raising their deeper, but in many cases only vaguely formulated, conflicts of principle orrationale. As he puts it, their arguments can be 'incomplete!y theorised'. Second, analogical reasoning overcomes the vagueness and obtuseness of rules and principles when it comes to the complexities of pa:tticular cases and the identification of relevant differences. It encourages an attention to the 'totality of circumstances' and peculim·ities of the case at hand. It also allows multiple and diverse criteria to be taken into consideration. Many of these criteria may appeal to incommensurable moral views and factors. However, the casuistical approach does not require that all be measured or weighed according to some single metric. Instead an attempt is made to appreciate their distinctive force in the context of the pa:tticular situation. Finally, consistency in this approach comes not from the application of a rule or single principle, but through drawing pm·allels or contrasts with holdings in previous cases. Analogy invokes the doctrine of stare decisis, that we 'stand by things decided' to justify our decision by appealing to precedents in case law, while allowing incremental change to take account of new circumstances. Equality before the law dictates that we treat similar cases similm·ly. On this line of thinking, law does not consist of common rules or a consensus on principle. Rather, it provides a common resource of diverse considerations that defence and prosecution employ to debate the ways in which the current case is relevantly alike or unlike other cases. In other words, it becomes a means for expressing disagreements and differences within a common language. Although analogical reasoning shares the Dworkinian concern with 'fit', it does not require a 'morally best reading' of the legal system. In many respects, the method can be compm·ed to the way in which examiners in British universities classify borderline degree results as, say, a First or a 2:1. Generally, no body re-marks the scripts so as to fonn a new substantive judgement on them. Instead, they compare the distribution of mm·ks with those of other candidates and justify their views in relation to them. What makes one dist:tibution 'like' or 'unlike' another does not depend on any deep view of the substantive merits of the candidates. They are anonymous and most of those making the judgement have not read the papers in question. Nevertheless, certain peculiarities of the pa:tticular case can be referred to - common statements are that the candidate 'improved in the final yem·', 'did outstandingly on two papers', 'has a high average' and so on. Of course, there can be reference to the ends the rule is supposed to serve - such as increasing Firsts at one end and discouraging the overinflation of 2: ls at the other. There may also be reference to changing criteria- 'standm·ds have relaxed and we have to reflect that'. In these ways, linguistic ambiguities in the rules, their place in the system as a whole and even issues of justice and utility m·e addressed, but without opening up what would be an endless and, for the case at hand, not very fruitful discussion over the point (or pointlessness) of exams themselves or even what ideal standards of mm·king or whatever consist in.

The Rule of Law and the Separation of Powers

x.xxvii

Cases that pose issues of injustice might seem to require more than these rather formal discussions. Yet they need not. Take the British case of R v. R of July 1990, which led to a change in the law regarding rape in marriage. 1R In the original trial, CR pleaded guilty to the attempted rape of his wife and was sentenced to three years' imprisonment. He later appealed as far as the House of Lords and ultimately to the European Court of Human Rights on the 'Rule of Law' grounds that, under English common law, a husband had hitherto been immune from prosecution for rape of his wife because the marriage contract 'implied consent' to sexual relations. 19 Moreover, Section 1 (1) of the Sexual Offences (Amendment) Act of 1976 had specified that rape only arose when sexual intercourse was 'unlawful'. Given that the case predated the Human Rights Act, which arguably provides judges with wide-ranging discretionary rules that might legitimate such a decision, then - from a positivist and democratic point of view - this judgement surely represented a breach of the Rule of Law and could only be reconciled with the doctrine by something like a Dworkinian approach. One response would be to grant this point but retort that when the law offends our moral judgements we should seek to change it through democratic channels rather than via creative judgements by the comts. However well motivated, such creativity does undermine the Rule of Law and the various goods it serves, and could be exploited by those who do not share our moral views for purposes we would find objectionable. 20 Yet, this argument does not deal with the problem that courts will still find themselves having to take remedial action to avoid evils that come to light due to evolving circumstances- the legislature would be overwhelmed if they did not and decision-making by officials would be impossible. However, to fill such gaps judges need not articulate a complete moral philosophy or legal theory. They can simply engage in incremental adjustments that follow a general trend to be found in other cases (Shapiro, this volume, pp. 242-43). In fact, this is precise!y what the trial judge, Owen J, did in his judgment- as did Lord Lane in backing his decision in the Court of Appeal. They noted how the institution of marriage had altered considerably from when Sir Matthew Hale defended the notion that a man cannot rape his wife in 1736, that many exceptions reflecting these changes had gradually been made to this view and that this decision was in line with them. No reference was made, or had to be made, to any grand moral principles other than those that followed from the immediately related cases and what Lord Lane called 'the true position of a wife in present day society'. In other words, their reasoning paralleled that over exam classification outlined above. It may still be objected that analogical reasoning per se offers no genuine constraint on discretion. Every case is analogous to every other case in some ways and disanalogous in others. Consequently, you can find a precedent for pretty much any interpretation you might wish to offer. Logically this is true. However, we should not take the decisions and opinions of courts in isolation from those of other courts. Courts form part of a complex system of fairly autonomous units. They need to coordinate their activities, for the most part without relying on a centralized or hierarchical structure to do so. Judges do not derive their authority from being modern Solomons. Rather, they are expected to be experts on, and impartial appliers of, the law. Consequently, they have an interest in stressing that they are only saying what everyone else is saying. Lawyers share this interest. Their professional standing similarly rests on their legal expertise. Since they will appear before many comts, they want them to operate pretty much like one another. As a result, the likeliest way of winning a case is to stress that the court is acting like other comts rather than appealing to natural justice or some original, however brilliant, reading of the law. In other words, there is a strong systemic pressure towards

xxxviii

The Rule of Law and the Separation of Powers

stare decisis. Through an examination of tort law in the US and the UK, Mattin Shapiro has shown that, within this context, analogical reasoning and reference to precedent operates as a feedback mechanism (Shapiro, 1972; see also Shapiro and Sweet, 2002, ch. 2). The massive number and variety of cases dealt with by courts generate a great deal of 'noise' in their communications with each other. Referring to an excessively large number of cases provides a mechanism for ensuring that each comt has moved in line with the others. It allows for feedback to correct errors or to note adaptations to meet new environments. As a result, decision-making is small 'c' conservative, with change incremental and path-dependent. In certain respects, this system allows law to rule for similar reasons to the way democratic legislators place themselves under the law. For the most part, there is no rational-hierarchical centralized decision-maker trying to run the system according to some common plan. Instead, there are a number of lm·gely independent decision-makers who nevertheless wish to work harmoniously with each other. In other words, it is the balance of power among comts that renders them consistent, not the existence of a Supreme Court ensuring that all decide according to constitutional principles or rules. In fact, these bodies are far more likely to overthrow precedent and be subject to disagreement precisely because they lack the same incentives to coordinate with everyone else. Conclusion The Rule of Law has always seemed a paradoxical notion given that it relies on the rule of persons to uphold it. I have argued that we have to confront this dilemma head-on. No feature of law per se can secure its rule. Rather, it is precisely because persons (in the plural) rule that they place themselves equally under the law. Law operates within rather than above the 'circumstances of politics'. As such, it needs to be sensitive to difference and diversity. A balance of power between groups both within and competing to join the legislature encourages just that. Of course, there will always be asymmetries of power in certain areas. However, these weaknesses in the democratic system require democratic remedies. No other mechanism or metric exists for deciding fairly and legitimately what equal treatment requires. The role of courts is to uphold the resulting laws consistently and impattially. Since judges have the same biases and prejudices as anyone else, they cannot be expected to exercise discretion any more wisely or objectively than anyone else. Their partiality is constrained by the need to argue through the law in a way that is comprehensible and acceptable to the rest of the legal community. The doctrine of stare decisis and the discipline of arguing analogically via cases cmbs the temptation to engage in grander theorizing and allows those who disagree on many of these wider issues to agree on the patticular merits of a given decision. In both politics and law, a concern for the public interest and equality before the law comes from the balancing of particular interests and views rather than the imposition of a general or universal perspective from outside the political or legal system.

Notes In Britain, for example, the then Lord Chief Justice has criticized plans for legal and asylum reform as undermining the Rule of Law. See Woolf (2004).

The Rule of Law and the Separation of Powers

xxxix

See the round-up of press reactions to Woolf's remarks in 'His Ambition is Rule by Lawyers', The Guardian, 6 March 2004. 3 I take this term from Mackie, this volume, Chapter 8. Although intended as a characterization of Dworkin's theory, it captures the goal of a range of theorists, such as Fuller (1969) and Hayek (1973-79), with whom Dworkin otherwise disagrees. 4 For the contrast between 'formal' and 'substantive' approaches, see Craig (this volume, Chapter 4). 5 The issue of whether a democratic account of the Rule of Law requires a constitutional or other courts to uphold the rights and procedures intrinsic to democracy, as is argued by Ely (1980) and, in more qualified terms, by Dworkin (1986a, ch. 1) is bracketed here to be dealt with in a companion volume on Constitutionalism and Democracy. However, at least one reason why this proposal proves unsatisfactory is that these criteria are as disputable as any other law or right, so that placing their protection in the hands of a court suffers from parallel difficulties to them and is itself subject to the Hobbes challenge. For a discussion of some of the difficulties, see Bellamy and Castiglione (1997). 6 As Pettit (1999, pp. 52-58) has shown, we owe the highlighting of this third aspect of arbitrariness to the Republican political tradition. 7 Obviously the possibility of very occasional arbitrary acts by individuals could never be eliminated within the circumstances of justice. Attempting to do so for all but a race of angels would inevitably involve the use of arbitrary power. 8 Fuller (1969, p. 39) offers eight criteria, which Finnis (1980, p. 270) reorders and Raz (this volume, pp. 80--84) extends and elaborates. Rawls (1971, pp. 236-39) is content with four, which Margaret Radin reduces to two- that 'there must be rules' and that 'those rules must be capable of being followed'- (this volume, p. 41), the second of which Raz also sees as capturing 'the basic idea' of the doctrine (this volume, p. 84). My analysis draws on all these accounts. 9 The locus classicus of this discussion in the Anglo-American literature is the debate between Hart and Fuller in Hart (1958) and Fuller (1958). For an interesting reprise of this argument, see Kramer (this volume, pp. 3-5) and Simmonds (2004). 10 In their debate, Kramer rightly notes how Simmonds's ingenious examples ofTaliban-like authorities operating entirely outside the Rule of Law depend on the implausible assumptions that the Rule of Law is an all or nothing affair and that a regime could be concerned solely with achieving its ideological goals to the neglect of all other areas of social life. 11 For an excellent account of the advantages of rules, see Schauer (1991, especially ch. 7) and Sunstein (1996, pp. 110-15). 12 United States v. White 322 US 503, 515 (dissent) quoted in Greenawalt (1984, p. 107). 13 See the detailed account of this tradition in relation to Machiavelli and Rousseau in Holmes (2003, ch 1). See, too, Bellamy (200 1, especially pp. 241-49) from which this section of this Introduction draws. 14 Harrington expressed this point in terms of the fable of the girls and the cake. The best way of ensuring a fair division is to make sure that the one who divides the cake takes the last slice. Thus, he advocated having two councils, one to deliberate and the other to enact policy, with the first controlled by the aristocracy and the second by the whole populace (Harrington, 1992, pp. 22-25, 64-67). 15 This distinction between rules of 'absolute', 'strict' and 'discretionary' application comes from MacCormick (1998). 16 This is the problem with approaches such as those of Scalia (1989) or Campbell (1996, pp. 129-48). For criticism see respectively Sunstein (1997) and MacCormick (2000). 17 For these three broad categories of interpretative argument (linguistic, systemic, and justice and utility), see MacCormick (1993). 18 I am grateful to Sheldon Leader for drawing this case to my attention. I have taken the details from the judgment of the European Court of Human Rights inCR v. United Kingdom (1996) 21 EHRR 363. 19 The appeal to the ECHR claimed there had been a violation of Article 7 (1) of the European Convention on Human Rights, whereby 'No one shall be held guilty of any criminal offence on account of any act or omission which did not constitute a criminal offence under national or international law at the time when it was committed'. 20 These Rule of Law-type objections to the discretionary effects of Bills of Rights are made by Goldsworthy (2001) and Campbell (2001). 2

xl

The Rule of Law and the Separation of Powers

References Bellamy, Richard (1999), Liberalism and Pluralism: Towards a Politics of Compromise, London: Routledge. Bellamy, Richard (2001), 'The Rule of Law and the Rule of Persons', Critical Review of International Social and Political Philosophy (CRISPP), 4, pp, 221-51. Bellamy, Richard and Castiglione, D, (1997), 'Constitutionalism and Democracy: Political Theory and the American Constitution', British Journal of Political Science, 27, pp. 595-618. Berlin, Isaiah (1969), 'Two Concepts of Liberty', in Four Essays on Liberty, Oxford: Oxford University Press, pp. 162-66. Campbell, Tom (1996), Legal Theory of Ethical Positivism, Aldershot: Dartmouth. Campbell, Tom (2001), 'Incorporation through Interpretation', in T. Campbell, K.D. Ewing and A. Tomkins (eds), Sceptical Essays on Human Rights, Oxford: Oxford University Press, Ch. 5. Dicey, Albert V. (1959), Introduction to the Study of the Law of the Constitution (lOth edn), London: Macmillan. Dworkin, Ronald (1978), Taking Rights Seriously, London: Duckworth. Dworkin, Ronald (1986a), 'Political Judges and the Rule of Law', in Matters of Principle, Oxford: Clarendon Press. Dworkin, Ronald (1986b), Law's Empire, London: Fontana. Ely, John H. (1980), Democracy and Distrust: A Theory of Judicial Review, Cambridge, MA: Harvard University Press. Finnis, John (1980), Natural Law and Natural Rights, Oxford: Clarendon Press. Fuller, Lon L. (1958), 'Positivism and Fidelity to the Law: A Reply to Professor Hart', Harvard Law Review, 71, pp. 630-72. Fuller, Lon L. (1969), The Morality of Law (rev. edn), New Haven, CT: Yale University Press. Goldsworthy, Jeffrey (2001), 'Legislative Sovereignty and the Rule of Law', in T. Campbell, K.D. Ewing and A. Tomkins (eds), Sceptical Essays on Human Rights, Oxford: Oxford University Press, Ch. 4. Greenawalt, K. (1984), 'Policy, Rights and Judicial Decision', in M. Cohen (ed.), Ronald Dworkin and Contemporary Jurisprudence, London: Duckworth. Habermas, Jiirgen (1996), Between Facts and Norms: Contributions to a Discourse Theory of Law and Democracy, Cambridge: Polity Press. Hamowy, R. (1987), 'Law and the Liberal Society' ,Journal of Libertarian Studies, 2, pp. 287-97. Harrington, James (1992), The Commonwealth of Oceana, ed. J.G.A. Pocock, Cambridge: Cambridge University Press. Hart, H.L.A. (1958), 'Positivism and the Separation of Law and Morals' ,Harvard Law Review, 71, pp. 593629. Hart, H.L.A ( 1994), The Concept of Law (2nd edn), Oxford: Clarendon Press. Hayek, Friedrich A. (1944), Road to Serfdom, London: Routledge. Hayek, Friedrich A. (1960), The Constitution of Liberty, London: Routledge. Hayek, Friedrich A. (1973), Rules and Order, London: Routledge. Hayek, Friedrich A. (1973-79), Law, Legislation and Liber~v, 3 vols, London: Routledge. Hayek, Friedrich A. (1976), The Mirage of Social Justice, London: Routledge. Hobbes, Thomas (1991), Leviathan, ed. Richard Tuck, Cambridge: Cambridge University Press. Holmes, S. (2003), 'Lineages of the Rule of Law', in J.M. Maravall and A. Przeworski (eds), Democracy and the Rule of Law, Cambridge: Cambridge University Press. Kymlicka, W. (1995), Multicultural Citizenship, Oxford: Clarendon Press. McCormick, Neil (1989), 'Spontaneous Order and the Rule of Law: Some Problems', Ratio Juris, 2, pp. 41-54. McCormick, Neil (1993), 'Argumentation and Interpretation in Law', Ratio Juris, 6, pp. 16-29. McCormick, Neil (1998), 'Norms, Institutions and Institutional Facts', Law and Philosophy, 17, pp. 31118. McCormick, Neil (2000), 'Ethical Positivism and the Practical Force of Rules', in T. Campbell and J. Goldsworthy (eds), Judicial Powe1~ Democracy and Legal Positivism, Ashgate: Dartmouth, Ch. 2.

The Rule of Law and the Separation of Powers

xli

MacKinnon, Catherine (1989), 'On Difference and Dominance', in Feminism Unmodified: Discourses on Life and Law, Cambridge, MA: Harvard University Press. Pettit, Philip ( 1999), Republicanism: A Them~v ofFreedom and Government (2nd edn), Oxford: Clarendon Press. Rawls, John (1971), A Theory of Justice, Oxford: Oxford University Press. Raz, Joseph (1994), 'Rights and Individual Well-Being', in Ethics in the Public Domain: Essays in the Morality of Law and Politics, Oxford: Clarendon Press, pp. 5-24. Scalia, Antonin (1989), 'The Rule of Law as a Law of Rules', University of Chicago Law Review, 56, pp. 1175-88. Schauer, F. (1991), Playing by the Rules: A Philosophical Examination ofRule-based Decision Making in Law and in Life, New York: Oxford University Press. Shapiro, Martin (1972), 'Toward a Theory of Stare Decisis', Journal of Legal Studies, 1, pp. 125-34. Shapiro, Martin and Sweet, A. Stone (2002), On Law, Politics and Judicalization, Oxford: Oxford University Press. Shklar, Judith (1987), 'Political Theory and the Rule of Law', in Allan Hutcheson and P. Monahan (eds), The Rule of Law: Ideal or Ideology?, Toronto: Carswell. Skinner, Quentin (1996), Reason and Rhetoric in the Philosophy of Hobbes, Cambridge: Cambridge University Press. Skinner, Quentin (1998), Liberty Before Liberalism, Cambridge: Cambridge University Press. Simmonds, N.E. (2004), 'Straightforwardly False: The Collapse of Kramer's Positivism', Cambridge Law Journal, 63, pp. 98-131. Sunstein, Cass (1996), Legal Reasoning and Political Conflict, New York: Oxford University Press. Sunstein, Cass (1997), 'Justice Scalia's Democratic Formalism', Yale Law Journal, 107, pp. 529-67. Viroli, M. (1998), Machiavelli, Oxford: Oxford University Press. Waldron, Jeremy (1989), 'The Rule of Law in Contemporary Liberal Theory', Ratio Juris, 2, p. 82. Waldron, Jeremy (1990), The Law, London: Routledge. Waldron, Jeremy (1999), Law and Disagreement, Oxford: Oxford University Press. Weale, A. (1999), Democracy, Basingstoke: Macmillan. Wittgenstein, Ludwig (1953), Philosophical Investigations, Oxford: Blackwell. Woolf, Lord (2004), 'The Rule of Law and a Change in the Constitution', Squire Centenary Lecture, University of Cambridge, 4 March.

Part I Defining the Rule of Law

[1] ON THE MORAL STATUS OF THE RULE OF LAW MATTHEW

H.

KRAMER*

BY the phrase "the rule of law", I mean nothing more and nothing less than the state of affairs that obtains when a legal system exists and functions. In Lon Fuller's felicitous formulation (somewhat modified and amplified), the rule of law is the subjection of human conduct to the governance of legal norms through the operations of a legal system. Indeed, the rule of law as understood throughout this article is admirably encapsulated in Fuller's eight principles of legality. 1 Under those principles, (1) a system of governance operates through general norms, and all or most of the norms partake of the following properties: (2) they are promulgated to the people who are required to comply with them; (3) they are prospective rather than retrospective; (4) they are understandable rather than hopelessly unintelligible; (5) they do not contradict one another and do not impose duties that conflict; (6) they do not impose requirements that cannot possibly be fulfilled; (7) they persist over substantial periods of time, instead of being changed with disorienting frequency; and (8) they are generally given effect in accordance with their terms, so that there is a congruence between the norms as formulated and the norms as implemented. Although these basic precepts of legality are never perfectly satisfied by any regime of law, the satisfaction of each of them to a substantial degree is essential for the existence of any legal system. The state of affairs constituted by the substantial fulfilment of those precepts is the rule of law. Some of my previous writings in support of legal positivism have aimed to show that the rule of law is not an inherently moral ideal. The first main section of the current article will attempt to * Professor of Legal 1

& Political Philosophy, University of Cambridge; Fellow of Churchill College. This article is excerpted, with many modifications, from a long chapter of my bookin-progress Where Law and Morality Meet. For a full-scale exposition of the eight principles of legality, see Lon Fuller, The Morality of Law, revised edition (New Haven 1969), chap 2. I have examined Fuller's work at length in my In Defense of Legal Positivism (Oxford 1999) [hereinafter cited as IDLP], ch. 3.

4

The Rule of Law and the Separation of Powers

The Cambridge Law Journal

66

[2004]

elucidate this point further and to parry some criticisms that might be leveled against it. Then, in the second main section, we shall examine some objections that have recently been lodged against my past arguments about the rule of law. Some powerful rejoinders by Nigel Simmonds to those arguments will provide occasions for clarifying and strengthening the legal-positivist position. I.

REASONS RATHER THAN PROBABILITIES

My challenges to the status of the rule of law as an inherently moral ideal have often focused on heinous legal systems in which the officials are driven by purely prudential motivations relating to the consolidation of their own power and the exploitation of the citizenry. 2 Nevertheless, as will become clear below, those challenges do not commit me to any view about the frequency with which the officials in iniquitous legal regimes are actually motivated by purely prudential considerations. On the one hand, the prevalence of prudential factors among the actual determinants of the conduct of officials in wicked regimes is almost certainly underestimated by many legal philosophers. On the other hand, any claim about the actual determinants of officials' behaviour-whether in accordance or at odds with the view stated in the preceding sentence-is an empirical assertion that cannot be substantiated or disconfirmed through philosophical reasoning. To ascertain the considerations that in fact motivate evil officials, we would have to undertake a lot of social-scientific research within the legal systems of a wide range of countries. Such a project would undoubtedly involve complex questionnaires and interviews as well as careful observation of the patterns of official decisions and actions. After all, exactly what typically motivates nefarious legal officials is a matter that will hinge on contingent features of human psychology and sociocultural influences. Neither any legal-positivist philosopher nor any philosophical opponent of legal positivism has ever engaged in extensive research of the sort that would be required for the adequate investigation of those psychological and sociocultural phenomena. Yet the fact that jurisprudential theorists have not pursued the complicated empirical enquiries just mentioned is hardly due to remissness on their part. Rather, any enquiries along those lines would be largely or wholly beside the point, since the 2

Through most of this article, I rely on a standard distinction between morality and prudence. Somebody's prudential reasons-for-action are focused exclusively or primarily on his own interests and only derivatively if at all on the interests of other people Somebody's moral reasons-for-action are focused exclusively or primarily on other people's interests and only derivatively if at all on his own interests (apart from interests, such as a concern for acting in a morally proper fashion, which are themselves defined by reference to the well-being of other people).

The Rule of Law and the Separation of Powers

C.L.J.

Rule of Law

5

67

controversies between the proponents and the foes of legal positivism are philosophical rather than empirical. To be sure, the arguments mustered in those controversies do often rely on some empirical premises. However, the operative premises are at a very high level of abstraction and generality, and are therefore not the points at issue between the supporters and detractors of jurisprudential positivism; although those premises figure m controversies, they are not themselves controversial. Instead, the chief points of contention in debates over the soundness of legalpositivist theses are conceptual. More specifically for our present purposes, the chief points of contention pertain to reasons-for-action. Here the phrase "reasonsfor-action" refers not only to factors that actually do motivate people, but also to factors that would motivate them if they were to understand the serviceability of those factors for the furtherance of their general objectives. Because of this counterfactual element in many reasons-for-action, my reflections on the motivational underpinnings of the rule of law are not empirical. Those reflections are focused primarily on the sundry considerations that are promotive of the ends of wickedly self-interested officials, whether or not the considerations are actually perceived by the officials as so promotive. Irrespective of whether the considerations get taken into account when the evil officials choose how to behave, they are among the officials' reasons-for-action if they would get taken into account by anyone who is unremittingly clearsighted in pursuing the officials' aims. At the same time, the conception of reasons-for-action that informs my discussions of the moral status of the rule of law does not encompass external interest-independent reasons-that is, reasons which stand as reasons for a selfish person P regardless of whether they serve any of the objectives that are actually desired by P. Although interest-independent reasons not accepted as such by P would of course have to be included in any comprehensive account of reasons-for-action, no such account is needed or envisaged here. I concentrate in this article on strictly prudential considerations simply because we are here pondering whether such considerations are sufficient to warrant a substantial degree of adherence to ruleof-law principles by selfishly power-hungry officials. For the contemplation of that issue, a restrictive conception of reasons-foraction is singularly appropriate. Let us look, then, at the paramount question-a conditional question-that is addressed by my critique of the moral status of the rule of law. If the officials who operate a morally deplorable system of governance are motivated by purely prudential

The Rule of Law and the Separation of Powers

6

68

The Cambridge Law Journal

[2004]

considerations relating to the consolidation of their own power and the exploitation of the citizenry, will they have solid reasons for abiding by rule-of-law requirements to a significant extent? In the light of what has been said in the last couple of paragraphs, we know that the solid reasons need not be grasped and heeded as such by the evil officials; but we also know that those reasons, insofar as they exist, are purely prudential. (As I have indicated elsewhere-IDLP, 66-we further know that the relevant prudential reasons must consist in something more than opportunities to deceive others by appearing to be acting morally. If one's only prudential reason for engaging in a certain mode of behaviour B lies in the likelihood that one will be perceived as acting on the basis of solicitude for the interests of others, then one's devious conduct is in effect a testament to the moral significance of B.) What is of most importance for us to notice afresh is that the conditional question about the wicked officials' reasons-for-action is not an empirical inquiry about their actual motivations. One can answer the question perfectly well through philosophical argumentation, without speculating about the frequency with which the officials in flagitious regimes are in fact animated by purely selfish concerns. An affirmative answer to my conditional question is perfectly consistent with the empirical claim that the officials in nefarious regimes are always or almost always motivated by nonprudential considerations (morally warped non-prudential considerations, of course, such as a concern for the purity of some racial or ethnic group). Although I do not subscribe to such an empirical thesis, a denial of it here would be wholly superfluous. My legal-positivist critique of the moral tenor of the rule of law does not depend on any such denial, for it does not contend that the factors which actually drive the conduct of repressive officials are of this or that particular type. Whatever those factors may be in any given setting, there will obtain strong prudential reasons for the officials to exert their sway through the rule of law-at least if they aspire to exert their sway for a sustained period over a society larger than a handful of families. That such strong prudential reasons-for-action will indeed obtain is a proposition which my In Defense of Legal Positivism has endeavoured at length to substantiate-in order to bear out my positivist denial of the inherently moral character of the rule of law-with some lines of argumentation that will be laconically summarised here. 3 Those lines of argumentation centre on the efficacy of law in guiding and channelling human conduct. For 3

The lines of argument are presented chiefly in chapters 3 and 4 of JDLP, but they undergo elaboration in some of the subsequent chapters

The Rule of Law and the Separation of Powers

C.L.J.

Rule of Law

7

69

unjust officials and also for just officials, the advantages of law are threefold: it provides clear-cut direction; it fosters incentives for obedience; and it enables officials to coordinate their activities. A system of legal governance presents citizens with definite indications concerning what is demanded of them. Equally important, it creates strong incentives for citizens to comply with their rulers' demands, because it tends to ensure that disobedient people are punished and that obedient people are left unpunished. Furthermore, it greatly enhances the coordination of the myriad instances of interaction among the officials themselves, who could not carry out a largescale enterprise of governance over a sustained period of time without the regularity and the complicatedly interlocked expectations which a functional legal system makes possible. In sum, if power-hungry rulers are determined to exert and reinforce their repressive sway for a long period over a sizeable society, their efforts will be severely set back if they do not avail themselves of the coordination and the incentive-securing regularity made possible by the rule of law. Whether the requirements imposed and the objectives pursued by officials are products of moral concern or of exploitative selfishness, the officials can most effectively achieve their ends through the operations of a legal system. Although the preceding paragraph supplies no more than a terse conspectus of some of the chief points for which I have argued at length elsewhere, it is sufficient for present purposes. Instead of elaborating those points anew, my discussion here will explore an ostensible problem or shortcoming which arises from the fact that my analysis of the moral status of the rule of law leaves open the possibility that wicked legal officials are never or almost never driven by purely prudential concerns. A critic might retort by questioning whether my conclusions about that moral status are really justified. Such a critic may well accept that there are solid prudential reasons for the exploitative officials to conform with rule-of-law requirements, but he will query the importance of such an observation. Given that my arguments concerning the officials' reasons-for-action do not rule out the possibility that their motivations are in fact entirely non-prudential, the critic may feel that I have not genuinely cast doubt on the inherently moral character of the rule of law. If the animating concerns of evil officials in their compliance with rule-of-law precepts are not the prudential reasons for such compliance, and if as a matter of fact those animating concerns are always non-prudential, then why should we not regard the rule of law as a morally pregnant state of affairs? Though such a state of affairs is not always associated with morally worthy effects, its actual motivational underpinnings are

The Rule of Law and the Separation of Powers

8

70

The Cambridge Law Journal

[2004]

(ex hypothesi) always moral in the sense of being non-prudential. Why then, the critic may ask, should the legal-positivist assessment of the rule of law not be discounted as something that lacks any purchase on the real world? One quite minor weakness in an objection of this sort, of course, is that it treats a possibility as if it were an established actuality. My legal-positivist approach to the rule of law-which addresses the conditional question specified at the beginning of the antepenultimate paragraph above-has left open the possibility seized upon by the hypothetical critic, simply because my approach does not commit me to any view whatsoever on the actual promptings that lead the officials in a repressive regime of law to operate it as a regime of law. Not dealing with that matter at all, I scarcely have to concede that the critic's supposition about the invariably moral character of those promptings is correct. Still, that weakness in the critic's challenge can be put aside here, for his retort founders in a more serious respect. The unacceptable crudeness of his challenge becomes manifest when we consider some of the conclusions that would be generated if his line of reasoning were to be applied to other patterns of behaviour. Let us ponder the example of shooting a gun at somebody else at point-blank range. Suppose that, as a matter of empirically verifiable fact, everybody or virtually everybody who deliberately engages in such a mode of conduct is motivated by non-prudential considerations (albeit often deeply misguided and iniquitous considerations). Perhaps because of the horrific consequences of shooting someone else in cold blood and because of limits on human callousness, no one or virtually no one who deliberately fires a gun at point-blank range is prompted to engage in such conduct by selfish urgings. We may suppose that only someone driven wholly or predominantly by non-prudential motives is psychologically capable of pumping a bullet into a nearby human being. Only such motives can give rise to the requisite degree of unrelenting sternness. Although I believe that such an empirical thesis is false, it might conceivably be true as a contingent matter of human psychology, and in any event its truth-value is beside the point when we gauge the moral status of the act of shooting another person. For the task of apprehending that status, we have to ask a conditional question parallel to the one which I have posed about the rule of law. If someone were motivated by unalloyedly prudential concerns, would there be strong reasons in highly credible circumstances for him to fire a gun directly at somebody else?

The Rule of Law and the Separation of Powers

C.L.J.

Rule of Law

9

71

Especially if we ignore the prospect of a person's being punished for committing such a misdeed-a prospect that has no bearing on the moral character of the misdeed itself-the answer to this conditional question is plainly "yes". There will clearly be numerous credible contexts in which somebody driven solely by self-interested considerations will have strong reasons (strong prudential reasons) for shooting some other person(s), at least if there is little or no chance of his undergoing sanctions for his homicidal conduct. Perhaps no one will ever actually act upon those prudential reasons, because of the putative facts of human psychology hypothesised in my last paragraph; and thus maybe acts of homicidal shooting will occur only on the basis of nonprudential zeal. Nevertheless, the weighty prudential reasons for many such acts will obtain as such. Precisely because those reasons will indeed obtain, the pattern of conduct that consists in firing a gun at somebody else at close range is not inherently moral. It is not a mode of behaviour that can hardly ever be credibly described as strongly promotive of the selfish interests of somebody who resorts to it. If the hypothetical critic's animadversions on my analysis of evil officials' adherence to rule-of-law principles were correct, they would apply as well to any comparable analysis of the act of shooting a person at point-blank range. If one rejects the animadversions in application to the act of firing a gun, and if one maintains that the ample serviceability of that mode of conduct for the selfish interests of any thug who undertakes it is enough to justify our denying that it is intrinsically moral, then one should adopt the same position with regard to the wicked officials' fulfilment of rule-of-law requirements. Of course, to say as much is not to say that we can safely ignore all differences between the two patterns of behaviour. We manifestly have to take some such differences into account. Most notably, the tendency of an instance of homicide to further the selfish aims of an evil person who perpetrates it is immediately obvious, whereas the corresponding tendency of the rule of law to further the selfish aims of nefarious officials who operate a legal regime is not nearly as evident on first glance. Exactly because of this dissimilarity between the slaying of someone with a gun and the sustainment of the rule of law, I have felt a need to offer lengthy arguments that highlight the expediency of the latter mode of behaviour for the furtherance of the powerhungry interests of repressive officials. Only through such arguments, which have been presented on a full scale in In Defense of Legal Positivism and summarised fleetingly above, can we recognise that the correct answer to my conditional question about

The Rule of Law and the Separation of Powers

10

72

The Cambridge Law Journal

[2004]

the rule of law is affirmative. Nevertheless, what those arguments enable us to grasp is that the answer to that question is indeed affirmative-just as surely (though not as obviously) as the answer to the parallel question about the act of shooting a gun at someone at point-blank range. Note that cognate conditional questions can pertinently be posed about countless other modes of conduct that are perhaps never pursued by people except on the basis of non-prudential concerns. Let us contemplate, for example, the activity of gambling at cards. Maybe, as a contingent matter of human psychology, such an empty and degrading mode of conduct never sufficiently entices anyone who is not inspired by solicitude for others to seek a means of acquiring money rapidly. Though such an empirical claim is hardly very convincing, it could conceivably be true. Possibly, every card-playing gambler is like the grandfather portrayed in The Old Curiosity Shop, whose reckless attempts to win at cards were motivated entirely by his love for his granddaughter Nell. As Charles Dickens wrote, the old man "gambl[ed] with such a savage thirst for gain as the most insatiable gambler never felt, [yet] had not one selfish thought!" 4 Even in the highly unlikely event that all card-playing gamblers are similarly motivated, we scarcely have to accept that the pursuit of gain through gambling at cards is an inherently moral endeavour. About that pursuit we should ask a conditional question similar to the ones that we have asked about adhering to rule-of-law requirements and firing a gun at a nearby person. If somebody prone to gambling were motivated exclusively by selfish promptings, and if his chances of winning at cards were as high as he himself believes, would there be strong reasons for him to follow his card-playing propensity? Like the answer to each of our other conditional questions, the answer to this latest conditional question is plainly affirmative. If someone with a decent chance of faring well at cards is driven wholly by prudential concerns, there will be many credible contexts in which he will have solid reasons for gambling at the card table. To be sure, perhaps nobody ever actually acts on the basis of those prudential reasons; perhaps every gambler instead resorts to betting on cards solely for the furtherance of non-prudential aspirations. Nonetheless, even if that exceedingly implausible empirical suggestion is true, anyone with good prospects of success will still have solid prudential reasons for indulging in such betting. His inclination or disinclination to be moved by such reasons is independent of their existence as such. Precisely because those reasons will indeed exist 4

Charles Dickens, The Old Curiosity Shop (Everyman edition, London 1995), p. 231.

The Rule of Law and the Separation of Powers

C.L.J.

Rule of Law

11

73

in many credible circumstances, the pattern of behaviour that consists in the pursuit of winnings through cards is not inherently moral in its motivational underpinnings-even if its actual motivational underpinnings are always moral as a matter of fact. We could go through homologous lines of analysis in respect of sundry other types of conduct. The basic point would of course be the same in each case. If the hypothetical critic were correct in taking exception to my account of the rule of law, then it would also be true that firing guns at people and gambling at cards-and countless other dubious activities-are intrinsically moral. In other words, the critic's rejoinder does not tell against my account of the rule of law at all. By focusing on a question about actual motivations (or, rather, about motivations that supposedly are actual) when the germane question is about reasons-for-action, the rejoinder would lead us to classify absolutely every kind of behaviour as inherently moral. It would thereby remove all significance and content from the category of "inherently moral kinds of behaviour", at least insofar as our assessments of behaviour are oriented toward its motivational underpinnings. Far from having shown that the bearings of the rule of law are in any way distinctive, the hypothetical critic's retort would in effect be contending that officials' sustainment of the rule of law is on a par with every other way in which human beings might act. Naturally, the hypothetical critic might respond along the following lines. With regard to acts such as homicidal shootings and activities such as gambling at cards, we would be not only wrong but plainly wrong if we were to assume that the actual underlying motivations are always non-prudential. By contrast, so the critic might maintain, such an assumption about actual motivations is not so evidently wrong when applied to the conformity by authoritarian officials with rule-of-law requirements. Although the critic might allow that the assumption is incorrect even in this last-mentioned case, he would add that it is not manifestly so. Whereas we need no detailed empirical enquiries in order to feel confident that the actual motivations for homicidal shootings and for wagers on cards are not invariably nonprudential, some detailed enquiries will be necessary if we are to feel similarly confident about the actual motivations for evil officials' adherence to rule-of-law principles. Such, at least, is what the hypothetical critic might affirm in order to redeem his rejoinder to my positivist reflections on the rule of law. Any response of this sort by the critic would be intensely problematic. Obviousness is a scalar property which obtains in

The Rule of Law and the Separation of Powers

12

74

The Cambridge Law Journal

[2004]

varying degrees. 5 Hence, the critic's latest retort would suggest that the difference between the moral status of wicked officials' cleaving to the rule of law and the moral status of gun-shooting or cardplaying exploits is at most a matter of degree. Having conceded as much, the critic is then vulnerable to arguments which indicate that the difference of degree is quite meagre. In this connection, let us mull over a couple of points. We should first recall that, although compliance with Fuller's precepts of legality is prerequisite to the existence of a legal system, the degree of compliance can be considerably less than unfailing perfection. Even a benignly liberal-democratic regime of law will fall quite a way short of adhering perfectly to the Fullerian precepts. (Indeed, given that crimes of many types are probably detected more readily by a stiflingly repressive regime than by a liberal-democratic regime that leaves its citizens ample room for individual freedom, the satisfaction of Fuller's eighth principle-the effecting of a congruence between the law as formulated and the law as enforced-is probably often realised to a greater extent within illiberal systems of law.) A repressive and exploitative regime will frequently fall even further short than a liberal-democratic regime in the degree of its conformity with most of the precepts of legality. After all, the repressive rulers will be disposed to depart from those precepts on occasions when liberal-democratic rulers would feel morally inhibited from doing so. (On the other hand, the repressive rulers will be disposed to show inflexibility in circumstances that would probably elicit greater complaisance on the part of liberal-democratic rulers.) Still, although a grimly exploitative regime will not be meticulously unwavering in its fulfilment of the principles of legality, its officials will have solid prudential reasons for abiding by those principles to quite a substantial degree. If they act in accordance with those reasons, then the level of their compliance with rule-of-law requirements will be sufficient to endow their system of governance with the character of a legal system. Now, what should be highlighted here is that the status of their system of governance as a legal system is quite compatible with the fact that the wicked officials are not punctiliously unswerving in their adherence to Fullerian principles. It is quite compatible with the occurrence of their unscrupulous deviations from those principles, as long as the deviations are not so wide-ranging and numerous as to plunge their regime below some threshold of regularity and coordination. Thus, when we ponder whether the actual motivations for evil officials' sustainment 5

For a discussion of the difference between scalar and non-scalar properties, see Matthew H. Kramer, The Quality of Freedom (Oxford 2003), pp. 169-184.

The Rule of Law and the Separation of Powers

C.L.J.

Rule of Law

13

75

of the rule of law are not always non-prudential, we should keep in mind that heinous incarnations of the rule of law are very likely marked by aberrations that would not occur-or would not occur on the same scale-under a liberal-democratic scheme of governance. Though it would probably not be plausible to maintain that officials who cleave with extreme fastidiousness to the Fullerian principles are motivated by prudential concerns, the nefarious regimes that receive most attention in my discussions of the rule of law are not regimes wherein officials are generally fastidious. They adhere to the Fullerian principles to a significant extent (if their systems of governance are indeed systems of law), but their adherence is combined with ruthless deviations when their interests are thereby served. When we have in view this combination of oppressively efficient governance and unconscionable departures from norms, the evil officials' sustainment of the rule of law will no longer seem so different from homicidal shootings and betting on cards, in respect of its actual motivations. When we recognise that the legal systems on which we should be concentrating for present purposes are hardly characterised by much squeamishness, we can discern that the difference of degree mentioned in my last paragraph is not sizeable. A second reason for doubting the sizeableness of the difference is that the hypothetical critic has failed to distinguish adequately between two sets of actual motivations: motivations for governing, and motivations for governing in accordance with the rule of law. It might be that the actual motivations for governing, even among iniquitously power-hungry officials, are always non-prudential at least in large part. It might be, in other words, that no array of officials driven predominantly by prudential considerations could muster the tenacity and self-confidence needed for imposing their baleful dominion over the other people in their society. Perhaps only when the evil officials are captivated by the allure of some twisted non-prudential objective, such as racial supremacy or religious purity or socialist revolution, will they be filled with sufficiently ruthless determination to place themselves in charge of the direction of their society. So, at least, someone might conjecture. Although a hypothesis along those lines would in my view be far from compelling, a rejection of it would require some elaborate empirical investigation. We need not reject it here, in any event. Much more dubious, however, would be a kindred hypothesis that relates not to the actual motivations for governing tout court but to the actual motivations for governing in accordance with the rule of law.

The Rule of Law and the Separation of Powers

14

76

The Cambridge Law Journal

[2004]

In order to train our attention on the latter set of motivations, we should take as given that some wickedly repressive officials have gained sway over a society. We should also take as given, for the sake of argument, the truth of the empirical hypothesis broached in the preceding paragraph. That is, we should assume that the wicked officials are motivated chiefly by some hideous non-prudential aspirations rather than mainly by purely selfish concerns. Are we obliged to infer that the actual motivations for the officials' sustainment of the rule of law (if indeed their regime is a regime of law) are likewise non-prudential? Especially when we recall that the sustainment of the rule of law by autocratic officials will typically not be marked by fastidiousness, we can see that the answer to this question is negative. Even if ex hypothesi the aims that have prompted the officials to take or retain control of their society are non-prudential, the motivating factors that lead them to abide by rule-of-law principles can very credibly be the strictly instrumental considerations that also render the rule of law serviceable for the ends of exploitative officials who are motivated to govern by thoroughly prudential concerns. That is, when repressive officials driven to govern by non-prudential objectives have opted to adhere to the rule of law, their actual motivations for doing so can very credibly reside in the direction-providing and incentive-promoting and coordination-facilitating advantages of the rule of lawadvantages that would be just as noteworthy in the eyes of the officials if their motivations for governing were instead entirely prudential. They need not look upon the rule of law as per se expressive of their objectives or indeed of any non-prudential aspirations. They might instead very credibly regard it as valuable only instrumentally. To be sure, their reliance on the rule of law for the tightening of their grip over their society is ultimately in the service of non-prudential goals, or so we are assuming. Nevertheless, they can very credibly view the rule of law not as an integral element of those goals but simply as an instrument for the effective realisation thereof. In adopting such a view, they will be embracing the rule of law on the basis of the same considerationswholly instrumental considerations-that would militate in favour of their embracing it even if their underlying concerns were utterly selfish. In any such credible circumstances, the non-prudential tenor of the evil officials' actual motivations for governing is not matched by the tenor of their actual motivations for governing through the rule of law. In short, the relevant empirical issues are by no means as straightforward as the hypothetical critic's latest riposte has implied. The difference of degree highlighted in that riposte is

The Rule of Law and the Separation of Powers

Rule of Law

C.L.J.

15

77

indeed at most a difference of degree, which turns out to be much more tenuous than it might at first seem. My response to the critic's original challenge therefore stands. His challenge can salvage the inherently moral status of the rule of law only by trivialising that status. Unless we are prepared to attribute such a status to homicidal shootings and to wagers on cards, we should not attribute it to the conformity of officials with rule-of-law precepts. II.

SIMMONDS AND THE RULE OF LAW

In a textbook which in its second edition has become an estimably imaginative and perceptive work of legal philosophy in its own right, 6 Nigel Simmonds replies forcefully to my strictures on his account of the rule of law. That powerful account, presented in the course of his discussion of Fuller's writings, came under critical scrutiny by me in In Defense of Legal Positivism. Indeed, some of my central arguments concerning the moral status of the rule of law have evolved partly in reaction to Simmonds's ruminations on that status. Accordingly, his defence of his position against my queries will have to be countered if my positivist theory of law is to be vindicated. We shall not herein examine all of his lines of reasoning, but we shall probe a few of his principal retorts. Let us begin, however, by noting that Simmonds often tends to obscure one of the principal distinctions highlighted in the preceding section of this article: the distinction between actual motivations and reasons-for-action. As has been observed, reasonsfor-action obtain as such even when they do not actually motivate people to behave in certain ways. Simmonds obfuscates this potential divergence between motivations and reasons when he writes that "[m]uch of the debate relating to Fuller's theory ... revolves around rival assertions regarding the probable or possible behaviour of wicked regimes. Would they have good reasons for establishing and respecting the institutions of law?" 7 As should be apparent from what has been said already in this article, only the second sentence in this quotation identifies the key question addressed by my critique of the moral status of the rule of law. The first sentence, by contrast, adverts to an empirical question from which my critique prescinds. To be sure, Simmonds speaks of probable or possible behaviour rather than of probable or possible motivations. However, that difference makes no difference in the present context. Largely because reasons-for-action obtain 6 7

N.E. Simmonds, Central Issues in Jurisprudence, 2nd. edn (London 2002) [hereinafter cited as Simmonds, Issues]. Ibid., p. 226 For a passage that similarly elides the distinction between actual motivations and reasons-for-action, see ibid., p. 230

The Rule of Law and the Separation of Powers

16

78

The Cambridge Law Journal

[2004]

independently of their efficacy in actually motivating anyone to seek to adopt certain courses of conduct, they likewise obtain independently of anyone's actual adoption of those courses of conduct. Thus, for example, the officials in vilely repressive and exploitative regimes have solid reasons (solid prudential reasons) for abiding by rule-of-law requirements to a considerable extent, whether or not they grasp those reasons and act on the basis of them. Even if the officials in all or most such regimes are somehow too obtuse to heed those reasons-and even if they attenuate and abbreviate their grip on power as a result-the strong prudential reasons for their compliance with rule-of-law principles will exist. Because those reasons will exist as such, the sustainment of the rule of law is not an inherently moral pattern of behaviour. This conclusion about the moral status of the rule of law derives entirely from reflections on evil officials' reasons-for-action, and not from any surmises about the ways in which the officials will be inclined to behave. Consequently, only the second sentence in the quotation from Simmonds accurately pinpoints the paramount issue around which the debates arising from Fuller's work have revolved. A. Simmonds's First Major Retort: The First Strand

In his first main rejoinder to my exploration of the moral bearings of the rule of law, Simmonds conjures up a scenario in which some Taliban-like despots have imposed a stifling religious orthodoxy on the society over which they exercise power. He adduces this scenario in order to recount a situation in which a grimly repressive regime can hold sway on a large scale for a long period without abiding by rule-of-law precepts to any significant extent. A "pure regime of terror may work perfectly well" under the circumstances outlined in the scenario, or so Simmonds contends (Simmonds, Issues, 230). Quite a lengthy quotation from his argument is necessary here in order to set the stage for my rebuttals: Suppose, for example, that a group of religious fundamentalists has the sole aim of discouraging the propagation of atheism, being content to leave all other aspects of social life (such as the operation of markets and the protection of property rights) to be sustained (or not, as the case may be) by informal customary practices unsupported by official rules and sanctions. Unregulated brutality towards anyone suspected of sympathy for atheism would serve such a regime perfectly well. Atheists and potential atheists would be deprived of the opportunity to exploit the interstices of liberty that are inherent in any meticulously enforced set of rules; they would know that the only way to be safe was to avoid doing

The Rule of Law and the Separation of Powers

C.L.J.

Rule of Law

17

79

anything that might annoy the regime's supporters, or arouse their suspicions. (Simmonds, Issues, 230.) Let us begin our examination of this passage by considering a few minor points, before we turn to the chief weakness in the argument. In the first place, as should be evident from some of my remarks in the first main section of this article, Simmonds's use of the word "meticulously" near the end of the quoted passage is illadvised.8 Although in some respects the officials in an oppressive regime may well have solid prudential reasons for being more meticulous than a liberal-democratic regime in following the Fullerian precepts-for example, in enforcing prohibitions with stern unftinchingness-they will very likely have solid prudential reasons for being less meticulous in other respects. Nowhere in any of my writings have I intimated that wickedly exploitative officials will have strong reasons for complying punctiliously with every one of the Fullerian principles. They must comply to quite a considerable degree with each principle if their regime is to operate as a system of legal governance at all, but the extent of the conformity can fall notably short of punctiliousness. Hence, insofar as Simmonds is arguing that the officials in the Taliban-like regime will not have weighty prudential reasons for adhering to rule-of-law precepts assiduously, he is knocking down a straw man. What he instead should be trying to show is that the officials will have no such reasons for adopting the looser level of adherence which I have described as promotive of the ruthless effectiveness of authoritarian rulers. Another minor shortcoming in Simmonds's example lies in the nature of the regime on which it focuses. Whereas my discussions have trained attention primarily on rulers for whom starkly prudential considerations are of dominant importance, all or most of the Taliban-like rulers are driven predominantly or exclusively by heinous non-prudential concerns of religious purity. Still, as should be apparent from the first main section of this article, the inappositeness of Simmonds's example in the respect just mentioned can be pretermitted here. After all, by taking due account of the distinction between motivations for governing and motivations for governing through the rule of law (or between reasons for governing and reasons for governing through the rule of law), we can perceive that even officials whose retention of their grip on power is animated by non-prudential aspirations will have 8

Equally unfortunate is the following sentence, which relies on the word "assiduous" for its truth: "A regime that makes an assiduous effort to follow Fullerian precepts is motivated by a moral aspiration, not by a desire to advance wicked goals with maximum efficiency" (Simmonds, Issues, 233).

The Rule of Law and the Separation of Powers

18

80

The Cambridge Law Journal

[2004)

instrumental reasons for abiding by rule-of-law requirementsinstrumental reasons that might constitute their actual motivations for upholding the rule of law (if the officials do indeed uphold it). As was argued earlier, the fulfilment of rule-of-law requirements can be directly motivated by strictly instrumental considerations even when it is ultimately in the service of non-prudential objectives. Thus, if we concentrate on the direct reasons for governing through the rule of law, and if we are careful to distinguish them from the ultimate motivations for governing, my critique of the moral status of the rule of law will work as well with Simmonds's example of the Taliban-like despots as with the examples of selfishly exploitative officials on which I have generally focused. Of somewhat greater significance is another dubious feature of Simmonds's example. It is not clear in what sense the example's religious zealots have formed a governing regime. They are not regulating most aspects of life in their society in any way, and they are not providing an infrastructure that facilitates any basic governmental functions. Instead, they are much more like bandits who are endeavouring to enforce their will in one quite narrow area of life while leaving everything else ungoverned. (Indeed, given the extreme unlikelihood that any sizeable society will stay together if it is virtually anarchic, the example could more realistically have been presented as a depiction of a situation in which the overbearing zealots' campaign against atheism is conducted within the framework of the operations of some genuine governmentperhaps a government just as harshly repressive as the coterie of religious zealots.) Accordingly, the fact that the anti-atheistical militants do not run a legal regime is due to their not running any veritable regime at all; the fact that they do not govern in accordance with the rule of law is due to their not governing, period. Of still more importance is a tension between two of the statements in the passage quoted above. On the one hand, Simmonds declares that the religious fanatics are pursuing "the sole aim of discouraging the propagation of atheism". On the other hand, at the end of the passage he submits that the fanatics' policy of unregulated brutality will lead people to "know that the only way to be safe [is] to avoid doing anything that might annoy the regime's supporters, or arouse their suspicions". The fanatics' sole aim of snuffing out atheism is by no means most effectively achieved through the engendering of a state of affairs along the lines described in the second of these quoted statements. That state of affairs is fully consistent with a situation in which the

The Rule of Law and the Separation of Powers

C.L.J.

Rule of Law

19

81

unregulated brutality is very frequently inflicted on devout believers who have been falsely accused by the regime's supporters (maybe out of personal animosity and vindictiveness, or perhaps out of ignorance and hysteria, or maybe out of sheer corruption). If the brutality of the religious fanatics is indeed unregulated, then the opportunities for efficacious false accusations will abound. Insofar as the persecution of many devout believers does ensue, it will certainly not strengthen the fanatics' hold over the populace-since it will very likely antagonise and eliminate people who would otherwise be inclined to support the harsh imposition of religious uniformity. Even more clearly, such persecution by the coterie of religious zealots will not help them to attain their sole aim of discouraging the propagation of atheism. On the contrary, although that aim might still be realised if the terrible penalties inflicted on pious believers are combined with the infliction of terrible penalties on atheists, the former set of penalties will at best do nothing to advance the desired objective. What is more likely is that the penalties imposed on believers will set back the coterie's sole objective. For one thing, as has just been noted, people who would otherwise have been favourably disposed toward the cadre of zealots will have become alienated from them. They might still cooperate with the fanatics out of fear, but they will no longer be inclined to cooperate in an enthusiastic manner bred of sympathy. Moreover, the atheists in the society may well adopt a "nothing to lose" attitude when they perceive that they will stand a significant chance of undergoing severe punishments even if they abstain from any atheistic practices. The prospect of escaping such punishments by forgoing such practices is constitutive of a strong incentive to forgo them; inasmuch as that prospect is impaired, the concomitant incentive is likewise impaired. If an atheist who might refrain from giving expression to his views would not very much lower his chances of being subjected to ferocious treatment, he may well wonder why he should refrain. Of course, it is not inevitable that the upshot of the use of unregulated brutality by the religious militants will include the frequent perpetration of violence by them against pious believers. Nor is it inevitable that their conformity with rule-of-law requirements would avert such a state of affairs. Nonetheless, the probability of the occurrence of counterproductive violence by the fanatics against devout individuals is considerably higher in what Simmonds describes as "a pure regime of terror" than in a regime of law. Consequently, there are strong reasons for the militants to abide by rule-of-law principles to quite a substantial degreeinstrumental reasons concerning the effectiveness with which the

The Rule of Law and the Separation of Powers

20

82

The Cambridge Law Journal

[2004]

militants can achieve their sole aim of discouraging atheism. Whether or not the religious despots recognise that there are those reasons for them to adhere to rule-of-law precepts, and whether or not they act on the basis of those reasons, the reasons obtain as such. If the lone objective of the despots is to stanch the propagation of atheism, they are most apt to succeed if their ruthless pursuit of that objective is channelled through the rule of law. (This conclusion would of course be underscored if we were to take account of the fact that a pure regime of terror which operates through unregulated savagery will not only be prone to punish many compliant people but will also be prone to leave many disobedient people unpunished. More precisely, its proneness to errors in each direction will in typical circumstances be greater than the error-proneness of a legal regime. Nevertheless, throughout his discussion of the religious extremists, Simmonds appears to envisage a situation in which the enforcement of anti-atheistical norms is markedly over-inclusive without being significantly underinclusive. Since my replies to his discussion are adequate without my having to challenge his assumption on that particular point, I shall indeed leave the assumption unimpugned here.)

B. Simmonds's First Major Retort: The Second Strand Simmonds allows that the anti-atheistical zealots will very likely extend their sway into areas of life other than that of religious practices. He therefore allows that they will establish a full-fledged legal system. Furthermore, he readily accepts that the propagation of atheism will be dampened if the legal system includes norms that prohibit atheistical activities and if those norms are stringently enforced when breached. However, he maintains that the religious fanatics will have no reasons to refrain from punishing people for engaging in various atheistic practices or utterances that do not infringe the prohibitory norms. On the contrary, he contends, they will have strong reasons for disregarding the terms of such norms by cracking down on any manifestations of atheism irrespective of whether the manifestations violate those terms. At least in their regulation of the spiritual side of people's lives, then, the religious extremists will have strong reasons for declining to satisfy Fuller's eighth principle of legality (relating to the need for a congruence between the formulation and the implementation of legal norms). Simmonds argues for this claim as follows: But why should the regime treat as immune from officially organised violence those atheistic activities that do not violate any published and prospective rule? After all, some atheistic activities may emerge in forms not anticipated by the religious

The Rule of Law and the Separation of Powers

C.L.J.

Rule of Law

21

83

regime, but they may be just as obnoxious to the regime (and just as much to be discouraged) as those activities that were indeed anticipated and prohibited. By ordering or permitting ad hoc official violence against the perpetrators of such obnoxious activities, the regime would not only discourage the particular activities in question, but would also discourage the expenditure of intellectual effort in dreaming up further innovative forms of atheistic practice aimed at escaping the existing prohibited categories. (Simmonds, Issues, 231.) Now, although this quoted passage should be read in conjunction with the paragraph that immediately follows it (on which we shall concentrate presently), we should pause here to note en passant two questionable features of this extract. First, in arguing that punishments for lawful atheistic activities will be serviceable for keeping people docile, Simmonds strangely assumes that the legal norms which prohibit atheistic undertakings will each be quite narrowly framed and will thereby leave open the possibility of atheistic practices that are not covered by the norms. Much more likely is that a broad prohibition on atheistic practices will be included among those norms. Just as the Hebrew Scriptures lay down both narrowly targeted bans on idolatrous practices and wholesale bans on idolatry, so too the Taliban-like extremists will almost certainly issue blanket proscriptions of atheism as well as more narrowly defined proscriptions of specific activities associated with godlessness. Such an approach by the extremists need not involve any departures from Fuller's precepts of legality; and, in comparison with the approach sketched by Simmonds, it will more directly and perspicuously convey to potential atheists the message that their proclivities will not be tolerated. Accordingly, the Taliban-like tyrants have strong instrumental reasons for adopting the strategy outlined here, in preference to the strategy recounted by Simmonds. Whether or not the tyrants grasp those reasons and act upon them, the reasons obtain as such. Second, in any event, I have never maintained that no deviations or hardly any deviations at all from rule-of-law principles will be serviceable for the entrenching of an evil regime's power and the fulfilment of its objectives. Weighty though the direction-providing and incentive-promoting and coordinationenabling advantages of the rule of law are, a vile regime will typically have to leaven them with the advantages of flexibility that come with departures from rule-of-law requirements. As I have readily acknowledged elsewhere (IDLP, 70), even very substantial departures from such requirements during quite a brief period or within quite a confined geographical or substantive area can consolidate the dominance of an iniquitous regime without

22

The Rule of Law and the Separation of Powers

84

The Cambridge Law Journal

[2004]

depriving it of its general character as a regime of law (and therefore without depriving it of the advantages attendant on that general character). Thus, if Simmonds's only point were that transgressions of rule-of-law precepts can reinforce a heinous regime's grip on power, his argument would not really clash with anything said here or in my previous writings on legal positivism. However, it straightaway becomes plain that Simmonds is putting forward a much bolder thesis. He brands as "straightforwardly false" my claim that a wicked regime will tend to sap incentives for compliance with its mandates if it does not abide by rule-of-law requirements to a considerable extent. He declares that "such a situation would not be brought about by the scenarios [relating to the Taliban-like fanatics] that I have described above; nor does the avoidance of such an undermining of incentives require compliance with Fuller's eight principles" (Simmonds, Issues, 231-232). He endeavours to support these assertions by asking us to contemplate two situations. In the first situation, the citizens governed by a repressive regime are punished for violating the regime's promulgated norms, "but they also suffer on a very frequent but irregular basis, random acts of violence perpetrated by officials of the regime". In the second situation, the citizens are punished for violating the regime's promulgated norms, "but, with equal frequency, they are also punished for activities obnoxious to the ruling powers, although not prohibited in any published and prospective rule" (Simmonds Issues, 232). Having delineated these two sets of circumstances, Simmonds proceeds to deliver his central line of argument against me: In neither of these situations would incentives for compliance with the rules be reduced by the occurrence of official acts of violence not provided for in the rules, or provided for only in retrospective rules. For, by violating the rules, I would be greatly increasing my chances of incurring a punishment, or my chances of incurring an additional punishment. Even if unregulated official violence is so widespread that citizens can all expect to be its victims from time to time, it is difficult to see how this would undermine their motives for complying with the rules. The fact that you will beat me if I break the rules gives me a good reason for complying with them even if you frequently beat me at random: after all, the beating in consequence of breaking the rules will be an additional beating that I could have avoided by compliance. (Simmonds, Issues, 232.) Before assailing the line of reasoning in this extract, I should mention a couple of preliminary points on which there is no disagreement between Simmonds and me. First, in any thorough

The Rule of Law and the Separation of Powers

C.L.J.

Rule of Law

23

85

examination of punishment-centred incentives for conformity with legal mandates, two main factors are of key relevance: the probability and the severity of the threatened punishments for nonconformity. Incentives for obedience will be increased if either of those two factors is increased, and the incentives will be diminished if either factor is diminished. Nonetheless, purely for ease of exposition, my discussion here (like Simmonds's discussion) will concentrate on the probability of one's undergoing punitive measures and will put aside the matter of the severity of those measures. In other words, my rejoinders to Simmonds will proceed as if the probability of punishment were the only variable factor and as if the severity were constant. Nothing whatsoever of substance hinges on this simplifying assumption, which is operative solely for reasons of stylistic clarity. A second preliminary point on which there exists no disagreement between Simmonds and me is that strictly speaking the incentives for one's compliance with mandates are determined not by the probability of one's undergoing penalties for noncompliance, but by the perceived probability of such penalties. Unlike reasons-for-action, which obtain whether or not the person for whom they are reasons is aware of them, incentives are awareness-dependent. A person will not be induced to modify his behaviour by reasons-for-action of which he remains ignorant. All the same, like Simmonds, I shall for present purposes set this matter aside and shall assume that the actual probability of punishment is the perceived probability thereof. Such an assumption is warranted again chiefly on stylistic grounds of clarity and smoothness. In addition, the actual probability and the perceived probability will correspond pretty well under most circumstances, especially if the actual probability is high. At any rate, once again, nothing of any philosophical importance hinges on my simplifying assumption. My analyses below could stand up perfectly well even if they took explicit account of the potential for discrepancies between actual probabilities and perceived probabilities; however, nothing of substantive interest would be added to the analyses, and they would become stylistically much more gnarled. Let us, then, disregard the potential for such discrepancies. Now, in coming to grips with Simmonds's argument, we should straightaway observe that the relevant probability is comparative. That is, the probability of one's undergoing some punishment as a result of one's disobedience is only one key element; also of crucial importance is the probability of one's undergoing some punishment even though one has been obedient. Suppose that, if Joe violates a

The Rule of Law and the Separation of Powers

24

86

The Cambridge Law Journal

[2004]

legal norm during a certain period, the probability of his being apprehended and imprisoned during that period is 90%. Suppose further that, if he does not violate any such norm during that period, the probability of incarceration is only 5%. In these circumstances, Joe manifestly has a strong punishment-centred incentive for compliance with the norms of the legal system under which he lives. The incentive resides not simply in the highness of the 90% level but in the disparity between 90% and 5%. To see as much, let us suppose that the latter figure were instead 90%. In that case, Joe would have no punishment-centred incentive at all to adjust his conduct in order to comply with the prevailing legal mandates. The probability of his undergoing some punishment would not be lowered even slightly by his compliance. If the original 5% level were increased to 95%, then Joe would actually have a punishment-centred incentive for flouting the norms of his legal system. By disobeying those norms, he would in fact lower the probability of his being penalised. In short, as can easily be discerned, the punishment-centred incentives for anybody's conformity with legal norms will vary directly with the probability of his being penalised for nonconformity and inversely with the probability of his being penalised despite his conformity. The overall punishment-centred incentive consists in the gap between those two probabilities. In my arguments concerning the tendency of the rule of law to promote punishment-centred incentives for obedience, one of my claims has been that those incentives decrease as the aforementioned gap decreases. The smaller the disparity between the two specified probabilities, the weaker the inducement for complying with the prevailing legal norms. 9 If for example the second of the two probabilities relating to Joe has increased from 5% to 50% (say), his punishment-centred incentive for restraining himself from transgressing the mandates of his legal system will have waned considerably. If his restraint of himself is burdensome, then the impairment of his punishment-centred incentive for obedience may well have tilted the balance of his motivations in favour of noncompliance. Much the same will be true, of course, in comparable situations involving other people. Thus, given that a legal system with its regularised effectuation of its mandates will generally keep low the probability of the infliction of punishments on people who heed those mandates, and given that it will generally keep high the probability of the infliction of punishments on people who 9

For the sake of stylistic simplicity, I am assuming here that the disparity will remain positive; that is, I am assuming that the probability of undergoing punishment for non-compliance will remain higher than the probability of undergoing punishment despite compliance.

The Rule of Law and the Separation of Powers

C.L.J.

Rule of Law

25

87

contravene the mandates, it will tend to foster incentives for obedience. It will typically fare much better in fostering such incentives than will a system of governance that persistently fails to effect a congruence between its norms as articulated and its norms as enforced. On the basis of the last two paragraphs, we can now detect the central fallacy in Simmonds's line of reasoning. Before we look at that fallacy, however, we should glance at two minor objectionable aspects of his pronouncements. First, in statements that have been quoted above and in one statement that has not been quoted, Simmonds three times refers to the undermining of incentives for obedience. In so doing, he conveys the impression that my arguments about the incentive-promoting advantages of the rule of law have suggested that substantial deviations from Fuller's eighth principle of legality will do away altogether with each person's punishment-centred incentives for compliance. In fact, my arguments have adverted to the sapping or weakening of such incentives rather than to the wholesale destruction of them. Only in extreme circumstances is a system of governance so sweepingly erratic or so wildly incompetent that it will have thoroughly undermined everyone's punishment-centred incentives for conformity with its directives. In respect of such extreme circumstances, indeed, we can rightly query whether a system of genuine governance would be operative at all. Much more credible are situations in which the punishment-centred inducements for obedience have been attenuated rather than eliminated. Such situations-rather than the far less credible circumstances in which the inducements have been entirely eliminated-are what I have usually had in mind when pointing to the eventualities which the incentive-promoting tendency of the rule of law will help to avoid. My arguments which highlight that very tendency have not implied that a regime's significant departures from the rule of law will perforce reduce to zero the disparity between the probabilities specified in my last paragraph. Rather, those arguments have more mildly and realistically indicated that such departures will lessen that disparity. A second minor shortcoming in Simmonds's discussion surfaced briefly at the end of my last subsection. In his quotations from In Defense of Legal Positivism and in his own analyses, Simmonds glosses over the fact that the disparity just mentioned can be diminished not only by the frequent imposition of penalties on obedient people, but also by the frequent failure to impose penalties on disobedient people. This point is of some importance because a regime that falls markedly short in its adherence to rule-

The Rule of Law and the Separation of Powers

26

88

The Cambridge Law Journal

[2004]

of-law principles will tend to impair incentives-for-compliance in each of these two directions. Both because of the relatively low level of coordination among its officials and because of the relatively abundant opportunities for corruption on their part, 10 such a regime will typically be significantly under-inclusive as well as significantly over-inclusive in its punitive measures against uncompliant people. Hence, even if Simmonds were correct about the effects of over-inclusiveness, he would still have to confront the incentive-sapping effects of under-inclusiveness. In fact, however, Simmonds is wrong about the effects of overinclusiveness. His analysis of the matter, let us recall, is presented by reference to the two situations which he delineates-situations in which people are very frequently punished even if they have committed no infractions of the norms which they are called upon to obey. Let us suppose that for each person the probability of undergoing some punishment for disobedience is 90% while the probability of undergoing some punishment despite obedience is 60%. We should now look more closely at two of the sentences from the passage quoted most recently above: "[B]y violating the rules, I would be greatly increasing my chances of incurring a punishment, or my chances of incurring an additional punishment .... The fact that you will beat me if I break the rules gives me a good reason for complying with them even if you frequently beat me at random: after all, the beating in consequence of breaking the rules will be an additional beating that I could have avoided by compliance". When Simmonds proclaims that a violation of the rules will "greatly increas[e] my chances of incurring a punishment" and that the beating inflicted for the violation will be "an additional beating that I could have avoided by compliance", he has slipped into assuming that the chances of his undergoing a beating if he abides by the rules will be slim. He could not confidently characterise the punishment as "an additional beating that I could have avoided by compliance", if the probability of his undergoing that beating in spite of remaining compliant is 60%. Yet, by implicitly presuming that the chances of his suffering the punishment will be very low if he conforms with the rules, he is assuming away the most salient characteristic of each of the two situations which he has sketched: the fact that in each situation the probability of anyone's undergoing a punishment in spite of 10

I should note that corruption, in the sense of a willingness by an official to accept payments or other inducements in return for declining to enforce the norms which he is supposed to enforce, will in some contexts be morally preferable to his unwavering performance of his role. When a guard in a Communist gulag or a Nazi concentration camp takes bribes in return for conniving at transgressions of the odious rules which he is supposed to enforce, he acts less appallingly than if he were to give effect to the rules with unflinching sternness.

The Rule of Law and the Separation of Powers

C.L.J.

Rule of Law

27

89

fulfilling the requirements of the prevailing norms is very high rather than low. Were that probability instead very low-5%, say, rather than 60%-each of the depicted situations would be marked by few anomalies in its punitive procedures and would therefore exemplify the incentive-promoting regularity of the rule of law. In other words, in the course of denying that either of the situations would involve the dampening of punishment-centred incentives for obedience, Simmonds has slipped into assuming that each of those situations instantiates the rule of law. It is little wonder, then, that he does not find in either of them any of the incentive-impairing effects that are to be expected in the absence of the rule of law! It should go without saying that nothing in my last paragraph's riposte to Simmonds depends on the sequentiality or simultaneity of the multiple punishments which he envisions. He asks us to contemplate a set of circumstances in which some person undergoes a disobedience-triggered punishment and a random punishment. Given the numbers stipulated here, the probability of the infliction of the former punishment was 90%, and the probability of the infliction of the latter punishment was 60%. Thus, unless the occurrence of either punishment would affect the likelihood of the occurrence of the other-a possibility about which more will be said in a moment-the probability of the two punishments together was 54%. That probability has obtained irrespective of whether the two punishments occur at the same time or at successive times during some defined period. It is to be compared with the probability of the occurrence of two random punishments. Unless the imposition of one random punishment would affect the likelihood of the imposition of another, the probability of their joint occurrence is 36%. Again, that probability has obtained irrespective of whether the two punishments occur (if they do occur) at the same time or at successive times during some defined period. Like the disparity between 90% and 60%, the disparity between 54% and 36% is something to be taken into account by anybody who is deciding whether to flout a mandate of the regime under which he lives. However, the ratio of 3:2 between either of those pairs of figures is much smaller than would be the ratio if the probability of the infliction of each random punishment were much lower. Pro tanto, then, the incentives for submissiveness are weakened. Of course, the occurrence of one punishment might impinge on the likelihood of the occurrence of another, even though Simmonds has given no indication to that effect. Two observations are appropriate in response to this possibility. First, the basic point of my rejoinder to Simmonds would be wholly unimpeached by such a

The Rule of Law and the Separation of Powers

28

90

The Cambridge Law Journal

[2004]

possibility, though the nature and extent of the rejoinder's applicability to the amended scenarios might become complicated. My basic claim is that, so long as the probability of the infliction of a random punishment is high (as it is stipulated to be in each of his scenarios), the beating inflicted as a disobedience-triggered punishment cannot safely be characterised as a hardship that could have been avoided through compliance. That central claim of mine is true even if we somehow discover, in respect of an altered scenario, that the probability of a random punishment is low. My claim would not be applicable to the altered scenario, but it would still be true. Second, the potential ways in which the occurrence of one punishment could affect the probability of the occurrence of another are myriad. We certainly cannot know a priori which of those ways-if any-will prove to be actual in any particular system of governance. To be sure, if we know that each random punishment is indeed strictly random, then we can know that the probability of its occurrence will not be modified by the occurrence of any other punishment (whether that other punishment is disobedience-triggered or is itself strictly random). However, we can instead understand "random" in this context to mean "only incidentally or weakly correlated with disobedient conduct", in order to leave some room for the modification of the probability of a random penalty during a defined period by the imposition of some other penalty during that period. In that case, the occurrence of a punishment of either type (disobedience-triggered or random) could lower or raise the probability of the occurrence of another punishment of either type. Whether there will be any such effects in some particular system of governance, and whether they will be augmentative or diminutional, and how intense they might be, are plainly empirical questions that cannot be answered in the abstract. Fortunately, these untidy empirical questions can be put aside, since any answers to them will not detract at all from the correctness of the central contentions in my rejoinder to Simmonds. Albeit this or that answer to any of those questions may complicate the applicability of my contentions-as the last paragraph above has readily acknowledged-it will not cast any doubt on their truth. Let us JOin Simmonds, then, m exploring these empirical complexities no further. Before moving on to probe one additional line of reasoning by Simmonds, we should briefly examine his two scenarios in order to determine whether there are any significant differences between them. Most of my remarks in the past several paragraphs have concentrated implicitly if not explicitly on the situation outlined in

The Rule of Law and the Separation of Powers

C.L.J.

Rule of Law

29

91

his first scenano, a situation of very frequent acts of random punishment. His second scenario, which adumbrates a regime that punishes people for "activities obnoxious to the ruling powers" as well as for violations of the regime's promulgated norms, may be quite different. On the one hand, if the regime's categorisation of ostensibly permissible activities as obnoxious is itself unpredictably capricious and continually shifting, then the second of Simmonds's scenarios does not differ from the first. On the other hand, the regime's categorisation of ostensibly permissible activities as obnoxious might not be random at all. There might be clear and established patterns that are discernible by citizens who acquaint themselves with the regime's punitive measures. If so, then the regime's operations may well be sufficiently regularised and ascertainable to constitute a regime of law-at least after those operations have been running for some time. To be sure, as Simmonds indicates, the ostensibly permissible activities deemed by the regime to be obnoxious are "not prohibited in any published and prospective rule". Nevertheless, the practice of penalising people who engage in those activities might be patterned and predictable to such an extent that the penalties are manifestations of implicit norms which can be grasped by any minimally intelligent citizen who observes the practice. Officials' acts of imposing the penalties would in effect be the means by which those norms are promulgated. Even when punishing people for engaging in ostensibly permissible activities, then, the regime would be operating in accordance with the Fullerian purpose of subjecting human conduct to the governance of norms. In that event, its workings cannot serve as the grounds for any conclusions about the tenability or vibrancy of a regime that does not govern through the rule of law. After all, those workings in such circumstances would amount to governance through the rule of law. C. A Further Retort

Simmonds launches another piquant challenge to legal positivism by asking us to contemplate an imaginary society called "New Monia" (Simmonds, Issues, 244---249). Drawing creatively on the work of Max Gluckman, he recounts a very small and simple society in which the means of giving effect to settlements of disputes are highly limited. Because of the absence of police forces and prisons, and because of the dense multifacetedness of people's relationships with one another, the dominant manner of dealing with disputes is to press the parties toward compromise and reconciliation. The focus for defusing any conflict lies on the overall relationship between the parties and on the need for give-

The Rule of Law and the Separation of Powers

30

92

The Cambridge Law Journal

[2004)

and-take, rather than on the specific point of contention that has kindled strife between them. Each party will be expected to concede some ground on that specific point in order to sustain and improve his overall ties with the other party. Simmonds conjures up his scenario of New Monia in order to compare the arrangements for dispute-resolution in such a society with the corresponding arrangements in modern liberal-democratic societies. Parties to a dispute in a liberal-democratic country such as the United Kingdom or the United States might be able to resolve their differences through negotiation and compromise, but they might instead end up in court. If their conflict does become the object of legal proceedings, their respective entitlements will be settled by reference to norms that are applicable to the specific point of contention between them. The focus of the court will not lie on the overall relationship between the parties; rather, the specific controversy will be treated largely in isolation from broader contexts. Nor will the court's primary aim be to effect a compromise-based reconciliation between the parties. Although a settlement of the differences between them will remain a possibility especially in the early stages of the proceedings, the principal task of the court is to find in favour of one party or the other by fixing the respective entitlements of each. Instead of seeking harmony through a process of give-and-take, the court is in the business of producing winners and losers. Although Simmonds views the society of New Monia quite favourably, he observes that its approach to dispute-resolution is not without drawbacks. "New Monia pays a price for its desirable features, and the price is paid in individual liberty" (Simmonds, Issues, 246). Within a legal system like that of England, he maintains, individuals will be able to ascertain with considerable confidence and precision the things which they are at liberty to do. Moreover, they will be permitted to do those things irrespective of the feelings and sensibilities of other people. There is no need for compromise or concessions on the part of anyone, so long as she is within the scope of her entitlements. Things are quite different in New Monia. Whenever somebody there behaves in a way that annoys somebody else to more than a trivial extent, he or she may well be authoritatively pressured to desist from the behaviour in order to avert a deeper rift. New Monians favour congeniality and stability over individual freedom. "My ability to stand foursquare within my established rights and to tell the world to go to the devil is severely reduced" (Simmonds, Issues, 247). Simmonds proceeds to deliver a general statement of the dissimilarities between the

The Rule of Law and the Separation of Powers

C.L.J.

Rule of Law

31

93

posture of the individual in an advanced legal system and the posture of the individual in New Monia: Clear, published, prospective rules that are meticulously enforced by officials therefore serve liberty. No matter how narrowly the content of the rules may constrain my freedom, the very fact that they are ascertainable rules, and are reliably enforced, is likely to give me certain areas of entitlement within which I will be free from interference. In the world of New Monian compromise, by contrast, I will never be wholly free to behave in a manner that my fellows consider unacceptable or obnoxious. (Simmonds, Issues, 247.) Before we go on to consider why Simmonds has introduced the contrast that is summarised in this passage, we should investigate the contrast itself. Although Simmonds has written expertly elsewhere on Wesley Hohfeld's analysis of the four main types of entitlements that are commonly designated as "rights", 11 he here uses the term "rights" quite loosely. He describes the situation of the individual under the English legal regime as follows: "I know that I can behave in eccentric ways that my neighbours may object to: but, if I am acting within my legal rights, the law will protect me" (Simmonds, Issues, 247). This quoted sentence is correct if "rights" is a reference to claim-rights against interference with one's engaging in some activities, or (more likely) if it is a reference to such claim-rights combined with legal liberties to engage in those activities. So long as claim-rights against interference with one's performance of some action are operative, the law will indeed protect that performance. However, at this juncture in his discussion, Simmonds is focusing on legal liberties rather than on claim-rights. He is pondering a scenario in which two neighbours annoy each other by using their respective premises in irritating ways. Unable to reach a compromise, they wind up in court where each sues the other. "The court will seek to decide who is and is not acting within their rights" (Simmonds, Issues, 246)-by which is meant that the court will seek to decide whether each party is or is not legally at liberty to use her premises in the annoying manner for which she has opted. Thus, the reference by Simmonds slightly later to the protective role of rights is at best misleading. When entitlements designated as "rights" are legal liberties, they are protective only in the sense that legal enforcement mechanisms cannot successfully be mobilised to prevent their exercise.

11

Simmonds, Issues, chap. 8; N.E. Simmonds, "Rights at the Cutting Edge", in Matthew H Kramer, N E. Simmonds, and Hillel Steiner. A Debate over Rights (Oxford 1998), pp 113, 146-195.

32

The Rule of Law and the Separation of Powers

94

The Cambridge Law Journal

[2004]

The point made in the foregoing paragraph is more than a terminological quibble. Once we recognise that many legal liberties are not accompanied by claim-rights against all types of interference with their exercise, we can see that Simmonds's statements about the solid freedom of the individual under the rule of law are unduly sweeping. To discern as much, we should probe a lightly modified version of his own example. Suppose that a landowner Leonard plants a hideous shrub near the boundary between his plot of land and an adjacent plot that belongs to Michael. Michael complains bitterly about the eyesore that has been introduced by Leonard, but his expostulations come to nought. Michael therefore sues Leonard. He fails, however, as the court decides that Leonard is legally at liberty to plant and keep the shrub on his land. Unable to counteract Leonard's endeavours through the courts, Michael instead erects a high fence that blocks off the sunshine which the shrub needs in order to flourish. The shrub withers and dies, as Leonard fumes. He brings a legal claim against Michael, but he loses as the court decides that Michael is legally at liberty to build the fence and to obscure the sunlight. Thus, although Leonard has been found to be acting "within his rights" (that is, within his legal liberties), he has discovered that he cannot with impunity "behave in eccentric ways that my neighbours may object to" or "tell the world to go to the devil". Of course, the failure of the lawsuit by Leonard was hardly inevitable. Had the law been different, the court could have found him to be endowed with legal rights against interference with his shrub-growing (either against interference generally or against the specific sort of interference undertaken by Michael). The point of this example has certainly not been to indicate that there will be no protected leeway for eccentric behaviour by individuals under the rule of law. Rather, the point has been to highlight the fact that there are countless areas of damnum absque injuria within the law. Leonard is akin to other people in being frequently doomed to failure if he seeks to avail himself of the protective force of the law against somebody else's interference with his idiosyncratic pursuits. Although Leonard can know that he will not be prevented by legal officials from indulging his penchant for hideous shrubs, he cannot be similarly confident about a lack of obstruction from his fellow citizens such as Michael. Simmonds's neglect of this instance and the myriad other instances of damnum absque injuria in the law is one respect, then, in which his comments about the posture of the individual under the rule of law are too sweeping. (In an earlier

The Rule of Law and the Separation of Powers

C.L.J.

Rule of Law

33

95

section of his replies to me, 12 Simmonds himself offers a good example of damnum absque injuria. He imagines a situation in which the votaries of some political party are legally at liberty to search a dissident's home at any time. In such circumstances, her privacy will not be legally protected against their intrusions.) Perhaps even more important, however, are two other respectsoverlapping respects-in which Simmonds has exaggerated the connection between the rule of law and individual freedom. First, a legal system can bestow a suffocatingly extensive degree of protection on each individual against the eccentric behaviour of other people. Suppose that, instead of failing in his lawsuit against Leonard, Michael has prevailed. With the backing of the enforcement mechanisms of the state, he has effectively obliged Leonard to uproot the ugly shrub. Suppose further that Michael decides nonetheless to erect a garish fence on the border between his plot of land and Leonard's plot. Outraged by the fence, and thirsting for revenge in any event, Leonard seeks an injunction against the implementation of Michael's plan. Let us suppose that he succeeds. In this new version of the Leonard/Michael scenario, then, neither man is permitted to behave in an eccentric fashion that irritates his neighbour. Whereas the initial version of the scenario involves legal liberties on each side of the dispute, this new version involves legal duties. Each man is obligated to conform his behaviour to the sensibilities of the other. If the obligations extend over many potential facets of their behaviour, then each of them probably enjoys less latitude for idiosyncratic projects than is enjoyed by the typical inhabitant of New Monia. To be sure, it should go without saying that such a state of affairs under the rule of law is not inevitable; but such a state of affairs is indeed possible, and accordingly Simmonds's contrast between the promotion of harmony in New Monia and the promotion of freedom under the rule of law is simplistic. Second, equally important as a ground for querying Simmonds's stark contrast is the fact that the rule of law will typically be administered by an elaborate governmental apparatus, which can monitor and regulate people's conduct with stifling efficiency. If officials use that apparatus for the squelching of dissent and the exploitation of the citizenry, they may well limit the freedom of most people to a greater degree than would be true in New Monia. In the principal passage quoted above, Simmonds maintains that a system of legal norms will secure certain domains of freedom for 12

Simmonds, Issues, 239-240. Because of constraints of space, I have here had to leave some of Simmonds's replies undiscussed. I respond to them fully in the chapter from which this article has been excerpted.

The Rule of Law and the Separation of Powers

34

96

The Cambridge Law Journal

[2004)

each person regardless of how severely the norms curtail the range of things which each person is legally permitted to do. He juxtaposes that situation with the situation of individuals in New Monia, who are "never ... wholly free to behave in a manner that [their] fellows consider unacceptable or obnoxious". His comparison is eminently plausible in connection with liberal-democratic regimes of law, which undoubtedly intrude less into people's lives than does the conformity-inducing peer pressure in New Monia. However, the comparison is much less cogent in connection with repressively illiberal regimes, where the regularised enforcement of harshly restrictive mandates tends to ensure that individuals enjoy even less freedom than do New Monians. Because the relations among the New Monians are egalitarian to quite a considerable extent, each person will have reasons for putting up with some objectionable modes of behaviour on the part of his fellows; each can thereby incline his fellows to put up with his own peculiarities. Contrariwise, evil officials wielding dominance over hapless citizens do not have similar reasons to adopt a live-and-let-live stance of forbearance. They need not look to citizens for tolerance, and thus they have no reasons to endeavour to elicit it through reciprocal indulgence. Consequently, Simmonds's blanket comparison between the New Monian system and every legal system is far from compelling. We have no grounds for thinking that every credible legal system will endow individuals with more solid freedom than will the New Monian system of peer pressure. Let us now explore why Simmonds has drawn his contrast between the reconciliatory processes in New Monia and the rule of law. He explains that he is addressing the question whether officials' compliance with the Fullerian principles of legality is a genuine moral virtue. For that purpose, he rightly contends, we should not be comparing two full-fledged legal regimes of differing degrees of benignity. Instead, the baseline for comparison should be a mode of governance that does not operate through the rule of law. New Monia is ideally suited to serve as that baseline, for it is "a reasonably attractive society which is just and humane but which attaches little or no importance to the meticulous enforcement of clear, published and prospective rules" (Simmonds, Issues, 249). Simmonds is plainly correct in maintaining that the germane contrast is not between two legal regimes, and he is likewise correct in submitting that the New Monian system of governance is an excellent baseline for comparison. What is objectionable is the other side of his comparison. That is, although his resort to New Monia as a baseline for a comparison with a legal system is unexceptionable, his implicit resort to a morally commendable legal

The Rule of Law and the Separation of Powers

C.L.J.

Rule of Law

35

97

system as the counterpart for a comparison with New Monia is open to criticism. If we are seeking to gauge whether the rule of law is morally worthy, we must be careful not to mistake the worthiness of the contents of certain legal norms for the inherent worthiness of the rule of law itself. To avoid the prejudicing of our enquiry in that fashion, we should compare the arrangements in New Monia with an iniquitous legal system. Given that the contents of the norms in such a system are deplorable, we can be confident that any respects in which the arrangements of New Monia prove inferior will be due to the moral significance of the rule of law itself. Only by proceeding in this manner can we be sure that the outcome of our enquiry has not been distorted by the presence of substantive elements of worthiness (in a legal system) that are not intrinsic to the principles of legality. Yet, as has already been observed, a comparison between New Monia and an odiously repressive legal regime will by no means inevitably reveal that the scheme of governance in New Monia is less promotive of individual freedom than the dissent-stifling legal system with which it is being compared. On the contrary, the strict effectuation of the mandates in that legal system will tend to ensure that individuals within its jurisdiction are cramped even more severely in their activities than are the New Monians. If the officials in the system were considerably more relaxed in their enforcement of the especially draconian directives laid down by their regime, the situation of the people subject to those directives would perhaps compare favourably with the situation of the people subject to peer pressure in New Monia. Precisely because the officials are not relaxed in their enforcement of the draconian directives, however, the individuals governed by the repressive legal regime are very likely possessed of less leeway than the individuals in New Monia. In short, Simmonds's latest line of argument has turned out to be as unavailing as his previous rejoinders. His efforts to defend an anti-positivist account of the Fullerian principles of legality are dexterous and illuminating, as can be inferred from the length of my replies. Nevertheless, those efforts have foundered. Regardless of whether morality is understood in contrast with prudence or is instead understood in contrast with badness, the designation of the Fullerian principles as the "inner morality of law" is inapposite.

[2]

RECONSIDERING THE RULE OF LAW MARGARET JANE RADIN*

I.

INTRODUCTION

The ideal of "the rule of law, not of men" 1 calls upon us to strive to ensure that our law itself will rule (govern) us, not the wishes of powerful individuals. According to this traditional ideal, government must be by "settled, standing Laws," not by "Absolute Arbitrary Power." 2 Although the Rule of Law ideal is central to our legal tradition, it is deeply contested. Among those who affirm the traditional ideal there is no canonical formulation of its meaning, and critical theorists argue that the Rule of Law is mere ideology that should be jettisoned. In this essay I suggest that it is too soon to throw out the Rule of Law

* Carolyn Craig Franklin Professor of Law, University of Southern California Law Center. Thi.s article is an expanded and revised version of my lecture at Boston University School of Law presented on March 3, 1988. Earlier versions were presented to law faculty workshops at Tulane, Southern California, Michigan, and Yale. I am grateful to the hearers on these occasions; many of their comments range beyond the scope of the present essay, but they will continue to motivate me in further work. I am grateful, as well, to the many colleagues and friends whose sharp insight and critical honesty keep me trying to dispel the haze. 1 For obvious reasons, because I am considering the Rule of Law in today's context, I shall rephrase the ideal as "the rule of law, not of individuals." Yet we must not forget that when the ideal developed, and during most of its long history, it was inconceivable that any individuals who were not "men" could be a part of political life. 2 J. LOCKE, Of the Extent of the Legislative Power, § 137 in Two TREATISES OF GOVERNMENT (P. Laslett rev. ed. 1970) (3d ed. 1698) (emphasis omitted). The formulation of the term "the Rule of Law" is sometimes credited to the nineteenth-century British jurist Albert Venn Dicey, who elaborated upon three characteristics of "the rule or supremacy of law": absence of arbitrary power on the part of the government; ordinary law administered by ordinary tribunals; and general rules of constitutional law resulting from the ordinary law of the land. See A.V. DICEY, INTRODUCTION TO THE STUDY OF THE LAW OF THE CONSTITUTION 179-92 (1908).

The Rule of Law and the Separation of Powers

38 782

BOSTON UNIVERSITY LAW REVIEW

[VoL. 69: 781

wholesale, but that at minimum the concept should be reinterpreted. By way of concrete introduction to the complex of problems I mean to examine, consider Robinson v. Diamond Housing Corporation. 3 Robinson was one of Judge J. Skelly Wright's famous decisions remaking modern landlordtenant law in the early seventies. Judge Wright decided that the landlord, Diamond Housing Corporation, could not end Mrs. Robinson's month-to-month tenancy if its termination was motivated by retaliation against Mrs. Robinson for having successfully asserted the habitability defense of "illegal contract." 4 The result was that if a tenant lives in housing that seriously violates the housing code, he or she has a right to stay there, rent-free, until the landlord either fixes up the premises or proves economic compulsion to exit the landlord business. As Judge Wright put this, perhaps not fully mindful of the irony, precedent and the housing code "guarantee the right of a tenant to remain in possession without paying rent when the premises are burdened with substantial housing code violations making them unsafe and unsanitary." 5 In a ringing peroration, Wright declared that this result is "required ... by respect for the separation of powers and the rule of law." 6 It is this claim about the underlying ideal of the Rule of Law that prompts me to bring up the Robinson case here. 7 My guess is that a great number of legal observers thought the Rule of Law required the case to come out the other way, in favor of the landlord's previously unquestioned right to end a month-to-month tenancy on proper notice. Can the Rule of Law impose any limitation on the "arbitrary power" ofindividual decision makers? Is the Rule of Law a good way to go about trying to achieve government by law? I begin with an attempt to make clear exactly what is the Rule of Law, and the way in which its meaning is contested because of differences in emphasis of its instrumental and substantive aspects. Having noticed the fact that the main formulations of the Rule of Law do agree upon an assumption that law consists of rules, I move on to an inquiry into the nature of rules. Specifically, I consider a social practice conception of rules, commonly attributed to 463 F.2d 853 (D.C. Cir. 1972). The District of Columbia Court of Appeals had recently recognized the "illegal contract" defense in Brown v. Southall Realty Co., 237 A.2d 834, 837 (D.C. 1968). Successful assertion of this defense meant that a tenant's promise to pay rent was unenforceable. 5 Robinson, 463 F.2d at 865. Wright conceded that he would be "deeply troubled" if his ruling Jed to "many families living indefinitely in substandard housing without paying rent," id. at 869, but he apparently believed that giving the landlord the legal choice of either repairing the premises or being unable to charge rent would result in an increase in habitability at no increase in rent. 6 Robinson, 463 F.2d at 871. 7 Perceptive•law students ask me, "If Wright wanted to protect a tenant's right to safe and sanitary housing, why didn't he enjoin the landlord to repair the premises, rather than holding that the tenant could remain indefinitely without paying?" This is an excellent (and complex) question, although answering it would lead me afield here. See infra p. 818. 3 4

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

39

783

Wittgenstein, 8 which holds that rules can only be claimed to exist when there is community agreement in practice. To put this conception roughly, agreement in action does notfollow from there being a pre-existing rule; agreement in action is the only basis for claiming that there is a rule. I then consider the implications for the Rule of Law once this social practice conception of rules is accepted. First, I show how the traditional ideal of the Rule of Law is connected with a certain traditional conception of the nature of rules, which I shall call "formalist." Then I ask the more specific question which is the main focus of this essay: Does the social practice conception of rules render the Rule of Law impossible? I argue that the social practice conception of rules· does indeed render the traditional Rule of Law impossible, because of the link between the Rule of Law and the traditional conception of rules. I believe the Rule of Law can survive as a coherent ideal if it receives a modern interpretation, however, and I conclude by suggesting a possible avenue for such reinterpretation. 9

II.

THE RULE OF LAW AS A CONTESTED CONCEPT

Before reconsidering the Rule of Law in light of a social practice conception of rules, we must try to define the elements of the Rule of Law. The complex of ideas thought to comprise the Rule of Law is not completely canonical. We can, in fact, readily distinguish two main contested views: a primarily instrumental version and a more substantive version. The instrumental version holds that the Rule of Law is a prerequisite for any efficacious legal order. The substantive version holds that the Rule of Law embodies tenets of a particular political morality. 10 These versions are illustrated by two of the best modern 8 The view I discuss is often attributed to Wittgenstein, but no view can be attributed to him uncontroversially. As I discuss below, see infra pp. 797-81 0, my reading of Wittgenstein is broad and neo-pragmatic in character, but a narrower, more analytic reading is preferred by many. 9 The task of working out the required reinterpretation in detail, and exploring its connections with reinterpreted central aspects of political theory, awaits further work. The complete analysis would involve theories of the person, the state, and community, and theories of the proper roles of citizens, legislators, and judges in the well-developed polity. For another piece of the picture, see Michelman, Law's Republic, 97 YALE L.J. 1493 (1988) (presenting a modern interpretation of government by law through reinterpretation of the political theory of civic republicanism and exploration of the conundrum of government-of-laws versus government-by-the-people). 10 To avoid possible misunderstanding, I should say here that the view I refer to as substantive is substantive only in an attenuated sense. As elaborated below, see infra notes 22-33 and accompanying text, the view I call substantive merely asserts that having a system of rules of a certain type by itself serves substantive values, such as fairness and respect for persons. On its face, at least, the view I call substantive does not make the stronger claim that, in addition, in order to fulfill the Rule of Law ideal the rules must themselves be good rules. Ronald Dworkin's recent view of the Rule of Law is indeed "substantive" in this more robust sense, but I believe Dworkin's view departs from the main parameters in which the Rule of Law has previously been con-

40

The Rule of Law and the Separation of Powers 784

BOSTON UNIVERSITY LAW REVIEW

[VoL. 69: 781

attempts to address the Rule of Law, those of Lon Fuller 11 and John Rawls. 12 Although labels oversimplify, I think it will be appropriate to characterize Fuller's version as instrumentaP 3 and Rawls's version as substantive. 14 A.

The Instrumental Conception of the Rule of Law, or "How to do Things with Rules"

Lon Fuller claimed that the Rule of Law is part of the "internal" morality of law. 15 I interpret this to mean that the complex of ideas associated with the term "Rule of Law" is essential for the efficacy of any system of legal rules. This instrumental conception of the Rule of Law is presented in Fuller's engaging parable of King Rex, 16 who failed to make law by ignoring the eight elements of the Rule of Law. These eight elements compromise "the morality ceived. SeeR. DWORKIN, A MATTER OF PRINCIPLE 11-12 (1985) (the "rights conception" of the Rule of Law "requires, as part of the ideal of law, that the rules in the rule book capture and enforce moral rights"). 11 See L. FULLER, THE MORALITY OF LAW 33-94 (rev. ed. 1969). 12 See J. RAWLS, A THEORY OF JusTICE 235-43 (1971). 13 I do risk mischaracterizing Fuller's view by calling it instrumental, because Fuller himself apparently believed that his formulation embodied other values besides instrumental efficacy. See infra note 17. For another account of the instrumental conception of the Rule of Law, see J. RAZ, The Rule of Law and Its Virtue, in THE AUTHORITY OF LAW: ESSAYS ON LAW AND MORALITY 210-29 (1979), although Raz in the end turns toward the substantive conception in important respects. Cf Michelman, Justification (and Justifiability) of Law in a Contradictory World, in JusTIFICATION IN LAW, ETHICS, AND POLITICS 71, 72 (J. Pennock & J. Chapman eds. 1986) (XXVIII) (characterizing legalist justification as implying a law that is general, external to the case to be decided, rationally applicable to cases, and transcending individual deciders). 14 Other accounts that stress the substantive aspect of the Rule of Law are F. HAYEK, THE POLITICAL IDEAL OF THE RULE OF LAW (1955) (characterizing the liberty produced by the Rule of Law in terms of the laissez-faire market regime-private property and freedom of contract); J. LUCAS, THE PRINCIPLES OF POLITICS §§ 24-25 (1985); R. SARTORIUS, INDIVIDUAL CONDUCT AND SOCIAL NORMS (1975). Cf Moore, The Semantics of Judging, 54 S. CAL. L. REv. 151, 293 (1981) [hereinafter Moore, Semantics]; Moore, A Natural Law Theory of Interpretation, 58 S. CAL. L. REV. 277 (1985) [hereinafter Moore, Theory of Interpretation]. For a discussion of Moore's ideas, see infra notes 100-10 and accompanying text. 15 Fuller entitled his chapter on the Rule of Law "The Morality That Makes Law Possible." He did not explicitly elaborate on why he thought his analysis rested on anything properly called morality. He did seem to think that government by rules is necessary before the question of justice can be meaningfully posed: The internal morality of the law demands that there be rules, that they be made known, and that they be observed in practice by those charged with their administration. These demands may seem ethically neutral so far as the external aims of law are concerned. Yet, just as law is a precondition for good law, so acting by known rule is a precondition for any meaningful appraisal of the justice of law. L. FULLER, supra note 11, at 157. 16 For the parable of King Rex, see id. at 33-38.

The Rule of Law and the Separation of Powers

THE RULE OF LAW

1989]

41

785

that makes law possible": 1. Generality. Roughly, there must be rules, cognizable separately from (and broader than) specific cases, such that the rules can be applied to specific cases, or specific cases can be seen to fall under or lie within them. 2. Notice or publicity. Those who are expected to obey the rules must be able to find out what the rules are. 3. Prospectivity. The rules must exist prior in time to the actions being judged by them. 4. Clarity. The rules must be understandable by those who are expected to obey them. 5. Non-contradictoriness. Those who are expected to obey the rules must not simultaneously be commanded to do both A and not-A 6. Conformability. havior to the rules.

The addressees must be able to conform their be-

7. Stability. The rules must not change so fast that they cannot be learned and followed. 8. Congruence. The explicitly promulgated rules must correspond with the rules inferable from patterns of enforcement by functionaries (e.g., courts and police). I believe this list can be boiled down to two principles: first, there must be rules; second, those rules must be capable of being followed. The requirement that there must be rules encompasses what Fuller meant by generality. Generality implies that the rules are broader (more general) than specific cases or particulars, which can be brought within them, or seen to be comprehended, subsumed, or covered by them. All particulars of specific cases that fall under the rule are covered by the rule. If a rule commands, "No one under 21 is allowed in a saloon," then it applies necessarily to all people under 21 and to all saloons. Thus, here the familiar idea that like cases ought to be treated alike-(consistency) is seen to be an attribute of the formally general nature of rules. Consistency here means simply that every particular item that is within the extension of the operative words in the rule must be recognized as being within the extension of the words, and hence within the purview of the rule. We are not allowed to pick out only certain people under 21 or only certain saloons and say our rule applies only to them, simply because the rule's general nature precludes this. Similarly, the familiar idea that the discretion of judges and other functionaries should be limited in applying or interpreting the rules is, in Fuller's view, likewise seen in a formal guise. The generality of rules (that is, their very status as rules) is negated if functionaries are able to treat one item that comes within the extension of the operative words-hence falls under the rule-differently from another. If functionaries can treat different minors or saloons differently (without showing that they aren't "really" minors or "re-

The Rule of Law and the Separation of Powers

42

786

BOSTON UNIVERSITY LAW REVIEW

[VoL. 69: 781

ally" saloons), then we don't have the "rule" we thought we had. The requirement that the rules must be capable of being followed encompasses all the rest of Fuller's list. In order to be capable of being followed by the addressees, rules must have certain characteristics, which I think can also be boiled down to two. Let me colloquially call these two requirements "knowability" and "perform-ability." In order for those to whom the rules are addressed to know what they are commanded to do, the commands must be public, congruent, and non-contradictory, clear enough to understand, and they must not change too fast. In order for the addressees to do what is commanded of them, in response to the commands, the commands must be prospective (not retroactive), not contradictory or non-congruent, and not physically, mentally, or circumstantially impossible for the human beings addressed to follow. This is indeed an instrumental conception of the Rule of Law, 17 which could more colloquially be called "how to do things with rules." 18 All of Fuller's requirements are directed toward there being rule-like commands that can successfully induce desired behavior (whatever it is) in the addressees. 19 Substantive ideals, like fairness or democracy, and autonomy or dignity ofpersons, are not explicitly raised. Neither are democratic procedural traditions, like the separation of powers or the access to courts and jury trials. It seems that if a Nazi regime wanted to accomplish its heinous goals by means of rules, it would perforce make its commands conform to these requirements. 20 17 It is apparent, however, that Fuller himself objected to the type of instrumentalist reading I have given his work, for he chose to reply to the instrumentalist reading of his early reviewers in the revised edition of The Morality of Law. See L. FULLER, supra note 11, at 187-244 (chapter V is titled "A Reply to Critics"). Fuller tried to clarify his views after all of his reviewers argued that "the alleged internal morality of law is merely a matter of efficacy." ld. at 201. In his attempted clarification he spoke about the law as a "facility enabling men [sic] to live a satisfactory life in common," id. at 223 (emphasis in original), and about "'an interactional theory of law'" that characterizes law as a joint normative enterprise of lawmakers and law-followers. ld. at 237. Fuller's concluding plea for legal philosophy to turn away from conceptualism and positivism and toward "the social processes that constitute the reality of law," id. at 242, makes clear that he was groping for a pragmatic theory of law. 18 After so dubbing the instrumental conception, I discovered that there is a book with this title. See W. TwiNING & D. MIERS, How To Do THINGS WITH RULES: A PRIMER OF INTERPRETATION (1982). 19 What motivates the addressees to conform their behavior to these rule-like commands is a separate issue. Traditionally, of course, a primary motivator is penal sanctions. 20 See, e.g., J. RAZ, supra note 13, at 211 ("A non-democratic legal system, based on the denial of human rights, on extensive poverty, on racial segregation, sexual inequalities and religious persecution may, in principle, conform to the requirements of the rule of law better than any of the legal systems of the more enlightened Western democracies."). It should be noted that Fuller was unwilling to commit himself unequivocally to this implication of the instrumentalist conception. He often noted that the Nazi

43

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

787

This conception of the Rule of Law assumes that law consists of rules. In Ronald Dworkin's terms, this conception is premised upon a model of rules. 21 Implicit in the understanding of law as rules, and of the Rule of Law as a guide to the efficacy of those rules, are two sets of philosophical problems. The first focuses on the individual as a rule-follower. Fuller's instrumental conception of the Rule of Law presupposes that the law controls its subjects by providing incentives to structure behavior. Thus, although not included on Fuller's list, it is clear that there must be a ninth and tenth requirement: (9) addressees of rules must be rational choosers; (10) addressees must be suitably motivated, perhaps by penal sanctions, perhaps by opportunities for reward. The addressees must be such as can respond by following the rules, if the rules have the two characteristics of know-ability and perform-ability. The addressees must further be such as will respond, if they are motivated to do so by their desires to obtain rewards or avoid punishment. The problems of choice and motivation are not my concern here, however, and I set them aside. Instead, my focus will be on the second set of philosophical problems. These revolve around the very concept of rules itself. Does it make sense to think, as Fuller apparently did, that rules are general in the sense of being logically prior to action, and that they apply formally to particular actions that fall under them? These are the issues raised by the social practice conception of rules. But before addressing these issues, it will be helpful to examine the substantive conception of the Rule of Law. The Substantive Conception of the Rule of Law, or "How to Foster Liberty and Constrain Leviathan"

B.

John Rawls claims that the Rule of Law is an aspect of his overall scheme of "justice as fairness." The Rule of Law is formal justice-"the regular and impartial administration of public rules" 22-applied to the legal system. It promotes liberty, the prime value in justice as fairness. Indeed, perhaps Rawls may be read as making the stronger claim that the Rule of Law is required for liberty. Rawls proposes a rationalist model of law, from which he draws out one version of the traditional complex of ideas that comprise the Rule of Law. Here is his seminal definition of a legal system: "A legal system is a coercive order of public rules addressed to rational persons for the purpose of regulating their conduct and providing the framework for social cooperation." 23 This regime itself was not law-like, see, e.g., L. FULLER, supra note 11, at 40-41, and he may have thought that truly bad regimes would never be law-like. 21 R. DWORKIN, TAKING RIGHTS SERIOUSLY 14-80 (1977) [hereinafter R. DWORKIN, RIGHTS]. In his later work, Dworkin refers to the "rulebook model of community." See R. DwoRKIN, LAw's EMPIRE 209-15 (1986). Although he does not elaborate on the connections between this model of community-roughly, a positivist social contract position-and the notion that law consists of rules, they seem to be related. 22 J. RAWLS, supra note 12, at 235. 23

!d.

44

The Rule of Law and the Separation of Powers

BOSTON UNIVERSITY LAW REVIEW

788

[VoL. 69: 781

conception encapsulates many of the traditional precepts of the Rule of Law, in that, as Rawls explains, the precepts are derivable from it. 24 A lot is packed into these carefully chosen words. In the following paragraphs I shall unpack them slightly. "A legal system is a coercive order of public rules": In Rawls's conception (as in Fuller's), the law consists of rules. To regulate conduct, and thereby achieve the social cooperation necessary for justice, rules must have certain characteristics associated with the Rule of Law: l. "Ought implies can." The addressees must have the ability to conform, and the authorities must act in good faith. Impossibility of conformance, therefore, must be recognized as a defense.

2. Similar treatment of similar cases. This includes the requirement of consistency, and imposes limitations on judicial discretion. 3. "Nul/urn crimen sine lege" (no crime without law). Laws must be known and expressly promulgated, with clear meaning. There must be no retroactivity. Laws must be general and not bills of attainder. Severe offenses must be narrowly construed. 4. Natural justice. There must be structures for achieving truth and correct enforcement: trials, hearings, rules of evidence, due process. Judges must be impartial and independent. Trials must be fair, open, and not "prejudiced by public clamor." 25

Rawls advances two arguments for a connection between these precepts of the Rule of Law and the fundamental substantive value, liberty. First, he argues that absence of these characteristics will cause a chilling effect. To the extent that rules are vague, or that like cases are not treated alike, or that judicial process is irregular, "the boundaries of our liberties are uncertain." And when these boundaries are uncertain, "liberty is restricted by reasonable fear of its exercise." 26 Hence, if rational persons in the original position are choosing a legal structure to promote their paramount value-liberty-they will choose the Rule of Law because of the need for predictability, determinateness, and certainty of legal consequences. The second argument advanced by Rawls for a connection between the Rule of Law and liberty is characterized by Rawls as "Hobbes's thesis." 27 The scheme of social cooperation, which has as its purpose the enhancement of liberty over the conditions prevailing in the state of nature, requires the precepts of the Rule of Law. This is so because, under Hobbesian assumptions, the problem of social cooperation is the problem of the "n-person prisoner's dilemma." It benefits each person to restrain herself if all others cooperate, but it benefits her even more to act in self-interest while all others cooperate. Hence, there are strong incentives for each to act against the scheme of social 24 25

26 27

Id. at 236. Id. at 239. Id. Id. at 240.

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

45

789

cooperation, but only if she thinks she can get away with it while everyone else cannot. (If all can cheat and get away with it, the system reverts to the state of nature, which is everyone's least preferred choice.) A coercive sovereign is necessary to remove the incentives of self-interested individuals to break the rules that all should see are in their interest so long as all obey. "The existence of effective penal machinery serves as men's [sic] security to one another." 28 Hence, even an "ideal theory" of justice requires an account of penal sanctions as a stabilizing device. 29 Once the rational choosers in the original position have recognized the necessity of a coercive sovereign that can enforce penal sanctions; they must also recognize the necessity of restraining this Leviathan. In order to control Leviathan, the Rule of Law is needed: "the dangers to liberty are less where the law is impartially and regularly administered in accordance with the principle of legality." 30 Moreover, the principle of liberty leads to the principle of responsibility, which in turn leads again to the Rule of Law. The principle of responsibility requires the Rule of Law to include all the traditional defenses in the criminal law, not just the instrumental aspects of notice and non-retroactivity needed for instrumental efficacy of rules. "[U]nless citizens are able to know what the law is and are given a fair opportunity to take its directives into account, penal sanctions should not apply to them." 31 This description should make clear why I refer to Rawls's argument as a substantive conception of the Rule of Law (even though he too is "doing things" with rules). His conception of liberty-the negative liberty of traditional liberalism-is held to be required for achieving justice in society; the precepts of the Rule of Law are held to be strongly connected with that substantive value. Although the list of precepts-generality, consistency, notice, 28 /d. In later work, Rawls makes clear his view that if a liberal political order is to be stable over time, it must rest on moral consensus and not be merely a Hobbesian "modus vivendi." Rawls, The Idea of an Overlapping Consensus, 7 OxFORD J. LEGAL STUD. 1, 9-12 (1987). It is unclear, however, whether this development of Rawls's views would modify his treatment of the Rule of Law in A Theory of Justice. J. RAWLS, supra note 12, at 235-43. 29 Rawls characterizes his theory of justice as an "ideal theory" in the sense that everyone is presumed to act justly. In general, this means that all principles dealing with the errors, complexities, and injustices of the real world are not properly a part of the theory. This excludes civil disobedience and revolution, as well as the theories of corrective justice and punishment. J. RAWLS, supra note 12, at 8-9. I have expressed doubts elsewhere that "ideal theory" is a good way to reason about justice. See Radin, Risk-of-Error Rules and Non-Ideal Justification, in JusTIFICATION IN LAW, ETHICS, AND PoLITICS 33 (J. Pennock & J. Chapman eds. 1986) (Nomos XXVIII); Radin, Justice and the Market Domain, in MARKETS AND JUSTICE 165 (1. Chapman ed. 1989)(Nomos XXXI). 3o J. RAWLS, supra note 12, at 241. 31 !d. According to Rawls, the principle of responsibility is not based on the idea that punishment is primarily retributive or denunciatory, but that punishment exists for the sake of liberty itself. !d.

The Rule of Law and the Separation of Powers

46

790

BOSTON UNIVERSITY LAW REVIEW

[VoL. 69: 781

perform-ability, and congruence-is roughly the same as in the instrumentalist conception, the justification offered is quite different. The Rule of Law is grounded not on the bare claim of efficacy of behavioral control, but on the specific political vision of traditional liberalism. Liberty is the core value; overreaching by Leviathan is the danger on one hand, and disintegration of social cooperation because of the prisoner's dilemma is the danger on the other. 32 There is one traditional element of the Rule of Law in its substantive guise that is not explored by Rawls, yet it looms large in American jurisprudence. This element is the commitment to the separation of powers, and the connected ideas about judicial review and the constrained role of the judge. This commitment also figures in the instrumental conception from the instrumental point of view: rules as applied must not differ from the rules as made (Fuller's precept of "congruence"). Otherwise, the system gives conflicting commands and fails in its purpose of guiding behavior of the addressees. In a system committed to the institutions of courts and judicial review as ways to apply law, this instrumental point takes the form of insisting that judges be constrained so that they strictly "apply," and do not "make," the law. In the substantive conception of the Rule of Law the constrained judicial role is more central because it is held to be required for democracy, a core substantive value. The point is put in typical fashion by Rolf Sartorius:

If courts were to have the authority to make law, they would constitute a 32 For Rawls, the substantive connection between the Rule of Law and the core value of liberty is so strong that he seems to make the claim that evil rules are preferable to rulelessness: justice as fairness is served to some extent by having rules, even apart from the justice of the rules themselves. Rawls says that "other things being equal, one legal order is more justly administered than another if it more perfectly fulfills the precepts of the rule of law." J. RAWLS, supra note 12, at 236. This means that "[i]t will provide a more secure basis for liberty and a more effective means for organizing cooperative schemes." /d. Rawls does recognize, of course, the tension between this position and the need for justice in substantive law. He goes on to say that because the Rule of Law precepts "guarantee only the impartial and regular administration of rules, whatever these are, they are compatible with injustice." !d. Perhaps Rawls's apparent claim that there is some increment of justice in merely being rule-bound, wholly apart from the justice of the rules, is a momentary lapse in his commitment to "ideal theory." In a fully ideal theory one would think we should assume that rules are just. No need or opportunity to make claims about the justice of unjust, but rule-bound, regimes would arise. It seems more intuitively attractive to think that once we are in a non-ideal situation, we cannot say a priori or in the abstract whether rule-boundedness will make it worse or better. Rawls seems to turn towards this more pragmatic position when, a few pages later, he mentions the moral dilemmas of "partial compliance theory." In arguing that these dilemmas are to be resolved with the priority of liberty in mind, he notes that "[s]ometimes we may be forced to allow certain breaches of [the Rule of Law] precepts if we are to mitigate the loss of freedom from social evils that cannot be removed, and to aim for the least injustice that conditions allow." J. RAWLS, supra note 12, at 243.

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

47

791

legislative elite in a very pure form indeed. But the very arguments which favor majority rule as the form of legislative decision procedure as against some form of nondemocratic elitism imply that the law-making power should be the exclusive province of the majority. They militate against entrusting a judicial elite with the awesome power to make the law as well as apply it.JJ Those who are versed in the debate surrounding judicial review in American jurisprudence will recognize that this strong statement is simplistic and leaves untouched many famous puzzles. One is the counter-majoritarian difficulty-the problems surrounding the idea that the majority in creating a constitution can be understood to have constrained its own quotidian legislative actions by a broader vision, to be guarded by the judiciary. Another is the working-out of the idea that where there are "gaps" in the law, judicial law• making (subject to legislative override) may be perfectly consistent both with congruence and democracy. A third is the curious tendency to emphasize limiting the power of the judiciary rather than the equally nondemocratic organs of the executive branch. Nevertheless, Sartorius's statement well captures the force of the distinction between "making" and "applying" law, and the constrained conception of the judicial role, in the substantive conception of the Rule of Law. C.

A Contested Concept

At this point I hope it is clear that much of the complex of ideas called the Rule of Law is two-faced. My discussion so far has shown that the central precepts of the Rule of Law can be defended either instrumentally, as necessary to make a legal system work to structure behavior, or substantively, as necessary to fairness, human dignity, freedom, and democracy. The ambiguous precepts include generality (no bills of attainder); treat like cases alike (consis~ tency); notice; non-retroactivity; perform-ability (no impossible or conflicting commands); and congruence (limit discretion of enforcement functionaries). Hence, the Rule of Law itself is deeply ambiguous, a contested concept. Someone who stresses the instrumentalist aspect might affirm that a rule-bound dictatorship evidences the Rule of Law; someone who stresses the substantive aspect would not. Having elaborated the instrumentalist and substantive conceptions, we can also see the connection between the Rule of Law and liberal legalism. As far as I can tell, legalism means the decision of particular cases by means of general pre-existing rules. 34 Then liberal legalism is the variety of legalism exhibited by liberalism, or found in liberal ideology. If this is correct, the instruR. SARTORIUS, supra note 14, at 175-76. SeeR. UNGER, LAW IN MODERN SOCIETY 52-57, 176-81 (1976); cf J. SHKLAR, LEGALISM I (1964) (defining legalism as "the ethical attitude that holds moral conduct to be a matter of rule following, and moral relationships to consist of duties and rights determined by rules"). 33

34

48

The Rule of Law and the Separation of Powers

BOSTON UNIVERSITY LAW REVIEW

792

[VoL. 69: 781

mental conception of the Rule of Law is simply legalism, and the substantive conception is the (or a) liberal form of legalism. Aside from the surface correlation in the lists of precepts, it is apparent that the two conceptions do have much philosophical underpinning in common, including five important assumptions: ( l) law consists of rules; (2) rules are prior to particular cases, more general than particular cases, and applied to particular cases; (3) law is instrumental (the rules are applied to achieve ends); (4) there is a radical separation between government and citizens (there are rule-givers and appliers, versus rule-takers and compliers); (5) the person is a rational chooser ordering her affairs instrumentally. 35 Nevertheless, the two conceptions are different. The instrumental conception is a model of government by rules to achieve the government's ends, whatever they may be. The substantive conception is a model of government by rules to achieve the goals of the social contract: liberty and justice. The instrumental conception purports to be more general and ahistorical; the substantive is more clearly bound up with our particular modern ideological heritage. 36 III.

TRADITIONAL fORMALISM AND THE RULE OF LAW

Now I can zero in on the connection between the Rule of Law and traditional formalism-roughly the view that a unique answer in a particular case can be conclusively derived from application of a general rule. I begin by recounting several traditional senses of formalism, and considering their connection with traditional conceptions of the Rule of Law. Then I review the social practice ("Wittgensteinian") critique of the traditional conception of rules as formally or logically prior to particular~. This will bring me to the crucial issue of whether a switch to a Wittgensteinian view of rules makes any difference to our conception of the Rule of Law. 35 As I have said, in this essay I am concentrating on the first two assumptions. As the modern reconsideration of the Rule of Law unfolds, however, they will all have to be considered together. Then the issue will arise whether these five assumptions imply each other or bear strong affinities to each other. I have not begun to argue this here; but I am sure the reader can infer that my intuition is that they are strongly related, though probably not by entailment. If that is right, then if we drop (2) the traditional conception of rules, and (l) the notion that law consists of [such] rules, we would also drop (3) the means/ends or process/substance distinction, (4) the public/private distinction, or what I like to call the "we-they" theory of government, and (5) the individualist/instrumentalist view of the person. 36 If one is convinced that only those who embrace the particular substantive values of traditional liberalism would embrace the idea that law is a system of rules-as well as the other assumptions associated with the model of rules mentioned earlier-then the two conceptions ultimately coalesce. In such a view, the historical and ideological context that gives rise to the model of rules is what unites the two conceptions of the Rule of Law.

49

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

793

Senses of Traditional Formalism

A.

1.

Formal connection between rule and particulars

Traditionally, legal "formalism" is the position that a unique answer in a particular case can be "deduced" from a rule, or that application of a rule to a particular is "analytical." (The scare quotes indicate that philosophical arguments about language have made the traditional understanding of these concepts just as problematic as "rule" itself. I shall return to this issue in the next section.) The connection between the rule and its application is, in other words, formal. Pejoratively, this is mechanical jurisprudence, or the computer model of judging. In this model, judges do not judge; they are only black boxes, who function to juxtapose the rule and the particular so the formal connection can be declared.J7 2.

Formal connection between foundations and rules

There is another way that the term "formalism" traditionally has been used. This "formalism" is the view that there exists a mind-independent reality consisting of certain first principles either of fact or value. These first principles form a logical, analytical, "foundation" for the law. A natural law theory would be formalist in this sense if it claimed that legal and moral rules are real and are there for us to discover, for example, or if it claimed that legal rules are deducible from a foundational set of real values. In this foundationalist kind of formalism, the crucial formal connection is between the structure of the universe and the rules of law. It is thus different from the mechanical jurisprudence kind of formalism, where the crucial formal link is between rules and the application of those rules to particulars. 3.

Formal connection between words and things

In a famous metaphor, H.L.A. Hart suggested that legal rules have a 37 A commitment to traditional formalism in the sense of mechanical jurisprudence is compatible either with legal positivism or natural law, as a general theory of law. It doesn't matter where the rules come from-from the sovereign, or from revelation or natural reason-it only matters that their application to particulars is a matter of deduction or formal analytical connection. It seems, however, that this kind of formalism is in practice especially allied with legal positivism. Positivists assert that Ia w and ethics inhabit radically separate domains. The underlying motivation for maintaining such a thesis may be an ethical skepticism: the conviction that values are subjective or arbitrary. The traditional Rule of Law ideal carries an implied aspiration toward objectivity in law; it is the alternative we must pursue to avoid the evils of arbitrary power. For the skeptical positivist who desires to be able to affirm the Rule of Law by treating law as objective, then, law must be kept on the fact side of the factjvalue dichotomy, which also keeps it on the objective side of the subjective/objective dichotomy. If the rules are formally derived from facts and the particulars are facts, as they are in positivist views, formalism is the obvious way to connect rules and facts without an admixture from a subjective domain.

The Rule of Law and the Separation of Powers

50

794

BOSTON UNIVERSITY LAW REVIEW

[VoL. 69: 781

"core" of certainty and a "penumbra" of uncertainty. 38 Hart declared that in the penumbra, judges must "legislate." A common understanding of the core/ penumbra distinction-although probably not Hart's own understanding of it-is that in the core, formal deductive application of rules to particulars is possible. 39 If we do assume that formal application to particulars is possible within the core of a rule, then we can characterize a third sense of traditional formalism: formalism in semantics. 40 If there is a set of particulars, comprising the core meaning, to which a rule is applicable analytically or through deduction, it must be true that the words in the rule have an analytic connection with at least a subset of the particulars falling within these words' extensions. For example, if a rule contains the word "vehicle," then for the rule ever to be formally applicable there must be some subset of cases involving objects that are seen to be vehicles through deduction or some sort of analytic connection alone. 41 The notion of an analytic connection between a word and its extension raises thorny problems in the theory of reference. 42 Yet some logical positivists may have thought word-meaning could be analytic in this way. Formalism in semantics may assume that the extension of a word can be logically determined by connection with a list of necessary and sufficient criteria,· and thus that there is a formal connection between a word and the things to which it applies once the criteria are known. The necessary and sufficient criteria could be conventional artifacts of language. For a thorough reductionist, however, the criteria would be deducible from foundational sense-data, and formalism in se38 Hart, Positivism and the Separation of Law and Morals, 71 HARV. L. REv. 593, 607 (1958). 39 Hart said that "[i]f a penumbra of uncertainty must surround all legal rules, then their application to specific cases in the penumbral area cannot be a matter of logical deduction," id., leaving it open to question whether logical deduction is what we do in core cases. Hart's general view of rules and language should more likely be interpreted, however, as incorporating the Wittgensteinian critique. See, e.g., Hart, Definition and Theory in Jurisprudence, 70 LAW Q. REv. 37 (1954). See infra notes 76-77 and accompanying text. 40 See Moore, Semantics, supra note 14, at 152, 157. 41 The example of the word "vehicle" appearing in a statute was debated by Hart and Fuller in their Harvard Law Review exchange in 1958. See Hart, supra note 38, at 607; Fuller, Positivism and Fidelity to Law, 71 HARV. L. REV. 630, 662-63 (1958). It is drawn from the case of McBoyle v. United States, 283 U.S. 25, 26-27 (1931) (reversing defendant's conviction for theft of an airplane under the National Motor Vehicle Theft Act because the word "vehicle" as used in the statute could be read to include only those contrivances "capable of being used 'as a means of transportation on land'"). 42 Of course; we can retain our common understanding that words have a core of certain meaning (some objects are unmistakably vehicles) without a formalist explanation of the source of this certainty. One way to do this is through Wittgenstein's essentially pragmatic view that words have meaning by virtue of being embedded in a form of life.

51

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

795

mantics would lead back to formalism in metaphysics. 4.

Formal realizability and the formalist conception of rules

In the traditional conception of the nature of rules, a rule is self-applying to the set of particulars said to fall under it; its application is thought to be analytic. It is often said that rules are logically prior to the particular cases that fall under them. Another way of putting this is to think that somehow the applications to particulars are already present in the rule itself. For example, if a rule says that no one under 21 is allowed in a saloon, then the application of the rule to Sally, who is in Joe's Saloon on her sixteenth birthday, is analytic or formal. The result that Sally must be excluded follows immediately from the rule itself and Sally) specific circumstances, and is unaffected by any other previous applications of the rule or other circumstances. Under the traditional conception of rules, the property of analytic self-application, thought to inhere in rules, is sometimes called "formal realizability." 43 When a directive or command is formally realizable, its application is deductive or analytic. In other words, it deserves to be called a "rule." To the extent that legal directives are formally realizable they are to be implemented by mechanical jurisprudence. In my view, rules would be formally realizable to the extent that the words in them are formally realizable. Thus, traditional formalism in the conception of rules leads back to traditional formalism in semantics. Formal realizability is an asserted characteristic primarily of words, and only secondarily of rules.

B.

Traditional Formalism and the Rule of Law

The traditional understanding of formal realizability correlates with the demand for clarity in both the instrumentalist and substantive versions of the Rule of Law. The traditional view has been that in order for people to know what the law is so that they can conform their behavior to it, it is necessary for the law to consist, to the fullest extent possible, of formally realizable rules. Directives that are not formally realizable are sometimes called ."standards." 44 43 The notion of formal realizability is rooted in German jurisprudential thought. Duncan Kennedy finds the term in Rudolph von !bering's Spirit of Roman Law. Kennedy, Form and Substance in Private Law Adjudication, 89 HARV. L. REV. 1685, 1687-89 (1976), citing R. VON !HERING, DER GEIST DES ROMISCHEN RECHT 50-55, 84 (1883). Though Kant did not use the term "formal realizability," he apparently thought that law could not exist without it. See, e.g., I. KANT, Perpetual Peace, in KANT ON HISTORY 85, 91-92 n.4 (L. Beck ed. 1963) ("[T]he possibility of a formula similar to those of mathematics is the only legitimate criterion of a consistent legislation."). 44 This use of the term "standard" was apparently introduced by Duncan Kennedy, Kennedy, supra note 43, at 1688, and has been adopted by a number of critical writers. See, e.g., Schlag, Rules and Standards, 33 UCLA L. REv. 379, 382 n.l6 (1985). Unfortunately it causes confusion because other writers do not use the term "standard" in

The Rule of Law and the Separation of Powers

52

796

BOSTON UNIVERSITY LAW REVIEW

[VoL. 69: 781

If in the previous example of the rule about presence in saloons we substituted the words "no one who has not reached maturity and responsibility" for "no one under 21," it would be deemed a "standard" and not a "rule." The prevalence of "standards" in the law poses problems for the notion that clarity of commands depends upon formal realizability. The traditional understanding of formal realizability as the crucial property of rules also correlates with the traditional conviction that the Rule of Law demands that judges "apply" rather than "make" the law. If rules do not tie judges' hands with their logical or analytic application, the traditional view is that judges will have personal discretion in how to apply the law. This will undermine congruence and confuse those who are supposed to follow the rules. It will diminish the efficacy of the system (in the instrumental view) and the liberty of the rule-followers (in the substantive view). It will also confer on judges a realm of "arbitrary power" and undermine democracy. The Rule of Law as it comes down to us in the liberal tradition is committed to the model of rules, and this means, under the traditional conception of rules, that it is committed to traditional formalism. One reason for this, as Rawls saw, is that this model of law readily connects with aspects of the Hobbesian view of the nature and purposes of social cooperation. Because everyone is always motivated to "defect," government by majority rule reduces to accommodation among shifting coalitions of self-interested individuals. The government is a Leviathan to be restrained. Yet, if judges can ignore the legislative "bargains" of these shifting coalitions, then judges are even more in need of restraint than are legislatures. Hence, the role of the judges is tightly constrained. Judges should be mere tools of implementation of the underlying "contract" that guarantees social cooperation and prevents "defection" and degeneration into the state of nature. The underlying "contract," that justifies the existence of government at all, can be thought of as a "constitution." Most of the time judges should function as tools to implement what the legislature enacts, but when legislatures overstep the bounds of justified government, judges must function as tools of the underlying constitutional "contract" and intervene to strike down the offending legislation. In view of the very great danger posed by judges, the Hobbesian theorist needs an ironclad system of rules to restrain them. Rules, in the traditional formalist sense, must determine the decisions of judges to the fullest extent possible, on pain, otherwise, of losing the underlying justification for the whole legal system. Although Rawls is not a thoroughgoing Hobbesian theorist by any means, his conception of a legal system coheres with this motivation for commitment to the model of rules. 45 It appears that any form of traditional liberalism, with this way. See, e.g., R. DWORKIN, RIGHTS, supra note 21, at 22 (using "standard" as a general term referring to rules, principles, policies, and others). See infra note 83 and accompanying text. 45 Recall Rawls's quintessential definition: "A legal system is a coercive order of public rules addressed to rational persons for the purpose of regulating their conduct

53

The Rule of Law and the Separation of Powers

THE RULE OF LAW

1989]

797

its commitment to means/ends rationality and to methodological individualism, will be drawn to the traditional conception of rules as the means to structure incentives so as to achieve maximally the individual behavior desired by the state. 46 IV.

THE WITTGENSTEINIAN CONCEPTION OF RULES AND ITS CONSEQUENCES (IF ANY)

A.

The Wittgensteinian Social Conception of Rules

A Wittgensteinian view of words, and of the rules containing them, suggests that there is no such thing as traditional formal realizability, or in other words, that the traditional formalist conception of rules is wrong. The Wittgensteinian view of rules may be characterized as both a social and a practice conception. It is a social conception because in this view rules depend essentially on social context, and it is a practice conception because rules also depend essentially on reiterated human activity. If we assume that the traditional formalist conception of rules is what we mean by the word "rule," the Wittgensteinian perspective may also be characterized as a species of rule-skepticism. A number of assertions may be encompassed in skepticism about rules in the traditional formalist conception of them. These assertions are skeptical answers to a set of problems about rules and rule-governed action. One problem is in determining when it can be said that someone is "applying" a rule. How do we know the rule is being "applied"? The skeptic says, "We can't really tell." We cannot identify any formal or logical criterion by which we can determine whether or not someone is following a rule. The problem is important for a distinction that is crucial to the Rule of Law-the distinction between the existence of the rules and their application by rule-appliers, such as judges. A second problem is how we should understand the notion that rules are "binding." In what consists the "binding-ness" of rules? The skeptic says, "This bond cannot be shown." We cannot demonstrate any formal or logical nexus between the rule and the rule-follower's behavior in response to it. The problem is important for the coherence of the conception of a legal system as a and providing the framework for social cooperation." J. RAWLS, supra note 12, at 235. 46 I am not attempting to treat fully in this essay the broader connections between the Rule of Law and theories of politics and the person. It seems apparent, however, that the traditional conception of rules that apply self-evidently of their own logical force to one person in one particular set of circumstances can be connected to methodological individualism. Additionally, it appears that the reduction of self-government to the following of rules laid down by legislative representatives follows from the traditional liberal tendency to think of government and citizens as radically separate; and it appears that the model of the rational person as chooser of action based upon external incentives for bettering her own welfare can-as many devotees of law and economics have argued-be connected to a need for structuring her activities by clear rules laid down by an authoritative entity.

The Rule of Law and the Separation of Powers

54

798

BOSTON UNIVERSITY LAW REVIEW

[VoL. 69: 781

"system of coercive rules" binding on citizens. A third problem is how we should understand "ruleness" itself. I have suggested that traditional formal realizability is a conception of rules that asserts that rules are analytically applicable to particulars insofar as they contain words with precise, determinate extensions. The skeptic says, "Traditional formal realizability cannot be shown to exist." We cannot demonstrate that application of a rule is analytic or deductive. This problem is important for the idea that a rule is logically prior to particulars falling under it and can be seen to apply of its own force to any individual cases falling under it. I think Wittgenstein in Philosophical Investigations 47 is fairly read as rejecting the traditional conception of rules in favor of a social practice conception in which agreement in responsive action is the primary mark of the existence of a rule. 48 I also think Wittgenstein is fairly read as a rule-skeptic in the three senses I just suggested. 49 There is no way to tell deductively or analytically when a rule is being followed; there is no special state that describes the binding•ness of rules; and traditional formal realizability is not the right way to conceive of the nature of rules. The relevant passages in Philosophical Investigations are three sequences of remarks: on continuing an arithmetic series according to a certain formation rule; on reading; and on obeying a rule. 5° The questions Wittgenstein addresses are both internal and external. That is, with respect to continuing a series, he is interested both in what happens when someone can say, "Now I can go on" and what happens when others can say, "Now she's got it." Likewise with 47 L. WITTGENSTEIN, PHILOSOPHICAL INVESTIGATIONS (G.E.M. Anscombe trans. rev. ed. 1968) (hereinafter L. WITTGENSTEIN, PHILOSOPHICAL INVESTIGATIONS]. See also L. WITTGENSTEIN, REMARKS ON THE FOUNDATIONS OF MATHEMATICS 391-437 (G.E.M. Anscombe trans. rev. ed. 1978) (another version of Wittgenstein's investigation on rules). 48 The proper reading of Wittgenstein's later work is hotly disputed in the secondary literature. Among those whose readings seem most congenial to me are Peacocke, Reply: Rule-Following: The Nature of Wittgenstein's Arguments, in WITTGENSTEIN: To FOLLOW A RULE 72 (C. Leich & S. Holtzman eds. 1981); P. STRAWSON, SKEPTICISM AND NATURALISM: SOME VARIETIES 75-95 (1983); S. KRIPKE, WITTGENSTEIN ON RuLES AND PRIVATE LANGUAGE (1982). For opposing readings, see G. BAKER & P. HACKER, SKEPTICISM, RULES & LANGUAGE (1984); C. McGINN, WITTGENSTEIN ON MEANING (1984). 49 In characterizing Wittgenstein as a rule-skeptic, we must remember that he does not say there can be no such thing as a rule, only that the traditional formalist conception is the wrong way to think about how rules work. Wittgenstein's skepticism about rules has been compared with Hume's skepticism about induction. SeeP. STRAWSON, supra note 48, at 14-21; S. KRIPKE, supra note 48, at 62-68. In a reading that seems to me reasonable and interesting, Kripke understands Wittgenstein's Philosophical Investigations as propounding a skeptical paradox about rules. See also Yablon, Law and Metaphysics (Book Review), 96 YALE L.J. 613 (1987) (reviewing Kripke for a legal audience). 50 L. WITTGENSTEIN, PHILOSOPHICAL INVESTIGATIONS, supra note 47, §§ 143-242.

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

55

799

reading: What licenses "Now I am reading" or "Now she's reading"? The discussion of rules in the third sequence is a generalization of these two examples. The internal questions are: What is the inner experience of obeying a rule? In what consists the determined-ness or binding-ness on me? The external questions are: How can other people tell when someone is obeying a rule? In what consists the knowledge or recognition of rule-application? To all of these questions Wittgenstein gives skeptical answers. In the introspective portions of his analysis, Wittgenstein considers suggestions that the binding-ness might be a mental state or process or disposition, 51 a special experience or feeling or inner sensation, 52 a kind of derivation or being guided, 53 a causal or "super-strong" connection, 54 or a way of grasping in a flash or in a direct sense. 55 Wittgenstein rejects these attempts to capture the essence of the rule-follower's subjective experience of binding-ness. His conclusion is that the rule-follower acts without reasons, 56 without choice, 57 blindly, as a matter of course. 58 Following the rule is "simply what I do." 59 In external portions of his analysis, Wittgenstein replies to his introspective suggestions, from the point of view of how we observe someone else following a rule. His conclusions are similarly skeptical. All we can say when we think someone is following a rule is that it seems natural. 6 For someone to apply a rule correctly is simply to do it "as we do it." 61 The decisive factor in knowing whether someone has followed a rule "is not his [the rule-follower's] inner experience but the circumstances under which he had it." 62 The result of this skeptical deconstruction of the formalist notion of rules is that rule-following must be understood to be an essentially social phenomenon. Rule-following can only be understood to occur where there is reiterated human action both in responding to directives and in observing others respond. Only the fact of our seemingly "natural" agreement on what are instances of obeying rules permits us to say there are rules. The rules do not cause the

°

/d. §§ 146, 152-154. /d. §§ 155, 157, 159C160. 53 /d.§§ 162-163, 172-173, 175. 54 !d. §§ 169, 176-177, 197. 55 /d. §§ 191-192. 56 /d. § 211. That there is no identifiable reason for one's experience of "bindingness," no essence of "binding-ness," should of course not be misunderstood to mean that one cannot have reasons for following a rule. 57 /d. § 219. Similarly, that there seems to be no choice about whether or not one has the experience of "binding-ness" should of course not be misunderstood to mean that one cannot choose whether to follow a rule or to disobey it. 58 /d. § 238. 59 /d.§ 217. 60 /d. § 185. Cj. S. KRIPKE, supra note 49, at 92-96; P. STRAWSON, supra note 48, at 75-80. 61 L. WITTGENSTEIN, PHILOSOPHICAL INVESTIGATIONS, supra note 47, § 145. 62 /d.§ 155 (emphasis in original). Cf. id. § 180 (calling words a "signal" of mental state, not a description of it). 51

52

56

The Rule of Law and the Separation of Powers

800

BOSTON UNIVERSITY LAW REVIEW

[VoL. 69: 781

agreement; rather, the agreement causes us to say there are rules. 63 As Wittgenstein says, "The word 'agreement' and the word 'rule' are related to one another, they are cousins." 64 As Wittgenstein also says, this is "not agreement in opinions but in form of life." 65 Rules are not logically prior to uniformity of action in response to them; rather, uniformity of action is prior to the existence of rules. This view of rules is irreducibly social and rejects the traditional formal, logical, analytic, or deductive understandings of rule-application. 66 The form of life called rule-following is a practice, and the practice is a form of agreement among members of a social group. It is not possible to obey a rule "privately,"67 nor is rule-following something that can be done by "only one man, only once. " 68 In the Wittgensteinian view, an action is determined by a rule when the action seems a "matter of course," 69 and "disputes don't break out." 70 How do I know when I am following a rule? When I feel the answer is automatic or self-evident. But my subjective feeling of "binding-ness" could be mistaken.7 1 How do I know when I'm really "applying'' a rule? When everybody else also would have felt compelled to do as I did. How do we know someone else is following a rule? When we all would have felt bound to do as she did. Although the Wittgensteinian view thus certainly admits that there can be action determined by a rule, it is not the kind of determined-ness required by the traditional concept of formal realizability. Formal realizability (the traditional formalist conception of the nature of rules) asserts that only the words of the rule and the action of one person with regard to one set of particulars are relevant in questions of rule-following. This Wittgenstein denies. If formal realizability would allow for the possibility of obeying a rule privately or only once, then formal realizability is merely what Wittgenstein would call a "philS. KRIPKE, supra note 49, at 96. L. WITTGENSTEIN, PHILOSOPHICAL INVESTIGATIONS, supra note 47, § 224 (emphasis in original). 65 !d. § 241. Cf. S. KRIPKE, supra note 49, at 96-98. 66 I believe the appropriate Wittgensteinian conclusion is that, along with the concept of rules, the concepts of formality, logic, analyticity, and deduction cannot mean what we have thought them to mean; even these categories are irreducibly social. See infra note 77 and accompanying text. 67 L. WITTGENSTEIN, PHILOSOPHICAL INVESTIGATIONS, supra note 47, § 202. Wittgenstein also notes that "to think one is obeying the rule is not to obey the rule." 63

64

!d.

68 69 70

!d. § 199 (emphasis in original). !d. § 238. !d. § 240: "Disputes do not break out (among mathematicians, say) over the

question whether a rule has been obeyed or not. People don't come to blows over it, for example. That is part of the framework on which the working of our language is based (for example, in giving descriptions)." !d. 71 See supra note 67 and accompanying text (rule-following does not depend solely on one's private experience of it).

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

57

801

osophical superlative. " 72 There are still rules, of course; rules are an important practice of ours. But traditional formal realizability is not the right way to understand "rule-ness." Consequences (if any) for the Model of Rules and the Rule of Law

B. l.

Does Our Conception of Rules Matter?

One might say, "So what?" Under the Wittgensteinian view there are still obvious cases of rules and rule-following; whether we hold the traditional view of rules or the Wittgensteinian view makes no difference for our conceptions of the model of rules and the Rule of Law.7 3 The "So what?" response is particularly appealing to· those who read Wittgenstein more narrowly than I do. I take a view that aligns him with modern pragmatism. His insistence that meaning cannot be separated from "use"-from reiterated human activity, practices, embedded in and helping to constitute a "form of life"-seems, when given the central weight I believe it deserves, to make clear his intellec- · tual affinity with pragmatism. 74 Yet those who read Wittgenstein more narrowly understand him merely to have been making an analytic point about language, and rules couched in language, to the effect that no rule can determine the scope of its own application. They deny that Wittgenstein means that there is something essentially social or "communitarian" about rules. 75 If we thus interpret Wittgenstein narrowly we might think that everything remains the same. Law still consists (entirely or in important part) of a set of rules authoritatively laid down and implemented by judges who apply, but do not make, the rules. The only difference would be conceptual: we just change the way philosophers, inside their own heads, understand rules. In other words, if we think that all we change is our internal philosophical view of rules when we become Wittgensteinian, then the "So what?" response seems initially plausible. The "Sq what?" response also seems initially plausible under a broader view of the role of practices. If one is inclined to treat practices as so "brute" and L. WITTGENSTEIN, PHILOSOPHICAL INVESTIGATIONS, supra note 47l § 192. For a sophisticated view of rules that has fully incorporated the Wittgensteinian perspective, and yet sees no important practical effects for our conceptions of law and the Rule of Law, see F. SCHAUER, PLAYING BY THE RULES: A PHILOSOPHICAL ANALYSIS OF RULE-BASED DECISION-MAKING (unpublished manuscript 1989). 74 This view of Wittgenstein is explicit in the work of Richard Rorty, see R. RoRTY, PHILOSOPHY AND THE MIRROR OF NATURE 367-68 (1979) [hereinafter R. RORTY, MIRROR OF NATURE], and is implicit in the views of many other modern critics of foundationalism and the traditional dichotomies of fact and value, theory and practice, and the like. What may be distinctive in the modern wave of pragmatism is the degree of emphasis placed on the social construction of reality. For a lucid description of the basic tenets of pragmatism, as well as some elaboration of a modern pragmatic view of law, see Grey, Holmes and Legal Pragmatism, 41 STAN. L REv. 787 (1989). 75 See G. BAKER & P. HACKER, supra note 48, at viii (arguing against Kripke); C. McGINN, supra note 48, at 9 (same). 72

73

The Rule of Law and the Separation of Powers

58

BOSTON UNIVERSITY LAW REVIEW

802

[VoL. 69: 781

"given" that they cannot be interpenetrated by what we say or think about them, then the practice we call philosophy (or theory, or conceptual schemes) is radically separate from every other practice. When we reconceive meaning in terms of practice, all our practices (except philosophy) remain unchanged. So again, as with the narrow analytic view, any Wittgensteinian change is internal to philosophy.7 6 Wittgenstein's statement that we know there are rules when "disputes don't break out" provides a way to reinterpret the idea that rules have a "core" of certain application, as well as a "penumbra" of doubt, without resort to traditional formal realizability. In this reinterpretation, the core signifies our observance of rule-responsive behavior that feels compelled or blind or "a matter of course"; the penumbra signifies our observance of disputes breaking out or the absence of the feeling of compulsion. The core is roughly coextensive with our ability to say, "We all think that's self-evident." We can all recognize action under rule when we eject sixteen-year-old Sally from Joe's saloon in response to the directive, "No one under 21 is allowed in a saloon." Under the Wittgensteinian view someone who doesn't see this result as self-evident will be considered insane or from Mars; and this is just as strong a form of social control as being thought irrational or deficient in deductive powers. Even if we accept a slightly broader view of Wittgenstein, the philosophical change might still seem not to imply any significant differences in practice. A Wittgensteinian reinterpretation is just as possible, and just as necessary, for all the concepts allied with the traditional conception of "rule" as it is for the conception of "rule" itself. If the traditional conception of a "rule" amounted to a mistaken "philosophical superlative," then so too have our conceptions of "formal connection," "analyticity," "deduction," and so on. But it is possible to think of "formalism" not in the traditional sense of a logical connection

a

76 The most vociferous exponent of this view, at least in law reviews, is Stanley Fish. See, e.g., Fish, Dennis Martinez and the Uses of Theory, 96 YALE L.J. 1773, 1775 (1987) (discussing the dichotomy between theory and practice and arguing that the two cannot interact). Fish is clearly wrong that judging is a "brute" practice of this kind; judging as a practice also includes ideals about judging, critiques of judging, understandings about the role of judging in the polity, etc. See infra notes 89-90 and accompanying text (interaction between rhetoric and the world). I even think that Fish is wrong about the practice of baseball. Part of the reason Ore! Hersheiser is a better pitcher than Dennis Martinez is his attention to "theory"; his detailed notes about hitters' propensities and past performances enable him to make better decisions about how to pitch to them, and these decisions are a part of the practice of baseball. The primary complaint against these Fish-like views is that if practices are so "brute," so always-already given, we cannot make space for critique and social progress. On this reading, Wittgenstein seems like an apologist for the status quo. See, e.g., R. UNGER, PASSION: AN ESSAY ON PERSONALITY 11-12 (1984). But cf H. PITKIN, WITTGENSTEIN AND JUSTICE 46 (1972). Maybe, in !he end, whether we read Wittgenstein in the broad, holistic way I prefer, in which the theory /practice dichotomy is dissolved and the practices of conceptualization and critique can interpenetrate many (all?) others, is perhaps a matter of personal temperament.

59

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

803

existing independent of a social setting and history of application, but instead as decisionmaking by uncontroversial reasoning.7 7 It is possible to think of "analyticity," "deduction," and "logical connection," all as similarly embedded in an ongoing practice in which completely uncontroversial results are reached by what we "naturally" think of as rational procedures under the circumstances. Does a Wittgensteinian reinterpretation of all of these concepts prevalent in the traditional Rule of Law change anything? At least, doesn't it leave untouched everything practical, everything that we can observe or that has significance for us and our lives? When the phlogiston theory was rejected, it did not change the fact that fire will burn us. Even if we admit that the formally realizable rules required by the Rule of Law rest on a social practice, a brute unquestionable agreement in a form of life, don't all our reasons for requiring formally realizable rules (clarity, notice, congruence, stability, prospectivity, restraint of judges) remain the same? 78 Perhaps Wittgenstein himself would say, if we could ask him, that this is the correct way to understand his reinterpretations. After all, he was not one to think that the form of life called philosophy had much influence on our larger forms of life. Nevertheless, if this would be his position, I think it too modest.79 We should not hastily conclude that commitment to the Wittgensteinian understanding of rules, at least to the broader, nee-pragmatist view of Wittgenstein I prefer, would be irrelevant for our views about law and the Rule of Law. But neither should we hastily conclude that because there is no such thing as traditional formal realizability, everything is indeterminate or up for grabs. We shall have to proceed in a more cautious and more piecemeal manner. As Wittgenstein might say, we must "look and see." We can begin by noticing that the Wittgensteinian limitation of rules to situations in which "disputes don't break out" does not help lawyers much. Disputes do break out among lawyers and judges, and litigants do come to blows,. at least metaphorically. If our only way to find out what result is compelled by a rule is to be part of a community that recognizes action as rulefollowing, then there is no way to bring any truly disputed cases-in which neither side changes its position even after being confronted by what the other side believes to be completely uncontroversial reasoning-under pre-existing rules, as the Rule of Law requires. Perhaps a Wittgensteinian would not want to think of law as consisting only of rules. Of course, truly disputed cases can indeed come under reasonable judgment, or whatever faculty our practice recognizes as making appropriate judgments in disputed cases. But if by "our practice," we mean "our legal 77 See Grey, Langdell's Orthodoxy, 45 U. PITT. L. REv. 1, 8 (1983) (characterizing formalism in this way). 78 For a thoughtful argument to this effect, see F. SCHAUER, supra note 73, §§ 8.5, 8.6. 79 In other words, if Wittgenstein professed this Fish-like view, see supra note 76, he would be untrue to himself.

60

The Rule of Law and the Separation of Powers

804

BOSTON UNIVERSITY LAW REVIEW

[VoL. 69: 781

practice," then our practice is at odds with the Rule of Law ideal that is supposed to describe and govern it. A stubborn commitment to the model of rules as essential to the Rule of Law explains why legal theorists go to such lengths to try to show that even "hard cases" have pre-existing determinate answers. 80 Hart's declaration that judges "legislate" in hard cases is unorthodox in the liberal tradition because it tends to undermine the commitment to the model of rules. It is not surprising that Hart urges us to think such hard cases are few and far between. To the extent that law, when we "look and see," is the terrain of "disputed" cases, the pragmatic Wittgensteinian view seems to tell us we had better find another model (other than rules) for what law "is." 2. . Compromise Positions on the Model of Rules Perhaps, then, a Wittgensteinian reinterpretation requires us to divide law into domains, consisting of rules and other things, and to reconceive the Rule of Law in light of this complexity. A number of prevalent views (although they may have rested originally on the traditional conception of rules) present themselves as candidates for appropriate reinterpretation. The challenge, of course, is to characterize those "other things"-the domain of law that does not consist of rules-as "lawlike" in a sense that is politically acceptable to us. The Wittgensteinian interpretation of the core/penumbra distinction can divide legal directives into domains of formal realizability vel non, but, as Dworkin complains, waving one's hand at the penumbra and simply announcing that we should recognize that judges "legislate" is not politically acceptable. 81 One who believes that rules (and words) have cores and penumbras would argue that we need all attributes of the Rule of Law to make sure that in at least the core cases, the formalist result is properly implemented. This strategy supposes that part of the legal realm instantiates formalism and the precepts of the Rule of Law, while the balance is given over to the "discretion"-the "Absolute Arbitrary Power"-that the traditional Rule of Law, in both its instrumental and substantive conceptions, decries. In the "discretionary" areas we cannot be manipulating behavior by means of known pre-existing rules (the instrumentalist view) nor can we be fixing the structure that enhances negative liberty (the substantive view). Unless we declare by fiat that the penumbra is insignificant compared to the core, 82 this strategy gives up a good deal to the abyss of absolute arbitrary power, the elimination of which is supposed to be the raison d' etre of the social contract. Ronald Dworkin's bifurcated model of Jaw is aimed at answering this kind of criticism of Hart. When Dworkin attacked the model of rules he attributed to legal positivism, he did so by dividing up the general category of "standards" into "rules," on the one hand, and "principles, policies and other sorts 80

81 82

See, e.g., R. DWORKIN, RIGHTS, supra note 21, at 81-130. !d. at 44-45. This appears to be Hart's approach. See Hart, supra note 38, at 614-15.

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

61

805

of standards" that "do not function as rules," on the other. 83 His argument was that the model of rules does not fully describe what the law is, because law also consists of principles; and that rule-application does not fully describe what judges do, because judges also use principles. 84 Moreover, the use of principles determines one right answer to legal questions even though principles lack formal realizability, because in using principles the judge should apply the best interpretation of the entire body of past legal authoritative acts, where that interpretation rests on the best political theory available to explain and justify law as an enterprise. Can Wittgensteinian reinterpretation stop with this reinterpreted version of Dworkin? There ·are a number of important objections, only two of which I raise here. 85 First, in actual practice, in the form of life we call the law, it makes no sense to say that right answers are available in principle when in reality there is no agreement on the answers; the disputes stubbornly remain SeeR. DWORKIN, RIGHTS, supra note 21, at 22. Dworkin asserted that there is a "logical difference" between rules and principles. Rules are applicable either/or, ali-or-nothing, whereas principles have "weight." !d. at 26-27. Further, rules set out necessary conditions for their application but principles do not; in theory, a rule can be completely stated with all exceptions, but a principle cannot. !d. at 25-26. Although the notion of a "logical difference" can be given a Wittgensteinian reinterpretation, references to "logical difference" in conjunction with the idea of a complete statement of a rule with all its exceptions strongly suggest that Dworkin held the traditional conception of rules. In Dworkin's more recent work, LAw's EMPIRE, supra note 21, he has turned toward interpretive theory, which renders shaky the idea that any rules could be self-evidently applied in the way assumed by the traditional conception of rules. As far as I know, however, Dworkin has not explicitly reinterpreted his rules/principles distinction. 85 There is by now a large secondary literature on Dworkin and it is not my aim to add to it. In that literature, the lines of critique of Dworkin that I think are most significant are those relating to (1) the role of community and dialogue in legal meaning and decision making, see, e.g., Michelman, Foreword: Traces of Self-Government, 100 HARV. L. REv. 4, 66-73 (1986); (2) the role of political practice in law, see, e.g., Bruns, Law as Hermeneutics, in THE POLITICS OF INTERPRETATION 315 (W. Mitchell ed. 1983); and (3) the problem posed for coherence theory by indeterminacy and widespread conflict in the law. See, e.g., Altman, Legal Realism, Critical Legal Studies, and Dworkin, !5 PHIL. & Pus. AFF. 205, 217 (1986). I am sympathetic also to Stanley Fish's critique, suggesting that Dworkin's dimensions of "fit" and "soundness" take apart fact and value in a way Dworkin acknowledges should be rejected along with the old paradigm supporting positivism. See Fish, Working on the Chain Gang: Interpretation in the Law and in Literary Criticism, in THE POLITICS OF INTERPRETATION 271, 284-85 n.5 (W. Mitchell ed. 1983). In addition I believe there is an important critique that is underrepresented in the Dworkin literature. Contrary to the Critical Legal Studies critique that law is indeterminate and incoherent, it seems to me that law and its institutions may indeed exhibit integrity or coherence-for example, as they did with respect to racism and sexism-and yet be coherently wrong. A main task for nonfoundationalist theory is to find room for this kind of judgment. 83 84

The Rule of Law and the Separation of Powers

62

806

BOSTON UNIVERSITY LAW REVIEW

[VoL. 69: 781

when all the arguments are done. If those who are subject to judicial decisions cannot confidently predict the outcome, then Dworkin has not succeeded in preserving the Rule of Law virtues for the area in which "disputes break out"; he is susceptible to the same criticism leveled against Hart. When we really do "look and see" what law is like as a practice or a form of life, it is much more complicated than simply applying authoritative directives that are either rules or not-rules. At this point we might turn to the distinction between "rules" and "stan" dards" described by Duncan Kennedy and adopted by a number of critical writers. 86 Kennedy's claims are more complex than Dworkin's because Kennedy "continuum-izes" the dichotomyP A directive can tend toward "ruleness" (formal realizability) or be relatively standard-like. Thus the semantic underpinning of the formal realizability dimension must be complex: we must assume that words vary along a continuum from extremely vague (or otherwise indeterminate) to completely determinate. 88 This vision may be adequately Wittgensteinian in its understanding of language (or may be reinterpreted to be), but it still seems to leave out something important in its conception of law as consisting of verbal directives of one kind or another. In order to be true Wittgensteinians, at least on the neo-pragmatic understanding of him, we must find a way to say things about law as an activity or form of life containing practices other than the giving and following of verbal directives. 86 See supra note 44 and accompanying text; Kennedy, supra note 43, at 1687-1701; see also Schlag, supra note 44, at 429-30 (accepting the rules/standards distinction and arguing that the distinction has been institutionalized in legal argument so as to form a dialectic incapable of resolution). 87 See Kennedy, The Stages of the Decline of the Public/Private Distinction, 130 U. PA. L. REV. 1349, 1352-53 (1982). 88 Kennedy further claims that individualism is associated with rules and altruism with standards. This claim is obscure for various reasons we need not elaborate here. Although Kennedy asserts that we all know intuitively that individualism and altruism are associated with rules and standards respectively, polls among my first-year students show that this is not a widespread intuition. Standards may be associated with individualism because they allow decision-makers to treat persons as unique individuals; rules force decision-makers into a "collectivist" straight-jacket. From the communitarian point of view, one vice of standards is that an individualist judge might prefer the power of discretion a standard gives; a communitarian judge might count it a virtue of rules that they make her a rule-bound functionary. See Schlag, supra note 44, at 399426. If we press beyond Kennedy's critique of formalism in this article, so that even the conception of a "rule" cannot include the notion of traditional formal realizability, then we may be remanded to a social practice conception of rules that blurs the distinction that Kennedy made here. All rules are in a sense other-directed. In later work, Kennedy adopts a more Wittgensteinian view of rules and takes a more pragmatic view of law. See, e.g., Kennedy, Toward a Critical Phenomenology of Judging, in THE RULE OF LAW: IDEAL OR IDEOLOGY? 141 (A. Hutchinson & P. Monahan eds. 1987). See infra notes 97-99 and accompanying text.

63

The Rule of Law and the Separation of Powers

1989]

3.

THE RULE OF LAW

807

On the connection between philosophy and normative social activity: the repercussions of reconceiving rules

Although I speculated earlier that perhaps Wittgenstein himself wouldn't have thought that philosophical changes in 'our conception of rules would make much difference to us in real life, I think that view is too simplistic. At least on the pragmatic view of Wittgenstein that I find congenial, theory and practice cannot remain so separate; our conceptual schemes and human activity cannot be held apart. 89 Deep inner shifts in theoretical practices (like the "conversation" of philosophy 90) must have repercussions in other kinds of practices (like legal decisionmaking). In this section I suggest how far-reaching the inner shifts must be if we accept the Wittgensteinian view of rules. Later I speculate on how those shifts might be allied to a more pragmatic reinterpretation of the Rule of Law. The philosophical consequences we must accept if we accept the pragmatic Wittgensteinian view of rules are, I think, these: the tendency of "applying" rules to coalesce with "making" rules; the tendency of the "rule" to coalesce with the "particulars falling under it"; the idea that rules are contingent upon whole forms of life and not just specific acts of a legislature; and the essential mutability of rules. From a Wittgensteinian perspective, making rules and applying rules cannot be radically separate activities. A rule would cease to exist if we (the relevant community) stopped apprehending it as a rule and stopped recognizing ourselves and others as acting under it. This view of legal rules would contrast with one aspect of the positivist view formulated by H.L.A. Hart. Hart drew a sharp distinction between primary rules of obligation, which can be valid whether or not people obey them, and the rule of recognition, whose existence depends upon the observed actions and public commitments of the legal community.91 For a Wittgensteinian it seems the distinction cannot be quite so sharp. The rule of recognition itself certainly depends upon community agreement in practice, but so too (even if in a less direct sense) must the notion of validity. If a primary rule were never followed, at some point it would become questionable whether we were justified in thinking it to be a rule. · If we fail to follow a rule we undermine its existence as a rule. Thus when we do follow a rule we reaffirm its existence. Every time we apply a rule we also make it. Moreover, because a rule depends upon the existence of a social practice, it seems that an essential part of rule-application in a legal context is predicting the behavior of the community of judges and other legal actors, 89 See H. PUTNAM, REASON, TRUTH AND HISTORY 66-69 (1981); Radin, MarketInalienability, 100 HARV. L. REv. 1849, 1884-87 (1987) (exploring the consequences of using only market rhetoric to talk about human affairs). 90 R. RORTY, Pragmatism, Relativism, and Irrationalism, in CoNSEQUENCES OF PRAGMATISM 160 (1982) [hereinafter R. RORTY, Pragmatism]; R. RORTY, Mirror of Nature, supra note 74, at 389-94. 91 See H.L.A. HART, THE CONCEPT OF LAW 97-107 (1961).

The Rule of Law and the Separation of Powers

64

808

BOSTON UNIVERSITY LAW REVIEW

[VoL. 69: 781

including (it seems) one's own behavior if one is a legal actor. (At least this is true for the kind of rules Hart collectively called rules of recognition.) This is a very different picture of rule-application than the one embedded in the traditional formalist conception of rules, where the point of a rule is to close off contingencies related to one's own or others' behavior. If we accept the Wittgensteinian view of rules, there can also be no radical distinction between a rule and the particulars falling under it. This seems to deny the generality (in Fuller's sense) of rules. In a Wittgensteinian view, rules depend upon the practice of decisions that the relevant community accepts as rule-like. Rules do not wholly pre-exist the particular applications of them in practice, because people must actually follow rules before we can say rules exist. More ·broadly, a switch to the Wittgensteinian view of rules must profoundly affect our views about law and the Rule of Law because the Wittgensteinian view makes the existence of legal rules contingent not just upon the acts of legislatures or other authoritative entities, but also upon the surrounding social context, the content of an entire form of life. Of course, rules can come into being partly because we have specific practices (legislation, judging, agency promulgation) that formulate them. In order to find out what rules exist by virtue of these practices, we must "look and see" what the practice is. 92 But that which makes any directive rule-like relates not only to its promulgation according to an essentially social existing practice, but also to the content of surrounding social activities and understandings. 93 Rules are created and continue to exist not only because a legislature says so, but also by virtue of their being embedded in our nomos. 94 A judge's decision in response 92 According to Hart, for example, we would look and see if our rule of recognition validated the particular rules in which we were interested. Hart's rule of recognition, which has its being only in the acceptance of legal actors as manifested in their actions and in their continuing public commitments, is essentially pragmatic. (In fact, Hart wondered whether it even should be thought of as a "rule." See H.L.A. HART, supra note 91, at 107-14.) 93 Ultimately whether or not a rule exists for us depends upon a social commitment that is not rule-based. Cf F. SCHAUER, supra note 73, § 6.2 stating: But regardless of why a person might treat the existence of a rule as a reason for action ... the process of taking a rule to be applicable is a ful)ction of something external to that rule. It may be another rule, but that process ultimately runs out, and at that point we face the realization that something other than a rule, any rule, will determine the applicability of a rule, thus, to say that a rule exists for any agent is to say that that agent treats the rule, for non-rule-based reasons, as supplying a reason for action. And to say that a rule exists within some decisional environment is similarly to say that the decision-makers in that environment, again for ultimately non-rule-based reasons, treat the rule as relevant to the decisions they are called upon to make. (emphasis in original). 94 This use of the term "nomos" stems from the work of Robert Cover. See Cover, Poreword: Nomos and Narrative, 97 HARV. L. REv. 4 (1983).

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

65

809

to a rule responds necessarily to the community as a whole and not just to what the legislature has said. 95 Finally, if we accept the Wittgensteinian view we must recognize that rules are not immutable. Whether an activity is s.een as rule-like is contingent upon material social context and agreement. Over time the "same" action in response to the "same" directive can go from being compelled by a rule to not being compelled by a rule, or vice versa. 96 The traditional notion of law as rules cannot readily accommodate the idea that the contours of the law may shift through no legislative or official act but merely through social change. At what point in the process of coming-to-be-rule-like do we treat something as law? · The point of "the Rule of Law, not of individuals" is that the rules are supposed to rule. The easiest (most "natural") way to achieve that in our historical and philosophical context is to assume that rules apply to particula.r cases in an analytical or self-applying way. "Individual~"-judges, police, administrators-are needed to make sure these self-evident applications are carried out, but these individuals are not supposed to rule. They are to be rulebound, merely instrumental functionaries. Once we admit that rules are mutable and inextricable from material social practice, we will at least experience a psychological change in the way we perceive our roles as legal actors. This alone may have subtle but pervasive consequences in our practices. For example, suppose the majority of lawyers and judges come to see legal decisionmaking as pragmatic work in the sense described by Duncan Kennedy. 97 Kennedy describes the interaction between the views and desires of the legal actor and the \'normative power of the field. " 98 Even though the actor will sometimes find the field to be "impacted"-perceive it to be stubbornly rule-like-this cannot be read from the face of the legal materials. Often what looks at first to be rule-like will turn out not to be. I agree with Kennedy that on some not-quite-conscious level 95 And because acting under a rule re-affirms the nomos that treats any particular directive as binding, judging can never be choice-less or value free. When rules exist only because we continue to apprehend them as rules, it can be argued that we are morally responsible even for "mere" rule-application. Cover explored this responsibility in R. CoVER, JusTICE AccusED (1975), and Cover, supra note 94, at 53-68. 96 Compare this passage from Dworkin: We think the question whether someone may legally drive faster than the stipulated speed limit is an easy one because we assume at once that no account of the legal record that denied that paradigm would be competent. But someone whose convictions about justice and fairness were very different from ours might not find that question so easy; even if he ended by agreeing with our answer, he would insist that we were wrong to be so confident. This explains why questions considered easy during one period become hard before they again become easy questions -with the opposite answers. R. DWORKIN, LAW'S EMPIRE, supra note 21, at 354. 97 See Kennedy, supra note 88, at 150-67. 98 !d. at 157.

The Rule of Law and the Separation of Powers

66

BOSTON UNIVERSITY LAW REVIEW

810

[VoL. 69: 781

good lawyers know this. But if conscious awareness of this malleability in response to legal work-this non-pre-existing-ness of the law-replaces the prevailing positivist rhetoric about decisions predetermined by the plain meaning of rules, the character of our practice will change in ways we cannot now predict very well. 99 If we do come to think of rules in a different way-anti-formalist, nonfoundationalist, contingent, socially constructed-! think the change in thought would not be irrelevant, but rather would be incompatible with the traditional ideal of the Rule of Law. Will this shift in our conception of rules allow us to form a better conception or interpretation of the Rule of Law? Or will it force us to abandon the concept of the Rule of Law altogether? I think it is too soon to counsel abandonment. V.

A.

TOWARD REINTERPRETATION OF THE RULE OF LAW

Liberal Legalism and Non-Formalist Natural Law

In this section I consider a recent scholarly attempt to affirm the traditional Rule of Law while denying traditional formalism. In my view the attempt collapses back into liberal legalism and fails. Michael Moore lias recently argued for a "natural law" theory of legal interpretation and use of precedent, while denying the possibility of formalist rule-application. 100 Moore's theory is not the "natural law" of traditional foundationalism, but rather a reincarnation in terms of coherence theory. His claims are "(1) that there is a right answer to moral questions, a moral reality if you like; and (2) that the interpretive premises necessary to decide any case can and should be derived in part by recourse to the dictates of that moral reality~" 101 The connection between these notions and the Rule of Law is Moore's claim that "the rule of law virtues" constitute morally real values that must be part of judicial decisionmaking if judges are to reach morally right answers. 102 The six "virtues" Moore lists are: separation of powers, equality, liberty, substantive fairness, procedural fairness, and utility. 103 Although he cites Lon Fuller in this passage, it is clear that Moore's version of the Rule of Law is close to Rawls's substantive conception. Moore gives a liberal legalist gloss for each of his "virtues." For example, Moore intends the separation of powers virtue to refer to the traditional idea that judges should apply, not make, the law. He intends the equality virtue to refer to the idea of treating like cases alike ("formal justice"), and the liberty virtue to refer to curbing the chilling effect of 99 Cf R. RORTY, Pragmatism, supra note 90, at 174-75 (discussing the unforeseeability of the consequences of pragmatism). 100 See Moore, Theory of Interpretation, supra note 14, at 286-88; Moore, The Interpretive Turn in Modern Theory: A Turn for the Worse? 41 STAN. L. REV. 871 ( 1989) [hereinafter Moore, Interpretive Turn]. 101 See Moore, Theory of Interpretation, supra note 14, at 286. 102 Id. at 313-18. 103 Id. at 314.

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

67

811

legal uncertainty. Moore characterizes these "rule of law virtues" as "basic values that justify judges in a democratic society ... in accepting a humbler, more constrained mode of decisionmaking than that which they adopt when they view themselves as persons and not as functionaries of some kind." 104 According to Moore, "these six rule of law virtues together demand that judges give up some of the decisional freedom we each have as persons when deciding what, all things considered, it is best to do." 105 There are two salient results of coupling the claim of moral reality with the six asserted "rule of law virtues." First, the judge is to behave differently as a "functionary" tha.n as a "person," the difference being "constraint." Second, because these morally real values underlie stare decisis, sometimes a judge as "functionary" is morally required to stick with the wrong answers of the past (embodied in legal precedent) rather than what she as a "person" otherwise knows to be morally right. For Moore the Rule of Law "virtues'' are not only. "real," they are strong enough to turn (real) wrong into (real) right. Moore's theory echoes the traditional formalist model of rules in which it is the rules that are to "rule" or decide, not the person. Without satisfactory argument to show how the judge as "functionary" is simultaneously a responsible moral chooser (a "person"), this theory-in spite of itself-does not distance itself much from the traditional notion that the judge is a functionary who implements the pre-existing comi.ection between a rule and a set of particulars "falling under" it. 106 Moreover, because Moore's theory requires the judge as "functionary" to implement what she as a "person" knows to be wrong, his theory of the judge's role echoes the positivist separation of law and morals that he is at great pains to deny. 107 Without satisfactory argument about how the liberal legalist virtues turn wrong into right, the functionary judge's position suggests that whether the law is right or wrong is of no concern to the judge. Inherent !d. /d. at 318. 106 It might be thought that there is room for the person to be a inoral chooser because of the moral leeway afforded by the necessity of interpreting legal texts. In Theory of Interpretation, supra note 14, at 318 n.80, Moore distinguished legal reasoning from moral reasoning by saying that moral reasoning "does not involve the kind of textual exegesis involved in law, literary criticism, and theology." Because the Rule of Law virtues "demand that judges justify their decisions with reference to legal texts," legal reasoning "is more like dream interpretation, literary criticism, and theology than unconstrained moral reasoning." /d. at 318. According to Moore's view of moral reality, moral rules are there to be "discovered," whereas legal rules are not. But Moore believes that even though law is an interpretive activity, there can be authoritatively correct interpretations of it because texts have a metaphysically real meaning. "[T]o interpret dreams or statutes is to seek true explanations and evaluations, where 'true' is not a term internal to our practice but brought to it from general metaphysical and epistemological positions." Moore, Interpretive Turn, supra note 100, at 957. 1°7 See Moore, Semantics, supra note 14, at 153. 104 105

The Rule of Law and the Separation of Powers

68

812

BOSTON UNIVERSITY LAW REVIEW

[VoL. 69: 781

in this theory is a deep force toward conservatism, legitimation of the status quo, by the time-honored "natural law" strategy of writing it into the fabric of the universe. In spite of Moore's scorn for conventionalism, he does not adequately explain why the values we have inherited from liberal legalism are to be taken as "real" other than the fact that they are part of our legal culture. I am suggesting, in essence, that Moore's theory is at odds with itself. The tension at its heart is caused by his commitment both to a version of liberal legalism and to non-foundationalist views of meaning and justification. Ultimately these cannot coherently co-exist. The underlying commitment to liberal legalism-the traditional substantive conception of the Rule of Law-causes the traditional formalist characteristics of the judge's role as functionary and the traditional formalist interpretation of rules and precedent to re-assert themselves even as the non-foundationalist views of meaning and justification are destroying these ideas about rigidity of role and application of rules and precedent. It is no wonder that Moore argued-in an early article attacking all formalist theories of meaning-that formalism seems to be the theory of adjudication required by the Rule of Law. 108 For Moore the unstated problem is, how can we deny formalism and affirm the Rule of Law? His answer was that because formalism (by which he meant the traditional conception of formalism, not a Wittgensteinian reinterpretation) is impossible, and we wish to affirm the Rule of Law, the task for theory is to come "as close to formalism as knowledge of language will allow." 109 Apparently the strength of Moore's own commitment to liberal legalism does not allow him to see that his own arguments against traditional formalism also render meaningless the notion of coming "close to" it. If formalism is impossible, there is no "it" to which one can come close. 110 The Rule of Law and Interpretive Community: Toward a Pragmatic View

B.

Suppose we drop both the notion that law is mediated through formal rules in the traditional sense (accepting Wittgensteinian reinterpretation of formalId. at 153-58. Id. at 167. 11 ° Could a Wittgensteinian reinterpretation of formalism rehabilitate Moore's argument? One could contend that brute agreement on results, the socially evident compulsion of acting under a rule, does "come as close to formalism as knowledge of language will allow," and that, for all the traditional Rule of Law reasons-especially the separation of powers and constrained judicial role-we should formulate legal directives to maximize formal realizability (as reconceived). See, e.g., Schauer, Formalism, 97 YALE L.J. 509 (1988). Moore himself would not accept any Wittgensteinian view; to him Wittgenstein's reliance on human practice seems merely "conventionalist," the category his perspective opposes to the realism he espouses. See Moore, Theory of Interpretation, supra note 14, at 291-301. Nevertheless, the Wittgensteinian move here seems plausible. It leads, however, to the view that liberal legalism remains the same when we alter its philosophical underpinnings, a view I find too simplistic. 108 109

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

69

813

ism), and the broader idea that law consists essentially of verbal directives like rules and standards, apart from the practice in which they are embedded (accepting a more far-reaching reinterpretation). We will still find it deeply normatively appealing to conceive of ourselves as a people governed by its law rather than by arbitrary individual power, because conceiving of ourselves this way I take to be constitutive of ourselves as a political community. 111 The enduring normative appeal of the Rule of Law is the reason I pursue reinterpretation. The reinterpretation I want to pursue (barely beginning in this essay) turns toward a view of law that emphasizes practice as well as words. It is a view that turns toward pragmatism, seeking to view law as a pragmatic normative activity. In this view, hermeneutics-the view of interpretation and meaning as holistic and practice-based-forms the epistemology of law. The work of Robert Cover is suggestive for the reinterpretation I seek. Cover is perhaps the quintessential modern anti-positivist. He claims that the role of judges is rightly law-creating ("jurisgenerative"). Thus, he stands on its head the traditional slogan that judges should apply rather than make law. 112 When judges slip into the old rhetoric and claim to act not as people but as functionaries whose hands are tied, they "substitut[e] the hermeneutic of jurisdiction for the hermeneutic of the text." 113 In other words, they refuse to take responsibility for their actions by taking refuge in their role as rule-followers. "Judges are people of violence" 114 ; one way of being violent is to disavow responsibility for the consequences of their functionary behavior. In Cover's view, judges are not functionaries but rather constitute an interpretive community. Law-creation is not unique to them; it inheres in all interpretive communities. In interpretive communities, "applying" and "making" law coalesce. As a community applies its law, it continuously [re-]makes it. 115 Law is not the creature of the state but the mark of an interpretive community. Law is not made by legislatures (or judges) alone. Under the sway of the positivist model of law as rules, the role of the state has been to crush all law contrary to that of the state. This Cover deplores. For Cover, to act rightly as a judge is to refuse to use the violence of one's office to enforce the law of the state against the law of various dissenting communities, while at the same time recognizing and taking responsibility for the fact that the law of the community of judges can come into conflict with, and is not intrinsically privileged 111 See Michelman, supra note 9, at 1499-1503 (discussing the way we conceive selfrule and law-rule). 112 Cover, supra note 94, at 53-60. 113 ld. at 58. 114 ld. at 53. 11 5 Even the new Dworkin of LAw's EMPIRE, supra note 21, would, it seems, readily acquiesce to this last statement. "[Hercules] does not amend out-of-date statutes to suit new times . . . . He recognizes what the old statutes have since become." See, e.g., id. at 350. (Unlike Cover, however, Dworkin is resolutely statist; he seems to write on the heroic assumption that we are one interpretive community expressing itself through the state.)

70

The Rule of Law and the Separation of Powers 814

BOSTON UNIVERSITY LAW REVIEW

[VoL. 69: 781

over, the law of other interpretive communities. Like Cover's theory of law, hermeneutic social theories have rejected foundationalism, formalism, and the idea that a rule could be applied by one person, only once, apart from a group in whose interpretive practice the rule is embedded. Hermeneutic theories have also rejected the notion that there can be application without interpretation, or interpretation without politics and value, or politics and value without commitment. 116 These theories view interpretation as holistic, pragmatic, and historically situated. Because of these features, hermeneutic theory bears an affinity with the turn toward coherence theories and pragmatism in ethics and metaphysics. Hermeneutic theory thus seems more promising than the new non-foundationalist "natural law" for legal theorists who seek to exorcise formalism. A pragmatic reinterpretation of the Rule of Law would at least deny that law consists of formally realizable rules in the traditional sense. More controversially, perhaps, I believe such a reinterpretation would deny that law consists quintessentially of rules at all, as well as the notion that rules are separate from cases and logically pre-exist their application. Such a reinterpretation would also deny the strict division of people into rule-givers and rule-followers, and the conception of judges as rule-appliers rather than rule-makers. How would these broad theoretical features play out into a substantive conception of the Rule of Law that can supersede the traditional one? One task in answering this question is the one I left aside in this essay: the remaking of the supporting theories of the person and of politics. 117 In the meantime, we can set out to "look and see," readjusting our theory and thus the interpretation of our practice at each step along the way. Thus I suggest that we explore how each precept of the Rule of Law might be reinterpreted. If we can retain the entire complex, or most of it, but with new philosophical underpinnings, we shall retain the Rule of Law as a central normative commitment of our legal system. For those who accept the new interpretation, however, its meaning and import for the form of life we call law may be very different. I suggest we begin by reconsidering the Rule of Law precepts clustering around notice, non-retroactivity, and the separation of powers. In the instrumental view of the traditional model of rules, a law must not be retroactive because it cannot be the cause of conforming behavior unless it precedes the behavior. A law must be made public because it cannot be the cause of conforming behavior if the addressees do not know about it. In the substantive view of the traditional Rule of Law, fairness is the value underlying the requirements of non-retroactivity and notice. These precepts both rest 116 See HERMENEUTICS AND PRAXIS xiii-xvi (R. Hollinger ed. 1985); R. RORTY, MIRROR OF NATURE, supra note 74, at 315-21; Garet, Comparative Normative Hermeneutics: Scripture, Literature, Constitution, 58 S. CAL. L. REv. 35, 38-39 (1985). 117 See supra notes 9-46 and accompanying text. See also Michelman, supra note 9, at 1502 (arriving at something close to the same point as I do through reinterpretation of politics).

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

71

815

on the idea that it is unfair to base adverse government interference with individual interests on the individual's failure to conform to rules the individual cannot fairly be charged to have known in advance of the non-conforming behavior, either because the rule has not properly been made public (the notice precept) or because the rule has not yet been enacted at the time of the conduct (the non-retroactivity precept). What the pragmatic approach can add to this traditional view is a broader understanding of what constitutes sufficient publicity or notice, and a reinterpretation of retroactivity in light of the idea that rules are not made merely by legislatures or other authoritative entities. In the pragmatic view, a rule will be public whenever strong social agreement exists in practice, regardless of whether a legislature or a court has spoken.118 Similarly, if a rule exists normatively even without specific legislative enactment (as, for example, would a rule against intentional homicide), then later legislative confirmation would not necessarily mean that it would be un" fair retroactive application to punish earlier transgressions. 119 Moreover, where the line of evolution of legal interpretation is clearly foreseeable, it would not be unfair to hold people to what they can see is the emerging interpretation. 120 The Rule of Law precepts concerning separation of powers and the nature of the judicial role pose a more complex issue. If we accept the Wittgensteinian view of rules, we must reject the conception of the separation of powers that pictures a rigid distinction between the legislature as rule-maker and the judges as rule-appliers; indeed, we must reject, as well, the more general 118 In Keeler v. Superior Court, 2 Cal. 3d 619, 470 P.2d 617, 87 Cal. Rptr. 481 (1970), the California Supreme Court decided that a man who assaulted his ex-wife when she was nine months pregnant and killed the fetus, with intent to do so, could not be prosecuted for murder because common-law judicial precedent did not clearly give notice that killing an unborn baby could result in a murder prosecution. /d. at 636, 470 P.2d at 628, 87 Cal. Rptr. at 492. A pragmatic view would make us confront the unspoken social context-an abortion issue the Court did not wish to discuss-by asking: is our contemporary nomos one in which the killing of an unborn nine-month fetus by assault on the mother recognizable to us as murder? 119 For example, if defendant assaults with intent to kill and the victim dies fourteen months later, it is fair to abolish the common-law year and a day rule and prosecute defendant for murder. The defendant can be fairly charged with knowing that if the victim dies the crime is murder, and we do not mind the "chilling effect" on people's liberty to assault with intent to kill. See United States v. Jackson, 528 A.2d 1211, 1216-20 (D.C. 1987). 120 An interesting passage in the Hart and Sacks materials suggests that Buick Motor Co. could not object on notice and non-retroactivity grounds to Cardozo's decision in MacPherson v. Buick, 217 N.Y. 382, 389, 111 N.E. 1050, 1053 (1916) (holding the manufacturer liable to the ultimate consumer for injury caused by its negligence), because Buick could not reasonably maintain that it was justified in behaving negligently in light of the old (and decaying) doctrine limiting recovery to those in privity of contract with the manufacturer. H. HART & A. SACKS, THE LEGAL PROCESS 575 (tent. ed. 1958).

The Rule of Law and the Separation of Powers

72

816

BOSTON UNIVERSITY LAW REVIEW

[VoL. 69: 781

distinction between government as rule-maker and citizens as rule-followers. But this does not mean that there is no difference between the judicial role and the legislative role. The pragmatic view must certainly accept that the distinction is normatively important in the political tradition of our community. In the pragmatic normative understanding, our constitution is not merely a document but rather that which "constitutes" us as a political community. 121 We are constituted both by commitment to majority rule and by countermajoritarian commitments. One way to understand this constitutional dualism is to see it as a reflection of the tension between our understanding of our present state and our understanding of social ideals toward which progress is possible. In this pragmatic view of politics, we are always attempting to accomplish a transition from today's nonideal world to the better world of our vision, and it is a transition that never ends. 122 Moreover, our visions and our nonideal reality paradoxically constitute each other: what we can formulate as being better depends upon where we are now, and the way we understand where we are now depends upon our vision of what should be. To the extent we live in present imperfect reality, majority rule is normatively appealing, and so is conventionalism in our understanding of morals and politics. To the extent we live in our visions of a better world (even though such visions must be contingent upon our present circumstances), judgment against the majority is normatively appealing where the majority has transgressed those visions, and so is some form of transcendence of conventionalism in our understanding of morals and politics. Because our constitutional tradition partakes of both tendencies, we constitute ourselves both by conventional reality and by vision. Our understanding of the separation of powers can be reinterpreted in light of this constitutive tension. We might say that often legislatures are responsible for fidelity to convention, and judges for fidelity to vision. More important, perhaps, is recognition that legislatures and judges must interact in pursuing our vision. Consider, for example, civil rights in the modern era. The Court led the way in Brown v. Board of Education, 123 but in Bell v. Maryland, 124 it was able to let civil rights legislation carry the weight of progress toward the vision of social equality. The interaction of courts and legislatures in the modern development of the 121 An important issue is whether "we" are indeed one political community. See infra note 127. It seems to me that "we" are one community on the issue of whether 2 + 2 = 4, but perhaps we are many diverse communities on other issues more readily regarded as ethical, religious, or political. The pragmatic normative significance of our having a constitution is to deny at least the furthest reaches of such pessimistic irreconcilable pluralism. We are one in at least some sense(s). 122 See Radin, supra note 89, at 1875-76. 123 347 U.S. 483 (1954). It is important to recall that traditional concerns about neutral principles and the Rule of Law made Brown controversial when it was decided. For an eloquent pragmatic defense of the Court's decision, see Black, The Lawfulness of the Segregation Decisions, 69 YALE L.J. 421, 424-28 (1960). 124 378 U.S. 226 (1964).

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

73

817

law of landlord and tenant is an equally good although less celebrated example. Judges like J. Skelley Wright could move forward on their understanding of fair treatment of tenants in principle, but it was for legislatures to tell us what exactly might count as retaliatory eviction or breach of the implied warranty of habitability. 125 Independence and impartiality are traditional aspects of the judge's role as a functionary separate from the legislative power; these elements of the Rule of Law are supposed to ensure that judges stay within their task of rule-applying. But independence and impartiality can refer to moral autonomy and a commitment to judgment in light of one's own moral understanding of the nature of community-not just to formal separation from the interests of the legislature or the litigants. Robert Cover, for example, regards independence for a judge in the same way as the independence characterizing a person of moral integrity who faces choices that impact on other people's lives, not as merely formal independence of a functionary whose hands are tied. 126 In such an anti-formalist view of the judge's role, judges are an interpretive community conscious of their obligation to act as independent moral choosers for the good of a society, in light of what that society is and can become. 127 The law, as long as it is part of a viable and developing community, is neither "found" nor "made," but continuously re-interpreted. There are still rules. But there are no rules that can be understood apart from their context; nor are there rules that can be understood as fixed in time. 128 125 When Judge Wright decided Robinson v. Diamond Housing Co., 463 F.2d 853 (D.C. Cir. 1972) (the case with which I introduced the problem of the Rule of Law in this essay), he could point to legislation enacted subsequent to the events at issue to support his result. The legislation Judge Wright looked to was in turn based upon earlier court decisions. !d. at 857 n.l. For many other examples of pragmatic interaction between legislatures and courts, see G. CALABRESI, A COMMON LAW FOR THE AGE OF STATUTES (1982). 126 See, e.g., Cover, supra note 94, at 58-59 (advocating a "natural law of jurisdiction" which "implies the articulation' of a legal principle according to an independent hermeneutic"); R. CovER, supra note 95, at 171 (exploring the moral predicament of anti-slavery judges who nevertheless enforced the fugitive slave laws ·in the era before the Civil War). See especially id. at Part III ("The Moral-Formal Dilemma"). 127 There are problems I gloss over here that must be taken very seriously. Is it appropriate to think of ourselves in any sense as only one community or does that do violence to heterodox groups that are less powerful, hence less able to say who "we" are? And is it intolerable elitism to think of judges as exercising moral autonomy on behalf of the rest of us? See Sullivan, Against Republicanism, in AFTER THE BICENTENNIAL (M. Tushnet ed.) (forthcoming); and Sullivan, Rainbow Republicanism, 97 YALE LJ. 1713, 1716-18 (1988). 128 Indeed, this feature of continuous reinterpretation by an interpretive community of judges and those who argue to and with them help us understand the dynamic adaptability of the common law. It explains why the common law is traditionally characterized by organic metaphors like life and growth, and why it has resisted all theoretical attempts to bring it satisfactorily within a model of rules.

The Rule of Law and the Separation of Powers

74

818

BOSTON UNIVERSITY LAW REVIEW

VI.

[VoL. 69: 781

CONCLUSION

In Robinson v. Diamond Housing Corporation, Judge Wright did not see himself as a functionary implementing pre-existing formal rules. Yet he did see himself as acting in accord with the Rule of Law. If we are receptive to the reinterpretation I have begun to sketch, I think we may now see at least how he was trying to act in accord with the Rule of Law. Judge Wright recognized that in order even to know what his own prior cases "held" and their bearing on the case before him, he had to formulate a normative understanding of what the legislature and courts were doing by promulgating and enforcing a housing code. What they were doing depended not just on the words of the code, but also on the social context of housing in the District of Columbia community at the time, where (it seemed to Judge Wright, at least) landlords were commercial enterprises controlled by powerful groups and tenants were poor and powerless. What the code meant depended also on the broader social context of the Vietnam era. 129 Within this context, Judge Wright understood the normative force of the law as affirming not only that there should be safe and sanitary housing for people, but also that people themselves should be involved in bringing this about by enforcing their housing rights in court. If he had allowed a situation to continue in which tenants would immediately lose their homes whenever they tried to invoke the law to improve those homes, the aspect of the law calling for community participation would be a cruel hoax, and the group upon whom it was perpetrated would become even more alienated from the community as a whole. In light of his obligation to support his community, Judge Wright saw the law as working against this alienating result. The Rule of Law ideal itself persuaded Judge Wright to see the law this way. One could fault Judge Wright for not going further and enjoining the landlord to repair the house, rather than arriving at the ironic result that the tenant's right to habitable housing meant she could stay indefinitely in uninhabitable housing for free. Perhaps even the iconoclastic Judge Wright could not imagine bucking the traditional disfavor for affirmative injunctions. Perhaps it was relevant that in this particular case there was no need for such an injunction because Mrs. Robinson had in fact stopped living in the unsafe and unsanitary house before final disposition on appeal. One could also fault Judge Wright for taking too sanguine a view of the economic situation, if he thought that market forces would necessarily prevent landlords from sticking tenants with the increased costs of making housing habitable. 130 129 In a letter to Professor Edward H. Rabin, Judge Wright said, "I was indeed influenced by the fact that, during the nationwide racial turmoil of the sixties and the unrest caused by the injustice of racially selective service in Vietnam, most of the tenants in Washington, D.C. slums were poor and black and most of the landlords were rich and white." Rabin, The Revolution in Residential Landlord Tenant Law: Causes and Consequences, 69 CORNELL L. REv. 517, 549 (1984). 130 The citation of Bruce Ackerman's theoretical article on the economic effects of

The Rule of Law and the Separation of Powers

1989]

THE RULE OF LAW

75 819

I do not think, however, that we should fault Judge Wright merely because he did not adhere to the traditional conception of law as static formal rules. His action in trying to keep open tenants' avenues of redress and thereby keep the community open to tenant participation, while at the same time not depriving landlords of the expected monetary gain afforded by property ownership, seems quite law-like. It seems so, that is, when law is understood to mean not just a set of rules laid down but rather to include an evolving complex of political commitments to the flourishing of the community and the individuals in it. In the view of law as a pragmatic normative practice, law does not disappear. But it is always open to people to recognize, in various ways, that the law in the statute books is not the real law. 131 Some practices are so deeply accepted that they seem like immutable rules. Life according to rules is not impossible but quite routine, as Wittgenstein saw. It is only that, if we take the pragmatic and hermeneutic view of law, our understanding of rule-following must be reconstituted so that we know that rules are neither formal in the traditional sense, nor eternal, nor existing independently of us; and so that we know that every application of them is a reinterpretation. We must know that each time we feel ourselves to be rule-followers we are rule-creators as well. Whether or not the term "the Rule of Law" should be dropped becomes another pragmatic question, a matter of judgment about its ideological baggage. Is the term indelibly linked in our usage with the ideas that law is instrumental and consists of pre-existing formal rules applied in a value-free manner? Or can its precepts be reinterpreted in the way I have begun to suggest? If we can use the term in this new way, it seems that we would at least have to drop the slogan, "the Rule of Law, not of individuals." If law cannot be formal rules, its people cannot be mere functionaries. housing code enforcement as if it were an empirical study of the D.C. housing market is indeed a weakness in Judge Wright's opinion. See Robinson, 463 F.2d at 860, citing Ackerman, Regulating Slum Housing Markets on Behalf of the Poor: Of Housing Codes, Housing Subsidies and Income Redistribution Policy, 80 YALE L.J. I 093 (1971). 131 Even Dworkin now allows for this possibility by claiming that the Rule of Law includes fulfillment of moral rights even if they are not enacted as positive law. See R. DwoRKIN, supra note 10, at 11-13 (contrasting this rights conception with the "rulebook" conception); cf R. DwORKIN, LAw's EMPIRE, supra note 21, at 211-15 (the model of community as commitment to political integrity includes recognition that rights and duties are not exhausted by particular institutional decisions and enactments).

[3] THE RULE OF LAW AND ITS VIRTUE

*

Joseph Raz F. A. HAYEK has provided one of the clearest and most powerful formulations of the ideal of the rule of law: " stripped of all technicalities this means that government in all its actions is bound by rules fixed and announced beforehand-rules which make it possible to foresee with fair certainty how the authority will use its coercive powers in giVen circumstances, and to plan one's individual affairs on the basis of this knowledge." 1 At the same time the way he draws certain conclusiOns from this ideal illustrates one of the two main fallacies in the contemporary treatment of the doctrine of the rule of law: The assumption of its overriding importance. My purpose is to analyse the ideal of the rule of law in the spirit of Hayek's quoted statement of it and to show why some of the conclusions which he drew from it cannot be thus supported. But first we must be put on our guard against the other. common fallacy concerning the rule of law. Not uncommonly when a political ideal captures the imagination of large numbers of people its name becomes a slogan used by supporters of ideals which bear little or no relation to the one it originally designated. The fate of " democracy " not long ago and of " privacy " today are just two examples of this familiar process. In 1959 the International Congress of Jurists meeting in New Delhi gave official blessing to a s1milar perversion of the doctrine of the rule of law. " The function of the legislature in a free society under the Rule of Law is to create and maintain the conditions which will uphold the dignity of man as an individual. This dignity requires not only the recognition of his civil and political rights but also the establishment of the social, economic, educational and cultural conditions which are essential to the full development of his personality." 2 The report goes on to mention or refer to just about every political ideal which has found support in any part of the globe during the post-war years. If the rule of law is the rule of the good law then to explain its nature is to propound a complete social philosophy. But if so the * A draft of this paper was presented to a conference sponsored by tbe Liberty Fund and the Umversity of San Francisco. I am grateful to Professor Rolf Sartorms and Mr. David Libling for useful suggestions on ways to Improve an early draft of tbe paper. 1 The Road to Serfdom (London, 1944), p. 54. 2 Clause 1 of the report of Committee I of the International Congress of Jurists at New Delhi, 1959.

The Rule of Law and the Separation of Powers

78

The Law Quarterly Review

196

[VOL. 93

term lacks any useful function. We have no need to be converted to the rule of law just in order to discover that to believe in it is to believe that good should triumph. The rule of law is a political ideal which a legal system may lack or may possess to a greater or lesser degree. That much is common ground. It is also to be insisted that the rule of law is just one of the virtues which a legal system may possess and by which it is to be judged. It is not to be confused with democracy, justice, equality (before the law or otherwise), human rights of any kind or respect for persons or for the dignity of man. A non-democratic legal system, based on the denial of ltulnan rights, on extensive poverty, on racial segregation, sexual inequalities and religious persecution may, in principle, conform to the requirements of the rule of law better than any of the legal systems of the more enlightened western democracies. This does not mean that it will be better than those western democracies. It will be a:n immeasurably worse legal system, but it will excel in one respect: in its conformity to the rule of law. Given the promiscuous use made in recent years of the expression " the rule of law " it is hardly surprising that my claim will alarm many. We have reached the stage in which no purist can claim that truth is on his side and blame the others of distorting the notion of the rule of law. All that I can claim for my account is, first, that it presents a coherent view of one important virtue which legal systems should possess and, secondly, that it is not original, that I am following in the footsteps of Hayek and of many others who understood " the rule of law" in similar ways. 1. THE BASIC IDEA

"The rule of law" means literally what it says: The rule of the law. Taken in its broadest sense this means that people should obey the law and be ruled by it. 3 But in political and legal theory it has come to be read in a narrower sense, that the government shall be ruled by the law and subject to it. The ideal of the rule of law in this sense is often expressed by the phrase " government by law and not by men." No sooner does one use these formulae than their obscurity becomes evident. Surely government must be both by law and by men. It is said that the rule of law means that all government action must have foundation in law, must be authorised by law. But is not that a tautology? Actions not authorised by law cannot be the actions of the government as a government. They would be without legal effect and often unlawful. 3

Cf , on this sense of the phrase, Jennings, The Law and the Constitution (London,

1933), pp 42-45.

The Rule of Law and the Separation of Powers

APRIL 1977]

The Rule of Law and its Virtue

79

197

It is true that we can elaborate a political notion of government which is different from the legal one: government as the location of real power in the society. It is in this sense that one can say that Britain is governed by The City or by ithe trade unions. In this sense of government it is not a tautology to say that government should be based on law. If the trade union ruling a country breaks an industrial relations law in order to impose its will on tbe parliament or if the President or the F.B.I. authorise burglaries and conspire to pervert justice they can be said to violate the rule of law. But here the rule of law is used in its original sense of obedience to law. Powerful people and people in government just like anybody else should obey the law. This is no doubt correct, and yet does it exhaust the meaning of the rule of law? There is more to the rule of law than the law and order interpretation allows. It means more even than law and order applied to the government. I shall proceed on the assumption that we are concerned with government in the legal sense and with the conception of the rule of law which applies to government and to law and is no mere application of the law and order conception. The problem is that now we are back with our initial puzzle. If government is, by definition, government authorised by law the rule of law seems to amount to an empty tautology, not a political ideal. The solution to this riddle is in the difference between the professional and the lay sense of law. For the lawyer anything is the law if it meets the conditions of validity laid down in the system's rules of recognition or in other rules of the system.• This includes the constitution, parliamentary legislation, ministerial regulations, policeman's orders, the regulations of limited companies, conditions imposed in trading licences, etc. To the layman the law consists only of a subclass of these. To him the law is essentially a set of open, general and relatively stable laws. Government by law and not by men is not a tautology if "law " means general, open and relatively stable law. In fact the danger of this interpretation is that the rule of law might set too"strict a requirement, one which no legal system can meet and which embodies very little virtue. It is humanly inconceivable that law can consist only of general rules and it is very undesirable that it should. Just as we need government both by laws and my men, so we need both general and particular laws to carry out the jobs for which we need the law. The doctrine of the rule of law does not deny that every legal system should consist of both general, open and stable rules (the • I am here following Hart, The Concept of Law (Oxford, 1961), pp. 97-107.

The Rule of Law and the Separation of Powers

80

1-98

The Law Quarterly Review

[VoL, 93

popular conception of law) and particular laws Oegal orders), an essential tool in the hands of the executive and the judiciary alike. As we shall see, what the doctrine requires is the subjection of particular laws to general, open and stable ones. It is one of the important principles of the doctrine that the making of particular laws should be guided by open and relatively stable general rules. This principle shows how the slogan of the rule of law and not of men can be read as a meaningful political ideal. The principle does not, however, exhaust the meaning of the rule of law~d does not by itself illuminate the reasons for its alleged importance. Let us, therefore, return to the literal sense of the " rule of law." It has two aspects: (1) that people should be ruled by the law and obey it, and (2) that the law should be such that people will be able to be guided by it. As was noted above, it is with the second aspect that we are concerned: the law must be capable of being obeyed. A person conforms with the law to the extent that he does not break the law But he obeys the law only if part of his reason for conforming is his knowledge of the law. Therefore, if the law is to be obeyed it must be capable of guiding the behaviour of its subjects. It must be such that they can find out what it is and act on it. This is the basic intuition from which the doctrine of the rule of law derives: the law must be capable of guiding the behaviour of its subjects. It is evident that this conception of the rule of law is a formal one. It says nothing about how the law is to be made: by tyrants, democratic majorities or any other way. It says nothing about fundamental rights, about equality or justice. It may even be thought that this version of the doctrine is formal to the extent that it is almost devoid of content. This is far from the truth. Most of the requirements which were associated with the rule of law before it came to signify all the virtues of the state can be derived from this one basic idea. 2.

SOME PRINCIPLES

Many of the principles which can be derived from the basic idea of the rule of law depend for their validity or importance on the particular circumstances of different societies. There is little point in trying to enumerate them all, but some of the more important ones might be mentioned: (1) All laws should be prospective, open and clear. One cannot be guided by a retroactive law. It does not exist at the time of action. Sometimes it is then known for certain that a retroactive law will be enacted. When this happens retroactivity does not conflict with the rule of law (though it may be objected to on other grounds). The law must be open and adequately publicised. If it is to guide people

The Rule of Law and the Separation of Powers APRIL

1977]

The Rule of Law and its Virtue

81

199

they must be able to find out what it is. For the same reason its meaning must be clear. An ambiguous, vague, obscure or imprecise law is likely to mislead or confuse at least some of those who desire to be guided by it. (2) Laws should be relatively stable. They should not be changed too often. If they are frequently changed people will find it difficult to find out what the law is at any given moment and will be constantly in fear that the law has been changed since they last learnt what it is. But more important still is the fact that people need to know the law not only for short-term decisions (where to park one's car, how much alcohol is allowed in duty free, etc.) but also for longterm planning. Knowledge of at least the general outlines and sometimes even of details of tax law and company law are often important for business plans which will bear fruit only years later. Stability is essential if people are to be guided by law in their longterm decisions. 5 Three important points are illustrated by this principle. First. conformity to the rule of law is often a matter of degree not only when the conformity of the legal system as a whole is at stake, but also with respect to single laws. A law is either retroactive or not, but it can be more o~r less clear, more or less stable, etc. It should be remembered, however, that by asserting that conformity to the principles is a matter of degree it is not meant that the degree of conformity can be quantitatively measured by counting the number of infringements or some such method. Some infringements are worse than others. Some violate the principles in a formal way only which does not offend against the spirit of the doctrine. Secondly, the principles of the rule of law affect primarily the content and form of the law (it should be prospective, clear, etc.) but not only them. They also affect the manner of government beyond what is or can usefully be prescribed by law. The requirement of stability cannot be usefully subject to complete legal regulation. It is largely a matter for wise governmental policy. Thirdly, though the rule of law concerns primarily private citizens as subject to duties and governmental agencies in the exercise of their powers (on which more below) it is also concerned with the exercise of private powers. Power-conferring rules are designed to guide behaviour and should conform to the doctrine of rule of law if they are to be capable of doing so effectively. (3) The making of particular laws (particular legal orders) should 5 Of course uncertainty generated by instability of law also affects people's planning and action. If it did not, nor would stability have any impact The point is that only if the law is stable are people guided by their knowledge of the content of the law.

82

The Rule of Law and the Separation of Powers

200

The Law Quarterly Review

[VoL. 93

be guided by open, stable, clear and general rules. It is sometimes assumed that the requirement of generality is of the essence of the rule of law. This notion derives (as noted above) from the literal interpretation of " the rule of law " when " law " is read in its lay connotations as being restricted to general, stable and open law. It is also reinforced by a belief that the rule of law is particularly relevant to the protection of equality and that equality is related to the generality of law. The last belief is, as has been often noted before, mistaken. Racial, religious and all manner of discrimination is not only compatible but often institutionalised by general rules. The formal conception of the rule of law which I am defending does not object to particular legal orders as long as they are stable, clear, etc. But of course particular legal orders are mostly used by government agencies to introduce flexibility into the law. A police constable regulating traffic, a licensing authority granting a licence under certain conditions, all these and their like are among the more ephemeral parts of the law. As such they run counter to the basic idea of the rule of law. They make it difficult for people to plan ahead on the basis of their knowledge of the law. This difficulty is overcome to a large extent if particular laws of an ephemeral status are enacted only within a framework set by general laws which are more durable and which impose limits on the unpredictability introduced by the particular orders. Two kinds of general rules create the framework for the enactment of particular laws: Those which confer the necessary powers for making valid orders and those which impose duties instructing the power-holders how to exercise their powers. Both have equal importance in creating a stable framework for the creation of particular legal orders. Clearly, similar considerations apply to general legal regulations which do not meet the requirement of stability. They too should be circumscribed to conform to a stable framework. Hence the requirement that much of the subordinate administrative law-making should be made to conform to detailed ground rules laid down in framework laws. It is essential, however, not to confuse this argument with democratic arguments for the close supervision of popularly-elected bodies over law-making by non-elected ones These further arguments may be valid but have nothing to do with the rule of law, and though sometimes they reinforce rule of law type arguments, on other occasions they support different and even ·conflicting conclusions. (4) The independence of the judiciary must be guaranteed. It is f the essence of municipal legal systems that they institute judicial

The Rule of Law and the Separation of Powers APRIL

1977]

The Rule of Law and its Virtue

83

201

bodies charged, among other things, with the duty of applying the law to cases brought before them and whose judgments and conclusions as to the legal merits of those cases are final. Since just about any matter arising under any law can be subject to a conclusive court judgment it is obvious that it is futile to guide one's action on the basis of the law if when the matter comes to adjudication the courts will not apply the law and will act for some other reasons. The point can be put even more strongly. Since the court's judgment establishes conclusively what is the law in the case before it, the litigants can be guided by law only if the judges apply the law correctly. 6 Otherwise people will only be able to be guided by their guesses as to what the courts are likely to do-but these guesses will not be based on the law but on other considerations. The rules concerning the independence of the judiciary-the method of appointing judges, their security of tenure, the way of fixing their salaries and other conditions of service-are designed to guarantee that they will be free from extraneous pressures and independent of all authority save that of the law. They are, therefore, essential for the preservation of the rule of law. (5) The principles of natural justice must be observed. Open and fair hearing, absence of bias and the like are obviously essential for the correct application of the law and thus, through the very same considerations mentioned above, to its ability to guide action. , (6) The courts should have review powers over the implementation of the other principles. This includes review of both subordinate and parliamentary legislation and of administrative action, but in itself it is a very limited review-merely to ensure conformity to the rule of law. (7) The courts should be easily accessible. Given the central position of the courts in ensuring the rule of law (see principles 4 and 6) it is obvious that their accessibility is of paramount importance. Long delays, excessive costs, etc., may effectively turn the most enlightened law to a dead letter and frustrate one's ability effectively to guide oneself by the law. (8) The discretion of ihe crime preventing agencies should not be allowed to' pervert the law. Not only the courts but also the actions of the poUce and the prosecuting authorities can subvert the law. The prosecution should not be allowed, e.g. to decide not to prosecute for commission of certain crimes, or for crimes committed by certain classes of offenders. The police should not be s I am not denying that courts also make law This principle of the rule of law applies to them primarily in their duty to apply the law. As law-makers they are subject to the same principles as all law-makers.

The Rule of Law and the Separation of Powers

84

202

The Law Quarterly Review

[VoL. 93

allowed to allocate its resources so as to avoid all effort to prevent and detect certain crimes or prosecute certain classes of criminals. This list is very incomplete. Other principles could be mentioned and those which have been mentioned need further elaboration ang further justification (why-as required by my sixth principle-should the courts and not some other body be in charge of reviewing conformity to the rule of law? etc.). 7 My purpose:-in listing them was merely to illustrate the power and fruitfulness o,f the formal conception of the rule of law. It should, however, be remembered that in the final analysis the doctrine rests on its basic idea that the law should be capable of providing effective guidance. The principles do not stand on their own. They must be constantly interpreted in light of the basic idea. The eight principles listed fall into two groups. Principles 1 to 3 require that the law should conform to standards designed to enable it effectively to guide action. Principles 4 to 8 are designed to ensure that the legal machinery of enforcing the law should not deprive it of its ability to guide through distorted enforcement and that it shall be capable of supervising conformity to the rule of law and provide effective remedies in cases of deviation from it. All the principles directly concern the system and method of government in matters directly relevant to the rule of law. Needless to say many other aspects in the life of a community may, in more indirect ways, either strengthen or weaken the rule of law. A free press run by people anxious to defend the rule of law is of great assistance in preserving it, just as a gagged press or one run by people wishing to undermine the rule of law is a threat to it. But we need not be concerned here with these more indirect influences. 3. THE VALUE OF THE RULE OF LAW

One of the merits of the doctrine of the rule of law I am defending is that there are so many values it does not serve. Conformity to the rule of law is a virtue but only one of the many virtues a legal sy.stem should possess. This makes it all the more importan.t to be clear on the, values which the rule of law does serve. The rule of law is often rightly contrasted with arbitrary power Arbitrary power is broader than the rule of law. Many forms of arbitrary rule are compatible with the rule of law. A ruler can pro1 Similar lists of principles have been discussed by various authors English writers have been mesmerised by Dicey's unfortunate doctrine for too long. For a list similar to mine see Lon Fuller's The Morality of Law (2nd ed ), Chap 2 His discussion of many of the principles is full of good sense My main reason for abandoning some of his principles is due to a difference of views on conflicts between the laws of- one system

85

The Rule of Law and the Separation of Powers

APRIL 1977]

The Rule of Law and its Virtue

203

mote general rules based on whim or self interest, etc., without offending against the rule of law. But certainly many of the more common manifestations of arbitrary power run foul of the rule of law. A government subjected to the rule of law is prevented from changing the law retroactively or abruptly or secretly whenever this would suit its purposes. The one area where the rule of law excludes all forms of arbitrary power is in the law-applying function of the judiciary where the courts are required to be subject only to the law and to conform to fairly strict procedures. 8 No less important is the restraint imposed by the rule of law on the making of particular laws and thus on the powers of the executive. The arbitrary use of power for personal gain, out of vengeance or favouritism is most commonly manifested in the making of particular legal orders. These possibilities are drastically restricted by close adherence to the rule of law. " Arbitrary power " is a difficult notion. We have no cause to analyse it here. It seems, however, that an act which is the exercise of power is arbitrary only if it was done either with indifference as to whether it will serve the purposes which alone can justify use of that power or with belief that it will not serve them. The nature of the purposes alluded to varies with the nature of the power. This condition represents " arbitrary power " as a subjective concept. It all depends on ihe state of mind of the men in power. As such the rule of law does not bear directly on the extent of arbitrary power. But around its subjective core the notion of arbitrary power has grown a hard objective edge. Since it is universally believed that it is wrong to use public powers for private ends any such use is in itself an instance of arbitrary use of power. As we have seen the rule of law does help to curb such forms of arbitrary power. But there are more reasons for valuing the rule of law. We value the ahility to choose styles and forms of life, to fix long-term goals and effectively direct one's life towards them. One's ability to do so depends on the existence of stable, secure frameworks for one's life and actions. The law can help to secure such fixed points of reference in two ways: (1) by stabilising social relationships which but for the law may disintegrate or develop in erratic and unpredictable ways; (2) by a policy of self-restraint designed to make the law itself a stable and safe basis for individual planning. This last aspect is the concern of the rule of law. This second virtue of the rule of law is often, notably by Hayek, identified as the protection of individual freedom. This is right in s The rule of law itself does not exclude all the possibilities of arbitrary law-making by the courts.

The Rule of Law and the Separation of Powers

86

204

The Law Quarterly Review

[VoL 93

the sense of freedom in which it is identified with an effective ability to choose between as many options as possible. Predictability in one's environment does increase one's power of action. 9 If this is freedom well and good. The important thing is to remember that this sense of freedom differs from what is commonly meant by political freedom. Political freedom consists of: (1) the prohibition of certain forms of behaviour which interfere with perspnal freedom and (2) the limits imposed on the powers of public authorities in order to minimise interference with personal freedom. The criminal offences against the person are an example of the first mode of protecting personal freedom, the disability of the government to restrict freedom of movement-an example of the second. It is in connection with political freedom in this sense that constitutionally guaranteed rights are of great importance. The rule of law may be yet another mode of protecting personal freedom. But it has no bearing on the existence of spheres of activity free from governmental interference and is compatible with gross violations of human rights. More important than both these considerations is the fact that observance of the rule of law is necessary if the law is to respect human dignity. Respecting human dignity entails treating humans as persons capable of planning and plotting their future. Thus, respecting people's dignity includes respecting their autonomy, their right to control their future. A person's control over his life is never complete. It can be incomplete in any one of several respects. The person may be ignorant of his options, unable to decide what to do, incapable of realising his choices or frustrated in his attempts to do so, or he may have no choice at all (or at least none which is worth having). All these failures can occur through natural causes or through the limitations of the person's own character and abilities. Naturally, there are many ways in which one person's action may affect the life of another. Only some such interferences will amount to an offence to the dignity or a violation of the autonomy of the person thus affected. Such offences can be divided into three classes: insults, enslavement and manipulation. (I am using the last two terms in a somewhat special sense.) An insult offends a person's dignity if it consists of or implies a denial that he is an autonomous person or that he deserves to be treated as one. An action enslaves another if it intentionally practically denies him all options through the manipulation of the environment. (Though it may be for a 9 But then welfare law and governmental manipulation of the economy also increase freedom by increasing-if succes&ful-people's welfare. If the rule of law is defended as the bulwark of freedom in this sense, it can hardly be used to oppose in principle gvernmental management of the ecmony.

The Rule of Law and the Separation of Powers

A.PruL 1977]

The Rule of Law and its Virtue

87

205

length of time-as in real slavery-! mean to include here also coercing another to act in a certain way on a single occasion.) One manipulates a person by intentionally changing his tastes, his beliefs or his ability to act or decide. Manipulation-in other words-is manipulation of the person, of those factors relevant to his autonomy which are internal to him. Enslavement is the elimination of control by changing factors external to the person. The law can violate people's dignity in many ways. Observing the rule of law by no means guarantees that such violations do not occur. But it is clear that deliberate disregard for the rule of law violates human dignity. It is the business of law to guide human action by affecting people's options. But as we saw not every interference with the external circumstances of one's life, however unjustified, is a violation of one's dignity. Deliberate violation of the rule of law affects not only the external circumstances but also one's very ability to decide, act or form beliefs about the future. A legal system which does in general observe the rule of law treats people as persons at least in the sense that it attempts to guide their behaviour through affecting the circumstances of their action. It thus presupposes that they are rational autonomous creatures and attempts to affect their actions and habits by affecting their deliberations. (Though if the law institutes slavery or confers powers on officials enabling them to manipulate people, it can violate people's dignity even when observing the rule of law. It may also do so by sanctioning insults or in various other ways ) Violations of the rule of law affect one's fate by frustrating one's deliberations, by making it impossible for a person to plan his future to decide on his action on the basis of a rational assessment of their outcome. The rule of law provides the foundation for the legal respect for human dignity. 4. THE RULE oF LAw AND ITS EssENCE Lon Fuller 10 has claimed that the principles of the rule of law which he enumerated are essential for the existence of law. This claim if true is crucial to our understanding not only of the rule of law but also of the relation of law and morality. I have beeri treating the rule of law as an ideal, as a standard to which the law ought to conform but which it can and sometimes does violate most radically and systematically. Fuller, while allowing that deviations from the ideal of the rule of law can occur, denies that they can be radical 1o In The Morality of Law (2nd ed., Yale, 1969), Fuller's argument is complex and his claims are numerous and hard to disentangle Many of his claims are weak and unsupportable. Others are suggestive and useful It is not my purpose to analyse or evaluate them. For a sympathetic discussion see R. E Sartorius, Individual Conduct and Social Norms (Dickinson, 1975), Chap 9

The Rule of Law and the Separation of Powers

88

206

The Law Quarterly Review

[VoL. 93

or total. A legal system must of necessity conform to the rule of law to a certain degree, he claims. From this claim he concludes that there is an essential link between law and morality. Law is necessarily moral at least in some respects. It is, of course, true that most of the principles enumerated in section 2 above cannot be violated altogether by any l_egal system. 11 Legal systems are based on judicial institutions. There could not be institutions of any kind unless there are general rules set,ting them up. A particular norm can authorise adjudication in a particular dispute, but no number of particular norms can set up an institution. Similarly retroactive laws can exist only because there are institutions enforcing them. This entails that there must be prospective laws instructing those institutions to apply the retroactive laws if the retroactive laws are to be valid. In the terminology of H. L. A. Hart's theory one can say that at least some of the rules of recognition and of adjudication of every system must be general and prospective. Naturally they must also be relatively clear if they are to make any sense at all, etc. Clearly, the extent to which generality, ciarity, prospectivity, etc., are essential to the law is minimal and is consistent with gross violations of the rule of law. But are not considerations of the kind mentioned sufficient to establish that there is necessarily at least some moral value in every legal system? I think not. The rule of law is essentially a negative value. The law inevitably creates a great danger of arbitrary power-the rule of law is designed to minimise the danger created by the law itself. Similarly, the law may be unstable, obscure, retrospective, etc., and thus infringe people's freedom and dignity. The rule of law is designed to prevent this danger as well. Thus the rule of law is a negative virtue in two senses: conformity to it does not cause good except through avoiding evil and the evil which is avoided is evil which could only have been caused by the law itself. It is thus somewhat analogous to honesty when this virtue is narrowly interpreted as the avoidance of deceit. (I do not deny that honesty is normally conceived more broadly to incorporate other virtuous acts and inclinations.) The good of honesty does not include the good of communication between people for honesty is consistent with a refusal to communicate. Its good is exclusively in the avoidance of the harm of deceit-and not deceit by others but by the honest person himself. Therefore, only a person who can deceive 11 I am not adopting here Fuller's own conception of the law, but rather I am following my own adaptation of Hart's conception Cf Hart's The Concept of Law and my Practical Reason and Norms (Hutchinson, 1975), pp 132-154 Therefore, the discussion which follows is not a direct assessment of Fuller's own claims

89

The Rule of Law and the Separation of Powers APRIL

1977]

The Rule of Law and its Virtue

207

can be honest. A person who cannot communicate cannot claim any moral merit for being honest. A person who through ignorance or inability cannot kill another by poison deserves no credit for it. Similarly, that the law cannot sanction arbitrary force or violations of freedom and dignity through total absence of generality, prospectivity or clarity is no moral credit to the law. It only means that there are some kinds of evil which cannot be brought about by the law. But this is no virtue in the law just as it is p.o virtue in the law that it cannot rape or murder (all it can do is sanction such actions). Fuller's attempt to establish a necessary connection between law and morality fails. In so far as conformity to the rule of law is a moral virtue it is an ideal which should but may fail to become a reality. There is another argument, however, which establishes an essential connection between the law and the rule of law, though it does not guarantee any virtue to the law. Conformity to the rule of law is essential for securing whatever purposes the law is designed to achieve. This statement should be qualified. We could divide the purposes a law is intended to serve into two kinds: those which are secured by conformity with the law in itself and those further consequences of conformity with the law or of knowledge of its existence which the law is intended to secure. 12 Thus a law prohibiting racial discrimination in government employment has as its direct purpose the establishment of racial equality in the hiring, promotion and conditions of service of government employees (since discriminatory action is a breach of law) Its indirect purposes may well be to improve race relations in the country in general, prevent a threat of a strike by some trade unions or halt the decline in popularity of the government. Conformity to the rule of law does not always facilitate realisation of the indirect purposes of the law, but it is essential to the realisation of its direct purposes. These are achieved by conformity with the law which is secured (unless accidentally) by people taking note of the law and guiding themselves accordingly. Therefore, if the direct purposes of the law are not to be frustrated it must be capable of guiding human behaviour, and the more it conforms to the principles of the rule of law the better it can do so. In section 2 we saw that conformity to the rule of law is one among many moral virtues which the law should possess. The present consideration shows that the rule of law is not merely a moral virtueit is a necessary condition for the law to be serving directly any good 12 See further on this distinction J Raz, "The Functions of Law," Oxford Essays in Jurisprudence (2nd series, ed. A, W B Simpson, Oxford, 1973).

The Rule of Law and the Separation of Powers

90

208

The Law Quarterly Review

[VOL.

93

purpose at all. Of course, conformity to the rule of law also enables the law to serve bad purposes. That does not show that it is not a virtue, just as the fact that a sharp knife can be used to harm does not show that being sharp is not a good-making characteristic for knives. At most it shows that from the point of view 'Of the present consideration it is not a moral good. Being sharp is an inherent goodmaking characteristic of knives. A good knife is, among other; things. a sharp knife. Similarly conformity to the ·rule of law is an inherent value of laws, indeed it is their most important inherent value. It is of the essence of Jaw ,to guide behaviour through rules and courts in charge of their application. Therefore, the rule of law is the specific excellence of the law. Since conformity to the rule of law is the virtue of law in itself, law as law regardless of the purposes it serves, it is understandable and right that the rule of law is thought of as among the few virtues of law which are the special responsibility of the courts and the legal profession. Regarding the rule of law as the inherent or specific virtue of law is a result of an instrumental conception of law. The law is not just a fact of life. It is a form of social organisation which should be used properly and for the proper ends. It is a tool in the hands of men differing from many others in being versatile and capable of being used for a large variety of proper purposes. As with other tools, machines and instruments a thing is not of the kind unless it has at least some ability to perform its function. A knife is not a knife unless it has some ability to cut. The law to be law must be capable of guiding behaviour, however inefficiently. Like other instruments, the law has a specific virtue which is morally neutral in being neutral as to the end to which the instrument is put. It is the virtue of efficiency; the virtue of the instrument as an instrument. For the law this virtue is the rule of law. Thus the n~le of law is an inherent virtue of the law, but not a moral virtue as such. The special status of the rule of law does not mean that conformity with it is of no moral importance. Quite apart from the fact that conformity to the rule of law is also a moral virtue, it is a moral requirement when necessary to enable the law to perform useful social functions; just as it may be of moral importance to produce a sharp knife when it is required foi" a moral purpose. In the case of the rule of law this means that it is virtually always of great moral value. 5. SoME PITFALLs

The undoubted value of conformity to the rule of law should not lead one to exaggerate its importance. We saw how Hayek noted correctly its relevance for the protection of freedom. We also saw

91

The Rule of Law and the Separation of Powers APRIL

1977)

The Rule of Law and its Virtue

209

that the rule of law itself does not provide sufficient protection of freedom. Consider, however, Hayek's position. He begins with a grand statement which inevitably leads to exaggerated expectations. The conception of freedom under the law that is the chief concern of this book rests on the contention that when we obey laws, in the sense of general abstract rules laid down irrespective of their application to us we are not subject to another man's will and are therefore free. It is because the lawgiver does not know the particular cases to which his rules will apply, and it is because the judge who applies them has no choice in drawing the conclusions that follow from the existing body of rules and the particular facts of the case, that it can be said that laws and not men rule. . . . As a true law should not name any particulars, so it should especially not single out any specific persons or group of persons. 13 Then, aware of the absurdity to which this passage leads he modifies his line, still trying to present the rule of law as the supreme guarantor of freedom: The requirement that the rules of true law be general does not mean that sometimes special rules may not apply to different classes of people if they refer to properties that only some people possess. There may be rules that can apply only to women or to the blind or to persons above a certain age. (In most instances it would not even be necessary to name the class of people to whom the rule applies: only a woman, for example, can be raped or got with child.) Such distinctions will not be arbitrary, will not subject one group to the will of others, if they are equally recognized as justified by those inside and those outside the group. This does not mean that there must be unanimity as to the desirability of the distinction, but merely that individual views will not depend on whether the individual is in the group or not. 14 But here the rule of law is transformed to encompass a form of government by consent and it is this which is alleged to guarantee freedom. This is the slippery slope leading to the identification of the rule of law with the rule of the good law. Ha:Yek's main objection is to governmental interference with the economy: We must now turn to the kinds of governmental measures which the rule of law excludes in principle because they cannot be achieved by merely enforcing general rules but, of necessity involve arbitrary discrimination between persons. The most 13

F A Hayek, The Constitution of Liberty (1960), pp 153-154.

a Ibid. p. 154.

The Rule of Law and the Separation of Powers

92

The Law Quarterly Review

210

[VoL. 93

important among them are decisions as to who is to be allowed to provide different services or commodities, at what prices or in what quantities-in other words, measures designed to control the access to different trades and occupations, the terms of sale, and the amounts to be produced or sold. There are several reasons why all direct control of prices by government is irreconcilable with a functioning free system, whether the government actually fixes prices or merely lays down rules by which the permissible prices are to be determined. In the first place, it is impossible to fix prices according to long-term rules which will effectively guide production. Appropriate prices depend on circumstances which are constantly changing and must be continually adjusted to them. On the other hand, prices which are not fixed outright but determined by some rule (such as that they must be in a certain relation to cost) will not be the same for all sellers and, for this reason will prevent the market from functioning. A still more important consideration is that, with prices different from those that would form on a free market, demand and supply will not be equal, and if the price control is to be effective, some method must be found for deciding who is to be allowed to buy or sell. This would necessarily be discretionary and must consist of ad hoc decisions that discriminate between persons on essentially arbitrary grounds. 15 Here again it is clear that arguments which at best show that certain policies are wrong for economic reasons are claimed to show that they infringe the rule of law and the making of supposedly misguided but perfectly principled particular orders is condemned as an arbitrary exercise of power. Since the rule of law is just one of the virtues the law should possess, it is to be expected that it possesses no more than prima facie force. It has always to be balanced against competing claims of other values. Hence Hayek's arguments to the extent that they show no more than that some other goals inevitably conflict with the rule of law are not the sort of arguments which could, in principle, show that pursuit of such goals by means of law is inappropriate. Conflict between the rule of law and other values is just what is to be expected. Conformity to the rule of law is a matter of degree, and though other things being equal, the greater the conformity the better-other things are rarely equal. A lesser degree of conformity is often to be preferred precisely because it helps realisation of other goals. In considering the relation between the rule of law and other values the law s.hould serve, it is of particular importance to 15

Ibid pp 227-228

The Rule of Law and the Separation of Powers APRIL

1977]

The Rule of Law and its Virtue

93

211

remember that the rule of law is essentially a negative value. It is merely designed to minimise the harrtl to freedom and dignity which the law may cause in its pursuit of its goals however laudable these may be. Finally regarding the rule of law as the inherent excellence of the law means that it fulfils essentially a subservient role. Conformity to it makes the law a good instrument for achieving certain goals, but conformity to the rule of law is not itself an ultimate goal. This subservient role of the doctrine shows both its power and its limitations. On the one hand if the pursuit of certain goals is entirely incompatible with the rule of law then these goals should not be pursued by legal means. But on the other hand one should be wary of disqualifying the legal pursuit of major social goals in the name of the rule of law. After all the rule of law is meant to enable the law to prom0te social good, and should not be lightly used to show that it should not do so. Sacrificing too many social goals on the altar of the rule of law may make the law barren and empty. JOSEPH RAZ. * • D Phil , Fellow of Balliol College, Oxford

[4]

Formal and Substantive Conceptions of the Rule of Law: An Analytical Framework Paul Craig

Professor of Law, Worcester College, Oxford

There is a voluminous literature on the rule of law which exammes the concept from almost every conceivable perspective. The analysis which follows makes no pretence at being a complete survey of these differing approaches. It does however attempt to address the subject in a way that is both important for public lawyers, and of broader significance outside of any particular legal system. The central theme of the article is the distinction between formal and substantive meanings of the rule oflaw. This dichotomy is, as will be .seen below, of crucial importance in determining the nature of the specific legal precepts which can be derived from the rule of law. The difference between these conceptions of the rule of law will be explored fully below, but the essence of the distinction can be conveyed here. Formal conceptions of the rule oflaw address the manner in which the law was promulgated (was it by a properly authorised person, in a properly authorised manner, etc.); the clarity of the ensuing norm (was it sufficiently clear to guide an individual's conduct so as to enable a person to plan his or her life, etc.); and the temporal dimension of the enacted norm, (was it prospective or retrospective, etc.). Formal conceptions of the rule oflaw do not however seek to pass judgment upon the actual content of the law itself. They are not concerned with whether the law was in that sense a good law or a bad law, provided that the formal precepts of the rule of law were themselves met. Those who espouse substantive conceptions of the rule of law seek to go beyond this. They accept that the rule of law has the formal attributes mentioned above, but they wish to take the doctrine further. Certain substantive rights are said to be based on, or derived from, the rule oflaw. The concept is used as the foundation for these rights, which are then used to distinguish between "good" laws, which comply with such rights, and "bad" laws which do not. The structure of the analysis will be as follows. The first part of the article will consider the formal conception of the rule of law. This part of the argument will be divided into three sections. There will be an examination of the work ofRaz who articulates the formal conception of the rule oflaw most clearly and explicitly. This will be followed by an analysis ofDicey's conception

The Rule of Law and the Separation of Powers

96 468

Public Law

of the rule of law. It will be argued that he too was a formalist. In the final section of this part of the article Unger's challenge to the formal conception of the rule of law will be considered. The second part of the article will focus upon a thoroughgoing substantive account of the rule of law provided by Dworkin. The implications of adopting such a conception of the rule of law will be brought out, and will be exemplified through consideration of the work of Sir John Laws and Trevor Allan. The third and final part of the article will consider whether there is some middle way between the adoption of a purely formal conception of the rule of law and the fully substantive version of the doctrine.

1. The formal conception of the rule of law

(a) joseph Raz

It may be helpful at the outset to make clear why those who subscribe to the formal conception of the rule of law insist that the concept should bear this meaning. Raz provides the clearest explanation.

If the rule of law is the rule of the good law then to explain its nature is to propound a complete social philosophy. But if so the term lacks any useful function. We have no need to be converted to the rule of law just in order to discover that to believe in it is to believe that good should triumph. The rule oflaw is a political ideal which a legal system may lack or possess to a greater or lesser degree. That much is common ground. It is also to be insisted that the rule oflaw is just one of the virtues by which a legal system may be judged and by which it is to be judged. It is not to be confused with democracy, justice, equality (before the law or otherwise), human rights of any kind or respect for persons or for the dignity ofman. 1 What Raz is getting at here can be explained quite straightforwardly. We may all agree that laws should be just, that their content should be morally sound and that substantive rights should be protected within society. The problem is that if the rule oflaw is taken to encompass the necessity for "good laws" in this sense then the concept ceases to have any useful independent function for the following reason. There is a wealth ofliterature devoted to the discussion of the meaning of a just society, the nature of the rights which should subsist therein, and the appropriate boundaries of governmental action. Political theory has tackled questions such as these from time immemorial. To bring these issues within the rubric of the rule oflaw would therefore have the effect of robbing this concept of any function independent of such political theories. Laws would be condemned or upheld as being in conformity with, or contrary to, the rule oflaw in this substantive sense when the condemnation or praise would simply be reflective of attachment to one particular political theory. The message which Raz conveys is an important one. If you wish to argue about the justness of society do so by all means. If you wish to defend a ' "The Rule of Law and its Virtue" (1977) 93 L.Q.R. 195 at 196.

The Rule of Law and the Separation of Powers Formal & Substantive Conceptions of the Rule of Law

97 469

particular type of individual right then present your argument. Draw upon the wealth ofliterature which addresses these matters directly. Nothing however is to be gained by cloaking whatever conclusion you reach in the mantle of the rule of law, since this merely reflects the conclusion which has already been arrived at through the relevant political theory. It is for this reason that Raz insists that the rule of law should be seen in formal terms. The consequence of this reading is, as Raz readily admits, that the rule oflaw could be met by regimes whose laws are morally objectionable, provided that they comply with the formal precepts which comprise the rule of law. It is equally the case, on this view, that a democratic regime will not necessarily always have laws which do measure up to the rule of law. What then is the proper remit of the rule of law viewed in this formal manner? Raz makes it clear that it cannot just mean that government action is authorised by law since the concept would then be thin indeed. Any law properly passed by Parliament would meet the rule of law defined in this manner. That laws should be passed in the correct legal manner is none the less a necessary facet of a formal conception of the rule of law. It is not however sufficient. The other important aspect of the rule of law is that the laws thus promulgated should be capable of guiding one's conduct in order that one can plan one's life. It is from this general precept that Raz then deduces a number of more specific attributes that laws should have in order that they could be said to be in compliance with the rule oflaw. All are related to this idea of enabling individuals to be able to plan their lives. The "list" includes the following: that laws should be prospective, not retrospective; that they should be relatively stable; that particular laws should be guided by open, general and clear rules; that there should be an independent judiciary; that there should be access to the courts; and that the discretion which law enforcement agencies posses~ should not be allowed to undermine the purposes of the relevant legal rules. On this view the rule of law is essentially a negative value as Raz himself admits. Given that the law can empower the state to do all manner of things the rule of law minimises the danger created by the law itself.· It does so by ensuring that whatever the content of the law, at least it should be open, clear, stable, general and applied by an impartial judiciary. It would however be mistaken not to recognise the more positive side of the rule of law when viewed in this manner. Even if the actual content of the law is morally reprehensible, conformity to the rule of law will often be necessary to ensure that individuals actually comply with the demands which the law imposes. One final point which is of importance concerning this conception of the rule oflaw is that, as Raz emphasises, it is only one virtue of a legal system, and may have to be sacrificed to attain other desired ends. We may feel that the rule of law virtues of having clear, general norms must be sacrificed if the best or only way to achieve a desired goal is to have more discretionary, open textured legal provisions. This may be the case in circumstances when it is not possible to lay down in advance in the enabling legalisation clear, prospective rules in sufficient detail to cover all eventualities. Modifications to the rule of law in this manner are not somehow forbidden or proscribed. Given that it is only one

The Rule of Law and the Separation of Powers

98 470

Public Law

virtue of a legal system it should not prevent the attainment of other virtues valued by that system.

(b) Dicey

Dicey's conception of the rule oflaw2 is well known and it has been subjected to analysis from all of the diverse directions set out above. The focus of the discussion which follows will be upon the formal/substantive divide and the way in which this facilitates our understanding of his reasoning. Dicey's first limb of the rule of law was that: ... no man is punishable or can be lawfully made to suffer in body or goods except for a distinct breach oflaw established in the ordinary legal manner before the ordinary courts of the land. In this sense the rule oflaw is contrasted with every system of government based on the exercise by persons in authority of wide, arbitrary, or discretionary powers of constraint. 3 There are a number of well known critiques of this principle. One of the most important was that Dicey underestimated both the existence of discretionary power which existed at the time when he was writing; and also the fact that such discretionary power was often a necessary and legitimate consequence of the growth of governmental power in the nineteenth century. To return to the focus of the current discussion, the key question is whether Dicey's first principle is to be perceived in formal or substantive terms? Now there is no doubt that the words used by Dicey could bear a substantive meaning. This is particularly true of the word "arbitrary", and some have suggested that Dicey's vision of the rule of law should be viewed in this manner. 4 The weight· of evidence is, however, clearly against this reading of Dicey's work: his first principle is formalistic and not substantive. This is readily apparent from the first sentence of the above formulation. This requires that laws under which people are condemned should be passed in the correct legal manner and that guilt should only be established through the ordinary trial process. Nothing here speaks to the content of the laws which an individual will have to face when taken before the courts. But what then of the remainder of the first principle? Does this not have a substantive content? Would not laws which are "bad" or "evil" be labelled as arbitrary within the meaning of Dicey's first principle? If not, what then does this word connote? Now, as stated above, it would of course be possible for the word arbitrary to have a substantive content. On this view a law which was properly enacted by Parliament, in compliance with all correct procedures, which was pristinely clear in its application, and which was applied by an impartial judiciary, might

The Law if the Constitution (lOth ed., 1959). ibid., p. 188. 4 See, e.g. T.R.S. Allan, Law, Liberty and justice, The Legal Foundations (1993), p. 46. 2

3

if British

Constitutionalism

The Rule of Law and the Separation of Powers Formal & Substantive Conceptions of the Rule ofLaw

99 471

none the less be tainted as arbitrary if it was thought to infringe certain fundamental rights, or if it entailed excessive punishment. It is equally clear that the word arbitrary can have a formal meaning. When used in this latter sense the word arbitrary would provide the foundation for criticism of two kinds of norm. One category would comprise those allegedly legal rules which, when examined, do not in fact have any legal foundation. They might not have been enacted in the proper manner because, for example, they have not been passed through Parliament and do not come within the ambit of the prerogative. The other category of formal arbitrariness would be used to describe those norms which have been passed in the correct legal manner, but where the resulting law was impossibly vague or unclear, with the result that individuals had no idea how to plan their lives in the light of the relevant legal rule. Formal arbitrariness in either of these senses is independent of whether the content of the legislation was good or bad, just or unjust. So which of these two senses of arbitrary did Dicey have in mind when formulating his rule of law? Two arguments, one positive, the other negative, point strongly to the fact that he was using the term in the latter, formalistic sense. The positive argument is to be found in Dicey's own discussion within later sections of the LAw of the Constitution. When discussing freedom of the individual Dicey contrasts continental systems with that in England. He claims that the former were not free from arbitrary power. For Dicey, the Bastille was the "visible sign oflawless power", even though it had only a handful of people in it when it fell. This was because it was a symbol of arbitrary power, in the sense that the executive would incarcerate people there without any lawful authority, or for the commission of crimes which were impossibly vague. Dicey spends two pages lamenting the fate of poor Voltaire who was twice placed in the Bastille at royal or aristocratic whim. 5 In England, by way of contrast, the singularity of our law was not so much its leniency or goodness, but its legality. 6 Although we might have had harsh laws, a person's fate was not dependent upon the caprice of some other person who might happen to have power. Thus Dicey was under no illusion that all English laws were substantively just; nor does he attempt to claim that they were. His conclusion that England was not subject to arbitrary power, and that in this respect we fared better than those on the continent, was based on the formalistic sense of the term arbitrary considered above. The laws might have been harsh, but they had to be properly passed, and applied by the ordinary courts, before an individual could lose his or her freedom. Moreover, Dicey's discussion of the relationship between sovereignty and the rule of law further reinforces the view that his conception of the latter was formal and not substantive. 7 The negative argument which points to the same conclusion is that if Dicey had intended the term arbitrary to bear a substantive connotation then he provided absolutely no criterion as to how this sense of arbitrariness was to be 5 Dicey, op. cit., n. 2, pp. 189-191. " ibid., pp. 267-268. 7 ibid., Chap. 13.

The Rule of Law and the Separation of Powers

100 472

Public Law

determined. We shall see in the discussion which follows the difficulties which have to be faced if one wishes to adopt a substantive conception of the rule of law. Suffice it to say for the present that such a view of the rule of law necessitates the articulation of some criterion which will then provide the foundation for the conclusion that a particular law really was "unjust" or "bad". Now Dicey did of course have strong political views, as is well known. Yet at no stage is there any evidence to suggest that he intended these political and moral precepts to be used to determine that a properly enacted law which was clear, and applied by an impartial judiciary, could none the less be regarded as substantively arbitrary, and hence contrary to the rule oflaw, on the grounds that it infringed these or any other such precepts. The secoud principle of the rule of law concerns equality. Dicey's formulation of the principle is as follows. We mean ... when we speak of the 'rule oflaw' as a characteristic of our country, not only that with us no man is above the law, but (what is a different thing) that here every man, whatever be his rank or condition, is subject to the ordinary law of the realm and amenable to the jurisdiction of the ordinary tribunals. 8 Once again there are well known critiques of Dicey's second principle. His misunderstanding of the French droit administratif was legendary, as was his misapprehension of how much administrative law existed in nineteenthcentury England, with adjudication through specialist tribunals rather than the ordinary courts. 9 Our primary concern is, however, as to whether this second principle is formalistic or substantive. As with the first of Dicey's principles, so here too it will be argued that the weight of evidence clearly indicates that Dicey was thinking in formal rather than substantive terms when formulating his ideas about equality. Dicey's formulation is concerned primarily with formal access to the courts, riot with the nature of the rules which individuals will find when they get there. This point is captured well by Marshall: Equality before the law, understood as the equal subjection of all classes to a common rule, might at least be contrasted significantly with chaos or lawlessness, but it does not in itself imply any qualitative view about the sort of law to which all are subject. 10 Now to be sure it is true that Dicey was explicitly against officials being accorded any special privileges over and beyond those of ordinary citizens, and in this sense Dicey imported a substantive element into this aspect .of his rule oflaw. But beyond this Dicey's second principle does not touch on substantive equality at all. As Marshall states, speaking of this part of Dicey's analysis: " ibid., p. 193. " See generally, H.W Arthurs, "Without the Law", Administrative justice and Legal Pluralism in Nineteenth Century England (1985). '"G. Marshall, Constillltional Theory (1971), p. 137.

The Rule of Law and the Separation of Powers Formal & Substantive Conceptions of the Rule ofLaw

101

473

It omits, however, to register the truism that the law which all citizens find when they get to the common courts may make unequal provision for some as against others. The same law that bound all could say that the Crown could not be sued, and that policeman and state officials should have powers, privileges, or legal defences not open to private citizens. 11

A substantive conception of equality would require the articulation of principles through which the 'courts would then determine whether the application of one rule to Group A was compatible with the application of a different rule to Group B. Legal systems use varying criteria to resolve questions of this kind. Issues of considerable complexity are involved as courts attempt to decide whether, for example, the division between two groups as to the content of the rules which they face is based on some rational, intelligible difference between them. Any thoroughgoing theory of substantive equality will moreover be based, implicitly, if not explicitly, on some broader background political theory of which it forms but one important part. There is no indication that Dicey in his second rule intended to grapple with such matters, nor that he intended the second limb of the rule of law to have this type of substantive content. 12 · Dicey's third limb of the rule oflaw does not sit easily with the previous two. The essence of this precept can be stated as follows. We may say that the constitution is pervaded by the rule of law on the ground that the general principles of the constitution (as for example the right to personal liberty, or the right of public meeting) are with us the result of judicial decisions determining the rights of private persons in: particular cases brought before the courts; whereas under many foreign constitutions the security (such as it is) given to the rights of individuals appears to result from the general principles of the constitution. 13 This limb of the rule of law has caused considerable confusion. One common error is to read this aspect of the rule of law as demanding that a society must indeed possess certain individual rights if it is to conform to the rule oflaw. If this view were correct then Dicey would indeed be imbuing the rule of law with a substantive content. The reason why this is erroneous is quite simply that it is not what Dicey actually said. He might have said it. He might have said a number of things; but he did not say this. What Dicey actually said was crucially different. His argument was not that the rule of law demanded adherence to certain specific substantive rights. It was that if you wished to protect such rights then the common law technique was better than that employed on the continent. This is manifestly clear from Dicey's own formulation of the third principle of the ibid., pp. 138-139. Indeed, leaving aside the specific issue of the special powers allegedly possessed by public officials, there is little evidence that Dicey was in fact thinking of, or that he recognised, the real differences in the powers of particular groups within society, such as the police, diplomats or those who operated utilities. 13 Op. cit., n. 2, pp. 195-196, 203. II

12

102

The Rule of Law and the Separation of Powers 474

Public Law

rule of law, and from the ensuing discussion. In this discussion Dicey argued that the protection of rights on. the continent through Bills of Rights was ineffective, since such constitutional documents could so easily be abrogated at the stroke of a pen. Under the common law, where individual rights were the result of numerous judicial decisions indicating when the individual was at liberty to speak freely etc., it would be considerably more difficult for some authoritarian regime to sweep these rights aside. 14 Now this argument may or may not be true on its merits. Even if it did have some empirical validity when Dicey wrote, one might argue that matters are in any event different now and that constitutionally enshrined protections for rights would serve the individual better than the traditional common law methodology. But this species of argument is irrelevant to the point made above, which is that Dicey's third limb of the rule of law is no more substantive than the previous two. It no more demands the existence of certain specific substantive rights than do the earlier limbs of his formulation.

(c) Unger There have been frequent challenges to the rule of law. One of the most interesting is that advanced by Unger who argues, in effect, both that the formal conception of the rule of law was always a mask for substantive inequalities in power, and that in the modern day this formal conception is in any event increasingly unattainable. 15 The contention is that the rule of law provided a convenient legitimating mask for substantive inequality within liberal society. For Unger a liberal society was one in which there were many different groups, with no particular group being able to dominate the whole. There was no preordained caste system, nor was there any fixed hierarchy of the kind which prevailed during the medieval period. In liberal society a justification was required for the way in which society was ordered. In Unger's view the rule of law emerged to provide this justification. How did it perform this function? . For Unger it achieved this goal by making it appear that power was impersonal. His vision of the rule of law is not markedly different from that espoused by Raz and Dicey. In Unger's terms it enshrines commitment to the generality, autonomy, neutrality and predictability of legal norms. These ideas capture the same themes as those articulated in the earlier discussion. For Unger it was these very formal attributes which played such a large part in legitimating the existing power structures within society, by making it appear that power was impersonal. Unger's contention is that this legitimating function performed by the rule of law was always really a sham. There were two reasons why this was so. On the one hand, one of the premises underlying this legitimating function performed by the rule of law was that most power was concentrated in government. Yet he argues that in practical terms considerable power lay in other places, including the workplace, the family etc. Inequality in these areas

a

H

15

ibid., pp. 200-202. LAw in Modern Society (1976), pp. 176-181, 192-223.

The Rule of Law and the Separation of Powers Formal&: Substantive Conceptions of the Rule ofLaw

103 475

was not touched by a commitment to formal equality within the legal arena. On the other hand, and more importantly, Unger claims that the assumption behind the rule oflaw rhetoric was that power could be effectively constrained by rules. If rules were indeed general and impartially applied then it would be difficult for the ruler, or a particular class, to turn them to personal advantage. He contends that this assumption was not sustainable. This was in part because even if rules were general their content would reflect the power of the dominant class. It was in part also because even if a rule was general it would still have to be applied by the judiciary, and this could not be done in a manner that was truly value neutral. Notwithstanding these problems the rule of law could still appear to be a legitimating device within Unger's liberal society. Things are, he claims, different in a post-liberal society, in that the very aims of the rule oflaw itself, in terms of the generality, neutrality, had autonomy of law, are undermined. A post-liberal society as described by Unger has two characteristics. One is increased governmental involvement in the economy and in the regulation of soci~ty in general. The government is forced to intervene in more and more areas, from social welfare to planning, and from utilities regulation to health and safety, in order to ameliorate the problems which flow from unjustified hierarchy. The other related characteristic is that the boundaries between the public and private spheres become eroded. The idea that the government is merely a neutral guardian of the social order comes under increased strain. These changes in the nature of the societal order have ramifications for the type of legal norms which emerge. Legislation becomes more open-textured and is framed in broader, less precise terms. This is because the complex aims which government now seems to achieve can not be attained through clear and precise rules. More open-ended discretion has to be left to administrative agencies and to the courts. The style of legal reasoning alters. It ceases to be formal and becomes more purposive. Formal legal reasoning could be used when there were clear, general rules which were capable of mechanical application. The nature of the statutes which have to be interpreted in a postliberal society pushes the judiciary towards a more purposive style of reasoning, in which increased attention is placed on the ends which the legislation is intended to serve in order to determine its actual remit. Purposive legal reasoning tends to place a higher premium on the substantive justice of the outcome, and not to focus exclusively on concerns with formal justice. These characteristics of a post-liberal society are said by Unger to have profound implications for the traditional attributes of the rule of law. The generality of law is undermined. This is in part because the complex problems which have to be dealt with often render the formulation of general rules impossible. It is in part because the increased attention which is now placed on substantive as opposed to formal justice means that we might wish to have more particular rules which differentiate between groups to a greater degree. Autonomy is also said to be undermined. Courts will now be forced to apply open-textured, often vague, statutes, which leave many issues unclear. The judiciary will then be placed in a position where they have to weigh a

The Rule of Law and the Separation of Powers

104 476

Public Law

wide variety of factors. Their judgments will come to resemble more closely those which are made in the political forum, or by administrators. People will undoubtedly have different views on Unger's provocative analysis. Some might agree wholeheartedly with it, and see it as a timely unmasking or deconstructing of a "revered" legal myth. Others might regard it as an exercise in historical sociology which is insufficiently grounded in empirical evidence. Yet others might adopt some intermediate position, including the present writer. Exigencies of space preclude any detailed analysis of Unger's thesis, but three related comments may be of help in locating this thesis within the more general structure of this article. The first is to reiterate a point made earlier. Unger's view of what the rule of law actually means does not differ markedly from that of Raz or Dicey. He too adopts a formal version of the rule of law, as expressed in the ideas of generality, neutrality, clarity and autonomy. Tlie second comment concerns Unger's thesis about the rule oflaw within what he terms liberal society. This aspect of Unger's thesis has two related components: that the concept was employed as a device to legitimate power inequalities and that it could never properly fulfil this goal. Even if we accept the first component of this argument, the second is more open to question. We have already seen that one of the principal reasons why Unger claims that the objective for which the rule of law was being employed could n.ever be properly achieved was because of substantive inequalities in power. This meant that the content of the resulting norms would be weighted in favour of the dominant grouping within society. Let us assume that this was indeed the case, although such an assumption is contestable. There is none the less an element of circularity in the argument. The rule of law as used by Unger is a formal concept. Adherence to the concept has never been claimed to guarantee a just society, if that phrase is used to connote a society in which the substantive distribution of wealth and power is morally acceptable. Nor has the formal concept of the rule oflaw ever pretended to be a guarantee that the substantive content of particular laws will be just, in the sense of preventing any form of bias within the law for a dominant power grouping. To claim therefore that any legitimating function performed by the rule of law within liberal society was undermined because of substantive power inequalities is to condemn the rule of law for not combating issues which it, as a legal concept, never claimed to be redressing. The third comment relates to the decline of the rule of law which Unger perceives within what he terms post-liberal society. This third comment has both an empirical and a conceptual dimension. In empirical terms one might simply question how serious the problem actually is. How many statutes do actually take the open-textured, vague form which he identifies? How often are the judiciary in fact forced into making legal judgments which are said to replicate closely the type ofbalancing process which legislatures themselves have to undertake? Are all laws now like this? There are also interesting conceptual issues raised by Unger's analysis. One of the most important may be put in the following terms. We have already seen

The Rule of Law and the Separation of Powers Formal & Substantive Conceptions of the Rule of Law

105

477

that writers such as Raz emphasise that the rule oflaw is but one virtue which legal systems should possess. It may well have to be sacrificed if we wish to reach certain ends, the attainment of which is not possible while still adhering to the formal rule oflaw precepts. Viewed in this way the developments which Unger identifies within post-liberal society could be understood simply as instances in which society has decided that the pursuit of other virtues, such as help for particular disadvantaged groups, necessitates the sacrifice of formal rule of law values.

2. The substantive conception of the rule oflaw

(a) Dworkin

It is not fortuitous or surprising that one of the principal advocates of the formal conception of the rule of law, Raz, is also a leading exponent of legal positivism. The formal conception of the rule of law, and the desire to keep legal questions separate from broader issues of political theory in deciding what the content of the law actually is, fit naturally together. The view oflaw and adjudication espoused by Dworkin is very different. It is central to this thesis that, subject to questions of fit, the courts should be deciding legal questions according to the best theory ofjustice. 16 On this view broader questions of political theory are central to the resolution of what rights people currently possess. Given this theory, it would be odd, to say the least, to conceive of the rule oflaw in purely formal terms. That this is indeed so can be seen by considering Dworkin's own discussion. 17 He distinguishes between two different conceptions of the. rule of law. The first, which he terms the rule book conception, is in effect a version of the formal rule oflaw discussed above. It says nothing about the content of the laws which exist within a legal system, but merely insists that the government should never exercise power against individuals except in accordance with rules which have been set out in advance and made available to all. As Dworkin recognises, those who have this conception of the rule of law care about the content of the law, "but they say that this is a matter of substantive justice, and that substantive justice is an independent ideal, in no sense part of the ideal of the rule oflaw". 18 The second conception of the rule of law is termed by Dworkin the rights conception. He defines it in the following manner:

It assumes that citizens have moral rights and duties with respect to one another, and political rights against the state as a whole. It insists that these moral and political rights be recognized in positive law, so that they may be enforced upon the demand 2-3. Bell, Policy Arguments in Judicial Decisions (1983).

J.

The Rule of Law and the Separation of Powers Formal & Substantive Conceptions of the Rule of Law

111 483

courts can reason in such a manner. Quite the contrary. For positivists, such as Raz, courts should reason in this fashion when they are faced with cases for which the existing source-based law provides no answer: According to [the sources thesis], the law on a question is settled when legally binding sources provide its solution. In such cases judges are typically said to· apply the law, and since it is source-based, its application involves technical, legal skills in reasoning from those sources and does not call for moral acumen. If a legal question is not answered from legal sources then it lacks a legal answer the law on the question is unsettled. In deciding such cases courts inevitably break new (legal) ground and their decision develops the law (at least in precedent-based legal systems). Naturally, their decisions in such cases rely at least partly on moral and other extra-legal considerations. 40 Raz does, moreover, go considerably further in articulating a view as to the nature of common law adjudication. He distinguishes between the role of the courts in what he terms regulated cases, those which fall under a common law or statutory rule which does not require judicial discretion for the determination of the dispute, and unregulated cases, where there is some gap in the law applicable to the case. 41 The latter include cases where there is some indeterminacy of language or intention, or those where there are two conflicting rules potentially applicable to the case. 42 On this view courts are regarded as making law in cases of unregulated disputes; 43 when they do so they should adopt those rules which they believe to be best, in the same manner as a legislator; this may well entail taking into account moral considerations; there may, nonetheless, be constraints which cause courts to be less adventurous than legislators; the courts' law making function is not dependent upon the courts necessarily realising that this is what they were doing; and law application and law making may both be present within a particular case. 44 The purpose of this part of the discussion is not to demonstrate that the positivist view of adjudication is necessarily correct. The object is much more limited. It is simply to show that the examples which Allan does proffer simply do not prove the point for which he argues. The fact that one can point to cases where courts have adverted to considerations of justice, fairness or morality may be perfectly reconcilable with positivist legal theory. They may simply be

Raz, The Authority ~fLAw, Essays on LAw and Morality (1979), pp. 49-50. ibid., p. 181. 42 ibid., pp. 193-194. 43 And indeed in some cases of regulated disputes. 44 ibid., Chap. 10. This is necessarily the barest of outlines of what is a complex argument. For further discussion see, Raz, Ethics in the Public Domain, Essays on the Morality of L.aw and Politics (1994), Chaps. 10, 13. There is also a considerable debate within positivism about the nature of any connection between law and morality, see, e.g. Coleman, "Negative and Positive Positivism" 11 journal ~f Legal Studies 139 (1982); Soper, "Legal Theory and the Obligation of the Judge: The Hart/Dworkin Dispute" 75 Mich.L.Rev. 511 (1977). 4 "]. 41

112

The Rule of Law and the Separation of Powers 484

Public Law

regarded as cases where the courts have engaged in lawmaking in the positivist sense of that term. 45

3. A middle way? (a) Raz Given the preceding analysis it might well be thought that it is not possible for there to be a middle way between the formal and substantive visions of the rule of law. This may well be so. Before reaching any conclusion on this issue we should, however, consider the views of Joseph Raz who has returned to the topic in his more recent work. 46 His analysis concentrates upon the role of the rule oflaw within Britain, rather than focusing upon the more universal aspects of the concept as he had done within his earlier work. According to Raz the core idea is the "principled faithful application of the law". 47 The major features are "its insistence on an open, public administration of justice, with reasoned decisions by an independent judiciary, based on publicly promulgated, prospective, principled legislation". 48 The principle of the rule oflaw is addressed to the courts, legislature, and also other bodies such as the police and administrative authorities. While this vision of the rule oflaw requires the courts to be faithful to legislation which emerges from a democratic legislature, it also sets limits to majoritarian democracy, by requiring principled as well as faithful adjudication. 49 The core of the thesis is to be found in this very notion of principled adjudication. It requires that the courts make decisions which are reasoned and public. But the real nub of the idea is captured in the following quotation: In insisting that judicial decisions should be not only faithful but also principled, I am suggesting that the function of the rule of law is to facilitate the integration of particular pieces of legislation with the underlying doctrines of the legal system ... A particular reform of police powers to search for prohibited drugs ... should be applied in a manner which is both faithful to the legislative purpose and principled in integrating it with traditional doctrines of the liberties of the citizen. 50 This aspect of the judicial function is justified in part by the need to ensure that there is a coherence of purpose within the law. 5 1 It is also held to be 45 Nor can this point be met simply by a blanket rejection of positivism, cf Allan, op. cit., n. 4, p. 28. One may, of course, prefer a Dworkinian view of the adjudicative process and reject positivism. But then the excursus into case law analysis which Trevor Allan undertakes risks becoming either meaningless or circular. It becomes the former if it entails a refusal to consider the possibility that the cases under examination may be explained on positivist grounds. This definitional fiat would mean that all cases would necessarily be subject to a non-positivist reading, so why bother examining any cases at all? It becomes circular if the prefernce for the non-positivist reading so loads the case law analysis that one refuses to consider the positivist reading of this material. "'Ethics in the Public Domain, op. cit., n. 44, Chap. 17. 47 ibid., p. 373. •s ibid., pp. 373-374. 40 ibid., p. 374. 50 ibid., p. 375. 51 ibid., p. 375.

The Rule of Law and the Separation of Powers Formal & Substantive Conceptions of the Rule of Law

113

485

justified in order to mix 52 "the fruits of long-established traditions with the urgencies of short-term exigencies". It is precisely because legislatures can be susceptible to short-term influences, whether generated by elections or the need to respond quickly to public pressure, that the courts should have a role as the guardians of longer-term tradition. 53 What then of the place of civil rights within this vision of the rule of law? Raz is careful in this respect. He states that the protection of such rights is partly presupposed and partly implied by the preceding analysis. 54 The analysis presupposes such rights in that Raz's present discussion of the rule of law is confined to democratic societies, and a society cannot be democratic without the existence of such rights. The analysis implies the existence of such rights because "in insisting on the integration of legislation and other current measures with legal tradition enshrined in doctrine, the rule of law respects those civil rights which are part of the backbone of the legal culture, part of its fundamental traditions" .55 The discussion thus far has already touched upon a number of aspects of legal theory stricto sensu, and this is inevitable in any meaningful discussion of the rule of law. It would, however, take us beyond the remit of this article to consider the way in which the vision of the rule of law fits into the broader theory of law which Raz enunciates. 56

(b) Jowell We should not conclude our analysis without considering another attempt to delineate a middle ground between a purely formal conception of the rule of law, and the thoroughgoing substantive version of this doctrine. This is to be found in the work of Jeffrey Jowell. 57 He accepts that one must be careful about equating the rule oflaw with the substance of particular rules. He accepts also that a significant part of the rule of law is concerned with procedure or form as opposed to substance. Jowell does however believe that the rule of law has a substantive dimension. 58 He perceives the rule of law as a principle of institutional morality and as a constraint on the uninhibited exercise of government power and argues that it does possess a substantive aspect. This aspect is manifest in the judiciary's willingness to strike down agency action if it is unreasonable, arbitrary or capricious. Jowell recognises that in some instances judicial intervention is premised upon the fact that the agency has departed from. the sphere over ibid., p. 376. Loc. cit. ,. ibid., pp. 376-377. 55 ibid., p. 376. 50 Suffice it to say for the present that any such discussion would have to consider, inter alia, the version of positivism espoused by Raz, in the form of the strong social thesis, the nature of the sources of law posited by this thesis, and the extent to which this theory leaves room for moral considerations when identifying the contents of a particular legal rule. See Raz, op. cit., n. 40, pp. 47-48. 57 "The Rule of Law Today", inJowell and Oliver (eds), The Chan)!ingConstitlltion (3rd. ed., 1994), Chap. 3, esp. pp. 71-77. SX ibid., pp. 71-77. 52

53

The Rule of Law and the Separation of Powers

114 486

Public Law

which it has been given authority by the legislature. In other circumstances the courts do not even really pretend that they are enforcing legislative intent, and are in reality subjecting agency decision-making to substantive control based on the rule of law. It is reasonably clear from a general reading ofJeffrey Jowell's thesis that he seeks to tread a middle ground between the formal and substantive conception of the rule of law. 59 There are, however, two difficulties in holding to this middle ground. On the one hand, the principles of judicial review are not self-executing. A judicial decision holding that' administrative action should be struck because it is unreasonable or capricious will often, of necessity, involve the identification of various interests, and the assignment of normative weight to them. This has become more evident of late as the courts and commentators have begun to talk of applying these administrative law principles with differing intensity depending upon the nature of the interests· at stake within a particular case. The possible inclusion of proportionality as an independent head of judicial review reinforces this point, 60 since this doctrine, by its very nature, requires the identification of the competing interests in a dispute and the assignment of normative value to them. Given that this is so it is difficult to perform this exercise of normative evaluation without explicitly or implicitly relying on some background conception of justice. On the other hand, the exclusion of constitutional doctrine is problematic. Jeffrey Jowell argues that the substantive aspect of the rule of law stems primarily from the need to constrain the uninhibited exercise of governmental power. The administrative law tools of judicial review are the mechanism to achieve this end. Yet given the foundation of the argument it is difficult to see why tools more commonly associated with constitutional law, concerning rights such as expression, equality and the like, should not also be of relevance here. If the argument is based on the need to constrain the uninhibited exercise of government power, and the administrative law principles of judicial review are regarded as serving this facet of the rule of law, then why should the "limits of the rule oflaw" 61 be set so as to exclude constitutional constraints designed to serve the same end? Once constitutional constraints on governmental power are included it becomes even more difficult to avoid the taxing issues of justice and political theory adverted to above. The preceding argument should not in any way be taken as denying the real importance of constraints on governmental power. The object of the analysis is, rather, to question whether the inclusion of such constraints within the rule oflaw is readily compatible with the preservation of a middle ground between a formal and substantive conception of this doctrine.

'" ibid.• pp. 72-73, 76--77. "" Jowell and Lester, "Beyond l#dnesbury: Substantive Principles of Judicial Review" [1987] P.L 368. "' Jowell, op. cit., n. 57, pp. 76--77.

The Rule of Law and the Separation of Powers Formal & Substantive Conceptions of the Rule of Law

115

487

4. Conclusion There will be no attempt to summarise the entirety of the arguments presented above. Three connected points may, however, be made by way of conclusion. First, the rule oflaw is rightly regarded as a central principle of constitutional governance. It is, therefore, all the more important that we should be as clear as possible about its meaning. It has been argued that clarity in this respect cannot be attained unless public lawyers are aware of the issues of legal theory which underlie the concept. This should not come as a surprise. At the most basic level one might well expect that the meaning to be attributed to a phrase such as the "rule of law" would be dependent upon what one understands by the term "law". That this is indeed so is confirmed by the preceding analysis, since what ultimately divides the formalist and substantive conceptions of the rule of law is disagreement about the way in which we identifY legal norms. Secondly; it should be noted that both Raz and Dworkin actually agree on one important issue which is central to us here: the adoption of a fuliy substantive conception of the rule of law has the consequence of robbing the concept of any function which is independent of the theory of justice which imbues such an account of law. Their fundamental disagreement concerns the very nature of law and the role of the courts in adjudication. Thirdly, it is of course open to public lawyers, and indeed any one else, to choose between the contending views of the rule of law presented above. Debate on this issue is helpful. It is nonetheless important to understand the consequences of adopting a particular position on this matter. The phrase the "rule oflaw" has a power or force of its own. To criticise governmental action as contrary to the rule oflaw immediately casts it in a bad light. Such criticism may well be warranted depending upon the circumstances. Yet ifthe,nub of the ,critique is posited upon the substantive conception of the rule of law then intellectual honestly requires that this is made clear, and it also demands clarity as to the parti.cular theory of justice which informs the critique.

[5] IS THE RULE OF LAW AN ESSENTIALLY CONTESTED CONCEPT (IN FLORIDA)? JEREMY WALDRON

I. THE FLORIDA DEBACLE

One of the remarkable features of the turmoil surrounding the counting and recounting of votes in the State of Florida in the 2000 U.S. Presidential Election was the frequency with which "the Rule of Law" was invoked. Let me count the ways. 1 When the Florida secretary of State, Katherine Harris, announced her certification of the result, she said, "[t]he true victor in the Florida election is the Rule of Law"; but her critics said that her patent political bias - she was also co-chair of the Bush campaign in Florida- was an affront to the Rule of Law. When the two sides began their various campaigns of litigation, we were told that their willingness to take contentious political issues to court was a tribute to the Rule of Law in the United States; but we were also told that litigation undermined the Rule of Law, since it showed the parties' unwillingness to let any legal decision stand. The authorization of recounts by the Florida courts appeared to some the antithesis of the Rule of Law, so long as no clear rules were laid down to govern the counting. (Indeed, the Democrat-appointed majority on Florida's Supreme Court was accused of judicial usurpation of the Rule of Law.) But others said that the careful opinions of the Florida judges gave comfort to those who cherished the Rule of Law, because they'd evidently done the best they could to balance the competing interests of the voters, the campaigns, and the officials concerned. When the matter came before the Supreme Court of the United States, Vice-President Gore took the line that the Rule of Law I will not burden these paragraphs with footnotes. The use of "the Rule of Law" in the various phases of the Florida debacle can be documented from any number of contemporary newspaper accounts. 1

The Rule of Law and the Separation of Powers

118

138

JEREMY WALDRON

precluded any cnttctsm of the justices. But it quickly became apparent that many others thought that the best they could do for the Rule of Law was to condemn the decision in Bush v. Gore as shamefut and in this they were joined by members of the Court's minority, in words that have now become legendary: It is confidence in the men and women who administer the judicial system that

is the true backbone of the Rule of Law. Time will one day heal the wound to that confidence that will be inflicted by today's decision. One thing, however, is certain. Although we may never know with complete certainty the identity of the winner of this year's Presidential election, the identity of the loser is perfectly clear. It is the Nation's confidence in the judge as an impartial guardian of the Rule ofLaw. 2

The Rule of Law was not the only political ideal invoked on all sides in this way. Naturally there was also a great deal of talk about democracy. But just because law and law-suits were involved in every phase of the debacle, it was inevitable that the Rule-of-Law card would be played by the parties whenever it suited them.

II. ANALYTIC CONFUSION

As political philosophers, we like to keep our armory of concepts in good shape; that's why we devote so much energy to the analysis and clarification of terms like "liberty," "justice," and "law." So there are bound to be alarm-bells ringing in analytic circles when a term like "the Rule of Law" is invoked so frequently on so many sides of so many issues in a fraught political debacle. We know, of course, that common usage of a loaded term like the "Rule of Law" does not necessarily reflect careful philosophical analysis, and that good analysis can survive a rather loose fit with ordinary usage. Still, there must be some connection. Philosophical analysis of an important term cannot afford to distance itself too much from the use of that term on the streets: otherwise the analysis is not the analysis of anything interesting, from a political point of view. On the other hand, if street-level use is completely confused, we may have no alternative but to pronounce the term 2 531 U.S. 98, at 128-129 (Stevens J.). See also Breyer J. in dissent, ibid. at pp. 157-158.

The Rule of Law and the Separation of Powers IS THE RULE OF LAW AN ESSENTIALLY CONTESTED CONCEPT?

119

139

meaningless. That is, it may tum out to be impossible to say what any use of the term implies or is implied by, or what it presupposes or is presupposed by; its protean usage may make it nothing more than a site for equivocation. At that stage, the best we could do would be to substitute a well-defined technical term for it, which will serve us for analytic and theoretical purposes. 3 The technical term might be a homonym of the ordinary linguistic term. But the ordinary linguistic term, with its confused and confusing meanings, will in fact have been relegated to the realm of street-level rhetoric; and serious philosophers will have forgotten about it. Even before Florida, a number of theorists had come close to a conclusion along these lines so far as "the Rule of Law" was concerned. Judith Shklar's verdict was perhaps the most severe: It would not be very difficult to show that the phrase "the Rule of Law" has become meaningless thanks to ideological abuse and general over-use. It may

well have become just another one of those self-congratulatory rhetorical devices that grace the public utterances of Anglo-American politicians. No intellectual effort need therefore be wasted on this bit of ruling-class chatter. 4

On Shklar's account, "the Rule of Law" is just an empty slogan, useful perhaps as decoration for whatever else one wants to assert into a political dispute, but incapable of driving one's argument much further forward than the argument could have driven on its own. Our experience of its use in the Florida debacle might seem to reinforce that position. "The Rule of Law" sounded grand, certainly; but at the end of the day, many will have formed the impression that the utterance of those magic words meant little more than "Hooray for our side!" Joseph Raz has noted a tendency to use "the Rule of Law" as a general stand-in for everything nice one could ever want to say about a political system, or everything good one could want from it.

3 For a defense of this approach, see Felix Oppenheim, Political Concepts: a Reconstruction (Oxford: Basil Blackwell, 1981), pp. 177 ff. 4 Judith N. Shklar, "Political Theory and the Rule of Law," in Allan C. Hutcheson and Patrick Monahan (eds.) The Rule of Law: Ideal or Ideology (Toronto: Carswell, 1987). p. 1.

The Rule of Law and the Separation of Powers

120

140

JEREMY WALDRON

Not uncommonly when a political ideal captures the imagination oflarge numbers of people its name becomes a slogan used by supporters of ideals which bear little or no relation to the one it originally designated. 5

Raz, however, does not despair of the phrase. He maintains that it once did have a meaning that was more determinate than this, and he sets out in his essay to retrieve or reconstruct that more specific meaning. I shall examine the detail of the specific meaning Raz attributes to "the Rule of Law" in section 6. For the time being, however, we shall leave him with Judith Shklar in the camp of those who are sad but not surprised that the term is available for such indiscriminate use in a situation like the Florida debacle. Should we accept their diagnosis? I wonder. Raz and Shklar both write as though "the Rule of Law" once had a clear meaning, but now lamentably it has lost it, through misuse occasioned by wellmeaning but muddled enthusiasm. This lamentation over lost clarity is a common trope in the rhetoric of philosophical analysis. But it is usually a myth. The terms whose ordinary usage we now deplore as confused, the terms that we accuse our contemporaries of misusing, have almost always been perplexing. Although numerous attempts have been made to pin them down with clear definitions, or precisely because of this, they have presented themselves throughout their history as sites for contestation as to what counts as their proper use. 6 They have always resisted the analyst's corral.

III. A HISTORY OF CONTESTATION

"The Rule of Law" is no exception. Both in its origin and in its application to early-modern and modern political institutions, its content and implications have been uncertain and controversial. Let me give a few examples. Joseph Raz, "The Rule of Law and its Virtue," in his collection The Authority of Law: Essays on Law and Morality (Oxford: Clarendon Press, 1979), at p. 210. 6 The best example is the controversy about the proper use of "rights." For a fine account of the long historical pedigree of our modem disputes and confusions about what it means to say that someone has a right, see Richard Tuck, Natural Rights Theories: their Origin and Development (Cambridge: Cambridge University Press, 1979). 5

The Rule of Law and the Separation of Powers IS THE RULE OF LAW AN ESSENTIALLY CONTESTED CONCEPT?

121

141

As early as the fourth century B.C., we have Aristotle hesitating on the implications that the Rule of Law might have for the use of rules in politically fraught situations. On the one hand, he says in the Rhetoric that as many matters as possible must be settled in advance by general rules. 7 On the other hand, he says in the Politics that for hard cases, the Rule of Law consists in a legal specification of the individual or group that must take personal responsibility for the decision. (For those cases, said Aristotle, the law controls the decision indirectly by controlling the process of appointment and basis on which appointees are educated.) 8 We have him saying, too, in the Nichomachean Ethics, that if there are likely to be many hard cases, then the law should adopt a less rigid rule. 9 These hesitations in Aristotle's account mean that, although he's generally regarded as the founder of our Rule-of-Law tradition, no one quite knows what to draw from him. "I stand with Aristotle, then - which is a pretty good place to stand," wrote Justice Scalia in an article published more than ten years before his controversial intervention in Bush v. Gore. 10 But what Scalia draws from Aristotle - "Allocate as little personal discretion to judges as possible" - is different from what a jurist like Lawrence Solum draws from him - "Have faith in the virtue of the judiciary" - and different again from the conclusions of a formalist philosopher like Ernest Weinrib - "Trust the immanent reason of the common law." 11 Again, Aristotle is famous for his rather mystical observation that "the law is reason unaffected by desire":

7 Aristotle, On Rhetoric, trans. George A. Kennedy (Oxford: Oxford University Press, 1991), p. 31 (Bk. I, Ch. ii, 1354a). 8 Aristotle, The Politics, ed. Stephen Everson (Cambridge: Cambridge University Press, 1988), p. 78 (Bk. III, Ch. xvi, 1287a). 9 Aristotle, Nichomachean Ethics, translated by Sir David Ross (London: Oxford University Press, 1954), p. 133 (Bk. V, Ch. 10, 1137b). 10 Antonin Scalia, "The Rule of Law as a Law of Rules, University of Chicago Law Review, 56 (1989) 1175, at p. 1182. 11 Lawrence B. Solum, "Equity and the Rule of Law," in Ian Shapiro (ed.) Nomos XXXVI: The Rule of Law (New York: New York University Press, 1994) 120; Ernest Weinrib, "The Intelligibility of the Rule of Law," in Hutcheson and Monahan, supra note 4, 59, at p. 63.

The Rule of Law and the Separation of Powers

122

142

JEREMY WALDRON

[He who bids the law rule may be deemed to bid God and Reason alone rule, but he who bids man rule adds an element of the beast; for desire is a wild beast, and passion perverts the minds of rulers, even when they are the best of men. 12

But this means different things in the hands of those who equate law with reason and those who see law as the opposite of individual judgement. If law is reason, then legal structures must facilitate the exercise of reason, and that means trusting human discretion to be reasonable in certain circumstances. 13 From this point of view, the last thing you want is to be hedged around with rigid rules, for rules are necessarily arbitrary compared to the reasonable pursuit of the goals that underlie them. 14 On the other hand, the very same passage from Aristotle seems to support something like Scalia's case for determinate rules: unless a rule operates like a machine to exclude individual judgement, there is always a danger that individual passion or purpose - "the element of the beast" - will creep back into political decision-making. Such perplexities have helped lay the foundation for enduring controversy about whether judge-made law is to be regarded as the epitome of the Rule of Law or as part of the problem that the Rule of Law is supposed to solve. The concern about judicial legislation is a hardy perennial. We hear the concern in Montesquieu's insistence that where the judge becomes the law-maker, "the life and liberty of the subject [are] exposed to arbitrary control." 15 We hear it too in Jeremy Bentham's invective against the "dog law" made by "Judge & Co." 16 We hear echos of it in Robert Cover's attack on "jurispathic Aristotle, Politics, supra note 8, p. 78 (Bk. III, Ch. xvi, 1287a). "Liberals think the Rule of Law should be aligned with the rule of reason. You take a law and then ask how reasonable people would interpret it."- Richard Fallon, as quoted by Steven Thomma, in "Rule of Law or Rule of the People?" The Philadelphia Inquirer (December 12, 2000). 14 For the inherent arbitrariness of rules, see Duncan Kennedy, "Form and Substance in Private Law Adjudication," Harvard Law Review, 89 (1976) 1685, atp.1689. 15 Montesquieu, The Spirit of the Laws (1748), ed. Anne Cohler, Basia Miller and Harold Stone (Cambridge: Cambridge University Press, 1989), p. 157. (Bk. 11, Ch. 6). 16 See Jeremy Bentham, Of Laws in General, edited by H.L.A. Hart (London: Athlone Press, 1970), pp. 152-153 and 184 ff. See also Gerald J. Postema, 12 13

123

The Rule of Law and the Separation of Powers IS THE RULE OF LAW AN ESSENTIALLY CONTESTED CONCEPT?

143

courts" 17 and in Lon Fuller's nervous concession that common-law decision-making seems to partake of the retroactivity prohibited by his "internal morality of law." 18 We also hear a version of it in Thomas Hobbes' insistence that rule by judges is just another form of sovereignty, an argument that Hobbes thinks shows the hopelessness of the liberal position that a sovereign can be bound by law: A fourth opinion, repugnant to the nature of a Commonwealth is this: That he that hath the Soveraign Power is subject to the Civill Lawes. ... Which errour, because it setteth the Lawes above the Soveraign, setteth also a Judge above him, and a Power to punish him; which is to make a new Soveraign; and again for the same reason a third to punish the second; and so continually without end, to the Confusion, and Dissolution of the·common-wealth. 19

On this logic, either you have a final human authority or you do not. If you do, then all talk of being ruled by the law is fatuous, whether or not the final authority wears a wig and sits in a court. There must always be somebody - not some text but some body - who has the final word. The attempt to find a way around this absolutist logic is what exercised generations of jurists after Hobbes - from Harrington's republicanism, through the liberalism of Locke, Montesquieu and the Federalists, to the Victorian paradoxes of Albert Venn Dicey's attempt to reconcile parliamentary sovereignty and the Rule of Law. 20 It remains with us today in the United States as we swing uneasily between the position that the Rule of Law positively Bentham and the Common Law Tradition (Oxford: Clarendon Press, 1986), p. 277. 17 Robert M. Cover, "Nomos and Narrative," Harvard Law Review, 97 (1983) 4,atpp.40-44 18 Lon L. Fuller, The Morality of Law, Revised Edition (New Haven: Yale University Press, 1969), pp. 56-57. 19 Thomas Hobbes, Leviathan (1651), ed. Richard Tuck (Cambridge: Cambridge University Press, 1996), p. 224 (Ch. 29). 20 James Harrington, The Commonwealth of Oceana (1656) ed. by J.G.A. Pocock (Cambridge: Cambridge University Press, 1992), pp. 8ff.; John Locke, Two Treatises of Government (1689) ed. Peter Laslett (Cambridge: Cambridge University Press, 1988), pp. 326-330 (II, paragraphs 89-94); Montesquieu, supra note 15, at pp. 154 ff. (Bk. 11, Chs. 2-6); James Madison, Alexander Hamilton and John Jay, The Federalist Papers (1787-1788), ed. Isaac Kramnick (Harmondsworth: Penguin Books, 1987), pp. 318-322 and 436-442 (Numbers

The Rule of Law and the Separation of Powers

124

144

JEREMY WALDRON

requires that the Supreme Court have the final say in any constitutional crisis, and the position that judicial supremacy is as offensive to the Rule of Law as any other form of unreviewable hegemony in a constitutional regime. As with judicial authority, so with lawyers and litigation. Again the Rule-of-Law heritage is riddled with antinomies. Alexis de Tocqueville paid tribute, famously, to the disproportionate involvement of lawyers in American public life, believing that this protected the Rule of Law against the sort of impetuous reformism that always threatened to destabilize it. 21 For others, however, litigators are the enemy, their livelihood depending (in the words of Edmund Burke) upon whatever rendered property or other institutions "questionable, ambiguous, and insecure," as they struggle to find loopholes in the rules set down by society for the ambition or greed of their clients. Nothing permanent, said Burke, nothing with the stability normally connoted by the Rule of Law, could ever be expected from such men. 22 Accordingly, the use of litigation as a strategy for influencing and determining public outcomes may be seen at the same time as subservience to the ideal of legality and as the very thing that threatens to undermine it.

IV. RULE-OF-LAW ISSUES IN FLORIDA

These uncertainties as to what the Rule of Law consists of- litigation or self-restraint, judicial supremacy or judicial deference, rules or standards, mechanical judgment or reasoned discretion - are not just a matter of intellectual history. They are there waiting to crop up whenever a political system goes into crisis or controversy along 51 by Madison and 78 by Hamilton); and A.V. Dicey, Introduction to the Study of the Law of the Constitution (1886) (Indianapolis: Liberty Classics, 1982), pp. 268-273. 21 Alexis De Tocqueville, Democracy in America, ed. Alan Ryan (London: Everyman, 1994), Vol. I, p. 273 (Ch. xvi): "Men who have made a special study of the laws derive from this occupation certain habits of order, a taste for formalities, and a kind of instinctive regard for the regular connection of ideas, which naturally render them very hostile to the revolutionary spirit and the unreflecting passions of the multitude." 22 Edmund Burke, Reflections on the Revolution in France ed. Conor Cruise O'Brien (Harmondsworth: Penguin Books, 1968), p. 131.

The Rule of Law and the Separation of Powers IS THE RULE OF LAW AN ESSENTIALLY CONTESTED CONCEPT?

125

145

any of these dimensions. And when you think about it, the Florida debacle was a natural ground for the airing of these disputes. What was at stake was the most powerful office an individual can occupy in the United States, perhaps in the world. This office is allocated by indirect election on the basis of an extraordinarily complicated set of rules relating the democratic expression of individual preferences to political outcomes, with further layers of relatively untested rules as back-up should the first echelon of rules prove unsatisfactory. With these high stakes - power and democracy - the interpretation of the back-up rules was thrown in doubt, and the issue was litigated and came before the courts. Since the courts have a political complexion in the United States, the issue very quickly reduced itself to this: who controls (or what are the political allegiances of those who control) the laws that are supposed to control the allocation of the most powerful office in the land? No better ground for exploring the traditional antinomies of the Rule of Law can possibly be imagined. Moreover, when we look at the way in which the Rule of Law was invoked in the Florida debacle, we find that it was invoked in ways that were precisely calculated to raise these central issues. The rival invocations of the Rule of Law were not just a matter of" 'tis!' and 'tisn't!'" in regard to the legitimacy of each particular decision. They weren't simply products of legal or political disagreement about what specific laws or precedents required, with one side thinking that a particular provision of Florida electoral law required X while the other side thought it required or permitted not-X. They involved wholesale, not just retail disagreement about what the Rule of Law required. Quite apart from the meaning of any particular text or precedent or doctrine, there was disagreement among participants and commentators in Florida in four general areas: (i) the exercise of official discretion by partisan actors, (ii) the governance of discretion by vague or indeterminate standards, (iii) the parties' recourse to litigation, (iv) the role of courts in resolving substantive disputes. In each of these four areas, there were some who thought the practice in question was the epitome of the Rule of Law and some who thought it was the antithesis of the Rule of Law. Let me go through these in tum. (i) Some participants and commentators believed that an exercise of discretion could certainly count as a fine example of the Rule of Law, provided the discretion

The Rule of Law and the Separation of Powers

126

146

JEREMY WALDRON

was legally vested in the person purporting to exercise it. But aspects of American state practice cast serious doubt on this in the context of the Florida debacle. It is common in the United States for official positions to be held by political appointees; people appointed precisely because of their political partisanship. That such a partisan should be the one entrusted with discretion to certify an electoral result is problematic, to say the least, from the point of view of anyone interested in the outcome being determined by law rather than by the political clout of the parties. (ii) Apart from the question of who exercises discretion, there was also controversy about the way in which discretion is exercised. Discretion can be tightly bound by rules, so that the official is required to exercise it in one way if certain facts are present "a chad hanging by one corner" - and another way if other facts are present. Or the discretion may be more loosely governed by requiring the official to make a judgement about some factor "the intention of the voter" - on which opinions may differ. Many who criticized the Florida Supreme Court's empowerment of ballotcounters thought that in the absence of clear rules to govern the vote-counting, any legal authorization of discretion was itself antithetical to the Rule of Law. On this view, the Rule of Law is supposed to supersede the role of human discretion; though, as we have seen, there is also a view that the Rule of Law is supposed to frame the exercise of reasoned judgment and furnish it with an aura of legality. (iii) Even the recourse of the parties to litigation could be viewed as a consummation of the Rule of Law or as its opposite. "Gore and Lieberman are exercising their legal . . . right to contest the messiest election result in memory," wrote commentator Michael Kramer. He went on: "The way out of chaos is through the Rule of Law, and that means seeking any court in a storm.'m That's one view of the matter, and for the other we have James Baker's suggestion (as early as November 27) that an attempt to prolong litigation is contrary to Rule-of- Law values: "At some point the Rule of Law must prevail and the lawyers must go home." 24 From the latter perspective, persistent litigation is like a gambler insisting 23 24

Michael Kramer, in The New York Daily News (November 27, 2000). The Atlanta Constitution (November 27, 2000).

The Rule of Law and the Separation of Powers IS THE RULE OF LAW AN ESSENTIALLY CONTESTED CONCEPT?

127

147

on rolling the dice again and again, hoping against hope that blind fate will yield a random decision in his favor. But from the former perspective, litigation is essential to the Rule of Law: the entitlement to litigate embodies the law's respect for each individual, and litigation is the adversarial institution we use to ensure that all legal issues are addressed, one by one, and that layers of legal accountability are respected. To count litigation as something contrary to the Rule of Law whenever it exasperates us is simply to display the temperamental limits of our allegiance to the ideal. (iv) Some thought that the resolution of any issue counted as the Rule of Law provided it was done through the hierarchy of courts. Gore counsel David Boies put it like this, anticipating what turned out to be the final decision by the U.S. Supreme Court: "It may be a decision based on the rule of law that we agree with, it may be a decision based on the rule of law that we disagree with, but it will be based on the rule of law."25 Yet others evidently thought that the autonomy and unaccountability of judges was precisely the sort of problem that the Rule of Law ideal was supposed to confront. On the former view, a court deciding an issue is the same as that issue being decided as law. Of course it is possible that the courts may get the law wrong; but short of the fantasy that the laws themselves might rear up and render their own objective decision, this is the most that the Rule of Law could possibly entail. Realistically according to this view - the Rule of Law consists in issues being settled by ponderous legal processes, procedures of deliberation and reason-giving that are focused on antecedent legal materials rather than political advantage, and in a form of deference on the part of the contesting parties that is motivated by the stake they have, along with their fellow citizens, in the integrity of the legal and constitutional order. On the latter view, by contrast, there is always the danger that judges are taking advantage of the power and authority of their office to make themselves into the very autocrats whose rule the Rule of Law is supposed to supersede. On both sides in the Florida debacle, criticism of judicial decision-making resonated with what is now a settled feature of American political culture the suspicion that judges are elevating their own morality or their own political preferences above the law of the land. We want the 25

David Boies, as quoted in the New York Daily News (December 10, 2000).

The Rule of Law and the Separation of Powers

128

148

JEREMY WALDRON

rule of laws, not men, say most Americans, and "men" includes judges. 26 V. WHAT DOES "ESSENTIALLY CONTESTED" MEAN?

As I have said, these antinomies are not new. Whether the antagonists in Florida knew it or not, they are in fact aspects of a venerable heritage of contestation that comes down to us as part and parcel of the Rule-of-Law tradition. The fact that "the Rule of Law" has always evoked this contestation has led some to surmise that an old article by a linguistic philosopher, W.B. Gallie, entitled, "Essentially Contested Concepts" might afford the most fruitful basis for approaching the analysis of the term. 27 Gallie's claim was that there are certain concepts "the proper use of which inevitably involves endless disputes about their proper uses on the part of their users."28 I think the suggestion that this might apply to the Rule of Law is an interesting and fruitful one. More must be said, however, than has been stated so far about what essential contestability amounts to. 29 The idea is clearly vulnerable to overuse. In his original paper, Gallie suggested that it might help us understand debates about democracy, art, and Christianity. In political theory, the idea has also been used to analyze power, interests and freedom. 30 But in the law review literature, the use of the term 26 See the discussion in Ronald Dworkin, "Political Judges and the Rule of Law" in his collection A Matter of Principle (Cambridge: Harvard University Press, 1985), 9. 27 Richard Fallon, " 'The Rule of Law' as a Concept in Constitutional Discourse," Columbia Law Review, 97 (1997) 1, at p. 6 (citing W.B. Gallie, "Essentially Contested Concepts," Proceedings of the Aristotlean Society, 56 (1955-1956) 167). 28 Gallie, supra note 27, at p. 169. 29 All that Fallon said to elucidate the idea was this, supra note 27, at p. 6: "The Rule of Law is what some philosophers have called an 'essentially contestable concept': it has evaluative as well as descriptive elements, and its correct application cannot be fixed simply by appeal to ordinary usage. In more concrete terms, the 'true,' 'best,' or 'preferred' meaning of the Rule of Law depends on the resolution of contestable normative issues; disagreements are therefore to be expected." 30 See William E. Connolly, The Terms of Political Discourse (Lexington, Mass.: D.C. Heath, 1974); Steven Lukes, Power: A Radical View (London:

The Rule of Law and the Separation of Powers IS THE RULE OF LAW AN ESSENTIALLY CONTESTED CONCEPT?

129

149

has run wild, 31 with "essentially contested" meaning something like "very hotly contested, with no resolution in sight." But in Gallie's article, "essentially" is not just an intensifier. One way of getting at Gallie's idea is to say that the term "essentially" refers to the location of the disagreement or indeterminacy: it is contestation at the core, not just at the borderlines or penumbra of a concept. We all know about vagueness and legal indeterminacy based on vagueness: some things are green, some are blue; but on the borderlines there are blue/green cases of uncertainty. This and allied phenomena like "open texture" are quite familiar in jurisprudence. 32 By contrast, a concept such as democracy evokes disagreement not only about marginal cases (say, Kuwait or Malaysia) but also about paradigm or core cases. For some, the United States is a paradigm or core case of democracy: if it is not a democracy nothing is. For others, the U.S. is not a good paradigm, not only because of things like low voter-turnout, lack of proportional representation, and compromises with various aristocratic ideals (like judicial review), but also because it lacks important features of social democracy - the empowerment, in something more than merely formal terms, of the worst-off members of society. The disagreement between those who define democracy purely in terms of formal electoral arrangements and those who insist on a social element, is disagreement about the core or essence of democracy, and thus it seems to indicate that democracy might be described as essentially contested. Even this, though, does not quite capture Gallie's idea: it misses the element of value or normativity associated with this Macmillan, 1974); Christine Swanton, Freedom: a Coherence Theory (Indianapolis: Hackett, 1992). 31 A Westlaw check reveals that the following have been described as "essentially contested concepts": alienation, autonomy, author, bankruptcy, boycott, citizenship, civil rights, coherence, community, competition, the Constitution, corruption, culture, discrimination, diversity, equality, equal protection, freedom, harm, justification, liberalism, merit, motherhood, the national interest, nature, popular sovereignty, pornography, power, privacy, property, proportionality, prosperity, prostitution, public interest, punishment, reasonable expectations, religion, republicanism, rights, sovereignty, speech, sustainable development, and textuality. 32 For the "open texture" of ordinary language and its application to law, see H.L.A. Hart, The Concept of Law, Second Edition (Oxford: Clarendon Press, 1994), pp. 124-136.

The Rule of Law and the Separation of Powers

130

150

JEREMY WALDRON

contestation-to-the-core. The key to Gallie's idea of essential contestability seems to be a combination of normativity and complexity: only normative concepts with a certain internal complexity are capable of being essentially contested. Consider the following characteristics that a concept might have. Suppose a concept is used to refer to some sort of valued or outstanding achievement, which everyone agrees is a complex achievement, having to do with the presence of a number of different aspects or attributes. Though everyone agrees the achievement is complex, no one thinks it is just the sum of what it takes as its constituent parts: everyone says that the achievement is valued as a whole, even though they disagree about the parts. Moreover, as Gallie puts it, there is nothing absurd or self-contradictory in any one of a number of possible rival descriptions of its total worth. The achievement, as he puts it "is initially variously describable."33 Suppose too that everyone acknowledges that the achievement "admits of considerable modification in the light of changing circumstances; and such modification cannot be prescribed . . . in advance." 34 With this as background, we imagine people advancing and defending (and criticizing and modifying) rival conceptions of the concept. Various conceptions might cite lists of attributes that stand in a "family-resemblance" relation to one another, and they might offer rival accounts of how the distinct attributes go together to make up the valued achievement that all the conceptions are trying to give an account of. For example, in discussions of democracy, one conception might stress electoral rights and civil liberties (considered purely as adjuncts to electoral rights); another might stress economic equality and the effective enfranchisement of working people; a third might stress grass-roots participation and thoughtful deliberation among an educated populace; a fourth might characterize electoral competition simply as a peaceful framework for changeovers in power; and so on. Each conception is put forward as an attempt to outdo others in capturing an elusive sense, that we all share, a sense that somewhere in the midst of this contestation there is an important ideal that social and political systems should 33 34

Gallie, supra note 27, p. 172. Idem.

The Rule of Law and the Separation of Powers IS THE RULE OF LAW AN ESSENTIALLY CONTESTED CONCEPT?

131

151

aspire to. What that ideal is exactly none of us can say without participating in the contestation, i.e. without offering a conception of it that is bound to be controversial. However, it is precisely through successive efforts that we together - all the contestants - get at it in this way to sustain and develop a shared sense of the presence and importance of such an ideal. "Essential contestability," then, is about a way in which certain ideals are present to us. Some ideals are present in clear and well-defined form- the economists' "wealth-maximization," for example. In that case controversies tend to center on strategies and implementation, and the weight that the ideal should have against other competing values. But in the case of essentially contested concepts, they are present to us only in the form of contestation about what the ideal really is. The idea of essentially contested concepts has sometimes been understood pessimistically, as an imperfection that takes us in the direction of relativism. For example, William Connolly characterizes their role in this way: Politics involves the clash that emerges when appraisive concepts are shared widely but imperfectly, when mutual understanding and interpretation is possible but in a partial and limited way, when reasoned argument and coercive pressure commingle precariously in the endless process of defining and resolving issues. 35

It may lead us to adopt an attitude of humility, but it will be humility in the face of an intractable problem. I read Gallie's account as much more upbeat than that, and certainly, by suggesting that the Rule of Law is an essentially contested concept, I don't want to be taking a gloomy view of the "imperfect" or "precarious" nature of argumentation about the Rule of Law. Instead, I take seriously Gallie's suggestion that the contestation surrounding the use of an essentially contested concept characteristically advances the quality of argumentation using that concept: Recognition of a given concept as essentially contested implies recognition of rival uses of it (such as oneself repudiates) as not only logically possible and humanly "likely", but as of permanent critical value to one's own use or interpretation of the concept in question .... One very desirable consequence of the required recognition in any proper instance of essential contestedness might there35

Connolly, supra note 30, at p. 40.

132

The Rule of Law and the Separation of Powers

152

JEREMY WALDRON

fore be expected to be a marked raising of the level of quality of arguments in the disputes of the contestant parties. 36

In other words, the contestation between rival conceptions deepens and enriches all sides' understanding of the area of value that the contested concept marks out. Of course there is a danger that an idea like essential contestability will be used as a cop-out. We need to bear in mind that the claim that a concept is essentially contested is not self-certifying, and it is also not a way of immunizing oneself against analytic mistakes. It is tempting to associate essential contestability with a comprehensive skepticism about the analytic/synthetic distinction, along Quinean lines. 37 But that is a mistake. 38 Even if democracy is an essentially contested concept, still there are some propositions about it which are analytically false: for example, that democracy is a living animal, or that a private event in the life of a single individual can be democratic. Moreover, the claim that democracy is an essentially contested concept is itself an analytic thesis: it presupposes that analysis can establish certain truths about the concept in question - e.g., that it is complex (in roughly the way Gallie describes), evaluative, and so on. John Gray has raised the question of whether there might not be "a radical fault in the very notion of a contest which cannot by its nature be won or lost." 39 How can there be argument about something which (in Gallie's words) is "not resolvable by arguGallie, supra note 27, at p. 193. W.V.O. Quine, From a Logical Point of View (Cambridge: Harvard University Press, 1953), pp. 20-26. 38 Connolly, supra note 30 at pp. 7 and 41, says that essential contestability implies that, at some point, the distinction between analytic and synthetic truths (e.g. about democracy) breaks down. What essential con testability connotes is that some claims that look like they are analytic statements- "Nothing is a democracy unless it involves popular voting among freely nominated candidates for office" - tum out to be controversial in ways that our ordinary paradigms of analytic statements - like "A bachelor is an unmarried man" - are not. But Connolly is careful to add that this "is not the same as saying that there are no analytic statements." 39 John Gray, "Political Power, Social Theory, and Essential Contestability," in David Miller and Larry Siedentop (eds.) The Nature of Political Theory (Oxford: Clarendon Press, 1983). 36

37

The Rule of Law and the Separation of Powers IS THE RULE OF LAW AN ESSENTIALLY CONTESTED CONCEPT?

133

153

ment of any kind?"40 If it is true that "no one clearly definable general use of any of [the concept in question] can be set up as the correct or standard use,"41 then there doesn't seem to be anything for the disagreement to be about. Indeed, the characterization in terms of essential contestability seems to misrepresent what it is like to engage in such contestation. To a participant, surely nothing matters more than that his definition be sustained and his opponent's refuted. He is not saying simply, "Here's one more view about democracy to put in the catalogue." Towards the end of the paper, I want to consider the extent to which a concept can be essentially contested without all users being aware of that fact. But even at the level of the well-informed theorist, we may ask: Is it possible to engage in one of these debates as a partisan of a particular view but also as a theorist who knows why disputes of this kind are intractable? Can one acknowledge that a concept is essentially contested and still claim that one's own view is right and one's opponent's view wrong? Well, one certainly cannot expect to prevail comprehensively - that is, to have one's opponents slap their foreheads and say, "But of course! JW's definition of 'democracy' is correct. Why didn't we all see that before?" Realistically the hope one invests in one's participation in such a dispute is that the contestation - and the sense of the underlying ideal at subsequent stages- will be the better for one's intervention. But that is not the same as hoping or expecting to settle the matter.

VI. THE RULE OF LAW AS AN ESSENTIALLY CONTESTED CONCEPT

The Rule of Law seems to fit this bill quite well. I said at the beginning of section 5 that Gallie gives an account of essential contestability in terms of the complexity of certain evaluations. The Rule of Law is clearly an appraisive concept: it is deployed by almost all of its users to enter a favorable evaluation of the 40

41

Gallie, supra note 27, at p. 169. Ibid., p. 168.

The Rule of Law and the Separation of Powers

134

154

JEREMY WALDRON

regimes or situations to which it applies. 42 Its complexity is also well-known. Now, there are actually several levels of complexity associated with the Rule of Law. At one level, well known to students of AngloAmerican jurisprudence, the phrase "the Rule of Law" is used to conjure up a sort of laundry list of features that a healthy legal system should have. These are mostly variations of the eight desiderata of Lon Fuller's "internal morality of law": laws should be (1) general, (2) publicly promulgated, (3) prospective, (4) intelligible, (5) consistent, (6) practicable, (7) not too frequently changeable, and (8) actually congruent with the behavior of the officials of a regime. 43 Fuller explained that these desiderata work together as a system, and that each of them is a matter of degree, and he himself thought that some of them - public promulgation, for example were more important than others. Other jurists have made other lists. John Finnis gives us eight principles, which are more or less Fullers' in a different order. 44 John Rawls gives us four precepts, each a cluster of requirements, instead of the customary eight (or two): "ought implies can," "similar cases [are to] be treated similarly," "no offense without a law," and "natural justice (due process) is to be observed."45 Joseph Raz gives us eight principles, too; 46 but they are somewhat different from Fuller's and Finnis', and some of them cluster together several different principles. Raz also claims that his eight principles can be reduced to a single basic idea - "that the law should be capable of providing effective guidance."47 (Similarly, Margaret Radin "boils down" Fuller's eight principles to two: "[F]irst, there must be rules; second, those rules must be capable of being followed.") 48 But see Morton J. Horwitz, "The Rule of Law: An Unqualified Human Good?" Yale Law Journal, 86 (1977) 561, at p. 566. 43 Fuller, supra note 18, at p. 39. 44 John Finnis, Natural Law and Natural Rights (Oxford: Clarendon Press, 1980), p. 270. 45 John Rawls, A Theory of Justice (Oxford: Oxford University Press, 1971), pp. 236-239. 46 Raz, supra note 5, at pp. 214-219. 47 Ibid., at p. 218. 48 Margaret Jane Radin, "Reconsidering the Rule of Law," Boston University Law Review, 69 (1989) 781, at p. 785. 42

The Rule of Law and the Separation of Powers IS THE RULE OF LAW AN ESSENTIALLY CONTESTED CONCEPT?

135

155

Most of these precepts and principles relate to the ordinary working of a legal system. The approach they take is summed up by John Finnis when he says that "the Rule of Law" is "[t)he name commonly given to the state of affairs in which a legal system is legally in good shape . . . ."49 By itself. however, the complexity and difference exhibited at this level is not really sufficient to justify calling the Rule of Law an essentially contested concept, since Fuller, Rawls, Raz, Radin and Finnis do not present themselves as advocating rival conceptions. Their approaches seem quite congenial to each other; they are filling in the details of what is more or less the same conception in slightly different ways. We see this when we try to relate their laundry lists to a rather more challenging idea associated with the Rule of Law. This is the idea that we might be able to make a salutary addition to the classic menu of constitutional forms: as well as rule by one man, rule by the few, and rule by the many - monarchy, oligarchy, and democracy - there might also be the possibility of the rule of laws, not men. That the law might rule rather than any number of men is the theme of Aristotle's discussion of the Rule of Law in the Politics; it's a common topic in medieval political philosophy; 50 and it survives into the modem era in the discussion sparked by Hobbesian absolutism, which I mentioned in section 3. Now, many legal positivists regard the idea of the rule of laws, not men as a nonstarter. Since laws are made, interpreted and enforced by men, there is really no contrast, and to the extent that men are involved, we still want to know whether we are dealing with (say) a legislative democracy, oligarchy, or monarchy. 51 The most we can hope for, they say, is men ruling us through the medium of law, and that is what the laundry lists set out by Fuller and the others purport to describe. They are, certainly, quite demanding norms of legislation and public administration, but they appear to have given up on the idea that law can ever truly be ascendant over the indiFinnis, supra note 44, at p. 270. See, e.g., James ofViterbo, "Is It Better to Be Ruled by the Best Man than by the Best Laws?" (1295-1296) in Arthur Stephen McGrade, John Kilcullen, and Matthew Kempshall (eds.) The Cambridge Translations of Medieval Philosophical Texts, Volume Two: Ethics and Political Philosophy (Cambridge: Cambridge University Press, 2001), p. 322. 51 See also Jeremy Waldron, The Law (London: Routledge, 1990), pp. 32-36. 49

50

The Rule of Law and the Separation of Powers

136

156

JEREMY WALDRON

vi duals who have the responsibility of making and enforcing it. They constitute a second-best ideal, relative to the traditional Rule-of-Law aspiration. Other political and legal theorists have second-best ideals that are rather different in character from the Fuller-style laundry list. For example, we have jurists defending the view that the Rule of Law favors certain types (or sources) of law over others. The choice somewhat depends on the kind of "men" whose rule they want to contrast with the Rule of Law. For Justice Scalia, the rule of men means discretionary rule by judges, and he favors a body of law consisting to the extent possible of hard, textually grounded rules, which specify ex ante the outcomes required in particular situations.52 F.A. Hayek's apprehensions on the other hand, are directed towards rule by managerial legislators; for him, the Rule of Law means something like the rule of spontaneously emerging norms. 53 The ascendancy of legislation, on his account, would be at odds with the Rule of Law, even if the legislation were drafted and enforced in a way that satisfied the strictures of Fuller et al. For Ronald Dworkin, one gets the impression sometimes that the Rule of Law means the ascendancy in a political system of a certain culture of argumentation: so the "men" whose rule it opposes are those who would close the argument down or insist that something is to prevail just because they say it should. 54 Nor do legal theorists stand unanimously with Hobbes and Austin in despairing of the higher aspiration that law should have priority over sovereign authority, i.e. over the judges who interpret the law and the legislators who make it. The Rule of Law continues to inform the doctrine of separation of powers, one of whose rationales is to ensure that those who make the law will be bound by the laws they make, just like any other member of society. The idea of constiScalia, supra note 10, at pp. 1177 ff. F.A. Hayek, Law, Legislation and Liberty: Volume I - Rules and Order (Chicago: University of Chicago Press, 1983), at pp. 94 ff. 54 This might seem at odds with Dworkin's so-called "right answer" thesis, but it isn't. Since the right answer thesis is not associated with any method for determining conclusively who has the answer to the legal question right, it functions mainly as a regulative idea of objectivity, underwriting our insistence that there is still something to argue about even when some human authority claims to have had the last word. 52 53

The Rule of Law and the Separation of Powers IS THE RULE OF LAW AN ESSENTIALLY CONTESTED CONCEPT?

137

157

tutional government taking place entirely under the control of a body of fundamental law continues to attract and fascinate us, even when we acknowledge that it runs up against its limits in the power of the judges to interpret the constitution (judicial supremacy) or the power of the people to alter it (popular sovereignty). Even legal positivists these days admit a version of the thesis that law controls sovereignty, whether in the form of A.V. Dicey's emphasis on legislative due process or- more familiar- in the form of H.L.A. Hart's secondary rules. 55 Admittedly, Hart's secondary rules remain human artifacts, and Hart himself is scrupulous about not attributing to them any of the value usually associated with the Rule of Law ideal. Still, the status of secondary rules as human practices that emerge, rather than being the product of intentional enactment, does establish some sort of ascendancy for law as such over the deliberate initiatives of human politicians. Taking all this into account, what we then have in regard to the Rule of Law is a form of contestation which amounts to an on-going debate among jurists and political theorists about the practicability of law being in charge in a society. In the account of essential contestability that I gave in section 5, I followed Gallie in associating the idea with different ways of conceiving some valued achievement: a complex achievement, variously describable, is conceptualized in various and rival ways, under the auspices of a concept one of whose functions is to praise or accredit that achievement. However, with regard to the Rule of Law, I am now suggesting that the contestation is not so much about ways of characterizing an achievement that we already have, but rather about ways of answering a challenge - how to make law rule, rather than men - which many are convinced cannot be answered. The problem is conceived as urgent, and for that reason, the solution - if a solution could be found - would be highly valued. But much of the contestation is between rival proposals for responding to this urgent challenge: we respond to it with the separation of powers, or with an emphasis on common law, or an emphasis on rules, or on secondary rules, or the aura that surrounds constitutional norms, etc. In Gallie's original exposition, essential contestability was associated with the existence of an original exemplar, whose achieve55

Dicey, supra note 20, at pp. 268-269; Hart, supra note 32, at pp. 91 ff.

The Rule of Law and the Separation of Powers

138

158

JEREMY WALDRON

ment the rival conceptions sought to characterize and develop. For example, disagreements about the meaning of "the Christian way of life" are held together by reference to the Gospel accounts of the life of Jesus of Nazareth; each conception purports to provide the best account of what that complex exemplar stands for. 5 6 But I am suggesting, now, that reference back to the achievement of an exemplar may be too narrow an account of what gives unity to a contested concept. Perhaps there is no exemplar of the Rule of Law, but just a problem that has preoccupied us for 2,500 years: how can we make law rule? On this account, the Rule of Law is a solution-concept, rather than an achievement-concept; it is the concept of a solution to a problem we're not sure how to solve; and rival conceptions are rival proposals for solving it or rival proposals for doing the best we can in this regard given that the problem is insoluble. I have mentioned two levels of complexity: (i) the various precepts and principles on Fuller's laundry list; (ii) the different ways there are - of which Fuller's is one - of trying to solve (or respond with a second-best solution to the insolubility of) the problem of designing a political system in which the laws rule rather than men. There is also a third level of complexity: (iii) the several values which arguably might be served by the Rule of Law. Some theorists associate the Rule of Law with respect for fairness and human dignity; 57 others associate it with the provision for an environment hospitable to freedom; 58 still others see it as purely instrumental value, having to do with the effective pursuit of whatever other goals one is trying to use law to promote. 59 This provides an interesting variation on the account of essential contestability that I have taken from Gallie's work. Gallie's account involves contestation concerning the internal complexity of a valued achievement: the idea is that we know we value the achievement, but we are unsure what it consists of. An opposite form of contestation would be this: we know broadly what the achievement would consist of, but we are unsure why we value it. Moreover, different accounts Gallie, supra note 27, at pp. 17 6-181. See Finnis, supra note 44, at p. 273. 58 See F.A. Hayek, The Constitution of Liberty (London: Routledge and Keg an Paul, 1960),pp. 142-143. 59 For the contrast between instrumental and non-instrumental conceptions of the Rule of Law, see Radin, supra note 48, at pp. 784 ff. 56 57

The Rule of Law and the Separation of Powers IS THE RULE OF LAW AN ESSENTIALLY CONTESTED CONCEPT?

139

159

of the value of the practice give us different bases for elaborating and developing the practice. 60 In the case of the Rule of Law, I think both forms of contestation are present. There is contestation about the content and requirements of the Rule of Law ideal, and there is contestation about its point. The two forms of contestation of course feed off one another. To put it slightly differently, the lead idea of the Rule of Law is that somehow respect for law can take the edge off human political power, making it less objectionable, less dangerous, more benign and more respectful. But we disagree on how this can be done, and whether it can ever be done completely. We also disagree on the precise nature of the danger posed by human power in its unmitigated form, and on the values that would be served by introducing law into the picture. We disagree about the ailment, the medicine, and the character of the cure. All this makes for an extraordinarily complicated package of theoretical contestation.

VII. FINALLY, THE STREETS OF FLORIDA

It would be silly to portray the likes of James Baker and David Boies as political philosophers engaged in theoretical contestation about the essence of the Rule of Law. They used the slogan when it suited them, and there is little evidence that they had time or patience for much philosophical reflection on the nature or content of the concept they were invoking. They invoked it as though it were perfectly clear what it meant, and as though no one could imagine it being invoked on any other side of the dispute in which they were currently participating. I guess some of those who talked about "the Rule of Law" were a little more thoughtful than that. 61 But for the most part the 60 This, I think, is what is going on in Ronald Dworkin's three-step account of interpretive concepts in Law's Empire (Cambridge: Harvard University Press, 1986), pp. 46-49 and 65-68: first we identify a certain practice; then we attribute a point to it; and finally we revise the practice in the light of that value-attribution. 61 Justice Stevens' dissent in Bush v. Gore, supra note 2, offered some brief reflection, though he, too, wrote as though the Rule of Law ideal were straightforward and as though it were under straightforward attack. This was also true of the vast majority of newspaper and news-magazine commentaries. (An exception was Steven Thomma's account of conservative and liberal versions of the Rule of Law, cited supra in note 13.)

The Rule of Law and the Separation of Powers

140

160

JEREMY WALDRON

use of the phrase on the streets, in tones of categorical requirement ("We must accept the Rule of Law") and categorical denunciation ("They are violating the Rule of Law"), seems to belie a lot of what I have said about its essential contestability. Gallie thought that a concept was essentially contested when the contestedness was understood to be part of the very meaning of the concept. 62 His idea was that someone who does not realize that democracy, for example, or art are sites of contestation really doesn't understand the concepts he is invoking. As we have seen, this need not prevent a person from putting forward a firm view about the concept or taking sides in the controversy about its meaning. But anyone who says that one of these concepts has a perfectly clear meaning and that he cannot see why so many people get it wrong, shows (on Gallie's account) that he himself doesn't really understand what he is talking about. Now I think this suggestion needs to be qualified. As I said at the end of section 2, I don't think we should rush too precipitously, even in Florida, to the position that partisan political uses of "the Rule of Law" are meaningless pieces of ideological abuse; we certainly shouldn't do so just because the political partisans have not read their Gallie and show no understanding of essential contestability. Maybe we need to elaborate more carefully the relation between theoretical and non-theoretical uses of an essentially contested concept. We know, first, with regard to any piece of conceptual analysis that there is likely to be a loose and informal fit, at best, between the theorist's specification of necessary and sufficient conditions, etc. and the use of the concept on the street. The use of almost any concept on the street is a form of know-how63 - the street user will say things like "I know how to use it," "I know it when I see it," etc. - and that know-how may not be reducible even at the back of the user's mind into the form of propositional analytic knowledge that the concept has such and such implications and presuppositions. Still, unless street usage is horribly chaotic, the analyst can test his propositional knowledge against the user's know-how. Gallie, supra note 27, at p. 169. Cf. the distinction between knowing how and knowing that in Gilbert Ryle, The Concept of Mind (London: Hutchinson, 1949), Ch. II. 62 63

The Rule of Law and the Separation of Powers IS THE RULE OF LAW AN ESSENTIALLY CONTESTED CONCEPT?

141

161

Sometimes the user's know-how will be meta-theoretical: the user will show that he knows how to hesitate in a borderline case, for example. (He knows there are some things that are sort of green and sort of blue.) But this may not be so in the case of all meta-theoretical diagnoses. Some such diagnoses may take place only at the level of theory. Moreover, in the case of an essentially contested concept, there are two levels of theory to which ordinary street usage might orient itself (and vice versa). There is (a) the theorizing involved in the construction of a particular conception of a contested concept, and there is (b) the theorizing involved in the diagnosis of the concept as essentially contested. It is probably unrealistic to think that street-level usage involves anything much more than a glimpse of (b). The street user may show himself- to the theorist's eye- as an informal partisan of a particular conception of the Rule of Law; and the theorist will appeal to the street-user's know-how to test his analytic account of that particular conception. But the street-user's occasional use of "the Rule of Law" (or whatever the concept is) may show little awareness of the fact that this well-worked-out conception is challenged by other equally well-worked-out conceptions of the concep,t in the sort of rivalry we have described as essential contestability. The ordinary user may be puzzled and dismayed by the fact that other people seem to be citing "the Rule of Law" as though it meant something else altogether. But since he hasn't thought much about the well-worked-out conception that corresponds to his own use, since his use of the concept is sporadic anyway, and since he has little time for fancy theory, he is unlikely to dwell on the relation between the wellworked-out conception that corresponds to his informal use and the well-worked-out conception that corresponds to his opponent's informal use. Indeed the situation may be even untidier than this: the ordinary user may have little sense of which side he's on so far as these rival conceptions are concerned, and he may draw on various ones from time to time with little concern for their analytic consistency. This, I think, is what we should expect if a concept is essentially contested. People always use concepts with varying degrees of self-consciousness and reflectiveness. With an essentially contested concept, there are more layers available for reflection - for example,

The Rule of Law and the Separation of Powers

142

162

JEREMY WALDRON

the layer at which rival conceptions present themselves explicitly as such, and the layer at which the theorist becomes aware of how their rivalry enriches everyone's sense of what is going on- but just for that reason, there is likely to be a greater variety of reflective involvement in the use of the concept. The verdict of essential contestedness does not stand or fall with everyone's being aware of it; it stands or falls with the way it helps us understand all that goes on at the various levels at which people use a concept and reflect upon their use of it. I said in section 5 that, in Gallie's account, one of the things that distinguishes an essentially contested concept from concept-talk that is just "radically confused" is the fact that contestation between rival conceptions deepens and enriches our sense of what is at stake in a given area of value. Gallie describes this as a process taking place in the context of contestation among theorists who are aware that the concept they are contesting is essentially contested. But perhaps that awareness need not be present in order for the deepening and enriching effect to accrue. Maybe our understanding of the Rule of Law (of the values and possibilities associated with it) is deepened and enriched even by vigorous debate among partisans, each of whom believes - if he gives any thought to the matter at all - that the Rule of Law has a single, clear meaning. For example, someone who starts off convinced that the Rule of Law means that politicians must submit to the discipline of clear constitutional rules may be horrified when he hears someone else saying that the Rule of Law is actually promoted by people's willingness to engage incessantly in litigation. But as the argument goes on, he may modify his position somewhat, or those who listen to the dispute may come away with a somewhat more complex position than either of the ones held by the two disputants - and as a result subsequent interventions by these or other parties may be more thoughtful and sophisticated. Of course there's no guarantee that this will happen; in some cases, disputation might make things worse, not better. My point at this stage is just to emphasize- against Gallie's suggestion- that disputation can make things better whether or not the participants are in position to associate that process with anything like the idea of essential contestability. Perhaps it is best to say, then, that we should call a concept essentially contested when we find that contestation

The Rule of Law and the Separation of Powers IS THE RULE OF LAW AN ESSENTIALLY CONTESTED CONCEPT?

143

163

about its definition helps deepen and enrich our sense of what is at stake in a given area. We should not suppose that this deepening and enriching effect depends upon a prior characterization of the concept as essentially contested or that it depends upon the parties accepting such a characterization. Their arguments play a part in the process whether they do so self-consciously or not. As it happened, the Rule of Law was invoked in Florida on just the occasions one would expect it to be invoked if it were (at some level) essentially contested along the lines I have outlined. As I said in section 4, people appealed to the Rule of Law to defend and to criticize legally authorized exercises of discretion by political partisans, to defend and criticize the use of rules rather than standards as a basis for recounting votes, to defend and criticize innovative judicial decisions, and to defend and criticize the American tendency to litigate and go on litigating. No doubt it was all over-determined, and it is easy enough to say, as many commentators said at the time, that a lot of it was simply self-serving rhetoric. Still, that verdict is not inconsistent with the possibility that our understanding of the Rule of Law - or of the difficulties associated with the Rule of Law - was in fact enriched and deepened in the context of partisan, perhaps even cynical invocations of the ideal in Florida. (By the way, it would be consistent, too, with Justice Steven's lament that public confidence in the judiciary as guardians of the Rule of Law has declined as a result of the decision in Bush v. Gore; that might be true, and still people might come away with a better sense of the complexity of the ideal and of the nature of the challenge that it poses.) This optimistic assessment might be taken one step further (though I'm not sure whether I want to go this far or not). I can imagine someone saying that the fact that this dispute was fought in the context of contestation about the Rule of Law is itself a tribute to the presence of the Rule of Law in American politics. The idea would be that the Rule of Law is actually served by people asking and arguing about what counts as the Rule of Law in regard to some decision or set of decisions that the Rule of Law is supposed to control. (Analogously, we might say that we are sure we live under a Constitution in the United States, precisely because we argue so much about what the Constitution means.) There is something to

The Rule of Law and the Separation of Powers

144

164

JEREMY WALDRON

this. A society ruled by laws, not men, is bound to be a society in which there is constant debate about what the Rule of Law means. This is so not just because the Rule of Law is an essentially contested concept, but also because law and legalism are inherently garrulous and self-reflective practices. It is part of law and the practice of law to reflect and wonder what law is, 64 and for this reason it is part of any evaluative ideal involving law - such as the Rule of Law - that it will have this self-referential argumentative presence in society. We must be careful, though, not to confuse necessary and sufficient conditions. The Rule of Law probably cannot exist in a society unless people engage in constant argument what the Rule of Law amounts to; but it doesn't follow that the sheer fact that they engage in such argument means the Rule of Law exists. One can imagine a sort of fake legalism, which concocts argumentation about the Rule of Law purely to lend political decision-making a spurious legitimacy under this head. Still, when it seems not only that such argumentation is taking place but also that the outcome of the whole debacle somehow turns on it in a non-preordained way as in Florida- then there might be greater reason for optimism. Despite all the chaos and confusion of rival invocations of the Rule of Law, there certainly seemed to be a sense abroad that whoever won that argument- unwinnable as it might seem from the theorist's point of view- was entitled to win the election. I don't think we should denigrate that sense just because we, as philosophers, are irritated by the untidiness and unsophistication of the parties' invocations of the Rule-of-Law ideal. Columbia Law School 435 W. 116 Street New York, NY 10027 USA

Cf. Dworkin, Law's Empire, supra note 60, at p. 90: "[N]o firm line divides jurisprudence from adjudication or any other aspect of legal practice.... So any judge's opinion is itself a piece of legal philosophy, even when the philosophy is hidden and the visible argument is dominated by citation and lists of facts. Jurisprudence is the general part of adjudication, silent prologue to any decision at law." 64

Part II The Rule of Law and Judicial Discretion

[6]

Freedom and the Rule of Law The first of two talks by F.A. Hayek

0

NE of the first things the foreign student of English history must learn is that the English rarely change both the character of an institution and its name. A change of name does not necessarily mean that the institution has changed and a name may long persist after the institution which it described has vanished. This may become the cause of dangerous illusions, and I want here to discuss one such illusion. Englishmen are convinced that their individual liberty is protected by a tradition which is summed up by the phrase 'the rule of law '. But there can be no doubt that the rule of law was a peculiarly British institution, the foundation of that liberty of the Englishman which was once the envy of the whole world and which most western countries have attempted to imitate. It has long rightly been regarded as so essentially a British tradition that it must seem little short of impertinence for a person of foreign origin to question whether it still prevails in Britain. But the very fact that the name has been retained in current use appears to have prevented people from recognising that it no longer describes the same thing. And, as I shall show later, the very experiences which other countries have had in their attempts to imitate British institutions have sharpened the eyes of the outside observer to dangers which the British themselves are apt to overlook. The ideal of the rule of law that emerged from the struggles of the seventeenth century aimed at a limitation of government and legislation as such. When, today, we read in a widely used legal work that in Great Britain ' the rule of law is whatever Parliament as the supreme lawgiver makes it ', it has obviously ceased to have that meaning. The rule of law in the original sense meant much more than complete legality of government. The acts of a dictator may be strictly legal; but though it has been seriously contended for this reason that the rule of law was preserved even under Hitler, this would evidently not be true in the sense of the ideal that was contrasted with arbitrary government. If a government is given· authority to do whatever it regards as desirable, every act of such a government is legal, but this does not mean that it will act under the rule of law. This is putting the crucial issue in extreme terms. But an ideal may lose its power by slow attrition as well as in consequence of a revolution. What I shall be COI:lCerned with in these two talks will be to show that the slow cumulative change which in this field has taken place in this country over the past generation amounts to little less than a revolution. The consequences of this revolution are not yet generally recognised, and whatever discussion there has been has concerned such apparent technicalities as the limits of.Ministers' powers and the functions of administrative tribunals. These are the issues with which, ultimately, I shall be concerned. But in order to bring out their significance I must first try to show what the rule of law once did mean in this country and how other countries have met certain problems created by the modern developments of government.

The Rule of Law and the Separation of Powers

148

2

The manner in which the rule of law limits· the coercive powers of government is by tying these powers by previously announced general rules. Its aim is to limit coercion as far as possible. We can never eliminate coercion entirely, because the only way effectively to prevent it is by the threat of coercion. But we can make coercion in a high degree avoidable by placing the monopoly of coercion in the hands of the state and restricting it there to instances which are defined by general rules. These can then be taken into account by the individuals in making their plans. The evil of coercion is that it prevents the individual from .using his own gifts and knowledge to the best effect. If he knows in advance what he is allowed or will be required'·to do, he can shape his plans accordingly. It used to be the boast of free men that so long as they kept within the boundaries of the known law, they needed to ask nobody's permission and need obey nobody's orders. Is there anybody left today who can make this claim? If the difference between mere legality and the reign of law is today little understood, this is in large measure because we call every enactment of the legislature a law, while in the concept of the rule of law the term is used in an older, narrower sense. In this sense a law is, as Blackstone described it, ' a rule, not a transient sudden order from a superior or concerning a particular person; but something permanent, uniform and universal'. The greater part of the laws which Parliament now passes every year are in this sense not laws at all but particular instructions which it issues to the servants of the state about the manner in which the whole machinery of government and its property are to be administered. Though everything the government does is likely to affect the private citizen in some manner, the laws which direct the governfllent machine do not bind him in his actions. In a free country the private citizen is not supposed to be an object of administration, not to be subject to any orders or commands other than the general abstract rules which apply equally to everybody. The ideal that the ordinary citizen should be subject to coercion only under such conditions and in such a manner as is required by previously existing general laws demands· that the acts passed by the legislature satisfy cerqtin requirements. Any law which gave an authority powers to order citizens about in whatever manner it deemed necessary to achieve a particular purpose, would infringe the 1ule of law, though it would make such action perfectly legal. It is p•Jssible in a perfectly legal or constitutional manner to set up what in effect is, in a phrase which derives from Aristotle, a government of men and not a government of law-a phrase which received general currency in the great constitutional debates of the seventeenth century. If this is to be avoided the legislator himself must be guided by certain general principles about what the law ought to be. Yet the ultimate legislator cannot be bound by any laws in the technical sense of the word law, because he can change any law. The principles which must guide the legislator if the rule of law is to be preserved must therefore be metalegal principles, principles outside the law itself; they must be of the nature of a commonly accepted political ideal, a firm tradition which prevents the legislature from infringing the rule of law by the laws which it passes.

The Rule of Law and the Separation of Powers

149

3 Preventing Arbitrary Coercion Differently expressed, the aim of all this is to prevent arbitrary coercion. A man is free if he need not obey the arbitrary will of any other person; and arbitrariness, again, is not a question of whether the source of the order or its purpose are legitimate, but of whether or not the same requirement applies under the same circumstances to all· other people. A constitutional legislature may be as arbitrary as any tyrant if it unjustly discriminates, if, because of certain aims it wants to achieve, it requires one man to do or omit specific things which are not exacted from all according to some general rule. Nobody doubts that a bill of attainder or all retro-active legislation are as much arbitrary interference in individual affairs if they are passed by Parliament as would be true of a similar act of an autocrat. If the rule of law is to prevail, the private citizen must not be an instrument of governmental policy, required to serve its momentary purposes, but must be free within the kno·wn rules to pursue his own ends undisturbed by unpredictable interference of authority. The known general laws which tell the private citizen in which circumstances and in what manner the government will use coercion have to him much the same significance as the known laws of nature; and need restrict his freedom no more than his knowledge of the natural effects of his actions. Most law3 say that if he acts in such-and-such a manner, certain things will happen. If he enters into a contract or damages another person's property, he knows that he will be constrained to make good the obligation incurred. He similarly knows that liis partner in a contract will be made to fulfil it too. The laws of the state give him the same kind of information about what will happen as the consequence of his actions as the laws of nature, and they assist him in the same way to make plans which he can be confident to carry out. So far as his capacity of shaping his own life is concerned, there is no fundamental difference between the knowledge that if he builds a bonfire on the floor of his living room his house is likely to burn down, and the knowledge that if he sets his neighbour's house on fire he will find himself in gaol. The main purpose of the laws of the state is indeed to secure him a known free sphere within which he can act without having to fear that anybody will interfere with his plans. The state can achieve this only by threatening coercion to those who do not observe those lines of demarcation. But since in these instances coercion i:; the consequence of actions which are within the choice of the individual, he can avoid ever being coerced. Unavoidable Duties But though the majority of the laws of the state inay be of this kind, not all are. The state also must require some actions from the citizens which are not avoidable. The most important instance of this is the payment of taxes; and conscription in peace time is another. Here the citizen has indeed no choice. Yet the fact that these duties are known long in advance means that he can provide for them in all his plans, that they constitute as much a datum to which he can adjust himself beforehand as the change of the seasons or the necessity of keeping his roof in repair or clearing the snow from his doorstep. He can

The Rule of Law and the Separation of Powers

150

4 still in the known conditions shape his own life as he thinks best, is not dependent on anybody's arbitrary orders, and can pursue his own ends without fear of interference. Even when such duties are not only unavoidable but can in the nature of the case not be predicted, such as the duty to serve on a jury· or to help prevent a crime, or of military servic-e in war-time, we take good care not to leave the decisions in the arbitrary will of any person but to make them a risk which threatens all people equally, and so far as possible we compensate people fully for any damage they have suffered in the service of the community. This traditional English ideal of freedom under the law, in which the individual needs to obey nothing but the known law, is the only kind of freedom which men living in society can aspire to. That there can be no conflict between law rightly understood and freedom, but that law is an essential condition of freedom, was clearly seen by most of the great liberal thinkers of the past, from Cicero to John Locke. This insight was greatly obscured by the phantasma of ' natural liberty ' on which much of the utopian thought since the eighteenth century has been based. But liberty in civilisation cannot be natural liberty; it must always rest on institutions which protect the individual against coercion-against the coercion by other individuals as much as by the state. Since the state has successfully claimed the monopoly of coercion, the protection against coercion by the state has become the main problem of preserving individual liberty. Yet though liberty in civilisation must always be an artefact, a product of suitable institutions, this does not mean that it always has been or is even likely to be entirely the result of rational design. Indeed the institutions which secure freedom, like everything which freedom itself has created, have been mainly the product of a slow growth, the result of a step-by-step evolution and not of deliberate creation. It is just because they have not been rationally contrived but have gradually grown .that their role and significance are so imperfectly understood. As Sir Winston Churchill has recently reminded us, 'the liberties of Englishmen rested not on any enactment of the state, but on the immemorial slow-growing custom declared by juries of free men who gave their verdict case by case in open court '. This has many advantages but also its dangers. Though the original establishment of the rule of law was an entirely British achievement, it always remained a somewhat unsystematic affair, often secured only by somewhat vague traditions and beliefs. The interpretation of its meaning was largely left to foreigners, Frenchmen, Americans, and later Germans, who either endeavoured to explain to their compatriots the secrets of British freedom, as Montesquieu did, or who attempted deliberately to create in their own countries institutions which would achieve what traditions in Britain had done. In adapting British traditions to their own circumstances they had to face problems which in Britain became important only much later, and devised solutions which are highly relevant to difficulties which Britain faces today.

151

The Rule of Law and the Separation of Powers

5

Adapting British Traditions Ahroad Not the least gain of these foreign developments is that they often explicitly embodied in legal institutions or maxims ideals which in Britain had been only political doctrines guiding evolution. Montesquieu, for instance, has often been accused of having misinterpreted the British constitution when he asserted that it was based on the principle of the separation of powers. Some degree of separation of the powers which make the law and those who apply it is indeed essential for the maintenance of the rule of law. Yet it was clearly only partially realised in the British constitution. There can be no doubt, however, that when Montesquieu wrote it had been for at least a hundred years a widely held English political doctrine which profoundly affected evolution. Similarly, to take a principle which is even more fundamental to the rule of law: it has often been alleged that English criminal law did not recognise the maxim expressed in the Latin tag nulla poena sine lege, i.e., that there must be no punishment inflicted unless a previous law provided for it. The maxim may indeed not have been a recognised part of the law of the land. But there is no doubt that it had for long been a clearly recognised political ideal. Indeed, as early as 1740 we find Dr. Samuel Johnson, reporting a speaker in the House of Commons, formulating it as clearly as could be wished. The words put into the mouth of a Mr. Campbell are: 'Where there is no law there is no transgression is a maxim not only established by universal consent, but it is in itself evident and undeniable, and it is, Sir, surely no less certain, that where there is no transgression there can be no punishment'. I shall have to say more about the whole complex of doctrines which together make up and secure the rule of law in my next talk. My aim here has been mainly to draw attention to the curious relation between the slow spontaneous growth in this country which has so long been ahead of the world, and the deliberate efforts by which other countries have tried to profit from this lesson.-Third Programme

The second of two talks by F.A. Hayek

I

N my first talk* I said that the rule of law implies a set of doctrines all designed to secure that coercion of the individual by the state can be used only where it is required by previously existing and generally known rules. With few exceptions these principles have all been developed in England. The two important contributions which we owe to America are a bill of rights which defines the protected sphere of the individual; and the assurance that the rule of law is not infringed by routine legislation. This is provided by the written constitution enforced by a supreme court which can declare Jaws unconstitutional. There is also one point on which England has curiously Jagged for a long time and where the defect has been remedied only in recent years: the right of the individual to sue the crown, as __ distinct from its personal agents, for damage done to him by illegal acts. But otherwise all the major principles have been firmly established in this country for over 200 years and are still an unquestioned part of the political inheritance of the country-with one major exception to

The Rule of Law and the Separation of Powers

152

6 which the later part of my argument here will be devoted. I will rapidly run through the more important of those principles I have not yet mentioned in order to come quickly to the one crucial point where the rule of law is seriously threatened today and where other countries have developed important safeguards which Britain is lacking.

Independent Judges and State Policy In theory, at least, it is still unquestioned doctrine that the law

ought to be general, equal, and certain, and that it ought to be administered by independent judges. This involves not only, as we have seen before, some degree of separation of powers and the recognition of the principle of nulla poena sine lege, but also quite generally that government cannot coerce the private citizen in the service of the momentary goals of its policy, but only where it is required by the general rules of law. Indeed, since Lord Camden's famous dictum of almost 200 years ago, that ' public policy was not an argument in a court of law', the independent judge is not supposed to be concerned with the particular ends the government is pursuing or even to know about them. Though we sometimes use the phrase that it is the policy of the law, for example, not to recognise contracts for immoral purposes, the word policy is here used to describe a general rule and not in the specific sense in which we rightly distinguish between law and policy . and feel that a judge has no concern with policy. I will not say anything here about all those procedural safeguards of the rights of private parties in any legal dispute in which English law has always excelled. But I hope nobody will think that I underestimate their importance when I add that their value depends on a full recognition of the rights they are intended to protect. This seems to me to be overlooked by all those who hope that by imitating the forms of court procedure the substance of impartial jurisdiction can be preserved. They forget that such jurisdiction presupposes the existence of recognised general rules by which the dispute can be decided. The real essence of the rule of law, the purpose which all these safeguards serve, is to restrict discretion or arbitrariness of the authorities in wielding their coercive powers. In the process of government discretion of many different kinds is exercised, and these different kinds of discretion are often confused. we use the term discretion first with regard to the powers of the judge to interpret the rules of law. But authority to interpret a rule is not really discretion in the relevant sense: the task of the judge is to discover what is in the spirit of the whole system of valid rules of law, a task of finding what precedent or the imperfect wording of the legislator have not explicitly decided but what should and could be expressed as a general rule. That the task of interpreting the law is not one of discretion in the strict sense appears from the fact that the interpretation of the law can be and usually is made subject to review by a higher court. Indeed, it is probably the best test of whether a decision is bound by rule and not left to the discretion of the authority that the substance of the decision can be made subject to review by such an independent body which needs to know only the existing rules and the facts which can be supposed to have been known to the parties.

153

The Rule of Law and the Separation of Powers

7 The term discretion is also generally used with reference to the relation between principal and agent as it arises throughout the whole hierarchy of government, from the legislature down to the lowest administrative organ. At every step of this hierarchy the question arises how much any particular agency is entitled to decide without referring to a superior authority. There can be no question that in this sense the administrative agencies of government, in managing the affairs of government proper, need as much discretion as the managers of any business. It can also not be denied that the jealousy of the legislature sometimes unduly hampers. the necessary discretion of the executive. But this, though very relevant to the problem of parliamentary control, has little to do with the discretion which affects individual liberty. The problem of discretionary powers in relation to the rule of law is not merely how to limit the powers of particular agents of government but how to limit 1he powers of government as a whole; it is a problem of the scope of administration in general. That the government, if it is to make efficient use of the means at its disposal, must be able to exercise a great deal of discretion, is not disputed. The important point is that the private citizen and his property are in this sense not an object of administration by the government, not a means to be used by the government for its purpose and in the way its agents think most expedient; it means that the only legitimate way of inducing the private citizen to use his resources for the common good is to make him obey general rules. Only in so far as administration involves interference with this private sphere of the citizen is the question of discretion relevant in our connection; and the rule of law requires in effect that the administrative authorities should have no discretionary powers of this sort. This strict limitation of the coercive powers of the administration to what the law explicitly laid down was secured in this country in the eighteenth and early nineteenth century by subjecting all such action to control by the ordinary courts of law. When this principle was adopted an elaborate administrative machinery did not yet exist and it is probable that strict adherence to it has long delayed its growth in this country. David Hume described the position about two hundred years ago, in these words: ' though some inconveniences arise from the maxim of adhering strictly to the law, yet advantages so much overbalance them, as should render the English forever grateful to the memory of their ancestors who, after repeated contests, at last established that noble principle'. Dicey on Administrative Law The probletn presented itself differently to those continental countries which in the early nineteenth century, at a time when they had already evolved a complex and powerful administrative apparatus, endeavoured to follow the English example and to establish the rule of law. Until then the law which governed this machinery, the administrative law, had been outside the scope of any judicial control; most of it existed only in the form of internal and often secret regulations which were concerned more with regulating the duties of the officials towards their superiors than with limiting their powers over the citizen. ' Administrative law ' in consequence in the nineteenth century became a by-word

154

The Rule of Law and the Separation of Powers

8 for arbitrariness and was not without justification regarded as the opposite of the rule of law. This view can still be found in a justly famous English work of the end of the century, A. V. Dicey's Law of the Constitution. That work for the last time restated the traditional English view of the rule of law and exercised great influence. By that time, however, Continental conditions had greatly changed and there had developed, particularly in France and Germany, new safeguards against the arbitrary powers of the rapidly growing bureaucracy, safeguards which England lacks to the present day. As Dicey himself recognised in his later years, the growth of the powers of the modern state had made the growth of a separate body of administrative law as inevitable in England as it had been elsewhere. But Dicey insisted that control by the ordinary courts was the only proper safeguard; by that insistence he probably has delayed more than assisted the achievement of an effective subjection of the administrative machinery under the law. The first inclination of the Continental countries, in particular in Germany where the development is most interesting although least known in this country, had been to imitate the English example and to entrust the control of the administration to the ordinary courts. But this proved soon to be both impracticable and ineffective where the main task was not merely to apply an existing law but to enforce and formulate rules or to elaborate principles which had been laid down by the law only in general outline. The chief reason for entrusting this task not to the ordinary courts but to new courts created for the purpose was that it required special knowledge and competence which the ordinary judges, trained chiefly in civil and criminal law, could not be expected to possess. But the administrative courts which in France and Germany received definite shape about 1870 were intended to be fully independent courts of law, entirely separate from the executive and different from the ordinary courts only by the kind of cases which came before them.

A Political Argument It is true that there had been also another and wholly different argument in favour of separate administrative courts: an argument which ~s irreconcilable with the preservation of the rule of law. This is the argument that the legitimacy of some administrative actions cannot be decided simply as a matter of law but will always involve questions of administrative policy or expediency. This argument has on the whole been successfully resisted in the shaping of the continental administrative courts but it has almost entirely governed the evolution of whatever separate administrative jurisdiction this country has so far had. According to this view administrative ·courts will be concerned with the aims of the government of the moment and therefore cannot be fully independent but must form part of the administrative organisation, subject to directions at least by its head. They are suited to ensure a uniform policy and proper subordination of the lower authorities to the intentions of the government, but they do not give to the citizen the protection of the rule of law. This can be achieved only by an independent judiciary which, whether separate from the ordinary courts of law or not, is concerned with the law only.

The Rule of Law and the Separation of Powers

155

9 If such courts are effectively to safeguard the rule of law and the rights of the individual, they must have power to decide, whenever an infringement of the private sphere of the citizen is alleged, not merely whether the authority concerned had a right to act where it did but also whether the particular decision taken was required of them by law. The protection by the courts could of course be made entirely ineffective if the law gave the administrative agencies wide discretion to do within a given field whatever they think expedient and the courts had power only to ascertain whether the action was or was not within the discretion allowed. The continental administrative courts were created at least with the intention that they should ultimately control even the discretionary action_ of the administrative authorities. And although a change in the political climate which occurred soon after the establishment of those courts prevented this ideal from ever being fully achieved, a considerable degree of control even of what the law put in the discretion of the authorities has been achieved. In Britain the development has been altogether different. Dicey's misunderstanding of the Continental arrangements had for a long time entirely discredited the conception of separate administrative courts. At the same time, and partly because the ordinary courts could not be expected to show much understanding of the complex problems of administration, legislation proceeded to exempt larger and larger ranges of administrative action from judicial review. This process has now gone on for a long time. When more than twenty-five years ago the then Lord Chief Justice raised the cry of 'New Despotism' many people still regarded him simply as a reactionary alarmist. But even then he was speaking of a development which had already gone on for a considerable time; he cited the case of a fellow judge who had pointed out before the first war that ' Parliament had enacted only last year that the Board of Agriculture in acting as it did should be no more impeachable than Parliament itself'. Since then the danger of this development has come to be widely recognised: when we find one of the latest Fabian Tracts entitled Socialism and the New Despotism the concern is clearly no longer a party matter. The situation has in recent years become serious. After a long period in which the idea of separate and independent administrative courts was held in contempt in Britain we have seen a proliferation of socalled administrative tribunals within the administrative machinery; these tribunals, though more or less bound to the· forms of judicial procedure, are in essence courts of administrators enforcing a policy and not courts of judges administering the law. This seems to be regarded as inevitable and desirable. In- one of the most widely used treatises on English administrative law we find it even represented as an advantage that such a tribunal ' can enforce a policy unhampered by rules of law and judicial precedents . . . Of all the characteristics of administrative law none is more advantageous, when rightly used for the public good, than the power of the tribunal to decide the cases before it with the avowed object of furthering a policy of social improvement in some particular field; and of adapting their attitude towards the controversy so as to fit the needs of that policy'. And, to show that this view too is not a party matter, we find it echoed in a

The Rule of Law and the Separation of Powers

156

10 Conservative Party pamphlet on the Rule of Law which argues of these tribunals that ' flexible and unbound by rules and precedent, they can be of real assistance to their Minister in carrying out his policy'. If this applies, as it largely does apply, to the interference of administrative agencies in what used to be the private sphere of the ·citizen, in what John Milton called • the root of all liberty, the power to dispose and economise, in the land which God has given them, as masters of family in their own inheritance ', this is of course complete denial of the rule of law. It means that while the Continental countries have gradually striven to bring their administrative law under the rule of law, the British have accepted administrative law in the very sense which they once held in contempt. An eminent British jurist justly commented on this, in a recent study of French administrative law; 'it is not impracticable', he said, 'as in England is by some supposed, that the executive should be effectively subjected to a rule of Jaw: the French have succeeded in the undertaking. It seems to me essential to the survival of any rule of law in England ', he continued, 'that the executive should be speedily subjected to some rule of law '. There is today in many parts of the world, and particularly in those countries which have suffered or have been threatened by a totalitarian regime, a strong movement under \Vay for strengthening the rule of law. It found expression last year, at the first international congress of jurists held at Athens, in the adoption of the Act of Athens in which jurists from forty-eight countries solemnly urged the maintenance of the fundamental principles of the rule of law. Yet in the country which has been its original home and for 300 years has been leading in its development, the urgency of preserving it is still little recognised. The crucial issue to which I have been trying to draw attention is still regarded by most people as a legal technicality with \vhich they have little concern. Yet unless it is generally understood that those apparent technicalities of administrative law concern the most basic issues of individual freedom, there is real danger that the country which has first achieved it may find that it has preserved less of the rule of law than most other countries of the western world.-Third Programme

[7]

JUDICIAL DISCRETION

*

Ronald Dworkin 1. To the layman a lawsuit or a trial is an event in which a judge determines a controversy by application of established principles, rather than new principles invented to dispose of the case. He knows that individual judges may fail this ideal of justice; but he believes such failures to be aberrations, their occurrence marking injustice rather than its opposite. To him judges should and in general do, in the words of the admittedly metaphorical maxim, find the law and not make it. The layman's respect for law is founded in large part on his view that this is a fair method of deciding controversies. The academic branch of the legal profession seems now fairly agreed that the layman is mistaken. His mistake consists in not realizing that in our, as in every mature, legal system there are cases in which the rules of law dictate no result and which therefore force the judge to choose a solution, that is, to exercise judicial discretion. The belief that judges always find law, and never make it, is a tenet of a dogma called formalism, itself a result of the desire for something old-fashioned, unattainable, and bad, called mechanical jurisprudence. Some few lawyers have thought that the very idea of a legal rule which judges could apply to reach decisions was a myth. But the position presently in vogue insists rather that there are two sources of judicial decision: rule and discretion. It is thought a fallacy to ignore either of these sources, as the layman does in not allowing for choice. There remains controversy as to whether this duality in judicial technique represents only a necessary adjustment to human limitations or is itself an ideal of the judicial process, and as to its consistency with the theory of democracy. But the general proposition, that the exercise of judicial choice or discretion within areas circumscribed more or less tightly by rules is not an occasional misfiring but a characteristic feature of the legal process, is today almost a law school cliche. 1 *To be presented in a symposium on "Philosophy of Law" at the sixtieth annual meeting of the American Philosophical Association, Eastern Division, December 28, 1963. 1 A list of distinguished subscribers would include such different figures as Sir John Salmond (Glanville Williams, Salmond on Jurisprudence, 11th ed., Sweet & Maxwell, 1957, p. 44); Roscoe Pound (An Introduction to Philosophy of Law, Yale Paperbacks, 1959, esp. pp. 59-64); Benjamin Cardozo (The Nature of the Judicial Process, Yale Paperbacks, 1960, esp.

158

The Rule of Law and the Separation of Powers

SYMPOSIUM: PHILOSOPHY OF LAW

G25

Words like 'discretion' and 'choice' have different senses, and it is conceivable that some who say that judges have discretion mean simply that judges have d'ecisions to make, not already made by others for them, or that judges must reason or make judgments of one sort or another in making these decisions. The professional opinion of which I speak uses 'discretion' in a more characteristic sense to mean that the judge must sometimes reach his decision by means other than the application of standards, that such standards sometimes leave him free to choose. 2 I propose to argue that this professional opinion, grounded on the realization that some of the reasons courts give for decisions do not operate like rules, and that some of these decisions are difficult and controversial, is nevertheless an inaccurate and misleading way to describe these facts, and that the layman's opinion, while deceptive insofar as it may suggest that adjudication is simple, is, so far as it goes, closer to the truth. 2. In legal texts, law review articles, and popular presentations of the law, certain propositions like ''.An offer has been accepted, and can no longer be withdrawn, when an acceptance has been posted by the offeree,'' are described as rules of law. Some of these rules are described as statutory, meaning that a competent legislature has enacted them, others as common law rules, meaning that they have been developed by the courts apart from such legislation. Such' "textbook" rules are often given by courts as reasons for deciding a case or part of a case one way or another. It is possible, in some cases, to cast the entire argument of the court's opinion in the form of one or more syllogisms, in each of which Lecture III); Morris R. Cohen (Reason and Law, Collier, 1961, p. 11); Karl Llewellyn (The Common Law Tradition: Deciding Appeals, Little Brown, 1960); and H. L. A. Hart (The Concept of Law, Oxford, 1961, esp. pp. 124-132). I believe Hart's book to be one of the most important and provocative volumes in the literature of legal philosophy. As will appear, I adopt what I believe is a central thesis of that book to suggest that Hart's statements about judicial discretion and choice are less well taken. 2 The writers cited in n. 1 supra and other authorities speaking of judicial discretion do not mean that a judge should ever decide whimsically. They speak of judicial traditions or craft or other restraints which "limit" or ' 'hedge'' that discretion. And they speak of the ''sound'' judge who will exercise his discretion so as best to serve his society. But by speaking of a judge so constrained and so dedicated as nevertheless exercising discretion, or legislating, they suggest not simply that he must make predictions or judgments in applying standards, but that, in the area of discretion, he is not bound to apply particular standards at all.

The Rule of Law and the Separation of Powers

626

THE JOURNAL OF PHILOSOPHY

the major premise is such a rule, the minor premise a statement of fact either agreed upon by the parties or determined in the proceedings, and the conclusion a statement of the final or an interim decision in the case. Such cases may be called ''textbook cases.'' Two major arguments, very different one from the other, have been developed to show that the layman's view of law is wrong. The first-the "radical argument"-asserts that the layman would be wrong even if all cases were textbook cases. The second-the "argument from hard cases"-asserts that he is wrong because some cases are not textbook cases. 3. The fact that a judge's opinion is not inconsistent with a textbook rule does not indicate, according to the radical argument, that the judge is in any sense bound by such a rule. A rule is called a rule only because and so long as the courts do follow it. A lower court judge who decides to overrule prior cases, or even to disregard a statute, might be reversed by a higher court or he might not. In any event, if the highest court of jurisdiction chose not to follow a rule the textbook said was established, the disappointed party would have no option save to accept the decision, and the text's author no option save to alter the rule, or at least add a footnote, in the next edition. The radical argument invites the conclusion that antecedent principles never determine the result of a case. Professor H. L.A. Hart in his recent book, The Concept of Law, provides what seems to me a persuasive answer to the radical argument. 3 To say that we are bound by rules is not necessarily to say that we shall be punished for deviations or that our deviant acts will have no authority. Being bound by a rule, at least from what Hart calls the "internal" point of view, is a matter of doing what the rule prescribes, offering the rule as a reason (in the sense of a justification) for doing so, and offering it as a reason (in the sense of a standard) for criticizing others who do not. By way of illustrating the fallacy involved in the radical argument, Hart invites us to consider the difference between the game of baseball, in which the players are bound by the umpire's (or other scorer's) decision, no matter how wrong it may be, and the imaginary game of Scorer's Discretion, in which the umpire is free to rule strikes and balls, players out or safe, or runs scored as he wishes. We should be able to tell, after observation, whether a game being played was baseball or Scorer's Discretion, and we might well 3

Hart, op. cit., pp. 138-144.

159

160

The Rule of Law and the Separation of Powers

SYMPOSIUM: PHILOSOPHY OF LAW

627

sum up the difference between the two games by saying that baseball had a great many rules, including the rule that the decision of the umpire is final, and that Scorer's Discretion had no rules except the rule that the decision of the umpire is final. Of course, Hart continues, we could imagine baseball umpires beginning generally to make rulings as they wished without reference to the rules so that baseball if it survived at all would become Scorer's Discretion. But because that might happen is no reason for saying that it has happened, and unless it does it would be deceptive to describe baseball by saying that the umpires are free to rule as they wish and are not bound by the rules. I have mentioned the radical argument, and outlined Hart's reply, because that reply establishes some important distinctions: Discretion is not implied by the fact that an official's decision may be final or nonappealable, nor by the fact that it may be controversial or wr.ong. 4. The argument from hard cases begins by tabulating the sorts of cases that are in one way or another unlike textbook cases. (i) In many cases a court is pressed to, and in some cases does, overrule a textbook rule, and substitute a new one. (ii) Even when, as is more often the case, a court is determined to follow a particular textbook rule if it applies, that rule may be so ambiguous that it is not clear whether it applies, and the court cannot decide simply by studying the language in which the rule has been expressed. (iii) Sometimes two textbook rules by their terms apply, and the judges must choose between them. In some such cases the need for choice may be disguised, in that only one rule is mentioned, but research (or imagination) would disclose another rule that the court could have adopted as easily. (iv) Sometimes a court itself will state that no textbook rule applies to the facts. Often the gap may be cured by what is called "expansion" of an existing rule, but sometimes a wholly new rule must be invented. (v) A large, and increasing, number of cases are decided by citing rules so vague that it is often unhelpful even to call them ambiguous: the critical words in such rules are 'reasonable', 'ordinary and necessary', 'material', 'significant', and the like. It does no good to say that questions as to the application of these terms are questions of fact, because disputes will remain, and decisions be required, even after agreement on what has happened. It will not do, the argument continues, to say that the judge decides these hard casC\'3 by Delphic derivation of the ''purpose'' of a rule, for the purpose of a rule is usually de1ts ex machina. Nor will it do to say that such hard cases are decided in accordance

The Rule of Law and the Separation of Powers

628

THE JOURNAL OF PHILOSOPHY

with meta-rules which reveal how textbook rules are to be overruled, interpreted, expanded, contracted, or invented, or standards like ''material'' applied. The explanations courts give in hard cases typically take the form of appeals not to rules but to policy goals or the requirements of justice, and lawyers and judges will often disagree about the propriety of such an appeal in a particular case. .At best, characterization of such explanations as meta-rules only postpones the problem one step : we would need meta-meta-rules to select among meta-rules. .At some point the regress must pe broken by the admission that at that point a judge does not apply rules, but rather chooses. The most that can be said is that attention to the explanations judges give at that point may reveal patterns of judicial behavior, self-adopted guides which will aid a careful observer in predicting decisions, in making them more '' reckonable'' or perhaps less '' unreckonable. '' 4 The argument from hard cases thus proceeds from a survey of what we might call (with some license) the "logical" behavior of the types of explanation that judges give for their decisions. The conclusion, that judges have discretion, is drawn from a comparison of the role of rules in law as it is with what that role would be if all cases were textbook cases. The argument can be summarized in terms of Hart's games: The rules or other reasons that judges offer do not always specify the correct legal result, as the rules of baseball specify the correct baseball result, of agreedupon facts. To the extent to which law is in this way less like baseball, it is more like Scorer's Discretion, and law thus lies somewhere between the two games, involving elements of each. 5. The standards of law do indeed differ in logical behavior from the standards of baseball, but not in a way that makes law more like Scorer's Discretion, and the ascription .of discretion to judges is a poor way to describe the differences. It will be easier, I think, to see the error involved if we consider two more imaginary games. In the first-"Limited Scorer's Discretion"-the rules of baseball obtain in full, except that the umpire retains discretion to call 4 This is the burden of Llewellyn, op. cit. It is interesting to compare this book with Hart's. Hart is concerned to show the way in which judicial behavior is governed by standards, but abandons his inquiry at the point where such behavior is no longer governed by rules. Llewellyn begins where Hart stops, but gropes rather for techniques of predicting what judges will do. If one combined Hart's concerns with Llewellyn's research, the result would be a discussion not of guides to reckonability, but of legal standards other than rules.

161

162

The Rule of Law and the Separation of Powers

SYMPOSIUM: PHILOSOPHY OF' LAW

629

any pitch not hit by the batter either a strike or a ball, as he wishes, without regard to whether or not the ball is in the strike zone. The umpire's discretion is limited, in that it can be exercised only in certain events, and can have only certain consequences on the score, depending on the circumstances of the game. The umpire might adopt particular goals for himself, such as making himself a hero to the crowd, which would guide him in his decisions and upon which the players might begin to rely in predicting these decisions. But the umpire's freedom to set and apply such personal goals would not further restrict his limited discretion, but rather evidence it. The second game-" Policies "-is more complicated. .All the rules of baseball are initially in force, but the umpires (called "indexes" in Policies) are required in each case to consider whether alteration of the pertinent rule would bring the game closer to the realization of certain fixed policies, or make it more consistent with certain fixed principles, and, if so, to make the necessary alteration before deciding the play. Such policies and principles are set out, with assignments of relative importance, in an official manual. The policies include reducing the number of injuries to players, allowing individual players of greater natural strength and physical coordination to become stars and thus public idols, encouraging managers to exercise strategy in guiding their teams, and making the game more exciting to the spectators; and the principles include one that unsportsmanlike conduct should be penalized and another that the rules should not discriminate in favor either of home teams or visiting teams. The iudex is required, after each decision made on the basis of an altered rule, to issue a short written statement justifying that alteration. Such statements are published and made available to all the indexes and to the interested public, and the altered rule becomes binding until further changed in like manner. The iudex would have a difficult job. The policies of managerial strategy and crowd interest would ordinarily act as a brake on changing the rules, but the iudex could not refuse every suggested change on that basis. Some policies would be easier to apply than others: it would be easier to discover which sort of acts led to injuries than which players had better coordination or which situations a crowd would find exciting. Conflicts in policies or principles would present problems. When the manual indicated that one of the two competing policies was more important than the other, a decision might be justified that advanced the more important, but if failure to apply the less important policies would

The Rule of Law and the Separation of Powers

630

THE JOURNAL OF PHILOSOPHY

seriously undermine it, this might be a good reason for preferring it to the more important. The policy of reducing injuries might require that rules be framed so that players would be deterred from dangerous moves, such as running too close to the fence. The iudex would find it difficult, however, to justify picking any specific distance from the fence as a limit dictated by the policies-it would all depend on the speed with which the player was running, the type of play involved, and so forth. The iudex might develop a rule that a batter is safe, even though the ball is caught, if the fielder comes "unreasonably" close to the fence to make the play. How close is too close would depend on the circumstances of the particular play, and on the pertinence of such circumstances to other policies besides that of reducing injuries. The principle of unsportsmanlike conduct might be thought to present special difficulties, since questions of morals seem involved here, and such questions are believed to be less ''objective'' than questions of injury prevention. Indeed, if there were not broad agreement within the community of participants and spectators as to what sort of conduct is unsportsmanlike, the iudex would find this principle comparatively useless in making his decisions. It would be a breach of his role to make decisions on the basis of his personal standards or the standards of one interested player or team, just as it would be to make decision on the basis of a personal or debatable notion of what sorts of anatomical events constitute an "injury." But if there is such broad agreement, these standards may be used even though the "validity" of such standards might not be demonstrable in the same manner as judgments about injuryproducing conduct and even though the morals of the mob are not necessarily the only, or the best, morals. It might be, of course, that the index would be correct in a particular decision as to unsportsmanlike conduct, even though his decision was initially unpopular, if he succeeded in showing that, because of motives or other facts or consequences that the crowd had not adequately considered, the conduct in question was contrary to broadly accepted requirements of fair play. 6. We might say that Limited Scorer's Discretion lies between baseball and full Scorer's Discretion, that it contains elements of both. But would it not be misleading to say the same of Policies, to say, that is, that LimitE'd Scorer's Discretion and Policies are alike in that in each the decision maker has limited discretion 1 Is it not the pertinent distinction between these two games that the iudex in Policies has 110 discretion?

163

164

The Rule of Law and the Separation of Powers

SY1Jf.POSIUM: PHILOSOPHY OF LAW

631

·when we speak of someone 's discretion in making decisions, we mean that the standards, if any, of the activity in question grant him the right to make any decision he wishes, and deny any other participant the right to claim a particular decision from him. We might mean by "limited discretion" that the standards grant the actor complete choice in some phase of the activity, as the bridge player has complete freedom to arrange the cards in his hand. We might mean instead that the standards grant him a certain freedom within limits, as the bridge player must follow suit, but is free to play any card of that suit he may have. Even when an act is free in principle, it may not be permitted in special circumstances or when the actor's motives are improper (this is what we mean when we say such discretion must not be abused). If we should ask for a justification of a decision in terms of particular standards, we shbuld accept as sufficient a showing that the act lies in such an area of discretion. The standards of Policies do not grant the iudex a choice, even a choice limited in any of these senses. On every play the participants are entitled to the "correct" result; on no play is the iudex entitled to decide as he wishes. At no point and in no sense would it be a justification of a iudex 's decision to say that the decision was within his discretion. One who insists that the iudex has discretion must be assuming either that the principles and policies of the manual are not standards at all, which is wrong, or that such standards are somehow defective, in that they allow discretion to seep in unwittingly through, as it were, a leak. But what could it mean to say that standards are defective~ We might look back again, for an answer, to standards that are not alleged to leak. The rules of baseball, like the rules of most other games, specify precisely which events are to have which consequences, and the events designated in the rules are the only events having any significance whatsoever. Officials applying such rules therefore concern themselves almost exclusively with making authoritative reports, rather than characterizations, of what has happened. Infrequent disputes about the rules themselves are settled by reference to an authoritative rule book. An umpire can refuse to apply an alleged rule in a situation it embraces only on the basis that the rule is invalid; if he finds it invalid he cannot then apply it a few plays later. Rules do not have the dimension of weight. In the event that an inconsistency were discovered in the rules, the conflict could not be resolved by applying the ''more important'' rule. Such a eonflict, or a material ambiguity, would be regarded as intolerable, as an emergency calling for resolution by stipulation, or, in organized games like baseball, by decision

The Rule of Law and the Separation of Powers

632

THE JOURNAL OF PHILOSOPHY

of associations or committees with explicit rule-making powers. Not all games work like this-children's games that "are made up as they go along" are different-but this "logical" behavior is typical of the rules of most games. The principles and policies given to the iudex work differently. They do not specify precisely the events that might invoke them, and any event that occurs in the game may have pertinence in their light. The iudex typically concerns himself with characterizations rather than mere reports of what has happened. Such standards do have relative weight, and this dimension will often be the reason for applying or refusing to apply them. A policy need not be called invalid in order to be disregarded, and, once disregarded, may form the sole ground of a rule change in the very next play. Situations in which antagonistic policies or principles are at stake do not evidence inconsistencies, but are rather contemplated by the arrangement. They do not call for emergency resolution outside the confines of the game : their resolution is part of the game. There are, then, differences in the sorts of reasoning appropriate to decision in terms of rules and decision in terms of other standards. 5 These differences are not merely differences in the degree to which such other standards fail to match the model of rules, for they have models of their own. If there were not, within the community, general agreement on what constitutes good as distinguished from bad reasoning in terms of such standards, the iudex, if not altogether incapacitated, would have to rely on his own personal views of good argument, whatever that might mean. In that event, participants would no longer be able to make claims to particular decisions, and the game would become Scorer's Discretion. It does not follow, since there are such general standards, that the iudex is relying on such personal views when his decision is difficult or debatable. Then, as in every case, he intends his reasoning to be based on public, not private, standards of good argument, and the community receives and criticizes his decision on that basis. The claim that policies and principles are "loose" 5 The ''logical'' behaviors of principles and policies also differ to some extent from each other. Compare the kinds of reasoning appropriate to a policy of eliminating unsportsmanlike conduct with those appropriate to a principle that such conduct is wrong and should be penalized. And there are other kinds of standards at work in law, if not in Policies, which arguably have still different "logical" behaviors: for example, "maxims" that are appealed to ''all else being equal.'' Details of these distinctions might be interesting, even profitable, although it would be silly to argue about classification for its own sake.

165

166

The Rule of Law and the Separation of Powers

SYMPOSIUM: PHILOSOPHY OF LAW

633

standards might mean merely that decisions thereunder are often controversial. That fact is consistent with the equally important fact that the iudex, no less than the umpire in baseball, is charged with finding, and the participants are entitled to have, the decision in every case that constitutes the best resolution of the stipulated principles and policies. 7. It might be objected that, if the iudex has no discretion, it must follow that business executives and generals have no discretion either. Each of these officials is under an obligation to make every decision so as to further a specific policy: maximizing profits in one case, winning a war in the other. Yet it seems strange to say that they never have discretion; we might end by saying that no one ever had discretion. Perhaps, after all, discretion means merely that the actor has a decision to make, a decision not already made by another, and has nothing to do with the presence or absence of standards against which the decision must be made. This objection points up the importance of distinguishing between standards (which constitute arrangements such as games and contests and which entitle the participants to make claims as of right to certain privileges and awards and to official decisions sustaining such claims) and binding policies that are not standards. The general in war owes a duty to his nation to make decisions that he believrs most advance the victory. Yet neither the nation nor any of his men nor the enemy can be said to be entitled as of right to any particular decision. The policy of victory does not also act as a standard for determining claims, rights, or entitlements. The stipulated standards of Policies do have this function. Compare the position of a manager deciding which relief pitcher to call with that of the manager awarding a prize, donated by the public, to the "most valuable player" on the team. A good manager will approach each of these decisions on the basis that only one decision is the right one (although there are important differences in the sort of justification of his decision and the sort of resolution of ''ties'' that would be acceptable). Yet it would be correct to say that he has discretion in the first case, and not in the second. To regard the iudrx as having discretion, because the businessman or general has discretion, would be to miss the very essence of the game, to miss the point that it is an arrangement of entitlements. This imaginary game is not, of course, thr only example of such an arrangement operating under standards other than rules: other examples include certain moral arguments and certain sorts

The Rule of Law and the Separation of Powers

634

'l'HE JOURNAL OF' PHILOSOPHY

of contests. It would be strange to say that a person who seeks to decide a troublesome question of conduct in terms of the moral standards of his community or who acts on a committee named to award a prize for the best paper submitted on a given legal subject had discretion, no matter how difficult or controversial his decision may be. 8. My purpose thus far has been to challenge the unarticulated premise of the argument from hard cases: the assumption that the only true standards are rules and that, if the explanations given by judges in hard cases are to be regarded as standards at all, they must be regarded as rules manques, deteriorated versions which cannot be applied without a measure of discretion. I suggest, although I cannot here demonstrate, that hard cases are decided in law, as in Policies, by the application of standards other than rules. 6 I am not suggesting of course, that law and Policies are isomorphic, that every significant institutional or formal feature of law finds a counterpart in the imaginary game. 7 The most important difference, for the present purpose, is that the judge has no authoritative manual exhaustively listing the standards against which he must make his decision. If the judge were free to adopt his personal preferences as legal standards, then indeed his decisions would be chosen. But he is not. He is subject to the 6 I am speaking throughout of eases in which the judge is said to "make law" by the exercise of his discretion. There are eases in law, as in baseball, where the officials are explicitly granted discretion-the judge, for example, in fixing sentences and in certain other remedial or procedural matters, and the umpire in calling official times out and in certain disciplinary matters. This discretion is exercised "within limits" and "cannot be abused," but we would not say that the participants were entitled to any particular decision on such a question. I have ignored throughout the question of whether the judge, the jury, or the index has discretion in determining ''what happened,'' not because I want to say they do, but because there is no space in which to consider the question. Similarly, I wish to exclude the role of the Supreme Court in deciding constitutional eases from the scope of this essay. There is distinguished opinion that this judicial function is unique, and considera· tion of the alleged differences would not be manageable here. 7 In what follows I am assuming one similarity: that law__, like games and contests, is an arrangement of entitlements, and I am attempting to offer a picture of judicial practice consistent with that assumption. I articular trial or which transferred sentencing powers to the executive. 28 It is far from clear whether these decisions would be followed in this country if the Westminster Parliament were to enact such legislation for the United Kingdom. Behind these legal uncertainties lurk doubts whether the separation of powers principle is reconcilable with the ~democratic commitment to majority rule. They have often been felt in France, where as already noted the separation principle is regarded as fundamental. 29 Ever since the i791 Constitution, reluctance to shackle the legislature has been largely responsible for the French rejection of the U.S. theory of "checks and balances" and of judicial review of legislation. This perspective has persisted until the present day. Arguably it has merely been modified, and not abandoned, by the institution in the Fifth Republic of the Conseil constitutionnel which is only entitled to assess the constituConstitutional Theory, ch. V, at 124. There are some exceptions: see in particular C. Munro, "The Separation of Powers: Not such a Myth", [1981] P.L. and T.RS. Allan, Law, Liberty and justice, ch. 3. 28 See Liyanage v. The Queen [1967]1 A. C. 259, Moses Hinds v. The Queen [1977] A. C. 195, Ali (Mohammed Muktar) v. The Queen [1993]2 A. C. 93. 29 See Vile, Constitutionalism and the Separation of Powers, chs. VII and IX. 26 27

The Rule of Law and the Separation of Powers

Separation ofPowers and Constitutional Government 605

tionality of laws before they are promulgated. It is not a constitutional court in the fullest sense. 30

The doctrine defended A number of points can be made in reply to the argument deployed by Sir Ivor Jennings. First, he was too sceptical about the possibility of a coherent allocation of functions. It is possible to define in general terms the legislative, executive and judicial functions, which are allocated by a constitution to particular bodies or institutions. What is crucial is that this distribution is enforceable by the courts. They are entitled to take the final decision whether in practice a function is to be regarded as legislative, executive or judicial. 31 Of course, Jennings was right to point out that the theoretical criteria for determining, say, whether a decision is properly to be characterised as "administrative" or "judicial" are unclear and that there are many borderline cases. But it is perfectly coherent to claim, for instance, that decisions on personal rights and liberties are inherently suitable for judicial resolution, and so must be made by a court, while the distribution of other goods and benefits may be regarded as a matter for administrative decision. This particular distinction is captured by Article 66 of the Constitution of the Fifth French Republic: "[t ]he judicial authority, guardian of the liberty of the individual, ensures respect for this principle in conditions determined by the law;'. As a consequence a statute giving the police wide powers to inspect vehicles on the public streets, was declared unconstitutional-in the absence of either adequate standards to guide the exercise of these powers or provision for judicial control. 32 Similarly, English courts have held that a statute should not be interpreted to allow a final decision on a matter oflegalliabilit¥, for example, to pay taxes, to be taken by an administrative authority. 3 These decisions show that even within the context of an unwritten, or uncodified, constitution, courts are able to draw a clear line between administrative and judicial functions. Equally, courts are able to draw lines between legislative functions on the one hand, and judicial and executive functions on the other. The legislative function is broadly the function of framing general rules applicable to a potentially unlimited range of circumstances. In contrast, Acts of Attainder and retrospective criminal legislation of the type held invalid in the famous Liyanage case clearly usurp the judicial function. 34 The practice of executive legislation under broad powers delegated by 30 See]. Bell, French Constitutional Law (1992), ch. 1, esp., 55-56, and A. Stone, The Birth of Judicial Politics in France (1992) esp. ch. 9. 31 For a robust statement of this view in the context of the U.S. Constitution, see M. H. Redish, The Constitution as Political Structure (1995), ch. 4, esp. pp. 117-119. 32 Decision 77-75 ofJanuary 12, 1977, GD, 340. 33 e.g., seeCommissionersofCustomsandExciseLtdv. CureandDeeleyLtd[1962]1 Q.B. 340. 34 In Liyanage v. The Queen [1967]1 A. C. 259 the Privy Council invalidated on separation of powers grounds legislation in Sri Lanka which amended the criminal law and the rules of evidence for the trial of plotters of an unsuccessful coup.

281

The Rule of Law and the Separation of Powers

282 6o6

Public Law

the legislature does not show that there is no dilference between the legislative and executive functions, but only that in some countries courts have tolerated a departure from strict separation of powers principles. In Germany, however, the Constitutional Court has ruled that it is for the legislature to lay down the crucial (wesentlich) principles in all major regulatory areas, particularly where the rights protected by the Basic Law are at issue. 35 In these circumstances it is unconstitutional for the Bundestag to delegate legislative authority to a federal minister, state government or a private body. The application of this doctrine, known as the Wesentlichkeitstheorie, may be controversial, 36 but it shows that a principled line can be drawn between legislative and executive functions. The rationale for this German doctrine is that it protects democratic values. Citizens expect their elected Parliament to take the most fundamental decisions which affect them, rather than transfer responsibility to a minister or authority which is not directly answerable to the electorate. 37 The separation of powers in this context, therefore, reinforces democracy, and is not, as one argument in the previous ·section suggested, inimical to it. The requirement that the legislature itself lays down the most important general rules is also reflected in a number of United States and French civil liberties cases, a topic discussed later in this article. 38 But perhaps a more significant point to make in reply to the critique of Jennings and Marshall is that the separation of powers is not in essence concerned with the allocation of functions as such. Its primary purpose, as we saw earlier in this article, is the prevention of the arbitrary government, or tyranny, which may arise from the concentration of power. The allocation of functions between three, or perhaps more, branches of government is only a means to achieve that end. It does not matter, therefore, whether powers are always allocated precisely to the most appropriate institution-although an insensitive allocation would probably produce incompetent government and run counter to Locke's efficiency rationale. This point is perhaps most clearly appreciated if we consider what has become one of the most complex areas for separation of powers analysis: the organization, and control, of administrative authorities and agencies. These range from bodies which allocate social security and 35 See in particular 33 BVerfGE 125, 158-159 (1972) (Parliament must lay down principles governing professional qualifications, though details may be left to professional associations) and 49 BVerfGE 89, 126-127' (1978) (Parliament must take fundamental decision whether to permit use of nuclear energy). The development of the principle is particularly remarkable, since the Basic Law, Art. 80, I, authorises the delegation oflegislative authority, provided the content, purpose and scope of the authorisation are clearly spelt out. 36 SeeM. Kloepfer, "Der Vorbehalt des Gesetzes im Wandel", (1984) 39 ]uristen Zeitung 685, G. Nolte, "Ermachtigung der Executive zur Rechtsetzung", (1993) 118 AoR 378, and D. P. Currie, The Constitution of the Federal Republic of Germany (1993), ch. 3, esp. 125-133. 37 49 BVerfGE 89, 124-126 (1978). 38 See pp. 607-611 below.

The Rule of Law and the Separation of Powers

Separation of Powers and Constitutional Government

283

607

welfare benefits (such as public housing), to regulatory bodies, for ex~mple, the Independent Television Commission and the Monopolies and Mergers Commission, and finally to supervisory or investigatory officers, such as the Comptroller and Auditor-General and the Parliamentary Commissioner for Administration (PCA). Now it can be asked whether these bodies perform legislative (or rule-making, to use the American term), administrative, or judicial functions. But these are impossible questions to answer. For in truth many agencies perform at least two, and perhaps all three, functions. This is apparent in the United States, where it is common for an independent regulatory agency to engage in rule-making, to formulate and apply policies, and to take individual decisions, often after a formal hearing. Perhaps in the United Kingdom the only authorities which consistently discharge all three functions are local authorities, which may make by-laws, formulate planning, highways and housing policies, and decide applications for planning permission which might be characterised as judicial, or at least quasi-judicial decisions. 39 But certainly many agencies, including government ministers, exercise a variety of functions, some of which can be characterised as executive and some as judicial. 40 Does this phenomenon mean that separation of powers analysis should be abandoned as hopeless? It would seem so, if the pure theory is adopted, with its rigid insistence that each function of government is discharged by a separate institution. But the answer may be quite different if we see the principle as essentially concerned with the avoidance of concentrations of power. For then questions may be asked about the relationship of the agency to the three traditional branches of government. 41 Does Parliament or the government have sole right to hire and fire members of the authority and its staff? Does the government have exclusive power to issue directions or guidance to the agency? If the agency takes judicial or quasi-judicial decisions, how far is it subject to review by the ordinary courts? On this approach there would be a violation of the principle if the executive were entitled, without assent of the legislature, to give detailed directions to an agency, and appoint its members, when that agency takes decisions affecting individual rights and judicial review is (virtually) excluded. That would not be because an executive agency carried out judicial functions, but because it was so structured as to create or reinforce a concentration of power in the hands of the government. Finally, the pure separation theory says little or nothing about those 39 For a discussion of the powers oflocal authorities and administrative agencies, see P. P. Craig, Administrative Law (3rd edn., 1994) ch. 3. Perhaps the Civil Aviation Authority to an extent exercises all three functions: see G. R. Baldwin, "A British Independent Regulatory Agency and the 'Skytrain' Decision", [1981] P.L. 57. 40 See Jennings, The Law and the Constitution, 290-294. 41 For a full exposition of this approach in the United States context, see the articles by P. L. Strauss, "The Place of Agencies in Government: Separation of Powers and the Fourth Branch", 84 Col. LR 573 (1984), and "Formal and Functional Approaches to Separation of Powers Analysis-A Foolish Inconsistency", 72 Cornell L.R. 488 (1987).

The Rule of Law and the Separation of Powers

284 6o8

Public Law

·functions of government which cannot easily be characterised as legislative, executive or judicial. I have in mind the work of officers such as the Comptroller and Auditor-General and the PCA, or of bodies as diverse as the Audit and Boundary Commissions. Apart from the last-named, these institutions have been established to oversee government and reinforce political and financial accountability. The Boundary Commissions are responsible for drawing up and revising election constituencies, and are of course required to act independently of government and party political pressure. None of these functions can easily be identified as fall..: ing within one of the traditional categories. But it is important that these officers, and the members of the Commissions, are not required to act under the total control of either the legislature or executive, for that would create the danger of a concentration of power or (to put it another way) a distortion of the balance of power between these two branches of government. In fact the rules in these areas are often quite unsatisfactory. For instance, the Boundary Commissioners' recommendations must be submitted to the Home Secretary. A draft Order in Council, which may contain government amendments to the Commissioners' proposals must be submitted to, and be approved by, both Houses of Parliament. But the legislature, unlike the government, has normally no opportunity to change the proposals, and there is very limited scope for judicial review of the Orders once they have been approved. The procedure in short gives ample scope for executive delay and manipulation in favour of the governing party. 42 A second example has been afforded by the method of appointment of the Parliamentary Commissioner for Administration. An officer of Parliament charged with the investigation of maladministration, he has been appointed by the Crown-although he may only be removed by an address of both Houses. 43 However, the government has recently accepted the recommendation of the Parliamentary Select Committee that he should be appointed by the Commons, on a motion of the Prime Minister with the agreement of the Leader of the Opposition and the Chairman of the Select Committee. 44 The new procedure will respect the balance of power in this context between government and legislature.

The partial separation theory The argument in the previous section has shown that the separation of powers should not be explained in terms of a strict distribution offunctions between the three branches of government, but in terms of a network of rules and principles which ensure that power is not concen· 42 For full discussions of these points, see de Smith and Brazier, Constitutional and Administrative Law, 260-264; H. F. Rawlings, Law and the Electoral Process (1988), 50-62; R. Blackburn The Electoral System in Britain (1995), 133-142. 43 Parliamentary Commissioner Act 1967, s. 1(2)-(3). 44 For discussion, seeP. Giddings and R. Gregory, [1995) P.L. 45.

The Rule of Law and the Separation of Powers

Separation of Powers and Constitutional Government 609

trated in the hands of one branch. (In practice the danger now is that the executive has too much power, though it is worth remembering that at other times there was more anxiety about self-aggrandizement of the legislature. 45 ) That does not mean that the allocation of functions is wholly irrelevant. I will explain in the next section of this article how in civil liberties cases courts may properly insist that general rules be made by the legislature and that the executive does not act without legislative authorization to deprive individuals of their rights. But the importance of a correct definition and allocation of functions should not be exaggerated. Madison for instance was not troubled by these questions, though nobody has argued so cogently for the separation of powers principle. Outside the context of court rulings in civil liberties cases, the principle is most frequently applied in the architecture of the constitution itself. Powers are allocated to different institutions. The legislature is normally divided into two branches, a procedure recommended by Madison on the ground that otherwise it would be too powerful. 46 Each branch is empowered to check the others by exercising a partial agency or control over their actsY That is why, for example, in the United States Constitution the Senate must give its advke and consent to the appointment of ministers, ambassadors and judges, 48 and the President may veto Bills passed by the House of Representatives and the Senate, subject to an override by a two-thirds majority vote in each Bouse. 49 It is not very helpful to ask whether in the former instance the Senate is exercising an executive power and whether in the latter the President acts as a third branch of the legislature. What is important is that there is a system of checks and balances between institutions which otherwise might exercise excessive power. As Madison put it in Federalist Paper 51, the structure of government should be so arranged "that its several constitu.. ent parts may, by their mutual relations, be the means of keeping each other in their proper places."

Implications of the separation of powers It is worth exploring some of the implications of commitment to the separation of powers found in other legal systems before turning to the significance of the principle in the modern United Kingdom constitution. For it has been argued that the principle is too ambivalent to be useful. It is also said that it has been compromised by, for example, the need to tolerate extensive executive law-making under the welfare state. 50 Yet there are a number of cases in the United States, and rnore recently France and Germany, where separation of powers arguments 45 The Founding Fathers of the U.S. Constitution were primarily concerned by the extensive powers exercised by state Assemblies. 46 Federalist Papers Number 51. 47 ibid., Number 47. 48 U.S. Const., Art. II, s. 2. 49 ibid., Art. I, s. 7. 50 See in particular Marshall, Constitutional Theory, ch. V.

285

The Rule of Law and the Separation of Powers

286 6ro

Public Law

have been decisive in litigation. In particular (though not exclusively) they have been used to bolster claims that a citizen's fundamental rights have been violated. This coupling of the two arguments is no accident. For the litigant may be uncertain whether his right will prevail against the substantial public interest which, according to the government, justifies its infringement. Alternatively the contours of the right may be uncertain or controversial. It may be relatively easier to deploy an argument that the decision has been taken unconstitutionally by the executive when it . was only appropriate for the legislature to act. 5 Moreover, under this argument judicial review is less susceptible to the charge that it is fundamentally undemocratic. Indeed, when a court requires the legislature to impose clearer standards on the executive or the police, it would be reinforcing democratic values. Individuals and pressure groups enjoy readier access to legislators than they do to ministers, civil servants or police officers, and so may in theory exercise some influence on the drafting oflegislation. Equally, the courts should be particularly willing to strike down retrospective criminal legislation, insofar as it interferes with the ~eneralliberty of the defendants and also with their rights to a fair trial. 2 In these circumstances, the civil rights argument reinforces the case that the legislature is violating the structural limits on its powers. Courts are, therefore, right to look with particular suspicion on acts of the executive which affect individuals and which are entirely unauthorised by statute. This is exemplified by a number of United States cases, among them the famous Supreme Court decisions in the Steel Seizure and Pentagon Papers cases. In the former the Court held President Truman had violated the separation of powers when he issued an order directing seizure of all steel mills. 53 Put simply there was neither constitutional nor Congressional authority for the order, which at least two members of the majority, Black]. and Douglas]. regarded as an attempt on the part of the President to exercise legislative power. Moreover, it trespassed on the mill owners' constitutional property right guaranteed by the Fifth Amendment. The right at issue in the Pentagon Papers case was the First Amendment freedom of speech and of the press. 54 The federal government had applied for an injunction to restrain publication of classified Defence Department documents in the New York Times and the Washington Post. The most important reason for denial of the remedy was the hostility in the United States to any prior restraint on speech,

51 SeeP. E. Quint, "The Separation of Powers under Nixon: Reflections on Constitutional Liberties and the Rule of Law", [1981] Duke L] 1, and B. Neuborne, ''Judicial Review and Separation of Powers in France and the United States", 57 New York Univ. L.Rev. 363 (1982). The discussion in this section owes much to these articles. 52 See the argument in Liyanage v. The Queen, n. 34 above. 53 Youngstown Sheet and Tube Co. v. Sawyer 343 U.S. 579 (1952). 54 New York Times Co. v. United States 403 U.S. 713 (1971).

The Rule of Law and the Separation of Powers

Separation of Powers and Constitutional Government 6r r

but several members of the Court stressed the absence of any Congressional statute, which authorised the government to take proceedings in these circumstances or which prohibited publication of such documents. In contrast, the English courts have been quite willing to apply the law on breach of confidence at the request of the executive without prior Parliamentar~ enactment, as shown by the Crossman Diaries and Spycatcher cases. 5 In other First Amendment cases the Supreme Court has invalidated legislation conferring wide licensing powers on city officials or police officers. In Lovell v. Griffin it struck down an ordinance giving a City Manager unfettered discretion to ban the public distribution of circulars and pamphlets. 56 Other decisions have applied the same principle in the context of the allocation of permits to use loudspeakers, 57 and to hold public meetings. 58 In all these cases the official was free to formulate for himself the grounds on which he determined the applications. In effect the city ordinance gave him a legislative as well as an executive power. This is clearly undesirable. Without detailed standards to constrain official discretion, it is unlikely that due regard will be paid to the constitutional rights at issue. Moreover, such wide power can be applied selectively, say, to political groups favoured by the official or chief of police, without any opportunity for democratic control. There are similar decisions of the Conseil constitutionnel in France. 59 Under the Constitution of the Fifth Republic, there is a complex division of legislative authority between Parliament and the government. Broadly, the former has power under Article 34 to enact lois over certain specified topics, while under Article 37 other matters "fall within the field of rule-making". In other words, it would ap£ear that government enjoys a wide residuary power to issue reglements. However, the jurisprudence of the Conseil has interpreted these provisions in favour of Parliament, recognising inter alia that it alone may enact rules which affect the general principles of French law (principes generaux du droit), 61 and rules which impose penal sanctions involving a deprivation of personal liberty. 62 In addition, as already mentioned, it has insisted that statutes grant the judiciary the authority to determine issues of personal 55 See A.-G. v. Jonathan Cape Ltd [1976] Q.B. 752, and A.-G. v. Guardian Newspapers (No. 2) Ltd [1990]1 A. C. 109.

303 u.s. 444 (1938). Saia v. New York 334 U.S. 558 (1948). 58 Kunz v. New York 340 U.S. 290 (1951). 59 For general discussion, see Neuborne, "Judicial Review and Separation of Powers in France and the United States", n. 51 above, 394-398. 60 See B. Nicholas, "Loi, Reglement and Judicial Review in the Fifth Republic", [1970] P.L. 251, and Bell, French Constitutional Law, ch. 3. 61 Decision 69-55 ofJune-26, 1969, GD 228. For general principles oflaw, see Bell, French Constitutional Law, 73-75. 62 Decision 73-80 of28 Nov. 1973, Recueil1973, 45 (see B. Nicholas, "Fundamental Rights and Judicial Review in France", [1978] P.L. 82, 92) and Decision 82-143 ofJuly 30, 1982, GD 530. 56

57

287

The Rule of Law and the Separation of Powers

288 612

Public Law

liberty and not delegate such authority without clear standards to the police. 63 The separation of powers principle may also be relevant to constitutional disputes or litigation, even where there -is no immediate or direct impact on the rights and liberties of individuals. This may be illustrated by reference to two different sets of circumstances. The first concerns the organization and powers of broadcasting authorities. In a number of cases the German Constitutional Court has said that it is for the state legislature to lay down the fundamental rules about membership . of the regulatory bodies, its powers, and the standards by which public and private broadcasters are to operate. 64 These matters are too important to be left to administrative regulation or to the discretion of the authority itsel£ This is really a constitutional point: the legislature, and not the executive, nor even an independent agency, should take responsibility for framing these rules which touch on the exercise of fundamental liberties. The second set of circumstances involves political p;1rties and the rights of members of the legislature. Continental European constitutions recognise the distinctive role of political parties and other groups, a matter on which both United Kingdom and United States constitutions are silent. One of the earliest decisions of the Conseil constitutionnel held a rule of the National Assembly unconstitutional with regard to Article 4 of the Constitution. 65 That guarantees the right to form parties, though "[T]hey must respect the principles of national sovereignty and democracy". Article 19-3 of a draft Assembly regulation would have given it the power to decide whether a party respected these principles and hence should be recognised for parliamentary purposes. The Conseil clearly thought it undesirable to leave this sensitive matter to the judgement of the majority party or coalition in the Assembly, in effect to the will of the government. In Germany the Basic Law reserves this assessment for the Constitutional Court. Only it may decide that a political party is unconstitutional because it "seeks to impair or abolish the free democratic basic order or to endanger the existence" of the country. 66 In principle, it must be right to allocate this decision to the judicial branch, since both legislature and executive have an interest in its outcome, which may run counter to democratic values. Somewhat similar issues may be raised in the context of disputes about whether a member should be expelled from the legislature for financial or other misconduct. In the United Kingdom this has been left for the Houses of Parliament themselves to regulate, and it is inconceivable that the ordinary courts would interfere with a decision to expel or suspend a member. In contrast the United States Supreme Court has Decision 77-75 ofJanuary 12, 1977, GD 340. See 12 BVerfGE 205 (1961), 57 BVerfGE 295 (1981), and 73 BVerfGE 118 (1986). Also see E. Barendt, Broadcasting Law (1993), 34-36. 65 Decision 59-2 ofJune 17, 18 and 24, 1959, GD 34. 66 Basic Law, art. 21, II. 63

64

The Rule of Law and the Separation of Powers

Separation of Powers and Constitutional Government 6I3

held that it may review the exclusion of an elected member by the House of Representatives; his claim was not barred by the "political questions" doctrine and was justiciable. 67 Article 1, section 5 of the Constitution provides that a member may be expelled for disorderly behaviour, but that provision was not invoked by the House. Warren C.J. rightly concluded that for the House to claim further powers to exclude an elected member would be dangerous to a representative democracy. The Court has also upheld an appeal against conviction for contempt of Congress, imposing significant limits on its power to conduct investigations. These procedures carried the rjsk that Congress would cease to act as a legislature and would assume an executive or judicial role. 68 Under the Basic Law the Bundestag is free to determine election disputes, but is subject in this respect to the control of the Constitutional Court. 69 The separation of powers principle is therefore not simply a formal guide to the organization of state power. It can be given teeth by constitutional courts to reinforce the protection conferred by the constitution on individual rights, and to prevent one branch of government from accumulating excessive powers. Whatever its theoretical defects may be, case law in other systems shows that it is not as vacuous as its English critics have alleged.

The separation of powers in the United Kingdom One of Walter Bagehot's best known remarks is that "(t)he efficient secret of the English [sic] Constitution may be described as the close union, the nearly complete fusion, of the executive and legislative powers. " 70 He contrasted this with the Presidential system in the United States, where the President and Congress are totally separate from each other. The former is not chosen by or responsible to the latter. In the United Kingdom, on the other hand, the executive is formed from the majority party in the legislature. Ministers sit in Parliament, so there is no separation of membership of the legislative and executive organs as there is in the United States and in France. 71 The Prime Minister is able to control the House of Commons through his power to recommend to the Crown a dissolution of Parliament. For Bagehot this was entirely satisfactory, though it is hard to see why he reached this conclusion when he admitted that parliamentary government is in essence "sectarian government" 72 • When Bagehot wrote, there was perhaps still a system of" checks and Powell v. McCormack 395 U.S. 486. Watkins v. United States 354 U.S. 1978 (1957). 69 Basic Law, Art. 41. See Currie, The Constitution of the Federal Republic of Germany, 111-112. 70 The English Constitution, introd. by R. H. S. Crossman (1963), 65. 71 French Constitution, art. 23 provides ministers may not sit in Parliament, though art. 31 gives them access to be heard. 72 The English Constitution, 222. 67

68

289

The Rule of Law and the Separation of Powers

290

614

Public Law

balances" in the U.K. constitution. Admittedly it was a shadow of the eighteenth-century balance of powers, when the Commons, Lords, and the King (with his Court) still constituted distinct social interests, participating almost equally in the formation of legislative policy. But the House of Lords balanced the Commons to some extent until its legislative powers were reduced by the Parliament Acts 1911-49. During the nineteenth-century Members of Parliament were conscious of their independent representative role as advocated by Burke, and were more willing to act independently of political party loyalties and constraints than they are now. Bagehot argued that excesses of sectarian zeal could be checked by the Prime Minister's power to dissolve Parliament. He seems to have ignored the possibility that the government and Parliament might be party to the same excesses. Overall his attempts to find the constitution's "supposed checks and balances" (to use his term) were very unconvmcmg. The truth is that there is no effective separation of powers between legislature and executive in the United Kingdom in the sense of a system of "checks and balances". The advent of mass political parties has destroyed the semblance of such a system which existed a century ago. Except on the rare occasions when there is a significant party split, 73 the government effectively controls the legislature. For Members of Parliament are, it can be said, selected by party members or activists, and are certainly in practice answerable to them for their conduct and voting record. The chief concern of party loyalists is to support the government, or to bring it down when it is controlled by the other party. Hence the executive is able to control members of the legislature through the device of party. This phenomenon is of course not confined to the United Kingdom. Even in the United States, Jackson]. observed that the party system had significantly increased the effective power of the President by enabling him to apply pressure on the legislators who are supposed to check him. 74 In fact the constitutional settlement achieved by the Glorious Revolution nearly introduced an institutional separation of legislative and executive powers. During the 1690s there was considerable apprehension that a Hanoverian King would attempt to control Parliament. The Act of Settlement 1701, therefore, provided that, after the date of his succession, no minister or Crown appointee could sit in the Commons. But the provision was never brought into operation; instead, it was replaced by more moderate rules against placemen which precluded civil servants, and army and navy officers, from membership of the House. 75 At the beginning of the nineteenth century William Cobbett thought it 73 74

75

As has now happened in the Conservative Party over the European Union. Youngstown Sheet and Tube Co. v. Sawyer343 U.S. 579,651 (1952). See Holmes, The Making ofa Great Power,223.

The Rule of Law and the Separation of Powers

Separation ofPowers and Constitutional Government 615

desirable to re-introduce the stronger prohibition. 76 Even now some deference to institutional separation is shown by the limit on the number of ministers who may sit in the Commons. 77 There is, however, an effective separation of the judicial power from the other branches. Judges may not sit in the House of Commons and they are protected from summary removal under the Act of Settlement 1701. By convention lay peers may not take part in the deliberations of the House of Lords as an appeal court. On the other hand, ,the position of the Lord Chancellor and the freedom, increasingly exploited, of the Law Lords to participate in the legislative debates of the Upper House, contravene the principle, albeit moderately and perhaps acceptably. While there is in practice a fusion of legislative and executive powers, there is in principle a distinction between the two functions. 78 In other words, government may control the legislature (and certainly there is overlapping membership), but it must still legislate through Acts of Parliament. Since the Case of Proclamations the Crown has had no power to change the law without resort to parliamentary enactment. 79 Nor can it levy taxation, suspend a statute or dispense anyone from complying with its provisions. All three of these prerogative legislative powers were swept away by the Bill of Rights 1689. Recently the House of Lords had a marvellous opportunity to restate these principles in the Criminal Injuries Compensation Board case. 8 Clearly the Home Secretary was trying to legislate, when he used his prerogative power to introduce the tariff compensation scheme. The House of Lords should have simply held it was unconstitutional for the executive to legislate. The point is important. Individuals and pressure groups almost certainly find it easier to persuade M.P.s to move amendments to a Bill than to induce the government to change course. Judicial intervention on this straightforward constitutional principle would have struck a blow for the values of democracy. The Home Secretary's powers to determine the effective length of discretionary life sentences of imprisonment also raise delicate separation of functions questions. The European Court ofHuman Rights has twice ruled that a court must decide whether a prisoner may be detained again, subsequent to his release following expiry of the minimum tariff period of his sentence. 81 These rulings were determined by the text of

°

76 Cobbett and other contributors to the Political Register, Tom Paine and others were highly critical of the fusion of Cabinet and Parliament: see Vile, Constitutionalism and the Separation of Powers, 109-112. ' 77 The present limit is 95: House of Commons Disqualification Act 1975. 78 See F. S. Burin, "The Theory of the Rule of Law and the Structure of the Constitutional State", 15 Am. Univ. L.Rev. 313 (1966), for the argument that this separation is integral to the rule oflaw. 79 (1611) 12 Co. Rep. 74. 80 R. v. Secretary of State for the Home Department, ex parte Fire Brigades Union [1995] 2 W.L.R. 464, discussed by E. Barendt, [1995] P.L. 357. 81 Weeks v. United Kingdom (1988) 10 E.H.R.R. 293; Thynne, Wilson and Gunnell v. U.K. (1991) 13 E.H.R.R. ~66.

291

The Rule of Law and the Separation of Powers

292 616

Public Law

Article 5(4) of the Human Rights Convention, which provides that anyone suffering deprivation of liberty has a right to have its legality determined speedily by a court. The separation of powers principle is also at issue here. The executive is ill-fitted to decide issues of personal liberty. It is much more susceptible than a court to media and other pressure which is likely to give individual liberty less weight than public order and safety considerations. 82 The courts are generally anxious to safeguard the judicial function against attempts by the executive to take it over. This is clearly evidenced by the number of administrative law cases, where they have interpreted statutes not to oust their jurisdiction to determine questions oflegal entitlement or liability. In effect they disregard provisions whi~h appear on their face to preclude judicial review of administrative decisions. 83 One common justification for these rulings is that there is a presumption Parliament does not intend the courts' jurisdiction to be excluded, so the judges are giving effect to the will of the legislature. But this is very artificial. The separation of powers principle provides a much more cogent explanation, and in effect justifies what is in practice a restraint on the principle of parliamentary legal supremacy. It remains, of course, unclear whether United Kingdom courts would be prepared to follow the Privy Council rulings which have. invalidated retrospective penal legislations as contrary to the separation principle. 84 Adherence to the familiar doctrine of parliamentary sovereignty might lead to the application of even the most monstrous retrospective legislation. It can safely be assumed that the courts woUld strive as hard as they could to avoid that course, and it would be relevant in this context that the European Human Rights Convention precludes such legislation in most circumstances. 85 Lords Diplock and Scarman have observed that the constitution is based on the separation of powers. 86 But they made this point to reinforce the argument for judicial restraint in interpreting statutes; the role of the courts is not to make law, but to interpret the words used by the legislature. Interestingly, the only Law Lord to make explicit reference to the separation of powers in the Criminal Injuries Compensation Board case was Lord Mustill, who dissented from the majority conclusion that the Home Secretary had abused his powers. 87 For him the principle indicated that the courts should be hesitant to expand their rapidly developing power of judicial review. In a sense all these statements support the For general discussion, see A. Ashworth, Sentencing and Criminal justice (1992), ch. 2. e.g., Anisminic v. Foreign Compensation Commission [1969] 2 A.C. 147; Commissioner for Customs and Excise v. Cure and Deeley Ltd [1962]1 Q.B. 340. 84 See the Liyanage case and other decisions referred to in n. 28 above. 85 Art. 7. But the restriction does not preclude the trial for any act which at the time of its commission was criminal "according to the general principles oflaw recognised by civilised nations": Art. 7(2). For this reason the controversial War Crimes legislation would not be suspect. 86 Duport Steels v. Sirs [1980]1 W.L.R. 142, 157 and 169 respectively. 87 [1995]2 W.L.R. 464, 487-488. 82

83

The Rule of Law and the Separation of Powers

Separation of Powers and Constitutional Government 617

traditional French or pure theory of the separation of powers, which has emphasised the discrete and exclusive functions of the three branches rather than the "checks and balances" approach. The same attitude is surely implicit in those cases where the courts refuse to intervene because they fear that, if they did so, they would be trespassing on the province of the House of Commons. For example, Lord Browne-Wilkinson has recently said that Article 9 of the Bill of Rights, providing that freedom of speech in Parliament may not be questioned, manifests a wider constitutional principle that couhs and Parliament recognise their discrete roles. Judges should not allow any challenge to what is said or done within Parliament in pursuance of its legislative functions. 88 The effect of this pronouncement was that the defendant could not rely on speeches made by the plaintiff in the New Zealand House of Representatives to substantiate a defence of justification to the latter's libel action. That might lead to an injustice to the defendant (though the Privy Council conceded that in extreme cases proceedings should be stayed if in the absence of this evidence the issue could not be tried fairly). On the alternative "checks and balances" understanding of the separation principle, a court should have no hesitation in interfering here. Judicial control of parliamentary privilege is vital to prevent the legislature, or one branch of it, abusing its powers. The separation of powers and constitutional reform Separation of powers issues are as much ignored in discussion of constitutional reform ~s they have generally been in constitutional analysis. That is not surprising. There is no obvious constituency to whom they appeal. In comparison, there is national pressure in Scotland and Wales for devolution, while civil liberties groups campaign for incorporation of the European Convention on Human Rights. Moreover, the legal relationship of the legislature and the executive, much the most important and complex aspect of the principle, was for the most part settled three hundred years ago in the wake of the Glorious Revolution. Finally, it should not be forgotten that the agenda for constitutional reform is mostly drawn up by the two major political parties. The only reason why it is conceivable that the United Kingdom may soon experi-' ence devolution and incorporation of the Convention is that the Labour Party is now sympathetic to both causes. But it is most unlikely that the Labour party, or any other, would favour constitutional reform which would impose more effective checks and balances on the executive. After all it was the rise of the modern mass political party which has largely led to the erosion of that balance of power which characterised the constitution in the eighteenth, and to some extent even in the nineteenth, centuries. 88 Prebble v. Television New Zealand Ltd [1995]1 A. C. 321, 332; see G. Marshall [1994] P.L. 509. In July 1995 the High Court stayed two separate libel actions by M.P.s, on the principle the defendant newspapers could not establish their defence.

293

The Rule of Law and the Separation of Powers

294 618

Public Law

Nevertheless, there is one serious reform proposal, which is arguably inspired by the separation of powers principle: the irlstitution of fixedterm Parliaments, with consequent abolition of the Crown's prerogative to dissolve the House of Commons and call a general election. As Bagehot pointed out, the power in effect enables the executive, and in particular the Prime Minister, to discipline a legislature which might otherwise be too independent or unruly. During the nineteenth century it may have been desirable for the government to be able to check Parliament in this way. Bu_t it is now crucial to strengthen the latter. Removal of the dissolution power would make it easier for the House of Commons to resist what it considered to be unwarranted government measures, albeit by strengthening the position of dissident members of the majority party who would be immune from the threat of an immediate dissolution and possible loss of their seats. The legislature in short would be in a position to check the executive more effectively. A comparative point is that in Germany the Federal President has very limited discretion to dissolve the Bundestag. Under Article 68 of the Basic Law he may only do this if the Chancellor's-confldence motion is not passed by a majority of its members; further, the Constitutional Court may review the exercise of both the Chancellor's and the President's discretion in this context. 89 Another more controversial proposal would be to reduce the number of ministers entitled to sit in the House of Commons, At the moment the limit is 95, about 15 per cent of the total number of M.P.s. Questions should be asked about whether it is really necessary to have so many ministers in the Commons, or whether it would matter much if more were drawn from the Lords or if some, at junior level, were not even members of either House. It might be worth paying the price of a loss of ministerial accountability to Parliament to reduce the size of the pay-roll vote. More balance of a different kind would be achieved by increasing the powers of the Second Chamber. This desirable step could not, of course, be taken unless its composition were first changed, and its members elected on an acceptable basis, perhaps as in a federal system to represent constituent national or regional communities. What would clearly be unacceptable on separation of powers grounds is for all members of the Upper House to be appointed, even temporarily, by the government, once hereditary peers lose their voting rights. This would be a monstrous expedient, for the head of the executive would be claiming a right to choose one branch of the legislature. The courts could also take the separation of powers more seriously, though in the absence of a codified constitution it is hard for them to articulate or develop appropriate principles. In particular, they should 89 62 BVerfGE 1 (1983): see Currie, The Constitution of the Federal Republic of Germany, 113-6, and M. Louk, "The Constitutional Court Reviews the Early Dissolution of the West German Parliament", 7 Hastings Int. and Comp. L.Rev. 79 (1983). The President must dissolve the Bundestag if it fails to elect a Chancellor: see Art. 63. Otherwise there is in effect a fixed term Parliament.

The Rule of Law and the Separation of Powers

Separation of Powers and Constitutional Government 619

scrutinize even more carefully administrative decisions which impinge on individual rights without clear legislative authority. The applications for injunctions in the Crossman Diaries and Spycatcher cases should have been rejected on separation of powers grounds alone--as they would have been in the United States. The courts should also reconsider their pusillanimous attitude to parliamentary privilege and contempt cases, where abstention from review may lead to real injustice to individuals. Whatever the shape of future constitutional reform, M.P. s should be careful to ensure that legislation is precise and does not delegate power to the executive or to administrative agencies without any standards to guide its exercise. Such vigilance strengthens the democratic process, and further assists the courts to check that power is not abused. Firially, there should be a comprehensive review of the legal arrangements governing the Boundary Commissions, the National Audit Office, the Audit Commission and all other bodies which are concerned with acc.ountability. Their independence must be guaranteed against undue interference from either the legislature or executive. One means of achieving this end would be to provide that Parliament and government check each other, both at the stage of appointing members and at the time for implementation of their proposals. All these developments and arrangements are implicit in the separation of powers, a principle too long ignored in United Kingdom constitutional law.

295

[13] ON SPEAKING SOFTLY AND CARRYING BIG STICKS: NEGLECTED DIMENSIONS OF A REPUBLICATION SEPARATION OF POWERSt John Braithwaite* I

The Concentration of Private Power

A. TOWARDS THE 'PURE DOCTRINE' OF THE SEPARATION OF POWERS

The separation of powers may be the most central idea in the theory of institutional design. Yet this has only been true of thinking about public institutions. This paper extends the relevance of the doctrine into thinking about checking the power of private institutions. The practice of separating powers dates at least from the Code of Hammurabi 1 when laws were carved in literal stone that would constrain the actions not only of subjects but also of the king. There follows a more or less cumulative history of separations of powers that we see sedimented in the institutions of contemporary Western democracies. Among the important moments in this history were the mixed Spartan Constitution, 2 the Roman Senate and Justinian's Code, 3 Magna Carta, 4 the jury, the growth of universities

*

t

1 2 3 4

Head, Law Program, Research School of Social Sciences, Australian National University. I owe a debt to four different intellectual communities within the Australian National University that have shaped my thinking in this paper. Parts of the paper have been presented to seminars of the Economics Program and the Social Political Theory Group of the Research School of Social Sciences. From the former group, I am particularly grateful for Bruce Chapman's criticisms on the economic analysis of law. From the latter, I am particularly indebted to Philip Pettit's ideas on republican institutional design. Third, I have been stimulated by the group of scholars within the Law Program in RSSS interested in the separation of powers - Sir Anthony Mason, Christine Parker, Fiona Wheeler, John Williams, & Leslie Zines. Fourth, I thankfully acknowledge the large group of scholars of regulatory institutions from across ANU whose work is cited in the paper, particularly Stephen Bottomley. Finally, my thanks to Tonia Vincent for her relentless research assistance. Part VII of the paper relies heavily on things I learned with and from Brent Fisse during our fieldwork together. Thanks also to Colin Scott & Peter Grabosky for helpful comments. W.F. Leemans, Legal and Administrative Document5 of the Time a{ Hammurabi and Samsuiluna (Mainly from Lagaba) (Leiden: J.E. Brill, 1960). W.B. Gwyn, The Meaning of the Separation of Powers (The Hague: Martin us Nijhof, 1965) at 88 [hereinafter Meaning]. Justinian, The Digest o{ &man Law: Theft, Raj1ine, Damage and Insult, trans. and ed. C.F. Kolbert (Harmondsworth: Penguin Books, 1983). A. Pallister, Mab'7ta Carta: The Heritage o{Liberty (Oxford: Clarendon Press, 1971); W.I.Jennings, Magna Carta and Its Influence in the World Todr)y (London: HMSO, 1965).

The Rule of Law and the Separation of Powers

298

306

UNIVERSITY OF TORONTO LAW JOURNAL

as accumulators and communicators of knowledge that become progressively more independent of church and state, the rise of judicial independence, and bicameral parliaments. Notwithstanding the important contributions of Locke 5 and other enlightenment scholars, 6 the practice of separation of powers was ahead of the theory until Montesquieu published The Spirit of the Laws. 7 The richest development of these ideas flowed from the debates between the federalists 8 and anti-federalists 9 in the drafting of the us Constitution. The political philosophy of both the federalists and anti-federalists was republican. Philip Pettit has been the primary inspiration in a program of work at the Australian National University to excavate the foundations of the republican approach to the checking of power as a commitment to freedom as non-domination. 10 In that work, checking power under a rule of law designed to minimize the capacity of others to exercise arbitrary power over us is seen as the keystone of the freedom republicans cherish. While the republican theorizing and Constitution-writing of the late eighteenth century clarified thinking about the separation of powers, its legacy was also to narrow vigilance to the checking of state power. What Vile characterises as the 'pure doctrine' of the separation of powers illustrates It is essential for the establishment and maintenance of political liberty that the government be divided into three branches or departments, the legislature, the executive and the judiciary. To each of these three branches there is a corresponding identifiable function of government, legislative, executive or judicial. Each branch of government must be confined to the exercise of its own function and not allowed to encroach upon the functions of other branches. Furthermore, the persons who compose these three agencies of government must be kept separate and distinct , no individual being allowed to be at the same time a member of more than one branch. In this way each of the branches will be a 5 J. Locke, Two Treatises o_fGovernment (Cambridge: Cambridge University Press, 1960). 6 For example, James Harrington &John Milton on mixed government. See MJ.C. Vile, Constitutionalism and SejJaration of Powers (Oxford: Clarendon Press, 1963) at 29-30. 7 C. de Secondat Montesquieu, The SjJirit of the Laws, trans, and ed. A.M. Cohler, B.C. Miller & H.S. (Cambridge: Cambridge University Press, 1989). 8 A. Hamilton, J. Madison & J. Jay, The Federalist Papers (New York: Mentor Books, 1961). 9 R. Ketcham, ed., The Anti-Federalist PajJers and the Constitutional Convention Debates (New York: Mentor Books, 1986). 10 P. Pettit, RejJUblicanism: A Theory of Freedom and Government (Oxford: Oxford University Press, 1997); J. Braithwaite & P. Pettit, Not.fust Desserts: A Republican Theory of Criminal Justice (Oxford: Oxford University Press, 1990); S. Bronitt & G. Williams, 'Political Freedom as an Outlaw: Republican Theory and Political Protest' (1996) 18 A.L.R. 289.

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

299

307

check to the others and no single group of people will be able to control the machinery of the State. 11 B. BEYOND THE PURE DOCTRINE

The autocrat of the state was seen as the threat to our freedom. As a result, when we think of the separation of powers today, we think of separations among these branches of the state - the legislature, the executive, and the judiciary. Yet equally important in the history of the separation of powers has been the separation of church and state. More important in terms of contemporary effects is the separation of business and the state. Today, the largest fifty transnational corporations all have greater resources, stronger political connections in the world system, more practical coercive capabilities, and more sophisticated private policing technology than most of the world's statesY The technology that monitors all our financial transactions, the trace of our movements about the city as we make them, the actual things we do in the most private spaces of the metropolis are captured on video not by the state tyrant that George Orwell feared/ 3 but by private repositories of power. The legacy of the republican tradition is obsession with the powers of state police in societies with twice as many private as public police. 14 In nations like Australia, a Rupert Murdoch has more influence over the Prime Minister and Cabinet than any member of the judiciary. Moreover, that is only a tiny part of Rupert Murdoch's power compared to the influence he has in the United States, China, and beyond. Through influence in a number of states, such private actors sometimes shape global regulatory regimes in ways that make the citizens of all states subservient to them. 15 11 Vile, supra note 6. Distinguishing these three branches is no simple matter, as Sir Anthony Mason explains '[t]he lesson of history is that the separation of doctrine serves a valuable purpose in providing safeguards against the emergence of arbitrary or totalitarian power. The lesson of experience is that the division of powers is artificial and confusing because the three powers of government do not lend themselves to definition in a way that leads readily to a classification of functions (Anthony Mason, 'A New Perspective on the Separation of Powers,' Reshaping Australian Institutions: Australian National University Public Lecture, Lecture 1, 25 July 1996 at 5). 12 RJ. Barnet & R.E. Muller, Global Reach: The Power of the Multinational CorjJOrations (New York: Simon and Schuster, 1974); RJ. Barnet &J. Cavanagh, Global Dreams: Imjmial CorjJOrations and the New World Order (New York: Simon and Schuster, 1994). 13 G. Orwell, 1984 (New York: Harcourt, Brace, Jovanovich, 1949). 14 For the seminal collection on the privatization of policing, see C.D. Shearing & P.C. Stenning, eds., Private Policing. Sage Criminalfustice System Annualf, vol. 23 (Beverly Hills: Sage, 1987). Also see C.D. Shearing & P.C. Stenning, 'Modern Private Security: Its Growth and Implications,' Crime and Justice: An Annual Review of' Research, vol. 3, eds., M. Tonry & N. Morris (Chicago: University of Chicago Press, 1981). 15 For example, Peter Drahos and I are documenting the influence of 16 chief executives of American companies in reshaping the world intellectual property order

300

The Rule of Law and the Separation of Powers

308

UNIVERSITY OF TORONTO LAW JOURNAL

Todroru:h (London: Routledge & Kegan Paul, 1978). 33 ]. Madison, A. Hamilton &].Jay, supra note 9 at 321. 34 This is a classic Dahlian conception of power. See R.A. Dahl, 'The Concept of Power' (1957) 2 Behavioral Science. 201. 35 On this distinction, see S. Clegg, Power, Rule and Domination (London: Routledge & Kegan Paul, 1975) at 67-75. 36 S. Lukes, Power: A Radiml View (London: Macmillan, 1974).

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

305

313

will on the object of control, 37 some will be a Gramscian hegemony which constitutes individuals who cannot recognize that they are being dominated/8 some will be Foucauldian disciplinary networks partially advanced through the practices of agents rather than intended or willed as acts of power, 39 some will be Foucauldian shepherds governing and caring for their flock. 40 Normatively, what republicans of my stripe want preserved is freedom as non-domination; in different contexts different types of power exercised by plural agents of power will do that job best. So one wants such plural separations of disparate modalities of private and public power as will maximize freedom as non-domination. None of this is particularly novel; it is simply a somewhat radicalising extension of tendencies that can be found in the writings ofJames Madison. There is, however, novelty in the analysis of speaking softly and carrying sticks that is my reason for pushing far with this pluralising of Madison. There are a number of reasons why Madison and other advocates of the separation of powers found the doctrine attractive. There is the desire to limit the damage that one all-powerful bad ruler can do, to expand the diversity of perspectives that have influence in politics, to foster deliberative democracy by requiring one branch of governance to persuade another that it has exercised its power wisely, to constrain the rule of men by the rule of law, to empower those who might otherwise be powerless. This essay will not systematically evaluate the desirability or feasibility of these rationales for the separation of powers. The analysis will be limited to a fresh perspective on just one rationale for the separation of powers, albeit what republicans should regard as the most fundamental one 41 -checking abuse of power. Before we can reach the point of understanding why pluralities of checks is the reconceptualization of the separation of powers we need for the contemporary world, we must begin with an understanding of the empirical literature on why efforts to check abuse of power so often backfire.

37 B. Latour, 'The Powers of Association,' ed. J. Law, Power, Action and Belief.' A New Sociology of Knowledge? Sociological Review Monogmj1h 32 (London: Routledge & Kegan Paul, 1986). 38 A. Gramsci, Selections from the Prison Notebooks of Antonio Gmmsci, trans. and ed. Q. Hoare & G.N. Smith (London, Lawrence & Wishart, 1971). 39 M. Foucault, Di.w:ij1line and Punish: The Birth of the Prison (Pantheon: New York, 1977). 40 B. Hindess, Discourses ofPower: From Hobbes to Fouwult (Oxford: Blackwell, 1996). 41 It is the most important one for those who, like me, share Pettit's view that the most fundamental value in the design of political institutions is the assurance of freedom as non-domination (P. Pettit, supra note 10).

The Rule of Law and the Separation of Powers

306

314

UNIVERSITY OF TORONTO LAW JOURNAL

III

Why Big Sticks Rebound

A.DETERRENCEFAILURE

The starting point for reaching the conclusions promised in the Introduction is to abstract from what we have learned empirically about the way the regulation of private power works, or rather why it so regularly fails to work. The republican idea of checking power with countervailing power is often read as a deterrence model for controlling abuse of power. Indeed, deterrence will have an important place in the conclusion ultimately reached in the present analysis. 42 This section will explain first why deterrence often does not work well, drawing on research about emotion and rationality. Then we will see that deterrence often backfires and explain why deterrence has this capacity to defeat its own objectives. In subsequent sections, the argument will be that ifwe understand these problems properly, the separation of powers will prove relevant to their amelioration. The discipline that has grappled most systematically with why deterrence does not work well is criminology. People almost universally value their lives. So it is surprising that introducing capital punishment is not shown to significantly reduce the crime rate, nor does abolishing the death penalty increase it. 43 It is surprising that building more prisons and locking up more people in them for longer periods, tougher sentencing, does not have predictable effects in reducing the crime rate. 44 Even the 42 Abuse of power might of course be controlled by rewards rather than sanctions. Rewards are not as seriously considered as deterrence in this role because giving even more resources to the rich and powerful for doing what it should be their citizenship obligation to do is unappealing to republicans. While rewards should be taken more seriously than they are in the theory and practice of checking power, some of the empirical limitations of control by deterrence will also be found to apply to control by reward. 43 W.C Bailey & R.D. Peterson, 'Murder, Capital Punishment and Deterrence- A Review of the Evidence and an Examination of Police Killings' (1994) 50 Journal of Social Issues. 53; J.K. Cochrane, M.B. Chamlin & M. Seth, 'Deterrence Brutalization -An Impact Assessment of Oklahoma Return to Capital-Punishment' (1994) 32 Criminology. 107; B. Forst, 'Capital Punishment and Deterrence: Conflicting Evidence?' (1983) 74 Journal of Criminal Law and Criminology. 927; T. Sellin, The Penalty oj' Death (Beverly Hills: Sage, 1980); Schuessler, 'The Deterrent Influence of the Death Penalty' (1952) 284 Annals of the Academy of Political and Social Sciences. 54; R. Hood, The Death Penalty: A World Wide Suroey (Oxford: Oxford University Press, 1989). 44 A. Blumstein, J. Cohen & D. Nagin, eds. Deterrence and Incajmcitation: Estimating the ~jj'ects of Criminal Sanctions on Crime R11tes (Washington, DC: National Academy of Sciences, 1978). J.P. Gibbs, Crime, Punishment and Deterrence (New York: Elsevier, 1975); E.A. Fattah, 'Deterrence: A Review of the Literature,' (1977) 19 Canadian Journal of Criminology. 1; P J. Cook, 'Research in Criminal Deterrence: Laying the Groundwork for the Second Decade,' Crime and .Justice: An Annual Review of Research,

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

307

315

optimists about imprisonment do not conclude it has large effects. 45 It is surprising that we can experimentally double or halve the number of police patrol cars in neighbourhoods, without the experimental areas experiencing changes in crime rates in comparison to control neighbourhoods.4G It is surprising that people who perceive the expected severity of punishment for committing a crime to be high are not more likely to refrain from crime than people who expect the severity of punishment from committing a crime to be low. 47 The us, with a death penalty that vol. 2, eds. N. Morris & M. Tonry (1980) at 211. The latter review and an update soon to be published in the same series by Daniel Nagin ('Criminal Deterrence Research: A Review of the Evidence and a Research Agenda for the Outset of the 21st Century') are more optimistic about the existence of a deterrent effect. Nagin concludes that 'evidence for a substantial deterrent is much firmer than it was fifteen years ago.' With research on the deterrent effect of imprisonment Nagin concludes that the most recent study (S.D. Levitt, 'The Effect of Prison Population Size on Crime Rates: Evidence from Prison Overcrowding Litigation' (1996) 111 Quarterly Journal of Economics. 319) is both the most methodologically advanced and shows the biggest crime-reduction effect of any of these studies. However, there may not be a significant deterrent effect even in Levitt's study because Nagin reports that the estimated reduction in crime from imprisonment is 'not much larger than rates of offending of incarcerated populations reported in various studies of the incapacitation impacts' (thereby leaving little residual deterrent impact after the effect of the offenders' being off the street is taken into account). 45 James Q. Wilson, 'Crime and Public Policy,' Crirne, eds. J.Q. Wilson &J. Petersilia (San Francisco: ICS Press, 1995) at 489. 46 D.H. Bayley, Polir:efor the Future (New York: Oxford University Press, 1994). The classic study here, the Kansas City Preventive Patrol Experiment is not without its methodological problems as Lawrence Sherman & David Weisburd have shown. Sherman & Weisburd show, moreover, that increased police activity targeted on known hot spots, as opposed to random patrol as in the Kansas City experiment, can have a significant effect on crime. L. Sherman & D. Weisburd, 'General Deterrent Effects of Police Patrol in Crime "Hot Spots": A Randomized, Controlled Trial' (1995) 12 Justice Quarterly. 625-49. For the original study see G. Kelling et al., The Kansas City Preventive PatrolExjJenrnent: A Surnrnary Report (Washington, DC: Police Foundation, 1974). 47 While this is mostly the result in the following studies, the expected certainty of punishment is also in the majority found to have an effect, at least in some contexts. G.P. Waldo & T.G. Chiricos, 'Perceived Penal Sanction and Self-reported Criminality: A Neglected Approach to Deterrence Research' (1972) 19 Social Problems. 522; W.C. Bailey & R.P. Lot, 'Crime, Punishment and Personality: An Analysis of the Deterrence Question' (1976) 67 Journal of Criminal Law and Criminology. 99; R.E. Kraut, 'Deterrent and Definitional Influences on Shoplifting' (1976) 23 Social Problems. 358; M. Silberman, 'Towards a Theory of Criminal Deterrence' (1976) 41 American Sociological Review. 442; M.W. Spicer & S.B. Lundstedt, 'Understanding Tax Evasion' (1976) 31 Public Finance. 295; J. Teevan, Jr., 'Deterrent Effects of Punishment: Subject Measures Continued' (1976) 18 Cdn.Jo. Corr. 152;]. Teevan,Jr., 'Subjective Perceptions of Deterrence (Continued)' ( 1976) 13 Journal of Research in Crime and Delinquency. 155; J. Teevan, Jr., 'Deterrent Effects of Punishment for Breaking and Entering and Theft,' Law Rej(mn Cornrnission a{ Canada, Fear of'Punishrnent

The Rule of Law and the Separation of Powers

308

316

UNIVERSITY OF TORONTO LAW JOURNAL

other dev~loped nations do not have, with more private and public police than they, with imprisonment rates several times higher than the OECD average, has not the lowest crime rates, but the highest of any of the wealthy nations. How can this be so? B. EMOTION AND DEFIANCE

One reason is that the protection we get from many of the worst crimes is not bound up with calculative deliberation. For most of the people who (Ottawa: Law Reform Commission, 1976); L.S. Anderson, T.G. Chiricos & G.P. Waldo, 'Formal and Informal Sanctions: A Comparison of Deterrent Effects' (1977) 25 Social Problems. 103; R.F. Meier & Weldon T. Johnson, 'Deterrence as Social Control: The Legal and Extra-legal Production of Conformity' (1977) 42 American Sociological Review. 292; W. Minor, 'A Deterrence Control Theory of Crime,' Theory in Criminology: Contemj)()mry View.1; ed. R. Meier (Beverly Hills: Sage, 1977); L. Cohen, 'Sanction Threats and Violation Behavior: An Inquiry into Perceptual Variation,' Quantitative Studies in Criminology, ed. C.F. Welford (Beverly Hills: Sage, 1978); G.F. Jensen & M. Erickson, 'The Social Meaning of Sanctions,' Crime, Law and Sanction: Theoretical PersjJectives, eds. M. Krohn & R. Akers (Beverly Hills: Sage, 1978); R. Mason & L. D. Calvin, 'A Study of Admitted Income Tax Evasion' (1978) 13 Law and Soc'y Rev. 73; K.E. Waerneryd & B. Walerud, 'Taxes and Economic Behavior Some Interview Data an Tax Evasion in Sweden' (1982) 2 Journal of Economic Psychology. 187; R.L. Akers et al., 'Social Learning and Deviant Behavior: A Specific Test of a General Theory' (1979) 44 American Sociological Review. 635; H.G. Grasmick & GJ. Bryjak, 'The Deterrent Effect of Perceived Severity of Punishment' (1980) 59 Social Forces. 471; R.F. Meier, jurisdictional Differences in Deterring Marijuana Use' (1982) 12 Journal of Drug Issues. 51; R. Paternoster et al., 'Estimating Perceptual Stability and Deterrent Effects: The Role of Perceived Legal Punishment in the Inhibition of Criminal Involvement' ( 1983) 7 4 Journal of Criminal Law and Criminology. 270; R. Paternoster et al., 'Perceived Risk and Social Control: Do Sanctions Really Deter?' (1983) 17 Law and Soc'y Rev. 457; D.M. Bishop, 'Legal and Extralegal Barriers to Delinquency: A Panel Analysis' (1984) 22 Criminology. 403; F.P. Williams III, 'Deterrence and Social Control: Rethinking the Relationship' (1985) 13 Journal of Crim. Just. 141; K.A. Kinsey, 'Theories and Models of Tax Cheating' (1986) 18 CriminalJustice Abstracts. 403; R. Paternoster & L. Iovanni, 'The Deterrent Effect of Perceived Severity: A Re-examination' (1986) 64 Social Forces. 751; I. Piliavin et al., 'Crime, Deterrence and Rational Choice' (1986) 51 American Sociological Review. 101; S. Klepper & D. Nagin, 'The Deterrent Effect of Perceived Certainty and Severity of Punishment Revisited' ( 1989) 27 Criminology. 721; S. Klepper & D. Nagin, 'Tax Compliance and Perceptions of the Risk of Detection and Criminal Prosecution' (1989) 23 Law and Soc'y Rev. 209; H.G. Grasmick & R. Bursik, 'Conscience, Significant Others and Rational Choice: Extending the Deterrence Model' (1990) 24 Law and Society Review. 837; S.R. Burkett & E.L. Jensen, 'Conventional Ties, Peer Influence and the Fear of Apprehension: A Study of Adolescent Marijuana Use' (1975) 16 Sociological Quarterly. 522; C.R. Tittle, Sanctions and Social Deviance (New York: Preager, 1980); K. Williams & R. Hawkins, 'The Meaning of Arrest for Wife Assault' (1989) 27 Criminology. 163; S. Simpson, 'Corporate Crime Deterrence and Corporate Control Policies: Views from the Inside,' White-Collar Crime Reconsidered, eds. K. Schlegel & D. Weisburd (Boston: Northeastern University Press, 1992).

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

309

317

caused us problems last week, we did not deal with those problems by killing them. Our refraining from murder was not because we weighed up the benefits against the probability of detection and likely punishment; it was because murder was right off our deliberative agenda; murder was simply unthinkable to us as a way of solving our problems. It is understanding what constitutes that unthinkableness that is the key to crime prevention. Whether the penalty for murder is death or something else is mostly quite unimportant to that understanding. Moreover, when murder does become thinkable, it often does so in a way that is not rationally deliberative in a way the deterrence model assumes. It sometimes becomes thinkable in the context of emotions temporarily hijacking those more calculative processes in the brain on which deterrence depends. 48 This emotional short-circuiting of rational calculation in our brain had survival value in the history of our species. The emotion of anger sends blood rushing to our hands so we are ready to fight, to grasp a weapon. The emotion of fear sends blood to our feet so we are ready to flee. The emotion of lust sends blood rushing to a place in between. The short circuit gives our bodies the capacity to exploit windows of opportunity to attack, defend, flee, procreate before the opportunity has passed. In the contemporary world, as opposed to the world of our biological creation, the means of risk management and procreation are institutionalized in ways that make more doubtful the survival value of a brain that is liable to have its faculties for rational deliberation pre-empted by emotion. So the plight of modern humans is to experience regular regret for things we do during those moments when the emotions hijack the brain. For most of us, this is a weekly occurrence, for some a daily one. It does not take something as drastic as a man attacking us with a spear for threat to trigger anger; the brain makes connections of lower-level threats like an angry voice to fire up our anger. This is why emotional defiance to regulatory threats is relevant not only to 'crimes of passion' like murder. Overbearing threats by a government official can engender emotional defiance to what economists would expect to be rational business compliance with regulatory laws. Toni Makkai and I examined compliance by chief executives of 410 Australian nursing homes with 31 regulatory standards. We found that the subjective expected level of punishment did not predict compliance in any of a variety of more simple and more complex multivariate models we were able to

48 For a readily digestible account, if neurophysiologically dubious in parts, of this phenomenon see D. Goleman, Emotional Intelligena (New York: Bantam Books, 1995).

The Rule of Law and the Separation of Powers

310

318

UNIVERSITY OF TORONTO LAW JOURNAL

construc,t. 49 Diagram 1 shows, however, that this result conceals the fact that there were contexts where deterrence worked and other contexts where it not only failed to work, but where there was a counter-deterrent effect. A scale to measure the psychological trait of the emotionality of the chief executive was the only variable we could find which specified when the deterrence model would work and when it would not. When managers were low on emotionality, they responded to perceived increases in threat in a 'cold and calculating' way. But the CEOs who were high on emotionality responded to escalated threats by getting mad rather than by ceasing to be bad. In Diagram 1 if deterrence simply failed for high emotionality managers, the high emotionality line would be flat. In fact it slopes downwards, meaning that for emotional managers, the stronger the deterrent threat, the less compliance. Deterrence fails as a policy not so much because it is irrelevant (though it is for many) but because the gains from contexts where it works are cancelled by the losses from contexts where it backfires. The most concentrated research effort criminology has seen was on the deterrence of domestic violence?> It tends to bear out a similar picture. In a first randomized experiment, Sherman and Berk51 found that arrested domestic violence offenders were less likely to re-offend than those dealt with less punitively. This study had a major effect on public policy. However, subsequent experiments found no net effect of arrest in reducing violence. Again, this overall result concealed the fact that for employed men arrest reduced subsequent violence, while for unemployed men it escalated violence. Sherman interprets this as a result of underclass men who have experienced repeated stigmatizing and unfair experiences with the criminal justice system responding with anger and defiance when they are arrested. Unfortunately, while there are fewer unemployed than employed men in most communities, the violenceescalation effect of arrest for the unemployed was about twice the violence-reduction effect among the employed, giving a nil result overall in most of the studies. How to interpret the various methodological weaknesses of these studies is riddled with controversy. 52 What is clear, 49 J. Braithwaite & T. Makkai, 'Testing an Expected Utility Model of Corporate Deterrence' (1991) 25 Law and Soc'y Rev. 7 [hereinafter 'Testing']; Toni Makkai & John Braithwaite, 'The Dialectics of Corporate Deterrence' (1994) 31 J. of Res. Crime & Del. 347 [hereinafter 'Dialectics']. 50 For a review of the research effort by its most central participant, see L. Sherman, Policing Domestic Violence (New York: Free Press, 1992). 51 L.W. Sherman & R.A. Berk 'The Specific Deterrent Effect of Arrest for Domestic Assault' (1984) 49 American Sociological Review. 261. 52 See E. Stanko 'Policing Domestic Violence: Dilemmas and Contradictions' (1995) Austl. & N.Z.J. Crim. Criminology, Special Supplementary Issue. 31.

The Rule of Law and the Separation of Powers

311

ON SPEAKING SOFTLY AND CARRYING STICKS

319

Diagram 1 Effect of interaction between emotionality and deterrence on compliance with the law"' High Compliance

E

31

- - a - High emotionality

30

Low emotionality

29 m a t e d

c 0

m p I i a n c e

28 27 26 25 24 23 22 High deterrence

21 0

10

20

30

Median Mean

40 1 std dev

Scale for full deterrence model

however, is that never in my lifetime are we likely to see such a large investment of public money in systematic randomized tests at multiple sites of a deterrence hypothesis. If we are not likely ever to get better data than this, at least we need to settle on the conclusion that it is problematic to claim that deterrence is more likely to produce negative than positive effects on those who abuse power. C. COGNITIONS OF STIGMA AND PROCEDURAL INJUSTICE

Moreover, on the same nursing home compliance data set as in Diagram 1, Sherman's approach to interpreting positive and negative effects of deterrence receives support. It was found that nursing home inspectors with a highly stigmatizing approach to law-breakers reduced compliance by 39 per cent in Diagram 2, while inspectors with a reintegrative shaming philosophy for securing compliance improved compliance in the two years following their inspection by the same amount. With reintegrative shaming, the non-compliant act is disapproved, while those responsible 53 Ibid.

312

The Rule of Law and the Separation of Powers

320

UNIVERSITY OF TORONTO LAW JOURNAL

Diagram 2, Mean improvement in compliance for nursing homes where inspectors used high disapproval and high reintegration styles; high disapproval and low reintegration styles; low disapproval and high reintegration styles (N=129; F-Value=3.58; p=.03)

2

Improvement

in

0

Compliance

-1

-2

2

3 Low Disapproval/ High Reintegration

High Disapproval/ High Reintegration High Disapproval/ Low Reintegration

are treated with respect as responsible citizens. With stigmatization, they are shamed disrespectfully, labelled as bad people who have done the bad act. Inspectors who one might call captured, inspectors who were 'nice,' tolerant, and understanding of law-breakers, also made things worse, though the deterioration in compliance following their inspection was less than with the stigmatizing inspectors. Sherman's invocation of the literature showing that compliance is more likely when actors perceive regulation to be procedurally fair suggests that the reasons big sticks rebound are to be found in the psychology of cognition as well as the psychology of emotion. 54 More 54 See T. Tyler, Why Peoj;le Obey the Law? (New Haven: Yale University Press, 1990). A.E. Lind & T. Tyler, The Social Psychology ofProcedural]ustice (New York: Plenum Press, 1988). For the nursing home data discussed in Diagrams 1 and 2, only one of the standard facets of procedural justice - process control - had a significant effect in improving compliance, see T. Makkai & J. Braithwaite, 'Procedural Justice and Regulatory Compliance' (1996) 20(1) L. & Human Beh. 83.

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

313

321

police can increase crime if police are systematically procedurally unfair or stigmatizing in the way they deal with an underclass; more police can reduce crime if they are procedurally fair and reintegrative in their policing. When we increase the number of police, one reason we do not generally achieve a measurable reduction in crime is that we put on the beat a mixture of extra stigmatizing police who make things worse and extra fair, and reintegrative police who make things better. The cognitive mechanisms that produce a 'reactance' against threat have now been the subject of an enormous experimental research effort. This work shows how foolish it is to follow the institutional design advice of Hobbes 55 and Hume 5G of preparing for the worst - assuming that people are knaves. Unfortunately, when we treat people as knaves they are more likely to become knaves. The less salient and powerful the control technique used to secure compliance, the more likely that internalization of the virtue of compliance will occur. Experimental research on children and college students demonstrates the counter-productive effect salient rewards and punishments can have: long-term internalization of values like altruism and resistance to temptation are inhibited when they view their action as caused by a reward or punishment. 57 Over 50 studies examining the effect of extrinsic incentives on later intrinsic motivation indicate that inducements that are often perceived as controlling (e.g., tangible rewards, surveillance, deadlines), depending on the manner in which they are administered, reduce feelings of self-determination and undermine subsequent motivation in a wide variety of achievement-related activities after the reward is removed. 58

These findings seem to be of fairly general import, being supported in domains including moral behaviour, altruism, personal interaction, aggressive behaviour, and resistance to temptation?' Just as strong 55 T. Hobbes, De Cive (New York: Appleton-Century-Crofts, 1949). 56 D. Hume 'Of the Independency of Parliament,' Essay.1; Moral, Politir:al and Literary, vol. 1 (Oxford: Oxford University Press, 1963). 57 M.R. Lepper, 'Dissonance, Self-Perception and Honesty in Children' (1973) 25 Journal of Personality and Social Psychology. 65; M.R. Lepper & D. Greene, The Hidden Costs oj'Reward (Hillsdale, NJ: Erlbaum, 1978); T. Dix &J.E. Grusec, 'Parental Influence Techniques: An Attributional Analysis' (1983) 54 Child Development. 645; M.L. Hoffman, 'Affective and Cognitive Processes in Moral Internalization,' Social Cognition and Social DevelojJment, eds. E.T. Higgins, D.N. Rubble & W.W. Hartup (New York: Cambridge University Press, 1983). 58 A.K. Boggiano et al., 'Use of the Maximal-Operant Principle to Motivate Children's Intrinsic Interest' (1987) 53 Journal of Personality and Social Psychology. 866. 59 M.R. Lepper, supra note 57; R.A. Dienstbier et al., 'An Emotion-Attribution Approach to Moral Behavior: Interfacing Cognitive and Avoidance Theories of Moral Develop-

The Rule of Law and the Separation of Powers

314

322

UNIVERSITY OF TORONTO LAW JOURNAL

external incentives retard internalization, using reasoning in preference to power-assertion tends to promote it. GO D. REACTANCE

Such findings are an important part of an empirical grounding for why republicans should have a preference for dialogue over coercion as a means of checking power. Brehm and Brehmtil constructed a theory of psychological reactance on the basis of the kinds of studies we have been discussing. Diagram 3 shows that the net effect of deterrent threats 62 is the sum of a deterrence effect and a reactance effect. Diagram 3 also shows that reactance is least when we seek to restrict freedom to do something that is not very important to us, greatest when the freedom subjected to control is something the regulated actor deeply cares about. Hence, if freedom to park our car where we want is not an especially important freedom, the way we react to the size of parking fines will be rather like the left-hand panel in Diagram 3. The net effect of threat on compliance will be close to the prediction of a crude rational actor model. If freedom of religion is a vitally important freedom to Christians, then throwing more Christians to the lions may only strengthen their commitment to martyrdom, adding rather than detracting from the growth of Christianity, as in the right-hand panel of Diagram 3. For republicans, the pattern of empirical results summarized in Diagram 3 has an ominous implication. Countervailing threats to check

ment' (1975) 82 Psychological Review. 229; T. Dix &J.E. Grusec, 'Parental Influence Techniques: An Attributional Analysis' (1983) 54 Child Development. 645; A.K. Boggiano et rd., 'Use of the Maximal-Operant Principle to Motivate Children's Intrinsic Interest,' ibid. 60 J.A. Cheyne & R.H. Walters, 'Intensity of Punishment, Timing of Punishment, and Cognitive Structure as Determinants of Response Inhibition' (1969) 7 Journal of Experimental Child Psychology. 231; R.D. Parke, 'Effectiveness of Punishment as an Interaction oflntensity, Timing, Agent Nurturance and Cognitive Structuring' (1969) 40 Child Development. 213; M.L. Hoffman, 'Moral Development,' Carmichael~\· Manual ol Child Psychology, ed. P.H. Mussen (New York: Wiley, 1970); D.Baumrind, 'The Development of Instrumental Competence Through Socialization,' Minnesota SymjJOsium oj'Motivation, vol. 7, ed. A.D. Pick (Minneapolis: University of Minnesota Press, 1973); C.Z. Zahn-Waxler, M.R. Radke-Yarrow & R.A. King, 'Child Rearing and Children's Prosocial Initiations Towards Victims in Distress' (1979) 50 Child Development. 319. 61 Ibid. 62 The theory posits the same form of relationship as in Diagram 3 for reactance to rewards as to punishments. However, the data suggest that reactance to punishment is stronger than to rewards. See S.S. Brehm & J.W. Brehm, Pryr:hological Reactance: A Theory of'Freedom and Control (New York: Academic Press, 1981) at 229-46.

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

315

323

Diagram 3 The interactive effects of force and importance of freedom"' LOW IMPORTANCE OF FREEDOM

Reactance

MOOERA TE IMPORT .>.NCE OF FREEDOM

HIGH IMPORTANCE OF FREEDOM

/'

Reactance

Low

High

High Low

High Low Force Against Exercise of Freedom

Force Against Exercise of Freedom

Force Against Exercise of Freedom

the power of tyrants will work well when it is a power the tyrant cares little about. But try to check the power of a tyrant to enslave his people (when he sees that enslavement as central to his power), try to force a patriarch to desist from family violence when domination of family is a prerogative of utmost importance, then one confronts maximum defiance. At the same time, we will find an encouraging republican implication of the pattern of results in Diagram 3. If deterrence works well (without reactance) for freedoms that actors do not care deeply about, what we can do about organizational abuse of power is target the actors within the organization who care least about the freedom being deterred. In the next section we will find that such deterrable soft targets with the capacity to prevent abuse of power can usually be uncovered by processes of dialogic regulation. Empirical research about how big sticks rebound has been much more manageable (and amenable to random assignment) in relatively micro63 S.S. Brehm &J.W. Brehm, Psyr:hological Rear:tanr:e: A Theory oj'Freedom and Control, ibid. at 60.

The Rule of Law and the Separation of Powers

316

324

UNIVERSITY OF TORONTO LAW JOURNAL

contexts, such as families and nursing homes. Yet there are now the beginnings of an historical scholarship applying the theoretical framework to how stigmatization and punitiveness that is perceived as unjust can threaten violence between nations, indeed how it can risk global tyranny.G 4 The way Hitler exploited the humiliation of the German people at Versailles is the most thoroughly researched instance of this kind of macro-reactance. G5 E. RATIONAL COUNTER-DETERRENCE At all the levels we have discussed - individuals, families, small organizations like nursing homes, transnational corporations and states which are major powers - the conclusion is that simple rational choice models of compliance do not work especially well because empirical results sum contexts where rational threats are productive, non-productive, and counter-productive. In addition to reactance effects that have emotional and cognitive dimensions not captured by rational choice models (rage, defiance, perceived procedural injustice, stigmatization, devaluation of the intrinsic virtue of compliance) the stronger deterrents are, the more rational actors will find countermeasures that effectively undermine the deterrents. Hence, in Diagram 3, we need to discount the rational deterrence effect not only by the reactance effect, but also by measures which rationally subvert the deterrent effect. These grow with the size of the deterrent because the more severe the deterrents, the more reason regulated actors have to invest in counter-deterrence. And of course, the more powerful the regulated actors are, the more capacity they have to mobilize counter-deterrence. An unemployed person convicted of a crime is relatively powerless to do anything to duck the stigma, while a transnational corporation convicted of crime can hire a public relations firm to convince people through full-page counter-publicity that what it did was, all things considered, in the public interest, and that the prosecution was vindictive.3 92 L. Lin, 'The Effectiveness of Outside Directors as a Corporate Governance Mechnism: Theories and Evidence' (1996) 90 North U. L. Rev. 898 [hereinafter 'Effectiveness'). 93 The moral argument has been brought into sharp relief in recent years by laws in many jurisdictions around the world requiring parents and guardians to pay for the damage from the juvenile crimes of their children. My own view is that such laws do perpetrate an intolerable injustice on the parents. Yet there is a difference between this hard-edged parental liability for the crime and exposing a parent to moral censure by concerned citizens for, say, allowing their child to wander the streets at any time of night. One reason that dialogic conferences often seem to work better as a means of controlling delinquency than criminal trials is that the parental gatekeeper is a softer target than the young offender. In New Zealand and Australian family group conferences for juvenile crime I have often observed the following scenario. The victim of the crime, aided by the victim's supporters, explain the suffering the crime has caused them. The offender, a tough with a hard shell, is unmoved. His emotional shell is a shield that protects him from shame directed at him. But when the victim sends a shaft of shame across the room, that shield deflects it so that it pierces like a spear through the heart of the mother of the offender, or the father, or sister, who sit beside him. One of them begins to sob. It then becomes the shame and the sadness of his sister that gets through the offender's hardened exterior. He internalizes the pain he has caused, experiences shame, and from then on is quite unable to sustain the delusion that he has done nothing wrong. Dialogic regulation does mean sharing responsibility for those we care about and feeling it deeply. That is something natural and human, in a way that making the sobbing sister pay the compensation to the victim is not. Dialogic regulation picks up the gatekeeper vulnerability analysis in a subtle and decent way rather than in the barbaric way of legal liability for parents. According to New Zealand Maori traditions of justice, Western justice is barbaric for allowing the offender to stand alone in the dock as an individual accused of a crime. Individuals, in their view, must take responsibility for their wrongdoing supported by those who love them, with those who love them sharing vicariously in that responsibility and voluntarily offering to help as best they can to right the wrong. The offender thereby is not alone, alienated, at risk of the pathology of individualized guilt that can eat away at a person. Rather, they are meant to feel the shame of letting their loved ones down, of causing them to share in this pain. Unlike Western guilt, this shame (whakama) can be readily transcended. As the shame of letting one's family down, it is

The Rule of Law and the Separation of Powers

328

336

UNIVERSITY OF TORONTO LAW JOURNAL

B. TARGETING INSURERS AND SUPPLIERS

The most powerful empirical demonstration of the power of targeting gatekeepers rather than beneficiaries of the wrongdoing comes from the most global of regulatory problems - pollution from ships at sea. Ronald Mitchell has demonstrated how the International Convention for the Prevention of Pollution from Ships (MARPOL) was an utter failure. 94 Signatories were required under the convention to impose penalties for intentional oil spills. The most important targets- petroleum exporting nations - were committed not to enforcing these laws. Most nations simply did not care to invest in proving offences that were difficult to detect. It was only a few petroleum importing nations such as the US who took the requirement seriously. This simply meant that ships had to be a little careful to discharge pollution outside the territorial waters, of these few countries. Non-compliance with the regime was the norm. Then in 1980 the MARPOL regime was reformed in a way that Mitchell estimates has generated 98 per cent compliance. 95 This was a remarkable accomplishment given that the costs of compliance with the new regime were very high for ship owners, that predictions grounded in the economic analysis of regulation were for minimal compliance.% The key change was a move away from imposition of penalties on ships responsible for spills to an equipment sub-regime that enforced the installation of segregated ballast tanks and crude oil washing. One reason for the improvement was transparency; it is easy to check whether a tanker has segregated ballast tanks but hard to catch it actually discharging at sea. But the other critical factor was the role of third party enforcers (a) on whom ship operators are dependent, and (b) who have no economic interest in avoiding the considerable costs of the regulation. These third party enforcers are builders, classification societies and insurance acquitted by forgiveness and restoration within that family. Forcing a family to accept liability directly to a victim obviously cuts totally across this philosophy. (For a more detailed exposition of the gatekeeper dynamic in restorative justice conferences, see ]. Braithwaite & S. Mugford, 'Conditions of Successful Reintegration Ceremonies: Dealing With Juvenile Offenders' (1994) 34 Brit.]. Criminal. 139). 94 R. Mitchell, 'Intentional Oil Pollution of the Oceans,' Institutions for the Earth: Sourt:es of' Effedive International Environmental Protection, eds. P.M. Haas, R.O. Keohane & M.A. Levy(Cambridge, MA: MIT Press, 1993); R. Mitchell, 'Regime Design Matters: International Oil Pollution and Treaty Compliance' (1994) 48 International Organization. 425; R. Mitchell, 'Compliance Theory: A Synthesis' 2 Receil. 327; R. Mitchell, Intentional Oil Pollution at Sea: Environmental Policy and Treaty Comj1liance (Cambridge, MA: MIT Press, 1994). 95 R. Mitchell, Intentional Oil Pollution at Sea: Environmental Policy and Treaty Compliance, ibid. at 270-1. 96 C. Okidi, Regional Control of Ocean Pollution: Legal and Institutional Problems and Prospects (Alphen aan den Rijn, Netherlands: Sijthoff & Noordhoff, 1978) at 34.

329

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

337

companies. Builders have no interest in building cheaper ships which will not get certification by international classification societies nominated by national governments. Classification societies have no interest in corrupting the standards they enforce, which are the whole reason for the generation of their income. Finally, insurers will not insure ships that have not been passed by a classification society acceptable to them because they have an interest in reducing the liabilities that might arise from oil spills. The new MARPOL regime therefore achieves 98 per cent compliance in large part because the effective target of enforcement shifted from the ship operators who benefit from the pollution to builders, classification societies, and insurance companies who do not benefit from it. However, because the ship operators (and builders) are totally dependent on classification societies and insurers, they have no choice but to accept that the regime-compliant ships, which the classification societies have an interest in ensuring, are the only ones that get through the gate. C. THE PLETHORA OF THIRD PAR1YTARGETS

The best-known examples of separating enforcement targeting from the actor who benefits from the abuse is requiring employers to withhold tax from the taxable income of their employees, which they report; or banks to withhold and report tax on the interest earned by their customers. Little enforcement is needed against the employers and banks who withhold and report because they do not benefit from any under-reporting of income. Tax cheating is only a really major problem in those domains where it is impossible to harness such disinterested gatekeepers. Peter Grabosky has initiated a program of work which continually discovers new species of third party enforcers of regulatory regimes from volunteer divers who check compliance with South Australia's historic shipwrecks legislation to elected worker health and safety representatives. 97 Grabosky's work shows just how disparate are the possibilities for shifting enforcement targeting- from actors who benefit 97 P.N. Grabosky, 'Professional Advisors and White Collar Illegality. Towards Explaining and Excusing Professional Failure' (1990) 13 U. New So. Wales LJ. I; P.N. Grabosky, 'Citizen Co-Production and Corruption Control' (1990) 5 Corruption and Reform. 125; P.N. Grabosky, 'Law Enforcement and the Citizen: Non-governmental Participants in Crime Prevention and Control' (1992) 2 Policing and Society. 249; P.N. Grabosky, 'Green Markets: The Environmental Regulation by the Private Sector' + (1994) 16 Law and Policy. 419; P.N. Grabosky Beyond the Regulatory State ( 1994) 27 Australian and New Zealand Journal of Criminology. 192; P.N. Grabosky, 'Counterproductive Regulation' (1995) 23 International Journal of the Sociology of Law. 347; P.N. Grabosky, 'Using Non-Governmental Resources to Foster Regulatory Compliance' (1995) 8 Governance. 527.

The Rule of Law and the Separation of Powers

330

338

UNIVERSITY OF TORONTO LAW JOURNAL

from th~ cheating to actors who do not but on whom the cheat depends for something critical to their welfare. This simple shift is capable of making headway with some of our seemingly most intractable regulatory problems. Another colleague, Neil Gunningham, has long despaired about the way hazardous chemicals regulation succeeds in changing the practices of the top 20 chemicals transnationals, but barely touches thousands of little chemical companies who are too many, too unsophisticated, and too dispersed to be effectively supervised by state inspectors. ns More recently, however, Gunningham has realized that most of these little chemical companies are vitally dependent on TNCs as suppliers, distributors, customers, or all three. This has led Gunningham to the insight that a private or public regulatory regime, which requires major companies to ensure not only that its own employees comply with the regulations but also that the upstream and downstream users and suppliers of its products comply, may massively increase the effectiveness of the regime. Dn The reason is that a TNC that supplies a little chemical company has much more regular contact with them than any government inspector, more intimate and technically sophisticated knowledge of where their bodies are buried, greater technical capacity to help them fix the problems, and has more leverage over them than the state. D. BUILDING A THOUSAND GATES TO THE POWER OF CORRUPT OFFICIALS Privatising public gatekeeping can be one way of separating powers so that enforcement can be targeted on an actor who does not benefit from the abuse of public power. Most national customs services have a lot of corruption. Both senior managers and street-level bureaucrats benefit enormously from bribes paid for turning a blind eye to the under- or over-invoicing of goods. The fact that public customs services have an organizational interest in continuing to sell favours creates a market opportunity for a private organization set up to 'sell trust.' This is just what the Swiss company, Societe Generale de Surveillance (SGS), set out to do when it took over the customs service of Indonesia and other developing countries. It persuades nations to sell large parts of their customs work to SGS through a reputation for incorruptibility that 98 See, for example, N. Cunningham, 'Environment, Self-Regulation and the Chemical Industry: Assuming Responsible Care' (1995) 17 Law and Policy. 57. 99 I rely here on personal communications with Cunningham at ANU seminars. Also seminally relevant here are the insights in Fiona Haines work on regulation in primary and secondary capitalist markets, see F. Haynes, 'The Show Must Go On: Organizational Responses to Traumatic Employee Fatalities Within Multiple Employer Worksites' (PhD Dissertation, University of Melbourne, 1996).

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

331

339

enables it to deliver huge savings to governments. A 1991 Press Statement of the Indonesian Minister of Finance claimed that SGS had saved his country 4.5 billion dollars us of foreign exchange between 1985 and 1990 and earned it 1 billion dollars US in extra duties and taxes. Because it is such testimonials that bring SGS business, SGS has a financial incentive to catch cheats and weed out corruption in its own ranks. A major corruption scandal that would strike everyone as quite normal in the customs service of a developing country might cause financial ruin for SGS. SGS sets up its inspection gates in the country of export (where superior intelligence on over- or under-invoicing is available) rather than in the importing country. It accomplishes this by having over a thousand scrupulously audited offices at all the world's key exporting sites. The company constrains itself from engaging in any manufacturing or in any trading or financial interests that would threaten its independence. 'Selling trust' is profitable, so operatives are well paid. As the company's Senior Vice President,]. Friedrich Sauerlander confessed to me, in an organization of 27 000 people, his internal security organization had uncovered 'some slip-ups.' But in all major ways, it had been possible to sustain an organization with an incentive structure to reward trust. The beneficiaries of the old breaches of trust were left where they were. But through building a thousand gates to their power on the other side of the world, and guarding those gates with SGS units that flourished in proportion to how much abuse of trust they stopped, targeting enforcement on the bad guys inside the gates became mostly redundant. From Lockheed, to polluters from ships, to employers and banks withholding tax, to chemical companies, to outside (instead of inside) Directors targeted by public interest groups over corporate abuse of power, 100 to big adolescent boys exposed at family group conferences for assaulting their mothers and sisters, 101 we can see some promise in 100 The leading example here is the 'Corporate Campaign' against the J.P. Stevens company over their abusive labour practices. The top management team were very hard nuts here. But the campaign was able to so embarrass outside directors that they resigned from the board, a consequence which really did concern top management. See B. Fisse &J. Braithwaite, Imjmr:t, supra note 66 at c. 2. 101 Here an alternative target is an extended family with capacities to monitor, restrain, and shame the violent boy. See]. Braithwaite & K. Daly, 'Masculinities, Violence and Communitarian Control,' Just Boys Doing Business, eds. T. Newburn & E. Stanko (London: Routledge & Kegan Paul, 1994). Also see T. Lajeunesse, Community Holistic Circle Healing: Hollow Water First Nation, Aboriginal PeojJles Collection (Canada: Supply and Services, 1993). G. Burford & J. Pennell, 'Family Group Decision Making: An Innovation in Child and Family Welfare,' Child Welj(m in Canada: Policy ImjJliwtions, eds. B. Galaway & J. Hudson (Toronto: Thompson Educational Publishing, 1995); G. Burford & J. Pennell, Family GroujJ Det:ision Making: New RolesfiJr 'Old' Partner~· in

The Rule of Law and the Separation of Powers

332

340

UNIVERSITY OF TORONTO LAW JOURNAL

shifting enforcement targeting from actors who benefit from their abuse to actors who do not but on whom the abuser depends for something critical to their welfare. VI

Separating Powers Within and Between the Public and Private Sectors

A. SUMMARY SO FAR

This paper has shown that deterrence failure is a major impediment to effective regulation and made a case for two strategies that can beat reactance, the deterrence trap, and other sources of deterrence failure: 1. replace narrow, formal and strongly punitive responsibility (the 'find the crook' strategy) with broad, informal, weak sanctions (dialogic regulation); 2. separate enforcement targeting from identification of the actor who benefits from the wrongdoing. At the macro level of the polity, the combination of (1) and (2) means dialogic regulation combined with robust separations of powers both within and between the public and private sectors. The number of third party enforcement targets is greater to the extent that we have richer, more plural, separations of powers in a polity. For example, under a plural separation of powers, the media baron who sells editorial support and biased reporting to a politician in return for the promise of a television licence might in a more effective republic than the one in which we live have their power checked by • courts of law; • a statutorily independent broadcasting authority that allocates licences only to fit and proper persons and has the capacity to investigate in cases of non-compliance; 102 • industry association self-regulatory bodies; 10 ~ • the Press Council; 104 • corporate charters of editorial independence; 105

102 103 104 105

Resolving Family Violenr:e. Implementation Report Summary (Newfoundland: Family Group Decision Making Project, 1995). Broadcasting Services Act (Cth.) (1992). In Australia, the Federation of Australian Radio Broadcasters and the Federation of Australian Commercial Television Broadcasters are important. P. O'Malley, 'Regulating Contradictions: The Australian Press Council and the Dispersal of Social Control' (1987) 21 Law and Society Review. 83. John Fairfax Publications Proprietary Limited, Charter of Editorial Independence, adopted by the Board of Directors on 12 March, 1992.

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

333

341

• a vigilant journalists association that requires its members to comply with a journalists' code of ethics; 10G • parliamentary oversight committees that investigate abuses of power by the executive, and other (separate) committees that check diligent performance of the duties of independent regulators; • public interest groups that are granted standing to lodge complaints to all of the foregoing institutions; • audit committees of boards of directors (all of whom are outside Directors) with a remit to adjudicate complaints against management for ethical abuses; • corporate ombudsmen with public reporting capabilities; 107 • ethical investment funds with an investigative capacity they can use to put activist shareholders on notice about such abuses of power in media corporations. Separations of powers both within and between the private and public sectors are important to controlling such abuses of power, as is countervailing power from institutions of civil society that muddy any simple public-private divide. Moreover, the more potential targets of third party enforcement such separations of powers throw up, the better the chance that one of them will be a soft target with such leverage over the abuser of power that the simple device of a regulatory dialogue will move that third party to use their leverage to stop the abuse, trigger internal reforms to prevent recurrence, and trigger the private justice system of the organization to discipline those responsible for the abuse. Thus, the richer and more plural the separations of powers in a polity, the less we have to rely on narrow, formal, strongly punitive regulation targeted on the beneficiaries of abuse of power. The more we can rely on a regulatory dialogue wherein an appeal is made to the sense of social responsibility of all actors with a capacity to prevent the wrongdoing, the more persuasion can replace punishment. Reasons for this are that the more hands powers are separated into (a) the more likely that one of those actors with power to prevent will be a soft target, and (b) the more third parties there will be who do not benefit from the abuse themselves 106 In Australia, journalists comply with the Australian Journalists Association Section of the Media, Entertainment and Arts Alliance (formerly the Australian Journalists Association) Code of Ethics (1996). 107 Among the companies that have or have had ombudsmen are the Washington Post, General Electric, Dow Chemical, and American Airlines.]. Braithwaite, 'Taking Responsibility Seriously: Corporate Compliance Systems,' Crmigible Corj)()mtions and Unruly Law, eds. B. Fisse & P.A. French (San Antonio: Trinity University Press, 1985) at 46.

The Rule of Law and the Separation of Powers

334

342

UNIVERSITY OF TORONTO LAW JOURNAL

but who hold power over the abuser. Put another way, the more plural the separations of powers (a) the more overdetermined is the capacity to prevent abuse; and (b) the more cases there are of disjuncture between an interest in the abuse and a capacity to prevent it. B. PLURAL PRIVATE SEPARATIONS; PLURAL PUBLIC SEPARATIONS

This paper seeks to correct the bias of the republican tradition toward a focus on separations of public powers. Yet the arguments advanced are as relevant to abuse of power by the police as they are to a private media organization. The head of state who rigs electoral boundaries is a hard target because nothing is more important to her career than the election outcome. Citizens who ask a judge to overturn the head of state's electoral rigging approach a softer target because the judge does not benefit from the election result. The traditional separation of powers between executive government and judiciary can deliver the benefits revealed in our analysis of private sector disjunctions between interest and preventive power. At the same time that our novel rationale for the separation of powers shows traditional republican thinking about the separation of powers in the public sphere to be somewhat impoverished, its impoverishment arises from the fact that it is not as plural as it might be. Three ( executive, legislature, judiciary) is not a very big plurality. Moreover, antifederalist separations of powers (in some pre-revolutionary us state constitutions) aspired to avoid the concentration of power (as did its private sector analogue in anti-monopoly law) by having the executive responsible for X, the legislature for Y, the judiciary for z. In the most uncharitable reading of this arrangement, each branch is left alone to abuse power without too much interference within its own sphere from the other branches of government; a strict separation of powers simply assures that the sphere of each is not too broad. This would be uncharitable, however, because in all of the early us state governments, while each branch had spheres of independence from the other branches, they also had spheres where their power was checked by the other branches of government. 108 108 M.S. Flaherty, 'The Most Dangerous Branch' (1996) 105 Yale LJ. 1725. Yet, late in his career, no lesser a republican figure than Jefferson (much influenced by John Taylor's, An Inquiry into the Prinr:ijJles and Policy of the Government of the United States, 1814 [New Haven: Yale University Press, 1950] ) became an advocate of a total separation of the powers of three branches, rejecting the dominant view of John Adams and the Federalists that there should be some overlapping so there could be mutual checking of power. For Taylor, 'Instead of balancing power, we divide it, and make it responsible' (by which he and Jefferson meant the three branches all must be responsible to the people by direct election) (ibid. at 88). See Vile, supra note 6 at

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

335

343

Even so, the reconceptualization of the rationale for the separation of powers in this paper implies that in debates on the separation of public powers, attention is needed not just to assurance of the independence of honest judges from corrupt parliamentarians and corrupt executive governments (and vice versa). Attention is also needed to making corrupt, self-serving, nepotistic judges who flout the rule of law; patrimonial parliamentarians; and evil executives each vulnerable to the power of the other branches. One of the problems we must confront if we are to make progress with some of our tougher problems, like police corruption, is how to deal with pleas that any encroachment on the independence of the police will take us back to a world where the police lock up whomsoever the executive government tells them is a troublemaker. How do we get universities that are fearless in undertaking research of which the state disapproves, yet that do not use this independence to deflect responsiveness to a community which sees it as dominated by venal godfathers of disciplines of less value than other neglected spheres of scholarship? The answer proposed is to have a police force and a university that are sufficiently autonomous from state power, business power, church power, media power, and the power of professions like law not to be dominated by them. Part of their resilience to any single source of domination will come from their very dependence on all those sources of power. We need a police that is vulnerable to publicly reported surveys of citizen satisfaction with the respect police show for rights, Hl!l to meetings of the Police-Aboriginal Liaison Committee, to meetings of the Police-Gay Liaison Committee, to meetings with local businesses concerned about break-ins at their factories, to meetings of local Neighbourhood Watch groups, to criticisms made at family group conferences that the police officer involved was unnecessarily rough, to the Ombudsman, to Parliamentary Committees, to Royal Commissions to investigate matters of extraordinary malfeasance, to a free and fearless press, to the Council for Civil Liberties, to the judiciary, and yes, vulnerable to an executive government that will sack the Commissioner if there is reasonable 163-70. The late eighteenth-century French constitutions also rejected the idea of checks and balances in favour of a strict separation of powers, at least until the lessons of Robespierre and Napoleon Bonaparte had been learnt. These were lessons about the fragility, adversarialism, and vulnerability to tyrannical coup d'etat of purist democratarian separations of powers, see Vile, supra note 6 at 198-9. Madison had foreseen that the best way to preserve the separation of powers was 'by so contriving the interior structure of government as that its several constituent parts may, by their mutual relations, be the means of keeping each other in their proper places' (A. Hamilton,]. Madison &J.Jay, supra note 9 at 318-9, see also at 302-18.) 109 J. Braithwaite, 'Good and Bad Police Services and How to Pick Them,' Policing Australia, eds. P. Moir & H. Eijkman (Melbourne: Macmillan, 1992).

The Rule of Law and the Separation of Powers

336

344

UNIVERSITY OF TORONTO LAW JOURNAL

suspicion· that she is corrupt. Mter 200 years of ugly tyranny in nations with beautiful constitutions, it is no longer persuasive to suggest that a separation of state powers will ensure that the government 'will be controlled by itself.' 110 In other words, a police service that is enmeshed in many webs of dependency will be vulnerable to the many when it corruptly does the bidding of the one. This, I suspect, is the way of resolving the dilemma of independence for different branches of government versus checking of power between branches of government. Checking of power between branches of governm'ent is not enough. The republican should want a world where different branches of business, public, and civil society power are all checking each other. While the broad principles are clear here, the nuts and bolts of checks and balances, of independence and interdependence, require contextual deliberation for any given source of power. Clearly, there must be some sorts of power against which a police service must be protected by law ('Arrest this man or we will cut your budget.'). Republican theory of the sort Philip Pettit and I endorse 111 requires detailed empirical investigation of the different ways of organizing independence and interdependence so as to discover a set of institutional arrangements most likely to maximize freedom as non-domination. At least the principle of separating powers so that there are enough actors with the independence and preventative capacity so that one of them can be moved by dialogue to stop the abuse is clear enough. In the public arena, the literature on separation of powers bequeaths to us a variety of reasonably well-understood heads of public power that might be separated- different houses of parliament, levels of government in a federation, lower versus appellate courts, administrative appeals tribunals, and so on. While it is a tricky business to put together or tinker with a robust public architecture of powers, at least we have some sense of the elements we might play with. In contrast, the separation of private powers is comparatively under-researched. Here most readers will need some elementary sense of what might be involved in separating private powers. The next section is a preliminary foray into what might be at issue. VII How to Separate Powers in the Private Sector A. POWERS OF SHAREHOLDERS, DIRECTORS, MANAGERS

The law review literature on corporate governance has a deal of useful guidance on accomplishing separations of powers in the private sector, 110 A. Hamilton,]. Madison &].Jay, supra note 9 at 323 (James Madison). 111 P. Pettit, supra note 10;]. Braithwaite & P. Pettit, supra note 10.

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

337

345

but not useful enough. The concentration tends to be on the separate powers of shareholders, directors, and managers. 112 Important separation of powers issues are at stake here; such as whether a majority of members of the board of directors of a public company should be 'unrelated directors,' 113 meaning that they have no business dealings with, nor a management position in, the company; whether the nominating committee for the appointment of new directors should have no management directors on it; 114 whether it should be forbidden for the CEO to be Chairman of the Board; 115 whether there should be a bicameral board with a supervisory board as in Germany, France, the Netherlands, and Indonesia; 11 G and generally how to give outside directors a role more than that ofthe CEO's 'pet rocks.' 117 There is certainly merit from a republican point of view in engendering shareholder democracy, encouraging activist shareholders to call management to account, 118 securing representation for minority shareholders on the board, and effective monitoring of the board by institutional shareholders. m When the New York Stock Exchange first required a Board Audit Committee of non-Executive Directors as a condition of listing on the exchange in 1977, 120 this was an important step for the separation of private powers. It has spread to many parts of the world. Long before that, in 1862, a more important step was requiring companies as a matter of law to be audited by a pro-

112 M.A. Eisenberg, 'Legal Models of Management Structure in the Modern Corporation: Officers, Directors, and Accountants' (1975) 63 Cal. L.R. 363; K.A. Kirwan, 'The Use and Abuse of Power- The Supreme Court and Separation of Powers' (1995) 537 Annals of the American Academy of Political and Social Science. 76; B.H. Siegan, 'Separation of Powers and Economic Liberties' (1995) 70 Notre Dame L. Rev. 415; M. Hardt, 'The Non-Equivalence of Accounting Separation and Structure Separation as Regulatory Devices' (1995) 19 Telecommunications Policy. 69. 113 Toronto Stock Exchange Committee on Public Governance in Canada, 'Where Were the Directors?' GuidelinesfiJr Imj~roving Cr~rjwmte Governance in Canada (Toronto: Toronto Stock Exchange Committee on Public Governance in Canada, 1994) at 4. 114 Ibid. at 4-5. 115 Ibid. at 41.; S. Bottomley, 'From Contractualism to Constitutionalism: A Framework for Corporate Governance' (draft paper) [hereinafter 'Contractualisrn.' 116 S. Turnbull, 'Governance Flaws and Remedies,' Corporate Directors' diploma Course, University of New England, Topic 7.2 (Armisdale, NSW: University of New England, 1993). 117 GM outside directors are reported as having had a 'pet rock' self-image in L. Lin, 'Effectiveness,' supra note 92 at 898, notes 246, 940. liS A. Fraser, 'Reinventing Aristocracy: Corporate Governance in the Civil Constitution of a Modern Republican Society' (unpublished manuscript). 119 G.P. Stapledon, Institutional Shareholders rLnri Corjwmte Governance (Oxford: Clarendon Press, 1996); S. Bottomley, 'Contractualism,' supra note ll5. 120 L. Loss, Fundamental\· o{ Securities Regulation (Boston & Toronto: Little Brown & Company, 1983) at 484.

The Rule of Law and the Separation of Powers

338

346

UNIVERSITY OF TORONTO LAW JOURNAL

fessionally certified auditor; 121 This is the well-understood end of the separation of private powers. Useful literatures already exist on how to make these separations work better: how to improve shareholder accountability, restructure Directors' duties, bring board audit committees to life, and improve the performance and independence of auditors. The emphasis on separation of powers between management and the two other branches of corporate governance neglects the main game, however, which is separation within management. With private power, moreso even than with public power, the power in the hands of the other branches of governance is extremely modest compared with the concentration of power in the executive. B. AUDIT AND AUDIT OF AUDIT

The most important steps toward separating powers within management have involved internalising the outside audit by setting up internal reporting of audit-based accountability. This has not been restricted to financial auditing. In private companies today, environmental audit, safety audit, audits of compliance with the us Foreign Corrupt Practices Act (introduced after the Lockheed scandal), antitrust audit, and the like, are common. 122 During the past decade, there has been a us-based movement to partially integrate these functions under the rubric of legal audit. 123 Companies like Exxon have long had a functional equivalent of such an internal audit of the auditors. Instead of doing it under the auspices of a legal audit by the general counsel, from 1973 Exxon used the office of the Controller, a vice president at corporate headquarters in New York. Brent Fisse and I investigated the internal auditing of Exxon fifteen years ago following the payment of 46 million dollars us in bribes in Italy, more minor payoffs in Thailand, the Dominican Republic, Indonesia and Japan and the improper recording of political donations in Australia as legal fees. 124 By 1981 there were no fewer than 400 internal auditors working for Exxon. These were by no means all financial auditors; they included engineers, for example, who would work on teams with the 121 The English Registered Companies Act of 1862. See G.R. Brown, 'Changing Audit Objectives and Techniques' (October 1962) The Accounting Review. 696 at 697; DJ. Fraser & M.E. Aiken, Stettler's System Based Audits (Sydney: Prentice Hall, 1981) at 12-3; R. Tomasic & S. Bottomley, Corj;amtions Law in Australia (Sydney: Federation Press, 1995) at 15-17. 122 For example, T. Lamond & I. Watt, Environmental Management and Audit Manual (Sydney: CCII Australia, 1996). 123 L.M. Brown & A.O. Kandel, The Legal Audit: Carjwrate Internal Investigation (New York: Clark, Boardman, Callaghan, 1990). The journals Preventive Law Reporter and Corporate Conduct Quarterly are devoted to this topic. 124 B. Fisse &J. Braithwaite, Imj;ar:t, supra note 66 at c. 15.

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

339

347

accountants to ensure that the company's environmental policies were being implemented in all subsidiaries. Auditors in each Exxon subsidiary reported directly to their chief executive, but they also had a dotted line reporting relationship up to the Controller in New York. It was the latter who decided the size of each auditor's workforce. Auditors were therefore not tied to the purse strings of those whom they were auditing. Auditors could also bypass the Controller in circumstances where people in the Controller's office might be part of a cover-up. Indeed, the most lowly local auditor who detected, say, an improper bookkeeping practice, was required to ensure that this was reported through the general auditor to the audit committee of the Board in New York. A former Exxon general auditor, Ted Kline, explained: Say, for example - I was an auditor and uncovered something that was unsavory or should have been reported and told my supervisor and he said to let it go. Well, auditors are briefed that their obligations do not end there. The employee who makes the report knows that his supervisor should report it up the line and that if his supervisor does not, the auditor must. He cannot seek sanctuary, so to speak, just by saying, 'It was not my job. I told my boss and that was it .... ' Much to the consternation, I guess, of some employees perhaps who would say, 'Surely the Company does not want me to put this in writing'; and the answer is 'yes.' We want it to be put in writing, exactly what happened, and we will send it right up to New York to the board audit committee. There is no other way we know that we can get the message across that we are very serious about this. 125

Auditors have a responsibility to report on matters that are not within their area of direct responsibility. For example, auditors in most companies who discovered a letter suggestive of a price fixing conspiracy would regard this as none of their business. An effectively working separation of corporate powers tackles such a perception, as the same former Exxon general auditor explains: When an auditor reaches a situation where he needs to question whether we have violated antitrust laws, then he needs to go to the Law Department in order to ascertain that. Most of our auditors are not lawyers and they are not qualified to find out whether we have committed a violation. When he does contact the Law Department he makes a record of the fact that this was turned over to the Law Department and he sends word up through audit channels to New York that he has turned over· a situation to the Law Department. The Law Department then sends up through the Law Department channels the fact that they are handling the situation, so the New York Law Department knows that it is 125 W. Clifford & ]. Braithwaite, Cost Effective Business Regulation (Canberra: Australian Institute of Criminology, 1981) at 30.

The Rule of Law and the Separation of Powers

340

348 UNIVERSITY OF TORONTO LAW JOURNAL

not being covered up down in the individual territory someplace. Anything having to do with antitrust is both a complicated and very delicate matter so we make sure it is carried up through the Law Department. We do not let an auditor walk out on a limb and carry something all the way to the board audit committee and then the lawyers shrug their shoulders and say that it is not an antitrust violation at all. That is the way we handle things when we are not sure it is a law violation. 126 At the heart of Exxon's auditing system was a marrying of independence and interdependence designed to deliver an effective intracorporate separation of powers. The former general auditor explained the key concepts as follows First- the well known segregation of duties and responsibilities such that no single function, departme_nt or employee will have exclusive knowledge, authority or control over any significant transaction or group of transactions; Second- the proper documentation of transactions and business events; Third - the systematic and thoughtful supervision in documented reviews of managers and other employees' work; Fourth - the timely preparation of records, reports and reviews; Fifth - control measures designed in such a way that they are responsive to the nature and degree of risk and exposure .... ; Sixth - the various aspects of control should not be so interdependent that a serious deficiency in any one would make other controls also ineffective. 127 The separation of powers became explicit when the full Board of directors annually received the report of the board audit committee on compliance with Exxon's rather encompassing Business Ethics Policy. At the same meeting, independent reports on the same subject were received from the company's external auditors, the auditor general, the controller, the general counsel, and the CEO. The globalization of business has enabled new separations of powers, new answers to the question of 'Who audits the auditors?' At Exxon, IBM, and in many leading us and European pharmaceutical companies, the auditors from one country regularly audit the auditors from another; regional office auditors audit the auditors of national subsidiaries, and a national auditor or the Asian regional auditors may audit the head office auditing group in New York. Arranging guardianship in a circle is an advance on the hierarchical organization of guardianship that historically allowed companies like Lockheed and Exxon to rot, like fish, from the head down. In the hierarchical organization on the left side of Diagram 4, the only guarantee against corruption by an nth order guardian is an 126 Ibid, at 34. 127 Ibid. at 24-5.

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

341

349

Diagram 4 Formal models of hierarchical and republican conceptions of trust

Guardian 1

Guardian n+ 1

t

Guardian n

t t

Guardian 2 Guardian 3

Guardian 1 Heirarchical fiduciary conception of guardianship

Republican conception of guardianship

n +1 th order guardian; if the guardian at the top of the hierarchy is corrupt, then all assurance against abuse of power by the system collapses. Yet guardianship which is organized in a circle is still a reasonably closed system, even when the circle includes external auditors and an Ethics Committee of outside directors. C. ENTERING THE CIRCLE WITH OUTSIDE POWER

Progress to bring into the circle utterly outside watchdogs, who do not depend on the corporation for an income, is slow in the world of corporate governance and restricted to rather special domains of accountability. In Australia, consumer movement nominees on the Banking Ombudsman Council have access to the consumer complaint records of private banks and have a public reporting responsibility. The most widespread accomplishments in widening the circle of guardianship to give total outsiders a window to examine the audit performance of companies have arisen under the European Union's EMAS (Eco-Management and Audit Scheme). 128 These are voluntary standards, yet with considerable force in a world where many major purchasers (particularly state purchasers) and some insurers require

128 Council Regulation (EEC) No. 1836/93 (29 June 1993) Allowing Voluntary Participation by Companies in the Industrial Sector in a Community Eco-Management Audit Scheme, Official journal of the European Community L 168/1 (10.7.1993);

The Rule of Law and the Separation of Powers

342

350

UNIVERSITY OF TORONTO LAW JOURNAL

certification from companies in environmentally high-risk industries. EMAS requires companies to demonstrate continuous improvement in environmental impact and product stewardship to an outside environmental auditor. The report of that outside auditor is generally public and therefore can be examined by green groups on the lookout for environmental scandals. There is a much longer history of empowerment of constituencies internal to the corporation, which have very different interests from management, and affiliations to power bases outside the organization. The leading example here is elected union health and safety representatives who report both to the management, who pays them, and the union, which legitimates them. 129 Rights of access to safety data are often negotiated as a matter of contract between the union and the employer, 130 or as a matter of public law. These rights of access are sometimes checked by union-employed safety inspectors who conduct inspections of workplaces independently of state inspectors. In the case of large unions like the us United Mine Workers these can be many dozens in number. 131 A comparison of high-accident and low-accident coal mines by Pfeifer, Stefanski, and Grether found that both safety directors and miners in mines with low accident rates reported that in the low-accident mines the union put greater pressure on management for safety through bargaining and dialogue. 132 In Australian coal mines, elected worker inspectors do an independent check of a mine before a shift starts to double-check the assessment of company safety staff that levels of methane and other fundamental concerns are under control. Their assessment of the safety of the workplace is written in a record book at the entrance to the mine, which is available to all workers and to government safety inspectors. Union inspectors have a legal right to prevent or stop work at a mine on safety grounds until such time as a government safety inspector can get out to the mine to adjudicate whether the safety stoppage is justified. This is quite an impressive separation of private powers that has existed in the British Empire for more than a century EMAS

129 See N. Cunningham, Safeguarding the Worker: The Role r!f Law (Sydney: Law Book Company, 1984). 130 J. Braithwaite, To Punish or Penuade: Enfarr:ement of Coal Mine Safety (Albany: State University of New York Press, 1985) at 8 [hereinafter Punish]. 131 Ibid. at 8. 132 M.C. Pfeifer,Jr.,J.L. Stefanski & C.B. Grether, Psyr:hologir:al, Behavioral and Organizational Factors Affixting Coal Miner Safety and Health (Columbia, MD: Westinghouse Behavioral Services Center, 1976); see also D.R Biggins, M. Phillips & P. O'Sullivan, 'Benefits ofWorker Participation in Health and safety' (1991) 4 Labour and Industry. 138.

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

343

351

and that can also be seen in mine safety auditing in Japan, the United States, France, Romania, and Poland. 133 While the union is an independent external agent of certain categories of employees for certain types of problems, professions are more general external agents. A corporate circle of accountability is less closed to the extent that a general counsel who is in it feels an allegiance to the ethics of the legal profession that approaches in strength allegiance to the corporation, an accountant who feels ethical responsibility to the accountancy profession, and so on. Western nations are witnessing a proliferation of new professionalisms relevant to penetration of corporate accountability by allegiances to values from outside the company. These include emerging professionalism in environmental auditing, occupational health and safety, the Society of Consumer Mfairs Professionals in Business, pharmacology (especially important in the pharmaceutical industry), and now in Australia even an Association of Compliance Professionals, and internationally a Regulatory Mfairs Professionals Society. D. REPORTING ARCHITECTURE TO SECURE TRANSPARENCY AND INDEPENDENCE As the earlier comments about the Exxon system imply, reporting relationships are critical to an effective separation of corporate powers. Checking of power cannot work without a transparency that renders abuses in one area visible to another sphere of power. A recurrent abuse in the pharmaceutical industry occurs when the production manager of a plant, who is paid performance bonuses and promoted on the basis of getting product out of the plant, overrules a finding of her quality control manager that a batch of drugs does not m+eet specifications. 134 The chances of a batch of drugs that just fails to meet specs causing side effects that would be sheeted home to this failure are slight, while the payment of the production manager's bonus may be a certainty, but only so long as she gets the batch at issue out on time. Hence, the incentive of the manager to break the law. A simple solution to this problem was adopted by some of the more quality-conscious TNCs in the 1970s, and has now been mandated in the laws of a number of countries. This is that a production manager is not allowed to overrule a quality-control judgment on a batch of drugs. It can only be done over the signature of the CEO. The effect is to make the perverse incentives the firm creates for the production manager to break the law transparent. Another effect is to taint the CEO and people who advise her with knowledge (the reverse of a vice president responsible for 133 134

J. Braithwaite, Punish, supra note 130 at 9-10. J. Braithwaite, Pharmaceutir:al, supra note 68 at c. 3.

344

The Rule of Law and the Separation of Powers

352

UNIVERSITY OF TORONTO LAW JOURNAL

going tojail policy). The final effect is to strengthen the hand of qualitycontrol management against the normally more senior production managers. CEOs in practice are extremely reluctant to overrule qualitycontrol recommendations because the cost of redoing one batch of drugs is a comparatively minor matter to them, and is perhaps a good way of sending a message to production managers to improve their performance on quality. Yet the prospect of a batch causing a fatality, however remote, could also be fatal for the CEO, as could the CEO tolerating a culture of sloppiness about quality. This then is an example of how clever reporting architecture assures the separation of powers between quality control and production, prompts the need for dialogue on a quality culture that tolerates no errors on pharmaceutical specifications, and shifts decision making following the dialogue from a hard target with incentives to abuse power (the production manager) to soft targets, with incentives to uphold the law (the CEO and the quality-control manager). In a more generalized way, the post-scandal reporting policies Brent Fisse and I discovered at Exxon were exemplary in the way they ensured, if implemented, 135 that the soft targets in the company, who could be moved by ethical dialogue, got to know about the temptations to which hardened crooks within Exxon were succumbing. As the Controller explained to us, effective control means having an organization full of 'antennas.' All units of the organization had a responsibility not only to report confirmed violations of the 'Business Ethics Policy' but also 'probable violations.' 'Probable violations' were defined by corporate policy as 'situations where the facts available indicate that a violation probably occurred, even though there was insufficient information for a definite determination.' 13G Hence, a matter could not be sat upon on the strength of being 'under investigation.' However, an obligation to report 'probable violations' is a less potent protection than a responsibility to report 'suspected violations' as well. When a violation is reported, there is an obligation on the part of the recipient of the report to send back a determination as to whether a violation has occurred, and if it has, what remedial and/ or disciplinary action is to be taken. Thus, the junior auditor who reports an offence and hears back nothing about it knows that his or her report has been blocked or sat on somewhere. She then must use the safety valve channel direct to the board audit committee. 135 We have no data on how rigorously these policies were implemented throughout this massive organization, though we have some doubts. On the other hand, we are confident they were policies that top management took seriously at the time of our research in the early eighties. 136 Exxon Corporation, Australian Complicance Professionals Association (New York: Exxon, 1978).

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

345

353

If she does not, she is in breach of the Business Ethics Policy for failing to ensure that the problem gets either resolved or put before the board ethics committee. Many companies have policies requiring the reporting of ethics violations, but not many have policies that oblige the reporter to assure that the report is not blocked. This is important because one thing we know about criminal corporations is that they are expert at structuring communication blockages into the organization to protect top management from the taint of knowledge. This had been true when Exxon was paying off politicians around the world during the seventies. We cited memos from this era with statements like' [d]etailed knowledge could be embarrassing to the Chief Executive at some occasion on the witness stand.' E. THE STRATEGIES

The strategies of generic importance for separating private powers we have discussed are 1. better securing the separation of the powers of the three major branches of corporate governance - shareholders, directors, and managers; 2. better separating powers within management - quality and production, environment and production, for example; 3. expanding audit capabilities to a range of areas beyond finance safety, antitrust, ethics, for example; · 4. professionalizing audit so that internal auditors have an external professional allegiance to balance corporate loyalty; 5. abandoning hierarchies of accountability in favour of circles of accountability so that auditors audit auditors, ensuring that someone guards the guardians; 6. allowing outsiders with interests different from corporate interests into the circle of accountability - unions into safety management circles, consumer group representatives into consumer complaint handling circles, greens into environmental circles by mandated public reporting of corporate environmental objectives, and public reporting of audits of whether the objectives are attained; 7. guaranteeing transparency and tainting soft targets with knowledge by institutionalizing a safety valve reporting route direct to a board audit committee, a corporate ombudsman, or both; 8. in domains where serious abuse of power is at risk, independent reports on compliance to the board audit committee from separate powers - line management, legal, audit, unions; 9. obligations on all employees to report suspected violations of law and violations of all corporate policies that involve an abuse of corporate

The Rule of Law and the Separation of Powers

346

354

UNIVERSITY OF TORONTO LAW JOURNAL

power (e.g., corporate ethics and environmental policies). Obligations to report the suspected violations direct to the board when the employee does not receive back a written report that the matter has been satisfactorily resolved. Failure to meet this obligation being itself an ethical breach that colleagues have an ethical obligation to report. For each of these strategies there is a debate to have about whether they should be mandated by the state, left to business or professional selfregulation or seen as demands that social movements should seek to extract directly from private power. These debates will not be engaged here. VIII

Separation of Powers and Efficiency

A. TWO CONCERNS

Since the decline of communism and fascism, not many of us believe that totalitarianism is more economically efficient than democracy with checks and balances. 137 This is not to deny that there are trade-offs between the virtues of checking power and the inefficiency of doing so. Sometimes there are significant economic costs that we are happy to bear when constitutional courts hold up decisions of executive governments. On the other hand, economists are more convinced than ever that a strict separation of powers between the central bank and other branches of government, 138 and political independence for the authority that enforces anti-monopoly law, are essential for economic efficiency. We must consider, however, whether the reconceptualized rationale for a public-private separation of powers I have articulated raises a spectre of new inefficiencies which we may not happily bear. One worry arises from the separation of enforcement targeting from the actor who benefits from the abuse of power. According to the economic analysis of law, regulatory enforcement should be designed so as to deliver an optimal level of attainment of a regulatory objective. Pursuit of perfect compliance with, say, environmental law is an error. We should set sanctions at a level where inefficient non-compliance will always be 137 Advocates of the separation of powers such as Jefferson have always been concerned to separate powers in such a way as to allow the executive government to get on with governing. See W.G Gwyn, Meaning, supra note 2 at 33-5, 58, 118-20. For my own account of why a rich republican democracy is likely to be more economically efficient than totalitarianism, see, J. Braithwaite, 'Institutionalizing Distrust: Enculturating Trust,' Trust and Democratic Governance, eds. M. Levi & V. Braithwaite (New York: Russel Sage Foundation, 1997). 138 0. Issing, Central Bank Independence and Monetary Stability (London: Institute of Economic Affairs, 1993).

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

347

355

deterred. But when the economic benefits of non-compliance exceed the environment costs reflected in a fine the firm is willing to pay, we should want it to break the law. Economic welfare will be enhanced when it pays the fine to compensate for the environmental damage and goes ahead with the benefits of the economic activity. This kind of analysis is based on a variety of assumptions that prove to be false in regulatory practice as a result of vice presidents responsible for going to jail bearing costs, while presidents get benefits without costs, and other complications. The most basic complication is that managers routinely have little idea of the costs and benefits of regulatory compliance. A study of nursing home regulatory compliance, by Makkai and Braithwaite/ 39 found, for example, that while the dominant models in the economic analysis of law assume that the relationship between the expected cost of compliance and compliance is monotonic, it is non-monotonic (a curve with a turning point). The models assume that expected costs are determined by actual costs, whereas only 19 per cent of the variance in expected costs was explained by actual costs. Some of the limits of economic models in this case were explained by CEOs who were 'disengagers.' The behaviour of disengagers was to be understood not in terms of rational game playing but in terms of dropping out of the regulatory game. The disengagers were in the regulatory system but not of it and certainly not economically calculative about it. Makkai and Braithwaite show that the fit of economic models can be improved by adding attitudes of managerial disengagement to the models. What all this means is that if we use under-specified pure economic models to determine the optimum level of penalties to ensure that rational actors, who bear both the costs and benefits, make choices that maximize welfare, we will fail to do so. Economic models with a grain of truth and a gram of falsity will prescribe major under-compliance or over-compliance most of the time. But let us assume there are some contexts where the assumptions of economic models of law are not false. The worry is that by shifting enforcement on to gatekeepers who do not care about the benefits of the polluting activity, who care only about the costs of sanctions, compliance will be secured in circumstances where compliance is inefficient. We will get over-compliance. This worry also exists in domains where false economic models have set levels of standards and penalties for their breach that cause over-compliance even before cost-bearing is separated from benefit-taking.

139 T. Makkai &J. Braithwaite, 'The Limits of the Economic Analysis of Regulation: An Empirical Case and Case for Empiricism' (1993) 15 Law and Policy. 271.

348

The Rule of Law and the Separation of Powers

356

UNIVERSITY OF TORONTO LAW JOURNAL

Another worry about the rather radical separation of powers in the us system of government is that it nurtures fiscal irresponsibility. For example, when there is a clear need to reduce the deficit by increasing taxes or cutting spending, very often this does not happen. It is a kind of mismanagement the Americans are more able to live with than other nations because of the many other sources of resilience in their economy. When failure to cut a deficit occurs in the face of a transparent need to do so, the political tradition is for the President to blame the Congress, the Congress the President. A system of mutual irresponsibility, if we paint it in the worst light. Separations of powers can induce economic inefficiencies, either from over-compliance or from mutual fiscal irresponsibility. But there is a common solution to these two problems and it is a republican one. Both types of inefficiency arise from an adversarial approach to how regulation should work in the first case, and to how politics should work in the second. Concerns about over-compliance and mutual irresponsibility are ameliorated to the extent that we can replace adversarialism which fortifies self-regarding interests with dialogue that redefines interests in a public-regarding way. B. OVERCOMPLIANCE

In the form of regulation described in this paper and elsewhere in the work of new regulation scholars, 140 actors discuss regulatory problems in a way that is not totally self-regarding. In the conferences following nursing home inspections, the discourse is public-regarding on all sides. Of course there is underlying self-interest. But it is inconceivable that a nursing home manager would dismiss a plea to reduce the pain of a resident by suggesting that she did not care about the pain, only about the costs. The more genuine the dialogue, the more seriously management internalizes the concerns of the residents; the more seriously residents and regulators internalize the concerns of management. Under such a regime of other-regarding actors negotiating outcomes with recourse to sanctions only when good faith breaks down, any conception of an economically optimal level of regulatory sanctions makes no sense. The hope for avoiding over-compliance rejects the spurious quantification of the economic analysis of law in favour of a negotiated search for win-win solutions based on a regulatory culture wherein actors 140 E. Bardach & R. Kagan, Going by the Book: The Problem of Regulatory Unreasonableness (Philadelphia: Temple University Press, 1982); V. Braithwaite, Games ofEngagement: Postures within the Regulatory Community (forthcoming); V. Braithwaite et al., 'Regulatory Styles, Motivational Postures and Nursing Home Compliance' (1994) 16(4) Law and Policy. 363; E. Meidinger, 'Regulatory Culture: A Theoretical Outline' (1987) 9 Law and Policy. 353.

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

349

357

are seen as virtuous when they take seriously not only their own concerns but the concerns of others, including the concerns of business about over-compliance. Empirically, it often happens that nursing home managers do speak up when they fear over-compliance - regulatory demands which deliver small benefits at such a high cost that if only some of those dollars could be spent elsewhere in the home, both sides would be better off. Indeed, it is common for regulators and residents to speak up about shifting resources from over-compliance here to undercompliance there. Republican regulatory institutions foster the creative search for. optimality through sharing ideas and mutual identification with the interests of the other. The mutuality that engenders this creativity is destroyed by the adversarialism of deterrent-based regulation and, indeed, by the game-playing of tax-incentive-based regulation. 141 In the day-to-day work of little regulatory decisions on the ground, dialogue in which all actors empathize with concerns about one another's costs and benefits is the most practical protection we can suggest against over- and under-compliance. Going after 'soft targets' would be a dangerous strategy if it were not embedded in a plural dialogue among actors with different spins on costs and benefits. More precisely, it would be a contingently dangerous threat to the efficiency of a regulatory regimedepending on whether the economic analysis underpinning it was misspecified in a way that caused under- or over-compliance. Contextual dialogue among hard and soft targets moves the soft targets more than the hard, yet in a way that has the soft targets listen to the point of view of the hard targets. The irony in our observations of nursing home regulation is that this kind a dialogue about regulatory costs actually generates some contextual data about costs to replace the ill-informed 'gut feels' that managers have about regulatory costs when subjected to traditional command and control directed at the beneficiaries of cheating. C. FISCAL IRRESPONSIBILI1Y

To the extent that the game of American politics is totally adversarial and politically self-interested, parties will simply manoeuvre to make the other party impose unpopular taxes or spending cuts while avoiding doing so themselves. Fortunately, American political life does attenuate adversarial self-interest through dialogic mechanisms. The President meets with Congressional leaders to discuss public-regarding solutions to gridlock in 141 In as-yet-unpublished data on us nursing home regulation, Toni Makkai, Valerie Braithwaite, Dianne Gibson, and I will document perverse ritualism of limited or no benefit to residents in nursing homes where extra Medicaid payments accrue to homes that perform better on performance indicators. An example is wheeling in residents who are sound asleep to be present at an activities program when the performance indicator is the number of residents present at activities programs.

The Rule of Law and the Separation of Powers

350

358

UNIVERSITY OF TORONTO LAW JOURNAL

circuml)tances where the press and the people stand ready to condemn both sides if they do not manifest a more public-regarding approach to the dilemma. To the extent that the democracy succeeds in being dialogic in this way in the President's office and in the media debate, the admitted economic inefficiency from the separation of powers is attenuated. Considering the problem from the opposite direction, cabinet government in British parliamentary systems more or less solves the mutual irresponsibility problem. Cabinet is of the parliament and effectively controls it (or hands over to a new cabinet if it cannot). If a tough budget is needed, cabinet does in practice bear responsibility at the next election if the electorate perceives it as failing to have done what was needed. But the cost is that the budget is the creation only of the parties represented in cabinet to the exclusion of other interests represented in the legislature. A less republican concentration of power than in the American case. Again one remedy to this concentration is a dialogic one. Citizen groups have their pre-budget submissions tabled not before cabinet but before parliament. Perhaps they should be allowed to address the parliament on them. Then members of parliament from all parties have an opportunity to discuss these submissions together with Treasury forecasts relevant to the budget before the budget is framed. An alternative during the early years of the Hawke-Keating government (1983-1996) in Australia was that pre-budget submissions would be discussed by the Economic Planning Advisory Council. As a member of this council representing community and consumer groups, my job was to argue with the Prime Minister, Treasurer, and other members of the Cabinet and bureaucracy about the pre-budget submissions from that consumer-community group perspective. More importantly perhaps, the Council included state Premiers from parties other than that represented in Federal cabinet. In the private dialogue of the Council meeting, the Premiers were not performing for the media, not seeking to score points against the other party. Rather they seemed to seize a genuine opportunity to listen to the thinking of the Prime Minister and to seek to sway it through public-regarding reasoning. Either way, the checking of power that seems most valuable is dialogic. Adversarial politics in the glare of the television cameras is good for getting the issues before the people. Efficient economic resolution of those issues depends, however, on the quality of the economic dialogue in a democracy. Be the concern about economically inefficient outcomes articulated in an American or an Australian political context, in the context of framing budgets or implementing regulation, republicans are on the economic high ground because of their commitment to institu-

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

351

359

tionalizing dialogue that can transcend the deepest inefficiencies. These are, at root, the outcome of self-serving adversarialism. Bruce Chapman has raised a quite different economists' worry about stability. This is the 'danger in having an overlapping separation of powers with many sources of decision-making for the stability of any final organizational choice? One can certainly imagine that the plurality of interests might lead to a range of different and decisive coalitions (a bit like the voting paradox) cycling over a rich but unstable array of possible choices.' 142 The first reply is that the deeper worry is impoverished but stable choice - the economic choices of Breznev's Soviet Union or Marcos's Philippines. The fundamental protection against pathological policy instability for the republican is constitutional conventions that guarantee all the separated powers a legitimate voice in the deliberations, followed by a constitutional method for settling the matter until such time as new circumstances can be shown to prevail. Charles Sabel 143 sees 'the central dilemma' of economic growth as 'reconciling the demands of learning with the demands of monitoring.' The dilemma is that economic actors need to trust each other to learn by sharing know-how. At the same time, they must distrust each other by monitoring that the gains from the shared know-how are shared in the agreed way. Sabel views economic success as flowing from the design of discursive institutions that make discussion of know-how inextricable from discussion of apportioning gains or losses. Mutual dependence can resolve the paralysing fear of deceit by allowing both scrutiny and learning to be natural consequences of a joint enterprise. A fusion of identities (as we see in families even moreso than in work groups) means that untrustworthiness by one member causes other members to share the shame of the breach of trust while all share the joy of others' learning. Both learning and monitoring are products of discursive problemsolving and partial fusing of identities in joint ventures. 144 As in the public sector, so within the private sector, therefore, the inefficiencies from separations of power that risk uncoupling the demands of learning from the demands of monitoring are best resolved by the mutuality and deliberation commended by the republican ideal of governance.

142 Editor's comments on an earlier version of this paper. 143 C. Sabel, 'Learning by Monitoring: The Institutions of Economic Development,' The Handbook of Economic Sociology, eds. NJ. Smelser & R. Swedburg, (Princeton: Princeton University Press, 1986) at 137. 144 See also W. Powell, 'Trust-Based Forms of Governance,' Trust in Organizations: Frontiers of Theory and Research, eds. R.M. Kramer & T.R. Tyler (Beverly Hills: Sage, 1996).

The Rule of Law and the Separation of Powers

352

360

UNIVERSITY OF TORONTO LAW JOURNAL

IX Conclusion

A standard rationale for the separation of powers is deterring abuse of power with countervailing power. Deterring abuse of power, be it private or public, is not something we are good at. Problems like police corruption, dumping hazardous wastes and corporate fraud seem to bounce back after each wave of scandal and reform. 145 An increasingly coherent theoretical and empirical literature can now make sense of why deterring the abuse of power so often backfires. Emotive defiance, cognitions of stigma and procedural injustice, psychological reactance, the deterrence trap, and rational countermeasures are among the reasons big sticks often rebound. We have seen that all of these mechanisms apply to powerful actors; several of them have more force with powerful than powerless actors. Reactance and rational countermeasures are greatest with the powers the powerful care most about, which generally means the commanding heights of their power. It has been argued that solutions to these problems are to 1. replace narrow, formal and strongly punitive responsibility with broad, informal, weak sanctions; 2. separate enforcement targeting from identification of the actor who benefits from the abuse. Together 1 and 2 imply: (a) strong separations of powers within and between both the public and private sectors, combined with (b) another republican regulative ideal - problem-solving dialogue. The richer and more plural the separation of powers, the more over-determined will be the capacity to detect and prevent abuse of power. The more actors there are with this preventive capability, the more likely some of them will be soft targets who can be persuaded to check abuse of power by simple and cheap discursive appeals to their virtue. The more institutions for the control of abuse of power are based on moral reasoning rather than deterrence, the more public-regarding actors with preventive capability will there be. Deterrence is certainly needed when dialogue fails to control abuse of power, as is incapacitation when deterrence fails. 146 But the more we can succeed in keeping deterrence and incapacitation in the background, the better the prospect that the separation of powers will check abuse of power through moral suasion and the better the chance that it will do so in a way that enhances rather than hampers economic efficiency. 145 See L. Sherman, Scandal and Reform (Berkeley: University of California Press, 1978). 146 I. Ayres &J. Braithwaite, ResjJonsive, supra note 87.

The Rule of Law and the Separation of Powers ON SPEAKING SOFTLY AND CARRYING STICKS

353

361

An interesting implication of this for republican political theory is that the separation of powers and dialogic appeals to the virtue of citizens are not just separate republican ideals. The separation of powers creates a world where dialogue can displace sanctioning as the dominant means of regulating abuse of power. These republican prescriptions are not only coherent in the sense that both the separation of powers and dialogic reconstitution of interests help secure freedom as non-domination. Deliberative democracy is also causally dependent on the separation of powers. We have shown that if appeals to the virtue of soft strategic targets is to work, the form this separation of powers must take is much more plural than the traditional separation of legislature, executive, and judiciary. The more richly plural the separations of public and private powers, the more the dependence of each fiduciary on many other fiduciaries will secure their independence from domination by any one of them. This theory of republicanism amounts to a rejection of the radical Jeffersonianism of strict separation of powers that became influential in the early nineteenth century, that is represented in the French Constitution of 1795. 147 Simply dividing power and making it directly accountable to the electorate, preventing judges from meddling in the affairs of the legislature and vice versa, was a romantic theory even then, one that was bound to give birth to adversarial struggles for control that would deliver a Napoleon Bonaparte. The romantic theory of this century has been that antitrust law could democratize the new private power. 148 A pragmatic republicanism for the burgeoning private power of the twentyfirst century will give more emphasis to the checking of power part of the republican ideal; it will pluralize the separations of powers, while rejecting any aspiration that each divided power be fully independent. Many semi-autonomous powers recursively checking one another 149 rather than a few autonomous branches of governance. This means rejecting the status quo of the separation of powers, rejecting radical Jeffersonianism, and creatively radicalizing Madison for a world where new and disturbing concentrations of private power continually emerge. 147 See note 109. 148 Maximizing the breaking up of private power through antitrust law that creates inefficiently small firms is politically unsustainable in a world of intense international competition. Even if one could do it, why would one want to? In some senses, it is easier for state and civil society to demand the kinds of separations of powers and dialogic justice discussed in this article from one profitable large firm than from a dozen small struggling firms. A conception of the separation of powers as simply dividing or breaking up concentrations of power is neither attractive nor realistic in the contemporary world. 149 See Diagram 4.

[14] A REVISIONIST VIEW OF THE SEPARATION OF POWERS Geoffrey Brennan and Alan Hamlin ABSTRACT

The doctrine of the separation of powers attracts almost universal support as a central element of the liberal constitution designed to protect citizens against governmental power. However, there is little agreement on, or analysis of, the J?recise institutional requirements of the doctrine or the method by which the claimed benefit is achieved. We set out a simple model of the interaction between citizen-voters, the legislature and the executive to illustrate that the functional division of powers can operate systematically against the interests of citizen-voters. This case provides the basis both for a taxonomy of distinct senses of the separation of powers, and for the revisionist claim that there is a general liberal presumption against the functional separation of powers. KEY WORDS • constitution • separation of powers

I. Introduction Within the liberal constitutionalist tradition, the central problem of politics is to design institutions that, as far as possible, ensure that political power is exercised in the interests of citizens. The liberal element in this tradition refers to the centrality of citizens' interests as the normative test for government action; the constitutionalist element refers to the choice of institutions (rather than the attributes of kings, say) as the primary domain of normative concern. In this tradition, a characteristic feature of government is the possession of a monopoly in the legitimate exercise of coercive power. Such a monopoly is required both because citizens need protection from each other, and because citizens desire the provision of certain public goods that, acting individually, they will not provide in adequate quantities. A major problem for citizens, however, is that the assignment of coercive power renders them vulnerable to exploitation by their political agents. Accordingly, a major constitutional question is how to design institutions that limit such exploitation. One piece of institutional fabric that enjoys almost universal support in this context is the separation of powers. That is, there is a widespread belief, This paper was prepared while Hamlin was a visitor at the Research School of Social Sciences, ANU. Hamlin's research is also supported by ESRC research grant R000233782. We are grateful to two anonymous referees for comments on an earlier draft.

The Rule of Law and the Separation of Powers

356

346

GEOFFREY BRENNAN AND ALAN HAMLIN

both among political theorists and in democratic folklore, that the separation of assigned political powers somehow provides a check against exploitation and thereby promotes the interests of citizens. However, what exactly the separation of powers involves in terms of institutional arrangements, and how those institutional arrangements work to curtail exploitation are remarkably under-analysed questions. 1 One conceptualizing of this problem is as an instance of the principalagent problem studied by economists. 2 In problems of this type, a principal (in this case the citizenry) seeks to secure some desired end, but must do so through the mediation of agents (in this case politicians and bureaucrats). The agents are better placed than the principal to secure the desired end, but possess ends of their own - ends that are at variance with those of the principal. The principal's problem is to structure the institutional environment within which the agents operate so as to induce agents to behave more rather than less as the principal would wish. Given this conceptualization, the relevant question is whether the separation of powers contributes to the resolution of the principal-agent problem in favour of the principal. 3 That question is the central concern of this paper. Our aim is to investigate whether the widespread confidence in the separation of powers is justified. We want to know whether (or when) the separation of powers does, in fact, act as an effective constraint on the abuse of power, and if so, how and why. The answer to this question offered in the present paper is that everything depends on exactly how the cake is cut - specifically, on whether the separation is along 'vertical' or 'horizontal' lines. Our argument is that, while horizontal separations of power will often generate benefits to the citizenry, · vertical separations of power will not - and indeed may be expected to act against the interests of citizens. We take this conclusion to be particularly significant because the functional separation of powers as between the legislature, the executive and the judiciary is essentially vertical in its general 1. See, for example, Sabine (1973), Vile (1967), for accounts of the development of the doctrine. Rereading the traditional references to Federalist Paper 47 and The Spirit of the Laws certainly leaves open questions of the precise institutional requirement of the doctrine and the precise nature of the benefit derived from the separation of powers. It is also clear that one of the themes of the discussion of the doctrine of the separation of powers has been the fluidity of the theory and the variety of the practical arrangements that claim to embody that theory. 2. For detailed discussion of the principal-agent model, see Grossman and Hart (1983). 3. The liberal-constitutionalist version of the principal-agent problem may seem so central to the constitutionalist project that it exhausts the entire terrain. Not so. The problem of majoritarian cycling (that is, the failure of majority rule to generate an equilibrium) is logically distinct from the principal-agent problem, and the capacity of political institutions to suppress cycling in no way depends upon the resolution of the principal-agent problem. Thus, it might be that the separation of powers could play a part in the suppression of majoritarian cycles even if it plays no part in the resolution of the principal-agent problem - and indeed we shall suggest that this is the case with one variant on the theme of the separation of powers.

The Rule of Law and the Separation of Powers SEPARATION OF POWERS

357 347

structure. It is in this sense that our argument points to a revision of the traditional doctrine. At the very least, we believe that the analysis throws the burden of proof onto those who would claim that the functional separation of powers protects citizens from governmental exploitation. Our argument proceeds as follows. Section II analyses the separation of powers by appeal to an analogy with monopoly markets, exploiting the standard monopoly model of the microeconomics textbook. That analogy helps to establish what we see as the critical distinction between horizontal and vertical separations of powers. The reasoning in Section II represents the analytical core of the paper, and the results established there are, we believe, unexceptionable. The real work to be done in this paper is to show that the monopoly analogy is indeed an appropriate one. Accordingly, in Section III, we develop a model of government behaviour which retains the general structure of the monopoly model. This model specifically incorporates the constraining influence of electoral competition and enables the effects of the functional separation of powers between the legislature and the executive to be examined explicitly. Formally, this model is of a game with three types of players- citizen-voters, political parties and bureaucrats - and the central question is that of how citizen-voters fare under two contrasting institutional regimes: one in which the legislature and the executive are separated; and another in which they are integrated. The argument of Section III seeks to focus on the mechanisms at work when powers are functionally separated: the reasoning is abstracted from particular institutional versions of the separation doctrine as they arise in specific political systems (such as the US presidential/congressional system). Those specific institutional versions of the separation doctrine would clearly require correspondingly specific and detailed modelling. But, in our view, the basic ideas and arguments which we identify will be at work in all cases where functional separation as such is seen as a mechanism for constraining the exercise of power. In Section IV, we explore the robustness of the conclusions drawn from the model of Section III (and from the monopoly analogy) in the face of a variety of potential criticisms, and argue that the basic result is remarkably general. Section V then builds on the results to offer a novel taxonomy of the various senses of the separation of powers seen as ideal types and a sketch of the applicability of the analysis to specific political structures. Section VI offers some concluding remarks. II. The Monopoly Analogue

The intuition behind the separation of powers doctrine draws on the analogy with economic competition: separating monopoly powers and setting those

358

The Rule of Law and the Separation of Powers 348

GEOFFREY BRENNAN AND ALAN HAMLIN

powers in conflict with each other may act in the interests of consumers. The aim of this section is to explore this analogy in some detail. Begin with the standard economic model of monopoly. Specifically, consider a monopolist supplying some private good to a community of atomistic consumers. Suppose, for simplicity, that the good costs nothing to produce. We may think, following Cournot, of a mineral spring that is the sole source of mineral water. Such a monopolist will restrict output below the efficient level in order to drive prices and profits up and such action results in the loss of consumer welfare. These welfare losses represent potential gains from trade that are lost through the monopolist's action. These potential gains represent the focus of the economist's case for competition. But here we shall examine the effects of a single new entrant - that is, we shall separate the monopoly power into two. What we want to emphasize right from the outset is that there are at least two ways in which such a power may be divided - corresponding to horizontal and vertical separation which have very different implications for consumer welfare. The first case is that of horizontal or competitive separation. Here, the new entrant offers the product for sale just as the incumbent firm dries each firm might have access to the spring for 12 hours per day, for example. To simplify, we shall assume that each firm takes the other's output decision as given and rtxed in deciding on its own output. The result is the standard Cournot duopoly model, depicted in terms of reaction curves in Figure 1. The reaction curve, RA, for example, shows the output level that rll'IIl A would choose for each (expected) level of output chosen by firm B. For example, if B produces zero output, A will produce the monopoly output XM; but if B produces the monopoly output XM, A will maximize subject to this constraint and, in the case where the aggregate demand curve is linear, this will result in A's output being t XM. Similarly, in the linear case, if B produces the competitive output Xc, A will produce nothing. In short, if the aggregate demand curve is linear, RA will also be linear with vertical intercept XM and horizontal intercept Xc. And the reaction curve for B will be symmetric as shown in Figure 1. The unique Na,sh equilibrium now occurs at point E 0 , where the reaction curves intersect, since at that point each firm is maximizing profit given the output choice of the other. Since E0 lies outside and above the line XMXM we know that total industry output at E 0 must exceed the monopoly output so that the market price, industry profit and welfare loss must necessarily be reduced relative to the monopoly case. The separation of monopoly power, in this instance, has clearly worked to reduce that power and its exploitation and to benefit consumers. 4 4. Of course, the two firms may collude in an attempt to increase their profits - but, if they are successful, they can at worst take us back to the simple monopoly outcome. If there are ilXed costs of entry or some other economies of scale the two-producer cartel will be less efficient than the single monopolist, but we abstract from such considerations here.

The Rule of Law and the Separation of Powers

359

SEPARATION OF POWERS

349

A's Output

0

Xc

B's Output

Figure 1. The Standard Cournot Duopoly Model of Output Decisions.

This happy result depends crucially, however, on the particular way in which the monopoly power is divided. The second possible method of division that we will consider involves a functional or vertical division of the monopoly power. Suppose, in our natural spring example, that we allow one flrm to bottle the water and a second flrm to deliver it, both of these functions to be performed at zero cost, as before. But now each flrm, as an independent price setter, will charge a positive price in order to secure profit and the question is what these prices will be. Once again we may derive a reaction curve for each firm, showing, in this case, the price that each will set given (its expectation of) the price set by the other; the price to the final consumer will then be the sum of the two prices. Hence, in Figure 2, RA shows the price A will charge for each price charged by B. If B sets a zero price, A will set the full monopoly price PM. If B sets a positive price, then this price will be a marginal cost to A, and A will respond with the appropriate monopoly price. There is a reaction curve for B that is symmetrically derived, and the Nash equilibrium is again identified as the intersection of the two reaction curves at EF. And, since EF lies outside and above the line PMPM, the total price must be greater than the monopoly price PM. Thus, in equilibrium, output is lower than in the simple monopoly case and the welfare losses correspondingly larger, even though industry profits are necessarily reduced. In this second case, then, we have again created two firms out of the

The Rule of Law and the Separation of Powers

360

3SO

GEOFFREY BRENNAN AND ALAN HAMLIN

A's Price

0 B's Price

Figure 1. Price Setting under Vertically Separated Monopoly Power.

original one, we have again arranged things so that the interests of the two firms are in conflict, and again industry profits are reduced. However, in this case the separation of the monopoly power has worked to push up the market price of mineral water and increase the welfare losses. In a very real sense the separation of power has acted against the interests of consumers. s Of course, it is possible to think of relationships between the two vertically separated firms other than the independent price adjustment relationship analysed above. Two possibilities are worthy of explicit mention: two-part payments and bargaining. Under a two-part payment system, the upstream firm (the bottling plant in our example) could undertake to supply whatever quantity the downstream firm (the delivery firm) demands in exchange for payment of a price equal to true marginal cost plus a lump-sum payment. The lump-sum payment would then allow the upstream firm to capture at least some of the profit without distorting the marginal costs faced by the downstream firm. Indeed, in this simple case, if the upstream firm acts solely in its own interests, it will capture all of the profit associated with the industry and leave the downstream company indifferent between operating and closing down. In this way the mechanism of two-part payments will reach exactly the same equilibrium as the simple monopoly model. S. This case of 'successive monopoly' is discussed in the industrial context in Bresnahan and Reiss (198S); see also Tirole (1988, Ch. 4.2).

The Rule of Law and the Separation of Powers

SEPARATION OF POWERS

361

351

The bargaining approach to the relationship between the two firms would model the firms as engaged in bargaining over the prices and quantities to be determined. The details of such a model depend on the detailed structure of bargaining assumed, but it is clear that fully rational bargainers will seek to maximize the surplus received by the industry and then bargain over the division of that surplus between the firms. In this case, the bargaining model will also lead to the same outcome as the simple monopoly model with output and prices chosen to maximize total industry profits. In each of these two cases then, vertically separated monopolists may achieve the simple monopoly outcome rather than the inferior outcome associated with the independent price adjustment model. But two points should be emphasized. First, only under stringent assumptions will the outcome of two-part tariff or bargaining systems ensure the simple monopoly outcome - for example, the upstream firm would need perfect information about the downstream firm if it is to set the optimal lump-sum tariff, and bargaining would have to be perfectly efficient and costless with full information on both sides. Second, these mechanisms can never do better (in terms of consumer welfare) than the simple monopoly outcome. Indeed, whatever the actual relationship between the two profit maximizing firms in a vertical relationship of the type described, the outcome will generally be worse (and never be better) for the consumer than the outcome under a simple monopoly. In short, the vertical separation of the industry imposes costs by creating an externality between the two firms that, in general, will not be fully internalized. There are several morals that we want to draw from this simple tale of two separations. One is that, even in the familiar monopoly case, it is not self-evident that separating monopoly power necessarily reduces that power or its undesirable effects: whether division acts in the interests of consumers is a contingent matter depending, inter alia, on precisely how the division is effected. A second moral is to emphasize that there is no one-to-one relationship between limiting the rents enjoyed by the powerful and enhancing the lot of those who are subject to the power. In our mineral water case both separations of power can be expected to reduce profits - but only one, the horizontal separation, can be expected to benefit consumers: the vertical separation will, in general, harm consumers. And it is the provision of benefits to citizen-consumers rather than the reduction of rents accruing to the powerful per se that is the central normative idea in the liberal constitutionalist programme. A third and important moral for us involves the thought that if we are to cast about in economic analogies for a case similar to the classic doctrine of the separation of powers as it is understood in the political constitutional context then it is surely the case of the vertical, functional division that comes

362

The Rule of Law and the Separation of Powers 352

GEOFFREY BRENNAN AND ALAN HAMLIN

most readily to mind. The separation of powers between the legislature, the executive and the judiciary is, after all, functionally based. In order to produce policies it is necessary for each of these functions to be deployed sequentially - to determine policies, to have them implemented as intended, and to have them enforced appropriately. A governmental monopoly divided functionally in this way seems essentially similar to the functional division of the mineral spring monopoly into a bc?ltling frrm and a marketing firm. The failure of each firm to account fully for the costs that their independently maximizing decisions impose on the other firm leads to a potential loss of consumer welfare in the mineral water case and the same result is threatened in the case of the functional division of government: the separation of powers holds the prospect of exploiting citizens more fully rather than offering any protection from exploitative government. In what follows, we seek to explore this possibility more directly by dispensing with analogies and developing a model of the political environment that nevertheless retains the general analytic structure of the mineral springs case.

m.

The Model

Our purpose here, then, is to construct a simple but plausible model of the legislature-executive interaction so as to examine the properties of alternative institutional arrangements between these two. The model presented in this section is kept as simple as possible, consistent with this purpose. Arguments concerning the generality of the results derived from this simple model are deferred to the next section. We shall focus on the functional separation of powers and the particular instance of the separation of powers as between the legislature and the executive. The legislature is the body that retains the power to make the laws and determine policy: and we take it that any potential law or policy that fails to be presented to the legislature or fails to be ratified by a majority of the legislature's members (however constituted) will not be enacted. Although the legislature is the site of policy-making power, it does not exercise that power in a completely unconstrained manner. We shall model two potential constraints on the power of the legislature. The first derives from popular democratic elections contested by political parties; the second - which is the primary topic under consideration - relates to the interaction with the executive. The executive we shall understand to be the body that implements policy that actually supplies the outputs of public policy. Of course, we do not mean that the executive necessarily produces those outputs, but only that the executive is the channel though which outputs must flow. To be clear on this point, the executive in our model may be thought of as supplying the

The Rule of Law and the Separation of Powers

363 353

SEPARATION OF POWERS

legislature with a menu of choice - indicating the costs of the alternative policy options. The legislature is then charged with selecting from that menu and raising the finance to pay for the chosen policy. Our use of executive here excludes any direct role in the process of policy making. In many systems (including the US system), the executive also operates as a part of the policy making process- often as a steering committee of the legislature or as a legislative house in a multicameral system. But we wish to focus on the functional division of powers, and so exclude other possibilities (which will be taken up briefly below). In the context of the functional separation of powers, the classical separation thesis requires that the legislature and the executive operate independently, non-collusively, or at arm's length, in the sense that each pursues a course of action determined from its own perspective given the environment that surrounds it, an environment which explicitly includes the other body. The separation thesis amounts to the claim that such an arm's length relationship acts in the interests of citizens. It is this claim that we seek to investigate. For convenience, we shall first sketch out the model of each of these branches of government. We want to make these sketches as simple as possible so as to focus on matters we see as crucial but not so simple as to render the discussion implausible. Once these ingredients are assembled, we shall turn to the discussion of the form of the game in which the citizenvoters, the political parties and the executive are all players. (a) The Legislature

The legislature is elected; and, in the public choice tradition, we assume that voters in this election are rational in that they vote for outcomes that they expect to leave them better off. More specifically, our model of the legislature is one of two-party competition under majority rule. Parties are taken to be motivated by a desire to hold office - but this motivation is instrumentally derived. That is, parties do not seek office for its own sake, nor for the prestige associated with positions of power. Rather, parties seek the rents to be obtained from office-holding. Specifically, parties aim to maximize expected net surplus E(S;): E(S;) = P;S;

(1)

where P; is the probability that party i wins power S; is the net surplus that party i secures from victory. Parties compete by offering voters policy packages that include varying amounts of aggregate rent for citizens. For simplicity, we shall take it that these rents - and the surplus appropriated by the elected party - derive from the consumer surplus generated from the provision of some composite

The Rule of Law and the Separation of Powers

364 354

GEOFFREY BRENNAN AND ALAN HAMLIN

public good (G) and further that G is produced under conditions of constant average (and hence marginal) cost. We abstract from the concerns associated with the revelation of citizens' preferences for public expenditure and assume that the same aggregate demand curve is perceived by all political actors. Consider the role of interparty competition here. The probability of party A being elected is a function of the net surplus offered to voters in A's policy package, relative to the net voter surplus in party B's package. In the absence of party competition, a monopoly government would appropriate all the available surplus in political rents, offering no voter surplus. That is, the monopoly government would provide G at the optimal level but capture the full consumer surplus. But party competition constrains the appropriation of surplus. Each party bids for votes by offering to voters a policy package (we assume that the packages are announced simultaneously) with a particular net surplus for voters. Note that each party will undertake to provide G at the optimal level because, for any given level of political surplus for party A, (SA), A's probability of being elected is increasing in voters' rent (and similarly for B). Thus, the surplus that each party seeks to appropriate in the political arena can be seen as a simple net reduction in the total consumer surplus available to citizen-voters. So, party competition will deliver optimality in the level of the public expenditure, but, as we have already noted, that aspect of policy choice is also available even under monopoly government. Interest must focus, therefore, on the distribution of the available surplus between the citizen-voters and the legislature in equilibrium. It is important to note that, in our model, party competition will not necessarily drive political rents to zero. (The significance of this point will be discussed further in Section IV.) In a standard one-dimensional median voter model, competition between two parties will drive political rents to zero in equilibrium. In that model, any party that attempts to charge voters more than the marginal cost of supplying the public good will lose the election; A's probability of victory goes immediately to zero as soon as A departs from the price equal to marginal cost strategy. This is an example of the standard Bertrand-Hotelling analysis in which the payoffs to the parties are sharply discontinuous. Games of this sort may· be thought of in terms of the following simple example. Imagine two players competing for a prize of variable size, each player has to announce a number (simultaneously) and the lower number wins. But the winning number also directly determines the size of the prize; we may think of the prize as the relevant number of dollars (if both players announce the same number, they share the prize). The unique Nash equilibrium of this game is that both players choose the number zero, so that the equilibrium size of the prize is also zero. This outcome corresponds to the zero profit result in the standard Bertrand model of duopoly, the Hotelling result of zero surplus in spatial

The Rule of Law and the Separation of Powers SEPARATION OF POWERS

365

355

competition, and the zero surplus result in the simple median voter model. 6 The sharp discontinuity in the payoffs to political parties is, however, extremely implausible. So, in our model, when parties make their simultaneous bids, the fact that A's package involves less aggregate voter surplus than B's is not sufficient to drive the probability of a victory for A immediately to zero. More specifically, the probability of A's victory is an increasing function of the net level of surplus offered by A relative to that offered by B, and that function takes on a value oft when SA= SB, and decreases continuously with (SA- SB). To summarize this discussion, we keep the model as simple as possible whilst retaining the central feature that electoral competition imperfectly constrains government's ability to extract surplus. We therefore specify the probability of electoral victory functions: P1

=f(S1 - Si)

for i,j =A, B; i :l=j

(2)

and impose only the following conditions:

/(0) =

1

2' /'(.) < 0, /'(0)

=-a

(3)

An illustrative probability of electoral victory function is sketched in

Figure 3. Note that a, the gradient of the function at the point where SA = SB (so that the two parties offer the same platform), is not specified in detail. We shall return to this below. The simultaneous maximization of (1) by A and B yields the first-order conditions:

/(S1 - Si)

+ S;/1 (S;- Si)

= 0 for i,j =A, B; i :1= j

(4)

Or, in the standard elasticity formulation: -f(S1 - Si) S; = / 1 ( S; - Si)

(5)

In equilibrium, the two parties will make the same bid, so that the equilibrium bid, S*, is given by:

*

-/(0)

1

S = f 1(0) = 2a

(6)

This indicates that, in equilibrium, the surplus extracted by government is 6. For general discussion of such discontinuous games, see Dasgupta and Maskin (1986). For an example of a discontinuous game formulation of a competition between rent-seeking candidates for political office, see Hillmann and Samet (1987). Their model also differs from ours in a number of other respects, the size of the prize is ri.Xed and the 'bids' made by rival candidates are sunk costs.

The Rule of Law and the Separation of Powers

366 356

GEOFFREY BRENNAN AND ALAN HAMLIN

0

Figure 3. An Illustrative Probability of Electoral Victory Function.

a function only of a, the sensitivity of the probability of victory function in respect of small differences between the platforms of the two parties. If a is large, the constraint imposed by electoral competition is tight and equilibrium surplus extraction is slight. In the limit as a ~ oo we approach the Bertrand-Hotelling case in which two-party competition enforces the fully competitive outcome with all surplus accruing to the electorate. If a is small, however, the constraint imposed by electoral competition is lax, and the government is able to extract considerable surplus. In the limit as a ~ 0 we approach the case in which the probability of victory is independent of the bids made so that the government acts as a pure monopolist. As the model stands, the equilibrium appears to indicate a fiXed cost of government, in that the surplus extracted by government depends only on a and not on the size of the government as measured by total public expenditure, so that increasing the scope of government does not increase the surplus extracted by government. However, there is nothing in the specification of the model that ensures that a is independent of the size of government. Different detailed specifications of the probability of victory function will yield different detailed results, but two general possibilities seem particularly salient here. In the first, a is indeed independent of other aspects of the model and the legislature does impose a fiXed cost on society that depends only on the severity of the electoral constraint. In the second case, we might argue that a does depend on the scale of government. For simplicity, this possibility is captured by the case where the legislature appropriates a fiXed share of the available social surplus (R), so that any expansion in the scale of government would imply an equiproportional increase in the rent captured by the legislature. For convenience we shall consider the case where

The Rule of Law and the Separation of Powers

SEPARATION OF POWERS

this fixed share is one half, so that where s• = l/2a.

367

3S7

s• =

R/2 alongside the fixed cost case

(b) The Executive

As we have already indicated, we may think of the executive as providing the legislature with a menu of priced policy alternatives. This might seem to give the executive two potential sources of power - the power to select the alternatives that will be considered - that is, the power of agenda control and the power of setting prices. However, we are concerned with functional separation of powers and so wish to exclude any direct involvement by the executive in the process of decision making. Thus, we shall model the case in which the executive has no agenda control power. It might be that the legislature decides first on the set of alternatives to be considered and then commissions the executive to produce more detailed information and costings on each alternative prior to making a final decision. In our case, with a single composite public good being the relevant policy question, details of this sort are finessed since the only effective decision facing the legislature concerns the quantity of the composite public good (and hence the tax system), and the only information to be provided by the executive concerns the price per unit of the composite public good. The executive branch of government is also taken to be bent on maximizing its own rents subject to the constraints facing it. Again there are two possible types of constraint - one provided by the mechanism by which the executive is appointed and continues in office, and one provided by the environment in which the executive operates and, in particular, by the interaction with the legislature. In this section, we consider only the second type of constraint in order to focus attention on the contrast between two cases - when the executive is independent of the legislature and when the executive is effectively integrated with the legislature. More general discussion of the first type type of constraint is presented in the next section. The constraint to be modelled here is essentially a demand curve - the executive is limited in its ability to pursue its own ambitions by the legislature's behaviour once prices have been announced. Constrained only by the relevant demand curve, the executive will maximize its own rent by charging a premium on all public goods supplied, just as a simple monopoly might set a price-cost margin so as to maximize its profit. Again, for simplicity we take it that the aggregate demand curve for the public good is linear. This case is illustrated in Figure 4. In this Figure MC indicates the true marginal cost of supplying the public good, while PM indicates the price charged by the monopoly executive. The quantity provided is XM and the executive obtains a rent indicated by the area PMEDB in Figure 4 which, in this linear case, is one half of the maximum available

The Rule of Law and the Separation of Powers

368 358

GEOFFREY BRENNAN AND ALAN HAMLIN Price

Demand Output

Figure 4. Linear Aggregate Demand Curve for the Public Good.

consumer surplus (R), indicated by the area of the triangle ABC. Because the process by which the executive captures its surplus is distortionary, the net surplus remaining for the citizen-voters is R/4, with the remainder lost as excess burden (triangle CDE). We now turn to the interaction between the legislature and the executive. (c) The Game between Citizens, Parties and the Executive The doctrine of the separation of powers suggests that in modelling the interaction between the legislature and the executive, each should be viewed as independent of the other. Furthermore, the doctrine holds that such independence will yield desirable outcomes. However, in the cases modelled here the result of such a separation of powers is far from beneficial. The game between citizen-voters, parties and the executive under such separation of powers is in three stages. First the executive announces a price at which it will supply the public good. Second, given this price, the two parties simultaneously announce their platforms. Third, the citizen voters vote to elect one of the parties.' This election then determines the final distribution of payoffs. 8 7. Notice that if the order of moves were reversed so that parties had to commit to quantities of the public good prior to knowing its cost, the parties could not offer credible policy packages since they could not also commit to tax rates or other details of financing. Equally, such a commitment would allow the executive to escape the constraint of the demand curve and raise prices still further. Only if costs are known (at least approximately) in advance, can the electoral mechanism be viewed as a plausible restriction on government. 8. We shall not consider the commitment problem that arises once the election is over and

The Rule of Law and the Separation of Powers SEPARATION OF POWERS

369 359

The first point to note is that the legislature's demand curve for the public good is identical to that of the citizen-voters, regardless of the outcome of the election, since, as we have seen, the legislature will always want to provide the optimal quantity of the public good. Given this demand curve, the executive's behaviour at stage one of the game will be to set price PM (Figure 4) and so capture a rent of R/2. But, given this, the total available surplus at the second stage of the game involving competition between parties for control of the legislature is R/ 4, and our earlier analysis of that competition indicates that this will be passed on to the citizen-voters either net of a fixed cost, or divided between the legislature and the citizen-voters, depending on which case is considered. The surplus enjoyed by citizen-voters will be either (R/4 - 1/2a) or R/8. In both of these cases, then, the independently operating legislature and executive act to reduce the surplus accruing to the citizen voters - which is the acid test of normative desirability here. To see this most clearly, consider the joint rent maximizing strategy for a legislature-executive cartel in the case where the legislature imposes a variable cost. Let the demand curve for the composite public good be: P =A- bX

(7)

and the constant marginal cost of production = 0. Then R, the total surplus available, will be A2/2b. Under these circumstances an independent executive would set a price of A/2. But the joint rent maximizing problem can be written as: Max PX

1

+ "4(A- P)X

subject to (7). And it is straightforward to check that this requires that the executive's price be set at P * = A/3. With the result that the legislature-executive cartel captures 2/3 of the available surplus rather than the 5/8 captured in the non-cooperative equilibrium. But it is important to note that cooperative deals of this type that increase the total rents extracted by government (legislature and executive combined) may also be in the interests of the citizen-voters. And this is the case in this example. Under the cartel arrangement the surplus accruing to the citizenvoters is 2R/9 compared to the R/8 under the non-cooperative equilibrium characteristic of the full separation of powers. In this case, at least, the cooperation between powers provides a Pareto improvement, made possible before policy is enacted. Although we consider the onP.-shot version of this game, we may think of this as being embedded in a repeated game in which reputation effects may be expected to operate to enforce promise-keeping by elected parties. In any case, the pursuit of that issue here would take us too far from our main objective (see Alesina, 1988).

The Rule of Law and the Separation of Powers

370

360

GEOFFREY BRENNAN AND ALAN HAMLIN

by the internalization of the externality between the legislature and the executive and the consequent reduction in excess burden. Of course, the citizen-voters may be made still better off in this example by completely integrating the legislature and the executive. There is some ambiguity in this term. What we have is a model of a legislature constrained by majority voting and an executive constrained by the demand curve for the composite public good; if we are to integrate these two, which constraint will be effective? ~f we integrate by placing the executive (as well as the legislature) under the electoral constraint (only), the analysis of subsection III(a) above stands as a model of the integrated government. If, on the other hand, we integrate by placing the legislature (as well as the executive) under the constraint of the demand curve (only) the analysis of subsection III(b) above stands as a model of the integrated government. The conclusions in the case where the legislature takes a iiXed share of the available surplus are summarized in Table 1 which indicates that while integration under the electoral constraint acts most strongly in the interests of citizens, integration of any sort is better than the alternatives of either complete separation or separation with collusion. Similar results apply in the case where the legislature takes a fixed rent. Table 1. Payoffs under Alternative Constitutions Payoff to Full separation

Legislature

Executive

R

R

8

Collusion

{2:}

2

Citizenry

Excess Burden

R

R

8

4

2R

R

9

9

0

Integration under electoral constraint

{~}

R

Integration under demand constraint

{~}

R

R

4

4

2

This simple model offers the same fundamental conclusion as·was derived from the consideration of the monopoly analogy in the previous section. The creation of an arm's length relationship between the legislature and the executive creates an externality and this externality points to both a reduction in the rents received by government and in the surplus received by citizen-voters. More particularly, the functional separation of the executive from the legislature acts against the interests of the citizens.

The Rule of Law and the Separation of Powers

SEPARATION OF POWERS

371 361

IV. Extensions and Generalizations We now turn to consider some possible criticisms of the argument presented so far. One major objection (offered by a referee) is that the analysis simply misrepresents the doctrine of the separation of powers - that, properly interpreted, the doctrine requires only the separate and independent selection of office holders in the executive and legislature, rather than the independent action of executive and legislature. We have responses at two levels. First, it is of course an important theme of our argument that the proponents of separation are vague in specifying the precise institutional features they have in mind. If our attempt to avoid that vagueness focuses debate on the real content of the separation doctrine, so much the better. Further, if independence of appointment procedures is taken to be basic, the mechanism by which these appointment procedures bear on the level and distribution of rents must be made explicit. For our part, we cannot see how appointment procedures as such could do any relevant work except through their impact (if any) on the actions of office-holders once appointed and, in particular, the extent of collusion between executive and legislature office-holders. In other words, we retain our view that a major theme in the classical debate on the separation of powers revolves around the effects of non-collusive behaviour - and that the way in which separation inhibits collusion, and the resultant effects on citizens' political surplus, remain critically relevant. A second possible line of criticism is that our particular results relating to the effects of suppressing collusion on citizens' surplus are an artifact of the particular model employed, or that the opposite results could be generated from an equally plausible model. In fact, we believe our basic result to be remarkably general. The essential point is that the functional or vertical separation of powers introduces an externality as between the separated bodies, and it is the creation of this externality - rather than its precise specification - that drives the result. Seen in this way, the purely functional separation of powers acts to prohibit the internalization of externalities. We distinguish between two general classes of criticisms of our simple model. The first class concerns the nature of the particular constraints that are effective in limiting the behaviour of either the executive or the legislature; the second class concerns the nature of the interaction between the executive and the legislature. We will consider each in turn. The model as presented places an electoral constraint on the legislature and a demand curve constraint on the executive. It is clear that other constraints could be incorporated. For example, the executive might itself be elected or appointed and monitored in some other way that effectively constrains its behaviour. Or the legislature might be subject to additional constitutional restrictions. However, we believe that the basic result of our-

The Rule of Law and the Separation of Powers

372 362

GEOFFREY BRENNAN AND ALAN HAMLIN

model would survive all such modifications of the detailed structure of the model. Our reasons for this belief are as follows. Any mechanism or set of mechanisms that act to constrain an agent (either the legislature or the executive) may work either perfectly or imperfectly. Perfect mechanisms operate so as to render all of the theoretically available surplus to the principal (here, the citizenry); imperfect mechanisms leave some of the surplus in the hands of the agent, or reduce the surplus flowing to the principal by dissipating some of the surplus, or both of these. If a principal has access to a perfect constraining mechanism then it is clear that she/he should seek to place all possible powers under that constraint. Separating powers and placing them outside the control of the constraining mechanism can do no conceivable good. Even if the perfect mechanism can be employed on both of the separated bodies separately, the result cannot be better than perfect. Clearly then, if a perfect constraining mechanism is available, the separation of powers can have no work to do. It is therefore in the case of imperfect mechanisms that we might expect the separation of powers to offer some advantage. But the case we model is precisely of this sort- and points to the opposite conclusion. It is in this sense that our modelling of electoral competition as an imperfectly constraining mechanism was significant; if we had adopted the extreme model of competition in which electoral competition constrains perfectly, there could have been no possibility of the separation of powers working to further protect the citizens. By making the electoral mechanism imperfect we gave the separation of powers its best chance. But even where there is only an imperfect constraint operative on each body, the separation of powers - and the associated introduction of an externality - causes these imperfections to operate sequentially, so that each builds cumulatively on the other to the disadvantage of the citizen. And all of this is so whatever the precise nature of the constraining mechanism. Take, for example, the possibility that the separated executive is elected by the population at large, and that there is competition between two candidates for office which broadly parallels the discussion of competition between political parties in our earlier model. Suppose, in particular, that the candidates compete by offering prices at which they will stand ready to supply the public good if elected. If the electoral process is the imperfect constraint that we model above, the result will be that the price will be reduced below the level found in our earlier model, but will still be set above marginal cost. The externality is smaller, but it is still there, and it will still be eliminated by the integration of the legislature and the executive. The details change but the central result remains the same: the functional separation of powers will often act against the interests of citizens and will never act to their advantage. This final clause is important from the constitutional perspective. On some accounts, the functional separation of

The Rule of Law and the Separation of Powers SEPARATION OF POWERS

373

363

powers will harm the interests of citizens; and on all other accounts of the constraining mechanisms, the functional separation is at best harmless: there is a clear constitutional argument against the functional separation of powers. The second class of criticism to be considered concerns the nature of the interaction between the executive and the legislature if these are separated. Again, we have modelled one particular (and we believe plausible) form of interaction, and it might be objected that our result hinges on this particular specification. But this is not so. Recall the discussion of the monopoly analogy in which different models of interaction between the upstream and downstream firms were considered. The point there was that while some models of interaction (such as independent price adjustment) lead to outcomes worse (from the point of view of consumers) than simple monopoly, no plausible forms of interaction would generate the opposite result: at best, we could reproduce the simple monopoly outcome. This point carries over directly to the model of political process. Although other forms of interaction between the legislature and the executive could be modelled including forms of bargaining, for example - we know in advance that we will never achieve an outcome that is better for the citizens than simple integrated government. On some accounts, the functional separation of powers will harm the interests of citizens; and on all other accounts of the interaction, the functional separation is at best harmless: there is, again, a clear constitutional argument against the functional separation of powers. Our claim then is remarkably strong. It is no less than that wherever two (or more) functionally distinct, self-seeking agents are each subject to control mechanisms by the common principal, the principal would not lose - and will benefit in many instances - from integrating the two agencies and subjecting the integrated agency to some subset of the controls previously employed. To put the same point more directly and in constitutional terms: if a constitution is to be designed to protect citizens from exploitation by government, it should not incorporate a functional separation of powers. 9 V. The Separation of Powers We are now in a position to bring together our discussion in the form of a more general account of the various senses of the separation of powers and the effects that we might associate with each sense. The major analytic 9. Note that we do not rely on any arguments concerning the costs of monitoring agents, but to the extent that it is generally less costly to monitor one agent than two, and less costly to operate fewer constraining mechanisms, such second order considerations would reinforce our basic conclusion.

The Rule of Law and the Separation of Powers

374

364

GEOFFREY BRENNAN AND ALAN HAMLIN

distinction to emerge from our discussion is the distinction between the horizontal and vertical separation of powers and we shall summarize our account of each in turn. The horizontal (or competitive) separation of power occurs whenever a single power is allocated to more than one agent. There are two major cases to consider. In the first a power is allocated to two or more bodies who operate side by side in the manner of rival firms. In the second a power is allocated between two or more bodies which compete for that power. In the first case the competitive element is clear and the source of the benefit to the citizen-consumer is similarly clear. (This is the first separation of power discussed in Section II). In the second case, the competition- and the benefit - are slightly less obvious since ex post only one body wields the power and so it may appear that the power is not separated at all. However, the ex ante competition for the power may act as some constraint on the exercise of that power ex post. An example relevant to the present context is the competition between political parties or candidates for government office. Ex post there is a single powerful office holder, but the ex ante competitive separation of potential power between the candidates may be expected to constrain the use of this power. This form of separation of power is exemplified in our model of electoral competition. We argue, with the prevailing orthodoxy, that the horizontal separation of powers generally acts in the interests of citizens. The hallmark of the horizontal separation of powers is that a single power or bundle of powers is allocated competitively or in a manner that. promotes competition. By contrast, the hallmark of the vertical separation of powers is that functionally distinct but related powers are unbundled and allocated to different individuals or groups of individuals. When powers are functionally separated between bodies each body has the power to act independently over at least some restricted domain. It is as if each body becomes a specialist in the use of some particular power and wields that power independently, even though it is necessary for all of the powers to be deployed in sequence if any particular outcome is to be achieved. We have argued, against the prevailing orthodoxy, that the vertical or functional separation of powers generally acts against the interests of citizens. A critical feature of the desirable, horizontal kind of separation of powers is that it provides an 'exit option' for citizens. Each citizen is faced with an alternative (either ex ante or ex post) and the potential transfer of custom reduces the monopoly power of the agents. The vertical or functional kind of separation of powers provides no such exit option and, as we have argued, simply adds an external cost to the system. Citizens cannot gain from vertical separation which provides no competition, and will generally suffer from the externality. While the distinction between horizontal and vertical aspects of the separa-

The Rule of Law and the Separation of Powers SEPARATION OF POWERS

375 365

tion of powers is our primary concern, we might also identify a further sense of the division of powers which we believe to be distinct from either horizontal or vertical separation - we refer to this as the decentralization of powers. The decentralization of power is exemplified by the shift from a single decision maker to a group of decision makers - from a king to a council as it might be. In such a move, the power is decentralized in the sense that no agent commands the power unilaterally, but the power of any one agent is not rendered subject to competition. While it is clear that such a decentralization of power reduces the power held by any single individual, it is by no means clear that power overall is constrained, or that those subject to the power can expect to benefit. To be ruled by a committee may be just as unpleasant as - and perhaps even more unpredictable than - to be ruled by an individual. Of course, the precise effects of any decentralization of powers will depend on the decision making rules employed within the council. One salient possibility here is that the ruling council is split into two parts each of which has a veto over the other - in short, a bicameral arrangement of powers. In this way, bicameralism might be seen as a model of one type of decentralization of power. But again there is no reason to suppose that bicameralism can be expected to alleviate the principal-agent problem. 10 Decentralization is essentially a horizontal move. It takes a power, or set of powers, and spreads that power more thinly, rather than separating power from power as is the case in a vertical or functional separation. We therefore envisage a taxonomy of the separation of powers as depicted in Figure 5 which indicates the cross-cutting nature of the two distinctions - between horizontal and vertical on the one hand and separation and decentralization on the other - with the proviso that the case of vertical decentralization may be an empty box. Each of the non-empty boxes in Figure 5 represents an ideal type. Of course, in practice, many particular institutional arrangements thought of as exemplifying the doctrine of the separation of powers will contain some mixture of these ideal types. Clearly, on our analysis of the ideal types, such 10. We have argued in another paper that bicameralism can act to suppress majoritarian cycling {Brennan and Hamlin, 1992) although we are now convinced that the conditions required for bicameralism to induce stability are still more restrictive - see Tsebelis (1993). Note that it is particularly clear in this case that the benefits of the suppression of cycling are not uniquely associated with either the separation or decentralization of powers. If we were to start with a single council we might eliminate cycling by either separating the council into two houses or by reducing the council to a single member. Thus, while bicameralism may have some virtue in respect of the suppression of cycling we find no general reason for arguing that the decentralization of power can be expected to act in the interests of citizens. A further example of decentralization is provided by the movement from individual political candidates to political parties. In another paper (Brennan and Hamlin, 1993) we suggest that this move may be significant in improving the constraining mechanism of electoral competition.

The Rule of Law and the Separation of Powers

376

366

GEOFFREY BRENNAN AND ALAN HAMLIN

Horizontal

Vertical

Separate

Decentralize

Figure 5. Taxonomy of the Separation of Powers.

examples might involve both constitutional benefits and constitutional costs, and the balance of such costs and benefits is an open question in each case. The discussion in this paper is not directed at any particular set of institutional arrangements, or at determining the balance of costs and benefits in any particular case, but at the fundamental logic of the separation of powers. Nevertheless, we would suggest that our analysis provides an appropriate starting point for a re-examination of particular institutional arrangements. A wide variety of institutional arrangements and structures are often treated as falling under the general rubric of the doctrine of the separation of powers. Alongside the functional division of the legislature and the executive are grouped the independence of the judiciary, federalism, bicameralism and a variety of other arrangements that are all claimed to share the spirit of the separation doctrine. The taxonomy of the senses of the separation of powers offered here will help, we believe, in establishing which particular institutional arrangements really do embody the spirit of the doctrine. The simple but fundamental distinction between vertical and horizontal separations of power provides us with a useful device in the analysis of alternative institutional structures. But even with this device, analysis cannot be expected to be simple. Take the case of federalism as an example. Much of the work on federalism as a constitutional restraint on the power of government focuses on the competitive relationship between state governments. 11 It is reasonably clear that this aspect of federalism reflects a horizontal separation of powers and that this underpins the presumption that such competition between states may be expected to be welfare enhancing. 12 11. Mobility of resources is usually the key to such competition, following Tiebout (1956); discussion of federalism as a constitutional constraint can be found in Brennan and Buchanan (1980), Hamlin (1985, 1991). 12. Note that this is only a presumption; there will be cases where even horizontal competition is welfare reducing; most obviously when strong externalities arise as between states.

The Rule of Law and the Separation of Powers

SEPARATION OF POWERS

377 367

But federalism also incorporates a vertical separation of powers as between the federal and the state governments. This is illustrated by consideration of the case in which both levels of government have access to the same generalized tax base. With each government attempting to raise revenues, but ignoring the impact of its taxes on the revenue raising of the other government, it is clear that each government will impose an externality on the other so that cooperation between the levels of government - or integration into a unitary government - could lead to increased tax revenues and increased citizen welfare. (This aspect of federalism is emphasized by Flowers, 1988.) This vertical separation of powers underpins a presumption against federalism. A key issue in the constitutional analysis of federalism then is the balancing of these two aspects - the horizontal and the vertical - in the various practical institutionalizations of federalism. The general point here is that while the insight gained from our analysis of ideal types is transferable to a variety of areas of practical constitutional debate, the transfer requires some sensitivity to the differences between constitutional provisions that are sometimes too crudely treated together. VI. Concluding Remarks We have argued for two propositions: one general and one specific. The general proposition is that the notion of the separation of powers is open to many interpretations and that these various senses of the separation of powers, when analysed, turn out to be very different in their effects. The traditional language which emphasizes the family likeness of all constitutional provisions which seek to divide, separate, decentralize or otherwise diffuse powers obscures the very real and analytically significant differences that exist. Of course, any particular set of institutional arrangements can be expected to include a mixture of these various elements, so that analysis of those particular institutions will require the combination of various analytic elements; but this in no way reduces the importance of decomposing the general notion of the separation of powers into its various senses, and providing detailed analysis of each of the identified senses. Our second and more specific proposition relates to a sharp analytic contrast between two particular senses of the separation of powers. While there is an analytic basis for a presumption in favour of the horizontal or competitive separation of powers as protection for citizens, the same analysis provides the basis for a presumption against the vertical or functional separation of powers. In that sense, ours is a revisionist view of the separation of powers understood in its classic, functional form. In our view, the onus of proof now lies with proponents of the classic separation doctrine.

The Rule of Law and the Separation of Powers

378 368

GEOFFREY BRENNAN AND ALAN HAMLIN

The questions of what is at stake in the separation of powers, and how it works to protect citizens' interests should, at the very least, be declared open.

REFERENCES Alesina, A. (1988) 'Credibility and Policy Convergence in a Two Party System with Rational Voters', American Economic Review 78: 796-80!5. Brennan, G. and J. M. Buchanan (1980) The Power to Tax. Cambridge: CUP. Brennan, G. and A. P. Hamlin (1992) 'Bicameralism and Majoritarian Equilibrium', Public Choice 74: 169-79. Brennan, G. and A. P. Hamlin (1993) 'Rationalizing Parliamentary Systems', Australian Journal of Political Science 28: 443-!57. Bresnahan, T. F. and P. C. Reiss (198!5) 'Dealer and Manufacturer Margins', Rand Journal of Economics 16: 2!53-68. Dasgupta, P. and E. Maskin (1986) 'The Existence of Equilibrium in Discontinuous Economic Games, 1: Theory and ll: Applications', Review of Economic Studies !53: 1-26, 27-41. Flowers, M. (1988) 'Shared Tax Sources in a Leviathan Model of Federalism', Public Finance Quarterly 16:67-77. Grossman, S. and 0. Hart (1983) 'An Analysis of the Principal-Agent Problem', Econometrica !51: 7-4!5. Hamlin, A. P. (198!5) 'The Political Economy of Constitutional Federalism', Public Choice 46: 187-9!5. Hamlin, A. P. (1991) 'Decentralisation, Competition and the Efficiency of Federalism', Economic Record 67: 193-204. Hillmann, A. L. and D. Samet (1987) 'Dissipation of Contestable Rents by a Small Number of Contenders', Public Choice !54: 63-82. Sabine, G. H. (1973) A History of Political Theory, 4th edn. London: Harrap. Tiebout, C. M. (19!56) 'A Pure Theory of Local Expenditures', Journal of Political Economy 64:416-24. Tirole, J. (1988) The Theory of Industrial Organisation. Cambridge, MA: MIT Press. Tsebelis (1993) 'The Core, the Uncovered Set and Conference Committees in Bicameral Legislatures', Mimeo, Hoover Institution. Vile, M.J.C. (1967) Constitutionalism and the Separation of Powers. Oxford: OUP.

GEOFFREY BRENNAN is currently Director of the Research School of Social Sciences, Australian National University. He is author with James Buchanan of The Reason of Rules (198!5) and The Power to Tax (1980) and with Loren Lomasky of Democracy and Decision (1993). ALAN HAMLIN is Reader in Political Economy at the University of Southampton. He is the author of Ethics, Economics and the State (1986) and editor (with Philip Pettit) of The Good Polity (1989).

Paper submitted 14 July 1992; accepted for publication 14 May 1993.

[15] Institutionalizing the Public Interest: The Defense of Deadlock and Beyond

M

ROBERT E. GOODIN Australian National

University

any bemoan divided government and the consequent deadlock of democracy. The logic of The Federalist arguably defends it, in ways heretofore unappreciated, by appeal to a least-commondenominator definition of the "public interest." That quasi- Federalist logic is explored, and alternative political structures are assessed against the public interest criterion to which it appeals. Another and more defensible notion of the public interest is introduced, and its very different political styles, institutions, and policies are adduced.

I

n words that resonate in our own day, James MacGregor Burns (1963, 3) bemoans "the deadlock of democracy." We have often been too late, and we have been too late with

too little. Whether we can master depression in peacetime

is still in doubt. ... Currently baffled by a sluggish economy,

we seem unable to promote long, sustained economic

growth. Negroes still do not share the basic rights of citizenship promised in the 14th and 15th Amendments. We have done almost nothing about ... a coordinated and vitalized transportation policy. Our social welfare measures are inadequate, especially in medical care. We cannot play

our full economic role abroad because of inhibiting forces in Congress.

He traces those ills to America's constitutional structure. This model was the product of the gifted men who gathered in Philadelphia over 175 years ago, and it deserves much of the admiration and veneration we have accorded it But this

is also the system of checks and balances and interlocked

gears of government that requires the consensus of many

groups and leaders before the nation can act; and it is the system that exacts the heavy price of delay and devitalization that I have noted (Burns 1963, 6). 1 There is too much checking and too little balancing, too much advising and too little consenting, for the effective conduct of public affairs. Institutional gridlock has become the order of the day. Divided government has been much more the rule than tbe exception throughout the postwar period. 2 Revisionists urge a more balanced perspective (Mayhew 1991; Jones 1994, 1995). They point out, quite properly, that periods of divided government are no less productive of important legislation or major treaties than periods of unified rule (Mayhew 1991; Fiorina Robert E Goodin is Professor of Philosophy, Research School of Social Sciences, Australian National University, Canberra ACT 0200, Australia The author is grateful for comments from Geoff Brennan, Morris Fiorina, Russell Hardin, Calvin Jillson, Arend Lijphart, Vincent Ostrom, Bingham Powell, David Soskice, Gordon Tullock, and seminars of the ANU Institutional Design Project and the LSE Government Department 1 See similarly Burns 1984; Sorenson 1984; Committee on the Constitutional System 1987; Robinson 1985, 127-224; Cutler 1988; Sundquist 1988-89, 1993; Cox and Kernelll991; Thurber 1991; Harris 1993 2 Earlier periods, particularly the first half of the twentieth century, were characterized by strikingly unified government in contrast (Fiorina 1992, chapter 2) "Unified" government here means that the same party controls the presidency and both houses of Congress Taking account of the extent to which the third branch of government-the Supreme Court-is controlled by appointees of the other party would reduce those periods of unified government still further

1992, chapter 6). They point out, quite properly, that the three-cornered bargaining game among the separate branches of government in the United States may not be so very different, in its formal properties or practical effects, from a three-cornered bargaining game among minority parties in a coalition government anywhere else in the world (Laver and Shepsle 1991; Fiorina 1992, chapter 7). True and important though those observations are, they nonetheless mask real differences in the content of the enactments and the quality of the responsibility emerging from American-style divided governments. True, divided government gets things done. But the need to compromise among multiple centers of power typically leads either to incoherent policy packages (induced by the drive toward political log-rolling) or to watereddown legislation (induced by the drive toward the common ground) 3 True, parliamentary parties may face a formally analogous task in hammering out the legislative agenda for a coalition government. But where there is a formal coalition, collective agency has been created, and all parties to it will be judged at least in part by its successes or failures. Where, as in the United States, there is merely coalition-like governing, there is no collective agency and no shared responsibility. Even if in formal terms there were a core of policy compromises which would be Pareto-superior from the point of view both of a Democratic president and a Republican Congress, in terms of the larger reelection game it may be more in the interests of each to eschew compromise and try to lay the blame for consequent policy failures on the other. My focus here is not, however, on the pragmatics of divided government, its causes, or its consequences. Instead, my concern is with matters of principle. I shall examine a principled defense of the deadlock of democracy, not exactly Madison's own, perhaps, but one that can be pieced together out of fragments from The Federalist and plausibly ascribed to the Founders as one among several theories at work in their multilayered Those phenomena, familiar enough from House-Senate conference committees (Longley and Oleszek 1989), inevitably characterize interbranch negotiations as well Occasionally, interbranch competition may lead to "stronger" legislation than either alone would have championed, as in the case of the Clean Air Act of 1970 or the Economic Recovery Tax Act of 1981; but where that is the result simply of partisan outbidding of political opponents, there is no reason to suppose that stronger is necessarily better (Fiorina 1992, 91)

3

380

The Rule of Law and the Separation of Powers

Institutionalizing the Public Interest defense of the separation of powers. 4 Despite the homage to historical precedents, this should be clearly understood as an exercise in analytical rather than historical political philosophy. My focus is firmly upon the arguments, in stripped-down, cleaned-up, stylized forms, rather than upon the rich complexities to be found in the complete corpus of any particular historical author or authors. The centerpiece of the analysis which I thus wrench from The Federalist is a particular definition of the "public interest," introduced in the next section. Alternative political structures are then assessed against that criterion. Finally, I advocate another and more defensible notion of the public interest, indicating some of the sorts of political institutions, practices, and public policies toward which such a revised notion might lead.

THE FEDERALIST DEFENSE OF DEADLOCK Making a Constitution and Defending It Political history is one thing, political philosophy another. Madison's ([1840] 1966) Notes of Debates in the Federal Convention of 1787 speaks clearly to the former. As is evident from those transcripts, the constitutional separation of powers in the United States, and the checks and balances among them, largely emerged from political horsetrading of a perfectly ordinary sort: between small states and large, slave states and free, farmers and manufacturers, creditors and debtors; between advocates of states' rights and of central authority, of a strong and of a weak executive; and so on. A vaguely principled appeal was occasionally made, both in Philadelphia and elsewhere (Kenyon 1966; Storing 1981), to Montesquieu's ([1748] 1949, chapters 9-12) precepts about the separation of powers and checks and balances being necessary to ensure liberty and avoid tyranny. More often than not, those appeals looked much more like prudential reassurances (in the case of the federalists, the opposite in the case of the antifederalists) that the proposed political structures would give people adequate power to protect their own interests. The concern seemed to be less with liberty in the abstract, as a matter of high principle, than with each person's or faction's own liberty, as a matter of sheer personal or factional interest. Or at least so it seems to modern eyes: As has already been noted, these historical authors themselves typically ran a much more complex (nuanced, perhaps just confused) set of arguments alongside one another. If Madison's Notes recount the U.S. Constitution's history, The Federalist constitutes its principled defense. There, too, appeals to Montesquieu's notions of liberty and tyranny recur, in the same apparently principledcum-prudential way. In The Federalist, though, various other arguments of a much more forthrightly principled nature also emerge. Here I shall focus upon one in particular, the kernel of which can be found at least in the interstices of The Federalist's text. The interpretation All references, including specific page citations, are to the Cooke edition of The Federalist (Hamilton, Madison and Jay [1787-88] 1961)

4

332

June 1996 and elaboration of this argument is in many ways my own: It is a construction, arguably a concoction; perhaps it is more aptly described as "federalish" rather than fully "federalist." Nevertheless, it is an interestingly different argument (and, as I shall show, wrong in important respects) deserving of attention in its own.right. The most significant difference is that the Montesquieu-style argument provides no real defense of deadlock to those to whom it needs defending. Saying that checks and balances promote liberty by precluding tyranny is at least in one sense merely to say that politics is a zero-sum power struggle, that any power given to the government is power (cum liberty) taken away from the people. Where the balance should be struck-how separated the powers of government should be and how strong the checks-depends on the relative importance attached to the government's ability to act relative to the public's freedom from government actions. Those who bemoan deadlock, thinking that government is too tightly hamstrung, will be unpersuaded by being told that deadlock is necessary to avoid tyranny and protect liberty. They see the same tradeoff, they just strike the balance differently. To say deadlock is justified in order to protect liberty and avoid tyranny is, in that context, merely to restate precisely what is at issue. It constitutes reiteration, not argument. In contrast, the argument I shall adduce in the interstices of The Federalist appeals to an independent principle in defense of deadlock. Instead of merely trying to persuade people to weigh competing considerations differently, it points to a neglected reason for thinking that, at least in certain circumstances, deadlock may be the right outcome. Although the reasons it gives are in the end bad reasons, they nonetheless are well worth exploring and exposing.

Defining the Public Interest The positive defense of deadlock builds on a particular analysis of what constitutes the public interest. This analysis styles the public interest as the "common good" or (in the phrase the Preamble to the Constitution borrows from Article III of the old Articles of Confederation) as the "general welfare." The Founders themselves had relatively little to say explicitly on that topic, either in their Philadelphia debates or in their Federalist pamphleteering. Like virtually all political writers of the period, the Founders tended to employ the phrase "public interest" with relatively little reflection. Assuming as they did that the term's meaning was relatively transparent, writers of this period tended to devote little care or attention to its precise elaboration. The closest the Founders come to defining the public interest is Madison's famous Federalist 10 (p. 57) definition of the contrasting notion of "a faction": "a number of citizens, whether amounting to a majority or minority of the whole, who are united and actuated by some common impulse of passion, or interest, adverse to the rights of other citizens, or to the permanent and aggregate interests of the community." There are many ways those words could be read. Against the Enlightenment background so evident in

The Rule of Law and the Separation of Powers

Vol. 90, No.2

American Political Science Review Philadelphia, however, there is only one really obvious reading (Riley 1986). "Faction" is clearly a reference to what French philosophes and theologians from Malbranche through Montesquieu ([1748]1949, hook 11) to Rousseau ([1762] 1973, hook 2, chapter 3) had been calling the "particular will" and what today we might more readily term "personal (or sectoral) interests,'' The contrasting notion of the "general welfare" is clearly a reference to what they had been calling the "general will," which today we might more readily term the "common good." 5 Seen in that light, the Founders' contrast is clearly between the disreputable pursuit of narrow factional interests, particular to the members of that faction, and the enlightened pursuit of broad public interests, general in form and held in common across all segments of the community. That, in effect, amounts to a "least-common-denominator" definition of the notion of the public interest. The public interest is equated, literally, with the common good, with that which is common among my good and yours. In Locke's (1689, I, sec. 92) perspicacious phrase, "the public good" is equated with "the good of every particular Member of that Society." 6 As I shall show, that is not the only (or the best) way of construing the public interest. Yet, the Founders are naturally drawn toward that conclusion by their way of anchoring the conceptual contrast in view. Fixating on the contrast between the general and the particular inevitably makes the general that which is not particular, that which is common to all. A least-common-denominator notion of the common good/public interest follows quite naturally from that way of framing the problematique. That said, it must be added, of course, that the Founders gestured at many meanings. Most were subtly different from one another, some blatantly in conflict, all in ways largely unnoticed by these writers themselves. Thus, in ascribing this view of the public interest to the Founders I mean merely to say that this was one view of it, among others, that they held. My interest being analytical rather than purely historical, I intend simply to set those historical issues to one side. My aim is to explore the way this particular strand of thought about the public interest may articulate with thought on the separation of powers and other institutional mechanisms for promoting it.

Determining and Defending a Separation of Powers I shall query the wisdom of that analysis of the public interest below. For now, let us accept it uncritically and 5

A parallel distinction, familiar from Roman law via the republican

revival (Gunn 1989) and Blackstone's Commentaries (1769, book 4), is

between "public" and "private" interests, rights and wrongs 6 Locke adds at the end of this sentence: "as far as by common Rules, it can be provided for," thus forshadowing my reconceptualization below Locke's "i e " at this point indicates a strictly analytical connection, in contrast to the merely causal claims ("what is good for the community is good for each of its members") that characterize other (e g, Jacobin) versions Likewise, by saying "every" rather than "all," Locke makes his a distributed version of the argument, in contrast to the aggregative utilitarian forms that came later On these other versions, see Gunn ( 1989)

381

proceed to demonstrate how the separation of powers and checks and balances might be justified as a contribution to the public interest thus construed. The root idea, which might be visited upon if not literally adduced from Federalist 51, is this. The Federalist' Model

1. Separate the powers of government among as many distinct branches or subbranches of government as possible; 7 2. make each branch or subbranch of government independent of (i.e., not under the control or power of) one another, insofar as possible; 8 3. make each branch or subbranch of government responsible to the entire public but partitioned in a different way, insofar as possible;" and 4. give each branch or subbranch of government an effective veto over government action, insofar as possihle.IO 7 "Subbranches" alludes to possibilities such as dividing the legislature into two chambers Insofar as the executive itself is not unitary (Easterbrook 1994; Lessig and Sunstein 1994), it may be logically possible to parse powers within it as well, although The Federalist adamently opposes "a plurality in the executive" (Hamilton, Federalist

70)

s "The necessary partition of power must be supplied, by so contriving the interior structure of the government, as that its several constitutent parts may, by their mutual relations, be the ficaos of keeping each other in their proper places In order to lay a due foundation for that separate and distinct exercise of the different powers of government, each department should have a will of its own; and consequently should be so constituted, that the members of each should have as little agency as possible in the appointment of the It is equally evident that the members of members of the others each department should be as little dependent as possible on those of the others, for the emoluments annexed to their offices" (Madison, Federalist 51, 347-48; see similarly Madison, Federalist 47 and 48, and Hamilton, Federalist 71 and 73) 9 The constitution aims "to guard one part of the society against the injustice of the other part by comprehending in the society so many separate descriptions of citizens, as will render an unjust combination of a majority of the whole, very improbable, if not impracticable [TJhe society itself will be broken into so many different parts, interests and classes of citizens, that the rights of individuals or of the minority, will be in little danger from interested combinations of the majority" (Madison, Federalist 51, 351) This proposition emerges particularly clearly in Madison's discussion of the logic a bicameral legislature apportioned on different bases: "To divide the legislature into different branches[,] and to render them by different modes of election, and different principles of action, as little connected with each other, as the nature of their common functions, and their common dependence on the society, will admit" ensures that "no law or resolution can be passed without the concurrence first of a majority of the people, and then of a majority of the states" (Madison, Federalist 51, 350 and FederalL