Engaged Philosophy: Essays in Honour of David Braybrooke

Engaged Philosophy:Essays in Honour ofDavid Braybrooke

Engaged Philosophy:Essays in Honour of DavidBraybrooke

Edited bySusan Sherwin and Peter Schotch

UNIVERSITY OF TORONTO PRESSToronto Buffalo London

c© University of Toronto Press Incorporated 2006Toronto Buffalo LondonPrinted in Canada

ISBN 0-8020-3890-5

n∞Printed on acid-free paper

Library and Archives Canada Cataloguing in Publication

Engaged philosophy: essays in honour of David Braybrookeedited by Susan Sherwin and Peter K. Schotch.

Includes bibliographical references and index.ISBN 0-8020-3890-5

1. Philosophy and social sciences. 2. Ethics. 3. Braybrooke, David. I.Braybrooke, David II. Schotch, Peter K. III. Sherwin, Susan, 1947–

B63.E54 2006 170 C2005-90614-2

University of Toronto Press acknowledges the financial assistance to itspublishing program of the Canada Council for the Arts and the OntarioArts Council.University of Toronto Press acknowledges the financial support for itspublishing activities of the Government of Canada through the BookPublishing Industry Development Program (BPIDP).

Contents

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . xi

1 Introduction:About David BraybrookeSusan Sherwin 31.1 David Braybrooke, the Personal Story . . . . . . . . . . 51.2 David Braybrooke, the Scholar . . . . . . . . . . . . . 111.3 Overview of Essays . . . . . . . . . . . . . . . . . . . 141.4 References . . . . . . . . . . . . . . . . . . . . . . . . 20

Part One: Practical Engagement 21

2 Teaching Class: Justice and Privatization in EducationNathan Brett 232.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . 232.2 Privatization and Costs . . . . . . . . . . . . . . . . . . 26

2.2.1 Efficiency . . . . . . . . . . . . . . . . . . . . . 272.2.2 Benefits . . . . . . . . . . . . . . . . . . . . . . 28

2.3 Two Normative Contrasts . . . . . . . . . . . . . . . . 312.3.1 Exclusion . . . . . . . . . . . . . . . . . . . . . 312.3.2 Partiality . . . . . . . . . . . . . . . . . . . . . 322.3.3 The Public-Private Divide in the ‘Real World’ . 34

2.4 Education, Equality, and Choice . . . . . . . . . . . . . 352.4.1 The Demands of Justice . . . . . . . . . . . . . 372.4.2 Free-Market Choices . . . . . . . . . . . . . . . 38

2.5 The Priority of Justice . . . . . . . . . . . . . . . . . . 402.6 Education as a Human Need . . . . . . . . . . . . . . . 422.7 Fair Opportunity . . . . . . . . . . . . . . . . . . . . . 452.8 Teaching Class . . . . . . . . . . . . . . . . . . . . . . 47

v

vi CONTENTS

3 Determining Health Care Needs After the Human GenomeProject: Reflections on Genetic Tests for Breast CancerSusan Sherwin 513.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . 52

3.2 The World after the Human Genome Project . . . . . . 55

3.3 The Case for Treating BRCA Testing as a Need . . . . 57

3.4 Some Complicating Features of BRCA Testing . . . . . 61

3.5 Towards a Relational Understanding of Needs . . . . . 65

3.6 BRCA Testing Revisited . . . . . . . . . . . . . . . . . 69

4 The Mutual Limitation of Needs as Bases of Moral Enti-tlements: A Solution to Braybrooke’s ProblemDuncan MacIntosh 774.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . 78

4.2 The Appeal of Needs . . . . . . . . . . . . . . . . . . . 78

4.3 Braybrooke’s Problem . . . . . . . . . . . . . . . . . . 80

4.4 Possible Solutions to Braybrooke’s Problem . . . . . . 83

4.4.1 Questioning the Empirical Assumptions of the Prob-lem . . . . . . . . . . . . . . . . . . . . . . . . 83

4.4.2 The Defeasibility of Needs . . . . . . . . . . . . 84

4.4.3 Is the Location of Needs in Individuals or Groups? 84

4.4.4 Do People Really Need Indefinitely Long Lives? 87

4.4.5 Needs to Life Extension Not Really More Press-ing Than So-Called Less Basic Needs . . . . . . 88

4.4.6 Meeting Basic Needs versus Meeting the Needsthe Meeting of Which Make Life Worth Living . 89

4.4.7 Meeting Other People’s Needs as Not Really Be-ing a Cost to the Meeting of Our Own NeedsGiven That We Are Needs Altruists . . . . . . . 89

4.4.8 Needs as Relative to Projects Approved by Com-munities; Re-Choosing Needs . . . . . . . . . . 90

4.4.9 The Conceptual Analysis of Needs; the Axiomat-ics of Needs . . . . . . . . . . . . . . . . . . . 92

vii

5 Canadians and Global Beneficence: Human Security Re-visitedEdna Keeble 1015.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . 1025.2 The Ethics of State Responsibility . . . . . . . . . . . . 1065.3 ‘Freedom from Fear’ and ‘Freedom from Want’ . . . . 1115.4 The Ethics of Individual Responsibility . . . . . . . . . 1175.5 Rethinking Human Security . . . . . . . . . . . . . . . 122

6 Braybrooke on Public Policy: Precautionary and Fair; Fea-sible and AmeliorativeSharon Sutherland 1256.1 Introduction to Braybrooke’s Strategy . . . . . . . . . . 1266.2 Approaches to Making Public Policy . . . . . . . . . . 131

6.2.1 The Rational-Deductive, Comprehensively Ratio-nal, Synoptic, or Economics Model . . . . . . . 131

6.2.2 Satisficing, or Environmental Heuristics: Simon’sAlternative . . . . . . . . . . . . . . . . . . . . 134

6.2.3 Some of Simon’s Successors . . . . . . . . . . . 1396.2.4 Incrementalism . . . . . . . . . . . . . . . . . . 142

6.3 Braybrooke’s Citizen Management . . . . . . . . . . . 1456.3.1 Braybrooke’s Disjointed Incrementalism . . . . 1476.3.2 Virtues of Disjointed Incrementalism for the In-

dividual . . . . . . . . . . . . . . . . . . . . . . 1486.3.3 Virtues of Disjointed Incrementalism in a Social

Process . . . . . . . . . . . . . . . . . . . . . . 1506.3.4 Braybrooke’s Cautions . . . . . . . . . . . . . . 152

6.4 Policy Applications Compared . . . . . . . . . . . . . 1536.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . 161

7 Life of Pi and the Existence of TigersSteven Burns 1657.1 Author’s Note . . . . . . . . . . . . . . . . . . . . . . 1687.2 Toronto and Pondicherry . . . . . . . . . . . . . . . . . 1707.3 A Digression on Belief . . . . . . . . . . . . . . . . . . 1737.4 The Tiger in the Lifeboat . . . . . . . . . . . . . . . . 175

viii CONTENTS

7.5 Ockham’s Razor . . . . . . . . . . . . . . . . . . . . . 1777.6 Ministry of Transport Investigators . . . . . . . . . . . 1827.7 Which Is the Better Story? . . . . . . . . . . . . . . . . 1857.8 Is the Tiger REAL? . . . . . . . . . . . . . . . . . . . 187

Part Two: Theoretical Engagement 193

8 David Braybrooke’s Philosophy of Social ScienceMeredith Ralston 1958.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . 1968.2 Traditional Philosophy of Social Science . . . . . . . . 1978.3 Braybrooke’s Contribution . . . . . . . . . . . . . . . . 2038.4 Feminist Epistemologies . . . . . . . . . . . . . . . . . 2128.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . 222

9 Empathy and EgoismSue Campbell 2239.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . 2239.2 Sherman’s Account of Empathy . . . . . . . . . . . . . 2279.3 Default Egoism in Sherman . . . . . . . . . . . . . . . 2299.4 Imagining Empathic Selves . . . . . . . . . . . . . . . 2339.5 Moral Sentiment . . . . . . . . . . . . . . . . . . . . . 2369.6 Common Purpose and Relational Interests . . . . . . . 2399.7 Re-imagining the Self . . . . . . . . . . . . . . . . . . 243

10 The Problem of Moral JudgementRichmond Campbell 25110.1 The Problem . . . . . . . . . . . . . . . . . . . . . . . 25210.2 Moral Realism: The Default Position . . . . . . . . . . 25410.3 Externalist Moral Realism . . . . . . . . . . . . . . . 25610.4 Internalist Moral Realism . . . . . . . . . . . . . . . . 25810.5 The Belief-Desire Theory of Moral Judgment . . . . . . 26110.6 Two Objections Briefly Considered . . . . . . . . . . . 26510.7 The Opposition between Reason and Emotion . . . . . 266

ix

11 Moral Claims and Epistemic ContextsMichael Hymers 27311.1 Moore’s Open Question . . . . . . . . . . . . . . . . . 27511.2 Deflationism . . . . . . . . . . . . . . . . . . . . . . . 27911.3 Contextualism: Default and Challenge . . . . . . . . . 28111.4 Contextualism: Rejecting Epistemological Realism . . . 28511.5 Ethical Contexts and Meta-ethical Contexts . . . . . . . 28911.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . 301

12 Braybrooke and the Formal Structure of Moral Justifica-tionTom Vinci 30312.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . 30312.2 Braybrooke on the Open-Question Argument . . . . . . 30512.3 Explorations . . . . . . . . . . . . . . . . . . . . . . . 30812.4 Three Conceptions of Meta-ethics . . . . . . . . . . . . 31212.5 A Side Trip to Epistemology . . . . . . . . . . . . . . . 31412.6 The Main Line of Advance Resumed . . . . . . . . . . 31712.7 An Alternative Briefly Considered . . . . . . . . . . . . 323

13 David Braybrooke On the Track of PPEPeter K. Schotch 32513.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . 32513.2 How Track Came About . . . . . . . . . . . . . . . . . 32713.3 The Basic Set-up . . . . . . . . . . . . . . . . . . . . . 32913.4 How to Apply Moral Philosophy . . . . . . . . . . . . 33013.5 Actions . . . . . . . . . . . . . . . . . . . . . . . . . . 330

13.5.1 When good actions go badly . . . . . . . . . . . 33413.6 Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . 33613.7 Utilitarianism . . . . . . . . . . . . . . . . . . . . . . 33813.8 Where’s the politics? . . . . . . . . . . . . . . . . . . . 341

14 Bootstrapping Norms: From Cause to IntentionBryson Brown 34514.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . 34514.2 Rules inLogic on the Track of Social Change. . . . . . 349

x CONTENTS

14.3 Some Worries . . . . . . . . . . . . . . . . . . . . . . 35214.4 The Bootstrap Argument . . . . . . . . . . . . . . . . . 35514.5 The Roots of Normativity . . . . . . . . . . . . . . . . 360

14.5.1 Evolutionary Advantage . . . . . . . . . . . . . 36014.5.2 Naturalism . . . . . . . . . . . . . . . . . . . . 36114.5.3 Meta-language. . . . . . . . . . . . . . . . . . 36114.5.4 The bootstrap . . . . . . . . . . . . . . . . . . 36214.5.5 Interpretation of self and others . . . . . . . . . 364

Appendices 367

Appendix AAnother—Literary—Side of David Braybrooke:The ComicDialectician 369Excerpt from Faculty Council Minutes, 29 September 1983 . 372Excerpt from Faculty Minutes, 11 October 1983 . . . . . . . 373An Open Letter from David Braybrooke

Faculty Council Arts & ScienceDalhousie University . . . . . . . . . . . . . . . . . . . 373

Appendix BDavid Braybrooke’s Publications 1955-2005 377Books . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377Journal Articles . . . . . . . . . . . . . . . . . . . . . . . . 378Chapters in Books . . . . . . . . . . . . . . . . . . . . . . . 382Book Reviews . . . . . . . . . . . . . . . . . . . . . . . . . 387

Notes on Contributors 391

References 395

Acknowledgments

The editors and contributors to this volume are grateful for the gener-ous support received from various individuals and offices of DalhousieUniversity, especially the Vice President (Academic and Research), theDean of Arts and Social Sciences, and the Department of Philosophy.

ENGAGED PHILOSOPHY

Chapter 1

Introduction: About DavidBraybrooke

SUSAN SHERWIN

Abstract

This chapter introduces the reader to the life and work of David Bray-brooke. It identifies key themes in his extensive list of publications andexplains the significance of the essays in this collection.∗

This collection of original essays has been produced by faculty mem-bers who have been colleagues or students of David Braybrooke in hisyears of full-time teaching at Dalhousie University (1963-90).1 Our in-tention is to express our collective affection, admiration, and respect forthis important philosopher on the occasion of his eightieth birthday byreflecting on aspects of his life and work as they have inspired our ownphilosophical thinking.

By happy coincidence, the publication of this volume coincides withthe momentous publication of his own book,Analytical Political Phi-losophy: From Discourse, Edification, the fourth in a series of booksBraybrooke has published with University of Toronto Press since 1998.

∗Thanks to Richmond Campbell and Steven Burns for editorial advice on this chapter.1He has been a very active Emeritus Professor since 1990, retaining a summer home in Halifax

and spending three to four months per year in Nova Scotia, enthusiastically engaged in philosoph-ical activities. Hence, all who are themselves situated in Halifax still count themselves among hiscurrent colleagues

3

4 CHAPTER 1. INTRODUCTION

The others—Moral Objectives, Rules, and the Forms of Social Change(1998),Natural Law Modernized(2001), andUtilitarianism: Restora-tions; Repairs; Renovations(2004)—together form an integrated, mu-tually reinforcing corpus of many of the major ideas he has contributedto the literature of political philosophy and the philosophy of social sci-ence throughout his career. So weighty is this four-volume collection,and so absorbing of his philosophical energies over the past decade,Braybrooke has come to think of it, unofficially, as theSumma Philo-sophica Latirivuli (where ‘latus’ stands for ‘bray’—which is dialecti-cal (Northamptonshire) English for ‘broad’—and ‘rivulus’ for ‘brook’).This working title is vintage Braybrooke, combining his lifelong loveof classics, his offbeat, irreverent sense of humour, and an acknowl-edgment of the collective and far-reaching significance of his influenceon philosophical thought. The designation signifies the interconnectedcomplexity of Braybrooke’s philosophical works and makes visible theover-arching program that ties the various pieces together. We fondlyshare in adoption of this designation and add our own essays to comple-ment theSummaby offering reflections on the themes he explores thereand in his earlier publications.

To help set the stage for the chapters that follow, I shall review someof Braybrooke’s significant intellectual and scholarly writings in orderto provide the reader with a sense of how his many publications weavetogether to form a general theoretical system; I hope also to providesome insight into why this system is so valuable. I shall approach thistask by first reviewing his personal journey through political philosophyand then by briefly discussing three key themes that structure his work.2

Finally, I shall briefly explain the organization of this book and highlightthe role of each chapter.

2I am very grateful to David for his assistance and patient cooperation in piecing togethersome of the elements of a rich and complex life in a couple of personal interviews in June 2004.

1.1. DAVID BRAYBROOKE, THE PERSONAL STORY 5

1.1 David Braybrooke, the Personal Story

David3 was born 18 October 1924 in the United States and came ofage during the Second World War. He started his university educationat Hobart College, where he majored in classics with the intention ofpreparing for a career in the United States Foreign Service. His under-graduate study was interrupted, however, by the war. He volunteeredfor service in the United States Army early in 1943 and found himselfstationed in Antwerp several months after D-Day. (He notes that Hitlergave up shortly after he appeared in Europe and suggests that perhapsHitler calculated that if the United States was prepared to send the likesof Braybrooke into war, it must have been very confident indeed.) Forthe first month and a half in Antwerp, he experienced frequent bom-bardments of the city by German buzz bombs and rockets. A rocketlanded a mere block away from where he was waiting for a tram oneday, causing houses and their inhabitants to completely disappear be-fore his eyes. That experience gave him a rich appreciation for life andgratitude for each day he is alive, an attitude that has served him well inall subsequent years; he truly does embrace each day with obvious joyat the opportunities it affords.

In the end, his time in the army, like the rest of his life, was largelydedicated to intellectual pursuits. While still in Antwerp, he managedto find a tutor to help him study Latin and Greek poetry. The army alsosent him to Louisiana State University then to the University of Illinoisto study basic engineering, and, since he found himself on a universitycampus, he managed at LSU to take courses in literary criticism andsociology. He was even able to persuade the army to send him to Cam-bridge, England, for a term before he was discharged.

After a little more than three years in the army, he returned to com-plete his BA and prepare for academic work in the area of social and

3Readers will notice a shift in style of reference to David Braybrooke in this Introduction. Itreflects my own struggle, and that of most of the other contributors to the volume, in deciding howbest to refer to a man who is both an influential scholar and a personal friend. We have resolvedthis problem by using the formal ‘Braybrooke’ whenever we are discussing his work, as meritsrespectful discussion of any author’s work; when speaking personally about the man, as I do inparts of this Introduction, we shift to the more informal ‘David’ to capture the warm regard inwhich we hold him.


political theory. He was torn between studying psychology at Yale oreconomics at Harvard; his decision to pursue economics at Harvard washeavily influenced by Professor Brooks Otis, his mentor from HobartCollege and an early role model for his life as a scholar. Otis was agreat classical scholar who later became chair at Stanford and then uni-versity professor at Chapel Hill; among his other claims to fame is thefact that he introduced David to the work of John Maynard Keynes.David reckoned that it was necessary to have a firm grounding in eco-nomics to work effectively in social and political thought. (As well,David did some personal utility calculations and determined that sincehe would be unlikely to read much economics purely for pleasure, itwas important to approach it in a disciplined way.)

After graduation, and with the continuing support of Professor Otis,David was invited back to Hobart to teach as an instructor in historyand literature within an excellent general education program dealingwith Western civilization. But he grew restless in this role and soonembarked on graduate work at Cornell University in the departmentof philosophy. At that time, Cornell required philosophy doctoral stu-dents to specialize in two subfields of philosophy and also to study onefield outside of philosophy. David chose ethics and epistemology as hisphilosophy subfields and economics for his ‘outside’ interest. (Thosechoices have served him well throughout his career and are well rep-resented within the essays of this collection.) His dissertation was onwelfare and happiness, and the economist on his committee directedhim to include a chapter dealing with Arrow’s work on social choicetheory (much to the chagrin of the philosophers on the committee whothought it was ‘the hardest thing they ever had to read’—‘serves themright!’ in David’s opinion). This led to his first publication, ‘Farewellto the New Welfare Economics,’ a very heartfelt farewell at the time,though he has found himself returning to this subject in later years andnow sees it as one of the foundations of his work.

He completed his PhD in 1953 and took up an instructorship at theUniversity of Michigan. While this opportunity was considered oneof the best jobs in the country for a new philosophy graduate, Davidthought that the industrial design of the Ann Arbor campus had little


charm: buildings were ‘depressing on the outside and graceless on theinside.’ He also felt somewhat alienated from his colleagues there sincethey seemed principally occupied with denigrating the merits of ordi-nary language philosophy while David’s approach was (and is) muchmore pragmatic: use ordinary language philosophy where it is helpfuland employ other methods where it falls short. He soon left Michiganfor Bowdoin College, a small liberal arts college in New England whichhe found to be more to his temperament; indeed, he remembers Bow-doin as being his version of heaven—just the sort of place he had alwayswanted to teach.

Alas, he was not to stay in this idyllic setting for long. By chance,returning by ship from Oxford in 1953, David had met a professor fromYale who proposed him for a special position that would recognize hisdistinctive talents, teaching philosophy in an interdisciplinary programand working with Professor Charles E. Lindblom in the honours pro-gram in economics and politics for juniors and seniors. This was anoffer David found too good to refuse; an ambitious young professorsimply did not turn down offers from Yale then (or now). One of thehappy consequences of this position was a long-term, fruitful collabora-tion with Professor Lindblom. Not all of his colleagues at Yale were sopleased with his versatility and tendency to cross traditional disciplinaryboundaries, however. It seems that he was judged to be ‘too much of aphilosopher’ for the social scientists and ‘too much of a social scientist’for the philosophers. Moreover, the timing of his stay at Yale was mostunfortunate: the entire philosophy department was severely divided be-tween analytical and continental theorists at that time. As a result ofthese struggles over method and philosophical orientation, David wasdenied tenure at Yale.

Fortunately, he won a Guggenheim Fellowship at the very same timeas his tenure decision was being announced, and that honour and oppor-tunity took some of the sting out of the tenure news. (It also helped easethe pain to learn that all of the other analytical philosophers in the de-partment decamped the following year to join the department of philos-ophy at the University of Pittsburgh; clearly, he was not the only one tofind the atmosphere inhospitable to analytically oriented philosophers.)


As a result of having ‘publishedand perished’ at Yale, David de-cided he would ‘retire early’ by moving to a site where he would be freeto teach and live as he liked with little of the pressure to publish thatcharacterized life in an Ivy League institution. His goal now was to en-joy other aspects of his rich, self-directed intellectual life. Fortunatelyfor the contributors to this volume and for thousands of others he hastaught and inspired, he selected Dalhousie University, in Halifax, NovaScotia, for that retirement from the fray of academic intensity. Whenhe arrived in 1963, he found Dalhousie to be a small, provincial uni-versity by the sea that concentrated on undergraduate education and thetraining of doctors, lawyers, and other professionals. The philosophydepartment had only recently been separated off from psychology, andits few other members had little interest in publishing; they exercisedno research demands on him. Moreover, he was able to secure a posi-tion that was to be half-time in philosophy and half in the departmentof political science; in other words, his interdisciplinary interests wereacknowledged and valued at his new academic home.

At that time, there were only about a half-dozen serious scholars inthe humanities and social sciences at Dalhousie, and it was easy to get toknow each of them well. It made for a stimulating, interdisciplinary, in-tellectual community in which David thrived. And while the Dalhousielibrary was small compared with what he had become accustomed to atYale, at least it had the advantage of having the books it owned ready athand since there was so little competition for the books that interestedhim. It is worth noting, though, that this calm backwater atmosphere didnot last very long. Co-incident with his time at the university, Dalhousiehas evolved into a serious research institution with strong departmentsof philosophy and political science. David played no small part in thistransformation, and he is rightly proud of his contribution to a stimulat-ing and supportive environment.

I can speak most authoritatively of his impact on the philosophy de-partment at Dalhousie since it has been my own academic home for thepast thirty years. Here, David has shown both intellectual and personalleadership. On the intellectual front, he has always shared his work withcolleagues and students and encouraged critical discussion with those


interested in philosophy at every stage of learning. Most of his bookshave been at least ‘tried out’ in the classroom if not also initiated there.He believes universities should make more of the opportunity for under-graduates to share in a professor’s active research. They have things tocontribute partly because they are not already steeped in the literature.He has also been remarkably generous in his willingness to engage withthe work of others; no matter how busy he is, he always offers prompt,insightful, and supportive feedback.

Moreover, he has insisted on promoting an atmosphere of true col-legiality and has conscientiously role-modelled the behaviours that sup-port it. Probably influenced by his painful time at Yale, where he foundhimself struggling to launch his own career within a deeply divided de-partment, David has worked hard to foster healthy departmental rela-tions at Dalhousie University. Within the philosophy department, thistakes the form of a few institutionalized rituals. Faculty members meetfor Wednesday lunches at the University Club, which allow us to shareideas and concerns in an informal atmosphere outside of more struc-tured department meetings. Even more important, however, is our weeklyphilosophy colloquium series, which meets all year round, stoppingonly for Christmas and the Labour Day weekend. These sessions arescheduled for Friday afternoons and are routinely followed by beer atthe Graduate Students’ Club, where students, faculty, and interested oth-ers can continue to debate the topic of the weekly paper.

Much to our local benefit, David became so content at Dalhousiethat he was able to resist several tempting efforts to recruit him to joinhis analytical colleagues from Yale at their new home at the Universityof Pittsburgh, despite the fact that their arrival there helped to make itthe strongest philosophy department in North America at the time. Al-though David sometimes wonders if his work would have received evenmore attention if he had accepted one of those offers and situated him-self in a leading American university, he generously credits Dalhousiefor providing a stimulating environment that allowed him to be produc-tive and creative. Moreover, it was a place in which he was, for themost part, happy, a condition he thinks played an important role in hisachievements.


Relieved of the pressure to produce that defined life in an Ivy Leagueuniversity, he found himself free to pursue whatever research projectscaptured his imagination. He soon settled into a very productive pub-lishing pace that seems only to accelerate with age. On reviewing hislong and growing list of significant publications, one might easily getthe sense that David is a man totally occupied with work. That wouldbe a big mistake. He has always understood that while meaningful workis very important to everyone, it is not the whole of life. It is also im-portant to make room for travel, music, literature, art, romance, and,generally, for fun. (He loves to laugh, and his laughter is infectious.Finding tidbits of absurd news stories to amuse him has become some-thing of a competitive sport among his colleagues and friends. Readerscan get a pretty good impression of his mischievous side by reading Ap-pendix A.) Without such activities, he believes, one becomes less ratherthan more productive, and definitely less creative.

As an academic, he takes seriously all dimensions of a professor’s re-sponsibilities: research, teaching, and administrative work. His researchaccomplishments are readily available to all and form the basis of mostarticles in this book. His teaching and administrative accomplishmentsare less visible to the larger world, but they have had a profound impacton the thousands of students and colleagues who have directly benefitedfrom the opportunity to work with him (and on those who have the op-portunity of studying/working with others he has inspired). He lovesteaching and it shows. (Some recent and cherished evaluations report‘Braybrooke rocks!’ and ‘Professor Braybrooke is fun to be with.’) It iswith much ambivalence that he will finally cease offering formal classesin 2005 as he turns eighty-one.

Indeed, it is this love of teaching that eventually took him to Texas.Dalhousie University has long had a rule of mandatory retirement atage 65. David was by no means ready to give up teaching in 1989—it will be difficult enough to do in 2005—so he decided it was finallytime to accept one of the many offers he received over the years fromthe United States and took up the position of Centennial CommissionChair in the Liberal Arts (Professor of Government and Professor ofPhilosophy) at the University of Texas at Austin. Fortunately for us,

1.2. DAVID BRAYBROOKE, THE SCHOLAR 11

he maintained a home in Nova Scotia and has arranged to spend a fewmonths each summer back at Dalhousie, where he remains an activepresence in our summer reading groups and weekly philosophy collo-quium sessions. His philosophical engagement with his immediate anddistant colleagues shows no signs of slowing, despite the fact that he isnow approaching yet another retirement (this time from the Universityof Texas at Austin).

1.2 David Braybrooke, the Scholar

Braybrooke’s work centres around the interrelated themes of needs,rights, and rules with particular attention to the appropriate processesfor making changes in existing social rules. He argues that proper un-derstanding of these three concepts helps to constitute a meaningful andpractical approach to social justice. Hence, he has been concerned withmaking clear how to interpret and apply these familiar, but often mis-used, concepts that figure so prominently in political debates about pub-lic policies.

While deeply respectful of alternative philosophical schools, Bray-brooke situates his own work within the analytic tradition that has domi-nated English-language philosophy for many decades; it relies on rigourand conceptual analysis as principal tools. His aim is to help guide pol-icy debates by allowing participants to determine appropriate rules forattending to the needs of citizens of nations and of the world in a fair andachievable way. His staunch support of analytic philosophy has meantthat throughout his career he has frequently found himself embroiled inwhat might be dubbed ’ ‘philosophy wars’ (akin to the recent strugglesover the nature of science and scientific activities commonly referred toas the ‘science wars’). Indeed, his most recent work, the forthcomingAnalytical Political Philosophy, has been written with the principal in-tention of responding to charges from pro-Straussian colleagues at theUniversity of Texas at Austin who claim there is no value in analyticalapproaches to political philosophy.4 Braybrooke’s work clearly demon-

4It is ironic that David discovered Leo Strauss long before any of the people who now chal-lenge David’s philosophical approach in the name of Strauss: After one year of undergraduate


strates the error of this thesis.His efforts at clarifying and applying the core concepts of needs,

rights, and rules add up to a program for democratic action. It incor-porates three basic assumptions: (1) meeting certain needs is the basicpurpose of public policy, (2) rules are required to make sure appropriateneeds are met, and (3) these rules must include rules governing rights.Braybrooke has long been willing to stare down the challenge of post-modernism; indeed, he is even willing to accept the somewhat unfash-ionable label of ‘grand program’ to describe the intent of these effortsto ‘render political life more coherent, more just, and even more no-ble.’5 His own grand program is particularly valuable by virtue of thefact that it has been developed in tandem with (and, hence, attentive to)other important philosophers offering their own versions of the grandprogram in political philosophy, principally, John Rawls, Robert Noz-ick, and David Gauthier. Like them, he sets out his understanding of thebasic structure available to a politically organized society. Braybrookediffers from these others in an important way, however: he situates theseterms in the context of real-world concerns of daily politics in an indus-trialized Western democracy (and, hence, the title of this collection ofessays). Moreover, he is not one to forget his debts to the history ofphilosophical ideas; he makes great efforts to relate his own programwithin two important, and usually considered to be widely different,historical traditions: utilitarianism and natural law theory.6

Let us look, then, more closely at his work on these three key con-cepts, beginning with needs.7 Braybrooke argues that surely the princi-pal question for all social policies is to consider their impact on humanwelfare. Such a project requires some way of measuring the relevantimpact of policies, and that requires us to identify appropriate criteriafor our focus. As countless critics have demonstrated, utility as it is

study, David headed off to the New School to take some graduate classes. There he had the oppor-tunity to take a class from the soon-to-be-famous Leo Strauss on the work of Socrates, Machiavelli,and Hobbes.

5This phrase is taken from the Introduction toAnalytical Political Philosophy, draftmanuscript.

6These historical connections are made most clearly and fully in his recent books,Utilitari-anism(2004) andNatural Law Modernized(2001).

7The first comprehensive discussion of needs appears in his 1987 book,Meeting Needs.

1.2. DAVID BRAYBROOKE, THE SCHOLAR 13

generally understood is too problematic a measure for this task, thoughits principal insight of considering the welfare of everyone is sound.Braybrooke argues that a conception of human needs can better servethe role of evaluating the social worth of public policies. Such a con-cept must distinguish between mere wants or preferences and the thingsthat are necessary to a life worth living. Thus, public policies should beheld accountable for first meeting the latter and, then, for allowing asmuch room as possible for individuals to pursue their own preferences.Needs may be either course-of-life (what is required for life or health)or adventitious (what is required for meeting particular goals), and itis important to understand the distinction to ensure that policies do notlose track of course-of-life needs in responding to vocal demands formeeting particular adventitious needs.

While some course-of-life needs are clearly universal by virtue ofbiological requirements for nutrition, hydration, shelter, sleep, and soon, others are not (e.g., medical care), and adventitious needs may bevery diverse indeed. Hence, public policy must be evaluated in terms ofthe needs of a specific reference population; those responsible for deter-mining these needs will belong to a policy-making population, whichmay not be the same as the reference population.8 It is, ultimately, amatter of general public debate what shall be deemed needs that mustbe met for a given population, the minimum standard of provision ofthose needs, as well as the policies that should be pursued to meet thoseneeds. Clearly, then, the rules for that debate become very important.So, too, is a measure for comparing the effectiveness of different poli-cies at meeting the needs of the population in question. Braybrookeintroduced the census notion (described below) to solve that problem.9

Braybrooke’s work on the concept of rights is intertwined with hiswork on the concept of rules.10 He argues that rights attach to individu-

8I follow Braybrooke’s terminology in his recent work here; it is far more accessible than thealternative form ‘Selfgovliset’ which he used in the originalMeeting Needsbook.

9The census notion requires agreement on minimal standards for provision of a need (e.g.what would constitute adequate nutrition for a 150-pound adult), and then everyone in the popula-tion is to be surveyed to see whether they are receiving enough calories and vitamins to meet thisminimum; priority is then placed on bringing everyone up to the minimum (for their particularsize).

10The first substantive discussion of rights appears in his 1968 book,Three Tests for Democ-


als, and their administration is largely left to the individual in question.Individuals are not alone in their responsibility for enacting their rights,of course. There is a collective responsibility to ensure that the rights ofothers are respected. This shared responsibility often requires more thanjust formal acknowledgment of the rights of others; it may also requireus to ensure that the conditions are in place to permit them to exercisetheir rights. Governments are to be judged according to the respect andprotection they provide of individual rights (both our own and those ofour fellow citizens). His theory of rights is based on what he has de-veloped as a census notion of welfare. According to the census notion,policies should aim to push as many people as possible into higher cat-egories of welfare, leaving as few as possible behind. In this way, he isable to set measurable tests for democracy that link both rights and wel-fare (as he has argued since his 1968 book,Three Tests for Democracy:Personal Rights, Human Welfare, Collective Preference).

On his account, rights are, ultimately, social: they only make sensewithin a defined social context. More specifically, they are best under-stood as a particular type of settled social rule. As such, rights apply andare exercised within particular societies. Things are a bit more compli-cated yet, since rights actually involve two kinds of rules: those gov-erning what it means to have and exercise a right and those determiningwho actually has a particular right. Because rights represent a kind of(double) rule, it is useful to understand their formal structure if we wishto make evaluations about governments’ ability to respect rights. Thisinterest in getting clear on the nature of social rules has led him into theformalism of deontic logic (inLogic on the Track of Social Change).As Peter Schotch demonstrates in chapter 13, the formal work on rulespermeates multiple dimensions of Braybrooke’s grand program.

1.3 The Structure of the Volume: Overview of Essays

The rest of the essays in this book take up several of the major ideasin Braybrooke’s work and are divided into two major blocks reflecting

racy. His work on rules appears in many places but is most thoroughly discussed inLogic on theTrack of Social Change.

1.3. OVERVIEW OF ESSAYS 15

the general theme of the book and of Braybrooke’s work. Part I, Practi-cal Engagement, deals with ways in which Braybrooke’s ideas connectup with the practical world of public policy or personal responsibili-ties. Part II, Theoretical Engagement, consists of essays that explorethe philosophical underpinnings of his work and look more closely atsome of the theoretical debates that occupied him.

More specifically, chapters 2 through 4 focus on his very importantwork on needs. In chapter 2, Nathan Brett brings this framework to bearon the subject of private education. He explores the threat to justice ofincreased moves to privatization of education for families who can af-ford it. Braybrooke views education as a basic need of human beingsand warns that justice requires us to ensure that this need is met to areasonable standard for all young people. With this concern for equaleducational outcomes, Brett examines some of the current debates re-garding private versus public education services and examines the nega-tive impact on equality of private schooling. He concludes that the trendto increased use of private schooling is resulting in reductions of socialjustice.

In chapter 3, I take up Braybrooke’s work on needs to explore apressing question about how we should address the potentially growingset of medical demands for services that test for genetic dispositionsto contract life-threatening illnesses. I look specifically at the topicof access to tests for genetic susceptibility to inherited forms of breastand ovarian cancer. By reflecting on Braybrooke’s distinction betweencourse-of-life needs and adventitious needs, and drawing on feministwork on relational understandings of personhood and needs, I am ableto help refocus the debate and find a basis for distinguishing genuinehealth needs from health desires; while the latter may be strongly felt,only course-of-life needs should be given priority in allocating limitedpublic health resources.

Duncan MacIntosh (chapter 4) takes up the arguments of chapter 3and some in Braybrooke’s book,Meeting Needs(1987). He consid-ers several possible solutions to the problem Braybrooke first identifiedas the ‘bottomless pit’ of medical needs for technological support ofthose whose lives could be extended. MacIntosh argues, contra Bray-


brooke, that there is no clear and reliable distinction to be drawn be-tween course-of-life needs and adventitious needs. He proposes a strat-egy that will constrain the bottomless pit of medical demands by onlypermitting communities to recognize as needs with attached rights thoseprojects that are in fact co-feasible within that community.

Edna Keeble (chapter 5) pieces together some of Braybrooke’s ideasabout democratic theory and questions of state and personal responsi-bility in order to address some urgent questions in the philosophy ofinternational relations. She reviews his proposals regarding the per-sonal responsibilities of individuals living in rich countries to relievethe miseries of people living in less fortunate countries. Keeble drawson his philosophical ideas about responsibility to explore the meaningand possibility of global justice in the post-9/11 era, when national se-curity has become the preoccupation of governments. She appeals to hisideas about cosmopolitan ethics while discussing some of his proposalsfor practically deploying such an ethic and draws from his work a moreoptimistic view of international relations than is evident within currentgovernment policies.

In chapter 6, Sharon Sutherland traces Braybrooke’s insight into theimportance of incrementalism to debates in public policy that he takesup throughout his career, beginning with his very first book,A Strat-egy of Decision: Policy Evaluation as a Social Process, published withC.E. Lindblom in 1963. In his ongoing efforts to make public policyresponsive to needs, Braybrooke has always recognized the importanceof developing institutional processes that ensure debate and revisabilityas policymakers evaluate the impact of their efforts. She shows howthis approach is central to his major contribution to specific policy inhis work on traffic congestion (a subject of lifelong interest to him asthe son of a highway designer). This dynamic strategy is taken up andvigorously argued most recently in his 2004 bookUtilitarianism.

The section concludes in a different mood entirely, with Steven Burns’sessay ’Life of Pi and the Existence of Tigers’ (chapter 7). While Burnsdoes not discuss Braybrooke’s publications per se, he does demonstratean important aspect of Braybrooke’s well-known views about how anindividual ought to live one’s life. As noted above, Braybrooke has


long understood that a productive intellectual life depends on living arich life overall, engaging in many areas of stimulating activity, espe-cially the arts. His early studies involved work in literary criticism,and his first appointment was in a program of study of great books ofWestern civilization. Literary examples populate and illuminate all ofBraybrooke’s work. Moreover, as we see in Braybrooke’s own descrip-tion of life as an academic (Appendix A), he thought it important tofinish out a day of scholarship by turning to other forms of intellectualstimulation besides philosophy; he was, himself, particularly partial toliterature and satire. Thus, Burns’s discussion of philosophical themespresent in the important novelLife of Pi pays tribute to Braybrooke’sideas about how a philosopher should engage with the world.

In Part II, Theoretical Engagement, we move from essays that ex-plore some of the implications of Braybrooke’s work in practical orworldly matters to ones that concentrate on some of the more theoreticalquestions he has published on in greater abstraction. As a bridge, we be-gin this section with Meredith Ralston’s helpful review of Braybrooke’swork on the philosophy of social science, ranging from his 1965 col-lection, Philosophical Problems of the Social Sciencesthrough to his1998 book,Moral Objectives, Rules, and the Forms of Social Change.Ralston explores Braybrooke’s efforts to distinguish three overlappingschools (naturalist, interpretive, and critical social science) that he pro-poses should be understood as offering ‘mutual support and stimula-tion’; each has a role to play in particular types of analysis. Ever thepeacemaker, Braybrooke argues that we are best served by a philosophythat permits use of multiple methods according to the problem before usrather than a single approach for all circumstances. In reviewing Bray-brooke’s contribution to this literature, Ralston finds within it a positionthat resonates with contemporary feminist epistemology and philosophyof science.

In chapter 9, Sue Campbell picks up a theme that is central to Bray-brooke’s work on needs and all of his discussions of political philoso-phy, namely his rejection of philosophical egoism. She looks, in par-ticular, at the way Braybrooke challenges the traditional philosophicalapproaches to egoism in his bookNatural Law Modernized(2001). Sue


Campbell joins him in asking why egoism is the default account of hu-man motivation; she argues for an alternative starting point that beginswith reliance on moral sentiment as the basis for a more social theory ofmotivation. Like Sherwin and Ralston, she finds within Braybrooke’swork a conceptual apparatus that is very close to the work of feministphilosophers, though she also sees that his work would benefit frommaking these connections more explicit and holding more firmly to thevision feminists have developed.

In chapter 10, Richmond Campbell continues the exploration of ques-tions regarding the role of sentiment in moral judgment and focuseson an important question in moral epistemology and meta-ethics. Hediscusses some of Braybrooke’s ideas regarding moral justification andmoral judgment. Richmond Campbell revisits a familiar philosophi-cal impasse between cognitivists, who argue that moral judgments area form of belief that can be true or false, and non-cognitivists, whoclaim that moral judgments are expressions of feeling or desire. Theseare always represented as mutually exclusive options. He seeks to re-solve this apparent impasse by drawing on an idea found inNaturalLaw Modernized(2001), where Braybrooke suggests that David Humecan be viewed as someone who embraces aspects of both positions. Bybuilding on this insight, Richmond Campbell is able to offer a novel in-terpretation of moral judgments as constituting a complex, hybrid state,thereby dissolving the apparent opposition between reason and emotion.

Chapters 11 and 12 continue to examine the implications of Bray-brooke’s work for questions in moral epistemology. Both Michael Hymersand Tom Vinci use a recent essay of Braybrooke’s titled ‘What TruthDoes the Emotive-Imperative Answer to the Open-Question ArgumentLeave to Moral Judgments?’ to discuss themes about the nature ofmoral knowledge. Hymers (chapter 11) discusses the importance ofBraybrooke’s claim that ethical justification does not have to be tracedback to meta-ethical justification. Taking a Wittgensteinian approachto the nature of moral propositions, Hymers proposes that justificationand, in particular, certainty of moral propositions are a function of epis-temic context, not semantic or empirical content. Vinci (chapter 12)takes a different approach to the issues raised in Braybrooke’s stimulat-


ing paper. He argues that Braybrooke situates his account of moral ruleswithin a theory of meta-ethics. Vinci discusses the structure of this ac-count within the broader theory of formal foundationalism with analogyto epistemic justification. This connection provides a useful model forunderstanding Braybrooke’s account of moral justification, but it is notwithout difficulties of its own.

We then follow Braybrooke in his decision to turn to formalism tocapture some of the important ideas regarding the ways in which soci-eties actually come to decisions and set (and change) policies. In chapter13, Peter Schotch explores the intersection between Braybrooke’s inter-ests in moral philosophy and social science. Through graphic repre-sentations, Schotch helps to make accessible the formal work he, Bray-brooke, and Brown have done on rules in order to render more visibleBraybrooke’s understanding of the place of political economy in a rule-based picture of morals. This essay draws together the ways in whichthe concept of rules runs through a substantial body of Braybrooke’s lifework.

In chapter 14, Bryson Brown continues to expand on the importantand difficult role that rules play in Braybrooke’s approach to moral andpolitical thought. He reminds us that inLogic on the Track of SocialChangethe authors explain how rules combine both descriptive andnormative features. In this essay, Brown draws on the work of Wil-frid Sellars to help elaborate the nature of normativity that is built intotheir shared understanding of the nature of rules.

Together, these essays pay our respect to David Braybrooke in thebest way we know how, by critically engaging with his work. It reflectshis teaching that philosophy is a conversation, best pursued throughopen discussion. David describes his principal interest in all of his workas that of ‘being useful.’ In that, he has admirably succeeded.

I trust this Introduction has given the reader a sense of the impor-tance of David Braybrooke’s contributions to the literature in many di-mensions of philosophy. That sense will surely be confirmed by readingthe full volume. I may not, however, have succeeded in conveying theother side of David Braybrooke: his irreverent, irrepressible determina-tion to have fun. At his urging, we have included, as Appendix A, some

excerpts from an essay he published in theDalhousie Reviewin 1985. Itdescribes the saga of his efforts to have a description of the way a pro-fessor spends his time in the summer months included in a documentbeing prepared by the faculty of arts and science for the government. Init, he reveals his sense of how he himself spends his days and also hisdistinctive sense of humour as he pokes gentle fun at colleagues who,to his mind, take this question altogether too seriously.

1.4 References

Braybrooke, David. 1965.Philosophical Problems of the Social Sci-ences. New York: Macmillan.

— 1968.Three Tests for Democracy: Personal Rights, Human Wel-fare, Collective Preference. New York: Random House.

— 1987.Meeting Needs. Princeton: Princeton University Press.— 1987.Philosophy of Social Science. Englewood Cliffs, NJ: Pren-

tice Hall, Foundations of Philosophy Series.— 1998.‘Philosophy of Social Science, Contemporary Theories (Schools).’

In E. Craig, ed., RoutledgeEncyclopedia of Philosophy, 8: 838-47.London: Routledge.

— 1998.Moral Objectives, Rules, and the Forms of Social Change.Toronto: University of Toronto Press.

— 2001.Natural Law Modernized. Toronto: University of TorontoPress.

— 2004.Utilitarianism: Restorations; Repairs; Renovations. Toronto:University of Toronto Press.

— 2005.Analytical Political Philosophy: From Discourse, Edifica-tion. Toronto: University of Toronto Press.

Braybrooke, David, Bryson Brown, and Peter Schotch. 1995.Logicon the Track of Social Change. Oxford: Clarendon Press, ClarendonLibrary of Logic and Philosophy.

Braybrooke, David, and C.E. Lindblom. 1963.A Strategy of De-cision: Policy Evaluation as a Social Process. New York: The FreePress.

Part One:Practical Engagement

Chapter 2

Teaching Class: Justice andPrivatization in Education

NATHAN BRETT

Abstract

This paper looks at some of the issues raised by recent moves towardprivatization in education. I begin the debate over private-versus-publicservices by considering the claim that a private system is a more efficientway of achieving the goals of education. The next section takes up thisquestion of differences in goals by examining some basic differences inthe normative structures of the public and private spheres. Using (and de-fending) David Braybrooke’s principle of equality in meeting needs, thepaper concludes by arguing that the privatization of educational servicesundermines in important ways the pursuit of equality of opportunity.∗

2.1 Introduction

David Braybrooke has been a stern critic of moves toward privatiza-tion that began with the Thatcher and Reagan regimes in Britain andthe United States respectively. In his work he has treated education

∗An early draft of this paper was present at the Society for Applied Philosophy Meetings:Education in the 21st Century, Oxford, England, June 2002. I would like to thank Steven Burnsand Susan Sherwin for their helpful editorial comments. Thanks also to Caroline Sequeira for herhelp in editing.

23

24 CHAPTER 2. NATHAN BRETT

as a need that requires a public response. This paper explores the im-plications of moves toward privatization in education and attempts toshow why education should not continue down this path toward priva-tization. This will involve considering some basic differences betweenpublic and private enterprises. It will also explore the implications ofBraybrooke’s views that education is one of the basic needs of humanbeings and that a just society aims at equality in meeting needs.1

Judged from the viewpoint of equality, the political regimes thathave held power in much of the West over the past twenty-five yearshave been a disaster. We have witnessed a widening of the gap be-tween the rich and the impoverished countries of the world. Wealth hasbecome concentrated in the hands of fewer individuals. While somepeople are enormously wealthy, average real wages and employmentin full-time work have decreased. In North America (and elsewhere)more wage earners have been moved into ‘flexible’ categories of work,where they do not have job security or health and retirement benefits.Meanwhile, the power of labour unions to resist these changes has beengreatly diminished by the globalization of production and markets.

These changes toward greater social inequality have also been evi-dent in the reorganization of our educational establishments. Over thepast ten years, for example, the amount of public money going to Cana-dian universities has (in real terms) been reduced, while the costs ofthis education have continued to rise. One result has been dramatic in-creases in university tuition. In my own province (Nova Scotia), tuitionhas increased by 100 per cent over the last ten years. This is not ex-ceptional. A recent study of state-supported universities in the UnitedStates found somewhat lower rates of tuition increases (107 per centover twenty years) but reflected the same shift toward university educa-tion accessible only to those who can afford it.2

1See (Braybrooke, 1987a) Chapter 2 Section 2 ‘Course of Life Needs,’ and Chapter 4 Section5 ‘Equality in Meeting needs,’ 143-50. See also Section 6 below.

2Jacques Steinberg, ‘More Family Income Committed to College,’ New York Times, 2 May2002. ‘While the average amount each state spent on higher education per student rose by 13percent over that period, to $6,747, state institutions raised their tuition and fees by 107 percent, to$3,512, the study [‘Losing Ground’] found.’ SeeLosing Ground: A National Status Report on theAffordability of American Higher Education, p. 9,http://www.highereducation.org/reports/losing ground/affordability report final bw

2.1. INTRODUCTION 25

It does not take an expert to see that these changes will foster greatersocial inequality. Given that education is a central need of human beingsand that it connects in dramatic ways with the meeting of other needs,this inequality is an important source of injustice.

There is a corresponding differentiation of opportunitywithin theuniversities, as administrators (like their counterparts in industry) havesought a cheaper and more flexible workforce by increasing the pro-portion of untenured and part-time employees, who lack security andbenefits and work for about one-third of the salary of tenured faculty.3

At the same time, universities have become more and more dependenton private corporations to fund research programs. This has meant thatprivate enterprise has sometimes been in a position to determine whatquestions will be investigated and even what findings will be made pub-lic.4 Though this is a direct affront to the ideals of a liberal university,it may not directly contribute to inequality. But indirectly it does. Sincethese very corporations are the chief sources of the hierarchal dispersionof wealth, research that contributes to their power will ultimately havean inegalitarian impact.

Public schools, too, have experienced undermining cuts to their fund-ing. The Halifax School Board recently decided to cut the funding forFrench immersion schools. Private schools will be the beneficiaries ofthis change, as parents who can afford it move their children to the pri-vate schools. In some provinces (and many states in the United States),such moves to private schools are being reinforced by tax policies thatallow tuition to count as a tax deduction. There is a corresponding re-moval of funds from the public system. Like the universities, publicschools have sought ‘partnerships’ with private industry as a replace-

3See Wesley Shumar,College for Sale: A Critique of the Commodification of Higher Educa-tion (London: The Falmer Press, 1997), 52-6. Shumar claims that underpaid part-time faculty nowdo 50 per cent of the teaching on average in U.S. universities. Canadian faculty have been resistingthis trend. (Though it took a protracted strike to settle all of the issues, the Faculty Association atDalhousie University was successful in holding the line at 10 per cent.)

4In Canada the names David Healy and Nancy Olivieri come to mind in connection withcorporate attempts to control findings about dangerous side effects of products. (These researchersbrought to light attempts to control their negative findings about pharmaceutical drugs.). See J.Thompson, P. Baird, and J. Downie, The Olivieri report:The Complete Text of the Report of theIndependent Inquiry Commissioned by the Canadian Association of University Teachers(Toronto:James Lorimer, 2001),http://www.dal.ca/committeeofinquiry.


ment for public funding. The results can be alarming: corporations withpoor environmental records, for example, are sometimes supplying thevideo materials used in classroom discussions of environmental issues.5

I begin the debate over private-versus-public services by consider-ing the claim that a private system is a more efficient way of achievingthe goals of education. This brief discussion supports two (related) con-clusions. The claim of greater efficiency is (often) used in a way thatmakes it true a priori, though it poses as an empirical claim. Second, ef-ficiency comparisons will not be valid if public and privatized systemshave different goals—that is, assume different answers to the question,‘Efficient at what?’ The next section takes up this question of differ-ences in goals by examining some basic differences in the normativestructures of the public and private spheres. The account contrasts theimpartiality of public norm structure with the partiality of that which isprivate. This account is then applied to the question of equality in educa-tion. I argue that the privatization of educational services is an importantstep away from conditions of social justice. In defending education asa public service (comparable to a public health system) I utilize DavidBraybrooke’s views that education is one of the basic needs of humanbeings and that a just society aims at equality in meeting those needs. Iconclude that the shift of educational services into the private sector isan important step away from conditions of social justice.

2.2 Privatization and Costs

Privatization occurs when some practice or enterprise that has beenwithin the public domain, and financed through tax resources, is movedto the private sector and subjected to the competitive environment ofthe market. Advocates and opponents of privatization of services, suchas education and health care, both make claims of greater efficiency in

5For examples, see Maud Barlow and Heather Robertson,Class Warfare: The Assault onCanada’s Schools(Toronto: Key Porter Books, 1994): ’A Dow Chemical sponsored video,Tracesof Today, created in conjunction with an educational marketing company, Modern Talking PicturesService, claims “scientific studies and practical experience around the world have shown thatincinerators are an environmentally safe method for disposing of combustible material, includingplastics”’ (81).

2.2. PRIVATIZATION AND COSTS 27

defending their positions. The former see increased efficiency as a nec-essary result of removing a non-competitive public bureaucracy. Thelatter see privatization itself as necessarily introducing extra layers ofprofit-makers and hence expenses.

2.2.1 Efficiency

The ideology of ‘neoconservatism,’ currently dominant in the politi-cal decisions of Western democracies, views privatization as progress,a change toward increased efficiency, reduced taxes, and ultimately ahigher standard of living. Bringing services into the private sector is,for many proponents, a moral imperative. Even if public operation ofthe education or health care systems provided care equivalent to thatprovided by a private system at less cost,6 many proponents of privati-zation would be against it because opportunities for private enterpriseare lost. But many advocates of this ‘new economy’ would find theidea of greater efficiency in a public system simply incredible. On thisview, a public service bureaucracy is inefficient by definition. Moreover,state-run monopolies on health care or education involve an absence ofcompetition, which protects inefficiencies. Adherents of this politicalview seem not to notice that the market itself can produce monopoliesand provide room for vast corporate bureaucracies—organizations thatimpose costs and exercise power but contribute nothing of value to oth-ers in society.7

There is an ideology of the left that might be described in equivalentterms. According to this view, whatever is done ‘for profit’ is necessar-ily done at someone’s expense and is therefore less efficient. Hence, adirect provision of services through the public sector is inherentlymore

6According to the Council of Canadians (www.canadians.org) total health care spending inCanada - under public health care—is $3,298 per person as compared with $7,000 per personunder the private system in the United States. The Canadian system covers all persons; one personin seven is not covered under the U.S. system. In terms of life expectancy, Canada rated second-highest worldwide, the United States twenty-fifth.

7Many private sector institutions can be described in these terms—e.g., those involved incurrency speculation, misleading advertising, and organizations of the sort revealed by the col-lapse of WorldCom and Enron in the United States. Scrutiny of such cases reveals that some-times the removal of red tape—deregulation—has contributed remarkably to the inefficiency ofour economies.


efficient. In both of these positions one can find a claim that is truea pri-ori. Clearly, eliminatingunnecessarybureaucracy will reduce costs. Ofcourse, there is a cost in providing government services that we ignoreat our peril when we speak of health care or public education provided‘free of charge.’ On the other hand, it is true that privatization involvesthe introduction of profit-makers—or perhaps layers of them—into asystem that provides these services; and these changes necessarily in-troduce costs. The huge cost of private medical health insurance in theUnited States is a case in point. But in neither account is it a necessarytruth that the overall cost will be greater.

2.2.2 Benefits

Surely we can compare a publicly run British Rail with its privatizedcounterpart, and compare them across several dimensions, in arriving atan assessment of greater or lesser efficiency. But to do this, we will haveto ensure that we take into account any change in benefits. If privatizedtrains no longer run on time, or no longer service outlying areas, or aremore prone to deadly crashes, the fact that less is being spent showsus nothing about efficiency. Similarly, if changes in the provision ofmedical or educational services turn out to be less expensive becauseonly those who can afford them get treated or educated, then we canhardly say that this is more efficient until we have reached agreementabout this change of benefits. Typically, changes toward privatizationend up serving a smaller part of the population and differentiating in thelevel of benefits received by those who are served.

The opposition between these alternatives, therefore, is not merely aquestion of fact about the best way of doing the same things. The centreof the dispute is inevitably a normative dispute about goals. In the caseat hand, it involves disagreement about the proper aims of education.To privatize is to move in the direction of exclusion and competition. Apublic solution, on the other hand, demands inclusion and cooperation.Moreover, there are different criteria of desert and entitlement that liein the normative backgrounds of the two approaches. Even the ideal ofequality is differently interpreted, depending on whether the approach

2.2. PRIVATIZATION AND COSTS 29

is a public or private one.8

This claim that public and private educational services involve es-sentially different goals may seem to be contradicted by facts of whichwe are all well aware. Public and private schools have existed side byside in many communities for years. It is probably true, as well, that thepublic system evolved out of private beginnings.9 Students often movefrom one system to the other; exam results from the two systems canbe compared; and often the curricula are very similar. It would seem tobe nonsense of the sort I have just depicted to insist ona priori groundsthat these schools have fundamentally different aims.

This point is helpful in clarifying the object of this critique. In rais-ing doubts about the privatization of education, I am primarily interestedin changes that aim to move educational services more fully under thecontrol of private corporations. Traditional private schools (as opposedto schools run by businesses such as the Edison Corporation) do notraise all of the same worries. Typically, the traditional private school isnot run with the primary goal of making and increasing profits. Theseschools are voluntary associations aimed at providing distinct and oftensuperior educational opportunities for the students they serve. However,even traditional private schools give rise to some questions of justicethat are of concern here. Because they give wealthier parents a chanceto provide educational opportunities that are unavailable to others, theyexist in conflict with a sense of justice that demands equality in educa-tion. I will return to this.

In the reassignment of education more fully to the market, schoolingbecomes a ‘commodity’ which students consume, and the logic of themarket governs essential decisions about what will be taught and howinstruction will be carried out. Promoters of privatization argue that thecompetitive conditions of the market provide the best environment forthe preservation and enhancement of what we value independently ofthe market. A Mercedes is a better car than a Lada despite (and perhaps

8The sense of equality endorsed by a public system goes beyond mere equality ofopportunity—if that is construed to mean only an equal chance to compete; it includes adjust-ments aimed at equalization relative to need.

9See Myron Lieberman,Public Education: An Autopsy(Cambridge, MA: Harvard UniversityPress, 1993), 14-16.


because of) the fact that the Mercedes was manufactured by a profit-seeking corporation and the Lada was not. The process of extracting aprofit does not entail that automotive standards were sacrificed. On thecontrary, the successful producer is efficiently producing what peoplewant. And what people want must meet certain standards. Likewise,it may seem, the goals of education could be pursued and enhanced bycorporations dedicated to making a profit by advancing the same goalsendorsed by public education. On the other hand, a system of publiceducation may be producing the educational equivalent of a Lada, notdoing a good job of serving the goals of liberal education.

One problem with this argument is the assumption that profit-drivenexchanges in the market can provide and enhance whatever we value.10

If this were true, then prostitution would be the ideal form of sexual inti-macy. Not all that we value can be reduced to or mirrored by commodityvalue.11 Education is a part of what makes us fully human, and to allowprofitability to dictate its goals is to treat humanity as a mere means. Asecond problem is the oversimplified conception of the market that isassumed when we claim that the markets can provide superior meansto whatever we want. A missing element in this picture is the modernmarket’s role in producing the ‘educational needs’ that it then satisfies.To survive in the competitive modern economy, vendors of a productor service must ‘stimulate consumer demand’ through advertising. An-other key to success in modern markets lies in selling products that mustbe replaced and services that will need to be repeated. Of course, adver-tising and obsolescence can work together. Building obsolescence intoproducts may backfire unless one also builds it into the thinking of ‘con-

10’Why isn’t everything for sale?’ asks Elizabeth Anderson inValue in Ethics and Economics(Cambridge Harvard University Press, 1993). Her answer is that we have a plurality of values,many of which cannot be captured in commodity values (chap. 8).

11But here a common point of the defenders of privatization arises: teachers are paid for whatthey do. If profiting were incompatible with what we value about education, then it is hard to seewhy it is not already devalued. That teaching is financed and organized under a market systeminstead of a tax-based public system will not alter the fact that teachers are already teaching ’forprofit.’ One reply to this is that teachers would be doing something else if their motivation weresimply to profit from what they are doing. This reply has some salience: teaching is not (most ofthe time) alienated labour; it is generally valued intrinsically. But the point seems to remain: if paycheques do not corrupt teaching, why would it be changed for the worse through full incorporationinto the market? We will have to dig deeper.

2.3. TWO NORMATIVE CONTRASTS 31

sumers’ by getting them to believe that whatever is not new is obsolete.As education moves further into the private sector, these aspects of themarket will become increasingly important. Students and potential stu-dents will be further targeted by advertisers. In the competitive market,forms of education that are profitably served will be ‘naturally selected’over those that are not. The market may not place a premium on criticalthinking or the exploration of history. Nor will it be sensible for thosedirectly motivated to increase profits to convey those skills that liber-ate by enabling a person to learn and think for herself. Rather, marketsuccess will be achieved and sustained by teaching quite specific skillsthat rapidly become obsolete. This will mean that a market system canbe expected to move toward specialization of the curriculum. Not onlywill specific skills be more easily standardized and tested, they will alsoprovide an education market that sustains itself as people return to beretooled to cope with the changing conditions of the job market.

2.3 Two Normative Contrasts between Public and Private†

To advance the discussion of privatization in education, I want to con-sider at a more abstract level some differences between the public andprivate approaches. On the view that I defend, exclusion and partialityare essential features of the domains we refer to as ‘private.’ This meansthat we should be particularly vigilant where social changes involve pri-vatizing what has been a public matter. It does not mean that we shouldalways prefer public to private solutions of problems. On the contrary,being exclusive and partial is sometimes a virtue. Let me explain.

2.3.1 Exclusion

A private ‘sphere’ is one in which—in some respects—a person orgroup is to be left free from interference and from public review orscrutiny. When we use the contrast between public and private nor-

†This section derives from a discussion of the public-private divide in relation to patentson living things in my ’Private Life: Reflections on Privacy, Patents and Biotechnology,’written for the Law Commission of Canada. SeeNew Perspectives on the Public—PrivateDivide (Vancouver: UBC Press, 2003), chap. 3.


matively to mark the sphere of individual autonomy, the private sphereconsists of those types of decisions and actions from which the controlof the state ought to be excluded. Likewise, the right to privacy is theright to exclude others from information about ourselves. The most ob-vious connections between privacy and exclusion, of course, are thoseinvolved in private property. Property is private when there is a rightto use and control something together with the right to exclude othersfrom its use. Private enterprise is structured around such private prop-erty rights. Apart from some rights to exclude, we would not have anynormative sense in which anything could beprivate.

2.3.2 Partiality

Private matters are also, in some respects, immune from generalizationtests that apply to public matters. Along some dimension of decision-making, it is permissible, within the various private spheres, totreat likecases differently. That is, the various private spheres involve relationsof partiality among persons.12 This needs explanation by means of ex-amples. Let me take relations of family and friendship as paradigmsof private relations. In being part of a family, or in having friends, Iam (normatively) connected with specific people in ways that I am notrelated to others. These relations between certain specific people andmyself are relations of partiality. If I treat my friends and family as ifthey mattered to me no more than anyone else, I mistreat them. I canrecognize that two children are equally needy and still be obliged totreat one preferentially, because one child is mine. I am partial to myfriends in the same way, and rightly so, according to the moral outlookthat prevails in most societies. In entering into promises and contracts, Iestablish further relations of partiality. I can no longer treat the interestsof A and B as of equal importance in all respects, if I have promisedsomething to A but not to B. Much of our normative world is comprisedof such relations of partiality. A good chair of an academic departmentwill act in a way that favours her department over others, just as a good

12See Thomas Nagel,Equality and Partiality(Oxford: Oxford University Press, 1991), esp.chap. 2.

2.3. TWO NORMATIVE CONTRASTS 33

leader of a corporation will work for its goals and those of its share-holders over the interests of others. Within the outlook that the privacyof property establishes (taken by itself), the owner of private propertyneed not concern himself with the fact that others have far greater needfor these resources than he does.

All of these partial viewpoints exist in contrast with—and often intension with—public viewpoints that demand impartiality. In some con-texts, the types of loyalties that define the relationships just consideredbecome matters of unfairness and injustice. A judge cannot take pridein giving lighter sentences to her friends. A teacher treats his studentsunfairly if he plays favourites. There is, in fact, a perspective of pub-lic reason from which all such specific relations of partiality can appearto be vices. This viewpoint has sometimes been thought to define theoutlook of reflective morality.13 But it is a mistake to suppose that par-tiality can be eliminated from a defensible reflective morality. It is truethat loyalties can easily be, and often have been, taken to excess. A nor-mative world constructed entirely from such partial relations would beone from which justice had disappeared. But the partialities involved infriendship and citizenship are not—like (say) racism—something thatwe should strive to overcome. A world of wholly utilitarian impartial-ity, in which people are rendered free from commitments to specificothers, and in which caring (in this sense) and loyalty disappear, is itselfnot a worthy ideal of human perfection. A justifiable social system willincorporate and protect both of these normative structures.

In claiming that partiality is a constitutive feature of what is private, Iam not saying that impartiality plays no role within areas of life that arepredominantly private. It has often been said that a good parent treatsher children equally. But this is not to deny that relations of partial-ity are a constitutive element of the family. My contention is that suchrelations of partiality are essentially involved in the private spheres un-

13This claim represents a tradition that includes most of moral philosophy, including philoso-phers as diverse as Plato and Hume. For a critique of the view, see, e.g., Bernard Williams’sdiscussion of integrity in ’A Critique of Utilitarianism,’ in J.J.C. Smart and B. Williams,Utilitar-ianism: For and Against(Cambridge: Cambridge University Press, 1973), and Nagel,Equalityand Partiality. The view has also been questioned by some feminists who defend an ethic ofcare, e.g., Carol Gilligan,In a Different Voice: Psychological Theory and Women’s Development(Cambridge, MA: Harvard University Press, 1982).


der discussion. On the other hand, partialities may be involved in the‘public’ spheres that comprise our social world. But partiality is not adefining feature of what is public; and it is not what we are after whenwe designate something as a matter of public interest or responsibility.14

It seems clear that we can end up with the various lines between pub-lic and private drawn in the wrong place. We are prepared to condemna totalitarian state in which there is little by way of protection for au-tonomous decision or informational privacy. The special relationshipsof partiality are an important aspect of any life worth living. But a socialworld built entirely (or predominantly) out of partialities, where thereis no public perspective from which human actions are constrained byreference to impartial standards, is also not a plausible ideal.

2.3.3 The Public-Private Divide in the ‘Real World’

In our everyday social world the public and private do not exist as neatlyisolated ‘spheres’ (despite our use of this term). The normative differ-ences between the public and private that I have described in this sec-tion are ideal types that never correspond precisely to our actual situa-tion. In the ‘real world’ the public and private exist in combination witheach other. Systems of private property and enterprise, for example, arealways dependent upon public law to provide a basis for a certain or-ganization of ownership and competition. Public services, on the otherhand, are always provided in ways that involve some dependence onprivate markets in goods and services. Moreover, public services aregenerally15 provided by decisions of government, and the government(in a democratic state) is the representative of the public interest. Butgovernments can (of course) fail in this task of representation. Govern-ments can become the representative of private corporations, for exam-

14I am also not arguing or assuming that partiality (and hence privacy) necessarily connectswith egoism, with the pursuit of self-interest. Not all forms of partiality involve the promotion ofself-interest, although, of course, this is a highly important manifestation of partiality. The kind ofpartiality that is involved in care—for example, care for the members of one’s family—can involvegreat sacrifices of the time and energy that might have been devoted to one’s own projects. Thechair of a department can impose costs on herself—for the sake of the department. In times ofconflict, the partiality of loyalty to a state can cost a person his life.

15There are also non-governmental organizations.

2.4. EDUCATION, EQUALITY, AND CHOICE 35

ple, and act in ways that are contrary to the public interest. In light ofthese features of our actual situation, one might wonder whether thereis any point in drawing the sharp contrast between public and privatethat I have elaborated. One might infer from the complex integration ofpublic and private that there is no real division between these two andthat ‘privatization’ is itself intelligible only as an abstraction.

This would be a mistake. It would be comparable to the mistakeof thinking that there is no useful division between utilitarian and Kan-tian (or respect for person) ideals, because in our world of everydaydecisions and institutions these ideals are integrated with each other incomplex ways. These different norm systems sometimes complementand sometimes combat each other; but there is no adequate understand-ing of the normative bases of our institutions that does not include bothperspectives. There is a difference between policies that move us furthertoward an entrenchment of individual rights, for example, and those thatsubordinate such rights in the pursuit of a higher level of general wel-fare. The 1982 entrenchment of a Charter of Rights in the CanadianConstitution is an example of the first sort of change (though the Char-ter itself aims to achieve a careful balance between consideration ofindividual rights and the public good). More recently, we have seen theintroduction of anti-terrorist laws in both the United States and Canadathat impose new security measures at some cost to individual rights toprivacy and freedom from search. Understanding the abstract differ-ences between these two theories is essential to understanding what isgoing on in our social world despite the fact that the real world doesnot provide examples of institutions which exemplify either in isola-tion from the other. It is in this same spirit that the normative divisionbetween public and private should be understood.

2.4 Education, Equality, and Choice

Obviously no answers to questions about privatization in education fol-low directly from these abstract differences between the public and pri-vate spheres. But these contrasts suggest importantly different visionsof education, both of which capture matters of importance to us. One


is the image of public education in a just and democratic society. Fromthis viewpoint, social justice requires that opportunities for educationbe provided on a basis that is substantively equal. Education is impor-tant to the success of the lives people live. A society that shows equalconcern and respect to its citizens will not allow wide disparities in theeducational opportunities that it provides. Those who do not know whatis going on in their world and in their society will choose from an im-poverished range of options. They will compete with others for jobs andopportunities at a disadvantage. They will not have an effective voicein the selection of leaders and policies in their community. The factthat a child comes from a poor family should not mean that she alsohas to contend with a poor school. Even if the poverty of her familyis a function of the choices and lack of effort of her parents (which itoften isn’t) this is no fault of hers, and it can provide no justificationfor thinking that she deserves a less adequate education than children ofmore well-to-do parents.

Set against this egalitarian image is a nest of intuitions about the re-sponsibility of the family in relation to education. We actually have asocial system that is not egalitarian, but dramatically hierarchal in manyrespects, in which there are better and worse jobs, houses, neighbour-hoods, universities, and so forth. Viewed from within this hierarchy, itwould be odd to suppose that, although in other respects well-to-do par-ents are free to confer advantages on their children that are unavailableto less fortunate members of the community, they can provide no advan-tages in relation to education. Education is one of the keys to successin upward mobility, and those who have the means often regard them-selves as having a right to spend their income in ways that give theirchildren an advantage.

The educational systems that have developed within liberal democ-racies are an uneasy compromise between these two ways of thinkingabout educational opportunities. State-funded public education is un-derwritten by arguments that lead in the direction of equal access andequal opportunity. But private money can often be utilized to obtainaccess to alternative superior schooling in private institutions.


2.4.1 The Demands of Justice

It will be useful to briefly contrast two extremes on this continuum ofpublic-versus-private norm systems in their application to education.On one side (call it ‘the left’) is a system from which parental discretionis removed. This could be done on the model of public health care as itis provided in many countries (including Canada)—where a ‘two-tier’system is deemed to be unjust and private medical services are effec-tively excluded by legal regulations. Applied to education, this wouldmean the elimination of private schools. The sort of social system de-fended by Marx and many modern socialists requires that educationalopportunities be allocated equally and in a way that reflects need andability. Of course, this view does not require exactly the same treatmentof every child, irrespective of their differences—any more than a pub-lic health care system requires the delivery of the same medicines ortreatments to everyone regardless of their medical condition. But a so-cialized system of education does reject the idea that opportunities opento a child should depend on the wealth or the choices of her family.

One could also generate arguments for a completely public systemfrom a liberal theory such as that of Rawls.16 A decision from behind aveil of ignorance would provide no good reasons for a caste system inwhich the resources and quality of early education available to childrendepends on the wealth and choices of their parents. As we move up fromprimary education, emerging differences in effort and talents of studentswill provide some basis for differentiation under the difference princi-ple,17 since investing more in bright and willing students will sometimesprovide benefits that can make everyone better off. But, again, there isno reason to suppose that differences selected in this way will correlatewith the wealth and choices of parents. A system of public scholarshipswould be necessitated to implement this use of the difference principle.

Of course, the degree to which public schools actually provide sub-stantively equal opportunity is limited. Often, there are quite substantialfunding differences between one school district and another (e.g., be-

16John Rawls,A Theory of Justice(Cambridge, MA: Harvard University Press, 1971), 136-42.17The difference principle permits inequalities only where they are to the advantage of the least

well-off member of society.


tween middle-class, suburban neighbourhoods and the inner-city schools,or between urban and rural schools). Recently in the United States,Fourteenth Amendment guarantees of equal treatment have provided thebasis for legal decisions against the inequalities that have often resultedfrom differences in the property tax base from which schools gener-ally have derived their funding; but other judgments, in both the UnitedStates and Canada, protect parental choices in relation to the educationof their children.18

2.4.2 Free-Market Choices

On the other side (call it ‘the right’) is the complete elimination of pub-lic schools in favour of a free market in education.19 Education, on thisview, should be opened up to a free market and treated as a product likeany other within that market. Like food and housing (which, of course,are necessary to meet basic needs), and like automobiles and computersystems (which satisfy less vital interests) under a private system, ed-ucational products would not be available at uniform price or quality.Opened to freedom of choice, the educational market can be predictedto disperse educational opportunities as widely as the market has dis-persed the quality of other products that we consume (think of the rangein quality and price of food, houses, computers, automobiles). Propo-nents of free-market education might ask, rhetorically, what crude formof materialism would permit individuals to reflect their values in thechoices they make over wines or automobiles but deny them this oppor-tunity in relation to the education of their children?

There are versions of both libertarian and conservative political moral-ities that support such an educational policy. Consider a libertarian view.If children are taken to be the property of their parents (in a relativelystrong sense), then parents are free to invest in their children—or toput their money elsewhere—just as they are free to invest in improve-

18A 1973 U.S. Supreme Court case,Rodriguez v. San Antonio Independent School District,actually held that inequalities of funding were not a violation of equal protection, since educationis not a fundamental interest subject to judicial review and states do not have to provide educationat all. But subsequent decisions in many state supreme courts have found reason to disagree andhave held that unequal funding is unconstitutional. See Lieberman,Public Education, 206.

19Lieberman defends such a proposal inPublic Education.


ments to (other!) real estate.20 This form of libertarianism attempts toarticulate a political morality within the private perspective from whichsubstantive universalization tests are excluded. (All adult human beingshave the same negative right to choose lives for themselves. But nonehas any positive right that these choices be realizable.) The libertar-ian theory takes individual choice to be its central concern, but it findsno objection to a social system that offers a wide range of choices tosome that it effectively excludes from others. In this respect, it presentsa politics of partiality. The social system that emerges from the pat-terns of individual choices is one in which the natural differences (e.g.,in talents) among human beings are likely to beamplifiedbecause thetalented will be rewarded (by the choices of others) with additional re-sources and powers. As long as this is the result of choices that adultshave actually made without coercion, there is nothing wrong (accord-ing to this view) with a system that offers widely varying educationalopportunities to children21 (including none at all).

A defender of free enterprise in education need not accept this radi-cal version of libertarianism with its counter-intuitive implications aboutchildren. The view can be defended on the basis of a conservative po-sition founded centrally on ‘family values.’ This view credits parentswith a right to control the education of their children, a right that issufficiently strong that it trumps competing social considerations. Ofcourse, such a right would be no guarantee that there are educationaloptions that match the choices that parents will want to make. But (onthis view) it is assumed that this is more likely under policies that placeeducation more fully within the economy of the free market. Familiescan be expected to want to provide their children the best education theycan afford.

20See Jan Narveson ,The Libertarian Idea(Philadelphia: Temple University Press, 1988),chap. 19, for a discussion of ‘The Problem Children,’ in which Narveson explores some of thetroubling implications of the view that children are private property.

21Narveson, chap. 20.


2.5 The Priority of Justice

By what standards are we to reconcile in a principled way this opposi-tion between individual choice (and privilege) in education versus de-mands of justice and equal treatment? This question is central to debatesover privatization. The question also presents a dilemma with whichmany of us have struggled personally, in trying to do the best that wecan for our own children while recognizing that there are issues of jus-tice at stake and that it is not possible to justify a system that privilegessome children over others. Ultimately, I believe the clash that I have de-scribed rests on differing conceptions of the relation between the rightand the good (in the sense that Rawls uses this dichotomy).22 The lib-ertarian and conservative positions described above subordinate publicmatters (at least with regard to education) to individual adult decisionsabout ‘the good,’ about what is worth doing with one’s life. The justsociety is one in which the rules and institutions provide a frameworkthat protects the rights of individuals to make these choices. This viewis combined with a thesis about the relation between parents and chil-dren. Until a child has matured to the point where she can make rationalchoices for herself, it is the parent’s right (libertarian view) or authorityand responsibility (conservative view) to make these life choices for her.

On the other hand, the egalitarian position that I have described as-sumes the priority of the right over the good. Individual freedom to actand the scope of the freedom to make private commitments that involvepartiality to certain others are always subject to prior constraints of jus-tice. Ronald Dworkin sums up the conflict that has been central to thispaper in the following way:

Notice what might seem to be a contradiction between twoethical ideals most of us embrace. The first dominates ourprivate lives. We believe that we have particular responsibil-ities toward those with whom we have special relationships:ourselves, our family, friends and colleagues ... We believethat someone who showed equal concern for all members of

22Rawls,A Theory of Justice, 446-52.

2.5. THE PRIORITY OF JUSTICE 41

his political community, in his private life would be defec-tive. The second ideal dominates our political life. The justcitizen, in his political life, insists on equal concern for all.23

The liberal answer to this conflict follows from the priority of justice:

A competent overall ethics must reconcile these two ideals.They can be reconciled adequately, however, only when pol-itics actually succeeds in distributing resources in the waythat justice requires. If a just distribution has been secured,then the resources people control are morally, as well aslegally, theirs; using them as they wish, and as special at-tachments and projects require, in no way derogates fromtheir recognizing that all citizens are entitled to a just share.24

Applied to the issues raised by privatization in education, this reso-lution of the conflict between public and private values would cut fairlydeeply into the (supposed) right of individuals to provide privileged ed-ucation for their children. It would also provide grounds to limit theways in which the market can disperse the resources of a communityand thereby lose the initially fair distribution of opportunities.

But here we should again return to the real world. We are not at alllikely ever to arrive at a new beginning in which the resources to whichindividuals have access are evenly divided. In practice we will have tolive with incremental change, and we will have to be vigilant about in-cremental changes that take us even further from conditions of justice.In devising social policy we will have to assume a wide dispersion ofwealth. It is against this background that we must consider how to struc-ture the institutions that provide educational opportunities. Privatizationwill move us further in the direction of choices that reflect private loyal-ties and interests. We have witnessed these changes in many industries,from coal mining to airlines, and we can presume that these changeswill not be reversed. Is there something special about education that

23Ronald Dworkin,Sovereign Virtue: The Theory and Practice of Equality(Cambridge, MA:Harvard University Press, 2000), 235.

24Ibid.


should lead us to set it apart from full integration with the competitivepressures of the market?

2.6 Education as a Human Need

In much of his work, David Braybrooke aims to reinstate the conceptof need in theoretical discussions of social policy.25 As Braybrookepoints out, need has never been displaced from the discourse of policy-makers. Contemporary theoretical work in economic theory, politicalphilosophy, and ethics, on the other hand, has generally displaced talkof needs by the concept of utility, and utility itself is analysed in termsof preference.26 This is unfortunate, since the resolution of many prob-lems of social policy requires a distinction between what is needed andwhat is merely preferred. We can utilize this division in arguing that amarket-driven educational system will give us the wrong resolution ofthe conflict between public and private opportunity in education.

In addition to the claim that education is one of the basic needs of hu-man beings, there are three claims defended by Braybrooke of particularrelevance to the present problem. The first is a criterion of what countsas a need. We can identify what is needed most easily by reference to itsabsence or insufficiency: failure to meet a need detracts from a person’sability to function in the social roles that are appropriate to him or heras a human being.27 The second is a normative thesis that Braybrookedenominates the ‘Priority Principle.’28 According to this principle (andcommon sense in matters of allocation) providing conditions in whichhuman beings can meet their needs takes precedence over permitting orenabling people to satisfy preferences for what is not needed. A thirdclaim concerns distributive justice: to the extent that it is practical, ajust society will aim to achieve equality in meeting needs.29

25Braybrooke,Meeting Needs(Princeton: Princeton University Press, 1987).26‘Where Does the Moral Force of the Concept of Needs Reside, and When?’ inAnalytical

Political Philosophy: From Discourse, Edification(Toronto: University of Toronto Press, 2005).27Meeting Needs, chaps 2 and 3.28Ibid., chap. 2.6, ‘A Principle of Precedence.’29Ibid., chap. 4.5, ‘Equality in Meeting Needs.’

2.6. EDUCATION AS A HUMAN NEED 43

Consider the first of these claims. In the sense at issue here, whatpeople genuinely need is what is necessary to their thriving as humanbeings; and, conversely, the absence of what is needed undermines theirability to function physically and in the social roles that make us human.This is a thesis of modern natural law theory (natural law naturalized,one might say).30 But this claim, too, captures what is common knowl-edge and common sense.31 As human beings, we need food and waterbecause without them we will cease to function at all; and with inade-quate supplies we’ll be impaired. Similarly, we need to acquire somelanguage abilities in order to function in the social roles that give usour identities as individuals and members of cultures. A lack of somelinguistic abilities—for example, the capacity to read and write—leavesa person unable to function well in the social roles required by modernsocieties. The need for education is a generalization over a wide rangeof capacities of this sort. But it also recognizes that a general level ofshared information is itself a condition of functioning adequately withina modern society such as our own.

Second, given this conception of needs, the priority principle seemsto require little by way of justification. It is more important that we re-tain the capacity to function as human beings than to get exactly whatwe want. The utility calculus of individuals will reflect this. But in ag-gregating preference satisfaction, there is no guarantee that the needs ofa few will not be outweighed by the preferences of many. That an auto-mobile factory contaminates its immediate environment, leaving thosenearby without an adequate supply of clean air, may be deemed a smallprice to pay for the greater wealth and economic opportunity that isgenerated by growth in an unregulated industry—if we are simply try-ing to maximize the satisfaction of preferences. Similarly, if we have afree market in educational services, the costs of providing for those withspecial needs will be deemed too high and the benefits to low to be jus-

30Braybrooke defends a version of natural law theory inNatural Law Modernized(Toronto:University of Toronto Press, 2001). The role of human needs in relation to what is universal inmorality is summed up at 239-41.

31Obviously, the claim does not attempt simply to capture ordinary language. We are proneto say such things as ‘I need a drink’ or ‘I need a second car.’ A view designed to explain andelucidate the contrast between really needing something and merely preferring or wanting it mustexclude such claims.


tified in terms of economic good gained by the stockholders of compet-ing private businesses. Of course, we could introduce legislation—andsome communities have done so—requiring educational corporationsto take the responsibility of meeting such needs. But this is just to saythat we could hold the line on privatization by retaining public controlover this aspect of education. Neither the pursuit of utility nor the pur-suit of profit in a free market can be expected to capture the (intuitivelyplausible) policy of giving priority to needs over mere preferences.

Finally, what justification can be given for the idea that a just societywill afford its citizens an equal opportunity to meet their needs? I willassume, as the context of this discussion, a level of affluence like thatof modern Western societies, in which there are more than sufficientresources to meet the needs of the people. It is not necessary for someto die in order that others live or for some to remain at home in order thatothers receive an education. In these circumstances the argument is verysimple. The requirement of equality in meeting needs follows directlyfrom the priority principle. If there is enough to meet needs, but theneeds of some are not met, the allocation system is permitting some tosatisfy mere preferences while others do without what they need. Givenwhat needs are, this system will treat some people as if it did not matterthat they lack the resources to lead fully human lives while others areable to enjoy luxuries. Braybrooke’s own argument against this degreeof imbalance involves showing that it constitutes ideal conditions forexploitation.32 Some will be unable to meet their needs without servingthe interests of others. Far from being consistent with social justice, asystem of affluence that does not provide equal opportunity for meetingneeds maintains the essential conditions of oppression.33

But here we must consider an objection. Food is obviously a needof human beings. It clearly does not follow directly from this that foodshould be equally distributed or that there is something wrong with thesystem in which food is made available by private enterprise and madeavailable in a way that generates a huge range of options in type, quality,

32Meeting Needs, 152-5.33Ibid., 152-5. Charity is a possible alternative to exploitation. It is not, in the circumstances

described, an escape from oppression, since it maintains the relation of dominance. For this reasonit is resented by those who are aware of the injustice of their condition.

2.7. FAIR OPPORTUNITY 45

and price. A system of private enterprise in food may, in fact, enableeveryone to meet their needs in this respect. This is not to say, of course,that our present system of private enterprise in food does adequatelymeet the needs of all human beings. Clearly it doesn’t. The point of theobjection is that, though food is a commodity that we need, justice doesnot require that it be available on an equal basis. This is because fooddoes not just meet our needs; it also satisfies a wide range of preferencesthat go beyond what is necessary to avoid starvation or incapacitation.Similarly, while education is a human need, its acquisition satisfies awide range of preferences that go beyond what is needed for one to bea fully functioning member of a modern society.

This point is obviously right. It is important to understand the impli-cations of the fact that as human beings we have a need for education. Itis also important to see why a just society will provide the opportunityfor a minimum level of education on an egalitarian basis. The aboveaccount provides some of the argument behind the ideals that led ourcommunities to a system of public education, one that is protected fromthe operation of a free market. Braybrooke’s argument also enables usto see how the absence of such a system creates the condition underwhich people can be marginalized, exploited, and oppressed. But thatthere is a need for education with these implications does not establishwhat minimum is adequate. Nor will these considerations, by them-selves, provide the basis for a thoroughgoing equality of opportunity ineducation that extends from kindergarten to university.

2.7 Fair Opportunity

Given that there is not going to be a one-time redistribution of resources(and if there were, it would soon be needed again), attempts to achieveconditions that are consistent with a liberal conception of distributivejustice will have to rely on tax policies that redistribute wealth whichhas been gained in circumstances that do not meet the conditions of jus-tice. But whether tax policies move us closer to justice, are neutral, orpromote an even greater differentiation between economic classes willdepend upon the uses to which tax revenue is put. Systems of public


welfare and medicine - as well as education—do redistribute wealth.But often these systems fall far short of achieving conditions of equalityeven with respect to needs. In many cases it would be absurd to sup-pose that they were intended to do so. But there are exceptions. WhenCanadians defend universal medicare and resist a ‘two-tier’ system (asa large majority of Canadians do) it is because they see justice as requir-ing equality in meeting needs. Perhaps because the needs that are metby education are less immediate, we do not find the same consensus thata two- (or multi-) tier educational system is marked by injustice.

In education, meeting needs is necessary but not sufficient for meet-ing the conditions required by justice. Reflective comparison of publicinstitutions of welfare and medicine with education will reveal that theformer do not carry the same potential to move those who benefit fromthem toward a more fair or just distribution of resources. A safety net isjust that—a device that (when it functions successfully) prevents peo-ple’s situations from becoming dramatically worse. But lives of meresubsistence, in which one does not perish or suffer incapacitation fromlack of food or insulin, should not be the lot of any members of an af-fluent community—as Braybrooke himself would be the first to insist.Given what we know about human motivation, however, a system thatprovided affluence irrespective of a person’s contribution to the commu-nity would not work. Even if it did work, it would not be fair to thosewho are productive contributors. So, differences in levels of wealthcan be expected even in a just community. But this does not mean, ofcourse, that the opportunities available to individuals in a just societyshould themselves be quite uneven so that some struggle to gain even asmall portion of what others have handed to them.

What is needed is a system that moves us toward equality of op-portunity in which differences in power and compensation are not sim-ply reflections and amplifications of an arbitrary distribution of wealth.Here public education could play a centrally important role. Given theimportance of training and knowledge in gaining access to careers, thisaccess should not be limited by factors that have nothing to do with aperson’s effort or choices. In the real world—as opposed to that of theo-ries in which we can contemplate a total reallocation of wealth—public

2.8. TEACHING CLASS 47

education is one of the few institutions that has the potential to move ustoward conditions of distributive justice. Arguably, no other institutionhas the same potential to redistribute resources in a way that achievessome main aspects34 of the ideal of distributive justice that liberal theo-rists such as Rawls and Dworkin have defended.

At the centre of arguments for privatization in education, of course,is a rejection of this liberal ideal. But (to translate this into one of themetaphors of advocates of free enterprise) it is hard to see how onecould defend as fair a race in which some participants are allowed tostart well ahead of others. Pessimism will lead some to view this sit-uation as inevitable. Self-interest may lead those who are well placedto defend their advantage (perhaps by embracing this pessimism). Butwhat excuse shall we find for those philosophers who manage to con-vince themselves that it is a matter of entitlement that from the begin-ning some individuals have opportunities for self-improvement that oth-ers are denied?35

2.8 Teaching Class

If the argument of this paper is right, the normative division betweenproponents and opponents of privatization is a deep one. It is an illusionto suppose that this is really a debate over the most efficient means ofpromoting goals that are held in common. But where does this leave us?Is there some common ground on which to stand in order to decide (orsplit the differences) between these two views?

According to Thomas Nagel inEquality and Partiality, there are nogrounds within contemporary morality to resolve this conflict in a prin-

34A system in which people have an equal opportunity to compete—e.g., for marks and hencefor positions in higher education—does not solve the problem that human capacities are them-selves unevenly distributed. The reduction of that form of arbitrariness—to the limited extent thatit is possible—is dependent on need-based policies.

35Here I am appropriating the cadence and spirit of David Hume’s critique of the ancientphilosophers, whose metaphysical views (e.g., of nature’s horror of a vacuum) he deemed to bethe products of overactive imaginations. It ends: ‘We must pardon children, because of their age,poets because they profess to follow implicitly the suggestions of their fancy; But what excuseshall we find to justify our philosophers in so signal a weakness?’ SeeA Treatise of HumanNature, book 1, part 4, section 3.


cipled way.36 I am not convinced that this sceptical thesis is right. But Ido see the difficulty involved in finding any resolution that does not begcentral questions. My own view accords with that of Dworkin (above),which accords with the views of Kant and Rawls and Braybrooke—thatthe impartial demands of justice must set limits to our freedom to pur-sue our own goals in ways that are exclusive and partial. Moreover, Ibelieve that it is a mistake to defend privatization on the basis of greater‘choice’ in education, where the choices of some will be options thatthey can’t refuse, for services they would not otherwise accept. I amaware, however, that this objection relies on an appeal to an impartialstandard, which proponents of privatization may refuse to concede.

On the other hand, I am confident that in the social world to whichour theories must be applied, we will continue to find both a sense ofjustice and an acknowledgment that certain basic choices cannot rightlybe dictated by the state. Some compromise will have to be reached be-tween the sense that justice rules out inequality in education and theview that parents should be free to choose, within their means, the ed-ucation that their own children will receive. Most reflective people willfind that they cannot avoid either of these perspectives, and, therefore,they will be willing to accept ways of integrating them. The argumentsof this paper aim to reveal the ways in which moves toward privatizationbring with them important changes that conflict with the sense of justicethat lies in the background of our public systems of education.

Our theories must also face the reality of imperfection in the socialcircumstances to which we apply them. We know that governments areoften very imperfect representatives of an impartially conceived publicinterest. We know that some public schools will be good and others willnot, and that some public schools are sufficiently bad that parents willbe justified in taking their children out of them, if they can. But there areways of cheating public schools that will continue to offend our senseof justice—if we are not duped by them. One of these ways of cheatingis to impoverish public schools by cutting their resources and then usingtheir faults as enticements to private alternatives. Another is to utilizetax credits for those who pay for private education as ways of removing

36Nagel,Equality and Partiality, 3-9.

2.8. TEACHING CLASS 49

funding from public schools so that public schools are left with fewer re-sources and required to teach those with special needs whom the privateschools will not accept. These are familiar strategies of conservativegovernments committed to privatization.

But (you may object) what if these alternative schools actually doa better job? My answer to this question is given above. If those ar-guments are sound, they show that privatized schools will not be doingthesamejob. They will be leading us away from the ideal of equalitythat is a necessary condition of justice and democracy. In amplifyingthe differences among us, they will be teaching class.

Chapter 3

Determining Health Care NeedsAfter the Human Genome Project:Reflections on Genetic Tests forBreast Cancer

SUSAN SHERWIN

Abstract

This chapter appeals to Braybrooke’s conceptual work on needs and triesto pick up where he leaves off at the end ofMeeting Needswhen he throwsup his hands in an honest admission that his theory cannot deal with the‘bottomless pit of medical needs.’ I explore how the new genetics is ex-panding the scope of perceived medical needs and look particularly at thegenetic test for hereditary breast cancer now on the market. Building onBraybrooke’s category of course-of-life needs, this chapter proposes de-veloping a relational understanding of persons and of needs as a way toavoid the deeper bottomless pit of medical need attached to the availabil-ity of genetic testing. If, as Braybrooke advises, we are careful in ourapplication of the concept of ‘need,’ the pit takes on a different structureand requires different strategies than is commonly assumed.∗

∗I had the opportunity to present earlier drafts of this paper at the University of BritishColumbia (W. Maurice Young Centre for Applied Ethics), the University of Toronto (Joint Centrefor Bioethics), and Dalhousie University (Philosophy Department). I am very grateful for all thethoughtful feedback I received on these occasions. I am especially grateful for the careful attentionprovided by Lori d’Agincourt-Canning, Jeffrey Nisker, and Christy Simspon.

51

52 CHAPTER 3. SUSAN SHERWIN

3.1 Introduction

In his important bookMeeting Needs, David Braybrooke (1987) restoresthe concept of human needs to a central place in policy considerations.He is responding to a widespread tendency among modern economistsand liberal political theorists to substitute individual preferences for thetraditional role assigned the concept of needs in social policy deliber-ations. Braybrooke seeks to stem this tide and restore the concept ofneeds to its rightful place where it can provide ‘firm guidance in thechoice of social policies’ (5). He argues that without a clear under-standing of what are legitimately classified as needs, it is difficult toknow how to organize government priorities and how to evaluate spe-cific policies according to norms of justice and beneficence. The aim ofhis book is to develop and defend a substantive, and usable, account ofthe concept of needs.

Rather than producing an analytic definition of needs in terms of aset of necessary and sufficient conditions, Braybrooke offers an accountof how the concept can (and should) be used in the practical evalua-tion of social policies. He finds it necessary to distinguish between twotypes of needs: course-of-life needs and adventitious needs. The formerdeals with the needs associated with normal functioning in society andthe latter addresses conditional needs that derive from specific projects.Thus, course-of-life needs are common to all and adventitious needs re-flect personal preferences. He believes that social policies should try toensure satisfaction of both types of needs, but, in most circumstances,priority should be granted to course-of-life needs.

This comprehensive and persuasive account of the concept of humanneeds makes a very important contribution to the role of ethics in pub-lic policy. Unfortunately, as Braybrooke is well aware, his insightfulaccount faces serious limitations. At the conclusion of the book, withcharacteristic honesty and humility, he cites two problems he finds in-tractable. In each case, the concept of needs ‘breaks down’ in the faceof needs so enormous that they threaten to drain all public resources.The first area concerns efforts to apply a single, normative concept ofneeds to the global community because the scale of scarcity and human


suffering shifts dramatically once we leave the borders of affluent indus-trialized nations. The second problem area is that of medical services,particularly those associated with expensive, life-prolonging care forthe critically ill: providing the full range of medical services to every-one whose life might be extended by such services might easily exhaustall public resources.

I should, probably, follow the wisdom of Braybrooke’s leadershipand join him in throwing up my hands at the problem of applying hissubstantive conception of human needs to the messy terrain of medicine,but this does not seem to be an option for those of us who work squarelyin the area of health care ethics. According to the crumbling (but stillstanding) pillars of the Canada Health Act, Canadians are entitled topublicly funded medical and hospital treatments that are judged to bemedically necessary. Provinces and territories, responsible for imple-menting the unwieldy requirements of this Act, face the daunting taskof determining which services fall under the category of ‘medical ne-cessity’ or ‘medical need’ and of deciding what level of service countsas discharging this duty. These are moral, as well as political and eco-nomic, decisions, and a conception of human need that makes sense inthe context of high-tech medical care is an essential element for ethicalpublic policy in this realm.

In the mid-1980s, when Braybrooke was writingMeeting Needs,the type of medical care that worried him was the sort associated withresource-intensive, life-saving treatments: organ transplants, prolongedtime in the ICU, long-term dependence on expensive technological sup-ports, and the nursing care required to maintain such efforts - that is, thevarious sorts of life-extension opportunities medical science has beengenerating over the last half-century. As if these problems were notbad enough, we now find ourselves facing many new demands on ourhealth care system. One large set of these is associated with the grow-ing knowledge that arises from the Human Genome Project and the re-lated promise that access to genetic knowledge may sometimes lead toactions that could significantly increase life expectancy for particularindividuals. This is a more indirect approach to life extension than theinterventions Braybrooke found overwhelming in the 1980s, but for the


people in question, the opportunity to avoid or delay life-threateningillness may seem every bit as necessary as traditional life-prolongingmedical interventions. Arguably, these needs are even more importantto address, since if it is possible to avoid critical illness altogether, notonly will the lives of affected individuals be extended, but their healthwill also be improved (unlike many in critical stages of illness who canhope only for extended quantity of rather poor-quality lives).

My aim in this chapter is to reflect on how, as a society still (barely)committed to a publicly funded health care system that is expected toprovide people with the health care they need, Canadians should reflecton the added complexities that emerge from the potential explosion ofknowledge about individuals’ genetic disposition to develop particulardiseases. I shall focus, in particular, on the case of genetic testing formutations associated with some forms of hereditary breast cancer asan illustration of how to think about the connection between geneticknowledge and the concept of needs. I assume that we may soon seemany more options of genetic tests for susceptibility to other adult-onsetdiseases; therefore, it is important to understand how to approach anysuch test within a conceptual framework of medical needs. As well,I hope that exploring the question of medical needs in the context ofgenetic testing will bring us closer to addressing Braybrooke’s initialworries about how to approach other forms of medical technology.

I shall argue that in order for an account of human needs to playa meaningful role in public policy deliberations it must support an un-derstanding of human beings as relational entities rather than the morefamiliar conception of persons as distinct, self-contained beings. First,I shall explain how the problem that confronted Braybrooke’s specificaccount of human needs in 1987 could now be significantly more threat-ening as we move rapidly into the era of individualized genetic knowl-edge, and then I shall show why I believe that the resolution of thisconundrum rests with a reinterpretation of the concept of the individualand, correspondingly, of the needs we should perceive as attached tosuch an entity.

3.2. THE WORLD AFTER THE HUMAN GENOME PROJECT 55

3.2 The World after the Human Genome Project

The Human Genome Project (HGP) is now complete. From the begin-ning, its promoters have promised that the genetic knowledge revealedby this massive research program would lead to an array of therapeu-tic interventions designed to correct genetic anomalies associated withserious illnesses. To this point, that promise remains unfulfilled; no sig-nificant cures have yet emerged from the genetic knowledge generatedfrom the HGP, and most experts believe that any such innovations are,at best, decades away (Kitcher 2001). If and when such therapies doappear, we may have to revisit the question of medical need yet again indeciding which of the possible genetic ‘corrections’ should be classifiedas medically necessary. For now, however, there is more than enoughto concern ethicists if we limit our focus to the immediate impact of theHGP. So far, its principal fruit is the ability to develop tests for geneticsusceptibility to a small but growing number of particular diseases.

Heralding this new era of genetic knowledge is the BRCAnalysis testdeveloped and marketed by Myriad Genetics Inc. to identify many ofthe thousands of mutations that occur at BRCA1 or BRCA2 that havebeen associated with some forms of breast and ovarian cancer. BRC-Analysis is one version of a number of tests that have been developed todetect mutations at BRCA1 or 2. (For simplicity, I shall speak generi-cally of ‘BRCA testing’ to refer to all of the genetic tests designed to de-tect the many types of identified mutations at either BRCA1 or BRCA2and I shall speak of those contemplating testing as ‘women,’ althoughmen also carry the BRCA genes and those who do face elevated risk ofdeveloping cancers and of passing on the genetic mutations to their chil-dren.) BRCA testing, and especially Myriad’s version of BRCAnalysis,is particularly significant since it is one of the first genetic dispositiontests to reach the marketplace. As such, it is the forerunner of manyother genetic tests that will likely come on stream if the corporate prof-its and/or social benefits of this test prove encouraging. As researchprograms pursuing further genetic tests proliferate, it is essential thatwe pay careful attention to the question of what should be done with thehost of genetic secrets that will soon be revealed.


Although genetic testing has been around for some time, BRCA test-ing represents a major shift. Where earlier tests were primarily used onfetuses (or on potential parents worried about the possibility of geneticdisease in future children), this test is targeted at adults regarding theirown future health status. As well, where some earlier tests did investi-gate future illness in adults, such as Huntington’s disease, they tendedto be limited to a small, identifiable group of people at risk; breast can-cer, in contrast, threatens more than half the population. With a one innine incidence rate, breast cancer affects many Canadian families, andpopular media have ensured widespread awareness of the threat of thedisease. A genetic test associated with such a widely distributed andhighly feared disease is potentially of interest to a very large segmentof the population (even though the test is meaningful only for thosewomen who come from high-risk families—that is, less than 1 per centof women).1 Myriad’s decisions to aggressively market its version ofthe BRCA test and to try to enforce its patent right internationally alsosignify important changes to the ways in which genetic tests may beaffected by corporate agendas.

Because BRCAnalysis raises so many serious ethical questions thatwill surely be compounded when it is joined by other genetic suscepti-bility tests, I propose to explore some of these questions as they relate tothe place of dispositional genetic tests within the complex array of med-ical services that should be understood as meeting significant needs. Ifwe do determine that BRCA testing meets a medical need that shouldbe addressed within a public (or private, for that matter) health sys-tem, we must be prepared for the possible arrival of a host of additionalgenetic tests for susceptibility to adult-onset diseases. Together, suchtests could easily swamp an already leaky health care system and gen-erate enormous confusion among a population increasingly encouragedto look for single causes and straightforward, technology-based solu-tions to complex, multi-factorial diseases. It is, therefore, urgent thatwe seek to identify and resolve the complex question of whether or not

1Those who might carry the BRCA mutation are thought to come from families with a historyof two or more closely related family members with the same or related cancers, a pattern of cancerin two or more generations, and cancers diagnosed at an early age.

3.3. THE CASE FOR TREATING BRCA TESTING AS A NEED 57

to view such tests as meeting an important need before we go muchfurther down the path of developing additional genetic tests.

3.3 The Case for Treating BRCA Testing as a Need

Breast cancer is indeed a serious disease: all too often, it results in se-vere illness and death. At present, statistics indicate that one in nineCanadian women will develop breast cancer over the course of theirlives—a 13 per cent lifetime risk (www.cancer.ca). It is worth notingthat many of these cases involve older women (over 60); however, thisdisease is no respecter of age. Young women also contract breast cancer:it is the leading form of cancer affecting women under 45 (accountingfor 33 per cent of the cancers affecting this group) and the major causeof death among women aged 15 to 40 (www.youngsurvival.org ).Moreover, it appears that when breast cancer does appear in youngwomen, it tends to be more aggressive and results in lower survival ratesthan when it is first detected in older women (www.youngsurvival.org).

Although breast cancer is a threat to all women, it has long beenknown that some women face especially high rates of developing breastcancer. As Pat Kaufert (2003) observed, well before the development ofmedical genetics, women who grew up in families with frequent casesof breast cancer understood that they faced a high personal risk of de-veloping the disease. In the early 1990s, researchers determined thatmutations in the BRCA1 and BRCA2 genes were associated with someforms of hereditary breast cancer or ovarian cancer. (It is estimated that5 to 10 percent of women with breast cancer have a hereditary form,and a significant proportion of these have been tied to an identifiableBRCA mutation.) On the basis of knowledge about the role of BRCAmutations, BRCAnalysis and other forms of BRCA tests have been de-veloped to allow some women from families with high rates of thesecancers to determine whether they carried one of these genetic muta-tions, though tests do not yet exist for all apparently hereditary forms ofbreast cancer. Women who discover that they carry a detectable BRCAmutation will know that they face a lifetime risk of developing breastcancer that is at least one in three and possibly much higher. They will


also learn that they have a 50 per cent chance of passing on this muta-tion to their children. And if they do develop breast or ovarian cancer,they face a significant risk of dying an early and difficult death.

It is not hard to see, then, why women who come from families withhigh rates of breast cancer might want to know their personal BRCAstatus. Those who undergo testing face one of three possible outcomes:a positive result indicating that they have the specific BRCA mutationthat has affected family members; a definitive negative result indicatingthat they do not have the BRCA mutation found in their close relatives;or an indeterminate result that could mean either they do not have a ge-netic risk of breast cancer or that there is no definitive test for the geneticanomaly that affects them (Sutcliffe et al., 2001). If they are fortunateenough to have a definitive negative result, they will likely be relievedof an enormous emotional burden. They may feel freer to have childrenof their own, and they may have an easier time getting employment,health and life insurance, and perhaps even a spouse. Those who testpositive have available to them an option that could make a significantdifference to their life expectancy: if they decide to undergo prophy-lactic surgery to remove both breasts and both ovaries before any ofthese organs becomes diseased they will be able to dramatically reduce(though not eliminate) their likelihood of developing breast or ovariancancer (Eisen and Weber, 2001; Meijers-Heijboer et al., 2001). It is lessclear how those with indeterminate results will respond.

If we turn to Braybrooke’s account of needs, we can make a case forthe claim that access to BRCA testing should be classified as meeting acourse-of-life need and, therefore, that society is responsible to providefor this option. Braybrooke’s account is explicitly social and situated.He does not try to derive an exhaustive, timeless, universal set of humanneeds that every society is responsible for meeting. Rather, he situatesneeds within the context of a particular linguistic community and tries tocapture the sorts of things members of that community recognize someobligation towards addressing. Hence, course-of-life needs are the sortsof needs that members of an existing society (or linguistic community)recognize as important to their pursuit of almost any project one mightchoose in that society. They are the sorts of conditions necessary for

3.3. THE CASE FOR TREATING BRCA TESTING AS A NEED 59

living a life that members of that community recognize as characteristi-cally human. In the most straightforward sense, this includes the abilityto continue to live and function normally. Thus,

Being essential to living or to functioning normally may betaken as a criterion for being a basic need. Questions aboutwhether needs are genuine, or well-founded, come to the endof the line when the needs have been connected withlife orhealth. (31; emphasis added)2

Since being connected to life or health generates a course-of-lifeneed, we can make the case that, at least for women from high-riskfamilies, BRCA testing satisfies such a need, since those who test pos-itive may seek medical interventions that will dramatically reduce theirrisk of contracting breast or ovarian cancer. In this way, we can connectBRCA testing quite explicitly with the core intuition of course-of-lifeneeds.

Further, Braybrooke extends the notion of course-of-life needs be-yond mere survival. Like many others, he recognizes that human beingsare essentially social creatures and we need to attend to the conditionsunder which most people live and the contexts in which they satisfy theirneeds. Human existence involves some core relationships that help tostructure our lives and shape the ways we seek to satisfy core needs.Braybrooke identifies four ‘basic social roles’ that he claims are suffi-ciently generic and central to the majority of adults in modern societyto be considered as constituting the sites for addressing course-of-lifeneeds: parent, householder, worker, and citizen (48).

Braybrooke reviews some of the many lists others have offered ofbasic human needs and seeks areas of consensus. Rather than chooseany single account, or offer any particular definition (set out by a list ofnecessary and sufficient conditions), he offers us a criterion for inclu-sion that captures the intuitions underlying the family of lists of human

2He then goes on to offer a rather puzzling but intriguing comment in the context of a discus-sion of breast cancer. He claims, ’Indeed they may already have come to the end of the line [forjustifying something as a course-of-life need] when the needs have been connected with keepingimportant parts of the body’ (31).


needs. ‘The criterion,’ as he calls it, for a course-of-life need is ‘beingindispensable to mind or body in performing the tasks assigned a givenperson under a combination of basic social roles, namely the roles ofparent, householder, worker, and citizen’ (48). The criterion classifiesas needs the things necessary to allow people to function ‘without de-rangement in the[se] tasks and roles’ (49). He believes it is a questionof fact whether or not something is a course-of-life need:

Something is a Matter of Need, by the Criterion, if thereare good grounds in causal evidence for presuming that anymember of the Reference Population would be unable tocarry out the tasks assigned her by one or more of the rolesmentioned if the matter in question were not provided for atthe Minimum Standard appropriate to that person’s physique,temperament, and circumstances. (49)

Arguably, women at high risk of having BRCA mutations may feelthat their ability to function responsibly in some core social roles, es-pecially that of parent, may be facilitated by access to knowledge oftheir BRCA status. This knowledge may influence their willingness tobear children genetically related to them (since they have a 50 per centchance of passing on the mutation to their children) and their franknessin discussing the subject with any children they do bear. And for thosefortunate women for whom the test provides a definitive negative re-sult, such tests may also facilitate their ability to qualify for health andlife insurance and, thereby, for employment in so far as these things arerelated (as they are in the United States).

It would seem, then, that according to Braybrooke’s account of course-of-life needs the case can be made for establishing access to BRCAtesting for women who are at significant risk of carrying BRCA mu-tations on two grounds: both for the contribution such knowledge canmake to their prospects for life and health and also for the ways suchknowledge may facilitate their ability to function in core social roles. Ifwe decide that this type of genetic test represents a course-of-life need,then we have a social obligation to ensure that our policies reflect itsimportance. That means we should try to provide the means for satisfy-

3.4. SOME COMPLICATING FEATURES OF BRCA TESTING 61

ing this need for all who are likely to benefit significantly from the test.Similar arguments would have to be considered for any other geneticsusceptibility test that is developed.

Let us turn now to some of the complexities of BRCA testing andexplore reasons (beyond the potentially high financial costs) why wemight want to resist the implication that such tests be seen as falling inthe category of a course-of-life need.

3.4 Some Complicating Features of BRCA Testing

The case in favour of treating access to BRCA testing as a course-of-life need was based on the potential of allowing women who know theirBRCA status to take steps to reduce their risks of contracting breast can-cer if they test positive or more confidently pursuing some of their coresocial roles if they do not. The situation is complicated, however, as somany things in medicine are, by the fact that there is plenty of uncer-tainty attached to the actual impact of a BRCA mutation in a particularwoman’s life. Women who undergo testing must come to grips with thefact that they may receive misleading information. Like other medicaltests, BRCA tests are vulnerable to the possibility of inaccuracy: theymay provide false positives or false negatives. In this case, the costsof the test, the rush to market, and the significant emotional burden at-tached to such testing have created a situation in which there has notbeen enough experience with the test or enough research to indicate theprecise degree of accuracy of the various BRCA tests (Kaufert, 2003).

Moreover, they face the daunting task of making sense of the prob-abilistic information that derives from identifying such a mutation. Un-like purely predictive tests for single-gene disorders such as Hunting-ton’s disease, there is no perfect correlation between breast cancer andBRCA mutations. Even women who test definitively negative are notassured that they will not develop breast cancer, only that they will notface a higher rate of risk than other Canadian women (that is, a ‘normal’rate of one in nine, or perhaps one in ten when we remove statistics ofwomen carrying BRCA mutations).3 Breast cancer is multifactorial,

3From a medical perspective, the absence of a BRCA mutation does not confer the absence


and genetic mutations are just a small part of the causal story (less than10 per cent), though it is a very important part for the minority of womenwith family histories of BRCA mutations.

For those who test positive, there is no way of determining theirpersonal level of risk because gene penetrance (that is, the percentage ofindividuals affected who will go on to develop the disease) varies widelywith particular mutations (ranging from 36 to 85 per cent) in ways thatare not easily predicted. For nearly all of the thousands of variations ofBRCA1 or 2 mutations, it is impossible to make accurate predictions ofpersonal cancer risk.4 In addition to the mysteries of penetrance, there isthe further complexity that these numbers speak of lifetime risk; thereis a set of probabilities to interpret to estimate whether the particulargenetic variation will affect particular women in their youth, middle, orold age.5

Adding to the difficulties of determining the value of BRCA test-ing is the fact that there is no particularly satisfactory response avail-able in the face of a positive test. The options are primarily those ofsurveillance, chemoprevention, or surgery. Surveillance can be pursuedthrough breast self-examination (BSE) along with frequent mammo-grams. Besides the fact that neither technique actually prevents breastcancer, both face other obstacles: BSE has been shown to be not particu-larly effective as an early detection strategy and there are risks attachedto performing frequent mammograms on young women, especially inlight of the fact that there is speculation that BRCA-associated breastcancer may be radiosensitive. Chemoprevention - that is, taking to-moxifen or other powerful drugs—is unproven with BRCA-associatedbreast cancer and carries high risks of iatrogenic illnesses (current data

of risk. If a person tests negative, it could be because the mutation is not present, but it could alsobe because the person has a mutation that cannot be detected by current technology or becausethere is a mutation in another gene not yet identified. So negative BRCA tests are meaningful onlywhen the specific BRCA mutation has been identified for a high-risk family. Hence, most negativetests are considered ’indeterminate’ and the person is still considered at risk because of the familyhistory (Sutcliffe, 2001).

4In fact, penetrance levels are so uneven with BRCA mutations that the NIH warns that ‘Thepenetrance of BRCA1 and BRCA2 mutations has yet to be determined for other than mutationscommonly found in Ashkenazi Jews or Icelanders.’www.cancer.gov/cancerinfo .

5Statistically, however, women with BRCA1/2 mutations develop breast cancer at an earlierage than other women.

3.4. SOME COMPLICATING FEATURES OF BRCA TESTING 63

suggest it is of no benefit to women with BRCA1 mutations, though itmay be helpful to women with BRCA2 mutations) (King et al., 2001).The most effective preventative strategy is surgical intervention, requir-ing bilateral prophylactic mastectomy as well as removal of the ovaries;needless to say, this is a drastic intervention with serious physical andpsychological repercussions. Moreover, even this major interventiondoes not provide complete protection against breast cancer.

Women who receive testing face further difficulties in deciding whatto do with the results, especially if they are positive or ambiguous. Withwhom should they share the information and from whom should theywithhold it? What should they tell their daughters (and sons) and when?There are no long-term studies evaluating the medical and psycholog-ical impact of test results or of the various interventions that followpositive (and negative) test results, though existing data do indicate thatwomen who undergo testing are affected more deeply than they antici-pated (Bish et al., 2002; Dorval et al., 2000).

The data that do exist are disturbing. Consider the fact that the mosteffective preventative option available to those with BRCA mutations isradical surgery. A 1997 study in theNew England Journal of Medicinereports that, ‘on average, 30-year-old women who carry BRCA1 orBRCA2 mutations gain from 2.9 to 5.3 years of life expectancy fromprophylactic mastectomy and from 0.3 to 1.7 years of life expectancyfrom prophylactic oophorectomy’ (Schrag et al., 1997).6 While expertsconsider this a good outcome, many women clearly expect more fromthe decision to undergo radical surgery. One study found that ‘at-riskwomen showed a 70 per cent time-tradeoff value, indicating that thewomen were willing to sacrifice 30 per cent of life expectancy in orderto avoid prophylactic mastectomy’ (Unic et al., 1998) This may explainwhy a 1995 study found that as few as 10 per cent of women at high

6Another study found that ’for a 30-year-old woman, according to her cancer risks, prophy-lactic oophorectomy improved survival by 0.4 to 2.6 years; mastectomy, by 2.8 to 3.4 years; andmastectomy and oophorectomy, by 3.3 to 6.0 years over surveillance. The QALYs saved were 0.5for oophorectomy and 1.9 for the combined procedures in the high-risk model. Prophylactic surg-eries were cost-effective compared with surveillance for years of life saved, but not for QALYs.CONCLUSION: Among women who test positive for a BRCA1 or BRCA2 gene mutation, pro-phylactic surgery at a young age substantially improves survival, but unless genetic risk of canceris high, provides no benefit for quality of life’ (Grann et al., 1998)


risk of developing breast cancer choose prophylactic mastectomy (Ste-fanek et al., 1995). It is worth noting, however, that more recent studiesthat take into account our ability to identify many BRCA1 and BRCA2mutations are giving a different picture. For example, 2001 studies thatspecifically address the effectiveness of prophylactic surgery for thosewith identified BRCA1/2 mutations indicate there is definite benefit at-tached to undergoing surgery; unfortunately, these studies are limited toa three-year follow-up period but most women seek longer-term infor-mation (Meijers-Heijboer et al., 2001; Eisen and Weber, 2001).

So, here we have a technology that is promoted as serving the func-tion of a course-of-life need, yet we find many women uncertain whetherto even avail themselves of the test. In fact, Myriad Genetics, the com-pany that patented the BRCAnalysis test, has been surprised to find thatwomen have not been demanding testing in the large numbers antici-pated despite an active direct-to-consumer advertising campaign. If somany of the women who might be BRCA-positive are resisting the test,should it really be classified as a course-of-life need? It is worth notinghere that Braybrooke does not require that every person avail themselvesof all of the opportunities classified as matters of need, even course-of-life need; there are often good reasons why individuals choose not toclaim something that falls under the category of need. So the fact thatmany women who are at particular risk of BRCA mutations choose notto undergo testing is not necessarily proof that genetic testing shouldnot be classified as a course-of-life need. It should, however, give uspause: the ambiguity in determining the specific harms and benefits oftesting for particular women make it difficult to situate BRCA testing inthe realm of course-of-life needs.7

7Perhaps, in light of the uncertainty of outcome and the weighty combination of benefits andburdens attached to testing, we should, for the present at least, treat such tests as responding toadventitious needs, according to how they fit within particular women’s values and life plans.Adventitious needs are captured by a relational formula of the form: x needs y in order to do z;that is, they are relative to the particular ends an agent has adopted, and, as a result, they functionlargely like preferences. The urgency of an agent’s need for y to accomplish z depends on theimportance of z to that agent. Given the wide variety in goals individuals set for themselves, andthe wide variety of responses to those goals adopted by their fellow citizens, Braybrooke quitewisely avoids recommending making satisfaction of the large range of adventitious needs a moralrequirement of social policy.

3.5. TOWARDS A RELATIONAL UNDERSTANDING OF NEEDS 65

3.5 Towards a Relational Understanding of Needs

We can find helpful direction in Braybrooke’s account as to how to orga-nize our social policy under these sorts of conditions of uncertainty andanxiety—that is, are Canadian provinces and territories morally obli-gated to provide genetic testing and related medical services for BRCAand similar anomalies under the claim that such services are ‘medicallynecessary’? Braybrooke makes clear that we cannot restrict core needsto those required for bare survival; many important course-of-life needsare intimately connected to the societies in which they occur. As notedabove, he argues that we must make room for meeting needs that arisein the context of playing core social roles. In other words, some needs,even course-of-life needs, are inherently relational in that they arise inthe context of pursuing relationships with others.

I must disagree with him, however, on the proposal that we can limitthe relevant roles and relationships to the basic set of parent, house-holder, worker, and citizen. I do not believe that we can capture all themorally significant social dimensions of human life within this smallset of social roles. For one thing, this list refers only to roles held by‘responsible’ adults and thereby privileges the socially generated needsof this group while ignoring many other aspects of a richly textured re-lational life. Further, his account treats these generic roles as if they ap-plied comparably to all social groups without attention to gender, race,class, or other important dimensions of the lives of persons. I propose,therefore, to build on the leadership Braybrooke has provided by tyingcourse-of-life needs to social roles and also to expand on his conceptionof relational significance.

In my own work, I seek to make clear the importance for policy-making of replacing the reigning liberal conceptions of persons as ra-tional, independent beings with an understanding of persons as thor-oughly relational beings (Sherwin, 2003). While Braybrooke offers usa significantly more socially complex moral subject than most liberaltheorists, he still thinks, ultimately, that our unit of analysis should beunderstood as generic persons with common characteristics. I believethat this picture needs feminist re-visioning—supplanting his partially


relational conception of persons with a fully relational account.The relational conception of persons that emerges from feminist

theory portrays a subject who is (at least partially) constituted by so-cial interactions. Rather than understanding persons as ontologicallyprior to society, a relational account envisions persons as beings whoare created—in large part—through their social relations.8 AlthoughBraybrooke distances himself from the traditional liberal view of per-sons as transparent, rational deliberators with well-ordered preferences,he promotes in its stead a view of persons as the locus for a set of needsand personal projects, many of which can be discussed abstractly ingeneric terms. Where traditional liberal theorists encourage us to re-spect the existing values of individuals, Braybrooke assigns such in-dividual variation to the sphere of adventitious needs (or preferences)and assigns priority to common course-of-life needs (though even herethere is room for preferences in how those basic needs are satisfied).Like other liberal theorists, however, he seems to assume that personalvalues are either apparent or readily discovered by the individual whoholds them through a process of introspection. In contrast, relationaltheorists believe it significant that most people come to develop manyof their values in the social process of dialogue and action. In otherwords, personal values are formed through social engagement, not priorto such interaction. For this reason, it is essential that we pay atten-tion to the circumstances in which these values are forged. If injusticestructures the conditions under which they are generated, then policiesthat simply respond to those values are likely to at least perpetuate, andperhaps even exacerbate, the original injustice.

Needs, like values, emerge from particular social circumstances. Al-though some needs deal with basic animal survival, the effectiveness ofa society’s infrastructure may determine whether or not these crucialneeds even become visible (for example, the need for non-toxic air andwater). Further, Braybrooke rightly recognizes that the social needs thatare attached to core social roles must be seen as developing in partic-

8As Annette Baier (1985) explained, persons are always ‘second persons,’ created throughsocial processes; we become persons through learning from other persons how to be persons our-selves.

3.5. TOWARDS A RELATIONAL UNDERSTANDING OF NEEDS 67

ular social contexts. In addition, he notes that some needs are derivedfrom scientific discoveries that support practices capable of sustainingcourse-of-life needs; medical care seems clearly to fall under this cat-egory (83). It is important, then, that we examine the contexts of sci-entific discovery that generate specific course-of-life needs if we are toresolve the difficult sorts of questions associated with determining med-ical needs. For example, in the case at hand, it is worth reflecting on theprocesses that have led us to the current levels of genetic knowledge andthose that have led us away from other types of approaches to breast can-cer and other diseases (as well as those that generate technology-ladenservices to extend life for the critically ill and ignore such factors as thevalue of literacy in increasing life expectancy).

There are additional insights feminists can build on within Bray-brooke’s theory. He offers explicit recognition that persons are essen-tially embodied beings. Feminist relational accounts add to this insightthe fact that many of the bodily characteristics of persons carry impor-tant social meanings. Within contemporary Western society, social andpolitical significance is attached to such features as sex, race, disability,sexuality, and age. These features are the basis of social groups that ex-ist in relations of oppression or dominance with other social groups, anda person’s classification within the complex array of social groups willaffect her or his life experiences. In other words, relational persons areinseparable from their particular historical, social, and political circum-stances; they cannot be properly understood in abstraction from thesecontexts.

If we accept the need to understand persons—the focus of liberalconcern—as relational rather than self-contained beings, then we mustmodify the ways we understand our principal moral and political obli-gations to them. In particular, we need to revise the ways in which weunderstand the traditional liberal concerns of autonomy and justice inorder to reflect the fact that persons are now conceived of as a differenttype of entity. (Feminism, in my view, does not require us to abandoncommitment to the core liberal values of autonomy and justice, but itdoes demand significant reinterpretation of each concept.) Because so-cial group membership is a principal component of feminist relational


accounts of personhood, efforts to promote autonomy and justice mustconsider ways in which these values should be addressed at the level ofsocial groups as well as at the more familiar level of individuals.

In particular, we must make explicit the fact that autonomy is bothdefined and pursued in social contexts. This requires us to reflect onways in which increased options may sometimes limit real autonomy.For example, feminist researchers have explored ways in which the veryavailability and normalization of prenatal testing of fetuses, while ap-parently expanding the control women have over their pregnancies, havethe effect of making many women feel compelled to participate in suchtesting even though it distorts their experience of pregnancy and doesnot fit well with their deepest values regarding acceptance of childrenwith disabilities (Rothman, 1986; Sherwin, 2001). On traditional views,autonomy is achieved if the women concerned have full information andare free from explicit compulsion; relational interpretations make visi-ble ways in which social practices of normalizing prenatal testing makerefusal of such ‘services’ very difficult for many women.

Hence, where liberal individualist theories either look at the genericconditions required for any individual to exercise autonomy (in polit-ical theory) or focus on the quality of decision-making by individualsas they are presently constituted (in practical ethics), feminist relationalaccounts explore the social conditions that support—or inhibit—eachperson’s ability to identify and pursue her own concerns. Feminist ac-counts are particularly sensitive to the ways in which oppressive struc-tures limit the types of choices available to members of particular socialgroups. Feminists are also concerned with the fact that social condi-tions may affect an individual’s very ability to exercise autonomy inthat some people who face significant oppression may have fewer op-portunities than others to develop the skills necessary to exercise au-tonomy (Meyers, 1989). Oppressive conditions may interfere with theopportunities people have for developing skills necessary for exercisingautonomy generally or in particular types of circumstances.

Similarly, feminist relational accounts of social justice look beyondconsiderations of fair distributions of benefits and burdens to considerhow various policy options will affect existing patterns of oppression

3.6. BRCA TESTING REVISITED 69

or privilege. They consider not only what share of society’s limitedresources an individual can justly lay claim to (the underlying goal ofBraybrooke’s account of needs) but also how patterns of social orga-nization may construct needs and roles differently for particular socialgroups and how these differences might affect recognition of, and re-sponse to, claims of need. The Canada Health Act promises equal ac-cess to medical and hospital-based treatments, but it does nothing toprovide for equal access to appropriate levels of the various social andeconomic conditions (such as housing, employment, and education) thataffect an individual’s health status and need for medical treatment. Totreat all individuals with comparable medical course-of-life needs asmaking equivalent demands on the health care system is to treat theircurrent situations as if they arose from a level playing field. A feministrelational understanding of social justice would have us look beyondefforts to fairly meet current needs and also examine the social condi-tions that generate those needs. It would also direct us to ensure that thepractices and policies that address current needs do not unjustly worsenthe condition of members of disadvantaged social groups, lest we createfuture distortions in justice.

3.6 BRCA Testing Revisited

I am finally ready to return to some of the moral complexities attached tothe availability of BRCA testing with respect to the question of whethergenetic susceptibility testing should be classified as a medical (course-of-life) need and treated accordingly by public policy. My claim is thatrevising Braybrooke’s proposal of socially situated course-of-life needsin accordance with feminist relational theory can help us decide how torespond to the sorts of demands genetic dispositional testing is begin-ning to generate within a system committed to meeting medical needs.

The traditional liberal conception of personhood would have us con-struct this problem as one of individual preferences and focus on themoral principle of respect for patient autonomy. Indeed, this is the waythat genetic susceptibility testing is being promoted: as a matter for indi-vidual choice in a context of thorough, reliable information. (In Canada,


BRCA testing is provided in settings where trained genetic counsel-lors are available.) Braybrooke’s approach is a significant improvementsince it urges us to develop an understanding of course-of-life needs thatwould take precedence over matters of individual preference and choice;hence, it allows us to see that not all decisions can be reduced to mat-ters of consumer choice. Unlike adventitious needs that are a form ofpreferences, course-of-life needs are meant to capture something moreurgent than the things individuals happen to choose.

From what we now know of BRCA testing, it seems plausible to re-gard it as a course-of-life need for the small minority of women whosefamily history suggests there is a high likelihood they are at risk of car-rying an identifiable BRCA mutation (less than 10 per cent of the 13per cent of women who will contract breast cancer in their lifetimes).This means that those women who meet strict eligibility criteria, basedon scientific evaluation that they are at high risk of having such a mu-tation, should be granted access to the test (should they so choose) andany follow-up therapy they decide upon in consultation with medicalexperts. Other women—that is, the vast majority, including many withbreast cancer in their families—have such a low likelihood of carryingan identifiable BRCA mutation, however, that testing should not be seenas a course-of-life need. Nor should it be described as an adventitiousneed, to be available upon informed request. Rather, it should be seenas not meeting any need at all.9

A relational understanding of personhood, autonomy, and justicehelps us to understand why it is not acceptable to allow individuals whodo not meet scientific eligibility requirements for testing to exercise per-sonal choice in seeking out such tests. Relational theory directs us topay close attention to the prevailing climate in which women are askedto make these difficult decisions. It is a climate in which all women,especially those considered vulnerable to hereditary breast cancer, areregularly warned about the threat of breast cancer and advised to pursueevery avenue that might reduce their personal risk no matter the actualimpact of specific measures in their circumstances. In the current con-

9In fact, this has been the way BRCA testing has been handled in Canada to date. It isavailable only to women who meet strict scientific eligibility requirements.


text of heightened fear, inadequate information, poor understanding ofthe meaning of available statistical data, and the manipulating effectsof marketing, it is problematic to speak blithely of ‘voluntary informedchoice.’ Many of these factors actively undermine the ability of womento engage in voluntary, rational deliberation - the basic requirements ofpatient autonomy. Moreover, these conditions are not ones that can befully ameliorated by good genetic counselling, since genetic counsel-lors cannot themselves alter the background social conditions that maybe distorting understanding.

We need to reflect on how these background social conditions haveevolved and how BRCA testing itself fits into a particular, problematicmodel of medical care. Breast cancer is a highly visible and emotionallycharged disease. It has become emblematic of concerns about women’shealth to the point of eclipsing attention to many other conditions thatthreaten women. Public awareness of the availability of BRCA testscontributes to the public’s perception of the pervasiveness of the threatof breast cancer and also to the priority given to individual strategiesfor breast cancer avoidance in a competitive health research and healthservices environment.

Moreover, as BRCA tests are joined by other genetic susceptibil-ity tests, the growing array of genetic tests is likely to feed increasingpublic sentiment that most health risks reside in our genes - a tendencythat Abby Lippman (1991) has dubbed ‘geneticization,’ whereby com-plex phenomena are reduced to simple expressions of single genes. Thisattitude works in conjunction with the complementary perception of ge-netic determinism (that is, the view that genes determine experience)to foster public beliefs that improvements in health will flow directlyfrom increased genetic knowledge. Lost from these ways of viewingthe world is the understanding that genes operate within particular en-vironments and that genetic tendencies result in specific outcomes onlyin the presence of additional causal factors. Already, the search for ge-netic susceptibilities is obscuring the need and undermining support forresearch into other conditions which contribute to particular diseases,including breast cancer. When BRCA testing is joined by an array ofsusceptibility tests for other diseases that many adults fear (heart dis-


ease, Alzheimer’s, diabetes), popular understanding is likely to shifteven further in the direction of geneticization and genetic determinismand away from other types of programs that can reduce the incidence ofmany forms of disease.

If BRCA and other genetic susceptibility tests are judged to be course-of-life needs for growing segments of the population, then as the num-ber of tests proliferates, we must ensure that the resources needed todeliver the tests in a responsible manner are available to potential users.This will require a significant social investment in training many moreprofessional genetic counsellors. Provincial health plans will need toset criteria that determine what genetic tests will be covered by pub-lic insurance and with what degree of support. And, clearly, they willhave a moral obligation to ensure that those who receive positive (andperhaps also indeterminate) test results will receive the therapy (bothmedical and psychological) deemed appropriate by patients and theirphysicians. Those engaged in setting health research priorities must en-courage long-term studies to fill in the current knowledge gaps aboutoutcome measures for various responses to BRCA test results, and sim-ilar studies will be required for other types of genetic tests.

Before we get too caught up in building the infrastructure for widespreadgenetic testing, then, we need to recognize that decisions (or non-decisions)about proliferation of such tests made at the policy level have profoundimplications for patients and providers in all their deliberations. Be-cause breast cancer, like most diseases, is multifactorial, it is essentialthat we not lose track of other course-of-life needs attached to this andother diseases. This will require policies that provide research and treat-ment resources that correspond to the various factors that contribute toeach disease, perhaps in rough proportion to the degree of involvementof each. Hence, we must not trade an enthusiasm for genetic informa-tion for recognition of the enormous significance of environmental andsocioeconomic factors in the incidence and severity of disease.

Yet, as scientific understanding of the role of genetic mutations in-creases, those who set health research priorities seem to be losing sightof the well-established fact that most illness and suffering result fromconditions that can be ameliorated by social and political actions. For


example, there is growing evidence that we can probably prevent farmore cases of breast cancer by identifying and altering the environ-mental factors that lead those who are susceptible for genetic and otherreasons to develop breast cancer than we can through BRCA testingand radical surgical preventions (Chernomas and Donner, 2004). Look-ing beyond the impact of breast cancer, we see overwhelming evidencethat reducing the economic, social, and political inequalities that plaguemembers of disadvantaged social groups—for example, aboriginals, im-migrants, and people who are homeless—will have a major effect ontheir health and life expectancy. Our collective moral responsibilitiesrequire that health research and health services policies not be deterredfrom these important dimensions.

It is significant that the move to geneticization and overinvestmentin genetic strategies occurs as individual patients, researchers, and med-ical specialists pursue paths that make good sense from their particularsocial locations, each acting responsibly in the context of their own in-terests and expertise. Given the choice, as individuals we are each likelyto pursue any option that offers to reduce the risk of contracting diseasesthat frighten us, and researchers and medical caregivers will feel thepressure to provide these options. Thus, these social effects cannot beavoided by insisting on more responsible decision-making by individualplayers. Nor will they be resolved by more complete consumer infor-mation. We must, therefore, address the problem at the level of socialpolicies, not individual autonomy. Otherwise, geneticization will con-tinue to threaten efforts to address other, important course-of-life needs.Justice demands that we focus our health policies on strategies that willcontribute most effectively to the health of the whole population, espe-cially the health needs of the most disadvantaged. Indeed, it is the factthat these two approaches—individual and collective pursuit of healthprotection strategies—do not always come to the same conclusions thatmakes it urgent that we continue with Braybrooke’s project of tryingto develop an understanding of needs that can justly guide social pol-icy, lest we end up trying to respond to whatever individual preferenceshappen to present themselves.

It may appear, however, that we are now even deeper into the ‘bot-


tomless pit’ of medical needs for expensive life-prolonging therapiesthat caused Braybrooke to despair of a solution within his systematicaccount of needs. I believe we have made some progress, though. First,it should now be clear that we can deploy Braybrooke’s conception ofcourse-of-life needs to legitimately limit access to genetic susceptibil-ity tests to those for whom the proposed test is indeed likely to meeta course-of-life need: that is, to those at significant risk of carrying anidentifiable mutation for a life-threatening disease with a potential life-saving therapy available. We can, in good conscience, deny it to otherswho may very much want to know their BRCA status: they have noneedfor such testing.

Second, it now seems clear that we should not concentrate solely ondetermining whether or not to treat BRCA and future genetic suscep-tibility tests as meeting course-of-life needs for some individuals; weshould also reflect on how we have arrived at this difficult decision. Inother words, we need to think about how our health care system hasevolved and where it is heading. We cannot make satisfactory progressin determining how to categorize individual types of medical technol-ogy on a case-by-case basis alone, even if the technologies in questiondo promise to save or prolong lives. Instead, we need to reflect on how,collectively, we ought to approach matters of life and health throughoutour social policies.

Such reflections make clear that we are obligated to ensure that thescientific projects our society funds and the delivery systems it devisesreflect the general health needs of the population. Our goal should beto foster those policies most conducive to protecting and promoting thelives of all citizens. Social justice requires us to work particularly hardat addressing the health needs of those members of society whose healthis seriously harmed by pervasive patterns of injustice in society. Hence,we cannot wait to respond to the demands of frightened individuals whoseek access to technologies that private corporations have determined tobe profitably marketable. In fact, we should make clear from the startthat we will not allow our health system to respond to requests for teststhat meet, at best, adventitious needs in a consumer-driven climate. Oursupport for genetic susceptibility testing should be limited to those tests


that can be shown to meet course-of-life needs for a defined segmentof the population. Moreover, we need to take steps to ensure that thehealth interventions that are being developed are the ones that will bestpromote health in our society.

Braybrooke is correct that health and life are core course-of-lifeneeds. We must be very careful, however, not to read him as directingus down the dangerous path of trying to determine, technology by tech-nology, what medical innovations will be classified as meeting course-of-life needs. These decisions are best made in terms of overall socialpolicy objectives that keep social justice and relational autonomy con-siderations squarely in view.

Chapter 4

The Mutual Limitation of Needs asBases of Moral Entitlements: ASolution to Braybrooke’s Problem

DUNCAN MACINTOSH

Abstract

David Braybrooke argues that meeting people’s needs ought to be the pri-mary goal of social policy. But he then faces the problem of how to dealwith the fact that our most pressing needs, needs to be kept alive withresource-draining medical technology, threaten to exhaust our resourcesfor meeting all other needs. I consider several solutions to this problem,eventually suggesting that the need to be kept alive is no different in kindfrom needs to fulfill various projects, and that needs may have a structuresimilar to rights, with people’s legitimate needs serving as constraints oneach other’s entitlements to resources. This affords a set of axioms con-straining possible needs. Further, if, as Braybrooke thinks, needs are cre-ated by communities approving projects, so that the means to prosecutethe projects then come to count as needs, then communities are obligedto approve only projects that are co-feasible given the world’s finite re-sources. The result is that it can be legitimate not to funnel resourcestowards endless life-prolongation projects.∗

∗My thanks to Heidi Tiedke, Susan Sherwin, and Richmond Campbell for helpful discussion.My thanks also to Sherwin and to Peter Schotch for their editorial work on this volume, and, ofcourse, to David Braybrooke, for the stimulation of his ideas and for the example of philosophicalcollegiality which he has constituted in the philosophy department at Dalhousie University.

77

78 CHAPTER 4. DUNCAN MACINTOSH

4.1 Introduction

In his bookMeeting Needs(1987), David Braybrooke argues that meet-ing people’s needs ought to be the primary goal of social policy. Hedistinguishes two types of needs. First, there are course-of-life needs.These are needs which everyone has and which must be satisfied forpeople to function normally in society when performing any of the rolesof which the society approves and advancing any of the projects it per-mits people to choose. Second, there are adventitious needs, ones vary-ing from project to project, and which must be satisfied in order forpeople to advance the specific projects for which they have individualpreferences. Social policy should be organized to try to satisfy needs intheir order of importance—course-of-life needs first, adventitious needssecond.

In this paper I moot several possible solutions to a problem that wor-ries Braybrooke. The problem is that using costly, resource-intensivemedical technology to meet the basic needs people have to be kept aliveat what would otherwise be the end of their lives will consume all theresources available for meeting needs in general. If grounding entitle-ments in needs is correct, we must find within the concept of needs away to limit this crisis in a morally acceptable way. Many of the solu-tions I shall consider have difficulties, but discussing them will refinethe problem and point the way to a workable solution.

4.2 The Appeal of Needs as the Basis of Moral Entitlementsto Resources

It is attractive to see needs as the basis of moral entitlements. It seemsmore plausible, for example, to see it as entitling someone to somethingthat they need it, rather than just that they want it, or would be madehappy by it. That they have a right to it might seem an even better ar-gument. But what grounds the right? Again, it is plausible to say thata person has a right to something in proportion to her need for it. Thisseems more acceptable than saying that her right owes to her havingfound the thing, or her having claimed it, invented it, contracted for it,

4.2. THE APPEAL OF NEEDS 79

or having mixed her labour with it. For these latter pretexts presupposeprinciples of justice in the acquisition of goods characteristic of insti-tutions of private property, institutions themselves needing grounding;and, arguably, they are defensible only if their existence results in peo-ple’s needs being met. True, such institutions have been defended byNozick as being required in order to afford everyone unfettered liberty(Nozick, 1974). But on reflection, surely there are things more impor-tant than that kind of freedom, specifically, the meeting of people’s ba-sic, undeniable needs. There seems, then, to be something foundationaland irrefutable about someone’s having a title to something on the basisof needing it.

Taking entitlement to ground in needs has the further advantage thatthe question of whether someone has a need seems to be empirical,something that can be ascertained by investigation, and on which therecould be a consensus compelled by the non-moral facts. Thus it wouldmake entitlement objective. And this would mean that, in one stroke,we had answered three questions of meta-ethics:

1. Question: how can morality be factual? Answer: the question,what, morally, ought to be done, reduces to the question, whoneeds what, and so who should be given what?

2. Question: how can we infer from statements about what is (true)to statements about what ought to be (true)? Answer: by the me-diation of needs: if it is true that someone needsx, it ought to bemade true that they are givenx.

3. Question: how can apprehending moral facts be reason-giving inthe sense of motivating? Answer: our moral duties are to meetneeds, and seeing that someone needs something tends to inclineone to procure it for them.1

Taking needs as the basis of moral entitlements also provides a ground-ing for forming policy on how to distribute resources among people, forwe could then compute which policies are correct from knowledge of

1For more on this constellation of issues, see Richmond Campbell’s essay in this volume.


people’s needs in a needs calculus, much as was hoped by utilitariansfor the felicific calculus.

4.3 Braybrooke’s Problem

Unfortunately, there are at least two problems for this view, problemsderiving from the possibility of two kinds of ‘needs monster,’ on anal-ogy with the notion of a utility monster. Utilitarians hold that a worldis morally better the more happiness there is in it. There are at leastthree versions of utilitarianism. On one, what matters morally is howmuch happiness there is in the world, not how many people share in thathappiness. But on this version, if there were a creature who, by beinggiven all of the world’s resources, would be made more happy than thesum total of everyone else’s happiness on any other distribution of re-sources, that being should get all the resources; and this seems unjust.(This creature is hyper-efficient at converting resources accorded to itinto its own happiness, and therefore hyper-appetitive for resources.)

Suppose we say instead that a world is better the more happinessthere is in it, provided everyone is made equally happy in that world—total quantity of happiness is not all that matters; it matters too how thehappiness is distributed. But then there could be another kind of util-ity monster, a creature so ineffective at converting resources accordedto it into its own happiness that, even distributing most of the world’sresources to it at our expense, so that the rest of us have only enoughresources to live lives just barely above being miserable, this creature’slife will still just barely be above being miserable. (If the first creature isthe efficient utility monster, the second is the inefficient utility monster,and it too is hyper-appetitive of resources.)

This suggests a third kind of utilitarianism, one in which both the to-tal amount of happiness and its distribution matter, but the distributionrequired is not perfect equality. Instead, if vastly more people could bemade vastly happier by not trying to make the inefficient utility monsterhappy, then that is how resources should be distributed. Equality mat-ters, but only to a degree; it can be outweighed by other considerationsusing some discounting factor. People do not have an absolute right to

4.3. BRAYBROOKE’S PROBLEM 81

participate equally in the total amount of happiness, only a weightedright. Their being equal participants is only allowed to weigh down to acertain degree the total amount of happiness. The degree, of course, isproblematic to formulate or justify.

At any rate, until recently, the existence of such monsters was alargely abstract problem: most of us are in fact pretty similar in ourappetites for resources and in our efficiency at converting resources intoour own happiness. True, relative to me (a non-disabled person), a dis-abled person is in some degree an inefficient utility monster; and rela-tive to a disabled person, I am in some degree an efficient utility mon-ster.2 But the resource cost to the rest of us for helping the disabled isrelatively small. For many of the disabled are fairly easily helped, sohelping them is cheap (we put in wheelchair ramps and so on), whilethe remainder are fantastically hard to help (they have irreparable spinalcord damage, say), so hard that there is almost nothing we can do (ex-cept give them nursing care, say), and so, even doing what we can, itis cheap to help. So barring a massive increase in the number of dis-abled people, or in the severity of their disabilities, or in the availabilityof hugely costly means of helping them, redistributions of resources toaccommodate the disabled will probably not overly drain the total re-sources available.

But medical technology is getting to the point where, with great costin resources, it is possible to extend individual lives; and the longer livesare to be extended, the more resource-costly it is to do the extending.Already the lives of some persons are being extended at great cost inresources, the cost growing as technologies become available for ever-increasing prolongation. (Think of the epidemic of diabetes and thecost of the dialysis sometimes needed to treat those in diabetes-induced,end-stage renal failure—hundreds of thousands of dollars per year perpatient.) And if ever more resources are devoted to producing thesetechnologies in order to extend people’s lives, ever fewer resources will

2I am assuming for the sake of argument that, on average, it would take more resources toprovide a disabled person with a given amount of happiness than would be required for a non-disabled person; for I am assuming that extra resources would be required to mitigate the effectsof their disabilities. But in a given actual case, this assumption could be false: notoriously, howhappy people are can be quite independent of whether or not they have perceived disabilities.


be available to support the activities that make people’s lives worth liv-ing. Pretty soon, we will all be inefficient utility monsters. We will allbe said to require for our happiness more and more of the ever-more-available but ever-more-costly technologies needed to prolong our lives.

The analogous problem for a needs-based morality is obvious, and itis a problem Braybrooke admits he is unsure how to solve (Braybrooke(1987a) chap. 8): if people need anything, they need the prerequisitesfor being alive; and as our bodies inevitably age, we will all come to apoint where only longevity-prolonging technology can keep us alive—we will all need this technology. But not all of us can have it, or atleast not without other people, at some point in their lives, not gettingthe resources to have a decent life. Meeting people’s needs for life-sustaining technologies can only be met at the cost of not meeting manyother needs. Perhaps these other needs are less important. And yet theyare needs which, if they go unmet, make it dubious whether life is muchworth living.

On the face of it, Braybrooke’s problem is not solvable. Surely needsshould be met in the order of their importance; and surely it is moreimportant to meet someone else’s need to be kept alive (assuming theywill have some minimally decent quality of life) than to meet my need tohave a life more than minimally decent in quality. It then seems morallyobligatory to distribute resources in such a way that as many people aspossible are kept alive as long as possible, even if this is to be at theexpense of the overall quality of people’s lives.

Indeed, the problem seems even more pressing for needs-based ethicsthan for utilitarianism, or at least the third form of utilitarianism. Forthe utilitarian can say that all extending a person’s life by a day doesis increase one person’s happiness by one day; and if there is a moreefficient way to increase more people’s happiness, then since the totallevel of happiness would be higher that way than by using otherwise de-ployable resources in heroic medical efforts for one person, that is howthey should be used. But it seems that, on needs-based ethics, the needa person has to be alive trumps all other non-life-and-death needs of anyand all other persons.

4.4. POSSIBLE SOLUTIONS TO BRAYBROOKE’S PROBLEM 83

4.4 Possible Solutions to Braybrooke’s Problem

Braybrooke’s problem in a nutshell is this: given that our social andinstitutional priorities should be meeting people’s needs, in their orderof priority, how is this not to oblige us to live lives of drudgery sup-porting the longevity of ourselves and others? There seem to me to beseveral possible solutions. I now consider them in order of increasingconceptual ambitiousness.

4.4.1 Questioning the Empirical Assumptions of the Problem

One solution is to claim that the problem makes empirically false as-sumptions. Maybe it will not prove all that expensive to extend lives;maybe this can be done without much sacrifice in the meeting of otherneeds, and in the meeting of the needs of other people who do not yetrequire medical technology to live. For one thing, extended lives maywind up being extended in productivity as well, so that, as the need forlife-extension technology expands, so will those resources constitutedof a robust workforce. And so we will have resources to meet theseneeds: people with longer lives will be able to work to pay for ever-longer lives.

It may also be possible to innovate not just in inventing life-sustainingtechnologies, but also in their resource cost—maybe they are likely tobecome cheaper and more efficient, especially as they are pursued in anever-expanding market.

Further, maybe meeting some people’s prolongation needs will re-sult in others’ needs being met—the person whose life we save can nowcontinue in the workforce; and with her productivity there, she can con-tribute not only to maintaining her own life, but to meeting the needsof other people, and to meeting needs of other sorts than those of lifeprolongation. This solution, however, banks on empirical hopefulnessand so cannot be counted upon.


4.4.2 The Defeasibility of Needs

Another proposal recognizes that the needs of a given person or pop-ulation may be defeasible and may have to yield to the more pressingneeds of other persons or populations, so much so that those whose so-called needs are defeased never really had them. To take an example,all women might have been thought to need to be tested for the genefor breast cancer. But in fact, unless there is a specific pattern of breastcancer in one’s family, it is unlikely that one has the gene. So the peo-ple with breast cancer in their families need the test more than peoplewho do not; and if the test requires resources that are also in demand formany other, more pressing needs, maybe the people who do not havethis known risk factor do not really need the test at all.3

There are attractions to the proposal: it seems right to meet morepressing needs first and to expend resources where they will do the mostgood. However the proposal presupposes that we have some generalprinciple for balancing out the distribution of resources to people, andthis has not yet been deduced from the concept of needs as bases ofmoral entitlement. Besides, there is no guarantee that, even giving re-sources first to those who need them most, there will not come a timewhen resources will be exhausted even by this principle. After all, ev-eryone’s life comes to an end if it is not extended by medical technol-ogy, and no one’s need to live a longer life is, other things equal, morepressing than anybody else’s. We will all eventually need costly heroicmedical technology. We will all still become inefficient needs monsters.

4.4.3 Is the Location of Needs in Individuals or Groups?

A related solution is to reconceive the locus of needs. It is generallyassumed that the thing which has needs is the individual person. Butperhaps needs are had not by individuals in isolation, but by popula-tions; or perhaps individuals do have needs, but only derivatively fromthe needs of the populations of which they are members. Either way,

3This is an aspect of Susan Sherwin’s proposal which is presented in her essay in this volume.In addition to owing this and the next proposal to her paper, I am grateful for its succinct and lucidsummary of Braybrooke’s book and of what I am calling Braybrooke’s problem.


and returning to the problem of what to do about an aging person whorequires costly medical technology to remain alive, our first questiontherefore should not be whatthis older person needs, but whatolderpeople need. The answer to this is perhaps whatever will keep as manyof this population’s members alive for as long as possible with as gooda quality of life as possible. And to that end, perhaps, what this popu-lation needs is not heroic, resource-costly new medical technology, butbetter nutrition, community-based, preventive medicine, better educa-tion, and a higher minimum income. We could do much more for thepopulation of older people with low-tech preventive medicine and so-cial reorganization than with the limited number of high-tech machineswe could produce.4

Notice that, by taking populations rather than individual persons asthe bearers of needs, we benefit conceptually by being able to considerwhat makes a population as a whole well off in terms of needs sat-isfaction; this notion seems already to embed some of the conceptualapparatus we require in order to do needs balancing: the mere fact thata given member of a group is faring poorly does not automatically meanthat the group is faring poorly and so may not mean that there has beena failing of moral duty in looking after that population’s welfare. Thiscould allow us to justify not meeting what would have been the pressingneeds of an individual on an individualist analysis of needs.

Unfortunately, it is not obvious that groups matter except so far astheir members matter individually, so that the needs of the group mayultimately be nothing but the sum of the needs of its members; and ifwe think of the parts of this sum as constituted of one individual at atime, we are back to each of these people needing heroic technology,their need for it obliging us to drain away resources from meeting othersorts of needs of theirs and of other people.

We might reply by saying that the very identity of the members asindividualsderives relationally from the identity of the group, so thatit makes no sense to see any of the members as individuals except inrelation to the group. Or at least it makes no sense to see them as having

4This proposal is probably a conceptual generalization of a specific proposal Sherwin makes.


needs except derivatively from the needs of the group.5 Thus your needfor aid in having your life extended merely gives you title to what apolicy that advances the longevity of older people in general would giveyou. So you are specifically entitled only to what, on average (to takeone example of a criterion for assessing the state of a population), bestserves older people. And since having expensive technology for youwould mean not having community-based medicine for the many otherswho could benefit by it compared with your having your technology,you do not have a need for that technology, but only for the services of,say, community-based medicine.

Problems remain, however, for we can still face conflicts of needsbetween groups—the quality-of-life needs of young people versus thelife-and-death needs of old people, for example. We might try to evadethat by treating everyone as members of one giant group, therapizing theneeds of the population as a whole rather than individual by individual,and giving to each individual only the form of treatment which benefitsthe most members of the population as a whole.

But while no doubt we are all in some sense defined relationally toeveryone else, we also have sub-relations in which inhere some needs;and these, again, can conflict with the needs inhering in other sub-relations, or even in ourselves as between the several sub-relations inwhich we participate. (For example, I am now relatively young but willone day be old.) Moreover, the larger we make the population in thegroup, the more the problem of what to do for its members becomes ouroriginal problem; for the more inclusive we make the group, the moreheterogeneous we make its membership, and so the more potential forconflict of needs defined in sub-relations there is among its members.If the group contains the young and the old, for instance, it seems morelikely that there will be a conflict of needs between the group’s mem-bers. While if we say that what each member is entitled to is determinedby the needs of the group, then this presupposes that we have figured outhow justly to balance the distribution of resources to people generally.

Thus in reconceiving the locus of needs onto groups and away from

5Again, I am indebted to Sherwin for her work on relational conceptions of persons; there isa close connection between what I am considering here and aspects of her proposal.


individuals, I have perhaps illicitly assumed that, once needs are locatedas inhering in groups, the notion of needs would permit helping as manypeople as possible, and helping them as much as possible; and this maybe problematic for allowing us to ignore someone whose inclusion inour policy would reduce the total number of people we could help, orthe degree to which we could help them. For how is this to be justi-fied? After all, they too are members of the population. So why are weallowed to ignore them? The answer is presumably that we have not ig-nored them—they are getting the same treatment as everyone else (com-munity medicine, say)—it is just that it is not really doing them muchgood. But then it seems we should have a way of deducing from theconcept of needs itself, not just that needs inhere in populations ratherthan individuals, but also the correctness of using a certain conceptionof what it is for a population’s needs to be adequately met. For whyshould not the test be, say, that the least well-off person in the groupis made better off by the correct policy, rather than that most people inthe group are made better off? But if the former is the correct test, wehave our problem back, for the least well-off people will again becomeinefficient needs monsters relative to us. While if the latter is the correcttest, we are efficient needs monsters relative to them.

Finally, even waiving this conceptual issue, and even if we can go along way towards meeting the needs of a group conceived at the grouplevel with low-tech manoeuvres, ones low in resource cost, sooner orlater, the only way to go further will be with expensive technology andresearch, and we will have our problem back; we will once again all beinefficient needs monsters.

4.4.4 Do People Really Need Indefinitely Long Lives?

Another way to solve Braybrooke’s problem might be to challenge di-rectly the idea that the mere fact that something would be a means toextending a person’s life means she needs that thing; for it is disputablewhether a person needs to live a very, very long life. Suppose therewere a way to keep a rabbit alive indefinitely: does the rabbit reallyneed to live forever, especially if the means involves virtually starvingall the other rabbits? Now suppose we can do this for a person. Does


she really need to live indefinitely? Why? Perhaps it would be nicefor her—she might like it or want it, and so might her friends, family,colleagues. But that it would make her and various other people happydoes not entail that she needs it. No doubt there are various projects herbeing alive would allow her to prosecute. But does she need to completeindefinitely many projects? Is there any need that she be the one to doall of them?

This line of argument may not work, however, for our conceptionof what someone needs to have had a full, satisfactory life is probablysomething whose boundaries gradually expand as the possibility of alonger life expands. Perhaps when life spans were shorter it would havebeen thought extravagant to have a life much more than seventy years,and it would have been doubted whether anyone needed to live muchlonger—a seventy-year span would have been long enough to find love,raise a family, have a career or two, be a decent citizen, experimentwith alternative lifestyles, see what there was to see of the world, cul-tivate one’s potentials and talents, make some mistakes and atone forthem. What need to live much longer? But now that people can liveto be eighty, ninety, a hundred, there seems to be all that much moreto experience and do—there is now the prospect of a long, healthy, andpeaceful retirement, extended grandparenting, travel, new hobbies, ex-tended education, seniors’ political activism. Our conception of what itis to have lived as long as one needed to live for a good and full life hasexpanded, and will keep expanding. And why should it not?6

4.4.5 Needs to Life Extension Not Really More Pressing ThanSo-Called Less Basic Needs

But maybe we are wrong to think of the means to life prolongationas representing greater or more pressing or higher-priority needs thanthose that would have to go unmet to meet them. Why, after all, does

6I should acknowledge that Braybrooke was more interested in the problem of the resourceneeds of those experiencing medical difficulties in mid-life, a time when, by any measure, theyhad not yet lived a full life; but the logic of the problem is the same for older people, especiallyas our conception of a reasonably full and long life expands. Indeed, the logic is starker for olderpeople, since aging and its infirmities are inevitable and so ubiquitous.


a given person need to live another day? Answer: in order to do thethings persons trying to live a good life do in a given day—go to amovie, maybe. But a given younger person not on death’s door withoutexpensive technology to rescue her might need to go to the movie too.Looked at perhaps a bit cynically, a life is nothing but a string of daysfull of miscellaneous projects, and so the need to live, and the needfor the means of continuing to live, is nothing but a need to advanceprojects. But we all have projects, no one’s automatically more valuableor pressing than anyone else’s. So why should the fact that a givenperson requires a lot of resource-costly medical technology to engage inher next project mean that its requirements should have higher prioritythan, say, mine, which need nothing but a movie ticket? That you willdie without the machinery sounds like a big, trumping argument, but allyou are going to do with your technology is, say, go to a movie. So nowwe are comparing your going to a movie, figuratively speaking, with mygoing; and why should you win? Why should I renounce my ticket tofinance your machine just so you can go to the movie instead of me?

4.4.6 Meeting Basic Needs versus Meeting the Needs the Meetingof Which Make Life Worth Living

Relatedly, if you are not even going to go to the movie—maybe no onecan afford movies after we all pay for your machine—why stay alive atall? What is the point if you cannot, say, go to movies? It is implausiblethat we should have to trade off what makes life worth living just tomake more life, whether the trade-off is within one person’s life, orbetween the lives of two.

4.4.7 Meeting Other People’s Needs as Not Really Being a Cost tothe Meeting of Our Own Needs Given That We Are NeedsAltruists

Yet another way we might go is to observe that people’s needs overlap inthis sense: if it is true that one person’s seeing that another person needssomething inclines the first person to ensure that the second has her needmet (as was hypothesized of needs as part of what makes needs-based


grounding of entitlements attractive), then perhaps the first person hasa need to see the second person’s need met. But if one of our needs isto see other people’s needs met, so that meeting their needs is part ofmeeting ours, then funnelling resources to them is not seen by us as aresource cost to us and the meeting of our needs. Returning to utili-tarianism, relative to me, a disabled person is not an inefficient utilitymonster if I derive happiness from the happiness of the disabled; forthen I would not experience the funnelling of resources to the disabledas a cost to my resources. And relative to a disabled person, I am notan efficient utility monster if she derives happiness from the happinessof the non-disabled, for then she would not experience the funnelling ofresources to them as costs to her resources.

This can only take us so far, however. For at some point, my needsbeing defined as needing the meeting of your needs must ground out ina need defined independently of other people’s needs. And in any case,it is certainly false from the start that our only needs are for meetingothers’ needs. But if we have any other needs, and if meeting them couldrequire an indefinitely long life, it will be possible for life-prolongationneeds to swamp all other needs.

4.4.8 Needs as Relative to Projects Approved by Communities;Re-Choosing Needs

Braybrooke himself in effect provides the materials for a more promis-ing solution. He sees needs as functions of the extension of the term‘needs’ in a linguistic community. This extension is determined by theprojects which the community endorses as permissible or important,with the most basic needs being things that are means to the pursuit ofvirtually any project. But then it might be an option to have a com-munity alter its conception of appropriate projects, thence to alter whatwill count as needs; and perhaps one factor in any such self-chosencultural evolution would be the co-tenability of its projects given finiteresources. A community might choose, then, to have the project of liv-ing a very long life not be among the projects it recognizes as important.To be sure, there are issues this proposal raises:


1. It implies that communities choose their cultures reflectively anddeliberately; and while this may be true to a degree, much of aculture’s evolution occurs without such self-direction.

2. Even if communities choose their cultures—choose the projectsthey find acceptable as options for people within the culture—it seems intelligible to imagine a community making a mistakeabout whether a project should be permitted. But then, in makingdecisions to put accepted projects into equilibrium with availableresources, it is presumably possible for a community unjustly tolimit or permit projects. And if the genuineness of a need in turndepends on the correctness of cultural decisions about whetherto embrace or eschew projects, it follows that there are culture-independent standards on something’s being a need. In that case,however, it is not—or not just—the community approving or notapproving certain projects that means people do or do not need themeans to them, but these independent standards. So it is the stan-dards which are doing the work; and so for us to use this solution,we must figure out what the standards are.

But let us waive this worry for a moment; let us suppose then thata community cannot make a mistake: someone needsx just if thecommunity has approved her project andx is instrumental to herprosecuting it. Then there is another problem:

3. Needs then seem less empirical, and more stipulative, and so lesssuitable to answering the meta-ethical questions the answers towhich I advertised as among the attractions of seeing moral en-titlements to ground in needs. At the very least, needs would beempirical only in that it is empirical which stipulations a commu-nity has made.

Another difficulty with this approach is:

4. It solves the problem by an illegitimate kind of fiat: if we findthat our community has tacitly embraced the project of peoplehaving indefinitely long lives, and if we find that this is drain-ing away resources from every other project, we just decide no


longer to embrace that project. But this seems no different fromsimply refusing to meet people’s extant needs. It is too easy a wayout, not to say an immoral way. At the least, surely principles ofprocedural justice are violated. People are entitled to know whatthe rules are and not to have the prospects of their presumptivelycommunity-approved projects yanked out from under them. Wecould get around some of this by giving notice that, from now on,no new beginnings of projects of a certain sort will be approved;but there would still be all those people who have well begun theproject of extending their lives, people who will continue to drainall the resources. A final difficulty with the proposal is:

5. If there are constraints of procedural justice on its implementa-tion, then unless these can be analysed from the concept of needs,again, it is not needs that are grounding entitlements, but some-thing else, or needs plus something else.

4.4.9 The Conceptual Analysis of Needs; the Axiomatics of Needs

I do not have room to deal with all of the difficulties with the previ-ous proposal. But I see a way to begin rehabilitating it if we combineit with some conceptual analysis of the notion of needs. The problemhas turned out to be that of how to resolve conflicts of needs, the con-flict between basic and less basic needs, and between the basic needs ofsome persons and the less basic needs of others. It is obvious from themedical technology problem that there is no purelypractical solutionto the conflict: if living a long time is an acceptable project, then thebasic needs of people seem expandable indefinitely as life-prolongationtechnology improves. The only hope of a permanent solution, then, isconceptual. And the form of conceptual solution I shall suggest is thatthe logic of needs is more like that of rights: just as your rights and minestand as mutual limits on each other, so my needs and yours mutuallylimit each other. This could work in either of two ways: it may be that itcannot be true that you have a certain need if some need of mine wouldhave to go unmet in order to meet yours. Or it may be that, while we canhave needs in conflict, the status of one’s needs as serving as a basis for


one’s moral entitlement to resources is limited by the status of the needsof others as serving as a basis for their moral entitlement to resources.Along these lines, we can articulate, by conceptual analysis, some ax-ioms from the idea that needs are the basis of resource entitlements. Allother things equal, surely

a It is better that the needs of more people are met than less.

b It is better that the needs of a given person are met well rather thanpoorly.

c No one’s needs have automatic title to be met if the cost is that noone else’s needs would be met.

d A person has title to have her most pressing needs met even if thisis at the cost of other people having their less pressing needs fullymet, provided their needs are still met fairly well.

e If two people equally need something, and there are resources togive it to only one, a morality of needs is silent on who should getit—something else must break the tie.

f If a certain distribution of resources would meet all of everyone’sneeds, it is the morally correct distribution (and if more than onedistribution would do this, any of them can count as correct, thoughperhaps depending on the implementation of an appropriate symmetry-breaking technique and the official community acceptance of thechosen distribution).

g If, compared with any other arrangement (and failing the possi-bility in (f)), meeting a given person’s needs more fully than theneeds of all others are met would result in more people’s needsbeing met, and in these needs being met more fully, then this is acorrect arrangement (again, with a clause to handle tied systems ofdoing this, where the other systems do it by privileging a differentperson).

h No person is such that, independently of other considerations,their needs deserve to be met rather than those of some other per-son.


i No person is such that, independently of other considerations,their needs deserve to be met more fully than those of some otherperson.7

These axioms afford a solution to our problem, at least if I am rightthat the need to live another day is no more pressing than the needsof a person not immediately at risk of dying. For then the need to liveanother day is just the need to do whatever one was going to do that day;and both of these people have that need. But since this means that thesetwo people have needs tied for being basic and pressing, then, by axiom(e), needs ethics has nothing to say about which of the two should havetheir need met. We may use a symmetry-breaking technique. Politicalnegotiation might be part of the process of breaking the tie, perhaps thesort of negotiation we see during elections about how much priority toassign health care in the national budget and in the process of interestgroups lobbying policymakers. (Axiom (e), then, goes a long way toderiving procedural justice from the notion of needs as entitlements.)

Furthermore, if needs are to base moral entitlements to resources, ifresources are finite, and if needs come into existence by communitiesapproving potential projects, and approving the choosing of them bycertain individuals, then perhaps something is only a legitimate projectif its foreseeable resource requirements are not incompatible with thosethat will result from other people in the community makingtheir ap-proved choices among approved projects. For recall axiom (f), that aresource distribution is correct if it meets all of everyone’s needs. Thereappear to be two ways to bring about such a distribution: we can findthe resources to meet all needs; or, considering the situationab initio,as if prior to any needs existing, we can be careful to create only needsthat the totality of available resources can meet.

Let me develop this second strategy: suppose that, by approvingprojects and their selection by individuals, we in effect allow to comeinto existence needs not all of which can be met given finite resources;then we have brought into existence an unjust pairing of needs and re-

7Probably more axioms could be given than these; these ones could be given a more hierar-chical structure, and they could be more perspicuously shown to follow from the notion of needsas entitlements. Projects for another time.


sources compared with a possible arrangement in which all people’sneeds could all be fully met given available resources. So projects mustbe co-feasible given resources, and people can be allowed to chooseamong those projects only in ways co-feasible given the permitted choicesof others.

This is something we already take care to ensure in our culture. Werestrict the number of people who can be members of certain profes-sions, or who can be funded artists; sometimes we require or allow peo-ple to compete for the privilege of having a certain project, in part tomake sure that people having it is not an excessive drain on resourcesneeded for other people’s projects—think of grant and scholarship com-petitions. All of this is consistent with the axiomatics of needs. Andit solves our problem, provided, again, that life-prolongation needs arejust project needs (or provided that living a long life is a separate projectable to be approved or not independently of certain other projects); forthen life-prolongation needs are limited in their permissible extent bythe co-feasibility constraint. That is, if life-prolongation needs are,project-wise, no more pressing than those associated with any otherproject, then they may be permissibly limited.8

This would be consistent with the possibility I mooted earlier ofits being intelligible for a community to make a mistake in approvingprojects, and so in allowing supposed needs to come into existence bythese approvals. For, arguably, no community has successfully createdgenuine needs if its so-called needs violate the needs axioms. In particu-lar, no community has successfully created genuine needs if its so-calledneeds violate the constraint (axiom (f)) that needs, and so projects andthe numbers of people allowed to participate in certain projects, must

8It might be objected that, while we can avoid failing to meet people’s needs by failing toapprove projects, or by failing to approve individual people’s choosing certain projects, thus pre-venting the needs from coming into existence, we may also be harming people by doing this; forwe are in effect limiting various opportunities of people. The reply is that we are doing this justly,and that allowing these opportunities, because it would result in reductions in the social capacityto meet needs, extant or potential, would itself be unjust; and if Braybrooke is right that meetingpeople’s basic needs is the first moral priority in entitlement considerations, then failing to meetneeds would be worse than not allowing opportunities. Finally, arguably we are not being unjust inrestricting opportunities provided we follow procedural justice in distributing them—for example,provided we break ties in unbiased ways.


be co-tenable given finite resources. We can then say that people whoplead that they have a need, where their having the so-called need metwould involve a total drain on resources, never really had that need inthe first place, because the project that is costing all of these resourceswas never correctly approvable.

This meets the concern of another point raised above that it is wrongto refuse to meet people’s extant needs just because they have becomehuge resource drains. For the needs in question are now revealed neverto have been legitimate, since they violate the needs axioms. But whatabout the procedural justice issue, people having a right not to havetheir projects cancelled without notice? Again, axiom(e) to the res-cue: the projects of those who are having resources drained away bythe longevity project of others also have such rights; it is sad that theresource drain was not anticipated, but we now have a tie on claims;and this can be resolved, again, by political negotiation and symmetry-breaking.

Note that all of this solves the problem as I formulated it earlier interms of needs monsters. For this solution makes both kinds of mon-ster impossible, since both drain all resources in ways we now see areincompatible with the axioms that conceptually define needs as basesof moral entitlements: no one can be allowed projects whose existencewould create needs which would consume all the resources required tomeet other people’s needs, on pain of violating axioms (f), (h), and (i).

Still, there are troubles. For one thing, this solution relies on theclaim that something’s being a means to your not dying does not makeit something for which you have an especially pressing need. But if thatyou will die withoutx does not constitute your having a pressing needfor x, what on earth does? Surely there is something special about needsfor things required in order for one to go on living. Well, perhaps whatis special about continuing to live is that the end of life is not just thefailure to advance a project, but the end of all possibility of projects. Onthe other hand, if life-prolongation needs consume all resources, thatconstitutes, in its own way, the end of all possible projects too for thosedenied resources by their deployment on life-prolongation needs. Somaybe there is nothing special about life-prolongation needs after all.


Again, we have a mere tie, one we may resolve by axiom (e).

Another objection might be this: are there not biologically given,unrenounceable needs, ones not just in common to all actual culturallyapproved projects, but to all possible ones so far as we are biologicallybased, resource-dependent beings? This a plausible worry. But it ineffect introduces a different conception of what a need is from the oneBraybrooke gave us. His account implies that, given biology and thedecisions society makes about how to deal with it, needs are, at leastin part, constructed by the community approval of projects and theircommunity-approved adoption by individuals; and so needs are, there-fore, hugely plastic.

It confirms Braybrooke’s conception over the conception of needs asbiologically given, that many people have had projects whose prosecu-tion required renouncing the so-called biologically given needs—thinkof kamikaze warriors eager to sacrifice their lives in what they see asa just cause, or athletes who knowingly compromise their longevity byusing ultimately toxic but performance-enhancing drugs in the projectof athletic excellence. Staying alive is not part of every project; so theneed to stay alive is not universal, and so not biologically given either.True, many possible projects require one’s being alive for one’s pros-ecuting of them. But that does not automatically make staying alive auniversal, biologically determined, inevitable need; it only makes it ameans to certain ends—ends at least in some degree, for some people,optional.

There may be a way to accommodate the nerve of this objectionmore fully in Braybrooke’s conception of needs, however. For it may bethat most people, perhaps for biological reasons, value certain things—for example, a long and healthy life—and this would induce peopleto endorse these things as projects, and so these projects would befound as part of virtually every culture. Nevertheless, the needs ax-ioms still provide both the conceptual framework of needs and a con-straint on morally permissible community endorsements of projects—there is still the requirement of the co-tenability of needs given finiteresources. While resources should be matched to needs, needs shouldalso be matched to resources.


The co-tenability constraint no doubt leaves a lot of latitude; prob-ably there are many ways to meet it, and so many different kinds ofcommunities’ approved projects could pass it. But it is also a real con-straint. And it may be that, given this, certain projects may be legiti-mately community-endorsed only by the consent of all needs claimantsin the community. Imagine, for example, a community that has becomeobsessed with astronomy: by consensus vote, it approves the projectof devoting most of its resources to producing an extremely long-livedastronomer; the hope is that she can be kept alive long enough, withextraordinary medical and technological efforts, to witness the remain-ing history of the universe, even unto its eventual heat-death billions ofyears from now. This would be a case where a co-tenable collection ofneeds has been created: the astronomer needs to live indefinitely to wit-ness the end of the universe; and all others in the culture are like workerbees who need to do their part in making sure the astronomer lives to seethe end. (Offspring are raised to have the same projects.) But had thecommunity not come to a consensus on this enterprise, the astronomerwould have been an inefficient needs monster relative to some people inthe community, and her needs would not have had automatic title to bemet.

There are further complications here, however. For one thing, itnow seems that needs are not the ultimate foundation for entitlements.Rather, preferences ground everything; for it is preferences that inducemembers of communities to approve of possible projects, and to approveindividuals choosing projects. So the need for something, x, can cometo exist only if people in communities come to a consensus on projectsand the having of them by selected persons, projects to whose prose-cution x is then a means. And this must make us ask whether, apartfrom the constraints of the axioms of needs, there are constraints onmorally permissible preferences, and therefore projects, and thereforeneeds. It must also make us wonder how it is to be determined whata community is; who is in it; whether communities are obliged to besuch that the projects one community approves are co-tenable, given the


world’s finite resources, with the projects other communities approve;9

and whether it is permissible to attain consensus in a community by theexile of dissidents.

There is no room to deal with these matters here, beyond observ-ing that, while needs may require preferences in order to come to ex-ist, preferences cannot make just anything a need. For possible needsare constrained by the axioms of needs; and if it is needs that createmoral entitlements, then at the very least, not just any preferences canaccrue moral entitlements to be satisfied. Given the range of possiblepreferences, a range which has, since Hume, been thought to embed allmanner ofprima faciedistasteful, and yet supposedly ultimately uncrit-icizeable, preferences—from the purely self-interested to the positivelymalevolent—this limiting of moral entitlements by the required media-tion of needs must be seen as moral progress. For it means that merepreferring does not make right.10

9These questions figure in another problem which worries Braybrooke: as part of the interna-tionalization of culture, we in the developed countries have begun to see those in the undevelopedcountries as members of our linguistic community, so that the extension of our use of the term‘needs’ now includes their needs. But then we are obliged to provide resources for meeting theirneeds, again, possibly at the expense of making everyone’s life just barely above miserable interms of met needs.

10This may afford a start on a solution to yet another worry of Braybrooke’s, namely, that someneeds are prima facie immoral, or at least morally embarrassing; but yet as needs, surely they havetitle to be met, possibly in competition with prima facie moral needs. How can we justify meetingthe latter over the former from the concept of needs? Well, the needs axioms, particularly theco-feasibility constraint, may so constrain genuine needs as to rule out the immoral ones. Thisconnects with work I have done trying to prove that all of people’s possible preferences mustbe such as to be co-tenable in the sense of being co-advanceable, this ruling out preferences toexploit other people in the sense of arranging the non-satisfaction of their preferences as a meansto the satisfaction of one’s own. Note that this would require all beings with preferences to see allother such beings as in the same community, and this would help answer some of the questions ofthe preceding paragraph in the main text. Similar sorts of moves could be used to rule out for amorality of needs, various extreme, selfish conceptions of ethics, e.g., needs egoism. For more onthe required co-tenability of preferences, see (MacIntosh, 1998).

Chapter 5

Canadians and GlobalBeneficence: Human SecurityRevisited

EDNA KEEBLE

Abstract

This contribution utilizes some of Braybrooke’s philosophical ideas aboutresponsibility to rethink the concept of human security, a notion rootedtheoretically in a cosmopolitan ethic. I begin with a discussion of theethics of state ‘responsibility to protect’ doctrine as a new norm in theinternational arena. Next, I bring out the distinction made between ‘free-dom from fear’ and ‘freedom from want’ in human security discussions inorder to illustrate that the distinction is a confusing one and that the con-fusion is propagated in foreign policy. I explain the need to look beyondquestions of the state’s responsibility and examine the place of individ-ual responsibility through the lens of Braybrooke’s schema of personalresponsibility for global beneficence. I conclude with what it means torethink human security as a shared responsibility between the state andindividuals, particularly in the Canadian context.∗

∗My thanks to Stella Gaon for generously giving her time to read the first draft meticulouslyand help hammer out my argument (literally), to Susan Sherwin and Peter Schotch, for their lead-ership and comments, and to David Braybrooke, not simply for his response to this piece but alsofor his exemplary ‘personal responsibility for global beneficence.’

101

102 CHAPTER 5. EDNA KEEBLE

5.1 Introduction

In ‘A Progressive Approach to Personal Responsibility for Global Benef-icence,’ David Braybrooke opens with the question ‘What personal re-sponsibilities do we, people living in rich countries, have for relievingmiseries in the less fortunate countries?’1 With this one question, Bray-brooke presents an immediate challenge to students of Canadian foreignpolicy both to reflect on the notion of responsibility in the internationalarena and to focus on the role of individual Canadians who live in arich, developed state to alleviate poverty and hardship abroad. On theone hand, Braybrooke’s challenge is a familiar one: students of Cana-dian foreign policy have become acquainted with what responsibilitymay mean, given the prominent role that has been played by the Cana-dian government in establishing and promoting the work of the Inter-national Commission on Intervention and State Sovereignty (ICISS) tocreate a new international norm appropriately labelled ‘the responsibil-ity to protect.’ On the other hand, Braybrooke’s challenge is a novelone: students of Canadian foreign policy have looked at the responsibil-ity of the Canadian state, not that of individual Canadians, in addressingglobal problems and in understanding problems as global in scope. In-deed, the extent to which the individual figures prominently is in termsof the shifts in government thinking (but not, so far, in continued gov-ernment action) away from simply making the state and individuals athome secure to making individuals and their communities abroad se-cure. This ‘people-centred’ rather than ‘state-centred’ approach to se-curity is what distinguishes thinking about human security, which theCanadian government has espoused but has failed to pursue fully, fromthe more familiar national security thinking.

My contribution to this volume is reminiscent of Bernard Weiner’sarticle inThe Monistwhich appeared in the same issue as Braybrooke’snoted above. Weiner, a social psychologist, points out that he is not aphilosopher but seeks to bring empirical data that might inform philo-sophical debates regarding responsibility inferences.2 Trained as a po-

1Braybrooke (2003a) 3012Weiner (2003) 165


litical scientist, specifically in international relations and Canadian for-eign policy, I also find value in the intersections between academic fieldsbut seek to do the reverse of Weiner and utilize some of Braybrooke’sphilosophical ideas about responsibility to rethink the notion of humansecurity with respect to its empirical application. This is an importanttask because human security thinking, which can be arguably groundedwithin a framework of cosmopolitan ethics,3 has lost currency withthe Canadian government. Indeed, the turn to human security think-ing could be seen as a harbinger of global justice with Canada, if notin the lead, at least marching in step to create new norms in the in-ternational arena, not merely with governmental and intergovernmen-tal actors but more importantly with non-governmental ones. The 1997ban on anti-personnel landmines, referred to as the ‘Ottawa convention,’stood out as a prime example of the realization of Canada’s human secu-rity agenda. This was followed by the 1998 Rome statute to establish anInternational Criminal Court in which Canada also played an importantrole.

It has been commonplace in the field of international relations to dis-tinguish between the post-Cold War era and the post-9/11 era. It maybe that the post-Cold War era was indeed a time of optimism for statesbecause the national security question became less pressing with theend of the nuclear confrontation between East and West. But that erawhich began with the fall of communism in Eastern Europe in 1989 andquickly followed by the final dissolution of the Soviet Union on Christ-mas day 1991 came to an abrupt end with the events of 11 September2001. The world has changed, it is argued, and in the post-9/11 era, na-tional security trumps all. I make the claim that it is precisely becausewe are in an age of terrorism that human security as opposed to na-tional security thinking becomes even more important. With the UnitedStates in the lead, the focus on retribution as opposed to distribution inthe international arena will only serve to undermine the security of all,including Canadians. Human security is rooted theoretically in a cos-mopolitan ethic, but this is not being realized by current governmentpractices. My central argument is that Braybrooke’s ideas about per-

3See Penz (2001)


sonal responsibility help us to think about how this ethic might be moreeffectively put into practice, and in so doing, point to a rethinking of hu-man security as a shared responsibility between the state and individualsin the society.

The literature on cosmopolitanism, as David Hollinger reminds us,has exploded in recent years in critical and democratic theory.4 SamuelScheffler states that despite different conceptions of cosmopolitanism,at its root it is ‘the idea that each individual is a citizen of the world.’5

The problem, according to Scheffler, is that the idea of world citizenshipcalls into question the normative status of an individual’s attachmentsto particular individuals and groups, so that an apparent incompatibilityarises between an individual’s commitment to equality and his or herspecial responsibilities to family, friends, and communities. However,Scheffler argues that this incompatibility only arises in extreme formsof cosmopolitanism, but moderate cosmopolitanism ‘insists only thatone’s local attachments and affiliations must always be balanced andconstrained by the interests of other people’ and that ‘substantive normsof global justice [exist] in addition to the norms that apply within asingle society.’6

This chapter is based on Scheffler’s conception of moderate cos-mopolitanism which allows for multi-layered citizenship so that an indi-vidual can have ethical commitments to those both within his or her stateand outside of it. Beyond that, I also take as a given that there is state re-sponsibility, particularly if the state is a democratic one.7 The problem

4Hollinger (2001) Hollinger presents ‘the most prominent features of the new cosmopoli-tanism [even] at the risk of rendering the movement more unified that it is’ (236) in order todistinguish cosmopolitanism from Martha Nussbaum’s ‘universalism’ and Will Kymlicka’s ‘plu-ralism.’ An interesting contrast to Hollinger’s review of some of the literature is to be found inNyers (2003). Nyers uses the term ‘cosmopolitanisms’ because ‘to think of cosmopolitanism inthe plural is to upset much of the received knowledge we possess on the subject’ (1072). Nyersprefers the term ‘abject cosmopolitanism’ not only to bring out the ‘familiar us/them relations’ ofWestern cosmopolitan theorists versus the global poor of the Third World, but also to problematizethe cosmopolitanism of the abject to bring out their voice.

5Scheffler (1999) 2586Ibid., 260. Scheffler points out that the ‘conviction that we have responsibilities to our fam-

ilies, friends, and communities is so deeply embedded within common-sense moral thought’ thatthe idea of moderate cosmopolitanism appears to be stating the obvious (262-3).

7According to David Miller, the concept of ‘outcome responsibility’ as opposed to moralresponsibility allows us to talk about national as opposed to individual responsibility, particularly if


is that after leading the way in making individual (personal) securityrather than national security the focus of foreign policy (prime exam-ples being the anti-personnel landmines treaty and the establishment ofthe International Criminal Court), the Canadian government has shiftedback to a national security focus. This has been a retrograde move: itdisregards the radical change in the security problem which has shiftedfrom state-to-state conceptions to the dominance of terrorism and insur-gency within states and the crossing of borders without the reliance onstate sponsorship. This move also disregards the opportunities offeredto deal with the changed security problem through engagement with in-dividual persons in developed states, like Canada, via non-governmentalorganizations (NGOs). These civil society organizations have been ca-pable of mobilizing individual persons in Western states. Braybrookeargues that support of NGOs, or even better fuller participation in them,offers an effective way for individual persons in Western countries toexercise responsibilities toward people in the developing world. His ar-gument is predicated on the recognition that it will be confusing andself-defeating to demand optimal exercise. Indeed, Braybrooke high-lights that individuals in developed countries just doing something now,and then periodically doing more following comparisons with expendi-tures on time and money on luxuries, is what a reasonable ethics will de-mand. Nor is Braybrooke’s argument simply hopeful thinking: the anti-personnel landmines treaty and the International Criminal Court amplydemonstrate how effective NGOs can be in agitating for internationalmeasures to protect individual or personal security. At the same time,Braybrooke is not as expressly mindful as he might have been aboutthe reasons that champions of human development might have had forthinking his appeal to needs is too restrictive. However, his conceptionof needs intersects with Amartya Sen’s conception of capabilities.8 InMeeting NeedsBraybrooke uses a criterion like Sen’s of functionings

the political community is open and democratic. By virtue of enjoying the benefits of membershipin the group, particularly those in the cooperative practice model, dissent alone from the policiesof the group does not allow individuals to evade responsibility for the outcome of those policies.Interestingly, Miller points out that national responsibility is even stronger than state responsibilitybecause the concept of national responsibility holds present generations responsible for the actionsof past ones. See Miller (2004)

8Sen (1999)


(in Braybrooke’s case, as citizens, householder, worker, parent) and,accordingly, includes needs for companionship, social acceptance, free-dom from harassment, and recreation (including also sexual activity).9

This should do something to restore Braybrooke’s credit with champi-ons of human development. In that way, Braybrooke offers a way forus to rethink human security and find avenues for individuals (and notonly states) to exercise their responsibilities abroad.

5.2 The Ethics of State Responsibility

The extent to which human security thinking can be grounded in aframework of cosmopolitan ethics is found in the recognition that thereferent (or subject) of security should be the individual, and that indi-viduals have an equal claim to fundamental human rights, to lives freeof violence, and to minimum necessities of food, shelter, and health.These claims are to be met in the first place by an individual’s fellowcitizens through the operation of their government, but failing that, in-dividuals outside of their country are obligated to meet these claimsthrough their governments. In the parlance of international relationspredicated on the separation of ‘domestic’ and ‘foreign’ policy, humansecurity is first and foremost a domestic matter because it ‘raises theobligations of states to their own citizens’ to protect their rights, ensuresafety from threats both at home and abroad, and provide a minimumstandard of living.10 When states fail in meeting these obligations, itbecomes a foreign policy matter for other states because these statesmust then step in to assist such individuals in need. This stems fromobligations of the citizens of one country to those of another with statesacting as agents for their populations.11

Thinking about security in terms of the individual is a clear depar-ture from the dominant conventional understanding of security in inter-national relations. To the extent that ethical considerations enter intothe state’s calculus, such considerations would apply only to its own

9Braybrooke (1987a)10Penz (2001) 4611Ibid. 43

5.2. THE ETHICS OF STATE RESPONSIBILITY 107

citizens. If ‘politics stops at the water’s edge’12—a phrase that impliesthat national security issues are not subject to debate and the democraticstate must proceed with one voice in the face of the outside world—thenethics is also territorially bounded, and this suggests a Machiavellianunderstanding for rulers when the state’s interests are threatened. Se-curity in this sense is defined primarily in military terms with the statedealing with external threats in an interstate system characterized byself-help and anarchy. National security must be a foreign policy pri-ority in order to ensure the survival of one’s own state. The state mustprocure weapons, build armies, and enter alliances in order to preventforeign attacks. The Cold War that protected Western states, includingCanada, under the American nuclear umbrella against Eastern countries,which were in turn protected by the Soviet nuclear arsenal, was predi-cated on nuclear deterrence. The ‘order’ created by the balance of powerbetween the two sides was also referred to as mutual assured destruc-tion (MAD), given the devastating consequences of a full-fledged warbetween the Soviet Union and the United States. Accordingly, consid-erations of international order necessarily preceded international justicebecause the ‘high politics’ dealing with the survival of the state (andthe planet) in the event of a third world war between East and Westtook precedence over different philosophies regarding the organizationof political and economic life. Essentially, it was a good thing that thetwo sides could co-exist. With each side having second-strike capabil-ity, no first-strike took place; and when the Cold War ended with theWest emerging victorious, a ‘new world order’ infused with liberal op-timism began to renew considerations of international justice and, morespecifically, to reorient the purposes of national militaries.13

As the likelihood of interstate warfare involving Western countriesdeclined in the 1990s, intrastate warfare in many parts of the worldbecame more prominent. From northern Iraq to Bosnia to Somaliato Rwanda to Kosovo, internal conflicts erupted onto the international

12This oft-quoted phrase comes from Michigan Republican senator Arthur Vandenberg (1884-1951), an isolationist turned internationalist, who was a key supporter of the North Atlantic TreatyOrganization (NATO). Vandenberg sought to create a bipartisan foreign policy for the UnitedStates.

13See Rosemary Foot and Hurrell (2003)


scene in ways unprecedented. The distinction between combatants andnon-combatants was no longer respected, and the horror, viciousness,and inhumanity of such conflicts were revealed by media outlets cov-ering conflict zones. Gone were the rules of international humanitar-ian law which governed interstate warfare and protected innocent civil-ians. These were not wars of one country’s armies against another, butrather conflicts arising out of ‘ethnic cleansing,’ warlordism, and geno-cide, where the enemy was the state or specific ethnic groups withinit, as governments or self-proclaimed warlords terrorized, maimed, andkilled citizens within their territories. These violations of human rightsgave rise to humanitarian actions in which foreign militaries intervenedon behalf of innocent civilians and challenged the sovereignty of Iraq,Bosnia, Somalia, and Serbia. But such acts of international interventionwere arguably justified.14 Drawing from the just war tradition, and theethics of the use of force, Jean Bethke Elshtain develops a doctrine ofequal regard as opposed to a model of international victimization as thebasis for ‘the use of force as a remedy under a justice claim.’ She states:‘human beings qua human beings deserve equal moral regard ... [which]means one possesses an inalienable dignity that is not given by gov-ernments.’ When governments commit ‘[a]cts of aggression, whetheragainst [their] own people or against those who cannot defend them-selves, [these] are stipulated as cases of injustice that warrant the use offorce.’15 In essence, these governments had failed in their obligations toprovide human security to their citizens, thus triggering justice claimsby their citizens on other states.

The focus on the rights of individuals is the cornerstone of the ‘re-sponsibility to protect’ doctrine. Created in September 2000 and releas-ing its report in December 2001, the twelve-member International Com-

14The idea of humanitarian intervention is very controversial. As Nicholas Wheeler pointsout, some question the rights of states to risk the lives of soldiers and other non-military person-nel to ‘save strangers’; others point out that the international order among states is predicated ondiffering conceptions of justice and that states, not individuals, are the primary bearers of rightsand duties in international law. Mohammed Ayoob highlights the critiques of postcolonial statesof essentially the actions of the West against the rest. Wheeler provides a powerful case for hu-manitarian intervention, whereas Ayoob carefully examines its major shortcomings.See Wheeler(2000); Ayoob (2004)

15Elshtain (2003) 67

5.2. THE ETHICS OF STATE RESPONSIBILITY 109

mission on Intervention and State Sovereignty, co-chaired by GarethEvans and Mohamed Sahnoun, has shifted international discourse fromthe ‘right to intervene’ by foreign militaries to the ‘responsibility to pro-tect’ individual citizens.16 Although this responsibility is borne, firstand foremost, by governments, thus reaffirming the sovereignty of thestate, governments can no longer do whatever they please to their ownpopulations. The Commission extends the responsibility of protectingvulnerable and innocent populations outside of the state’s borders toother states if the state fails in its obligations or, indeed, is the sourceof the violations. Recognizing that military force is a deadly instru-ment, the Commission sees it as but one element of the ‘responsibilitycontinuum’: the ‘responsibility to prevent,’ the ‘responsibility to react,’and the ‘responsibility to rebuild.’ The focus on the justness of mil-itary intervention—that is, the ‘responsibility to react’—comes aboutbecause the use of militaries for humanitarian purposes is a relativelynew norm, exemplifying human security thinking. Indeed, the person-nel of foreign militaries who may have to fight in order to protect vul-nerable populations are arguably ‘just warriors’ because the actions oftheir governments have been legitimated by the fulfillment of six crite-ria: right authority, just cause, right intention, last resort, proportionalmeans, and reasonable prospects.

The ideal of responsibility of states conceived in this way, however,brings up questions of how it would come into force. One of the primaryreasons for the establishment of the Commission was to rectify the in-action of the international community for the massacres in Rwanda, inwhich 800,000 people were killed as other states stood by and did noth-ing to help the population. The prevention of another Rwanda impelledthe Commission to think ‘realistically’ about how the responsibility toprotect doctrine could become a guiding principle for state action; itled the Commission to place the authority in the hands of the UnitedNations Security Council and to reaffirm the centrality of certain statesand their militaries. In that way, the responsibility to protect doctrine,which would nominally obligate each state, was in reality a principle forlarge states and was therefore not unlike the collective security principle

16See RTP (2001); RTP2 (2001)


already enshrined in the United Nations Charter, which placed respon-sibility and authority in the hands of the five permanent members ofthe Security Council (United States, United Kingdom, France, Russia,and the People’s Republic of China). As Mohammed Ayoob points out,the ‘unrepresentative character of the Security Council and the threat ofveto that hangs over humanitarian decisions makes the Security Councila less-than-satisfactory medium for the determination of internationalwill regarding humanitarian emergencies.’17

Moreover, with the end of the Cold War only one state in the worldhas the military capability to undertake the sorts of interventions thatcould ensure some reasonable amount of success.18 That is why Elsh-tain argues that international justice, ‘construed as an equal claim to theuse of coercive force’ deployed on behalf of individuals victimized bytheir own governments, can only be realized with the United States dis-proportionately undertaking the burden.19 Invoking what she refers toaffectionately as the Spiderman ethic that ‘with great power comes greatresponsibility,’ Elshtain sees the United States, ‘itself premised on a setof universal propositions concerning human dignity and equality,’ asthe only country able ‘to enforce international justice as an equal regardnorm.’20 Although the United States will likely often act in coalitionwith others (but not likely through the United Nations because Elshtainsees the power of veto crippling UN action), American commitment tothe enforcement of the equal regard norm will result, she argues, in thereversal of the lesson of the Melian dialogue: ‘The strong do what theymust in order that the weak not suffer what they too often will.’21

After 11 September 2001, the apparent vulnerability and compla-cency of the United States galvanized a Republican administration un-der George W. Bush to undertake a war on terrorism predicated onthe principle that ‘you’re either with us, or you’re with the terrorists.’

17Ayoob (2004) 10218Some might argue that France and the United Kingdom also have the ability to intervene,

but it is not altogether clear that they would act alone without the United States. The examples ofBosnia and Rwanda stand out in this regard.

19Elshtain (2003) 6420Ibid. 73-421Ibid. 75

5.3. ‘FREEDOM FROM FEAR’ AND ‘FREEDOM FROM WANT’ 111

American military interventions are occurring, as is clearly evident inAfghanistan and Iraq, but these are being justified, not as exercises ofthe responsibility to protect, but as pre-emptive attacks to protect Amer-ican national security from, for example, al-Qaeda, who were being ter-ritorially housed by the Taliban and Saddam Hussein, who apparentlyhad weapons of mass destruction.22 Although the United States govern-ment has argued that the Taliban and Hussein regimes were oppressive,and Afghanis and Iraqis are now better off, humanitarianism was not thereason for going to war. The American government also highlights thatyoung Afghani girls are now in schools after having been denied theright to education by the Taliban, but again the oppression of womenwas not the reason for military action. As Mary Kaldor has pointed out,the Bush administration describes itself as on a moral crusade in which‘sovereignty is conditional for other states, but unconditional for theUnited States because the United States represents “good.”’23 Liberaloptimism regarding the humanitarian purposes of national militaries hasbeen eclipsed as the post-Cold War era has given way to the post-9/11era. The 1990s may simply have been a hiatus, ‘an interregnum, madepossible because Western militaries had spare capacity and time to dohuman rights work.’24 This was a period reminiscent of the two decadesbetween the two world wars, when ‘idealism’ abounded in the interna-tional community. But these periods seem to be anomalies; the normalstate of affairs appears to be one in which responsibilities of states donot extend beyond their own borders.

5.3 ‘Freedom from Fear’ and ‘Freedom from Want’

The responsibilities that I have been discussing have been very narrowin scope. Human security is not merely about the use of coercion to

22Kaldor (2003b) This was a forum on whether the U.S.-led action against Iraq constitutedhumanitarian intervention. A compelling voice in the forum is that of Mary Kaldor, who supportshumanitarian intervention ‘if it is understood as cosmopolitan law enforcement,’ which the waragainst Iraq was not but rather a ‘recipe for more violence and for global polarization.’ Kaldormade these comments as part of the forum. My thanks to Stella Gaon for directing me to this.

23Kaldor (2003a) 1224Michael Ignatieff, as quoted in Tanguy (2003) 148


enforce claims to protection of vulnerable populations. Human secu-rity is a much broader concept that includes meeting the basic needs ofindividuals worldwide and that is predicated on ‘an ethical responsibil-ity to reorient security around the individual in a redistributive sense,’obligating those who have to those who have not.25 This conceptual-ization draws our focus away from world conflict to world poverty, andespecially to the extent to which the world’s resources are controlledand consumed by a small, rich minority who should take responsibil-ity for the impoverished, desperate, and often fatal circumstances of theworld’s majority. This rich minority, of course, lives primarily in theWest, and the problem is that most of us in the West do not feel theweight of responsibility for the world’s poor. Indeed, Thomas Poggeprovides compelling explanations as to why global structural and in-stitutional arrangements that benefit the rich harm the poor, assertingthat ‘most of us do not merely let people starve but also participate instarving them,’ by ‘failing to fulfill our more stringentnegativedutynot to uphold injustice, not to contribute to or profit from the unjustimpoverishment of others.’26 Human security defined broadly encom-passes global economic injustices and includes demands for changes tothe existing socioeconomic order. This more expansive understandingof human security is found in the work of the United Nations Devel-opment Programme (UNDP), which introduced the concept of humansecurity in its 1994Human Development Reportas encompassing eco-nomic, food, health, environmental, personal, community, and politicalsecurity.27 Referred to as the ‘sustainable human development’ view ofhuman security, it is predicated on ensuring ‘freedom from want’ and onfusing the broadened security discourse with the development agenda toaddress fundamental socioeconomic inequalities and the lack of socialjustice in the global arena.28 In this case, the obligations of states arenot solely or even chiefly met by deploying their militaries for humani-tarian purposes but by using development assistance, trade, investment,monetary, and other policies to address global economic injustices.

25Newman (2001) 24026Pogge (2002) 214, 19727Programme-UNDP (1994)28Hampson (2002)


The Canadian government has focused its efforts on the safety ofpeoples, defining human security along the lines of ‘freedom from fear’as opposed to ‘freedom from want.’ It has identified five areas of prior-ity in its human security agenda: (1) protection of civilians; (2) peacesupport operations; (3) conflict prevention; (4) governance and account-ability; and (5) public safety. The government has argued that humansecurity thinking is not merely the development agenda in a new guisebut rather a recognition that the nature of conflict has changed in fun-damental ways: wars are being fought within states as opposed to be-tween them; threats emanate not merely from military but also fromnon-military sources, such as human rights violations, terrorist acts, en-vironmental degradation, transnational crime, and small arms prolifera-tion; and the safety and well-being of individuals (as opposed to merelystates) are integral to achieving global peace and security.29 The focusmust be on the human costs of violent conflict and the means to remedythem. That is why the Canadian government became the principal spon-sor of the ICISS when UN Secretary-General Kofi Annan saw the issueof humanitarian intervention as too divisive within the United Nationsframework.30 In so doing, the ICISS report is a direct reflection of ‘theCanadian government’s efforts to influence normative developments ininternational society ... (and indeed) a great deal of the Commission’slanguage and concepts reflect the human security agenda that was soprominent a part of Canadian foreign policy in the 1990s.’31

Human security efforts on the part of Canada also reflect a changedpolitics by working with civil society actors in the global arena. For ex-ample, the advancement of the treaty on anti-personnel landmines andthe Rome statute establishing the International Criminal Court couldnot have happened without the participation of key NGOs, such as theInternational Campaign to Ban Landmines (ICBL) and the NGO Coali-tion for an International Criminal Court (CICC), which were themselvesacting as umbrella groups for hundreds, if not thousands, of smallerNGOs.32 For Canada, working with NGOs and other like-minded states

29SFP (1999); FFF (2002); McRae and Hubert (2001)30Axworthy (2003) 20231RTP (2001) 49132McRae and Hubert (2001)


became central to advancing its human security agenda. This was adifferent type of politics founded on rethinking state-society relationsin which the process was opened up to groups that had been tradition-ally marginalized in foreign policy development: peace activists, socialjustice activists, women’s groups, cultural groups, minorities, students,and so forth. Many of these groups found common cause with the gov-ernment’s human security agenda as the government adopted notions ofeconomic justice, gender equality, and human rights in its policies. Inthat way, the revisioning of security along human security grounds in-cluded changing the process to bring in societal voices, if not values, tothe making of state policy.

Civil society engagement is crucial for bringing about change be-cause NGOs represent the realignment of political identities that arebased not simply on national but also on transnational attachments. Forexample, although the United States remains a non-signatory to the anti-personnel landmines convention, it was an American, Jody Williams,who co-founded the U.S.-based ICBL as six NGOs in 1992 and ex-panded it to over 1300 members in over 85 countries; who fought tire-lessly to ensure a successful treaty by working with other NGOs, likethe International Committee of the Red Cross (ICRC), and with states,like Canada;33 and who won the Nobel Peace Prize in 1997 for her in-ternational humanitarian efforts. Similarly, although the United Statesvoted against the Rome statute to establish the International CriminalCourt, it was also an American, William Pace, the executive directorof the New York-based World Federalist Movement-Institute for GlobalPolicy, who headed the CICC and who approached then foreign affairsminister Lloyd Axworthy to enlist Canada’s support for the establish-ment of a new court to hold individuals responsible for war crimes whennational judicial systems fail.34 Thus, human security ‘reflects a politicsof empowerment’ brought about by the spread of democracy, the growth

33As Mark Gwozdecky and Jill Sinclair point out, the anti-personnel landmines treaty suc-ceeded because of civil society leadership, highlighting that ‘paragraph eight of the preamble ofthe Ottawa Convention gives formal recognition to the ICBL and ICRC for their efforts, an honournot normally granted in a treaty instrument developed by a state-focused international community’(37). See Gwozdecky and Sinclair (2001)

34Robinson (2001) 172


of information technologies, and the impact of non-governmental orga-nizations, particularly in the humanitarian and development fields, ondomestic and global issues.

What has become of Canada’s human security agenda? On 27 April2004, the Liberal government under Prime Minister Paul Martin re-leasedSecuring an Open Society: Canada’s National Security Policy.35

Unabashedly realist in tone, the document identifies threats to the Cana-dian state with a sense of urgency and impels the government to act.The document spells out distinctions between ‘personal security,’ ‘na-tional security,’ and ‘international security’36 in order to bring out thecentrality of national security. Only when personal and international se-curity issues have an impact upon national security do these two mattersbecome important. The government states explicitly: ‘National securitydeals with threats that have the potential to undermine the security of thestate or society,’37 thus invoking, as evident throughout the document, amore traditional conception of security predicated on the state as the ref-erent. The concept of human security has apparently disappeared fromCanada‘s lexicon. Precipitated by the events of 11 September 2001,this shift to national security language by the Martin government hasled to the identification of threats ’at home.’ This has meant that theDepartment of Public Safety and Emergency Preparedness, essentiallyCanada’s equivalent to the U.S. Department of Homeland Security, hasdisplaced the Department of Foreign Affairs38 and has taken the lead indefining Canadian security.39

I would argue that the apparent disappearance of human securitylanguage in the current government’s thinking stems from the miscon-ception that national security and human security do not go hand inhand. Domestically, the government’s function in ensuring freedom

35SAOS (2004)36Ibid., 437Ibid., 338When Paul Martin took over as prime minister, the Liberals undertook major reorganizations

of federal departments. Apart from creating the Department of Public Safety and EmergencyPreparedness, the government also separated out International Trade from Foreign Affairs, creatingtwo new departments.

39I take up what this means for Canadian foreign policy and Canada-U.S. relations in Keeble(2004).


from fear on the part of its citizens goes alongside its function in en-suring freedom from want. Ensuring freedom from fear stems fromthe government’s coercive power: at home, that function is undertakenby the police and other enforcement authorities. Abroad, as we see inthe responsibility to protect doctrine, that function is undertaken by themilitary in protecting individuals. Ensuring freedom from want stemsfrom the government’s social welfare responsibilities: at home, this isoften referred to as ‘social security’ or ‘income security.’ Abroad, thisis manifested in terms of the government’s foreign aid policies. Theproblem is that despite the recognition that conflict may be rooted indeep-seated socioeconomic inequalities between groups, and that ad-dressing these ‘root causes’ is fundamental to conflict prevention, theCanadian government has seen its human security agenda in a very nar-row way. By continuing to make distinctions in human security think-ing between ‘freedom from fear’ and ‘freedom from want,’ the govern-ment does not fully address the ‘development’ side of the equation andthe responsibility therein. What the government does not appreciate isthat policies to ensure ‘freedom from fear’ abroad are actually aboutrestoring the national security of a state by other states like Canada. Totruly address its responsibilities to individuals abroad, the governmentmust ensure ‘freedom from want.’ It is true that the record for donorstates like Canada in alleviating world poverty, reforming internationalfinancial institutions, renegotiating trade arrangements, crafting invest-ment policies, and the like is not good. Since the end of the SecondWorld War, official development assistance (ODA) policies, includingthose of Canada, have had less to do with humanitarian objectives andmore with political and economic ones, despite the rhetoric to meet ba-sic human needs and help the poorest of the poor. That is why CranfordPratt’s recent survey of Canadian foreign aid policy finds little role forcosmopolitan ethical considerations despite the Canadian government’scommitment to human security, particularly during Lloyd Axworthy’stenure as minister of foreign affairs from January 1996 to September2000.40 The Canadian state has clearly not lived up to its responsibilityto individuals abroad. How might we revive attention to human secu-

40Pratt (2001)

5.4. THE ETHICS OF INDIVIDUAL RESPONSIBILITY 117

rity in Canadian foreign policy thinking with the effect of giving greaterweight once again to what person-to-person responsibility can amountto in the international arena?

5.4 The Ethics of Individual Responsibility

A modest proposal for rethinking human security is to locate respon-sibility at the level of the individual as well as at the level of the stateby drawing from the work of David Braybrooke on meeting needs andtaking personal responsibility for extending the common good beyondstate boundaries.41 Braybrooke provides philosophical arguments forthree sorts of claims which, I will argue, may help to revive human se-curity in Canadian foreign policy thinking by placing individual Cana-dians front and centre in our understanding of responsibility. I addressBraybrooke’s arguments here and then conclude in the last section onwhat his arguments would mean for Canadian foreign policy.

Braybrooke provides philosophical arguments for the following threesorts of claims:

1. The concept of needs has moral force for all individuals who livein rich, safe countries of the West, and some choose to accept thisclaim.42

2. These individuals take on responsibilities in relation to social or-ganizations, and in this case organizations dedicated to globalbeneficence. As private persons, individuals take on these respon-sibilities by supporting non-governmental organizations, such asAmnesty International or Medecins sans Frontieres, as well asspecific agencies of the United Nations, such as the United Na-tions Children’s Fund (UNICEF).

41 My specific points of departure from David Braybrooke’s voluminous work are ‘A Progres-sive Approach to Personal Responsibility for Global Beneficence’ (cited earlier), ‘The CommonGood beyond State Boundaries’ (October 2003), made available to me in manuscript by DavidBraybrooke, pending publication, and ‘Where Does the Moral Force of the Concept of NeedsReside and When?’in Braybrooke (2005).

42Braybrooke makes this claim, I would argue, explicitly in his ‘Where Does the Moral Forceof the Concept of Needs Reside and When?’ and implicitly in his ‘A Progressive Approach toPersonal Responsibility for Global Beneficence.’


3. Such individuals have taken on the sorts of responsibilities thatmove in the direction of a world community.

In ‘Where Does the Moral Force of the Concept of Needs Resideand When?’ Braybrooke juxtaposes his concept of needs with AmartyaSen’s concept of capabilities. I discussed in the last section that it wasthe UNDP that first introduced an expansive understanding of human se-curity by conceptualizing it in terms of sustainable human development.The UNDP’s human development approach, encapsulated in itsHumanDevelopment Reportpublished annually since 1990, is predicated onSen’s ideas on capabilities: ‘the idea that the purpose of development isto improve human lives by expanding the range of things that a personcan be and do, such as to be healthy and well-nourished, to be knowl-edgeable, and to participate in community life.’43 Sen’s influence ondevelopment thinking has been pronounced,44 and his capabilities (alsoreferred to as ‘freedoms’) approach has clearly informed understand-ings of human security as ‘freedom from fear’ and ‘freedom from want.’Indeed, Sen, along with Sadako Ogata, former UN Commissioner forRefugees, co-chaired the Commission on Human Security, which pre-sented its final report to United Nations Secretary-General Kofi Annanin May 2003.45

Braybrooke argues that the needs and capabilities approaches shouldbe combined, but that the concept of needs has greater moral force evenwith those who might be less willing or indeed outright sceptical ofgiving aid to those less fortunate than themselves. He reasons that indi-viduals will be more willing to place others’ needs ahead of their ownpreferences but most likely quite unwilling to place others’ preferences

43Fukuda-Parr (See 2003, pg. 303)44Stewart and Deneulin (2002)45The Commission on Human Security was established under the leadership of Japan. Whereas

Canada took the initiative in establishing the International Commission on Intervention and StateSovereignty, as discussed earlier, with the focus on redefining humanitarian intervention—and thuswithin the context of ‘freedom from fear’—Japanese leadership with regard to the Commissionon Human Security signalled that country’s commitment to a broader understanding of humansecurity. Not surprisingly, the Japanese government asked Sen to co-chair the Commission inorder to help bring out the fuller ‘freedom from want.’ For more on the difference between theCanadian and Japanese positions on human security, see Acharya (2001). The final report of theCommission on Human Security is available online athttp://www.humansecurity-chs.org/finalreport .


in terms of leading to further capabilities (e.g., ‘enlargement of theirchoices’) before their own. Braybrooke points out that needs, often notas expressed as such but rather in ‘surrogate’ ways such as thirst, starva-tion, exposure, or illiteracy, point to a sense of urgency that often movespeople to act by essentially ‘pulling on their heartstrings.’ Indeed, needsexpressed in this way tend to eliminate considerations of desert becausethe needy could not possibly be responsible for their present situations.That may be why Sen’s capabilities approach, which does allow forgreater agency on the part of the needy, may be less appealing to thosewho would address the need. Given the international disparities in theglobal context, we are talking about persuading those in the rich, safecountries in the West, like Canada, to take responsibility for meetingthe needs of those abroad. Every human being does have a right to betreated the same, but the ‘treatment’ (e.g., food, shelter, health) underconsideration must be seen in light of those who need it and those whohave it and also can provide it to others. The argument for the moralforce comes in when we focus not on those in need but on those whocould respond to the need.46

In what ways do individuals in rich countries like Canada take re-sponsibility? In ‘A Progressive Approach to Personal Responsibilityfor Global Beneficence,’ Braybrooke argues that individuals take on re-sponsibilities in relation to social organizations. He focuses on charita-ble organizations as opposed to governments or corporations in settingup a climate that encourages individuals to give. Braybrooke assertsthat, given an inclination to give (particularly evident when the moralforce of needs is invoked), an individual will assign herself or himselfpersonal responsibility to act beneficently, not in terms of a duty be-

46This statement will appear counter-intuitive to international development specialists whohave seen the change in approach to development from ‘basic needs’ to ‘human development.’Although the basic needs approach places people at the centre of development, it has been seen asdefining human well-being in terms of commodities as opposed to capabilities, therefore denying(or being silent on) human rights, freedoms, and agency of those in the developing world, placingthem in a further subservient position to donors. See Fukuda-Parr for comparisons between theneoliberal, basic needs, and human development approaches. However, as I stated in the introduc-tion, by bringing Braybrooke’sMeeting Needsinto the discussion, we can see more clearly howBraybrooke’s and Sen’s conceptions intersect, thus restoring in part Braybrooke’s credibility withchampions of human development.


ing imposed, but rather in terms of ‘rules that the agent makes up forherself [or himself], for instance, rules about the size of current contri-butions.’47 The point that Braybrooke makes is that individuals livingin a rich country such as Canada, which has a tradition of philanthropylike the United States (but unlike Sweden or Japan), may be (ethically)guided to give in order to help remedy global inequalities. Thus, asprivate persons connected to organizations which work toward aspectsof the global good, individuals may be encouraged to continue giving,and indeed giving more as years progress by simply asking them ‘to domore—not optimally more, just more.’48 According to Braybrooke, itmight be that individuals would find, for example, that a second sportsutility vehicle is simply a luxury that they do not need, and they canredirect such monies to contributions the following year, setting thestandard for additional beneficence. Given active members of a tra-dition of philanthropy, some will give more and others may drop outbut new contributors will join, so that with each succeeding round morewill be given to charitable organizations such as Amnesty Internationalor Mdecins sans Frontires or specialized agencies of the United Nations,such as UNICEF. Braybrooke states that ‘on the progressive approach,there is no neat end to the question about responsibility and the size ofcontribution’ because the problem of global disparities never goes away,forcing those of us living in rich countries ‘to bear with the uncomfort-able feeling that we are each of us falling short of doing what we coulddo.’49

Moreover, by taking on these responsibilities, individuals are par-ticipating in ways that move in the direction of a world community.In ‘The Common Good beyond State Boundaries,’ Braybrooke pointsout that NGOs and semi-autonomous UN agencies, particularly thoseunder discussion, are pursuing aspects of the world common good by,for example, providing clean water, increasing literacy, redressing hu-man rights violations, or instituting peace processes. When individ-uals support these organizations, they help to bring about a sense of

47Braybrooke (2003a)310, 316.48Ibid., 31249Ibid. 315-16


the common human good beyond their own state, and ‘if participationcould be strengthened ... we would have on the international scene ro-bust elements of a world community, people organized to pursue publicgoods on the world scene.’50 This would necessitate what Braybrookerefers to as a movement from ‘the mailbox relation to active commu-nity practices,’51 predicated on a multi-stage program that would in theend bring together donors, the national staff of the NGO, staff in thefield, and recipients, creating robust attachments. Braybrooke acceptsthat our special responsibilities to families, friends, and fellow nationalsmay work alongside our general responsibilities to others, particularlythose in need, throughout the world. He states that the ‘multiplica-tion of personal activity,’ along with the works of NGOs, would helpto bring into fruition Samuel Scheffler’s conception of ‘moderate cos-mopolitanism,’52 predicated on new allegiances which transcend statesand national cultures. This may unsettle the self but in the context ofcreating amelangeof identities that allows the individual to flourish.53

In the end, individuals taking responsibility for global beneficence bysupporting organizations that work for the world common good maybe more likely to pressure their government to act in the same way.As Braybrooke states, ‘Judiciously developed, these motivations mightachieve a delicate balance between private giving and support for gov-ernmental beneficence, so that the very people who gave most gener-ously privately would be most vigorous in supporting humanitarian po-lices on the part of the government.’54

50Braybrooke, ‘The Common Good.’51Ibid.52Braybrooke applauds Scheffler’s work as do I. According to Scheffler, in ‘Conceptions of

Cosmopolitanism,’ ‘moderate cosmopolitanism about justice will be a compelling position only ifit proves possible to devise human institutions, practices and ways of life that take seriously theequal worth of persons without undermining people’s capacity to sustain their special loyalties andattachments’ (275).

53Braybrooke also applauds Jeremy Waldron’s work. Waldron fears the ‘cultural exclusivenessof the identity politics of community’ (113-14) and argues instead for the identity of the self predi-cated on the management of a ‘variety, a multiplicity of different and perhaps disparate communalallegiances’ (110). See Waldron (1995)

54Braybrooke (2003a) 318


5.5 Rethinking Human Security

The connection that Braybrooke makes is crucial to rethinking humansecurity: individuals matter as responsible agents moving in the direc-tion of a world community. By supporting private charitable organi-zations and UN semi-autonomous agencies working towards the worldcommon good, they fulfill their moral obligations not only as privatepersons but also as citizens to others outside of the state. They maydemand that more of their taxes be directed to foreign aid, or more ofthe state’s efforts be placed in reforming global financial structures thatfavour rich countries. If human security thinking is truly to be groundedin a cosmopolitan framework, Braybrooke’s arguments bring out thatwe should be paying more attention to the relationship between indi-viduals and non-governmental organizations as opposed to simply therelationship between states and NGOs.

Braybrooke’s arguments present a challenge to students of Canadianforeign policy who see the impact of the Canadian population on thecountry’s international affairs as negligible. The government is able tomake foreign policy according to its own objectives and goals so thatwhen the government sets up public consultation processes or, in thespecific case of the Liberal party when it was elected into power in1993, institutes a ‘democratization’ of Canadian foreign policy, theseefforts are merely to legitimize the state’s interests not those of the so-ciety. Those interests may play out in terms of vying for prestige onthe part of governmental actors (e.g., Axworthy wanting to win a NobelPeace Prize) or promoting the government’s economic agenda under asmokescreen of humanitarianism. The point is that the government isin control of foreign policy processes so that even when civil societyactors play a more prominent role, they are subject to co-optation, and,in that way, their cosmopolitan understandings may not have full force.The maintenance of national attachments is central to when the govern-ment acts. This helps to explain why former Foreign Affairs MinisterAxworthy, the strongest proponent of human security in Canadian for-eign policy, had seen it not only as ‘new way forward’ but also a way ‘to

5.5. RETHINKING HUMAN SECURITY 123

maintain Canada’s role as a leading voice on the world stage,’55 a voiceseparate from that of the United States. Axworthy stated explicitly that‘the best Canadian foreign policy remains an independent policy’ andthat ‘Canadians want their country to be more than a junior partner tothe United States.’56

However, Braybrooke’s arguments bring out that conceptualizationsof human security should include relationships between individuals andcharitable organizations. By locating some degree of responsibility atthe level of the individual in this way, we begin to see that Canadi-ans through their actions may continue to keep alive the goals of hu-man security despite the government’s apparent reluctance to make it apriority—such reluctance seems to be a response either to pressure ex-erted by the United States or to its own (mis)perception of greater threatsat home. Ensuring ‘freedom from fear’ and ‘freedom from want’ can beachieved both abroad and at home. National security may not trumpall if Canadians as private persons continue to take on responsibilitiesthat move in the direction of a world community. Many of the organiza-tions that are part of the Canadian Council for International Cooperation(CCIC), an umbrella organization of approximately one hundred Cana-dian non-profit organizations working to end global poverty and achievesustainable human development, receive support from the contributionsof individual Canadians. If human security is to be revived in Canadianforeign policy, we must see that responsibility is a shared undertakingbetween the state and individuals and that the individual’s connectionto the world common good through NGOs is perhaps the path to cos-mopolitanism.

55Axworthy (2001, 2003)56Axworthy (2001) 8

Chapter 6

Braybrooke on Public Policy:Precautionary and Fair; Feasibleand Ameliorative

SHARON SUTHERLAND

Abstract

David Braybrooke’s continuing exploration of how the strategies underwhich public policies are formulated might be made more responsiveto need, and more successful, is the subject of this chapter. The pa-per situates his work on decision-making alongside two other influentialcontributions: first, the synoptic vision, alternatively called the ‘compre-hensively rational’ approach, the ‘economics model,’ and the rational-deductive model; and, second, Herbert Simon’s boundedly rational ap-proach. Both alternatives apply only to decision-making by individuals.The paper then takes up incrementalism in the hands of Charles E. Lind-blom and Braybrooke before examining Braybrooke’s overall dynamicstrategy that provides for decision-making by collectives (as well as byindividuals), as set out in his 2004 book on utilitarianism. The papercloses with two recent examples of successful incremental policy in traf-fic management in Britain and France, a subject that has long interestedBraybrooke as a natural object for incrementalist policy strategies.∗

∗David Braybrooke is the most collegial of colleagues. I want to thank him for his ready anddiscerning assistance with sources, for his willingness to share ideas, for solving my puzzles, andfor his gifts of excitement, patience, and encouragement for thirty years.

125

126 CHAPTER 6. SHARON SUTHERLAND

6.1 Introduction to Braybrooke’s Strategy

This chapter is an analysis and appreciation of David Braybrooke’s con-tinuing exploration of how to improve the strategies under which publicpolicy is formulated in regard to their responsiveness to needs and theirattentiveness, in deliberative stages, to intelligibility, feasibility, andpublic safety. HisUtilitarianism: Restorations; Repairs; Renovations(2004) recommends an incrementalist-utilitarian policy and decision-making strategy that would encourage citizen examination and discus-sion of policy problems, interventions, and outcomes—activities thatcould improve and amend appreciation of the problem environments.He continues his ‘renovations’ in his further work on needs, rights,and consequences (Analytical Political Philosophy, 2005), in which heshows how the concept of needs can inform deliberation.

In an encyclopedia entry on decision-making systems for individualsand collectivities, Braybrooke notes that there is a public good in havinga political system that meets at least the minimal conditions for popularparticipation and control. Control by citizens can only come about ifthey participate; if they see their opportunities and feel it their duty tolearn about and correct ‘short-sighted policy results’ from the market orfrom political bargaining:

A polyarchal system of decision-making will work as in-tended by its advocates only if this control in all its aspects iseffective (an empirical condition) and a large portion of thepopulation participates (a definitional condition ...). But nosystem for decision-making will work unless the people whohave parts to play in the system actually take part. (Bray-brooke (2000), 3316.)1

His strategy is constituted of his prescription to increase participa-tion in deliberation, plus his alternative to the dominant approaches todecision-making in policy literature.2 This latter is an alternative in sev-

1Polyarchy is a term originated by Robert Dahl to describe political regimes of the real worldthat are, comparatively, inclusive and open to public contestation. See Dahl (1971), 7-9).

2‘Does Utilitarianism Require Perfect Information about Consequences, Leaving Coordina-tion Problems Aside? No.’ Braybrooke (2004), chap. 2).

6.1. INTRODUCTION TO BRAYBROOKE’S STRATEGY 127

eral ways. First, it is a decision-making approach for collectivities andfor individual analysts, whereas the alternative decision-making meth-ods are all designed for individuals. Second, it proposes that policychoices at first stage should meet needs and that the evaluative standardshould be whether needs have been met. (Only after needs have beenmet should policy move to supporting ‘commodious living.’)3 Third,the Braybrooke strategy is able to incorporate, as usable ‘message,’ allcontributions short of violence (although civility is to be desired as avalue in itself). Therefore those participating and observing can learnfrom expressions of emotion, evidence-based or logically well-informedinferences, and from experimental exploration. Fourth, it should be ex-plicitly brought to mind that Braybrooke’s strategy, in use in deliber-ation at either individual or collective levels, is dynamic, proceedingin stages and making use of new information in evaluation and furtherrounds of deliberation and decision.

Braybrooke is a theorist of participation in the developmental democ-racy tradition, seeing participation as having as much substantive util-ity for the collectivity as for the individual, with particular insight intohow discussions develop concepts. Participation promotes the increas-ing intelligibility of policy alternatives afoot in the collective discus-sion, which makes democracy fuller, and it meets individual demandsfor recognition and reciprocity as well as developing individual capacityfor deliberation.4

It is perhaps important to communicate to a reader without a back-ground in empirical political science or policy studies that contributionslike those made jointly and singly by Braybrooke and Lindblom—workthat is attentive and empirical without being scientistic—are importantadditions that are also corrective. Braybrooke’s confidence in the gen-eral public, although ‘inexpert’ in subjects of which it has little experi-ence, is remarkable, not least in contrast to a stream of American ‘sci-entific’ political science that was mainstream in what we can now seeas the first half of Braybrooke’s career. That is the so-called ‘empirical

3The term is from the last chapter ofUtilitarianism, Restorations, Repairs, Renovations.4See hisMoral Objectives, Rules and the Forms of Social Change, particularly chap. 4, ‘The

Meaning of Participation and of Demands for It,’ 54-84.


democratic theory,’ perhaps most clearly represented by B.R. Berelsonand Herbert McClosky, which held that the undereducated and poorlyinformed probably could not be safely brought into the polity, and fre-quently argued that citizen apathy was functional for the system, theopposite of Braybrooke’s position. The ‘average’ citizen was found inempirical studies of democratic beliefs in mass populations to be in-efficacious, anomic, lacking information, alienated, submissive, cyni-cal, low in citizen duty, and irrational. In Herbert McClosky’s words,‘Democratic viability is ... saved by the fact that those who are mostconfused about democratic ideas are also likely to be politically apa-thetic and without significant influence. Their role in the nation’s de-cision process is so small that their “misguided” opinions or non opin-ions have little practical consequence for stability’ (McClosky (1964),376; see also Berelson (1952), 313-30). On the other hand, the 5 to10 percent of the population that makes up the activists, drawn over-whelmingly from educated and well-off sectors of society, had almostuniformly positive personality traits, as dubiously measured by the testsof the era. A typical item of a standard five-item test of political effi-cacy (said to predict participation) is ‘People like me don’t have any sayabout what the government does.’5 This statement might appear onlyrealistic to an individual who had never been brought into public life inany capacity.

In strong contrast, Braybrooke has developed a plan for democraticdeliberation as a part of policymaking. He is confident that interactionbetween people in general, in the ‘received’ formal institutions of gov-ernment, or in spontaneous groups that form around problems, will im-prove the intellectual quality and feasibility of policy proposals. Partic-ipation surely has the effect of supporting the received decision-makinginstitutions, but it can also importantly correct and deepen what they do.Braybrooke believes that participatory processes work best for policies

5See, for example, Robinson et al. (1972), particularly chap. 12, ‘Attitudes toward the Po-litical System,’ 441-82. L.W. Milbraith and M.L. Goel Milbraith and Goel (1977) present anexcellent summary of this literature. For another view, also empirically based, see Sutherland andTanenbaum, ‘Submissive Authoritarians’ Sutherland, S.L. and Tanenbaum (1980) and also their‘Irrational vs Rational Bases of Political Preference: Elite and Mass Perspectives’ (Sutherland,S.L. and Tanenbaum (1984), 173-97).

6.1. INTRODUCTION TO BRAYBROOKE’S STRATEGY 129

for which local authorities have jurisdiction, such as traffic manage-ment, as opposed to policies like air pollution that exhaust or exceed theresources and authorities of the jurisdiction nominally dealing with thatissue, but even there he believes that participation is feasible.

It is also important to take into account how pervasively if fantas-tically the synoptic or rational-deductive ideal once saturated fledglingpolitical science, with lasting impact in the policy area. Here the workof Braybrooke, Lindblom, and Simon is importantly corrective. Synop-tic ambition was particularly striking after the early computers seemedto promise unlimited, instant, and cheap computational labour, and sys-tems theory modelled on biological models took hold. One sample ofHarold Lasswell’s thinking, as he laid it out in his 1930Psychopathol-ogy and Politics, expresses a typical form of confidence:

As it is today we are on the threshold ... of assessing the re-alism of the intelligence maps upon which people base their[political] choices. As an aid in this huge enterprise it willbe essential to improve the coverage and the depth of thecurrent flow of information that appraises unconscious aswell as conscious factors throughout the world social pro-cess. For the United States the following questions suggestsome of the major points to be covered ... Is the weight ofthe super-ego becoming less severe upon Americans than itwas a generation ago? ... Is the rate of change different inNew England, the Middle Atlantic States, and other regions?(Lasswell (1968), 318)

In an earlier article, ‘Chicago’s Old First Ward: A Case Study inPolitical Behavior,’ (Lasswell (1923), 127), Lasswell wanted to firstdescribe and then predict human political behaviour as a first step tocontrol of political behaviour.6

6See Robert Horowitz’s essay, ‘Scientific Propaganda: Harold D. Lasswell,’ in Storing (1962).Horowitz cites Bernard Crick’sThe American Science of PoliticsCrick (1959) on Lasswell’s rest-less futurism: ’Professor Lasswell is not a settler, he is a pathfinder and a border scout ... as thewagons stop moving and the ploughs are dragged out, he always gallops off to yet more virginsoil, leaving the Social Science Research Council to organize a Territorial Government’ (180-1 inCrick; 127 in Horwitz).


While not everyone was interested in mapping and regulating rel-ative weights of the collective superego from one region to another,the same idea that cheap computational power could generate power-ful predictive theory from empirical observations spurred speculativework in social indicators theory and systems theory. Both social indi-cators theory and the most comprehensive systems theory were takenup in government through the 1960s and 1970s. In the managementand allocation aspects of public policy work, this era saw trials acrossthe industrialized world of ‘planning, programming and budgeting sys-tems’ (PPBS), which were abandoned and modified without being thor-oughly understood by practitioners. PPBS is by design synoptic andcentralizing, intended to increase productivity in hierarchical programsbut also to discern and manage ‘horizontality,’ including redundant pro-gramming, all based in a fantastic efficiency calculus of both nested andhorizontal cost-benefit results.7 The thinking behind the development ofPPBS and its comprehensive, scientific, results-oriented style of evalua-tion (to inform budgetary decisions) is still important in the public pol-icy process, although many practitioners assume it has been replaced bythe New Public Management. But the New Public Management, whichdominates the modern industrialized world, more nearly incorporatedPPBS than it superseded it, and one sees synoptic assumptions under-lying the haphazard prescriptions and practices in results measurementin contemporary federal Canada. Because hazard and small ‘p’ poli-tics by administrators can fill in where the synoptic method invariablyfails—and in principle must fail—political space is lost in public policyareas.

The general purpose of this paper is to put Braybrooke’s highly orig-inal contribution into the context of what I take to be, as a student of gov-ernment, the most influential work on how decisions are and should bemade. I present, first, an indication of what it is to try to be synopticallyrational in decision-making and then move to the first developed alterna-tive to synoptic or comprehensive rationality: Herbert Simon’s idea thatanalysts and decision-makers could make efficient strategic choices in

7On PPBS, it is difficult to best Herman van Gunsteren’sThe Quest for Control: A Critiqueof the Rational-Central-Rule Approach in Public Affairs(van Gunsteren (1976)).

6.2. APPROACHES TO MAKING PUBLIC POLICY 131

information seeking, leading to tolerable outcomes as opposed to max-imizing. Incrementalism is the only really seriously developed alterna-tive to satisficing, and Charles Lindblom and David Braybrooke are thedominant figures in incrementalism. In the parts of its development thatthey do not share, their paths are complementary rather than contradic-tory. Braybrooke’s ‘strategy,’ drawn from chapter 2 of his 2004 book onutilitarianism, is summarized to give the reader a fair idea of its com-prehensiveness and richness. The paper then closes with brief examplesof successful incremental policies in an area of Braybrooke’s interest—traffic congestion—contrasted with policy failures that attempted majorchange in one concentrated effort.

6.2 Approaches to Making Public Policy

For more than half a century, it has been broadly accepted by those whowork in and write on public policy that the acceptable recommendationsfor how policy should be made are the synoptic or comprehensively ra-tional method and Herbert A. Simon’s recommendation of ‘satisficing’or bounded rationality. Both are relevant to the single decision-makeror analyst, not to collectives. Because much of the synoptic prescriptionhas been implicit, each of Simon, Braybrooke, and Lindblom has felt itnecessary to describe its general requirements.8

6.2.1 The Rational-Deductive, Comprehensively Rational,Synoptic, or Economics Model

Brian Fry, in hisMastering Public Administration(Fry (1989), 189-90), finds in Simon himself a fairly complete version of the ‘classic’ setof requirements for the individual rational decision-maker who wouldmaximize chosen or assigned values. To select the alternative that willbe followed by the most desired (valued) array of consequences, theanalyst knows at least the following:

all the relevant aspects of the decision environment;

8Braybrooke does credit P.H. Wicksteed’sThe Common Sense of Political Economy(Wick-steed (1910)) with a paradigmatic formulation of rational-comprehensive reasoning but obviouslycannot rely on reader familiarity with it.


all the alternative courses of action;

all the consequences of those alternatives, or, the probability dis-tributions for the consequences;

a stable rank ordering for all possible consequences based on astable set of values, such that all consequences can be rank orderedand that ordering will endure; and

all the relevant computational routines.9

In their bookA Strategy of Decision(Braybrooke and Lindbloom(1963)), Braybrooke and Lindblom identify another feature of the ‘ideal’synoptic policymaking process. Citing Tinbergen’sOn the Theory ofEconomic Policy(Tinbergen (1952), 1), they note that the decision-maker or policymaker must ‘accept the responsibility of coordinatingthe policy to be chosen with other policies’ (38). They also acknowl-edge a similar point made by Marshall Dimock in hisA Philosophy ofAdministration(Dimock (1958), 140) that the synoptic approach shouldtake into account larger institutional goals and objections.

In Strategy, Braybrooke and Lindblom adopt the term ‘rational-deductivemethod’ for the full-scale, one-off intellectual attack on a problem. Theyimagine the form of its precision:

Ideally, the system [for ranking possible policies] would becomplete, not necessarily in the sense that it mentioned ev-ery contingency but in the sense that its ultimate principleswere rich enough to supply the intermediate principles orthe sequence of intermediate principles that would decideany case that might come up. With such a system, the un-certainties of evaluation would have been mastered on thevalues side. For, on the values side, determination of policybecomes simply a matter of calculation, a question of feed-ing in the observed facts and thinking consistently through a

9Fry draws from Simon’sModels of Man(Simon (1957), 241), March and Simon’sOrga-nizations(March and Simon (1994), 138), and Simon’s ‘Rational Decision Making in BusinessOrganizations’ (Simon (1970), 500).


sequence of transformations. One discovers the facts, looksup (or derives) the relevant hypotheticals, and deduces bystrict logic which policy is to be selected. (10)

Braybrooke and Lindblom remark on the apolitical and, indeed, anti-political nature of accepting the prospect that one analyst or a singleteam could ever execute an authoritative ‘single-pass’ solution of a pub-lic policy problem. The rational-deductive ideal, if it worked, or even ifit were only believed to work by enough people, would transfer an idealof science to the field of values and constitute an attempt to determinethe content of the public interest (9, 16).

They identify the important difficulties in intendedly comprehen-sively rational prescriptions such as crude utilitarianism: we have nointerpersonal calculus; hence we will have no agreement on conclusive-ness; we have no rule for drawing accounts of consequences to an end;and we have no rule for identifying reference groups for whom conse-quences should be taken into account (vii).

Before their joint work on incrementalism in Strategy was under-taken, Lindblom published his astoundingly accessible but equally as-toundingly misunderstood article ‘The Science of “Muddling Through”’(Lindblom (1959)), still reprinted in policy and public administrationtextbooks. This piece directly challenged the synoptic tradition.10 Lind-blom calls the synoptic option the ‘rational-comprehensive’ or root method,and refers to incrementalism as the strategy of ‘successive limited com-parisons,’ the branch method. As well as setting out the characteristicsof the two methods in text, he also provides a text table that summarizestheir main features. For the rational-comprehensive method, he notesthat a preliminary exercise in value clarification was necessary beforemoving to empirical analysis of alternative policies, with the empiricalanalysis developing the means to seek the ends that have been isolated;that the test of a ‘good’ policy is that it can be shown to be the mostappropriate means to the desired ends; that every factor is taken into

10 In the 1979 follow-up and clarification, ‘Still Muddling, Not Yet Through,’ Lindblom (1979)Lindblom says that the 1959 essay had then been reprinted in some forty anthologies. The 1959paper is still found in collections in organization theory, public policy, and public administration,whereas the 1979 paper is rarely even mentioned.


account in analysis; and that theory is important to the exercise. For thebranch method, he notes the direct contrasts: that value goals and em-pirical analysis are closely intertwined and not really distinguishable,and thus means-end analysis is problematic; that the test of a good pol-icy is that it will be recognized as such by analysts who neverthelessmay not agree that it is the one best way to reach an agreed objective;that analysis is limited and a succession of empirical comparisons willreduce the importance of theory. Despite the clarity with which he ex-pressed the idea that while, ideally, the root method does not exclude,‘in practice it must’ (80), and that successive limited comparison is amethod in itself, not a fallback, commentators misrepresented his work.The piece was often used in discussions in policy courses as proof thatincrementalism was another word for drift, and that there was no al-ternative to attempts to achieve full understanding. Perhaps the maindifficulty was that professional policy analysts working in governmentfelt uncomfortable about analysing means as containing their own val-ues; this would be not value clarity but a threat to their perception ofthemselves as value-‘neutral’ and to the elected or political leadership.

6.2.2 Satisficing, or Environmental Heuristics: Simon’sAlternative

Herbert Simon, in the previous generation of scholars, also observedthat the rational-deductive model (or ‘economics’ model as he calls it)does not directly help the decision-maker in a large corporation to per-form successfully. It is Simon’s alternative to synoptic rationality thatstill dominates the applied literature and that has inspired many vari-ations. Simon does not discard the synoptic model, believing that itremained an appropriatenormativemodel for decision-making, a beliefhe stated at least as late as 1980 in his article ‘The Behavioral and SocialSciences’ (Simon (1980), 75). But he rules out the synoptic model fordirect application. On his own, very clearly, inAdministrative Behav-ior (Simon (1976), xxiv), and, with March, inOrganizations(Marchand Simon (1994)), Simon described the ‘bounded’ rationality of theindividual decision-maker, who is ‘intendedly rational [in the sense ofkeeping the economics model firmly in mind], but only limitedly so’


(161-2) lacking the cognitive power to attain full rationality.11 The cog-nitively limited decision-maker should therefore find and be satisfiedwith an outcome that is better to some extent than the status quo, towhich Simon gave the name ‘satisficing’ (in contrast to optimizing).

Individual decision-makers, limited in resources of knowledge, tech-nical skill, and time, are however supported and guided by their environ-ment of givens in the organization, which has greater reach and agency.The organization provides, as Richard Scott summarizes in hisOrgani-zations: Rational, Natural and Open Systems(Scott (1992)), ‘integratedsub goals, stable expectations, required information [or at least informa-tion that is accepted as usable], necessary facilities, routine performanceprograms, and in general a set of constraints within which required de-cisions can be made’ (47). Scott here consolidates Simon’sThe NewScience of Management Decision(Simon (1977), 51.)

Reinhard Selten (Selten (2001)) picks out three ideas in Simon’searly writings on bounded rationality in an individual decision—thesearch for alternatives, satisficing, and aspiration adaptation:

He [Simon] described decision making as a search processguided by aspiration levels. An aspiration level is a value ofa goal variable that must be reached or surpassed by a satis-factory decision alternative ... Decision alternatives are notgiven but found, one after the other, in a search process. Inthe simplest case, the search process goes on until a satisfac-tory alternative is found ... Often, satisficing is seen as theessence of Simon’s approach. However there is more to it... Aspiration levels are not permanently fixed but are ratherdynamically adjusted to the situation. They are raised if itis easy to find satisfactory alternatives, and lowered if satis-factory alternatives are hard to acquire. This adaptation ofaspiration levels is a central idea in Simon’s early writingson bounded rationality. (13-14)

11Oliver Williamson, in a chapter in his edited book,Organization Theory(Williamson(1995)), titled ‘Chester Bernard and the Incipient Science of Organization,’ says that once Simonhad developed his ideas on satisficing, Simon argued that economists should abandon notions ofmaximizing for satisficing. However, no research tradition on satisficing developed in economics,and satisficing became instead identified with ‘aspiration level mechanics’ (79).


Simon aimed to operationalize observable definitions of aspects ofdecisions such that his findings would be seen to be empirically founded,leading to a science. He also distinguished facts from values, therebyisolating values by virtue of their subjectivity and lack of a base in rea-son. (This is an element of continuity with the synoptic model.)

The part of Simon’s work that resonated most with working decision-makers was what they understood as Simon’s promise of an efficient orsummary analytical procedure that could be defended as ‘rational.’ InAdministrative Behavior(Simon (1976), 75) Simon stipulates how hissummary procedure is ‘rational,’ within bounds or limits. The word‘bounded’ describes the process of decision, not its substance. Simondefines achievable or bounded rationality as ‘the selection of preferredbehavioral alternatives in terms of some system of values [inter-subjectivelyaccepted] whereby the consequences of behavior [decision outcomes]can be evaluated’ (75). He has also stipulated that the satisficing pro-cess is rational in that the procedure and its limitations would be setout. In an administrative setting this would point to an individual mak-ing a choice, rationally using a selection of the received views of theorganization. This action could be reviewed by other officials, probablywith a higher likelihood of review if outcomes were poor. Simon alsoaccepts that an unforeseen event or factor can cut into the bounded prob-lem and change the outcome. This is to accept theoretical and empiricalemergence that will make a heuristic inapplicable to a model that hadappeared to fit a problem.

Therefore Simon has been forthright that the process of bounded ra-tionality or satisficing makes no guarantees of satisfactory outcomes.In other words, a bounded solution is a routinized decision, where a‘satisficing’ procedure in analysis is efficient but still contains an ele-ment of risk. Repeating in 1990 what he had stated in essence as earlyas 1956 in his article ‘Rational Choice and the Structure of Environ-ments,’ Simon described bounded rationality in a metaphor: ‘Humanrational behaviour ... is shaped by a scissors whose two blades are thestructure of the task environments and the computational capabilitiesof the actor.’ And not even the structure of the task environment canalways provide good information because it can be misunderstood: ‘de-


cisions begin with goals and values, and are strongly conditioned byorganizational loyalties and perceptions’ (Simon (1990), 7). Simon hadnevertheless created an expectation that there were discoverable envi-ronmental heuristics that could guide routine decisions to good results(in contrast to definitionally uncertain unprogrammed or one-off deci-sions).

There are objections. Those that are most applicable in the publicsector have to do with Simon’s steering attention to what appears tobe most readily doable, in the efficiency calculus of an organization,whereas the most serious problems that administrators, analysts, andpolitical leaders seek to ameliorate are defined by and exist in a value-based or political framework. Theodore Lowi, in the course of his ad-dress (Lowi (1992)) to the American Political Science Association onassuming his presidency, directed a stinging attack on Simon’s work.He calls it reductively apolitical, because of its minute focus on the de-cision or decision premise and says it had aided and given comfort tothe shift in political science from judgmental and evaluative analysis toa rationalist reduction of government to just one aspect of the economicsystem.

Braybrooke and Lindblom for their parts each generously acknowl-edge Simon’s role in encouraging theorists and practising analysts tomove away from the economics model. But they are not wholly uncrit-ical of Simon’s bounded rationality. Lindblom, in his ‘The Science of“Muddling Through”’ (Lindblom (1959)), offers a succinct explanationof why the choice of an aspiration level cannot do much to limit themargins of a current problem to create a good fit of a received modelwith a new situation, even if it appears of a type: the problem of sortingthrough values is ‘always a problem of adjustments at the margin [and]there is no practical way to state marginal objectives or values except interms of particular policies’ (86)—particular decision situations. Thisis an example of his settled view, as I understand it, that feasible policydesigns are more limited in number than the problems to which theymight be applied, so that analysts could readily recognize an attractivepolicy design but support it for different reasons. And to be sure, it isalways possible that the analyst is wrong about which environmental


heuristic is a good model for a given problem.Braybrooke, in his article in theInternational EncyclopediaBray-

brooke (2000), sees the utility in Simon’s approach of looking for stan-dard features within problems over time and across problems. But henotes ‘the gap that Simon leaves as to where to start with a new [non-routinized] problem’ or how to start recasting an old problem into amore usefully approachable form.12 This is, to be sure, a point of de-parture for his own approach to decision theory, his judicious additionof features of utilitarianism to a careful version of incrementalism.

Braybrooke of course is harder on synoptic rationality or maximiz-ing because of its absolute infeasibility, in one place saying that ‘atbest, it offers an idealized hypothetical basis for criticizing omissionsin deliberation’ (3315). For me, however, Simon’s satisficing has neverseemed convincing. Given that it is impossible to satisfy the substan-tive requirements for synoptic rationality because of cognitive, resource,and time limitations among others,anysummary procedure for analysisapplied to the real world is, in being incomplete, potentially a satisfic-ing procedure. The problem seems to be that satisficing might be anyincomplete procedure whatever. Alternatively, its claims to superioritydue to strategic completeness might be impossible to prove, because ofthe lack of a substantial standard as well as for measurement reasons.Even if Simon is not making a claim for his strategic incompleteness ashidden optimization, his recommendation of strategic incompleteness inmodelling heuristics to problems seems to suggest that decision-makerscan and should rationally tackle tasks in routinized areas by restrict-ing themselves to the big ‘found’ key features, the noses on the face,which the organizational task environment, over time, will have identi-fied. Here it seems too convenient that a cognitively limited decision-maker would live in a world where causality within task environmentscould collaborate to deliver moderately good results over time.

Mary Douglas (Douglas (1995)) suggests that Simon made an infer-ential leap from Chester Barnard’s idea of the ‘zone of indifference’ in

12Personal communication, 12 December 2004. Braybrooke’s treatment of utilitarianism takentogether with incrementalism can address both new problems and be useful for recasting andreworking older problems.


the preferences of any employee in the organization (allowing for theirperformance of directives without hesitation and thus the accomplish-ment of organizational goals) to what would appear to be the real orga-nizational world such that desires can be satisfied within the organiza-tional context ‘without optimization, and therefore without the calculus’(106). I take Douglas’s remark to mean that Simon extended Barnard’snotion that persons would execute organizational duties most efficientlywhen duties are ‘indifferent’ to them in the light of their own convic-tions, to another idea that the environment behaves similarly: that mostfactors in the decision environment itself would be irrelevant or neutralin relation to what one could call ‘core causation’ in the environment,and so could usually be safely ignored. Douglas should surely not be un-derstood as saying that Simon would have either believed or argued thatsatisficing could be a more efficient way of optimizing, but she can per-haps be read as saying that Simon made a leap to a certain isomorphismbetween the limitations of the decider and the apparently massive cluesin the decision environment under routine conditions. Overall, practi-tioners do tend to accept satisficing as a procedure for operationalizingpragmatism, a local theory of usefulness, in their decision-making envi-ronments. Thus, in my opinion, Simon left his field without providingactual robust substantial assistance to practitioners. And, as Lowi said,Simon’s endorsement of the view that facts and values can and mustbe separated is a problematic contribution to policy studies, althoughintended to separate politics or values leadership from administration(Lowi (1992), 105).

6.2.3 Some of Simon’s Successors

As one might expect, the hopeful train of thought begun by Simon—who worked in psychology, computer science, economics, and politicalscience—has been as widely taken up. One sees his ideas in use byscholars who try to apply it to describing environments and also by re-searchers working to discover acceptably reliable prescriptive heuristicsin particular environments. As a colourful example of the descriptivepreference, one can cite Cohen, March, and Olsen’s influential article,‘A Garbage Can Model of Organizational Choice’ (Cohen and Olsen


(1972)), in which problems and solutions circulate in the decision envi-ronment to be sometimes brought together by an astute decision-maker.

Gigerenzer and Selten (Gigerenzer and Selten (2001)) illustrate theheuristics thrust. Their collection includes work by anthropologists,economic theorists, mathematical economists, cognitive and other psy-chologists, computer scientists working in artificial intelligence, zoolo-gists, biologists, and neuroscientists, among others. The many investi-gators seem agreed that any rationality we may find will be embeddedin survival strategies observed in local environments and that there willnever be a general-purpose algorithm. (But in his larger project, Si-mon was notably dedicated to a science of administration.) Only oneof the nineteen contributions to the volume considers how a collectivitymight reach a decision. In Thomas Seeley’s ‘Decision Making in Super-organisms: How Collective Wisdom Arises from the Poorly InformedMasses,’ (Seeley (2001)) the masses reach what is said to be an ob-jectively correct solution for the collectivity. The masses happen to behoneybees, and their problem is to choose the best site from those avail-able for a new hive. It is worth probing Seeley’s observation-rich essayon a hive as a ‘super organism,’ because it models social behaviour innon-human organisms or ‘systems,’ the kind of work that systems theo-rists readily extend to human organizational needs (Seeley restricts hisconclusions to his bees).

In a beehive, Seeley judges that ‘the swarm as a whole is a sophisti-cated and skilled decision-making agent’ because the cognitive skills ofeach bee ‘appear to be surprisingly simple’ (249). Seeley does not spec-ulate how the honeybee scouts process the four or five relevant site vari-ables to report in dancing to the other scouts what looks like a summedscore over several variables. The dancing by individual scout bees forthe sites they have visited among alternative sites varies only in persis-tence and intensity. Eventually ‘consensus’ (all dancing scouts perform-ing for the same site) is reached. The scouts and the rest of the hivingswarm then leave to establish a new hive.

The scout honeybees’ collective choice is, in all the occasions ob-served by the scientists, objectively the absolutely best choice—on thesurvivalist criteria the human scientists use to rank honeybee housing,


and impute to the scout honeybees, to be sure. The persistence of site-oriented dancing correlates with the presence of a number of featuresdescribing the site: such features as entry dimensions that afford thehive and its food protection, adequate food storage capacity for the win-ter, and proximity to a source of food. Scout honeybees stay on topicin the hive site evaluation. No scout honeybee has concerns other thanthose that define a workable site; no uncle is in bee real estate. Non-functional criteria would manifest as ‘irrational’ scout dancing persis-tence for a site that is mediocre on the criteria necessary to the hive’ssurvival. There is no such ‘irrational’ persistence promoting sites thatare poor, again as assessed by the scientists.

I conclude, despite the title of Seeley’s engrossing article, that itis irrelevant whether honeybee masses are ‘poorly informed.’ What isimportant is that the scouts accurately communicate to one another rel-evant empirical information for a closed model that defines the best siteand that their collective dancing achieves an accurate site comparison.(It is also fortunate in this problem that scouts whose preferences fall bythe wayside switch to dancing for sites still in contestation, and that therest of the hive swarms with departing scouts.) Each site factor in theequation is necessary but not sufficient; all factors must be present forthe site to be chosen by the scouts of a hiving swarm. Thus the hiving‘problem’ unfolds by formula. Seeley is cited above as saying that theskill of the hive or the aggregate is in part the outcome of the apparentsimplicity of each bee, and there is a lesson in this remark: simplicityis a form of discipline and it surely works for the scouts’ hiving deci-sion. The aggregated wisdom of all scouts is not a simple extrapolation,but the differences in the preferences of the scouts can be negotiated andsummarized in the danced ‘discussions’ without serious loss—as judgedby outcomes. The success in choosing a good hive seems almost lessimportant than the idea that the bees have a formal public process bywhich the scout bees obviously—but only ‘somehow’—aggregate andcommunicate their information. It is not fully intelligible to the humanobservers as to how the scouts emerge with the maximizing decision.

But, as Seeley’s presentation hints, because of the complexity andundisciplined variety of human individuals’ preferences and motiva-


tions, human collectives face important uncertainties in outcomes ineven highly repetitive situations. In hisMicromotives and Macrobe-haviour(Schelling (1978)), Thomas Schelling shows that outcomes thatare inconvenient and harmful and thus ‘irrational’ in the aggregate of-ten result from the situationally rational behaviour of individuals. If wehumans ever have leaders as intendedly rational as the honeybee scouts,they are alarmingly absent in many situations where they could be use-ful. Schelling notes that the easy cases in decision-making are those inwhich the aggregate is an extrapolation from an individual. More dif-ficult are those cases, of which the small ‘p’ political environment ina firm or a government bureau could be examples, where ‘People areresponding to an environment that consists of other people respondingto their environment, which consists of people responding to an envi-ronment of people’s responses’ (14). Dynamics can be sequential orreciprocal or both. Schelling uses audience behaviour as an initial il-lustration. Under voluntary seating arrangements, an audience will failquite miserably in its use of space. Individual seating choices typicallyresult in a distribution of spectators such that rows of seats will be leftvacant in the very front of a hall, creating crowding in the back and thusdisturbance by new arrivals of those already seated toward the back.Performers may have the impression of playing to an empty hall, or toone that is sliding away from them. The only process we have devel-oped is to give power to ushers, who take individual preferences intoaccount only insofar as a developing seating pattern permits. The situa-tion where we collectively recognize a settled problem and know whatour goal must be and how to implement it seems unproblematic and rarein comparison to dynamic situations, individual and collective.

6.2.4 Incrementalism

For problems that change over time, where information is not completeand may always be misleading to some degree, a strategy that can makedoubt part of the decision process is necessary. What is available isincrementalism, for which Lindblom and Braybrooke are the leadingproponents. One can briefly sketch their shared and distinguishing in-terests by way of introduction. Lindblom, like Braybrooke, is interested


in content-rich policy evaluation and choice, with features of fragmen-tation and seriality. He has, however, placed less emphasis on delib-eration in comparison to Braybrooke, who wants to suggest ways toimprove deliberation in collectives, in respect to content and procedure,admitting argumentative elements such as provocation.13 And, althoughLindblom has clearly shown persistent interest in and respect for peo-ple’s everyday ideas, in his insightful book with David Cohen,UsableKnowledge: Social Science and Social Problem Solving(Lindblom andCohen (1979)) and in his own tour de force,Inquiry and Change: TheTroubled Attempt to Understand and Shape Society(Lindblom (1990)),Braybrooke alone places emphasis on the concept of needs, for whichthe most concentrated example is probably hisMeeting Needs(Bray-brooke (1987a)).

Lindblom is keenly aware of the cognitive limitations and the so-cial givens that weigh on individuals. His bookPolitics and Markets(Lindblom (1977)) emphasizes the indoctrinated character of the citi-zenry and the weight of business in politics. Elsewhere, more recentlyin Inquiry and Change, he has described the ordinary person’s cognitive‘impairment’ and how we might work toward freeing ourselves of someof the received and personal doctrines that burden our thinking. Lind-blom believes we may invoke habitual patterns of ideas unhelpfully atmoments of decision or expression of preference. Thus, for Lindblom,the ‘givens’—the noses on the faces—that Simon believed would pro-vide helpful frameworks may well be foolish and destructive ‘impaired’thinking—although, as noted earlier, Lindblom does not engage Simon.

In Intelligence of Democracyand elsewhere Lindblom works care-fully to distinguish what he calls ‘partisan mutual adjustment’ from in-crementalism. In the former, participants who are at least somewhat in-dependent of one another take part in fragmented, decentralized decision-making. The eventual winning policy is more like a happening or spon-taneous occurrence than a reasoned decision for which relevant reasonscan be given. It is unplanned in its content.

Neither scholar claims, therefore, that incrementalism is more than

13For example, see his article ‘Toward an Alliance between the Issue-Processing Approachand Pragma-Dialectical Analysis’ Braybrooke (2003b).


a process that holds out modest hope for analysts and collectives thatwant to ameliorate some of the problems that beset them—although itis surely the only decision-making theory that gives defensible reasonsfor even modest hope and which can be applied to dynamic problems incollectives. Incrementalist approaches depend on people’s capacities forobservation and attentiveness while manifesting doubt constructively asevaluation; continuously thinking, discussing, and amending beliefs,singly in Lindblom’s formulation, and in groups in Braybrooke’s de-velopment of ‘discussions with a direction.’ Individual analysts willlook at a problem strategically, disaggregate its elements, choose anelement that appears to be manipulable in a useful way, and recom-mend adjustments. To the extent that repetitiveness can be seen as apossible platform for decision-making, this system might be seen toresemble Simon’s bounded rationality. But, in my view, incremental-ism is importantly different in that it harnesses doubt and uncertaintyas a motor for continued attentiveness. It recommends frequent evalu-ative phases, continuously ‘probing’ and thus disturbing and amendingobservers’ grasp of the problem context, amending common wisdomabout the problem as the work proceeds. Thus both leading advocateswant to see problem stability approached as an empirical question. Bothsee perceptual fluidity as a good quality.

The small number of incrementalist thinkers who contribute to pol-icy journals tend to put their emphasis on the feeling-forward perspec-tive of incrementalism. Terry Connolly (Connolly (1988)) provides aparticularly communicative metaphor in comparing tree-felling models—‘heroic single-pass one-time synoptic analyses’ (38)—with highly inter-active meandering analysis, what he calls hedge-clipping models. ‘Clip-ping’ is a feasible approach to a task or problem when the defining ele-ments of a problem situation are decoupled. The advantages of a situa-tional approach that is at the same time closely evaluative are that con-sequences can be corrected; solution criteria will typically be tolerant(allowing emergent problems to be taken into account in the process);action sequence can be flexible; and the implementation of a sketchedplan can safely be interrupted.14 There are many ‘hedge’ problems that

14See also Connolly’s ‘Uncertainty, Action and Competence (Connolly (1980)). Michael

6.3. BRAYBROOKE’S CITIZEN MANAGEMENT 145

we explore only after having made the first small intervention. Overall,however, it seems safe to say that while incrementalism is present inpolicy journals, and sometimes interestingly so, all subsequent adher-ents of incrementalism as a tactic for the individual analyst have foundit difficult to add to the substance of Lindblom’s 1959 article.

6.3 Braybrooke’s Participatory, Precautionary CitizenManagement

Braybrooke’s contribution to decision-making theory has long struckme as ground-breaking in that it is an open, political, and highly partic-ipatory approach to the possibilities incollectivedecision-making. Inhis recent work, he has pointed to the gap in decision-making theory forthe problem of collectives:

Unfortunately, at the collective level, unless an autocrat ora perfect bureaucracy ... is authorized to act for the collec-tivity, a problem arises about different criteria with differentthresholds being invoked at the same time by different peo-ple. Which are to guide the decisions of the collectivity?(Braybrooke (2000), 3315)

He invites participation in a variety of forums as a way of assistingpeople to refine their information and values relating to a problem, andto encourage people to forge their beliefs into intelligible—and there-fore usable—statements for use in discussions with a direction, includ-ing intelligible statements about needs. He also sees participation asencouraging reasonableness in representation of needs and, of course,as a way of identifying and helping a collective to work through in-ternal opposition in non-violent ways that do not involve suppressionor exclusion.15 There is no attempt to separate facts from values. Onthe contrary, he is clear that means cannot be separated from ends. He

Hayes’s Incrementalism and Public Policy(Hayes (1992)) also takes on ‘the rational-comprehensive ideal.’

15See chap. 16, ‘Scale, Combination, Opposition: A Rethinking of Incrementalism,’ in hisMoral Objectives(Braybrooke (1998a), 311-30).


wants to multiply opportunities for people to voluntarily participate andbecome significant contributors to long-term deliberations about issues,with a view to increasing the intelligibility of problems and of whatothers make of them, both for themselves and others. As a theorist ofparticipation, he emphasizes the intellectual aspects of human develop-ment through deliberation. While group discussions may on occasionbe taken very seriously by the society and by the polity, Braybrooke hasno plan for a formal and political machinery along the lines of Dahl’s‘Chinese boxes,’ in which jurisdictions are multiplied and nested suchthat all continuing discussion forums lead into the highest state forumswheredecisionsthat affect public policy and define public goods aretaken (Braybrooke and Bailey (2003)). Braybrooke, like Lindblom, pro-motes his alternative only to the extent that it will work better than thestatus quo in bringing attention and evolution to received settlements.

A superficially paradoxical feature of Braybrooke’s prescription thatis also the key to its richness is his combination of incrementalismand a version of utilitarianism. Incrementalism is by definition a con-tinuing strategy for sequential passes addressing comparatively well-understood problems, yielding an improved probability that the trend ofchange in the problem will be ameliorative and that consequences can bemanaged—improved over ‘one-time’ or heroic solutions. The strategyis prudent, but not conservative as to ends. On the other hand, the ethicsBraybrooke adopts—utilitarianism—emphasizes that the author of anaction is an agent who will take responsibility for consequences. The li-cence to act under an ethics of consequentialism is said to arise from thepreliminary, one-time and heroic, assumption that consequences flow-ing from an act will bear out the choice. Since there is little possi-bility that potential consequences can be anticipated, and certainly notin one pass, there is no reason to assume that any well-intentioned ac-tion will turn out well within the period of judgment. Braybrooke setsout how the strategy of making maximally prudential changes underdisjointed incrementalism—well-considered small moves designed tomeet needs—constitutes a more satisfying substitute for the moral per-mission that would in theory be conferred by working through the sumof benefits over costs in human happiness, the (synoptic) ‘felicific calcu-


lus.’ No such calculus is realizable. Braybrooke therefore incorporatesthe fallibility of human foresight and the tendency of the world to sur-prise and to provoke blame as conditions of normal decision-making.For him, a determination to address human needs through reversibleactions undertaken in a mood of cautious humility is the moral and po-litical possible.

Chapter 2 in Braybrooke’s 2004 volumeUtilitarianism is perhapshis most fully developed view of what might be done better in demo-cratic planning for public policy, and the material that follows has beendirectly paraphrased from this source. Stripped to its bones, Bray-brooke’s recommendation is that both individual decision-makers andcollectives (groups, jurisdictions, and so forth) in social and politicallife might adopt disjointed incrementalism as both a cognitive style andas a strategy, and to do so under an ethics of utilitarianism whose firstpriority is the provision of needs.

6.3.1 Braybrooke’s Disjointed Incrementalism

The summary features of Braybrooke’s formulation of disjointed incre-mentalism as an ameliorative decision strategy are set out in point formbelow. The ‘problems’ are those habitually arising from the continueddominance of the synoptic reflex in decision theory.

Problem: too many value criteria confuse and overwhelm analysis.Solution: move to a higher-order value for guidance; here needsprovide focus.Problem: attempts to take into consideration all consequences beforeacting will postpone ameliorative activity in difficult problem areas.Solution: align the theory of policymaking with the ‘revisionaryprocess’ active in current machinery and processes of policymaking inwhich stage-by-stage revisions inch toward a better status. Once aproblem has been nudged in the right direction, the revisionary processwould engage to pursue further improvements or, alternatively, torevoke unsatisfactory changes.Problem: designing policy for large-scale or complex problems forcesanalysis into ambitious attempts to cover ever more ground in the


belief that more work will bring an optimal solution closer; suchanalysis is still merely incomplete, and still to an unknown degree.Solution: study difficulties that are close at hand and familiar, adoptdisjointed incrementalism to separate facets into still smaller problems,make marginal changes to these small problems, and subject them tosteady attention (above). An incremental change is defined as changein an apparently relatively unimportant variable that is a feature of afamiliar problem, or a less important change in an apparently importantvariable. A second definition is that an incremental change is one thatis planned so that it is readily reversible in substance and in itsconsequences by current institutional resources and powers, and is byvirtue of close examination more intelligible to everyone.Problem: general lack of time and lack of research in the implementingbody.Solution: develop public attentiveness strategically. Start with a modestchange, invite participants (individuals and groups) to join the processto suggest other alternatives or further adjustments and to bringforward other values. This process allows for values achieved, but notoriginally intended, now to be recognized and pursued further(supersession): the revisionary process can work with any gain thatturns up.Problem: precaution is often limited to the front end ofdecision-making. Intellectual investment in conventional policy goalscan dominate thinking and stymie action; traditional thought patternscan develop into doctrines.Solution: precaution should be part of each phase. One can take anevaluative stance, then deal with things as they stand in cautious action,where attempts are made anew to anticipate possible differencesbetween expected and actual outcomes.

6.3.2 Virtues of Disjointed Incrementalism for the Individual

The individual policymaker or citizen, trying to foresee and evaluate thepotential consequences of the possible new policies, following Bray-brooke’s formulation, considers the following:


Peremptory values: These are sometimes deal breakers; more gener-ally, they establish constraints on how smaller problems can be brokenout of larger ones or on what policy alternatives will be carried for-ward. The analyst considers various possible policies in their light, withthe goal of excluding any proposal whose consequences are intolerable(those that infringe minority rights, those that distribute risk unfairly,and so forth). The possibility of deadlock must be accepted. Otherwiseit may be believed utilitarianism might be mobilized to argue for over-riding minority rights. The task of the collective is to discuss values andto choose policies whose values are ameliorative, within peremptoryconstraints.

Remedial orientation: Because in this approach the analyst will en-dorse one-step remedies that move in the right direction and are withinthe range of acceptable practice, Braybrooke conceives of this as a satis-ficing or ameliorative and remedial approach rather than an optimizingapproach.

Incremental alternatives: The single agent can realize a primarybenefit of incrementalism by working on small problems and separableparts of larger problems, with the purpose of retaining known benefitsof present policy and alleviating known problems. Small moves consti-tute a form of probing the environment to test its reactions to dosage.On the other hand, if a major change is wanted, it will be brought aboutin stages by a series of smaller moves, each phase preceded by a fresh-eyed evaluative phase.

Institutional insurance: A single agent can assess whether currentinstitutions have the capacity to correct for a mistake made within theframework under construction. If security is not assured, the analyst canlook again at institutional features, or simply not recommend the pro-posed move. A manageable policy move is defined by the fact that theinstitution or organization(s) of primary interest has or have the abilityto reverse the move and undo the harm. Braybrooke notes that the ca-pacity to remedy the consequences of moves seen in hindsight to havefailed is definitional of ‘an adequate political organization.’ Overall,even where insurance is ample, small-scale innovations are obviouslymore intelligible and easier to absorb and correct. To bring about a new


state of affairs in a major social institution such as health care, for exam-ple, many nested series of moves, decoupled to the extent possible, andalert to necessary coupling, will allow the ‘insurer’ to monitor change,while reviewing the rate at which insurance is being depleted.

Redesigning, supersession: These are the mechanisms of option re-finement. Although a single policymaking agent will be less rich inideas than a group, single agents do hit on ideas that bring efficiencygains, and they can identify workable compromises; they redesign. Asecond mechanism is supersession of values. Braybrooke uses the termto cover situations where events outrun a policy framework, incitinggoal revision. His example is that metered parking may have been in-troduced to give visitors fairer and faster access to services and shops.But, once in place, metered parking might noticeably reduce the num-ber of private vehicles entering the city centre. Metered parking is thenseen as a tool to reduce congestion. Supersession as a mental event is orshould be familiar to all of us. The word identifies a useful opportunityfor re-imagining one’s area of interest that we often call rationalization.Analysts can likewise experience perspective-shifting insights on causalmechanisms and results within even well-established policies.

6.3.3 Virtues of Disjointed Incrementalism in a Social Process

The social processes of discussion, evaluation, and choice are whereBraybrooke’s strategy is best fulfilled. Needs will, in the first stage, andonly in part, take the place of the ‘greatest happiness principle,’ of tradi-tional utilitarianism. (Once needs have been met, people can move intomore abstract deliberations, based in preferences, or additional com-forts can be deliberated.) All features of the strategy applicable to theindividual are carried over to the collectivity: peremptory values; reme-dial orientation; incremental alternatives; institutional insurance; andredesigning and supersession.

In addition, there are other benefits that arise from situating the strat-egy in a collectivity. Most are part of the shift itself. The traditionalethical-theory focus on the single agent is superseded. In the place ofthe single agent, one finds, more fittingly in social and political life, a


multiplicity of actors with roles and information, and communal discus-sion and negotiation.

Just as the criterion of need guides means and ends for the substanceof policy decisions, the spirit of the constituting rules of the overarchingsystem is protected by respect. Participants in ordinary political activityshould not exploit by stealth constitutive features of the political settle-ment.16

Recommended features for collective policy discussions include:Partisanship and fragmentation: Fragmentation in the body politic

becomes an advantage for bringing forward information and alterna-tives. First, under the strategy, the community makes use of what Bray-brooke calls the revisionary process (the various levels of evaluative de-liberation available) to generate richness for future redesign phases. Thecollective will work on a problem until it is as intelligible as possible,design rules, and devise institutional insurance.

With many groups at work, the stimulation could lead groups to vig-orous analysis of different policies. Such constructive competitivenesswould uncover bias. And bias itself contains information about fearsand desires. Many analysts, deliberating, will also be alert to superses-sion, as discussed above.

Fixing jurisdictions: Subsidiarity suggests the approach. The juris-diction that is closest to the manifested problem and has the authorityto address it will be the focus for group contributions. Ideally, the au-thority will implement public policy changes and take responsibility forthe prudential considerations and consequences, any program of actiontaking place through many iterations. Thus the ‘formal’ or state jurisdic-tions, including legislatures and their committees, will, as now, be theagent of change drawing from public resources. But voluntary associ-ations and private organizations will want continuously to evaluate andto press for further movement as it becomes intelligibly feasible, andas they see that institutional insurance is likewise affordable and feasi-

16In developed democracies, intense partisanship is a considerable danger to institutional per-formance. For an argument that actions that evade the deliberative phases of judgment constitutea moral offence against the collectivity, see (Thompson (1993)). For an extended application toother cases of Thompson’s formulation, see (Sutherland, S. L. (1995)).


ble. Citizens and associations will make use of all their opportunities toengage with the authorities.

Any received jurisdiction’s powers and responsibilities can be chal-lenged and shaken by events beyond its control. But Braybrooke doesnot foresee a great deal of evolutionary reshuffling of powers and dutieshorizontally and vertically. He believes that old societies with stableworking constitutions will by and large continue to work with their re-ceived jurisdictions. To be sure, the existing dispensation can well con-tinue to ignore interests not organized into the discussion. Here againBraybrooke signals his intention to press needs as guides to prioritiesfor collectives.

6.3.4 Braybrooke’s Cautions

One would expect that, in a process designed with both individual rightsand group welfare in mind, there would be a good number of cautions.Centrally, some interests will be ignored or given a low priority. Thedangers of passionate contention are also real. Institutionalized interestsare vigorously promoted to the public. As well, partisan misrepresen-tation of policies may capture popular understanding, blocking changethat could almost incontestably increase the welfare of most people.The certainty of realization of desired outcomes is always in doubt. Theonly comfort is that, on the basis of previous experience, there is a goodprobability that a change that comes through such a process will help tomeet needs. But a probability is just that. Institutional insurance mightalso fail. Braybrooke insists that the capacity of a political organizationto produce remedies for its own failed policy, one it undertook and ‘in-sured,’ is ‘one of the tests of an adequate political organization—indeed,of an adequate social organization of any kind’ (Braybrooke (2004), 60-1).

Braybrooke identifies clusters of situations in which the open pro-cess of incrementalism cannot be successful: when there is a factuallyor perceived overwhelming external threat; when an initially promis-ing path leads to decremental change (see below on bovine spongiformencephalopathy); when social contestation becomes so intense that par-ticipation and the open process is itself impeded; when there is a sud-

6.4. POLICY APPLICATIONS COMPARED 153

den perception of explosive issues in what had seemed anodyne to oth-ers; when there are irreconcilable differences about whether a changeis incremental or non-incremental; and, in some cases, when a non-incremental policy may work better than the incremental strategy.

Nevertheless, Braybrooke believes that continuous work in open pro-cesses by social groups and political institutions will, overall, produceenough consensus, richness, and redundancy to reduce the potential ofmissing a factor that could lead to disaster. If consequences are signalledin public processes, there is a better chance that the public jurisdictionwill draw in enough resources for institutional insurance. Therefore thestrategy in its fullest use at the collective level will work better thanother approaches in modestly stable environments.

Just as single agents can salt away resources to buy future freedom ofaction, so can communities and agents acting for them behave prudently.The tools are now familiar: retreat, redesign, or reconceive the problemfor a higher jurisdiction with more resources.

6.4 Policy Applications Compared

One of the shortcomings of people as political animals is that publicpolicy successes seem to us to be only ‘normal’; we do not reflect thatthese are the result of careful planning and are maintained by steady co-ordination and superior vigilance. On the other hand, most of us readilyattribute blame when policy is unexpectedly costly or when it fails de-cisively.

One can find successful policy reforms, and in exactly the kind ofproblem areas that Braybrooke has long thought should most easilyyield to incrementalist approaches. In 1974 he published his book ontraffic congestion in London,Traffic Congestion Goes through the Is-sue Machine, whose focus is the serial nature of a ‘conversation,’ asubstance-laded debate among the members of a stable discussion net-work. Braybrooke has continued to use traffic problems as illustrationsof difficulties that invite incrementalist experimentation.

In February 2003, the three-year-old Greater London Authority, whosepredecessor government had been abolished fourteen years earlier by


Prime Minister Thatcher, introduced the world’s first traffic conges-tion charge on individual vehicles entering a charging zone covering22 square kilometres in central London. The mayor had taken up a pro-posal that was already seen to be an option in 1969, but was set aside, asBraybrooke recounts inTraffic Congestion. The 2003 implementationhas many features of a collective deliberation and series of decisions, aswill be seen. Preparations were made for the commuter fee to be paidby telephone, on the Internet, or at some shops and gas stations. Cam-eras at key intersections filmed the stream of traffic, and car ownerswho had not paid the fee were mailed penalty charge notices. A six-month review by the London Transport Authority said that the chargehad stopped about 50,000 drivers from entering the congestion zoneeach day, for a reduction of about 16 per cent. The drop in cars reducedthe number of traffic jams by a third and accidents by a fifth. Deterreddrivers switched to public transport or skirted the zone. It is estimatedthat road journey times were between 10 and 15 per cent faster.

In both administrative and political insurance perspectives, the roadcharge is thoroughly incrementalist from the Braybrooke perspective.For a factor of interest to political analysts, the policy is led by a protag-onist whose personal political insurance is good. The new and formermayor, Ken Livingstone, has never owned a car, and, as mayor, he re-fuses official cars and takes the underground to work (Finnegan (2004)).

The innovation itself has many substantial insurance features.17 First,the road charging policy is the centrepiece of decoupled but enablinginitiatives in other areas. For only two examples, bus services wereimproved notably before the charge was applied, and other initiativesimproved safety and movement flow for two-wheelers and pedestrians.To allow for the public’s typical slowness to absorb official informa-tion, a call centre that could respond to hundreds of thousands of callsa week was established. Another salient point is that the charge doesnot affect those living in London—that is, the voting base—but it doesaffect people who were adding to the congestion but paying their prop-

17Information for this summary is taken from Transport for London’s Press Centre. The fullreportSix Months Onis available atwww.tfl.gov.uk/tfl/cc-intro.shtml . See alsoMayor of London (2002).


erty taxes elsewhere. Because almost 90 per cent of people coming intoLondon’s congestion zone already travelled by public transport, reduc-ing car traffic by less than a fifth affects only a small proportion of thetotal commuter flow (although a sizable portion of drivers). Thus thevast majority of people felt the road charge was reasonable. Once thishad been ascertained, it could be anticipated that any appetite for protestwould be manageable.

The close involvement of affected communities in decision-makingand planning for evaluation is also incrementalist. The business com-munity, which experienced sales losses, was engaged in what mightbe called supersession activity, helping to work out the overall impactby including benefits, such as lower costs to them for goods transportthrough the congestion zone.

The Authority’s evaluation of its initiative has been continuous, bul-letins being indicator-based and issued regularly. Scrutiny of the im-plementation has aimed at continual ‘squeezing all the error out of thesystem’ (Men (2003), 2). One can see that the insurance calculationsbefore implementation have worked out: at six months in, only 110 per-sistent non-payers have had their cars impounded. Even ‘outsourcing’provided opportunities to show firmness in evaluation. The companyhired by the Greater London Authority to implement and administer thepolicy ‘failed’ its targets in some respects (targeted revenue was bledby the unanticipated heavy compliance). Stringent penalties were ap-plied in contract renegotiation and were well publicized. In the firstsix months of operation, compliance with fees increased strongly, pay-ment rates for ‘penalty charge notices’ increased, calls for informationdropped from 167,000 per week in the first weeks of operation to 70,000a week, and the penalized public’s use of the appeal system droppedfrom more than half of all those who received fines to 16 per cent. Aftersix months, people were being educated and were mostly reconciled.

Transport for London’s 2005 document ‘Central London Conges-tion Charging Scheme Impacts Monitoring Summary Review: January2005’ covers the period between the previous review and the autumnof 2004. Overall, traffic levels in 2004 were very similar to the periodimmediately following the introduction of charging, when the numbers


of cars entering the charging zone during charging hours plummeted,with most of the displaced commuters moving to the bus service, for atotal increase in bus patronage of close to 50 per cent from 2002 to late2004. The average reduction in congestion remains at 30 per cent, withreduction in delays for peak-hour car journeys still significantly reduced(compared with the night travel rate). Road traffic accidents have beenreduced somewhat further over the gains registered in the year 2002-3.Journey times for buses improved, with a further 15 per cent reductionin 2004 on the indicator, excess waiting time over the 24 per cent reduc-tion achieved in 2003. The enforcement process in 2004 resulted in stillfurther increases in compliance and payments, and in reduced numbersof appeals over the previous period. Interestingly, the Authority contin-ued to press through 2004 for improvements in call centre performanceand enhancements to the options for payment of the charge. A surveyin the period found that more than 80 per cent of those who used thesystem were now satisfied with the service, although the media reportsome sentiment that the initiative involves too much revenue-seeking.

Another incremental policy with several related but decoupled facetswas initiated in France in late 2002 to reduce deaths and injuries onroads. The president of France, elected in May of that year, had sig-nalled his determination to reduce road deaths throughout France. Thegovernment established an inter-ministerial committee, assisted by aplanning and policy secretariat made up of members with a significantinterest in road security. In February 2003, a bill was presented to Par-liament that envisioned three main semi-independent initiatives aimedat the same goal: steeply increased penalties for road offences, par-ticularly for repeat offenders; an automatic system for identifying andfining driving offenders; and a general and evolving reinforcement ofsevere prevention strategies. At the end of February, the governmentannounced that judges would be given the authority to double the cur-rent penalty in aggravating circumstances. Drunk driving and recklessspeeding would be punished by imprisonment for as long as ten years.18

As the reforms took hold, the prime minister revised and improved the

18www.premier-minstre.gouv.fr (2002)). See also the French site,www.zeroaccident.net .


policy with the collaboration of lower-level jurisdictions.The most stunning impact was made by photo radar. With its imple-

mentation at the end of October 2003, a fixed and fully automatic fine of90 Euros for speeding was implemented. The flat fine is supplementedby a heavily graduated points system that gives drivers demerits propor-tional to their excess speed. Evaluating the initiative after two months,Le Monde (Le Monde(2003), 6) reported that the French, who habit-ually drove at speeds of 130 kilometres per hour on even the busiestroutes, had lowered their speed by an average of 10 kilometres per hourper driver. The results were a 25 per cent reduction in road deaths and20 per cent fewer accidents and injuries. The greatest decrease monthover month was in high-speed driving, defined as a minimum of 50 kilo-metres per hour over the limit. The following year, 2004, saw a further10 per cent decrease in deaths; the incidence of speeding at more than10 kilometres per hour over the limit also dropped, more than 15 percent from the 2003 figures (Le Monde(2004a), 8). Also in 2004, an‘accidentologue,’ a specialist in the study of accidents, described thereduction in deaths over the two years as ‘prodigious’ and as having re-sulted from the exertion of steady and varied pressures on drivers. Theexpert had never seen the state succeed in such a complex task (Got(2004), 8).

Some aspects of the first phase of implementation were not as ef-fective as they could have been. For example, the telephone informa-tion service intended to familiarize the national driving public with thephoto radar system was implemented with a mere five operators. Thisalmost negligible educational effort made for initial public confusionand disaffection, and of course means that initial call volumes cannotbe compared with later volumes to yield information for use in futurechanges.19 Nor did the first year of the measures decrease deaths to thelevel in Great Britain: in 2003, estimated road deaths in France were6,000 compared with about half that in Britain, whose road density isvery similar. (The further reduction of 2004 does however bring Francecloser to the British figures.) And French politicians are taking positionson the ‘radars’ that reflect their constituents’ impatience over the limi-

19Le Monde(2003), 6, and Gaetner (2003).


tation of freedom, the severity of the fines, and the apparent lack of anymargin in speed before a fine is levied, such that some fine-payers char-acterize the income aspect as a racket (Le Monde(2004b), 8). Thus thepolicy’s gains require steady discussion and marginal adjustments. Butover time it becomes clearer to all that it would be next to impossiblefor policymakers and politicians to ‘insure’ against the resulting certainloss of life that removing the radars would cause. In other words, thepolicy has succeeded because the public understands it and wants it andit lowers the death rate.

Incremental policies are more often accumulated over time in publicorganizations, than they are politically led, intricate, multi-stage initia-tives such as the two described above. Each of the two examples de-pends on a complicit if not convinced public, although the two reformsare both attractive and good examples of incrementalist policy. Speak-ing generally, it is difficult for officials to incite politicians to launch andlead full-scale reviews of accumulated policy in areas that seem routine,are technical, and affect large numbers of the public, because the workwill be detailed and based in legal and management concepts that aredifficult to master. Also, some of these policies are invisible, until some-thing goes wrong. But developments in other technical areas like foodsafety make it clear that it could be important for politicians to interestthemselves in accrued policy in apparently settled areas. Who wouldhave anticipated, for example, that an established policy of feeding ani-mal waste to ruminants (in this case, cattle), in operation in Britain since1926, would have become consequential for humans more than half acentury later, when temperatures for rendering animal waste into fertil-izer and animal food (an almost unregulated activity within the industry)were lowered to save fuel? A single decrement, the lowered tempera-ture that did not kill a particular organism, eventually led to the formof bovine spongiform encephalopathy (BSE) that is able to move frommeals to humans. Unfortunately BSE is now an all-too-familiar exam-ple of dramatic policy disaster brought about by apparently incrementalshifts in practice that are now seen to be pivotal decrements.

Political leaders are more likely to try for a dramatically compre-hensive policy accomplishment than to work in near invisibility to un-


derstand and improve accumulated policy. Politicians prefer a full-scaleanalysis and decision followed by a decisive act to move from ‘a well-specified initial state or unsolved problem to a well-specified final stateor the solved problem’ (Connolly (1980), 37). Because the aim is of-ten to build up a career or party legacy of dramatic accomplishmentsthat communicate a message of competence and sometimes of ideol-ogy, such ‘mega’ policy is often called, interchangeably, ‘agenda pol-icy’ and ‘legacy policy.’ Examples can range from physical projectslike the exploitation of tar sands in northern Alberta, to the construc-tion of the Millennium Dome with its anticipated program of continuedrevenue-generating activities in London, England, to paradigm shiftsin established policies, such as privatization of government bodies andservices, the American president’s interest since 2004 in implementingprivate accounts for social security provision, or the implementation ofa major symbolic change via the ‘personal community charge’ or polltax in Britain in 1988 under Prime Minister Thatcher.

Political and policy field analysts cluster around these dramatic, po-litically led policies.20 It is fair to say that they are typically perceivedas failures. Michael Moran (Moran (2001)) covers several families ofpolicy failures that began as full-scale, drastic plans. As one exampleof the failure of a policy intended to be realized in a single pass, thetotal cost of the Millennium Dome was projected in the second half ofthe 1990s at 200 million and eventually cost more than 700 million forthe physical structure alone. The first projection seemed so unrealisti-cally low that the responsible government accounting officer asked fora written instruction before he would issue the initial payments to getconstruction underway. The early warning was ignored.

Other failures resemble the reverse engineering of a received policy.One sees a shared world-view between an interest, such as the agri-cultural industry, and its nominal regulator, the government ministryresponsible, that allows both to perpetuate their construction of reality.The world-view in itself enforces institutional blindness and thus resis-tance to put anything into question, with consequences for the publicgoods jointly provided. For example, the dramatic failure to prevent

20See Moran (2001); Hood et al. (1999); Bovens and ’THart (1996); Hall (1981).


BSE, due to what looked at the time like an inconsequential efficiency-seeking change in rendering temperatures, occurred in such a setting.Shared monetary interest between regulated industry and the regulatoris enough to ensure their collaboration and disable critics. Interventionfrom the top of the political system, brought about by connections, caneven suppress free flows of crucial information and prevent the publicfrom looking after its own interests. In such silent decremental fail-ures, the resulting policy catastrophe emerges from the daily routinesof an organization. Thus, using Simon’s concepts in the wrong direc-tion, environmental perversity, often rooted in institutional incentives,carries some factor forward to disaster. One of Moran’s examples isthe Barings Bank failure, where irresponsible trading was allowed by anon-interfering, trust-based regulation style. Another is BSE in Britain.

BSE is also a worry in Canada. The Alberta Auditor General re-ported after the first half of 2004 on Canada’s testing record. Testingfor BSE in Canada depends on volunteerism. Individual ranchers decidewhether to submit for BSE testing the head of an animal slaughtered ontheir land—usually a ‘downer,’ an animal too sick to get up. Obviously,ranchers also use their personal judgment about whether or not to senda near downer to a licensed facility, where it could be spotted by inspec-tors and pulled out of the food chain for BSE testing. But the viabilityof ranchers and of the Canadian beef industry clearly depends on theabsence of physical evidence of BSE. The dynamics were made appar-ent in the fall of 2003, when Alberta Premier Ralph Klein seemed to tella meeting of U.S. governors and Western Canadian premiers that ranch-ers should ‘shoot, shovel and shut up’ when faced with a potential BSEcow (Globe and Mail(2003)). Fortunately, the publicity given to Mr.Klein’s remarks, plus the timeliness of the Alberta Auditor General’sReport at mid-year (Alberta (2004), combined to raise consciousnessand thus compliance. A failure was averted for 2004, but the incentivesystem for testing remains worrying.

6.5. CONCLUSION 161

6.5 Conclusion

In a world characterized by under-examined great leaps and receivedpolicy that contains unexamined problems, Braybrooke’s forward-lookingpolicy strategy looks helpful and important. He wants to engage morepeople in active deliberation on public policy in a style of analysis thatamounts to taking apart mentally and incrementally policy elements thatwere often put together in small increments over long periods of time inthe real world. Braybrooke’s claims are modest: relatively incrementalpolicies are more democratic in daily practice because they are moreintelligible and therefore more amenable to supervision by those mostaffected. They are relatively safe in comparison to policies undertakenas tightly coupled ‘single pass’ undertakings and, further, persistent in-cremental changes in a consistent direction can lead to great changeover time. Incremental policies are closely planned to be reversible, andcontingencies are closely monitored so that damage can be repaired.The attentiveness to detail in policies planned under incremental strate-gies becomes more attractive and important as the scale of operationsincreases (Braybrooke (1998a), chap. 16), because with detail the prob-ability of a smooth implementation improves. The French approachto lowering the number of road deaths is a large-scale policy that tookplace along with several decoupled but complementary policy innova-tions, each aspect requiring interventions with individuals. With care-ful planning for intelligibility, public acceptance is more readily earned(328). As Bailey and Braybrooke note in their essay on Dahl, frustra-tion with unintelligible policy issues feeds ‘the tendency of politics as ameaningful activity to vanish from the lives of more and more citizens’(Braybrooke and Bailey (2003), 111).

Over time, the alienation of citizens from the democratic process un-dermines genuine self-government and autonomy. The loss of people’scivic faith and activism—their intelligent support, energy, resources,goodwill, and insight applied to meeting their own needs and those ofothers—causes incalculable harm to the quality of public life and in-curs danger. Still more serious, and under-noticed by client-citizens,is the fact that the national political space addressable by those whom


the vote ‘puts into power’ is much less than it was. Political partieswere, until at least the final third of the twentieth century, reasonablyeffective mechanisms for organizing, aggregating, and creating citizenpreferences for fundamental state policies. Now central banks directeconomies in a globalized world. The efficient part of apparatuses thatgovern us, to nod to Bageot, has moved beyond national borders, andeconomic power is not itself subject to national governments. To quoteGunter Grass, ‘Parliament is degenerating into a subsidiary of the stockmarket.’

What is interesting, even in a world with fewer points of access andinfluence, is how constructive Braybrooke’s strategy is. Braybrooke isclear that people in groups can with effort and intelligence create newpolitical space around workable ideas—solutions that attract the interestof people whose values differ. Braybrooke would have citizens organizearound a problem without partisanship or fixed interests (the process ofdeliberation will identify interests in any case), and proceed by investi-gation and debate to parse their problem into the smallest possible un-coupled elements until good starting points can be identified and plansand actions developed. His strategy does not assume that all or evenvery many citizens must be mobilized for a reasonable start. Nor doesit assume that issue-oriented groups must have the ear of ‘power’ fromthe beginning. Rather it is more likely that a group’s work will attractthose with resources once it is well formulated and workable.

So-called ordinary people accomplish great things when they col-laborate. The Canadian prairies, particularly Saskatchewan, had a largeand thriving system of cooperatives in the middle of the twentieth cen-tury. Town populations, whose ‘knowledge society’ consisted of a doc-tor, a banker, and the school teachers, found in themselves the abilityand persistence to understand commodities markets, run businesses, andprovide leadership to professionals. And they still have it. One sees in-ternational renewed interest in health cooperatives, where citizens havea voice in the priorities of medical clinics and provide feedback to theirpractitioners.

Changes in scale, like feelings of impotence, can erode the empathycitizens feel for one another, reducing potential cooperation. In refer-

6.5. CONCLUSION 163

ence to scale, Braybrooke holds that our local participation, which wemay undervalue, is not to be seen as a first step toward exerting influ-ence in a higher jurisdiction, but rather as a life-long base for our prac-tice of ‘robust civility’ (Braybrooke and Bailey (2003), 104, coined inreference to Dahl). Robust civility is a quality that Braybrooke knowswell. He has investigated it in his work and lived it in his capacities ascitizen and colleague. In the 1960s and early 1970s, Braybrooke andLindblom’s prescription for prudently reversible and intelligible policychange was often summarily dismissed by so-called progressive intel-lectuals. It was interpreted as an endorsement of, or an apology for,the status quo, and therefore as getting in the way of large-scale, com-prehensively rational, radical policies. One can hope that Braybrooke’smore complete formulation of incrementalism with utilitarianism ad-dressing needs will encourage more citizens to take up tools to chipaway purposefully at public evils.

Chapter 7

Life of Pi and the Existence ofTigers

STEVEN BURNS

Abstract

As a tribute to David Braybrooke’s interest in the arts and his ecumenicalwork in philosophy, this essay gives a critical reading of Quebec novelistYann Martel’s Life of Pi. Martel offers in his novel both argument andillustration of the thesis that, not a principle of parsimony, but a principleof richness and persuasiveness governs which explanatory entities are tobe given status in an account of the truth. Pi, as an extreme case, is abelieving Christian, Muslim, and Hindu who claims that belief in Godis justified by the excellence of the stories that the religions tell. In thisessay, attention is focused on the existence of Pi’s tiger companion, whichis both denied and held to be necessary ‘in the story.’ Even in the faceof contrary evidence, belief in the existence of the tiger is said to be aprerequisite for appreciating the story at its best. Objections to Martel’s‘best story’ hypothesis are developed against the background of AnthonySavile’s parsimonious account of ‘best explanation’ in aesthetics, and anontologically more modest reading of Pi’s story is proposed.∗

∗The first version of this paper was delivered to the Visiting Speaker Series at the UniversityCollege of Cape Breton, January 2004. Subsequent versions have been discussed with colleaguesat theZentrum fr Kanada-Studienat the University of Vienna and at Dalhousie University’s Phi-losophy Colloquium. I am indebted to the habitus of the Boys’ Book Club with whom I firstdiscussed Martel’s novel, and I am grateful to many others who have also helped me to make mystory clearer.

165

166 CHAPTER 7. STEVEN BURNS

David Braybrooke surprised me above all by the range of his interests.I first saw the book he co-authored with Lindblom displayed on themantel of a colleague.1 I was being welcomed to the Dalhousiedepartment as a new junior department member. Not noting theauthors’ names, but being surprised by the sociological cast of theirtopic, I made a lame joke about the ubiquity of little red books withpractical ambitions (this was in the 1960s, when Mao Tse-Tung’s ‘littlered book’ was on everyone’s shelf). David was present, and unfazed bymy tactlessness. He was soon introducing me to new string quartetsand new colleagues in other disciplines, to local as well as universitypolitics, to new artists and novelists, and of course to a wide range ofphilosophical topics pursued with rigour and enthusiasm. He set astandard of multinational, multilingual multidisciplinarity that I andmy colleagues could barely recognize let alone emulate. It was he whomade it immediately clear to the new appointees that we were expectedto publish, and to get us started he sent our names and areas of interestto the book review editor of the Canadian Philosophical Association’sjournal Dialogue. David may have led by example, but he was notabove shoving from behind at the same time.I had adopted from my doctoral supervisor, Peter Winch, the view thatphilosophers’ schematic examples and their logical formalizationscould obscure as well as illuminate philosophical problems. Here is anexample: You are one of two people in a lifeboat that will hold onlyone person. Should you jump overboard or push the other personoverboard? In matters of ethics and philosophy of mind, Winch held,the examples described in subtle detail by good novelists were likely toreveal things to the philosophical analyst that were really worthknowing. I was pleased to find that there was room in Braybrooke’svision of philosophy for this rebellious view, and indeed that he wasmaster of the detailed example himself.2 It is as a tribute to theseaspects of his extraordinary influence on a department in which it hasbeen a privilege and a pleasure to work that I dedicate the following

1Braybrooke and Lindbloom (1963)2In one apt case he relished the details of a real-life example, the lifeboat cannibalism of the

survivors of the shipwrecked Mignonette. See Braybrooke and Fingard (1985)

167

adventure to David Braybrooke. It concerns yet another lifeboatadventure, and was first conceived for a general audience. As a result,it begins with a novel and sneaks up on the philosophy, rather than theother way around.There is precedent for this, too, in Braybrooke’s work. He oncecomposed a comic piece for the Arts and Sciences Faculty Council ofDalhousie University, ‘A (Summer’s) Day in the Life of a Professor ofPhilosophy.’3 It had the serious aim of indicating that professors ofhumanities do not take four-month vacations between April andSeptember. It had comic consequences: at its next meeting the Councilvoted to have the text removed from the minutes, apparently on thegrounds that some members did go walkabout in the summer and werehumiliated by the comparison with the fictional Professor Green’sindustriousness. What I remember enjoying most about the ‘Day in theLife’ was that by mid-evening Prof. Green turned to ‘lighter reading,’though Braybrooke was quick to add that ‘even this may have a bearingon his work.’ It was in just such a moment that I took upLife of Pi, andit has indeed had a bearing on my philosophical work. That I have beenmoved to write this paper is evidence that David was right yet again.Dalhousie University has a long tradition of teaching a class onphilosophical ideas in literature. Begun by J. Gould Schurman (whowas professor of English and then professor of philosophy in the1880s), the class was made famous by Herbert L. Stewart from the1920s to the 1940s4 and revived by Robert M. Martin in the 1970s and1980s. I have several times been allowed to teach it, and were I toteach it again I would includeLife of Pi, because it is an extraordinaryliterary experience with ambitious philosophical content.Yann Martel had a degree in philosophy and two earlier works offiction to his credit whenLife of Pibrought him international fame.5

3Braybrooke (1985) (Editor’s note: the essay appears in this volume as Appendix A.)4‘Stewart’s course ... which brilliantly analyzed the writings of Hardy, Meredith, Carlyle, Mrs.

Humphry Ward, H.G. Wells, George Bernard Shaw, and others, was an educational experiencethat had few equals in the country,’ wrote a former student, T.A. Goudge, in Goudge (7 68), (542).Goudge was by then chair of the department of philosophy at the University of Toronto and hadbeen invited to write this pioneer essay on the history of Canadian philosophy by none other thanDavid Braybrooke.

5Yann Martel,Life of Pi (Toronto: Random House, 2001). The book won the Hugh MacLen-


The book has four parts. There is an ‘Author’s Note’ at the beginning,and other, smaller, author’s notes scattered through the text. They areprinted in italics. Up to a point the author’s story matches Martel’sbiography, and I shall call him ‘Martel,’ but the more sophisticatedreader will insist that this ‘author’ is also a fictional character. Thebody of the book is in three parts; it is the story of the life of a personnamed Pi, which is short for Piscine Molitor Patel. Part I is called‘Toronto and Pondicherry,’ because Pi grew up in the south of India, inthe French-colonial city of Pondicherry, where his father was thezoo-keeper, and then emigrated to Canada, where he studied atuniversity and now lives as a middle-aged man with his wife, twochildren, a dog, and a cat. As the author says at the end of Part I, ‘Thisstory has a happy ending’ (103). It is a story that can make your spiritsoar. Part II is called ‘The Pacific Ocean.’ This is the most famous, andlongest, part of the novel, in which the sixteen-year-old Pi spends morethan seven months in a lifeboat accompanied by a 450-pound Bengaltiger and very nearly starves to death. It can strike you with terror andmake your heart ache. Part III takes place in the infirmary in Mexicowhere Pi finally came to shore after his extraordinary ordeal. It is onlythirty pages long, but it makes your head swim. Together these foursections constitute a remarkably unified novel, made of strikinglydiverse parts. I shall begin at the beginning.

7.1 Author’s Note

The author’s note tells us two key things. First, he went to India towork on his next novel. It was to be set in Portugal, but India was acheaper place to live, and besides, ‘That’s what fiction is about, isn’t it,the selective transforming of reality? The twisting of it to bring out itsessence? What need did I have to go to Portugal?’ (vi). That novel waseventually abandoned. It had many virtues:

nan Prize, the 2002 Man Booker Prize, the 2004 German Book Prize, and in 2003 was selected bya CBC panel for the ‘Canada Reads’ national radio program. In what follows, citations by pagenumber alone are to this edition.

7.1. AUTHOR’S NOTE 169

Your theme is good, as are your sentences. Your charactersare so ruddy with life they practically need birth certificates.The plot ... is grand, simple and gripping. You’ve done yourresearch, gathering the facts—historical, social, climatic,culinary—that will give your story its feel of authenticity.The dialogue zips along, crackling with tension ...[However] an element is missing, that spark that brings tolife a real story ... Your story is emotionally dead. (vi, vii)

This episode gives us a hint about the importance of fiction and itsnature. It is to deviate from reality in order to show the essence ofreality. As he concludes: ‘If we, citizens, do not support our artists,then we sacrifice our imagination on the altar of crude reality and weend up believing in nothing and having worthless dreams’ (xi). Thesecond key thing is that he meets in Pondicherry a man who learns thatMartel is a writer. This leads immediately to the familiarphenomenon—have I got a story for you!—and he promises him ‘astory that will make you believe in God’ (viii). It is Pi’s story. Martelthen goes to Toronto to meet Pi, hears his story first-hand, and retells itto us. Clearly this story, unlike the one that was discarded, will not be amere description of crude reality, but will have emotional life. So itdoes. And ‘it will make you believe in God.’ My central question is,how is the story supposed to make us believe in God, and can we agreethat it does?We may think that the answer is that Pi’s survival is a miracle. If youbelieve that a child can live for seven months in a boat on the openPacific, hallucinating from starvation and dehydration, andaccompanied by an adult tiger, then you have to believe in divineprovidence. That may be what the man believed who told the authorabout this amazing story. It is not the point that Martel wants to make.Pi may think that God was with him on the boat.6 But Pi was a believerbefore his adventure; it was not the adventure that made him believe or

6Chapter 52 consists of a complete list of the contents of the lifeboat. It ends: ’1 boy with acomplete set of light clothing but for one lost shoe / 1 spotted hyena / 1 Bengal tiger / 1 lifeboat /1 ocean / 1 God’ (162).


should make us believe. In fact, I think that Martel has a very differentidea in mind, so I shall propose a much longer answer.

7.2 Toronto and Pondicherry

A word about our hero’s name is in order. His uncle had lovedswimming and had studied in Paris. ‘Piscine Molitor’ was the name ofhis favourite swimming pool, so he suggested that name for hisnephew. Piscine Molitor Patel’s schoolmates, however, found ‘Piscine’an odd name; they teased him and called him ‘Pissing.’ So one schoolyear, Piscine renamed himself ‘Pi’—‘equals 3.14,’ he added. It stuck.(So, too, did the association with water and the idea of mathematicalscience.)Two important themes are established in the first one hundred pages:God and nature. Nature takes the form of wild animals. Pi studies bothreligion and zoology when he finally gets safely to Canada and attendsuniversity. In fact we first meet him explaining his research on thethree-toed sloth. He found its calm and introspective demeanoursoothing to his shattered self. In fact, ‘[s]ometimes I got my majorsmixed up ... [The sloth] reminded me of God’ (5). Much moreinstructive is Pi’s childhood experience. His father was owner of thePondicherry Zoo, and he thought his childhood ‘was paradise on earth’(15). He was wakened by roaring lions in the morning, soothed by thebenevolent gaze of bison and orangutans, and delighted by the brilliantcolours of the birds. In the midst of this Garden of Eden he learned that‘the most dangerous animal in a zoo is Man’ (31). It is not that thegreat carnivores are not dangerous; in fact the reader will not quicklyforget the scene in which Pi’s father takes his sons to watch a live goatbeing released in the cage of a hungry Bengal tiger. ‘Tigers areverydangerous,’ his father shouted above the snarling of the beast. ‘I wantyou to remember this lesson for the rest of your lives’ (37). Pi latercomments: ‘It was enough to scare the living vegetarian daylights outof me’ (39).We are then given a lesson on how humans can create an environmentfor a wild animal that will allow it to be emotionally stable and

7.2. TORONTO AND PONDICHERRY 171

stress-free even in confinement. In these pages of ingenious lore aboutzoos and lion-tamers, and about wild animals that are more interestedin comfort and safety than in what we think of as ‘freedom,’ Martel isclearly preparing us for Pi’s great adventure, but we have as yet no ideawhy or how this is going to be relevant. Even less do we know how the‘God’ theme is going to be relevant. When he visits Pi in Toronto, theauthor describes his house as a temple. It has shrines to Hindugods—Shiva, Krishna, Shakti, and Ganesha; it has icons of thecrucified Christ and the Virgin Mary, and it has an Arabic prayer rugnext to a veiled Koran (chapter 15). His university degree in religiousstudies was not something that he forgot about after getting hisparchment. But this interest in religion begins in his youth. The youngPi is a kind of religious genius. He is raised a Hindu and loves its‘sculptured cones of red kumkum powder and baskets of yellowturmeric nuggets’; he feels at home in its rites and rituals, and ‘theuniverse makes sense to [him] through Hindu eyes’ (53). There is amoving lecture on this. We are introduced to the concept of Brahmannirguna, the original oneness, without qualities, beyond understandingand beyond language. There follows Brahman saguna,with qualities,with moral attributes and taking the forms of the distinguishable godsmentioned above. But Brahman is also atman, the breath of the spiritwithin each of us, a pilgrim spirit which seeks to find its karma, to be atone with the Absolute.Pi is a well-content Hindu when at fourteen he encounters Jesus Christ.In his community Christians have a ‘reputation for few gods and greatviolence’ (56), but he meets a kind priest who teaches him theextraordinary story of a God who became a humble human, washumiliated, and put to death. Imagine, thinks Pi, if my father had saidto me,

‘Piscine, the lions are killing the other animals in the zoo.The situation has become intolerable. Something must bedone. I have decided that the only way the lions can atonefor their sins is if I feed you to them.’ ‘Yes, Father, thatwould be the right and logical thing to do. Give me amoment to wash up.’ (58)


What a weird story. Pi asks for a different story, but finds that theChristians only have the one Story; it is about love and humility. So Pifinds himself haunted by this new god and becomes a Christian.

‘Islam had a reputation worse than Christianity’s—fewer gods, greaterviolence’ (64), but a year later, at fifteen, Pi meets a Muslim teacherwhose humble friendship leads him to discover the postures of Islamicprayer, the ritual recitation of the ninety-nine names of God, thesoothing beauty of a ‘religion of brotherhood and devotion’ (67). Pihas not, however, abandoned Hinduism or turned his back onChristianity. He believes all three. He is in love with God. ‘I was apractising Hindu, Christian and Muslim’ (71). Of course he is attackedby the representatives of all three faiths; you can only be one of those,they say; the others are heresies or superstitions! But the youngsterquotes Mahatma Ghandi: “‘All religions are true.” I just want to loveGod’ (76). And so he does.

It is not at this point clear what this aspect of Pi’s character has to dowith his great adventure. It is not obviously his faith in God that getshim through his ordeal on the Pacific. Nor is it clear what his survivaladventure has to do with his faith. He believed in the Divinity of allthree religionsbeforehe left India for Canada. So Martel has created avivid character as a comment on our relation to nature on the one hand,and our relation to the spirit of religion on the other, but has given usfew clues about how this is to relate to the story that will make usbelieve in God.

There is one minor theme in Part I that should be mentioned. Recallthe author’s opening remarks about fiction, the story that transformsreality, and his warning that we should not sacrifice our imagination onthe altar of crude reality or we will end up believing in nothing. Thattheme is given some new words by the adult Pi as he retells his story.‘Dry, yeastless factuality’ is one phrase. ‘The better story’ is another(70). The two phrases appear in what we might consider an argumentin favour of believing in God. Imagine an agnostic’s last words as heslips from life: ‘if he stays true to his reasonable self, if he staysbeholden to dry, yeastless factuality, he might try to explain the warmlight bathing him by saying, “Possibly a f-f-failing oxygenation of the

7.3. A DIGRESSION ON BELIEF 173

b-b-brain,” and, to the very end, lack imagination and miss the betterstory’ (70).

7.3 A Digression on Belief

A philosopher cannot read about Pi’s belief in these varied godswithout recoiling. It is bad enough to believe in any unprovable entity,but to believe in three incompatible ones is intolerable. Yet Pi is awonderfully believable character. How is this? What does logic tell usabout belief? That one cannot believe bothp and not-p in the same wayand at the same time; that is a contradiction. But this tells us preciouslittle. The dialecticians will ask, ‘What is not-p?’ and answer,‘Anything butp.’ So if p = ‘god exists’ then ‘roses are in leaf outsidemy window’ is an instance of not-p; and one can easily believe in bothp and not-p. That is not the point! By ‘in the same way’ we meant thatone cannot believe both that God exists and that God does not exist. Orshould that be: “‘It is true that God exists” and “it is not true that Godexists”’? Better still, is the problem not this: that one cannot bothbelieve that God exists and not believe that God exists? I am skippingpast distinctions of great logical import, but in any case it can beargued that belief comes in degrees. This fact should show that toclaim that there are no options other than to believep or disbelievep isan irresponsible simplification of the mental life. I can believe inp alittle bit or a lot. No doubt I can at the same time and in the same waybelieve not-p. Surely this is what most of our minds are like most ofthe time. Philosophers, however, insist on preserving the originalinsight that it is a contradiction, an impossibility, both to believe andnot believe the same thing in the same way at the same time. So theywrite: ‘Those who prefer to think in terms of degree of belief shouldread such expressions as “S believes thatp” as shorthand for “Sbelieves thatp to a degree greater than 0.5 (on a scale from 0 to 1).”’7

Does this really help? If it was implausible that one both believep andnot-believep, is it equally implausible that one be 0.51 certain ofp and0.51 certain of not-p (or 0.51 uncertain of p)? Surely maintaining that

7Mele (2001), 10.


this is a flat contradiction belies our real inner lives. Only a DivineKnower could parse our beliefs that finely! And we are doingepistemology, not theology.If an account that allows for degrees of belief makes matters morecomplex, consider what happens when we allow that some people aremore capable of believing than others. Some people have bighearts—there are scarcely limits to how much they can love, and thoseclose to them might well wish that they love less smotheringly. Othersmay love as much as they can, and one might wish that they could haveloved more. So it is with the capacity for belief. On an Aristotelianscale, those deficient in belief go through life as sceptics or ironists,while those with an excess of the capacity are as it were gluttons ofbelief. Pi seems to be a belief glutton. I think that if he believed 0.51 p,he would believep a great deal more than would an ironist whobelieved 0.51 p. These are two different scales of ‘degrees of belief.’Let us construct a paradox: Pi is undecided about something,q; hebelieves 0.5q and 0.5 not-q. An ironist, let us say, believes only half ashard as Pi. If she believed 100 per centq she would be in the samepsychological state as Pi with regard toq. She is capable of that. She isalso capable of 100 per cent disbelief thatq. Now if she were inexactly the state that Pi is in, she would not be undecided, as he is, butwould be believing 1.0q and disbelieving 1.0q. That is plainly aproblem, indeed approaching absurdity. Yet the state is possible,because Pi is in it. Perhaps an Aristotelian prescription for this muddlewould be apt. A virtuous believer would be one who sought the meanbetween these extremes, who believedp to a degree supported by theevidence forp, but no more than that. She would be neither abelief-ascetic nor a belief-glutton, but a moderate or virtuous believer.Aristotle, of course, wisely warns us not to treat the mean as amathematical point, and says that we should expect no more precisionin such matters than is appropriate.8 Nonetheless, on this sort ofaccount we might conclude that Pi is not a virtuous character; he

8See, for instance, Aristotle,Nicomachean Ethics, I:3. See also II:7, on truthfulness as amean, and III:9, where after a long discussion of courage he concludes that from this we can format best a rough conception of its nature (1117b).

7.4. THE TIGER IN THE LIFEBOAT 175

believes much too much, too freely, and too heartily; but of course weshould not doubt that he is possible.

7.4 The Tiger in the Lifeboat

In the second part of Pi’s story, titled ‘The Pacific,’ we have the heartof the novel. The ship carrying Pi and his family and the family’s zooanimals to Canada suffers an unexplained explosion and sinks in roughwind and water in about twenty minutes. I shall be very sketchy aboutthis story, for it contains much magic that I do not want to spoil. Pifinds zoo animals thundering about the deck.9 Some crewmen throwhim into a lifeboat towards which he sees a favourite tiger desperatelyswimming. The tiger is named Richard Parker because of a clericalerror, but that is an anecdote which I shall leave for the reader to enjoy.Suffice it to say, this character effectively does have a birthcertificate.10 Pi throws him a lifebuoy and he climbs in the boat beforePi realizes how stupid he has been. He also finds in the boat a zebra,which has broken its leg jumping in, and a spotted hyena. Then anorangutan floating on a bale of bananas reaches them. So he is in alifeboat with four dangerous, wild animals!Over the course of the next while, as the storm abates and beautifulsunrises break the darkness of the nights, the animals hold the stage.The hyena, vividly described, begins to eat the injured zebra. Then itfights the orangutan and kills her. Pi is about to defend himself byattacking the hyena when the tiger, who has been under a tarpaulin,reappears. Pi flees to a tiny raft that is dragging behind the boat. Thetiger takes care of the hyena.

9‘I clearly heard monkeys shrieking. Something was shaking the deck. A gaur—an Indianwild ox—exploded out of the rain and thundered by me, terrified, out of control, berserk. I lookedat it, dumbstruck and amazed. Who in God’s name had let it out?’ (115).

10I am referring here to the pedigreewithin the novel. Martel has explained that he chosethe tiger’s name because of several external coincidences. The cabin boy who was killed andeaten by the survivors of theMignonettein 1884 was named Richard Parker (see note 2, above).Remarkably, so was the man eaten by the survivors of a fictional shipwreck in Edgar AllenPoe’s only novel,The Narrative of Arthur Gordon Pym, written nearly fifty years earlier, in1837. Seehttp://www.amazon.com/exec/obidos/tg/feature/-/309590/ (ac-cessed 27 December 2004).


By now Pi has found the stock of water and food in the lifeboat. Heregains his spirits and then figures out how to train the tiger torecognize a division of territory.11 For instance, he uses a lifebuoywhistle as a sort of whip. He even marks his end of the boat with hisurine. The tiger eventually settles into his comfort zone at the other endof the boat. And so they travel. Pi catches fish and turtles, and feedsRichard Parker. He figures out the water treatment equipment, andkeeps them in a minimum of fresh water. Eventually, ‘It was RichardParker who calmed me down. It is the irony of this story that the onewho scared me witless to start with was the very same who brought mepeace, purpose, I dare say even wholeness’ (179). When a school offlying fish, who are in turn being hunted by predators, fly over and intothe boat,

Richard Parker ... raised himself and went about blocking,swiping and biting all the fish he could. Many were eatenlive and whole, struggling wings beating in his mouth. Itwas a dazzling display of might and speed. Actually, it wasnot so much the speed that was impressive as the pureanimal confidence, the total absorption in the moment.Such a mix of ease and concentration, such abeing-in-the-present, would be the envy of the highestyogis. (201)

They develop routines. There are the inevitable sharks, some hugestorms (chapter 83!), a ship that steams by without seeing them(chapter 86), cycles of boredom and terror, elation and despair. Butscarcity of food and water eventually wears them down. They seemabout to die. The tiger, gaunt and barely panting, lies in the bottom ofthe boat. Pi goes blind, begins to hallucinate. There is a conversationwith another lifeboat that seems to be imaginary.Then the most wonderful delirium takes over. Pi sees an island, anisland of floating algae, with ‘trees growing out of pure vegetation’(285). There is edible seaweed, pools of fresh water, and marvellous

11See especially the detailed instructions for establishing dominance over a wild animal inchapter 71.

7.5. OCKHAM’S RAZOR 177

inhabitants, the gentle, sociable meerkat. This story, with its rejoicingand rejuvenation, its secret dangers and surprises, has to be read to bebelieved. It is an astonishing tour de force, a chapter of thirty pages ina book whose other chapters average three pages each. Within anotherpage their boat is washed up on the shores of Mexico. Richard Parkerjumps right over Pi.

I saw his body, so immeasurably vital, stretched in the airabove me, a fleeting, furred rainbow. He landed in thewater, his back legs splayed, his tail high, and from there, ina few hops, he reached the beach. He went to the left, hispaws gouging the wet sand, but changed his mind and spunaround. He passed directly in front of me on his way to theright. He didn’t look at me. He ran a hundred yards or soalong the shore before turning in ... [He] looked fixedly intothe jungle. Then Richard Parker, companion of my torment,awful, fierce thing that kept me alive, moved forward anddisappeared forever from my life. (315-16)

7.5 Ockham’s Razor

Here we must introduce another philosophical theme. Ever since Platothere has been a debate about whether going beyond yeastlessfactuality can be justified. In Parmenides Plato criticizes himself fortrying in his earlier dialogues to explain the phenomena of the world interms of his theory of Forms.12 There are enough things in the worldalready without postulating an unchanging Form for every quality ordistinction that we see among phenomena. That just adds a multitudeof new things to the universe that need explaining themselves. Worse,he adds arguments to show that there are not just many new things, butinfinitely many of them! And if that were not enough, he tries to showthat even if they did exist they would be so disconnected from theworld of apparent things that they could not explain them anyway. It isas if he invented the slogan, ‘Do not multiply entities needlessly.’

12More properly, Plato portrays an elderly Parmenides criticizing a young Socrates for holdingthese views.


That is the principle known as Ockham’s razor.13 Ockham posed anenormous challenge to the philosophy of his time. Its centralcharacteristic is known as ‘realism.’ ‘Realism’ is a peculiar term inphilosophy. In ordinary language we know that it means roughly this:the world around us is real, we can grasp it with our minds, the worldthat we are conscious of is not a figment of our imaginations, it is notdependent on our minds at all. It is there whether we think of it or not.Reality is independent of our ideas of it. When philosophers speak ofrealism in connection with the Middle Ages, however, they meansomething quite different. They mean that Platonic Ideas are real. Thishas its roots in Plato’s teaching that the Forms are the reality of whichthe empirical world is just the appearance. Forms are the stable andknowable objects of the intellect. What we perceive through the sensesis constantly changing; they are objects of belief, but never ofknowledge. By Ockham’s time the doctrine was that

the human intellect discovers in the particulars apprehendedby sense experience an intelligible order of abstractessences and necessary relations [which are] ontologicallyprior to particular things and contingent events and thatfrom this order the intellect can demonstrate necessarytruths concerning first causes and the being and attributes ofGod.14

Ockham rejected this realism, maintaining that our knowledge is basedon direct experience of particular things and events, not on grasping anintelligible essence. He was realist in the ordinary sense, for he thoughtthat the world is real and that we can have knowledge of it, but he wasa nominalist in the medieval sense. He thought that the separate forms

13William of Ockham is arguably the greatest philosopher England has produced. He was atleast the greatest philosopher of his century (c. 1285-1349). He studied and taught at Oxford inthe 1310s, spent four years in Avignon in the 1320s being examined for heresy by agencies ofPope John XII, and lived in the court of the Emperor Louis of Bavaria in the 1330s and 40s, wherehe defended the divine right of kings against the traditional view that only the Pope had a divineright to power and that kings got their power from the Pope. He is believed to have died in 1349of the black plague.

14‘William of Ockham,’ Encyclopedia of Philosophy, ed. Paul Edwards (London: Collier-Macmillan, 1967), 8: 307


or essences are not real. The names we give them are just that, names(nomina), words that do not stand for real entities.The name Ockham’s razor is given to a principle that Ockham did notinvent but often used. It is often called the principle of parsimony,15

and is usually given as ‘Do not multiply entities needlessly.’ Ockhamlikely did not use exactly this formulation,16 for the good reason that itseems to imply an ontological interpretation—that a logical principlecan lead us to conclusions about the existence or non-existence ofthings—but he frequently used near neighbours: ‘Plurality is not to beassumed without necessity’ and ‘What can be done with fewerassumptions is done in vain with more.’ These slogans are intended tomean that if you have some things to explain, and two possibleexplanations of them, the one that requires fewer extra principles, orwhich is simpler, is the better explanation, the one more likely to betrue.What makes a story simple? Compare what makes an explanatorysystem simple. Leibniz thought that this is the best of all possibleworlds because God chose, among all those possible worlds, the onethat had the greatest variety of phenomena with the least number oflaws. Imagine a world in which when you drop a pencil it falls most ofthe time. The other times it rises, or floats, or dances a jig. That wouldbe a world with more phenomena, more odd things happening, butmuch more complicated natural laws—gravity would be overruled bywho knows what other varied explanatory principles. (Imagine a worldin which the human female had her teats where a cow does, or in whichthe human male had his penis where a pig has its tail. That wouldcertainly be odd enough. I suppose that the missionary position wouldcease to be a preferred form of reproductive activity. That would besad, since having sex face to face is one of the things that makeshumans a more loving species, but perhaps we could get used to it. Butwhat if some men had theirs rear-mounted and others forward; or ifsome women had theirs up and others down? That would not only be a

15Probably first found in Aristotle,PhysicsI:4, 188a, 17-18: ‘it is better to assume a smallerand finite number of principles.’

16As a Dalhousie student once argued. See Schlosser (1983).


world with weird standards of beauty, but one in which biology was aneven weirder science; either the laws of genetics would varyinexplicably, sometimes producing one type and sometimes the other,or else they would be more complicated in ways I shall leave the readerto imagine.)On the other hand, Leibniz thought that if there were fewer explanatoryprinciples—for example, that all bodies seek rest and there is noprinciple of inertia to keep motion going—then soon all motion wouldcease. That would be a world with vastly less variety of phenomena.So, he thought: this actual world is the one chosen by the creatorbecause it is the best solution to the problem of how to get the greatestvariety of phenomena with the simplest explanatory principles.Ockham and Leibniz, of course, used God as a first principle. They didnot intend that Ockham’s razor be usedagainstthe idea of a creator, letalone to reduce the number of entities in the universe by one, but infact that is what has happened. The modern world discovered that itdid not need God as a hypothesis to explain what could be explainedmore simply by natural science. So the principle of parsimony is usedto adjudicate between competing scientific theories: given equalpredictive power, the theory with simpler laws or presuppositions is thebetter one; and it is used to show that we do not need the concepts ofPlatonic Form, soul, or God, any more than we need the concepts ofphlogiston or aura. Adding God to our explanatory system does nothelp to explain anything more than what we can explain by othermeans—it just indicates that we have not discovered the explanationyet. So, to put it in the familiar (if perhaps mistaken) ontological form,God is an extra item in the list of things in the universe, one that is notnecessary, and ‘we should not multiply entities beyond necessity.’It is in this spirit that Anthony Savile wrote his account of bestexplanation in aesthetics.17 His main aim in this paper is to explain theplace of the artist’s intention in appreciating a work of art. Hisconclusions are that so far as determining the full basic text of the work

17Anthony Savile, ‘The Place of Intention in the Concept of Art,’Proceedings of the Aris-totelian Society69, n.s. (1968-9). I shall refer to it as reprinted in Harold Osborne, ed.,Aesthetics(Oxford University Press, 1972), 158-76.


is concerned the artist has the last word, but so far as determining thecorrect reading of the text is concerned the artist only has the firstword.18 The idea that there is a ‘correct reading’ of a work of art is ofcourse controversial, but invariably some readings are better thanothers, and typically one is better than all the others. We can call it thebest available reading and use this as a gloss on ‘correct.’ This view,that the artist does not have the last word about what is the correctunderstanding of his or her work, requires Savile to provide an accountof how such a judgment is to be justified. His idea, borrowed fromWittgenstein, is that judgments in a court of law can provide a usefulmodel.19 When a sentence in a contract needs to be interpreted by acourt, the signatories’ intentions are not decisive; they are what is atissue. What the court tries to determine is what the sentence in itscontractual context means, what the correct reading of it is. HereSavile writes:

The closest I have been able to come to finding an accountwhich meets these various demands is to say that theobjective meaning of a string of words uttered in a givencontext is given by that reading of those words which (a)accounts for the presence in that string of as many relevantfeatures of that string as possible and in the best of casesaccounts for all of them; (b) which is as simple as any otherequally complete reading; (c) which gives as unitary anaccount as possible of that string; and (d) which makes theproduction of the string appropriate in the intersubjectivelyidentifiable circumstances of its utterance.20

The correct reading of a work of art, that is, is the one that accounts forthe greatest number of relevant features of the work, while being assimple and unitary and appropriate as the competing readings. Thisgives for aesthetic judgment a set of principles very much like the

18Ibid., 164.19See G.E. Moore, ’Wittgenstein’s Lectures in 1930-33,’ in Philosophical Papers (London:

George Allen and Unwin, 1959), 315.20Savile, ’The Place of Intention in the Concept of Art,’ 169-70.


razor. If you are trying to account for the greatest number and varietyof phenomena with the most parsimonious set of principles possible,then you are trying not to multiply entities beyond necessity.There are problems with such an account. I shall not pursue them here,but it should be mentioned that Savile gives us no guidance about howto choose between two readings of equal explanatory power, one ofwhich is simpler and the other more appropriate or unified. Nor doeshe explain why we should prefer a reading that accounts for a fewmore relevant details but is much less simple and unified, to one ofmuch greater elegance but slightly less explanatory power.Nonetheless, he is clearly in the tradition of Ockham and Leibniz. Thebest reading of a work of art will not multiply entities needlessly. Ihave introduced this to my discussion of Pi because Martel explicitlydisagrees with this sort of view. He favours the baroque, he admiresmultiplicity, he is tolerant of contradictions and incompatibilities, hedenies that necessity is a virtue, and he champions the imaginativetranscendence of mere factuality. The remainder of my reading of Piwill be aimed at adjudicating this dispute.

7.6 Ministry of Transport Investigators

In the final section of the novel, Martel creates a second framing devicefor his story. Two officials from the Japanese Ministry of Transport, MrChiba and Mr Okamoto, assigned to investigate the lost ship, are sentto Mexico to interview the lone, and entirely unexpected, survivor. Piis now in the local infirmary, beginning to recover from his ordeal. TheJapanese officials have imperfect English, so they record theirconversations with Pi. These exchanges are very amusing, but alsovery disorienting. Chapter 96 reports them as asking Pi to tell themwhat happened to him, ‘with as much detail as possible’ (323). Chapter97 just says, ‘The story’ (324). That is, the whole story of the ordeal inthe lifeboat (chapters 37 to 94) is first told here. What we have read inPart II is Martel’s version of an older Pi’s version of the story that hetold to the Transport Ministry men. Presumably, Chapter 97 wasshorter and less detailed than the version we have just read. The big

7.6. MINISTRY OF TRANSPORT INVESTIGATORS 183

shock, however, is that the officials do not believe the story. Suddenlywe as readers are shifted from our suspension of disbelief. We haveread the story, of course, as a believable adventure. Reality is strangerthan fiction, we say, and when we read a story we are in believingmode. Suddenly someone is saying, ‘What about this algae island yousay you came upon? ... [It] is botanically impossible’ (326). ‘Nowabout the tiger, we’re not sure about it either’ (328), and so on. Pi inthese pages is a very self-assured teenager. He retorts, ‘What you don’trealize is that we are a strange and forbidding species to wild animals.We fill them with fear. They avoid us as much as possible’ (329).

‘In a lifeboat? Come on, Mr. Patel, it’s just too hard tobelieve!’‘Hard to believe? What do you know about hard to believe?... Love is hard to believe, ask any lover. Life is hard tobelieve, ask any scientist. God is hard to believe, ask anybeliever. What is your problem with hard to believe?’‘We’re just being reasonable.’‘So am I! I applied my reason at every moment. Reason isexcellent for getting food, clothing and shelter. Reason isthe very best tool kit. Nothing beats reason for keepingtigers away. But be excessively reasonable and you riskthrowing out the universe with the bathwater.’‘Calm down, Mr. Patel, calm down.’ (330-1)

This sceptical inquisition continues for some time, and the reader mustnow feel quite disoriented. What is the status of Pi’s story? Theinvestigators finally say,

‘We would like to know what really happened.’. . .‘So you want another story?’‘Uhh . . . no. We would like to know what really happened.’. . .‘I know what you want. You want a story that won’tsurprise you. That will confirm what you already know.


That won’t make you see higher or further or differently.You want a flat story. An immobile story. You want dry,yeastless factuality . . . You want a story without animals.’(335-6)

This challenge of course returns us to the theme that was broached inPart I.21 So Pi tells them another story, in whichheswam to thelifeboat, only reaching it because the French cook threw him alifebuoy, and his mother made it to the boat clinging to a bale ofbananas. The sailor was already aboard, with his broken leg. But thecook soon went out of control, butchering the injured sailor for foodand bait, and plundering the boat’s supplies of water. He killed Pi’smother when she attacked him for hitting her son. Pi killed him thenext day, when the cook—perhaps mad, perhaps overcome withremorse—did not bother to resist. Pi concludes, ‘Solitude began. Iturned to God. I survived’ (345). This story is sometimes viscerallybrutal. I shall not quote the gory bits, but I am reminded of a scene inMartel’s earlier novel,Self, in which the female narrator is raped.22 Heis capable of evoking profound horror and disgust.The Japanese investigators ignore the horror and quickly point outwhat the reader has probably already noticed: the two stories match.The zebra and the sailor both have a broken leg. ‘And the hyena bit offthe zebra’s leg just as the cook cut off the sailor’s’ (345). So the zebrais the sailor, the orangutan is Pi’s mother, and the hyena is the cook,which means that the tiger is Pi! This is a version of the shipwreckstory with four people in the lifeboat instead of one person and fourwild animals. We are reminded of several hints in the story itself thatthis may be the truth. In Pi’s first version, he sees wild animals runningabout the sinking ship. It is quite unlikely that the zoo animals wouldhave been freed from their cages. It is quite likely that hystericalpeople would have galloped about the deck in terror, seeming quite

21See p. 70, as quoted above at the end of section 2.22Yann Martel,Self (Toronto: Alfred A. Knopf Canada, 1996), 282-312. Martel often pursues

cosmic themes. Ways of dying is the subject of at least two of the four stories in his first book,The Facts behind the Helsinki Roccamatios(Toronto: Alfred A. Knopf Canada, 1993), includingthe title story. The other two deal, likeLife of Pi, with the relation between art and reality.

7.7. WHICH IS THE BETTER STORY? 185

beside themselves. We are being led right from the start into Pi’shallucination, his sense of the incredible story that he is going toexperience, and his disconnecting from his everyday sense of reality inorder to cope with the extremes into which he is about to be plunged.Again, Richard Parker disappears for three days at the beginning of thestory and reappears when Pi has been without food, water, or sleep forall that time. Perhaps it is no wonder that he hallucinates a Bengal tiger.My first reaction, at this point, was to think that the gentlesixteen-year-old Pi is traumatized by these events to such an extent thathe must cast the story in terms of animals. This would achieve twopurposes: it would allow him to think of the terrible cruelty of the cookas the natural behaviour of a hyena; it would help him to acknowledgethe murder of his mother while treating it as more like the killing of afavourite zoo animal. So there are psychological reasons for his havingtold his tale as one involving animals. That, however, would not makethe story true.

7.7 Which Is the Better Story?

Now we are given a blunt choice. Pi asks: ‘So tell me, since it makesno factual difference to you and you can’t prove the question eitherway, which story do you prefer? Which is the better story, the storywith animals or the story without animals?’ Mr Chiba and MrOkamoto agree that the story with animals is the better story. ‘Thankyou,’ Pi replies, ‘And so it goes with God’ (352).That is the punch line of the novel, or at least of Pi’s and Martel’sargument. We are to accept the story that weprefer, rather than the onethat is parsimonious.23 It is not dry, yeastless factuality that we shouldbelieve, but the better story, and if you are thinking of the ordeal in thelifeboat, it is the story with the tiger. If you are thinking of the meaningof life, it is the story with God. That is why this is a story that will

23In his own account of how the story came to him, Martel writes: ‘reality is a story and wecan choose our story and so why not pick “the better story” (the novel’s key words)?’ It will beclear that I do not entirely agree with the first two clauses of this metaphysics. See Yann Martel,‘How I Wrote Life of Pi,’ http://www.powells.com/fromtheauthor/martel.html(accessed 4 May 2004).


make the reader believe in God; the stories with God, the stories of theHindus and Christians and Muslims of the earlier discussion, makebetter stories than the alternative theories. We remember the agnosticwho ‘to the very end, lack[ed] imagination and miss[ed] the betterstory’ (70).We need to ask ourselves what it is that makes a story better? When wediscussed Ockham we assumed that it would be something likesimplicity. If the tigers and gods areunnecessary, then surely weshould not accept them as true just because we wouldprefer to have itthat way! Wittgenstein has a remarkable comment, towards the end oftheTractatus Logico-Philosophicus, that seems to make room forexplanations that are less than the simplest:

The process of induction is the process of assuming thesimplestlaw that can be made to harmonize with ourexperience.This process, however, has no logical foundation but only apsychological one.It is clear that there are no grounds for believing that thesimplest course of events will really happen.24

That is a purely logical remark. What is best for logic is not necessarilytrue of the world. Wittgenstein had an elegant metaphysics in mind.The world is made up of facts. Each fact is independent of all theothers, so it can be true regardless of the truth or falsity of the otherfacts. The way the world actually is, then, is a contingent matter.Nothing is necessarily the case. The natural laws, or the patterns wesee in events, do not determine how things are, for the facts couldalways have been otherwise. This is a clever account of reality and itslaws. Surprisingly, it leaves us not with Ockham (for there are manyinteresting similarities between those two philosophers), but withMartel. Ironically, it leaves us exactly in Pi’s position. If the twostories are equally complete—they both explain the murders and thesurvival ordeal, and neither explains why the ship sank in the first

24Wittgenstein,Tractatus Logico-Philosophicus, tr. C.K. Ogden (London: Routledge andKegan Paul, 1922), 6.33-6.331.

7.8. IS THE TIGER REAL? 187

place—what makes us think that the simpler story is the better one? Infact, there are many reasons for thinking that the story with the tiger isthe better one. It is more inventive in its appeal to the imagination, it ismore instructive about the various characters of its participants, it ismore astounding in its demands on our credibility, but having won usover it is more gripping. I think that we are actually more inclined tobelieve the story with the tiger than we are to think that the bare-bonesstory is the ‘real’ one. Those are powerful qualities. It is important forthe argument that we have not given up explanatory power; each storyexplains as much as the other. Where we gain, however, is that onestory is more gripping, more imaginative, more instructive, moreinventive, more demanding, more insightful, and more satisfying. Theother story has fewer entities.

Martel ends his novel with the official report of the Transport Ministryinvestigators. It concludes: ‘Very few castaways can claim to havesurvived so long at sea as Mr. Patel, and none in the company of anadult Bengal tiger’ (354). This sentence is presumably deliberatelyambiguous. Literally it says that no castaway has survived in thecompany of a tiger (and thus, not Pi either). As English is colloquiallyunderstood, however, it means that noothercastaway has survived inthe company of a tiger, as Pihas. In its context (the report containsphrases like ‘an astounding story of courage and endurance’;‘unparalleled in the history of shipwrecks’), the sentence is bound tobe taken in the latter sense. If the Ministry of Transport officials acceptPi’s story of Richard Parker, how could we readers not recognize it asthe one worth accepting? Similarly, the story with God. We believe inthe tiger, and we should believe in God.

7.8 Is the Tiger REAL?

Must we believe in God? Must we believe in Richard Parker? I thinkthat the answer is no. At least I do not accept Martel’s argument thatwe should. But does that mean that I think we have to make do with the


yeastless story?25 Not at all. I think that we can appreciate the betterstory and have our tiger, but without being eaten by him, so to speak.Pi did not make up the Richard Parker story just to amuse or confoundthe Ministry officials. I do not doubt that this is how he experienced hisadventure. Thus there is an important sense in which I believe hisstory. I have mentioned some of the psychological pressures that wouldhave made it easier for him to see his trials and tribulations through thefilter of the zoo animals. I now want to propose that Richard Parker isPi’s doppelganger. He is Pi’s double. He is a being whom Pi thinks ofas separate, but who is a projection of himself. The doppelganger isoften a ghostly second person who goes about with you, and advisesyou, but is invisible to others. This is an old literary device, that of thecharacter who sees himself or herself as someone else. A recentexample of it is the alter ego of the Russell Crowe character (JohnNash) in Ron Howard’s filmA Beautiful Mind. A classical example isEdgar Allen Poe’s story ‘William Williams,’ which remains ambiguousuntil the end about how many Williams there are. The doppelgangeralso marks a venerable psychological state: sometimes a person seemsto have multiple personalities, sometimes it takes the form of anout-of-body experience; in any case, one sees oneself as an other andnot as oneself. In every case there is really only one person eventhough there may appear to be two.This is not the first time that Martel has explored this idea. I havementioned the rape scene inSelf, where he uses the technique ofwriting two columns per page. In one column are expressed the fearand pain and horror of the suffering woman; in the other column arethe lucid reflections of a disengaged intelligence trying to figure outhow to escape with her life. Martel expresses a very differently dividedconsciousness using the same technique in the last of the storiescollected inThe Facts behind the Helsinki Roccamatios, ‘The VitaÆterna Mirror Company.’ In this case a young man listens to hisgrandmother’s recollections (in one column), drifting off at times (in

25Andrew Reynolds (of the University College of Cape Breton) pointed out to me that al-though ‘yeastless’ is a wonderful metaphor, it can be deceiving if taken to mean that bread ismagically raised by yeast. There is a good biochemical account of what the yeast accomplishes;the dichotomy between a ‘yeasty’ explanation and a dry ‘scientific’ one is a false dichotomy.


the other column) to his own reflections. These experiments withportraying a divided consciousness can be seen as preparation fortelling the story of a deeply divided Pi.

The doppelganger often can do things that serve as guidance to the realperson. Think of what Richard Parker does that Pi thinks he cannot do:swim to the lifeboat, eat raw turtle, kill the cook, do whatever it takesto survive regardless of the point of it. Pi, the gentle vegetarian whobelieves that all gods are one, is but part of the whole person. He couldnot have continued long on his own. We are all animals as well asspiritual beings, and Pi implicitly realizes that Richard Parker is anaspect of him which he cannot do without. His animal nature is ofcourse an aspect that must be tamed and kept in its place, as RichardParker quite literally is. But it is also an aspect of himself withoutwhich he would not be complete and would certainly not havesurvived.

On this reading, no special explanation is needed for the disappearanceof the double. A projection that has outlived its usefulnesswilldisappear into the jungle. On the other hand, a motive is required forits appearance. Martel obliges. Clearly it was the terror that reignedduring the sinking of the ship. It was Pi’s fear that can be credited withhaving created the tiger. It was a tiger, after all, that had first scared‘the living vegetarian daylights’ out of him. Now when he reallyconfronts the fear of death it takes the tiger’s form. When things havesettled into a routine, and the tiger has been suitably subdued, we get aknowing chapter on fear (chapter 56). It is followed by a line that Ihave already quoted: ‘It was Richard Parker who calmed me down,’says Pi. ‘It is the irony of this story that the one who scared me witlessto start with was the very same who brought me peace, purpose, I daresay even wholeness’ (179). That wholeness, I am proposing, is thereunification of Parker and Pi.

I have told a story about a story about a story—my story about Martel’sstory about Pi’s story. It is not the only story that could be told about it.A reader ofLife of Pican surely appreciate it as a simple adventuretale, as a story of survival in a lifeboat with a tiger. Or as a coming ofage story: a story of growing up, leaving childhood, going through the


tribulations of leaving one’s family behind and struggling with teenageabandonment, before arriving safely at adulthood. Or as a religiousallegory: a story of being forced out of paradise, put through theagonies of earthly life, and received at the end in a heavenly afterlife.Or as diasporic biography: a story of leaving India and suffering hugedislocation and disorientation, before finding a new home as animmigrant to Canada. It sustains all of those readings quite effortlessly.The reader need not see it as a doppelganger story, with the tiger as Pi’salter ego, in order to have appreciated the book.

My story, however, has advantages. My account of what is going on inLife of Pidoes not have real animals in the lifeboat, and it requiresfewer miracles, fewer amendments to the laws of nature, than does theone that Martel recommends.26 My story requires some extrapsychological complexity, but overall it is simpler, it is moreappropriate to our other convictions, and it explains at least as many ofthe novel’s relevant details. It does not reduce the story to ‘yeastlessfactuality,’ but it does not accept Martel’s claim that we should believewhatever flight of imagination commands our fancy. It should followthat I also do not accept Martel’s claim that Pi’s story should make usbelieve in God. I shall leave the drawing of that conclusion, however,to my reader.

Finally, I wish to emphasize a different but very important lesson: thatthe real interest and life in such an investigation begins with the work ofliterature. Margaret Atwood says that this is ‘a terrific book. It’s fresh,original, smart, devious, and crammed with absorbing lore.’27 She isreminding us that it is the novel that is of first importance. I think thatMartel is right that you miss the better story if you are not gripped byPi’s predicament, and so share his terror and his courage, and believe inRichard Parker. But the considerations that have led me to read it as I

26One should not discuss a story of a boy and a tiger without mentioningCalvin and Hobbes.Bill Watterson’s cartoon Hobbes is a stuffed animal tiger who comes to life only when he andCalvin are alone together. Hobbes is much more than a figment of the child’s imagination; indeedhe takes the initiative and often raises the philosophical tone of their discussions. Nonetheless, heis not a real tiger.

27Margaret Atwood in a Sunday Times review, quoted on the front cover of the paperbackedition of Life of Pi.


have, and to argue that it makes a false claim about metaphysics, have alife of their own. They are the stuff of philosophy.

Part Two:Theoretical Engagement

Chapter 8

David Braybrooke’s Philosophy ofSocial Science

MEREDITH RALSTON

Abstract

This chapter will examine David Braybrooke’s contribution to the phi-losophy of social science, specifically, and to feminist epistemology, in-directly, through an assessment of his 1987 bookPhilosophy of SocialScienceand his 1998 updates inMoral Objectives, Rules and the Formsof Social Change(chapter 13) andThe Routledge Encyclopedia of Phi-losophy. His detailed examples from naturalist, interpretive, and criticalsocial science illustrate how the three ‘sides’ of social science contain andcomplete the others. The chapter will show how his articulation of the‘mutual support and stimulation’ within the philosophy of social sciencehas paralleled, and in some ways anticipated, feminist critiques of episte-mology and the development of feminist methodologies themselves. Todo this, I will look at the traditional field of the philosophy of social sci-ence, examine what Braybrooke has added to the debate, and look at howhis ideas resonate with the latest ideas in feminist epistemology and phi-losophy of social science. I will argue that his belief in a unity of methodbetween naturalists and interpretive social scientists melds with a natural-ized feminist epistemology of the kind espoused by Richmond Campbelland Lynn Nelson, and may help overcome what seems to be the inherentrelativism of Sandra Harding’s feminist standpoint theory.

195

196 CHAPTER 8. MEREDITH RALSTON

8.1 Introduction

In the famous phrase of the now-cancelled television showThe X Files,‘The Truth Is Out There.’ In TV land, the phrase is meant to show thatconspiracies are all around us and are being covered up by sinister andunknown forces. With enough investigation, the truth will be revealedand all will be known. This idea of a universal, knowable truth is a pop-ular idea, but how real is it? What does this idea mean for understandingour reality and the world around us? Is this a fair representation of howwe know and what we know? Is knowledge just waiting to be foundlike another planet, perhaps, beyond Pluto? Has it always been there,or do we have to go and create knowledge? These are the issues andquestions dealt with by epistemology—the study or theory of the natureand foundations of knowledge—which asks how do we know, what canwe know, and, more recently, who can know. These are also the centralconcerns of the philosophy of social science.

This chapter will examine Braybrooke’s contribution to the philos-ophy of social science, specifically, and to feminist epistemology, in-directly, through an assessment of his 1987 bookPhilosophy of SocialScienceand his 1998 updates inMoral Objectives, Rules and the Formsof Social Change(chapter 13) and TheRoutledge Encyclopedia of Phi-losophy. His detailed examples from naturalist, interpretive, and criticalsocial science illustrate how the three ‘sides’ of social science containand complete the others. The chapter will show how his articulationof the ‘mutual support and stimulation’ within the philosophy of socialscience has paralleled, and in some ways anticipated, feminist critiquesof epistemology and the development of feminist methodologies them-selves. To do this, I will look at the traditional field of the philosophy ofsocial science, examine what Braybrooke has added to the debate, andlook at how his ideas resonate with the latest ideas in feminist episte-mology and philosophy of social science. I will argue that his belief ina unity of method between naturalists and interpretive social scientistsmelds with a naturalized feminist epistemology of the kind espoused byRichmond Campbell and Lynn Nelson, and may help overcome whatseems to be the inherent relativism of Sandra Harding’s feminist stand-

8.2. TRADITIONAL PHILOSOPHY OF SOCIAL SCIENCE 197

point theory.

8.2 Traditional Philosophy of Social Science

How do we know that the tree is ten metres high? How do we knowthat Mary is sad? Is knowing that the tree is ten metres tall the samesort of thing as knowing why Mary is sad? How do we know whethersomething is ‘true’ or ‘false’? How do we know that our best friendwasn’t abducted by aliens when she reports that she was? How do weknow any sort of crazy thing we see in the tabloids? Do we know any-thing? Our answers to these questions generally depend on our view ofknowledge itself and the relevance of the knower. Is knowledge univer-sal, rational, and objective or is it socially and historically specific? Isknowledge independent of the knower (because the particular knoweris irrelevant) or is it subjective and dependent on the knower? Ratio-nalists and empiricists have answered these questions differently andopposed each other particularly in the area of how we know: rational-ists relying on reason and a priori knowledge and empiricists relying toa greater extent on the senses, observation, and the scientific method.Today, most philosophers and social scientists are empiricists of somesort, though their emphases on what is knowledge and who can knowdiffer according to what philosophy of social science they follow. Arethey objectivists and believe that the idea of knowledge they pursue (atleast in the ideal) should be absolute and foundational, or are they rela-tivists and believe that knowledge is contextual, culture-bound, and thatthere is no ultimate measure of truth?

Traditionally, the philosophy of social science has been divided intothree main schools: naturalism, interpretive, and critical theory. Nat-uralists, who take their name from the natural sciences and the studyof the natural world, believe that we can know the social world in ba-sically the same way as we do in the natural sciences. We can findpatterns in social life; the cause and effect of events and behaviour canbe determined. There will be laws of human behaviour to discover, justas there are laws of gravity and motion, and causal regularities to bediscerned. There is a reality to be known independent of the knower,


and this reality can be verified through observation and measurement(Nielsen (1990)). The methods of the physical and scientific world,therefore, can be applied to the social world, and to naturalists, the sci-entific method ensures that results are tested, replicable, verified, or fal-sified. This doesn’t mean that scientists believe that they have neces-sarily discovered the truth but that they are getting ever more truthfulaccounts of the world and explanations for natural phenomena (though,according to Kuhn, not necessarily through a linear process of accumu-lation). Everything has a cause, but then that thing has a cause and so onand so on. The scientific method is a spiral process according to manyscientists (Kournay (1987)). An experiment starts with observations ofa certain phenomenon and then the scientist searches for regularities,trying to discover regular patterns and explanations for why these regu-larities occur. These explanations help to develop the theory or hypoth-esis, which is tested and used to make further predictions. The processthen starts over with more observations, further predictions, which areverified or falsified and then adjusted (Hempel 1966). A simple chem-istry experiment that both tests a law and makes predictions would beas follows: according to the law of conservation of mass, substancescannot lose or gain mass during a chemical change; that is, the mass ofa substance before reaction will equal the mass of the substances afterthe experiment. If mercury oxide (x), for example, is heated and trans-formed into liquid mercury (y) and oxygen gas (z), one can predict howmuch oxygen is in the compound by knowing the weight of the mer-cury oxide before decomposition and the weight of the liquid mercuryafterwards (x - y = z). The law itself can also be tested by weighing thetwo remaining substances and confirming that the mass has indeed beenconserved, within experimental error (y + z = x).

Many naturalists believe that a logical argument in the form of testablestatements of facts and laws creates explanations, predictions, and sta-tistical regularities. For Carl Hempel, a good scientific explanation re-quires asking the question: according to what general laws and by virtueof what antecedent conditions does the phenomenon occur (Hempel(1966))? A deductive argument gives the strongest possible groundsfor believing the conclusion is true and gives strong grounds for be-


lieving the conclusion if the argument is inductive. For example, ifall metal conducts electricity (the law), and this is a copper, thereforemetal, wire (the facts or antecedent conditions), then this wire conductselectricity (thing to be explained). Or (more problematically) in the so-cial sciences, if 80 per cent of street prostitutes were sexually abusedas children, and this woman is a prostitute, then there is an 80 per centchance that this woman was sexually abused as a child. Though mostnaturalists recognize that the social world is a good deal more complexand unpredictable than the natural world, they believe we can still findcausal regularities and statistical generalizations within the social world.

Interpretive theorists claim that naturalists fail to recognize humanintentionality, motives, social context, and people’s ability to change.They believe that the social world and hence social sciences are not likethe natural sciences. To explain some event or behaviour in the socialworld requires understanding the meaning behind the actions, and thatdepends on context and particularly on social relationships. In oft-givenexamples, what does it mean to put one’s hand in the air? Is the per-son voting, waving, signalling, or saluting? Understanding which it isand why requires knowing the context of the vote, wave, or salute. Nocausal regularity can cover all the possible interpretations and possiblemeanings; that is, if a person raises his hand it means x, he has raisedhis hand, so therefore it means x. In the above example outlining thepercentages of street prostitutes who have been sexually abused, whatreally can be said about the particular prostitute? Though it might bepart of the explanation for why this particular woman ended up as astreet prostitute, what about all the women who have been abused butwho don’t end up on the streets? And what if she wasn’t sexually abusedor is in denial about the sexual abuse? The fact that 80 per cent of streetprostitutes are sexually abused says little about the individual prostitute,according to the interpretive point of view. How did she become a pros-titute? Why does she think she became a prostitute? Did she choosethis life? What is life like for her? These questions, argue interpretivetheorists, are not dealt with by looking for a logical argument in thenaturalist tradition. Many more examples can be found to illustrate theneed for understanding contexts. Are you chopping wood for a fire, for


exercise, for building a house, or for working out frustrations? Why isa piece of paper a bookmark, and not scrap? Why is this death a murderand that one a suicide? Is the cause of death a brain hemorrhage or is itdriver negligence, slippery roads, or drunk driving?

Interpretive theorists object to the scientific method on at least fivegrounds: controlled experimentation, they believe, is impossible whendealing with humans and makes prediction and repeatability difficult;social phenomena are culturally and historically determined, and so it isvery difficult to predict changes in communities or individuals; humanbeings modify their behaviour as a result of acquiring additional knowl-edge; humans have goals and motivations that can’t be accounted for;and the inherent subjectivity of human behaviour cannot be predicted(Nielsen (1990)). Prediction in the social world is a particular problemfor naturalist social scientists, according to interpretive theory, becauseof human will and decision-making. We may be able to predict withsome accuracy that if we mix two parts of hydrogen with one part oxy-gen we’ll get water (H2O), or that one part carbon and two parts oxygenequals carbon dioxide (CO2), but can we predict with any reliability thatcombining poverty and poor education in one individual will result incriminal behaviour? And does prediction explain something anyway?A barometer may predict rain, but it’s the change in atmospheric con-ditions that explains why it is raining. Similarly, spots on a child maypredict measles, but it’s the virus that explains why the child is sick.Finally, interpretive theorists argue that the value-neutrality of the sci-ences is impossible in the social sciences because of the value-orientedbias of social inquiry. This will be developed more fully below.

Naturalists counter that controlled experiments are not crucial to theproject of the sciences or social sciences. Astronomy and astrophysics,for example, don’t use experimentation but rely on observation, math-ematics, and computer simulations. The common denominator is thescientific method, which creates a logical set of procedures for observ-ing change, or predicting behaviour by keeping some factors constantand changing others. For example, though not experimenting per se,astronomers rely on the scientific method’s goals of prediction, testa-bility, falsifiability, and verifying data through simulation. A computer


program, for instance, is used to show how two galaxies might look af-ter colliding with each other, based on the known elements of the twogalaxies. Astronomers then use observational tools to look for similarpatterns in the cosmos to see what, if anything, most closely resemblesthe modelled simulation.

Naturalists agree that prediction is difficult in the social sciencesbut believe that common patterns can be found and human subjectivity,goals, and motivations can be used as social variables and part of theantecedent conditions of the explanation. The goal of a rigorous sci-entific method should be adhered to because, they argue, it is still thebest way that we know of at this point in history to understand naturalphenomena. As one scientist friend reminds me, no matter who doesthe testing, if you drop something it will fall and he’ll be able to predicthow long it will take to hit the ground, just as one of Newton’s lawspredicts. Or put another way, naturalists believe there are facts aboutthe world, regardless of whether we have discovered them yet (that is,regardless of who is the knower). Pluto (as either a planet or asteroid)would continue to exist whether we had found it or not. Most scientistshave faith that though the scientific method isn’t perfect it eventuallyweeds out theories and ideas that are found to be false—for instance,that the sun revolves around the earth, that the earth is flat, that the uni-verse was created only 5000 years ago. The scientific method, then, is auseful tool to undermine ideas that can’t be tested, like creationism, orfantastical ideas such as the popular science fiction theme in movies likeThe Matrix that the world has only actually existed for a few minutesand that all our memories, lives, and ideas are continuously re-created.Though it is still the goal of some scientists to elaborate a grand the-ory that will unify, for example, the fields of electricity, magnetism, andgravity, most scientists recognize that different views are debated untila consensus in the community is reached and that they come into theexperiment or simulation with preconceptions and assumptions aboutcertain laws and theories (Kuhn (1962)). Further, we should not for-get that in terms of justification of scientific knowledge, philosophers ofscience like Popper and others argued that since theories can never bepositively confirmed, scientists can only falsify theories one at a time.


Theories that are not eliminated then may be true or may be very prob-ably true. This indicates a far more nuanced view of truth than interpre-tive and critical theorists usually give scientists credit for. Perhaps themost one can say, even for naturalists, is that truth, or ‘properly justifiedtrue belief’ in philosophical terms, is to be seen as an attainment goal orconcept, not as truth itself, and always relative to the particular scientificcontext (Kuhn (1962)).

The third traditional school in the philosophy of the social sciencesis critical theory. Developed from the Marxist critique of ideology, crit-ical theorists see themselves as doing emancipatory work, uncoveringwhat had been hidden and finding the realities masked behind dominantideologies (Nielsen (1990)). Knowledge is socially constructed and thecontext of knowledge is all important to understanding, explaining, anduncovering knowledge. Critical theorists reject the idea of objectiveknowledge and the disinterested, neutral knower of naturalism, and theemphasis on meaning and understanding in interpretive theory. Thepoint, as Marx famously put it, is not merely to understand the worldbut to change it. Critical theorists want to unmask ideologies that goagainst many people’s best interests but are seen to be the only possibleideas or systems, for example, the idea that capitalism is the only effi-cient mode of production. Or that human nature is naturally aggressiveand competitive.

There is no value-free knowledge, according to critical theorists, be-cause knowledge is socially constructed and situated. Undermining nat-uralism’s fact-value distinction is crucial for critical theorists becauseseparating the two out makes it appear that there is such a thing as‘facts’ on the one hand (objective and true) and mere values or opin-ions on the other (subjective and perhaps false). This separation, ac-cording to critical theorists, hides the ideology and biases that occureven in so-called facts and within the underlying values of the researchitself. For instance, if a quantitative research study concluded that be-cause of the massive abuse of the welfare system by welfare recipientspeople on welfare should be cut off from their support, critical theoristsmight question what were the essential values and world-view of theresearchers in the first place. Do they believe that people on welfare are

8.3. BRAYBROOKE’S CONTRIBUTION 203

lazy or, alternatively, do they see them as victims? Is the study simplyconfirming what the researchers think about poor people and/or, morebroadly, what a capitalist mode of production needs in order to functionefficiently—that is, have a large reserve army of labour?

Critical theorists also began the work of standpoint—that is, that dis-advantaged groups have the potential for a more holistic view of realitybecause of their oppressed position in society. As a tool for survival, ac-cording to critical theorists, the proletariat, the black slave, and womenhave a better or more complete view of the world than their bourgeois,white, or male counterpart. The connections between critical theory,anti-relativism, and feminism will be developed below.

8.3 Braybrooke’s Contribution to the Philosophy of SocialScience

Outlined so starkly, there seems to be little overlap and agreement be-tween the three schools within the philosophy of social science, and thisis Braybrooke’s starting point. He also indicates the helpfulness of aninfluential article by Brian Fay and Donald Moon, called ‘What Wouldan Adequate Philosophy of Social Science Look Like?’ (Fay and Moon(1977)). In this article, the authors ask three questions aimed at estab-lishing that the division between naturalism and interpretive theory isunnecessary: What is the relationship between interpretation and expla-nation? What is the nature of social scientific theory? And what is therole of critique? These questions provide the framework for a sophisti-cated argument outlining how each side is lacking: the naturalists don’tsee the need for interpretation of meaning, and the interpretive theoristsdon’t see the need for explanation and theories. An adequate accountof the social sciences needs both, according to Fay and Moon, and alsoneeds the questions asked only by critical theory.

Braybrooke agrees. In his 1987 book, he begins by outlining thedistinguishing characteristics of the three sides of the social sciences.Naturalism, as stated above, relies on the methods from the natural sci-ences and looks for causal generalizations and regularities. Interpretivetheory attempts to bring to light what the actions that people do signify,


looking for ‘settled social rules.’ Critical theory ‘refuses to take at facevalue the rules cited by the interpretive view’ or the causal regularities ofnaturalism (Braybrooke (1987b), 4). Regardless of the real differencesbetween the three sides, his aim is to show ‘how robust are all three andhow complexly interconnected’ (4). None of them has exclusive truthand the ‘differences coexist with a great number of parallels’ (5). Crit-ical theory, he argues, asks different questions but its methods reduceto those of the other two. Therefore, there is a unity within the socialsciences. In his 1998 update, he goes further, suggesting that a ‘settledsocial rule presupposes the existence of corresponding causal regular-ities in social phenomena and vice versa’ (Braybrooke (1998a), 251).Moreover, he argues ‘regularities found in social phenomena would nothave the foundation in personal choices and actions that they must haveto be fully explained in our eyes as producers of social science or asconsumers of it if they did not have a foundation in choices and actionsguided by rules’ (257). In his 1998 essay in theRoutledge Encyclope-dia of Philosophy, he also takes on the contribution of postmodernismto the debate about the philosophy of social sciences and claims thatthough aspects of postmodernism are helpful in understanding the lim-its of knowledge, some of the claims are so radically sceptical as to un-dermine the possibilities of knowledge and social science itself, and thishe does not want to do (Braybrooke (1998b), 846). Therefore, thoughpostmodernism can be seen as a separate (fourth) option to naturalism,interpretive, and critical theory, Braybrooke takes postmodernism as avariation of critical theory, in terms of his claims about unity (Bray-brooke (1998b)).

These are strong claims, and adherents of all three schools (and par-ticularly of postmodernism) will object to the explicit idea of a unity.How does he make these claims? First, he shows how naturalism tendsto concentrate on different questions, facts, and concerns than interpre-tive theory, specifically in looking at ‘group facts,’ that is, how groupsoperate as opposed to how individuals operate within those groups.As mentioned above, naturalism is definitely in the minority amongphilosophers and social scientists today: as Braybrooke says, it’s ‘quiteout of fashion’ (Braybrooke (1987b), 7). He asks the questions, Are all


group actions, in fact, reducible to individual actions within the group?And are these group facts inherently less interesting than the interpretiveemphasis on the individual? Surely not, he says, on both issues.

Braybrooke outlines an example of naturalistic social science thatrelies on group facts and causal regularities: Leon Epstein’s article onthe differences between Canadian and American political parties andtheir cohesion or lack thereof, respectively (Epstein (1964)). In the ar-ticle Epstein asks the question, Why do Canadian political parties, un-like American parties put in disciplined and cohesive performances innational legislatures? His answer—by looking at the ways in whichCanada and the United States are similar and different—is that it is theBritish parliamentary system (the major difference) which accounts forthe differences in the behaviour of their political parties. As Braybrookesays, using Mill’s method of agreement and/or difference (Braybrooke(1987b), 8), Epstein would be correct in arguing that though Canadaand the United States share several features in common (including alarge land mass, a federal system, and federal and provincial parties),it is the main difference between them that explains the difference inpolitical parties, that is, Canada’s use of the British parliamentary sys-tem and the United States’s use of the presidential system. Conversely,though Canada and the UK are quite different in some respects (withthe UK having a small land mass, a unitary system, and only nationalparties), Canada and the UK share the British parliamentary system andtherefore both have cohesive political parties.

Braybrooke shows that Epstein’s reasoning is naturalistic becausethe solution is a causal statement (whenever a country has the Britishparliamentary system, it has cohesive parties; Canada has the parlia-mentary system, therefore, Canada has cohesive political system), and itimplies a counterfactual (if Canada didn’t have the British parliamentarysystem, we wouldn’t have cohesive parties) and a causal generalization(if you want cohesive parties you need the British parliamentary system)(9). We will come back to this example shortly. For an example of inter-pretive theory, Braybrooke uses Elliot Liebow’s research on Tally’s Cor-ner, a famous study of a group of unemployed African-American menin Washington, DC (Liebow (1965)). In the study, Liebow observed and


interviewed the men. He concluded that they start out their adult liveslooking for jobs, but because they lack higher education, they are onlyable to get menial jobs. The menial jobs don’t pay very much and themen don’t get much respect. Their future looks bleak so they give upand don’t work at all. This affects their relationships with the womenin their lives and particularly the mothers of their children, whom theydon’t treat very well. Liebow then asked the question, Why do thesemen act as they do? His answer: because they can’t be good familyproviders, they set up alternative rules, and act in ways that support amacho image of themselves, to deal with their failures as husbands andfathers. It’s an interpretive study, according to Braybrooke, because thestudy is qualitative—Liebow deals with the meaning behind the actionsof the men—and because it’s about social rules and conforming or fail-ing to conform to rules.

For critical theory, Braybrooke doesn’t articulate a specific real-lifeexample but instead uses critical theory to critique the naturalistic andinterpretive examples. From a critical theorist’s point of view, Epstein’sexample would be banal or uninteresting at best, an ideological obfus-cation at worst. Epstein takes the current system at face value, doesn’tlook at whose interests the political system serves. And the study couldrationalize the current system because it fails to show that any opposi-tion is a sham anyway, because no parties will threaten the interests ofthe property-owning groups. Braybrooke says Liebow would fare bet-ter, but, even so, critical theorists would most likely have issues withLiebow’s assumptions about capitalism and the welfare state—that is,that the market is imperfect but generally beneficial—and his lack ofrecognition that alternative arrangements are available but hidden fromthe masses by a dominant ideology. None of this deeper critique is doneby Liebow, so his example would be considered limited as well, by crit-ical theorists (Braybrooke (1987b), 17-18).

Though critical theorists ask different questions and see themselvesas delving deeper into critique than naturalist or interpretive theorists,Braybrooke argues that they use the same methods as interpretive the-ory and naturalism. The interpretive side of critical theory lies in theemphasis on comparing ‘what social scientists aim at doing, under cer-


tain rules that express their recognized aims and their activities, withwhat they are doing unawares or unawares omitting to do’ (70). Crit-ical theorists tend to rely on qualitative methods and try to understandthe meaning behind or underneath something, like interpretive theorists.Support for pluralism within political science, for example, accordingto a critical perspective, leads either to ‘ignoring unorganised groupsor to postulating that they have special means of representation’ (70).The naturalistic side of critical theory comes out in the Marxist critiqueof ideology in that it ‘upholds a specific theoretical explanation of theomissions and distortions which it finds in subcritical social science’(76). That is, a causal connection is used to explain how change occurshistorically because of changes within a particular class structure. Themajor problem for admitting the naturalist bent of critical theory willoccur with the fact-value distinction and critical theory’s opposition tothis distinction. Braybrooke argues that this should be a problem forcritical theory only because it is a ‘fallacy,’ according to him, to arguethat because a study can be found to be ideological, that ‘every state-ment in that body ... has an ideological cast and import’ (81). Put evenmore starkly, he argues that ‘every single statement in a set of state-ments about social phenomena can be true, yet the whole set can haveideological, distortive consequences for people’s thinking’ (81). Thedifference between the two (the possibility of statements of fact vs. theideological impact of those facts and of the study as a whole) must besorted out and supported by evidence. This will be important later whenwe look at Sandra Harding’s work.

Braybrooke also argues, convincingly, that Marx himself used bothnaturalist and interpretive methods. Marx’s social science was, accord-ing to Braybrooke,

naturalistic in explaining the ideas of subcritical social sci-entists as determined ultimately by the roles given them insocial arrangements that are causal consequences of the cur-rent system of production. It is interpretative in the detailswith which this determination works out, including the rulesthat social scientists accept and conform to because of theirclass position. (83-4)


Braybrooke leaves critical theory at this point to concentrate on hismain argument that there is a unity within the social sciences, not justa unity of method between critical theorists and naturalist and interpre-tive theorists. He returns to his examples to show how inadequate theyare on their own and how they need elements of the others to be com-plete. While it may be true, Braybrooke argues, that Epstein’s exampleis a logical argument (that is, whenever a country has the British parlia-mentary system, it has cohesive parties [the law]; Canada has the par-liamentary system [the facts]; therefore, Canada has cohesive politicalsystem [the thing to be explained]), without more information into whatexplains the law itself (that is, what explains why the parliamentary sys-tem creates cohesive political parties), the example is inadequate. Theargument, though a logical explanation, still then fails to explain whyCanada has cohesive political parties. To do this, Braybrooke argues,we need individual person facts as well as the group facts, and rulesas well as regularities. For instance, individual members of Parliamentin the Canadian system vote as a bloc because they don’t want to bekicked out of the party, and they don’t want to run as independents. Butwhy don’t American members of Congress worry about the same thing?Because of the rule in parliamentary systems that says that if you are amember of the governing party, you must not vote against it. Again,why not in the presidential system? It is because of elements of theparliamentary system that need to be explained, that is, that in a parlia-mentary system there are no fixed election terms and a potential vote ofnon-confidence could bring down a government. If a government falls,an election must be called and MPs could lose their seats. Because ofthe separation of powers in the American system, and because there isno vote of non-confidence, even the impeachment of a president wouldhave little effect on the majority party in the Congress. Therefore, in-dividual members of Congress are far less likely to vote as a bloc. Allof this information is needed in order to explain why Canadian politicalparties, unlike American political parties, act in disciplined and cohe-sive ways in our national legislature. We need to know group facts andperson facts, social rules, and causal regularities. The regularity showsthat Canadian political parties are more cohesive than American; the


rule that tells individual MPs to vote as a bloc explains why. And viceversa. Canadian political parties are more cohesive than American onesbecause individual MPs vote as a bloc, and individual MPs vote as abloc because of elements of the Canadian political system that makepolitical parties more cohesive than American ones.

Braybrooke outlines how even loose regularities can be helpful, howlaws can be developed, even if transitory or specific to a community, andhow the model-theoretic view in naturalism could be used by interpre-tive social scientists. In his 1998 essay, he particularly concentrates onthe model-theoretic view and shows how this approach ‘relieves socialscience of the burden of finding regularities that hold for all societies’(Braybrooke (1998b), 840). As he states, even the natural sciences haverevised the view that laws hold for all time and now use a system thatsets up models that are true for that model only. The model is then usedto show how other models ‘resemble one range or another of phenom-ena closely enough to be more useful than other models in predictionand explanation’ (840).

In Liebow’s study, we see the development of social rules that areconformed to, denied, or changed. The men on Tally’s Corner start byfollowing a rule that says they cannot have respect unless they are regu-lar providers, and in order to retain self-respect (because they can’t fol-low this societal rule) they oppress the women in their lives. Braybrookethen shows how this example, though clearly interpretive, has naturalisttendencies. The conformity or lack of conformity to rules are solutionsto problems (how the men cope with their circumstances); the solutionsare the explanation and can be put into the form of a logical argument (ifyou are a married man and want respect, you must provide for your fam-ily (the law); this man is married but can’t provide (the facts); therefore,he gets no respect and treats his family badly (thing to be explained);and Liebow’s findings are empirical, meaning he observes the men, heis looking for repeated patterns of behaviour, and the study can be repli-cated with a similar group and verified or falsified—all characteristicof naturalism, not interpretive social science. Further, these facts couldhave been quantified (say, more than 50 or 75 per cent of the men inthis area act in similar ways), and therefore, statistical regularities can


be developed. Braybrooke uses examples of other quantitative researchstudies on unemployment and friendships to demonstrate how use ofstatistics could help to explain why the men on Tally’s Corner have nojobs (39) and why their friendships are ‘so loose and fitful’ (41). Thus,to complete the picture of the men on Tally’s Corner, group facts (soci-etal rules and regularities) are needed as well as interpretive person facts(how the particular individual gives meaning to his actions). Therefore,Braybrooke gives a very convincing argument that ‘there are both per-son facts and group facts on both sides’ (60) and that both theories’methods, questions, and ‘facts’ are needed for an adequate explanation.

But as stated in the beginning of this section, Braybrooke doesn’twant to just say that both naturalism and interpretive theory are neededin order to have an adequate explanation for a certain phenomenon. Inhis final chapter, then, he goes further by arguing that the ‘key idea ofeach side presupposes the key idea of the other’ (94). Rules, which hedefines as ‘prescriptions or prohibitions to which clear forms of punish-ment attach as deterrents to deviation’ (111), draw on regularities, andregularities (causal explanations like the rates of inflation, unemploy-ment, or divorce) draw on rules for their very definition. From every so-cial rule ‘an observable regularity may be deduced’ (116), and from reg-ularities of inflation, for instance, we can deduce rules for exchanginggoods or setting prices. A naturalist’s ideas about rates of employmentcannot exist ‘without accepting some interpretative account of what itmeans to have a job’ (112). In the same way, interpretive social scien-tists are ‘interested in whether husbands try to be steady providers. Itcannot find whether they do without counting how often they turn upfor a job when they have one, and how often they bring home their pay-checks’ (112). As an article in the local paper described recently, fatalcollisions on Nova Scotia highways have steadily decreased since 1984(the statistical regularity) because of the laws against drunk driving andthe increased use of seatbelts (the rules). If there hadn’t been a problemto be addressed (the rate of fatal collisions) there would be no reason forthe rules against drunk driving or in favour of seatbelts. If there wereno rules against drunk driving, there would probably be the same ratesof fatal collisions, or then another cause would be identified, and so on.


In the case of this article, the increased rate of fatal collisions this par-ticular year was an anomaly to be explained by the increased speedingof drivers, itself a rule that has been violated (Lipscombe (2004), A1).Rules and regularities, therefore, presuppose the other.

Braybrooke concludes his 1987 book with a caveat about generaliza-tions: ‘the fact that the generalization applies only to one agent does notprevent it from being a generalization’ (126), and he accepts as fact thatgeneralization is the ambition of social science (128). In his 1998 up-date, Braybrooke addresses the postmodern challenge inherent in anyclaim about generalization by undermining the claim of universalityeven in naturalism. As mentioned above, scientists no longer think ofthemselves as searching for the truth but argue that ‘scientific successhappens in local contexts and only for a time’ (Braybrooke (1998b),839). The postmodern concern with ‘grand narratives’ and ‘globalizingdiscourses’ (844) has led to recognition that discourse is power and thatclaims for knowledge ‘become conflicts of power or to “wills to truth”’(844). Postmodernists attack the idea of a unified view of the world,yet even Foucault’s theory comes ‘with claims of truth, or at least someapproximation of truth’ (844). According to Braybrooke, even these‘sweeping pronouncements’ can be accommodated by the three schools(845). Braybrooke argues that postmodernists

like Foucault and Lyotard bring back the study of rules ...with an emphasis more thorough-going than any other philoso-pher[s] ... But if there can be models of systems of rulesjust as there can be models of regularities then how far amodel fits a real system, in regard to any universally quan-tified statement that figures in the model, is a matter forquantitative—statistical—study just as much with rules aswith regularities. (845)

Methodologically, therefore, both interpretive and naturalist ques-tions are in play with postmodernism and now ‘in a form that is readyto receive the insistence of postmodern writers on truth being local, amatter of limited context that is always subject to supersession’ (845).Though Braybrooke acknowledges the impact of postmodernism on the


philosophy of social science and accepts certain of their limiting claimson knowledge, he would ‘resist any further-reaching scepticism’ (839).Because some of the postmodern critique is so sceptical of even the pos-sibility of knowledge, Braybrooke argues, we ‘may thus have to allowin the philosophy of social science for a school that (unlike any of thethree schools in the framework) holds that social science is not possible.We do not have to join it’ (846).

In the next section, I want to show how Braybrooke’s claims abouta unity of method, and his arguments about universality and generaliza-tions, anticipate feminist critiques and also how these arguments couldbe used to overcome some of the polarizing debates currently happeningwithin feminist philosophy of social science.

8.4 Feminist Epistemologies and Critiques of Social Science

The links to feminist epistemology in Braybrooke’s work may not seemobvious at first glance. Braybrooke does not engage directly with fem-inism in his 1987 book or in the 1998 updates and, therefore, does notthink of feminism as a ‘school’ (in its own right) within the philosophyof social science. Braybrooke’s example of naturalistic social sciencedoes not on the surface have anything to do with women, and his inter-pretive example deals only with the male perspective and doesn’t men-tion at all how the women feel about the men’s behaviour towards themand their children. Nevertheless, I would argue that there are links,overlaps, and resonances with feminist social science and epistemol-ogy in four areas. First, his argument that critical theory shares a unityof method with interpretive and naturalist theory is similar to the cur-rent debate within feminism about whether there is a distinctive femi-nist method or whether what makes feminism unique are the distinctivequestions and focus on women. In some ways, we could see the devel-opment of feminist epistemology as a parallel to critical theory. Bray-brooke’s comments about critical theory, though obviously different incontent, are therefore applicable at the level of the distinctive questionsand deeper critique. Second, Braybrooke defends quantitative methodsand acknowledges that interpretive and critical theorists tend to disdain

8.4. FEMINIST EPISTEMOLOGIES 213

quantitative methods and statistics. As Braybrooke writes, ‘they dislikeand distrust quantitative methods ... so they are inclined to scorn quanti-tative methods and denounce them as inappropriate to the study of socialphenomena’ (Braybrooke (1987b), 35). This rings true with some fem-inist researchers and again resonates with current debates. Third, hisviews on the problems of universality and generalization can be appliedto feminist debates about feminist empiricism and feminist standpointand, particularly, the extent to which research in the social sciences canbe generalized without essentializing some aspect of it. Finally, theproblem of the fact-value distinction calls for examination in the light offeminist concerns about objectivity and subjectivity, and Braybrooke’swork can be useful here, too.

Philosophers and social scientists who have been concerned withfeminism have tried several strategies for ensuring women’s voices areheard. One of the first methods in most of the social sciences and hu-manities was simply adding women in to rediscover women’s historyand contributions, to recognize that women’s experiences had been leftout, and to ask questions that had not been asked before or which hadnot been seen as important. Historians discovered women’s lost history,political scientists examined women’s political engagement, and soci-ologists looked for connections between family violence and women’seconomic dependence in the home. The focus shifted over time fromadding women in to putting women at the centre of analysis and lookingat what life was like for women. So, for example, doing a feminist anal-ysis of Liebow’s study would mean interviewing the women on Tally’sCorner, examining how they felt about their lives, about why their menacted as they did and why they accepted or rejected this behaviour. Likethe discovery of X-rays, what was invisible became visible with the newfocus on women’s experiences of abuse, family life, work, sexuality, andmore. As feminist critiques became more common, new methodologieswere created and old, ‘male’ methods were criticized. The scientificmethod was particularly critiqued for its supposed neutrality, objectiv-ity, and universality.

Feminist ways of knowing came to be identified with anti-naturalismand anti-science. Qualitative, participatory, and collaborate research


methods were identified as feminist, and science and scientific methodwere rejected as inherently male. Feminist philosophers like SandraHarding began to question whether epistemology’s concern for objec-tivity hid a subjective point of view and asked whether a politicizedinquiry, ironically, might be less distorted than a so-called point-of-viewlessness. She argued that the scientific method for all its heraldedobjectivity and universality was actually developed with a particularknower in mind, that is, a white, middle-class, highly educated male,and that its supposed neutrality could not withstand scrutiny. Manyother feminist philosophers criticized the maleness of philosophy andthe underlying male-as-norm ideology (Lloyd (1984); Westcott (1990)).The knower or researcher became crucial to the project of feminism be-cause it was at this level that feminism identified the distortion and bi-ases affecting the outcomes of research. As well, the separation betweenresearcher and the thing being researched became another problematicof the scientific method, in particular, as feminists saw the separation asobjectifying and distancing, not as ensuring objectivity as did traditionalresearchers.

But is it the method that is the problem or is it the underlying method-ology and epistemology—that is, are feminists actually doing some-thing different empirically from traditional sociologists or political sci-entists, or are they asking different questions, looking at different phe-nomena and coming to different conclusions? The main point of con-tention between feminist and traditional social scientists seems now tobe the question of objectivity, which we will look at in more detail be-low. Feminist researchers, generally, don’t accept the notion that theresearcher should be ‘on a different plane’ from the researched. Theyreject the idea that in order to keep the research as objective as possiblethe researcher can’t give opinions or engage in conversations with theresearched (supposedly to remove bias and prevent so-called subjectiv-ity) and therefore contaminate the research. But this position is becom-ing more common even in naturalist social science, as stated above inthe analysis of Braybrooke’s work. In a working paper for Status ofWomen Canada, on the state of feminist research, Damaris Rose paral-lels Braybrooke when she states that ‘developments in quantum physics


have established that even in natural science there is no such thing aspure observation uninfluenced by the observer’ (Rose (2001), 6) andthat claims for ‘validity based on researcher/research object separationshave less credence than previously’ (6).

Recent work in feminist epistemology and methodology suggeststhat there is not a separate feminist method but that feminists tend touse a multi-method approach, somewhat like what Braybrooke is sug-gesting for the social sciences as a whole. Multi-method means mix-ing individual interviews with focus groups and perhaps surveys andarchival work, that is, qualitative and quantitative research methods(Rose (2001); Reinharz (1992)). The feminism becomes the under-pinning of the study, of the questions and critiques that the researcherbrings to it. As Braybrooke stated, critical theory may ask differentquestions and delve deeper into the roots of a problem, but in methodit reduces to the methods of the other two. Feminism can bring aboutan awareness of the pervasiveness of gender and the links between gen-der and other forms of oppression, but in actually carrying out the re-search program, the methods reduce to some form of naturalism (quan-titative information and surveys) or interpretive methods (interviews,focus groups, and participant observation) or both.

What was said above regarding the question of different methods canalso be said for the question of quantitative methods. Though tradition-ally feminists have rejected quantitative methods, for similar reasons asdid interpretive and critical theorists, there has been an increased inter-est in mixing methods and looking for causal regularities in feministresearch. Many more social scientists are seeing the utility of knowingrates of divorce or family violence or sexual assault in order to makecases for services, illustrate to the public and funding agencies the se-riousness of feminist claims, and empirically prove that women havebeen systematically discriminated against in various ways. Rose arguesthat the dichotomy between qualitative and quantitative methods

is too simplistic and may be counter productive. While epis-temological differences are real, and do lead to differenttypes of questions being asked in relation to the same broadresearch topic, many of the standard dichotomies become


quite blurred when placed under closer investigation. Thisopens up the possibility of developing research strategiesthat recognize the potential complementarity of certain quan-titative and qualitative techniques. (Rose (2001), 4)

As I’ve stated above, complementarity was Braybrooke’s argumentnearly fifteen years earlier. Further, as feminist researchers’ resistanceto quantitative methods fades, Rose argues ‘it is possible to adopt quan-titative techniques (that is, methods in the narrow sense) without buyinginto all of the tenets of positivist epistemology associated with quanti-tative methodology in the large sense’ (12). Braybrooke and naturalistswould concur, I’m sure, but is this view tenable given feminist concernswith objectivity and universality?

Philosophers like Sandra Harding would probably still disagree aboutthe question of a unity of methods in the social sciences and particularlythe use of the scientific method. Harding is an advocate of feministstandpoint theory and has dismissed feminist empiricism as unable toovercome the androcentric bias inherent within the scientific method.I turn to her work now in greater detail and hope to show how Bray-brooke’s arguments can withstand her critique of traditional social sci-ence.

Harding rejects ‘objectivism’ because of the dispassionate, disinter-ested, scientific knower and his value-free, genderless, objective sci-entific procedures (Harding (1996), 302). She thinks science devaluesanything that doesn’t take an inherent point-of-viewlessness. She re-jects ‘interpretationism’ as too relativist because, in her analysis of in-terpretive theory, conflicting interpretations become equally defensibleand because they don’t recognize the existing power relations of maledominance (303). Neither school, she argues, works for feminists, noteven an explicitly feminist empiricism. Feminist empiricists or femi-nist scientists have argued that the androcentric bias in some scienceprojects is because it is ‘badly done science’ and that stricter adherenceto the scientific method will overcome these biases in the knower (305).It is not the method that is the problem; it is the researcher himselfwho has not been able to filter out his androcentric biases or stereo-types in the interpretive stage of the research. Harding distinguishes


between the ‘spontaneous’ feminist empiricists of the 1970s with themore sophisticated philosophical feminist empiricism discussed below(Harding (1993), 51). Harding (and many others who advocate fem-inist standpoint and the critical theorists before them) argues that thescientific method cannot overcome the problems of its so-called objec-tivity and universality. It is an illusion to think that science can removebias with a ‘transhistorical, unitary individual knower’ (Harding (1996),308), because science has not been ‘particularly welcoming to issues ofrace, class or cultural differences in women as subjects of knowledge—that is, between women as agents of knowledge’ (307).

Instead, Harding advocates feminist standpoint, taking as a startingpoint the experiences of women to develop ‘potentially more completeand less distorted knowledge claims than do men’s experiences’ (309).For example, looking at “‘men’s work” can provide only partial, dis-torted understandings’ of the nature of work. Because there is a ‘faultline’ between ‘women’s real experiences and the dominant conceptualschemes,’ phenomena like housework won’t fit traditional definitions ofwork or leisure and therefore should be analysed through concepts ofwomen’s experiences (310). Standpoint overcomes these problems bystarting from marginalized lives and taking ‘everyday life as problem-atic’ (Harding (1993), 50). The scientific method is too weak to detectsexism, so stronger standards for ‘good method’ are needed, ‘ones thatcan guide more competent efforts to maximize objectivity’ (52). Shetakes from intellectual history the ideas of Hegel (master-slave) andMarx (bourgeois-proletariat) to start from the voice of marginal groups(53-4). In this way, her views parallel the ideas of critical theory, asoutlined by Braybrooke. Since societies are stratified by race, class,and gender, the ‘activities of those at the top both organize and set lim-its on what persons who perform such activities can understand aboutthemselves and the world around them’ (54)—essentially, the Marxistcritique of ideology as stated by Braybrooke. Putting women at thecentre of analysis will lead to less distorted accounts of those women,according to Harding, because starting at the bottom of social hierar-chies ‘can provide starting points for thought—for everyone’s researchand scholarship from which humans’ relations with each other and the


natural world can become visible’ (54).

Though I haven’t done her ideas justice in such a short treatment, themain problem with her work is her claim that standpoint is not relativist(the interpretive view that we can’t judge between conflicting claims) oressentialist (the postmodern assertion that we can’t generalize) (Hard-ing (1993), 61-2). She spends much of her later work (until Harding(2004)) defending her theory against the claims of postmodernists thatby giving up ‘the goal of telling one true story about reality, one mustalso give up trying to tell less false stories’ (Harding (1996), 314). Andhere lies the contradiction: She doesn’t want to abandon objectivitysince she wants to reject relativism. As she states, ‘standpoint theorydoes not advocate—nor is doomed to—relativism. It argues against theideas that all social situations provide equally useful resources for learn-ing about the world and against the idea that they all set equally stronglimits on knowledge’ (Harding (1993), 61). She doesn’t want to saythat your account is as good as mine, especially if your account is an-drocentric. More controversial, and harder to sustain within her ownaccount, is her argument that ‘some social situations are scientificallybetter than others as places from which to start off knowledge projects’(61). But what makes one situation ‘scientifically better’ than another?On what basis does one make this judgment without relying on someform of empiricism? How do we judge between different accounts andavoid the problem of essentialism, that is, making assumptions aboutgroups of people (women, aboriginals, disabled) that they share essen-tial qualities? How do we avoid the hierarchies of oppression set upby standpoint (which oppression are we privileging: race, class, or gen-der?), and how do we avoid idealizing oppression of different groups?(On (1993)). Here Harding reaches an impasse and argues for ‘strongobjectivity,’ making the context of discovery part of the research and,therefore, acknowledging one’s own subjectivity to make one more ob-jective. ‘Strong objectivity requires that the subject of knowledge beplaced on the same critical, causal plane as the objects of knowledge.Thus, strong objectivity requires what we can think of as “strong re-flexivity”’ (Harding (1993), 69). Strong reflexivity means ensuring thatthe underlying assumptions and positions of both researchers and re-


searched are observed and reflected upon. The subject of knowledgebecomes part of the object of knowledge and the context of discoverybecomes part of the method itself (70).

But is this really incompatible with feminist empiricism? How canwe have the notion of ‘strong objectivity’ without recourse to someform of empiricism? And how is strong objectivity not intimately con-nected to the individual knower, as in feminist empiricism? As Roseputs it: feminists who work with quantitative methods also want to ‘re-define “objectivity” in terms of the need to make one’s position knownrather than invisible, and by limiting claims to universal applicability offindings—whether these stem from quantitative or qualitative research’(2001, 6). Harding, however, explicitly rejects the work of philosophersof science like Helen Longino and Lynn Nelson (Harding (1993), 51).She argues that though they offer ‘sophisticated and valuable feministempiricist philosophies of science,’ they are not really empiricists be-cause they accept ideas that would be ‘anathema’ to traditional empiri-cists, such as the ‘inescapable but also sometimes positive influence ofsocial values and interests in the content of science’ (51). Perhaps Hard-ing should read Braybrooke’s account of the changes within naturalismor look more closely at Longino and Nelson.

Helen Longino’s work suggests that we should detach ‘knowledgefrom an ideal of absolute or unitary truth’ (Longino (1993), 114). LikeBraybrooke’s recognition of the limits of naturalism, she uses the ‘se-mantic or model-theoretic’ view of science and in very similar termsas Braybrooke. Comparisons are drawn between the real world andthe theoretical world, and metaphors and semantics are used to assistwith the understanding. ‘The adequacy of a theory conceived as amodel is determined by our being able to map some subset of the re-lations/structures posited in the model onto some portion of the expe-rienced world’ (115). Thus, various sub-communities within a larger(scientific or social scientific) community will create numerous theo-ries, models, and ideologies. Knowledge becomes active, not static, andmultiple ways of knowing are validated (Longino (1993, 1996)).

Lynn Nelson’s work on epistemological communities could also bevaluable to the acceptance of feminist empiricism (Nelson (1993, 1996)).


She argues that communities construct knowledge because communi-ties are prior to individuals. What I can know is only what we know(Nelson (1993), 124). She draws on Quine’s work on naturalized epis-temology to show how radically interdependent methods and knowl-edge are, and how we begin knowledge with the historical, social, andscientific practices of a community. She argues that ‘knowledge is so-cially constructed and constrained by evidence’ (129). Her example ofman-the-hunter theories is illustrative here. Quoting Longino, when an-thropologists or archaeologists have examined chipped stones and thentheorized that the stones were used by the men for hunting (as opposedto the women for gathering), they are filling in the gap between theoryand data with their own assumptions, stereotypes, and biases about therelations between men and women (143-4). There is no direct evidencefor either theory. Therefore, we have to make gender relations part ofthe research project itself. Her conclusions suggest that we do not needto reply on ‘one timeless truth’ even as we ‘demand more empiricallyadequate knowledge’ (151).

Similarly, Richmond Campbell and his work on naturalized episte-mology would be helpful to Harding, as he seems to accept her view that‘the influence of feminist values can make the process of testing morelikely to track the truth and hence more objective than it would other-wise be’ (Campbell (1998), 2-3). He wants to give a ‘realist conceptionof objectivity [that] is compatible with feminist political goals,’ that, infact, requires them, and he ‘doesn’t want to reject the idea of truth or thepossibility of impartial inquiry’ (6). What he does that she doesn’t (andprobably should) is accept that values and knowledge can be sociallyconstructed and objective. For example, don’t feminists want to be ableto say that women are oppressed by such and such and here’s how? That‘women’s subordination exists and is wrong’ (socially constructed andobjective)? (2). Campbell, similar to Braybrooke, recognizes there isa ‘corresponding interdependence between facts and meanings and thatmeaning and value can’t be ascertained independent of each other’ (3).Many of the arguments he develops over the course of the book alignwith Braybrooke and would probably help Harding’s own arguments:political values don’t make objective inquiry impossible; science is not


always value-neutral; realism (‘a world existing independently of howwe think of it’) can be reconciled with the belief there isn’t one truedescription of reality; and finally, as I’ve outlined above, even realistscan see inquiry as socially constructed and value laden (2). His natu-ralized epistemology would recognize the ‘social and reflexive natureof knowledge’ (4), and would ‘reinforce and extend the present femi-nist critiques of androcentric science and the fact/value dichotomy butit promises to do so by contributing to a politically useful account ofobjectivity that is uncompromising in its realist implications’ (2). Butfor the realism, this sounds a lot like Harding.

Ironically, Harding may herself be moving in this direction. In hermost recent anthology (Harding (2004)), she seems to be suggestingthat perhaps epistemological relativism is not the horror that was pre-viously thought. She gives four reasons why ‘relativist fears can beset aside’ (not that feminist standpoint is not relativist, which she hasargued in the past) (11). Similar to Braybrooke (and Campbell), Hard-ing argues that just because a research project was shaped by values orinterests doesn’t mean that ‘empirical or theoretical quality of the re-search’ is necessarily contaminated (11), which seems quite differentfrom her earlier writings on androcentric research projects. Second, sheargues that ‘claims of any sort only [have] meaning in some particularcultural context,’ which is certainly an interpretive claim that she haspreviously dismissed. Third, she recognizes that we all have to makechoices ‘between value-laden and interested claims’ and yet we ‘gatherall the information we can from every kind of source available, weighit and tentatively choose’ (11). I’m not entirely sure how this helps herargument but I think she means that even if it is relativist, we still act,and therefore judgments and decisions are still made because of rela-tivist information. ‘We would regard as mentally disturbed someonewho let himself be paralysed by relativist considerations’ (11). Finally,and most importantly for the purposes of this paper, ‘we need to workout an epistemology that can account for both this reality that our bestknowledge is socially constructed and also that it is empirically accu-rate’ (12), which resonates with Longino, Nelson, Campbell, and Bray-brooke. Maybe there is a unity after all.


8.5 Conclusions

In this chapter, I have tried to show how Braybrooke’s work has influ-enced the development of the philosophy of the social sciences and howit has paralleled certain feminist critiques of science, social science, andepistemology. Braybrooke argues that there is a unity of method in thesocial sciences, and I would argue that though a debate still goes on insome circles, many feminists accept that claim as well. He gives goodgrounds for not dismissing quantitative methods out of hand, and showshow both can and should be used in research projects. He recognizesthe changes within naturalism and the natural sciences, which wouldmake feminist criticisms of the fact-value dichotomy more mainstreamthan before, and less problematic for the dilemma of relativism. Thecomplementarity between naturalism and interpretive theory resonateswith the multi-methods of feminists, and a naturalized feminist episte-mology would go some distance in addressing standpoint theorists’ con-cerns about the philosophy of social science. Though not his definitivewords on the subject, I’m sure, I’ll leave the final words for Braybrooke:

Once the relations within and between the schools have beenassessed again, with the lessons about postmodernism al-lowed for, the relations can be described as relations of inquiries—inquiries concerned with rules, concerned with regularities,concerned with texts, concerned with subjective experience.If one likes, the distinction between the schools, having servedall the purposes mentioned, can then be left behind as a his-torical curiosity. (Braybrooke (1998b), 840)

Chapter 9

Empathy and Egoism

SUE CAMPBELL

Abstract

In Natural Law Modernized, David Braybrooke raises the important ques-tion of why default egoism—that is, egoism as the default account of hu-man motivation—continues to dominate the imagination of philosophicalethics. In this paper, I use recent work by Braybrooke and by Nancy Sher-man to argue that reliance on moral sentiment theories can re-inscribeegoism in accounts of moral motivation that attempt a more social start-ing place, and that this re-inscription blocks our understanding of groupidentifications. Default egoism slips into Braybrooke’s own account ofmotivation with his use of Humean moral sentiment. Braybrooke, how-ever, also offers a naturalistic picture of our moral development as per-sons, which picture is independent of his use of moral sentiment theory.Braybrooke’s conception of moral education does dislodge default ego-ism, and I highlight the conception of persons that allows him to moveaway from egoism.

9.1 Introduction

An admirable feature of the political ethics of David Braybrooke is hisinsistence on a resolutely social starting place for ethics: the convic-tion that core questions in ethical theory concern and are addressed tomembers of communities who share or can be educated to share an in-terest in communal thriving. The conviction is evident in his attention

223

224 CHAPTER 9. SUE CAMPBELL

to needs as the human imperative to which our political proposals oughtto respond (Braybrooke (1987a)), and in an account of moral motiva-tion which supports the centrality of meeting needs in his work. Towit, Braybrooke contends that the rational egoist of traditional ethicsis a philosophical golem who distracts theorists from deriving a soundaccount of moral motivation as founded in fellow feeling and sharedpurposes. Our relationships exhibit ‘mutual commitment transcendingthe demand that others be able to make a productive contribution toa society of mutual advantage,’ and a complete ethics needs to discernfrom whence this commitment derives (Braybrooke (2001), 199). Bray-brooke writes:

The party that wants to found ethics on rational egoism asksthis, without any conviction that it can be answered, or evenneeds to be, as a question about how the commitment canbe added to rational egoism as something that goes beyondegoism. The party that does not believe in rational egoismas a foundation for ethics thinks the question needs to beanswered, but may understand it in the same way, assumingthat rational egoism is at the very least the default case of(mature) human motivation and that we need an explanationof how anything more generous in motivation appears ... Itis not rational egoism, however, that comes first, or serves asthe default case. (199)

Braybrooke in fact contends that as persons are socially shaped to-wards engaging in common projects and thus towards sharing interests,egoism would have to be understood as, at best, a degraded form ofhuman motivation.

I share Braybrooke’s commitment to an account of moral motivationgrounded in an understanding of persons as shaped through their com-munal lives. Psychological egoism is implausible as a description ofa widespread psychological orientation. There thus seems little reasonto defend egoism as the best that we are capable of morally. More tothe point of Braybrooke’s remarks, I believe that habitual philosophicalengagement with egoism is not a good use of time for theorists like my-


self who are concerned with issues of group hierarchy, group conflict,and the need to theorize cosmopolitan alternatives to present relation-ships of group exploitation and harm. Braybrooke raises the importantquestion of why default egoism—that is, egoism as the default accountof human motivation—continues to dominate the imagination of philo-sophical ethics.

A number of ethical theorists have seen our emotional capacities forempathy or fellow feeling as an alternative to grounding ethics in ourundoubted capacities for self-interest. In this paper I test Braybrooke’sinsights and concerns about default egoism against two such theories ofmoral motivation. I draw the first from recent work by prominent virtuetheorist Nancy Sherman (Sherman (1998b,a)). The second account isBraybrooke’s own, fromNatural Law Modernized(2001). Both Sher-man and Braybrooke have insisted that philosophers move away froman exclusive focus on reason and pay more attention to emotion in ac-counts of moral motivation. Indebted to the history of moral sentimenttheories, both focus on human capacities to sympathetically engage withothers’ feelings and argue that our sociable capacities befit us to supportthe good of others.1 I pose a potential dilemma for Sherman and Bray-brooke: highlighting our capacities for fellow feeling in order to directan alternative to an ethics of self-interest may simply continue to re-inscribe egoism as the default case of moral motivation.

I use Sherman’s work on empathy and cosmopolitanism to highlightand expand on the importance of Braybrooke’s concerns about defaultegoism. Sherman’s theory of empathy represents persons as egoists inways that conflict directly with the empirical research on which she re-lies. Default egoism is an assumption of her view. Braybrooke’s in-sight is that default egoism thwarts our attempts to properly conceivethe conditions of communal thriving. Sherman offers empathic altruismas a solution to egoism; in her theory, however, it is merely egoism’sJanus face. I shall argue that the motivational frame of egoism/altruismboth fails to capture our collective identifications and blocks attentionto their presence. If we do not understand these identifications, there

1Sherman defends the possibility of moral cosmopolitanism; Braybrooke defends the moremodest possibility of a widespread commitment to common goods.


is no possibility of offering a normative ethical theory that can supportcommunal thriving.

Sherman’s theory is an instructive case study of the kinds of con-ceptual and metaphorical structures in which accounts of empathy andfellow feeling tend to be embedded and which make egoism difficult todislodge. I shall show that default egoism in fact slips into Braybrooke’sown account of motivation with his use of Humean moral sentiment.Braybrooke, however, also offers a naturalistic picture of our develop-ment as persons that is independent of his use of moral sentiment theory.I argue that Braybrooke’s conception of moral education does dislodgedefault egoism, and I highlight a very different kind of representation ofthe self implicit in his thinking.

I offer two preliminary remarks to clarify the scope of this paper.First, philosophers distinguish between different kinds of commitmentto egoism in a theory of moral motivation, and I am interested in onlyone type. Philosophers have distinguished between advancing the de-scriptive claim that people act in predominantly or wholly self-interestedways—psychological egoism—and the normative claim that actions areright and/or rational only insofar as they maximize self-interest—variously,ethical and rational egoism (Feinberg (1978); Shaver (1998)). Commit-ments to psychological and normative egoism are closely linked. Adescription of people as predominantly egoistic places pragmatic limitson what can be expected as ethical or rational action, and a defence ofethical or rational egoism typically depends for its plausibility on theassumption that people are predominantly psychological egoists. Sher-man and Braybrooke are interested in deriving normative ethical posi-tions via realistic naturalized accounts of human motivation. I thus takethem to be primarily denying that people are predominant psychologicalegoists.2 The relation of empathy to psychological egoism (to which Irefer simply as ‘egoism’) is the focus of this paper.

Second, in assessing feminist challenges to the contracting individ-ual of traditional Western political theory, Genevieve Lloyd argues that‘the exercise of imagination is often central’ to challenges to moral the-

2Braybrooke explicitly denies the normative claim that rational action is co-extensive withself-interested action, but does so by showing the implausibility of psychological egoism.

9.2. SHERMAN’S ACCOUNT OF EMPATHY 227

ory (Lloyd (2000), 113). Lloyd points out that when we confront a fa-miliar theoretical model of the self with an undertheorized area of expe-rience, we often become aware of the inadequate fit between the modeland experience. This dissonance can make salient certain unexaminedfeatures of our models of the self and press us towards understandingsmore adequate to the experiences we are discussing. My intent is tomake salient how we represent or imagine persons as egoists throughaccounts of empathy and fellow feeling, and to use Braybrooke’s workon moral education to suggest an alternative.3

9.2 Sherman’s Account of Empathy

In recent work, Nancy Sherman seeks a naturalized approach to moralmotivation ‘at once sensitive to the rigors of a philosophical account andrespectful of the increasing empirical evidence we have of how we areconstructed as psychological creatures’ (Sherman (1998a), 109). Sheargues that our capacities for empathy provide an alternative or addi-tional basis for a cosmopolitan ethics to that offered through Stoic orKantian conceptions of a universal moral community founded on rea-son: ‘Studies suggest there are more ubiquitous and fundamental wayswe step beyond our egocentric perspectives. These involve empathy andits precursors ... What seems increasingly confirmed is that we step outof our egocentric, and by extension, culturally parochial worlds throughmechanisms such as empathy’ (Sherman (1998b), 80-1).

The concept of empathy has developed through a long and complextheoretical history, and Sherman makes considerable use of this historyto explain empathy and map its variations. What seems central to em-pathy throughout its lineage is the somewhat vague idea of being ableto take on another’s perspective to a situation and thereby respond as

3This paper is generally indebted to Lloyd’s work. It is also indebted to Khadija Coxon’swork on empathy (Coxon (2003)). Coxon advances the interesting position that focus on empathicaccuracy as the primary problem for empathy theory is in fact symptomatic of a non-relationalunderstanding of persons. The focus assumes that others’ thoughts and feelings are fully formedand there for us to take up or ignore and does not attend to how our efforts to understand others mayshape their experiences and subjectivities. Following Coxon’s lead, I am concerned in this papernot with empathic accuracy but with broader moral considerations involving how we conceivepeople through theories of empathy and fellow feeling.


they are responding or would respond. Adam Smith, Sherman’s pre-ferred touchstone, writes: ‘By the imagination we place ourselves inhis situation ... we enter, as it were, into his body and become in somemeasure the same person with him; and thence form some idea of hissentiments, and even feel something which, though weaker in degree,is not altogether unlike them’ (Smith (1759), 47-8, quoted in Sherman(1998a), 88).

There are two familiar aspects to Smith’s attempt to make concreteempathy as a mechanism for gaining access to others’ thoughts and feel-ings: an offer of spatial metaphor involving change of location (‘weenter, as it were, into his body’) and reference to the human capacitiesor tendencies—in this passage, imagination—which might defend theappropriateness of such metaphors. The dual way of concretizing em-pathy occurs throughout Sherman’s account. On the one hand she refersto empathy as a kind of ‘active transport’ (Sherman (1998a), 87). Onthe other, and crucially given her commitments to a naturalized moralpsychology, she offers an extensive survey of recent empirical work onchildren’s capacities for motor mimicry, shared attention, and imagina-tion, specifically role-playing.

For example, three-day-old neonates can and do mimic mouth open-ing and tongue extension, as well as the crying of similarly aged neonates.Children and their caretakers engage in mimetic synchronies, pickingup shared rhythms of behaviour (Sherman (1998a), 104). Moreover,children are ‘robust pretenders,’ and this, Sherman writes, ‘undoubtedlycontributes to their learning to role play in a way required for under-standing others’ beliefs’ (106). Sherman takes the studies she surveysto support claims about the biological nature of our empathetic capaci-ties and the ubiquity of our empathetic responses. In addition, Shermanrefers to studies that allegedly support the thesis that empathy predis-poses us to sympathy towards others and thus motivates us to behavealtruistically. Altruism is sometimes understood as acting on others’ in-terests in ways that compromise one’s own welfare. Sherman uses theterm to mean simply that we act on others’ interests while putting asideour own. This latter use is relatively standard in philosophical texts thatcontrast egoism and altruism, and I follow it in this paper.

9.3. DEFAULT EGOISM IN SHERMAN 229

Finally, there is a distinction in Sherman’s work between two modesof empathic engagement, which distinction is important to understand-ing her representation of the empathic self. In imagining Ella’s situa-tion, for example, I might try to imagine myselfas Ella, attempting tograsp her response through an act of self transformation. Alternatively,I might imaginemyself in Ella’s situation, an act of projection, thenanalogizing from my own response to Ella’s likely response. This lat-ter mode of inference remains empathy-like in its objective of trying togain access to others’ perspectives in order to share and thus understandtheir responses. Sherman allows a greater role for projective empathythan most theorists yet concedes that this mode may at best give us in-formation about how we ourselves would respond to a situation. Trans-formative empathy remains Sherman’s ideal of empathetic engagement,a normative heuristic for our empathy projects, and she suggests ourearly capacities for mimicry can be seen as its developmental buildingblocks (101-2).

Sherman joins a number of other recent theorists in hypothesizingthat empathy is a fundamental way of understanding others. She sug-gests that the empirical work she discusses gives credence to claimsthat empathy grounds our ability to form shared conceptions, to formattachments with others, to achieve the self-knowledge that comes fromunderstanding how we are seen, and the self-esteem that comes fromfeeling understood ((1998a), 84). Most important, empathy should holdpride of place in our accounts of moral motivation. Through empatheticengagement we overcome both egoism and cultural parochialism anddevelop altruistic tendencies that extend beyond family, associates, andfellow citizens to ‘those outside our borders’ (Sherman (1998b), 103).

9.3 Default Egoism in Sherman

Notable in Sherman are persistent and unreflective references to egoismand altruism as the frame both for understanding the nature of empa-thy and for persuading us of its moral importance. To use Braybrooke’slanguage, Sherman’s project is to understand our commitments to oth-ers, and she sees this task as one of describing how our commitments


transcend or go beyond egoism. Sherman argues that empathy is theprimary way in which we move from an egoistic perspective to a moregenerous altruistic one. Because altruism is the motivational structurethrough which Sherman understands our will to be other-interested, re-search into our empathetic capacities so as to encourage their develop-ment is key to normative ethics. I thus take Sherman to be committed todefault egoism in an account of moral motivation. Note that we movefrom an initial, presumably natural egoism to a developed altruism. Inow sketch some of the ways in which egoism creates problems for thecoherence of her account and subverts its objectives.

Most strikingly, though the importance of empathy allegedly de-rives from the necessity of overcoming egoism, Sherman’s descriptionof empathy undermines her supposition that we are egoists. Shermancommits herself to a strong position on the epistemic function of empa-thy. It is, she suggests, a primary way of understanding others’ minds;from the age of five on, we empathize with others in our daily inter-course with them in order to explain and predict their behaviour. OnSherman’s view, the egoist could not get on in the world at all, as sucha person would not be able to understand or predict the behaviour ofthose around her. Sherman’s account arguably renders egoism impossi-ble. Yet, in Sherman, it is the morally problematic place of our naturallyegoist perspective to which the moral importance of empathy answers.What I take to be an incoherence in Sherman’s account—that egoismis necessary to generate our commitment to empathy theory, but is nottenably maintained in the face of that theory—obviously threatens oneof her objectives. We cannot understand how empathy responds to theproblemof egoism if our natural tendencies to empathize make egoisma psychological non-starter.

It is important to note that we have no reason to suppose that any ac-count of the empathic egoist would be similarly incoherent. The prob-lem in Sherman is clearly tied to her epistemic use of empathy as ourprimary mode of access to others’ minds. The problem does serve tofocus our attention squarely on egoism and to compel us to ask why itis there as an assumption. In other words, the uncomfortable role ofegoism in Sherman’s account can show us why default egoism can be

9.3. DEFAULT EGOISM IN SHERMAN 231

so difficult for empathy theorists to avoid. Before turning to this issue, Iwill outline what further damage egoism does to Sherman’s aspirations.

Sherman postulates that through empathy we can overcome our ten-dencies towards cultural parochialism and, thus, move towards moralcosmopolitanism. Sherman uses the phrase ‘cultural parochialism’ togesture towards the morally problematic aspects of our collective livesthat come from identifying too exclusively with the interests of thosewhom we take to be similar to ourselves. Some philosophers have wor-ried that empathy exacerbates rather than ameliorates parochialism, apossibility of concern to Sherman:

There is one final point of linkage between empathy stud-ies and moral theory that needs to be discussed. And thisis the theme I began with, of cosmopolitanism. Can empa-thy break out beyond parochial boundaries of similarity andfamiliarity? ... In the end, do we merely trade egocentrismfor ethnocentrism, or something like it? (Sherman (1998a),113)

The important discussion is never held. Because egoism is struc-turally situated as the primary problem for moral motivation, the prob-lem of parochialism is simply deferred. It is Sherman’s ‘final point oflinkage’ and she addresses it by noting that ‘these are complex matters,but empirical studies can further advance our understanding’ (113).

There are also serious and subtle ways in which default egoism mis-leads us about the problematic dimensions of our collective identifica-tions. Because Sherman positions egoism as the default case of ma-ture human motivation, egoism is the core motivational structure fromwhich other problematic psychological tendencies are held derivative.It is thus an assumption of Sherman’s view that cultural parochialism isan extension of egoism. We have no reason to accept this assumption.As egoism itself has no real descriptive grounding in the account, it isnot available to explain group allegiance. Without some understandingof the problem to which empathy responds, there is nothing on which torest the claim that empathy will ameliorate parochialism.


Relatedly, as Sherman pictures individuals as egoists, they are with-out complex motivational structures derived from their relations to oth-ers. Imagining culturally parochial worlds as extensions of egocentricworlds homogenizes the character of cultures, distracting our attentionfrom their internal complexities. The term ‘cultural parochialism’ it-self paves over differences within culture. When Sherman talks of thewill to ally with those outside our borders, she neglects the existenceof group hierarchy and harm that is internal to complex and diverse so-cieties. I call these problems serious and subtle because I do not thinktheorists have properly seen that the tendency to homogenize group lifeis sometimes linked to beginning a discussion of moral agency fromthe standpoint of the egoist individual. I will have more to say aboutthis tendency when I examine Sherman’s imagery and when I turn toBraybrooke’s account of moral education.

Finally, psychological egoism is a theory of individual moral psy-chology. Joel Feinberg made the point decades ago that the assumptionof universal psychological egoism leaves us without a way to make adistinction between people who really are selfish and those who are not(Feinberg (1978)). Sherman’s claim that parochialism is an extensionof egoism may lead us to think that group selfishness is as universalas the individual egoism from which it somehow derives. Sherman’saccount displaces our attention from the real global problem of moralselfishness which resides largely in the consumerism, greed, and will toexploit others manifest in Western nations and which requires an analy-sis of distinctive group characteristics and relations.

Sherman’s strong endorsement of an empathy-based account of moralmotivation is obviously and explicitly programmatic in intent. Thisis not the basis of my criticism. My concern is that normative ethicsshould respond to a description of real moral concerns. Default egoismtakes up a place in Sherman’s account that should be occupied by anaccount of the real-world harms and injustices for which empathetic en-gagement with others might be a helpful response. David Braybrookewrites: ‘In the world at large, the excesses of group feeling are far moredestructive of ethics than the excesses of self-interest. In the face ofthese excesses, philosophers and moralists are arguing for a possibility

9.4. IMAGINING EMPATHIC SELVES 233

of moral community that often fails to be realized, sometimes it seems,on every hand’ (Braybrooke (2001), 146). Sherman offers us no insightsinto moral failures that derive from our allegiances. Unlike the golemof default egoism, these allegiances are very clearly a problem for thecosmopolitan ethicist and they take forms whose variety can only beobscured by the claim that when we see how empathy can help us tran-scend egoism, we will see that it can help us with parochialism. I takeSherman’s inability to deal at all realistically with group identifications,allegiances, and hostilities to be the most serious cost of default egoismin her account. Whether Braybrooke can deal realistically with theseallegiances will be a test of his account.

The unstable positioning of default egoism in Sherman’s account,and the problems it causes both for coherence and the promise of anapproach to ethics that highlights our capacities for empathy, raise thequestion of why she is committed to egoism/altruism as a conceptualframework for thinking about empathy. Is there something about Sher-man’s understanding of empathy that tends to drag egoism in its under-carriage? It is important to address this question if we are to avoid thecommon philosophical tendency to re-inscribe egoism in our theorizing.

9.4 Imagining Empathic Selves

Like most philosophers, Sherman pays little attention to the metaphorsthrough which empathy is concretized. The history of describing empa-thy is strikingly spatial, and Sherman depends on this history as a pri-mary way to locate and elaborate the concept of empathy and to mouldher distinctive use of it for moral cosmopolitanism. The metaphorsinvolve both change of place and the idea of empathy as overcomingdistance. I suggest these metaphors reinforce a dichotomous motiva-tional structure of self-interest versus other–interest that leaves egoismin place.

Many theorists have described taking on others’ perspectives as achange in location. Adam Smith described what he called ‘sympathy’as a case of trading places with others in fancy. One then brings the caseback ‘to one’s own bosom,’ comparing the others’ feelings and thoughts,


available through sympathy, to one’s own, and judging the appropriate-ness of the others’ responses (Smith (1759), 141, quoted in Sherman(1998a), 89). Edward Titchener, who first used the term ‘empathy,’ con-ceived it as an exploration of foreign environments: ‘as we read aboutthe forest, we may, as it were,becomethe explorer; we feel for ourselvesthe gloom, the silence, the humidity, the oppression, the sense of lurkingdanger; everything is strange but it is to us that strange experience hascome’ (Titchener (1915), 198, quoted in Sherman (1998b), 110). Thepsychologist Carl Rogers, who developed empathy theory as a thera-peutic model, spoke of ‘entering the private perceptual world of anotherand becoming thoroughly at home in it . . . To be with another in thisway means that for the time being you lay aside the views and valuesyou hold for yourself in order to enter another world without prejudice’(Rogers (1975), 4, quoted in Sherman (1998a), 92). Sherman herselflinks the development work that she surveys to a five-year-old’s abilityto make accurate predictions of others’ behaviour through metaphori-cal description that continues the theme of changing place: ‘Five yearolds can put themselves in someone else’s shoes. In figuring out others’behaviour, they can leave behind the egocentric point of view ... Thisdevelopmental transition is important, because it marks the milestone atwhich we break out of our egocentric world, and begin to see, in a sub-stantive way, through others’ eyes’ (Sherman (1998b), 112-13). Manymetaphors of change of place, Titchener’s for example, also involve asense of distance, and Sherman often adds the connotation of distanceto her descriptions by characterizing the places visited as alien and re-mote. She comments, for example, that ‘several points indicate widelyheld views about empathy in the clinical domain. First is the metaphorof empathy as perceiving another’s mind in a way that allows one to be-come familiar with its alien habits and terrain’ (Sherman (1998a), 92).

The possibility of transformative empathy that is Sherman’s norma-tive heuristic is made concrete through the idea of a change of place.Empathy as active transport is conceived as leaving the self, picturedas one spatial territory, in order to enter a second self conceived as adifferent spatial territory: a body, environment, world, pair of shoes, orpair of eyes. I can achieve another’s perspective because in empathizing

9.4. IMAGINING EMPATHIC SELVES 235

I have left behind the values and responses that characterize my ownperspective and might, if carried with me, bias an empathic encounter.

The spatial metaphor of leaving oneself to occupy another is not in-dissoluble from empathy as a concept, as some empathy theorists havemodelled what Sherman refers to as a projective rather than a transfor-mative ideal. However, Sherman’s use of this metaphor to support atransformative ideal makes it difficult to avoid the conceptual frame ofegoism/altruism. On Sherman’s model, self-interests are placed in oneworld bounded and distinguished from a second world characterized bythe others’ interests. When we engage someone’s interests through em-pathy, we transport ourselves to them, leaving our interests behind. Thispicture leaves egoism and altruism, traditionally conceived as beingself-interested versus being other-interested in a way that excludes self-interest, as the most tempting frame for conceiving the interest structureof empathy. Any serious moral engagement with others’ interests meansleaving the self and its interests behind, so the basic egoist individual,one characterized solely by self-interest, is embedded in empathy as itsstarting point.

Sherman utilizes the spatiality of empathy to make it seem especiallysuitable to moral cosmopolitanism. Sherman describes moral growth asa matter of moving steadily outward from the ego/self towards fam-ily, friends, countrymen, and finally the distant others beyond our bor-ders, first gaining access to their interests through empathy and thenbeing moved to action. Cultural parochialism is a result of empathy’snot being sufficiently ‘far-flung,’ and Sherman urges us to overcomethis distance: ‘the remoteness of a culture ... may not weigh heavilyagainst understanding the shared dimensions of human suffering andloss’ (Sherman (1998b), 113).4

Sherman’s exploitation of distance allows us to press further on theinadequacy of her treatment of group affiliation. The distance that em-pathy metaphorically traverses makes it seem like the problem to whichour ethics must respond is a lack of engagement with others rather thanrelations that exploit and harm them. Moreover, the metaphor of dis-tance fixes our attention in the wrong place, away from our local inabil-

4’Far-flung’ is how Sherman characterizes moral respect.


ity to treat others well and thus it contributes to the homogenization ofcultures in her account. Many of the harms that result from our groupallegiances and identifications—the harms of racism, sexism, ethnocen-trism, homophobia, and so on—are harms to those with whom we arein daily contact in diverse societies. The representation of the spatialself imported through empathy obscures the often destructive role ofour collective identifications, specifically issues of group hierarchy indiverse societies.

9.5 Moral Sentiment

I have used Sherman’s work to illustrate the force of Braybrooke’s con-cern about the tendency to understand moral commitment in ways thatleave egoism intact. Although Sherman is committed to a naturalizedaccount of moral motivation that shows how we develop towards so-ciability and concern for others, the egoist is firmly entrenched in Sher-man’s view through spatial metaphors for empathy that offer self-interestversus other-interest as the only options for moral motivation. Throughspatial metaphor we gain a representation of persons as originally egoistand at a distance from others. Moral growth is seen as the problem ofengagement achieved through sympathetic altruism.

Although moral sentiment theories are by no means the exclusivehome of spatial metaphors for the self, Adam Smith is the primary philo-sophical influence on Sherman’s account of empathy, and Smith andHume are key reference points in the history of the concept. I now turnto Braybrooke’s endorsement of Hume to raise my caution about thelegacy of moral sentiment theory. Braybrooke musters Hume in supportof his interesting thesis that the natural law tradition he defends has hada robust presence in the history of modern ethics. He also uses Humeto develop his own attention to emotion in the theory of moral motiva-tion and he implies that Hume’s theory of moral sentiment is compatiblewith his own account of moral education. I will go on to suggest thatBraybrooke’s imagining of the self is very different than Hume’s.

In both theTreatiseand theInquiry, Hume defends versions of thethesis that people are naturally disposed to sympathy with others, ‘and

9.5. MORAL SENTIMENT 237

in accordance with this concern, moved to support whatever is useful tothem’ (Braybrooke (2001), 129). Our approval of what benefits othersis referred to by Braybrooke as the sentiment of humanity or humanefeeling. Because the sentiment of humanity is universal (common to alland extending to all) and directed to what is useful for others, it seemsparticularly suitable to an account of motivation that makes a case forour capacity to support communal thriving.

There is a great deal of commentary and dispute about the best ac-count of both humane feeling, the Humean ‘moral sentiment,’ and ofthe sympathy on which it is based. Hume’s account of sympathy (hiscontribution to empathy theory) suggests that when we are in the pres-ence of another’s passion, we form such a lively idea of it as becomesthe passion itself (Hume (1978), 575). Commentators have struggledto make sense of this mechanism. My interest here, however, is not inexegesis of Hume but in how Braybrooke’s reading of Hume representsmoral agency. In Hume, capacity for humane feeling may be hamperedeither by self-interest or partiality for family and friends, or by a dis-tance from others in time and space that challenges the possibility ofsympathy. Developing the sentiment of humanity requires compensat-ing for these tendencies.

On Braybrooke’s reading, self-interest and partiality are compen-sated for by our trained capacity to assume a fully disinterested per-spective, that is, to bracket or put aside self-interest. Distance remainsa problem (Braybrooke (2001), 146). The mechanism of fellow feel-ing requires real spatial contiguity, and the extension of fellow feelingto those at a distance is, as Braybrooke acknowledges, perhaps nevermade adequately viable in Hume’s account. Nevertheless, Braybrookeexpresses optimism about Hume’s account of human feeling. ‘Withhearts and minds unclouded by self-interest or partiality for family andfriends, we feel for the sufferings of people long ago or on the otherside of the planet; and correcting our judgements to take into accountthe distance, we find those sufferings a great deal more important thana transient scratch on one of our fingers’ (131).

I have no conclusive sense of how committed Braybrooke is to Hume’ssentiment of humanity, but, with Hume, Braybrooke assumes the famil-


iar picture of the egoist self moving through distance towards others viathe development of sympathetic bonds, first easily to family and friends,finally with more difficulty to distant others. First, we find again the im-mediate separation of self-interest from other-interest. We are moved tosupport what is useful to our fellows ‘when [our] own interest is not inquestion’ (Braybrooke (2001), 129). Braybrooke implies that our inter-ests will either conflict with others’ interests or lead us to be indifferentto others. Second, we put aside our own interests in order to assumethe perspectives of others. The sentiment of humanity depends on ourbeing normal human beings, ‘fully equipped with the capacity to feel ina disinterested perspective’ (139).

Finally, because distance attenuates sympathy, we are reinforced inthe idea that the main impediment to communal thriving is disinterest indistant others, a preference for our near and dear. Hume’s moral senti-ment theory does not involve trading places in imagination. But it doescontinue to represent the egoist self as at a spatial distance from others,moving towards them through bonds of sympathy. Moreover, Hume’suse of distance takes on a metaphorical cast. He remarks that we sympa-thize more with ‘our acquaintances than with strangers; with our coun-trymen than with foreigners,’ but given his account of fellow-feeling,we should sympathize equally with all who are spatially contiguous tous, be they strangers or friends (Hume (1978), 580). Felt separatenessfrom others becomes troped once again through the notion of spatialdistance. Hume’s account may at best explain why we must make ex-ceptional effort to become emotionally engaged with distant others, butnot why those spatially contiguous to us, who may differ from us inclass, ethnicity, religious affiliation, sexual orientation, and so on, donot have our sympathy.

Joan Tronto has argued that if we think of the actual historic con-text of moral sentiment theory, it is not odd that the sympathetic selfand the egoist self are reciprocally reinforcing representations. In theeighteenth century Europe saw significant transformations in collectivelife. ‘While humans grew more distant from one another and the bondsbetween them became more formal and more formally equally, theyalso had to expand their gaze beyond the local to the nation, and indeed

9.6. COMMON PURPOSE AND RELATIONAL INTERESTS 239

sometimes to the global level’ (Tronto (1993), 31). Because of the ero-sion of close communal forms of life, ‘the strength of self interest asa regulator of human activity became increasingly powerful’ (50). Themetaphor of overcoming distance had a multi-layered significance formoral sentiment theorists. On the one hand, it spoke to the loss of com-munal forms of life whose regulatory function was being replaced byaccounts of enlightened self-interest; on the other, of the need to engagewith a wider range and variety of people and moral contexts.

Sherman and Braybrooke both draw on the tradition of moral sen-timent theory to find a description of moral psychology that is realisticand optimistic about our sociability. Both objectives are to some degreethwarted by their use of a tradition preoccupied by the representation ofthe self as at a distance from others, an egoist capable of fellow feeling.

9.6 Common Purpose and Relational Interests

The description of empathy as distance-reducing and as thereby encour-aging other-directed interests is so familiar as an account of moral so-ciability that we may wonder what our options are. In the chapter onmoral education inNatural Law Modernized, Braybrooke elaborates anaccount of universal childrearing that shows how we develop into per-sons via our enlistment into common projects and purposes. I shallargue that Braybrooke’s account of moral education offers us the op-tion of relational interest as well as an account of human developmentthat supports its centrality. In the final section of this chapter, I alignhis writing with recent feminist accounts of the self and emotion to ar-gue that it provokes and requires a very different kind of philosophicalimagining for the self than that found in moral sentiment theory, onemore adequate to theorizing group hierarchy in complex cultures.

Like Sherman’s, Braybrooke’s theory of motivation is modestly nat-uralized. But it is an account of our natural capacities as trained. Bray-brooke’s theory of motivation is a description of moral education, anaccount of how children are raised to be sociable in ways that may leadthem to support rules and laws that have as their end the good of thecommunity. The core of the account is that, universally, we raise chil-


dren to be persons by continually enlisting them in joint activities. Thesuccess of these cooperative activities requires that the participants de-velop mutual intentions and commitments towards a common end, un-derstood as the good of that activity.

Briefly, parents and other caretakers initially give significance to theinnate responses of their infants; that is, they attribute to their infantsdesires, intentions, frustrations, and so on, and do so in order to devisepatterns of interaction that will eventually make room for the child as apartner. As the child develops, the child and the adult come to rely oneach other to fulfill their responsibilities in the partnership. Braybrookewrites:

Children, like dogs and horses, can normally rely on com-mitments forthcoming from their mentors in return for thecommitments that they themselves make. If Kate desistsfrom throwing her porridge upon the floor, she will be leftto eat in peace; it will be supplied in due quantity at duetimes. If Frank stays out of the street, the walks to the parkwill continue. Moreover, when things are working well, thecommitment leads to achievement of common goods—theCommon Good of each joint activity or ritual: Both motherand child act with the object of having the child regularlyfed under a form of civility that both respect; both father andchild act with the object of walking to the park and doing soin decent order and safety. (Braybrooke (2001), 204)

Through these common activities, children’s needs are met in waysthat enlist them into partnerships in which they share intentions andgoals with others. Children develop not only pride in themselves fortheir contributions, but also trust in and concern for their partners, avaluing of mutual pleasure in the success of realizing shared intentionsand interests, and cooperative dispositions towards other such activities.

Parents and children have interests in common, but these interestsare alsosharedinterests. Braybrooke’s description of types of intereststhrough which infants are eventually socialized into persons is not to beconfused with a familiar philosophical picture of ‘parallel but solitary

9.6. COMMON PURPOSE AND RELATIONAL INTERESTS 241

interests served by a common scheme of cooperation’ (64), that is, ofself-interested individuals who act together in narrowly pragmatic ways.Participants in a shared activity take on responsibility for their contri-bution to the activity and hold themselves to be answerable to othersfor their contribution. But they also often take some responsibility forthe contributions of others. They develop mutual care and concern foreach other through their interactions and develop emotional dispositionsthat take as their object the activity as shared. For example, they takepleasure in shared achievement. Finally, they have a commitment to theshared activity. To think of my interests as parallel rather than sharedsuggests that when my hypothesized solitary interests would clearly bebetter met by an alternative arrangement, I would shift to that arrange-ment. But my commitment to a shared activity may make me reluctantto undertake such a shift. Put simply, if Frank’s father is sick and cannotgo to the park, Frank may well decide to stay home and keep him com-pany instead of going for his walk with someone else. As I understandBraybrooke, the responsibility, emotion, and commitment that charac-terize shared interests are neither conceptually nor developmentally in-dependent. The emotional attitudes developed in the course of sharedactivities—for example, Frank’s concern for his father’s well-being—are partly constitutive of our having shared interests and commitmentsand act as evidence of them.

Braybrooke is aware that too much of a disposition towards coop-erativeness can lead to exploitation and that a sensible upbringing willencourage independence and autonomy as well as cooperativeness. Butindependence and autonomy are not conceived as deriving from ego-ism: ‘the point of departure . . . is not a person who is bifurcated be-tween being sociable and being self-striving . . . The independence willbe an independence that has a cast reflecting the development of coop-erativeness’ (207). I gloss this remark by pointing out that our indepen-dence and autonomy as persons are cultivated in the kinds of develop-mental partnerships that Braybrooke describes. Braybrooke also allowsthat persons have egoistic as well as cooperative tendencies, and that ahealthy adult community will include parallel interests.

What Braybrooke insists on is that the kind of engagement in joint


activities that he describes, engagement through which we develop aspersons, puts the partners to the activities ‘outside of egoism’ in their re-lations with each other and in their commitments. Braybrooke’s accountof the fundamental basis of our capacity to care for each other’s interestsdoes not incorporate egoism as one of its stages. In other words, thoughBraybrooke responds to egoism, this response does not re-inscribe ego-ism within his view of motivation. Shared interests will form a dominantpart of most individuals’ motivational matrix. Developmental activities,role-related activities, activities in institutions like sports or the militarythat specifically inculcate shared interests and emotions that bind par-ticipants, activities in which we participate as friends and lovers, andcommunal endeavours that continue over some period of time, can allbe expected to give rise to such interests.

Over the last two decades, feminist theorists have elaborated an ac-count of persons as relationally constituted, as developing in and throughtheir relationships with others. They have focused on how the facts ofour interdependence should cause theorists to reconceptualize humancapacities and values. Braybrooke’s work makes a distinctive contri-bution to this body of theory by attending to motivation. Braybrooke’sfocus on shared interests forces us to see that we must look to collec-tive structures and activities first in order to talk about an individual’sagency. This point can be sharpened by looking at a similar suggestion.

Elliot Sober and David Sloan Wilson, in their classic study of altru-ism, recognize that egoism or altruism may be too spare a binary forthinking about the structure of human motivation and propose an al-ternative, ‘relationalism,’ as a way to think about many of our desires:‘Relationalism is the view that people sometimes have ultimate desiresthat certain relational propositions (connecting self and certain others)be true’ (Sober and Wilson (1998), 226). Sober and Wilson’s wording,in response to classic formulations of egoism, is somewhat awkward;however, a great many of my desires are clearly relational under theirdescription. I want it to be the case that the apples be divided equallybetween us, that I be a trustworthy mentor to my students, that our de-partment excel at undergraduate teaching, that Jan and I have a gooddinner tonight and a good trip to Australia in November—in general,

9.7. RE-IMAGINING THE SELF 243

thatweachieve certain objectives through activities that I participate inas a member of various formal and informal partnerships and collec-tives. Nevertheless, Sober and Wilson’s use of desire to sketch an al-ternative to egoism/altruism leaves open the possibility of an agent withsolitary interests in the context of shared activity. Braybrooke offersa model where the motives of individual agent derive from the sharedinterests formed through their collaborations with others.

9.7 Re-imagining the Self

To use the felicitous phrasing of Genevieve Lloyd, Braybrooke forcesthe philosophical imagination to run from the collective to the individ-ual, rather than, as is more common, the reverse. Lloyd’s writing onre-imagining the self can help us see that this move allows quite a differ-ent model of the self than that complicit with default egoism, and I useher work to more sharply distinguish the view of the self adumbrated inBraybrooke from that involved with default egoism. Lloyd talks abouttwo levels of experience that the shift in imagination highlights: thespecific experience of taking oneself to be bound up with others whenwe act and the complex experience of subjectivity that results from ourshared activities over time. She writes that ‘spatial metaphors are pow-erful in our prevailing political ideals of selfhood, helping to reinforcethe grip on the philosophical imagination of the sharply bordered in-dividual’ (Lloyd (2000), 121). But such models are an ill fit with theexperiences she describes.

One of Lloyd’s striking illustrations of feeling bound up with othersinvolves Pat Barker’s fictionalized account of poet Siegfried Sassoon’sreturn to the front in the First World War. In the first novel of Barker’sRegenerationtrilogy, Sassoon no longer supports the war and is able toavoid returning to the trenches. His decision to return cannot be under-stood, Lloyd thinks, as either egoism or as patriotic altruism. Instead,she writes: ‘Solidarity with the dead sustains the character’s sense ofself, and with it a responsibility to those ... doomed to death: “TheBattalion in the Mud”,


Out of the gloom they gather about my bedThey whisper to my heart, their thoughts are mine.

‘Why are you here with all your watches ended?From Ypres to Frise we sought you in the line.’In bitter safety I awake, unfriended. (117)

The poem is by the historic Sassoon. Lloyd’s point is that to try andpry apart self-concern or interest from other-concern or interest in orderto give one or the other explanatory priority would violate the charac-ter of Sassoon’s experience of selfhood. She writes that Sassoon actsout of a sense of himself ‘as bound up with others, in ways that resistclear-cut divisions between self and other. [His] taking of responsibilityis explicable only through relations that forge identities, in situationswhere there can be no clear cut distinction between self and other, be-tween egoism and altruism’ (117). Sassoon’s identity is that of a friend,an identity forged through the joint project of men trying to keep eachother safe in battle. This is not a project that can be described as havingparallel but solitary interests in safety that are best met by joint activity,or we would have no explanation for why Sassoon returns to the front.

Lloyd picks an example of shared interest that some might read as in-volving action in a context of extremity. Braybrooke’s account of moraleducation provides the basis for understanding the experience of feelingbound up with others when we act and take responsibility as an ordi-nary and familiar part of our moral lives. Like many ordinary featuresof our lives, it may only become phenomenologically apparent whenit becomes a site of tension, when particular shared interests seem toconflict with our own welfare or with the welfare of others. Sassoon’sshared interests threatened his own survival, and friends urged him notto return to the front. A way of thinking about at least some cases ofenmeshment, or over-identification with others, is as situations in whichour shared interests may come to conflict with what is best for others.Though philosophers rarely discuss enmeshment as a moral problem, itis common in those relationships, like parent/child or mentor/student,in which more senior partners must eventually shift out of the joint ac-tivities that have characterized the relationship and allow those younger


to take control.5 Though such examples bring shared interests to thefore as a point of tension, they can help us see the wide base of sharedinterests out of which we act.

Philosophers who want to defeat the egoist have sometimes imag-ined this figure as a psychopath,6 but this imagining has been not beeneffective in locating and dislodging default egoism, the kind of egoismof Sherman’s theory that is compatible with the claim that people aren’treally egoists because they have transcended egoism. To think aboutour experience of self and agency in ways made possible by Lloyd andBraybrooke gives us a better way to ask what we are imagining when weunderstand persons through a motivational structure of egoism/altruism.The interest structure offered by egoism/altruism and offered in Sher-man’s account of empathy gives us a portrait of people who have thecapacity to sharply distinguish self- and other-interest in ways that failto reflect common experiences of self and moral agency.

The alternative model of motivation that I have forefronted in Bray-brooke becomes richer when we follow Lloyd’s advice and move explic-itly from picturing selves spatially to explicitly contemplating the tem-poral dimensions of selfhood. Like Braybrooke, Lloyd regards takingon responsibilities and forming bonds of identification, bonds throughwhich our interests become shared, as reciprocal movements in our de-velopment as persons. But she would argue that Braybrooke’s accountof moral education gives only a snapshot of human sociability; it cap-tures at a time an ongoing process of developing allegiances that shapeone’s sense of self. We may think of this process narratively. We cometo think of ourselves as persisting subjects by interpreting and appro-priating the events of our pasts through narratives about who we areand why we acted thus (Schechtman (1996); Walker (1998b)). Many ofthese narratives will, like Sassoon’s, be narratives of relation, and eachof us will have many such narratives. ‘The fact that a self is intrinsicallysomething that has a past means that there is an internal multiplicity toselfhood’ (Lloyd (2000), 122). Lloyd affirms with Braybrooke that the

5For some discussion, see Michael Stocker (with Elizabeth Hegeman (1996); Piper (1991).Sherman makes reference to Piper’s work (Sherman (1998a)).

6See, for example, Thomas (1980)


formation of selves ‘is a collective matter’ (118) but adds that we mustunderstand persons as bearing this complexity in the way that they ex-perience their identities, responsibilities, interests, emotions, and oppor-tunities for agency. We thus no longer have a picture of moral progressas one of moving outward from a self that has a uniform motivationalstructure to engage with the interests of similar others, which interestsare kept distinct from self-interest on pain of empathic bias.

It should be clear that this relational model of the self renders Bray-brooke’s additional use of Hume’s moral sentiment theory retrograde.Braybrooke’s reading of Hume offers the options of self-interested orother–and–not–self-interested, excluding room for the very forms of re-lational desire that are so important in the account of common purposeand moral development. Moreover, for Hume, normal moral agencyrequires that we can clearly distinguish our own self-interests from oth-ers’ interests. I have suggested, however, that Braybrooke’s account ofmoral education leads to a view of self whose interests are bound upwith those of others in ways that call into question how far we are capa-ble of distinguishing our own interests from those of others, especiallygiven the complexity of our relational development over time.

The alternative, relational representation of the self that I have triedto make vivid through Braybrooke and Lloyd, does, I think, offer op-portunities to better understand the moral problems of group hierarchyin complex and diverse cultures. First, in recognizing that individualsthemselves have complicated relational identities, we will not inadver-tently homogenize cultures by imagining them (as Sherman does) asextensions of single egocentric worlds. Second, Braybrooke and Lloydencourage us to think of how individuals’ motivational structures areshaped through collective purposes and interests in ways that depend onand shape their emotional development. They stress that our coming tofeel certain ways is interwoven with our engagement in joint activities,and the emotional dimension of our engagement is part of what consti-tutes shared interests. When we picture the self as sharply bordered andat a distance from others, our attention remains on the development ofsympathy—the question is how a self becomes engaged with the inter-ests of others. The focus on sympathy, moreover, tends to bring with it


this view of the self. If we view the self as shaped already through col-lective activities and common purposes, we do not have the bare egoistself for whom the main challenge is one of engagement, and we can turnour attention from sympathy to the wider range of emotions that play arole in our group life.

Margaret Urban Walker has argued, for example, that our emotionalresponses form one of the primary ways in which we are collectivelyshaped to maintain political hierarchies (Walker (1998a)).7 She drawsattention to the work of Lillian Smith, an anti-racist activist in the Amer-ican south in the 1940s and 1950s. In her autobiography,Killers of theDream, Smith writes, ‘I do not think our mothers were aware that theywere teaching us lessons. It was as if they were revolving mirrors re-flecting life outside the home, inside their memory, outside the home,and we were the spectators entranced by the bright and terrible imageswe saw there’ (Smith (1950), 83). These words, which invoke Lloyd’saccount of our complex subjectivities, are part of Smith’s introductionto ‘The Lessons.’ In this chapter, Smith details how white southern chil-dren learned lessons about cleanliness, food, sin, and sexuality in waysinterwoven with training in racial segregation:

The lesson on segregation was only a logical extension ofthe lessons on sex and white superiority and God. Not onlyNegros but everything dark, dangerous, evil must be pushedto the rim of one’s life. Signs put over doors in the worldand over minds seemed natural enough to children like us,for signs had already been put over forbidden areas on ourbodies ... Each lesson was linked on to the other drawingstrength from it. (90-1)

Moreover, these lessons involved the inculcation of emotional re-sponses that would maintain this segregation as a joint activity of whitesouthern culture. ‘The mother who taught me what I know of tendernessand love and compassion also taught me the bleak rituals of keeping Ne-gros in their place’ (27, quoted in Walker (1998b), 70).

7have previously linked Lloyd and Walker (Walker (1998a)) on this point in Campbell (Camp-bell).


The emotional responses that characterize racist and ethnocentric hi-erarchies often have nothing to do with physical distance or emotionaldisinterest, but shape hierarchy in the context of proximity. David Haek-won Kim argues that contemporary racial segregation in the UnitedStates is maintained through a structure of white contempt for racialothers (Kim (1999)). Kim describes this contempt as a sense of offenceat the ‘base and potentially debasing’ qualities of those racialized (111).The response to this threat is the reinforcement of hierarchy and status,and this can take a variety of collective forms. Kim does use spatiallanguage to describe the kinds of status preservation that ward off the‘debasing contagion’ of the contemptible: avoidance, marking a bound-ary, fending off, circumscribing (112). Kim’s work suggests that ourfelt separateness from others should not be taken as a sign that we arenatural egoists who need to move outward to a collective. We may in-stead be expressing our emotional training in collective purposes. Inother words, when we are inclined to describe persons as bounded andat a distance from others, this tendency should direct us to the specificmodes and contexts of socialization that reinforce alienation from, dis-trust of, or contempt for certain others, often those close by. Walker,Smith, and Kim gesture further in the direction that Braybrooke andLloyd take us. Though neither Braybrooke nor Lloyd completely es-capes the traditional focus on collective bonds as sympathetic bonds,together they provide a model of the self that allows for a sophisticatedtreatment of the links between joint interests and emotional develop-ment.

In this chapter, I have strongly affirmed Braybrooke’s concerns aboutdefault egoism in moral theory that tries for a more social starting point.While I endorse Braybrooke’s and Sherman’s rehabilitation of the emo-tions in models of moral motivation, I have questioned whether priv-ileging empathy or fellow feeling is a good way to add needed com-plexity to these models. Braybrooke and Sherman are both indebtedto the history of moral sentiment theories in formulating their accountsof empathy. Though this chapter leaves open the possibility that notall theories of empathy re-inscribe egoism, Braybrooke’s work stimu-lates us to scrutinize these theories for the problematic assumptions that


they may embed. Genevieve Lloyd comments that there is no one rightmodel of the self but that we must work back and forth between ourexperience and our models to find ways of theorizing adequate to thepresent context. I suggest that the empathic egoist is too old–fashioneda self for the challenges that now inform our reflections on moral mo-tivation and that Braybrooke’s writing on moral education describes amore useful alternative.

Chapter 10

The Problem of Moral Judgement

RICHMOND CAMPBELL

Abstract

Is moral judgment a state of belief or a state of feeling and desire? Cogni-tivists about moral judgment answer: true or false belief; non-cognitivistsanswer: feeling and/or desire. What is at stake in this disagreement is thepossibility of moral knowledge, for if moral judgment is not a state oftrue or false belief, it cannot embody moral knowledge. Those who de-fend the possibility of moral knowledge, however, face a dilemma. Theymust reject, on pain of inconsistency, one of two plausible Humean views:(1) that moral judgments themselves contain motivation to act, or (2) thatmotivation cannot arise from belief alone. InNatural Law ModernizedDavid Braybrooke portrays Hume as defending the possibility of moralknowledge, thus apparently committing Hume to an inconsistent triad ofpropositions. This essay defends a novel interpretation of moral judg-ment as a complex, hybrid state of belief and feeling/desire that wouldundermine the long-standing dichotomy between cognitivism and non-cognitivism and demonstrate how Braybrooke’s understanding of Humedoes not lead to inconsistency. In doing so, the essay attempts to bring outthe moral significance of the underlying false opposition between reasonand emotion that has been with us since Plato and continues to bedevilcontemporary moral theory.∗

∗I am grateful for the extensive comments on an earlier draft provided by David Braybrooke,Rochney Jacobsen, Duncan MacIntosh, William Rottschaefer, and Susan Sherwin.

251

252 CHAPTER 10. RICHMOND CAMPBELL

10.1 The Problem

Is a moral judgment a state of belief or a state of feeling and desire?We speak, reason, and feel as if a moral judgment can embody moralknowledge. We thus imply that moral judgment is a state of belief. If,for example, I say, ‘It was wrong of you not to tell her that you are mar-ried,’ and you reply, ‘That’s true’ or ‘I know that,’ you imply agreementin belief. Yet at the same time our judgments convey negative feel-ings about, and even motivation to refrain from doing, what we judgeis morally wrong. Yours convey feelings of regret, the desire not to re-peat the mistake; mine convey feelings of disapproval and the desire topersuade you not to repeat it. Such are part and parcel of the judgmentthat one of us has acted wrongly. In sum, a moral judgment appears tobe belief and desire, belief and feeling, at once. But is it really possiblethat a moral judgment can be belief and desire, or belief and feeling?

Humean considerations argue against this possibility. Famously Humemaintained that morals move us and excite our passions, while judg-ments of fact, in themselves and apart from what we desire and feel,are inert. Put into contemporary terms, the argument derives from thefollowing inconsistency. For the reasons just given, we may be inclinedto hold:

1. Moral judgments are (or express)1 states of belief;

and also

2. In making moral judgments, people feel the force of moral normsthat guide their judgments and are thereby moved (to some de-gree) to act in accord with them, independently of any antecedentdesires or feelings that they may have.

Yet, according to a widely accepted view of motivation,

3. Beliefs never motivate just by themselves; instead, beliefs moveus in combination with antecedent desires and feelings.

1‘Judgment’ can refer to a mental state or an utterance expressing it. This essay focuses onthe first case, but in ‘What Is Moral Judgment?’ I argue that the cases are parallel.

10.1. THE PROBLEM 253

Since these views together lead to contradiction, (2) and (3) providea Humean argument for rejecting (1). If moral judgments are intrinsi-cally motivating [by (2)—see Hume (1978), 413-18] but beliefs are not[by (3)—455-76], then moral judgments cannot be beliefs [contrary to(1)].

The inconsistency, however, also constitutes a problem with the con-cept of moral judgment, since each of the three views has merit on itsown. View (3) is embedded in the dominant and widely shared belief-desire theory of action, according to which both belief and desire areneeded to explain behaviour. I might believe that food is before me butwithout any desire to eat. The fact of my belief by itself will not pre-dict my eating. On the other hand, if I desire food but do not believethat food is before me, the desire by itself will not predict my eating.In general, only the combination of belief and desire can predict andexplain behaviour on this theory. View (2), that moral judgments carrywith them their own motivation, is equally plausible. I might desire todo what is right for extrinsic reasons, say to protect my reputation, butthinking something is right is, it would seem, motivating just in itself.Unfortunately, views (2) and (3) entail that my thinking something isright cannot be a belief that it is right and hence that view (1) must befalse. But to reject (1) is surely contrary to common sense. It appears, insum, that one of these views must be wrong, yet it is hard to understandhow any one of them can be given up.2

In his Natural Law ModernizedDavid Braybrooke proposes to in-terpret David Hume as a natural law theorist and thus as a philosophercommitted to the possibility of moral knowledge (Braybrooke (2001),130-5) and therefore the truth of (1). It is clear that the problem justposed is as pressing for Braybrooke’s Hume as it would be for anyphilosopher who defends the possibility of moral knowledge. Indeed, itis more so, since Hume endorses the very propositions that seem incon-sistent with that possibility! Is Braybrooke’s non-standard interpretationof Hume incoherent in light of this apparent inconsistency?

2 Michael Smith (Smith (1994)) formulates a closely related problem in terms of rationalbelief and motivation. See my ‘What Is Moral Judgment?’Campbell (2005) for how it bears onthis problem.


In this chapter I argue that it is not, since moral judgments may beconstrued as embodying two different functions, as beliefs representingpossible moral truths and as emotion-laden motives to act. This ac-count of moral judgment developed in myIllusions of Paradox(Camp-bell (1998) is close to Braybrooke’s understanding of Hume’s theory.While he concedes that Hume may be read as rejecting the realist propo-sition that moral judgments can embody moral truths, he contends, withsome plausibility, that Hume’s theory of morals is best interpreted asbeing at bottom realist. My aim will not be to defend Braybrooke’sinteresting reading of Hume (though I hope to indicate why it is plausi-ble), but to explore whether a hybrid conception of moral judgment canbe defended on its own merits and to bring out its significance for theunderlying false dichotomy between reason and emotion that has beenwith us since Plato and continues to bedevil contemporary moral theory.

10.2 Moral Realism: The Default Position

I will be assuming that moral realism, the position that moral knowledgeis possible, is the default position in moral theory. Moral realism agreeswith common sense and is backed by powerful theoretical considera-tions that are well developed in the literature (see, e.g., Copp (1995),15-19). For example, moral realism simplifies our understanding oflogical inference in moral reasoning by allowing us to apply the stan-dard of truth preservation for valid inference. It also adds coherenceto our understanding of normative claims in general, such as claimsabout validity of inferences or mistakes in language. We think we knowthat some inferences are valid, that some uses of languages incorrect,despite the evident normative character of these claims. Why shouldnormative claims in morals have a different epistemological status fromnormative claims about logic and meaning? Braybrooke presses sim-ilar questions in his article ‘What Truth Does the Emotive-ImperativeAnswer to the Open-Question Argument Leave to Moral Judgments?’Braybrooke (2003c)

Critics of moral realism attempt to counter these defences and toraise objections of their own. Space does not permit even cursory dis-

10.2. MORAL REALISM: THE DEFAULT POSITION 255

cussion of all the main lines of objection to moral realism, though I havereviewed them elsewhere (Campbell (2003)). Here I confine myself toobjections that bear on the central problem for moral judgment just out-lined. Prominent among these objections is Mackie’s charge that moraljudgment is subject to systematic error, with the result that all moraljudgments are false (Mackie (1977)). Clearly, if Mackie’s error theoryis true, moral realism cannot be. His theory is that moral judgments areabout facts that are intrinsically prescriptive and as such imply what ishighly implausible from a scientific perspective. Moreover, nothing inscience would explain how we could know such facts. For these reasonscommonsense moral thinking is dubious metaphysically and epistemo-logically. Does Mackie’s theory serve to reinforce the doubts just raisedabout the coherence of moral judgment?

The inconsistency among (1), (2), and (3) is more reasonably viewedas being as much a problem for Mackie’s theory as it is for moral real-ism. The reason is just that his theory is committed to the same threeinconsistent propositions. His theory entails that moral judgments arefalse; hence, his theory entails that moral judgments are beliefs ratherthan merely states of desire or feeling. Moreover, his theory does notdeny either that moral judgments are intrinsically moving or that theleading scientific account of motivation is false. Indeed, his theory im-plies that even if no intrinsically prescriptive facts exist, we are movedin judging that they exist and he believes that we should accept whatscience says about motivation. Thus, he is committed to (2) and (3) aswell as (1), at least until science offers a better account of what movesus to act. In sum, he shares with moral realism the central problem withmoral judgment outlined above.

Moral realism has, of course, other ways to address Mackie’s objec-tion. It can, for example, deny that moral realism entails a commitmentto intrinsically prescriptive moral facts. Naturalistic forms of moral re-alism attempt to do precisely that. Nevertheless, even if they can suc-ceed (as I believe they can), the central problem remains a problem formoral realism. Moral realism seems committed to (1), since it is hardto see how there can be moral knowledge without moral belief. It must,therefore, reject either (2) or (3). It must, that is, answer Hume’s argu-


ment from (2) and (3) to the denial of (1). The next two sections willconsider two approaches prominent in the literature on moral realism.

10.3 Externalist Moral Realism

The externalist moral realist rejects (2) in favour of (1) and (3). Whyreject (2)? The standard answer, perhaps best developed by David Brink(Brink (1989)), is that we can easily imagine someone who judges some-thing to be morally wrong but just doesn’t give a damn. Not only canwe imagine such a person, sociopaths exist who apparently agree thatwhat they want to do is wrong but feel no compunction about doingit. Then there are ordinary folks who at times feel no inclination todo something that on balance is right from their point of view, such asrendering a small courtesy. How can moral judgments be intrinsicallymotivating in the face of these cases? The standard reply in defence of(2)—the ‘internalist’ reply—is that the cases are described incorrectly.The person who doesn’t give a damn doesn’t really believe that the actin question is wrong (or right). Instead the person thinks that the actionis ‘wrong’ (or ‘right’) only in the non-literal sense that this is what mostpeople regard as wrong (or right).

Admittedly, it may be difficult or even impossible to settle this dis-pute, since the cases present conflicting evidence about which descrip-tion is correct. The person says one thing but the person’s behaviour andfeeling are reason not to treat what is said as being intended literally.The externalist, however, provides a way out. She can present a positiveaccount of how motivation arises in cases of moral judgment withoutnecessarily conceding that (2) is true. An example from Braybrooke’sdiscussion of natural law theory illustrates how such an account mightunfold. According to this theory, to judge something to be morally un-acceptable is to judge it to be contrary to rules that must be obeyed byand large in order for members of society to thrive (Braybrooke (2001),3). The concept of thriving, let us suppose, can be given empirical con-tent through reference to meeting basic human needs, say accordingto Braybrooke’s theory of needs (Braybrooke (1987a)). Thus, on thisunderstanding of what is meant in judging something morally unac-

10.3. EXTERNALIST MORAL REALISM 257

ceptable, activities like mugging, fraud, and extortion would be judgedmorally unacceptable, since they would violate norms that enable a so-ciety to thrive. Indeed, on the theory of natural law that he defends, weknowthat these kinds of behaviour are wrong. But in knowing this mustwe be motivated not to engage in these kinds of actions?

According to the externalist the motivation and the belief are distinctlogically and in fact but are linked together causally, for contingent rea-sons, for example concerning human nature. Let us suppose with Humethat most of us are moved by considerations that impinge on the well-being of others, especially when our own needs are not so pressing asto prevent us from having a clear view of the needs of others. Suppose,that is, that we have the sentiment of humanity, to use Hume’s phrase. Inthat case, normally we will be moved to some degree by what we judgemorally because of our independent interest in the thriving of people ingeneral. Notice that the reason that we would be moved is that we havean antecedent desire that people thrive and our moral belief, on Bray-brooke’s theory of the content of moral belief, informs us of the meansto achieve that end. No entailment is assumed between having a moralbelief and having a desire to be moral. Moreover, the states in questionare distinct. We care about whether moral rules are obeyed because wecare about the consequences of their being obeyed for the thriving ofsociety. For these reasons we will react negatively when others flout therules and feel badly when, yielding to temptation, we do what we knowto be wrong. The position is, therefore, a moral realist position thatis at the same time externalist regarding moral motivation. It endorsespropositions (1) and (3) but rejects the internalist position contained in(2).

The difficulty with this story, and indeed any account of moral mo-tivation that is inconsistent with (2), is that it does not fit with whatwe know about moral development. As children we learn morals bylearning how to feel about certain ways of acting and how to behavein accordance with those feelings. In other words, we learn how to bemoved by moral considerations directly. We learn to desire not to docertain things and to feel badly when we do them or see others do them,and also to desire to do other things and to feel good about doing those


things. We learn, that is, how to be motivatedintrinsically by moralconsiderations. This fact about moral development is exactly in accordwith (2).

Furthermore, we do not first learn some highly theoretical beliefsabout the relation between, say, obeying rules and thriving and then, be-cause of our antecedent desires and feelings about matters contingentlyconnected with these beliefs, come to be motivated to act and to feel ap-propriately. Not only is such an account of moral development far fromwhat we know through casual observation and scientific study, it is alsocontrary to the way moral sensibility would have evolved according toour best understanding of the evolution of morals. Moral sensibility, inorder to exert immediate and powerful control sufficient to counter con-trary impulses, could not have developed in such an indirect way. Set-ting aside hard cases where moral feelings pull in opposite directions,countless other cases make us react immediately and unreservedly, justas one would expect if (2) were true. In the hard cases, we are unableimmediately to make a moral judgment, but when we are able to makethe moral judgment, we normally are disposed to feel, desire, and actin characteristic ways just in virtue of making that judgment. An ade-quate view of the relation between moral judgment and motivation mustaccount for this fact.

10.4 Internalist Moral Realism

The defender of (1) and (2) must reject (3). The difficulties facing thosewho take this option are equally grave. One that has been mentionedalready is that (3) is part of a well-established theory of motivation ac-cording to which beliefs motivate only when coupled with desires. AsMcNaughton puts it: ‘Desires without beliefs are blind; beliefs with-out desires are inert’ (McNaughton (1988), 21). Many would say thatthis understanding is part of our folk psychological theory of why wedo anything. It is this theory that is reflected also in Bayesian modelsof rational decision-making. There subjective probabilities of possibleoutcomes stand in for degrees of belief that those outcomes will resultif a given decision is taken and individual utilities of outcomes stand

10.4. INTERNALIST MORAL REALISM 259

for the degree of preference for those outcomes relative to alternatives.Although there is no consensus as to the best theory of motivation, noother approach is so well entrenched in ordinary thinking or in effortsto represent motivation mathematically. To reject (3) is to fly in the faceof this formidable tradition.

An even more serious difficulty arises from elements in the internal-ist realism that appear to be incompatible. Consider the precise sensein which the internalist realism, in accord with (1), implies that a moraljudgment is a state of belief. A moral judgment is to be a state of beliefas distinct from a state of emotion, desire, or attitude. Though the lattercan have a cognitive element (the object of the emotion can engage ourcognitive faculties), it is only belief that can be fully cognitive in all es-sential respects. This position is often characterized by our saying thatthe moral judgment or claim is ‘purely cognitive.’ Consider the follow-ing passage from McNaughton, a particularly lucid defender of internalmoral realism (see also Dancy (1993); Platts (1988)):

To be aware of a moral requirement is, according to the re-alist, to have a conception of the situation as demanding aresponse ... The requirement will only be satisfied if theagent changes the world to fit it. But the realist also wishesto insist that the agent’s conception of the situation ispurelycognitive[my italics]. That is, the agent has a belief that heis morally required to act. (McNaughton (1988), 109)

But now we must ask ourselves how it is possible for a state of mind tobe purely cognitive and at the same time be internally connected to feel-ing and desire as required by (2), since internal moral realism endorsesboth (1) and (2). I want to urge that it is not possible.

Notice that the internalism expressed in (2) does not entail that amoral judgment is a belief. In fact, expressivists deny (1), but becausethey hold that moral judgments are or express states of emotion, desire,or attitude, they are able to endorse (2). The reason is that by denyingthat moral judgments are beliefs, expressionists can have an unmediatedconnection between judgment and motivation. (If a moral judgment is,say, a state of desire, then that desire is ‘internal’ to the judgment just as


a matter of logic.) Internalist moral realism, however, is not in the sameposition. It cannot argue that the connection between judgment andmotivation is internal in the same way, for this form of moral realismrequires, as we have just seen, that moral judgment is a purely cognitivestate of mind and thus is not in any part a state of emotion, desire, orattitude. The state of belief that constitutes the moral judgment must beinternally connected to emotion, desire, or attitude without being itselfan emotion, desire, or attitude.

McNaughton thinks that the internal connection can be made intelli-gible if we understand that a belief about an object of desire can explainwhy the object is desired. He doesn’t think, for example, that a saucerof mud is an intelligible object of desire, given ordinary beliefs aboutmud. Those beliefs would not explain why anyone would desire mud.By contrast, doing what is believed to be morally required is an intelli-gible object of desire, just in virtue of what we believe about it. Whatwe believe about it, that it is morally required, explains why we desireit. In these cases the desire that arises from the belief becomes cognitiveitself and is thus dependent on the belief. McNaughton puts it thus:

The radical upshot of this line of thought is that ... the de-sire, as an independent element in action explanation, dropsout as redundant. Once we have understood the attractive-ness of the desired object, as the agent conceives it, we al-ready understand why he is motivated to act the way he does.Nothing needs to be added to his conception to complete theexplanation. (113)

Since the desire is not an independent element in the explanation ofmotivation and is redundant once the belief is fully understood, it ispossible, McNaughton thinks, to endorse (1) in the required sense—to understand moral judgment as purely belief - and at the same timeendorse (2).

This attempt to explain how desire can be internal to a belief statethat is purely cognitive and hence be itself cognitive trades on an ambi-guity. What makes the desire ‘cognitive’ on this account is the fact thatthe object of desire can also be an object of belief. For example, that I

10.5. THE BELIEF-DESIRE THEORY OF MORAL JUDGMENT 261

am helping someone in need can be both something I believe that I amdoing and something that I desire that I am doing. Moreover, I can de-sire to do what I am doing because it is truly a case of helping someonein need. I have no quarrel with this conception of the cognitive natureof desire. It is another matter, however, to claim that the desire ispurelycognitive. What would that mean? It could be purely cognitive if the de-sire is presentsolelybecause of what is believed, independently of anyother desires or feelings, but this understanding of desire is exceedinglystrange. It might be urged that the element of desiring (distinct fromthe object of desire) isitself cognitive, but it is entirely unclear what itwould mean for desiring to be purely a state of cognition. Ostensibly,the desiring taken as a whole is not just a matter of thinking of the ob-ject of the desire according to a certain belief. It is the distinct state ofdesiringthis object. Were that not so, something would be necessarilydesired provided only that one has certain beliefs about it. It would beas if certain facts about a thing that do not logically entail its being de-sired necessitated its being desired just in virtue of their being believed.Such facts would be akin to Mackie’s intrinsically prescriptive facts andno less mysterious.

10.5 The Belief-Desire Theory of Moral Judgment

At this point we need to take stock. The problem is framed between twomajor alternative views: that moral judgment is a belief and that moraljudgment is something else, such as a desire, emotion, or attitude. Thesealternatives may appear not only to be exclusive but to be exhaustive.That is, they may appear to be contradictories rather than simply con-traries. When judgment is understood linguistically as an utterance ora written statement rather than as a psychological state, the correspond-ing views are called ‘cognitivism’ and ‘non-cognitivism’ respectively(Brandt (1959), 205; Sayre-McCord (1988), 1-23). Thus, cognitivismsays that moral claims are true or false and non-cognitivism denies thatmoral claims are either true or false. Of course, to say that someone’smoral claim is true or false is just to say that it expresses a true or falsebelief. Cognitivism and non-cognitivism are, therefore, nothing more


than the semantic forms of the dichotomy that generates the problem,and we can easily reformulate the inconsistent triad of positions in se-mantic terms (see my ‘What Is Moral Judgment?’). Interestingly, thelabels used for the semantic distinction imply that cognitivism and non-cognitivism are to be understood as contradictories. We need only addthat the terms ‘internalism’ and ‘externalism’ applied to forms of real-ism are meant to be contradictories too, so that there is no in-betweenposition possible. Once we have set aside error theory, only three pos-sibilities appear to remain: externalist moral realism, internalist moralrealism, and non-cognitivism. Moreover, each view faces serious objec-tions given the inconsistent triad, since each view must reject a memberof the triad despite the independent support enjoyed by each member.

What is at fault, I now want to argue, is not any of the claims com-prising the triad but the underlying framework used to generate the prob-lem. In particular, I contend that contrary to appearances and contraryto the assumption made throughout a century of meta-ethical discus-sion, the view that moral judgments are or express states of belief (cog-nitivism) and the view that moral judgments are or express states offeeling, desire, and attitude (non-cognitivism) are either not contrariesor else not contradictories. On either alternative the problem of moraljudgment dissolves and room is left for a more adequate understandingof the complexity of moral judgment. Let me explain.

Once the debate around how to resolve the moral problem movesbeyond simple statements of the competing alternatives, it turns outthat positions corresponding to (1) invariably are interpreted to entailthat moral judgments are or expressonly beliefs. Positions correspond-ing to the denial of (1), correspondingly, are invariably interpreted sothat they entail that moral judgments are or expressonly feeling, de-sire, or attitude (not belief). The alert reader will already have notedthis phenomenon in the above exposition of the reasons for and againstthe main forms of moral realism. Of course, with the key qualification‘only’ made explicit, it is abundantly clear that the positions are con-traries and not contradictories. Even though they are officially definedwithout this qualification so that the truth of one entails the denial ofthe other, with the qualification the situation is logically different. Al-

10.5. THE BELIEF-DESIRE THEORY OF MORAL JUDGMENT 263

though the positions so qualified are inconsistent, they do not exhaustall the possibilities. Might it be possible that a moral judgment is (orexpresses) not only belief butalso feeling, desire, or attitude? This isa possibility that I will explore directly. Note what happens, though,if the qualification is dropped so that (1) is read just as stated and itsso-called denial read as the view that a moral judgment is or expressesfeeling, desire, or attitude. Then the positions are not even contraries!Either way, new possibilities emerge. No longer are we forced to chooseamong unpalatable alternatives.

What I propose is that we understand a moral judgment as a com-plex state (or mode of expression) comprising (or expressing) both be-lief and desire (or feeling or attitude). I should emphasize that what I amsuggesting is not a variation of expressivism, where moral judgment isfundamentally a non-belief state that has various cognitive aspects, forexample, a commitment to give a rational justification for that state. Ex-pressivism is still, in traditional terminology, a form of non-cognitivism,since moral judgments in this view are not states of belief (Gibbard(1990); Blackburn (1998)). The same can be said for the position thatmoral judgments are prescriptions rather than true or false statements(Hare (1952, 1989)). The view that I am proposing, however, is neithercognitive nor non-cognitive as traditionally conceived. My position isthat a moral judgment is a fully hybrid state of mind (or mode of ex-pression) that functions normally both as a belief and as a non-beliefstate, such as a non-belief state of feeling, desire, or attitude. Take thejudgment that stealing is wrong. Traditionally, a non-cognitivist (e.g.,Ayer (1936)) takes the judgment to be a combination of a cognitive rep-resentation of stealing coupled with a non-cognitive emotional reactionto what is represented. Though this understanding of moral judgment ina weak sense combines both cognitive and non-cognitive elements, thepresent suggestion is fundamentally different. The cognitive element isnot the representation of stealing per se, but of stealing as being morallywrong. The judgment that stealing is wrong contains the beliefthatstealing is wrong. It follows that the position is not non-cognitivist. Onthe other hand, it is not cognitivist, since it holds that the judgment thatstealing is wrong is at the same time normally a state of feeling and mo-


tivation. In making the judgment that stealing is wrong, I am disposed(normally, to some degree) not to steal, to deter others from stealing, tofeel shame, embarrassment, and regret if I steal and indignation towardsothers if they steal, and to defend these feelings and dispositions as be-ing justified. All these elements are combined in the moral judgment, atleast in a normal case.

The position advocated here allows an important qualification. Thesehybrid elements can pull apart in certain cases. Suppose I am raised tothink that homosexuality is wrong. When I am mature, my attitudechanges; I no longer believe that it is wrong. Yet before I am fully com-fortable with this new moral perspective, I may have many of the sameemotional responses that I had before I changed my belief. In short, Imay believe one thing and feel another. On the theory I advocate, thisis not a paradigmatic case of moral judgment. My judgment that ho-mosexuality is not wrong does not contain all the elements that such ajudgment normally would. By the same token, it could be said that I amstill judging it wrong on an emotional level, and that would be true, eventhough the moral judgment in this case too does not contain all the nor-mal elements. Here it is lacking the appropriate belief. In these respectsthe theory is like externalist moral realism because it draws no essen-tial connection between moral belief and moral motivation. However, itis also like internalist moral realism because in the normal case, wherebelief and desire, feeling and belief are present in the moral judgmenttaken as a whole, there is a direct and unmediated connection betweenthe moral judgment and the feeling and motivation that goes with it.When I make a moral judgment, I am normally motivated to act andmoved to feel just in virtue of making that judgment, independently ofany antecedent feelings or desires. The theory offered breaks down theexternalist/internalist dichotomy. It is easy to see that if this theory iseven roughly on the mark, the problem of moral judgment dissolves.How it dissolves depends on how view (1) is interpreted. Without thequalification that makes (1) imply that moral judgment is only a stateof belief, there is no incompatibility between (1) and (2) even whenthey are combined with (3). For a moral judgment in being not merelya belief can easily accommodate (2). Moreover, the implication of (3)

10.6. TWO OBJECTIONS BRIEFLY CONSIDERED 265

that beliefs by themselves are inert creates no difficulty. Though beliefsare inert by themselves according to (3), moral judgments are not nor-mally just beliefs. As well, the objections raised against external moralrealism and internalist moral realism do not apply, since in each casethose objections presuppose the qualification. On the other hand, if thequalification is included and made explicit, then it becomes clear that(1) cannot be true if the hybrid theory is true. In this case (1) can berejected without sacrificing moral realism.

10.6 Two Objections Briefly Considered

One objection arises from a distinction made by Anscombe between twoopposite ‘directions of fit’ ascribed to beliefs and desires (Anscombe(1963)). Anscombe compares a grocery list with a list made by a detec-tive reporting groceries bought (56). The same list can be the content ofa belief or a desire (that certain groceries are bought). It functions dif-ferently depending on whether it represents items desired to be boughtor items believed to be bought. In the first case, what is bought shouldfit the items desired to be bought; in the second case, what is believedto be bought should fit what actually was bought. In short, the worldshould fit the desire and the belief should fit the world. The directionsof fit are precisely the opposite, and for this reason when the list fails tofit what is bought, we blame the detective’s list in one case (belief) andthe grocery store in the other (desire).

What might be thought to be objectionable about the hybrid view isthat it requires that a single state of mind comprising both belief anddesire have opposite directions of fit. This worry has been raised inthe literature for internalist moral realism because on this view a moraljudgment is nothing more than a belief, yet it is supposed to function asa desire as well. Whatever its merits as an objection to that account, ithas no force against the hybrid theory. Let us grant for the sake of ar-gument that beliefs and desires do have opposite directions of fit. Whatis missing in the objection is any reason for thinking that because thedirections of fit are different, a state of mind, especially a complex one,cannot function both ways at once. Consider intention. Arguably, if I in-


tend to meet you tomorrow for lunch, I both believe that I will meet youfor lunch and desire that I will (Millikan (1996)). If I don’t meet you,say because I am unexpectedly detained, then there are two separatefailures, one of belief and another of desire, with different implications.The intention, however, was not impossible. It would be especially oddto think the difference in directions of fit is a problem for moral judg-ment, since what is desired on the hybrid account (that we do not steal)is not the same as what is believed (that stealing is wrong).

Another worry is that I do not specify the content of moral belief.Notoriously it is difficult to specify the content of moral belief in a waythat will win wide support. Of course, any adequate defence of moralrealism needs such an account. But the point of the present theory isto offer an alternative framework that resolves the problem of moraljudgment. Admittedly, the alternative leaves other matters concerningmoral realism exactly where they were. The hybrid theory is offeredas a better analysis of moral judgment. It can be right in this respecteven if moral realism is false. Indeed, the present suggestion actually iscompatible with error theory, since the hybrid theory is not committed tothe position that moral beliefs are true. It would be unfair to saddle thistheory with the task of fully defending moral realism. It does, however,provide an advance over the standard internalist and externalist formsof moral realism simply because it avoids the inconsistency with whichwe began. In the next section, though, I will suggest how the content ofmoral beliefs might be specified in a Humean form of moral realism.

10.7 Hume and the Opposition between Reason andEmotion

We return finally to the interpretation of Hume as a natural law theoristand hence a moral realist. Can Braybrooke’s interpretation be recon-ciled with Hume’s opinion that morals are based on sentiment ratherthan reason? Or with Hume’s claim in theTreatisethat to judge charac-ter as virtuous is to have a particular feeling of pleasure that ‘constitutesour praise and admiration’ (Hume 1978, 471)? Given his assertions,Hume may appear to be more naturally read as a non-cognitivist who

10.7. THE OPPOSITION BETWEEN REASON AND EMOTION 267

identifies moral judgment with feeling and desire. In light of the belief-desire theory, however, it may pay us to take a second look.

Consider first the question whether the belief-desire theory of moraljudgment can be squared with Hume’s understanding of moral percep-tion and judgment as involving sentiment. We have framed the formermainly in terms of belief and desire rather than belief and feeling, but itshould be evident that nothing in the structure of the theory proposed orin the arguments supporting it should stop us from incorporating feelingand emotion of specific kinds into the dual conception of moral judg-ment. If I condemn the gratuitous, senseless cruelty of some action, Inot only want it to stop and not to be repeated, but also feel displeasure,perhaps revulsion and disgust, in thinking about it, and my emotionalreaction arguably is as inseparable from condemnation as my desire toprevent the action. It is the same when I condemn my own action andfeel shame and remorse for what I have done. These feelings are asmuch a part of my condemnation as my desire never to repeat it. Wecan of course imagine special cases where we make the negative judg-ment but feel little or nothing, just as we have imagined cases where thedesire is absent. In the normal case, though, feeling and desire togethergo with the judgment.

What about moral belief? The argument can be made that feelingdoes not exclude moral belief as part of the judgment any more thandesire does. The hybrid theory accommodates both belief and the com-bination of feeling and desire. In this respect Hume’s recognition ofsentiment as part of moral perception is compatible with the alterna-tive approach to moral judgment that I am advocating. Hume did arguethat, since morals excite us to feel and move us to act, morals are notbased on reason, which tells us only what is true or false. But this ar-gument is just the argument, in slightly different words, from (2) and(3) to the rejection of (1). We have already seen that this argument issubject to more than one interpretation. If the rejection of (1) is prop-erly interpreted as rejecting the extreme position that moral judgment isonly belief, then the argument is sound but does not establish that moraljudgment does not contain belief in addition to desire and feeling. Inshort, Hume’s argument can be taken to show that morals are not based


solely on reason, not that reason plays no part in the content of moraljudgment or in its appraisal. It is true that Hume doesn’t explicitly for-mulate his position in these terms, but it would be uncharitable to saddlehim with the logically stronger, less reasonable position since he doesnot explicitly exclude it either.

We should examine, in any case, Hume’s understanding of moraljudgment when it is couched in the context of his more general the-ory of morals. Two parts of that theory stand out as directly relevant.First, Hume holds that the feelings that are internal to moral judgmentare feelings that arise when contemplating the object of those feelingsimpartially and therefore are separate from feelings aroused by self-interest. In a familiar passage Hume writes:

Nor is every sentiment of pleasure or pain, which arises fromcharacters and actions, of that peculiar kind which makes uspraise or condemn. The good qualities of an enemy are hurt-ful to us, but may still command our esteem and respect. ‘Tisonly when a character is considered in general, without ref-erence to our particular interest, that it causes such a feelingor sentiment as denominates it morally good or evil. (Hume(1978), 472)3

The complementary part is Hume’s understanding of human naturesuch that when we do contemplate aspects of character, bracketing ourown interests, we are in fact moved with pleasure in contemplating thoseaspects that are useful to persons (either to themselves or others, ei-ther directly or indirectly) and we are moved with displeasure by theopposite. When the first part, a key ingredient of his theory of moraljudgment, is put together with the second part, his idea of a sentimentof humanity that is generally shared, the result is a broadly utilitarianaccount of moral virtue and vice. Aspects of character that are usefulare judged properly to be virtuous and those that are hurtful are judgedproperly to be vicious.

3 See also Hume’s appendix on moral sentiment inAn Enquiry Concerning the Principles ofMorals (Hume (1966)).


Braybrooke would question the description of Hume’s view as util-itarian, since Hume’s focus is not on maximizing utility narrowly de-fined, but the issue of what to call Hume’s view need not detain us. Theunderstanding of Hume I am proposing is about the status of moral judg-ment rather than the details of his normative theory. It does not dependat all on how we interpret usefulness as applied to character. All that isessential to assume is that the usefulness of a character is real enoughthat almost everyone will respond to it positively when they consider itimpartially and the opposite when a character is hurtful. For the sake ofargument, then, let us grant this Humean assumption. What follows forthe status of moral judgment? Can it embody moral knowledge?

My argument in support of Braybrooke’s reading of Hume has threesteps. The first is to note that Hume’s understanding of moral judgment,that it is a feeling of a certain kind, obviously allows the feeling to besubject to rational appraisal. This can happen in two general ways, asHume makes clear. Someone can make the case that the person makingthe judgment is not viewing the subject matter impartially. If the feelingof pleasure or dissatisfaction arises from self-interest rather than fromtaking a general perspective, then the feeling is not justified as a moraljudgment. Whether self-interest explains the feeling is, moreover, a fac-tual matter that can be debated on rational grounds. Second, the feelingcan be faulted because it is based on a false view of the facts. The feel-ing of displeasure can arise from believing that a certain person is inpain when he is not. Again, the validity of the feeling would be under-mined based on a matter to which reason can be applied. No defenderof non-cognitivism should want to deny any of this, since it is all com-monplace and doesn’t by itself entail that the feeling embodies moralknowledge.

The second step, however, takes us significantly further along theroad to moral knowledge. Suppose our moral judger is not making anymistake about the facts (facts that anyone would concede are facts) andis thinking of the subject matter in a frame of mind that allows his feel-ings to respond to what is judged apart from its bearing on self-interest.Then, given the broad Humean assumption one paragraph back that weare accepting for the sake of argument, there exist facts that account


for the feeling of pleasure or displeasure, and this feeling is the moraljudgment. Although this implication does not obviously reduce to say-ing that the judger is having feelings aboutmoral facts, it isn’t clear whywe should resist this conclusion if we take Hume seriously. The feelingsare moral feelings on his view; moreover, they are feelings that wouldarise in exactly the same way for virtually anyone who is not mistakenabout what these facts are and views them apart from their bearing onself-interest. Why not say, then, that the feelings are about moral facts?

The answer that the non-cognitivist will urge is that the facts are notreally moral facts but quite ordinary facts. We have a situation in whichsomeone has (by Hume’s theory) moral feelings about non-moral facts.This response is, of course, classical non-cognitivism, but we don’t haveto read this situation that way. To insist that we do would be question-begging. One alternative is familiar from the literature. It is to treatHume as an ideal observer theorist who holds that a moral fact aboutvirtue (say) is a fact about what someone would feel who is in an idealposition to observe the subject matter (has an accurate view of the gar-den variety facts, is thinking about them impartially, is normal, and soon). Thus, Henry Aiken writes: ‘Hume holds that anymorally goodactis one which,in the last analysis, would be approved by an impartialspectator as useful or agreeable to ourselves or others’ (Aiken (1948),xxxviii). This alternative has familiar problems, for example, that thejudgment is not directly about the object judged but about someone’sreactions to it, even though the moral judgment appears to be directlyabout the object judged. A better alternative is to say just that the com-monplace fact cited earlier about the usefulness of a certain character isa moral fact; and it is a moral fact because it explains what is for Humealmost a universal feeling of pleasure when this fact about usefulness isconsidered impartially. One could say without too much exaggerationthat the moral feelings experienced represent the moral fact to whichthey are a natural response, at least if one accepts Hume’s understand-ing of moral judgment and human nature.

We are now in a position to take the third step. When we feel thepleasure in this way we also normally believe (rightly or wrongly) thatthe feeling experienced is justified or warranted by the subject matter.


We can be mistaken, because we have misunderstood the facts (in aperfectly ordinary sense) or are thinking about what we judge in a biasedway. Suppose, however, that we are not mistaken in these ways. Thenthe belief we have when we experience these feelings is factually true.That is, the belief that we are not in error in these ways is true. Ourjudgment thus has two elements, one of belief and another of feeling.We can recognize additional ingredients on the side of motivation, suchas the desire to encourage what we find pleasurable. Why not also allowon the belief side the further belief that the useful character that wemorally admire is a moral character, that is, is truly virtuous as a matterof moral fact?

Hume does not go quite this far, at least not explicitly, but the impli-cation is consistent with the rest of his theory. Morals are still based onsentiment, not on reason in the manner of theorists who would nowa-days be called ‘non-naturalist.’ Reason alone tells us nothing, but reasonis not completely absent either. Moral judgment springs from reasonand sentiment; it begins with and is ultimately based on sentiment. Itis, after all, through justified feeling that we know what is virtuous andwhat is vicious. Why isn’t the knowledge embodied in our justifiedfeelings moral knowledge?

If this interpretation of Hume counts as moral realism, it is a natu-ralistic form of moral realism, far from the examples of moral realismthat place emphasis on moral belief in contrast to moral feeling and thatlocate moral facts in a non-natural realm. It is, nevertheless, moral real-ism in precisely the sense that it entails moral knowledge of moral facts.Though I have not given a full-dress defence of this interpretation ofHume, the foregoing suggests how we can see Hume as a moral realistthrough the lens of a new conception of moral judgment. Hume appearsthrough this lens neither cognitivist nor non-cognitivist—a surprisingconclusion if these terms are contradictories. I have argued, though,that they are at best only contraries. Thinking otherwise has confusedus not only regarding the relation between moral belief and moral desirebut also about the possibility of moral knowledge based on moral emo-tion. The age-old dichotomy between reason and emotion dies hard, butwe need to move beyond it if we are to understand the complex nature


of moral judgment and moral knowledge.

Chapter 11

Moral Claims and EpistemicContexts

M ICHAEL HYMERS

Abstract

In ‘What Truth Does the Emotive-Imperative Answer to the Open-QuestionArgument Leave to Moral Judgments?’ David Braybrooke claims that thejustification of a moral claim is independent of the justification of moral-ity generally—that ethical justification does not have to be traced back tometa-ethical justification. I support this claim by appealing to a contex-tualist theory of epistemic justification. Drawing on the work of MichaelWilliams and Robert Brandom, I contend, first, that every claim is justi-fied by default and requires articulated reasons only when it is challenged.Not every challenge is a reasonable challenge—only those that share theburden of proof are. Second, I hold that sceptics about moral truth andjustification, such as J.L. Mackie, are committed to a substantive philo-sophical position, closely linked to foundationalism, that Williams hascalled ‘epistemological realism,’ insofar as they hold that moral claimsareintrinsically less certain than claims about non-moral facts on whichmoral facts might be taken to supervene. But there is no reason to be-lieve that any propositions are intrinsically more certain than any others.Certainty is a function of epistemic context, not semantic or empiricalcontent. Therefore, although moral claims may presuppose the truth ofsome meta-ethical claims, the justification of moral claims does not de-pend on the justification of meta-ethical claims.∗

∗In writing this essay, I have benefited greatly from conversations with and presentations by

273

274 CHAPTER 11. MICHAEL HYMERS

In ‘What Truth Does the Emotive-Imperative Answer to the Open-QuestionArgument Leave to Moral Judgments?’ Braybrooke (2003c), DavidBraybrooke arrives at a conclusion that may provoke some disquietamong those with strongly objectivist intuitions about moral judgments:

We have, in this conception of moral realism, as explainedmeta-ethically, ‘true’ used within one structure of ideas thatfigures among many logically possible structures, in this case,a structure of ideas ultimately invoking thriving, personaland social, and empirical evidence about thriving. It is aloose enough sense of thriving to accommodate some mi-nor variations in conceptions of thriving. At the same time,‘true’ is used unselfconsciously, without thinking of meta-ethics, within moral discourse itself with no apprehension,and no compelling reason to apprehend, that the basis forusing it is liable to being supplanted. Just so we can habitu-ally rely on one calendar, knowing that it is just one amongmany possible calendars, many of which have bizarre fea-tures, and none of which may be accurate astronomically.(350)

The objectivists whom I have in mind will regard Braybrooke’s use ofthe distinction between ethics and meta-ethics as artificial and illegiti-mate, contending that applications of the truth-predicate to moral claimsare justifiable only if meta-ethical claims about basic moral principlesand objective grounds of value are also justifiable. Such objectivist im-pulses will drive these philosophers either to become Platonists aboutvalue or to become subjectivists about value (because they concludethat the conditions necessary for true moral judgments are absent).

I shall argue here that the objectivist impulse that leads to theseresults arises, at least in part, from a commitment to what MichaelWilliams has called ‘epistemological realism’ (Williams (1996), 89-134; Williams (2001), 84, 193), the doctrine that there are natural rela-tions of epistemic priority among our beliefs and claims. Rejecting this

Mason Cash, Richmond Campbell, and Jenna Woodrow. Thanks also to Tom Vinci, Nathan Brett,and David Braybrooke for their comments on an earlier draft of this essay.

11.1. MOORE’S OPEN QUESTION 275

doctrine in favour of a variety of epistemic contextualism, I shall arguethat Braybrooke’s use of the ethics/meta-ethics distinction is reasonable,comparing his moral realism to a kind of conventionalism about mathe-matical truth that sees realism as lying in our ‘deep need’ (Wittgenstein(1978), 65) for conventions.

11.1 Moore’s Open Question, Expressivism, and Truth inEthics

Braybrooke begins his discussion with the observation that expressivisttreatments of moral language can be regarded as a response to G.E.Moore’s ‘open question’ argument. Moore contends that any attempt todefine ‘good’ is destined for failure because, for any candidate defini-tion, it will always make sense to say of a given action or outcome thatit conforms to the definition but is not good. Suppose we define ‘good’as utility, for example. A critic of utilitarianism can always intelligiblycontend that a given outcome has a higher utility than its rivals withoutbeing good.

However, if ‘the expression of emotions and imperatives’ is ‘themain thing about moral judgments, perhaps the only thing’ (Braybrooke(2003c), 343), then there is clearly no danger of dogmatically closingoff debate by insisting that good is pleasure or conduciveness to a thriv-ing society or the display of respect for humanity, whether in one’s ownperson or that of another. Insofar as moral ‘judgments’ are merely ex-pressions of emotion or imperatives to act or respond emotionally in aparticular way, there is a sense in which the question of what is goodis left wide open, for no definition of ‘good’ is given by such a theory.‘The emotive and imperative aspects of your [moral] judgment ... willalways fight free of being tied down,’ as Braybrooke puts it, ‘to one ofthe exemplary uses or definitions that you offer’ (342).

The emotive and imperative character of moral judgments is not byitself cause for any alarm, Braybrooke observes, because any reasonablestory about moral language must account for the motivational force ofmoral judgments, and expressivism does just this. ‘Moral judgmentshave to be moving, at least in suitable contexts and for a suitable au-


dience, and both emotions and imperatives are essential to their beingso’ (343). This much, indeed, remains compatible with giving an ac-count of which moral judgments are justifiable or right, for althoughthe expressivist holds that moral judgments are not truth-apt, she mayyet agree without contradiction that the attitudes expressed or impera-tives issued can be measured against some standard of evaluation. Forexample, some attitudes and imperatives may be held justified becausethey ‘promote the thriving of a society’ (343) and others unjustifiablebecause they detract from that thriving.

So, at a glance, it seems that there is no role for the concept of truthin moral theory. ‘Ethics can ... work, both at the level of immediatemoral judgments and at the deeper level on which these judgments arejustified, without invoking the concept of truth’ (344). However, mattersare more complicated than this in Braybrooke’s view. He argues thatattempts ‘to dismiss the concept of truth from ethics’ (344) lead theirproponents into two kinds of awkwardness.

First, there is the awkwardness involved in the fact that truth is theconcept to which we appeal when we are engaged in any sort of ‘se-rious endorsement’ (344), whether such endorsement should arise inthe context of particle physics, of history, of small talk about middle-sized dry goods, or of big talk about ethics. It is, says Braybrooke,a feature of ‘ordinary moral discourse’ (345) that we judge to be truethose moral claims to which we commit ourselves. And so, devia-tions from such ordinary practice demand some extraordinary reason.Such an extraordinary reason is not provided by expressivist accountsof moral discourse, such as emotivism and prescriptivism. These pos-itive views concerning moral language arose in response to prior cri-tiques of the truth-evaluability of moral claims and do not themselvesprovide any such critique. Indeed, Braybrooke argues, a moral realist—someone who holds minimally that moral claims are truth-evaluable andthat some of them are true—can happily endorse much of what the ex-pressivist wants to say about the pragmatic force of indicative sentencescontaining morally evaluative terms without supposing that the mean-ing of such terms as ‘good’ has thereby been settled. What is needed issimply a plausible story about how to distinguish ‘the distinctive crite-

11.1. MOORE’S OPEN QUESTION 277

ria for endorsing’ moral judgments from ‘the grounds for endorsing ormaking scientific propositions or common sense reports of perception’(345). If ‘truth’ is a term of serious endorsement, then we are entitledto apply it to moral judgments so long as we try to ‘work out a generaltheory of warranting judgments, in which warranting moral judgmentswould proceed in some respects differently from warranting reports ofperception, and warranting aesthetic judgments more differently still’(345). We might, for example, say that just those moral judgments arewarranted that are certified by ‘justified moral rules, most plausibly,rules that promote the thriving of societies and the people who belongto them’ (345).

Second, says Braybrooke, when we assert that particular moral rulesare justified, we seem clearly to be committing ourselves to the truth ofsuch statements. He summarizes the point:

A rule cited to warrant a moral judgment praising an honestaction is endorsed as according with the appropriate groundsor criteria for adopting such a rule, the grounds or criteriafor justifying it. Here the place of truth is secure, whether ornot what is justified by the justification are themselves truth-bearers, as in some case, they clearly are not, as long as whatare to be taken as appropriate grounds or criteria have beenidentified, even if only by fiat, and accepted. (345)

So, even if we were to retreat and allow the expressivist point that moralevaluations are not genuine assertions, but expressions of (dis)approvalor exhortations to act or refrain, and so not truth-value candidates,1 itwould not follow that such expressions cannot be candidates for justifi-cation, nor that there is no truth of the matter concerning whether or notthey are justified.

Now it may seem that in trying to hang on to truth in ethics Bray-brooke runs afoul of the very open-question argument that he wanted to

1As Rockney Jacobsen in Jacobsen (1997) has argued, this inference is suspect. The meaningof a sentence is unaltered by the pragmatic force with which it used on a given occasion, but ifmeaning and truth-conditions are inter-related, then it is not clear that a change in pragmatic forcefrom assertion to exhortation should have any effect on whether or what truth-value a sentenceexpresses.


treat expressivism as a response to, because his appeal to the certifica-tion of moral judgments by ‘rules that promote the thriving of societiesand the people who belong to them’ (345) looks suspiciously like a def-inition of ‘good.’ Does such a definition not dogmatically forestall de-bate on goodness in just the way Moore argued was unreasonable? No,says Braybrooke, because Moore failed to distinguish between ethicsand meta-ethics (347)—between the making of moral judgments andthe discussion or consideration of them without endorsing them—andone can quite coherently express one’s commitment to a particular def-inition of ‘good’ in one’s ethical judgments, while still acknowledgingat the meta-ethical level that such a definition is just one among manypossible definitions (347-50).

There may, for example, be variation among people concerning ‘whatthriving amounts to’ (348), even though most agree that conducivenessto thriving is a key criterion for the justification of moral claims. Or,indeed, there may be disagreement about whether the notion of thriv-ing has any role to play in defining ‘good.’ But neither of these facts,thinks Braybrooke, need impair ‘our commitment to the specific crite-ria on which we wish to take a stand’ (349). The long course of humanexperience, he seems to want to say, makes certain criteria of thriv-ing ‘uniquely compelling’ (349) grounds for the justification of moraljudgments. And so, Braybrooke concludes, we can endorse the truthof moral claims, even while allowing that the principles that may be in-voked to justify them are contingent principles that we might never haveadopted—that there are, indeed, other principles that reasonable peoplemight well prefer.

But it is at this point that the objectivists I alluded to in my open-ing remarks will want to interject with the following dilemma: eitherthe uniquely compelling nature of Braybrooke’s principles of thrivingconsists in our believing them to be the correct definition of ‘good,’ orwe cannot possibly be moral realists, for the kind of warrant requiredif moral judgments are to count as true must derive from the truth ofBraybrooke’s principles of thriving. In short, the distinction betweenethics and meta-ethics that Braybrooke relies on is illegitimate becauseit masks the fact that one can get moral truth only by denying that the

11.2. DEFLATIONISM 279

question of how to define ‘good’ is really open.

11.2 Deflationism

As I remarked earlier, I plan to defend Braybrooke’s reliance on theethics/meta-ethics distinction by appealing to a contextualist account ofepistemic justification and to an analogy with Wittgenstein’s conven-tionalist treatment of mathematics. I want to begin, however, with someremarks about the notion of truth most suitable to Braybrooke’s moralrealism.

Proponents of various deflationary accounts of truth may well giveBraybrooke a commendatory pat on the back for his proposal that wepreserve the place of truth in ethics by ‘work[ing] out a general theoryof warranting judgments, in which warranting moral judgments wouldproceed in some respects differently from warranting reports of percep-tion, and warranting aesthetic judgments more differently still’ (345).This is because a recurrent—though perhaps not universal—feature ofsuch accounts is their acknowledgment that the application of the truth-predicate is (a) something that can be appropriately done to any well-formed declarative sentence of natural language that can play a rolein inferences and be a patient of the standard logical operations, and(b) a distinctive speech act whose pragmatic force consists in the ‘se-rious endorsement,’ to use Braybrooke’s words, of the claim at hand.2

However, Braybrooke sees in such intellectual chumminess a disguisedcriticism—namely, that truth has no interesting role to play in ethics be-cause the application of the truth-predicate to moral claims amounts tono more than the reiteration of those self-same claims (348).

At least one relevant contrast here involves some sort of correspon-dence theory of truth. For a correspondence theorist, to call a moralclaim true is to commit oneself to there being some objective fact in

2I want to avoid an issue that divides deflationists—namely, whether a sentence or what asentence expresses counts as the appropriate vehicle of truth. My own sympathies lie in the lattercamp, but I do not thereby wish to be committed to an ontology of abstract, sentence-like entitiescalled propositions. Trying to clarify how that is possible would take me too far afield to beworthwhile here. (See Glock (2003), chap. 4, especially at 134). My fellow deflationists shouldbe able to recast what I say here in terms of Brandom’s version of the prosentential theory of truth.See Brandom (1994), chap. 5).


the world that makes the moral claim true and which does so in a waythat explains or grounds the truth of the moral judgment. Truth hasa substantive nature, and moral truth does, too. Starting from these as-sumptions, the deflationist may seem to be saying that insofar as nothingin general explains truth, nothing in general explains the truth of moraljudgments, and so attaching the truth-predicate to such judgments isunconstrained by any moral reality. Indeed, the judgments themselvesare similarly unconstrained. They are, perhaps, mere expressions offeeling or personal preference or prescriptions by which we hope to in-fluence the behaviour of others, and the fact that we can apply the truth-predicate to them does nothing to make them more cognitively robust.

It seems to me that this interpretation of deflationary treatments oftruth reduces them all to ‘some form of redundancy theory of truth’(347), according to which (b) above exhausts both the pragmatic andsemantic roles of truth. However, there is no reason to suppose thatdeflationism is committed to such a theory, which is doubly implausi-ble. First, it ignores the fact that ‘true’ sometimes has other pragmaticroles to play than endorsement, as when I caution someone that whatshe believes might well be true, but I see no justification for it,3 oras when I judge that if a valid argument’s premises are true, then somust its conclusion be, without thereby endorsing any of the premisesor the conclusion. Second, it collapses an account of the pragmatic useof ‘true’ into an account of the semantic role of ‘true’—a mistake en-tirely analogous to the one Braybrooke accuses expressivist accounts ofmoraldiscourse of having committed (i.e., supposing that an account ofthe pragmatic force of moral claims tells us about the semantic role ofmorally evaluative terms). Even someone who thinks that pragmatics islogically prior to semantics should be wary of that mistake.4 A wiserdeflationism insists on keeping these two dimensions of truth distinct.

Still, it may seem that deflationism has little to say about the appli-cation of ‘true’ and ‘false’ to moral judgments beyond the observationthat moral judgments are expressed by well-formed sentences that can

3 See Rorty (Rorty (1991), 128) for this ‘cautionary use’ of ‘true.’4 Robert Brandom (Brandom (1994), chap. 5), who holds just such a priority-of-pragmatics

thesis argues that the mistake of the classical pragmatists was precisely to confuse the pragmaticsof truth with the semantics of truth.

11.3. CONTEXTUALISM: DEFAULT AND CHALLENGE 281

be embedded in conditionals, and subject to negation, conjunction, anddisjunction. It may seem that way because it is that way. Deflationistsdo not think that truth has a nature, anda fortiori they do not think thatmoral truth has a nature. They do not think that truth is an explanatoryconcept that gives us some non-trivial insight into why some of our sci-entific theories work so well, anda fortiori they do not think that thetruth of moral judgments can be enlisted to explain why we hold someof the moral views that we do. But this need not prevent them fromagreeing with the spirit of Braybrooke’s proposal, already encountered,that ‘we can keep the concept of truth in place and work out a generaltheory of warranting judgments’ (345).

A generaltheory of warranting judgments seems to me to be too am-bitious a goal because I do not think that there is anything both generaland non-trivial to be said about the nature of epistemic warrant. But, in away, it is precisely such a lack of ambition that best serves Braybrooke’sdesire to see warrant—or justification, as I shall henceforth prefer—asvariable according to the context in which a judgment is made.

11.3 Contextualism: Default and Challenge

By the term ‘contextualism’ I mean to designate a view about epistemicjustification that is hinted at in Wittgenstein’sOn Certainty(1972) anddeveloped more explicitly by such philosophers as David Annis, RobertBrandom, and Michael Williams.5 I think that Williams’s view is thebest thought out of any of these (though it overlaps considerably withBrandom’s), and it dovetails significantly with my own more inchoateviews,6 so I shall focus on it here.

5 Jane Duran (Duran (1995)) and some other feminist epistemologists have made some con-vergent suggestions about the relation between justification and context. Williams is, unfortu-nately, rather dismissive of feminist epistemology—or, rather, I hypothesize that he would be ifhe bothered to mention it by name in hisProblems of Knowledge. There the closest he comesto it is to refer critically to ‘standpoint epistemology’ (Williams (2001), 220), which he confuseswith versions of relativism. The term ‘contextualism’ has also been applied to somewhat differentaccounts of justification advanced by David Lewis, Stewart Cohen, and Keith DeRose. I shall notbe concerned with such views here.

6 See (Hymers (2000), chaps 1, 3, and 4).


Williams defines contextualism as the view that ‘standards for cor-rectly attributing or claiming knowledge are not fixed but subject to cir-cumstantial variation’ (Williams (2001), 159). This view is best under-stood in the context of considering how to respond to two kinds of scep-ticism, which I follow Williams in labelling ‘Agrippan’ and ‘Cartesian.’Agrippan scepticism is so named because of its relation to ‘Agrippa’strilemma’ (Williams (2001), 62; Williams (1996), 60), also known asthe ‘Munchhausen trilemma.’ For any knowledge claim that is ad-vanced, the Agrippan sceptic demands some justification in support ofthat claim, but the presentation of evidence or argument in favour ofthe original claim results only in a further challenge from the Agrip-pan, who incessantly demands of us how we know that which we cite asjustification—and so the conversation continues. The Agrippan’s chal-lenge seems to leave us with only three options: embark on an infiniteregress of reasons for reasons (and so never complete the task of justi-fying the initial claim), argue in a circle (and so, fail to give a genuinereason), or dogmatically assert the truth of some claim (and so, faileven to try to give a reason). Because none of these seems to consti-tute actually justifying our original knowledge claim, we are placed inthe uncomfortable position of having to concede that we are ignorantwhere we thought we knew. Because such a sceptical argument can belaunched against any knowledge claim, the scope of our ignorance isimmense.

Both foundationalism and coherentism are often presented as at-tempts to answer the Agrippan sceptic. The foundationalist contendsthat our beliefs can be classified into two kinds: (i) basic beliefs, whichdo not rely on any other beliefs for their justification and which are ‘in-trinsically credible’ (Williams (2001), 82)—justified by their very na-ture; (ii) non-basic beliefs, which acquire what justification they can getfrom being inferentially connected to one of more basic beliefs. Suchbasic beliefs serve as stopping points for the sceptic’s infinite regresswithout being dogmatic (or so it is contended).

The coherentist, by contrast, contends that the foundationalist ad-heres to an incorrect linear model of justification—a model shared withthe Agrippan sceptic. This linear model is mistaken because it is linked

11.3. CONTEXTUALISM: DEFAULT AND CHALLENGE 283

to a dubious ‘semantic atomism’ (102) without which the idea of a ba-sic belief cannot be made intelligible. In order for the foundationalist’sbasic beliefs to wear their justification on their sleeves, they must be in-telligible in radical detachment from any other beliefs. If their contentis determined, even in part, by their inferential connections with otherbeliefs, then it seems that their justification must also be, and so theyfail to be basic beliefs.7

Doubts about this sort of semantic atomism make the coherentistdubious that there are any such things as the basic beliefs described bythe foundationalist, to which the epistemic credentials of all other be-liefs can allegedly be traced. On the contrary, justification must proceedholistically, according to this view. A belief counts as justified onlyinsofar as it belongs to a justified system of beliefs, and a system of be-liefs counts as justified, in turn, only insofar as it is a coherent system.But coherence is not mere consistency. A coherent system of beliefs isrich in inferential connections (both deductive and non-deductive), andit is more coherent to the degree that it is both comprehensive and ex-planatory. All local attempts at justification, the coherentist holds, areultimately beholden to a broader global coherence of this sort.

Williams’s contextualist thinks that foundationalism and coheren-tism are both committed to two controversial assumptions that lie atthe respective hearts of our two kinds of scepticism—Agrippan andCartesian. The first of these Williams calls the ‘Prior Grounding Re-quirement’ (24). (The second, to which I shall return below, Williamscalls ‘Epistemological Realism’.) The Prior Grounding Requirement,in brief, says that one is not being a responsible epistemic agent unlessone’s beliefs are based on ‘adequate evidence’ (24). This entails, onthe one hand, a thoroughgoing internalism, according to which the onlyjustifiers are those within the ken of the would-be knower, and, on the

7 Thus Descartes is led to the view that he has an immediate and utterly simple intuition ofhis own existence as a thinking thing, and empiricist foundationalists like Schlick find themselvesclinging to ostensibly atomistic judgments or expressions of immediate sense experience such as‘Red here now.’ In order for suchKonstatierungento serve as justifiers for any other beliefs, theymust be propositional in form. But it seems that as soon as they are propositional, they involveconcepts whose meaning or content is determined in part by their applicability on other occasions.This makes them vulnerable to error, and such vulnerability to error threatens their status as foun-dations (recent attempts to articulate a ‘modest’ or ‘moderate’ foundationalism notwithstanding).


other hand, a commitment to the view that one’s belief is justified only ifno further evidence would serve as defeating evidence. A consequenceof setting the standards of justification so high is that it is easy for acritic to challenge one’s justification. All she need do is ask, ‘How doyou know that?’ and the Agrippan trilemma is immediately brought intoplay. The burden of proof lies entirely and squarely on the shoulders ofthe knowledge claimant so that every epistemic claim is, as Williamsputs it, ‘guilty unless proved innocent’ (149).

This legal analogy suggests an alternative. If we adopt the viewthat beliefs are to be considered ‘innocent until proven guilty,’ then weget something like what Robert Brandom calls a ‘default and challenge’(Brandom (1994), 177; Williams (2001), 36) model of justification. Ac-cording to this view, every belief has the ‘normative status’ (Brandom(1994), 16-17) of being justifiedby default, and it loses this status onlyif it is brought into question by a reasonable challenge—a challengethat either gives positive reasons for thinking that the claim is false or,at a later dialectical stage, gives positive reasons for doubting whateverjustification may be marshalled in defence of a claim that has been con-fronted by a reasonable challenge. This model, Williams submits, doesbetter justice to our actual practices of giving and demanding reasons,according to which the burden of proof is something to be shared by theparties involved in a critical exchange. It denies the sceptic the unde-fended right to enter ‘naked challenges’ (Williams (2001), 150).

It might be wondered what, in principle, prevents the Default andChallenge model from leading disputing parties into an infinite regressof reasons. The answer is, ‘Nothing—in principle.’ If people had un-limited lifespans, few practical concerns, and vast intellectual resources,they might continue to debate forever. (Indeed, some debatesseeminterminable—we politely call them ‘philosophy.’) But in practice manydebates do come to an end - an end at which no further reasonable chal-lenge is forthcoming, or an end at which the knowledge claimant failsto respond to a reasonable challenge. Such ends are not the intrinsicallycredible basic beliefs of the foundationalist, but they are not cases ofacquiescing in dogma either. They are reasonable resting points, some-times temporary, sometimes very stable, and they vary from one context

11.4. CONTEXTUALISM: REJECTING EPISTEMOLOGICAL REALISM285

of debate to another.8 There is no reason,prima facie, to expect all rea-sonably resolved debates to end up in the same place. It makes no moresense to ask which points are intrinsically stopping points than it doesto ask ‘which points in Ohio are starting points?’ (Quine 1980, 35).

Such resting points help to define a context of debate or inquiry.Unlike the classical foundationalist’s basic beliefs, they are not intrin-sically beyond doubt, but to entertain doubts about them is to mistakeor change the context of inquiry. Thus, to respond without joking tomy claim that I left my wallet in my office with the challenge ‘How doyou know? You might just be dreaming’ is to enter a challenge that isinappropriate in the context, although it may well have a place in someother context of debate—a class on epistemology, for example.

But the latter-day foundationalist will demur, insisting that insofaras all knowledge of the world around us comes from the senses, we canexpect that all debates one way or another can be properly resolved onlyby tracing reasons back to basic beliefs about sense experience. And thecoherentist will insist that the temporary or stable resting points at whichmany debates actually stop do not count as anything more than justifi-cation for practical purposes. Such local justification becomes justifica-tion proper, only when it has been pursued to the global level (Williams1996, 290).

11.4 Contextualism: Rejecting Epistemological Realism

Both of these objections lead us—if not obviously—to our second kindof sceptical worry. If Agrippan scepticism gets its force from the PriorGrounding Requirement—from the permission of naked challenges toknowledge claims—the same cannot be said of Cartesian scepticism.For Cartesian sceptical problems are problems of underdetermination.To put this in my own favoured vocabulary, the sceptic about the exter-nal world assumes that the relation between experience and the worldis paradigmatically an external relation—that it is such that we couldidentify and describe our experience without thereby being assured that

8 There are also unfinished debates that reach no resolution, because of, for example, eitherintractable disagreement or the death of one or more parties.


we could identify and describe the reality that we assume to lie beyondappearances. This is because the Cartesian sceptic offers a rival hypoth-esis for the explanation of experience. Unlike the Agrippan sceptic, shedoes not simply ask, ‘How do you know that?’ Rather, she suggeststhat the hypothesis that there is a world beyond my mind, populated byobjects, events, and other minds, lacks explanatory power because theevidence in its favour is also evidence in favour of the hypothesis thatI am the victim of some evil demon’s deceptive activities. Cartesianscepticism is thus committed to what Williams refers to as ‘epistemo-logical realism’ (Williams (2001), 84, 193; Williams (1996), 89-134).This is the view that beliefs get their epistemic status from their con-tent, so that some beliefs are by their very nature epistemically prior toothers. It is the alleged intrinsic epistemic priority (as opposed to causalpriority) of beliefs about immediate experience that makes the Cartesianunderdetermination problem pressing. If epistemic priority is a matterof context and not content, then the Cartesian problem can be avoidedin many contexts.

Williams thinks that both foundationalism and coherentism are alsocommitted to epistemological realism. This commitment is most ob-vious in the case of the foundationalist, for whom basic beliefs - typi-cally beliefs about the immediate contents of experience—have an in-trinsic epistemic priority over non-basic beliefs. But Williams arguesthat it also characterizes the case of the coherentist, who is committedby a more circuitous route to granting epistemic priority to principlesof logic (needed for making assessments about coherence) (Williams(2001)1, 135; Williams (1996), 301) and to ‘meta-beliefs’ (Williams(1996), 302) concerning the coherence of one’s own total system of be-liefs (Williams (2001), 136-7). (Either the meta-belief that my systemof beliefs is coherent is justified, or it is not. If it is, then that mustbe because it belongs to a more encompassing system of beliefs, aboutwhose coherence I now require a further meta-belief whose justificationis in question. If it is not, then I have no reason to think that my beliefsystem is really coherent after all and no justification, ultimately, forany of my beliefs.) Coherentism is ‘foundationalism in disguise’ (137).

Contextualism, then, is characterized in part by a rejection of episte-

11.4. CONTEXTUALISM: REJECTING EPISTEMOLOGICAL REALISM287

mological realism (and with it, foundationalism, coherentism, and vul-nerability to Cartesian underdetermination problems) and of the PriorGrounding Requirement, in favour of a Default and Challenge Modelof epistemic justification. Can more be said in the way of a positivecharacterization of the view? Briefly, Williams outlines five kinds ofconstraints on epistemic contexts that determine whether a given beliefor claim should be taken to be justified.

First, if we are to be engaged in meaningful inquiry or debate at all,we must assume as justified a significant bundle of background beliefs.‘A doubt that doubted everything would not be a doubt’ (Wittgenstein1972, 450). This is just a corollary of the default and challenge model,which holds a belief justified unless challenged and requires that chal-lenges be reasonable. Williams refers to constraints of this kind as ‘in-telligibility or semantic constraints’ (159).

Closely related to, but less general than, such semantic constraintsare ‘methodological’ (160) or ‘topical’ or ’ disciplinary’ constraints (Williams(1996), 117). Such constraints treat certain propositions as ‘method-ological necessities’ (Williams (2001), 160; Williams (1996), 123), ex-empting them from doubt in a way that serves to define a context ofinquiry. Thus, it is a methodological necessity of particle physics thatthereare subatomic particles. If one doubts this, then one has ceasedto do particle physics and has moved to a new epistemic context—thecontext of some radically new physical theory, perhaps, or, more likely,the context of instrumentalist philosophy of science. Similarly, it is aprecondition of historical inquiry—Williams’s favourite example—thathistorians assume that the Earth existed long before we were all born.Such methodological necessities, as we saw earlier, are not themselvesindubitable. But to attempt to cast doubt on them is either to misunder-stand the context of inquiry or to try deliberately to change it.

Once one has embarked on a particular inquiry, one’s claims andquestions become subject to a third kind of constraint—‘dialectical’constraints (Williams (2001), 161; Williams (1996), 117). Whethermy claim that Arthur Wellesley’s failed night attack at Seringapatampersuaded him never again to launch a night attack is justified or notdepends on what has already been said by military historians on the


subject. If evidence has already been presented that offers some rivalexplanation for Wellesley’s preferences, then my claim will not enjoythe status of being default-justified. But if I am the first one to broachthe matter, then default justification is mine.

Of course, how stringent we are about demanding justification de-pends on how much time we have, and what the stakes are in a givendebate. The more severe the practical consequences of being wrongare, the tighter the ‘economic’ constraints (Williams (2001), 161) willbe. When my life hangs in the balance, I apply a higher standard of jus-tification to the claim that the cup on the left, not the cup on the right,contains the poison than I would if I were doing a high school chemistryexperiment.

Finally, if my claims about Arthur Wellesley at the battle of Seringa-patam are to be justified, it must not only be the case that there is nocontrary claim already in play and that sceptical (or, in the case ofpre-history, evangelical) doubts about the age of the Earth are held inabeyance; it must, additionally, actually betrue that the Earth existedlong before we were born. If the Earth is not more than two hun-dred years old, then no claim about Arthur Wellesley’s early exploitsis justified at all. Such ‘situational’ constraints (Williams (2001), 162;Williams (1996), 117) make it clear that contextualism is not a purelyinternalist picture of justification because my belief’s being justified de-pends in part upon the instantiation of facts that may be well beyond myken. Similarly, it must be the case that my eyes are functioning suffi-ciently well if my claim that the driver who almost ran me over was atthe wheel of a sports-utility vehicle and not a sedan is to be justified—even though the proper functioning of my eyes may be in certain re-spects beyond my own ability to discover. Contextualism thus acknowl-edges that reliabilism points to something important in our practices ofjustification, but declines any attempt to reduce normative concepts likejustification toostensiblynon-normative ones, like causation by a reli-able process.9

9As Williams points out, ‘reliability’ is a covertly normative term (Williams (2001), 33), sincea process is reliable to the extent that it produces results that we value. I take this to be much thesame point that John Pollock makes (Pollock (1986), 118-201), when he complains about the

11.5. ETHICAL CONTEXTS AND META-ETHICAL CONTEXTS 289

The contextualist thinks that both epistemological realism and thePrior Grounding Requirement are contentious theoretical commitments.One may be epistemically responsible without having adequate evi-dence for one’s belief, and beliefs get their epistemic priority or pos-teriority not from their content, but from theircontext. If this is so, thensceptical worries about our ‘knowledge of the external world’ can getno purchase, as long as we are not committed to the epistemological re-alism or the Prior Grounding Requirement that they rely on. Scepticismis not, of course,refutedby such considerations, but its reliance on con-tentious theoretical premises entails (i) that the sceptic is herself com-mitted to knowledge claims to which we need not be committed—anembarrassing position for a thoughtful sceptic to be in—and (ii) that weneed not be troubled by the sceptic’s peculiar worries precisely becausethey are peculiar—they are of concern only for someone who must ac-cept commitment to the theoretical presuppositions on which scepticaldoubts rely.10

11.5 Ethical Contexts and Meta-ethical Contexts

It should be clear that this sort of contextualism is friendly to Bray-brooke’s suggestion that we view justification in such a way that ‘war-ranting moral judgments would proceed in some respects differentlyfrom warranting reports of perception, and warranting aesthetic judg-ments more differently still’ (Braybrooke (2003c), 345). Contexts ofmoral judgment differ from contexts of perceptual judgment and fromcontexts of aesthetic judgment. It is, for example, a methodological ne-cessity of moral judgments that there are agents and patients who can beharmed or benefited by one’s actions and the actions of others. Withoutthis assumption, morality would have no foothold. In another context—

arbitrariness that faces any attempt to identifytheprocess that produces a belief and then evaluatethat process for its reliability.

10 As Williams points out, contextualism also has the unexpected but harmless result thatwhen we take the sceptic’s worries seriously by at least provisionally entertaining epistemologicalrealism—as in the context of discussing them in a class on epistemology—we actually cease toknow things about the world around us. But as soon as we emerge from our discussion into theworld of practical affairs, the knowledge we had lost comes back again. Williams calls this the‘instability of knowledge’ (Williams (1996), chap. 8; Williams (2001), chap. 4).


the context of perceptual judgments, for example—it is not clear thatthe potential harms or benefits of my actions for others have the sameconstitutive status. Indeed, that same judgment may be called into doubtin some other context (e.g., the problem of other minds).

The affinity between this sort of contextualism and Braybrooke’s po-sition is made more manifest if we reconsider his remark that Moore’sopen-question argument neglects the distinction between ethics and meta-ethics. The argument has an important place, he thinks, but its impor-tance is meta-ethical rather than ethical:

In meta-ethics, the possibility of alternative definitions of‘good’ and other moral terms is firmly established. Askingthe open question of what is seized upon to define ‘good’as in ‘Is this really good?’ amounts to asking ‘Could “good”not be defined otherwise?,’ which is the counterpart in meta-ethics of ‘Is this really good?,’ and displaces it. The an-swer to ‘Could “good” be defined otherwise?’ is always‘Yes.’(347)

But, thinks Braybrooke, from the possibility of alternative definitions of‘good’ we ought not to infer that no definition of ‘good’ is ever of anyvalue to us, or, indeed, that every definition is as valuable to us as everyother.

I think that the situation here parallels a certain form of conven-tionalism about mathematics that we find in the work of Wittgenstein.According to this kind of conventionalism, certain principles of math-ematics are implicit in our practices before we ever articulate them,perhaps for reasons having to do with ‘the natural history of humanbeings’ (Wittgenstein (1968),§415), in particular, our evolutionary his-tory. That we have the mathematics (or, perhaps, even the logic) that wedo is a contingent matter of fact. Thus that certain propositions countas necessary propositions—propositions thatseemto tell us about theessence of things—is, ultimately, the result of an implicit convention,but this does not make itarbitrary. ‘[T]o the depth that we see in theessence there corresponds the deep need for the convention’ (Wittgen-stein (1978), 65). And given the convention, we have ways of sorting


the justified mathematical claims from the unjustified ones.Likewise, having accepted a certain contingent definition of ‘good’—

perhaps because it is implicit in our behaviour—we may find ourselvesunder the illusion that it is the only possible one and, so, react with dis-satisfaction to any attempt to make that definition explicit, once we real-ize that such explicit formulation opens the door to alternative possibil-ities. (Schlick and Waismann, under the spell of theTractatus, reactedsimilarly to the idea of alternative logics.) But, again, given a particulardefinition of good, we have ways of sorting the justified moral claimsfrom the unjustified ones.

A comparison with mathematical conventionalism is not likely tostill the unease that a critic may feel upon encountering Braybrooke’scontention that the justification of ethical judgments is a matter en-tirely distinct from the justification of meta-ethical judgments.11 Here iswhere I think contextualism can be of greatest use to Braybrooke. If weare inclined to think that the justification of a moral claim presupposesthe justification of some meta-ethical principle, then that, I think, is be-cause we are tacitly committed to what Williams calls ‘epistemologicalrealism.’ An example may help to illustrate and justify my contention.

Imagine two academic philosophers engaged in debate about the oldRepublic of South Africa’s policy of apartheid. One philosopher insiststhat the policy is manifestly wrong and holds that it would be inappro-priate, therefore, to extend an invitation to the South African vice-consulto visit campus and speak on the topic ‘Apartheid: The Other Side.’12

The other philosopher responds by asking for some reason to believe inthe immorality of apartheid. ‘Well,’ says the first, ‘apartheid is a form oftreatment that discriminates on the basis of skin-colour.’ ‘But why,’ asksthe second, ‘is it wrong to discriminate on the basis of skin-colour?’—‘Because skin-colour is a morally irrelevant difference, and it is alwayswrong to discriminate on the basis of irrelevant differences.’ ‘But what

11 Braybrooke (personal communication) objects that I do not get him quite right here, becausethere may be cases in which meta-ethical judgments effectively undermine moral judgments—forexample, when someone asserts that ‘Moral judgments are nothing but the expression of emotion’in response to my claim that ‘It is wrong to boil your neighbours’ babies for dinner.’ I am inclinedto assimilate such cases to the debate about apartheid that I describe below.

12 I’m not making this up.


is the justification for such a rule?’ continues the second in his siftinghumour. At this point, the first philosopher may respond by citing somegeneral principle—the categorical imperative, perhaps, or the greatesthappiness principle, or, as Braybrooke would have it, a principle whichsays that moral rules are licensed insofar as they contribute to a ‘thriv-ing society’ (Braybrooke (2003c), 342; my little dialogue, here, followsthe pattern of Braybrooke’s). ‘But why,’ responds the second, ‘shouldwe care about the thriving of a society or about the greatest happinessof the greatest number or about treating others as ends in themselves?Why think that any of these things tells us about moral goodness? Why,indeed, think that there is any such thing as moral goodness? Isn’t good-ness, after all, a rather queer property to expect to find being instantiatedby anything in the world?—A property that has the power to motivatethose who apprehend it?’13

This debate over the rightness of a particular social practice quicklyturns into a debate about whether there are any such things as moralproperties or facts. It moves from the ethical level to the meta-ethicallevel as if that were the most natural dialectical move in the world, andup until the last step the second philosopher employs the very spare toolkit of the Agrippan sceptic, asking ‘why?’ whenever the first philoso-pher makes an assertion and insisting that the first philosopher bear theburden of proving these assertions.

Non-philosophers surveying the debate will, I contend, think that thesecond philosopher has missed the point or is merely playing games ofsome sort,14 and if the context of the exchange is as part of a publicdebate about whether or not to invite the South African vice-consul tocampus, then they will be right. Justification of a moral claim does notdepend on the justification of any general meta-ethical principle, andin this context it is precisely the moral claim that is at issue, not themeta-ethical principle. Why?

Let us apply the default and challenge model of justification to thedebate. According to that model, remember, the challenger cannot sim-ply enter ‘naked challenges.’ Challenges must, rather, be reasonable,

13 See Mackie (Mackie (1977), 38-42) for this ‘argument from queerness.’14 As I said, I’m not making this up.


where a reasonable challenge is one that either shows fault with somejustification already given or provides some reason for thinking thatthe conclusion being advanced is false. By this standard, the secondphilosopher is obliged to show either that appeal to a rule or generalprinciple fails to support the claim that apartheid is wrong or that thereare positive reasons for thinking that apartheid is morally acceptable af-ter all. In doing so, his position becomes vulnerable to criticism, too,and the Agrippan road to moral scepticism about particular principles,or about the objectivity of morality generally is blocked.

However, the second philosopher may well agree to these terms andstill try to make a sceptical case (i) about the particular moral claim(apologists for apartheid used to try to meet this criterion by arguing,for example, that the abolition of apartheid would lead to a civil warthat would produce far greater suffering than the policy of racial dis-crimination itself produced), or (ii) about the general principle (‘Thenotion of thriving is too vague,’ might be the complaint) or (iii) aboutthe very existence of moral facts (‘Mackie’s argument from queernessshows us that no moral property is ever instantiated’).15 The defaultand challenge model of justification offers us no ground for complainthere because the sceptic is in fact shouldering some of the burden ofproof. These challenges are clothed, not naked, and in this respect theyresemble the challenges of the Cartesian sceptic, who does not simplyask ‘How do you know?’ but tries to offer positive reasons for think-ing that we do not have much of the knowledge that we naively assumeourselves to have.

That much resemblance is too little for us to draw any interestingconclusions about the plausibility of such moral scepticism. But thereare deeper resemblances, and they deserve our attention. In particu-lar, the philosopher who tries to push the debate from stage (i) to stage(ii) assumes that claims made at stage (i) lack proper justification un-less a justification has been given of general moral principles. And the

15 Of course, there is another kind of moral scepticism, which has no obvious or plausiblecounterpart in the case of global scepticism about the senses, corresponding to the question ‘Whyshould I be moral?’ That’s a philosophical chestnut that does not need roasting here, thoughit is importantly connected with the internalism assumed by Mackie’s argument from queerness(Mackie (1977), 38-42), to which I shall return.


philosopher who tries to push the debate further to stage (iii) assumesthat neither particular moral claims nor general moral principles haveproper justification, unless it has been shown that there really are suchthings as moral facts. In both these moves, I detect epistemologicalrealism.

Consider the move from stage (i) to stage (ii). If I am to know someparticular moral claim to be true, then, the sceptic seems to be saying,I must know some general moral principle that certifies that particularclaim (a necessary condition, not a sufficient one). The justification ofthe particular claim is only provisional until we have presented the cer-tifying general principle. This is to say that in virtue of their content,particular moral claims are dependent for their justification on anothercategory of claims, also demarcated bytheir content—general moralprinciples. And this is to commit oneself to the idea that there are ‘natu-ral relations of epistemological priority’ (Williams (2001), 192). A rea-son to doubt epistemological realism is thus a reason to doubt that thejustification of particular moral claims is incomplete until those claimshave been traced to general moral principles.

This is not to deny that whatever general principles my particularmoral claims may presuppose must betrue if I am to have knowledgeof those particular claims. As we saw earlier, justification is subjectto various ‘situational’ constraints. But it is enough—subject to ourother contextual constraints on justification—that those principlesbetrue for my claims to be justified. I do not also have to have demon-strated them.16

The situation here is analogous to one that arises in the context ofour knowledge of cause and effect. Humeans about causation hold thatwherever there is some cause-and-effect relationship, there is also somegeneral causal law of which the particular case is an instance. But itis not obvious that from this we should conclude that in order to haveknowledge of some particular instance of that relation I must have a

16 As Tom Vinci reminds me, a foundationalist can make sense of my not having to give ajustification for a general principle in order for a particular claim to be justified by invoking a dis-tinction between therebeinga justification that could be given and that justification’shaving beengiven. Such free-floating justifications that wait around to be given smell too much of Platonismfor my contextualist nose, but that is another debate.


grasp of the causal law itself. It is enough—subject to other contextualconstraints—that there be such a law.17

Of course, if a reasonable challenge suggests that some general prin-ciple is false, then that does undermine my justification for particularclaims—just as evidence for thinking that documentary reports of thebattle of Seringapatam had been falsified would undermine whateverparticular claims I tried to make about that battle.18 But I need not firstundertake to show that any general principle is correct in order to makejustified particular claims. That is, general principles are not by theirvery nature (by their very content) better known than particular claims.(Nor are particular claims by their very nature better known than generalprinciples.)

A comparable case can be made concerning some sceptical doubtsabout the very reality of moral facts. The logical positivists, for ex-ample, doubted the reality of moral facts on grounds that no senseexperience would count for or against the truth of a moral claim like‘Apartheid is wrong.’ The plausibility of such a proposal is, of course,directly linked to the plausibility of holding a verification theory ofmeaning, according to which the meaning of a statement is its method ofconfirmation and disconfirmation. Such reductionism not only fails toaccommodate the fact that hypotheses are never tested in isolation frombackground assumptions, as Quine argued (Quine (1980), 42), but italso is clearly committed to epistemological realism, for it treats reportsand beliefs about the deliverances of our senses as having an intrinsicepistemic priority over reports and beliefs about any would-be moraldata—indeed, over any non-observational reports and beliefs.

However, other kinds of sceptical doubts about the reality of moralfacts may appear on the surface not to be guilty of anything like this.Take, for example, the general kind of moral scepticism that the sec-

17 See Davidson (Davidson (1980), 16-17).18 But contrast this with the case in which I have no reason for thinking the general principle

false, but in which a sceptic undermines my positive justification for the principle. Then I lackjustification for whatever particular judgments presuppose the principlein the context of discussingthe general principle, but it does not follow that in the context of discussing the particular claimsI lack justification for them, so long as the general principle they presuppose is, in fact, true.Knowledge may be unstable. See note 9 above and Williams (Williams (1996), chap. 8; Williams(2001), chap. 4).


ond philosopher in my little dialogue tries to press by appealing toMackie’s ‘argument from queerness’ (Mackie (1977), 38-42). Thereare, as Mackie notes, really two components to this argument. First,Mackie claims that objective moral properties would be completely un-like anything else whose existence we are justified in believing in:

Plato’s Forms give a dramatic picture of what objective valuewould have to be. The Form of the Good is such that knowl-edge of it provides the knower with both a direction andan overriding motive; something’s being good both tells theperson who knows this to pursue it and makes him pursue it.An objective good would be sought by anyone who was ac-quainted with it, not because of any contingent fact that thisperson, or every person, is so constituted that he desires thisend, but just because the end has to-be-pursuedness some-how built into it. (40)

Thus, Mackie thinks that objective moral properties must be intrinsi-cally motivating, so that their mere apprehension will automatically dis-pose us to act in a way that we take to be morally praiseworthy and suchthat their motivational power has nothing to do with contingent factsabout general or individual human psychology. This is the ‘queerness’of moral properties, for they seem in this respect quite unlike any otherproperties recognized by mature science.

Second, properties that have this queerness and are, moreover, sup-posed to be objective, present a challenge to epistemology. How couldwe come to know about such things?

[N]one of our ordinary accounts of sensory perception orintrospection or the framing and confirming of explanatoryhypotheses or inference or logical construction or concep-tual analysis, or any combination of these, will provide asatisfactory answer; ‘a special sort of intuition’ is a lame an-swer, but it is the one to which the clear-headed objectivistis compelled to resort. (39)

Let me focus on Mackie’s complaint that we do not obtain moral knowl-edge by way of ‘explanatory hypotheses or inference.’ His point, I take


it, is that we are not justified in believing in the existence of objectivemoral properties by an argument which takes the best explanation ofsome better-known fact to be that there are such properties. Thus, thefact that we have moral convictions—beliefs that seem most plausiblyendorsed by praising them as ‘true’—is not best explained by the exis-tence of moral properties, the apprehension of which induces in us suchconvictions. On the contrary, says Mackie, there is a rival hypothesisthat handles all the data about human moral psychology quite well. Ourbelief in the objectivity of value is really the result of our tendency toobjectify our own ‘wants and demands’ (43). Moral belief involves akind of fetishism:19

[B]oth the adjective ‘good’ and the noun ‘goods’ are usedin non-moral contexts of things because they are such as tosatisfy desires. We get the notion of something’s being ob-jectively good, or having intrinsic value, by reversing thedirection of dependence here, by making the desire dependupon the goodness, instead of the goodness on the desire.And this is aided by the fact that the desired thing will indeedhave features that make it desired, that enable it to arouse adesire or that make it such as to satisfy some desire that isalready there. (43)

Thus, because there is no objectivity of value, on Mackie’s account, allof our moral claims—which we make with the conviction that there areobjective values—prove to be false, and since we cannot know some-thing which is not the case, we have no moral knowledge, despite ourmany moral beliefs.

Mackie relies here on an underdetermination argument, for he con-tends that my moral experience—my conviction that certain actions andoutcomes are right or wrong—is explained as well by the hypothesisthat no moral properties are instantiated—just projected—as by the hy-pothesis that moral properties are instantiated. This might remind us of

19 Mackie (Mackie (1977), 42) compares his error theory to Hume’s view that we projectcausal relations onto the world, even though they are no objective part of the world, but we mightextend the comparison to Feuerbach’s discussion of God and Marx’s treatment of commodities inthe first volume ofCapital.


the Cartesian sceptical argument: my experience as if of a world of ob-jects and events beyond my mind is explained just as well by my lack ofperceptual contact with such a world (by my being a brain in a vat, forexample) as it is by mybeingin perceptual contact with such a world,and so, I am not justified in accepting the hypothesis of my perceptualcontact with the world. But this tempting analogy cannot be trusted todeliver the conclusion that Mackie’s error theory relies on epistemologi-cal realism. Underdetermination arguments do not by themselves entailepistemological realism. It is part of the nature of explanation that whatis to be explained must be assumed to be better known than that whichexplains it. However, this assumption may well be provisional, and inanother context that which was assumed earlier may now be cast intodoubt or treated hypothetically. Only if we think that that which weare trying to explain must be regarded as better known in all epistemiccontexts are we committed to epistemological realism.

However, Mackie’s view does assume that there is a kind of unity tomoral experience, such that it makes sense to suppose that such experi-ence can be identified and described independently of the moral prop-erties to which it ostensibly refers. Consider again Mackie’s contentionthat we come to believe things to be morally good because we first dis-cover that they satisfy certain of our desires. We then mistake the factthat we desire them for their intrinsic desirability (where intrinsic desir-ability means that ‘the end has to-be-pursuedness somehow built into it’(40)). What this story does not tell us is why we do not make the samemistake in the case of non-moral goodness. Ordinarily (there are excep-tions), we do not suppose that someone who does not share our taste indesserts is failing to respond properly to some objective feature of theworld. We do not suppose that preferring hockey to rugby is a moralfailing. But if something’s seeming to be good is just an illusion causedby our desiring it, then we ought, in fact, to be subject to the same il-lusion regardless of the content of the desire. Mackie’s argument thuspresupposes that we can identify certain of our own desires as havingmoral content, quite independently of being able to identify any moralproperties in the world around us (since, after all, there are not supposedto be any such properties instantiated). But how does anybody ever get


to do that? Mackie seems to be relying here on the very Cartesian modelof inner experience that the sceptic about the external world is inspiredby. According to that model, my inner experience is known to me inde-pendently of the way the world is in the sense that I could have exactlythe same inner experience in the absence of what we normally regard tobe its typical causes. Likewise, it seems, on Mackie’s account I couldhave the same inner moral experience, regardless of how the world is,and I could tell my desires with ‘moral’ content from my desires withnon-moral content without having to encounter actual moral situationsand without being trained and initiated into any particular set of moralpractices. All of my knowledge of the world, including my would-bemoral knowledge, is ultimately beholden to that which by its very na-ture is better known—my inner experience.

Now Mackie might try to respond that I have misunderstood his po-sition. There is, after all, some social or biological explanation for howit is that we come to make the error only in the case of ‘moral’ goodsand not in the case of non-moral ones:

Moral attitudes themselves are at least partly social in ori-gin: socially established—and socially necessary—patternsof behaviour put pressure on individuals, and each individ-ual tends to internalize these pressures and to join in requir-ing these patterns of behaviour of himself and of others. Theattitudes that are objectified into moral values have indeedan external source, though not the one assigned to them bythe belief in their absolute authority. Moreover, there aremotives that would support objectification. We need moral-ity to regulate interpersonal relations, to control some of theways in which people behave towards one another, often inopposition to contrary inclinations. (42-3)

We learn to distinguish our desires with moral content from those withnon-moral content by means of social training in ‘patterns of behaviour’that serve to ‘regulate interpersonal relations,’ not by acquaintance withthe form of the good or by way of ‘a special sort of intuition’ (39).But it seems to me that any plausible explanation of this sort will come


precariously close to granting precisely those facts about ‘the naturalhistory of human beings’ (Wittgenstein (1968),§415) that contemporarymoral realists think are all they need to make their case.20 Mackie’sconfession that weneedthis kind of regulation seems to me to concedethe most important point: the deep need for moral conventions is whattempts us to look for moral essences, but that makes the need no lessdeep.

Of course, Mackie would demur. Showing that moral desires arisefrom facts about our evolutionary history or that they have their genesisin a shared, inculcated concern for the thriving of a given social order(or some combination of both perhaps) would not count for Mackie as‘a reasonably robust version of moral realism’ (Braybrooke (2003c),347). This is because when Mackie tries to bolster the case for his errortheory by undermining the moral realist’s hypothesis, he does so by in-voking an account of moral properties drenched in the heady spirits ofPlatonism, supposing that only such real essences could account for the‘to-be-pursuedness’ (Mackie (1977), 40) of moral ends. But there is noreason why a moral realist has to accept either the full-blown motiva-tional internalism of such a view or the existence of real universals. Torepeat: ‘[T]o the depth that we see in the essence there corresponds thedeep need for the convention’ (citetRFM, 65). That morality is in somesense contingent may provoke a certain attitude of irony, but it does notdeprive morality of its internal necessity.21

As I have been not-so-subtly hinting, the case here is rather likethe case of mathematical truth.22 What the mathematical convention-alist offers us is a way of making sense of the truth of mathematicaljudgments that does not require us to go looking for mysterious math-ematical facts in the world or beyond it. Likewise, what Braybrookeoffers us is a way of making sense of truth in ethics that does not re-quire acquaintance with other-worldly moral facts whose epistemic ac-

20 See, for example, Campbell and Woodrow (Campbell and Woodrow (2003)) and Campbell’schapter in this volume.

21 One might take this as a way of trying to rephrase what I think is right about Rorty’s (1989)view—and perhaps what Braybrooke’s view has in common with it.

22It is no coincidence that Wittgenstein describes mathematics as ‘a network of norms’(Wittgenstein (1978), 431).

11.6. CONCLUSION 301

cessibility is mysterious and whose motivational force seems ‘queer.’23

The motivational force of moral judgments, according to Braybrooke,is something that derives from a prior commitment to ‘criteria of socialthriving’ (Braybrooke (2003c), 346), which in turn is the ‘historicallycontingent’ result of ‘evolutionary social adaptation to the conditions ofthriving’ (350). (And that certain practices and situations promote thethriving of individuals and communities is, thinks Braybrooke, largelya matter of empirical fact.)

11.6 Conclusion

I conclude that Braybrooke’s invocation of the ethics/meta-ethics dis-tinction is defensible. Moral claims can be justified by reference tomoral principles even though no justification has been given for thoseprinciples because moral principles are default-justified and cannot bechallenged without changing the context of inquiry or discussion. Con-sequently, if the proper application of the truth-predicate to claims isgoverned by contextually specific criteria of justification, ‘true’ maybe used of moral claims ‘unselfconsciously, without thinking of meta-ethics’ (350), even though it may be applied ‘within one structure ofideas that figures among many logically possible structures’ (350). Thecase of ethics is thus much like the case of mathematics, according toone kind of conventionalism. What contextualism helps to show us isthat one can be a conventionalist (at the meta-ethical level) and a realist(at the ethical level) at one and the same time.

23 I think the problem is exacerbated both here and in the mathematical case by commitmentto a correspondence theory of truth, but I already have too many axes in the fire (to coin anexpression).

Chapter 12

Braybrooke and the FormalStructure of Moral Justification

TOM V INCI

Abstract

In his paper ‘What Truth Does the Emotive-Imperative Answer to theOpen-Question Argument Leave to Moral Judgments?’ David Braybrookedevelops several themes in ethical theory, including an account of the na-ture of moral justification. I discuss his elucidation of the structure ofmoral justification within a broader account of how formal foundation-alism, a doctrine of the structure of epistemic justification, can providea useful protocol for understanding his account of the structure of moraljustification. I conclude by raising some difficulties.

12.1 Introduction

In his recent paper, ‘What Truth Does the Emotive-Imperative Answerto the Open-Question Argument Leave to Moral Judgments?’1 DavidBraybrooke develops several themes in ethical theory, including an ac-count of the nature of moral justification. Playing supporting roles arearguments showing that the account of justification thus developed cangive a measured response to the challenge posed by Moore’s ‘open-question’ argument. The account draws upon insights in Braybrooke’s

1 Braybrooke (2003c). Hereafter page numbers in parentheses refer to this text.

303

304 CHAPTER 12. TOM VINCI

own natural law approach to morality2 and in the classical emotive-imperative accounts of Stevenson3 and Ayer.4 There is an apparent ten-sion between these two theories regarding how to handle the conceptof truth in contexts of moral justification: natural law theory seems tofavour its application to moral claims; emotive-imperative theory doesnot. One of Braybrooke’s concerns is to show how a theory that incor-porates elements of both approaches assigns truth to moral claims inways that does violence to neither.

I approach Braybrooke’s paper as an epistemologist rather than anethical theorist. For much of the twentieth century, epistemology andethical theory have been concerned with apparently distinct subject mat-ters, each developing its own set of problems and methods—ships pass-ing in the night. Yet there have been some close encounters in both di-rections: epistemology looking to ethical theory and ethical theory look-ing to epistemology. An example of the former is Roderick Chisholm’sprogram in epistemology;5 another example is virtue epistemology.6

In the other direction, one project looks at whether models of knowl-edge developed within contemporary epistemology can be applied toethics—‘moral epistemology’ as it has come to be called. For exam-ple, Robert Audi has recently catalogued various models of knowing inepistemology—empiricism, rationalism, intuitionism, and non-cognitivism—and considered which best fit various positions in ethical theory. (Heproposes that a moderate version of intuitionism is the best fit for themost plausible version of ethical theory.)7

My own contribution to this volume falls within the general head-ing of moral epistemology. I am not, however, as Audi is, interested intaking a substantive position in epistemology but rather in looking at aformal account in epistemology about the way in which justification canbe transmitted from the general to the particular, Sosa’s formal founda-

2 David Braybrooke, Natural Law Modernized (Toronto: University of Toronto Press, 2001).3Stevenson (1944)4Ayer (1936)5Chisholm (1966), 11ff.6 See, for example, Axtell (1997)7Audi (1999)).

12.2. BRAYBROOKE ON THE OPEN-QUESTION ARGUMENT 305

tionalism.8 The exercise in moral epistemology that I undertake here isto suggest that formal foundationalism is embodied in Braybrooke’s ap-proach to ethical justification, to develop in detail a formal foundation-alist interpretation of Braybrooke’s account of justification, to considerhow some substantive problems arising within Braybrooke’s substan-tive ethical theory can be formulated with the model, and to show, withthe help of the model, how these problems might be resolved.

12.2 Braybrooke on the Open-Question Argument of G.E.Moore

Braybrooke’s paper begins with a discussion of the open-question argu-ment. Moore thought that if you find that something has a certain prop-erty that is not in itself evaluative—that it is pleasurable for example—itis then an ‘open question’ whether that is good. It is notoriously diffi-cult to give a precise definition of what Moore may have meant by this,but Braybrooke attempts one:

If you purport to define ‘good’ as pleasure or anything else,you will be cutting off any opponent who tries to disagree,saying to her, ‘This is not an open question ... no one canthink otherwise except through confusion.’9

The connection between an ‘open question’ and the possibility of giv-ing a definition for ‘good’ seems to be that if a question like ‘Is pleasuregood?’ is an open question in the sense intended then one cannot answerit by saying ‘Yes it is; and that is true by definition.’ Moore’s explana-tion for this is that ‘good’ is a primitive property; that is, it is a gen-uine property, something that truly characterizes things, but it is a basicproperty, not analysable into—hence not definable in terms of—naturalcomponents. Of course, classical emotivists also think that ‘Pleasure isgood’ does not express a definitional truth, but they think so not because‘good’ is a primitive term but because the sentence does not express atruth (proposition) at all—it expresses an emotion or an injunction.

8Sosa (1980), at 12ff.9 The passage is from Braybrooke (341); the internal quotation is from G.E. Moore, Moore

(1903), 21.


Braybrooke considers several proposals among emotivist-imperativist(‘EI’ for short) theoreticians10 about what is expressed by an utteranceof a declarative sentence concerning the moral status of doing X: (1)negative feelings I have about doing X; (2) an ‘injunction, softer per-haps than an ordinary imperative’ for others to feel the same way; (3) aninjunction for me to act appropriately, e.g., by way of making ‘repara-tions’ (341). The combination of these three is ‘what may be describedas the emotivist-imperative theory’ (342).

Braybrooke is not very sympathetic to the idea that when we say‘That’s good’ truth or falsity is simply inapplicable. He notes that wequite properly and naturally respond to moral claims by saying ‘That’strue’ (344-5) and is not sympathetic to the idea that this is a sense oftruth less robust that we apply to factual assertions (345). But Bray-brooke also finds in the EI approach something that is right and valu-able both in its commitment to the role of emotions and injunctionsin moral judgment and in its giving a satisfactory response to Moore’sopen-question argument. What is a bit tricky here, and requires carefulnavigation by Braybrooke and his readers, is combining the idea thattruth is applicable in a robust sense to the objects of moral evaluationwith the idea that those objects are to be understood on the emotive-imperative model. Braybrooke does this by connecting truth and en-dorsement: ‘The word “True” is the normal form of serious endorse-ment in ethics as in other branches of discourse’ (344). Now there is ahigh road and a low road between truth and endorsement. The low roadis the truth-deflationary road: ‘That’s true’ just stands in for whateverexpression ‘that’ refers to anaphorically. In the case where the expres-sion ‘That’s good,’ for example, is understood as expressing an emotionor an injunction rather than a proposition, that status is preserved evenwhen the expression is given in proxy form: ‘That’s true.’ Indeed, Bray-brooke mentions some deflationary accounts, including one of the clas-sic redundancy versions, the pro-sentential theory,11 but does not thinkthat these accounts have all the goods on truth: ‘there still seems to besome room for endorsement that cannot be occupied by merely repeat-

10 He cites Stevenson, Ayer, and Hare.11Grover et al. (1975)

12.2. BRAYBROOKE ON THE OPEN-QUESTION ARGUMENT 307

ing the original judgments or anaphorically signifying their repetition’(348).

But what is the high road? Soon after formulating the open-questionargument Braybrooke shifts from consideration of goodness to right andwrong, the moral justification (or lack thereof) of actions. Retainingfrom emotive-imperativism the objects of moral judgment, principallyinjunctions, Braybrooke then notes that injunctions, like other actions,can be justified or not. It is in an account of the nature of injunctionjustification that Braybrooke sees truth coming into the picture as morethan an element in a redundancy use of ‘That’s true.’ Endorsement isthe justification (possessing or being willing to provide same) and justi-fication requires truth: this is the high road connecting endorsement andtruth.

For Braybrooke, moral justification attaches in the first instance toparticular injunctions, ‘You, do x now.’ This attachment is secured bythree things: (1) particular injunctions are instances of general injunc-tions (rules) like ‘Everyone, always do x now,’ (2) rules can themselvesbe justified, and (3) justification somehow flows from the general to theparticular. In saying that a certain injunction is an instance of a generalrule, what we say will be either true or false. This is one way in whicha non-redundancy notion of truth appears in the account. Braybrookedoes not spend much time explaining (3) (though discussing it will oc-cupy a major portion of the second half of this paper), focusing insteadon (2), the locus of much of the substantive ethical theorizing that oc-curs in Braybrooke’s paper. Here, his account is broadly rule-utilitarian:the justification of a rule consists in the positive consequences if mostof us follow the rules most of the time. The ‘positive consequence’ issocietal thriving. Truth (in a non-redundancy sense) comes in here, too,for the second time: it will be either true or false that a certain rule,universally followed, will lead to thriving, given a determinate accountof what thriving is.

But this is the end of the line for truth and facticity: there is no factof the matter about which conception of thriving is the right one. Bray-brooke makes a case that moral justification depends on commitmentsto a conception of human thriving which can in principle vary as widely


as the limits of the human will and imagination. No one commitment, orset of commitments, is given ‘exclusive status’ (347). What I take this tomean is that, if, after giving the matter as much and as good thought asI am able, I say that I am committed to a conception of societal thrivingin which the weak and infirm are left to die, there is no replying, ‘Youare wrong: that is not a moral commitment.’ If that is Braybrooke’sview, it is one that I am sympathetic with: I also think that room mustbe found for individual, ‘existential’ choices in the foundations of moralconduct. But that is not the issue I propose to explore here; rather, I wishto explore the proposal Braybrooke has made about the formal structureof moral justification and his claim that a solution to the open-questionargument can be found in the emotive-imperative theory of normativediscourse.

We will find that the structure of moral justification Braybrooke pro-poses fits the model known in epistemology as ‘formal foundationalism’and that the solution which the EI theory affords for the open-questionproblem lies in the doctrine of persuasive definition advanced within aversion of the EI theory by Stevenson. I will argue that employing aversion of the doctrine of persuasive definition endorsed by Braybrookein the logical context of formal foundationalism leads to a difficultyfor Braybrooke’s account of moral justification akin to Moore’s origi-nal naturalistic fallacy. This result, if correct, has special significancefor Braybrooke since he states, in the very first sentence of the paper,12

that although there is life left in Moore’s open-question argument, prop-erly understood in emotivist-imperativist terms, the naturalistic fallacyis dead.

In the next section I begin the exploratory task.

12.3 Explorations

We have already seen that on Braybrooke’s reading, Moore treats anopen question as one that we cannot respond to by saying that a certain

12‘The emotive-imperative theory of moral judgments invites treatment as a response to theproblem exposed by G.E. Moore’s “open question” argument, which remains in the field after theassociated charge is refuted that it is a “naturalistic fallacy” to identify good with any descriptivecharacteristic’ (341).

12.3. EXPLORATIONS 309

claim is true by definition. In this sense Moore claimed that questionsof the form ‘I agree that x is F, but is it good?’ where F is any descrip-tive predicate, constitute an open question: we cannot settle the mattersimply by saying that ‘F things are good’ is true by definition. More pre-cisely, Moore argued that (1) any definition of ‘good’ of the form ‘X isgood = df. X is F,’ where F is a descriptive predicate, is either incorrector circular. Assuming that (2) ‘is good’ is agenuine predicate(genuinepredicates assign properties to the subjects of propositions) occurringin a genuine subject-predicate proposition, Moore argued validly thatgoodness is a primitive property, our understanding of which must comefrom some source other than explicit definition, for example, a specialkind of ‘apprehension’ of goodness. Braybrooke finds the notion of aspecial kind of moral apprehension objectionable—he does not say whyin detail, mentioning a preference for Wittgenstein (346). In any case,let us accept that moral apprehension in Moore’s sense is not a resourceavailable to us in our attempt to resolve Moore’s point about defining‘good.’

Traditional naturalists would, of course, simply reject premise (1),claiming, instead, that there is asingle correct definition of the sortMoore denies. If I understand Braybrooke’s response to both traditionalnaturalists and to Moore, it is that Moore is wrong when he says thatnocorrect, non-circular, naturalistic definition of evaluative language ispossible, and naturalists are wrong when they say that there is a singlesuch naturalistic, non-circular definition of evaluative discourse. Bray-brooke’s point takes advantage of the logical fact that in addition tooneandnonethere is the possibility ofsome. Braybrooke’s position, then,is that there could be arange of naturalistic definitions of evaluativelanguage. In the case of the language of rule justification, the range isdetermined by the conceptions of thriving to which various people usingsuch language are committed.

Now, this account may not, technically, be a form of moral rel-ativism,13 nor does Braybrooke describe it as such, but it is a closecousin. Like relativism, the definition form, if not each definition in-dividually, requires specification of a conception of thriving that is sup-

13 Thanks to Michael Hymers for helpful suggestions around the issue of relativism.


plied freely by the agent. The effect of this is that the truth of specificjudgment formulations, ‘Doing X is wrong,’ will vary relative to the dif-ferent extraneous commitments of moral agents. Whether we say thatthis variability is due to the assertions taking on different meanings ina way that systematically varies with these commitments or that it isdue to assertions having the same meanings but with an extra parame-ter in underlying logical form systematically varying with these com-mitments, the point is still that truth varies with parameters suppliedfrom outside. As evidence that Braybrooke recognizes the relativism-like characteristics of his account, I note that he spends considerabletime in mitigating the consequences of his account for those who wouldbe squeamish about moral relativism (348-9).

Finally, we come to the matter of the justification of particular in-junctions. On Braybrooke’s account these get their justification deriva-tively, by being instances of justified rules. So, showing that a particularinjunction, ‘You, do X now,’ proceeds in two stages. First, there is thejustification of the basic injunction, ‘You, do X now,’ what Braybrookecalls ‘moral judgments of the first instance,’ by appeal to the fact thatthis injunction falls under a general rule, ‘Always do X in circumstancesC’ (345). Then there is the justification of the rule itself. The followingtwo propositions represent these two stages:

A The injunction ‘You, do X now’ is morally justified in circum-stances C if there is a morally justified general rule, ‘Always do Xin circumstances C.’

B The rule ‘Always do X in circumstances C’ is morally justified iffthe general practice of following the rule will promote the thrivingof society according to a conception T of thriving.

Following Stevenson, proposition B is intended by Braybrooke to in-volve a ‘persuasive definition’ (347n22). A persuasive definition doesnot purport to report logical equivalences in a set of antecedently ex-isting concepts but is astipulationof how one will use language, theupshot of which is that we are committed to replace occurrences of thedefiniendumwith occurrences of thedefiniensin ordinary (non-meta-linguistic, non-modal, etc.) contexts. Persuasive definitions are not,

12.3. EXPLORATIONS 311

therefore, true or false. We can then take B as the definition form forwhich various specific versions are possible depending on which con-ception of thriving we specify. Different people will now choose dif-ferent definitions depending on which conception of thriving they arecommitted to. Proposition A does not seem intended as a definition butserves, rather, to express the way in which particular injunctions receivejustification from general injunctions.

There is, however, some question of exactly how proposition B ‘in-volves’ a persuasive definition. We have been operating on the assump-tion that what is being stipulated in B is the meaning of the words on theleft side of the biconditional, ‘The rule, “Always do X in circumstancesC,” is morally justified.’ This is what the solution to the open-questionproblem suggests—many possible definitions of the evaluative languageon the left side of a definition form, ‘x is EV iff x is F,’ where EV is asuitable evaluative term and F is a suitable descriptive predicate. Thisreading is also suggested by the following passage:

what . . . Stevenson called a persuasive definition, [was] inthis case, that wrong is at bottom a matter of running againstthe thriving ... of one’s society. There could be other persua-sive definitions, for example, unattractive ones like the onethat makes wrong at bottom a matter of interfering with theplans and activities of theUbermenschen, or innumerablesilly ones, like making wrong at bottom a matter of drinkingtar water and passing up beans. (343)

I will call this the ‘full-substitution’ reading of the role of stipulativedefinitions in Braybrooke’s account of moral justification. There is,however, a second reading on which the scope of substitution by stip-ulation is less than the entire left side of the definition but is, rather,confined to replacing occurrences of T in the definition.14 If I stipulatethat by ‘thriving’ I mean ‘non-interference with the plans and activitiesof theUebermenschen,’ then the only substitution that gets made is a re-placement of the former by the latter. Call this the ‘partial-substitution’reading.

14 Thanks to Richmond Campbell for suggesting it.


Each reading has its advantages and disadvantages. The full-substitutionreading gives a justification for statement B—it is true by stipulation-but generates a problem with the justification of particular injunctions,a problem I develop later. The partial-substitution reading avoids thevalidity problem but raises another: what justification can now be of-fered for the biconditional comprising statement B if it is no longer trueby stipulation? The latter problem seems the more damaging to Bray-brooke’s program, at least to its meta-ethical dimension, a dimensionto which Braybrooke assigns considerable importance (347, 350). So,until further notice, I shall operate with the full-substitution reading.

I now turn to a discussion of meta-ethics, its various forms, and whatrole, if any, it plays in Braybrooke’s account. This topic, of interest inits own right, will also serve as a bridge to epistemology and the no-tion of formal foundationalism, an account of the structure of epistemicjustification that I plan to use in section 6 as a model for interpretingBraybrooke’s account of the structure of moral justification.

12.4 Three Conceptions of Meta-ethics

I begin with a general discussion of the notions of meta-ethics and meta-language. A meta-language is a language that contains terms that referto words and contains sentences that say various things about the wordsthat are referred to by the terms. Thus, the sentence ‘This sentence hasmore than one word’ is a sentence in a meta-language because the term‘this sentence’ refers to words and this sentence says something aboutthose words. What is the relation between the conceptsmeta-ethicsandmeta-language? Braybrooke’s account of the former is brief:

In meta-ethics, the possibility of alternative definitions of‘good’ and other moral terms is firmly established. Askingthe open question of what is seized upon to define ‘good’ asin ‘Is this really good?’ amounts to asking ‘Couldn’t “good”be defined otherwise?’, which is the counterpart in meta-ethics of ‘Is this really good?’, and displaces it. (347)

The second question is meta-linguistic, involving mention of theword ‘good’ in formulating the definition; the first question is not. It

12.4. THREE CONCEPTIONS OF META-ETHICS 313

looks as if, on Braybrooke’s treatment, the distinction between ethicsand meta-ethics rests on the distinction between a first-order languageand its meta-language. Call this ‘the first conception of meta-ethics.’

But now suppose that we reformulate the definitions in a way thatuses but does not mention ‘good’; thus, ‘X is good iff ...’ Because this isnot a meta-linguistic statement, the development of this definition wouldnot be occurring within meta-ethics on the present suggestion. Thisseems unduly restrictive of the conception of meta-ethics Braybrookeis invoking since a successful outcome to the definitional task wouldstill not entail the truth of particular judgments of the form ‘Doing Xis morally unjustified/justified.’ Determining the truth of such ascrip-tions is an important task traditionally assigned to what has been called‘normative ethics,’ a discourse usually contrasted with meta-ethics. Soperhaps we can say that normative ethics is concerned with determin-ing the truth of propositions assigning the justification predicate to par-ticular actions; meta-ethics is concerned with arguing for the truths ofvarious propositions, definitions, and other such things, that do not en-tail the truth of categorical propositions assigning moral justification toparticular actions. We shall call this ‘the second conception of meta-ethics.’

It will prove helpful to compare meta-ethics so understood with an-other parallel enterprise, meta-epistemology. Meta-epistemology is of-ten regarded as the theory that deals with questions of whether andhow first-order knowledge or justification ascriptions are themselvesknown or justified. For example, if the way things look yieldsprimafacie justification for the way things are, how is the statement of thatproposition—‘If it looks as though p to S, then S isprima facie jus-tified in believing that p’—itself justified? Various answers could beproposed; and the discourse in which answers are discussed is usuallycalled ‘meta-epistemology.’ Notice that meta-epistemological discourseneed not be meta-linguistic15 (although in the present case it is), thusshowing a disanalogy with the first conception of meta-ethics; but no-tice also that the propositions established in meta-epistemology ascribe

15 For example, we can ask,If I know that p do I know that I know that p?This is a centralquestion in meta-epistemology but is not formulated meta-linguistically.


properties of (epistemic) justification to subjects, thus showing a dis-analogy with even the second conception of meta-ethics. The distinc-tion I have in mind is best characterized in terms of the language of‘first-order’ and ‘second-order’ discourse. This same language can alsobe applied to ethics. I will employ the terminology ‘second-order ethi-cal discourse’ when the issue is whether a first-order claim that doing Xis justified is itself justified. Since definitions need not be second-order(though of course they may be), definitions can be framed and deployedin first-order ethical or epistemological discourse. Second-order ethicaldiscourse therefore comprises a third conception of meta-ethics.

12.5 A Side Trip to Epistemology

In epistemology we might be interested in whether someone’sassertionis epistemically justified, whether he has a right, epistemically speaking,to say what he says. We might also be interested in whether a certainproposition is justified in relation to other statements that potentiallycould serve as evidence. Whichever subject we choose, the propertyof epistemic justification has to come from somewhere. There are twopotential sources of justification. (1) The first source consists of a trueconditional of the form ‘Given that X meets certain conditions, then X isepistemically justified,’ where X is a proposition or an assertion, takentogether with a clause affirming that the antecedent is true. The princi-ples can take the form of (a) biconditionals, in which case they count asdefinitions, or (b) one-way conditionals, in which case they are calledwarrant principles. (2) The second source for justification attachingto an assertion or a proposition is bytransmissionfrom other justifiedpropositions. The most basic candidate for a transmission principle issome variant onIf x entails y and x is justified then y is justified. Princi-ples of this kind areentailment transmission principles. I shall bypasssome of the most intensely argued controversies in contemporary epis-temology and simply postulate that the basic entailment transmissionprinciple (the one given in italics) is true.

Formal foundationalism16 is the doctrine that any demonstration that

16 The term and account is due to Sosa. See Sosa (1980) at 12ff.

12.5. A SIDE TRIP TO EPISTEMOLOGY 315

an entity is justified must proceed deductively in an orderly fashion em-ploying warrant principles to establish non-inferential justification andthen applying transmission principles recursively, first to non-inferentiallyjustified entities to produce inferentially justified entities, and then to in-ferentially justified entities to establish more inferentially justified enti-ties. Formal foundationalism is broader than traditional foundationalism—coherentism, for example, counts as a version of formal foundational-ism, but not ‘substantive foundationalism.’ (Suppose that a coherentistemploys the following definition of justification:x is justified iff x isan element of a maximally coherent set of beliefs.17 This definition canthen be used as one premise in a deductive proof that a given beliefis justified; the other premise would assert that the belief is in fact anelement in a maximally coherent set of beliefs.)

Now consider the following reasoning typically proposed by an epis-temologist interested in the epistemology of perception:18

Reasoning Pattern A

(1a) (S)(p) If it looks to S as though p, then S isprima faciejustified inbelieving that p.

(2a) It looks to Jones as though there is an apple on the table.

(3a) No counter-reasons against the proposition that there is an appleon the table are available to Jones. Therefore,

(4a) Jones is justified in believing that there is an apple on the table.

This argument is valid (given a suitable understanding ofprima faciejustification) and line (1) is a warrant principle. If all the premises aretrue we have demonstrated the truth of the first-order epistemic assertionat line (4) in accord with the protocol of formal foundationalism. (Thejustification at line [4] is non-inferential since no transmission principlehas been employed in the proof.) Of course the question arises whether

17 The position that I presenting here is a simplified version of Laurence Bonjour (1985), 87-93.

18 John Pollock (1974): the position he calls ‘descriptivism,’ 56 ff.


the premises are true. If pressed as to the principle at line (1), for ex-ample, I could, if I am schooled in epistemology, give a reply, perhapsalong the lines of

5a (1a) is justified because it leads to consequences that best accordwith our commonsense epistemic intuitions.

Line (5a) is a second-order epistemic claim, the argument for whichoccurs in meta-epistemology. But I do not need to be able do meta-epistemology in order to have demonstrated that Jones’s belief that thereis an apple on the table is justified: saying in response to a request for adefence of (1) ‘That’s another question’ is a legitimate way to terminatea dialogue without abandoning the legitimacy of one’s claim to (4a).Now consider the following pattern of reasoning:

Reasoning Pattern B

(1b) Whenever it looks to S as though p and there are no countervailingindications, then p is true.

(2b) It looks to me as though an apple is on the table.

(3b) No countervailing indications are present. Therefore,

(4b) There is an apple on the table.

Suppose that Jones believes (4b) and we ask her to justify the claim.Suppose that initially she offers the premises (2b) and (3b). But now Iam dissatisfied with the answer because the argument as it stands is notformally valid and I want to find out what premise Jones might offer toremedy this. Jones provides (1b). Finally, suppose that all the premisesare true and that Jones believes them all. Is Jones now justified in be-lieving that there is an apple on the table? Notice that the conclusion ofthis reasoning does notassertthat she is, unlike the case with pattern A.If she is justified, what would the source of the justification be? In thiscase it is an inference. Formal foundationalism says that justificationhas to come from one of two sources—a warrant principle or by infer-ence from a justified belief in accord with an acceptable transmissionprinciple. A consequence of this view is that all steps in an inference, I,

12.6. THE MAIN LINE OF ADVANCE RESUMED 317

to a conclusion, C, must be justified if C is to be justified. This meansthat the explicit representation of a case of inferential justification mustascend one level from that of pattern B (no justification property at-tributed: 0-level) to an argument containing premises that do attributejustification to statements (first-order epistemic statements):

Reasoning Pattern C

(1c) Statement (1b) is justified for Jones.


(3c) Statement (3b) is justified for Jones.Therefore (by a transmission rule),


Pattern C is the pattern of argument that is needed if inferential justi-fication is to be demonstrated of a proposition. It is important to notethat although the premises of this argument have to be true, there is norequirement that there be second-order justification—for example, thatstatement (1c) be itself justified for Jones. So, all the steps in patternC fall within first-order epistemology. Also, note that because of thedifference betweenbeing justifiedandgiving a justification, there is norequirement that Jones give the justification alluded to in each of (1c)-(3c).

12.6 The Main Line of Advance Resumed

We are now in a position to return to the main line of advance. Recallthe two stages in Braybrooke’s account of moral justification:

A The injunction ‘You, do X now’ is morally justified in circum-stances C if there is a morally justified general rule, ‘Always do Xin circumstances C.’

B The rule, ‘Always do X in circumstances C,’ is morally justified iffthe general practice of following the rule will promote the thrivingof society according to a conception T of thriving.


Sentence A does not express a definition; sentence B does express adefinition, a stipulative definition. Recall that a stipulative definitionis neither true nor false but expresses a commitment on the part of theperson making the stipulation to replace occurrences of thedefiniendumwith occurrences of thedefiniensin normal contexts. Braybrooke relieson both A and B in his account of moral warrant. First, there is the stageof an agenta endorsing the particular injunction, ‘You, do X now,’ ac-complished by appeal to the general rules. This is the first place wheretruth comes in. Endorsement expresses warrant, in this case, moral war-rant:

1 The particular injunction, ‘You, do X now,’ is morally justified incircumstances C.

Where does the warrant come from? Braybrooke’s answer: inferencefrom a general rule. Although injunctions are not propositions and donot, therefore, admit straightforwardly of logical implication, I shallsuppose that there is something analogous to the relation of logicalimplication holding between the rule, ‘Always do X in circumstancesC,’ and the injunction, ‘You, do X now,’ when circumstances C obtain.With this in hand in the theory of moral warrant there will be a proto-col analogous to formal foundationalism in epistemology. (There is onedifference: factual premises—for example, those asserting that condi-tions in a rule are met—are not, of course, justified in the same categoryof justification as is being transmitted, namely, moral justification ofinjunctions.) We can see from Braybrooke’s account that he in effectis relying on this protocol in his theory of moral warrant because hesupposes that the warrant of particular injunctions is inferred from thewarrant of general rules, and that the warrant of the latter is asserted bya warrant principle, specifically a definition, sentence B. The first stageof the reasoning can be expressed as follows:

Reasoning Pattern D

(1d) The rule ‘Always do X in circumstances C’ is morally justified.Therefore,


(2d) The particular injunction, ‘You, do X now,’ is morally justified incircumstances C,

in the presence of a principle asserting that justification is transmittedby the analogue for injunctions of logical implication. This pattern ofreasoning fits pattern C—occurring within first-order moral discourse.That is, no ascent to meta-ethics, third conception, occurs here.

Then there is the stage of endorsing the rule itself: this amounts toshowingthat (1) is true. Here Braybrooke appeals to its ability to pro-mote thriving (of the right kind) in society. If the rule does indeed pro-mote thriving in society—if it is true that the rule promotes thriving—then, according to Braybrooke, that is a second place in which truthcomes into the account:

A rule cited to warrant a moral judgment praising an honestaction is endorsed as according with the appropriate groundsor criteria for adopting such a rule, the criteria or groundsfor justifying it. Here the place of truth is secure, whetheror not what is justified by the justification are themselvestruth-bearers, as in some cases, they clearly are not. (345)

Here, there is also an argument. Taking into account the role of agentcommitment as a parameter in statements of the argument, we have thefollowing:

Argument E

(1e) The rule, ‘Always do X in circumstances C,’ promotes the thrivingof society on conception T of thriving.

(2e) The definition,The rule ‘Always do X in circumstances C’ is morallyjustified = df The general practice of following the rule will pro-mote the thriving of society according to a conception T of thriv-ing, is selected by agenta.Therefore,

(3e) The rule ‘Always do X in circumstances C’ is morally justified foragenta.


Line (2e) is the warrant definition; line (1e) provides an auxiliaryempirical premise, yielding the conclusion, line (3e). Line (3e) is non-inferentially justified in the formal foundationalist sense. (No trans-mission principle is used here.) Note that even at line (2e), the placewhere Braybrooke finds an excursion into meta-ethics (350), there isno second-order attribution of justification, hence no meta-ethical dis-course conceived in the third way.

We now combine the two stages into a single argument. A moralagent, making this argument explicit, would say the following:

Argument F

(1f) The rule, ‘Always do X in circumstances C,’ promotes the thrivingof society on conception T of thriving, which I accept.

(2f) I hereby stipulate that the words, ‘The rule “Always do X in cir-cumstances C” is morally justified,’ will mean the same as thewords, ‘The general practice of following the rule will promotethe thriving of society according to a conception T of thriving.’Therefore,

(3f) The rule, ‘Always do X in circumstances C,’ is morally justified(1f, 2f, commitment inherent in making the stipulative definition).

(4f) I hereby make the injunction ‘You, do X now.’

(5f) Circumstances C obtain.Therefore,

(6f) The injunction, ‘You, do X now,’ is morally justified. (3f, 4f, 5fand a suitable transmission principle for injunction-warrant)

This is the theory of moral warrant that Braybrooke set out to achievecharacterized in terms of formal foundationalism. Apart from relativelyminor disagreements around the issue of meta-ethics, my verdict is, ‘Sofar so good.’ But I now wish to make a more serious criticism of theproposal, centring on the special characteristics of stipulative defini-


tions.19 Stipulative definitions, as we have seen, are conventions, com-mitments to replace one set of words (thedefiniendum) with another (thedefiniens) in normal contexts. The underlying logical form of a stipula-tive definition is not a biconditional but an injunction. In the case of thestipulations regarding the language of rule justification at issue here, thecommitment is this:

Commitment made in stipulating a definition of rule justifi-cationThe stipulator is committed to replacing, in normal contexts,the words, ‘The rule, “Always do X in circumstances C” ismorally justified,’ with the words, ‘The general practice offollowing the rule will promote the thriving of society ac-cording to a conception T of thriving.’

It follows from this commitment that our agent is committed to re-placing the words, ‘The rule, “Always do X in circumstances C” ismorally justified,’ in line (3f) with the words, ‘The general practice offollowing the rule will promote the thriving of society according to aconception T of thriving,’ yielding

(3f*) The general practice of following the rule will promote the thriv-ing of society according to a conception T of thriving.

But now the inference expressed in argument F*—argument F* is ob-tained from argument F by replacing (3f) with (3f*)—aborts. Note thatargument F successfully confers inferential justification on the partic-ular injunction at line (6f), as intended. This is because argument Fcorrectly employs a transmission principle to get from (3f) to (6f). Butargument F* fails to confer inferential justification at line (6f) for the

19 Mike Hymers has raised doubts about whether I have the notion of stipulative definitionright here. He suggests four alternative accounts: that the definition might be simply ‘latent’ inour moral practices; that we might invoke a theory of ‘implicit definitions’; that we might simplybe able to re-substitute the left side of the definition for the right after having done the reversefollowing our commitment in making the stipulation in the first place; and that the definitionmight be understood as a ‘meaning postulate’—an analytic statement—in Carnap’s sense. All ofthese are interesting suggestions but it is not clear that they are primarily suggestions for me ratherthan for Braybrooke. In any case, they all require careful responses which shall have to be left foranother occasion.


reason that there is no justification present in line (3f*) to be transmit-ted.20 (Readers will note that this problem is akin to that identified byMoore under the heading of ‘the naturalistic fallacy.’) So Braybrooke’saccount of a structure of reasoning by means of which we can showhow particular injunctions acquire their moral justification, formulatedwithin the protocol of formal foundationalism and in terms of the full-substitution reading, fails.

There is, however, as noted above in section 2, an alternative read-ing of the scope of substitution in the stipulative definitions, thepartial-substitutionreading. On this reading we are required to replace, not thewhole left side by the whole right side of a definition, but only occur-rences of the thriving parameter (T) with specific kinds of thriving towhich our speaker is committed. In the case of argument F this meansthat the speaker is not committed to replacing 3f with 3f*, so a crucialmove in the objection fails: the account of how particular injunctionsacquire justification on the formal foundationalist model still stands, inthe form of argument G:

Argument G

(1g) The rule, ‘Always do X in circumstances C,’ is morally justifiediff the rule, ‘Always do X in circumstances C,’ promotes societalthriving.

(2g) I stipulate that by ‘societal thriving’ I mean T.

(3g) The rule, ‘Always do X in circumstances C,’ is morally justifiediff the rule, ‘Always do X in circumstances C,’ promotes T. (sub-stitution by stipulation)

(4g) The rule, ‘Always do X in circumstances C,’ promotes T.

20 I anticipate the following objection: ‘A correct definition, “A = df B” is alwayssymmetrical—one can replace the left side by the right, but also the right side by the left. So, in thepresent case, we simply replace f* by f, thus re-introducing the required injunction-justificationpremise.’ I reply that this holds only for reportive definitions, not stipulative definitions. The for-mer are, if successful, true necessary bi-conditionals; the latter are not true statements at all, henceare not necessarily true biconditionals. They are, rather, commitments to act in certain ways—andthe ways are asymmetrical.

12.7. AN ALTERNATIVE BRIEFLY CONSIDERED 323

(5g) The rule, ‘Always do X in circumstances C,’ is morally justified.(3g, 4g)

(6g) I now say: ‘You, do X now’; and circumstances C obtain.

(7g) The injunction, ‘You, do X now’ is morally justified (5g, 6g, anda suitable transmission rule).

However, this benefit is achieved at a cost—as mentioned in section 2,we no longer have as an answer to the question, ‘What justification canyou offer for (1g)?’, that it is true by stipulation. The challenge, arisingwithin meta-ethics (third conception) is,With what are we to replace it?

12.7 An Alternative Briefly Considered

I wish to pursue, very briefly, an alternative formulation of the theoryof warrant being pressed by Braybrooke that avoids this problem. Thisformulation takes note of the fact that we often attribute instrumentallyevaluative predicates to individuals in relation to certain sub-categorieswithin a general category that they exemplify. Thus we can speak ofa ‘good sports car,’ meaning that it is goodas a sports car(thoughnot, perhaps, as a touring car.) The notion of ‘good car’simpliciter isundefined. In the spirit of this we might think of defining what it is for agiven general injunction to be a good moral injunction. On the presentsuggestion this means that the injunction would be goodas a moralinjunction, good for the purposes which such injunctions serve. At thispoint one might plausibly have recourse to the conception of naturallaw underlying Braybrooke’s program in the present paper, namely, thatmoral injunctions, when followed, serve to promote societal thriving.An injunction, ‘Always keep promises,’ for example, might plausibly bethought to serve such purposes, and thus be good as a moral injunction.This could be phrased alternatively by saying that the rule, ‘Always keeppromises,’understood as a moral injunction(as opposed to some otherkind—prudential, authoritarian, etc.) is justified. Unfortunately, relativeinstrumental justification of this sort does not seem to be transmissibleto particular injunctions, for it does not follow that because a general


rule, when followed, will promote societal thriving, so will a particularinjunction when it is followed.

I conclude that the task of providing a theory of moral warrant withinthe protocol of formal foundationalism, based on the EI theory of ethicsand adequate to dispelling Moore’s allegation of a naturalistic fallacy,remains still prospective.

Chapter 13

David Braybrooke On the Track ofPPE

PETER K. SCHOTCH

Abstract

David Braybrooke has long been interested in the project of applyingmoral philosophy to social science. In this essay I attempt to show howvarious bits and pieces of his life’s work actually might be seen as accom-plishing that far-reaching aim. Part of the difficulty which others haveencountered with such central pieces as the theory of rules presented inLogic on the Track of Social Change, has more to do with the formal-ity of the presentation of that material (which the authors admit demandsat least some technical facility on the part of the reader and a consider-able store of good-will besides). In order that these same difficulties notreappear, the Track theory is presented in a (new) graphic format whichsimplifies the concepts but not to the point of triviality. Within this newlyminted approach, various issues are raised including a novel presentationof some well-trodden moral theory and the place of political economy ina rule-based picture of morals.

13.1 Introduction

When I came to Dalhousie in 1972, my relationship with David Bray-brooke went through a shake-down period during which we were bothwondering what to make of each other. I was a brash young man and

325

326 CHAPTER 13. PETER SCHOTCH

David was the established major figure—the power of the departmenteven though he wasn’t then (nor was he ever) the administrative leader.

We each had misgivings about the other, although perhaps misgiv-ings is too strong a word since it implies a lack of openness, or perhapseven of goodwill and neither of those obtained. For his part, David hadbeen dismayed by certain trends in the technical end of philosophy—trends away from applications and toward the philosophical equivalentof so-called pure mathematics. For my part, I had been similarly dis-mayed by the eagerness of some moral and other philosophers to dis-miss my work as not being ‘real philosophy.’ So the stage was set forus to start off on the wrong foot.

Instead, I discovered that David was glad to have my counsel oncertain matters that involved deploying formal methods in some of hisprojects, and David discovered that I was delighted to make whatevermutations seemed necessary in order that some existing formalism ac-tually have a useful application. Both experiences were novel to each ofus.

At a later stage in our cooperation, we moved into an entirely newmode in which the formal work did not consist of applying (perhapswith some modifications) an off-the-shelf theory, but the constructionof a new theory virtually from scratch, the shape of which was dictatedby the application, or series of applications. In a project likeLogic onthe Track of Social Change(referred to as Track in the sequel) someformalism would be proposed to fit initial requirements and then mightchange substantially as the requirements were refined over time.

What I intend to do in this essay harks back to the very first time Iheard David Braybrooke speak publicly about his philosophical aims.On that occasion he said that he was interested in the possibility of ap-plying moral philosophy to social science. I was struck then by thescope of this project and remain stricken to this very day.

In what follows I shall revisit this ambition to apply philosophy. Inthe version I preach, it will be the philosophy mainly expanded in Trackand some of its successor pieces. It might be a stretch for some to acceptthat this philosophy is in fact moral philosophy though there is this tobe said for it: For the most part it will be philosophy in which David

13.2. HOW TRACK CAME ABOUT 327

Braybrooke had a hand or at least an interest.Though much of what I say will bend in the direction of the techni-

cal, I will present the latter in graphical, rather than the more forbiddingalgebraic, form. In the interest of readability, if not comprehensibil-ity I will content myself with a gloss whenever that is possible withoutproducing a mere caricature.

13.2 How Track Came About

It’s all very well to give an account of how the various bits of Track andsome of its later exfoliations connect with David Braybrooke’s work inthe larger sense, but what sort of position should we grant Track itself?Is it

a project to which Braybrooke committed himself at a fairlylate stage in his career—a project, moreover, that can (ar-guably) be detached (at least very largely) from the projects(in moral, social, and political philosophy) which preoccu-pied him in earlier periods.

No it isn’t. In fact this view of things, though representative of abody of opinion against which Track labored in vain, is remarkably outof touch with Braybrooke’s work. We knew on the day that Track wouldchallenge a number of accepted ways of doing things but we hoped thatwith enough of what we called ‘good will,’ people could be brought tosee that the struggle with the formalism was worth-while in the end. Un-fortunately, it must be admitted, we seriously over-estimated the amountof good will upon which we might draw.

The Track project grew out of Braybrooke’s long time interest in thenotion of a rule in general, and of a social rule in particular. Anybodywho thinks this interest may be ‘detached’ from the rest of Braybrooke’sthought without any great violence to the whole, hasn’t done his home-work. But let us do ours.

That David Braybrooke’s interest in rules dates from his earliestwork, needs no further documentation. Instead we need to show thata willingness to employ formal methods in pursuing an account of rules


is not something which belongs only to some hypothetical ‘later period.’Providing one is not prepared to date such a period to include 1969—avery brave exegetical hypothesis that would be, we need only notice theuse that Braybrooke makes of the formalism of G.H. von Wright1 in hispaper Braybrooke (1969), and in his Braybrooke (1970).

In a nutshell, the program of Track is to bolster these earlier attemptsat showing how the notion of a rule can figure in certain kinds of his-torical explanation. In fact the various case studies that form the core ofTrack are very much to that point.

But there is another reason for Braybrooke to look for aid and com-fort among the formal philosophers: Though many and various are thosewho seek to call upon the concept of a rule to do this or that theoreticalwork, fewer by far are those who are prepared to offer adefinition. Nowof course there are bound to be those who say that we know perfectlywell what a rule is and don’t require a definition, especially a formalone, but Braybrooke is not among their number.

The first attempts to define a rule, as for instance ‘a pattern of norms’or ‘a context dependent family of imperatives’ would seem, on no verydeep appreciation of the concepts involved, to be circular. So rule talklooks to be all of a piece with norm talk and imperative talk, whichmakes things messy, to say the least. It also explains why so few philoso-phers attempt a definition. It is at exactly this point, the point wherethings are messy, that a formal approach can be helpful. It is quite truethat rigor is sometimes the enemy of clarity, but when there is no clarityto begin with, it is sometimes wise to fall back on rigor. And claritymay appear in the end after the rigorous approach has been pursued farenough.

In fact, we contend that this is precisely what happened in the case ofrules. It is not so easy for the casual reader to detect this since in Trackthe ‘order of reasons’ gets scrambled, as it nearly always does once wego from the research to the write-up. The chapter, largely informal,which introduces our (non-circular) account of rules comes before theformal chapters which introduce the logic of rules. Historically it wasthe other way about, and it is no exaggeration to say that we wouldn’t

1As developed in his von Wright (1963)

13.3. THE BASIC SET-UP 329

The Set of Alternative Social States

Figure 13.1: The Space of Alternative Social States

have hit on the former were it not for our struggles to get the lattercorrect, or at least not obviously incorrect.

13.3 The Basic Set-up

So let’s do some social science. We shall conceive of our subject aswidely as possible in terms of a certain set of objects—the objects stud-ied by social scientists. Now of course this is a considerable idealiza-tion, that all of the social sciences are really concerned with one sortof object, but perhaps the stretch isn’t quite as broad as one might firstexpect. These objects we shall call, following K. J. Arrow, the set ofalternative social statesunderstood as the objects with respect to whichthe state-variables, however we conceive these, receive a value.2

2There is an analogy to be made here with thephase spaceapproach to physical science.


13.4 How to Apply Moral Philosophy

Morality, at bottom, is an account of what we ought to do. Obviously itis situational—it would be a bizarre moral code which, in every set ofcircumstances demanded precisely the same action on our part. Some-body might say that morality does exactly that, by demanding that wealways do the right thing, but this is disingenuous in the extreme. Half,or more of the battle is the determination of which thing (or things)areright—in this or that circumstance.

There are different kinds of morality and hence moral philosophysprings into being. It might be better to say that there are many appar-ently different moral codes. Or perhaps “decodes” would be better still.One single moral principle ”The greatest good for the greatest num-ber”, or ”Harmony with nature” etc. might be understood, at the levelof ordinary life, in dozens of ways which at least seem distinct. Theseunpackings of over-arching principles are what we typically mean whenwe talk of moral codes.3

Now we get an inkling of how moral philosophy can be applied tosocial science: The latter describes the ‘phase space’ of alternative so-cial states, while the former tells us what we ought to do in this or thatcircumstance. This is to prescribe an action (or series of actions)giventhat we find ourselves in an alternative social state which satisfies somespecified condition or set of conditions.

13.5 Actions

But doing things, either what we ought to do or some other thing, re-sults in moving from one alternative social state to another. So knowingwhat to do (or at least what we ought to do) is tantamount to knowingwhich alternative social state is to be (or at least ought to be) selected asour new home, given that we find ourselves in such-and-such a presentalternative social state.

3And it might be no exaggeration to say that how exactly a principle is to be unpacked is thefundamental problem of practical ethics.

13.5. ACTIONS 331

Starting State

Intermediate Set

Termination Set

Figure 13.2: The Ice-Cream Cone Semantics of Action

The picture just presented demands clarification and qualification,much of which will require another occasion, but this we can say rightaway: It is not, in general, possible to specifyin advancewith precisionjust what the consequences of (a given agent) performing a certain ac-tion (actiontype) will be. The best we can do in ordinary circumstancesis to specify somerangeof states within which the result of performingthe action type in question will lie.

The semantics of action is thus a complicated business and not, per-haps, for the faint of heart or the impatient. But in keeping with ourpresent graphical slant on things, one picture will answer all but themost determinedly technical questions.

Indeed this picture explains why the approach is often described asthe ice cream cone semantics.

The apex of the cone, it’s originating point, is what we call thestart-ing stateof the action. The cone proper contains the only states which (ifany can) can be active while the action is being performed. We some-


times say that these are the only states in which the action might berunning once it has been started (if it starts at all) in its starting state.This should not be taken to mean that each of the cone stateswill in factbe visited, only that no other state will be. Similarly, the ice-cream partcontains the only states which can result (if any do) after the action isperformed. We might think of the ice-cream states as the ones that canbe reached (if any can) by performing the action beginning in the start-ing state. Similarly the cone states are those through which we mightpass as a result of performing the action beginning in the starting state(if indeed the action can begin there). This picture of the action is de-cidedly non-deterministic but that can be taken to reflect only our lackof knowledge about what the other actors will do in the circumstancescontemplated by the picture we here present.4

Compared to the literature (on the formal theory of actions), thisaccount is new in adding the cone. In other words, previous accountscharacterize actions entirely in terms of which states result once the ac-tion had been performed (providing that it can be performed—there isno guarantee of this in general). So for these theories, actions are essen-tially and only what we might termbringings about. We, on the otherhand, characterize actions not simply in terms of what they bring about,but also in terms ofhow they bring about whatever it is, if anything,that they do. Our view is perhaps closer to common sense, even thoughsuch a remark is sometimes thought pejorative in connection with for-mal work. For us, two distinct actions might bring about precisely thesame result. This would also seem to be the burden of a number of oldsayings many of which involve violence to cats.

Just what happens when we string a series of actions together (byconcatenationas we say—with the resulting object often referred to asa concatenate) is nearly but not exactly what you think will happen.Another picture will help.

In this we see that the cones (intermediate statesto give them theirtechnical title) get aggregated into a kind of super-cone for the con-catenate. In addition, all the ice cream of the last actions in the se-

4This leaves open the question of whether or not the non-determinism runs deeper than theepistemic—an important question no doubt but one that we have been glad to dodge.

13.5. ACTIONS 333

��

��

INTermediate set for concatenation

TERMination set for concatenation

not part of the concatenation

��

��

��

��

��

...

...

...

Figure 13.3: Concatenating Actions

quence (termination statesin the parlance of Track) gets lumped intoone object—a sundae perhaps? This leaves open only the fate of the in-termediate bits of ice-cream, the ones that are not shaded in the picture.

They, it might seem a trifle odd to say, are not part of the concatenateat all. There is a technical reason to prefer things this way5 but we shallappeal instead to the intuitive principle that a concatenation can onlybe said to be running when at least one of its ‘factors’ is running. Butat the states in question one action is terminating and its successor isstarting. Since the concatenate is not running then, these states shouldn’tbe counted as intermediate for the concatenation.

The entire semantic account of an action type, sayr, is the specifi-cation of its ice-cream cone given the selection of any alternative social

5The technical reason is that only by casting out these states will the virtual actionskip,mentioned in the next section, serve as what the algebraists term anidentity. This identity, alongwith the fact that the operation of concatenation is associative, endows the set of actions with aninteresting algebraic structure.


Figure 13.4: An Non-Terminating Action

state as a starting state. “But why” you will want to know ‘should anaction type evenhavea cone for every social state?’ Indeed that wouldmake action types rather more abstract than we would like. We mustthus allow cases in which a given starting point does not have a cone, orat least one of the usual sort.

13.5.1 When good actions go badly

The standard pathology of actions is confined to those whichdo not ter-minate. These are actions with no ice-cream so-to-speak and that’s theirgraphic representation. And if actions are conceived as the bringingsabout (of the truth or falsehood of propositions for instance), then theseare the only actions which fail.

But in our account of action types there is the cone to consider aswell. In this case there is clearly the possibility of another form of‘misfire.’. It might be that the action in question cannot even be started.

13.5. ACTIONS 335

Figure 13.5: A Non-Starting Action

Such an action has no cone at all, and is represented graphically by abackwards facing cone.6

An important property of actions which is studied in Track but notelsewhere in the literature (at least in the literature on formal accountsof action) is that one action mayblockanother.7 The basic idea is simpleenough: the actionr1 mightBLOCK the actionr2 in one of two ways.In the first way, attempting to performr2 after performingr1 (which isto say graphically at any point in the ice-cream portion ofr1) either willnot terminate or will not start. In the second way, attempting to perform

6This raises the issue of whether or not an action with no cone might nevertheless have someice-cream. Such a thing sounds bizarre and in Track such actions are calledvirtual. In point of factthe literature contains at least one such action called (in Track and elsewhere)skip. This actionalways terminates in the unit set of the state in which it was started and traverses no intermediatestates in its journey. To put the matter whimsically,skiphas always and everywhere just finishedbeing performed although nowhere is the performance of it actually taking place.

7This idea appeared in the early stages of research which was later reported in Track, and atfirst seemed nothing more than a minor curiosity. As we began to exercise the Track account ofrules however, it soon became clear that blocking is a central and crucial notion.


r2 while at the same time performingr1, which is to say graphically inthe cone portion ofr1 does not terminate or does not start.8

13.6 Rules

Which brings us to the center of Track and to the notion that must bear agreat deal of the weight when morals meet science—the theory of rules.

The theory as presented in Track might be termed ‘forbidding’ inmore than one sense, alas. In the first sense, the introduction of the ideaof actions blocking one another allowed a simplification of the theoryin that all rules can be cast into the form of rules of forbidding. Anearlier effort9 had already reduced all rules to those in the top row of thesquare of opposition—forbidding and permitting. But with blocking inour arsenal, we can parse rules which permit the actionr as rules whichforbid all those actionsr∗ which blockr.

Unhappily the theory also seems have been forbidding in that it pre-sumed too much in the way of formal science to be easily digestible byits target audience.

Given our current graphical turn however, we can see what a rule is,at a glance. Consider figure 13.6.

Here we see the universe of alternative social states from the per-spective of a particular subset of them, those which satisfy the conditionC, of a certain collection of rulesFi. These are the states, theC-states,where the rulesapply. We refer to actions that start inC asC-state ac-tions. What eachFi does, in simplest terms, is to forbid certainC-stateactions (to a certain class of agents, those within the so-called demo-graphic scope ofFi. To avoid complicating the exposition, we typicallysuppress mention of these agents).

The darkened areas of the social universe, represent the forbiddenactions according to the collectionFi in that the only way to reach any

8The illustration of these two in Track is in terms of tying shoe laces. You can block mytying my shoe laces by untying whichever lace I have just tied. So long as I perform my untyingaction, your tying action cannot terminate. This is an example of the second way of blocking.Alternatively you might make away with my shoes while I take a reckless nap which will ensurethat my tying action cannot start.

9By Charles Hamblin, see Hamblin (1972)

13.6. RULES 337

��

��

��

The set of states which satisfy the condition C

Figure 13.6: Instantaneous Picture of a Rule

darkened state is by means of an action which someFi rule forbids. ‘Ah’you say, ‘but what if the actual social state, our current state lies in oneof the dark zones?’ (see figure 13.7) Well, that would be a bad thing—so bad indeed, that we have a name for it—aquandary. Quandaries areto rules what contradictions are to propositions.10

It must be emphasized here that figure 13.6 is not a picture or dia-gram of the effect of a set of rules—it is instead a snapshot of the actionsforbidden by the rulesrelative to a certain conditionand acertain col-lection of agents. One might postulate the existence of certain rules thathold for all agents under all conditions—but this is no part of the Trackdoctrine.

As we trace out a certain trajectory in the social space which is tosay, move from one social alternative to another, we also pass from one

10In practical terms, one would be as bad off as somebody in a quandary if one’s social statewere entirely surrounded by a dark zone as a single candle burning in the darkness. In such asituation there is only one action not forbidden, but it is the virtual oneskip.


��

Figure 13.7: A Quandary Situation

condition to another. As this happen the class of forbidden actions and,it goes without saying, the class of agents bound by the rules, change.So a picture like figure 13.6 is really only one frame in a movie—and itis the movie which is the true representation of the rule.

13.7 Utilitarianism

For much of his philosophical career David Braybrooke has been inter-ested in utilitarianism. Given his interest in rules, we can see at least onereason for such an interest. To put the matter in the whimsical idiom ofthe previous section, one wants to know whether or not the movie has ahappy ending.

The thing is, a given society of individual agents who are bound bya sequence of rules (which come into effect as they progress from onealternative state to another) will, since portions of the state space areforbidden by the rules, tend to move to the other states, the ones which

13.7. UTILITARIANISM 339

are not forbidden. At least this will happen if sufficiently many peoplewho are bound by the rules, actually comply with them.

Waiving that (interesting) consideration for a moment, suppose thatthe society in question follows the rules as they evolve from one frameof the movie to the next. The question is: what should happen? Orbetter: If the rules are properly constructed, what should happen? Toput a moral slant on things: what would be a good outcome to followingthe rules?

This is a bit tricky but only only a bit and only because we aren’t yetequipped with the concepts we require in order to adjudicate the matter.The first thing we need is the notion of agoal. In this context a goal is asubset of the space of social alternatives. Which subset? Evidently, onein which we would like to live, one in which we want the actual socialstate to be. We shall shortly take up the issue of how we decide on ourgoal, but first we may need to be convinced that such a notion actuallyhas a role to play.

It should be clear that the function of rules is to restrict which alter-native social states are ‘accessible,’ which is to say that following therules amounts to staying within a certain corridor. But what if that cor-ridor doesn’t lead anywhere? What if following a certain set of rulesmeant only that any alternative social state was as likely as any other asour destination (subject only to the laws of nature)? To put the questionmore bluntly, what incentive would we have to actually bind ourselvesto following such rules?

There has been and remains an intuition on which it may turn out thatfollowing a rule is less advantageous in the short-run than not followingit, but even so, following is the right thing to do. An extreme versionof this view has it that some rules are just correct, pure and simple.Following them is always the right thing to do and not following themnever is. It would seem then such folk as these would answer the bluntquestion that our incentive for following rules can never be more or lessthan the rightness of doing so.

At this point we might ask for a list of the right-making propertiesand a great controversy will ensue. Instead, we shall take the positionthat even if the journey be worth more than the destination, the latter has


to count forsomething. In particular, there must actuallybea destinationas opposed to a random walk through social space. This separates usfrom the extremists just canvassed.

It also brings us closer to something resembling utilitarianism. Oncewe restrict our attention to those collections of rules which have whatwe called above, a goal,11 there remains the question of how to adju-dicate between rival collections. Evidently we can raise the questionof whether one goal is preferable to another and as soon as we do that,we have become some flavor of utilitarian. Strictly speaking there is awhole range of alternatives here only some of which are ‘really’ utilitar-ian. What we require is some evaluation of the alternative social stateswhich is to say a preference ordering.12 How utilitarian we are dependsupon how this preference ordering is used.

I think David Braybrooke’s contributions to this part of the accountare numerous. One which we might highlight first is the notion that thegoal, whatever subset it eventually turns out to be, has a kind of core.In other words in order to belong to the goal set of social alternativesthere is some list of minimal conditions which, while it may not, indeedprobably doesn’t, fully characterize the goal, yet gives us a starting pointfor calculating which states are likely candidates. Perhaps we can saythat although we may not be in a position to identify the goal with muchprecision, we are prepared to say immediately that social states of such-and-such a kind will definitelynot belong.

I look at Braybrooke’s work on needs as being of this sort. What-ever other properties a state satisfying a (social) goal might turn out tohave, within the goal the needs of the individual members of societywill be met according to some well-defined standard of provision. Infact the book on needs makes a useful beginning as a specification ofsuch standards and indeed spelling out just what sort of thing is a need,and telling us why the concept of needs and paying attention to them canbe seen as central to what we (but not Braybrooke) have been calling a

11There is another technical matter concerning goals. A goal should be not only a subset of thestate space which we eventually reach, but a subset which once entered, we do not subsequentlyleave—at least not without a change in the rules.

12Utilitarians, of the later generations at least, came to regard preference which a jaundicedeye. They it were who coined ‘Better Socrates unsatisfied than a pig satisfied.’

13.8. WHERE’S THE POLITICS? 341

social goal.

13.8 Where’s the politics?

Even if we can agree on the core properties of the goal states, it mightwell be necessary to go beyond those properties to specify an additionalcollection of features. In particular, there may be political-economicreasons for such a thing.

Think of it this way: We describe a subset of the space of social alter-natives in which we shall (eventually) arrive provided enough membersof society act in accord with the rules in a certain collection. We cansee, in the abstract perhaps that the rules should be such as to secure theprovision of needs, but will that be enough? Enough, that is, to ensurethe amount of compliance required to reach the destination subset. Itmight fail—human nature being what it is.

In order to ensure a sufficient level of compliance we must then takefurther measures. Moral education of a more or less strenuous sort hasbeen popular in this regard (one might think of the so-called culturalrevolution in China as an example) though not always efficient. Beyondthe softer measures are rank upon rank of harsher, and more expensivemethods. After a certain point, the game is no longer worth the candle—once the Gulags and the Dzerzhinsky squares appear, for instance.

In such a case we must find another way to convince (sufficientlymany) members of society to follow the rules. And what other way isthere, when coercive measures have been put aside13, but to appeal tothe preferences of sufficiently many? Evidently what has to be done is tofurther shape the goal (which will also require tuning up the collectionof rules).

But havent we now set ourselves an impossible task? After all K.J.Arrow has shown that blah blah blah and we all know whatthat means.Individual preference orderings cannot be aggregated in any convincingway. Or does it?

I am inclined to think that we have been coming at the problem thewrong way around. In the first place we cannot (or at least should not)

13not that they are ever entirely put aside


prove that social choice is impossible since it happens all the time. In-stead of first trying to define the social preference ordering , and thenchoosing alternative states on the basis of that ordering, we should askhow we might go about choosing the states. We can then see what socialpreference ordering (if any) is implied by this choice, and let the chipsfall where they may.

So how do we actually decide on social policy? How do we decidewhich kinds of alternative social states to aim for? It is a complicated,but not entirely impenetrable matter. Let us idealize it a bit but not toan absurd degree. And once more we can call upon the vision of DavidBraybrooke who has emphasizedissuesand made interesting conjec-tures concerning how issues are processed in a social setting.

The basic idea in our idealization, is that of anagendawhich isa collection of issues, each of which is represented by a proposition(which is part of the idealization). The propositions are considered inturn by a number of parties either all the individuals in the society or,more usually, their representatives. These representatives consult theirown preferences in order to judge whether or not any given propositionin the agenda is preferred to its negation.

This is another tricky technical matter since we have hypothesizedonly the preference orderings defined over states and we have now toparlay that into an ordering defined forsetsof states (which is what,semantically speaking, propositions are). Suffice it to say that such a“type-raising” is indeed possible.14 So each member can decide forherself whether she would prefer that the proposition P, for instance, betrue rather than false or vice versa or finally whether she is indifferentbetween the two.15

If we were to presume, which we assuredly do not, that perfect in-formation holds, then we could allow an agenda to contain every propo-sition. But that would be an idealization extending into what we called

14this matter is treated in more detail in Schotch (2001).15An advantage to considering propositions—sets of states as opposed to individual alternative

social states is that a number of different individuals can have wildly varying preference orderingson the states and yet have wide agreement on the orderings of sets. This fact makes it moreplausible that a relatively small number of representatives can accurately reflect the preferences ofquite a large constituency.

13.8. WHERE’S THE POLITICS? 343

the absurd.In our picture, a group of adjudicators, decision makers, or the rep-

resentatives of the interested parties (who are themselves, we presumealso interested parties) confront the list of propositions (issues) on agiven agenda. Some of the propositions might well be agreed to by ev-eryone (or their negations agreed to) and those go on to form part of thegoal. Many, it may be presumed will be preferred by some the nega-tions of which will be preferred by others. Here there is the possibilityof so-called log-rolling. You care nothing for P one way or the other butare committed to Q while for me the two attitudes are reversed. Thereis the material for a bargain here. Further bargains might result frommy weak preference for P retiring in the face of your strong preferencefor not P, so long as my strong preference for Q similarly trumps yourrelatively weaker preference for its negation.

Eventually we get as far as we can go. What are left are the issuesconcerning which we can not secure agreement. Some of us prefer Pas a matter of deeply held principles while the others prefer not-P on asimilar basis.

Various means have evolved to produce a decision in these casestoo but we shouldn’t ignore the possibility of simply dropping the morecontentious issues from the agenda. If the value of the bargain con-sisting of the package of issues already agreed to is sufficiently high,then there is that much incentive to find a way to complete the bargain.When strongly held principles stand in the way of a valuable bargain,we don’t discard the principles (for the most part) but that need not pre-vent us from discarding the occasion on which the principles becomeinconvenient.

An added incentive to pull back from a game of ‘duelling principles’is that this whole procedure, from agenda formation to final decision, inone that is going to take place again and again. If our principles clashtoday, who’s to say that tomorrow you won’t be converted to my pointof view, or I to yours. This highlights another of David Braybrooke’skey insights—the value of an incremental approach to policy formation.

If we allow the goal to be formed by a sequence of processes like theone just described, then it will be a moving target, although there must


be some limit on the velocity of its movement or the process won’t work.If all of our preference orderings changed wildly every day, then giventhe cost of computing (a stage of) the goal and of setting up a set ofrules which conduce to that goal, it would be too expensive to attempt.But, in real life, our preferences do not change wildly and the idea offorming a goal in these stages while at the same time reconstructing thecollection of rules so that its goal matches the one we are computing,seems if not a clear match to some actual procedure, then at least aninteresting idealization.

It seems that in the new science of issue processing coupled withthis new account of rules, we have come at last to something like thefoundations of policy analysis.

Chapter 14

Bootstrapping Norms: FromCause to Intention

BRYSON BROWN

Abstract

In Logic on the Track of Social Change, David Braybrooke, Peter Schotch,and I developed a theory of rules, including both a definition of rules anda logic explicitly designed to enable us to express their contents in an il-luminating way. This paper sketches an account of the relation betweenthe normative and the descriptive elements of that theory. The accountdraws on a bootstrap argument due to Sellars, who used it to ground anaccount of how an agent can come to be aware, and justified in asserting,that she is a reliable observer. Sellars’s argument relies on a distinctionbetween descriptive assertions about episodes and their normative inter-pretation. This paper relies on the same distinction to explain how anagent can come to be aware, and justified in asserting, that she isnorma-tively engagedwith a system of rules.

14.1 Introduction

In Logic on the Track of Social Change(1995; hereafterTrack), Bray-brooke, Brown, and Schotch present a definition of rules that includesboth causal and intentional elements. Our definition appeals to a tech-nical notion ofimperatives, which we identify asblocking operations.Blocking operations must be causally efficacious—that is, they must ac-tually (in some relevant and substantial range of cases) prevent actions

345

346 CHAPTER 14. BRYSON BROWN

of a certain type on the part of some agent(s). But they must also beintendedto have this effect. A rule is identified with the class of impera-tives (whether actual or merely possible) targeting the actions forbiddenby the rule.

How this blocking is accomplished varies with the blocking opera-tion employed—with young children, direct physical intervention (forinstance, holding on to a child to prevent her running out into the street)is fairly common. Even with adults, this form of blocking may be used,for instance, when trying to correct a flaw in how an athlete performs aswimming stroke, a dance step, or a gymnastic move. But saying ‘don’t’or ‘no’ or ‘stop’ can also be blocking actions, since some agents in somecircumstances respond to these utterances (on the part of certain otheragents) by not doing something they would otherwise have done. Andof course there are many more examples. In all these cases, however,the action qualifies as a blocking action if and only if it is both of a kindwhich, with a fair degree of regularity, causally prevents agents fromdoing actions of the targeted type, and is intended to have that effect.

Our principal aim inTrackwas to develop and apply a rich, formallystructured account of rules. The resulting account, linking rules andcausal relations without reducing the first to the second, links two dis-tinctive approaches to the social sciences. An emphasis on the impor-tance of rules is characteristic of what Braybrooke callsinterpretiveso-cial science.1 But a focus on regular correlations and causal hypothesesinvoked to explain them is characteristic of what he calls thenaturalisticapproach. Drawing on both, we proposed an approach to rules that inte-grates the descriptive evidence of regular correlation and causal analysiswhile allowing an important role for intentions as well.

However,inTrack we left the relation between these causal and in-tentional elements unexplored. In this paper, I examine both sides ofour proposed definition and sketch a bootstrap argument that tries toaccount for how causal interactions underlie but do not constitute the

1 For more on these two approaches to social science, as well as a third (the ‘critical’), seeBraybrooke (1987b), chap. 1). Braybrooke’s case for a more ecumenical view of these approachesto social science (as contrasted with the widespread view that they are mutually exclusive) is animportant inspiration for this paper’s efforts to understand the relations between the natural andthe normative.


intentional/normative component of rules. A striking epistemologicalargument due to Wilfrid Sellars is the main inspiration for both the ac-count I will give of the distinction between these elements and the ac-count I offer of how the intentional, normative dimension arises fromthe natural causal order.2 For reasons of space and time, I will not at-tempt a detailed reading of Sellars’s argument. My aim at this stageof the work is only to indicate how some aspects of Sellars’s views,especially the possibility ofretroactivelyassigning a normative statusto past events,3 contribute to a promising understanding of the roots ofnormativity. I believe the resulting account, though only an outline atthis point, has a virtue characteristic of Sellars’s approach to many is-sues: It is conceptually rich, but also metaphysically economical (evenaustere).

There are two entangled but distinct purposes for which we mightuse an account of social rules. They correspond to two very differentsorts ofexplanationsthat we can give of human behaviours. For thepurposes of naturalistic explanation, human behaviour is just part of thenatural goings-on in the world around us. Like other cases of relatedphenomena that constitute what Dudley Shapere (Shapere (1982)) hascalleddomains, behaviour is distinguished from other phenomena bya commitment we have made to treat it as calling for an explanationin terms of a common set of principles.4 Broadly, behaviour is distin-guished from other kinds of goings-on by the different regularities andprinciples we invoke to predict and explain it. But the regularities andprinciples at work in such behavioural studies are not differentin kindfrom the regularities and principles applying to (for instance) what wa-

2 For Sellars’s distinction between the natural and the logical order see, e.g., ‘Being and BeingKnown’ and ‘Some Reflections on Language Games.’ Sellars (1960, 1954)The bootstrap argumentpresented here is modelled on Sellars’s well-known bootstrap defence of an empiricist/inductivistaccount of perceptual knowledge in ‘Empiricism and the Philosophy of Mind.’

3 Something that is only possible because such assignments are interpretive rather than de-scriptive.

4See Sellars’s remarks on behaviourism in ‘Philosophy and the Scientific Image ofMan,’(Sellars (1962)) and inScience, Perception and Reality(Sellars (1963)) (hereafterScience),23ff., where behaviour is distinguished from other phenomena involving living things. Sellars’sdiscussion of worm behaviouristics, together with the distinction between nature and character inthe prediction/explanation of human behaviour (seeScience, 13ff.), reflects an interesting perspec-tive on the contextual constraints that circumscribe our use of certain explanatory principles.


ter does when cooled to 0o Celsius, or once heated (at standard pressure)to 100o Celsius as net heat is removed or supplied, respectively. Theyare commitments regarding how we shall project from what we observeof people and their doings to a fuller, but still purelydescriptive, storyof people in the world.

It’s worth noting that these behavioural principles are limited to con-texts in which something like standard conditions hold for the organ-isms involved; further, not just anything that humans or other organ-isms do is grist for the mill of behavioural science. We learn to drawdistinctions at mother’s knee, recognizing circumstances and kinds ofgoings-on that are part ofbehaviouras separate from other aspects ofthe nature of things. It’s worth noting that these distinctions are richand complex in ways we don’t often notice.5 But the point I want tomake here is just that social science in this vein is an entirelynatural-istic affair. It treats us as part and parcel of the natural world, subjectto the same kind of scientific description and explanation as any otheraspect of that world. This explicitly monist approach has a substantialprima facieedge over dualism, both of conceptual economy and coher-ence: Dualism (at least when taken at its metaphysical face value) positsmysterious, non-physical aspects of reality that are declared to be im-portantly connected both to the natural world and to a separate realm ofnorms. Deep and troubling puzzles have arisen in philosophers’ effortsto explain how a mere addition (however exotic) to descriptive real-ity can bridge the gap between descriptions and norms. But howeverthese questions are to be resolved, they are no concern of naturalisticsocial science. The aim of that endeavour is to capture and explain be-havioural regularities. The normative interpretation of behaviour playsno role there.

On the other hand, when our aim is interpretive, we want to embed

5 For our purposes, not much hangs on exactly where we draw the line between what is and isnot behaviour. What we would normally describe as deliberate actions certainly count behaviour;so are typical responses to sensory stimuli (shutting or covering the eyes when suddenly exposedto a very bright light, pulling one’s hand off after touching a hot surface, etc.). But falling tothe ground when thrown out a window does not (though a lot of behaviour would probably occuralong the way): this is something that happens because of features we share with all heavier-than-air physical objects, while behaviour turns on certain kinds of features that we share (broadly) withliving things.

14.2. RULES INLOGIC ON THE TRACK OF SOCIAL CHANGE 349

what people are and do in anormativeframework. Interpretive socialscience takes behaviour to be goal-directed and norm-guided. Such ef-forts, I will suggest, are analogous totranslation: they connect our ownunderstanding of norms in general and of familiar systems of norms inparticular with the doings we have observed, ‘regularizing’ them in therich sense of fitting them into a normative framework.6 The result, if weare successful, is an understanding oftheir behaviour cast in the kindsof terms we use to explain andjustify our own behaviour: meaningful,purposive, and at least sometimes rule-guided. Success in these effortsallows us to say, ‘Yes, we see why they do these things; they do thingsthat, were we situated as they are, with their beliefs and commitments,we ourselves would (or should) do.’

Sellars’s views onmeaningare in the background here. For Sellars, ‘“. . . ” means “—” ’ expresses a translation-relation, in which (normally)‘—’ is expressed in a language the speaker and audience share (andat the very least, in a language the audience understands) and theroleof ‘. . . ,’ in the language being translated, is identified with the role of‘—’ in the familiar language (1956, 163) These roles, in turn, are spec-ified in terms of the norms of use for ‘. . . ’ and ‘—’ in their respectivelanguages, including norms of inference in the language (and the differ-ence that ‘. . . ’ makes to the inferences that are endorsed by them) andthe links between language use and circumstance that are involved inobservation (what Sellars calledlanguage entry) and deliberate action(language exit). So what I’m suggesting is that a similar, translationaleffort to map behaviour into normative systems (whether normative sys-tems we are actually committed to or systems described with the helpof a meta-language of norms) underlies interpretive social science.

14.2 Rules inLogic on the Track of Social Change

In Track we specified thecontentof a rule in terms of three compo-nents: volk is the set of agents to whom the rule applies,wennis the

6 Here Davidsonian remarks regarding charity and the presupposition of rationality come in—although in our presentation, they arise explicitly as part of the methodological task of connectinghypotheses about rules in force with observations of behaviour, rather than as presuppositions ofinterpretation.


set of circumstances in which the rule applies, andnono is the set ofactions forbidden (to those agents in those circumstances) by the rule.But our definitionof rules begins with a technical notion labelled (forobvious reasons)imperatives. An imperative, in our sense, is ablock-ing operation, verbal or otherwise. A blocking operation, in turn, is anaction kind whose instances characteristically (in some important rangeof circumstances) prevent some agent(s) from performing an action(s)of a certain kind, and which is performedwith that intent(34). Blockingoperations need not always succeed; a blocking operation ‘qualifies forthe title if it works sometimes for some people . . . when conditions aresuited to itssucceeding’ (34; emphasis added). A rule, in the abstract, isa set of such imperatives—that is, a set of blocking actions, directed atthevolk, under thewenn, and against thenono. Finally, a rule in forceis a rule sufficiently many7 of whose imperatives actually occur.

This definition appeals to natural facts, but it combines them withnormative/intentional facts. The range of possible blocking operationsis determined by the natural causal influences that some available actiontypes have, in various circumstances, on the actions of certain agents.But they are not blocking operations just in virtue of these natural causalpowers. To be a blocking operation, what someone does must be donewith the intentionor purposeof preventing someone from doing some-thing.

This picture of rules leads us to the final image of a triptych, threepanels representing different stages in the application of a rule for agiven subject: In the first panel, the subject is exposed to a range ofimperatives aimed at ensuring that her behaviour accords with the rule.In the second panel, subjects in whom the rule has been successfully in-culcated govern themselves: They now possess an internal system that(often enough, under what we might callnormal psychological oper-ating conditions)8 prevents them from violating the rule. It may do soby producing internal imperatives blocking such actions when thewenncondition of the rule is met,9 or by replacing or altering the original

7 Sufficiently many, that is, for the purposes of instruction and enforcement.8 It is clear that there is a normative side to this notion—but we will not try to untangle it from

its descriptive aspects yet.9This is reminiscent of Hume’s remarks on abstract ideas, according to which the idea of

14.2. RULES INLOGIC ON THE TRACK OF SOCIAL CHANGE 351

system for generating actions in some way that has the effect of ensur-ing general compliance with the rule. These mechanisms either involveconscious obedience to the rule or lie in a middle ground10 betweenconsciouslyobeyinga rule and merely conforming to it.11 In the thirdand final panel, rules that early training does not suffice to inculcate (toa sufficient degree, in a sufficient number of subjects) areenforcedby(a wide range of) penalties directed against violations.

There is obviously a lot more to be said here. To forestall one misun-derstanding (arising from the assumption that our account ofimperativeaims to be an account of the familiar grammatical notion), I will pointout that our use of the word ‘imperative’ is technical: Its applicationto these action kinds is inspired by the ordinary understanding of theimperative voice as directing or giving orders—but some uses of thatvoice arenot imperatives in our sense (consider ironic and other non-standard uses of the imperative voice) and many other things (gestures,expletives, merely declarative sentences uttered in certain ways and incertain circumstances, and a broad range of actions like closing a dooror turning away from someone)are imperatives in our sense.

A second concern is the problem of deciding which, and how many,imperatives must occur for a rule to be in force. It depends on the kindof imperatives that occur, on the range of relevant rule alternatives thatthe imperatives must bring us to distinguish between, and the kinds ofcues that enable those subject to the rules to learn them. Here we rely ona rich background of natural facts about developmental and behaviouralpsychology. Given that background, it may turn out that certain typesof imperatives are particularly effective in driving the transition frompanel 1 in our triptych view of rules in force, where imperatives arise inthe course of training, to panel 2, in which the rules have been success-fully inculcated in someone subject to them. Further, some rules are so

the counter-example This to a mistaken generalization is called to mind automatically when wecontemplate the generalization; the result is that acceptance of the generalization isblockedby ourawareness of the counter-example.

10 Sellars identifies this middle ground in ‘Some Reflections,’ inScience, 321-58.11 In this middle ground, though it is false that we are consciously following the rules, the

subject’s behaviour is still correctly interpreted in terms of the rule: The subject does as she doesbecause the rule is in place, even though she does not explicitly represent the rule to herself in ameta-language and constrain her behaviour by means of that representation.


naturally and thoroughly inculcated (and any violations that persist oflittle enough concern) that panel 3 plays almost no role in keeping vi-olations down. On the other hand, others obviously require systematicand expensive enforcement efforts.

14.3 Some Worries

Epistemically this is tricky business. After all, asettledor establishedmatter-of-fact violation of a principle of descriptive science is arefu-tation of that principle: Although a refuted principle might still applyusefully in a special subset of cases, it can no longer be defended as afundamental principle. But an established, matter-of-fact violation of anorm does not show that the norm itself isn’t in force or didn’t apply tothe action violating it. Further, the fact that establishing that there hasbeen a violation of a (proposed) natural principlefalsifiesthat principleis widely, and rightly, viewed as a critically important source of the epis-temic standing our commitments to such principles do have: A principlethat cannot be tested in this way cannot be confirmed or vindicated bypassing such tests either.

Even with such tests in hand, the epistemology of natural laws ischallenging enough. They reach far beyond the goings-on that we haveactually observed; their acceptance commits us to claims about eventswe did not and may never observe. Further, such commitments are usu-ally thought to commit us to counterfactuals, constraining things or cir-cumstances that never actually occur. But, at least for an empiricist,the evidence we draw on to justify such commitments is limited to theobservations we have actually made. Hume’s concerns about induc-tion loom large here. Our evidence often seems far too thin to justifya serious cognitive commitment to general principles of nature. Theseconsiderations raise even deeper doubts about the epistemology of rulesand norms, which seem to be more tenuously linked to actual observa-tion.

Suppose for now that we know what we are claiming when we as-sert that a certain rule is in force, and invoke that rule’s being in force toexplain some actions performed by certain agents. (Recall that our ac-

14.3. SOME WORRIES 353

count inTrackemphasizes the regular appearance of imperatives aimedat inculcating and enforcing compliance.) Two questions come imme-diately to mind:

1. How can we justify the choice of which rules to invoke, whensocial scientists pursue interpretive social science? (The problemis that there is always a wide range of alternative rules that would‘fit’ the observed behaviour equally well, i.e. which the behaviourconformsto.)

2. . How does a rule’s being in forceexplainfeatures of the actionsthe rule applies to?

In Track, we are fairly lax about these concerns. Our position thereis that ‘it suffices for holding that people know what the rules are tofind them distinguishing behaviour that accords with the rules from be-haviour that does not. It suffices for holding that they are followingthe rules to find them either acting so as to avoid blocking, physicalor verbal, or modifying their course of action suitably when a form ofblocking is produced (or begun) by way of caution’ (39). And for thepurposes in view there, this is reasonably satisfying. We clearly havea successful practice in which we both identify certain rules that somepeople are following, and interpret behaviour, our own and others, byappeal to various goals and rules. Appealing to that practice is fair ball.But it does not illuminate what justifies our applying such concepts atall. Neither does it help to explain how we acquire them. As an ac-count of our grounds for distinguishing merely natural description andexplanation from the normative, and choosing, as we do, to apply thenormative to a wide range of human (and perhaps some non-human)behaviour, it doesn’t go far enough. So here we will spend a little moretime dwelling on the difficulties.

A convincing general answer to question 1 will be difficult to comeby. As Goodman showed long ago (Goodman (1965)), different waysof reportingour evidence to date regarding the colour of emeralds seemto be equally correct as descriptions of the evidence, while suggestingvery different inductive conclusions regarding the colour of emeralds


observed in the future. But this problem is clearly a problem for bothnatural laws and norms, and so not specifically a problem for norms.12

The first step towards an answer to question 2 is to say that it is notby the same means as natural principles explain regularities in our ob-servations. In the case of natural laws, we have a direct inference fromtheir holding to the regularities that the laws logically impose on thephenomena (this is just the converse of the falsificationist’s inference).In standard cases a rule’s being in force ‘explains’ why some actionscomply with it, but not by appeal to a direct inference from the rule’sbeing in force to the action’s complying with it. Natural things obey thelaws of gravity—whatever these actually are—with no need of thoughtor calculation. But (in the central cases that extended or even metaphor-ical cases are modelled on) agentsobeyrules in virtue ofconsciouslyrecognizingthat those rules apply to their action(s), and deliberatelyconforming their actions to the requirements of the rules. Full-fledgedobedience to a rule combines awareness of the rule’s demands, a con-scious commitment to respect them, and the ability to fulfill that com-mitment. Accidentally conforming to a rule’s demands lies at the otherextreme—here the rule plays no explanatory role at all. In between, wefind pattern-governed behaviour, which accords with the rule not merelyby accident, but for reasons that link thevolk, wenn, andnonocompo-nents together in a reliable way, ensuring that the subject will generallyrespond to circumstances in which he falls withinvolk andwennof therule by avoiding actions in thenono.13

There are slips between the inferential cup and lip for rule expla-

12 I do not mean this as a counsel of despair. Goodman’s puzzle passes over the role thatmaterial inference rules (‘Some Reflections,’ inScience, 352ff.) play in making the predicates ofour language the predicates that they are. This approach makes induction a fundamental part ofwhat we are committed to in speaking a language. Of course this does not show that the world mustbehave as induction leads us to expect. But it does undermine the simple equivalence Goodman’sargument assumes between ‘grue’ and ‘observed before t and green or observed after t and blue.’The inductive practice Goodman envisages as associated with ‘grue’ is clearly distinct from thatassociated with the language of this proposed definition.

13 See ‘Some Reflections’ for more details. Crucial to Sellars’s account of rule-governedbehaviour is the identification of a middle ground between this strong sense of obedience and‘mere’ conformity to a rule’s demands. According to Sellars, behaviour can be guided by a rulewithout the subject’s having to represent the rule in some meta-language. This middle groundallows Sellars to escape a regress that threatens the thesis that learning a language is learning toobey the rules of that language.

14.4. THE BOOTSTRAP ARGUMENT 355

nations, unlike explanations in terms of natural principles. Failures toobey a rule in force can be due to ignorance (whether culpable or not),to calculational errors and limits (especially in complex cases),14 andeven to willful disobedience: a rule’s being in force may even explain,by means of awareness of the rule’s being in force combined with adeliberate response to that fact, actions that don’t comply with it (weoften call such actions defiant or rebellious). These aspects are part ofwhat makes rule explanations normative—and their impact weakens theepistemic ties between observed regularities and tests of rule hypothe-ses. But there is a positive side to these differences as well, picked upin both panels 1 and 3 of our triptych: justifying rule hypotheses alsorequires evidence of background conditions eliciting and enforcing obe-dience to the rules. That is, rules are accompanied by efforts to instructand enforce rules—which appear at the descriptive level as blockingactions (in the causal sense), and characteristic patterns of response toviolations of the rules.

14.4 The Bootstrap Argument

Thus far this discussion has focused on the relation between hypothe-ses regarding rules in force and descriptive facts about the behaviour ofthose subject to them. But the specifically normative aspects of suchhypotheses about rules relate not just to direct behavioural evidenceand how it fits the hypothesized rules, but to interpretivecommitmentsregarding how we will construe (some aspects of some people’s) be-haviour. These commitments can be vindicated by successful applica-tion, which involves not merely examples of behaviour conforming tothe rule, but also giving an account of how conformity to the rule iselicited and enforced. The evidence in hand goes beyond whether ornot certain behaviour violates the rules to include an examination of thebehaviour in its social context, including efforts at teaching, inculcation,and enforcement of the rules.

14 Closely related to this is the injunction to legislators that laws must be such that it is clear toanyone aware of a law’s provisions whether what they are doing violates it.


We will not begin with rules explicitly formulated in words. Fo-cusing on this admittedly central case seems to produce an irresolvableregress: we cannot communicate the contents of such rules to each otherwithout presupposing a shared system of rules. This leads to an obvi-ous question: How could we communicate the contents of these rules,so that communication can get underway? Assuming the existence of abuilt-in system that we all shareab initio might provide an escape fromthis worry, but empiricists will want to avoid simply cutting the Gordianknot.15

Instead, we will begin with development. The key question willbe: How doindividualsmake the shift from merely behaving to actingdeliberately, in recognition of rules that apply to them and to others?Our answer must account for how we acquire concepts including theconcepts of a rule, of authority, and of commitment, and how we cometo bejustifiedin applying these concepts to ourselves.

Along the way I will draw explicitly on Sellars’s work to illuminatethese difficulties and point towards a strategy for dealing with them.Sellars aims to take empiricist scruples about evidence seriously, whilefinding room for rationalist ideas about the fundamental importance ofrules and concepts to our understanding of the world and of each other.This aim is apparent in Sellars’s careful examination of the descriptive,naturalistic underpinnings of his account of how someone can come tobe an observer. In his epistemic bootstrap argument, Sellars exploreshow an individual knower can appeal to a kind of induction to justifyher claim to be an authority on certain perceptible aspects of her sur-roundings. In doing so, she enters into the normative game by meetingits demands. We begin with an account of Sellars’s view ofsensationsand their relation to the philosophical notion of ‘sense data’:

It certainly begins to look as though the classical conceptof a sense datum were a mongrel resulting from a cross-breeding of two ideas: (1) The idea that there are certain in-ner episodes—e.g. sensations of red or C] which can occur

15 Note that this does not rule out sharing certain matter-of-fact dispositions that make it easierto train individuals to obey rules of a particular kind. See ‘Some Reflections’ for a very carefulapproach to escaping this regress.


to human beings (and brutes) without any prior process oflearning or concept formation; and without which it wouldbe impossible to see, for example, that the facing surface ofa physical object is red and triangular, orhear that a certainphysical sound is C]. (2) The idea that there are certain innerepisodes which are the non-inferential knowings that certainitems are, for example, red or C]; and that these episodes arethe necessary conditions of empirical knowledge as provid-ing the evidence for all other empirical propositions. (Sci-ence, 132)

The division Sellars makes here is between mere sensations as episodesin the natural order and the non-inferential knowings he holds somehave confused them with. For Sellars, knowings are episodes in thenormative order, playing a certain role in the give and take of justifica-tion. These episodes presuppose, but are not reducible to, certain factsabout the natural order. In particular, facts we are able to report mustcause (in certain conditions) states in observers that are behaviourallydiscriminable, that is, states that (once the observers are trained) cancause the observers to make certain noises or write certain inscriptions,noises and inscriptions that they will not make or write when not inthose states.16 These behaviourally discriminable states are what Sell-ars callssensations: they are posited natural states of observers that arenormally triggered by sensory exposure to a particular kind of stimu-lus. But they can also be triggered by other circumstances as well (thusexplaining, in outline, how perceptual illusions occur).

On Sellars’s account, the ability to have sensations is not acquired.17

16 This condition clearly failed for Rene Blondlot’s N-rays when a visiting scientist put the‘N-ray’ source in his pocket and the unknowing observers went merrily along making more ‘ob-servations’ of N-ray phenomena. This clever little jape put an end to the study of N-rays. Notethat meeting this condition alone is not enough to constitute a knowing, though it does make the‘observer’ a reliable source of information; an observer trained in this merely stimulus-responseway, no matter how reliable, does not properlyknowor report anything. On Sellars’s account, toknow or report, the observer must not merely be a reliable reporter—she must know that she isreliable. This reflexive, normative condition leads to a regress that Sellars avoids by appeal to thebootstrap argument we are about to consider.

17 Though it arises in the course of normal neuronal/sensory development, and so might rea-sonably be said to be acquired in a broader sense.


It constitutes the natural basis that allows us to learn to make percep-tual reports. But the ability toperceive—that is, to know certain factsabout our surroundings by means of a non-inferential response to oursensations—isacquired. Knowing requires not just reliable noise-making,but an awareness of one’s own authority as knower. This in turn requiresdeveloping not just the ability to respond reliably to different sensations,but also the concepts and skills required to producejustificationsof theclaims we make.

The transition from mere ‘thermometer’ reliability to having epis-temic status as an observer in one’s own right requires that one learnto defend a claim to theauthority to make such assertions in such cir-cumstances.18 Suppose that our ability to come to know facts about ourenvironment by means of our senses is something that we learn. Empiri-cism requires that the recognition of this ability in oneself be groundedin a kind of induction: what grounds our claim to authority in a presentreport is our demonstrated past reliability as reporters of certain facts(the redness or triangularity of an object, or that a certain note is C]).Now consider thefirst occasion on which one can correctly claim suchauthority. It seems that the past episodes we need to draw on for this in-duction couldn’t be cases of knowledge themselves:Ex hypothesitheyoccurred before we had the conceptual sophistication and/or a sufficientbody of pastexperience of our reliability as knowers of such facts insuch circumstances to give an inductive justification of our claim to au-thority.19 But without past cases of perceptual knowledge to appeal to,we seem to be stuck: until we have the inductive base in hand, we can-not use induction to justify our claim to authority, and we cannot buildan inductive base of past justified knowledge claims to generalize on

18 My concern here is with the second hurdle of the two Sellars considers. Passing the first hur-dle requires only that others be in a position to treat the observer as an authority; being a ‘reliablethermometer’—that is, being disposed to make correct reports under a range of circumstances—isenough. But the second hurdle requires that the observer be in a position to recognize her ownauthority.

19 A theme I am short-changing here is Sellars’s holism, according to which having a singleconcept (redness, for instance) requires having a more-or-less complete set of concepts includingconcepts of the conditions of observation, and related concepts (of coloured items and of othercolours, for instance). The bootstrap argument may well also be aimed against the broader regresspuzzle that this view of concept acquisition suggests.


unless we have a justified claim to authority already in hand.Sellars’s solution to this regress is to claim that the status of such an

episode as a knowing is not an occurrent, naturalfactabout the episode.Instead, it results from locating it, conceptually, in anormative, eval-uative context. And we can do this after the fact. That is, we canretroactively recognize a past report as a case of justified perceptualknowledge even though at the time that the report was made, we lackedeither (or both) a sufficient grasp of the normative business of justifyingour claims or a sufficient inductive base of similar cases in which our re-ports proved correct. Thus Sellars’s account of our status as perceptualknowers holds that at a certain point in our development we come tohave both the conceptual sophistication to justify our knowledge claims(in certain kinds of circumstance) by appeal to inductive evidence oftheir reliability, and sufficient experience of our past success, derivedfrom a retrospective application of these inductive standards, to providethe inductive evidence required. Then and only then do we meet thenormative standards applying to knowers—but the cases in which wethen recognize ourselves as meeting the standards extend back into thepast, as well as forward into the future.20

What I want to do here is to apply this kind of argument morebroadly, to sketch the origin of normative concepts including standardsof evidence, commitments, and contents. There will be an evolutionaryelement in the series of steps I will describe, as well as a prospective,commitment-based view of the openness of norms to interpretation. Fi-nally, there will be a meta-linguistic element in the account of our talk ofnorms (rather than a metaphysical one). The key element that I will bor-row from Sellars’s bootstrap argument is the possibility of retrospectiveinterpretation: As in Sellars’s epistemic bootstrap, we need to pass twohurdles here—the first demands a justification for interpreting some-one’s behaviour as a case of rule-following. Here I will assume thatevidence of a sufficiently reliable tendency to avoid actions breakingthe rule, grounded in the effective training and enforcement that occupy

20 If this were claimed to be a descriptive change in the facts about these earlier episodes (as areductive account would demand), it would require a bizarre and implausible kind of retro-causalinfluence by means of which a subsequent state or activity of the knower could causally alter thefacts about his own epistemic past.


panels 1 and 3 of our triptych, is enough. But the second requires thatthe rule-follower be capable ofjustifiably interpreting her behaviour asrule-guided. This means that she must have acquired both the concep-tual sophistication required and good inductive grounds for regardingherself as meeting the conditions those concepts impose. Here a regressarises which I take to be parallel to the regress Sellars addresses in hisbootstrap argument, and resolvable by invoking the same contrast be-tween description and interpretation. Unlike description, interpretationcan beretrospective, ascribing after the fact a normative status to anepisode that the episode did not have at the time that it occurred.

14.5 The Roots of Normativity

The proposal I offer here is tentative and sketchy, but I believe that itis suggestive. I believe that its components make crucial contributionsto a correct understanding of norms. As I present it here, the proposalhas five components, forming a series of steps beginning on the natural-istic side, but arriving finally at an account of the normative. The keynormative concept that I will focus on is the notion of acommitment. Ibelieve that the other normative concepts we have been considering canbe subsumed under this notion and, more narrowly, under the notionof a negativecommitment, a commitment not to do something meet-ing a certain description. In particular, I assume here that paradigmaticrule-governed behaviour is just behaviour that is constrained by a full-fledged, consciouscommitmentnot to do anything forbidden by the rule.But our story begins with a much weaker, merely behaviour-predicting,form of ‘commitment.’

14.5.1 Evolutionary Advantage

Being able to signal willingness to make and engagement in a com-mitment has substantial evolutionary advantages. For example, we candeal with coordination problems including some forms of prisoners’dilemmas by signalling willingness to cooperate with anyone willingto reciprocate—the signal could of course be such that it only takeseffect—that is, is only to be relied on—if a signal indicating recipro-

14.5. THE ROOTS OF NORMATIVITY 361

cation is received. At its simplest, this kind of signal could simply bea ‘flag’ flown at all times, indicating an inclination to cooperate withanyone else flying a suitable ‘response’ flag. Suppose we have a subjectwho begins the process here by making a friendly nod towards someonewhose cooperation she needs to enlist in food gathering. By workingtogether, the two can gather more than twice what each could gatherindividually. But each will have an opportunity to make off with all thefood as the end of the day approaches. The nod is returned, and the twogo off to gather food. If these nods are sufficiently reliable indicators ofcooperative behaviour, the benefits of relying on them to establish trustwill outweigh the risk of losing all the food to a sneak. Evidently, when‘commitment’-signals, linguistic and otherwise, are reliable symptomsof reasonably durable behavioural dispositions, they can be of consid-erable value to social organisms. Although the effectiveness of suchsignals creates opportunities for cheaters to prosper by manipulating orfaking them, enforcement efforts can make cheating risky enough thatthe value (and so the use) of the signals can be sustained.

14.5.2 Naturalism

The signals involved here are to be thought of as natural-linguistic sig-nals of ‘commitment’; they need only predict behaviour with sufficientreliability. The reliability of this prediction can in turn be groundedin the sort of behaviour the naturalistic side of our account of rulesinvolves—the inculcation of ‘rules’ (in the sense of behavioural regular-ities) by interaction between trainees and teachers using natural block-ing operations, together with enforcement if needed.

14.5.3 Meta-language.

Having developed a simple language in which we can give and acknowl-edge a variety of such ‘commitments,’ we can go on to develop languagein which we talk about this practice. Words like ‘commitment,’ ‘rule,’‘enforcement,’ and so forth, come into use here, as part of a complexlinguistic response to the various behaviours involving such signals,others’ responses to them, and the broader responses (of teachers and


enforcers, in particular) to behaviours that accord with or violate the sig-nals’ regular accompaniments. In our example, we may hear our subjectand her partner saying that they have ‘an agreement’ to work togetherand share the proceeds; others may be disposed to comment in variousnegative ways should our subject abscond with the day’s proceeds.

14.5.4 The bootstrap

Here we finally begin to approach the realm of norms. The retrospec-tive element of the bootstrap considers past behaviour and interprets it interms of newly acquired normative concepts. Normative concepts them-selves originate in meta-linguistic reflection on the language in whichwe undertake and acknowledge commitments. Like other aspects oflanguage use, there are complex natural regularities connecting the useof these words with other aspects of our behaviour.21 Normative wordsare regularly connected to a number of other aspects of language use,including the language of psychological explanation, allowing forget-fulness or inattention to serve as explanations of failures to act on com-mitments made. These also include the language of excuses, connectedin turn to various conditions which (when brought to the attention ofconcerned parties) have the effect of reducing various social signals,linguistic and otherwise, that serve to enforce the natural practice of‘commitment’ indication and acknowledgment. At this point, reflectionon patterns of behaviour in our object-level use of descriptive languagemay also lead us to apply this meta-language of rules, commitments,and so on, to it as well—that is, we may come to conceive languageitself in terms of this proto-normative vocabulary.

It is very difficult, at this point, not to just slide over the edge intosurreptitiously and then explicitly normative accounts of these phenom-ena. I ask the reader to cling tightly to a purely descriptive, naturalisticview of what I’ve said to this point. What is being described above are,broadly, systematic links between behaviour and various signals thatthe parties involved display and respond to. By carefully insisting on

21 See Sellars, Sellars (1960), especially 50ff. on the relation between such descriptive regular-ities (linking language use to other features of the world) and the normative ‘order of signification.’


the distinction between the normative and the descriptive until the lastturn, we can hope to shed some real light on the roots of the normative.

What I suggest is that our cooperative hunter-gatherer crosses overthe edge to full-fledged use of normative concepts when she learns toapply this meta-language of commitment to herself,justifiablyconceiv-ing herself as making commitments and keeping commitments she hasmade.22 As in Sellars’s epistemic bootstrap argument, she must havein hand inductive evidence justifying this application of commitment-talk to herself, that is, evidence of her ability to make and act on com-mitments in a reliable way, as well as evidence that she can reasonabout both her own commitments and those of others reliably.23 Justas in the epistemic case, this inductive requirement seems to producea regress: The past cases she appeals to in her first application of thismeta-language to herself mustby definitionpredate the time at whichshe first applies these concepts to herself. But she can apply themretroactively, recognizing past ‘commitments,’ and interpreting them,using her new vocabulary, ascommitments. Recognizing subsequentpast behaviour as having regularly fulfilled her (now recognized) com-mitments’ demands, and carrying this interpretation through to comple-tion, she applies her use of this vocabulary for rules, commitments, andso forth, reflexively to itself as well as to object-level linguistic practice.

From a purely naturalistic, descriptive point of view, all that she hasdone is to extend her natural-linguistic practice. But as a user of this ex-tended language, she now conceives herself as making commitments, asdistinguishing actions in accord with those commitments from actionsviolating them, as teaching and enforcing rules, and as pursuing vari-ous goals. By conceiving herself in these ways, employing these terms,and reasoning about them in the interpretation and evaluation of her be-

22 Without evidence of commitments kept, seeing oneself as making commitments is (at best)unjustified.

23 In fact this needs some care, since mere reliability is not enough. I will reliably meet theconditions of a promise to fall to the ground if thrown out a window. But this is evidence that I amsubject to a certain natural principle, not that I can make and meet commitments. Broadly, whatshows that I can make and meet commitments is evidence that, if I have made a commitment, Iwill reliably (where this is qualified by considerations about my being aware that the commitmentis in play and that the circumstances are such that it now demands something of me) act in certainways that I couldnot be predicted to act if I had not undertaken the commitment.


haviour (and that of others), she demonstrates a grasp of normative con-cepts grounded in, but not reducible to, natural facts about agents andcommunities. The contents of these concepts extend beyond the merelydescriptive natural facts about our subject and her fellows by virtue ofthe rich embedding of the associated inferences with the practical goalsand activities of her life, goals that extend beyond the very narrow goalsgoverning the business of description. And in the light of her and theirsuccess vis--vis these goals, together with the bootstrap considerationsbrought to bear above, shejustifiably interprets herself (and others) inthese terms.

This broader self-interpretation in terms of the language of commit-ments reflects something that I think is implicit in Sellars’s original,epistemic bootstrap argument: a view of oneself, as an observer, mak-ing commitmentsto the reports one makes, commitments that includean undertaking of the justificatory burden Sellars’s bootstrap inductionaims to meet.

14.5.5 Interpretation of self and others

The account of norms proposed here applies to us becausewe view our-selves(as our use of this kind of language demonstrates) as subject to thenorms (and to the explicit linguistic expressions of norms) that emergefrom our reflective engagement with each other and with actual lan-guage use. It is tempting to say, echoing some versions of relativism,that we make ourselves subject to norms solely by agreeing to speak ofourselves in normative terms, and that what those norms demand of usis simply a matter of what our community decides they demand. But tosay this is to ignore the naturalistic, descriptive facts out of which ourmeta-linguistic talk of norms arises. We cannot state, in advance, whatlimits this background imposes on the range of successful practices—there is no conceptual reduction of the normative to the descriptive tobe had. But the stability of our commonsense world-view, and the con-tinuity of its standards of evidence (including, as I argued in Brown(2003, 2004), a central role for explanatory coherence), give us robustassurance that we do not, and cannot, simply make things up ‘at will.’


Still, the contents of our commitments to such norms are open-ended. I cannot demonstrate here and now a full grasp of what thevarious norms I am committed to demand of me; part and parcel of thenotion of a norm is that its demands vary with circumstance and arejudged not just by me, but by others to whom I make and (at least inprinciple) can express my commitments. When I make commitments tosuch norms, what I am implicitly committed to includes all the conse-quences of those commitments. But I cannot, even in principle, workthese consequences through entirely on my own. Others’ responses tothe commitments I undertake are part of what determines the demandsa norm imposes on me. Ultimately, these implicit commitments areconceived normatively, in terms of what an ideal agent or communitywould take them to be: There is no non-normative ground in which wecan anchor the contents of norms.

One striking example of this arises in science, where our interpreta-tion of what is reported in previous observations (i.e., the inferences weare prepared to draw from such reports) shifts as our theories change.Reliable observations of masses, relative velocities, and so on, providedby classical physicists now sustain (given our present conceptual re-sources) inferences those physicists would have rejected. Our commit-ment to these norms and to our fellows demands openness to this kind ofdevelopment of the content of our present commitments over time. Theimplications of this openness are of great philosophical interest, but wewill not pursue them further here.

Appendices

367

Appendix AAnother—Literary—Side of DavidBraybrooke:The ComicDialectician

24

People will not be especially surprised to learn that professors inchemistry and physics spend the summer working on experiments intheir laboratories. In geology, oceanography, and biology, many of themare away from the campus doing research in the field. So are many pro-fessors in the social sciences and the humanities. Professors of French,for example, try to spend recess time in France, which would be a usefulthing for them and their students if they just practiced speaking Frenchthere. In fact, they have reading, writing, and research in French li-braries to do as well. Summers are the time to do reading of scholarlymaterial that isn’t tied to their classes, the time, too, to get ahead with theprojects—the books and articles that they cannot concentrate upon writ-ing under the distracting pressures of term-time. Many professors taketheir stipulated month’s vacation during the four-month recess, spend-ing the other three months on research—in many cases, on teachingas well, because even during the recess, teaching goes on with grad-uate students and in two consecutive summer sessions for undergrad-uates. When their research consists of reading and writing, however,their working time tends to spread out over the whole recess. Whatevervacation is taken is taken little by little, in an afternoon here, a day offthere.

Among the professors for whom this is true we find a certain profes-

369

370 APPENDIX A

sor of philosophy. Let’s call him Green. Green is perhaps an exceptionin keeping up his writing all through the year, in term-time and out; andan exception, too, in doing his writing entirely in his office, to whichhe repairs every morning, seven days a week, holidays included. Manyother professors, however, put in the same amount of work year by year,though they work on different patterns. They may, for instance, do theirreading during term-time; when the term has ended, they put in days at atime at their typewriters. Green does his intensive reading of scholarlymaterial, except for the reading (rereading) tied to his classes and theminimum necessary to keep on with his writing, in the summer.

A typical summer day for Green begins at 7:30, after exercises, ashower and breakfast. Until about 9 o’clock, with some short breaksfor household chores, he reads some philosophical work—chapters ina book; an article in a journal—related in at least a general way to thetopic that he is writing about. If the topic is the relation between mindand body, he will be reading an intricate discussion in theJournal ofPhilosophyof the extent to which computers can perform mental func-tions that in human beings can be performed only consciously. At 9o’clock or thereabouts, Green goes to his office, sits down at his type-writer and writes or rewrites 3 or 4 pages of a projected book or article.Now and then, a student will drop in to ask for a bit of advice; in thesummer, he or she is likely to be a graduate student, working on a the-sis. Or maybe a colleague, maybe a secretary, will come in to do somebusiness, about a calendar entry, or a regulation to be reformulated. Byand large, however, Green sticks with his project until he has finishedhis daily stint, and the ideas no longer flow freely—until sometime be-tween 11:30 and 12.

He will then spend half an hour or an hour on his correspondence,which might on a given day include writing recommendations for col-leagues elsewhere or for former students, ordering some books, ex-changing comments on current work with a colleague at another uni-versity. At 12:30 or thereabouts, Green walks home to have lunch. Hedoes a little light reading afterwards, then goes back to the departmentof philosophy to look through the day’s mail. Several items may requireimmediate attention: A prospective participant in a week-long confer-

371

ence on philosophy and psychology that he is organizing for later in thesummer writes to ask whether room in the program can be found forhim to read two papers; a publisher writes to ask whether he will assessa manuscript; Green drafts replies to both and leaves them with a secre-tary for typing. Then, just before 2 o’clock, he goes off again to spendan hour in the library, first looking up some books in the card catalogue,then tracking them down in the stacks, picking out some to check outat the circulation desk. Others he puts back after a glance, as not rel-evant to his work; still others, missing from the shelves, he will haveto ask the library to recall from other borrowers. Leaving the libraryat 3 o’clock, he might, some days, have a committee meeting to go to,perhaps one dealing with revisions to the rules of admitting mature stu-dents, perhaps one occupied with measures to encourage more facultymembers to apply for research grants from Federal agencies.

A committee meeting might in fact have taken up all the time that haselapsed since lunch. Today, however, Green has no committee meetings.So he goes back, to the department to see whether the secretary has fin-ished his letters for signature and to proofread the half-dozen pages thatshe has typed since yesterday of an article that he finished last week. Atquarter to 4, he walks home again and settles down for an hour with aphilosophical work ona priori knowledge. He goes for a swim at Sandthen is occupied until 7 or 7:30 with preparing dinner, eating it, andwashing up afterwards. Some evenings he will not be doing any workat all, not even any serious reading. But typically he will be readingone or another book from the pile beside his living-room chair; tonightit is Fernand Braudel’sThe Wheels of Commerce, a history of everydaypractices during the development of capitalism, that occupies him. (Hewill have a chance to draw upon some of the information in this booknext week, in a debate with other philosophers about the power that pri-vate property gives some people over others.) An interruption occurs at8:20: another member of the team with which he is collaborating on apicture book about 19th Century architecture in Halifax calls to discussthe current stage of this project. At 9:30, he turns to lighter reading,though even this may have a bearing on his work. He is currently read-ing Edmund Wilson’s diary for the 1940’s, and he takes that up again,

372 APPENDIX A

putting it down after a while to look through a book of 20th Centurywatercolors. At 10:30 Green is ready for bed; and to bed he goes.

Excerpt from Faculty Council Minutes, 29 September 1983

Professor Y. said that he thought Appendix B, A (Summer’s) Day inthe Life of a Professor of Philosophy, was undesirable. Its formulationwas objectionable if indeed this was intended as a strategic document.People had read this and thought it ridiculous as a portrayal of a seriousperson’s existence. It was a mistake, which worked against the generalimpact of the brief. Professor B. said that members of Council wouldrecall that this passage had originally been included in the brief. It wasnow presented as an appendix so that it could easily be dropped. TheDean called for a straw vote. Professor W. said that he thought it wasimportant to indicate that work did go on in the summer. Professor D.said she was in favour of removal. Aside from the frivolity, the appendixsuggested that Professors had no family or domestic concerns. ProfessorF. said that he remembered saying at an earlier meeting that this sectionwas the part that he had enjoyed most. If it were removed, he hopedthat a suitable footnote might yet take its place. The Dean called for amotion to deal with the matter. It was moved by Professor L., secondedby Professor T.,

that Faculty Council recommend to Faculty that AppendicesB and C be omitted from the Faculty Brief to the Royal Com-mission on Post Secondary Education.

Professor W. asked what was the objection to having somethingabout what Professors do in the summer. He did not see the harm inaddressing these questions seriously. The Dean said that was the com-mittee’s view. Professor B. said there were no paragraphs in the briefitself which bore on this question.

The motion was carried with one dissenting vote and three absten-tions.

EXCERPT FROM FACULTY MINUTES, 11 OCTOBER 1983 373

Excerpt from Faculty Minutes, 11 October 1983

The Secretary said that the report had been discussed on two occa-sions by Faculty Council and as a result there was a recommendationfrom Faculty Council to Faculty which he would move. The Secretarymoved, seconded by Professor D.,

that Appendices B and C be deleted from the document.

The Secretary said that Faculty Council had felt that these appendicesdid not achieve their objective. It was recognized that what the com-mittee had tried to do was to represent the fact that Professors workduring the summer and while on sabbatical leave; however, the form ofthese appendices it was felt by Council would be counterproductive inachieving this objective.

Professor B. said that he felt it was necessary to defend Professorsagainst assumptions of idleness. It was a natural reaction to want toexplain what we do during the summer. He said that one member ofCouncil had said that this was the only part of the brief he had enjoyed.Professor B. said that the appendices were anecdotal and personal andmight achieve some sympathy for professors. Professor 0. said that ifit came to revisions, was it a question of dropping the appendices orkeeping them. He thought they could be moderated somewhat. If thiswere not possible, he would support the motion. Upon being put to thevote, the motion was carried by fourteen votes to eleven.

An Open Letter from David BraybrookeFaculty Council Arts & ScienceDalhousie University

Friends: I showed the minutes of your meeting of 29 September to Pro-fessor Green and held a conversation of sorts with him. He told me thatin general he had felt, turn by turn, amazed, mortified, and amused bythe reception of the literal (though selective) account of one of his typi-cal days. ‘Amazed,’ he said, ‘at the variety of reactions. Some enjoyedit, but thought it gave a picture too enviably agreeable. Others thought

374 APPENDIX A

it unbearably grim.’ He went on, ‘I was mortified to find so extensivea mismatch between my colleagues’ sense of humor and my own. Notonly, it appears, do we differ in the number and kinds of things we findfunny. We evidently strongly disagree on whether the least bit of lifeand humour has a place in a public document.’ ‘Really,’ I said; ‘I sus-pect you’re ready to put in a joke anywhere.’ ‘Whenever I think of one,’he replied, without, it seemed to me, a trace of penitence.

‘But what,’ I asked, ‘amused you about the reactions?’ ‘Why,’ hesaid, ‘they were at such cross purposes. Some said the account wasridiculously far from giving an accurate picture of a serious person’s ex-istence. Others evidently thought the degree of industry that it describedwas wildly improbable. Where was the time needed for family andhousehold concerns, not to speak of relaxation? Some specially giftedpeople managed at one and at the same time to think it too frivolous andtoo serious besides.’

‘Ah,’ I said, ‘maybe that’s because you yourself make a mixed im-pression. Come clean: Which is the real you, too frivolous, or too seri-ous?’ He rejoined, emphatically, though rather unhelpfully, ‘Both. Butthe frivolous side hardly showed up in this account. Like G.E. Moore I’drather give close study toThe National Lampoonany day than write phi-losophy, and there was nothing about that.’ I protested, ‘Didn’t Moorespeak of reading novels?The National Lampoonwasn’t even around inhis day, for goodness’ sake.’

Green said, ‘I read novels, too; besides, I’m much more frivolousthan Moore was. I thought whoever wrote the account was a bit heavy-handed. In fact, it looks to me as though he himself had an overdose ofpublic solemnity. I understand he left out a lot of colorful data as notbeing to the point. Even so he didn’t succeed in being solemn enough,did he? Nothing about romance, kisses, fond embraces, making music,sharing an apple with Cherubino (the canary), playing fetch and tugwith the Corgi—though he did get something in about dishwashing.’ Ipointed out, ‘If any of that stuff had been in the account, it would havestruck even more people as completely frivolous.’ ‘Exactly,’ he said;‘yet it would have been truer to life.’

I put it to him: ‘Do you mean to say that the picture as it stands is

AN OPEN LETTER FROM DAVID BRAYBROOKEFACULTY COUNCIL ARTS & SCIENCEDALHOUSIE UNIVERSITY 375more serious than frivolous?’ ‘I do,’ he said. ‘But,’ I asked, ‘don’t yourealize that either way it has jeopardized your chances of exercising anyinfluence in faculty affairs? What chance, for example, do you now haveof being elected to the tenure committee, or the promotions committee,or Faculty Council?’ He grinned; but it was not merely a grin; it was agleeful, even exultant expression. It was as if, to everybody’s surprise,including his own, he had just pulled off a minor coup.

‘It may be worse for you,’ I said, ‘if you’re seriously taken to be liv-ing by such a standard of industry.’ A look of concern crossed Green’sface. ‘I wouldn’t want to be taken,’ he said, ‘to be leaving no room forjoie de vivre. Shouldn’t any serious academic find joy in his work, re-search as well as teaching? ‘We work,’ declares Delacroix, ‘not only toproduce but to give value to time.’ If enjoying work at least as muchas I do is uncommon at Dalhousie, then I must say it will be a betteruniversity when the enjoyment is more common. I hope some peoplewho feel the same way will continue to creep into those committees, ifonly by accident.’

I asked, ‘Where did you get that nugget from Delacroix?’ He an-swered, ‘From my frivolous side; I was reading a book of aphorisms,when I should have been doing something more ambitious.’ ‘Jokingagain,’ I commented; ‘what are you going to do now, spared for therest of your career the prospect of serving on any of those committees?’This time he smiled, just like Moore, seraphically. ‘Ah,’ he said, ‘I’llnever want for something to do.’

Appendix BDavid Braybrooke’s Publications1955-2005

Books

With C. E. Lindblom. A Strategy of Decision: Policy Evaluation as aSocial Process. New York: The Free Press, 1963. Paperback edition,1970. Translated into Portuguese asUna Estrategia de Deciso Social.Rio de Janeiro: Zahar Editores, 1972.

Philosophical Problems of the Social Sciences. New York: Macmil-lan, 1965. Edited, contributing an introduction and an integrating run-ning commentary.

Three Tests for Democracy: Personal Rights; Human Welfare; Col-lective Preference. New York: Random House, 1968.

Traffic Congestion Goes through the Issue Machine (in the UnitedKingdom, 1963 1968): A Case Study In Issue Processing, Illustrating aNew Approach. London: Routledge & Kegan Paul, 1974.

Ethics in the World of Business. Philosophy and Society Series. To-towa, N.J.: Rowman and Littlefield, 1983. Edited, contributing severalchapters (see below, under “Chapters in Books”) and an integrating run-ning commentary.

Philosophy of Social Science. Englewood Cliffs, N.J.: Prentice Hall,1987. [Foundations of Philosophy Series].

Meeting Needs. Princeton: Princeton University Press, 1987.With B. Brown and P. K. Schotch.Logic on the Track of Social

Change. Oxford: Clarendon Press, 1995. [Clarendon Library of Logicand Philosophy].

377

378 APPENDIX B

Social Rules: Origin; Character; Logic; Change. Boulder and Ox-ford: Westview Press (HarperCollins), 1996. Edited, contributing sev-eral chapters (see below, under Chapters in Books) and an integratingrunning commentary.

Moral Objectives, Rules, and the Forms of Social Change. Toronto:University of Toronto Press, 1998.

Natural Law Modernized. Toronto: University of Toronto Press,2001.

Utilitarianism: Restorations; Repairs; Renovations. Toronto: Uni-versity of Toronto Press. 2004.

Analytical Political Philosophy: From Discourse, Edification. Toronto:University of Toronto Press, 2005.

Journal Articles

“Farewell to the New Welfare Economics.”Review of Economic Studies22 (June 1955): 180 193.

“Berkeley on the Numerical Identity of Ideas.”Philosophical Review64 (October 1955): 631 636.

“Professor Stevenson, Voltaire, and the Case of Admiral Byng.”Jour-nal of Philosophy53 (November 1956): 787 796.

“Vincula Vindicata.”Mind 66 (April 1957): 222 227.“The Expanding Universe of Political Philosophy.”Review of Meta-

physics11 (June 1958): 648 672.“Diagnosis and Remedy in Marx’s Doctrine of Alienation.”Social

Research25 (Autumn 1958): 325 345.“The Relevance of Norms to Political Description.”American Polit-

ical Science Review52 (December 1958): 989 1006.“Authority as a Subject of Social Science and Philosophy.”Review

of Metaphysics13 (March 1960): 469 485.“The Ethical Control of Politics.”Ethics70 (July 1960): 316 321.With Others. “Some Questions for Miss Anscombe about Intention.”

Analysis22 (January 1962): 49 54.“Collective and Distributive Generalization in Ethics.”Analysis23

(December 1962): 45 48.

JOURNAL ARTICLES 379

“Personal Beliefs without Private Languages.”Review of Metaphysics16 (June 1963): 672 686.

“The Mystery of Executive Success Re-examined.”AdministrativeScience Quarterly8 (March 1964): 533 560.

“How Are Moral Judgments Connected with Displays of Emotion?”Dialogue4 (September 1965): 206 223.

“The Choice Between Utilitarianisms.”American Philosophical Quar-terly 4 (January 1967): 28 38.

“Un Esempio Americano di Gestione Operaia.”Biblioteca della Lib-ert (Gennaio Febbraio 1968): 23 38.

“Taking Liberties with the Concept of Rules.”The Monist52 (July1968): 329 358.

“Refinements of Culture in Large Scale History.”History and The-ory, Beiheft 9 (1969): 39 63.

“Three Separable Tests” (Reply to review essays by Robert Ginsbergand Joseph Margolis onThree Tests for Democracy), Philosophy Forum9 (March 1971): 140-147.

“Dialectic in History: Rejoinder to Professor Mulholland’s Com-ments.”Dialogue10 (March 1971): 103 108.

“The Firm but Untidy Correlativity of Rights and Obligations.”Cana-dian Journal of Philosophy1 (March 1972): 351 363.

With Alexander Rosenberg. “Getting the War News Straight: theActual Situation in the Philosophy of Science.”American Political Sci-ence Review66 (September 1972): 818-826.

With Alexander Rosenberg. “Anti-Behaviourism in the Hour of ItsDisintegration.” (A review essay on Explanation in the Behavioural Sci-ences, ed. Robert Borger & Frank Cioffi.)Philosophy of the SocialSciences, (December 1972), 355-63.

“The Logical Character of Verdicts: A Case Study.”University ofToronto Law Journal(Autumn 1972): 221 236.

“Two Blown Fuses in Goldman’s Analysis of Power.”PhilosophicalStudies24 (November 1973): 369 377.

“Utilitarianism with a Difference: Rawls’s Position in Ethics.” (Acritical notice of John Rawls, A Theory of Justice).Canadian Journalof Philosophy3 (December 1973): 303 331.

380 APPENDIX B

“The Philosophical Scene in Canada.”Canadian Forum53 (January1974): 29 34.

“From Economics to Aesthetics: The Rectification of Preferences.”Nos8 (March 1974): 13 24.

“Comment mettrea l’epreuvre le soupcon selon lequel une croyanceest ideologique?”Bulletin de la Societe de Philosophie du Quebec(aout1975): 47 49.

“The Insoluble Problem of the Social Contract.”Dialogue15 (March1976): 3 37.

“A la recherche d’une justice de structure au Nouveau Brunswick.”Philosophiques3 (avril 1976): 123-129.

“Still at the Fuse Box, Arguing with Goldman about the Power Sup-ply.” Philosophical Studies30 (October 1976): 269 272.

Critical notice of Peter D. McClelland,Causal Explanation and ModelBuilding in History, Economics And the New Economic History. In His-tory and Theory(1977): 337 354.

“Our Natural Bodies, Our Social Rights.” (A comment on a paperby Samuel C. Wheeler III),Nos14 (May 1980): 195 202.

“The Maximum Claims of Gauthier’s Bargainers: Are the Fixed So-cial Inequalities Acceptable?”Dialogue21 (September 1982): 411 429.Rejoinder to Gauthier’s reply: 445 448.

“The Possibilities of Compromise.” (A review essay onCompromisein Ethics, Law, and Politics (Nomos XXI), edited by J. R. Pennock andJ. W. Chapman.)Ethics93 (October 1982): 139 150.

“Preferences Opposed to the Market: Grasshoppers vs. Ants on Se-curity, Inequality, and Justice.”Social Philosophy and Policy2 (Autumn1984): 101 114.

“Work Related Ethical Attitudes: Commentary.”Business & Profes-sional Ethics3 (Winter 1984): 41 42.

“Scale, Combination, and Opposition: A Rethinking of Incremen-talism.” (A review essay on Jennifer Hochschild,The New AmericanDilemma.) Ethics95 (July 1985): 920 933.

“Marxism and Technical Change: Nicely Told, But Not the FullContradictory Story.” (Critical notice of two books relating to Marx-

JOURNAL ARTICLES 381

ism by Jon Elster.)Canadian Journal of Philosophy16 (March 1986):123 136.

“La cicala e la formica: una parabola sul mercato.”Biblioteca dellaLiberta 93 (Aprile Giugno 1986): 5 22. (Italian translation of ”Pref-erences Opposed to the Market: Grasshoppers vs. Ants on Security,Inequality, and Justice”.)

“Economic Thinking in Political Science: Maximizing on the Wayto Tautology.”Manuscrito10 (1987), 19-42.

“Social Contract Theory’s Fanciest Flight.”Ethics97 (July, 1987):750-764.

“Thoughtful Happiness.” (A review essay on James Griffin,WellBeing, and Richard Warner, Freedom, Environment, and Happiness.)Ethics99 (April 1989): 625 636.

Review essay on A. Sen et al., The Standard of Living. InEconomicsand Philosophy6 (Fall 1990): 339-350.

“Exaugural Lecture: Conceptual Tunnelling B A Mole’s Eye Viewof Philosophy.”Dalhousie Review70 (Fall 1990): 288-306.

“The Natural Law of Moral Education.”The Vital Nexus1 (May1991): 26-32.

“No Virtues Without Rules; No Rules Without Virtues.”Social The-ory and Practice17 (Summer 1991): 139-156.

“Liberalism’s Claim to Culture.” (Critical notice of Will Kymlicka,Liberalism, Community and Culture.) Dialogue30 (Spring 1991): 117-121.

Preface to “The Future of Post-Modernism.” InTransactions of theRoyal Society of Canada, 6th series, 3 (1992): 99-106.

“Inward and Outward with the Modern Self.” (A critical notice ofCharles Taylor,Sources of the Self.) Dialogue33 (Winter 1994): 101-108.

“Economic Theory Stalled: Model Theoretic Institutionalism as aWay Forward.” (A critical notice of books by A. Rosenberg and D.Hausman.)Dialogue34 (Summer 1995): 623-631.

With Herbert A. Simon. “In Memoriam: Lewis Anthony Dexter.”PS: Political Science and Politics28 (September 1995): 545-546.

382 APPENDIX B

“Justice in Jeopardy If Needs Not Met: A Reply to Gillian Brock.”Dialogue37 (Fall 1998): 799-804.

“The General Theory of Absurdities in Daily Life and Politics.”Dal-housie Review80 (Spring 2000): 45-63.

“Moral Rigidity Inside and Outside the Law.”Public Affairs Quar-terly 16 (April 2002): 173-188.

“A Progressive Approach to Personal Responsibility for Global Benef-icence.”The Monist86 (April 2003): 301-322

With Michael Bailey. “R. A. Dahl’s Philosophy of Democracy, Ex-hibited in His Essays.”Annual Review of Political Science6 (2003):99-118.

“Toward an Alliance Between the Issue-Processing Approach andPragma-Dialectical Analysis.”Argumentation17 (2003): 513-35. (DBserved as honorary editor of the special issue on “Conversation, Argu-mentation, and the Resolution of Issues,” in which this article appearedand also wrote a brief introduction to the issue.)

“What Truth Does the Emotive-Imperative Answer to the Open-Question Argument Leave to Moral Judgments?”Journal of Value In-quiry 37 (2003): 341-52.

“The Relation of Utilitarianism to Natural Law Theory.”The GoodSocietyto appear 2004 or 2005.

“Where Does the Moral Force of the Concept of Needs Reside andWhen?”Philosophyto appear 2004 or 2005.

Chapters in Books

“The Public Interest: The Present and Future of the Concept.” In CarlJ. Friedrich, ed.,The Public Interest (Nomos V). New York: AthertonPress, 1962, 129-154.

“An Illustrative Miniature Axiomatic System.” In Nelson W. Polsby,Robert A. Dentler, and Paul A. Smith, eds.,Politics and Social Life.Houghton Mifflin, 1963, 119-130.

“Marx on Revolutionizing the Mode of Production.” In Carl J. Friedrich,ed., Revolution (Nomos VIII). New York: Atherton Press, 1966, 240-246.

CHAPTERS IN BOOKS 383

“Economics and Rational Choice.” In Paul Edwards, ed., TheEncy-clopedia of Philosophy, vol. II. New York: Macmillan & Free Press,1967, 454-458.

“Ideology.” In Paul Edwards, ed.,The Encyclopedia of Philosophy,vol. IV. New York: Macmillan & Free Press, 1967, 124-127.

“The Duality of Politics and Ethics.” InThe Seventh Inter AmericanCongress of Philosophy: Proceedings of the Congress. Quebec: LesPresses de l’Universite Laval, 1967, 150-161.

“Skepticism of Wants, and Certain Subversive Effects of Corpo-rations on American Values.” InHuman Values and Economic Policy(N.Y.U. Institute of Philosophy for l966), edited by Sidney Hook. NewYork: New York University Press, 1967, 224-239.

“Let Needs Diminish that Preferences May Prosper.” InStudies inMoral Philosophy, American Philosophical Quarterly Monograph Se-ries, no. 1. Oxford: Basil Blackwell, 1968, 86-107.

“Values and Managers: Private Production of Public Goods.” In KurtBaier and Nicholas Rescher, eds.,Values and the Future. New York:Free Press, 1969, 368-388.

“The Logic of the Succession of Cultures.” In Howard E. Kiefer andMilton K. Munitz, eds.,Mind, Science, and History, Vol.II of Contem-porary Philosophic Thought: The International Philosophy Year Con-ferences at Brockport. Albany: State University of New York Press,197O, 270-283.

“Revolution Intelligible or Unintelligible.” In Robert H. Grimm andAlfred F. Mackay, eds.,Society: Revolution and Reform, Proceedingsof the 1969 Oberlin Colloquium in Philosophy. Cleveland: Press ofCase Western Reserve University, 1971, 93-118. Reply to comment byMarshall Cohen, 129-134.

Comment on re apportionment and political democracy. In NelsonW.Polsby, ed.,Re apportionment in the 70’s. Berkeley: University ofCalifornia Press, 1971, 112-118.

With Alexander Rosenberg. “Vincula Revindicata.”, In Tom Beauchamp,ed.,Philosophical Problems of Causation. Encino & Belmont:CA Dick-enson, 1974, 217-222.

“The Meaning of Participation and of Demands for It.” In J. R. Pen-

384 APPENDIX B

nock and J.W. Chapman,Participation in Politics (Nomos XVI). NewYork: Lieber Atherton, 1975, 54-88.

“Variety among Hierarchies of Preference.” In C. A. Hooker, J. J.Leach, and E. F. McClennen, eds.,Foundations and Applications of De-cision Theory, vol. I.. Dordrecht, Holland: D. Reidel, 1978, 55-65.

“Policy Formation with Issue Processing and Transformation of Is-sues.” In C. A. Hooker, J. J. Leach, and E. F. McClennen, eds.,Founda-tions and Applications of Decision Theory, vol. II.Dordrecht, Holland:D. Reidel, 1978, 1-16.

“Self Interest in Times of Revolution and Repression: Comment onProfessor Tullock’s Analysis.” In H. J. Johnson, J. J. Leach, and R. G.Muehlmann, eds.,Revolutions, Systems and Theories. Dordrecht, Hol-land: D. Reidel, 1979, 61-74.

“Through Epistemology to the Depths of Political Illusion.” In MariaJ. Falco, ed.,Through the Looking Glass: Epistemology and the Con-duct of Political Inquiry. Washington, D.C.: University Press of Amer-ica, 1979, 59-82.

With Peter K. Schotch. “Cost Benefit Analysis Under the Constraintof Meeting Needs.” In Norman E. Bowie, ed.,Ethical Issues and Gov-ernment. Philadelphia: Temple University Press, 1981, 176-197.

“Work: A Cultural Ideal Ever More in Jeopardy.” In P. A. French,T. E. Uehling, Jr, and H. K. Wettstein, eds.,Social and Political Phi-losophy: Midwest Studies in Philosophy, 7. Minneapolis: University ofMinnesota Press, 1982, 321-341.

“Making Justice Practical.” In D. Braybrooke and M. Bradie, eds.,Social Justice. Bowling Green, Ohio: Bowling Green Studies in Ap-plied Philosophy, 1982, 1-14. Rejoinder to comment by James P. Sterba,20 23.

“Can Democracy Be Combined with Federalism or with Liberal-ism?” In J. R. Pennock and J. W. Chapman, eds.,Liberal Democracy(Nomos XXV). New York: New York University Press, 1983, 109-118.

“Justice and Injustice in Business.” In Tom Regan, ed.,Just Busi-ness: New Introductory Essays in Business Ethics, edited by Tom Re-gan. New York: Random House, 1983, 167-201.

“Putting Things Together: Three Leading Subjects of Moral Con-

CHAPTERS IN BOOKS 385

cern in the World of Business,” in D. Braybrooke, ed.,Ethics in theWorld of Business. Totowa, N. J.: Rowman & Allanheld, 1983, 461-470.

“Moral Questions About the Business System as a Whole.” In D.Braybrooke, ed.,Ethics in the World of Business. Totowa, N. J.: Row-man & Allanheld, 1983, 471-477.

“La science et la technologie nous permettront elles de continuertravailler? Question que l’ideologie engendre dans la pensee bourgeoiseet aussi dans celle de Marx.” In C. Savary et C. Panaccio, eds.,L’idologieet les strategies de la raison. La Salle, QC: Hurtubuise, 1984, 171-200.

“Contemporary Marxism on the Autonomy, Efficacy, and Legiti-macy of the Capitalist State.” In Roger Benjamin and Stephen L. Elkin,eds. The Democratic State. Lawrence: University Press of Kansas,1985, 59-86.

“A Public Goods Approach to the Theory of the General Will.” In J.Porter, and R. Vernon, eds.,Unity and Plurality in Politics: Festschriftfor Frederick Barnard. London: Croom Helm, 1986, 75-92..

“The Sorites: Logic Astray for Want of Observation.” InThe Factof the Matter: Essays in Honour of Lawrence Resnick. Burnaby, B.C.:Department of Philosophy, Simon Fraser University, 1988, 1-4.

“The Conditions on Which Rules Exist.” In P. A. Schlipp and L.E. Hahn,The Philosophy of Georg Henrik von Wright, The Library ofLiving Philosophers La Salle, IL: Open Court, 1989, 187-209.

“Campaigning for Increased Measures of Safety: Paradox, The Shadowof a Paradox, and a Conceptual Way Out.” In J. P. Rothe, ed.,Challeng-ing the Old Order: Towards New Directions in Traffic Safety Theory,edited by J. P. Rothe. New Brunswick, N.J.: Transaction Books, 1990,169-182.

“How Do I Presuppose Thee? Let Me Count The Ways: The Re-lation of Regularities to Rules in Social Science.” In P. A. French, T.E. Uehling, Jr, and H. K. Wettstein, eds.,The Philosophy of the Hu-man Sciences: Midwest Studies in Philosophy, vol. XV.Notre Dame,Indiana: University of Notre Dame Press, 1990, 80-93.

“Gauthier’s Foundation for Ethics Under the Test of Application.”In Peter Vallentyne, ed.,Contractarianism and Rational Choice. New

386 APPENDIX B

York: Cambridge University Press, 1991, 56-70.“Meeting Needs: Towards a New Needs-Based Ethics.” In K. Aman,

ed.,Ethical Principles for Development. Montclair, N.J.: Institute forCritical Thinking, 1991, 80-90.

With Arthur P. Monahan. “Common Good.” In L. Becker and C.Becker,Encyclopedia of Ethics. New York: Garland Publishing, 1992,Vol. I, 175-178.

With L. A. Dexter. “Corruption.” In L. Becker and C. Becker, eds.,Encyclopedia of Ethics. New York: Garland Publishing, 1992, Vol. II,218-220.

“Needs and Interests.” In L. Becker and C. Becker, eds.,Encyclope-dia of Ethics. New York: Garland Publishing, 1992, Vol. II, 894-897.

“Two Conceptions of Needs in Marx’s Writings.” In R. Beehler etal., eds.,On he Track of Reason: Essays in Honor of Kai Nielsen. Boul-der and Oxford: Westview Press, 1992, 119-133.

“The Social Contract and Property Rights Across the Generations.”In Peter Laslett and James S. Fishkin, eds.,Justice Between Age Groupsand Generations[Philosophy, Politics, and Society, 6th Series]. NewHaven: Yale University Press, 1992, 107-126.

”“Y-a-t-il une connaissance philosophique de la societe?” In J. Proustand E. Schwartz, eds.,La connaissance philosophique: essais sur l’oeuvrede Gilles-Gaston Granger. Paris: Presses Universitaires de France,1995, 81-90.

“Interests and Needs.” In P. H. Werhane and R. E. Freeman, eds.,Blackwell Encyclopedic Dictionary of Business Ethics. Oxford: Black-well, 1997, 341-3.

A“The Representation of Rules in Logic and Their Definition.” InDavid Braybrooke, ed.,Social Rules: Origin; Character; Logic; Change.Boulder and Oxford: Westview Press/HarperCollins, 1996, 3-19.

“Changes of Rules, Issue-Circumscription, and Issue-Processing.”In David Braybrooke, ed.,Social Rules: Origin; Character; Logic;Change. Boulder and Oxford: Westview Press/HarperCollins, 1996,75-83.

“Logic and Historical Enquiry.” In D. Woolf, ed.,A Global Encyclo-pedia of Historical Writing. New York: Garland, 1998, Vol. II, 570-1.

BOOK REVIEWS 387

“Philosophy of Social Science, Contemporary Theories (Schools).”In E. Craig, ed.,The Routledge Encyclopedia of Philosophy. London:Routledge, 1998, Vol. VIII, 838-47.

“The Concept of Needs, with a Heart-Warming Offer of Aid to Util-itarianism.” In G. Brock, ed.,Necessary Goods: Our Responsibilities toMeet Others’ Needs. Rowman and Littlefield, 1998.

“Common Good.” In C.B. Gray, ed.,The Philosophy of Law: AnEncyclopedia. New York: Garland, 1999, 125-7.

With Arthur P. Monahan. “Common Good.” In L. Becker and C.Becker, eds,Encyclopedia of Ethics, 2d ed. New York and London:Routledge, 2000, Vol. I, 262-66. (A revision of the article on the sametopic in the first edition, 1992.)

“Interests.” In L. Becker and C. Becker, eds.,Encyclopedia of Ethics,2d ed.New York and London: Routledge, 2000, Vol. II, 868-71.

“Needs.” In L. Becker and C. Becker, eds.,Encyclopedia of Ethics,2d ed.New York and London: Routledge, 2000, Vol. II, 1220-24. (Thisarticle and the one just mentioned give separate, revised treatments oftwo topics that the first edition combined in one article.)

(The second edition of Encyclopedia of Ethics carried forward un-changed, except for an addition to the bibliography, an article on “Cor-ruption” written together with the late Lewis Anthony Dexter.)

“Decision Making Systems: Personal and Collective.” In N.J. Smelserand P. B. Baltes (eds.)The International Encyclopedia of Social and Be-havioural Sciences. Amsterdam and New York: Elsevier, 2000, Vol. V,3315-18.

“Sidgwick’s Critique of Nozick.” On line in a web site on utilitari-anism established by Paul Lyon.

Book Reviews

Of W.D. Lamont,The Value Judgment. In Philosophical Review(April1957): 255 8.

Of Peter Laslett, ed.,Philosophy, Politics, and Society. In Philo-sophical Review(July 1958): 418 421.

388 APPENDIX B

Of Richard W. Sterling,Ethics in a World of Power: the PoliticalIdeas of Friedrich Meinecke. In Ethics(July 1959): 292 294.

Of Elting E. Morison, ed.,The American Style: Essays in Value andPerformance. In Ethics(October 1959): 73 75.

Of Felix E. Oppenheim,Dimensions of Freedom; An Analysis. InPhilosophical Review(October 1963): 524 528.

Of Arthur S. Banks and Robert B. Textor,A Cross Polity Survey. InCanadian Journal of Economics and Political Science(February 1965):142-144.

Of Lon L. Fuller,The Morality of Law. In Dialogue(March 1965):441 444.

Of Samuel I. Shuman,Legal Positivism: Its Scope and Limitations.In University of Toronto Law Journal(1965): 190 194.

Of Neil A. McDonald, Politics: A Study of Control Behavior. InCanadian Journal of Economics and Political Science(February 1966):108 109.

Of R. S. Downie,Government Action and Morality. In Journal ofPhilosophy(June 1966): 363 367.

Of Arthur C. Danto,Analytical Philosophy of History. In Philoso-phy of Science(December 1967): 388 392.

Of Herbert Marcuse,NegotiationsandAn Essay on Liberation. InTrans action(October 1969): 51 54.

Of Gordon Leff,History and Social Theory. In History and Theory10 (1971): 122-134.

Of Virginia Held, The Public Interest and Individual Interests. InJournal Of Philosophy(April 1972): 192 202.

With Alexander Rosenberg. “Anti Behaviourism in the Hour of itsDisintegration.” (A review essay onExplanation in the Behavioural Sci-ences, edited by Robert Borger and Frank Cioffi.) InPhilosophy of theSocial Sciences(December 1972): 355 363.

Of Leon Dion,Societe et politique la vie des groupes. In CanadianJournal of Political Science(September 1973): 522 523.

Of J.J.C. Smart and Bernard Williams,Utilitarianism: For and Against.In International Studies in Philosophy(Fall 1975): 522-523.

BOOK REVIEWS 389

Of Alan Montefiore, ed.,Philosophie et relations interpersonnelles,jointly with a compte rendu by G. G. Granger of the companion vol-ume of essays by anglophone philosophers,Philosophy and PersonalRelations, edited by Alan Montefiore. InDialogue(March 1975): 159167.

Of S. Korner, ed.,Explanation. In Philosophical Quarterly(January1977): 74 78.

Of J. Manninen and R. Tuomela, eds.,Essays on Explanation andUnderstanding. In Philosophical Review(July 1978): 485 489.

Of S. I. Benn and G. W. Mortimore, eds.,Rationality and the SocialSciences. In Australian Journal of Philosophy(December 1978): 256261.

Of Dennis H. Wrong, Power: Its Forms, Bases and Uses. In Cana-dian Journal of Political Science 14 (June 1981): 437 438.

Of Stanley A. Hoffman,Ethics in International Relations. In Cana-dian Philosophical Reviews(April/June 1982): 107 110.

Of Jane J. Mansbridge,Beyond Adversary Democracy. In Ethics(October 1982): 153-155.

Of William Dray, Perspectives on History. In Dialogue(December1982): 782 783.

Of L. Armour and E. Trott,The Faces of Reason. In Queen’s Quar-terly (Autumn 1983): 688 692.

Of Allan Moscovitch and Glenn Drove, eds.,Inequality: Essays onthe Political Economy of Social Welfare. In Labour/Le Travailleur(Au-tumn 1983): 290 291.

Of Richard W. Miller,Analyzing Marx: Morality, Power, and His-tory. In Dalhousie Review(Spring 1984): 190 193.

Of W. L. Robison et al., eds.,Profits and Professions: Essays inBusiness and Professional Ethics. In Ethics(October 1984): 193 194.

Of Frederick Schick,Having Reasons: An Essay on Rationality andSociality. In Dalhousie Review(Fall 1984): 626-629.

With J. Fingard, of A. W. B. Simpson,Cannibalism and the CommonLaw. In Ethics(April 1985): 745 747.

Of Richard Ashcraft,Revolutionary Politics and Locke’s Two Trea-tises of Government. In Dalhousie Review66 (Fall 1986): 382-385.

390 APPENDIX B

Of Gregory Kavka,Hobbesian Moral and Political Theoryand JeanHampton,Hobbes and the Social Contract Tradition. In TheDalhousieReview66 (Winter 1986/87): 540-545.

Of Amartya Sen et al.,The Standard of Living, In Economics andPhilosophyCambridge: Cambridge University Press, 1987.

Of Arthur P. Monahan,Consent, Coercion and Limit. In DalhousieReview67 (Summer/Fall 1987): 367-369.

Of Alan Irwin, Risk and the Control of Technology. In Risk Abstracts(July 1988): 118 119.

Of David R. Shumway,Michel Foucault. In Dalhousie Review69(Summer 1989): 292-293.

Of Ian Shapiro,Political Criticism. In Canadian Journal of PoliticalScience24 (September 1991): 670-671.

Of Maurice Cranston,The Noble Savage: Jean Jacques Rousseau,1754-1762. In American Political Science Review86 (September 1992):778.

Of L. W. Sumner,Welfare, Happiness, and Ethics, in Philosophy inReview, 17, 2(April 1997), 141-4.

Of Shadia Drury,Leo Strauss and the American Right, in Globe &Mail (Toronto), 4 April 1998, D10.

Notes on Contributors

Nathan Brettis Professor and Chair of Philosophy at Dalhousie Univer-sity in Halifax, Nova Scotia. He teaches philosophy of law and politicsand the history of modern philosophy. Publications include many paperson Hume’s views and discussions of political and legal issues relatingto consent, discrimination, punishment, collective rights, property rightsand justice, and patents on living organisms. ([email protected])

Bryson Brownis Professor and Chair of Philosophy at the Univer-sity of Lethbridge. His philosophical interests center on logic, espe-cially paraconsistent logic, and philosophy of science. He is co-author,with David Braybrooke and Peter Schotch, ofLogic on the Track ofSocial Change(1995). Recent articles include ’Chunk and Permeate’(with Graham Priest) in theJournal of Philosophical Logic, and ’ThePragmatics of Empirical Adequacy’ in theAustralasian Journal of Phi-losophy. ([email protected])

Steven Burnsis Professor of Philosophy at Dalhousie University andProfessor of Contemporary Studies at the University of King’s College,and has served as Chair of both programmes. He has recently publishedinvestigations of Otto Weininger’s influence on Wittgenstein (see, e.g.,Wittgenstein Reads Weininger, eds. D. Stern and B. Szabados, CUP,2004), and is currently working on some studies of moral philosophyin literature. He has been a colleague of David Braybrooke since the1960s. ([email protected])

Richmond Campbellis Professor and Munro Chair in Philosophy atDalhousie University. He is the author ofIllusions of Paradox: A Fem-inist Epistemology NaturalizedandSelf-Love and Self-Respectand co-editor of Naturalized Moral EpistemologyandParadoxes of Rational-ity and Cooperation. His publications include articles in moral theory,epistemology, feminist theory, logic, philosophy of biology, and philos-

391

392 NOTES ON CONTRIBUTORS

ophy of mind. He is currently preparing a book on moral judgment.([email protected])

Sue Campbellis Associate Professor of Philosophy and Coordinatorof Women’s Studies at Dalhousie University. She is the author ofInter-preting the Personal: Expression and the Formation of Feeling(1997),and Relational Remembering: Rethinking the Memory Wars(2003).([email protected])

Michael Hymersis currently Associate Professor of Philosophy atDalhousie University. He received his PhD from the University of Al-berta in 1993 and held Postdoctoral Fellowships from the Social Sci-ences and Humanities Research Council of Canada and the Calgary In-stitute for the Humanities. He has published numerous articles on ques-tions concerning epistemology, philosophy of language, metaphysics,Kant, Wittgenstein, and neopragmatism. His bookPhilosophy and ItsEpistemic Neuroses(Westview Press) appeared in 2000. ([email protected])

Edna Keebleis an associate professor of political science at SaintMary’s University and holds a Ph.D. from Dalhousie University (1994).As one of Braybrooke’s many graduate students who has been influ-enced not only by the weight of his ideas but also by the strength of hischaracter, she regards both his scholarly work and his life well-lived asa continuing inspiration to be generous to others in time, monies andfriendship. ([email protected])

Meredith Ralstonis an Associate Professor in the Departments ofWomen’s Studies and Political Studies at Mount Saint Vincent Univer-sity. She is also a documentary filmmaker and has directed two filmson women and politics for the National Film Board of Canada. KieferSutherland narrates her latest film, on Filipino bar girls and sex tourism.She has just been included in the 2005 edition of Canadian Who’s Who.([email protected])

Peter Schotchis a Professor of Philosophy at Dalhousie University.He received his university education at the University of Waterloo. Bytraining, he is a formal logician with an interest in modal and many-valued logics (which were then not very respectable areas of research—a fact which may go some way to explaining his interest). Over the yearsSchotch has become much more interested in applications of formal sci-

393

ence which explains his interest in collaborating with David Braybrookeand Bryson Brown on the bookLogic on the Track of Social Change.([email protected])

Susan Sherwin, FRSC, is a University Research Professor at Dal-housie University. She is a Professor of Philosophy, Bioethics, andWomen’s Studies. Her principal research interests are in feminist ap-proaches to health care ethics. Her publications includeNo LongerPatient: Feminist Ethics and Health Care(1992), andThe Politics ofWomen’s Health: Exploring Agency and Autonomy, with the CanadianFeminist Health Care Ethics Research Network (1998). ([email protected])

S. L. (Sharon) Sutherland, FRSC, is Visiting Professor at the Schoolof Political Studies at the University of Ottawa, working closely withthe Public Administration Program of the School. She is also adjunctprofessor in the Department of Political Science at Dalhousie, where shebenefited from David Braybrooke’s collegiality in the 1970s, and a con-sultant to government. Beginning her career in attitude theory in polit-ical and social psychology (for example,Patterns of Belief and Action,University of Toronto Press, 1981), she later switched to study of therepresentative institutions and bureaucracy (Bureaucracy in Canada,Control and Reform, University of Toronto Press, 1985), ministerial re-sponsibility, machinery of government (the politics of institutional de-sign), and management policies and their political implications. She haspublished a number of journal articles in these areas, as well as in theethics of office holding.

Tom Vinciis Professor of Philosophy at Dalhousie University in Hal-ifax, arriving there upon completion of his Ph.D in Philosophy at theUniversity of Pittsburgh in 1977. He has been a colleague and friendof David Braybrooke since then. Professor Vinci’s publications includea book on Descartes,Cartesian Truth(Oxford, 1998), and articles inEpistemology, Philosophy of Science, Decision Theory and the Historyof Modern Philosophy. He lives in Halifax with his wife and three chil-dren. ([email protected])

References

Acharya, A. (2001, Summer). Human security: East vs west.Interna-tional Journal 56(3), 442–60.

Aiken (1948). Hume’s Moral and Political Philosophy. New York:Hafner.

Alberta, A. G. (2004, July 27).Report of the Auditor General on theAlberta government’s BSE-related assistance programs. Governmentof Alberta.

Anscombe, G. (1963).Intention(2nd ed.). Ithaca: Cornell UniversityPress.

Audi, R. (1999). Moral knowledge and ethical pluralism. In J. Grecoand E. Sosa (Eds.),The Blackwell Guide to Epistemology. Oxford:Blackwell.

Axtell, G. (1997). Recent work on virtue epistemology.AmericanPhilosophical Quarterly 34, 1–26.

Axworthy, L. (2001). Introduction. In McRae and Huber (Eds.),Hu-man Security and the New Diplomacy: Protecting People, PromotingPeace. Montreal and Kingston: McGill-Queen’s University Press.

Axworthy, L. (2003).Navigating a New World. Toronto: Knopf Canada.

Ayer, A. (1936).Language Truth and Logic. 2nd(1946). London: VictorGollancz.

Ayoob, M. (2004). Third world perspectives on humanitarian interven-tion and international administration.Global Governance 10, 99–118.

395

396 REFERENCES

Berelson, B. (1952). Democratic theory and public opinion.PublicOpinion Quarterly 16(3), 313–330.

Bish, A., S. Sutton, C. Jacobs, et al. (2002, January). Changes in psy-chological distress after cancer genetic counselling: a comparison ofaffected and unaffected women.British Journal of Cancer 86(1), 43–50.

Blackburn, S. (1998).Ruling Passions. Oxford: Oxford UniversityPress.

Bonjour, L. (1985).The Structure of Empirical Knowledge. Cambridge,MA: Harvard University Press.

Bovens, M. and P. ’THart (1996).Understanding Policy Fiascoes. NewBrunswick, U.S.A.: Transaction Publishers.

Brandom, R. (1994).Making It Explicit. Cambridge, MA: HarvardUniversity Press.

Brandt, R. (1959).Ethical Theory. Englewood Cliffs, N.J.: Prentice-Hall.

Braybrooke, D. (1969). Refinements of culture in large scale history.History and Theory 9, 39–63.

Braybrooke, D. (1970). The logic of the succession of cultures. InH. E. Kiefer and M. K. Munitz (Eds.),Mind, Science, and History,Volume II of Contemporary Philosophic Thought: The InternationalPhilosophy Year Conferences at Brockport, pp. 270–283. Albany:State University of New York Press.

Braybrooke, D. (1985). Appendix b, or a (summer’s) day in the life ofa professor of philosophy.Dalhousie Review 65.

Braybrooke, D. (1987a).Meeting Needs. Princeton University Press.

Braybrooke, D. (1987b).Philosophy of Social Science. Foundations ofPhilosophy Series. Englewood Cliffs, N.J.: Prentice Hall.

397

Braybrooke, D. (1998a).Moral Objectives, Rules, and the Forms ofSocial Change. Toronto: University of Toronto Press.

Braybrooke, D. (1998b). Philosophy of social science, contemporarytheories (schools). In E. Craig (Ed.),The Routledge Encyclopedia ofPhilosophy, Volume VIII, pp. 838–847. London: Routledge.

Braybrooke, D. (2000). Decision making systems: Personal and col-lective. InThe International Encyclopedia of Social and BehaviouralSciences, Volume V, pp. 3315–3318. Amsterdam and New York: El-sevier.

Braybrooke, D. (2001).Natural Law Modernized. Toronto: Universityof Toronto Press.

Braybrooke, D. (2003a, April). A progressive approach to personalresponsibility for global beneficence.The Monist 86, 301–322.

Braybrooke, D. (2003b). Toward an alliance between the issue-processing approach and pragma-dialectical analysis.Argumenta-tion 17, 513–535.

Braybrooke, D. (2003c). What truth does the emotive-imperative an-swer to the open-question argument leave to moral judgments?Jour-nal of Value Inquiry 37, 341–352.

Braybrooke, D. (2004).Utilitarianism: Restorations; Repairs; Renova-tions. Toronto: University of Toronto Press.

Braybrooke, D. (2005).Analytical Political Philosophy: From Dis-course, Edification. Toronto: University of Toronto Press.

Braybrooke, D. and M. Bailey (2003). R. A. Dahls philosophy ofdemocracy, exhibited in his essays.Annual Review of Political Sci-ence 6, 99–118.

Braybrooke, D. and J. Fingard (1985, April). Review of A. W. B. Simp-son,Cannibalism and the Common Law. Ethics, 745–747.

Braybrooke, D. and C. E. Lindbloom (1963).A Strategy of Decision:Policy Evaluation as a Social Process. New York: The Free Press.

398 REFERENCES

Brink, D. (1989).Moral Realism and the Foundations of Ethics. Cam-bridge: Cambridge University Press.

Brown, B. (2003, Fall). Notes on hume and skepticism regarding thesenses.Croatian Journal of Philosophy.

Brown, B. (2004, June). The pragmatics of empirical adequacy.Aus-tralasian Journal of Philosophy 82(2), 242–364.

Campbell, R. (1998).Illusions of Paradox: A Feminist EpistemologyNaturalized. Maryland: Rowman and Littlefield.

Campbell, R. (2003).Moral Epistemology. Stanford University.

Campbell, R. (2005). What is moral judgement.

Campbell, R. and J. Woodrow (2003). Why moores open question isopen: The evolution of moral supervenience.Journal of Value In-quiry 37(3), 353–372.

Campbell, S. Remembering who we are: Responsibility and resistantidentification. Forthcoming.

Chernomas, R. and L. Donner (2004). The cancer epidemic as a socialevent.

Chisholm, R. M. (1966).Theory of Knowledge. Englewood Cliffs, NJ:Prentice Hall.

Cohen, Michael D., J. G. M. and J. P. Olsen (1972). A garbage canmodel of organizational choice.Administrative Science Quarterly 17,1–25.

Connolly, T. (1980). Uncertainty, action and competence: Some al-ternatives to omniscience in complex problem-solving. In S. Fiddle(Ed.), Uncertainty: Behavioral and Social Dimensions. New York:Praeger.

Connolly, T. (1988). Hedge-clipping, tree-felling and the managementof ambiguity: The need for new images of decision-making. In L. R.

399

Pondy et al. (Eds.),Managing Ambiguity and Change, pp. 37–50.New York: John Wiley and Sons.

Copp, D. (1995).Normativity, and Society. Oxford: Oxford UniversityPress.

Coxon, K. (2003). Empathy, intersubjectivity, and virtue.

Crick, B. (1959).The American Science of Politics. Berkeley: Univer-sity of California Press.

Dahl, R. A. (1971). Polyarchy: Participation and Opposition. NewHaven: Yale University Press.

Dancy, J. (1993).Moral Reasons. Oxford: Blackwell.

Davidson, D. (1980). Actions, reasons, and causes. InEssays on Actionsand Events, pp. 3–19. Oxford: Oxford University Press.

Dimock, M. (1958). A Philosophy of Administration. New York:Harper.

Dorval, M., A. Patentaude, K. Schneider, et al. (2000). Anticipatedversus actual emotional reactions to disclosure of results of genetictests for cancer susceptibility: findings from p53 and brca1 testingprograms.Journal of Clinical Oncology 18(10), 2135–42.

Douglas, M. (1995). Converging on autonomy: Anthropology and insti-tutional economics. In O. E. Williamson (Ed.),Organization Theory:From Chester Barnard to the Present and Beyond, pp. 98–115. NewYork: Oxford University Press.

Duran, J. (1995). The possibility of a feminist epistemology.Philosophyand Social Criticism 21(4), 127–140.

Eisen, A. and B. Weber (2001, July 19). Prophylactic mastectomy forwomen with BRCA1 and BRCA2 mutations—facts and controversy.New England Journal of Medicine 345(3), 207–8.

Elshtain, J. B. (2003). International justice as equal regard and the useof force. Ethics and International Affairs 17.

400 REFERENCES

Globe and Mail(2003, 16 September). Farmer should have covered upmad cow, klein says.Globe and Mail.

Le Monde(2003, 31 decembre). Les conducteurs ont change de com-portement sur les routes.Le Monde, 6.

Le Monde(2004a, 28 decembre). 10% de tues en moins en 2004.LeMonde, 8.

Le Monde(2004b, 28 decembre). Le succes des radars confronte auxpressions deselus.Le Monde, 8.

Epstein, L. (1964). A comparative study of canadian parties.The Amer-ican Political Science Review 58, 46–59.

Fay, B. and D. Moon (1977, September). What would an adequatephilosophy of social science look like?Philosophy of the SocialSciences, 21–33.

Feinberg, J. (1978). Psychological egoism. In J. Feinberg (Ed.),Reasonand Responsibility: Readings in Some Basic Problems in Philosophy.Belmont, CA: Wadsworth Publishing.

FFF (2002). Freedom from fear: Canada’s foreign policy for humansecurity. Technical report, DFAIT, Ottawa.

Finnegan, W. (2004, February 9). Underground man.New Yorker, 62.

Fry, B. R. (1989).Mastering Public Administration: From Max Weberto Dwight Waldo. Chatham, New Jersey: Chatham House Publishers.

Fukuda-Parr, S. (2003). The human development paradigm: Opera-tionalizing sen’s ideas on capabilities.Feminist Economics 9(2-3).

Gaetner, G. (2003, 19 juin). Les vrais chiffres : securite routiere.L’Express.

Gibbard, A. (1990). Wise Choices, Apt Feelings. Cambridge, Mass:Harvard University Press.

401

Gigerenzer, G. and R. Selten (Eds.) (2001). Cambridge, Mass.: MITPress.

Glock, H.-J. (2003).Quine and Davidson on Language, Thought andReality. Cambridge: Cambridge University Press.

Goodman, N. (1965).Fact, Fiction and Forecast(2nd ed.). Indianapo-lis: Bobbs-Merrill.

Got, C. (2004, 28 decembre). La baisse de la vitesse a produit unepacification general. Le Monde, 8.

Goudge, T. (1967-68). A century of philosophy in english-speakingcanada.Dalhousie Review 47(4).

Grann, V.R., K. P., W. Whang, et al. (1998). Decision analysisof prophylactic mastectomy and oophorectomy in brca1-positive orBRCA2-positive patients.Journal of Clinical Oncology 16(3), 979–85.

Grover, D., J.L. Camp, Jr, and N. Belnap (1975). A prosentential theoryof truth. Philosophical Studies 27, 73–125.

Gwozdecky, M. and J. Sinclair (2001). Landmines and human security.In McRae and Huber (Eds.),Human Security and the New Diplo-macy: Protecting People, Promoting Peace, pp. 28–40. Montreal andKingston: McGill-Queen’s University Press.

Hall, P. (1981). Great Planning Disasters. London: Weidenfeld andNicholson.

Hamblin, C. (1972). Quandaries and the logic of rules.Journal ofPhilosophical Logic 1(1), 74–85.

Hampson, F. O. (2002).Madness in the Multitude: Human Security andWorld Disorder. Don Mills, ON: Oxford University Press.

Harding (2004).The Feminist Standpoint Theory Reader. New York:Routledge.

402 REFERENCES

Harding, S. (1993). Rethinking standpoint epistemology: ‘what isstrong objectivity’? In L. Alcoff and E. Potter (Eds.),Feminist Epis-temologies. New York: Routledge.

Harding, S. (1996). Feminism, science, and the anti-enlightenment cri-tiques. In A. Garry and M. Pearsall (Eds.),Women, Knowledge andReality: Explorations in Feminist Philosophy. New York: Routledge.

Hare, R. (1952).The Language of Morals. New York: Oxford Univer-sity Press.

Hare, R. (1989).Prescriptivism: The Structure of Ethics and Morals.Oxford: Oxford University Press.

Hayes, M. (1992).Incrementalism and Public Policy. New York: Long-man.

Hempel, C. (1966). Philosophy of Natural Science. New Jersey:Prentice-Hall.

Hollinger, D. A. (2001). Not universalists, not pluralists: The new cos-mopolitans find their own way.Constellations 8(2), 236–48.

Hood, C. et al. (1999).Regulation inside Government: Waste-watchers,Quality Police and Sleaze-busters. Oxford: Oxford University Press.

Hume, D. (1966). An Enquiry Concerning the Principles of Morals(1751 ed.). New York: Free Press.

Hume, D. (1978). A Treatise of Human Nature. Oxford: Oxford:Clarendon Press.

Hymers, M. (2000).Philosophy and Its Epistemic Neuroses. Boulder,CO: Westview Press.

Jacobsen, R. (1997). Semantic character and expressive content.Philo-sophical Papers 26(2), 129–146.

Kaldor, M. (2003a). American power: From compellance to cosmopoli-tanism?International Affairs 79(1).

403

Kaldor, M. (2003b, July 14). Humanitarian intervention: A forum.TheNation, 11–20.

Kaufert, P. (2003).From the laboratory to the clinic: The story of ge-netic testing for hereditary breast cancer, Volume 1, pp. 487–506.Toronto: National Network on Environments and Women’s Health.

Keeble, E. (2004). Whither human security?: Politics, privilege andpossibilities. InConference on the Canada-United States Relation-ship: Convergence or Divergence?, Plattsburgh. Center for the Studyof Canada: State University of NY at Plattsburgh and McGill Univer-sity Institute for the Study of Canada.

Keeble, E. (2005, Spring). Defining canadian security: Continuities anddiscontinuities.American Review of Canadian Studies 35(4), 1–24.

Kim, D. (1999). Contempt and ordinary inequality. In S. Babbitt andS. Campbell (Eds.),Racism and Philosophy. Ithaca, NY: Cornell Uni-versity Press.

King, M., S. Wieand, K. Hale, et al. (2001). Tamoxifen and breastcancer incidence among women with inherited mutations in BRCA1and BRCA2.Journal of the American Medical Association 286(18),2251–6.

Kournay, J. (1987).Scientific Knowledge: Basic Issues in the Philoso-phy of Science. Belmont, CA: Wadsworth Publishing Company.

Kuhn, T. (1962).The Structure of Scientific Revolutions. Chicago: Uni-versity of Chicago Press.

Lasswell, H. D. (1923, March). Chicago’s old first ward: A case studyin political behavior.National Municipal Review, 127.

Lasswell, H. D. (1968). Psychopathology and Politics. New York:Viking Press. First published in 1930.

Liebow, E. (1965).Tally’s Corner. Boston: little Brown.

Lindblom, C. E. (1959, Spring). The science of ”muddling through.Public Administration Review 19, 79–88.

404 REFERENCES

Lindblom, C. E. (1977).Politics and Markets. New York: Basic Books.

Lindblom, C. E. (1979, November/December). Still muddling, not yetthrough.Public Administration Review 39(4), 517–527.

Lindblom, C. E. (1990). Inquiry and Change: The Troubled Attemptto Understand and Shape Society. New Haven and London: YaleUniversity Press.

Lindblom, C. E. and D. K. Cohen (1979).Usable Knowledge: SocialScience and Social Problem Solving. New Haven and London: YaleUniversity Press.

Lipscombe, K. (2004, 29 July). Death on our roads.Halifax Chronicle-Herald, A1.

Lloyd, G. (1984). The Man of Reason: Male and Female in WesternPhilosophy. Minneapolis: University of Minnesota Press.

Lloyd, G. (2000). Individuals, responsibility and the philosophicalimagination. In C. Mackenzie and N. Stoljar (Eds.),Relational Au-tonomy: Feminist perspectives on Autonomy, Agency and the SocialSelf. Oxford: Oxford University Press.

Longino, H. (1993). Subjects, power and knowledge: Description andprescription in feminist philosophies of science. In L. Alcoff andE. Potter (Eds.),Feminist Epistemologies. New York: Routledge.

Longino, H. (1996). Can there be a feminist science. In A. Garry andM. Pearsall (Eds.),Women, Knowledge and Reality: Explorations inFeminist Philosophy. New York: Routledge.

Lowi, T. (1992). Lowi and simon on political science, public adminis-tration, rationality and public choice.Journal of Public Administra-tion Research and Theory 2(2), 105–109.

MacIntosh, D. (1998). Categorically rational preferences and the struc-ture of morality. In P. Danielson (Ed.),Modeling Rationality, Moralityand Evolution, Number 7 in Vancouver Studies in Cognitive Science,pp. 282–301. New York: Oxford University Press.

405

Mackie, J. L. (1977).Ethics: Inventing Right and Wrong. Penguin.

March, J. and H. Simon (1994).Organizations(2nd ed.). CambridgeMA: Blackwell. First published by Wiley, 1958.

Mayor of London (2002). Proposed london-wide transport strategy im-plementation targets. Technical report, London.

McClosky, H. (1964). Consensus and ideology in american politics.American Political Science Review 52(1), 376.

McNaughton, D. (1988).Moral Vision. Oxford: Blackwell.

McRae, R. and D. Hubert (Eds.) (2001).Human Security and theNew Diplomacy: Protecting People, Promoting Peace, Montreal andKingston. McGill-Queen’s University Press.

Meijers-Heijboer, H. et al. (2001, July 19). Breast cancer after pro-phylactic bilateral mastectomy in women with a BRCA1 or BRCA2mutation.New ENgland Journal of Medicine 345(3), 159–164.

Mele, A. R. (2001).Self-Deception Unmasked. Princeton and Oxford:Princeton University Press.

Men (2003, May/June). The men behind ken.London Bulletin, 2.

Meyers, D. (1989). Self, Society, and Personal Choice. New York:Columbia University Press.

Michael Stocker (with Elizabeth Hegeman (1996).Valuing Emotions.Cambridge: Cambridge University Press.

Milbraith, L. W. and M. Goel (1977).Political Participation(2nd ed.).Chicago: Rand McNally.

Miller, D. (2004, January). Holding nations responsible.Ethics 114,240–68.

Millikan, R. (1996). Pushmi-pullyu representations. In L. May,M. Friedman, and A. Clark (Eds.),Mind and Morals. Cambridge,MA: MIT Press.

406 REFERENCES

Moore, G. (1903).Principia Ethica. Cambridge: Cambridge UniversityPress.

Moran, M. (2001, October-December). Not steering but drowning:Policy catastrophes and the regulatory state.The Political Quar-terly 72(4), 414–427.

Nelson, L. (1993). Epistemological communities. In L. Alcoff andE. Potter (Eds.),Feminist Epistemologies. New York: Routledge.

Nelson, L. (1996). Who knows? what can they know? and when?In A. Garry and M. Pearsall (Eds.),Women, Knowledge and Reality:Explorations in Feminist Philosophy. New York: Routledge.

Newman, E. (2001). Human security and constructivism.InternationalStudies Perspectives 2.

Nielsen, J. (Ed.) (1990).Feminist Research Methods: Exemplary Read-ings in the Social Sciences. Boulder: Westview Press.

Nozick, R. (1974).Anarchy, State, and Utopia. New York: Basic Books.

Nyers, P. (2003). Abject cosmopolitanism: The politics of protection inthe anti-deportation movement.Third World Quarterly 24(6), 1069–93.

On, B.-A. B. (1993).Marginality and Epistemic Privilege. New York:Routledge.

Penz, P. (2001). The ethics of development assistance and human secu-rity: From realism and sovereigntism to cosmopolitanism. In R. Irwin(Ed.), Ethics and Security in Canadian Foreign Policy, pp. 38–55.Vancouver: UBC Press.

Piper, A. (1991). Impartiality, compassion, and moral imagination.Ethics 101, 726–757.

Platts, M. (1988). Moral reality. In G. Sayre-McCord (Ed.),Essays onMoral Realism. Ithaca: Cornell University Press.

407

Pogge, T. W. (2002).World Poverty and Human Rights: CosmopolitanResponsibilities and Reforms. Cambridge: Polity Press.

Pollock, J. (1974).Knowledge and Justification. Princeton: PrincetonUniversity Press.

Pollock, J. (1986).Contemporary Theories of Knowledge. Totowa, NJ:Rowman and Littlefield.

Pratt, C. (2001). Moral vision and foreign policy. In R. Irwin (Ed.),Ethics and Security in Canadian Foreign Policy, pp. 59–76. Vancou-ver: UBC Press.

Programme-UNDP, U. N. D. (1994).Human Development Report. NewYork: Oxford University Press.

Quine, W. V. O. (1980). Two dogmas of empiricism. InFrom a Log-ical Point of View(2nd ed.)., pp. 20–46. Cambridge, MA: HarvardUniversity Press.

Reinharz, S. (1992).Feminist Methods in Social Research. New York:Oxford University Press.

Robinson, D. (2001). The international criminal court. In McRae andHuber (Eds.),Human Security and the New Diplomacy: ProtectingPeople, Promoting Peace. Montreal and Kingston: McGill-Queen’sUniversity Press.

Robinson, J. P. et al. (1972).Measures of Political Attitudes. Ann Arbor,Michigan: Institute for Social Research.

Rogers, C. (1975). Necessary and sufficient conditions of therapeuticpersonality change.Journal of Consulting Psychology 21, 95–103.

Rorty, R. (1991). Pragmatism, davidson and truth. InObjectivity, Rel-ativism and Truth, pp. 126–150. Cambridge: Cambridge UniversityPress.

Rose, D. (2001).Revisiting Feminist Research Methodologies: A Work-ing Paper. Ottawa: Status of Women Canada.

408 REFERENCES

Rosemary Foot, J. L. G. and A. Hurrell (Eds.) (2003).Order and Justicein International Relations, New York. Oxford University Press.

Rothman, B. (1986).The Tentative Pregnancy: Prenatal Diagnosis andthe Future of Motherhood. New York: Viking Press.

RTP (2001). The responsibility to protect: Report of the internationalcommission on intervention and state sovereignty. Technical report,International Development Research Centre, Ottawa.

RTP2 (2001). The responsibility to protect: Research, bibliography,background, supplementary volume to the report of the internationalcommission on intervention and state sovereignty. Technical report,International Development Research Centre, Ottawa.

SAOS (2004). Securing an open society: Canada’s national securitypolicy. Technical report, Government of Canada, Ottawa.

Sayre-McCord, G. (1988). Introduction: The many moral realisms. InG. Sayre-McCord (Ed.),Essays on Moral Realism. Ithaca: CornellUniversity Press.

Schechtman, M. (1996).The Constitution of Selves. Ithaca, NY: CornellUniversity Press.

Scheffler, S. (1999, November). Conceptions of cosmopolitanism.Util-itas 11(3).

Schelling, T. (1978).Micromotives and Macrobehavior. New York:W.W. Norton.

Schlosser, W. S. (1983). Ockham’s razor.

Schotch, P. K. (2001). Elements of goal theory. In J. Woods andB. Brown (Eds.),New Studies in Exact Philosophy: Logic, Mathe-matics and Science, pp. 245–259. Oxford: Hermes Science.

Schrag, D., K. Kuntz, J. Garber, et al. (1997). Decision analysis—effects of prophylactic mastectomy and oophorectomy on life ex-pectancy among women with BRCA1 or BRCA2 mutations.NewEngland Journal of Medicine 336(20), 1465–71.

409

Scott, W. R. (1992).Organizations: Rational, Natural and Open Sys-tems(3rd ed.). Englewood Cliffs, N.J.: Prentice Hall.

Seeley, T. (2001). Decision making in superorganisms: How collectivewisdom arises from the poorly informed masses. In G. Gigerenzerand R. Selten (Eds.),Bounded Rationality: The Adaptive Toolbox,pp. 249–262. Cambridge, Mass.: MIT Press.

Sellars, W. (1954, July). Some reflections on language games.Philoso-phy of Science 21(3), 204–228.

Sellars, W. (1960). Being and being known.Proceedings of the Ameri-can Catholic Philosophical Association, 28–49.

Sellars, W. (1962).Philosophy and the Scientific Image of Man, pp.35–78. Pittsburgh: University of Pittsburgh Press.

Sellars, W. (1963).Science, Perception and Reality. London: Routledgeand Kegan Paul.

Selten, R. (2001). What is bounded rationality? In G. Gigerenzer andR. Selten (Eds.),Bounded Rationality: The Adaptive Toolbox, pp.13–36. Cambridge, Mass.: MIT Press.

Sen, A. (1999).Development as Freedom. New York: Knopf.

SFP (1999, April). Human security: Safety for people in a changingworld. concept paper, Department of Foreign Affairs and Interna-tional Trade—DFAIT, Ottawa.

Shapere, D. (1982, De3cember). The concept of observation in scienceand philosophy.Philosophy of Science 49, 485–525.

Shaver, R. (1998).Rational Egoism: A Selective and Critical History.Cambridge: Cambridge University Press.

Sherman, N. (1998a).Empathy and Imagination. Notre Dame: Univer-sity of Notre Dame Press.

Sherman, N. (1998b). Empathy, respect, and humanitarian intervention.Ethics and International Affairs 12, 103–119.

410 REFERENCES

Sherwin, S. (2001).Normalizing Reproductive Technologies and theImplications for Autonomy, pp. 96–113. Boulder: Westview Press.

Sherwin, S. (2003). The importance of ontology for feminist policy-making in the realm of reproductive technology.Canadian Journalof Philosophy Supplementary volume 26.

Simon, H. (1976). Administrative Behavior(3rd ed.). New York:Macmillan.

Simon, H. (1980, July). The behavioral and social sciences.Sci-ence 209(4), 72 – 78.

Simon, H. A. (1957).Models of Man. New York: John Wiley.

Simon, H. A. (1970, September). Rational decision making in businessorganizations.American Economic Review 69(4).

Simon, H. A. (1977).The New Science of Management Decision(3rded.). Englewood Cliffs, N.J.: Prentice-Hall.

Simon, H. A. (1990). Invariants of human behavior.Annual Review ofPsychology 41, 1–19.

Smith, A. (1976/1759).The Theory of Moral Sentiments. Indianapolis:Liberty Classics.

Smith, L. (1950).Killers of the Dream. London: The Cresset Press.

Smith, M. (1994).The Moral Problem. London: Blackwell.

Sober, E. and D. S. Wilson (1998).Unto Others: The Evolution andPsychology of Unselfish Behavior. Cambridge, MA: Harvard Univer-sity Press.

Sosa, E. (1980). The raft and the pyramid. InStudies in Epistemology,Volume 5 ofMidwest Studies in Philosophy, pp. 3–25. Minneapolis:University of Minnestoa Press.

Stefanek, M., K. Helzsouer, P. Wilcox, et al. (1995). Predictors ofand satisfaction with bilateral prophylactic mastectomy.PreventiveMedicine 24(4), 412–419.

411

Stevenson, C. (1944).Ethics and Language. New Haven, CT: YaleUniversity Press.

Stewart, F. and S. Deneulin (2002). Amartya sen’s contribution to de-velopment thinking. InSymposium on Development as Freedom byAmartya Sen, Volume 37, pp. 61–70. Studies in Comparative Interna-tional Development.

Storing, H. J. (Ed.) (1962).Essays on the Scientific Study of Politics.New York: Holt Rinehart and Winston.

Sutcliffe, S. (2001). Guest editorial on pulbic health aspects of breastcancer testing in canada.Chronic Diseases 20(1).

Sutcliffe, S., P. Pharoah, D. Easton, and B. Ponder (2001, July 1). Ovar-ian and breast cancer risks to women in families with two or morerelatives with ovarian cancer.International Journal of Cancer 87(1),110–117.

Sutherland, S. L. (1995, September). The problem of dirty hands inpolitics: Peace in the vegetable trade.Canadian Journal of PoliticalScience 28(3), 497–507.

Sutherland, S.L. and E. Tanenbaum (1980, February). Submissive au-thoritarians.Canadian Review of Sociology and Anthropology 17(1),1–24.

Sutherland, S.L. and E. Tanenbaum (1984, June). Irrational vs rationalbases of political preference: Elite and mass perspectives.PoliticalPsychology 5(2), 173–197.

Tanguy, J. (2003). Redefining sovereignty and intervention.Ethics andInternational Affairs 17.

Thomas, L. (1980). Ethical egoism and psychological dispositions.American Philosophical Quarterly 17, 73–78.

Thompson, D. F. (1993). Mediated corruption: The case of the keatingfive. American Political Science Review 87, 369–381.

412 REFERENCES

Tinbergen, J. (1952).On the Theory of Economic Policy. Amsterdam:North-Holland Publishing.

Titchener, E. (1915).Beginner’s Psychology. New York: Macmillan.

Tronto, J. (1993).Moral Boundaries: A Political Argument for an Ethicof Care. New York: Routledge.

Unic, I., P. Stalmeir, L. Verhoef, et al. (1998, July-September). As-sessment of the time-tradeoff values for prophylactic mastectomy ofwomen with a suspected genetic predisposition to breast cancer.Med-ical Decision Making 18(3), 268–77.

van Gunsteren, H. (1976).The Quest for Control: A critique of therational-central-rule approach in public affairs. London: John Wileyand Sons.

von Wright, G. H. (1963).Norm and Action. London: Routledge.

Waldron, J. (1995). Minority cultures and the cosmopolitan alternative.In W. Kymlicka (Ed.),The Rights of Minority Cultures, pp. 93–119.New York: Oxford University Press.

Walker, M. (1998a).Ineluctable Feelings and Moral Recognition. NotreDame: University of Notre Dame Press.

Walker, M. (1998b).Moral Understandings: A Feminist Study in Ethics.New York: Routledge.

Weiner, B. (2003, April). A nave psychologist examines bad luck andthe concept of responsibility.The Monist 86(2).

Westcott, M. (1990).Feminist Criticism of the Social Sciences. Boulder:Westview Press.

Wheeler, N. (2000).Saving Strangers: Humanitarian Intervention inInternational Society. Oxford: Oxford University Press.

Wicksteed, P. H. (1910).The Common Sense of Political Economy.London: Macmillan.

413

Williams, M. (1996).Unnatural Doubts: Epistemological Realism andthe Basis of Scepticism. Princeton: Princeton University Press.

Williams, M. (2001). Problems of Knowledge: A Critical Introductionto Epistemology. Oxford: Oxford University Press.

Williamson, O. E. (Ed.) (1995).Organization Theory: From ChesterBarnard to the Present and Beyond, New York. Oxford UniversityPress.

Wittgenstein, L. (1968).Philosophical Investigations(3rd ed.). Oxford:Blackwell.

Wittgenstein, L. (1978).Remarks on the Foundations of Mathematics(Revised ed. ed.). MIT Press.

www.premier-minstre.gouv.fr (2002). Violence routiere: pour une re-sponsabilisation accrue du conducteur. Technical report, premier-ministre.gouv.fr.

Documents

Engaged Philosophy: Essays in Honour of David Braybrooke