andrew levine - the general will, rousseau, marx, communism (1993)

This bold and unabashedly Utopian book advances the thesis thatMarx's notion of communism is a defensible, normative ideal.However, unlike many others who have written in this area Levineapplies the tools and techniques of analytic philosophy to formu-late and defend his radical political program.

The argument proceeds by filtering the ideals and institutions ofMarxism through Rousseau's notion of the "general will." OnceRousseau's ideas are properly understood it is possible to con-struct a community of equals who share some vision of a commongood that can be achieved and maintained through cooperation orcoordination that is at once both voluntary and authoritative. Thebook engages with liberal theory in order to establish its differ-ences from Rousseauean-Marxian political theory.

This provocative book will be of particular interest to politicalphilosophers and political scientists concerned with Marxism, so-cialist theory, and democratic theory.

" . . . unrelentingly intelligent and innovative. . . . What givesthis book its particular quality is the author's consistently sensibledealing with important and controversial questions, his ability toshape difficult claims persuasively to the argument in hand."

Alan Gilbert, University of Denver

The general will

The general willRousseau, Marx, communism

ANDREW LEVINEUNIVERSITY OF WISCONSIN - MADISON

CAMBRIDGEUNIVERSITY PRESS

Published by the Press Syndicate of the University of CambridgeThe Pitt Building, Trumpington Street, Cambridge CB2 IRP

40 West 20th Street, New York, NY 10011-4211, USA10 Stamford Road, Oakleigh, Melbourne 3166, Australia

© Cambridge University Press 1993

First published 1993

Library of Congress Cataloging-in-Publication Data

Levine, Andrew, 1944-The general will : Rousseau, Marx, communism / Andrew Levine.

p. cm.Includes bibliographical references (p. ) and index.

ISBN 0-521-44322-9

1. General will. 2. State, The. 3. Communist state.4. Democracy. 5. Rousseau, Jean-Jacques, 1712-1778 - Contributions

in political science. 6. Marx, Karl, 1818-1883 - Contributions inpolitical science. I. Title.

JC328.2.148 199332o'.oi'i - dc2o 93-2768

CIP

A catalog record for this book is available from the British Library.

ISBN 0-521-44322-9 hardback

Transferred to digital printing 2003

Contents

Preface page ix

introductionThe end of the stateExculpationsAn overview

CHAPTER 1The general will in theory

Are all wills private?Nonwelfarist interestsIndividualsA free-rider problem?Internal relationsThe reality of the general will

CHAPTER 2The "origin" of the private will

The original state of natureBefore the private willFrom self-love to amour propreDe facto statesFreedomEphemeral Hobbesianism

CHAPTER 3Solidarity

A point of departureWhat is solidarity?The practice of solidarityBeyond solidaritySolidarity forever?

I37

15

182O

2 2

253O

3334

36384 1

44505457

6060

63666873

Contents

CHAPTER 4Democracy in the Age of States

DemocracyModeling democracyDemocracy in a world of statesLimits to decentralizationRepresentationRepresentation versus delegationElectionsDemocratization

CHAPTER 5The last state

Democracy and dictatorshipRelative autonomyBeyond liberal democracy?The ideological state apparatusCultural revolution

CHAPTER 6The liberal state and/versus the last state

Romanticism/CareerismEnhancing autonomySelf-realization and the goodHigher and lower pleasuresExperimentalismThe liberal state and the last state

CHAPTER 7Rousseauean Marxism and/versus liberalism

What do liberals want?AutonomyHarmonyCommunity

CHAPTER 8Communism

Beyond scarcity?Beyond self-interest?Liberalism and the general willVirtueFrom virtue to justiceEquality and "bourgeois right"Basic liberties

75768o858890949899

1 0 1

1 0 2

1061 1 2

1151 2 0

123128

133136138141142

*47152159162165

168169173175179182185188

VI

Contents

Justice internalized 190Responsibility 191Justice and the general will 193The exercise of the general will 196Aggregating autonomy 198

CHAPTER 9After Communism, communism? 203

Small-c communism 206Socialism 209Marxism today 210Historical materialism 214Must we bring social democracy back in? 217

Index of names 221

Vll

Preface

In the spring of 1988, when I formed the intention to write thisbook, (big-C) Communism reigned over more than a third of theworld's population, the Cold War still simmered, and the prospectof nuclear annihilation continued to threaten. Then, in a remark-ably short time, Communism was overthrown as an economic sys-tem and mode of governance in Eastern Europe and the formerSoviet Union, the Cold War ended, and the threat of nuclear warreceded dramatically. It is not at all clear what impact these eventsought to have on philosophical reflections on politics or on socialtheory. But, by accelerating a process evident for some time and invarying degrees throughout the world, they have had a definiteeffect on our intellectual culture. By 1992, as I write these words,forms of theorizing linked, however tenuously, with the BolshevikRevolution and its continuations, with Marxism, and even withsocialism have fallen into greater disrepute than at any time inthis century.

To set out to defend (small-c) communism at a time when theGorbachev Revolution was unfolding was already to go against thecurrent. In the present conjuncture, it may seem incomprehensi-ble. I maintain, however, that doing so is reasonable, timely, andeven urgent. The discussions of Rousseauean political philosophyand Marxian communism that ensue will, I hope, partly vindicatethis contention. However, the reader should be aware at the outsetthat what I shall try to defend has at least two interconnected di-mensions that I have purposely confounded. It was, in part, to sig-nal this intention that I have made Rousseau the central figure inthe chapters that follow, even more than Marx himself, and that Ihave made the general will its organizing concept.

At one level, I argue that Marx's notion of communism or,more precisely, his core idea bereft of some of the unrealistic (and

ix

Preface

undesirable) claims he made in its behalf is a defensible normativeideal and that some ideas of Rousseau's about rational coopera-tion - understood in a way that renders them inoffensive to main-stream, liberal "individualism" - help to elaborate Marx's idea.Were I to have framed my positions exclusively in these terms,what I would go on to argue would be politically out of line withthe current consensus but theoretically consistent with it. Then thegeneral will, far from being central to my argument, would be, atmost, an intriguing but inapt concept - from which a few usefulinsights could be teased out and reconstructed. For several years,in the course of an interminable and (for me) always illuminatingdebate, Erik Wright has urged me repeatedly to adopt this frame-work. He has convinced me, despite considerable resistance onmy part, that most of what I have to say about Marxian commu-nism can be accommodated within it.

However, in the end, I have not followed his advice. No doubt,in making the general will central, I do throw an extrapolitical ob-stacle in the way of acceptance of my principal claims, at least forsome readers. Nevertheless, I believe that the advantages that fol-low from according the general will pride of place more than offsetthis loss. It will emerge that mainstream theorizing - which, inRousseauean terms, acknowledges only private volition - is inhos-pitable to Marxian communism. It is so, I shall argue, in conse-quence of its tacit reliance on an underlying metaphor that depictsindividuals, like atoms, as radically independent of each other. Onthe other hand, the general will, as Rousseau conceived it, rests ona different metaphor, according to which, as I shall go on to explain,individuals are related "internally." It may indeed be possible totranslate, more or less "without remainder," from a theoreticalframework based on one of these metaphors to one based on theother. But in political philosophy forms of expression matter - notleast for what they suggest heuristically and, above all, for thekind of politics they promote, for what we might call their "polit-ical tendency." This is why I think it worthwhile to dislodge a met-aphor that tends, as it were, in the wrong direction and to adopt,when appropriate, a more felicitous idiom, resting on a differentmetaphor, in its stead. Thus, at the same time that I offer a brief indefense of Marxian communism, I also argue for the coherenceand possible applicability of Rousseau's notion of the general will.

It is, in the main, readers influenced by contemporary forms ofeconomic modeling who will find the general will problematic.Others may feel that I pay too much attention to defending Rous-

Preface

seau's idea and wonder if what I say is not, in consequence, overlydefensive. I shall go on to argue, however, that individualist chal-lenges to Rousseauean political philosophy and Marxian politicaltheory have merit and that no defender of Marxian communismcan neglect them without peril. In the end, however, the generalwill withstands critical scrutiny and has a crucial if not strictly in-dispensable role to play in the normative theory of communism.

There is a sense in which the debate with mainstream ' 'individ-ualism" that runs throughout the following pages resembles thedispute between atheists and those liberal theologians who effec-tively concede the atheists' case against traditional believers butwho nevertheless represent their views in a theistic idiom. It couldbe said that what is at issue between them is only a form of ex-pression. But both sides would agree that this difference has mo-mentous consequences - for both theory and practice - and thattheir respective positions can only be assessed in light of theseconsequences. Similarly, my quarrel with individualists whowould dispute the coherence or applicability of the general will re-volves ultimately on the consequences of taking up one or anotherform of expression. It is, in short, because I think the case for com-munism compelling that I opt for the idiom I do. The analogy,however, is not perfect - unless one believes, as some "postmod-ern" relativists do, that scientific naturalism is just one form of dis-course among others with no particular purchase on truth. If wereject this account of science, as I think we should, a nontheisticconstrual of what, by hypothesis, both the atheists and the theo-logians believe has a clear mark in its favor. For reasons that I willnot attempt to elaborate, I do not believe that the kinds of norma-tive discourse deployed here can track truth with the authoritythat science does. In this instance, therefore, neither governingmetaphor can be defended on a priori grounds; their heuristic im-plications and political tendencies are decisive.

I conceived the idea for this book in 1988, but it continues a line ofinquiry that I have pursued intermittently since 1965 when, as anundergraduate at Columbia University, I attended a remarkablecourse on Rousseau and Marx taught by Professor Robert PaulWolff. In the time that has elapsed since then, political philosophyhas changed almost as dramatically as the political climate inwhich it operates. Until well into the 1970s, Marxism, for Ameri-can philosophers, was still a barely naturalized European immi-grant. With the rise of "analytical Marxism," that situation has

xi

Preface

altered beyond recognition.1 What follows here attests to thesetransformations. Twenty-five years ago, before A Theory of Justice2

and all that has followed in its wake, the liberal egalitarianism thatfigures so prominently in some of the chapters that follow wasmore implicit than self-conscious. This too has changed pro-foundly; today, liberal egalitarian social philosophy is a flourishingand ongoing concern. Only Rousseau studies exhibit a certain con-tinuity. However, in recent years Rousseau has not much engagedpolitical philosophers, in contrast to scholars interested mainly inthe history of political thought. It should be obvious that I considerthis neglect unwise, and I will be pleased if what follows herehelps in some small way to rectify it. In any case, I would notethat, despite all the changes of the past quarter century, my think-ing on Rousseau has remained remarkably constant. My lodestarsremain Wolff's own investigations of Rousseauean politicalphilosophy3 and what seems to me, in retrospect, a European im-port that has held up remarkably well over the years, Louis Alth-usser's still little-known lectures on The Social Contract.4

I have not hesitated to draw on earlier work of mine on Rous-seau, Marx, and Marxian communism. However, so far as pos-sible, I have relegated references to my previously publishedwritings to footnotes, and then only when assertions made in thetext call for extensive argumentation that would be off-putting torehearse in the present context. I trust that readers unfamiliar withwhat I have written elsewhere will find that what I shall go on todiscuss here stands well enough on its own.

I cannot begin to thank everyone who has helped in one way oranother to shape my views on the themes of this book. In addition

Cf. Erik Olin Wright, Andrew Levine, and Elliott Sober, Reconstructing Marxism(London: Verso, 1992), Chapter 1.John Rawls, A Theory of Justice (Cambridge, Mass.: Harvard UniversityPress, 1971).Unfortunately, very little of this has found its way into print; but see Robert PaulWolff, In Defense of Anarchism (New York: Harper, 1970).Louis Althusser, "Sur le Contrat Social (Les Decalages)," in Cahiers pour {'analyse,no. 8, L'Impensee de Jean-Jacques Rousseau (Paris, n.d.), translated as "Rousseau:The Social Contract," in Louis Althusser, Politics and History (London: New LeftBooks, 1972). In my first sustained investigation of Rousseau's political philos-ophy, I attempted, in effect, to deploy Althusserian positions on Rousseau tosome of the theoretical problems identified by Wolff; see The Politics of Autonomy:A Kantian Reading of Rousseau's "Social Contract" (Amherst: University of Massa-chusetts Press, 1976).

Xll

Preface

to countless conversations, Erik Wright has read and commentedin heroic detail on drafts of this manuscript. Daniel Hausman hasalso provided extensive and insightful comments at various stagesof its composition. Others who have provided comments on par-ticular chapters or on talks that have eventually become chaptersinclude Richard Arneson, G. A. Cohen, Gerald Doppelt, JaneMansbridge, Debra Satz, Elliott Sober, Robert Ware, and DanielWikler. I must note that few if any of these friends and colleagueswould want to sign on to all or even very much of what I shall goon to maintain. A version of Chapter 1 was read at a conference onBiology, Behaviour, and Society at the University of Alberta, Ed-monton, in May 1991; an article based on that talk will appear in aspecial supplement of the Canadian Journal of Philosophy. What fol-lows also draws on talks I have given at the University of Wiscon-sin - Madison, the University of Massachusetts - Amherst, theUniversity of California - San Diego, the University of British Co-lumbia, the University of Turin, Iowa State University, and theUniversity of Chicago, and at the 1989 International Meeting of theConference for the Study of Political Thought in Williamstown,Massachusetts. I have benefited greatly from discussions at theseevents. I would also like to thank my students at the University ofWisconsin - Madison, some of whom helped enormously in thegestation of my thinking on Rousseau and Marx. Finally, specialthanks are due to Terence Moore at Cambridge University Pressfor his assistance in seeing this book through to publication.

Xlll

Introduction

For nearly four centuries, a principal concern of political philo-sophers has been to justify the coordination of the activities ofradically independent individuals by the state, a complex of insti-tutions that monopolize the right to compel compliance by the useof threat of force. In what follows, I will challenge aspects of thisproject - the better to accommodate a way of thinking about soci-ety and politics that, following Marx's lead, envisions a social or-der, communism, constituted by people who have substantiallyovercome a need for state-compelled coordination. My aim is todefend a nonutopian but still Marxian version of this idea.

In an intellectual culture as prone to historicism as ours now is,it should be widely acknowledged that such major historical trans-formations as the expansion of (individualizing) market relationsand the emergence of the nation-state have profoundly affectedphilosophical reflections on politics and society. Nevertheless, theextent to which prevailing understandings of the individual andthe state depend upon transitory real-world phenomena is com-monly overlooked. For many purposes, the coexistence of a wide-spread awareness of the historicity of these concepts with theirahistorical treatment in contemporary political theory has been be-nign. But in order to defend Marxian communism, it is especiallyimportant to bear in mind how marked political philosophy todayis by the historical specificity of its fundamental concepts.

The individualistic aspect of modern political philosophy is atvariance with earlier understandings of politics and also, to some de-gree, with political currents, whether of the Left or Right, that priv-ilege communal values and perspectives. A case in point is Marxismitself. Since its inception, Marxism has developed partly in opposi-tion to, partly in isolation from mainstream, "bourgeois" social sci-ence and philosophy. According to each side's self-representations

The general will

and in their polemics against each other, bourgeois theorists en-dorse and Marxists oppose "individualism/' However, recentwork in the social sciences and philosophy and new styles ofMarxist theorizing have transformed this shared understanding ir-reversibly. Partly in consequence, many of the methodological andsubstantive positions that were once thought to distinguish Marx-ism from mainstream views have come to seem less salient andmany alleged differences now appear illusory.1 Indeed, so many ofthe old certainties have given way that it has become apt to askwhat, if anything, remains that is distinctively Marxist. In thechapters that follow I investigate this question - with a focus onnormative concerns. It will become evident that Marxist positionsin political theory and philosophical accounts of the state are notnearly so orthogonal as they formerly appeared. But it will alsoemerge that there remain fundamental differences with far-reaching implications that distinguish Marxist political theoryfrom mainstream positions.

I shall assume as a working hypothesis that an idea introducedby Rousseau, the general will, bears importantly on the possibilityand desirability of Marx's notion of statelessness. Rousseau's idea,very generally, is that in some contexts individuals can and shouldcoordinate their activities by seeking to advance their interests asintegral members of the collective entities they freely constitute. Inparticular, Rousseau maintained that it is reasonable for individu-als in a "state of nature" to subordinate their "private wills,"which aim at "private interests," to the "general will," which aimsat the "general interest," the interest of "the whole community" -in order to become the autonomous agents they potentially are.

My intent in representing this idea in Rousseauean terms ispartly reflationary, partly deflationary. Rousseau is a central figurein today's "canon" of political philosophers, yet his notion of thegeneral will, surely his principal contribution to Western politicalthought, plays virtually no role in contemporary political philos-ophy. This situation is not nearly so odd as may at first appear.Rousseau's account of the general will invites misunderstandings,and much of what he used the concept to express can be repre-sented in more familiar terms. Nevertheless, it will become evi-dent that there are good reasons for taking Rousseau's account of

Facets of this thesis are defended, implicitly and explicitly, in Erik OlinWright, Andrew Levine, and Elliott Sober, Reconstructing Marxism (London:Verso, 1992).

Introduction

the general will seriously and for attempting to retrieve the con-ceptual insights and empirical speculations that constitute its "ra-tional kernel." It is therefore worth attempting to reflate thiscurrently defunct idea. However, at the same time, if the generalwill is to be put to use in defense of statelessness, it is crucial thatthe misunderstandings it elicits be dissipated. Doing so will de-flate the apparent novelty of Rousseau's idea and recast its func-tion in political arguments. But it will not render the idea otiose.What I shall say about Rousseau is relevant to the defense of com-munism, and what I shall say in behalf of communism bears onthe case I shall make for the pertinence of Rousseau's idea to philo-sophical and political concerns. Thus I shall enlist Rousseau in de-fense of communism and communism in defense of the cogencyand timeliness of the general will. The principal justification forbringing the general will to center stage in the chapters that followis to facilitate these mutually reinforcing objectives.

In addition to Rousseau and Marx there will be a third point ofreference in several of the chapters that follow: liberalism. Not longago, Marxists were eager to deride liberalism's defects. However,in one of the more ironical turns of recent political and intellectualhistory, many progressive philosophers today find in liberalism avehicle for waging the struggle for equality associated historicallywith Marxian socialism. I believe that in the present conjuncture itis important not to forsake Marxian insights but also to acknowl-edge Marxism's flaws in contrast to aspects of liberal theory andpractice. My contention - modest in comparison with what manyon the left once believed, outrageous in light of today's receivedwisdom - is just that communism, understood generally as Marxintended, implements a sounder normative vision than anythingliberalism can contemplate and that general will coordination,an idea liberalism resists, has an important role to play in Marx-ian communism.

THE END OF THE STATE

Rousseau's aim in The Social Contract can be succinctly put: "[T]ak-ing men as they are and laws as they might be," he sets out todemonstrate the possibility of legitimate political authority andobligation.2 In his terminology, supreme authority over a given

2 The Social Contract, Book I, Introductory Note. Rousseau's Social Contract is di-vided into relatively brief chapters and is available in English in many editions,

The general will

territory or population is "sovereignty"; what the sovereign rulesis a "state."3 Rousseau then contends that "sovereignty is nothingbut the exercise of the general will"4 or, equivalently, that a state isestablished when individuals place themselves "under the su-preme direction of the general will."5 Since the general will is thetrue will of each of the state's members, the sovereign and its sub-jects are one.6 Rousseau succeeds in demonstrating, at least to hisown satisfaction, that the general will is real and therefore that po-litical authority and obligation are possible - in short, that the ideaof a just state is coherent and applicable under human conditions.But it is one thing for a just state to be possible and something elsefor it to be a feasible political objective. Rousseau was deeply pes-simistic about the latter prospect. The world is divided into defacto states; authority is everywhere asserted and acknowledged.A coercive apparatus superintends and reinforces a civil societybased on political and material inequalities. In consequence, hu-man mentalities have become corrupted, perhaps irreversibly, andthe de jure state, where legitimate authority and obligation exist,may never be realizable anywhere.

Thus, in Rousseau's political thought, the exercise of the generalwill is more nearly a regulative idea than a plausible political as-piration. Sovereignty is possible in principle but almost certainlynot in real history. Already in the Discourse on the Sciences and Arts,The First Discourse, Rousseau faulted "progress" for corrupting hu-man nature and burdening the human race with institutions det-rimental to its well-being. The Second Discourse extended thisdiagnosis by attributing the detrimental effects of progress to theemergence of inequalities within human communities. In The So-cial Contract, the toll exacted by these developments is depicted inan even more devastating light. Identifying essential humanitywith freedom of the will (autonomy) and arguing that the exerciseof the general will is a necessary condition for actualizing this es-

none of which is standard. For the convenience of the reader, therefore, I shall notcite a particular edition in referring to it but will instead indicate the chapter inquestion. All translations from The Social Contract are my own and are based on thetext edited by Bernard Gagnebin and Marcel Raymond, published in Jean-JacquesRousseau, Oeuvres completes, vol. 3 (Paris: Gallimard, Bibliotheque de la Pleiade,1964). The English translation that I have consulted most frequently is by DonaldCress. It may be found in Jean-Jacques Rousseau, The Basic Political Writings (Indi-anapolis and Cambridge, Mass.: Hackett, 1987).3 Ibid., Chapter 6. 4 Ibid., Book II, Chapter 1. 5 Ibid., Book I, Chapter 6.6 Ibid.

Introduction

sential trait, he contends that obstacles in the way of realizing ajust state are mortal dangers to humanity itself.7

Several years ago, in The End of the State, I argued that, in con-ditions of material abundance and under socialist property rela-tions, Rousseau's regulative idea or some close approximationactually is historically feasible.8 I maintained that communism inMarx's sense is Rousseauean sovereignty attained or rather over-come, since what Rousseau's social contract establishes is not, inthe end, a state at all but, as Kant would have it, a "republic ofends." To be sure, without considerable uncoerced compliance, node facto state could long endure. But, ultimately, states coordinateindividuals' behaviors "externally" - through force. On the otherhand, in republics of ends, behaviors are "internally" coordinatedby what Kant called a harmony of rational wills.9 The use or threatof force may sometimes be necessary even in a republic of ends,just as uncoerced compliance is indispensable in states. "The ad-ministration of things" that, according to Engels (and, before him,Saint-Simon), replaces "the governance of men" after the demiseof class society may still need to use force to overcome collective"weaknesses of will." Thus public coercive force is unlikely to dis-appear entirely under communism. But the use of force in genu-inely communist societies would be, as it were, an administrativeimperative, not an exercise of state power. I argued that the differ-ence is not merely definitional. The idea that the state under com-munism would "wither away" is distinct from other claims that

7 Ibid., Chapters 4 and 6. See also Andrew Levine, The Politics of Autonomy: AKantian Reading of Rousseau's "Social Contract" (Amherst: University of Mas-sachusetts Press, 1976), passim.

8 Andrew Levine, The End of the State (London: Verso, 1987).9 There is reason to think that what Kant intended by a "harmony of rational

wills" is, at least in part, the ideal of rational cooperation implicit in Rousseau'saccount of legitimate political association. A historical and conceptual connec-tion between Rousseau and Kant has long been acknowledged. Hegel discernedit in The Phenomenology of Mind in the section entitled "The Moral View of theWorld." In this century, the idea that an important motivation for Kant's workin moral philosophy was precisely to provide foundations for Rousseaueanpolitical philosophy has been pressed most forcefully by Ernst Cassirer; see es-pecially, The Question of Jean-Jacques Rousseau, trans, and ed. Peter Gay (Bloom-ington: Indiana University Press, 1954), and Rousseau, Kant, Goethe, trans. JamesGutmann, P. O. Kristeller, and J. H. Randall (Princeton: Princeton UniversityPress, 1945). I argue in support of Cassirer's thesis by exhibiting conceptual af-finities joining Rousseau's political philosophy with Kantian moral philosophyin The Politics of Autonomy.

The general will

can be made for Marxian communism - and less susceptible todisputation than many of them.

These claims depend on the plausibility of the core theses ofMarx's theory of history, historical materialism. In the historicalmaterialist scheme, communism is the end of a determinate pro-cess of historical change. In The End of the State, I argued that adefensible historical materialism implies only that the materialconditions for a communist economic structure can come intobeing.10 I also argued that there are reasons for thinking commu-nism more than just materially possible. In addition to "discover-ing" the model of rational cooperation communism supposes,Rousseau, despite himself, provided support for the conclusionthat communism actually is feasible. The aspects of Rousseau's po-litical philosophy that encourage faith in the transformative effectsof political institutions - above all, in the institutions that imple-ment democratic deliberation and collective choice - support theidea that the motivations of citizens in genuinely radical democra-cies can change; that individuals can become less self-interested,more solidary - in short, more inclined to support general overparticular interests. I maintained that a similar faith in democracyis evident in Marxist political theory or, at least, in an importantstrain of it, and that this conviction plays a pivotal role in Marx'scase for communism. Thus the idea that the state can and shouldwither away is underwritten by venerable and persuasive, thoughhardly incontrovertible, arguments.

In an era of diminished expectations, when ideas of this sort areeverywhere dismissed as Utopian, this conclusion will seemanachronistic, if not quaint. Nevertheless, "the withering away ofthe state" merits serious consideration. In Chapters 8 and 9, in re-flecting on communism and on the significance of continuing touphold the idea at a time when leftist politics has devolved into amotley of good causes devoid of any unifying vision or aim, I willtry to show why this conclusion is of more than passing theoreticalinterest. I concede, however, that a study in political philosophycan only issue in tentative conclusions about these matters. Theprospects for organizing political communities as republics of endsdepend on speculations about human nature and the likely out-comes of institutional arrangements that theoretical considerations

See Levine, The End of the State, Chapter 5. See also Wright, Levine, and Sober,Reconstructing Marxism, Chapters 3 and 5. The claims that follow in this para-graph are supported in The End of the State, Chapters 6-8.

Introduction

can never finally resolve. It remains for real-world political devel-opments themselves to establish or refute the claim that commu-nism, in something like the sense Rousseau anticipated and Marxintended, is a realistic and desirable political objective. However, ifcommunism is to be brought back into political theory, as I willcontend it ought to be, there is no choice but to proceed in the faceof vulnerability to empirical confutation.

It is now clear that the most sustained attempt to date to movehumanity toward communism, the Bolshevik Revolution, ended infailure. Partly in consequence of this historic defeat, political initi-atives with similar ambitions have effectively vanished from thepopular imagination. However, capitalism and the state system re-main in place. Can the aspiration to replace them with a commu-nist order remain suppressed indefinitely? If, as I shall argue,these core features of our civilization are impediments to auton-omy and self-realization, and if communism is a feasible and de-sirable alternative to them, this question answers itself. I wouldtherefore venture that, before long, the communist project will re-sume - not just in economically backward and politically under-developed countries but in the vastly more propitious conditionsMarx himself envisioned. It will be of some help in this endeavorif, when the time again comes, philosophy is not caught unawares.

EXCULPATIONS

I remarked at the outset that contemporary political philosophersgenerally fail to take due account of the historical specificity of cur-rently dominant notions of the individual and the state. The caseI will go on to construct in behalf of Marxian communism will infact appeal to different understandings of these and other politicalconcepts. Nevertheless, there is a sense in which what followshere is itself ahistorical in the manner of contemporary politicalphilosophy. In this section, I shall advance a brief comment on thisfeature of the ensuring discussion. Then, still in an exculpatoryspirit, I shall excuse my neglect of two important issues - themetaphysics of volition and the "post-Marxist" challenge to com-munism - that some readers might expect to see addressed in abook on the general will that defends key Marxian positions.

Philosophy outside history

Rousseau, Marx, Hobbes, Locke, Mill, and others figure promi-nently throughout the discussions that follow. However, this book

The general will

is not a study in the history of ideas. It is intended instead as acontribution to contemporary political theory. From a historicalpoint of view, its use of past philosophy therefore requires com-ment and justification.

It is commonplace for philosophers to approach the history ofphilosophy by focusing only on major figures, abstracted fromtheir historical contexts and inserted into an imaginary moment inwhich they somehow join in dialogue with one another and withphilosophers today. In this way, most of philosophy's past is ex-cluded from the history of philosophy: Only a handful of philos-ophers - indeed, only a handful of these philosophers' writings -are accorded legitimacy; everything else is ignored. Curiously too,the words of these co-investigators are then in varying degreesvenerated - like oracles from whose writings Truths can be teasedout or, more precisely, read in. To account for this unlikely config-uration of attitudes, it is well to recall the peculiar position of phi-losophy - as a traditional, humanistic discipline but also as acomponent of an emerging scientific culture. Many philosopherstreat the history of philosophy - or at least the fragment of it theyrecognize - in much the way that Renaissance scholars and theirsuccessors treated the writings they recovered from the Greekand Latin traditions. At the same time, like scientists, most phi-losophers believe in the growth of understanding and know-ledge. Practicing scientists characteristically have little use forthe history of science. In philosophy, where progress has provenmore elusive, the history of philosophy, reduced to some majorworks of a few master thinkers, is a more timely source of insight.From this point of view, the philosophical canon is not so muchsomething to be studied for its own sake as it is a resource amongothers. With progress sufficiently slow, past philosophy remainscontemporary for as long as it continues to be relevant to issuesstill in dispute.

Humanistic and scientific attitudes toward the history of philos-ophy hardly cohere and, together or separately, they offend a gen-uinely historical sensibility. From a historical point of view, ideas,like everything else, should be understood in context. It is this re-quirement that motivates some contemporary investigators to readpast philosophers, "great" and minor alike, against the back-ground of their time and place and to focus on the ways in whichtheir positions are addressed to their own contemporaries, ratherthan to philosophical giants of different times and places or to is-sues of contemporary concern.

8

Introduction

As noted, my treatment of historical figures, Rousseau espe-cially, is more in line with contemporary philosophical practicethan with "contextualist" historical research. Thus I identify Rous-seau's "private will" with the wills of individuals in a Hobbesianstate of nature, and I represent Rousseau as if he were engaged ina protracted dialogue with contemporary understandings of Hob-besian moral and political philosophy. This picture is plainly atodds with a contextualist account of the (major) texts I discuss.Hobbesian philosophy was not in fact a central concern of Rous-seau's, especially in the form I shall present it.11 Indeed, from amore historical point of view, a discussion of the "general will"would have to engage theological disputes about sin and the na-ture of grace as much as Hobbes's attempt to extricate hypotheticalindividuals from a devastating "war of all against all." It would fo-cus on theological controversies aroused by the Jansenist revival ofthe doctrine of predestination - in particular, on attempts to join abelief in God's "general will" that humanity be saved with the"particular" salvation of the elect.12 But French theological and po-litical theory in the century before Rousseau will hardly intrude onthe discussions that follow. Along with Rousseau, Hobbes, Locke,Mill, and, above all, Marx are the figures that matter for my pur-pose, not Pascal, Malebranche, Bossuet, Fenelon, Bayle, Montes-quieu, or Diderot.

In short, what follows is not intended as a historical study. It isinstead an investigation of the historical possibility of the generalwill, and of the forms and limits of its desirability. To this end, theodd and even contradictory attitude philosophers assume towardthe history of their subject actually is appropriate. I would readilyacknowledge that humanist postures are of dubious value in themodern world, especially when they encourage the veneration ofauthors and texts. However, I shall maintain that it is reasonable atthe present time to continue to identify with the tradition Marx in-augurated, and especially with that strain of Marxism that bears a

The Hobbes who figures in the following pages is very much a creature of con-temporary political philosophy. For a trenchant attack on what she calls "thestandard philosophical interpretation" of Hobbes, see S. A. Lloyd, Ideals as In-terests in Hobbes's "Leviathan": The Power of Mind over Matter (Cambridge: Cam-bridge University Press, 1992).Cf. Judith N. Shklar, "General Will," in Philip P. Wiener (ed.), Dictionary of theHistory of Ideas (New York: Charles Scribner's Sons, 1973), vol. 2, pp. 275 ff.;and Patrick Riley, The General Will before Rousseau: The Transformation of the Divineinto the Civic (Princeton: Princeton University Press, 1986).

The general will

conceptual affinity with Rousseauean political philosophy.13 In ad-dition and more importantly, Rousseau and Marx and the otherfigures I discuss still do speak to us instructively on the issues incontention here. If it is conceded that a philosophical investigationof the general will is worth undertaking, then even a committedcontextualist should concede that there is no harm in using his-torical figures ahistorically for this purpose. This is not to say thathistorical sensibilities can justifiably be set aside. Misrepresenta-tions, deliberate or not, are unlikely to be helpful or even benign.It is therefore advisable to be scrupulous in respecting the resultsof historical research. But within this constraint, so long as there isno pretense of advancing intellectual history, the noncontextualuse of the writings of the "immortals" is surely unobjectionableand even wise.

In any case, I have tried to treat the writings I discuss as morethan suggestive inkblots in a philosophical Rorschach test. What Iextract from them may not be quite what their authors intended orwhat contemporary readers saw. But the texts still constrain whatthere is to see. So understood, these writings advance theses andcontain arguments that cast light on the topics they address. This,I think, is what explains the exclusion of most of philosophy's pastfrom the history of philosophy as philosophers typically construethe subject. Some philosophical writings remain sources of insighteven in circumstances remote from their conditions of origin. Oth-ers, the vast majority, are of interest today primarily as artifacts oftheir time and place.14 Thus a properly historical history of phi-losophy frequently will diverge from a philosophical use of philos-ophy's past. Again, it is well for the sake of good work in bothdomains that historical and philosophical investigations informone another as much as possible; and it is crucial that the two notbe at odds. But good intellectual history is not always good phi-losophy or vice versa.

It is frequently the case, for both historians and philosophers,that God, as the saying goes, is in the details. But sometimes, es-pecially in philosophical uses of the history of philosophy, the in-sights that underlie particular theses and arguments are moreimportant than the formulations through which they are ex-

13 See Chapter 9 herein.14 Cf. Ian Hacking, "Five Parables," in Richard Rorty, J. B. Schneewind, and

Quentin Skinner (eds.), Philosophy in History: Essays on the Historiography of Phi-losophy (Cambridge: Cambridge University Press, 1984).

1 0

Introduction

pressed. For the present purpose, it will often be useful to distin-guish the core ideas of the writers I discuss from their actual po-sitions, even supposing that they can be uncontroversiallyascertained. I register this observation not so much to excuse whatmight otherwise be considered a cavalier disregard of the facts ofthe matter but to signal the reader that, throughout much of whatfollows, it is Hobbesian, Rousseauean, Millean, and Marxian po-sitions that are at issue, rather than the versions of these positionsactually developed by Hobbes, Rousseau, Mill, or Marx.

An ahistorical use of the history of philosophy is, of course, atodds with the spirit and letter of Marx's writings. But it is defi-nitely in accord with Rousseau's self-conscious practice. "Let ustherefore begin by putting aside all the facts," Rousseau declaredat the beginning of The Second Discourse, "for they have no bearingon the question." To be sure, Rousseau exaggerated his disregardof the facts. But, in truth, he was not a speculative anthropologistor historian. He sought not to advance understanding of what ac-tually happened in the past but to develop a philosophical accountof inequality and its consequences for moral, social, and politicallife. In view of what is now known, many of the ostensibly histor-ical claims made in The Second Discourse are false. But from his ownvantage point, the story Rousseau told was as plausible as anyother. It respected accepted views about the first human beingswithout worrying over their accuracy. A similar attitude will be as-sumed here with respect to Rousseau himself and the other phi-losophers, including Marx, whose ideas I shall deploy. To advancereflection on the general will - indeed, to insert the idea back intoits real historical context - I consider the general will ahistorically,much as Rousseau removed his reflections on the development ofinequality from the real history of the human race. Like Rousseau,I will doubtless fail to add anything to what is already known ofthe historical facts. But, again like Rousseau, I shall try to respectwhat the best histories tell us and to take advantage of the under-standings gained through historical research, whenever they canbe put to philosophical use.

The will

The distinction Rousseau drew between the private and the gen-eral will is central to his political philosophy and to his politics -conceived as an eternal struggle, waged in and over institutions,

1 1

The general will

to bring about the supremacy of the general will within each citi-zen of a just state.15 But, in Rousseau's writings, the concept ofthe will is taken over from received ways of thinking withoutmuch scrutiny or elaboration. In this respect, I shall follow Rous-seau's lead.

The will figures in many everyday reflections and philosophicalexplanations, but especially in attempts to conceive the differencebetween actions and other events, including mere behaviors. Someevents are also actions, executed by agents who are in pertinentsenses responsible for what they do. Agency is then held to involvevolition essentially. How agency is possible, what its forms andlimits might be, and what its implications are for a host of meta-physical, moral, and legal concerns are matters Rousseau ignores.Perhaps, as already suggested, his political philosophy implicitlyassumes an account of volition, supplied retrospectively by Kant.But Rousseau himself had little if any interest in these issues, andneither did Marx. What follows here will suppose, as they did,that human agency is a fact and that the standard view of it, ac-cording to which volition plays an essential role, is sound. I trustthat these assumptions are uncontroversial. However, should theyeventually prove faulty, the case for general will coordination willlikely not be impugned. Where the general will is in question, verylittle weight is borne by whatever the term "will" suggests to phi-losophers with more directly metaphysical concerns.

There are perhaps accounts of volition that would be incompat-ible with Rousseau's idea. I would venture, though, that generalwill coordination is compatible with any plausible metaphysicalview. But I shall not attempt to defend this claim here. Many po-sitions in moral, social, and political philosophy rest on notionsof freedom, agency, and the self about which sensible people canreasonably disagree. However, in doing moral, social, or politicalphilosophy it is usually unnecessary to focus on underlying meta-physical assumptions or even to make them explicit. There is noreason to expect otherwise in this case.

For the present purpose, it is enough to hold that wills are dis-tinguished by their objects - by the interests toward which theyaim. The private will aims at private or particular interests; thegeneral will aims at general interests. I shall try to provide an ac-count of these interests, and of their role in normative social andpolitical theory. This task is sufficiently daunting - and suffi-

15 See Levine, The Politics of Autonomy, Chapters 4 and 5.

12

Introduction

ciently independent of metaphysical worries about volition - to betaken up independently. My subject is "the general will" becauseRousseau introduced the idea into political philosophy in theseterms, not because I have anything metaphysically contentious tosay about the will.

Post-Marxism

Neither will I directly engage challenges to the positions advancedhere by "post-Marxists," including some feminists, who maintainthat power relations not grounded in class divisions obstruct au-tonomy and self-realization in the way that Marxists think class di-visions do.16 I shall, however, venture two brief and speculativecomments intended partly to excuse this absence and partly to in-dicate how objections of this sort might be met.

My first comment depends upon a conclusion that will emergefrom the chapters that follow. It will become evident that whatmakes general will coordination materially possible at the level of"whole communities" is not the elimination of interindividualconflicts per se but the removal of systemic factors that set individ-uals at odds, conflicts inherent in social structures themselves. ForMarx, class divisions are systemic in this sense. Although he neverdirectly addressed the issue, it is plain that, in his view, nothingelse is. I will go on to argue that, if Marx was right, the eliminationof class divisions is necessary (though not sufficient) for a societyto be directed, when appropriate, by a general will.

If there are other obstacles to "the exercise of the general will"that are systemic in character, whether in addition to the one Marxidentified or instead of it, then they too would have to be removedfor general will coordination to become materially possible. Couldthis "correction" be accommodated without fundamentally chang-ing the case for Marxian communism? It is not possible to answerthis question in general; we would need to assess particular

16 Despite the risk of concocting a "straw man," I have characterized "post-Marxism" with a broad brush in order to encompass a wide range of views. Onthis characterization, the designation "post-Marxist" would aptly describe,among others, Richard Wolfe and Steven Resnick, Knowledge and Class (Chi-cago: University of Chicago Press, 1988); Barry Hindess and Paul Q. Hirst,Marx's "Capital" and Capitalism Today (London: Routledge and Kegan Paul,1977); Heidi Hartman, "The Unhappy Marriage of Marxism and Feminism," inLydia Sargent (ed.), Women and Revolution (Boston: South End Press, 1981); andAnthony Giddens, A Contemporary Critique of Historical Materialism (Berkeley:University of California Press, 1981).

13

The general will

claims. But it is plain, even in the absence of an exhaustive inves-tigation, that there is no obvious reason why basic changes in theaccount that follows would be necessary. Post-Marxists only addto or alter the one-item list of systemic impediments to autonomyand self-realization that Marxists proffer. However, we shall findthat what matters for communism is more the existence of such alist than its content. I would therefore suggest, tentatively but inthe absence of a good reason for thinking otherwise, that the caseto be developed here in communism's behalf would still generallyhold even if a post-Marxist challenge to Marxism could somehowbe sustained. Perhaps in those circumstances it would not quite beMarxian communism that would realize Rousseau's ideal. Butsomething very much like it would.

How likely is it, though, that there are systemic obstacles in theway of general will coordination other than class divisions? Need-less to say, this too is an empirical question that cannot be settledby philosophical speculation. But it is surely reasonable to de-mand, before we admit the existence of systemic oppositions, thatthere be something that underwrites claims for their systematicity.Liberals effectively deny the systematicity of any interindividualoppositions because they do not acknowledge a theoretical basisfor privileging some social divisions over others or for accordingspecial weight to any of the particular oppressions they generate.On the other hand, the Marxist conviction that class divisions aresystemic and that nothing else is follows from Marx's theory of his-tory. It is because of the role class divisions play in the historicalmaterialist account of history's structure and direction that the endof class oppression is indispensable for the realization of the nor-mative vision communism embodies.17

In contrast to liberals who deny that any oppositions are sys-temic and to Marxists who hold that class divisions alone are, post-Marxists assert a plurality of systemic oppositions. But is there anywarrant for this claim? Again, this question cannot be answered inthe absence of a careful study of particular cases. However, it is fairto lay the burden of proof on the post-Marxists. For if there is a"functional equivalent" to historical materialism, it is not at allclear what it might be. It appears, in fact, that post-Marxists infer

17 Thus it will be argued in Chapter 7 that Marxists and liberals disagree moreabout social theory than about normative concerns. On the connection be-tween Marxism as an explanatory project and its emancipatory objectives, andon some fissures in the assumed connection between these aspects of Marxisttheory, see Wright, Levine, and Sober, Reconstructing Marxism, Chapter 8.

14

Introduction

the existence of systemic oppressions from the very beliefs thatlead liberals to deny their existence altogether. If this observation iscorrect, it is plainly the liberals, not the post-Marxists, who drawthe right conclusion.

My second comment, then, is that post-Marxists typically advertto the kinds of social theories that liberals do and then deploythem in a confused and misleading way. If I am right, instead ofaltering or adding to the Marxists' one-item list of systemic oppres-sions, they should, like the liberals, deny that there is any list atall. But then this challenge to Marxian communism would effec-tively devolve into the liberal challenge. Of course, post-Marxistsmight still differ from liberals in other respects. They might evenbe justified in concentrating on particular social divisions and theoppressions to which they give rise. But no social or historical the-ory would ground this focus. It would be determined instead bypolitical agendas or extrapolitical interests or perhaps by the em-pirical salience of one or another opposition in particular in-stances. These factors figure too in explaining Marxists' recourseto class analysis. But Marxists also have a theory of history thatwarrants their privileging class divisions. For what will follow,this reason makes all the difference.

To be sure, these very brief and speculative assertions are un-likely, separately or together, to allay the qualms of readers in-clined to take post-Marxist positions seriously. I offer them hereonly to account for my own neglect of this currently influentialstrain of theorizing and to alert the reader to the "old-fashioned"but, I think, more than ever timely nature of the chapters thatfollow.

AN OVERVIEW

This book consists of essays on a variety of topics that bear on thegeneral will and communism - moving from a focus on Rousseauto a focus on Marx.

As remarked, general will coordination is difficult to reconcilewith current ways of thinking about individual and group inter-ests. Therefore some readers are likely to find the idea problematicif not incoherent. Chapter 1 addresses this issue directly by distin-guishing general from private will coordination in a way that ac-commodates many contemporary misgivings but that neverthelessaccords with the intuition that led Rousseau to draw the distinc-tion initially. Chapter 2, an essay on Rousseau's Second Discourse,

15

The general will

argues indirectly for the cogency of general will coordination byunderscoring some peculiar aspects of its putatively unproblem-atic rival, the private will. Chapter 3, "Solidarity," probes the lim-its of general will coordination by analyzing different forms ofsocial interaction that resemble it in crucial respects. A focus onsolidarity effectively historicizes Rousseau's treatment of generalwill coordination, a consequence of some moment for the norma-tive assessment of this model of rational cooperation. In Chapter 4,the state, the distinctively modern political form described at theend of The Second Discourse and throughout modern political phi-losophy, is faulted in light of the expressly political value of democ-racy. Since this political form instantiates private will coordinationat the level of "the whole community," a demonstration of itsshortcomings with respect to a defensible and widely acknowl-edged standard provides a further reason for reflecting on socialorders that do not assume the ineluctability of private volition.Then, in Chapter 5, an essay on the state that superintends thetransition to communist statelessness in Marxist political theory, Ishow how Marxism's "last state" fares better with respect to dem-ocratic values than do actually existing states.

To investigate these issues is to put the nature and limits of lib-eral and Marxist moral philosophy in question and therefore to re-flect on their affinities and differences. I have already suggestedthat received understandings of the differences separating Marx-ists from liberals cannot be sustained. Nevertheless, there are re-spects in which Marxist political theory is non- or extraliberal andtherefore at odds with what has become the dominant political ide-ology. Much of what follows is about Marxism, liberalism servingmainly as a point of reference. But it is not always the case thatMarxist positions are preferable to liberal ones. Chapter 6, an essayon the state liberals endorse, identifies elements of liberal theorythat appear indispensable in any defensible normative account ofsocial and political arrangements. I argue there that, despite a sur-feit of opinion to the contrary, the form of the state Marxist polit-ical theory envisions need not offend defensible liberal values.Chapter 7 then investigates affinities and differences between lib-eralism and Marxism - focusing on their respective understand-ings of autonomy, human flourishing, and equality - and theirdiffering accounts of the role(s) political institutions can rightfullyplay in realizing these objectives.

After identifying some real differences distinguishing liberalismfrom Marxism, it will be appropriate finally to investigate commu-

16

Introduction

nism as a normative ideal. Chapter 8 presents and defends crucialelements of Marx's idea, focusing directly on the role of generalwill coordination and on its limits. What I defend there will seempale in comparison with traditional Marxist accounts. To some de-gree, traditionally unacknowledged affinities between Marxismand liberalism explain this divergence from standard expectations.Indeed, throughout what follows, a number of the positions I takeconcede a great deal to liberal views widely considered incompat-ible with Marxism. In Chapter 9, an essay on "communism afterCommunism," after the self-destruction of the Communist systemin the Soviet Union and Eastern Europe, I offer some reflections onthis situation.

At a time when liberalism is virtually hegemonic, Marxism in cri-sis, and communism in disrepute even among those who continueto call themselves Marxists, I do not expect the case I shall presentto alleviate many doubts about communism as a feasible projectand worthwhile ideal. I will be content if I can convince skepticalreaders that this ostensibly defunct idea, an inspiration for somany generations of socialist militants, is worth taking seriouslyagain. I do so with the understanding that political conjuncturesand intellectual climates change, sometimes with astonishing ra-pidity, and that what today appears anachronistic may soon seemtimely again. What follows is an essay in political theory, moti-vated by the conviction that it is important, perhaps now morethan at any time in the recent history of the Left, to defend hope ina communist future.

Chapter 1

The general will in theory

The Social Contract contrasts general wills with private or particularwills. A general will is said to be the will of "the whole commu-nity"; private wills are ascribed paradigmatically to individuals ina state of nature. Genuine authority exists only when laws aremade by individuals expressing their opinions about what the gen-eral will is, rather than their private wills as registered in theirpreferences. Collective choices that aggregate preferences repre-sent "the will of all," not the general will. The distinction is ex-haustive: Any will that is not general is private. Thus the wills ofgroups less than the whole community and even the will of all areassimilated conceptually to the category defined by the primordialcondition that motivates the establishment of de jure states. Prac-tical politics, in Rousseau's view, is a struggle for supremacy of thegeneral over the private will, a struggle waged at the level of"opinion" for the character and judgment of each individual.

However, despite the centrality of this distinction for his politicaltheory, Rousseau's account of the difference between general andprivate willing is poorly elaborated and easily misconstrued. Likeother eighteenth-century contractarian formulations, Rousseau'sappears to suggest that the state of nature and the political com-munity that succeeds it designate real historical conditions. Thisconstrual implies the actuality of general will coordination in ex-isting political communities - obscuring the distinction, implicitin The Social Contract, between de jure and de facto authority andconveying a misleading sense of the obstacles in the way of generalwill coordination for modern states. More seriously, Rousseau'sformulations support the idea that the general will is somehowborne by a supraindividual entity, the whole community, ratherthan by the individuals who constitute that social entity. It needhardly be said that neither understanding is correct. The contrac-

18


tarian story, even in its eighteenth-century versions, is onlymetaphorically temporal: Legitimacy is established not by anevent occurring at a determinate time or place but by a "contract,"ascribed by way of a rational reconstruction, that undergirdsparticular institutional arrangements. And, whatever Rousseau'slanguage may imply, the bearers of the general will, as of theprivate will, are always and only individuals. If the distinctionRousseau intends is unsustainable, it is not because it impliesimplausible ontological commitments or otherwise violates soundindividualist intuitions.

Wills, as Rousseau understands them, are distinguished not bytheir bearers but by their objects or, more precisely, by the inter-ests toward which they aim. Private wills aim at private interests;general wills aim at general interests. Thus Rousseau is committedto the idea that these interests are qualitatively distinct.

Rousseau's express formulation suggests that an individual whoaims at a private interest seeks what is individually best; an indi-vidual seeking to realize a general interest aims at what is best for"the whole community." But even if it is conceded that there al-ways is some outcome that is unequivocally best for the wholecommunity, this way of distinguishing general from private inter-ests admits of two distinct understandings, neither of them ade-quate for marking off a genuinely qualitative difference. On onereading, what is best for the whole community just is what is in-dividually best. But, then, general interests would be "private" inthe sense that they are supposed not to be. Rousseau's account ofindividuals' interests would differ from other theorists' only inso-far as he identifies what is individually best with what is best forthe collective entities individuals constitute, whereas others tendto endorse more "individualistic" understandings of the individ-ually best. Undoubtedly, this is an important difference separatingRousseau from Hobbes and his successors in the liberal tradition.But it is not a difference in kind that would constitute a departurefrom mainstream normative individualism. Moreover, on this con-strual, Rousseau's position is extremely implausible. To defend hisview, one would have to show, not just for the social contract itselfbut also for the issues that arise in the popular assemblies of thestate, that the interest of the whole community actually is individ-ually best for each citizen. It seems unlikely that such a claim couldbe sustained. Certainly, Rousseau made no attempt to do so.

The second reading identifies "individually best" with "best foroneself as an individual" and therefore regards general and private

19

The general will

interests as proper subsets of an individual's interest all thingsconsidered. The general interest would then be the individual'sinterest qua citizen; the private interest would be his or her interestqua individual or, what comes to the same thing, the individual'sinterest all things considered with the general interest, as previ-ously defined, abstracted out. Were general interests only propersubsets of individuals' interests all things considered, it couldstill be argued that it is reasonable, under some conditions, toprivilege this proper subset over others. Arguably, Rousseau'saccount of the establishment of de jure states by a social contractin which each individual "puts his person and all his powersin common under the supreme direction of the general will"1

could be interpreted this way. It is arguable too that, whateverRousseau intended, he ought to have had this understanding inmind. But, again, on this reading, general will coordination wouldnot differ in kind from the now standard understanding of rationalcooperation.

For reasons that will become apparent shortly, I shall call theo-rists who deny a qualitative difference between private and gen-eral interests substantive atomic individualists. Substantive atomicindividualism is pervasive in our intellectual culture. But it is nota core conviction in the sense that its abandonment would renderother, more defensible, intuitions about moral and political lifeuntenable. Rather, substantive atomic individualism is a dogmathat blocks progress in political philosophy. Ultimately, the powerof this dogma rests on the appeal of a metaphor of very doubtfulpertinence.

ARE ALL WILLS PRIVATE?

It is plain that only individuals bear interests. But it does not fol-low from this fact of nature that interests can only be predicated ofindividuals and not of sub- or supraindividual entities of animalsor even of things. But the wills that aim at such interests are in-dividuals' wills. It is tempting to conclude from this truism that allinterests are, in Rousseau's terms, private - in one or the other ofthe senses just sketched. Whoever values the interest of his kid-neys over his body altogether or of her political community overher own well-being or whoever accords preeminence to the inter-ests of whales or trees is either seeking to realize an objective that,

i The Social Contract, Book I, Chapter 6.

2O


for whatever reason, he or she considers individually best or elseis privileging some proper subset of what is individually best overother subsets and even over what is individually best tout court.

To dispute this claim for the ineluctability of private volition it iswell to reflect on the intuition that motivates the idea of generalwill coordination. Consider, therefore, situations in which it is nat-ural not to expect reasonable individuals always to do what is in-dividually best. Small group interactions and deliberations amongpersons involved in intimate relationships provide ready exam-ples. Imagine, for example, a couple deciding where it is best forthem to reside. John wants to live in X. Mary prefers to live in Y.However, both can perhaps agree that, insofar as what matters isthe endurance and quality of their relationship, it would be bestfor them to live in W. W, by hypothesis, is not where either John orMary, conceived individually, wants to live all things considered,nor is it in any obvious sense a compromise between X and Y.However, there is a sense in which W is best for both John andMary conceived as "indivisible" parts of a "moral and collectivebody," the couple John-Mary. If we add that John and Mary areequal in morally relevant senses, this John-Mary interest, irreduc-ible to the interests of John and Mary respectively, is paradigmat-ically a general interest.2

To make sense of the intuition upon which this paradigmrests, it will be helpful to realize that each of the construals ofRousseau's position that I have sketched assumes that individuals'interests have to do exclusively (or, at least, mainly) with personalwell-being or, in contemporary terms, welfare. "Welfarism," theconviction that only welfare need be taken into account in "ratio-nal choice" explanations or in normative assessments, is nowa-days widely assumed. But it is hardly a position that bears careful

2 In attempting to make sense of Rousseau's idea, Kant recognized the impor-tance of what has come to be called "the moral equality of persons" for imple-menting a genuine "republic of ends." Following Kant, we can say that personsmust be equal as moral agents. But equality in this sense is unlikely to be real-ized in real-world situations unless there is considerable equality of status andcondition among the deliberating parties. Thus the distinction between privateand general willing, despite Rousseau's insistence to the contrary, is not exhaus-tive. There can be deliberations that are not private will deliberations that arealso not general will deliberations in Rousseau's sense. Parents, for example,can privilege the interests of the family of which they are an integral part overtheir own well-being and even over the well-being of their children without de-liberating in accord with the Rousseauean model, because of the inequalities ofstatus and condition that the parent-child relationship entails.

21

The general will

scrutiny even in the limited contexts in which it typically figures.3

If nothing else, it is plain that welfare is not the only basis for theinterests people actually act upon. Thus personal integrity andcommitment to principles can also motivate human behavior,along with a host of other factors. It is plainly not irrational forpeople to act, say, for the sake of their own integrity or for prin-ciples to which they are committed, and no plausible moral theorycould maintain that individuals ought not to act in these ways.

Welfarism is insidious precisely because, having become so per-vasive, "welfare" tends to be confounded with value generally. Inconsequence, it frequently serves as a placeholder, standing in forwhatever individuals happen to value. But when the idea is actu-ally deployed, as in the unsuccessful attempts just sketched to de-pict general will coordination as a distinct kind of rationalcooperation, more particular understandings of welfare are effec-tively assumed.

Historically, there exist at least two different accounts of whatwelfare is. For the first utilitarians and other early welfarists,but also for some utilitarians today, welfare is identified withstates of mind - specifically, with pleasure or happiness. Morerecently, some writers have come to identify welfare with prefer-ence satisfaction.4 It is plain that, on either view, welfare hardlyincludes everything that people value or ought to value. Butfor individuals directed by private wills, it is fair, though anachro-nistic, to identify rationality with welfare maximization and valuewith personal well-being in the sense(s) the welfarist traditionprovides.

NONWELFARIST INTERESTS

The preceding account of John's and Mary's deliberations ap-pealed implicitly to considerations that are distinct from interestsin personal well-being. When John or Mary privileges a John-Mary interest over their John or Mary interests respectively, theyevince a commitment to the couple John-Mary that is distinct fromtheir concern with their own or even each other's welfare. When

3 Cf. Amartya K. Sen, "Rational Fools: A Critique of the Behavioral Foundationsof Economic Theory/' in Jane J. Mansbridge (ed.), Beyond Self-interest (Chicago:University of Chicago Press, 1990).

4 Cf. Richard B. Brandt, "Two Concepts of Utility," in Harlan B. Miller andWilliam H. Williams (eds.), The Limits of Utilitarianism (Minneapolis: Universityof Minnesota Press, 1982).

22


John or Mary acts out of a commitment to the collective entityJohn-Mary, they are acting in a way that is, as it were, welfare-indifferent. But then what is it that John or Mary seeks to advancewhen they privilege their John-Mary interests?

Some purchase on the nature of general interests can be gainedby asking what it is to be a nonwelfarist interest. Supra- and sub-individual entities as well as things and (most) animals, inasmuchas they are incapable of having states of mind or desires that canbe satisfied, can only have nonwelfarist interests. Therefore thegeneral interest, the interest of the supraindividual entity "thewhole community," must be nonwelfarist.5 In this respect, it dif-fers from the interests registered in "the will of all" - the aggre-gated private interests of the individuals who constitute "thewhole community." The will of all may be a combination of wel-farist interests. But the will of all is not the general will, a pointRousseau insisted upon repeatedly.6

At the very least, we can say that x is in the (nonwelfarist) in-terest of y, where x is a state of affairs and y is anything of whichan interest can be predicated, if x contributes to the survival of y.Thus, if John and Mary are committed to their relationship even atthe expense, if need be, of their own welfares, they evidentlywant, above all, that their relationship should endure. It is difficultto imagine why mere survival should be a paramount concern inthis case. But no matter how unwise it might be, there would benothing incoherent in John or Mary making the survival of theirrelationship their overriding interest. Arguably, "deep ecologists"evince commitments of this sort for endangered species,7 as do op-ponents of "cultural genocide" and some nationalists.

In order to impute more likely motives to John and Mary, and inmost other cases in which nonwelfarist interests figure, we can saythat for x to be in the interest of y, x must do more than contributeto y's survival; x must also be good for y. But good in what sense?This question cannot be answered in general. It depends on thenature of the entity y of which x is predicated. Thus it is prema-ture, at this point, to specify more precisely what a general interest

5 With respect to the nature of (nonwelfarist) interests, nothing of consequencehinges on how large communities are or on how they are individuated. Thus,for the present purpose, the couple John-Mary counts as a "whole community"in just the way that Rousseau's de jure state does.

6 See, for example, The Social Contract, Book II, Chapter 3, and Book IV, Chapter 2.7 Ecologists of this sort are more likely to invoke rights than interests, but their

ideas would seem to lend themselves more naturally to the idiom just proposed.

23

The general will

is. Perhaps this is also why Rousseau himself was reticent on thistopic. If, as I maintain, his model of rational cooperation has spe-cifically political applications only under material conditions thathe never imagined, he could hardly have had more than the barestintimations of the good for such communities. In any case, fornow, it will suffice to say that x is in the (nonwelfarist) interest ofy if it contributes to y's survival and also (usually) if it is in someappropriate sense good for y.

Needless to say, for purposes of normative theory, reasons mustbe given for according weight to the interests people have. In thisrespect, welfarist interests are no less problematic than interestsof any other kind. Even supposing that more than the survival oftheir relationship is at stake, it is not at all obvious in the case ofJohn and Mary that it is wise for either of them to privilege theirJohn-Mary interests over their interests as John or Mary.8 Butthe reality - that is, the coherence and possible applicability - ofJohn-Mary interests should be clear enough to anyone not blindedby welfarist assumptions. Tentatively, we can regard the generalwill at the level of the whole community as a John-Mary interestwrit large. It may be doubted whether such interests actually existin collectivities larger than intimate associations or small groups.Perhaps individuals' abilities to commit to collective entities can-not extend to communities that include nonintimates and strang-ers. It may seem even more doubtful whether commitments to"whole communities," even if they can be made, ought ever togovern individuals' deliberations. But the idea is surely cogent. Wecan imagine what a John-Mary interest writ large would be evenif, pace Rousseau, we doubt its relevance for political life or its rolein normative political philosophy.

However, those who are skeptical about general will coordina-tion could concede that interests can be predicated of supraindi-vidual entities or of anything else and still maintain that allinterests are private. They could argue that when John or Marygives precedence to a John-Mary interest over their own or eachother's well-being, they seek to realize what, in the circumstances,

8 In this case, it is not even plain that only the interests of John and Mary andJohn-Mary bear on their choice of domicile. If their place of residence appro-priately affects the interests, however construed, of individuals other than Johnand Mary, these other persons ought presumably to have a role in the decisiontoo. Arguably, even the judgments of individuals not directly affected may alsobe relevant.


they take to be individually best; or else that they privilege aproper subset of what they take to be individually best over othersubsets and over what is individually best tout court. Thus theyaim at private interests, albeit of a nonwelfarist kind.

But this rejoinder is inadequate. It fails because it supposes, in-correctly, that the notion "individually best" is unequivocal. It as-sumes, in fact, an understanding of "individually best" thatsupports, without quite implying, welfarism and that, even morethan welfarism itself, has come to pervade our political culture andits theory. To expose this assumption, it will be well to acknowl-edge that the expressions "individually best" and its equivalentslike "best all things considered" are ambiguous, depending, as itwere, on how the individual is conceived. As already remarked,this ambiguity is obscured by a metaphor, deployed throughoutthe past three centuries with varying degrees of explicitness by po-litical philosophers including Rousseau, that describes the funda-mental constituents of social and political life in a peculiarlytendentious way.

INDIVIDUALS

According to the ancient atomists, matter consisted of indivisibleparts or atoms bearing only external relations to one another.Therefore, to account for matter at higher levels of organization,atoms - or, according to the refinement introduced after Galileo'srevolution in physical theory, atoms in motion - are all inves-tigators need to countenance. Of course we now know that the at-oms scientists eventually identified are not, in fact, the ultimateconstituents of matter. But this discovery does not impugn the at-omists' guiding idea; it only betokens a mistake in the original ex-ecution of their project. In any case, what matters for our purposeis not the viability of atomism in physical theory but the lesson so-cial and political theorists have taken from the atomists' program.

For atomists, science is about atoms in motion. But it does notfollow from this tenet that scientific explanations must refer toatoms in motion only. The problem, for atomists, is not that phys-icalism is conceptually flawed. It is that the human mind is inca-pable of processing accounts of the complexity that physicalistexplanations would require. This pragmatic consideration sufficesto justify explanations that make reference to matter at higher lev-els of organization. Were our minds more like God's is supposed to

The general will

be, it would be feasible to explain everything atomistically. But inview of human limitations, scientific explanations cannot alwaysbe framed at the level of the ontology they propose.9

However, political philosophers are free from this pragmaticconstraint and can therefore be more atomistic than the atomiststhemselves. Insofar as political philosophy purports not to explainphenomena but only to justify institutional arrangements, there isnothing that stands in the way of full-fledged atomistic accounts.For philosophers captivated by the atomists' program, it is there-fore natural to take the individual conceived atomistically as apoint of departure for moral, social, and political theory. It haslong been recognized that Hobbes's political philosophy assumes aview of individuals conceived by analogy with atoms in motion.We will see in Chapter 2 how Rousseau arrived at a similar posi-tion. For Hobbes and Rousseau and for the majority of theoristsafter them, political philosophy is about forms of association en-tered into by atomic individuals.

Before atomizing market relations came to structure economicand social life in early modern Europe, it could hardly have seemedintuitive to make the individual a point of departure in accounts ofsocial order or institutional arrangements. When, as throughoutmost of human history before the triumph of capitalism, more sol-idary, distinctively social entities - villages, clans, estates - actu-ally were the fundamental constituents of social life, they, notatomic individuals, must have seemed a more natural basis for re-flecting philosophically on social and political arrangements.

In view of the historical particularity of the intuitions that now-adays seem commonsensical to us, it is wise to be cautious inacceding to them. However, I shall not question them or their im-plications here. It is important, though, to distinguish what I shallcall methodological atomic individualism, the idea that social andpolitical philosophy starts from a conception of the individual andhis or her interests, from substantive atomic individualism, the viewthat individuals are radically independent centers of volition. It issubstantive atomic individualism not methodological atomic indi-vidualism, that blocks recognition of the cogency of general willcoordination. Therefore it is substantive not methodological atomicindividualism that must be dislodged in order to find a place forthe general will.9 Some nonpragmatic reasons for eschewing the atomists' inclination to explain

everything in terms of atoms in motion are provided in Wright, Levine, and So-ber, Reconstructing Marxism, Chapter 6.

26


Unfortunately, the name "methodological atomic individualism"invites confusion with "methodological individualism," a well-known position in the philosophy of the social sciences accordingto which social facts are explained by facts about individuals, notsocial groups.10 I have chosen to use it nevertheless because thedesignation is apt in a way that its near namesake is not. Method-ological individualism is a thesis about social scientific explana-tion. Properly understood, it is not a methodological thesis at allbut a claim about what is, as a matter of fact, explanatory. Meth-odological atomic individualism, on the other hand, genuinely is amethodological thesis - about the nature of normative theory.

It is compatible with methodological atomic individualism thatan individual can will anything on any basis whatsoever, welfaristor not. An individual can even will outcomes that fail to repre-sent the radical independence of atomic individuals. Thus onemight seek to maximize some value distributed across a socialgroup, regardless of its distribution among the individuals whoconstitute that group. Atomic individualist philosophers typicallysuppose that individuals seek to advance their own welfare inter-ests. In this regard, Hobbes's case is exemplary. But, as Hobbes'sexample attests, self-interestedness follows from what is supposedto be true of human nature, not from atomic individualism perse.11 Methodological atomic individualism can coexist with anyconception of individuals' interests.

I would venture that, for those who are not wedded to wel-farism, it is methodological atomic individualism that covertly sus-tains the conviction that all wills are private. Since only particularor "private" individuals bear wills, and since this fact is registeredas a methodological tenet, whatever one wills, for whatever rea-son, is private by definition. However, if atomic individualism isunderstood as a claim for the volitional independence of individ-uals, for the substantive reality of an atomic individualist method-ology, the universality of private willing is anything but assured.

It is one thing for a radically independent "atom," bearing onlyexternal relations to other atomic individuals, to seek to advance

10 For a critical assessment of methodological individualism, see ibid.11 Thus, in the Leviathan (see, especially, Chapter 13), Hobbes depicts individuals

as "diffident," "vainglorious," and "competitive" and therefore inclined toseek security, domination of others, and untrammeled acquisition of resources.It is these aspects of human nature, not the fact that individuals' wills are rad-ically independent of one another, that motivate Hobbes's claims about the na-ture of individuals' wants.

The general will

its own interests (welfarist or not); and it is something else for anintegral part of a supraindividual entity to seek to advance its in-terests as an "indivisible part" of that entity. Each of these casescan be conceived to accord with the requirements of methodologicalatomic individualism. But only in the former case is an individual'swill substantively independent of the wills of other individuals.

This difference in willing is easily overlooked because it neednot be evident in the content of the ends individuals seek to bringabout. An atomic individual in the substantive sense can will anyoutcome whatsoever. Therefore a story can always be told that ex-plains individuals' choices which accords with substantive atomicindividualist assumptions. Thus even paradigmatic examples ofgeneral will coordination, like John and Mary's deliberations overa choice of domicile, can be construed in a substantively atomic in-dividualist way. In view of what is usually supposed true of hu-man nature and the human condition, there is reason not to expectindividuals to take anything other than their own well-beings intoaccount in their practical deliberations or normative assessments.But substantive atomic individualism does not preclude this pros-pect a priori.

Similarly, general will deliberations can result in outcomes in-distinguishable from those undertaken by individuals intent onpromoting their own welfare. A person concerned to advance theinterests of a "moral and collective body" of which he is an inte-gral part might aim to realize precisely what he would seek tobring about were he deliberating as a substantive atomic individ-ualist. The general interest can coincide with the welfare interestsof particular constituents of "the whole community." In fact, sincethere is an evident (causal) connection between what is best for thewhole community and what is best for the individuals in it con-ceived independently, it is fair to expect that the general interestfrequently will coincide with the welfare interests of many of theindividual members of the whole community.

Substantive atomic individualism is a claim about the self and itsrelation to other selves in the context of individual and collectivedeliberations. A substantive atomic individual assesses alterna-tives as if she bears no internal relations to other individuals; as ifher valorization of others' well-being, whenever her preferencesdo involve the well-being of others, is strictly contingent. Substan-tive atomic individualism is suggested by methodological atomicindividualism. But it is not entailed by it. Similarly, it suggests wel-farism, but it does not imply it.

28


Welfarism, in turn, suggests but does not imply egoism, the ideathat, in deliberating on what is to be done, one ought to take onlyone's own welfare into account. Even dedicated welfarists, if theyare not also amoralists, would deny that individuals ought to beconcerned with their own well-being exclusively. They would atleast concede that nonegoistic willing is reasonable and even com-monplace in moral contexts. Utilitarianism, for example, enjoinsan attitude of benevolence - according to which individuals taketheir own welfares into account, while according equal weight tothe welfares of all relevant others. Utilitarianism is therefore notegoistic. But it is substantively atomic individualist and also wel-farist; its fundamental principle, to maximize (aggregate or aver-age) utility, requires agents to assess the consequences of theiractions for their welfare effects on substantively independent in-dividuals whose respective utilities are then added together.

Implicitly, four senses of individualism in normative contextshave now been distinguished: (i) methodological atomic individ-ualism, (2) substantive atomic individualism, (3) welfarism, and (4)egoism. According to (1), the individual, conceived atomistically,is the point of departure for moral, social, and political theory. Foradherents of (2), individuals assess alternatives as if they are re-lated to other individuals only externally, though from within thatperspective they may seek to bring about any state of affairs what-soever. Proponents of (3) acknowledge only welfare interests. Ac-cording to (4), individuals take only their own welfare interestsinto account. (1) requires that social and political institutions ac-commodate individuals' interests; (2) holds that individuals' inter-ests are radically independent; (3) asserts that these radicallyindependent interests are always only welfare interests; and (4)maintains that these welfare interests are always self-regarding.For Hobbes and Rousseau, self-regarding welfare interests aremortally antagonistic in the "state of nature"; for philosophers,liberals especially, who follow Hobbes but not Rousseau, interestsremain in competition even after the state of nature is laid to rest.(4) is therefore nowadays widely assumed. But support for (4) iseasily shaken. (3) is more secure; at least it can be maintained with-out abandoning the moral point of view. However, it too is emi-nently vulnerable. Support for (1) is very deeply entrenched,perhaps for good reason. But (1) tends to be conflated with (2). Inthis respect, the liberals and Rousseau are of one mind. In Chapter2, we will find that Rousseau assumes (1), conflates (1) with (2),and then recounts the emergence of (4) and implicitly therefore

29

The general will

also of (3). His notion of general will coordination in The Social Con-tract is expressly a "solution" to a problem (4) generates. Fortu-nately, Rousseau's account - or rather the idea it expresses -stands independently of the problem it purports to solve.

Individuals who coordinate their behaviors according to theRousseauean model advance their interests qua methodologicalatomic individualists. In this sense, they aim at what is individu-ally best. But what is individually best in cases like the ones Rous-seau depicts is what is best for individuals conceived as integralparts of the "moral and collective bodies" they form. To advancetheir own interests, they therefore advance the interests of the col-lective entities that join them together. It is this idea that suggestsitself intuitively when we think about (some) deliberations in smallgroups or intimate associations. Thus, when John or Mary accordspriority to their John-Mary interest over their interests consideredatomistically, they do what is best for themselves as members of anentity they relationally comprise. When they place themselves"under the supreme direction of the general will," it is the inter-ests of the couple - not of the individuals who constitute it, con-ceived independently - that matters for their deliberations. Thecouple John-Mary is an internal relation formed by John andMary. It cannot be decomposed into two radically independent en-tities, the (atomic) individuals John and Mary, because in this caseit is the relation itself, not the parties related, that is fundamentalfor the proposed deliberation.

For substantive atomic individualists, on the other hand, it is in-conceivable that individuals could be related internally. Their pic-ture of the atomic individual precludes this conception. But weshall need to (re)introduce precisely this idea into political philos-ophy in order to put to use the insights implicit in Rousseau'saccount of legitimate political arrangements and in Marxiancommunism.

A FREE-RIDER PROBLEM?

Altering or, in this case, restricting the scope of a governing met-aphor can have consequences for more than just the language withwhich the situations it represents are described. In this instance,doing so affects the way conflicts between (putative) general andprivate interests are conceived, and the way resolutions of theseconflicts figure in normative theories.


Since the difference between general and private volition is notunequivocally evident in the content of the ends agents will, asubstantive atomic individualist could in principle account forthe same phenomena as proponents of general will coordination.At issue is the reality of general interests, not the content of indi-viduals' choices. Consider, then, how a substantive atomic individ-ualist would describe a conflict between an (alleged) generalinterest and a private interest. An extreme case will put the matterinto focus. Imagine an individual who would give up her life in theinterest of some "whole community" - a family, a tribe, a nation,a people, or a (de jure) state. It is plain that there are imaginableinstances of such behavior that substantive atomic individualistswould consider rational (or not irrational) and even morally praise-worthy. How can they reconcile these assessments with the indi-vidualism they profess?

One way would be to construe the individual's self-sacrifice asan act of altruism undertaken to benefit particular others. How-ever, this interpretation becomes increasingly unlikely the larger"the whole community" involved. It is one thing to sacrifice one'slife for one's children and something else to do so for (unknown)compatriots or (unborn) future generations. In any case, imaginean act of self-sacrifice - perhaps an act of exceptional bravery un-dertaken in time of war - that cannot plausibly be interpreted al-truistically. It would seem that the only way to maintain asubstantive atomic individualist perspective in cases of this sort isto suppose that the self-sacrificing individual is intent on not be-coming a "free rider."

People have interests in their own well-being. But individualwell-being may depend upon the existence - and perhaps also thegood (however conceived) - of one's family, tribe, nation, people,or state. A substantive atomic individualist might then claim thateach individual would be best off not making sacrifices for thesecollective entities but letting others do whatever is necessary tosustain them and make them flourish. Individuals would do best,in other words, if they "free-ride" on the efforts of others. But, bynearly all accounts, free riding is reprehensible, and a refusal tofree-ride is correspondingly praiseworthy. Thus it could be arguedthat the esteem accorded those who make sacrifices for collectiveentities is nothing more than the esteem that attaches to those whorefuse to obtain benefits without incurring the costs required toproduce them.

The general will

This construal of self-sacrifice, like the interpretation that as-similated self-sacrifice to altruism, assumes a welfarist under-standing of individuals' interests. But this is not the principalreason why, in imaginable examples, an individual who wouldgive up her life for "the whole community" evinces somethingdistinct from a reluctance to free-ride. In the cases in question,she would not free-ride, even if she could somehow be releasedfrom her (acknowledged) duty to contribute toward "publicgoods," because, by hypothesis, her true interest is the interest of"the whole community" of which she is an "indivisible part."12

She has, in other words, made the general interest her own.Therefore, even in giving up her life, she does what she trulywants, not what she feels she must do despite her desires.13 Ifshe feels conflicted, it is not because she faces a free-rider prob-lem. What she faces is an intrapersonal conflict in which herinterests qua atomic individual and her interests qua part of theirreducibly collective entity for which she would give up her lifevie for control.

Rousseaueans are not alone in advancing a notion of internalconflict that resists a free-rider construal. The deep ecologist whowould have the interest of a forest take precedence over his ownwelfare interests is, in this respect, in the same situation as the in-dividual who would sacrifice her life for the state. If his position isflawed, it is in virtue of the cogency of his reasons for favoring theforest's interests over his private interests, not because it is inco-herent to make a forest's interests one's own.

Needless to say, a case must be made for adopting a generalinterest or a forest's interest or indeed any interest whatever.The reasonableness of acting on welfare interests is relativelyuncontroversial. On the other hand, persuasive reasons for mak-ing the interests of supra- or subindividual entities - or of animalsor plants or inanimate objects - one's own are more difficult todefend in an intellectual culture in which substantive atomic indi-vidualism, welfarism, and even egoism predominate. But the co-herence of these nonegoistic, nonwelfarist, and substantivelynonindividualistic interests and their possible applicability shouldbe beyond dispute. In this (Kantian) sense, the general will is real,

12 For the same reason, the social contract Rousseau envisions is not the outcomeof a protracted negotiation. It is, as it were, each individual's initial negotiatingposition, the outcome desired most by each contracting party.

13 A similar argument is provided by Rousseau in The Social Contract, Book II,Chapter 5.

32


whether or not it has an important role to play in political philos-ophy or normative theory generally.

INTERNAL RELATIONS

It was to draw a contrast with the atomic individual that I invokedthe idea of internal relations. Historically, the term has designateda variety of positions, but its role here is straightforward and un-controversial. A property is essential to some entity if the entitywould not be what it is in the absence of that property. Relationscan be essential properties. Thus being the day after Sunday is es-sential for being Monday. When entities are related in a way that isessential for their being what they are, they are related internally;relations that are not internal are external. Intermittently since theeighteenth century, internal relations loomed large in metaphysi-cal discussions - intersecting debates about universals, analytic-ity, essentialism, and the theory of truth. The idea has provenparticularly useful in attempts at rendering Hegelian insightsclear. Thus it was widely invoked in Britain and America from the1890s until the First World War - in the work of Bradley, Royce,Bosanquet, and other neo-Hegelians. For much the same reason, ithas also figured in some recent attempts at Hegelian reconstruc-tions of Marxist positions, for example in Bertell Oilman's analysisof alienation.14

However this may be, the notion deployed here is innocent ofportentous methodological or metaphysical implications. The termis intended only to signal an alternative to atomism in normativetheory. The idea is that, in thinking about how social and politicalarrangements ought to be, it is sometimes appropriate to consider

14 See Bertell Oilman, Alienation: Marx's Conception of Man in Capitalist Society(Cambridge: Cambridge University Press, 1971). In Oilman's use, the notionbears an affinity with and arguably clarifies the Althusserian concept of "ex-pressive totality," ascribed by Althusser to Hegel and the young Marx andthen, in Althusser's view, abandoned by Marx after his "epistemologicaJbreak" with the "problematic" of his early writings. See Louis Althusser, ForMarx (London: Allen Lane, 1969); Louis Althusser and Etienne Balibar, Reading''Capital" (London: New Left Books, 1970). On this understanding, each part ofa social totality somehow contains within it ("expresses") the essential prop-erties of the whole. On the Ollman-Althusser view, the universe can in prin-ciple be apprehended in a grain of sand. However, in practice, as Oilmanconcedes, it is necessary to know a great deal about a social system in order toascertain precisely what a particular case expresses. Indeed, it is far from clearwhat this doctrine, so understood, adds to standard accounts of social scien-tific practice - beyond an obscure metaphysical redescription.

33

The general will

individuals to be related internally. Thus John and Mary, in thecouple John-Mary, are what they are in virtue of their role in thatrelation. The John and Mary who exist independently of the John-Mary couple are not essentially the John and Mary related inter-nally. Needless to say, the names designate the same humanbeings, and the independent spatiotemporal existence of Johnand Mary is not in dispute. What is at issue is how their interestsare conceived - whether atomistically or as integral parts of "thewhole community" they constitute. For the larger communitiespertinent to political philosophy, the same consideration applies.

THE REALITY OF THE GENERAL WILL

I first described general interests as nonwelfarist and communal.That characterization was apt but incomplete inasmuch as privateinterests can be nonwelfarist and communal too. I therefore pro-posed a characterization that more incisively differentiates generalfrom private willing. Private interests are individuals' interests quaatomic individuals; general interests are individuals' interests quamembers of collective entities in which they are related internally.

To hold that there can be "moral and collective bodies" withinwhich individuals are internally related is to maintain that collec-tivities can have interests irreducible to the interests of their con-stituent parts conceived atomistically. General will coordinationexists whenever individuals deliberate and act with a view to re-alizing such interests. I have noted that within intimate associa-tions and in small groups there exist phenomena that Rousseau'sidea plainly models, and I have argued for the superiority of theRousseauean account of these phenomena over the most likelysubstantive atomic individualist descriptions. But, so far, I havesaid nothing in defense of its applicability to "the whole commu-nities" of interest to political philosophy. Neither have I yet ad-dressed the question of its desirability.

Liberals have always argued for principled limitations on therightful use of state power.15 Because "the exercise of the generalwill" seems inimical to individuals' sovereignty over themselves,people with liberal sensibilities are likely to question the desirabil-ity of general will coordination in the political arena, even if theyconcede its feasibility. Similar hesitations, though for different rea-sons, are likely to resonate for anyone influenced by Marx. Marx

15 See Chapter 6 herein.

34


believed that true general interests cannot exist in class-dividedsocieties even notionally, and he thought that appeals to the gen-eral interest are always veiled invocations of private interests -typically of those who benefit from the prevailing system of classdomination.16 But, as I have already claimed, we shall find thatMarx's assessment is different for societies that have overcomeclass divisions. Not unrelatedly, we shall also find that liberalism isnot quite so inhospitable to general will coordination as liberals areinclined to believe.

Questions about the reality and desirability of general will coor-dination are conceptually distinct. However, treatments of thesequestions intersect - in part because the exercise of the generalwill at the level of whole communities is desirable just in those ma-terial conditions and under those social relations in which it be-comes feasible; in short, in the conditions that define communismin the Marxian sense. Of course, Rousseau hardly envisioned thisimplication of his model of rational cooperation. Nevertheless, itwill emerge that in crucial respects he was an important, if unwit-ting, contributor to its defense.

16 See, for example, the essay On the Jewish Question in Karl Marx and FriedrichEngels, Collected Works (New York: International Publishers, 1975), vol. 3, pp.146-174. A similar claim is advanced by Rousseau in The Second Discourse; seeChapter 2 herein.

35

Chapter 2

The "origin" of the private will

The term "private will" is rarely used nowadays except in schol-arly discussions of Rousseau. Nevertheless, what it designatespervades moral, social, and political theory. As remarked, eventhose who would concede the existence of the general will, theprivate will's rival in the Rousseauean scheme, believe that ithas little, if any, applicability in political contexts. On the otherhand, the reality and applicability of private volition are beyonddispute. I have suggested that this understanding is partly a con-sequence of a lingering inclination to consider individuals to berelevantly like radically independent atoms. But the idea that theprivate will is unproblematic and ubiquitous is also encouraged bymore direct confusions.

Individuals moved by private wills seek, insofar as they are ra-tional, to do as well for themselves as they can. This claim appearsinnocuous. But we have seen that it is ambiguous and potentiallymisleading. It is also contrary to the views of virtually all moralphilosophers before Hobbes. Practical reason, on the formerlydominant view, was substantive, not just instrumental; rationalityconcerned the ends individuals' willed much more than the ade-quacy of the means they adopted. To be sure, some philosophersalong with theologians and moralists continue to resist instrumen-talist accounts of practical reason. Still, the idea that rationality isa matter of advancing one's own interests (where "interest" is aplaceholder term, awaiting substantive elaboration) is so seldomdisputed nowadays - and so integral to the tradition of economicmodeling in contemporary social and political theory - that it haseffectively become part of common sense.

However, the common-sense view does not fare well underscrutiny. On the most natural interpretation of it, value is utility,


and rational agency is utility maximization.1 But it is doubtfulwhether, even for purposes of normative theory, individuals canbe said to have well-defined utility functions. It is also plain thatthe rationality of straightforward utility maximizing can be chal-lenged, particularly in contexts of interdependent action, wherethe "payoffs" to agents depend, in part, on what other agents do.In recent years, insights adapted from the mathematical theory ofgames and from decision theory have shed light on the nature andlimits of this notion of rational choice. But it has been known atleast since Hobbes's account of the "state of nature" (in which in-strumentally rational agents pursue their own ends without con-straint) as a "war of all against all" that unconstrained utilitymaximization can be problematic. In any case, whatever we con-clude about the dominant view of rational agency, what matters,for now, is that it is neutral with respect to claims about the contentof individuals' interests. It is one thing to hold that rational agentsseek to advance their own interests and something else to say whattheir interests are.

Nevertheless, the complex of ideas included in current notionsof private volition - and in Rousseau's earlier elaboration of theidea - includes claims about interests. Thus it has become com-monplace to construe individuals as economic agents intent on max-imizing their shares of resources, including leisure. To join thisview of interests to the utility-maximizing view of rational agency,it need only be supposed that individuals' utility yields rise withincreasing amounts of material resources and diminishing laborinputs. But this causal connection is not implied by the theory ofrational agency with which it is historically associated. To say thatrational agents maximize utility is not to say anything about theutility consequences of particular states of the world. So far as thereceived view of practical reason is concerned, individuals could beacting rationally when they seek to increase the shares of others,to implement particular distributive outcomes, or even to diminishtheir own resource bundles while maximizing their labor inputs. Ifactions of these kinds seem "irrational," it is because standardeconomic models are so taken for granted that it affronts commonsense to call their predictions into question.

i A forceful statement of this position is provided by David Gauthier in "Reasonand Maximization," Canadian journal of Philosophy, vol. 4, no. 3 (1975), pp.411-433.

37

The general will

These observations will be developed in the reflections on Rous-seau's Discourse on the Origin of Inequality among Men or Second Dis-course that follow. The lesson I shall draw is that, despite what iswidely believed, the private will - especially but not only as Rous-seau conceived it - is problematic too. What is of continuing philo-sophical importance in The Second Discourse is obscured if weassume, as readers today are wont to do, that Rousseau's aim wasonly to describe the genesis of a form of volition that no one canreasonably fault. In fact, it was Rousseau's contention that theemergence of the private will and then its exclusive reign is a tran-sitory episode in human history. My aim in this chapter is to bringthis claim into focus and to strengthen it by disentangling some ofthe ideas about practical reason and human interests that Rous-seau inadvertently confounded.

THE ORIGINAL STATE OF NATURE

According to Rousseau's express declaration, The Second Discourse,in its first part, aims to provide an account of "the natural state ofman." However, what Rousseau intended by this expression is eas-ily misunderstood. Perhaps it was obscure even to Rousseau him-self. At the risk of overlooking some facets of the story The SecondDiscourse tells, I shall here provide a clarifying sketch of this nat-ural state, the original "state of nature," that is faithful to most ofwhat Rousseau says. Since there is no familiar category into whichthe concept fits, it will be helpful, at the outset, to distinguish itfrom other notions with which it might be confused.

Despite references to indigenous peoples in the Pacific islandsand the Americas, Rousseau's account of "the natural state ofman" is not a composite picture of primitive societies. His depic-tion of human beings' natural state is based largely on his ownimagination, constrained only by a few, widely accepted "facts."As speculative anthropology, therefore, the state of nature thatRousseau describes hardly warrants serious attention even in com-parison to the travelers' journals and tales of exotic customs andplaces that appeared in Rousseau's time. In short, The Second Dis-course is not a history of the origins of inequality.

For readers today, Rousseau's state of nature may suggest the"original position" deployed by contemporary contractariantheorists.2 Within this genre of theorizing, the idea figures in

2 This designation was first deployed by John Rawls in A Theory of Justice (Cam-

38


thought experiments in which institutional arrangements are jus-tified by demonstrating that they would be chosen by rational in-dividuals in a setting, the original position, in which all vestiges ofthese arrangements have been abstracted away. In its account ofthe establishment of sovereignty by contract, The Social Contract ar-guably does depict the state of nature as an original position. ThusRousseau maintained that just political arrangements are con-structed by rational individuals at the point at which the state ofnature has become a Hobbesian "war of all against all" - where"obstacles that are harmful to [each individual's] maintenance"come into mortal contradiction with "the forces that each individ-ual can bring to bear to maintain himself."3 But we shall see thatthis unfortunate condition is, in fact, the culmination of a processthat The Second Discourse recounts. In The Second Discourse, there-fore, "the state of nature" refers to more than the original positionof The Social Contract.

Unlike the prototypical "war of all against all" in the Leviathan,the original position of The Social Contract is not an eternal con-dition lurking beneath the fragile surface of political order. Itis socially produced. The state of nature that the social contractends is therefore not a description of human nature and the hu-man condition per se. It is an account relativized to particular cir-cumstances - specifically, to conditions that develop after theintroduction of private property in land, the principal nonhumanproductive asset in preindustrial economies. This monumentalevent and the ineluctable transformations that follow in its wakeare recounted in Part 2 of The Second Discourse. For politicalphilosophers focused on The Social Contract, The Second Discoursecan therefore be read as a prologue. From this perspective, it isnatural to construe Part 1, Rousseau's account of the time beforethe introduction of private property, as stage setting. However this

bridge, Mass.: Harvard University Press, 1971). It aptly characterizes only oneof the uses to which Rousseau and other earlier contractarians put "the state ofnature."The Social Contract, Book I, Chapter 6. Rousseau expressly denied that this con-dition is a state of war, insisting that only states, not individuals, can be at war.But the universal opposition among individuals that Rousseau describes is pre-cisely what Hobbes's expression designates. Since "the war of all against all" isa widely known and well-understood description of the idea Hobbes and Rous-seau share, I shall, where appropriate, use the Hobbesian term despite Rous-seau's objections to it. For further discussion, see Levine, The Politics ofAutonomy, Chapter 1, esp. pp. 20-28.

39

The general will

impression is misguided; Rousseau's depiction of the original stateof nature bears considerable philosophical interest.

Thus it is well to reflect on Rousseau's contention that, in thosefirst days, human life was solitary but happy and secure. What heevidently intended to convey was the idea that human nature andthe human condition are as compatible with peace as with a war ofall against all or, more precisely, that "the natural state of man" istoo indeterminate to generate any conclusions about war or peace.To arrive at a precise determination, the social conditions that per-tain must be specified.

Rousseau's aim in describing "the natural state" of the first hu-man beings was precisely to characterize human nature and thehuman condition apart from any particular social conditions.Hence the solitude of the first men and women. According toRousseau's story, in those first days individuals encountered eachother infrequently and by accident. There was no language; there-fore there were no linguistic communities. Families did not exist.Of course, there was sexual intercourse, but men and womenstayed together only long enough to copulate; then they parted,perhaps never to see each other again. Children stayed with theirmothers only until they were old enough to fend for themselves.At that moment, the mother-child bond dissolved. In short, intheir natural state, "with neither a fixed dwelling nor any need forone another, [people] hardly encountered one another twice intheir lives."4

This graphically atomistic picture might be motivated by a con-viction that human beings really are disposed to be solitary; thatcommunities of any sort, even the most intimate, are at odds withhuman nature. But Rousseau believed nothing of the sort. Thestate of The Social Contract is a community united by a general will,and a sense of community on the part of its citizens is necessaryfor its maintenance and well-being. Unless we suppose that the ca-pacity to join together in this way arose out of nothing, its rootsmust lie in human beings' innate capacities. Thus Rousseauclaimed that pity (pitie), empathy, and compassion for the suffer-ing of one's fellows moved even the first human beings. Pity is theroot from which the sociability that makes general will coordina-tion possible grows. Inasmuch as pity, by Rousseau's account, is a

J.-J. Rousseau, Oeuvres completes, vol. 3 (Paris: Pleiade, 1964), p. 146. Subsequentreferences to The Second Discourse, hereafter SD, will be to this edition. In allcases, the translations are my own.

40


sentiment human beings share with the higher animals, its intro-duction is not quite ad hoc. But its inclusion in a socially unme-diated state of nature is certainly fortuitous. It is because humanbeings are capable of pity that, though naturally solitary, they caneventually join together to form a just state.

Thus Rousseau abstracted society away, while allowing for its(re)constitution. His aim was to examine the will under natural -that is, pre- or nonsocial - conditions. One of the conclusionshe derived from this exercise is even more startling today than itwas in his own time: that the private will is not a direct expressionof human nature. It is a consequence of very particular socialcircumstances.

BEFORE THE PRIVATE WILL

Individuals moved by private wills seek to realize their privateinterests as "efficiently" as possible. As already noted, this ideaincludes distinct claims about rational agency and individuals'interests. Perhaps one could not succeed in realizing private inter-ests, except capriciously or by accident, without being an instru-mentally rational agent. But one can be a rational agent withoutaiming at private interests.

Rationality, in the sense in question, is tantamount to consis-tency or, more strictly, transitivity. Suppose, as is commonplace to-day, that interests are represented by preferences for alternativeoutcomes.5 Then, for any rational agent, if x is preferred to y andy is preferred to 2, x is preferred to 2. Since reason does not rule onthe content of our ends, a preference for x over y or y over z is initself neither rational nor irrational. But, given these preferences, xmust be preferred to z. If we require in addition that preferences becomplete - that, for any agent and for all x and y, either x is pre-ferred to y or y is preferred to x - preferences can be representedby ordinal numbers. Additional conditions must be satisfied forpreferences to be represented by cardinal numbers. Unless cardi-nal numbers can be assigned to utilities in a theoretically mean-ingful way, individuals' utilities cannot be added together as, forinstance, utilitarianism requires. But even ordinal rankings are be-yond the capacities of individuals in the original state of nature.

5 For greater verisimilitude in modeling actual choice situations, it would be bet-ter to describe the indicated relation not as "preference" but as "preference orindifference." However, to simplify the exposition, I have omitted this precision.

41

The general will

Rousseau did not conceive the problem of rational agency so ab-stractly; he therefore never quite impugned the rationality of thefirst human beings in these terms. But he was eloquent in describ-ing the absence of arts and sciences in the original state of natureand in ascribing this (for him) mixed blessing to the underdevel-opment of peoples' intellectual capacities. "The progress of themind [is] precisely proportionate to the needs received by peoplesfrom nature or to those needs to which circumstances have sub-jected them/'6 In the original state of nature, desires are so limitedthat they can only be satisfied directly and immediately. Thus cir-cumstances militated against the development of individuals' (in-nate) abilities to adapt means to ends. Even ranking alternativeswas beyond the capabilities of the first men and women.

To seek to do as well for oneself as one can is to seek to advancethe well-being of a unitary entity enduring across some span oftime. Rational agents may have time preferences, represented bythe rate at which their utility functions discount future outcomes.7

But reason itself is indifferent to temporal stages. This view of ra-tionality therefore presupposes that persons are capable psycholog-ically of conceiving of themselves as beings who exist continuouslyover time (though not necessarily for entire lifetimes).8 Thusagents must have minds sufficiently integrated to join particularactions into coherent projects. But, in Rousseau's view, for an in-habitant of the original state of nature, this capacity does not yetexist. Of this benighted individual, Rousseau writes:

His soul, agitated by nothing, is given over to the single feeling of his ownpresent existence, without any idea of the future, however near it may be,and his projects, as limited as his views, hardly extend to the end of theday. Such is, even today, the extent of the Carib's foresight. In the morn-ing he sells his bed of cotton and in the evening he returns in tears to buyit back, for want of having foreseen that he would need it that night.9

In other words, in the original state of nature, human beingslacked the very capacity to will coherently. Their "wills," if they

6 SD, p. 144.7 A minority current within the mainstream, represented most famously by Sidg-

wick, denies even the rationality of time preferences. See Henry Sidgwick, Meth-ods of Ethics (New York: Dover, 1966), Book III, Chapter 13, section 3.

8 This position has been contested by a number of writers recently, most famouslyby Derek Parfit in Reasons and Persons (Oxford: Clarendon Press, 1984). A similarobjection was, of course, raised long ago by David Hume; cf. A Treatise of HumanNature, Part 4, Chapter 6.

9 SD, p. 144.

42


can be so designated, were merely successions of impulses. Thefirst men and women could therefore act only out of immediateinclination, never in a calculated way. Thus their wills were not yetfully private, if only because they were not rational agents in therequisite sense.

Like other living things, the inhabitants of the original state ofnature were moved by "self-love" (amour de soi), by a concern fortheir own well-being. But lacking the capacity to plan for an ever-receding future, they could only act unreflectively - to protectthemselves and to satisfy their wants. They were incapable of for-bearance or planning or strategizing. Therefore their "interests," ifthe term properly applies to the objects of the impulses that con-stitute their "wills," were more ephemeral than those towardwhich true private willers aim. This fact has important implica-tions for the nature of their interests. Prior to the social contract,competition gives rise to a war of all against all because bearers ofprivate wills, according to Rousseau's account, are accumulators,not just consumers; and accumulators, unlike consumers, can pur-sue their passion indefinitely without encountering natural limitsto mitigate their conflicts. Lacking means to coordinate their ac-tivities, competition among rational accumulators therefore takeson the deadly aspect Rousseau imputes. On the other hand, inas-much as their interests did not extend beyond what they were im-mediately able to grasp, the first men and women were notaccumulators - nor could they be. For these primordial human be-ings, the motivational principle that so many today consider uni-versal was literally inconceivable.

Bereft of rationality and of desires beyond their immediate ho-rizons, the first men and women differed in nearly all respectsfrom the individuals who end the state of nature by a social con-tract. In each case, the people depicted are substantive atomic in-dividuals. But their dispositions differ, and therefore their socialinteractions differ too. In the original state of nature, lacking thecapacities necessary for protracted antagonisms, individuals werenecessarily at peace. In the circumstances that make the socialcontract necessary, individuals' wills are opposed and conflict isuniversal. But, pace Hobbes, this situation is not a consequence ofhuman nature and the human condition after everything distinc-tively social is subtracted away. In fact, universal conflict is un-thinkable for men and women in their truly natural condition.

Nevertheless, a Hobbesian war of all against all is humanly pos-sible in certain conditions, and so too is the harmonious, internally

43

The general will

coordinated order that the social contract establishes. It is theburden of The Second Discourse to demonstrate the social conditionsnecessary for realizing these supposedly exhaustive possibilities.But the conditions Rousseau adduces are not equally attainable atall moments of the history he recounts. General will coordinationbecomes possible only after private volition has worked its effects.

FROM SELF-LOVE TO AMOUR PROPRE

The private will emerged in consequence of a process begun by asingular event. According to Rousseau's account, "the first manwho, having enclosed a plot of land, took it into his head to say thisis mine and found people simple enough to believe him, was thetrue founder of civil society."10 It was this act that inaugurated ahistorical dynamic culminating in a war of all against all waged bybearers of full-fledged private wills.

From primitive times through Rousseau's day, land was, alongwith labor, the principal means of production. Of course, in theoriginal state of nature neither labor nor land is owned. But thereis an important difference. Labor cannot be separated from the in-dividuals who deploy it; land is attached to no one by nature.Therefore, even without private property, people effectively con-trolled their own labor and appropriated its product. A successfulassertion of self-ownership in the original state of nature wouldonly formalize what was already the case. On the other hand, theprivate appropriation of land marked a profound departure from"the natural state of man." It is not just that a social relation - pri-vate property in alienable means of production - was then intro-duced into an otherwise nonsocial condition. What is crucial is thenature of this social relation. Making a plot of land one's ownposes a revolutionary challenge to the ancien regime of Rousseau'sstory. The revolution succeeded according to Rousseau, and theconsequence was an epochal transformation in the condition hu-man beings confront.

In Rousseau's view, the successful enclosure of plots of land -or, more generally, the private appropriation of alienable means ofproduction - might not have occurred. In contrast to what was tocome, the original state of nature, like the social orders of nonhu-man species, was reproducible indefinitely. But there were poten-tially destabilizing aspects of this condition that distinguished it

10 SD, p. 164.

44


from the situation of other animals. To a degree unprecedented inthe animal kingdom, the first humans beings were potentially ra-tional in the Hobbesian sense and intelligent. Without language orother means for cultivating innate capacities, rationality and intel-ligence were undeveloped and might have remained so forever.But, as a matter of fact, "progress" did occur. Needing to compen-sate for physical shortcomings, people fashioned tools to aid inhunting and gathering. With means for transmitting knowledgelacking, what was invented was usually lost only to be reinventedby others elsewhere. But some imitation took place. With so littlesocial interaction, imitation served the accumulation of knowledgepoorly. Still, over time, a few advances came to be widely shared.Above all, knowledge of metallurgy became part of the technolog-ical legacy of humankind. Through gradual accretion, these ad-vances eventually made settled agricultural production materiallypossible. But this technology could not be productively deployed.Without private ownership of land, the products of agricultural la-bor are available for anyone to seize. Therefore, despite its poten-tial for augmenting yields per unit of land, agricultural technologyin the absence of property relations would probably be less pro-ductive than hunting and gathering; it would certainly be less pro-ductive than it could otherwise be. In short, in the conditionsRousseau described, settled agriculture was impossible; the tech-nology was "fettered." It remained so until it was liberated by "thetrue founder of civil society."

In Marxist terms, this change in social relations of production facil-itated the development of productive forces. Labor power becamemore productive, the product of labor increased, and scarcity -understood in absolute terms, as an inverse measure of sociallydesired things - diminished. Marx's theory of history purports toexplain epochal social transformations by appeal to a universal hu-man interest in diminishing scarcity. However, this is not quite theinterest that explains the demise of the original state of nature.Since the first men and women lacked the capacity to desire any-thing that was not immediately present, they made no demandson nature that nature could not fulfill. For them, the miserliness ofnature passed unnoticed. Scarcity was neither a spur to develop-ment nor a source of social instability. It is not absolute but relativescarcity, the supply of socially desired things relative to the de-mand for them, that unleashes the dynamic process Marx identi-fied. Absolute scarcity was never more acute than in the originalstate of nature. But in those first days, relative scarcity did not

45

The general will

exist at all. Paradoxically, it was development itself or, more pre-cisely, development under the property relations introduced by"the true founder of civil society" that made scarcity a salient andcausally efficacious aspect of the human condition. Once relativescarcity was socially introduced, history became invested with adeterminate structure and directionality. The subsequent course ofhuman affairs then became inevitable.

Those who succeeded in turning plots of land into private prop-erty and who farmed their plots successfully were, of course, bet-ter off than their neighbors who continued to hunt and gather onunowned land. They became models to imitate. Those who failedto follow their example became even worse off than they were be-fore, if only because they had less land upon which to roam insearch of food. Insofar as the remaining hunters and gatherershad, by that time, become somewhat intelligent and (instrumen-tally) rational, they were moved to follow the lead of their moreproductive neighbors. Thus a process was set in motion that led tothe private appropriation of virtually all available land. In time,hunting-and-gathering was no longer a viable way of life. Produc-tivity increased and absolute scarcity diminished. But relativescarcity intensified inexorably.

In the original state of nature, the size of the human populationand the capacity of the land to support it were in equilibrium. It isnot clear why Rousseau thought that population pressures de-velop as productivity increases. Perhaps he believed that agricul-tural production runs up against inexorable physical limits - areasonable expectation, especially in a preindustrial age. In anycase, it is plain that Rousseau did not believe that the primordialharmony of the original state of nature could be maintained withrising productivity. The more there was to distribute, the largerthe number of individuals demanding a share. Scarcity thereforebecame increasingly important. Indeed, it came to shape individ-uals' interests and to transform their wills.

With settled agriculture came the family, patriarchically struc-tured and based on blood ties. In Rousseau's account, the family isan institution for organizing the labor process and reproducing la-bor power. It is also an institution for transmitting land and otherproductive assets from one generation to another. With rising pop-ulations, family plots were divided among surviving sons, each ofwhom established his own family in turn. Thus, with land re-sources remaining constant and population rising, the amount ofland per family - and per person - diminished from generation to

46


generation. This situation motivated the intensification of laborand the introduction of new technologies, increasing overall pro-ductivity and diminishing absolute scarcity. But these palliativeswere insufficient to offset population growth. Though vastly moreproductive than the hunting and gathering technology that pre-ceded it, settled agriculture brought relative scarcity into being;additional "progress" then intensified relative scarcity, turning itinto the aspect of the human condition that, more than any other,makes a social contract necessary.

But relative scarcity also plays a role in the emergence of theconditions that make a social contract possible. In order to survivewith scarcity intensifying, individuals were obliged to develop ca-pacities dormant in the original state of nature. Thus they becamethe full-fledged (instrumentally) rational agents they potentiallyalways were - setting long-range goals, deliberating prudently,and acting strategically. Their behaviors were increasingly gov-erned by a determination to anticipate future wants, a momentousdevelopment with far-reaching implications for individuals' inter-ests and dispositions. The fleeting impulses that directed the be-haviors of the first men and women gave way to sustained projectsorganized around the relentless accumulation of resources.11 In asense, acquisitiveness is only a continuation of self-love after pri-vate property has transformed the human condition and humanbeings' capacities to adapt to it. But the disposition to accumulateeventually took on a life of its own. "Human nature" became ac-quisitive, just as Hobbes thought it always was.

Before economists appropriated and developed this idea, it wastacitly if unreflectively supposed that individuals always wantmore of everything. But, of course, no one does. People becomesatiated, and their desires for particular things usually diminish asthese things are consumed or acquired. For greater verisimilitude,therefore, it has become standard to hold that leisure is a univer-sally desired good that agents trade off against other goods thatthey can only acquire through effort. Thus, for most goods, indi-viduals' utility functions will be such that, beyond a certain point,additional units will be desired in diminishing degrees - and lei-sure desired more.

11 For adherents of the view of rational agency assumed by those who considerprivate volition ubiquitous, accumulation per se is neither rational nor irratio-nal. But, given the human condition and individuals' "natures," rational hu-man agents will, as a matter of fact, want (virtually) unlimited quantities ofresources.

47

The general will

It will be useful to represent Rousseau's position not in the vagueway that he advanced it but with this clarification in mind. Then,if we introduce a universal measure of exchange, money, we cansay that more money will always be desired. Leisure too is alwaysdesired. What will vary is the degree of intensity of these desires.The shape of an individual's utility function is, strictly speaking,indeterminate, even for full-fledged private willers. But, typically,when additional amounts of money can only be acquired by sac-rificing leisure, individuals' utility functions will at some point ex-hibit diminishing marginal utility for money increments. In otherwords, the utility of yet more money will never be negative.

In any case, it was only after the introduction of private propertythat individuals became rational economic agents.12 For this out-come to result, it was not necessary that everyone be burdenedwith acquisitive natures at the inception of civil society. It sufficesif only some individuals are. A process analogous to natural selec-tion can do the rest. We have seen how the presence of even a fewaccumulators among persons moved by self-love worsens thesituation of the majority. As selection proceeds, this process fuelsitself. The more rational economic agents there are, the more rap-idly the processes that engender rational economic agency worktheir effects. Accumulation becomes the order of the day, pro-pelled by ever-increasing (relative) scarcity but also by an evermore acquisitive "human nature."

12 It is worth remarking, again, that this destination of the protracted processRousseau depicts was, for Hobbes, a universal feature of human psychology.Hobbes's picture of relentless accumulation effectively secularized an aspect ofSt. Augustine's account of human psychology after the Fall, according to whichan inexorable passion to accumulate is one of the afflictions visited upon hu-mankind in consequence of Original Sin. See Herbert A. Deane, The Politicaland Social Ideas of St. Augustine (New York: Columbia University Press, 1963),Chapter 2. Thus what Hobbes considered normal and even desirable August-ine thought a sign of human degradation. In effect, The Second Discourse revivesthe older Augustinian understanding. In Rousseau's account, the mentalitythat emerges in consequence of private property is an affliction and a result ofa fall from an earlier condition of felicity. However, for Rousseau, the OriginalSin of the first men and women was not Pride or Disobedience but the intro-duction of private property in alienable means of production. Although Rous-seau had no specifically economic alternative to private property in mind, hedid think that rational economic agents could - and would - rectify the antag-onisms it engenders at the political level by means of a social contract. Thus, inhis view, in contrast to Augustine's, "salvation" is possible through human ef-forts - thanks, ironically, to the very circumstance, the predominance of pri-vate volition, that makes it necessary.

48


Had "the true founder of civil society" not succeeded in pri-vately appropriating a piece of land, amour de soi might have sur-vived indefinitely. But with the end of the original state of natureand the ensuing transformation of human beings' interests anddispositions, the demise of self-love became inevitable. Rousseaucalled its successor amour propre, rational egoism. Individualsmoved by amour propre are essentially the idealized agents ofnineteenth- and twentieth-century economic theories, the rationalcalculators emblematic of "the spirit of capitalism." Strictly speak-ing, one need not be a rational egoist in order to count as a rationalagent. But the rational agents who enter into the social contract aremoved by amour propre. In fact, Rousseau's usage effectively con-flates amour propre with (Hobbesian) rational agency. We can, ofcourse, disaggregate these claims conceptually. But it remains truethat the paradigm case of private volition is precisely the rationalegoist whose origins The Second Discourse describes.

It should be noted, finally, that Hobbes's conception of individ-uals' interests is not quite the same as Rousseau's. In Hobbes'sview, individuals have interests in accumulating resources. Butthey also have interests in dominating others and in diminishingthe unrelenting fear that results from recognizing their vulner-ability to mortal assault. People are not just "competitive" butalso "vainglorious" and "diffident." Thus scarcity is less impor-tant in Hobbes's account of the war of all against all than inRousseau's. Its role in the state of nature is to exacerbate antago-nisms and to direct them into struggles over things. But Hobbes'sindividuals would find themselves at war even in a world ofabundance. For what mainly sets them at odds is the fear ofviolent death. In a state of nature, relatively equal and instru-mentally rational agents driven by this fear have an incentive toattack each other preemptively whenever an appropriate occasionarises. Thus any collection of human beings without a sovereignto keep their natures in bounds would evince a generalized dispo-sition to combat which, Hobbes tells us, is precisely what a stateof war is.13

13 "Hereby it is manifest, that during the time men live without a common Powerto keep them all in awe, they are in that condition which is called Warre; andsuch a warre, as is of every man, against every man. For Warre, consisteth notin Battell onely, or the act of fighting; but in a tract of time, wherein the Will tocontend by Battell is sufficiently known; and therefore the notion of Time, is tobe considered in the nature of Warre; as it is in the nature of Weather. For asthe nature of Foule weather, lyeth not in a showre or two of rain; but in an

49

The general will

However, for Rousseau, human beings are not, by nature, diffi-dent or vainglorious or even acquisitive. When they act as if theyare it is because they live in the grip of a civil society based on theprivate ownership of alienable productive assets. Inexorably, indi-viduals in these conditions lock themselves into a condition of gen-eralized conflict. As it was for Hobbes, this situation is dire. "Thehuman race would perish," Rousseau insists, "unless it changedits mode of existence."14

DE FACTO STATES

Rousseau's warning is susceptible to at least two readings. Bothare warranted and perhaps even intended. In The Second Discourse,what I shall call the neo-Hobbesian reading predominates.

The use of Hobbes's name in this context requires qualification.In recent years, a number of writers have deployed game-theoreticconcepts to reconstruct and assess Hobbesian positions.15 Doingso is, of course, anachronistic. But as a way of productively joiningHobbes's concerns with current issues in political philosophy, thefruitfulness of this approach is plain. It is in this spirit, even at therisk of distorting Hobbes's views, that I shall describe a Hobbesianstate of nature as a game.

But what game? There is no uncontroversial answer. However,for our purposes, a deeper understanding of Hobbes matters lessthan a clearer grasp of Rousseau's purchase on the emergence ofthe mentality Hobbes described. Therefore, without any pretenseof getting Hobbes right or even of siding with the stronger case,I will stipulate that a Hobbesian state of nature is a generalizedPrisoner's Dilemma in which individually rational (maximizing)choices lead to outcomes that are worse, in terms of each individ-ual's own interests, than outcomes that might otherwise be

inclination thereto of many dayes together; So the nature of War, consisteth notin actuall fighting; but in the known disposition thereto, during all the timethere is no assurance to the contrary." Leviathan, Book I, Chapter 13.

14 The Social Contract, Book I, Chapter 6.15 Cf., among others, David Gauthier, The Logic of '"Leviathan" (Oxford: Clarendon

Press, 1969); Gregory S. Kavka, Hobbesian Moral and Political Theory (Princeton:Princeton University Press, 1986); and Jean Hampton, Hobbes and the Social Con-tract Tradition (Cambridge: Cambridge University Press, 1986). This genre ofphilosophical reconstruction epitomizes "the standard philosophical interpre-tation" of Hobbes referred to in the Introduction, note 11.

50


Bcooperate defect

cooperate 2nd, 2nd 4th, 1stA

defect 1st, 4th 3rd, 3rd

Figure 1

achieved.16 The situation can be represented as a game with two(or more) players and two moves - "cooperation" and "defec-tion." To cooperate is to refrain from choosing what is individuallybest in order to produce outcomes that are better, given each play-er's private interests, than the outcomes that would result fromstraightforward maximizing choices. Cooperation is self-regulatedcoordination; it entails that some players sometimes not do what isindividually best. To defect is to fail to cooperate - to act, move bymove, with a view toward realizing one's own private interests.For players with private wills competing for scarce resources, it isdesirable that there be cooperation - if only because the continu-ation of the war of all against all necessitates defensive measuresthat deflect efforts away from accumulation. For Hobbes, cooper-ation would also mitigate the pursuit of vainglory and diminishthe anxiety that weighs on each individual in a state of war. But,for each player, it is better if only the other players cooperate. Thus,if there are two players A and B, A is better off if B cooperates andA defects than if both A and B cooperate; and similarly for B. BothA and B are better off, however, if both cooperate than if both de-fect. Finally, A is worst off if A cooperates and B defects; and sim-ilarly for B. Schematically, representing A's choices in order ofpreference on the left, and B's on the right, the payoff matrixshown in Figure 1 obtains. The payoff to A depends on what Bdoes and vice versa. But whatever B does, A is better off defecting.Since the payoffs are symmetrical, whatever A does, B is better offdefecting too. Therefore, if A and B are rational players, both willdefect, and each will do less well than they might have done hadthey cooperated. But they cannot cooperate. Insofar as they are ra-tional, when they confront Prisoner's Dilemma payoffs, they canonly act in ways that produce suboptimal outcomes.

If Hobbes was right, the rational pursuit of private interests inthe absence of political arrangements leads to the virtual certaintyof outright annihilation in a war of all against all, a supremely

16 For a different and probably sounder view, see Hampton, Hobbes and the SocialContract Tradition, passim, but esp. pp. 58-63, 132-138.

51

The general will

suboptimal outcome. In this plain sense, "the human race wouldperish unless it changed its mode of existence/7 Thus a neo-Hobbesian reading would construe Rousseau's warning literally. Ithink there are good reasons to resist this construal, but it is in-structive to see where it leads.

For Hobbes, there is a qualitative gap in individuals' orderingsof outcomes in the Prisoner's Dilemma they confront. Thirdchoices rank somewhat higher than fourth choices, inasmuch asindividuals can at least hope to emerge as victors in a generalizedstate of war but face certain annihilation if they cooperate alone. Toa much greater degree, each player's most preferred outcome, lonedefection, dominates universal cooperation. Lone defectors enjoythe benefits of cooperation without having to pay the costs. But theintensity of these preferences pales before each player's preferencefor universal cooperation over universal defection. Since playerscannot expect to be lone defectors but will instead remain in thewar of all against all if they attempt to bring about what they deemindividually best, it is overwhelmingly in the interest of each ofthem to implement mutual cooperation, their second-best choice.But how can they do so? Individuals who choose to cooperate, solong as they play against rational agents, will find their least pre-ferred outcome realized. Their dilemma is plain: To advance theirown (private) interests, they cannot allow defection to continue,but neither can they act rationally and avoid this result.

To "solve" this otherwise intractable problem, Hobbes wouldhave individuals change the game they play by institutingsovereignty, supreme authority, through a covenant concoctedto be advantageous to each contracting party. The sovereignwould then organize his subjects into a "police" for enforcinghis commands. With enforceable sanctions in place, defection or,as we then should say, disobedience becomes irrational, given in-dividuals' interests. In virtue of being able to issue enforceablecommands, the sovereign coordinates individuals' behaviors -and individuals attain the peace that they cannot achieve directlythrough cooperation.

Although The Second Discourse describes the emergence of an es-sentially Hobbesian state of nature, there is an important differ-ence between Hobbes's story and Rousseau's account of the originof de facto states. Because Hobbes's state of nature is ahistorical,and because he viewed relative equality (not just in physical andmental endowments but also in possessions) as given, Hobbesimagined that the individuals who establish "a common power to

52


hold [themselves] in awe" do so from roughly the same startingpoint. Rousseau, on the other hand, depicted the onset of the warand all against all as a consequence of processes that result in con-siderable inequalities of condition. For Rousseau, therefore, the es-tablishment of de facto sovereignty is not a solution contrived byequals. It is a response to a problem confronted by individuals ofunequal wealth, where the burdens of war weigh most heavily onthose who have the most to lose. States are formed whenever therich are able to convince the poor that it is in their interest to endthe war of all against all by political means. In Rousseau's view, itis the guile of the well-off and the gullibility of the rest that givesus de facto sovereignty.

In The Second Discourse, this trick, foisted on the vast majority, ispromoted by Hobbesian arguments tempered by professions ofconcern for those who are destined to become the principal vic-tims of the new order. Thus Rousseau imagines an individual "en-riched by industry," invoking "the horror of a situation whicharmed them all against each other and made their possessions asburdensome as their needs," entreating his propertyless neigh-bors in these words:

Let us unite . . . in order to protect the weak from oppression, restrainthe ambitious, and assure everyone of possessing what belongs to him.Let us institute rules of justice and peace to which all will be obliged toconform, which will make special exceptions for no one, and which willin some way compensate for the caprices of fortune by subjecting thestrong and the weak to mutual obligations. In short, instead of turningour forces against ourselves, let us gather them into one supreme powerthat governs us according to wise laws, that protects and defends all themembers of the association, repulses common enemies, and maintains usin eternal concord.17

In short, let us establish "a common power to hold us in awe," aHobbesian sovereign, for our mutual advantage.

Rousseau suggests, as Marx would half a century later, that thecitizenship the rich offer the poor masks and reinforces the under-lying inequalities of civil society, and that the institutions thattransform possessions (biens) into property (propriete) institutional-ize the poverty of the asset-poor. But the argument the rich ad-vance, though disingenuous and disadvantageous to "crude andeasily seduced men," is nevertheless sound. The untrammeledreign of private interest in a world of scarcity does threaten even

17 SD, p. 177.

53

The general will

those who have nothing more to lose than their lives, and the es-tablishment of de facto states does remove this mortal danger. IfRousseau's warning that "the human race would perish unless itchanged its mode of existence" is understood in the neo-Hobbesian way, as a prediction of universal physical annihilation,the people of The Second Discourse have indeed contrived asolution.

FREEDOM

The de facto state arises in consequence of a process set in motionby the introduction of private property in alienable productive as-sets. As private property increasingly determines social life, thestate of nature turns into a Hobbesian war of all against all, acondition that everyone has an incentive to transform. By autho-rizing a sovereign to act in their behalf - agreeing to obey his com-mands and to enforce his will against disobedients - individualsconstitute a force sufficiently powerful to ensure the order thatthey all desire.

But what order? For Hobbes, the function of the state is to keephuman nature from destroying humankind. Thus he would havethe state suppress anything that threatens to disrupt the fragilepeace it establishes. Revolutionary challenges to existing regimesare particularly dangerous, but any destabilizing political inter-vention can unleash a course of events likely to devolve into awar of all against all. The order that states exist to protect istherefore the existing order, whatever it may be. Hobbes didvalue freedom in the sense of "negative" liberty, "the absence ofexternal Impediments."18 But, in his view, order takes precedenceover freedom and everything else. For Rousseau, on the otherhand, disorder is not so odious in itself, nor is it inscribed inthe "natural" course of events. Political arrangements serve a dif-ferent purpose.

In Rousseau's view, there exists a fundamental interest thatHobbes failed to acknowledge and that legitimate political in-stitutions must protect. This interest is intimately bound to

18 Leviathan, Chapter 13. This is, roughly, the understanding of freedom that lib-erals would later make their own. I discuss aspects of post-Hobbesian freedomin Liberal Democracy: A Critique of Its Theory (New York: Columbia UniversityPress, 1981), Chapter 2, and Arguing for Socialism, 2nd ed. (London: Verso,1988), Chapter 1.

54


"one's dignity as a man, the rights of humanity and even itsduties/'19 If we focus on this interest and reflect on its impli-cations, a nonliteral - Rousseauean - reading of Rousseau's warn-ing of the impending demise of the human race suggests itself.On this understanding, what is put in jeopardy by a Hobbesianwar of all against all is not human life as such but essentialhumanity.20

What Hobbes ignored and what individuals moved by amourpropre fail to see is what Rousseau proclaims in the first words ofThe Social Contract: that "man is born free."21 Of course, no one is"born free" from "external Impediments." But negative liberty isnot what Rousseau had in mind. What he intended by "freedom"was autonomy, self-determination, "obedience to a law of one'sown making." What is essential to being human, according toRousseau, is the possibility of becoming the master of one's ownprojects and undertakings. For reasons Kant would later explore, itis reasonable to value autonomy absolutely because it is the con-dition for the possibility of moral agency, which alone makes hu-man beings "ends in themselves."22 Autonomy therefore takesprecedence over other interests - including the universal humaninterest in order and the more particular but still pervasive interestin rational accumulation that has developed after the introductionof private property.

Even if the human race could physically survive a war of allagainst all, Rousseau's concern is that essential humanity, auton-omy, cannot survive this condition. "Man is born free, but is ev-erywhere in chains." The humanity of the human race will perishif these chains - made more burdensome by the establishmentof de facto states - extinguish the very hope of genuine self-determination. But this is just where the process set in motion bythe introduction of private property tends. The Second Discourse re-counts the transformation of the human race into a collection ofunfree, heteronomously determined individuals - in Rousseau'sterms, into "slaves." When Rousseau insists that the preservationof life is not a sufficient reason for an individual to enter into the

19 The Social Contract, Book I, Chapter 4.20 See Levine, The Politics of Autonomy, Chapter 1.21 The Social Contract, Book I, Chapter 1.22 Cf. Immanuel Kant, The Foundations of the Metaphysics of Morals, passim, but

esp. Part 2. What follows draws on Levine, The Politics of Autonomy, pp. 20-29.Autonomy is discussed in Chapters 6 and 7 herein.

55

The general will

social contract, he is simply drawing out an implication of thisview.23 If autonomy is the source of all value, life itself is of novalue, except as a condition for exercising autonomous choice.Thus in a war of all against all, where autonomy is at risk of be-coming irretrievably lost, the physical continuation of the humanspecies is not the issue. The issue is humanity itself.

This is why Rousseau insists that autonomy can never be advan-tageously exchanged:

Renouncing one's liberty is renouncing one's dignity as a man, the rightsof humanity and even its duties. There is no possible compensation foranyone who renounces everything. Such a renunciation is incompatiblewith the nature of man. . . . [I]t is a vain and contradictory convention.24

Autonomy is priceless; its value knows no bounds. There is noth-ing, including peace or wealth, that can compensate for its loss.Any social contract that grounds legitimate political institutionsmust therefore respect autonomy. Thus the de jure state is an as-sembly of autonomous agents. In contrast, the de facto states TheSecond Discourse describes, the states whose origins Hobbes ex-plained, are associations of slaves. "The fundamental problem ofpolitical life" is therefore not quite the problem Hobbes engaged:the institution of a regime capable of rendering impossible the warof all against all. It is Hobbes's problem with a very non-Hobbesian proviso added on. Individuals must somehow contrivea way to avoid a war of all against all while retaining or, more pre-cisely, realizing essential autonomy; they must, in short, reconcileautonomy with authority.

So, far from solving this problem, Hobbesian sovereignty exac-erbates it. This is not the place to assess Rousseau's attempt at asolution: his insistence that each individual "place himself and hispowers under the supreme direction of the general will."25 Whatmatters for now is just that "the exercise of the general will" is in-troduced to address a problem private willing generates and inten-

23 See, especially, The Social Contract, Book I, Chapter 4, where Rousseau arguesthat pacts of involuntary slavery, in which individuals conquered in wars agreeto become the slaves of their conquerors in return for their lives, do not estab-lish legitimacy. Arguably, Rousseau's claims for the worthlessness of life in theabsence of freedom are exaggerated for rhetorical effect. But this conclusion lit-erally does follow from his view of autonomy as a necessary condition for thevalue of everything else. See Levine, The Politics of Autonomy, pp. 9-19.

24 The Social Contract, Book I, Chapter 4.25 Ibid., Chapter 6. For an extended discussion of the merits of Rousseau's case,

see Levine, The Politics of Autonomy, passim.

56


sifies. According to Rousseau's story, general will coordinationbecomes possible, desirable, and eventually indispensable only af-ter the private will has transformed the human condition.

EPHEMERAL HOBBESIANISM

The individuals we encounter by the end of The Second Discourseare Hobbesian in the sense that, without a sovereign to constrainthem, they would find themselves in a war of all against all. Buttheir Hobbesianism is triggered by particular social conditions andis destined to pass from the scene. Hobbes thought mutual antag-onism an inevitable consequence of human nature. For Rousseau,human nature can support this state of affairs, but it hardly ne-cessitates it. In fact, essential humanity points in a different direc-tion. As The Social Contract completes the story begun in The SecondDiscourse, the essence Rousseau imputes to human beings finallyworks its effects. Then what had previously been only a notionalpossibility becomes increasingly feasible and even necessary. Thusthe Hobbesianism that is the destination of The Second Discourse isephemeral. It is a "dialectical" moment in the history of the hu-man race, "overcome" when its immanent development undoesthe conditions for its possibility.

It is tempting to conclude that Rousseau's solution to "the fun-damental problem of political life," the subordination of private togeneral interests, is a "self-effacing" Hobbesian solution, accord-ing to which Hobbesian individuals undertake to end the Prison-er's Dilemma they confront by ceasing to advance the privateinterests they acknowledge.26 But this characterization is mislead-ing because Rousseau's political philosophy is non-Hobbesian atits core. Its fundamental commitment to human autonomy is for-eign to Hobbes's conception of political life. True Hobbesianswould deny that "man is born free." For Rousseau, however, au-tonomy is the condition sine qua non of humanity itself.

From a Rousseauean point of view, Hobbes concocted a generaltheory of political life by focusing on only one phase of human"history." This phase, the age of private interests, is a Dark Age forhumankind. It is a time in which the passion to accumulate forgeschains of unfreedom. So pervasive are its effects that human na-ture itself stands increasingly at odds with essential humanity. But

26 On self-effacing theories, see Parfit, Reasons and Persons, pp. 23-24, 43-45. Seealso Joshua Cohen, "Reflections on Rousseau: Autonomy and Democracy,"Philosophy and Public Affairs, vol. 15, no. 1 (1986), pp. 281-284.

57

The general will

as the reign of private property unfolds, a realization of unfreedomdawns. Individuals become aware that the social order they haveconstructed is a mortal threat to their most fundamental interest,to their own autonomy and therefore to essential humanity itself.In The Social Contract, this prise de conscience leads Hobbesian indi-viduals to end the unchallenged reign of private volition and toput the Hobbesian "moment," whose origins and ascendance TheSecond Discourse recounts, definitively to rest.

Since these reflections on The Second Discourse have stayed closeto Rousseau's own meandering thoughts, it will be helpful, byway of conclusion, to draw together some of the principal claimsthat can be teased out of the tale Rousseau concocted in that textand its continuations in The Social Contract. Thus we can say thatRousseau holds:

(1) that what it is to be a human being is to be free (autonomous);and therefore that autonomy takes precedence over all other endsthat human beings might will. Essential autonomy is asserted as ametaphysical claim about the nature of human beings. Autonomyis therefore a condition that all legitimate institutional arrange-ments must accommodate.

(2) that human beings' essential freedom has no direct implica-tions for human psychology. Thus, in humanity's "natural state,"neither private nor general wills govern individuals' lives. Individ-uals are instead moved by amour de soi, "self-love," in just the waythat higher animals are. This "natural state" is a happy conditionand also a stable one.

(3) that this condition is alterable. Unlike higher animals, humanbeings are capable of developing (instrumental^) rational capaci-ties; and, as a matter of fact, these capacities did develop. Amongother things, people discovered ways to augment the productivityof land through agriculture. But the first men and women wereunable to deploy agricultural technologies in the social conditionsof the original state of nature.

(4) that, thanks to a change in "social relations" that mightwell not have occurred, the successful introduction of privateproperty in land, it became socially possible to shift from huntingand gathering to settled agriculture. The implementation of thisadvance in productive technology had momentous consequences:augmenting population and diminishing the stock of land avail-able per person. But scarcity - in a relative, not absolute, sense -intensified as productivity increased. This change in the human

58


condition had profound ramifications for human nature. Self-love gave way to amour propre, and instrumentally rational indi-viduals developed properly private wills. Thus human beingsbecame indomitable accumulators, and the happy and peacefulstate of nature turned increasingly into a Hobbesian war of allagainst all.

(5) that this process is replete with dire consequences for hu-mankind. Among other things, it explains "the development ofinequality among men." More important for a philosophicaltheory of the state, the war of all against all puts essential auton-omy in peril.

(6) that the story The Second Discourse tells therefore culminatesin a condition that makes the establishment of de jure politicalcommunities, resting on a supremacy of the general will, both nec-essary and possible, if the human race is not to forfeit essential hu-manity irretrievably. "The human race would perish if it did notalter its mode of existence" - that is, if it did not become what italways essentially was, a union of autonomous beings.

Thus, after the process begun with the introduction of privateproperty has run its course, each individual has in principletwo wills: a private will and a general will. Each is compatiblewith transhistorical constraints on human psychology, and eachcan be actualized under different social and historical conditions.Paradoxically, the prolonged supremacy of the private will even-tually makes the supremacy of the general will possible; and theessential nature of humanity makes it desirable. What is morallyimperative converges on what is historically feasible. The em-blematic expression of this idea is the social contract itself: theearthly instantiation of the notional possibility of genuine self-determination.

The general will plays no role in The Second Discourse, and its co-gency is assumed in The Social Contract. Thus neither text ade-quately addresses the hesitations of those who find the ideaincoherent or obscure. But The Social Contract does help clarifywhat is intended by the contrast between an individual's generaland private will; and The Second Discourse helps even more. By tell-ing a story about the "origins" of the private will, an account thatemphasizes the historicity of private volition, it works to diminishthe influence of this idea over contemporary thinking. General willcoordination is certainly problematic. But if The Second Discoursestory is at all plausible, its ostensible rival is as well.

59

Chapter 3

Solidarity

To gain a better purchase on what general will coordination is, itwill be helpful to contrast it with a form of social interaction it ap-pears to resemble: acting in solidarity. In our political culture,"solidarity" is often used imprecisely and in a variety of vaguelyrelated senses. However, what follows focuses on only a very fewlegitimate uses of the term. An investigation of solidarity is usefulespecially for the insight it provides into the central political prob-lem Rousseauean political philosophy confronts: the problem oftransforming a world where private interests hold exclusive swayinto a social order in which "the exercise of the general will" playsan important role.

A POINT OF DEPARTURE

It is seldom enlightening to investigate political phenomena by re-flecting on obscure and oracular dicta. But, in this case, there is avenerable and celebrated saying that, though virtually featureless,suggests what is essential. According to the portion of the Mish-nah called the Pirke Aboth, "The Sayings of the Fathers," RabbiHillel is supposed to have said: "If I am not for myself, then whowill be for me? And when I am for myself only, what am I?"1 Theseoften quoted words are sufficiently vague to support a variety ofplausible interpretations. They might be held, for instance, to as-sert the moral importance of self-respect. But in the popular ser-monizing directed to those who seek wisdom in this saying, a

1 Hillel's dictum then continues, irrelevantly for our purpose, "if not now,when?"

60

Solidarity

consensus interpretation has emerged according to which Hillel'swords are taken to recommend what is nowadays understood by"solidarity." This is a defensible and fruitful interpretation, even ifit is not an incontestable rendering of Hillel's intent. It is this un-derstanding that I shall assume in utilizing these ancient words asa point of departure for investigating solidarity.

The question "when I am for myself only, what am I?" seemseasy for readers of Hobbes and Rousseau. I am the sort of indi-vidual who populates a Hobbesian state of nature, a being movedby amour propre. If Rousseau was right, such individuals existonly in determinate social conditions. Perhaps, however, amourpropre is even more historically particular than The SecondDiscourse suggests, enough so to make this answer an anachronis-tic response to Hillel's question. But even if individuals did notalways think of themselves (and others) as bearers of private willsin quite the way Rousseau intended, the behavior the term hascome to designate (or some very close approximation) surelyhas existed since time immemorial. "Being for oneself only" -what the moralists call selfishness - has long been considered amoral flaw. Hillel's words suggest that it is something else be-sides: that being for oneself only is often unwise or, as in aPrisoner's Dilemma, self-defeating. Individuals for whom this de-scription fits do less well for themselves (in just the sense thatthey intend by being for themselves only) than they otherwisemight.

The question "If I am not for myself, then who will be for me?"is more problematic. Construed as a rhetorical question, the an-swer is surely "no one." And the presumed implication is that thisis not a happy state of affairs; that it is better - in general but alsofor me - that others be for me too. Since the "me" in question isplainly intended generically, to designate everyone, the claim ev-idently is that it is good - or, more precisely, beneficial - for peo-ple to be for one another, to help one another, and that it isbeneficial also (apparently not just for independent reasons) thatthe people who help others also help themselves.

If it is fair to read this much into a question asked at a remotetime for an obscure purpose, it is only fair to note an ambiguity inthe inferences we are led to draw. The position Hillel asserts ad-mits of two distinct understandings:

(1) that a world in which people generally help one another ispreferable to a world in which they do not; and that a condition for

61

The general will

bringing such a world into being is that all people be to some de-gree (but not exclusively) for themselves; or

(2) that a world in which people generally help one another ispreferable to a world in which they do not; and that some degreeof being for oneself is an essential part of that state of affairs.

At issue is whether "being for oneself," but not for oneself only,is a means to a desirable end or whether it is also an end in itself,irrespective of its efficacy for promoting the practice of helpingothers. I would venture that, in our political culture, both readingsproperly attach to "solidarity" but that the instrumentalist concep-tion is more fundamental.

Since private volition today seems "normal," it is natural tosuppose that what Hillel's dictum endorses is "enlightened" self-interest - egoism, being for oneself, in conjunction with altruism,being for others. In a world in which it is generally advanta-geous to everyone that people help one another, it is well, Hillelmight be taken to have asserted, that egoism and altruism bejoined.

However, "altruism" is an ambiguous term. Sometimes what-ever is not egoistic is deemed altruistic. I suspect that this usageattests more to the prevalence of the mentality whose origin TheSecond Discourse recounts than to any analytical exigency. In effect,it categorizes motives as ordinary (egoistic) or extraordinary (non-egoistic) and then groups all the extraordinary motives togetherunder the same designation. This construal of altruism evidentlyobliterates important distinctions. In particular, it obscures the dis-tinction I want to draw between general will coordination and act-ing in solidarity.

An egoist seeks to advance his own welfare position. An altruist,then, will be an individual who seeks to enhance the welfare po-sition of at least one other individual. Anyone who acts for a non-welfarist motive would therefore be neither an egoist nor analtruist. The egoism-altruism distinction is not even exhaustivewithin a strictly welfarist frame of reference. What utilitarians call"benevolence," a commitment to enhance the welfare positions ofall relevant others (including one's own), would elude both cate-gorizations.

This understanding of altruism lends precision to Hillel's aph-oristic advice. Of course there is no reason to think that Hillel in-tended either "being for oneself" or "being for others" in awelfarist way or even that the intuitions behind contemporary wel-farism were available to him. But Hillel's dictum is sufficiently in-

62

Solidarity

determinate to accommodate this reading and, in so doing, toserve as a point of departure for contrasting solidarity with generalwill coordination.

WHAT IS SOLIDARITY?

Like most terms of political discourse, "solidarity" is used withenough imprecision to encompass a wide range of meanings. Anyaccount is therefore likely to exclude some legitimate uses of theterm. With this warning in mind, I shall characterize solidarity asa kind of conditional altruism. To be sure, Hillel's words appear toendorse unconditional altruism; he advises us to enhance the wel-fare position of others irrespective of whether or not they seek toadvance our own. However, for beings moved by amour de soi, un-conditional altruism would be difficult to sustain if unrequited. Tobe unconditionally for others when others are for themselves only,to give without receiving back, is likely to cause even the mostdedicated altruist to falter. Therefore lone altruists will likely notpersist in being for others very long. Instead, they will become liketheir neighbors. But then, as Hillel might have asked, "Whatwould they be?" Thus, if only to assure that altruism survives, it isreasonable to infer, according to the spirit if not the letter of Hillel'srhetorical questions, that the disposition he recommends is condi-tional altruism: where individuals are for others if and only if oth-ers are altruists too.

However, solidarity is only one form of conditional altruism.Conditional cooperation is another. A conditional cooperator is forothers if and only if others directly reciprocate. People cooperatewhen, for the sake of some common objective, they forbear fromdoing what they would do were they not cooperating. In a worldin which private volition predominates, the overwhelmingly mostimportant reason to cooperate is, of course, to advance one's ownwelfare. This reason pertains almost universally in the context ofinterdependent actions - where an agent's payoff depends, inpart, on what other agents do. Of particular importance in this re-spect are social interactions that have the structure of Prisoners'Dilemmas. In these circumstances, as we have seen, individualswould do better by cooperating than by not cooperating. But weknow that rational agents may find themselves unable to cooper-ate. Then their only recourse is to alter the conditions under whichthey interact - as individuals in a Hobbesian state of nature dowhen they successfully institute "a common power to hold them

63

The general will

in awe." It is now understood that, when Prisoner's Dilemmagames are iterated indefinitely, individuals can successfully adoptcooperative strategies.2 Being more successful in (iterated) Prison-er's Dilemma games, if they occur at all, they will eventually pre-dominate. In consequence, in Prisoner's Dilemma situations inwhich players interact indefinitely many times, conditional coop-erators will come to find the conditions under which they are dis-posed to cooperate fulfilled.

Conditional cooperators are conditional altruists in the sensethat they sometimes act to enhance the welfare positions ofothers - specifically, those individuals with whom they cooperate.But their altruism redounds directly to their own advantage; they(reasonably) expect those whom they help to help them in return.Genuine solidarity, however, is conditional altruism without re-gard to reciprocal aid. Thus North Americans or Europeans whoact in solidarity with workers in Central America or Africa havelittle reason to think that the individuals whose welfare they seekto enhance will ever be in a position to reciprocate their help.Solidarity therefore designates a more selfless form of other-regarding welfare-enhancing activity than conditional cooperation.

But this form of conditional altruism is not entirely selfless.Those who act in solidarity are indeed "for others" but also "forthemselves." It is not reciprocity but the general practice of help-fulness that these conditional altruists seek to promote; and theybenefit because they, like everyone else, are better off in a world inwhich people help each other than in a world in which they donot.

I argued in Chapter 1 that general interests and private interestssometimes do conflict and that the moral esteem often accordedthose who promote (putative) general interests at the expense ofprivate interests cannot be construed as a (morally praiseworthy)refusal to free-ride. However, the esteem accorded solidarity canbe depicted in these'terms. Helping is relevantly like a publicgood (as opposed to a general interest). A true public good spillsover to everyone regardless of contribution. If unconstrained, pureegoists would therefore be inclined to free-ride; that is, to ob-

2 See Robert Axelrod, The Evolution of Cooperation (New York: Basic Books, 1984).See also Michael Taylor, The Possibility of Cooperation: Studies in Rationality andSocial Change (Cambridge: Cambridge University Press, 1987), and Jon Elster,Ulysses and the Sirens (Cambridge: Cambridge University Press, 1979), pp. 141-146.

Solidarity

tain the benefits without incurring the costs of contributing tothe good's production. Helping is not a pure public good because,in principle, help can be withdrawn from nonhelpers. But it ishard to imagine, in practice, how information about an individu-al's contribution might be disseminated and how the selectivewithdrawal of help might be organized. We can therefore regardhelping as if it were a public good in the strict sense, and acting insolidarity as a contribution toward producing it. Whatever moralpraise attaches to promoting this beneficial practice - or to forbear-ing from free-riding on it - then accounts, in part, for the esteemaccorded helpers. However, I will suggest presently that this ex-planation, though apt, by no means exhausts the insights to begained by reflecting on the moral phenomenology of solidaristicactions.

Conditional altruism can also appear among individuals whosedispositions are better described as charitable than as solidary, in-dividuals who help (particular) others by improving their welfarepositions, conditional on others doing so too but without regard tobeneficial welfare consequences, direct or indirect, that might ac-crue to themselves. Charity can, of course, be given uncondition-ally. But if, in addition to wanting to help particular others,individuals also care about the efficacy of the help they provide,their altruism is more likely to be conditional. For, in many cases,charity will be futile or insignificant if it is not coordinated withthe charity of others, and few people will want to waste their ef-forts. Therefore even selfless givers are likely to be conditional al-truists most of the time. They are not genuinely solidary agents,however, because their generosity is not intended to promote apractice of mutual aid from which they themselves can expect toderive benefits. There are a variety of motives that can lead peopleto perform charitable acts - from the morally base (opportunism,social pressure) to the lofty (outrage at outrageous conditions, awish to identify with "the wretched of the earth," a desire to saveone's soul). There are, correspondingly, a host of reasons for beingaltruistic if and only if others are altruistic too. However, to explorethese situations is to investigate the psychology of charity, notsolidarity.

Solidarity and conditional cooperation are strategies for advanc-ing one's own welfare interests. Of course, in practice much ofwhat passes for charity is undertaken for self-interested reasons

The general will

too. But in principle the charitable strive to realize a differentideal. Charity is by its nature disinterested. Self-interested gener-osities that purport to be charity are, as it were, inauthentic -products of self-deception, disingenuity, or outright hypocrisy.Solidary agents, on the other hand, are not disinterested. They areauthentic creatures of the world The Second Discourse describes. Ifthey are not simple egoists it is because of an enlightened percep-tion of where their private interests lead, not because they aspireto the saintliness of the truly charitable.

Nor do they aspire to Rousseauean citizenship. Thus solidaryagents do not abandon private interests in the way that the indi-viduals of The Social Contract do. To act in solidarity is to seek toadvance the well-being of atomic individuals generally - in order,ultimately, to help oneself. Acting for the sake of a "moral and col-lective body" is, we have seen, a different matter altogether. In cir-cumstances where no one will be for anyone if people are not alsofor themselves, and where being "for oneself only" is generallydetrimental, it is rational even for Hobbesian individuals to be sol-idary agents. If the Hobbesian men and women of The Second Dis-course are not themselves solidary agents, it is because solidarity isdifficult to implement in practice, not because it is unreasonablefor self-interested rational agents to exemplify it.

THE PRACTICE OF SOLIDARITY

Real-world experience attests that solidarity is best promotedthrough its own practice. Curiously, this idea too can be teased outof a venerable and familiar teaching, originally advanced in areligious context but nevertheless relevant to real-world concerns.I have in mind Blaise Pascal's celebrated Wager on the existenceof God.3

Pascal thought belief in God reasonable, so long as there is anyprobability at all that the belief might be true. If I believe in Godand God exists, my "payoff," eternal salvation, is infinitely high; ifI do not believe in God and God exists, my payoff, eternal dam-nation, is infinitely low. On the other hand, if God does not exist,then whether I believe or not, my payoffs will be finite and roughlysimilar. In these circumstances, given the risks attendant on non-

3 Blaise Pascal, Pensees, trans. A. J. Krailsheimer (London: Penguin, 1966), pp.149-155.

66

Solidarity

belief and the possible gains, a reasonable person will believe.4 Butsince the subjective probability that God exists is sure to be low foranyone aware of the challenge posed to religious beliefs by the dis-coveries of the new sciences of nature, belief in God is difficult toinstill and sustain. Moreover, it is psychologically impossible to es-tablish this belief directly, simply by deciding to believe. Thus, forthose who would believe but find themselves unable to do so, Pas-cal advised the adoption of an indirect strategy. He suggested thatthey act as if faith were already established. Pratiquez d'abord: Prac-tice first. Then faith will follow; a faith that is unwarranted by cri-teria of rational or evidential support but that is neverthelessreasonable in light of the expected utilities of belief and nonbelief,respectively.

For fostering solidaristic dispositions, pratiquez d'abord is goodadvice too. Become more solidary by acting in solidarity. Success-ful solidaristic interventions will enhance trust and also transformpreferences. By helping others and being helped by others in turn,individuals become more helpful.

It is worth remarking that, despite his apparent conviction thathuman nature cannot be changed, Hobbes effectively anticipatedPascal's observation. To end the war of all against all, Hobbeswould have individuals change the payoff structures they confrontby authorizing a sovereign to make and enforce rules that imposesanctions on rule violators. But since individuals in a state of na-ture can only invest sovereignty in one of their own, and since in-dividuals are relatively equal in their natural endowments, theindividual chosen to be sovereign cannot have significantly morephysical power than anyone else. By himself, therefore, he is asunable as anyone else to issue commands that others will obey.But a Hobbesian sovereign is not just one individual among many.He is the authorized "representative" of each of his subjects. Inauthorizing the sovereign to act in their behalf, individuals in astate of nature agree, conditionally on others agreeing too, to be-come the sovereign's police: to enforce his commands againstothers as others would against them. In this way, the sovereignacquires the power to issue enforceable commands. Since thisagreement is a covenant, a promise to be executed in perpetuity,

4 Taken at its word, this argument is sound only if, contrary to what nonbelieversbelieve, the alternatives are partitioned in the way Pascal suggests; and if, con-trary even to what many believers think, salvation can be achieved by wageringon God's existence. It is therefore unlikely that many people will be persuadedby Pascal's argument.

67

The general will

individuals, upon entering into the pact, provide themselves theopportunity to form expectations about each other's behaviors.They come to see that the condition upon which they each transferall their rights to a common representative is in fact fulfilled.In this way, even Hobbes's social contract depends on a prioraptitude for cooperation. To be sure, individuals cooperate onlyenough to constitute a credible threat. It is because they cannot co-operate more extensively that sovereignty takes on the characterHobbes ascribes to it and even that sovereignty is necessary at all.Nevertheless, even for Hobbes, trust is indispensable for politicallife. Hobbesian commonwealths are possible only insofar as indi-viduals have more reason to fear the sovereign than they have tofear each other, and their fear of the sovereign is reasonable only tothe extent that they trust each other to execute his commands. Thecovenant Hobbes contrived builds on this trust and extends it.

To launch Hobbes's social contract, conditional cooperation, notsolidarity, is required. But Hobbes's account of the self-fulfillingconsequences of acting as if the conditional altruist's conditionswere fulfilled, even before they are, applies to conditional altruismgenerally and therefore to solidarity as well. So long as some in-dividuals are prepared to "practice first," to act as if it is alreadyreasonable for rational agents, directed by private wills, to be gen-erally helpful, then, other things being equal, it will in fact becomeincreasingly reasonable to be generally helpful. Barring (unlikely)countervailing factors peculiar to particular circumstances, soli-daristic struggles transform people and their preferences, makingthem more inclined to act to enhance each other's well-being. Inshort, the more solidarity there is, the more solidarity there willbe. The larger solidaristic movements are, the faster they grow.

BEYOND SOLIDARITY

Between solidarity and general will coordination there exists thedifference Kant identified between (enlightened) prudential delib-eration and moral agency. The "moral point of view," evident inKant's categorical imperative but emblematic too of non-Kantianmoral philosophy, is the standpoint of agent neutrality or impar-tiality. From this vantage point, individuals decide what to do onthe basis of interests moral personalities share, in disregard of fac-tors that distinguish particular agents from one another. Thus themoral point of view resembles the attitude Rousseau would havecitizens adopt in the assemblies of the people, the perspective of

68

Solidarity

generality. The difference is that Rousseau envisioned general willdeliberations occurring within particular communities, whereasthe moral order that Kant conceived encompasses all rationalagents.5 However, since The Social Contract nowhere addressescriteria for membership in particular communities and since iteffectively regards the just state as a universe unto itself, citizen-ship, as Rousseau conceived it, is tantamount to membership in a(Kantian) "republic" (Reich) of ends. What Kant called a "harmonyof rational wills" describes the self-regulated inner workings of aRousseauean state in just the way that it describes the moralorder.6

For Kant, following Rousseau, moral deliberations differ quali-tatively from prudential considerations, including those thatmotivate solidaristic politics. Thus Rousseau assimilated group in-terests (and coalitions of group interests) to the interests of egoistsin a state of nature, depicting them all as "private." In this sense,Rousseau would oppose solidarity as much as straightforwardegoism. But in The Social Contract, where the affinity with Kant'sdisparagement of prudential calculation is most pronounced,Rousseau's aim was to demonstrate the coherence of the idea of ajust state and its possible applicability. He sought to defend a time-less, Platonic ideal. But actual politics, as Rousseau knew betterthan anyone, is inextricably historical and something less thanideal.7 In the real world of politics, solidarity and morality do notstand quite so far apart as they do in the Platonic realm of The So-cial Contract.

In reflecting on the connection between solidarity and the generalwill, it will be useful to take the role the term plays in our politicalculture seriously. In contemporary political discourse, "solidarity"has an oppositional flavor. Historically, the term entered politicaldiscussion through the labor movement - in the course of an "in-dustrial revolution" that brought large numbers of people, fromdiverse backgrounds, into a common class position. Solidaritycontinues to bear the mark of its origin. In rhetorical contexts and

5 Kant's view of practical reason and therefore of rational agency is not, likeHobbes's, strictly instrumental. For Kant, reason does not prescribe particularends, but it does proscribe those that fail to meet the test of universalizability or,equivalently in Kant's view, that fail to respect persons as ends-in-themselves.On differences between Kantian and Hobbesian views of rational agency, seeLevine, Liberal Democracy, Chapter 4.

6 This claim is defended expressly in Levine, The End of the State, Chapter 1.7 Cf. Levine, The End of the State, Chapter 2, and The Politics of Autonomy, Part 2.

69

The general will

as a term of art, it is much more frequently invoked by militants incollective struggles of a progressive character than by the Right orby politicians executing normal administrative operations.8 Virtu-ally everywhere in the world today, there is more cooperationamong elites than among dominated peoples. But solidarity re-mains largely the province of subordinate groups. At least it is thissort of solidarity that inspires such ideals as "solidarity forever"and that elicits something close to the sentiment Kant, in reflect-ing on "the good will," called "reverence" (Achtung).9

I have already suggested that the (moral) esteem accorded soli-darity derives, in part, from the realization that solidary agentscontribute toward the promotion of a public good, the practice ofhelping others. But there is also, I think, another dimension to theesteem accorded solidarity.

It is instructive to follow Kant's lead by reflecting on the fact thatsolidarity, especially when practiced by subordinated individualsor groups, typically does elicit a sentiment akin to what "rever-ence" seems to denote, even among "impartial observers" uncom-mitted to the particular objectives of the individuals or groupsacting in solidarity. In investigating the conditions for the possi-bility of the good will, Kant used the phenomenology of moral ex-perience as a basis for reflection.10 This methodology is seldomdeployed today by moral philosophers, and it is surely mootwhether anything reliable about morality can be learned from thesubjective experiences of (potentially) moral agents. But the factthat the sentiments morality and solidarity induce are similar isat least suggestive. So too is the realization that general will coor-dination at the level of "the whole community" often seemsunattractive even when it may be feasible. These pieces of phe-nomenological "data" suggest a hypothesis that bears further con-

8 Of course, class-conscious elites can and do frequently act in solidarity withother elite individuals and class fractions. When they do, it is not inconceivablethat they would declare themselves to be acting in solidarity. However, this us-age is unlikely, especially in times and places where the labor movement hasnot already brought the expression into wide currency.

9 See Kant, The Foundations of the Metaphysics of Morals, Part 1.10 For example, in Part l of The Foundations of the Metaphysics of Morals, Kant sug-

gested that the feeling of reverence that the moral law inspires is the source ofits imperatival force - rendering the moral law something to do rather thanmerely to contemplate. Thus the conclusion that morality is a system of cate-gorical imperatives, to the extent that it depends upon the "regressive analy-sis" of moral experience Kant provides in the Foundations, rests on an accountof the feeling the moral law elicits in human beings.

70

Solidarity

sideration: that in the actual world genuine impartiality is notalways desirable and that, paradoxically, in existing or imaginablepolitical contexts respect for moral personality may actually be bet-ter promoted by solidarity than by "the exercise of the generalwill." Insofar as our sensibilities are informative, we recognizethat, in a world of moral recalcitrants, replete with obstacles in theway of the vision Rousseau and Kant elaborated, solidarity maysometimes be preferable even to morality itself. Thus it may some-times be well for solidarity to "stand in" for universal morality inthose less than ideal circumstances where, if worthwhile politicalaims are to be realized, real generality cannot stand in for itself.11

That there is a problem in reconciling political exigencies withprivate morality has been acknowledged at least since Machiavelli.Perhaps the best known modern formulation of the dilemma isMax Weber's observation that whoever would choose "politics as avocation" must frequently confront a sometimes irreconcilabletension between an "ethic of responsibility" and an "ethic of ulti-mate ends."12 In Weber's view, this conflict is inevitable if onlybecause responsible political action sometimes requires using per-sons as means, in violation of the demands of the (Kantian) "ethicof ultimate ends" that he endorsed. The tension between theseethics takes on a qualitatively greater dimension for political actorsintent not just on muddling through in the manner of politiciansin our political culture but on implementing a definite end inview - the Jacobins' republic of virtue, for instance, or Marxiancommunism.13 In addition, since revolutionary politics inevitablyinvolves extraordinary political interventions at odds with the de-mands of a universal morality, the tension is even more acute forpolitical visionaries who are also revolutionaries. This moral di-lemma carries over to historical assessments. Since there is seldomagreement about the desirability of the visions motivating political

11 The idea behind this contention is not peculiar to Kantian understandings ofwhat morality is. The same idea can be expressed, for example, in utilitarianterms. Thus it might be held that the way to advance well-being overall is not,as utilitarianism maintains for ideal cases, to seek to advance individuals' well-beings, counting all individuals' utilities equally, but instead to seek to ad-vance the well-being of individuals in some groups, if need be at the expenseof individuals in other social categories.

12 See Max Weber, "Politics as a Vocation," in H. H. Gerth and C. Wright Mills(eds.), From Max Weber: Essays in Sociology (New York: Oxford University Press,1958).

13 Cf. Steven Lukes, Marxism and Morality (Oxford: Oxford University Press,1985), Chapter 6.

71

The general will

actors, even in retrospect, sympathy for their aims will not alwaysmitigate judgments of their means. And, of course, condemna-tions of revolutionaries' ends abound as well. Nevertheless, dis-plays of solidarity, insofar as they can be distinguished fromcontentious aims and dubious means, remain remarkably immunefrom reproachful judgment. The otherwise unavoidable tensionWeber identified, even intensified to the extreme in revolutionarycontexts, seems to pale in the face of unequivocal manifestations ofgenuine solidarity.

In this respect, solidaristic politics differs instructively fromother means to (putatively) desirable ends that can also divergefrom the requirements of an ideal morality. Imagine a consensusfavoring peace in a world where relations between states (or indi-viduals) fall short of this ideal. Pacifists would have persons act asif the desired end were already established, the better to bring itabout. Nonpacifists, though they may acknowledge a presump-tion in favor of pacifist means, nevertheless maintain that violencemay sometimes be necessary for advancing the ideal they and thepacifists share. They might argue, as Sartre did, that, in some cir-cumstances, a person who acts as if peace were already attained isunwittingly complicitous in the violence pacifism is intended toredress.14 Or imagine individuals who value the emancipatory ef-fects of liberal tolerance in a genuine "marketplace of ideas" butfind themselves in a world in which formal tolerance functions"repressively," as Herbert Marcuse famously maintained - bymarginalizing opinions genuinely oppositional to a status quovery much in need of transformation.15 Civil libertarians, by anal-ogy with pacifists, might hold that the best way to bring about agenuine marketplace of ideas in these circumstances would be to

14 Cf. Jean-Paul Sartre's Preface to Frantz Fannon, Les damnees de la terre (Paris:Maspero, 1961); English translation by Constance Farrington, The Wretched ofthe Earth (New York: Grove Press, 1963).

15 Herbert Marcuse, "Repressive Tolerance," in R. P. Wolff, H. Marcuse, and B.Moore, A Critique of Pure Tolerance (Boston: Beacon Press, 1965). Marcuse didnot literally mean that tolerance was repressive but that, in societies relevantlylike mid- and late-twentieth-century America, it functioned in the way genuinerepression does - to enforce conformity to an oppressive social order. It is notentirely clear why Marcuse thought liberal tolerance had, as he put it, "turnedinto its opposite." Its practice in a world rife with economic inequalities is partof the story, but Marcuse also seems to have believed that the nature of com-munications technology in a television age also plays a role - by diminishingpeoples' moral and intellectual capacities for functioning as competent con-sumers in a marketplace of ideas.

72

Solidarity

act as if the ideal were already achieved. Marcuse disagreed. LikeSartre on violence, he acknowledged a presumption against inter-fering with speech but insisted that intolerance - toward positionsthat undergird the system that makes formal tolerance a sham -may sometimes advance the ideals tolerance ostensibly promotes.In each case, there is a consensus on ends and a disagreement overmeans. To adjudicate these disputes, it is necessary to assess therelevant evidence and to speculate about the likely outcomes of dif-ferent social practices. Those who are persuaded by Sartre or Mar-cuse plainly face the dilemma of reconciling what they think rightwith what they think they ought to do. Pacifists and civil libertar-ians also face a potential difficulty insofar as they concede thattheir positions depend on empirical considerations that mightsupport conclusions different from the ones they draw.16 However,in anticipating the ends they and their opponents share, pacifistsand civil libertarians do sometimes evoke emotions similar tothose elicited by displays of solidarity. They too anticipate a gen-uinely moral order - with appropriate consequences at the level ofmoral phenomenology. But because their political strategies are soeminently contestable, the sentiments they evoke are frequentlymitigated and sometimes take on a meretricious aspect. In con-trast, for most people most of the time, the solidarity of the op-pressed seems unequivocally worthy of esteem. Conceived as ameans for ending domination, it is also an anticipation of a worldwithout domination.

SOLIDARITY FOREVER?

Solidarity and general will coordination are distinct phenomena.But phenomenologically, the solidarity of "the wretched of theearth" foreshadows a time when the actual world approximatesthe ideal, when the real impartiality of general will deliberation,the antithesis of a generalized "war of all against all," can take itsrightful place in human affairs. I have used these phenomenolog-ical data, in conjunction with the feeling of apprehension that gen-eral will coordination sometimes evokes, to suggest that, in aworld still recalcitrant to the moral point of view, the ends of mo-rality may be better served by solidaristic politics than by the adop-tion of a more "general" perspective. This contention is consistent

16 They could avoid this problem by maintaining that violence or intolerance iscategorically proscribed regardless of its consequences. But then they wouldconfront the far greater problem of justifying these prohibitions.

73

The general will

with the widely acknowledged idea that political exigencies cansometimes force responsible political actors to diverge from thestrict requirements of private morality. It is, in effect, a partial elab-oration of that idea.

Will this conclusion always pertain, or will solidarity eventuallyundo the conditions that now make it both necessary and estima-ble? Traditional procommunists inclined toward the latter view.But the conviction that solidarity can "wither away" entirely is em-inently con testable.17 We will find that even under communismthe conditions that make solidarity indispensable cannot be alto-gether "superseded." Hillel's words therefore apply to communistmen and women, just as they do to us. But we will also find thatthe generalized practice of helpfulness that Hillel recommendsand the attention to individual well-being that I have imputed tohim will pale in importance in comparison with their current ur-gency. In other words, solidarity will wither away somewhat. Thusthe traditional view, despite its evident exaggeration, containsmore than a kernel of truth.


74

Chapter 4

Democracy in the Age of States

I have suggested that the "state" envisioned in The Social Contractis not a state at all but an internally coordinated "republic ofends." However, in this respect as in so many others, Rousseau isrevealingly equivocal. "The moral and collective body" foundedby the social contract does sometimes coordinate individuals' be-haviors through force and does concentrate the means of coercioninto a single institutional nexus. If the state is conceived in thestandard way as an institutionalized monopoly of the means of vi-olence, the social contract therefore does establish a state or, atleast, a social order superintended by institutions with statelikeproperties.

However, the ambiguity that surrounds the state of The SocialContract is submerged during what I shall call the Age of States,the time when states that accord no place for "the exercise of thegeneral will" reign everywhere on earth. These are the states Hob-bes had individuals establish in order to end the war of all againstall, the states whose "origins" Rousseau recounted in The SecondDiscourse. For both Hobbes and Rousseau, the opposition of pri-vate wills makes states necessary. But states that only coordinateprivate wills will not solve the "fundamental problem of politicallife." They will not "defend and protect with all common forcesthe person and goods of each associate" while assuring that eachindividual "obeys only himself and remains as free as before."1

Hence, for Rousseau, unlike Hobbes, the state of The Second Dis-course must give way to the state of The Social Contract.

There is, in addition, a related complaint against existing po-litical forms that is implicit in some of Rousseau's descriptionsof legitimate authority relations. This line of criticism is more

1 The Social Contract, Book I, Chapter 6.

75

The general will

pronounced, however, in Marx's writings on politics, especiallythose composed after the defeat of the Paris Commune. In thisstrain of Marxist political theory, the focus is not on autonomy perse but on democracy. States and the word system of nation-statesare faulted for failing to implement democratic values. This short-coming is not a contingent feature of particular de facto states. Onthis view, obstacles to the fullest possible realization of democracyare inherent in the very idea of the state as a form of political or-ganization. To be sure, lack of democracy can be ameliorated with-out abandoning the state altogether. Indeed, what Marxistsconsider the "last state," the state that superintends the transitionto communist statelessness, is distinguished from other states pre-cisely by its radically democratic character.2 But if this position issustained, democracy cannot be realized fully until the Age ofStates has finally passed from the scene.

DEMOCRACY

The autonomy Rousseau faulted de facto states for thwarting is aproperty of individuals. If we resume the atomic individualist pic-ture and conceive individuals as independent centers of volition,then an individual atom is autonomous if it determines its own ve-locity and direction. It is not enough that the atom's inertial mo-tion be unblocked. The freedom that matters for Rousseau is self-mastery, the setting of ends before oneself. We are autonomous orfree to the degree that we are self-directing. The idea of self-direction links autonomy with democracy. It is even fair to say thatautonomy and democracy represent the same idea - from thestandpoint of the individual and of the whole community, respec-tively. Since this affinity is obscured by contemporary and histor-ical uses of the term, it will be well to reflect briefly on what"democracy" has meant to its defenders and detractors. In doingso, the connection with autonomy will become evident. More im-portantly, we will gain a better purchase on the intuition that mo-tivates these values.

From the time of the ancient Greeks until roughly two hundredyears ago, "democracy" was universally despised. Nowadays, it isendorsed by nearly everyone - from celebrants of the systems ofgovernance in place in advanced capitalist countries to defenders


76


of popular democracy in the Third World and in the countries thatused to form the "socialist bloc." In short, the term has become"essentially contested," adopted by all sides, even as disagree-ments abound about its meaning. Support for the term, if not theidea, is so widespread that the core principle genuine democratshave always promoted, despite their differences, is now at somerisk of being lost in the morass. This situation is evidence for animportant transformation in human history. The universal supportnow accorded "democracy" attests to the entry of the demos, the(ordinary) people, into the political arena. It represents a generalrecognition that the claim of "we the people" to rule ourselves islegitimate and incontrovertible.

In earlier times, "democracy" meant rule by the demos. This wasthe democracy that political philosophers, if not the people them-selves, almost universally despised. Today it is more apt toconsider a political community democratic if its institutions imple-ment "peoples' power" in some procedural sense.

It has become commonplace for mainstream democratic theo-rists in the West and, more recently, throughout the entire worldto identify "democracy" with the governmental forms in place inthe so-called Western democracies.3 For reasons that will becomeevident shortly, this understanding is both tendentious and inap-propriate if "democracy" is to retain even a tenuous connectionwith its traditional meaning. A more plausible and historicallygrounded view of democracy is implicit in the theory of collectivechoice, in welfare economics, and in some contemporary politicalphilosophy. A political community is democratic, on this view, ifits collective choices are functions of the choices of its citizens,where citizens' choices, expressed in voting, count equally andwhere rights of citizenship are enjoyed regardless of such irrele-vancies as class, gender, ethnicity, or race. In other words, democ-racy is conceived as a mechanism for making social choices - forlegislating - that can be modeled by a "device" that combines in-dividuals' votes in an appropriate way. Thus we can depict thevotes of individuals 1, 2, . . ., n as inputs and the social choice asan output, and then construct a collective choice rule for aggregat-ing the votes as pictured in Figure 2. If voting rights are distrib-uted in such a way that no one is excluded for irrelevant reasons,

3 See Joseph Schumpeter's Capitalism, Socialism, Democracy (New York: Harper andRow, 1942), the locus classicus of this definitional move.

77

The general will

Collective choice rule Social choice

Figure 2

and if the collective choice rule is properly responsive to individ-uals' choices,4 we have a genuine democracy. This definition isroughly coextensive with the older idea of democracy, given thesize and composition of the demos in existing political communi-ties. To the extent that the two definitions overlap extensionally,the more modern understanding, in suggesting a picture of uni-versal political equality, obscures what the older idea takes forgranted: the existence of real social divisions in political commu-nities, and the intrinsic conflict in all socially divided societies be-tween popular masses and social elites. It will be important to myargument eventually to bring the class content of democracy backin. However, for now, if only to accord with how the word is now-adays used, it will suffice to think of "democracy" as rule by thepeople tout court.

What democrats value is collective self-mastery - rule by thepeople, not by elites. Rule by elites is a form of collective heteron-omous determination. In the spirit of Rousseau's definition, it is"slavery," a form of governance in which the other rules. Similarly,democracy is autonomy at the level of political communities. It isautonomy for "the moral and collective body" established by thesocial contract. Democracy is rule of the people by the people -and, presumably also, for want of a sufficient reason otherwise, forthe people.

Where freedom is understood as the absence of (coercive) re-straint, unfreedom may admit of degrees in the sense that onecan be more or less constrained from doing what one wants. Butone can be absolutely free. If freedom is held, in addition, to in-volve an agent's ability to do what he or she wants, then freedom

4 Collective choice theorists represent responsiveness to individuals' choices bystipulating conditions on collective choice rules that articulate "democratic" in-tuitions. It has been known since Kenneth Arrow's pathbreaking investigationsin the early 1950s that impossibility results can be generated even from veryweak sets of conditions. See Kenneth Arrow, Social Choice and Individual Values(New York: Wiley, 1951; 2nd ed., 1964). Recent, comprehensive accounts of col-lective choice theory include Thomas Schwartz, The Logic of Collective Choice(New York: Columbia University Press, 1986), and Norman J. Schofield, SocialChoice and Democracy (Berlin: Springer-Verlag, 1985).

78


too admits of degrees. Finally, autonomy, the ability to set endsbefore oneself, also admits of degrees. Autonomy is always con-ditioned by circumstances. We can make our own way, as itwere, but not, to paraphrase an expression of Marx's, in circum-stances of our own choosing. Among other things, institutional ar-rangements affect the level of autonomy individuals are able torealize.5

It is plain too that democracy, autonomy writ large, admits ofdegrees. It is not just that individuals collectively are unable to settheir own course apart from the circumstances they confront. Con-ditions internal to political communities affect the extent to whichthe people are able genuinely to rule themselves. The contentionhere is that the state itself affects self-rule detrimentally; thatso long as states coexist with material conditions compatible withstatelessness, economic, social, and political arrangements can bemore or less democratic, but always less democratic than theycould otherwise be. States are obstacles in the way of realizingdemocratic values. In the Age of States, the people in power is, atbest, a regulative ideal that cannot be very well approximated. Tomake democracy as real as humanly possible, it will be necessary,finally, to end the state.

I will go on to advance a number of considerations in supportof this claim. But it will become plain that the view of democ-racy it supposes - democracy as peoples' power - is misleading, notonly because it obscures the class nature of human societies inthe Age of States but also for its ahistoricity. In Chapter 5 I willargue, on broadly Marxist grounds, that static representations ofdemocracy are always, in some measure, deceptive. In the realworld of politics, in contrast to the Platonic realm of The SocialContract, a commitment to democracy is a commitment to theprocess of its enhancement: to democratization. Rousseau is equiv-ocal on this point; the strain of Marxist theory that envisionsthe end of the state under communism is more straightforward.From that vantage point, in any conceivable political formationthis side of communism, democracy, by its nature, is always inmovement, and its trajectory is what is especially of concern.Properly understood, what democrats favor - now and well intoeven the most distant future - is not quite the people in power,a static condition, but the process of enhancing power to thepeople.

5 These claims are defended expressly in Levine, Arguing for Socialism, Chapter 1.

79

The general will

MODELING DEMOCRACY

Contemporary democratic theorists typically ignore democracy'sdynamic aspect, just as they ignore the fact that, historically, de-mocracy meant rule by the demos. Collective choice models there-fore serve them well. If "democracy" is construed synchronicallyand understood to mean rule by an (undifferentiated) citizenry, itis fair to model it, as in Figure 2, by a flow diagram - with indi-viduals' choices as inputs and the social choice as the output -provided citizenship rights are properly distributed and fairlyrepresented, and provided the aggregating device is positively re-sponsive to individuals' choices and to nothing else.

All models abstract from features of the phenomena they repre-sent. However, the collective choice model favored by many con-temporary democratic theorists is excessively abstract. To itsdetriment, it incorporates a number of assumptions, endemic inmany strains of political theory and throughout the dominant po-litical culture, that more careful scrutiny cannot support. It there-fore blocks insight into the nature of democracy, even allowingthat democracy is properly construed synchronically and from aperspective that excludes class divisions. More importantly for ourpurposes, it impedes understanding the roles the state can play infacilitating or impeding the implementation of democratic values.

The most misleading of the assumptions that this model sup-ports is the idea that voting - in particular, majority rule voting -unproblematically implements collective self-rule. Majority rulevoting plainly does what its proponents intend in a wide range ofcircumstances. But the issue is more complicated than may at firstappear.

The temptation to identify majority rule with democracy is easyto explain. When the majority rules, social choices reflect individ-uals' choices in the sense that the outcome selected is preferred bymore voters than any other alternative in contention. Of course,the most preferred alternative would also win in any voting sys-tem in which a majority greater than 50 percent plus 1 is decisive,up to and including unanimity. But, in cases where more than sim-ple numerical majorities are required, the method of voting is bi-ased against new enactments; a minority can block any measure.Where unanimity is required, a single individual can stop any pro-posed enactment! In contrast, the method of majority rule intro-duces no bias for or against the status quo. Of all conceivablevoting schemes, it therefore seems best able to implement the idea

80


that the people should rule themselves. However, this conclusionholds only if it is assumed that an individual's vote represents agenuine choice that warrants recognition as a proper input to a col-lective choice rule, and if it is assumed that majority rule votingaggregates these choices in such a way that the outcome actually isdetermined exclusively by the choices individuals make. Neitherassumption can be sustained.

The case for majority rule voting is particularly compelling if itis also assumed that interests are private. Individuals' votes thenregister preferences for alternative, feasible outcomes. On this as-sumption, majority rule voting can be defended on grounds ofboth fairness and welfare. The procedure is fair in the sense thatindividuals' preferences are equally represented in the output.And it is welfare-promoting because each individual does as wellas possible, given the constraint that other individuals too areseeking to do as well for themselves as they can - in a worldwhere wills are radically independent of each other and where col-lective choices must be made.

Rousseau too advocated the method of majority rule,6 but not asa mechanism for aggregating private wills. A state where the ma-jority rules but where votes represent private interests would be,for him, at best a benevolent form of slavery. Arguably such a stateis preferable to the regimes in place in existing states. It would atleast exhibit greater sensitivity to the (private) interests of its sub-jects. But it is still a creature of the Age of States.

Implicitly, Rousseauean political philosophy therefore does faultthe identification of democracy with democratic collective choice.For Rousseau, the method of majority rule is necessary but notsufficient for collective self-mastery - still conceived statically andas if no fundamental social divisions obstructed equal membershipin "the whole community." For the people to rule themselves, it isalso necessary that the choices aggregated be of the right kind:specifically, that they represent judgments about what is best for"the moral and collective body" of which individual voters in dejure states are indivisible parts. For Rousseau, therefore, individ-uals do not ipso facto rule themselves whenever they arrive at so-cial choices by paradigmatically democratic procedures.

However, it is not necessary to invoke Rousseauean views aboutthe general will to arrive at this result. A similar conclusion can be

6 See The Social Contract, Book IV, Chapter 2.

8 l

The general will

reached from a perspective more consonant with mainstreamthinking about political legitimacy.

We can define imprudence as a failure to recognize or act upontrue (private) interests. There are, we know, many reasons why anagent in a particular choice situation might act imprudently. Indi-viduals can fail to reflect adequately on the alternatives in conten-tion or can reflect on the basis of false or inadequate informationor, for any number of other reasons, can hold false beliefs that im-pede rational choice. When choices represent imprudent prefer-ences, the social choice that results from combining them is notunproblematically an expression of individuals' interests. In thesecircumstances, the rationale for according moral weight to the ag-gregation of individuals' preferences fails to obtain, and the intu-ition that joins majority rule voting with democracy is impugned.

Even conceding that there is reason to combine preferences,does it follow that the best way - or even a good way - to do so isto count votes?7 Following the long-standing practice of welfareeconomists, collective choice theorists typically construe the ob-jects of individual and social choice as social states, exhaustive andmutually exclusive sets of (already individuated) items that consti-tute the context of choice. When collective choice models representvoting, alternatives in elections are conceived as social states. But,as such, they are not well partitioned for realizing democraticvalues. Welfare economists work with indifference curves that rep-resent individuals' preferences as sets of trade-offs between con-stituents of commodity bundles. Preferences are expressed by therelative prices individuals will pay for particular goods. But thispicture does not carry over well to interpretations of collectivechoice rules as voting systems. When individuals vote, they do notchoose among social states that can be decomposed and reassem-bled at will. In elections, alternatives are, so to speak, prepack-aged, and trade-offs are generally impossible. Voters are thereforenot able to choose among social states generally but are forced toselect from among only a few - having exercised little, if any, con-trol over their constitution.

This situation impugns the democratic character of many votingsystems. Let a and b be social states that voters will decide be-tween - say, by the method of majority rule. Even allowing thateach voter will have a clear preference, say, for a over b, the voter

7 Cf. John Burnheim, Is Democracy Possible? (Berkeley and Los Angeles: Universityof California Press, 1985), pp. 82-96.

82


will often prefer aspects of the less preferred social state b and mayalso be indifferent to much that a and b contain. The voter's pref-erence for a over b will then represent a preference all things con-sidered - for social states that are antecedently partitioned andpresented for social choice. The individual will prefer a to b on bal-ance and because a choice must be made, but that choice is not anunequivocal expression of the preferences that motivate it. In ide-alized economic interactions, on the other hand, agents partitionalternatives as they please. They set ends before themselves. Inthis sense, they enjoy a degree of autonomy that is elusive even intheory in electoral contexts.

There is yet another, more subtle, disanalogy. Indifferencecurves model individual-level cost-benefit analyses. Taking theirbeliefs and desires as given, individuals compute their own idealcommodity bundles by calculating how much of a particular goodthey would be willing to sacrifice for others. In contrast, voting istypically costless or very nearly so. Indeed, a motive for voting forone alternative over another is sometimes to help shift costs ontoothers. Therefore the information a vote conveys is different fromthe information conveyed by an agent's idealized economic be-havior. Through voting, we can discover which, among anteced-ently given social states, an individual prefers, taking all relevantconsiderations into account. In the cases economists regard as par-adigmatic, we learn what, given only their budget constraints, in-dividuals actually prefer.

There is an additional difficulty, finally, for anyone who wouldassume that votes generally represent preferences. In actual votingsituations, individuals often have incentives not to vote for the(prepartitioned) social state that they prefer on balance but insteadto misrepresent their preferences for strategic reasons. Supposethat a, b, and c are candidates in an election, and that individual iprefers a to b and b to c. Suppose too that i believes that a's chancesof winning are slight, whereas the contest between b and c is likelyto be so close that a single vote could be decisive. Suppose, finally,that i strongly prefers b to c. In these circumstances, i will havegood reason to vote for b over a, despite the fact the / prefers a to b.If i votes for a, the probability that c, the least preferred alternative,will win increases. If i votes for b, however, it is more likely that bwill win. In this case, one serves one's preferences best by not vot-ing for the alternative one most prefers. Incentives for strategicallymisrepresenting preferences increase as voting communities de-velop relatively stable populations. Thus i may prefer a to b and

83

The general will

even believe that a can win over b but will nevertheless vote for bover a in order to induce someone else, j , who strongly prefers b toa, to vote with i at some future time for y over z, a matter aboutwhich i cares more. As we move from highly idealized represen-tations of voting situations to real-world voting behavior, mecha-nisms inducing strategic misrepresentations of preferences arelikely to become increasingly important. But even in the most ide-alized cases, strategic interactions among voters are likely to causeelectoral systems to fail to represent individuals' preferences in theway that market transactions represent consumers' choices in themodels welfare economists propose.

Markets and electoral systems, conceived abstractly, are each in-terpretations of the "device" pictured in Figure 2. The difference isthat in markets individuals do as they please, subject to budgetconstraints and the exigency of interacting with others throughmarket relations, whereas in voting the majority determines theminority. But at a slightly less abstract level, the analogy betweenmarkets and voting breaks down. In economic models of marketarrangements, individuals order all the items in a particular choicecontext according to the prices they are willing to pay. Willingnessto pay is therefore a reasonable proxy for preference or, more pre-cisely, for preference subject to budget constraints. In voting, how-ever, because items are typically bundled into social states in waysthat lie outside a voter's control, because choices generally do notcome with "prices" attached, and because there are so many in-centives to misrepresent preferences strategically, voters' choicesare much less adequate proxies. There is therefore good reason toquestion the idea that majority rule voting combines individuals'preferences fairly and in a welfare-enhancing way even if, paceRousseau, we consider these objectives desiderata in their ownright.

Collective choice theorists have raised doubts mainly about collec-tive choice rules, supposing in effect that the inputs collectivechoice rules combine are unproblematic. They have focused on therationality of devices for combining individuals' choices, under-standing rationality as a weak (or "thin") normative standard -amounting to consistency. In general, the problem collectivechoice theorists since Arrow have identified is that collectivechoice rules that satisfy minimal and ostensibly unexceptionableconditions can take consistent orderings of alternative social statesas inputs and still generate inconsistent outputs. Inconsistencies

84


are manifest as cyclical or intransitive orderings. Given the intui-tions that drive democratic theory, inconsistencies are clearly un-desirable. If i prefers a to b and b to c but c to a, in violation of thetransitivity requirement, / is literally unable either to choose or toact exclusively in accord with his preferences when confrontedwith a choice between a, b, and c. A similar problem confrontspublic officials presented with inconsistent social orderings gen-erated from sets of consistent individual orderings. Inconsistency,then, is a severe, albeit a minimalist, reproach to both individualsand societies.

The objections raised here, however, assume a "thicker" stan-dard of rationality. I have faulted voting systems for admitting andeven encouraging imprudent choices and for failing to ensure thatvoters' choices, their "revealed preferences," coincide with thepreferences, prudent or not, that they actually hold.

The problems collective choice theorists have identified withaggregating mechanisms are relevant to political arguments thatdepend on the thin rationality of voting systems. They oughttherefore to be of grave concern to anyone intent on providing lib-eral democracy - or relevantly similar forms of the state - withjustifying theories.8 But the problems collective choice theoristsraise, serious as they are, arise only as a theoretical limit on ag-gregating choices. In real-world voting situations, paradoxicaleffects seldom, if ever, emerge. The problems with voting in ourpolitical culture therefore have less to do with the difficultiescollective choice theorists focus upon than with what collec-tive choice theory - with its penchant for abstraction - tends toobscure.

DEMOCRACY IN A WORLD OF STATES

States, again, exercise supreme authority over particular territoriesor populations. In a world divided into de facto states, politicalauthority is centralized intranationally and divided into distinct(though possibly interdependent) units internationally. Withineach unit, authority is supreme. Between units, no authority rela-tions exist at all - though there may be (largely) unenforceablerules that regulate the interactions of sovereign entities. For dem-ocrats, the division of the world into nation-states is, on balance, amisfortune because, at the international level too, states are obsta-

8 Cf. Levine, Liberal Democracy, Chapter 3.

85

The general will

cles to collective self-rule. This problem was not salient at the timeof the formation of the modern state system. But today it has be-come unavoidable.

As Marx and Engels insisted in The Communist Manifesto, thenation-state was instrumental for the development of nationaleconomies. In the period of its inception, as feudal solidaritiesgave way to market arrangements, decentralized feudal arrange-ments became "fetters" on economic development. Thus, as na-tional economies formed, human populations sorted themselvesout into discrete and unitary communities, a fact registered at thepolitical level by the development of a unitary nexus of state insti-tutions. Nation-states would, from time to time, interact - in dis-putes over territory or over commercial or security interests. Therewas diplomacy and war. But these "affairs of state" were marginalto political life. Politics was, in the main, an affair of individuals,not political communities. Intercommunal wars and their contin-uations by diplomatic means were, at most, impositions upon self-sufficient political systems. In monopolizing authority relationswithin a territory or over a population, states were supreme andcomplete unto themselves.

However, with capitalist development on a world scale and es-pecially with technological innovations in communications andtransportation, national economies have become increasingly su-perseded. In consequence, nation-states, as superintendents ofnational economies, have become inappropriate units of the worldpolitical order. With production and distribution organized on glo-bal lines, and with environmental problems from resource man-agement to pollution controls crossing state boundaries, the statehas become an obstacle in the way of addressing many fundamen-tal human concerns.

The problem is not just that the state system divides the worldinto units that are too small to address the challenges the world'speoples confront. The greater problem is that the dispersal of po-litical authority into nation-states, the existence of competing cen-ters of supreme authority, produces antagonisms that exacerbatethese problems. What may once have been a natural division hasbecome an anachronism and a source of potentially devastatinganimosities.

The obvious remedy, it seems, would be to expand the bound-aries of the nation-state to include the entire world. On this view,the early defenders of the state were essentially right; their error,from a contemporary vantage point, though not from their own,

86


was only in their conception of the state's proper boundaries.Needless to say, the world system of nation-states generates polit-ical facts that are self-perpetuating. Therefore a different politicalorganization of the planet may not be feasible in practice. It maynot even be desirable in any remotely foreseeable future, given theenormous social and cultural divisions the state system has gen-erated or exacerbated. But there is no hope for change without firstarticulating a different, more unitary vision of political order. Itwould therefore seem to follow that democrats and perhaps othersas well should promote the idea of One Big State. In doing so, itwould be necessary to rethink much of what has long been main-tained about political institutions. Among other things, one wouldhave to contrive, if only in the imagination, an administrative ap-paratus appropriate for a state of global dimensions. But the coreideas, articulated in the years since the nation-state first appeared,would remain robust under these transformations in political in-stitutions. If it was right in earlier times to construct nation-states,it is right today to devise a state to rule over the entire world.

However, democrats cannot be happy with this conclusion.Expanding the size of the state, ultimately to include the entireplanet, probably would render many problems more tractable ad-ministratively than they now are. But in some respects a worldstate would actually intensify conflicts between the state form ofpolitical organization and democracy. The problem is not just thateach citizen's control over the state would tend to decrease as thesize of the state expands. The more telling problem is that, no mat-ter what their size, states enjoin the partitioning of alternatives ac-cording to the exigencies of centralization, not the interests - noteven the private interests - of individuals. The larger the adminis-trative unit, the more acute the divergence would tend to become.A world state, because it is the largest possible administrative en-tity, is the limiting case. If the voting behavior of citizens in rela-tively small nation-states differs significantly from the idealizedmarket behavior welfare economists model, the behavior of votersin a world state would differ even more markedly - rendering theinsult to democracy more severe.

The problem, then, is not just that the nation-state is too small.It is also too large and, worse, too all-encompassing. It draws intoitself more authority than is required for realizing democratic in-tuitions, with detrimental consequences for the implementation ofdemocratic values. Just as some problems are too global to be ad-dressed at the level of the nation-state, many questions of social

The general will

choice are too parochial, too much an affair of those who areimmediately involved, to warrant relegating authority over themto centralized concentrations of power. Much more appropriatewould be a system that vested authority in functional entities ofvarying sizes positioned at the real interstices of social and eco-nomic life. Then social choices could be made by all and only thosedirectly affected. In doing so, the quality of these choices wouldsurely improve, as would the level of public deliberation and de-bate. More in point, decentralizing political authority would pro-mote the formation of collective choice problems that more nearlyresemble those that democratic theorists assume when they rep-resent voting systems in the manner of idealized market arrange-ments. But, of course, were authority relations decentralized anddiffused, we would no longer have supreme authority concen-trated in a unitary entity. In a word, we would no longer have astate.

The end of the state system at the international level, like theend of the state itself, is, at best, a distant goal. The problem is notjust the difficulty of getting from here to there. More important,from a normative point of view, is the nature of the goal itself.Among the consequences of humanity's long tenure in the Age ofStates is the unavoidable fact that national and other group iden-tifications have become intertwined with the existence of the statesystem. To be sure, the communal values that states now foster arenot unequivocally estimable. Even Rousseau, the ardent defenderof patriotism, would likely not endorse its many deleterious con-sequences: militarism, xenophobia, international animosities. Butthe values Rousseau sought to instill through a common civil re-ligion, public spectacles, and a patriotic education are not withoutmerit. In any case, the purchase on community that the state sys-tem currently provides cannot be legislated away. Therefore, in or-der not to unsettle needlessly the group identifications peopleactually make, it is wise to proceed with caution.

LIMITS TO DECENTRALIZATION

It is unlikely, in any event, that the state form of political organi-zation could be replaced completely. For in even the most radicallydemocratized system of governance, centralized authority - atthe level of the nation-state or perhaps a world state - will re-main necessary indefinitely to settle disputes authoritatively when

88


they arise between the larger or smaller administrative units towhich other state functions devolve, and to address fundamentalresource allocation problems.

Even today, some "judicial" functions are performed by institu-tions tied to both nonstate corporatist associations and to the for-mal court system of nation-states. Many professions, for example,police themselves through their own institutional mechanisms,and many intragroup conflicts are resolved by these means. To theextent that enforcement powers too devolve onto decentralized in-stitutions, the decentralization of "the administration of justice"could be expanded considerably. But then as now there would re-main a need for "courts of last resort" - to resolve issues that re-main intractable by other means or that have consequences thatspill over from one jurisdiction to another. It is likely, then, that athoroughgoing decentralization of the judicial system, even wereit desirable for the sake of democratization, would be impossible.Unless we were to assume, contrary to what can reasonably be ex-pected, that interindividual conflicts will someday disappear com-pletely, there will always remain a need for something like a"supreme" court and its associated institutional arrangements.

On the other hand, economic affairs can certainly be organizedwithout a central funding agency to allocate resources and directinvestments. Thus, throughout the world today, capitalist marketsallocate resources and individual capitalists make investment de-cisions. But these matters affect everyone in varying degrees. In-sofar as democrats are committed to the vague but compelling ideathat individuals ought to make the decisions that affect their lives,democrats therefore have at least a prima facie reason for keepingthese matters in the hands of "the whole community."9 But in or-der to allow basic questions about the allocation and use of pro-ductive resources to be made democratically, it would seem thatthere would have to be, as it were, a single site where at least themost fundamental budgetary decisions are made. In the formersocialist countries, economic affairs generally were centrallyplanned. In light of the economic performance of these regimes, itnow seems clear that efficiency considerations alone suffice to dis-credit this degree of control from the center. However, microlevel

9 Thus there is also a prima facie reason why democrats ought to oppose propertyrights that confer the power to make investment choices on particular individ-uals. In this way, democratic values weigh in against capitalism. Cf. Levine, Ar-guing for Socialism, pp. 127-134.

The general will

economic planning is not necessary for control to be exercised col-lectively over economic life. It should suffice if only the mostfundamental budgetary decisions are centrally planned. Even so,the extent to which it is desirable that institutions of a decentral-ized kind replace centralized, state institutions would be, to somedegree, constrained.

What can vary, however, is the impact of these constraints. Evenunder existing conditions, it may sometimes be feasible and desir-able, from a democratic point of view, to devolve social investmentdecisions to decentralized funding agencies with varying jurisdic-tions. Prototypes for this kind of institutional arrangement arestate and (sometimes) international institutions that supply capitalfor research projects or development schemes. Proposals are sub-mitted, evaluated by peers, and allocations are made. Of course,the funds to be allocated are ultimately set by state or internationalauthorities. But it is fair to speculate that, with the growth in pro-ductive capacities prerequisite for moving toward the end ofthe state under communism, these budgetary constraints will di-minish in significance and with them the impact of centralized al-locating mechanisms. Increasingly, ventures will be undertakenbecause peers acknowledge their promise; not because, as in ex-isting capitalisms, individuals have special access to capital or be-cause, as in previously existing socialist societies, state or partybureaucrats decide to pursue them. In short, centralized economicinstitutions may be unavoidable even into the remotest future, buttheir importance can diminish progressively.

In any case, this side of communism, the crucial question, fordemocrats, is how best to democratize the political regime. To thisend, a number of dogmas of our ostensibly democratic politicalculture must be rethought.

REPRESENTATION

Representative governments, Rousseau (sometimes) maintained,are not second-best approximations of impracticable direct de-mocratic institutions but insidious alternatives to genuine de-mocracy.10 When the people elect representatives to rule in theirbehalf, they forfeit essential autonomy - turning themselves into

10 See The Social Contract, Book II, Chapter 1, and Book III, Chapter 15, for Rous-seau's most celebrated strictures against representative government. WhenRousseau was engaged not in ideal theory but in speculation on feasible po-litical arrangements in a world he thought already irreversibly corrupted, his

90


slaves of the masters they have elected. However, in this regardtoo, there is no need to accept Rousseau's stringent and idiosyn-cratic conditions on political legitimacy to accord credibility to hisdismissive assessment of representative institutions. A similarconclusion follows from the observation that the best argumentsin favor of representative government are non- or even anti-democratic in character.

In general, these arguments are of two kinds.11 There are, firstof all, those which assume the characteristic liberal distinction be-tween civil society and the state and then support liberalism'scelebration of civil society at the state's expense. On this view, pol-itics is, at best, a necessary evil, wisely relegated to "public ser-vants," while the vast majority pursue their interests free from itsdeleterious effects. It is of course desirable that political affairs beconducted well, that public servants be competent and perform ac-cording to expectations. Politics should therefore be structured asa career like others, with appropriate penalties and incentives, thebetter to attract able personnel and to reward merit. In addition,there should be measures in place to assure that governance willaccord with established institutional norms. It is therefore useful,if only to facilitate accountability, that the people vote in periodicelections and otherwise participate minimally in political pro-cesses. Elections also enhance de facto legitimacy, particularly inpolitical cultures where democratic values have become estab-lished. But in no case should politics be, as it was for Rousseau andthe entire "classical" tradition in political theory, an inseparableand essential feature of human life. However necessary statesmay be in the face of an otherwise devastating war of all againstall, what makes life worth living transpires outside the politicalarena - in civil society. Representative government is justified, onthis view, because it makes the best of this bad situation. Thoughonly a pale facsimile of real democratic control, it is instrumentalfor implementing a feasible approximation of an apolitical andtherefore nondemocratic ideal in which individuals pursue theirinterests and express their natures in a civil society free from po-litical intrusions.

views on political representation were far more favorable. For a comprehensiveaccount, see Richard Fralin, Rousseau and Representation: A Study of the Develop-ment of His Concept of Political Institutions (New York: Columbia UniversityPress, 1978).

11 For elaboration of some of the points that follow, see Levine, Arguing for So-cialism, pp. 134-141.

91

The general will

Arguments that appeal to considerations like these are notstrictly antidemocratic. They do not fault democracy so much asexclude it from their conception of what is normatively desirable.In contrast, the other kind of argument adduced in support of rep-resentative government does challenge democratic values. Thus itis sometimes claimed that the peoples' representatives are likely torule better than the people themselves because the people are illinformed, mercurial, susceptible to demagogic manipulation - inshort, unfit for the tasks of governance. Representatives are betterqualified, if only because, with governance their principal occupa-tion, they are able to cultivate the capacities required for directingpolitical institutions. In opposition to the claims of participatorydemocrats, direct democratic institutions, for antidemocrats of thisstripe, undermine political competence, educating citizens awayfrom civility and sound judgment. Representative government isthen defended as an alternative. The case for representative insti-tutions can itself be less or more antidemocratic. According to theless antidemocratic version of the argument, representative insti-tutions are defensible because they foster capacities for gover-nance better than the feasible alternatives. According to the moreantidemocratic version, political representation is no better - andis perhaps worse - than other nondemocratic forms of governancebut is preferable to its rivals for fostering a sense of legitimacy in aworld where "democracy" is everywhere endorsed and whererepresentative government and democracy are commonly con-founded. In either case, the class bias that formerly motivatedhostility toward rule by the demos also motivates the case for rep-resentative government. At root is the conviction that the ordinarypeople are not to be trusted; that government "of the people, bythe people, and for the people" - the reality rather than the slo-gan - is a peril to be avoided.

Of course, it is precisely faith in the demos that motivates gen-uine democrats. But even the staunchest democrat must concedethat the arguments invoked in behalf of political representationcontain a grain of truth. Oscar Wilde is supposed to have said thatsocialism would take too many evenings.12 Obviously, it was dem-ocratic socialism he had in mind. Wilde was right; there is a dangerthat, in a genuinely democratic order, projects that might other-

Cf. Michael Walzer, "A Day in the Life of a Socialist Citizen/' in Obligations:Essays on Disobedience, War, and Citizenship (Cambridge, Mass.: Harvard Univer-sity Press, 1970), pp. 229-238.

92


wise be pursued constructively would be dissipated in endlessdeliberations and public assemblies. Therefore there is reason torelegate public tasks to public servants. It is also beyond disputethat governance requires special competences that, very likely,cannot be universally attained or equally distributed. Certainlythese competences are not now generally evident in the vast ma-jority of the population. Proponents of participatory democracymay be right when they insist that democratic participation is itselfan educator; that, in the right conditions, public deliberation anddemocratic collective choice transform individuals into democraticcitizens. But even if democracy is an effective means toward itsown enhancement, there is still less reason to invest faith in thecitizenry at large than in subsections of it selected for politicalskills.

These arguments must be weighed against the value democratsaccord collective self-determination. It is unlikely that anyonewould be persuaded to abandon democratic convictions on theiraccount, but democrats would nevertheless be wise to incorporatethe insights these arguments convey into their reflections ondemocratic institutions. To this end, it is well to note that plausiblehesitations about democracy seem less compelling if we imaginepolitical authority situated not as it now is in de facto states butin larger and smaller institutions corresponding to the naturaladministrative divisions of particular social orders. It has alreadybeen suggested that democratic values are better implementedunder such arrangements than within states. It also seems plainthat non- and antidemocratic hesitations about peoples' powerare less worrisome under these arrangements too. If individualsexercise power mainly in institutions that involve them directly,fewer evenings would be taken up with meetings, freeing timefor extrapolitical projects and interests. And people immediatelyinvolved with particular institutions are likely already to have alevel of expertise sufficient for governing these institutionssoundly. In any case, they are prime candidates for transformationthrough political participation. Even those who are otherwisedisinterested in public affairs are likely to have their attentionsfocused and to deploy their energies constructively on adminis-trative matters of immediate concern to themselves, when the im-pact of the decisions they make is likely to be directly experienced.Lenin's celebrated polemical (or genuine) declaration notwith-standing, cooks may not generally be suited for running thestate but, in the right conditions, they surely can run their own

93

The general will

kitchens well. Admittedly, in the Age of States, faith in the demoshas a Utopian flavor in light of what is everywhere observed.But if The Second Discourse story makes a sound point, the Age ofStates is only a moment in human history, and what is humanlypossible in any particular period depends, to a considerable de-gree, on what its institutional arrangements allow and on theeffects of political forms upon human beings' mentalities andcharacters.

REPRESENTATION VERSUS DELEGATION

Political representation is commonly promoted, incorrectly, as asecond-best approximation of direct democracy. But if the pointwere really to approximate this ostensibly impractical ideal, itwould be better to restrict the independence of legislators, requir-ing them to operate as delegates rather than independent repre-sentatives. Revolutionary institutions have typically gravitatedtoward systems of delegation - the Paris Commune and the 1905and 1917 revolutions in Russia are conspicuous examples - andMarxist political theory, or rather the strain of it epitomized by TheState and Revolution, endorses it emphatically. The practical diffi-culty of communicating the choices of large numbers of people isa mark against delegation and a reason for supporting the inde-pendence of representatives. However, modern communicationstechnologies substantially undermine this rationale. Systems ofdelegation, approximating direct democracy, are probably nowfeasible. With existing (or feasible) communications technologies,direct democracy itself may even be possible. Therefore represen-tation by independent legislators no longer wins out over its rivalson grounds of practicality alone. If a case is to be made for repre-sentation, delegation, or direct democracy, it must be on the re-spective merits of these political forms.

As remarked, justifications for representative government typi-cally appeal to non- or antidemocratic considerations. But there isan argument favoring representative institutions that accords wellwith democratic intuitions and that, paradoxically, has a vaguelyRousseauean flavor. It is instructive to draw this argument out inorder to clarify, by way of contrast, what a genuinely Rousseaueanposition would be.

The argument turns on the observation that political represen-tatives are freer than delegates to assess alternatives according totheir respective judgments about what is best for everyone, not

94


just their own constituents. Insofar as representatives are able torise above their constituents' desires, they approximate, it seems,the disinterestedness or impartiality that Rousseau imputes to cit-izens generally in a de jure state. Thus, if generalized impartialityis desirable but unfeasible, an approximation of this desideratumis provided by political representation.

It should be noted, first, that, despite its appeal to the interestsof the whole community, this argument is only superficially likeRousseau's. The will, Rousseau insists, must be general both in itsobject and in its source.13 It is therefore as crucial that the peoplethemselves decide as it is that the decision they make be in the in-terest of "the whole community." This is not an arbitrary stipula-tion. Rousseau would have "the whole people decide" for just thereason that he would have individuals subordinate their private totheir general wills - in order to enhance their autonomy. In The So-cial Contract, Rousseau twice claimed that political representationis tantamount to the alienation of liberty.14 We need not followRousseau to this dire extreme to see that a system of legislationbased on relinquishing decision-making power through the elec-tion of representatives who rule on the peoples' behalf is, on theface of it, autonomy-diminishing. In any case, it is plain that theaffinity of the argument in question with Rousseauean views islargely illusory.

More importantly, the argument is ambiguous. It might seem towarrant the conclusion that independent representatives, becausethey are able to adopt a global perspective, are better equippedthan delegates (or, for that matter, the people at large) for ascer-taining what the general will is. But a more natural - and verydifferent - reading of its conclusion would be that impartial rep-resentatives improve the probity of private will aggregation be-cause they are less bound than delegates to the parochial interestsof their constituencies.

Imagine that the assembly of the people must decide some ques-tion of social allocation: say, whether or not to invest in highwayconstruction and, if so, how much to invest and where and howthe money should be spent. Were the point to aggregate privatewills, the task would be to assure that all (private) interests arerepresented and combined fairly. According to the argument inquestion, independent representatives have a signal advantageover more rigidly controlled delegates in this regard. In addition to

13 The Social Contract, Book II, Chapter 6. 14 See note 10.

95

The general will

representing their constituents' immediate interests, they are ableto attend to broader concerns and ultimately to the common inter-ests of everyone. So long as these common interests are notalso the particular interests of some constituency, they might nototherwise register in the political process at all. Let us supposethat certain concerns about the environmental consequences ofhighway construction fall within this description.15 The claim,then, is that independent legislators would have a better purchaseon these interests than delegates would - for all the private inter-ests that bear on the collective choice, not just some subset ofthem, could be represented in their deliberations. Then, assumingthat votes are combined according to fair procedures, the out-comes they generate would be fairer than under a system of strictdelegation.

But we know that "the whole community" can have general in-terests that are not reducible to the interests of its constituentparts - even when all pertinent private interests are accorded theirfair weight. As remarked in Chapter 1, the nature of the interestsdifferent entities have will vary with the entities in question. Forpurposes of normative theory, welfare interests are of obvious im-portance for human beings - now and in any imaginable future.But they are not the only kinds of interests that matter. We will seethat for "whole communities" under communist conditions an in-terest in autonomy enhancement has a supremely important roleto play alongside welfare interests. General interests are tanta-mount to interests of this sort.16

Independent representatives can, of course, assess alternativesaccording to any standard whatsoever. Therefore they can seekto enhance autonomy, if they wish, even at the expense of welfare.It is also conceivable that autonomy is best served by an insti-tutional structure that prevents individuals from making collec-tive choices directly. The considerations adduced to this pointmake these outcomes unlikely but not impossible. It is not neces-sary, however, to rely on speculations about probabilities to dis-count these possibilities. We will find in subsequent chapters thatRousseau's insistence that the will be general in its source aswell as in its object is a theoretically well-motivated and defensibleclaim.

15 These interests are political analogues of externalities in (idealized) free mar-kets. Externalities are costs or benefits that are not represented in an item's ef-ficiency price.


96


In any case, it is plain that an interest in enhancing autonomy, ageneral interest, might not coincide with even the most fairly ag-gregated set of private interests, "the will of all." Thus it is con-ceivable that decisions about whether, where, and how to buildhighways, say, will differ according to which standard is used. It isalso conceivable that there will be a genuine general interest withrespect to none or only some of these questions. Thus it may be inthe general interest that a highway be built and that it be built inone of several ways but indeterminate with respect to which ofthese ways. In those circumstances, on that question, even a well-ordered communist society might rightfully accede to "the willof all."

Whatever standards are appropriate for assessing the respectivemerits of representation and delegation, it is clear that insofar asthe point is to approximate direct democracy delegation, not rep-resentation, is the better choice. Why not have the ideal directly?There are some obvious reasons: Whole communities cannot begathered together for deliberation and debate, and it would beburdensome, if not impossible, to aggregate the choices of everycitizen on all matters upon which decisions must be made. Butsuppose, as suggested, that these difficulties can be mitigatedtechnologically. Other hesitations about direct democracy wouldremain - in particular, those that motivate the non- and anti-democratic arguments that support representative institutions.The case for representation revolved around two claims: that directdemocracy and, to a lesser degree, its close approximations are toodemanding, and that they fail to draw on and cultivate special com-petences required for governance. In both respects, delegation mayactually be superior to the direct democracy that it approximates.A hierarchically structured network of delegatory institutionscould relieve individuals from the burden of too many meetings;and delegates are as well situated as independent representativesfor cultivating capabilities for governance. A system of administra-tive units composed of delegates could therefore address hesita-tions about democracy at least as well as the representativeinstitutions characteristic of existing liberal democracies. The dif-ference is that liberal democracy addresses these qualms at democ-racy's expense,17 whereas the alternative can better accord withdemocratic aspirations.

17 See Levine, Liberal Democracy Chapter 8.

97

The general will

ELECTIONS

It would be foolish to speculate on the nature of the institutionsthat could implement democratic values once the obstacles placedin the way of democracy by the concentration of authority into asingle administrative nexus are past. Rousseau thought that polit-ical institutions could not be specified, except in a very generalway, apart from contingencies of time and place. Marx voiced sim-ilar hesitations about communism. In this respect, as in so manyothers, it is wise to follow their lead. Thus it is hardly possible hereand now to decide where the interstices of social and economic lifein societies of the future will lie, or to advise administrative unitssituated at these points how to conduct their affairs. However, ifthe preceding discussion has been on track, there is at least onevery general suggestion that does seem apt for as long as any sys-tem of delegation or independent representation remains in place.This suggestion may appear paradoxical inasmuch as its objectiveis to advance democracy. The suggestion is that legislators, what-ever their degree of independence, not be elected but selected bychance.18

The ostensible reason to elect legislators is to assure that constit-uencies are well represented. The rationale is the same as formajority rule voting generally: The winning candidate is the indi-vidual most voters want to represent them from among the indi-viduals who present themselves for candidacy. But we have seenthat candidates, because they are individuals, are, so to speak,prepackaged social states. Thus there is even less reason than inthe general case to expect that the trade-offs voters make in select-ing from among candidates accurately reflect their choices.Moreover, in elections for candidates, more than in the generalcase, opportunities for strategic misrepresentations of preferencesabound. In addition, experience has shown that political represen-tatives, despite their differences, tend to constitute themselvesinto a political class with what Rousseau would designate "privateinterests" of their own. On the other hand, were members of leg-islative and administrative units selected periodically and for shortintervals - say, by lot from among appropriate constituencies as inthe ancient Greek democracies or as juries are selected today - the

18 In The Social Contract, Book IV, Chapter 3, Rousseau cites with approval Mon-tesquieu's contention that "voting by lot is the essence of democracy."


benefits of elections could be achieved without the obvious short-comings and without the unhappy consequence of fostering a po-litical class distinct from the general population.19

It might be objected that legislators chosen at random would beeven less preferred than those chosen by elections. Perhaps so. Butwhat matters, for a democrat, is that the measures enacted inlegislation represent the peoples' will, not that the individualswho make these enactments be preferred to their electoral rivals.Here, again, there is a plain advantage for delegation over repre-sentation. To be sure, since individuals have different skills anddifferent knowledge, one individual could not quite substitute foranother without consequence even in a delegatory system. But,whoever they are, delegates are required by their mandates to doas their constituents desire. For this reason, who the incumbent ismatters far less than in a representative system.

DEMOCRATIZATION

Throughout most of this chapter, democracy has been conceivedstatically, as a collection of institutional arrangements that imple-ment a certain (synchronic) understanding of collective self-rule.Also, as is appropriate in the Age of States, political forms havebeen assessed on the assumption that politics is exclusively ormainly an affair of private interests. Eventually it will be necessaryto abandon both assumptions. But my aim in this chapter has onlybeen to fault the state system from the standpoint of democracy.For that purpose, assumptions that are standard in our politicalculture suffice.

However, I have also contrasted the positions under consider-ation with Rousseauean views, and I have very tentatively pro-posed general strategies for democratization. Since the chaptersthat follow address these topics from a perspective closer to Marx'sthan Rousseau's, it is well to point out here that the strategies pro-posed in this chapter do not depend on the transformation ofcapitalism into socialism, communism's "lowest phase" in the his-torical materialist scheme. These measures merit consideration asmeans for moving toward communism. But they can also be as-sessed and implemented in their own right. In the present histor-ical conjuncture, this is at least a significant practical advantage.

19 See Burnheim, Is Democracy Possible?, pp. 9-12 and passim.

99

The general will

We will find, however, that there is an important sense in whichMarxists were right in joining socialism with democracy, even asthey failed to acknowledge the (relative) independence of the so-cialist and democratic projects. We will find too that there was asense in which Marxists were right in distancing themselves fromliberalism - both in theory and in practice. But, again, they erredtragically in failing to acknowledge the evident merits of liberal in-stitutions. Democratization and liberal concerns will be at issuethroughout the chapters that follow - independently and from theperspective of Marxist theory. Both themes come together perspic-uously in the Marxists' conception of the "last state," the state thatsuperintends the transition to communism.

1 0 0

Chapter 5

The last state

Even those who are convinced of the feasibility and desirability ofmoving beyond the Age of States to an earthly "republic of ends"might wonder if statelessness is, in fact, a viable political ideal.Rousseau's despair of the possibility of establishing a just state inreal history provides a cautionary note. If Rousseau was right, thedamage already done to the mentality of human beings and there-fore to their wills is irreversible. By instituting de facto states, theinhabitants of the state of nature concocted a seeming remedy forthe affliction diagnosed in The Second Discourse. But this palliativehas altered the condition of humankind beyond the possibility of agenuine cure. Rousseau's deeply pessimistic reading of the worldas it is deserves serious attention, as do the convergent judgmentsof conservative and liberal philosophers. Nevertheless, there isreason to think that humanity can move beyond the Age of States;that the road from here to there, however difficult to traverse, isnot insurmountably blocked.

It would be foolhardy to aim for a definitive theoretical defenseof this contention. Institutional arrangements, by their nature, aretoo closely linked to peculiarities of time and place for general con-siderations to be decisive. What follows, therefore, are only reflec-tions on one sustained attempt, the most important to date, tobroach this problem in theory and practice: the project for revolu-tionizing the state pursued by Marxists committed expressly to theend of the state under communism. To reflect on "the dictatorshipof the proletariat," as these Marxists called the "transitional" statethey envisioned, is to continue the discussion of institutionalmeans for implementing democratic values broached at the end ofChapter 4. If only for this reason, the designation Marxists use todescribe this form of the state is needlessly misleading. But sincethe expression "dictatorship of the proletariat" has standing in the

101

The general will

Marxist tradition and since it suggests claims worth defending, de-spite its unfortunate connotations, the expression will be retainedhere. What follows is not an attempt at a general account of theinstitutions appropriate for a state committed to its own witheringaway. An account of that kind is impossible, given the essentiallyconjunctural nature of political practice. The most that we can ex-pect is to appropriate lessons learned in ventures undertaken incircumstances quite different from our own, and bound to differeven more markedly from the conditions future partisans of state-lessness will encounter.

There is however a plainly philosophical problem, appropriatelybroached at this level of abstraction from actual politics, suggestedby the idea of a state concocted to implement its own demise. Sucha state, whatever its institutional forms, wields power in order toimplement a vision of ideal social arrangements. It does, in short,what virtually all political philosophers thought eminently properbefore the advent of the modern era but what hardly anyone todaythinks right - it uses state power to implement a particular andcontroversial conception of the good. Neutrality with respect atleast to controversial conceptions of the good is a tenet - somewould say a defining feature - of modern liberalism. The statecommunists envision to lead the transition to statelessness there-fore seems at odds with what had become the consensus view. It isa theme of this chapter and the next that this observation is itselfmisleading. But before examining the dictatorship of the proletar-iat in light of its differences from the state liberals advocate, andbefore reflecting on neutrality as an ideal, it will be well to say asmuch as possible, albeit not very much, about the nature of theostensibly offending polity. The burden of this chapter will there-fore be to reflect on the dictatorship of the proletariat as Marxistshave traditionally conceived it.

DEMOCRACY AND DICTATORSHIP

Throughout our political culture, democracy and dictatorship areopposed. In democracies the people rule; in dictatorships the peo-ple are ruled - by a dictator. But on even a cursory scrutiny of themost paradigmatic instances of real-world "dictatorships" - or oncommon-sense reflection on the nature of governance - it is evi-dent that dictatorship in this sense is impossible in a state of anysize or complexity. An individual can hardly rule alone. A "dicta-tor" can predominate among those who rule - prevailing, typi-

102

The last state

cally, in setting agendas and making decisions. But a dictatorcannot literally "dictate," and we saw in Chapter 4 that the peoplehardly rule at all in so-called democracies. Thus neither "democ-racy" nor "dictatorship," as these terms are widely used, desig-nates anything like what it suggests.

It might be thought that democracies are based on the rule oflaw, whereas dictatorships, like ancient tyrannies, are based on thewills of those who govern. This suggestion has merit. But the ideaof the rule of law is itself problematic. No putative dictatorship hasever dispensed with at least the appearance of legality - with "le-galism" as a means for legitimating existing power relations.1

What is called "the rule of law" plainly involves more than legal-ism. But it is far from clear that the difference will serve to distin-guish democracies from dictatorships.

In any case, current usage matters very little for the Marxist un-derstanding of "the dictatorship of the proletariat."2 The "dicta-torship" Marx discussed in his writings on the Paris Communeand that Lenin resumed in The State and Revolution was not the"dictatorship" of contemporary political culture and is not in-tended to contrast with "democracy." Representative govern-ments, the paradigm case of democracy in the current view, arebourgeois class dictatorships, according to Lenin's use of the term;and the dictatorship of the proletariat, as Lenin conceived it, isradically democratic. "Dictatorship," according to a celebrated for-mulation of Lenin's, is "rule based directly upon force and unre-stricted by laws."3 But, in this sense, for anyone persuaded by agenerally Hobbesian view of the state, all states are dictatorships.Marxists are Hobbesians in this respect. Indeed, they have a rea-son Hobbes never imagined for supporting the Hobbesian view.For Marxists, laws are not sufficiently independent of state insti-tutions to operate substantively, not just formally, as constraintson the sovereign's power. Indeed, in the Marxist view, both thestate and the law depend substantially on a third, genuinely in-dependent, factor - the economy. In traditional Marxist terminol-ogy, the state and the law are super structural. Both rest upon an"economic base." How then can the one also rest upon the other?

1 Cf. Judith N. Shklar, Legalism: An Essay on Law, Morals, and Politics (Cambridge,Mass.: Harvard University Press, 1964).

2 What follows resumes claims defended in my Arguing for Socialism, Chapter 6,and The End of the State, Chapters 7 and 8.

3 V. I. Lenin, "Proletarian Revolution and the Renegade Kautsky," in CollectedWorks (Moscow: Progress Publishers, 1972), vol. 28, p. 236.

103

The general will

This question is unanswerable as long as these notoriously obscuremetaphors remain unexplicated. This is not the place to pursuethis issue very far, but this much is clear: For Marxists, superstruc-tural phenomena are constrained by the economy in a way that -and to a degree that - the economy is not constrained by them.4

Moreover, superstructural phenomena are constrained in this wayand to this extent only by the economy, not by each other. It istherefore plain that, unless the Marxist metaphors are denuded oftheir intended sense, it is inconceivable, from a historical materi-alist vantage point, that the state could rest upon the law. Fromthis perspective, democracies are just as much "tyrannies" in theclassical sense as the "dictatorships" of mainstream politicaldiscourse.

We might conclude, then, that in the sense that Lenin intendedthe term, there is in fact nothing controversial in the contentionthat all states are dictatorships. This claim, it seems, asserts nomore than the widely held idea that states are, according to theconsensus view, institutionalized monopolies of the means of (defacto) legitimate violence, a definition meant to apply as well tostates with "autonomous" legal systems as to those without them.

4 Despite what Marxists, following Marx, have believed, the claim that the stateand the law are "superstructural" need not imply that all aspects of these in-stitutional arrangements and practices belong to the superstructure. The base-superstructure distinction is best construed as a claim about explanation. Verygenerally, then, the idea is that the base explains the superstructure more thanthe superstructure explains the base. As remarked, Marxists have tended to-ward the view that the economic base explains everything about legal and polit-ical superstructures - and also forms of consciousness. But it is plain that noMarxist, no matter how doctrinaire, could plausibly have maintained that liter-ally everything having to do with the state, law, and consciousness has an eco-nomic explanation. Judges in Britain wear wigs; in the United States they donot. Must Marxists be committed to the idea that this fact can be explained eco-nomically? Evidently not. The standard view, then, was that important or fun-damental political and legal phenomena and forms of consciousness belong tothe superstructure. But what counts as "important" or "fundamental"? It haslately become clear that there is no theoretically well-motivated way to answerthis question in general; and also that, strictly speaking, it is only necessary tohold that those aspects of the state, the legal system, and "forms of conscious-ness" that bear on economic structures need be explained economically to accordwith historical materialist precepts. Thus the superstructure may encompassmuch less than Marxists have traditionally supposed. These claims are ad-vanced and defended in G. A. Cohen, "Restricted and Inclusive HistoricalMaterialism," in History, Labour, and Freedom: Themes from Marx (Oxford: Claren-don Press, 1988). See also Wright, Levine, and Sober, Reconstructing Marxism,Chapter 5.

104

The last state

What is nonstandard and even distinctively Marxist in this strainof theorizing is the idea that states are class dictatorships, that statepower ensures and reproduces class domination.

However, this conclusion is easily misunderstood. As remarked,for political philosophers after Hobbes, the most fundamental con-cern has been to justify the coordination of individuals' behaviors.But from the perspective of historical materialism, social classes,defined by the social relations of production in which individualsare implicated, are the proper starting point for theorizing humansocieties and their political arrangements. With this crucial emen-dation, the core problem of modern political philosophy takes on adifferent cast. The interindividual coordination problem politicalphilosophers since Hobbes have addressed becomes an intraclasscoordination problem. This new problem is then "solved" in justthe way the old one was: by creating an apparatus that allows forthe preeminence of a single will and for its enforcement. This willcannot be a general will in Rousseau's sense, for in a class-dividedsociety there can be no general interest. Nor is it a sovereign's willin Hobbes's sense, since there is no individual or collection of in-dividuals who can genuinely represent the whole community. Inthe Marxist view, the will that directs the state is the will of theeconomically dominant class. The state allows the members of thisclass to act in concert - for the purpose of maintaining their dom-ination of subordinate classes.

There therefore is a sense in which Marxists go beyond theclaims of standard political theory in calling states dictatorships.Arguably, there is even a sense in which, in their view, democracyand dictatorship conflict. The contrast, however, is not with de-mocracy per se but with democracy purged of its class-centeredconnotations. When democracy is understood to designate rule bythe (undifferentiated) people, it is a form of what Hegel called aRechtstaat, a state grounded in the (supposed) universality of Right(Recht). However, a dictatorship, according to contemporary un-derstandings, is not a Rechtstaat. Like ancient tyrannies, thesepolities are based on the will of the dictator, a will that is arbitrarywith respect to Right. In calling states dictatorships, Marxists ef-fectively maintain that the Rechstaat is an illusion, that so long asclass divisions remain, the universality of Right is an impossibleand even harmful ideal.

In this respect, the characterization of states as dictatorships ispart of a broader challenge to any notion of morality or law thatstands beyond class divisions. For Marx, so long as class divisions

105

The general will

persist, class interests remain fundamental. They cannot be assim-ilated to private wills, as Rousseau proposed in The Social Contract,nor can they be dissolved into a putative general will. The stand-point of universality, emblematic of the moral point of view, istherefore flawed and deeply ideological. It works to impede theemancipation of subordinate classes by representing ruling classinterests as universal interests. The idea of a Rechtstaat, a univer-sal state, is correspondingly flawed and similarly ideological solong as class domination, organized by the state, persists. To de-pict states as dictatorships, then, is to continue a distinctivelyMarxist assault on a widely accepted conception of morality. It is,from the perspective of political theory, to fault the idea of Right.The insistence that states are class dictatorships is additionally in-formative. It helps explain why the idea of a Rechtstaat cannot besustained. More importantly, this insistence joins the theory of thestate with the idea that history is a history of class struggles. Theterm "dictatorship" in Marxist political theory is therefore neitherso contentious as may at first appear nor so standard as furtherinvestigation suggests.

But the fundamental Marxist claim is indeed that states are classdictatorships and therefore that they figure in a broader processof class struggle that shapes the nature and trajectory of humanhistory. In this process, there is no conflict between dictatorshipand democracy. Quite the contrary. The dictatorship of the prole-tariat is that form of the state that superintends the processthrough which the Age of States is finally superseded. To thisend, democracy - in the fullest sense of the term - is essential toits nature.

RELATIVE AUTONOMY

In The State and Revolution, Lenin is emphatic in insisting that theold regime be "smashed," not "seized," and replaced by a newtype of state. Following Marx's own reflections on the Paris Com-mune and its tragic aftermath, Lenin maintained that institutionsof the old form, by their very nature, promote bourgeois class ruleand that a proletarian class dictatorship, correspondingly, requiresinstitutional arrangements of a different kind. It is fair to general-ize this claim. If we accept the thesis, fundamental to the Marxistview of politics, that state power is always a class dictatorship,then, if Lenin was right, there exist relations of compatibility andincompatibility between types of state power or different class dic-

106

The last state

tatorships, on the one hand, and the institutions through whichstate power is achieved and maintained, on the other.5

Writing in the summer of 1917, between two great revolutionary"moments" and in the flush of rapidly changing and suddenlymomentous events, Lenin was disinclined to draw out many of theimplications of this claim, particularly those that focus on how rev-olutionary victories can become deflected. But although he doesnot quite acknowledge the point, it is a plain implication of hisview that a working class that seizes but does not smash the of-fending institutions of the class dictatorship it overthrows wouldnot, in fact, control a proletarian state. In time, if the old institu-tions remain in place, they would reestablish compatibility by re-constituting a bourgeois class dictatorship, perhaps administered byerstwhile revolutionaries or their successors.

Real history seems to have vindicated this unhappy inference.But in the Soviet Union and elsewhere where the "Soviet model"was imitated or installed, the failure to advance toward genuinecommunism was, at most, only partly explained by the survivalof old institutional forms. The indefinite prolongation and even-tual entrenchment of repressive and ideological arrangements in-appropriate for the permanent revolutionizing of the state were atleast as much to blame. Nevertheless, Lenin's implicit recognitionof a likely and profoundly debilitating danger facing revolutionaryRussia seems remarkably prescient.

It is unclear in general to what extent forms of state power limitinstitutional arrangements and vice versa. It is one thing to holdthat not all institutional forms are compatible with all class dicta-torships and something else to assert particular incompatibilities.Nowhere in the Marxist corpus is any general guidance offered onthis more vexing question. This is just what we should expect if itis true that no general theory of institutional arrangements is pos-sible, However, The State and Revolution makes much of one sort ofincompatibility - that alleged to hold between a "relatively auton-omous" repressive state apparatus, proclaimed the hallmark ofbourgeois class rule, and the dictatorship of the proletariat. It isnow plain that this claim is dangerously flawed.

What motivates the suggestion that an independent judiciary bereplaced by popular tribunals or that a specialized police and army5 To say that a type of class dictatorship and a set of institutional arrangements are

incompatible is to say that their coexistence is structurally unstable to an extentthat the configuration cannot be maintained indefinitely. One or the other mustgive way.

107

The general will

be superseded by peoples' militias is the same principle that un-derlies the injunction to smash parliamentary forms of governanceand establish workers' councils in their stead. It is a commitmentto enhance direct popular control over state institutions - in aword, to democratize them. But historical experience belies thisconclusion. It is plain that the "relative autonomy" of institutionsfor "the administration of justice" in existing liberal democracies isa signal advantage of these regimes that very much enhances in-dividual and collective well-being. It is equally plain that the com-parative lack of independence of the judiciaries and police of theformerly "socialist" countries worked to the detriment of the cit-izens of those states.

It could be argued that the superiority of liberal democracies inthis respect follows only from the failure of states constructed onthe Soviet model to democratize their repressive institutions. Theregimes that controlled those countries did "smash" the institu-tional forms of liberal democracy. But, virtually without exception,they did so preemptively, inasmuch as liberal institutions werenever actually established in the social orders they replaced. In anycase, they did not institute democratic arrangements in theirplace. Instead, the repressive apparatus of the former socialist pol-ities was directed by a party and bureaucracy even more remote frompopular control than the relatively independent judiciaries andpolice of liberal democracies. The repressive apparatus of a bureau-cratized state socialism in a political community bereft of liberaltraditions and with very limited democratic experience could hardlyhave been expected to realize the vision advanced in The State andRevolution. Sadly, experience has vindicated this prediction.

Nevertheless, it would be hard to deny that the history of theSoviet model from its inception to its demise tends to corroboratethat pessimism about human nature that pervades liberal politicalculture and that motivates its support for institutional safeguardsdesigned to save human beings from themselves and from eachother. It is important for democrats to address this wariness and toreflect on its implications.

Pessimism - some might say "realism" - about human nature isthe principal source of the zeal with which liberal democrats de-fend the relative autonomy of state institutions. Characteristically,radical democrats counter liberal doubts by ascribing transforma-tive effects to democratic participation itself. For radical demo-crats, liberal democrats err not so much in their view of thepsychology of present-day human beings but in thinking that the

108

The last state

human nature one observes must remain forever impervious to agenuinely democratic "education/' The citizens of liberal democ-racies and of other non- or only minimally democratic states dohave reason to fear their own natures and should therefore preferrelatively independent judiciaries and police to more democraticalternatives. For them, the tyranny of the majority is at least asmuch a menace as more conventional tyrannies. But, in the viewof radical democrats, the very real shortcomings of the human na-ture we observe can be altered profoundly. In this respect, Rous-seau was a pioneer participatory democrat. In his view, when "thewhole people rule concerning the whole people/'6 human beingsbecome citizens who, by the very nature of their wants, effectivelyrealize liberal values without recourse to non- or antidemocraticliberal constraints.7

However, the belief that democratic participation has these ef-fects is more an article of faith than a corroborated hypothesis. Par-ticipatory democracy has seldom been attempted even in small-scale organizations, and never broached at the level of the nation-state. Therefore evidence that would compel assent to or rejectionof participatory democratic claims is unavailable. But enough isknown about human nature to speculate with some confidence,and there is anecdotal evidence too. On these bases, it is fair toconclude, albeit tentatively and awaiting empirical confirmation,that faith in the positive transformative effects of democratic par-ticipation is eminently reasonable.

The tyranny of the majority is a plausible threat insofar as themajority of individuals are ill informed, mercurial, susceptible todemagogic manipulation, and generally lacking in the virtues ofcivility. For such a citizenry, it is appropriate - and even crucial -to limit the scope of democratic collective choice as liberalspropose and to move from direct democratic to representativeinstitutions as liberal democrats do. However, even the most irre-sponsible individuals, invested with real power over particulardecisions, can become responsible decision makers - especially ifthere are mechanisms in place that hold them accountable for thedecisions they make. The reason why is clear enough: Genuinecontrol encourages interest, and interest, in turn, encouragesknowledge, competence, and responsibility. In cases where indi-viduals' choices and activities are palpably effective, self-interest

6 The Social Contract, Book II, Chapter 6.7 See Levine, The Politics of Autonomy, pp. 72-80.

109

The general will

alone should suffice to focus individuals' interests in the workingsof the institutions they control. Thus we observe that the most po-litically ill-informed citizens are often knowledgeable about mat-ters like music or sports that, for one reason or another, engagetheir interest. We can therefore infer that, for most people, genu-ine control over institutions is likely to concentrate the mind andspark sufficient concern to address liberals' fears. Of course it re-mains unclear how much confidence to invest in these specula-tions. But it is plain that the deficiencies one observes in thecitizenry of existing states does not inevitably carry over to otheraspects of social experience and should not be assumed unalter-able under different institutional arrangements.

In contrast to many contemporary defenders of participatory de-mocracy, Marxists who support radical democratic positions focusless on individuals' activities in non- or quasi-political contextsthan on their behavior in revolutionary undertakings - the ParisCommune of 1871, the 1905 revolution in Russia, the events of thespring and summer of 1917, the October Revolution, the (failed)revolutionary eruptions in Central and Eastern Europe in the af-termath of the October Revolution, and the host of Third Worldrevolutionary ventures of the past several decades. They infer thatthe masses are indeed able to govern well from the fact that hostsof downtrodden and oppressed individuals have assumed controlof their destinies and participated competently in insurgent orga-nizations and revolutionary institutions in the course of these mo-mentous upheavals.

However, these "data" are unlikely to assuage liberals' fears.There is first of all the problem of maintaining competence and ci-vility as revolutionary events subside and give way to other, more"normal" concerns. It is far from clear that what is normal in ex-traordinary circumstances can be indefinitely sustained. In addi-tion, in revolutionary situations, revolutionaries commonly takeseverely illiberal measures against real or imagined enemies, andthis fact has weighed heavily on the liberal imagination. Indeed,from the last days of the French Revolution until the dread of nu-clear weaponry impinged on human consciousness, revolutionaryTerror was, according to Hegel's celebrated portrayal, the very em-blem of unthinkable horror.8 Arguably, excess is endemic to revo-lutionary politics; it has certainly been characteristic of it. In any

8 See Hegel, The Phenomenology of Mind, section BB, VI, B, III, "Absolute Freedomand Terror."

1 1 0

The last state

case, the history of revolutions and of popular rule within revolu-tionary situations can reinforce the fears that lead liberals to dis-tance political institutions from direct popular control as much asit can reinforce the hopes of radical democrats for bringing themcloser together.

There is little hope, at present, of intervening in this debate de-cisively or of changing settled convictions. No available evidence islikely to persuade those who believe human nature incapable ofrespecting liberal values "spontaneously" from thinking that lim-itations on self-rule can be safely relaxed for democratic reasons;and those who think otherwise are not likely to be dissuaded bythe host of contrary observations that can be adduced in supportof the need for keeping democracy within bounds. It is not justthat partisans of one or another view are obstinate; the bearingand even the relevance of the data that exist are too unclear. Theexperimental introduction of increasingly democratizing institu-tional forms along the lines suggested at the end of Chapter 4would probably generate evidence of a less equivocal sort. But it isunfair to prejudge the outcome of experimental ventures not yetundertaken - indeed, not yet on any actual political agenda.

Even if new or existing data were approached without prejudice,an unlikely assumption on so charged an issue, threshold effectsor significant discontinuities in the consequences of democratiza-tion are likely to make the interpretation of the evidence contro-versial. Not even the most ardent proponents of participatorydemocracy have offered principled reasons for thinking that the(alleged) beneficial effects of democratic participation rise or fallcontinuously with increasing or decreasing levels of participation.For participatory democrats, it would seem to be enough if some(presumably, far-reaching and radical) levels of participation en-hance democratic citizenship. Therefore it need not embarrass par-ticipatory democrats if, at some levels, enhancing or diminishingparticipation has no significant effects on individuals' capacitiesfor democratic citizenship and, at other levels, increasing partici-pation is actually detrimental.

In any case, the accumulation of relevant evidence ought to be ahigh priority for democrats, even if there is little prospect of set-tling debates about participatory democracy definitively. Demo-cratic commitments rest ultimately on faith in the demos and indemocracy's beneficial consequences. If this faith is not to degen-erate into a dogma, with potentially dire consequences for human-kind, it is well to construe it, by analogy with experimental

1 1 1

The general will

hypotheses, as a plausible but inconclusive conviction awaiting thetest of experience. Unlike those with otherworldly convictions,democrats can test their faith. Wise democrats will be eager todo so.

BEYOND LIBERAL DEMOCRACY?

The crucial test is to see if it is indeed possible to realize liberalvalues without recourse to liberal institutions - by relying on in-dividuals' characters or, as Rousseau would prefer, on the nature oftheir wills. Through popular self-rule - and in other ways too -individuals are to be educated to respect the moral equality of per-sons and all that it implies. No less a liberal than John Stuart Millalso emphasized the importance of character for implementing lib-eral and democratic values.9 But he foresaw no possibility of dis-pensing with liberal institutional safeguards. Historically, Marxistshave disagreed. I will argue that, ultimately, they may be right,but only if their disagreements with liberals are based on princi-pled democratic convictions and only if the nonliberal institutionalforms they envision are deferred to the very long run. Historically,Marxism has been deficient in both respects.

A conservative sensibility, attuned to the dangers inherent intampering with established political forms, would tend to sup-port Mill against Marx. In general, it is wise to act in accord withthis sensibility. However, conservatism, so conceived, need notentail immobility, nor need it even proscribe revolutionaryundertakings.10 It only militates against the impatience that seemsalways to afflict revolutionaries - advising caution in anticipatingthe pace of change - and warning against the idea that mentalitiesdeveloped over centuries and shaped, in part, by psychologicaland cultural factors remote from any "economic base" or political"superstructure" can be rapidly and profoundly transformed.11

9 The effects of institutional arrangements on individuals' characters is a centraltheme of both On Liberty and Considerations on Representative Government. Seealso Dennis F. Thompson, John Stuart Mill and Representative Government(Princeton: Princeton University Press, 1976).

10 See Levine, The End of the State, Chapter 4.11 Questions about the pace of change are distinct from questions about the ap-

propriateness of attempting revolutionary transformations (of real property re-lations) at particular historical moments. It is one thing, for example, to faultthe Bolsheviks for attempting too much too quickly and something else toclaim that the Bolshevik Revolution itself was premature, given prevailing con-

1 1 2

The last state

Respect for what some liberals would call the "rights" of others iseminently achievable for most people most of the time even now.In uneasy coexistence with the opposite inclination, the charactertraits from which such respect flows can flourish to an extraordi-nary degree in the exceptional, revolutionary conditions Marxistsinvoke as support for democracy's beneficial effects. But sufficientconfidence in individuals' capacities to respect "rights" in "nor-mal" circumstances and without institutional constraints wouldrequire, at the very least, a protracted period of character forma-tion - possibly lasting generations. Marx and Lenin and their co-thinkers went wrong in imagining that this process could be cutshort, that the old institutional forms could be changed quicklyand extensively without conjuring up the perils they protectedagainst. One need not quarrel, then, with their view of the natureof the changes needed for realizing democracy fully, or even withthe means they proposed for bringing the requisite changes about.They erred only in their estimates of the rate at which thesechanges can safely proceed. Marx, it seems, was clear-sighted to afault about what human beings could achieve but willfully blind tothe all-too-human obstacles standing in the way of his vision.

If, as Plato said, a virtue of a thing is that which makes it performits function well, and a vice is that which has a contrary effect, thenimpatience is a serious vice in a revolutionary. Perhaps more thanany other factor, impatience makes otherwise sound revolutionaryventures turn out badly. Marxists, including Marx, can be faultedfor this vice - in their practice and in its theory. Lenin's injunctionto smash rather than seize the old repressive state apparatus was,in context, a clear expression of impatience that has had dreadfulconsequences. It is true that the old repressive state institutions,relatively insulated from each other and from the main governingbodies of the state, are obstacles in the way of democratization andits (potentially) beneficial effects. It is also true, therefore, that thedictatorship of the proletariat, in order to fulfill its historical mis-sion, must eventually replace these institutions with more directlydemocratic forms. But it is wrong to undertake to democratize pre-maturely and without liberal protections. To do so is to foster thetyranny of the majority (or of those who purport to rule in their

ditions in Russia in 1917. Historical materialism, Marx's theory of history,bears on the latter question; the kind of conservatism that revolutionarieswould do well to incorporate is concerned, on the other hand, with the pace ofchange.

113

The general will

behalf) and to put individuals' "rights" - and the dignity they ex-press - in jeopardy. On even the most optimistic assessment of de-mocracy's transformative potentialities, it will be a long timeindeed before liberal values are sufficiently instilled that liberal"constitutional" protections can be significantly relaxed, as theymust eventually be to pave the way for the unprecedented and un-fettered democratization communism entails.

Those who acknowledge the possibilities historical materialismidentifies and who are eager to wager on democracy's transforma-tive effects should therefore forbear from calling for the last state,the state in transition to statelessness, to "smash" relatively auton-omous institutions for the administration of justice and the main-tenance of civil order. To issue such a demand is to neglect thecentral problem a state in transition must confront: the recalci-trance of deeply entrenched undemocratic and illiberal inclina-tions, and the difficulties these habits of mind pose in the processof transition. In well-functioning liberal democracies, programswith illiberal implications are unlikely to gain many adherents.Not so, however, in illiberal regimes where few have much to loseby relinquishing nonexistent liberal protections. Thus the call to"smash" liberalism preemptively - in conditions where liberal in-stitutions have not produced their own (limited) transformative ef-fects - is particularly reckless. Liberal democracy is a preschool forreal democracy. It is evaded at great peril. As history attests, thelikely outcome of attempts to pass over liberal democracy is tyr-anny - at first, perhaps in the guise of democratic forms, eventu-ally without any disguise at all.

Even with liberal democracy as an established point of depar-ture, the way ahead for the last state remains problematic and un-charted. It is therefore pointless to speculate on time frames for theprotracted transformative process of building securely liberal char-acter types, liberals without liberalism, even conceding the feasi-bility of the project. All that can be concluded with confidence isthat the process will be facilitated greatly if the institutional safe-guards liberal democracy has evolved to protect against the tyr-anny of the majority remain in force - now and for the foreseeablefuture, though not for an indefinite future. The dictatorship of theproletariat must find a place, as it were, for relatively autonomousrepressive state institutions. For a long time to come, perhaps avery long time, it must remain liberal, however incompatibleliberalism may ultimately be with its general democratizingtendency.

114

The last state

But then how can the last state advance toward statelessness?How can it be a state of a new type? To begin to answer these ques-tions, it is well to follow the lead of some Marxists after Lenin -Gramsci, for example, and Althusser - and to look back, again, toindications provided by political philosophers before Marx. Thuspartisans of the last state should take to heart Hume's celebrateddeclaration that "opinion is the true foundation of the state."12

This view is entirely in accord with Rousseau's thought too, inso-far as Rousseau conceived politics as a contest between the privateand the general will - a struggle for the "opinion" of each citizen.In short, it is insufficient for defenders of the last state to look onlyat repressive state institutions, as Lenin did. The real battle in andover the dictatorship of the proletariat is ultimately waged at thelevel of opinion.

THE IDEOLOGICAL STATE APPARATUS

It was perhaps with this idea in mind that Althusser identifiedwhat he called "the ideological state apparatus," a formulationplainly intended to parallel "the repressive state apparatus" of or-thodox Leninist theory.13 The expression is Althusser's, but theidea is Gramsci's. Among Marxists, it was Gramsci who first priv-ileged the role of ideology in the struggle for socialism and, by im-plication, in the process of moving socialist societies towardcommunism. And it was Gramsci who first conceived ideologicalinstitutions as state institutions, even when they are not juridically"public."14

Gramsci's aim was to provide a theory of socialist transforma-tion suitable to Western European conditions. Without quite say-ing so, he effectively acknowledged that the Russian Revolutionwas not an appropriate model. Because it took place in a societythat was underdeveloped economically and politically, it could bewaged, like the French Revolution more than a century before, asa frontal assault on the principal institutions of the old regime.12 David Hume, A Treatise of Human Nature, ed. L. A. Selby-Bigge (Oxford: Clar-

endon Press, 1888 and subsequently), Book III, Part 2, Chapter 8, pp. 539-549.13 Louis Althusser, "Ideology and the Ideological State Apparatus," in Lenin and

Philosophy and Other Essays (London: New Left Books, 1971).14 See Antonio Gramsci, Letters from Prison, ed. and trans. L. Lawner (New York:

Harper, 1973), and Selections from Political Writings, vol. 1 (1910-1920), ed. andtrans. Q. Hoare and J. Mathews (New York: International Publishers, 1977),vol. 2 (1921-1926), ed. and trans. Q. Hoare (New York: International Publish-ers, 1978).

" 5

The general will

The repressive apparatus of the czarist autocracy and then the Ke-rensky government, weakened by war and mired in a feudal past,proved incapable of withstanding the pressures of a popular insur-rection directed by a "professional" leadership organized on quasi-military lines. However, states that superintend more developedsocieties are likely to have considerably greater means for main-taining themselves in power. A frontal assault on such a state istherefore unlikely to succeed. What is required, instead, is a pro-tracted struggle - metaphorically, a "war of attrition." At stake inthis war is what Gramsci called "hegemony" - moral, cultural, andintellectual leadership. If the working class and its allies could gainhegemony, the old order would lose the battle for "opinion," thebasis for its (de facto) legitimacy and hence for its very existence.The conquest of the old state apparatus therefore ratifies an earlier,more fundamental victory won on the ideological front. Socialistrevolution, on Gramsci's view, is not essentially a matter of storm-ing a Bastille or seizing a Winter Palace. In developed countries,revolutionary events like these can be of only symbolic or expres-sive importance. The real task is to combat the old regime ideolog-ically, to assault its legitimacy, precipitating a crisis that is finallyresolved by the ascension to power of a new hegemonic class.

In his account of ideological struggle, Gramsci was an inno-vator. But his depiction of revolutionary politics as a struggle forhegemony recalls familiar Marxist accounts of the "bourgeois"revolutions through which capitalism decisively and irreversiblydefeated feudalism. In the traditional view, a new mode of produc-tion "matured" in the "womb" of the old. In other words, thegreat bourgeois revolutions cemented economic transformationsthat had already largely transpired within the old regime. Astrictly parallel development is inconceivable under capitalism.Capitalist economies can accommodate prototypes of socialist eco-nomic organizations - workers' cooperatives, for instance, or pub-licly controlled enterprises - but, so long as the economy remainscapitalist, these organizations must deal with one another andwith the larger economy in capitalist ways. Whatever their internalstructure or juridical status, they are therefore capitalist firms. Forthis reason, pockets of socialism cannot arise within capitalisteconomies.15 But it is possible for socialist "opinion" to form and

15 This formulation presupposes that capitalism and socialism are indeed histor-ical alternatives in the way that Marx's theory of history claims and that thebroader intellectual culture largely assumes. Eventually, this presupposition

The last state

gain hegemony. In this sense, a new form of civilization can in-deed grow to (near) maturity in the womb of the old - not in con-sequence of inexorable economic processes but as a matter ofconscious design. The venerable metaphor therefore remains apteven if the phenomenon it describes is not quite the same. Bour-geois revolutions became possible and even necessary because thepolitical superstructure of feudal societies grew increasinglyincompatible with emerging capitalist production relations. InGramsci's view, in order to prevail in developed capitalist econo-mies with "normal" states, socialist revolutionaries must create aparallel incompatibility - between the economy and the state, onthe one hand, and opinion or hegemonic ideology, on the other.

The signal achievement of state building in the modern periodhas been to consolidate all (de facto legitimate) means of violenceinto a single institutional nexus. No parallel consolidation has oc-curred for the institutions that shape opinion. Indeed, a much re-marked characteristic of capitalist societies - especially those withliberal democratic states - is the separation of the state from manyspheres of individuals' lives and concerns and the consequentplacement of many of the institutions Gramsci's analysis impli-cates outside the state - in "civil society." To be sure, this exclu-sion is seldom, if ever, complete. Some of the institutions involvedin the struggle for hegemony - schools, for instance - are typi-cally established, administered, and funded by the state. Thepress and other mass media are rarely public in this sense, partic-ularly in liberal democracies. Religious organizations, which stillplay a role in shaping the mentalities of many citizens, are, atmost, only tenuously affiliated with official state institutions andmay even be "separated" from the state by law - as required, for

may prove untenable. Then, despite what Marxists and also non-Marxists havelong believed, the capitalism-socialism distinction would have to be redrawn.It would be necessary, in other words, to disaggregate "capitalism" and "so-cialism" as these terms are traditionally understood. It is unclear what wouldfollow from so fundamental a revision of received ways of thinking about theeconomy and society, but one possibility might be, paradoxically, that the tra-ditional Marxist view of economic transitions - according to which modes ofproduction mature within the modes of production they are destined to super-sede - would be vindicated. At this point in the reconstruction of historicalmaterialist themes, these reflections are too speculative to warrant dwelling onthem. I do, however, venture some very tentative suggestions supporting aqualified version of the idea that it may be well to disaggregate traditional un-derstandings of "capitalism" and "socialism" in Chapter 9 herein; see alsoWright, Levine, and Sober, Reconstructing Marxism, Chapter 5.

117

The general will

example, by the United States Constitution. Despite these differ-ences, Althusser, elaborating Gramsci's idea, subsumed all ofthem or rather aspects of them under the concept "ideological stateapparatus," to which he added opinion-shaping organs emblem-atic of civil society - like the family.16 It is true, of course, that thestate, through the legal system, impinges upon family life andother ostensibly "private" matters, sometimes to the detriment ofliberal values, sometimes unexceptionably. Ideally, the state super-intends and supports the private sphere. In this sense, it is a con-dition for its possibility. But this fact hardly warrants collapsingthe distinction liberals draw between civil society and the state.The ideological state apparatus of Althusser's formulation strad-dles this difference and, by implication, discounts it.

Nevertheless, there is some justification for Althusser's designa-tion, and there is even a sense in which his formulation captureswhat is distinctive in Gramsci's political theory. For both Althusserand Gramsci, following Marx's depiction of the state as the set ofinstitutions that reproduce existing class relations, the institutionsstate functionaries administer directly could not possibly exhaustthe scope of the term. The "state," according to the assumed def-inition, would designate all the institutional arrangements thathave the requisite effects. Civil society, as liberals understand it,should therefore be replete with state institutions. Insofar as eventhose institutions that paradigmatically distinguish civil societyfrom the political realm work to reproduce existing relations ofclass domination, they are part of the state - effectively, though ofcourse not juridically. As remarked, this usage can confuse andmislead. However, it also suggests an important challenge tomainstream political theory, motivated by Marxist views of thestate's role in history. In its insistence that politics is not a realmapart but an inextricable aspect of all spheres of human life, Al-thusser's formulation points to a difference between Marxist and

16 For Althusser, the ideological state apparatus is, in effect, an ideal type, as isthe repressive state apparatus. In his usage, these ideal types are defined func-tionally. Actual state institutions typically exhibit features of both of these idealtypical concepts. Thus the police are mainly part of the repressive state appa-ratus, but their existence also has ideological effects. On the other hand,schools are mainly part of the ideological state apparatus, but they also func-tion repressively. It is unclear whether Althusser believed that these ideal typ-ical concepts are exhaustive. In any case, there plainly are state functions - forinstance, the regulation and coordination of productive activities and of fi-nance - that are neither ideological nor repressive in his sense.

118

The last state

liberal political theory, and an affinity with Rousseauean thoughtand the broader "classical" tradition in political philosophy.

The Althusserian-Gramscian insight concerns the relation be-tween civil society and the state. It can be appreciated if contrastedwith the superficially similar distinction, fundamental to historicalmaterialism, between a society's "economic base" and its "politi-cal superstructure." It is compatible with the historical materialistidea that a clearly differentiated "economy" is a relatively recentphenomenon in human history; that, as is often maintained, pre-capitalist societies conflated economic and political functions tosuch an extent that it was not until capitalism matured in the"womb" of European feudalism that the "economic" entered theconceptual scheme of individuals involved in production and ex-change. Nevertheless, in the historical materialist view, a distinc-tion between the economy and the state is of general explanatoryimportance throughout human history - even when it does notcapture the lived experience of the people whose history it pur-ports to explain. Similarly, liberals with a historical understandingof social and political ideas could concede the historical particu-larity of the distinctions they draw between civil society and thestate. But it is tempting for them too to conclude that this distinc-tion is of transhistorical normative significance, just as historicalmaterialism claims transhistorical explanatory relevance for itsown distinction between the economy and the state.

Historically, many liberals have succumbed to this temptationand for good reason. Without a transhistorically applicable pri-vate-public distinction, liberals would lack a critical tool for as-sessing nonliberal societies, and liberalism would be just oneideology among many. Marxists, of course, have long proclaimedthe historical particularity of all normative views. But, strictlyspeaking, Marxists can be agnostics on the question of the univer-sality of particular normative positions and even uphold the trans-historical standing of some normative standards.17 In any case,nothing Gramsci or Althusser claims about "the ideological stateapparatus" in any way impugns the transhistorical normativeimportance of the liberals' society - state distinction, This obser-vation is hardly surprising inasmuch as neither Gramsci nor

17 It is even consistent with historical materialist claims to defend the opposingview, as do historical materialists who are also moral realists. See, for example,Alan Gilbert, Democratic Individuality (Cambridge: Cambridge University Press,1990).

119

The general will

Althusser addresses the normative questions of interest to liberalpolitical philosophers. But they do challenge the underlying, de-scriptive position that liberal moral philosophy suggests: the ideathat there is a real difference between society and the state that isimportant for understanding the state transhistorically. Again, thisclaim is analogous to the historical materialist's distinction be-tween the economy and the state. But there is no reason in generalwhy a distinction drawn in the context of a moral philosophy,whatever its historical scope or metaphysical status, should be car-ried over into descriptive political theory. Nor is there any reasonin this case to consider the general rule contravened. Thus it is theview of Althusser, insofar as a substantive thesis can be teased outof a provocative formulation, and implicitly also of Gramsci andperhaps of Marx as well, that the liberal's distinction, but not thehistorical materialist's, is of only historically transient importance;that, however natural it may be to treat these distinctions similarly,the two are not, in fact, on a par.

As remarked, this position is an implication of the idea that stateinstitutions be defined as those which reinforce class domination.Is this view justified? In a word, it depends. For most explanatorypurposes, the Marxist definition is too encompassing to be ofmuch use. In class-divided societies, almost all institutional ar-rangements function, at least in part, to reproduce existing rela-tions of class domination. In calling everything that has this effecta state institution, many differences of interest to political theoristsare inevitably obscured. The juridical private-public distinction islikely to track political phenomena better. But, as Marxists in-structed by the Eleventh Thesis on Feuerbach, Althusser and Gram-sci were not so much interested in understanding the world in allits facets as in changing it. Their aim was to construct a "revolu-tionary theory," an understanding of the world capable of guidingrevolutionaries in the task of transforming society toward commu-nism. For this purpose, the more encompassing definition may in-deed be apt, for it suggests a political agenda appropriate to theMarxist political project that is likely to elude adherents of a nar-rower view of the state.

CULTURAL REVOLUTION

To endorse a process of socialist revolution, from the defeat ofcapitalism to the establishment of communism, and to conceivethis process as a struggle in and over both the repressive and ideo-

120

The last state

logical state apparatus is to imagine the state that overcomes theconditions that make states necessary, "the dictatorship of theproletariat/' engaged in a protracted - Marx might have said "per-manent" - ideological struggle. Marxist political movements inrecent decades have produced a name for this phenomenon:cultural revolution. The team captures the spirit of the Gramscian-Althusserian idea. In Althusser's case, the connection was almostcertainly intended. Althusser and his co-thinkers representedand defended a scarcely disguised Maoist politics,18 where Mao-ism was held, according to Western perceptions at the time, toprivilege ideological struggle as a means for overcoming the bu-reaucratic shortcomings and general immobility of Soviet Commu-nism. It was Maoism, more than any other Marxist tendency, thatpromoted the idea of cultural revolution.

Today, even more than Marxism itself, the idea of cultural revo-lution has fallen into disrepute, thanks largely to the misuse towhich the idea was put. Contrary to what was once widely be-lieved on the left, it is now clear that the Great Proletarian CulturalRevolution, launched in China in 1966 and continuing for nearly adecade, was an unmitigated disaster. Its costs in human life andwell-being, and its detrimental effects on economic developmentand political morality, are well known. Moreover, no one can plau-sibly claim that the Cultural Revolution moved China closer tocommunism. If anything, by its negative example and to compen-sate for the havoc it wrought, the Cultural Revolution encourageda turn back toward markets and private enterprise that may notend before the full restoration of capitalism in China.

Of course, underdevelopment and the unrelenting hostility ofworld capitalism - and, in China's case, of the Soviet Union too -were mitigating circumstances. But it is impossible to explain awaythe Cultural Revolution or any of the other unhappy vicissitudes ofCommunist experience without blaming Communism itself. Need-less to say, it is wrong to identify Communism with (small-c) com-munism or to see Communism as instrumental for communism'srealization. Communism is not irrelevant to communism, butneither is it as great an embarrassment as is widely believed.19

18 See, especially, "Contradiction and Overdetermination" and "On the Materi-alist Dialectic/' in Althusser, For Marx. Althusser's Maoism was always ex-pressed in arcane theoretical terms - partly in consequence of the pervasiveobscurantism of his intellectual style, partly to avoid direct conflict with the(pro-Soviet) French Communist party to which he belonged.


121

The general will

Nevertheless, in a period in which both Communism and commu-nism are in disrepute, in part thanks to Communism's demise,there is an evident futility in defending the idea of cultural revo-lution - just as there is in defending the dictatorship of the prole-tariat. Short of awaiting a new historical moment, then, one canonly argue for what is sound in these ideas in the hope of hasten-ing the time when a political practice will emerge in which theyassume their rightful place. One can hope too that the cultural rev-olutions of the future - and the class dictatorships in which theyoccur - will have little resemblance to the institutions and prac-tices that bore these names in the past. So long as liberal con-straints remain in place, and so long as the lessons of pastmisfortunes are taken into account, this hope is reasonable, nomatter how improbable it may nowadays appear.

As it wages a political and cultural revolution, the last state trans-forms its own repressive and ideological "apparatuses" by radi-cally democratizing them. But no matter how democratized theymay be, the dictatorship of the proletariat uses public coerciveforce and ideological authority to promote a particular and conten-tious conception of the good. Thus, as noted, the last state appearsto contravene the liberal desideratum of political neutrality. Inas-much as political neutrality in its own right and in light of histor-ical experience is a defensible ideal, it is crucial, in reflecting onMarxian communism, to confront this issue - and, more gener-ally, the compatibility of core liberal commitments with the Marxisttheory of the last state.

1 2 2

Chapter 6

The liberal state and/versusthe last state

For nearly two centuries, "liberalism" has designated a variety ofideas and institutional arrangements, sharing at least a family re-semblance and perhaps also an essential core. In advanced capi-talist countries, the dominant theory and practice of politicsnowadays generally are liberal. In some capitalist societies, like theUnited States, liberalism has been a dominant ideology from thebeginning.1 Among political philosophers and moral theorists inthe English-speaking world, liberalism has never been more ascen-dant than it now is. Part of the explanation for this phenomenonmust lie with the enormous impression made by the work of somecontemporary liberal philosophers. But their influence is itself ex-plained, in part, by the undeniable appeal of liberal tolerance. His-torically, liberals have sought to ensure tolerance by arguing forprincipled limitations on the state's - and also society's - powerover individuals' lives and behaviors.2 In contemporary accounts,liberal philosophers have propounded the idea of neutrality withrespect to "conceptions of the good." On this understanding of the

1 In the political cultures of some Western states, particularly in nineteenth-century Europe, "liberalism" took on more particular connotations, based uponsecularism (or anticlericalism) and tolerance, and a historical connection withlaissez-faire capitalism. Ironically, in twentieth-century North America (and tosome degree elsewhere), the term has come to designate (slightly) left-of-centerpolitical tendencies that endorse an affirmative use of state power to addresssocial problems, sometimes to the detriment of laissez-faire arrangements.However, secularism in public life and, of course, social tolerance are persistentthemes.

2 As remarked in Chapter 5, in contrast to Rousseauean thought and indeed tothe entire "classical" tradition in political philosophy, all liberalisms acknowl-edge a distinction between the state - and, more generally, the political arena -and an essentially apolitical "civil society." The first liberals were concerned al-most exclusively with limiting the power of states. John Stuart Mill was among

123

The general will

core liberal idea, the state - and, more generally, society throughwhat Mill called "the moral coercion of public opinion" - shouldnot seek to implement particular conceptions but should onlyguarantee that competing views, or at least those that are in anyway controversial, be treated fairly. This contention is ostensibly atodds with Marxist accounts of the last state.

I will argue that the cogency and appeal of this position depend,in part, on how "conceptions of the good" are conceived. For now,in accord with prevailing understandings, the expression can betaken to designate overarching systems of value and belief thatshape individuals' "plans of life" as well as some more limitedprojects. Not many centuries ago it was universally assumed thata legitimate function of political power was precisely to implementconceptions of the good in this sense. Even those who insisted ondelineating the respective jurisdictions of sacred and secularauthorities usually thought "defense of the faith" and the enforce-ment of moral restrictions legitimate political functions. Nowa-days, as already remarked, the contrary idea has becomehegemonic. Thus most "conservatives" in developed countries aretoday more liberal (in historically appropriate, if somewhat anti-quated, senses) than genuinely conservative. In this respect, mostradicals are liberals too. Not very long ago, it seemed clear thatMarxists at least were not liberals. Now, even this verity seemsproblematic. It is increasingly plain that most of the normativeconcerns of declared Marxists fall squarely within the problematicof liberal social philosophy. Perhaps what has long been the casehas only recently become evident. Nevertheless, it today seemsfair to conclude what would have been difficult to imagine only afew years ago: that nearly everyone is a liberal.

To be sure, liberalism has many professed critics. Some philos-ophers fault liberalism for failing to take proper account of thevalue of community,3 for its universalism,4 and for a host of othershortcomings. But even declared critics of liberal theory and prac-

the first to articulate a broader conception of liberalism's claims. Mill's breakfrom liberalisms that concerned only relations between the individual and thestate is registered in On Liberty, perhaps the principal liberal document of thenineteenth century.

3 See, among others, Michael J. Sandel, Liberalism and the Limits of Justice (Cam-bridge: Cambridge University Press, 1982). For a critical survey of the literature,see Amy Gutman, "Communitarian Critics of Liberalism," Philosophy and PublicAffairs, vol. 14, no. 3 (1985), pp. 308-322.

4 See, for example, Michael Walzer, Spheres of Justice: A Defense of Pluralism andEquality (New York: Basic Books, 1983).

124

The liberal state/the last state

tice generally uphold the fundamental ideals implicit in liberal tol-erance - neutrality in the political arena and, in civil society, theactive encouragement of diversity. For all practical purposes, lib-eralism's critics are liberals too.

It should be stressed, as some liberal philosophers recentlyhave, that liberal tolerance is a political value and that support forit does not entail neutrality at the level of personal morality.5 Tol-erance is a virtue of states or societies, and of individuals only in-sofar as they are citizens of liberal states and members of liberalsocieties. Thus liberals need not be moral relativists, nor need theydeny that there is a conception of the good to be discovered.Neither need they be agnostic on the question of knowing whatthe good is. Different views about the objectivity of moral dis-course are as open to liberals as to anyone else. The liberal idea isjust that states be neutral on questions of value and that "the moralcoercion of public opinion" not obstruct "experiments in living"undertaken to advance knowledge about the good or, indeed, forany reason whatsoever - provided, of course, that others are notharmed.

There is plainly merit in construing liberal tolerance as a politicalvalue. Doing so frees liberalism from dependence on contentiousmetaphysical claims while dissipating otherwise troubling ten-sions in the thought of important liberal writers. Thus John StuartMill, as a dedicated utilitarian, plainly believed that there was agood to be discovered and pursued. At the same time, he es-chewed deployments of state power or moral coercion in civil so-ciety aimed at realizing particular conceptions of the good. Thereis evidently no contradiction in these positions, so long as toler-ance is deemed a virtue appropriately confined to the politicalsphere.

However, when liberal tolerance is cast as a political value only,there is some risk of obscuring an aspect of liberal theory that iscentral to Mill's moral philosophy and to the thought of manyother prominent liberal philosophers too. Identifying this connec-tion and reflecting on its implications are indispensable if we areto address the tension that exists between the last state in theMarxist scheme, a state that aims at undoing the conditions for itsown necessity, and the core values liberal states aim to embody.

5 Cf. John Rawls, "Justice as Fairness: Political, Not Metaphysical," Philosophy andPublic Affairs, vol. 14, no. 3 (1985), pp. 223-251; and Charles E. Larmore, Patternsof Moral Complexity (Cambridge: Cambridge University Press, 1987).

125

The general will

Rawls suggests that liberal tolerance came into prominence inconsequence of a historical compromise reached implicitly in theaftermath of the Protestant Reformation and the religious wars towhich it gave rise.6 Individuals moved by the conceptions of thegood represented by Catholic and Protestant ideologies were in-clined to be intolerant of competing conceptions - in part thanksto the nature of their ideals, in part thanks to the psychology ofreligious conviction. But intolerance had, for generations, pro-moted devastating warfare - to the detriment of nearly everyone.Tolerance therefore emerged - as a more or less deliberate strategyfor promoting civil peace. It was a necessary evil in the face of thehavoc wrought by its absence. Then the necessary evil was trans-formed by liberal writers around the time of the Enlightenmentinto a positive ideal. What was originally only a modus vivendi,grudgingly adopted, became an indispensable feature of just po-litical and social arrangements.

No doubt, in broad outline, this story does account for the emer-gence of tolerance in recent European history. Arguably, it alsoprovides at least part of the explanation for the fact that toleranceis today most honored, if not always faithfully practiced, in thoseparts of the world whose history has been most directly touchedby the Reformation and the Enlightenment. However, it does notaccount for what we most need to understand if we are to discoverthe deeper affinity joining the liberal tradition - or, more precisely,a certain strain of it - with Rousseauean and Marxist political the-ory. It does not explain why, for at least some liberals, tolerancebecame a positive ideal.

To put the question more sharply: all liberals, by definition, de-mand as a matter of justice that public coercive force not be used toenforce particular (contentious) conceptions of the good. Some lib-erals, however, have made tolerance in this sense an aspect of theirvision of good social and political arrangements. Thus their view ofthe good coincides with their theory of the right.7 We can only

See Rawls, "Justice as Fairness." See also Robert Paul Wolff Jr., "Beyond Toler-ance," in Wolff, Marcuse, and Moore, Critique of Pure Tolerance.Individuals whose conceptions of the good are illiberal - individuals, say, whonot only oppose abortion but who want to live in a world in which abortions arelegally proscribed - will, of course, be frustrated in genuinely liberal polities.More generally, liberal regimes will, as a matter of right, always favor tolerance(at the political level) over intolerance. There is therefore a sense in which lib-eralism cannot, except on pain of incoherence, be neutral with respect to neu-trality itself. In this sense, liberalism effectively favors conceptions of the good

126


speculate about how this fortunate situation came about; perhapsmuch of the causal responsibility can be vested in the well-knownhuman propensity to make a virtue of necessity. In any case, inorder to understand the nature and importance of liberal tolerancefor this strain of modern liberalism - and to discern its bearing onMarxist concerns - it will be necessary to look beyond strictly po-litical motivations for neutrality and to bring more traditionalmoral philosophical concerns back in.

To this end, it will be helpful to reflect on aspects of the moraland political theory of John Stuart Mill. It need hardly be said thatnot all modern liberal theory is Millean in the sense I will identify;it could even be argued that Mill himself was not consistently oreven preponderantly Millean in this sense either. But Mill's writ-ings stand to this day as the preeminent liberal defense of toler-ance, and his arguments are most convincing if considerations ofthe sort I shall adduce are allowed. Strictly speaking, my conten-tion is just that there is, in fact, a strain of Mill's thought thatadvances the theoretical commitments I will, for the presentpurpose, impute to Mill tout court. I would hazard too that focus-ing on this dimension of Millean theory is useful for understand-ing his moral and political philosophy generally, and indeed forgrasping the nature of liberalism generally - but these claims areat some remove from the issue at hand, and I will not pursuethem here.

I will focus in this chapter on affinities between Millean andMarxist views of human nature. I must therefore repeat that whatI shall have to say about Marx's philosophical anthropology ishighly selective. It is plain that a good many of Marx's express re-flections on human nature and essential humanity cannot be sus-tained. But there nevertheless is a "rational kernel" that is bothsound and consonant with the aspect of Millean liberalism I willsketch. I will argue that, despite what is widely assumed, there isno deep incompatibility between the goals Mill sought to advancethrough liberal tolerance and the mission Marxist political philos-ophy imputes to the last state. Once what is misleading and un-essential is set aside, it will become clear that similar commitmentsunderlie both notions. Indeed, there is a sense in which Millean

that accord positive value to neutrality over conceptions of the good that are in-imical to neutrality. More precisely, with respect to the advocacy even of illiberalviews, liberalism is, by definition, neutral; what it cannot allow is their coerciveimplementation.

127

The general will

and Marxist political theory, from different vantage points andwith different concerns, each articulate the same core idea.

At the heart of both theories is, of course, a commitment to au-tonomy but also, less conspicuously though no less centrally, tohuman flourishing - to a natural condition of excellence that hu-man beings, properly cultivated and nurtured, can achieve. In thislatter respect, Mill and Marx are "Aristotelian." But unlike Aris-totle, Mill and Marx advance an essentially historical understand-ing of what human flourishing involves. For both writers, thenatural condition of humankind is not temporally fixed but histor-ically variable and generally progressive. They are each, as it were,historicized Aristotelians - assessing social and political arrange-ments according to how well they advance an ever-developingideal of genuine humanity. Some of the more implausible and un-appealing features of liberal and Marxist normative theory, as theyare usually understood, can be explained in part by the failure -even of Mill and Marx themselves - to recognize the implicationsof this general orientation in moral philosophy.

ROMANTICISM/CAREERISM

In this category, pride of place must be accorded two unlikely no-tions that have come to be considered emblematic of Marxist andliberal theory, respectively: Marx's idea that a fulfilled human lifeis best not identified with any particular occupation or activity butis instead irreducibly multifaceted; and the ostensibly opposed lib-eral notion of a "plan of life," of a life conceived as a career. Thelatter idea has become so deeply entrenched in liberal thought inrecent years that plans of life are commonly conflated with con-ceptions of the good - thereby entering into the definition of lib-eralism according to currently influential formulations.

It requires no particularly arcane psychological theory to faultthe ideal of universal, polymorphous self-realization that Marx ad-vanced in The German Ideology and that arguably continued to playsome role throughout his later work.8 To be sure, the idea doeshave critical force. The social inequalities capitalism generates ex-acerbate differences between individuals. Under communism orindeed in any regime more egalitarian than those that now exist,these differences would probably be less pronounced. But regard-less of the economic structure in place, "the natural lottery" would

8 Cf. Jon Elster, Making Sense of Marx (Cambridge: Cambridge University Press,1985), Chapter 2.

128


still distribute talents differentially. And even if, contrary to fact,anybody, potentially, could do anything well, so long as time re-mains scarce actually doing most things well would inevitably re-quire developing only a few capacities at the expense of others.Perhaps we could hunt in the morning, fish in the afternoon, andcriticize at night. But how many of us, even under the most be-neficent communism, could fill teeth in the morning, repair jet en-gines in the afternoon, and work on differential equations afterour evening meal?

Marx went wrong in The German Ideology in large part thanks tothe residual influence of Feuerbach and other Young Hegelian phi-losophers and, more generally, because he was immersed in an in-tellectual environment beguiled by Romantic notions of a unitaryhuman "essence." Following Hegel and his successors, Marx, inhis early writings, was tempted to suppose that All is One. It wasthen not too great a "dialectical" leap to conclude that one is all,that human beings can realize themselves by actualizing undiffer-entiated capacities virtually without constraint - or, in morepractical terms, that they can fulfill themselves by doing whateverthey feel inclined to do, turning from one activity to another asthey please, all the while performing well enough for the tasksthey undertake.

It need hardly be pointed out that this view radically underes-timates the time and effort required to cultivate the capacities hu-man beings might want to develop, even supposing that people,supplied with full and equal access to the means for cultivating ca-pacities, become vastly more able and multifaceted than the one-track individuals we observe in societies like our own. Plainly,Marx's proposal is psychologically unrealistic. It may also be log-ically flawed to the extent that some self-realizing projects neces-sarily involve doing some things better than others can do them orbetter than one can do other things oneself. Nevertheless, it can-not be denied that there is something attractive in the Marxistideal, whatever its shortcomings, in the face of the rigoristicsingle-mindedness of liberal writers who, in advocating neutralitywith respect to plans of life, assume totalizing conceptions of thegood. If the former view can be designated Romantic, the latterconception can be called Careerist because it depicts a human lifeas a career, structured around overarching objectives and dedi-cated to the realization of specifiable goals.

Needless to say, many human projects require extensive andlong-range planning; and some require lifetime commitments. But

129

The general will

it is hardly intrinsic to human endeavors that lives be conceived ascareers - that there be a single lifelong project with which individ-uals identify and against which their success or failure in life is as-sessed. Even single-minded dedication to a project requiring alifetime to fulfill is not per se Careerist, so long as the project - atits inception and throughout its duration - is undertaken and pur-sued in a genuinely autonomous way. It is characteristic of capi-talist societies that those who are more or less free to choose theirroles in the economic system come to be identified with their oc-cupational choices and to structure their lives, including theirextraeconomic projects, around their "chosen" careers. As is ob-served reproachfully in The German Ideology, under capitalism peo-ple who hunt, fish, and criticize become transformed into hunters,fishermen, and critical critics. That autonomy should be dimin-ished in this way, despite the existence of voluntary labor markets,is a deeply ironic offense to human freedom. Insofar as Marx, likeRousseau, conceived autonomous choice as a fundamental humaninterest that no social practice can legitimately infringe, this insultto freedom is a basis for indicting capitalist economic systems. Per-haps Marx's shamelessly Utopian account of life under commu-nism can be excused, to some extent, as a fervent expression of thisindictment.

It is fair to suggest that Careerism will appear natural only toindividuals leading forms of life that are already Careerist in broadoutline. Before the advent of modern capitalism, then, Careerismprobably would have seemed uncongenial and perhaps even un-intelligible to most people. Even today, it is likely to seem appro-priate only to those whose identities are effectively tied tocareers.9 Careerism is thus as historically particular as Romanti-cism is. And like Marx's Romanticism, it is an integument bestpeeled away from the "rational kernel" it partially obstructs.

What remains after this operation is performed is, in each case,an ideal of human life fused out of two complementary commit-ments: to freedom in the sense of autonomy and to self-realization;to the value Rousseau expressed in his insistence that "Man isborn free" and to the Aristotelian ideal of human excellence. It isbecause, appearance to the contrary, careers are more imposedthan chosen that they offend our interest in autonomy. And it isbecause polymorphous development is incompatible with the

9 Cf. Susan Moller Okin, "Justice and Gender," Philosophy and Public Affairs, vol.16, no. 1 (1987), pp. 42-72.

130


cultivation of human excellences that we cannot be quite so mul-tifaceted as The German Ideology would have us be. In the end, Ca-reerism and Romanticism conflict with human flourishing -with the fullest possible satisfaction of our fundamental interestsin autonomous choice and in freely developing our innate poten-tialities.

Strict Aristotelianism conceives human excellence in terms con-genial only to the lives of privileged individuals in the Athenianpolis or in close approximation of that condition. For Aristotle, thelife of an Athenian gentleman, dedicated to contemplative pur-suits, was ideal. This picture is static and undemocratic.10 ForMarx, in contrast, history is progressive in the sense that economicsystems and corresponding superstructures rise and fall to adaptto ever-increasing levels of development of productive forces, andthe resulting historical transformations expand the role of thedemos in public life. Liberals have less theoretically elaborated ac-counts of what Mill called "the permanent interests of Man as aprogressive being." Nevertheless, they too construe freedom andself-realization in roughly the same way. For Marx and Mill andthe broader traditions they represent, the principal objective of po-litical life is precisely to promote human flourishing by enhancingautonomy and self-realization. This commitment could, of course,be contested. But, at least in a modern and secular culture, it ishardly controversial in the way that particular conceptions of thegood are.

Even were liberal-Marxist valuational commitments more con-troversial or less plausible than they are, it would still be the casethat, from a liberal perspective, what justifies neutrality with re-spect to competing conceptions of the good would not enjoin thatthe state be neutral with respect to these core values. In philosoph-ical discussions, the issue is seldom if ever posed in these terms.But it is worth remarking that not even the most ardent proponentof neutrality would maintain that the liberal state has no legitimateinterest in protecting the autonomy of its citizens or the conditionsunder which they can become what they potentially are. In fact,

10 In other words, the Aristotelian picture of human excellence can be faulted notjust for its historical limitations but also because its realization for some, in his-torically pertinent circumstances, depends upon its nonrealization by others -the vast majority of slaves and free men and women whose labor is necessaryto sustain a gentlemanly existence for the few. Thus the Aristotelian idealflaunts the core liberal - indeed, modern - sense of justice and the notion ofmoral equality that underlies it.

131

The general will

for Mill and the strain of liberal theory he epitomizes, it is preciselythe idea that the state exists to advance these values that motivatesliberal tolerance. For Mill, diversity and the development of indi-viduality are not only effects of historical development; they areone of its most powerful motors. States are justified, in part, to theextent that, by encouraging diversity, they propel history along.Thus tolerance toward particular conceptions of the good is instru-mental for bringing about the most fundamental objectives Milland Marx endorse.

This instrumental connection provides a compelling reason forsupporting liberal tolerance based on a notion of the good life forhuman beings. It is therefore distinct from, though consistentwith, justifications for tolerance based on a notion of equal respectfor persons in a world in which individuals have different andsometimes conflicting conceptions of the good. Appeals to equalrespect are tantamount to appeals to justice in a distinctively mod-ern sense; appeals to the good life, on the other hand, have a moretraditional resonance, though conceptions of the good are histor-ically variable and, as both Mill and Marx would insist, "progres-sive." The modern sense of justice may indeed be shaped, asRawls and others maintain, by the need to accommodate to diver-sity in the face of the fundamental conflicts that divide modern so-cieties. But whatever its genesis, it is plainly compelling to personswith modern, egalitarian sensibilities. However, it does not sufficeto explain the positive valorization of heterogeneity and ideologi-cal diversity that liberals like Mill evince. To justify enthusiastic tol-erance - the active encouragement, not just the grudging acceptance,of differences - it is necessary to appeal, as Mill did, to a certainview of "the permanent interests of man as a progressive being."In doing so, Mill and Marx are again of one mind. If their sharedposition is sound, it will become plain that, appearance to thecontrary, the last state, the state that aims at overcoming the con-ditions that make states necessary, need offend neither the under-lying values tolerance aims to promote nor even, in any significantway, liberal neutrality itself. Indeed, under certain conditions -precisely the conditions radical democracy aims to foster - liberaltolerance is instrumental and perhaps even indispensable for pro-moting the aims of the last state. Paradoxically, but in accord withthe language traditionally deployed by Marxists to designate statesthat deliberately undo the conditions that make states necessary,liberal tolerance is a likely feature of any well-functioning proletar-ian class dictatorship.

132


ENHANCING AUTONOMY

Autonomy is only one sense of freedom, but support for it is im-plicit even in a commitment to the "negative freedom" with whichit is sometimes thought to contrast.11 Ostensibly, negative freedomis exemplified in Hobbes's definition of freedom as "the absence ofExternall Impediments," the forerunner of the classical liberalview;12 Rousseauean autonomy, on the other hand, epitomizespositive freedom. However, despite what is widely believed, theseideas are related. What reason could there be to value an absenceof external impediments except to advance human beings' capac-ities to do what they want?13 But doing what one wants is notyet acting freely, if wants are instilled by others. Thus the value

11 The idea that prevailing notions of freedom or liberty (using these words in-terchangeably) are distinct and even opposed was famously advanced by Isa-iah Berlin in his essay 'Two Concepts of Liberty"; see Four Essays on Liberty(Oxford: Oxford University Press, 1969). Berlin sets out to defend negative lib-erty against positive liberty, an idea that ostensibly includes autonomy in thesense of Rousseau and Marx. According to Berlin, negative liberty genuinelyadvances human well-being, whereas positive liberty contributes to oppressivestate domination and ultimately to the demise of negative liberty itself. How-ever, the distinction Berlin draws, whatever its pertinence for intellectual his-tory, rests on a metaphor that is easily overdrawn. Proponents of negativeliberty are said to be concerned with the area in which individuals are free,whereas proponents of positive liberty focus on the source of control. But thesource of control is always a concern for partisans of the liberties Berlin extols,and the area of control is an issue for even the most ardent proponents of pos-itive freedom. What Berlin has identified, at most, are strains of modern po-litical and social theory that emphasize different aspects of the same generalconcept - a point argued incisively by Gerald MacCallum in his paper "Neg-ative and Positive Freedom," Philosophical Review, vol. 76 (1967), pp. 312-334. Ifthere is any merit to Berlin's hesitations about positive freedom, it is becausethe strain of political theory he opposes does tend, in some of its varieties, toconflate autonomy with rational self-determination in a way that arguablydoes suggest "totalitarian" usurpations of freedom in the name of control ofthe less by the more rational. But it is incorrect to impute these views to Rous-seau or Marx, though perhaps not to some of their followers. More impor-tantly, it is hardly necessary to construe autonomy in the way Berlin does. See,for example, C. B. Macpherson, "Berlin's Division of Liberty," in DemocraticTheory: Essays in Retrieval (Oxford: Oxford University Press, 1973), and Levine,Liberal Democracy, Chapter 10.

12 It is commonplace, following Hobbes's example, to suppose that we are unfreein this sense only in consequence of what other human beings do - typically,in consequence of what they do deliberately. Bare nature does not render usunfree.

13 See Levine, Arguing for Socialism, Chapter 1, for an elaboration of some of theideas sketched in this paragraph and the next.

133

The general will

attached to doing what one wants supposes that one's wants areone's own choice. In short, whoever values Hobbesian freedomought also to value autonomy as well. For liberals as much as forRousseau or Marx, individuals are free, in the final analysis, inso-far as they are the authors of their own projects.14

Rousseau's use of "autonomy" suggests that autonomy is some-thing individuals either have or lack. Whether or not this was hisconsidered opinion, it is not a sustainable position. If freedom isconceived as the ability to exercise self-governing capacities likereasoning, imagining, or planning, it plainly admits of degrees.Autonomy is therefore something of which there is more or less;not all or none. It is even unclear what it would mean for agents tobe the sole authors of the ends they set before themselves. But au-tonomy can be enhanced or impeded. The degree of autonomyagents attain will be, in part, a consequence of the social order inwhich they live. Institutions foster the capacity for autonomousdetermination or thwart it and also limit the degree of autonomyan individual can be expected to achieve. The real world of auton-omy is socially relative and, since societies are irreducibly histor-ical, historically relative too.

As remarked, Mill supposed that there exists a long-range sec-ular tendency in human history for institutions to enhance indi-viduals' autonomy.15 In On Liberty, where this idea is most evident,it is more assumed than elaborated theoretically. But it is not dif-ficult to supply a rationale consistent with the overall commitmentof On Liberty to the Enlightenment idea of progress in human his-tory. With expanding technological capacities, the material condi-tions for expanding human beings' abilities to do what they wantare, in general, advanced. Of course, it is one thing to say that a

14 Most liberals, of course, do endorse autonomy, even if they also evince supportfor the distinction Berlin draws between negative and positive freedom. If, asI have proposed (see note 11), Berlin's distinction is understood descriptively,as a claim about how freedom has been traditionally construed, rather than, asBerlin would have it, a claim for the existence of two distinct concepts of free-dom, then the commonplace liberal view is consistent with the position I amproposing. It is not clear, however, that my position is in fact what liberals gen-erally intend: whether, in other words, they do not somehow conceive auton-omy as a value distinct from freedom or even whether they do not sometimesconflate autonomy and negative liberty.

15 Cf. Thompson, John Stuart Mill and Representative Government, Chapter 4. Mill'scommitment to progress is particularly evident in Chapter 2 of On Liberty, "TheFreedom of Thought and Opinion," where the overall trajectory of human his-tory is depicted as a story of technological and moral progress.


certain degree of autonomy is materially possible and somethingelse to maintain that this potentiality is in fact achieved in real his-tory. Mill believed that, in the long run, the ever-advancing ma-terial conditions for freedom prevail over impediments in the wayof freedom's advance. Therefore, Mill thought that, as a matter ofempirical fact, human beings become ever more free. Moreover,with the growth of individuals' abilities to do what they wantcomes an enhancement of their capacity to conceptualize possibil-ities and therefore to set ends before themselves. Thus the level ofautonomy feasible at particular moments in human history risestoo. This broadly technological account of history's driving forceis, of course, essentially the idea Marx developed into a full-fledged theory of history, historical materialism. In its classical for-mulations, historical materialism's focus is on the rise and fall ofeconomic structures, not on the enhancement of freedom per se.But autonomy is central to Marx's philosophical anthropology; andit is clear that, in Marx's view, the general human interest in ad-vancing autonomy is implicated in real-world class struggles.16

Like Mill, Marx thought that, in the long sweep of human history,despite setbacks and reversals, autonomy is, in fact, advanced. InMarx's Hegelian moments, especially in his early writings, theprogress of freedom is not construed linearly, as On Liberty sug-gests, but "dialectically." Thus, in On the Jewish Question,17 the

16 For an extensive listing of corroborative citations, see Rodney G. Peffer, Marx-ism, Morality, and Social Justice (Princeton: Princeton University Press, 1990),Chapter 1. Here follow two illustrative passages, the first from The German Ide-ology, reprinted in The Marx-Engels Reader, ed. Robert Tucker (New York: W. W.Norton, 1972), pp. 161-162; the second from Capital, vol. 3 (New York: Inter-national Publishers, 1967), p. 820

"With the community of revolutionary proletarians . . . who take their con-ditions of existence and those of all members of society under their control . . .it is as individuals that the individuals participate in it. It is just this combina-tion of individuals (assuming the advanced stage of modern productive forces,of course) which puts the conditions of the free development and movement ofindividuals under their control - conditions which were previously abandonedto chance and had won an independent existence over against the separate in-dividuals just because of their separation as individuals, and because of thenecessity of their combination which had been determined by the division oflabour, and through their separation had become a bond alien to them."

"The realm of freedom actually begins only when labour which is deter-mined by necessity and mundane considerations ceases. . . . Beyond it [therealm of necessity] begins that development of human energy which is an endin itself, the true realm of freedom, which however, can blossom forth onlywith the realm of necessity as its basis."

17 Marx and Engels, Collected Works, vol. 3.

135

The general will

French Revolution and its aftermath are depicted as an advance for"political liberty" but also as a cause of an intensification of un-freedom ("alienation") in the sphere of production and through-out "civil society."18 Ultimately, however, even this "negation"works toward freedom's further realization. It helps move human-kind toward communism, "the realm of freedom," where auton-omy is so far advanced in comparison to conditions in capitalistsocieties that it is fair to speak of a qualitative breakthrough.

Neither Mill nor Marx directly proposed a mechanism thatwould account for autonomy's progress in real history. But it isplain that an explanation consonant with the overall position ofboth thinkers - with Mill's historicized Aristotelianism and withMarx's declared philosophical anthropology - would advert, ulti-mately, to their shared conceptions of human nature. The idea thatautonomy is advanced as history proceeds is not, for eitherthinker, only a general hypothesis about the trajectory of historicalchange. It is not a "fact" that could have failed to obtain, "takingmen as they are and laws as they might be." For both Mill andMarx, this outcome is necessitated by a transhistorical disposition,inscribed in human beings' mentalities, that propels them to seekto become ever more free. This idea will be further elaborated inChapter 7. For now, it will suffice to remark that, for Mill and Marxas for Rousseau, it is a salient fact of human existence, of tremen-dous pertinence to political philosophy, that "man is born free."Rousseau, who declared this inexorable fact, doubted that humanbeings could ever exercise their birthright in real history; Mill andMarx thought otherwise.

SELF-REALIZATION AND THE GOOD

In just the way that claims about the value of autonomy depend,ultimately, on claims about human nature and other matters offact, views about the character and importance of self-realiza-tion are also empirical claims. As such, they are open to disputein light of new evidence or reinterpretation of evidence alreadyavailable. In principle, therefore, substantive assertions aboutself-realization are necessarily provisional and tentative. Never-theless, many conclusions can already be drawn with a high de-gree of confidence.

18 I argue in support of this claim in "Alienation as Heteronomy," PhilosophicalForum, vol. 8, nos. 2-4 (1978), pp. 256-268.

136


To this end, it will be helpful to distinguish spontaneousconceptions of the good from the totalizing, Careerist concep-tions widely invoked by liberal philosophers. A spontaneousconception arises "spontaneously" in the way Lenin thought"trade union consciousness" did - not without reflection butmainly in consequence of individuals' lived experiences. Careeristconceptions then integrate spontaneous theories into a single,overarching view - like class consciousness does, according toLenin. Careerism, like class consciousness, is an imposition onwhat arises spontaneously. But unlike class consciousness in theLeninist scheme, Careerism is not a welcome integration ofspontaneous conceptions. It can, in fact, be detrimental to humanfreedom.

Implicitly, the cardinal virtue of liberal societies, as Mill con-ceived them, is precisely tolerance for divergent spontaneousconceptions of the good. In practice, spontaneous conceptions dif-fer enormously. No doubt, inequalities in the distribution of re-sources and in opportunities for cultivating innate capacitiesaccount for some of this divergence. But it seems plain, even so,that to some degree people's values and judgments are irreduciblydiverse and that spontaneous conceptions of the good are there-fore bound generally to differ. It would be rash to conclude thatthey are therefore unconstrained. But they are surely not soconstrained that, other things being equal, they will invariablyconverge.

The idea that individuals' valuations and ultimately their spon-taneous conceptions of the good are bound to be (somewhat) atodds raises the question of the extent and nature of the constraintsthat govern judgments of value. Mill's discussion of higher andlower pleasures in Utilitarianism is not quite focused on this issue,but what he says about higher and lower pleasures is germane toit. We know with considerable assurance that there are, as Millbelieved, qualitative differences among the things we value.19 Inclarifying how this distinction can be made, drawing on Mill'saccount, some purchase can be gotten on the constraints thatlead spontaneous conceptions of the good to converge, while

19 In fidelity to the principles of (Bentham's) hedonistic psychology, Mill calledthe objects of value "pleasures." However, in the account that follows, nothinghinges on these things actually being pleasures in Bentham's sense. FollowingMill's practice, "pleasure" will here be used interchangeably with "value," butwithout any implication that the only thing valuable or the only thing peoplein fact value is pleasure.

The general will

remaining irreducibly heterogeneous. A deeper understanding ofthis situation will help, finally, to clarify the nature and historicityof self-realization.

HIGHER AND LOWER PLEASURES

To turn from spontaneous conceptions of the good to higher andlower pleasures in Mill's sense is, if nothing else, to focus on finer-grained judgments. In order to become clear about how thesejudgments are (psychologically) constrained, it will help to reflecton value judgments generally.

Judgments of taste provide an instructive example. Connois-seurs of, say, wine will usually agree about the ranking of wineswith respect to quality - at least in general. But within broad lim-its, particular judgments may vary, the variations becoming morepronounced the finer the discriminations involved. There is, inother words, a tendency toward consensus about gross differencesthat becomes less pronounced when finer judgments are made. Atthe limit, there is no disputing taste. But people with taste are gen-erally in accord - even if disputes about "fine points" cannot al-ways be resolved to everyone's satisfaction.

One might explain this phenomenon simply by the internaliza-tion of norms. The connoisseur would then be no better equippedthan anyone else to determine what is "really" better or worse.Perhaps there is nothing really better or worse to determine. Ex-perts would only have learned better than others what experts aresupposed to say. Then judgments of taste would ultimately be ar-bitrary even when there is a consensus of expert opinion. The factof consensus would only establish the success of socialization. Itwould not address the reality of qualitative differences.

At the other extreme, it could be held that there are real differ-ences in merit among wines that at least some human beings,connoisseurs, are equipped to discover. The acquisition of the req-uisite skills undoubtedly requires training of an appropriate sort.Connoisseurs are more made than born. But, properly trained,connoisseurs detect differences of quality. They do not prescribethem. Then disagreements about wines among connoisseurswould most likely arise in consequence of the uneven develop-ment of their evaluative skills. Unless there are real indetermina-cies in the world of wine, equally developed connoisseurs shouldbe in complete accord.

138


It will be helpful, at this point, once again to invoke an Aristo-telian idea by proposing, as it were, a "golden mean" betweenthese views of connoisseurship. For an Aristotelian, there are "ob-jective" standards of judgment, but standards have to do bothwith how the world is and with the nature of human beings. Injust the way that there is a natural standard for, say, an iris - thatis, roughly, what an iris would become under ideal (or nearlyideal) conditions - there is, an Aristotelian would say, a standardof taste in wines that an ideally cultivated connoisseur would de-velop. Insofar as the faculty of taste (in wine) is roughly the samethroughout the human population, judgments about wines amonghuman beings, properly cultivated, should be roughly similar. Notall irises cultivated under ideal conditions will in fact look alike.Similarly, in all likelihood, not all judgments of taste will convergeeven in principle. Beings of the same natural kind in their naturalconditions need not be similar in all important respects. The de-gree of variation will vary from kind to kind. But, among humanevaluators, there will be enough similarity to undergird a recog-nizable and objective standard of taste. Therefore judgments oftaste are not arbitrary on this view; they are determined both bythe properties of the things judged and by human nature itself.How constraining human nature is must, of course, remain anopen question. But it is fair to speculate at least with respect towine connoisseurship that it will limit evaluations severely, andtherefore that the area of convergence among individuals withfull access to the means for cultivating their capacities will beconsiderable.

If, as seems likely, fine-grained value judgments would some-times fail to converge completely even when people are providedwith equal and sufficient access to the means for cultivating theirevaluative capacities, this fact would not fully explain why peoplediffer in what they choose to do. Differences in people's choicessimply do not depend sufficiently on differences in judgments ofvalue. Many of the difficulties philosophers encounter in thinkingthrough the ramifications of the distinction between higher andlower pleasures arise because, following Mill, they assume anoverly facile way of linking evaluative judgments with choices.The underdevelopment of an individual's intellectual and aes-thetic capacities can explain a preference for pushpin over poetry,for one of Mill's lower over one of his higher pleasures; and so canweakness of will. But for a cultivated person, weakness of will is

139

The general will

seldom the best explanation. In ways that remain largely unex-plored in moral psychology, judgments of value do affect our de-sires and therefore our preferences and what we choose to do. Butdesire is not, and ought not to be, the slave of value judgments.Arguably, if objects of desire could be sorted out into naturalkinds, then, when an individual desires something of a certainkind, it would be irrational to prefer what one judges to be worse.But there do not appear to be many natural-kind divisions amongobjects of desire, at least if they are specified in a fine-trained way.An individual may, at time t, judge wine A better than wine B butnevertheless prefer to drink B at t. Doing so would hardly warranta charge of irrationality. It is unlikely that this assessment wouldchange if A and B were grouped together under some finer cate-gory - whether or not the categorical determination somehowindicates a real division among wines. Perhaps the finer the cate-gories are drawn, the less likely it becomes that a cultivated personwould prefer something judged inferior. But this is an idle specu-lation awaiting a more careful investigation of the relation amongjudgment, desire, preference, and choice. For now, I would onlysuggest that there is no good reason in general to censure thosewho, knowing better, drink inferior wine or play pushpin whenthey could be reading poetry. Mill's writings sometimes reflectthis high-minded attitude. Contemporary liberalism exemplifies it.But there is also the liberalism of On Liberty, which advocates tol-erance and laissez-vivre not just because justice requires it but be-cause the active encouragement of diversity is instrumental foreach individual's learning what is required for his or her ownflourishing.

It is fair to speculate that the range of variation for individuals'(freely chosen) projects and, more generally, for their judgmentsabout the kinds of lives they want to lead will be greater than forjudgments of taste. There is surely no "right way" for everyone tolive, and probably no "right way" even for any particular individ-ual to conduct his or her life. Any of a range of public or privatepursuits and active or contemplative projects - and any of a vari-ety of mixtures of these elements - will probably "satisfice" forself-realization. In the nature of the case, there is no point in look-ing for an optimal mix, even if some sense could be attached to theidea. In view of human differences and the contingencies of life,the only way to advance, as best one can and in the face of sub-stantial indeterminacies, is through trial and error; in short,through practical experimentation.

140


EXPERIMENTALISM

The exploratory probes and tests of spontaneous conceptions ofthe good promoted by Mill in On Liberty and elsewhere are indis-pensable sources of insight for promoting human flourishing.What can now be concluded tentatively about higher and lowerpleasures is based on such experiments and on extrapolationsfrom the data acquired through them. Of course, judgments madetoday must also be based on informed speculation about what wewould learn from experiments conducted under more favorableconditions, where the means for cultivating innate but historicallydeveloping human capacities are widely - and, in the ideal case,equally - distributed. It is pointless to try to say what experimen-tation in such conditions might reveal.

We can, however, speculate with some assurance that there willbe considerable convergence of opinion on general matters, just asthere is now convergence on gross judgments about wines amongconnoisseurs, and similarly significant variations of judgment onpoints of "detail." We will probably be able to make progress indistinguishing pleasures (and whatever else is of value) qualita-tively, and we may be able to develop a sounder philosophical the-ory of the good and a better and more comprehensive moralpsychology than we now have. Most in point, I would venture thatthe Careerism dear to contemporary liberals would eventuallydrop out of fashion. For whenever an entire life or a significantpart of a life is viewed as a career, a long-range project that struc-tures preferences and choices, practical experimentation is need-lessly constrained. Of course, an overarching plan of life is notnecessarily detrimental to the spirit of experimentation. A singlecareer, if freely (autonomously) undertaken, can be as worthwhile,from an experimentalist standpoint, as projects of shorter dura-tion. But when institutional structures themselves encouragesingle-mindedness, many of the constraints that liberals opposewhen supported by the state or the moral coercions of public opin-ion are effectively reimposed internally. Careerism is a kind of in-tolerance in its own right, mfrapersonal intolerance, resulting inthe self-regulated discouragement of the diversity experimentationpromotes.

It would be well therefore for liberals to resist the idea thatthere need be one thing that is necessarily and permanently one'sown and to revert back to that side of Mil lean liberalism that ac-tively promotes experimentation. Insofar as Mill's thought is

141

The general will

representative, there is nothing essential to liberalism that under-writes an aversion to hunting in the morning, fishing in the after-noon, and criticizing at night - or, more plausibly, to the idea thatone should try many things but not identify one's self with any ofthem in particular. This picture, in its Romantic elaborations, istransparently at odds with what we know of human psychologyand the requirements of self-realization. Human flourishing willelude the dilettantes Marx depicts, just as it will elude individualswho unreflectively constrain themselves by pursuing plans of life.In a word, a dearth of high-mindedness is as wrong-headed as asurfeit. The point, in proper Aristotelian fashion, is to find themean. If it is the right to hold that Careerism is not essential toliberal tolerance - indeed, that it is opposed to the spirit of liberalexperimentation - then that is an indication of where the meanlies.

In any case, the objective, for both Mill and Marx and the os-tensibly opposing traditions they represent, is to promote humanflourishing. What is to be done to this end is ultimately an empir-ical question that is better addressed experimentally than a priori.What can be concluded now is that it is well to cast impedances toexperimentation aside, and that the Careerism that has becomeemblematic of liberalism today - like the Romanticism that afflictsMarx's speculations on communism - is an impedance of justthis kind.

THE LIBERAL STATE AND THE LAST STATE

The standard view today is that equal respect for persons man-dates neutrality at the political level. I have just attributed to Milla somewhat different justification for political neutrality that ap-peals to its role in promoting autonomy and self-realization and,more generally, to human flourishing in determinate historicalconditions. Both rationales suffice to justify liberal tolerance. Bothare compatible, ultimately, with Marxist views. But the rationalethat appeals to equal respect, more than the appeal to autonomyand self-realization, is in evident tension with the Marxist idea ofthe last state. For that state is a "dictatorship" in which the work-ing class and its allies, the vast majority, deploy state power tosuppress capitalism and its vestiges and therefore also those con-ceptions of the good for which capitalism and its vestiges are in-tegral. For Marxists, equal respect is a feature of ideal social

142


arrangements. But its implementation in class-divided societies ispremature and even dangerous. It becomes apt only after class di-visions and therefore class dictatorships are superseded.

This conclusion is sound and important, but it is also misleadingbecause it exaggerates the sense in which a proletarian class dic-tatorship must be illiberal. To see why, it is well to reflect again onthe Millean rationale for tolerance, and on the hasty and unwar-ranted optimism evident in traditional Marxist formulations.

For Mill, neutrality in the political arena is justified because itpromotes autonomy and self-realization. But these effects are con-tingent and, by Mill's account, historically variable. Were the factsof the matter different, as Mill believed they had been throughoutmost of human history and as they still were throughout much ofthe world, the implications for liberal tolerance would differ ac-cordingly. Thus Mill expressly denies that tolerance, say, of speechis appropriate in societies in which institutions do not (yet) suffi-ciently cultivate individuals' capacities to allow them to functionas competent consumers in the marketplace of ideas, in short, inmost human societies. Neutrality, then, is not a matter of right, asin arguments based on equal respect, but a means for advancing"the permanent interests of man as a progressive being" underparticular historical conditions. From this perspective, it followsthat the Marxist complaint that universality in class-divided soci-eties is a sham and a mechanism for reproducing class domina-tion is more compatible with Mill's case for tolerance than firstappears. If Marx was right, before communism, a state of "thewhole people" is a dangerous illusion, and it may sometimes bejustifiable to counteract the dangers of this illusion by acting incontravention of what strict universality requires. But it does notfollow from this conviction that, this side of communism, tolerancemust be systematically withheld or restricted or selectively ap-plied. What follows is that "What is to be done?" is always an openquestion, to be addressed by considering the implications of alter-native proposals on "the permanent interests of man as a progres-sive being." If there is a disagreement between Mill and Marx, ithas to do with their respective views of what the evidence showsor is likely to show with respect to this question.

It is only because their empirical speculations differ that, in thefinal analysis, the liberal state on Mill's construal differs from thelast state of Marxist theory. The liberal state institutionalizes im-munities from state and societal interferences; the last state, like

M3

The general will

the state defended in The Social Contract, relies on transformativedemocracy to achieve a similar effect.20 Ultimately, this differencefollows from empirical speculations based on different views ofhuman nature and therefore of human beings' potentialities.

In all its variants, liberalism is pessimistic about human beings'prospects for radically altering their forms of political associa-tion. The "permanent interests of man as a progressive being" canbe advanced indefinitely and, in consequence, humankind canprogress morally. But men and women will never advance quali-tatively beyond their present condition in a way that would renderstates unnecessary. Human nature, though improvable, remainsrecalcitrant to "the great leaps forward" Marxists envision. Henceliberal safeguards, to save us from ourselves and from each other,are in order not just now and for the foreseeable future but forever.If these safeguards conflict with democratic values, then democ-racy must give way or be transformed beyond recognition. Thiswas effectively Mill's view in Considerations on Representative Gov-ernment, even as he marshaled what amounts to a brief in supportof participatory democracy. The permanent interests of man as aprogressive being sometimes support a fusion of democratic andliberal institutions, but wherever the one opposes the other, asthey are bound to do, the trade-off is struck in liberalism's favor.

In contrast, the strain of Marxist theory that envisions a last stateis democratic to a fault. Its faith in the transformative power ofdemocratic institutions underwrites the conviction that somethingapproximating the "state" Rousseau described in The Social Con-tract can be established on earth, "taking men as they are and lawsas they might be," if only, pace Rousseau, private property in alienablemeans of production is abolished. I have already suggested that thisview is not so much wrong-headed as naive; that it underesti-mates, at great cost to human well-being, the difficulties in the

20 Cf. Levine, The Politics of Autonomy, Chapter 2, for a discussion of Rousseau'sattempt to realize liberal values without recourse to liberal institutional ar-rangements. It should be noted, however, that, no matter how successful trans-formative democracy may be in ensuring the outcomes that liberals seek toachieve "constitutionally," the state Rousseau envisions would remain illib-eral - in content as well as in form - because it assumes a direct and unmedi-ated relation between the individual and "the whole community," and becauseit construes group interests as private interests. Rousseau's de jure state wouldtherefore proscribe rights of association, at least insofar as associations can beexpected to lead to the establishment of groups - and group interests - thatmediate between the individual and the state, infirming the generality of eachindividual's will.

144


way of transforming human nature. In consequence, and in con-trast to the received understanding but not the spirit of Rous-seauean and Marxist political theory, the last state should beliberal and should remain liberal - for as long as need be. Paradox-ically, liberalism is indispensable for the dictatorship of the prole-tariat, the state in transition to communism. It is only after thatstate has succeeded in its mission, by undoing the conditions thatmake it necessary, that it is safe to allow liberal safeguards towither away completely.

But if it is necessary to retain and even enhance liberal protec-tions in order to prevent dictatorships of the proletariat from de-volving into tyrannies, it is also necessary to retain a sense of whatdistinguishes the last state from the liberal state, in order to keepthe former true to its mission. The last state must proscribe the im-plementation of conceptions of the good detrimental to the transi-tion to communism. To this end, it may sometimes need to denytheir advocacy too. In this respect, it is potentially illiberal fromthe start. But so is the state Mill proposes. If it is to reproduce theprotections it accords its citizens, the liberal state must prevent theimplementation of conceptions of the good that undermine polit-ical neutrality. Therefore it too may be obliged, when circum-stances require, to proscribe the advocacy of views that can beexpected to have this effect. A thoroughgoing liberalism is possibleonly in states whose institutions are secure enough to withstandunflinching tolerance. In less propitious conditions, the liberalstate may have no choice but to save itself by illiberal means. If it isnot to lose its soul, it must, of course, act in a measured way anderr on the side of tolerance. But it cannot allow its liberalism to beits own undoing.21

In short, proponents of autonomy and self-realization cannot al-low these aims to be subverted. This rationale for illiberal depar-tures from liberal norms applies equally to the proscriptions thelast state enforces. If it is true that communism carries the promo-tion of autonomy and self-realization to the limit of what is hu-manly possible, and if the Marxist account of the prehistory ofcommunism is sound, then the use of state power to prevent the

21 Similarly, Rousseau maintained that, in the face of dire emergencies, the dejure state may find it necessary to suspend its usual institutional apparatus andinstall "dictators" to rule by decree; see The Social Contract, Book IV, Chapter 6.Rousseau's account draws on Macchiavelli's discussion of the dictatorship inancient Rome in The Discourses, Book I, Chapter 34. I discuss Rousseau's viewof emergency government in The End of the State, pp. 53-57.

145

The general will

restoration of capitalism is as justified as its use by a liberal state toprevent the imposition of, say, an intolerant religious orthodoxy.In other words, there are no principled, theoretical differences be-tween the liberal state and the last state in this respect. At most,there are differing assessments of the likely consequences of dif-ferent social practices.

So long as it is securely on track, the last state need be no lesstolerant than the liberal state of the advocacy of views inimical to itsexistence or opposed to its historical tendency. But, in practice, thelast state must, of course, proscribe capitalist social relations, aprospect Millean liberals can hardly endorse. In this sense, a gen-uine proletarian class dictatorship is indeed less tolerant than lib-eral states. But it is so in virtue of what its proponents supposeabout the possible futures of existing and feasible social and po-litical arrangements. I would therefore suggest that it is misleadingto describe the differences between the last state and the liberalstate in terms of a lesser or greater commitment to political neu-trality. It is more apt to conclude that what distinguishes Milleanfrom Marxian politics is the fact that, unlike Mill, Marx was com-mitted to a theory of history that foresaw communism as a histor-ical possibility and radical democracy under socialist conditions asthe means for bringing it about. Having no commitment to com-munism and therefore no plan for its realization, Mill could, as itwere, proscribe less and still remain faithful to the justificatoryprogram he effectively shares with Marx. In short, Mill had no rea-son to propose that liberals not tolerate practical experiments in-compatible with socialism and radical democracy; Marx did. Mill'sstate is therefore less restrictive than Marx's. But the two states areequally responsive to the exigencies that make neutrality desirableand in that sense equally neutral.

Which ideal is more appealing? Some purchase on this ques-tion and on related themes can be gained by reflecting moredirectly on the connection between liberalism and Marxism, orrather the Rousseauean strain of Marxism that envisions stateless-ness under communism.

146

Chapter 7

Rousseauean Marxismand/versus liberalism

Marxist theory is based in large part on the writings Karl Marxproduced from the early 1840s until his death in 1883. But Marxismdid not become a discernible tendency until around the time of theParis Commune (1871) and was not a fully consolidated currentwithin the workers' movement until after Marx's death. In con-trast, the origins of modern liberalism can be traced back at least topleas for religious toleration occasioned by the Protestant Reforma-tion and to theories of limited government that arose almost con-currently with the nation-state itself. By the mid-eighteenthcentury, more than a hundred years before Marxism's classicalperiod, liberalism was a self-conscious intellectual and politicaltendency.

It is therefore only within the past century that Marxism and lib-eralism have stood as ostensibly rival positions. However, it mightappear misleading to describe them as opposed, for throughoutthe past hundred years most Marxists have maintained that Marx-ism spurns the very idea of normative theory. If Marxists haveprincipled reasons for assuming this posture, then, strictly speak-ing, there would be no Marxist normative theory to rival liberalclaims. But at the same time that Marxists have disavowed norma-tive theory in general - and moral discourse in particular - theyhave also evinced an inclination, evident in Marx's own writingsand in the work of nearly all Marxists after him, to render moraljudgments. They have also insisted that normative discourse isideological and therefore ''superstructural," constrained, thoughnot strictly determined, by the "economic base" of the societiesthat produce and reproduce it. Throughout Marxism's history, ithas remained fatally unclear how constraining the economic baseis supposed to be. But it is compatible with even the strictest (plau-sible) construal that existing (capitalist) economic structures can be

M7

The general will

faulted by appealing to the normative standards they support.This position, like Marxists' penchant for moral condemnation ofexisting social, political, and economic arrangements, is perhapsconsistent with the claim that Marxism has no normative theory.But it is plainly in tension with it.1 We can therefore conclude that,whether or not Marxism includes a normative theory, Marxistshave always taken a normative stance. It is widely assumed thattheir stance, insofar as it is distinctively Marxist, is at odds withviews advanced by liberal social philosophers.

In the Second International and then under the aegis of officialCommunism, when orthodox historical materialism held sway, itcould be safely concluded that normative discourse was, atbest, an appendix to Marxist theory. For historical materialist or-thodoxy predicted capitalism's demise and communism's triumph.Normative arguments might be useful for encouraging class actorsto assume their historical destinies. They could hasten humanity'sfuture. But, in the end, they could not alter the outcome of aclass struggle driven by inexorable material interests, not moralconsiderations.

Needless to say, this conviction has lapsed as support haswaned for the orthodox account of the inevitability of capitalism'sdemise and communism's triumph. First, Western Marxists2 dis-tanced themselves implicitly from historical materialist orthodoxyin large part because of its claims for historical inevitability and thedevaluation of human agency that this claim was thought to

1 For accounts of Marx's view(s) of morality, see, among others, Lukes, Marxismand Morality; Richard Miller, Analyzing Marx: Morality, Power, and History (Prince-ton: Princeton University Press, 1984), pp. 15-97; a n d G. A. Cohen, "Review ofKarl Marx by Allen W. Wood," Mind, vol. 92 (1983).

2 I use the term "Western Marxism" in the widely accepted sense made currentby Merleau-Ponty and Perry Anderson. See Maurice Merleau-Ponty, The Adven-tures of the Dialectic, trans. Joseph Bien (Evanston, 111.: Northwestern UniversityPress, 1973), pp. 30-58; and Perry Anderson, Considerations on Western Marxism(London: NLB, 1976). Roughly, the term denotes that current of theorizing thatruns through the work of Georg Lukacs, Karl Korsch, Antonio Gramsci, the"critical theorists" of the Frankfurt School (Adorno, Horkheimer, Marcuse, etaL), existentialist Marxists (Sartre, Merleau-Ponty), structuralist Marxists (Alth-usser, Balibar), and so on. Politically, Western Marxism is oppositional with re-spect to the official Marxism of the Soviet Union and the Western EuropeanCommunist parties - though, in some cases, only implicitly. Philosophically,Western Marxism is shaped in varying ways by "continental" philosophicalcurrents - neo-Hegelianism, above all - and tends to focus programmaticallyon grand reconstructions of Marxian philosophy.

148

Rousseauean Marxism/liberalism

imply.3 More recently, analytical Marxists have identified internalshortcomings of the orthodox theory.4 In consequence, for anyonewho shares Marx's political objectives, normative issues havecome to take on a political urgency absent from classical Marxism.

Historically, liberalism - as a theory of principled limitations onthe use of public, coercive force - has served to defend and evenjustify (capitalist) property relations and (capitalist) markets. As anevolving political current, it has therefore usually opposed Marxistpolitical aspirations. However, liberals have long evinced a com-mitment to equality in at least some sense of that deeply contestedterm. By the time of the Enlightenment, if not before, liberal the-orists were committed in theory - if not always in practice - notjust to the idea that, in the respects that matter fundamentallyfrom a normative point of view, "all men are created equal"5 butalso to the conviction that the moral equality of persons entails thatthey be treated equally by basic political and social institutions.6 In

This inference can, however, be resisted. Cf. G. A. Cohen, "Historical Inevita-bility and Revolutionary Agency/' in History, Labour, and Freedom, pp. 51-82.Nevertheless, for virtually all Western Marxists except the Althusserians, his-torical materialism - insofar as it proclaimed the inevitability of socialism andcommunism - was at fault for diminishing the role of human agents in epochalsocial transformations. The Althusserians were second to none in discountingthe role of human agency. For them, historical materialism's flaw was its "evo-lutionism" - by which they seem to have meant its claim to have identified anendogenous dynamic internal to social structures. Thus, unlike most WesternMarxists, what Althusserians opposed in historical materialism actually is cen-tral to the theory.See Cohen, History, Labour, and Freedom, Chapters 6-9. For an extended discus-sion of orthodox historical materialism and some reasons to be wary of its ex-planatory pretensions, see Wright, Levine, and Sober, Reconstructing Marxism,Chapters 2-5.Liberals have sometimes encouraged and seldom actively discouraged societalnorms that relegated the laboring poor, women, and native peoples to an infe-rior standing. Also, liberals once were remarkably sanguine about the institu-tion of slavery and, later, about institutionalized racism. Contemporary liberalsare less blatantly hypocritical, but it can hardly be denied that a gulf betweenthe theory and practice of existing liberal states persists to this day.The (fundamental) moral equality of persons does not strictly entail equal treat-ment. Thus the medial church held all human beings equal "in the mind ofGod" - that is, with respect to the possibilities for salvation or damnation orother statuses deemed to matter fundamentally. But this conviction was neverthought to entail that individuals be treated equally by social and politicalinstitutions.

149

The general will

practice, equal treatment was usually understood to imply onlypolitical equality - equality before the law, equality in renderingcollective choices, in short, equality of citizenship. It was thisidea that Marx inveighed against so vehemently as he moved "be-yond" his own earlier liberal convictions.7 However, Marx's objec-tion was not to political equality per se but to the idea that it issufficient for treating persons as equals. Today, political equality isalmost everywhere endorsed. To be sure, inequalities of citizen-ship persist. But hardly anyone is so reactionary as to defend themin principle.

Most liberals are also committed to equality of opportunity. Butthe consensus around this idea is more deeply contested than theconsensus around political equality: Everyone is for equality of op-portunity, but what equality of opportunity is is very much a mat-ter of dispute. At one extreme, some "right-wing" defenders ofthe idea maintain that equality of opportunity is achieved when le-gal and customary impedances are removed against individualscompeting on a meritocratic basis for scarce and widely desiredbenefits. On this understanding, equality of opportunity is tanta-mount to political equality.8 At the other, "left-wing" extreme, arethose who think that to attain equal opportunity individuals mustbe provided with equal resources. On this understanding, equalityof opportunity is more or less tantamount to resource equality.9 In-termediate positions too can be imagined. But inasmuch as theright-wing view has come to seem increasingly hollow under real-

7 See, especially, On the Jewish Question (1843), m Marx and Engels, CollectedWorks, vol. 3. Marx argued that "political emancipation," establishing politicalequality, actually exacerbates inequalities in "civil society" - thereby enhancingoverall human unfreedom.

8 In practice, many on the right oppose state interventions in support of equalityof opportunity (for instance, state monitoring of private firms' employment pol-icies and state remedies against discrimination) more vehemently and moreopenly than they oppose similar remedies employed to redress violations of or-dinary citizenship rights. Typically, efficiency considerations are invoked insupport of proscribing state interventions in the economic sphere, consider-ations that would be deemed inappropriate in the political arena. In this sense,equality of opportunity, construed in the right-wing way, is not quite equivalentto political equality.

9 On some analyses, for instance Ronald Dworkin's, resource equality entails thatthere be (ongoing) compensation for the (morally arbitrary) unequal distribu-tion of talents across individuals. Then resource equality would differ from eventhe most left-wing construal of (meritocratic) equality of opportunity. See Ron-ald Dworkin, "What is Equality? Part Two: Equality of Resources," Philosophyand Public Affairs, vol. 19, no. 4 (1981), pp. 283-345, but cf. esp. pp. 310-311.

150


world conditions, liberal opinion has tended to converge towardthe left-wing pole.10

It has also become characteristic of liberal political culture in re-cent years to acknowledge a presumption in favor of full materialequality, albeit a presumption that is easily overridden.11 Thus it iswidely thought that, other things being equal, benefits and bur-dens should be equally distributed. But, of course, other thingsare seldom equal. In general, two kinds of considerations typicallyswamp the presumption for equal distribution: those that are in-ternal to a theory of justice broadly construed, and those that arenot. In the former category lie property rights and other "entitle-ments" and also merit, desert, effort, productive contribution, andany other substantive considerations that liberal writers mightinvoke to defend the justice of unequal distributions. Inasmuchas the boundaries distinguishing justice from other normativestandards can be vague, it is not always clear what would countas an external reason for overcoming the presumption for equaldistribution. Efficiency demands, on some construals, are goodcandidates;12 perhaps freedom is too.13

In any case, it is plain that liberalism has evolved a strain that isunabashedly egalitarian, and therefore in tension with the liberaltradition's long-standing support for capitalism and hostility toradical politics. Needless to say, egalitarian liberalism is more

10 Cf. Rawls, A Theory of Justice, pp. 75-90; an earlier, still classic formulation ofthis position is given by Bernard Williams in "The Idea of Equality," in Philos-ophy, Politics, Society, 2nd series (New York: Barnes and Noble, 1962).

11 Cf. Rawls, A Theory of Justice, pp. 100-108. The Difference Principle, accordingto which inequalities are justified if and only if they work to the advantage ofa "representative member of the least well-off group," illustrates this presump-tion perspicuously. It is, of course, controversial what full material equalitymeans. Rawls's account is expressly resourcist. His concern is with the distri-bution of certain resources - those that are instrumental for the realization ofparticular conceptions of the good likely to be entertained in actual societies.Rawls calls these resources "primary goods"; they include income and wealth,powers and offices, and the bases for self-respect. The "basic liberties" are alsoprimary goods, but, for reasons that need not detain us here, Rawls wouldhave them distributed equally and to the greatest extent possible.

12 Cf. Arthur M. Okun, Equality and Efficiency: The Big Trade-off (Washington,D.C.: Brookings Institution, 1975). But see Julian Le Grand, "Equity versus Ef-ficiency: The Elusive Trade-off," Ethics, vol. 100 (1990), pp. 554-568.

13 Strictly speaking, what freedom may offend are views of justice according towhich a just distribution is one that conforms to - or approximates - a certainstructure. Cf. Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books,1974), esp. pp. 160-164.

The general will

theoretical than practical. Few, if any, liberals are expressly anticap-italist, and even the most radically egalitarian liberal visions as-sume continuity with existing political and social arrangements.But, as a social philosophy, liberalism has developed in ways thatrender it more compatible with political positions traditionallyheld by radical egalitarians and Marxists than was previouslybelieved.

Thus (some) liberals have moved closer to Marxists politically,just as (some) Marxists have become more engaged with liberalphilosophical concerns. Given their respective histories, it is notsurprising that the emerging liberal-Marxist "dialogue," on itsphilosophical side, has unfolded mostly in a context familiar fromliberal political and moral philosophy. In consequence, the old is-sue about Marxist normative theory - whether it exists at all - hasgiven way to a new problem: whether it can still be maintainedthat Marxism and liberalism are at all opposed. Marxism may noteschew normative theory in the way that Marxists used to claim,but it has lately become doubtful whether in the domain of nor-mative theory anything distinctively Marxist actually exists -except perhaps as a wrinkle, among others, in a universe of dis-course shaped by liberal concerns.

As remarked in Chapter 6, Marxists differ from liberals in theirviews about the nature of society and the structure and directionof historical change. Thus Marxists envision a possible communistfuture that liberalism implicitly denies. Nevertheless, liberalism orat least the version of it articulated by Mill is nearly of one mindwith the strain of Marxist political theory that underwrites claimsfor the end of the state. But the two are not entirely in accord.Marxism bases its case for statelessness partly on faith in the trans-formative powers of genuinely democratic participation, a faithMill does not fully share. We have seen how this difference hasimportant consequences for political theory. It is the burden of thischapter to show that it also indicates an important underlying dif-ference at the level of moral theory - a difference having to do withconceptions of the person and ultimately with autonomy itself.

WHAT DO LIBERALS WANT?

Since its inception, liberalism in political philosophy has been de-fended alternately by appeal to utilitarian, contractarian, or rights-based moral philosophical arguments or some mixture of these

152


positions. Each kind of argument takes individuals' interests, con-ceived atomistically, as its point of departure. But each treats theseinterests differently. Utilitarians suppose that individuals' inter-ests are properly served when well-being is maximized.14 If anindividual's well-being is well defined and representable quanti-tatively, such that it can be meaningfully added to the representa-tions of other individuals' well-beings, then the utilitarian seeks tomaximize the logical sum of these magnitudes. Social and politicalarrangements are justified, on this view, to the extent that theymaximize this quantity.15 Contractarians treat individuals' inter-ests in a more complex way. In their view, social and political ar-rangements properly serve individuals' interests if and only if, insuitably characterized situations, rational individuals - aware oftheir own interests and capable of adapting means to ends -would choose them freely.16 Usually, the choice is made from astate of nature or, to use Rawls's more apt expression, an "originalposition."17 To imagine an original position is, as Rawls wouldhave it, to discover an "Archimedian point" - neither so richlycharacterized as to bias the outcome of individuals' choices nor soimpoverished as to render determinate choices impossible. The ex-istence of such a vantage point is an indispensable - and disput-able - condition for the deployment of this argumentative

14 Utility measures well-being. Insofar as utility functions can be constructed in-trapersonally - and then made interpersonally comparable - the point of util-itarian calculations is to realize as much utility as possible in a givencircumstance. As remarked in Chapter 6, Mill's principle of utility - with itsfocus on human flourishing - stands somewhat apart from the mainstreamutilitarian tradition.

15 Historically, utilitarians have sought to maximize aggregate utility. Some utili-tarians today favor maximizing average utility. The distinction can be of somemoment, especially for debates on population policy. A very large number ofpeople with very low levels of well-being might have greater aggregate utilitythan a smaller population with higher levels of well-being. But in the smallerpopulation the average level of utility would be higher. Cf. J. J. C. Smart andBernard Williams, Utilitarianism: For and Against (Cambridge: Cambridge Uni-versity Press, 1973); and Parfit, Reasons and Persons, Chapters 17-19.

16 Cf. Russell Grice, The Grounds of Moral judgment (Cambridge: Cambridge Uni-versity Press, 1967), and T. M. Scanlon, "Contractualism and Utilitarianism,"in Amartya Sen and Bernard Williams (eds.), Utilitarianism and Beyond (Cam-bridge: Cambridge University Press, 1982).

17 Cf. Rawls, A Theory of justice, pp. 260-263. On differences between Rawls's"original position" and Rousseau's account of the state of nature, see Chapter2 herein.

153

The general will

strategy.18 Rights-based theories, finally, consider certain inter-ests - those that rights claims express - to take precedence over allother considerations. Social and political arrangements are justi-fied, for rights-based theorists, if and only if they accommodate tothese antecedently acknowledged claims.

These argument types, though analytically distinct, are some-times jointed together. Thus one might hold, as many rights the-orists do, that utilitarian calculations ought indeed to governpublic policy debates, except where rights considerations interfere.Then rights take precedence over utilitarian calculations, trump-ing what the principle of utility instructs.19 Or one might argue, asLocke did in The Second Treatise of Government, that the state isfounded by a social contract but that the nature and extent of thecontract that founds the state are limited by the rights individualsantecedently hold in the state of nature. Each of these mixed po-sitions should be distinguished from arguments that assign rightsin consequence of utilitarian or contractarian principles.20 Virtuallyall modern political philosophies ascribe rights to individuals inorder to specify what they may or may not do, or what may or maynot be done to them. However, in genuine rights-based theories ortheories that fuse rights-based considerations with considerationsof other kinds, rights are fundamental. They are part of the prim-itive moral landscape, not consequences of principles that are yetmore basic.

Again, these arguments share the conviction that the individual,conceived atomistically, is a point of departure for thinking aboutsocial and political arrangements normatively. The atomic individ-ual is also a point of departure in The Second Discourse and in the

18 Marx, for one, can be taken to deny its existence. On this issue, see MichaelTeitelman, 'The Limits of Individualism," Journal of Philosophy, vol 69 (1972),PP- 545-556-

19 Therefore what is rightful can diverge from what is good. Thus rights theoristswho are also utilitarians could hold that distributions that maximize utility arerequired by their notions of the good but that it would be wrong to implementthis good by means that violate rights. Imagine, for example, a rights theoristlike Robert Nozick in Anarchy, State, and Utopia, who holds that (idealized) cap-italist markets generate distributions in which everyone has a right to his or herdistributive share. This (imaginary) Nozickian might also be a utilitarian whowould have resources distributed in accord with the principle of utility. If thisgood could not be achieved except by violating (property) rights - say, by state-organized redistribution of market-generated distributions - then the goodwould have to give way, as it were, to the right.

20 Cf. Russell Hardin, Morality within the Limits of Reason (Chicago: University ofChicago Press, 1988).

154


opening pages of The Social Contract. But in the state Rousseau'ssocial contract founds, the atomic individual, moved by a privatewill, gives way to the citizen, directed by a general will. For citi-zens, the atomistic metaphor is therefore inappropriate. The dif-ference is particularly salient when the Rousseauean model ofgeneral will deliberation is contrasted with the deliberations utili-tarians envision.

Both Rousseauean and utilitarian deliberations contrast with or-dinary ways of understanding political processes. In the main-stream political culture, legislation is conceived as a way tocombine the competing private interests of individuals andgroups.21 In the state of nature, according to Rousseau's story, atthe point that culminates in the establishment of political commu-nities, these interests are mortally opposed. But in settled politicalcommunities there is an overriding interest, on the part of nearlyall individuals and groups, in maintaining the political communityitself. There is therefore a common interest in supporting a frame-work within which everyone can compete. But within this over-riding consensus, individuals and groups are typically at odds.Thus the legislative process is like bargaining. Usually, all sides ina negotiation have an interest in striking a bargain, but within thatconsensus everyone is in competition with everyone else. As such,there is no fact of the matter about where a bargain should bestruck. The "right answer" is just whatever the parties agreeupon. The outcome will be fair if all interests are represented andthe bargaining reflects the relative strengths of the contendingsides, but it cannot be right or wrong. In short, justice can beserved in bargaining, but not truth. Insofar as legislation is con-ceived in this way, it can be more or less fair. But since there is nofact of the matter to discover, voting can never determine what weought to do independently of what we decide.

For Rousseau, on the other hand, voting in the assembly of thepeople is precisely a way to discover a matter of fact - what thegeneral interest is. It is not a way to combine private interests, evenwithin a general consensus favoring maintaining the politicalcommunity above all.22 Thus voting is more like a debate among

21 What follows concerns voting. If political representation (or delegation) is un-avoidable and if, as suggested in Chapter 5, representatives (or delegates)might better be chosen at random rather than by elections, then the selectionof representatives (or delegates) would, of course, fall outside the scope of theensuing discussion.

22 Cf. Levine, The Politics of Autonomy, pp. 59-72.

155

The general will

"disinterested" experts than like bargaining among self-interestedeconomic agents. It is a truth-discovery procedure. Rousseaueanvoters are not asked whether they prefer that a particular law beenacted but "whether or not it conforms to the general will that istheirs. Each man, in giving his vote, states his opinion on the mat-ter, and the declaration of the general will is drawn from the count-ing of the votes."23

For utilitarians, the aim is to maximize aggregate (or average)utility, where the utilitarian maximum is the logical sum of indi-viduals' utilities. Therefore utilitarian individuals, when they de-liberate about what they ought to do, express their opinion aboutwhat the principle of utility requires. But although Rousseaueanand utilitarian deliberations each aim at discovering a matter offact, the question utilitarians ask themselves is not the same as thequestion asked of citizens in a de jure state. For the utilitarian max-imum is the logical sum of individuals' utilities. In this sense, it isreducible or decomposable into the utilities of the individuals whoconstitute the political community. The general interest, however,is irreducible to the (private) interests of the individuals who con-stitute the state. Each individual, in entering into the social con-tract, becomes an indivisible part of a "moral and collective body."When an individual makes the general will his own, he is not con-cerned with what is best for some aggregation of (atomic) individ-uals that includes himself but with what is best for himself quacitizen and subject of a "moral and collective body" of which he isan "indivisible" part.

Imagine, again, a voter confronting a piece of legislation con-cerning highway construction: It must be decided, say, whethertwo cities should be linked by a new highway and, if so, where theroad should be built. On the mainstream view, the voters will havepreferences for the alternatives in contention and will registerthese preferences by their votes. If all (relevant) preferences arecounted and the voting is fair, the outcome will be just in the sensethat its content will reflect the distribution of preferences withinthe voting community. For a utilitarian, on the other hand, there isone outcome (or set of outcomes) that, if realized, would maximizesocial utility, the logical sum of individuals' utilities. It is this out-come that voting aims to discover.24 On this view, there is a right

23 The Social Contract, Book IV, Chapter 2.24 If there is reason to believe that a good way - or perhaps the only way - to

"discover" the utilitarian maximum is for individuals to make their own pri-vate maximizing choices, in disregard of their views about other individuals'

156


answer to the question "What ought we to do with respect to thishighway?" But the answer is right in virtue of the nature of indi-viduals' utility functions, not in virtue of any mind-independentfacts of the matter. For Rousseau, in contrast, there is a right an-swer to this question that is mind-independent in principle.25

There is, in other words, an outcome that is best for "the moraland collective body" individuals constitute; and these individualscould in principle fail to discover that outcome, even when theirpreferences are fully informed and adequately considered.

But is there, in fact, a right answer? Can we reasonably con-clude, say, that it is in the general interest that the highway be builtand that it be built through these fields rather than those? And, ifso, is it desirable that private interests be excluded from a decisionof this kind? I think that the answer to both of these questions isno. There is no reason in general why there must be a general in-terest on all questions and, still less, on all aspects of all questionsor why, when a general interest does exist, it alone should deter-mine what we ought to do. Especially where there are indetermi-nacies, there is no alternative, given Rousseau's exhaustive andmutually exclusive account of individuals' interests, but to accedeto private will determinations. Nevertheless, on many questionsthat could arise in the popular assemblies, there may be a generalinterest that bears on aspects of the issues in contention. In otherwords, there is reason to suppose that in some cases these questionscan be answered affirmatively.

Rousseau appears to hold that there always is a general interestand that it bears on all matters of legislation. But on this matter, asin so many others, he is equivocal. Thus it is suggested in The So-cial Contract that legislative assemblies should meet infrequently(though periodically) and that most questions of social policy -

utility payoffs, then there will be no practical difference between utilitarian vot-ing and mainstream voting. Economists since Adam Smith have held a similarview with respect to markets. For them, the unintended consequence of indi-vidual maximizing choices in idealized market arrangements is socially opti-mal (i.e., Pareto-optimal, taking preferences, production technologies, and theprevailing distribution of resources as given), and there is no other way reliablyto achieve this result. Thus they recommend reliance on the "invisible hand"of the market and fault economic agents who might be disposed to seek so-cially optimal outcomes by aiming for them directly.

25 Of course, there is likely to be a causal connection between what individualswant - or what they would want in ideal circumstances - and what is bestfor "the whole community" they constitute. There is, however, no logicalconnection.

157

The general will

perhaps including questions about highway construction andthe like - would be relegated to the magistrates who execute the(popular) sovereign's will, rather than to the sovereign peoplethemselves.26 Perhaps even Rousseau can therefore be enlisted insupport of the idea that the general will does not bear on all mat-ters of collective choice and in support of the inference that followsfrom this position: that private volition cannot be eliminated en-tirely from the administration of the de jure state. This conclusionis plainly compatible with the assignment of an indispensable rolefor general will coordination. But Rousseau is only slightly helpfulin ascertaining what this role might be. I would venture that, inorder to advance beyond his vague and equivocal suggestions, weneed to think more carefully about historically feasible approxima-tions of Rousseau's ideal; in short, we need to think more aboutcommunism and the role of general will deliberations in commu-nist societies. This issue will be resumed in Chapter 8.

What can be concluded at this point is that, by introducing theidea of general will deliberations into political philosophy, Rous-seau poured new wine into the old atomic individualist bottle.27 InThe Social Contract, the individual, conceived atomistically, is and isnot the fundamental constituent of the de jure state. On the onehand, the social contract is in the private interest of individualswith independent and antagonistic wills insofar as they aim to be-come the autonomous agents they potentially are. In this way, thede jure state is constituted by the wills of (atomic) individuals. Onthe other hand, the condition for realizing autonomy is member-ship in "a moral and collective body" in which individuals will thesame end, the general interest, and are therefore directed by a sin-gle will, the general will. When these conditions pertain, individ-uals relate to one another internally; they are no longer atomicindividuals but integral parts of "a common self" (un moicomtnun).28 Then their interests can no longer even be conceivedapart from the interests of the other individuals who compose this

26 See, for example, The Social Contract, Book III, Chapter 13.27 This is not the only example of Rousseau's penchant for representing new

thoughts in old, inherited forms. Even the social contract itself is not, strictly,a contract at all, for it is not an agreement among independently existing par-ties but, as Rousseau acknowledges (in Book I, Chapter 7, of The Social Con-tract), an agreement each individual "makes, so to speak, with himself." SeeLevine, The Politics of Autonomy, Chapter 1.


158


"self." Thus Rousseau articulated two quite distinct conceptions atthe same time - or rather weaved very different visions togetherinto a single story. In proffering an atomistic account of the genesisof political associations, he intimated a nonatomistic picture of thesocial order and of the state that superintends it.

According to this picture, individuals and their interests are stillfundamental. But individuals are not atomic individuals, and theirinterests are not conceived atomistically. This conception is foreignto utilitarianism, contractarianism, and rights theories. In brief,the general will - and the kind of person a general will directs -falls outside the conceptual horizons of the moral philosophiesthat undergird liberal political philosophy. In this crucial respect,liberalism and Rousseauean political philosophy part ways.Therefore, if I am right in insisting that Marxian communism isan essentially Rousseauean idea, liberalism and Marxism partways too.

AUTONOMY

Both liberals and Rousseauean Marxists accord autonomy pride ofplace. But, in consequence of the difference just identified, liberalautonomy is not quite the same as Rousseauean-Marxist auton-omy. The dissimilarity is easily overlooked, in part because liberaland Marxist conceptions have so much in common.29 But it isreal and can be made evident by focusing, again, on general willcoordination.

It is sometimes said that liberals are partisans of "negative" lib-erty only; that the freedom liberals value is just the absence of co-ercive restraint. On this view, individuals are free to the extent thatthere exists an area in which they are able to do as they want with-out the deliberate interference of others. No doubt, liberal writerscharacteristically do emphasize this aspect of freedom, sometimesto the detriment of more "positive" conceptions that focus on thesource of control. But, as argued in Chapter 6, it is unclear what thepoint of negative liberty is, except insofar as the absence of coer-cive restraints helps agents to do what they want. Thus even themost austerely negative conception of freedom suggests a more

29 Cf. Levine, Arguing for Socialism, Chapter 2. What follows corrects a misleadingimpression conveyed by the account I gave there of autonomy as a value aboutwhich procapitalists and prosocialists (including Marxists) agree.

159

The general will

positive view: that agents are free to the extent that they are able(so far as institutional arrangements affect their abilities) to do asthey please. But then even more positive conceptions suggestthemselves. The freedom to do as one pleases would hardly seemworth having if one's ends were instilled by others. Indeed, whyshould one value the ability to do as one desires except insofar asone values situations in which individuals are also the authors oftheir own ends? Thus a commitment to negative liberty, conceivedas an absence of political and social interferences in the pursuit ofone's ends, leads inexorably to a notion of autonomy, understood asthe ability, again so far as this ability is socially conditioned, to set(achievable) ends before oneself. Despite what may sometimes ap-pear, there is therefore no tension between liberals' characteristicfocus on negative liberty and the recognition, on the part of mostliberals, that autonomy is a fundamental value. The rationale fornegative liberty, far from opposing autonomy, actually supportsthe latter conception.

Liberal autonomy admits of degrees. Since the ends individualswill are conditioned by causal determinations, there are perhapsno entirely free (autonomous) acts. But individuals' choices can bemore or less autonomous; therefore social and political arrange-ments can be assessed according to how well they enhance or di-minish autonomy in this sense.

The autonomy liberals invoke is the autonomy that makes neg-ative liberty worth having. This autonomy bears the mark of itsatomic individualist origins. For partisans of negative liberty, it isindividuals, conceived atomistically, who are free if unrestrainedsocially in the pursuit of their ends. Similarly, for partisans of au-tonomy in the liberal sense, it is individuals, still conceived atom-istically, who set ends before themselves. On this view, there is nointernal relation between one individual's ends and another's andtherefore no reason, coincidence apart, why individuals, as auton-omous agents, should not find themselves at odds.

However, for Rousseau, since autonomy is maintained only ifthe citizen-subjects of a de jure state will the same thing, the gen-eral interest, autonomous agents cannot be at odds. Kant appro-priated this idea of Rousseau's political philosophy and developedit into the doctrine that rationality itself establishes a strong con-sensus on ends, a "harmony of rational wills."30 For Kant, reasondoes rule on the content of our ends - not directly but by proscrib-

30 Cf. Levine, The Politics of Autonomy, Chapters 1 and 2.

160


ing certain ends as un-universalizable or, what comes to the samething, incompatible with respect for moral personality as an end-in-itself.31 These proscriptions are stringent enough to preclude(antagonistic) oppositions. Thus practical reason by itself coordi-nates individuals' wills. To the degree that reason controls individ-uals' willing, individuals become part of a coherent, harmonioussystem. "The war of all against all," a consequence of unreason orrather of reason deformed and rendered private, gives way to anorder constituted by reason itself.

It is this idea that liberal opponents of positive freedom findmenacing. If freedom is identified with rational self-determination,with direction by a "rational self," it justifies control of the less ra-tional masses by their more rational leaders. At the limit, rationalself-determination underwrites "totalitarian" usurpations of free-dom for the sake of freedom itself.32 When the state or the partypresents itself as the privileged oracle of reason or the agent of adestiny reason requires, it establishes a claim to control the fate ofothers in order to make them free.

This is not the place to rebut this misunderstanding of Rous-seau's and Kant's idea.33 I would only register the observation thatthere are many, eminently avoidable steps on the way from a har-mony of rational wills to totalitarian usurpations of freedom forfreedom's sake and remark that the impetus for finding philosoph-ical origins for totalitarian politics seems to have been driven moreby Cold War preoccupations than by analytical exigencies. In anycase, it is not necessary to defend the side of Rousseau's thoughtthat Kant seized upon here because the aspect of general will co-ordination that is needed for defending communism as a norma-tive ideal does not rely on a distinctively Kantian notion of rationalagency. It is instead the non-(atomic)individualist side of Rous-seau's position that matters for our purpose - specifically, Rous-seau's conception of autonomy's social dimension.

31 Cf. Kant, The Foundations for the Metaphysics of Morals, Part 2.32 Berlin, "Two Concepts of Liberty/' comes perilously close to advancing this

view. It is more commonly associated with Jacob Talmon in The Rise of Totali-tarian Democracy (Boston: Beacon, 1952) and, of course, with Karl R. Popper inThe Open Society and Its Enemies, 2 vols. (Princeton: Princeton University Press,1962).

33 Some elements of a rebuttal can be found in Levine, The Politics of Autonomy,pp. 72-80, and Liberal Democracy, Chapter 10, and in an ancestor of that chap-ter, "The Political Theory of Social Democracy," Canadian Journal of Philosophy,vol. 6, no. 2 (1976). The relation between the position in question and Cold Warpreoccupations is addressed directly in that article.

161

The general will

HARMONY

A master's ability to set ends before himself owes in part to the la-bor performed by his slaves. Slaves relieve masters from burden-some toil, provide them with means for realizing their ends, and,most important, permit them to will ends that they could not fea-sibly entertain and might not even imagine were they bereft ofthese instruments of their will. At the same time, because they areslaves, the autonomy of these autonomy-enhancing agents is di-minished. Of course, liberals abhor slavery. Egalitarian liberalseven insist that basic rights and liberties be distributed equally(and to the greatest extent possible).34 But equal distribution of ba-sic rights and liberties is compatible with resource inequalities andinequalities of powers and offices that affect autonomy causally. Inprinciple, therefore, one individual can become more autonomousand another less so as an unintended consequence of inequalitiesthat even egalitarian liberals allow. This is not a conclusion liberalstypically acknowledge. But it is an implication of the individualis-tic view of autonomy that they hold.

Should a liberal seek to deny that someone's autonomy might beenhanced in consequence of the diminution of someone else's, itcould only be for reasons unconnected to autonomy itself. It couldbe that the world always or generally is such that, as Marx said ofcommunism, "the condition for the free development of each" is"the free development of all." But it would be incumbent on thoseproposing this claim to show why it is so here and now. Their bur-den would be to identify mechanisms capable of ensuring that theenhancement of one individual's autonomy cannot or generallywill not diminish anyone else's. It does not seem likely that suchmechanisms exist. It is more likely, as in the case of the master andslave, that the autonomy of some individuals will sometimes beenhanced at the expense of the autonomy of others.

In any case, so long as individuals are conceived atomistically, asindependent centers of volition, there can be no intrinsic reasonwhy autonomous wills should be in harmony - unless, as Kantmaintained, autonomous willing itself somehow guarantees har-monious outcomes. But there is little comfort for liberals in theKantian conception, despite its individualism, because liberals es-

34 For Rawls's account of basic rights and liberties, see A Theory of Justice, passim,but esp. Chapter 4.

162


chew substantive notions of rational agency. For liberals, followingHobbes, reason rules not on the content of our ends but only ontheir consistency and appropriateness for the extrarational ends weseek to bring about. In consequence, no matter how autonomousindividuals may become, their wills remain radically independentof one another's. Pending the discovery of extravolitional mecha-nisms capable of ensuring harmonious interactions among suchutterly distinct beings, it will always remain possible - and evenlikely - that individuals' wills will be opposed. Then the conditionfor the free development of each will generally not be the free de-velopment of all. Liberals would surely want to deny this eventu-ality. But they lack principled reasons for doing so.

Autonomy is perhaps as fundamental a value for liberals as forRousseau, and liberal egalitarians may be as committed as Marx-ists, perhaps more so, to advancing the autonomy of everyone.35

But if the world is bereft of mechanisms for ensuring equal respectfor the autonomy of all moral agents and if practical reason is too"thin" to secure a harmonious outcome, radically independent in-dividuals cannot in general be counted on to enhance their ownautonomy in harmony with other rational agents. Liberals find the (so-cial) world insufficiently mutable to seek harmony by transform-ing it, and they are averse to thinking of practical reason as morethan instrumental. They are also committed atomic individualists.Rousseauean Marxists differ from liberals on each of these counts.In Chapter 8, we will see how qua Marxists they envision a worldin which there is indeed reason to think that individuals will mu-tually support each other's autonomy. But qua Rousseaueans theyalso have resources for socializing (liberal) autonomy. We have seenthat one possibility, following Kant's example, would be to placethe burden for doing so on (practical) reason itself. However, as re-marked, this strategy, whatever its merits, raises a number of vex-ing questions that defenders of communism need not confront.This is why I have instead pursued a different strand of Rousseau'saccount of the general will and proposed, throughout these pages,

35 Marxists are committed to enhancing the autonomy of everyone under com-munism. But, as we have seen, the state Marxists envision as an indispensablemeans for arriving at communism is, like all preceding states, a class dictator-ship. It may therefore allow and sometimes actually encourage the dominationof the old ruling classes and also the economically dominant classes under so-cialism - to the detriment of the autonomy of the individuals who constitutethese groups.

163

The general will

that we relinquish atomic individualism and the exclusive supportfor private volition that it sustains. There is, however, an objectionto this strategy that must be addressed before we turn to commu-nism itself.

Individuals directed by a general will, who privilege the interestsof the "moral and collective bodies'7 of which they are indivisibleparts over their own "private" interests, could be faulted for ne-glecting "the cartes they owe themselves" and also the cares theyowe each other for the sake of "the whole community." For the in-terests of the whole community could sometimes require the dim-inution of the autonomy of some individuals, oneself or others, inviolation of the "equal respect" that is everyone's due. Unless thispossibility can be discounted, equal respect for the autonomy ofothers would seem ill served by general will coordination as byprivate willing.

But this conclusion is mistaken. To see why, it will be instruc-tive, once again, to contrast Rousseau's position with utilitarian-ism or, more precisely, with welfarist versions of utilitarianism.Were individuals' interests only welfare interests and were theseinterests measurable, as utilitarians believe, then the interests ofthe whole community would be the largest possible sum of indi-viduals' utilities. Schematically, if Us, social utility, represents thewelfare of the whole society and IT,, U2, . . . , Un represent thewelfares of individuals 1, 2, . . . , n, respectively,

Us = U, + U2 + . . . + Un.

Circumstances might then be such that in order to increase theutilities of some individuals and thereby to maximize the value ofUs, it is necessary to lower the welfares of others. Critics of utili-tarianism have long made much of this possibility. They havefaulted utilitarianism for failing to treat individuals' interests cor-rectly. According to their critics, utilitarians regard individuals'welfares - and therefore individuals too - only as means for max-imizing social utility, L/s, not as intrinsically valuable in them-selves. Those who would fault general will coordination for failingto respect the autonomy of each associate assume that general willcoordination is like utilitarianism in this respect; that its concernfor the interest of "the whole community" is at odds with respectfor persons as ends in themselves.

But the individuals Rousseau envisions aim to maximize auton-omy, not welfare. Does this difference cut against the claim that

164


Rousseauean political philosophy fails to accord individualsproper respect? I think it does - if autonomy is understood in theway that Rousseauean Marxists intend. There are two reasons whythis is so: The first has to do with the nature of the community au-tonomous agents form; the second with its historical trajectory.The first reason, developed by Kant, follows from Rousseau's ac-count of the social contract; since it is implicit in what has alreadybeen discussed, it can be briefly resumed. The second is a distinc-tively Marxist innovation that I shall only note here and develop inChapter 8. Both suggest a notion of community genuinely at oddswith liberal positions.

COMMUNITY

As remarked, Kant's notion of moral community was intended, inpart, to draw out the moral philosophical implications of "thewhole community" constituted by Rousseau's social contract. Tothis end, Kant deployed an account of practical reason richer thanthe instrumentalist conception liberals assume. But Rousseau'sview and relevant aspects of Kant's elaboration of it can be recon-structed without directly engaging issues about rational agency.The core idea, identified by Kant, is that autonomy is essentially so-cial because it can only be realized through interaction with othersin a "community" of equals, a moral community.

This claim is implicit in the idea that human beings, "born freebut everywhere in chains," only become the autonomous agentsthey potentially are in consequence of a social contract that estab-lishes full equality of citizenship. Generalizing this idea, Kantmaintained that individuals are equal qua moral agents preciselybecause they are (potentially) autonomous and that autonomy istherefore a necessary and sufficient condition for membership inthe community of moral agents. But, at the same time, individualscannot become autonomous outside the moral community. Forwhat it is to stand outside this community - in a "state of na-ture" - is just to act in a way that fails to respect moral personalityin oneself and others. It is to fail to respect the autonomy that ren-ders everyone, including oneself, a (potentially) moral agent.Therefore the autonomy of our fellow human beings is indispens-able for what matters supremely to each of us, and the enhance-ment of their autonomy, being necessary for the sustenance anddevelopment of the moral community that is our own first con-cern, must matter to us, just as our own autonomy does. For a

165

The general will

flourishing moral community is a condition for the enhancementof each individual's autonomy. It is what we all supremely value.36

The ability to set ends before oneself is a universal human ca-pacity. But the nature and limits of autonomy vary with levels ofmaterial and moral development. Very generally, with increasingmaterial development, the horizons for autonomous choice ex-pand. Marx's theory of history makes material development thefundamental structuring principle of human history. We have seenthat Mill also assumed a tendency for continuous materialprogress, even if he did not have an elaborated theory of it, andthat he maintained, more directly than Marx, that moral progressaccompanies the growth of productive forces. It was suggested inChapter 6 that this idea is congenial to Marx's theory of history aswell. It is therefore fair to say, for Marx too, that moral progressconsists in the enhancement of human autonomy, and that mate-rial progress is instrumental for its advance.

Does historical materialism provide special reason for thinkingthat moral progress involves the progressive socialization of (liberal)autonomy? This is not a question that Marx directly addresses.Nevertheless, historical materialism is relevant to it. Marx depictsa history of class struggles culminating in a proletarian class dic-tatorship and finally in classlessness under communism. Histori-cal materialism describes, in short, the demise of class divisions,the very obstacle that, in Marx's view, made Rousseau's de jurestate a Utopian fantasy. Only when economic forces no longerdraw people apart, according to Marx, can people finally treateach other as ends in themselves. At that point, real citizenship ina state truly based on Right (Recht) becomes possible. The moral

36 Readers who find the Kantian tone of these reflections off-putting may be per-suaded by the more prosaic, Humean idea that, in a direct confrontation be-tween moral agents and individuals pursuing their own self-interests, moralagents would not survive. In a word, they would be losers in a Prisoner's Di-lemma game in which they "cooperate" while their opponents "defect." It fol-lows from this consideration that those who value moral agency for themselvesought also to value it in others and ought therefore to do whatever is necessaryto help moral agency survive and flourish. If we add to this conclusion the ideathat morality supposes individual autonomy and then regard the collection ofmoral agents as a moral community, we have a good approximation of Kant'sposition. Cf. David Hume, An Enquiry Concerning the Principles of Morals, sec-tion 3, Part 1, p. 1987, in L. A. Selby-Bigge (ed.) Enquiries Concerning HumanUnderstanding and Concerning the Principles of Morals, 3rd ed. (Oxford: OxfordUniversity Press, 1975). The Prisoner's Dilemma is discussed in Chapter 2herein.

166


community Rousseau envisioned and Kant described is then trans-formed from an ideological snare and delusion to a genuine hu-man opportunity.

In short, the Marxist theory of history suggests that, at the endof the trajectory it describes, communities can exist in which "thecondition for the free development of each" is indeed "the free de-velopment of all." How defensible is this view? This is not theplace, once again, to subject historical materialism itself to criticalscrutiny.37 But it is apt to reflect on the plausibility of the (possible)end of the process it depicts. It is therefore time finally to reflectdirectly on communism and the conditions under which wholecommunities can exemplify "the exercise of the general will."

37 Cf. Wright, Levine, and Sober, Reconstructing Marxism, Chapters 2-5.

167

Chapter 8

Communism

Marx was famously loath to provide "recipes to the cookshops ofthe future/'1 and Marxists after him have generally followed hiscounsel. Of course all Marxists would agree that under commu-nism class divisions will be overcome and the historical materialistdynamic will have finally run its course. But beyond these verygeneral implications of Marx's theory of history, the Marxist cor-pus has little to say about the nature of communist societies. How-ever, caution in the face of an unpredictable future has not alwaysprevented Marxists from depicting communism in ways that todayseem blatantly Utopian. Revealingly, these descriptions are usuallycast negatively; thus we are told that communism is a society with-out commodity production, without alienation, and, of course,without a state. Marxists have suggested too that at the end of thehistorical materialist trajectory lies a society beyond scarcity andtherefore beyond justice. Marxists have also proposed that com-munist societies are populated by men and women of a radicallydifferent kind from those found in class societies, genuinely "so-cial" human beings. Some of these characterizations are exagger-ated, others flawed. But even were they entirely on target, theywould provide, at most, only vague intimations of what a commu-nist society might be like. In any case, despite their vagueness,there are few today, including dedicated Marxists, who would notfind them misleading. If nothing else, in their exuberant opti-mism, they are too much at odds with that skepticism abouthuman perfectibility that has come to pervade our intellectual cul-ture. This skepticism is exemplified perspicuously by mainstreamliberalism, the hegemonic political theory today.

i Karl Marx, Capital, vol. i, trans. Moore and Aveling (Moscow: Foreign LanguagePublishing House, 1959), Afterword, p. 17.

168

Communism

Marxists have generally paid scant attention to the affinitieslinking communism with the state Rousseau envisioned in The So-cial Contract. With few exceptions, Rousseau scholars have beensimilarly reluctant to join Rousseau with Marx.2 However, in TheEnd of the State, I argued that Marxian communism accords pride ofplace to the model of rational cooperation Rousseau imputed tothe de jure state and that distinctively Marxist views about historyand politics provide good, if inconclusive, reasons for thinkingthis ideal practicable. I maintained, in addition, that, despite Rous-seau's deeply paradoxical and equivocal formulations, the de jurestate is not strictly a state at all but an internally coordinated "re-public of ends." So too, therefore, is Marxian communism. If theseclaims are on track, it follows that, in its political dimensions, com-munism is precisely what traditional formulations have claimed: asociety without a state. It is so in just the way that the "state" ofThe Social Contract is.

It is time to elaborate on this contention and to investigate someof its implications. I shall begin by focusing on two claims some-times made about communism - that it is a society beyond scar-city, and that it is populated by distinctively social individuals. Aclearer understanding of these issues will help to clarify furtherthe Rousseauean-Marxist idea of autonomy and the notion of per-sonal responsibility it suggests. From that perspective, it will befeasible to broach questions about the nature and role of justice un-der communism. By pursuing this line of inquiry, a purchase oncommunism emerges that renders the idea normatively appealingand that shields it from charges of incoherence or utopianism.

BEYOND SCARCITY?

Distributive justice is morally compelling whenever fairness is ofconcern to those to whom benefits or burdens are distributed. Asociety beyond justice is therefore one in which fairness is of little orno concern. One way for fairness not to matter, for individualswho are disposed to care only (or mainly) about their own distrib-utive shares, is in conditions of abundance. When air is abundant,as it normally is, it does not matter how it is distributed. In this

2 Conspicuous among the exceptions are Galvanno Delia Volpe and Lucio Col-letti. See for example, Delia Volpe's Rousseau and Marx, trans. John Fraser (Lon-don: Lawrence and Wishart, 1978); and Colletti's From Rousseau to Lenin: Studiesin Ideology and Society, trans. John Merrington and Judith White (London: New-Left Books, 1972).

169

The general will

sense, even our society is "beyond justice" with respect to the dis-tribution of air. When it is claimed that communist societies are"beyond justice," what is evidently meant is that, under commu-nism, most of the things people want will be relevantly like airnow is. Then scarcity would not be a fact of the human condition,and the fairness of distributions would not be a concern. In aworld of abundance, justice would be of no normative significance.

The idea that communist societies will have overcome scarcity issuggested by Marx's theory of history. In the historical materialistscheme, modes of production rise and fall in order to accommo-date to ever-increasing levels of development of productive forces.This movement is propelled, ultimately, by transhistorical aspectsof human nature. As remarked, in Marx's view, human beingsseek to realize themselves, to become what they (potentially) are,through productive activity. But scarcity frustrates this impulse.History is then propelled by an interest in removing this frustra-tion, in making the world human beings confront more hospitableto the self-realization they seek. In a world of scarcity, an interestin self-realization is tantamount to an interest in developing pro-ductive forces. As scarcity gives way to abundance, this universalhuman interest is increasingly transformed into an interest in re-moving specifically human obstacles in the way of self-realization.Thus late capitalism is unstable not because it thwarts the expan-sion of productive capacities but because its way of developingproductive forces is detrimental to this fundamental humanconcern.3 The end of the historical materialist trajectory is reachedprecisely when scarcity no longer propels this movement andwhen an economic structure, communism in its "higher phases,"is in place - permitting, to the greatest degree possible, the real-ization of that universal human interest that scarcity had previ-ously thwarted.

3 See G. A. Cohen, Karl Marx's Theory of History: A Defense (Oxford: Oxford Uni-versity Press, 1978), pp. 302-307. Cohen's account of historical materialism dif-fers from more traditional formulations. In the standard view, capitalism isdestined to "break down" in consequence of an endogenous tendency for therate of profit to decline in capitalist economies. As this tendency works itscourse, investment possibilities dwindle and the economy, if capitalism has notalready been overthrown, grinds to a halt, On this understanding, capitalism isdoomed in just the way precapitalist economic structures are - because it isbound eventually to "fetter" the development of productive forces. In contrast,Cohen's account of "the distinctive contradiction of late capitalism" does notimply an inevitable capitalist breakdown, nor does it endorse the orthodoxMarxist political economic analysis from which this prediction follows.

170

Communism

The core intuition underlying this rationale has been elaboratedby philosophers in the liberal tradition for more than two hundredyears. In consequence, the so-called circumstances of justice, theconditions under which the requirements of justice are morallycompelling, are now widely understood.4 They involve both a"subjective" and an "objective" side. The subjective component,the idea that individuals are generally self-interested, will be dis-cussed in the next section. Objectively, justice matters whenevercooperative interactions are both possible and necessary for indi-viduals concerned to maximize their own distributive shares. Inconditions of absolute scarcity, where cooperation will not expandthe supply of goods to be distributed and interactions thereforehave a zero-sum character, awarding a benefit to one individualnecessarily deprives another of it, so that there is no advantage toanyone in interacting cooperatively. Thus two individuals on a life-boat with only enough water to assure the survival of one havenothing to gain from cooperation. These individuals have no stakein a fair distribution of the water; justice is not a (private) interestof theirs.5 Were (nearly) everything absolutely scarce, like wateron this lifeboat, the requirements of justice would pale in (norma-tive) importance. A society of this kind would be, as it were, be-neath justice, in contrast to a world of abundance which would bebeyond it. In short, justice matters when most of the things peoplewant are relatively scarce. But relative scarcity is not the only factabout the world that renders justice morally urgent. In addition,individuals must be sufficiently equal in natural endowments forcooperation to be mutually advantageous, and they must findthemselves in circumstances where, in consequence of geograph-ical proximity, they are obliged to interact.6

4 On "the circumstances of justice," see, among others, Hume, A Treatise of Hu-man Nature, Book III, Part 2, Chapter 2, and An Enquiry Concerning the Principlesof Morals, section 3, Part 1; H. L. A. Hart, The Concept of Law (Oxford: ClarendonPress, 1961), pp. 189-195; J. R. Lucas, The Principles of Politics (Oxford: ClarendonPress, 1966), pp. 1-10; and Rawls, A Theory of Justice.

5 Nevertheless, individuals whose sensibilities have been formed in the circum-stances of justice, finding themselves on a lifeboat with only enough water forone, might feel that a good way to decide who gets the (absolutely) scarce watermight be through some fair procedure - say, by drawing lots.

6 There is an evident connection between the circumstances of justice and tradi-tional contractarian accounts of the state of nature. Thus relative equality in thedistribution of natural endowments - in conjunction with geographical proxim-ity - helps to explain why, in Hobbes's view, individuals in a state of nature arevulnerable to mortal assault from one another, a fact which helps explain, inturn, why the state of nature is, for Hobbes, a "war of all against all."

171

The general will

Of these circumstances, relative scarcity is surely the most alter-able. Despite Rousseau's account of the early state of nature, itis implausible to think that human beings could ever live withoutcoming into contact with one other. It is even less likely that indi-viduals will ever become so unequal in natural endowments thatpersons with private wills would cease to find it in their intereststo cooperate. Certainly Marxist theory offers no reason to imaginethese eventualities, nor do any other theoretical traditions uponwhich defenders of communism might draw. But Marx's theory ofhistory does seem to provide a reason for predicting the eventualelimination of scarcity as a fact of the human condition.

Strictly speaking, what historical materialism envisions is an eco-nomic surplus so massive that class struggles over it and ultimatelyclass divisions themselves will expire for want of a sufficient rea-son. But, for its more orthodox adherents, historical materialismhas, by force of suggestion, if not strict implication, underwrittenan almost unlimited faith in the beneficent consequences of devel-oping productive capacities. Perhaps at no time was this faith moreevident than in the period when Stalinist orthodoxy held swayamong Communist militants and their sympathizers among theintelligentsia of Western and Third World countries. It was thisideology that underwrote the Stalinists' subordination of all as-pects of the revolutionary project, including the transformation ofproduction relations, to the development of productive forces. Itwas this ideology too that helps explain the long-standing disre-gard of the norms of political morality that afflicted the Sovietpolity for decades. The unhappy consequences of this political ori-entation are a matter of historical record: the brutal collectiviza-tion of agriculture; the imposition of a hierarchical structure and"productivist" ideology in the manufacturing sector; a selective,technocratic, and authoritarian educational system; a severe cen-tralization of political power; and most damaging of all, an indef-inite prolongation of police terror and the inexorable growth ofbureaucratic domination. Today, of course, Stalinism is in total dis-repute. But the idea that there can be a society beyond justice ispartly a residue of this abandoned faith or, more precisely, of itsone-sided exaggeration of historical materialist positions.

In any case, if the understanding of communism as a society be-yond justice must be revised or abandoned, it should be on itsmerits, not to avoid guilt by association with Stalinist enthusi-asms. There is surely much to fault in the received view. Never-

172

Communism

theless, the idea is not nearly so wrong-headed as may appear.So long as some goods, like being ranked first in an ordering, are"positional" such that not everyone who wants them can havethem; so long as others will in fact remain scarce no matter howoverwhelmingly developed productive capacities become; and solong as individuals' lifetimes remain finite so that time must al-ways be budgeted and some desires are bound to go unfulfilled,scarcity will remain a fact of human life. But relative scarcity cancease to structure and direct human life in the way it heretoforehas. The conviction that the requirements of justice diminish inmoral importance as productive forces expand is eminently defen-sible, even if scarcity cannot be "overcome" entirely and alto-gether. Thus talk of a society beyond scarcity and, for that reason,beyond justice is misleading, but the intuition that motivates it isbasically sound.

BEYOND SELF-INTEREST?

Another way that Marxists have envisioned moving beyond justiceis by maintaining that, in the higher phases of communism, indi-viduals will have "transcended" self-interested motivation. An ex-ample of this genre of theorizing is provided by the early Sovietlegal scholar Pashukanis.7 In contrast to a bourgeois order com-posed of individuals with wills in conflict, he envisioned "the so-cial person of the future, who submerges his ego in the collectiveand finds the greatest satisfaction and meaning of life in this act."Employing the expression of Feuerbach and other Young Hege-lians including Marx, Pashukanis foresaw "the development ofhigher forms of humanity as the transformation of man into aspecies-being."8 If "species-being" is identified with membershipin a "moral and collective body," Pashukanis might seem to beechoing Rousseau. But there is a more likely way to understandwhat Pashukanis and others like him had in mind. On this con-strual, under communism, individuals are altruists rather thanegoists. Self-interest is not so much "transcended" as replacedby its opposite. Since this understanding challenges neither thereceived view of rational agency nor the reigning metaphor of

7 See E. B. Pashukanis, Law and Marxism: A General Theory (1924), trans. BarbaraEinhorn, ed. Chris Arthur (London: Ink Links, 1978). See also Lukes, Marxismand Morality, pp. 35—36.

8 Pashukanis, Law and Marxism, pp. 159-160.

173

The general will

the atomic individual, it translates easily into the familiar liberalframework. Communism is a society "beyond" justice because oneof the (subjective) circumstances of justice, self-interest, fails topertain.

Liberals may find this position implausible, but they can hardlyfault its intelligibility. We have seen how the Hobbesian-liberal ac-count of rational agency is distinct from Hobbesian and liberal pre-dictions about individuals' interests. Therefore those who identifyrational agency with utility maximization have no groundsfor insisting that individuals maximize utility by seeking to ad-vance their own well-being.9 Altruism is as rational as egoism is.Altruism is as individualist as egoism is too. Egoists are concernedwith their own well-being conceived atomistically. Altruists areconcerned with the well-being of one or more other atomic indi-viduals. In each case, it is only the welfare interests of atomic in-dividuals that matter.

In Chapter 2, I suggested that Hobbes's account of the state ofnature provides an uncompromising picture of atomic individu-als with interests of this sort. Moved by a fear of one another ("dif-fidence"), an unceasing desire to accumulate ("competition"),and a yearning to dominate others and stand high in their es-teem ("vainglory"), each individual's utility function is utterlyself-regarding. We also saw how, starting from very different as-sumptions, Rousseau arrived at a similar understanding of themotivations of his contemporaries and, we may extrapolate, ourown. In his view, after the introduction of private property, thementality Hobbes thought universal emerges. If he was right, ra-tional egoism need not be the destiny of rational human beings;

9 However, on this view, once utility functions are formed, rational individualsmust then be assumed to be disinterested in each others' interests. Individualscan rationally seek to advance (or diminish) the welfare of other individuals. Butthese desires, if they exist, are represented in their utility functions. In otherwords, the Hobbesian-liberal theory of rational agency takes utility functions asgiven, but once they are in place the theory stipulates that individuals care onlyabout the utility payoffs to themselves. Egoism and mutual unconcern are there-fore distinct ideas. When individuals are rational egoists, it is not only in virtueof their being (mutually unconcerned) utility maximizers but also because of thenature of their utility functions. In a word, their interests themselves are ego-istic. On mutual unconcern, see Rawls, A Theory of Justice, p. 13. David Gauthier,following P. H. Wicksteed, has designated mutual unconcern "non-tuism" andasserted that it is a "dogma" of the prevailing conception of practical reason. SeeDavid Gauthier, Morals by Agreement (Oxford: Oxford University Press, 1986),p. 87, and "Reason and Maximization," pp. 413-415.

174

Communism

whatever Hobbes may have believed, it is not inscribed in humannature as such.

Hobbesian egoism, precisely because it is so stark and uncom-promising, provides an unappealing and unrealistic basis for think-ing about political association. But in consequence of its severity, itputs the subjective component of the circumstances of justice intoclear focus. These circumstances pertain, however, even if moremoderate accounts of self-interested motivation are assumed. It isenough, as writers after Hobbes have made clear, that, in assess-ing alternative courses of action, individuals generally adopt aself-regarding attitude. Their wills need not evince an exclusive orall-consuming desire to enhance their own well-being but only asignificant desire to do so. Altruists evince a corresponding desireto enhance the welfare positions of others.

I argued in Chapter 3 that conditional altruism is at some re-move from general will coordination. This conclusion holds withequal force for altruism generally. The social individual of Pashu-kanis's imagination differs from the egoist only in substituting al-truism for egoism. But individuals moved by a general will areneither egoists nor altruists. They are indivisible parts of the moraland collective bodies whose interests they seek to advance. Inshort, they are creatures of a universe that the atomic-individualistmetaphor denies. Therefore Pashukanis's social individual is moreat home than the Rousseauean citizen in an intellectual culturegoverned by the image of the atomic individual. As remarked, hisidea is perfectly intelligible from the standpoint of mainstream lib-eralism. But however compatible it may be with liberal theoreticalpositions, Pashukanis's vision is so demanding that, even were itfeasible, it would be unappealing if not chilling to anyone movedby liberal sensibilities. Paradoxically, the Rousseauean-Marxist al-ternative fares a good deal better in this regard.

LIBERALISM AND THE GENERAL WILL

The parties to Rousseau's social contract do not deliberate as "cit-izens" all the time. In non- or extrapolitical contexts they are ego-ists and perhaps sometimes altruists too. In order to understandhow general will coordination can coexist with forms of coordina-tion underwritten by the metaphor of the atomic individual, itwill be well to consider more carefully than we so far have whythe citizens of a just state are not and ought not to be directedexclusively by the general will. Inasmuch as communism is a

The general will

feasible approximation of this Rousseauean ideal, similar consid-erations will carry over to men and women living in communistsocieties.

According to Rousseau, when individuals place themselves un-der "the supreme direction of the general will" - in particular,when they deliberate as "indivisible" members of the "moral andcollective bodies" they (freely) constitute - they are equal qua cit-izens. It is for this reason that each individual has one and onlyone vote, and that declarations of the general will are made bycounting votes. Moreover, despite what citizens of liberal democ-racies have come to accept, equality of citizenship is not a strictlyformal requirement. For majority ruh voting to serve as a reliablemeans for discovering what the general will is, a condition abso-lutely necessary if the majority's decision is to obligate thosewhose positions were defeated in the popular assemblies, citizensmust actually be equal or very nearly so with respect to everythingthat bears on the "rectitude" of their deliberations and choices.Voters in a de jure state do not express preferences for alternativeoutcomes in contention but express opinions on what the generalwill is; they register judgments about a matter of fact that they arepresumed to be (equally) capable of ascertaining. As with jury vot-ing or the statistical pooling of (disinterested) expert opinions onpublic policy questions, voting is a mechanism for aggregatingjudgments that, even if they are not in fact equally sound, are rea-sonably treated as if they were. General will coordination or atleast the form of it that Rousseau expressly described cannot ob-tain in conditions that are detrimental to the development of ra-tional and independent judgments by all the members of thewhole community. Whatever detracts from real equality of citizen-ship is therefore inimical to general will coordination.

Ironically, general will coordination itself can have precisely thiseffect if it is allowed to supersede private will interactions com-pletely. Recall, for example, the John and Mary of Chapter i, de-liberating now about the division of labor within their household.Suppose that in this matter too they are determined to advance thegeneral interest of the couple John-Mary, rather than the privateinterests of the individuals John and Mary who constitute this"moral and collective body." Finally, imagine that they can some-how screen out particular interests that masquerade as general in-terests, so that John and Mary's general will deliberations have allthe rectitude Rousseau ascribes to the popular assemblies of the de

Communism

jure state. It could turn out, perhaps for reasons having to do withtheir respective life histories to that point, that it is the general in-terest of this couple that one of their members, say Mary, shoulddo all and only housework, while the other, John, should partici-pate actively in civil society and the state. By hypothesis, this di-vision of labor, because it is in the general interest, in no waydepends upon actual domination.10 Nevertheless, in consequenceof this choice, one member of the couple is consigned to menialand constricting tasks that, let us suppose, progressively under-mine her abilities to engage actively and autonomously in collec-tive decision making. At the same time, the other is able tocultivate these capacities at an accelerated rate thanks to the lifeopportunities he confronts. This result will, in turn, affect the out-comes of subsequent general will deliberations between John andMary; the reasons that warranted the initial division of labor willpertain with increasingly greater force. At some point, if this pro-cess continues indefinitely, it may become impossible to continueto regard John and Mary as equal citizens. The factors that bear ontheir abilities to engage equally and independently in collectivedeliberations and collective decision making will have become toounequally distributed between them. This sad turn of events willnot have come about in consequence of a contest of competing pri-vate wills; it will be a result of general will coordination itself. Theexercise of the general will will have undermined the conditionsfor its own possibility.

In this sense, general will coordination can be self-defeatingover time; it can also be undesirable, as even Rousseau recognized.Once the state of nature has become a war of all against all, thesocial contract is in each contracting party's interest. At that point,since general will coordination is indispensable for saving indi-vidual autonomy, it is best for each individual to "place himselfunder the supreme direction of the general will." But Rousseau'sargument pertains only to the state itself. In civil society, espe-cially in the commercial sphere, private interests continue to holdsway. Thus, in the economy that Rousseau imagined, individualswould be (largely) self-sufficient yeoman farmers or small-scale

10 Arguably, it does depend on residues of past domination. Those who wouldmaintain that true general will deliberations should not countenance pastdomination even in this attenuated sense may suppose that the same divisionof labor is in the general interest solely in consequence of differences in (in-nate) aptitudes.

177

The general will

artisinal producers who exchange their surplus products throughmarket transactions.11 Outside the political arena, people remainparadigmatic private willers, just as they were in the state of na-ture. Were they not so moved, they would, in Rousseau's expres-sion, "neglect the cares they owe themselves,"12 precisely the faultthat the social contract remedies in political life. Similar consider-ations apply to John and Mary in deliberations over their divisionof household labor. Even if general will deliberations in this casewould not undermine the equality that is the condition for its pos-sibility, there would have to be a compelling reason why Mary orJohn should make the interest of the John-Mary couple their ownin this case. On any likely elaboration of their story, it is virtuallycertain that they would be better off pursuing their own privateinterests. For should they fail to make their private interests theirown, they would put their own well-beings in jeopardy withoutobtaining adequate compensation in return. In contrast, the indi-viduals who alienate all their rights to the whole community getback nothing less than essential autonomy, thereby saving theirvery humanity from mortal peril.

Nevertheless, as with the idea of a world beyond scarcity, thereis something right-headed in the picture of communism as a soci-ety in which human beings have ceased to be generally self-interested. Communism surely does require something like the"transcendence" of the subjective component of the circumstancesof justice - insofar as it involves general will coordination, ofcourse, but also in those aspects of social life where private voli-tion continues to be appropriate. Marxists have generally deniedthe possibility of a "third way" between rational egoism and what-ever it is supposed to contrast with - whether it be "species-being" in Pashukanis's sense or the very different, Rousseaueanalternative that I maintain Marxian communism presupposes. Indepicting the private and the general will as mutually exclusiveand eternally opposed, Rousseau too effectively denied the exis-tence of a third way. But, in this case as in so many others, Rous-seau's reflections on politics provide resources for rectifying themisapprehensions his political theory suggests. It will thereforebe useful, again, to turn back to Rousseau, and to reflect on thecharacter of the individuals who constitute the society the de jurestate superintends.

11 See Levine, The Politics of Autonomy, Chapter 5, pp. 187-198.12 The Social Contract, Book I, Chapter 6.

178

Communism

VIRTUE

Rousseau's citizens, in their commercial interactions and moregenerally in the sphere of civil society, are moved by private wills,but they are not, after the social contract, the untrammeled egoistswhose Hobbesian dispositions make a social contract necessary.Like the public-minded citizens envisioned by other eighteenth-century political theorists, they are imbued with republican vir-tue - with a disposition toward moderation, simplicity of mannersand morals (moeurs), and a powerful longing for equality itself.13 Tobe sure, they are still generally self-regarding. But their egoism ismoderate and gentle. They are more nearly creatures of Hume'sTreatise than Hobbes's Leviathan or the concluding pages of The Sec-ond Discourse.

Virtue, for theorists in the republican tradition, is a dispositionto accord precedence to communal over individual concerns. ForRousseau, it is this disposition that allows the state, directed by ageneral will, to superintend a society composed of small indepen-dent producers who own their own means of production - andtherefore the products of their labor - and who, insofar as they arenot self-sufficient, trade on markets. Private property and marketsgenerate resource inequalities, and these inequalities, if they growtoo large, put liberty and therefore general will coordination itselfin jeopardy.14 Individuals must therefore be made virtuous in or-der to counter the mentality markets encourage; otherwise, thereis no hope of mitigating their effects. It is far from clear, however,that a virtuous disposition alone can suffice to stem the tide. ButRousseau, along with other republican theorists, imagines no al-ternatives to markets and private property. He is therefore unableto envision any more efficacious means for keeping inequality -and private volition - in bounds. Thus he proposes a host of "su-per structural" measures - from schooling to public spectacles tothe institution of a civil religion - for inculcating virtue in the faceof an economic "base" that militates against it.15

13 Cf. Levine, The Politics of Autonomy, pp. 168-172.14 Cf. The Social Contract, Book II, Chapter 11; and Levine, The Politics of Autonomy,

pp. 188-192.15 For discussion of some of the practical political measures Rousseau proposes,

see Levine, The Politics of Autonomy, Chapter 4, and The End of the State, Chapter2. Rousseau's reflections on political institutions are always sensitive to con-junctural factors and peculiarities of geographical and historical circumstance.Thus the institutional arrangements he endorses are not, strictly, technicalmanipulations deemed necessary for any de jure state, but measures that are

179

The general will

It is one thing to have a virtuous disposition and something elseto subordinate one's private will to the general will. Rousseau es-teemed virtue in its own right; he had only praise for the simplicityand equality of the yeoman society of his imagination. And his cel-ebrated contempt for luxury and the vanities of life in the cities andgreat estates is evident throughout his writings. But in The SocialContract, Rousseau defended virtue mainly for its efficacy in pro-moting general will coordination. Virtuous people, he maintained,are more likely than unabashed egoists to subordinate their pri-vate wills to the general will. From this perspective, it is immate-rial whether or not virtue is an end in itself. What matters is thatit is instrumental for general will coordination. Virtue's domain iscivil society; the proper arena for general will coordination is thestate. In Rousseau's view, it is (partly) because men and women incivil society are virtuous that they are able to become citizens, in-divisible parts of the sovereign, when they come together as leg-islators in the assemblies of the people.

It has been widely observed that revolutionaries from Cromwellthrough Lenin and beyond and, to a lesser degree, radicals gen-erally share a common political style, and that this style is rootedin the republican tradition.16 A modern revolutionary or militantradical is likely to be virtuous in roughly the republican sense, ifonly because anyone whose life is directed toward realizing fun-damental political and social transformations, typically at consid-erable personal cost, almost certainly places public values abovepersonal concerns. But there is an affinity too at the level of indi-vidual sensibilities. People who dedicate their lives to revolution-ary objectives are characteristically inclined, as seventeenth- andeighteenth-century republicans were, to lead lives of personal aus-terity and strict discipline. Indeed, many revolutionary groups,since the seventeenth century, have gone so far as to model them-selves on military organizations. To some degree, the latter phe-nomenon can be explained as a functional adaptation to the tasksrevolutionaries confront. But it is probably also explainable by ap-

generally instrumental for rendering de jure institutions feasible in conditionslikely to be confronted by the (contemporary) readership of The Social Contract.In this respect, they are like the measures advanced by Marx and Engels in TheCommunist Manifesto - programmatic policies appropriate for Germany in theprerevolutionary situation leading up to 1848 but not necessarily for commu-nists in other times and places.

16 See, for example, Michael Walzer, The Revolution of the Saints: A Study in the Or-igins of Radical Politics (Cambridge, Mass.: Harvard University Press, 1965).

180

Communism

peal to the mentality of those disposed to become "professionalrevolutionaries/' a cast of mind relevantly similar to the tempera-ment republican theorists esteem.

Typically, revolutionaries do not just strive to be virtuous them-selves; they seek to establish virtuous communities. In this re-spect, the more radical Jacobins like Robespierre and Saint-Just,like the Puritans before them, were exemplary.17 But even the Bol-sheviks and their successors, despite their official antipathy to-ward moralistic assessments, evince a similar sensibility. Theclassical Communist aversion to bourgeois excess and to that pre-occupation with the self that bourgeois society fosters reflects thisattitude.18 These class hatreds stemmed, in part, from the samesource as did Rousseau's abhorrence of luxury, his animosity to-ward city life, and his glorification of simple peasant ways.

Bolshevik attitudes reflect a similar sensibility in Marx himself.Marx was an ardent, though critical, admirer of the Jacobins, andhis enthusiasm has been recognized and appropriated by his fol-lowers. To be sure, Rousseau's express declarations in behalf of vir-tue are foreign to Marx's immediate theoretical concerns, andRousseau's use of virtue as a "corrective" within essentially cap-italist societies must appear anachronistic to any defender ofhistorical materialist positions. Rousseau could not conceive of al-ternatives to private property and markets; Marx could and did.Thus his liking for the mentality republicans sought to foster wasnot nearly so driven by theoretical exigencies as Rousseau's. It wasmore a matter of sentiment than intellectual conviction.

Nevertheless, there is something defenders of Marxian commu-nism can learn from the use Rousseau made of virtue. By restrict-ing general will coordination to the political sphere and thenseeking to mitigate but not entirely eliminate private willing else-where, Rousseau's example suggests a way to save the idea ofcommunism from the excesses of its Marxist defenders. Their com-munism seems Utopian and unappealing because they depict it asa society beyond justice or, more precisely, beyond the circum-stances of justice; because they insist that neither scarcity nor self-interest pertains under communist conditions. Taken literally,

17 I discuss the place of virtue in radical Jacobin politics in The End of the State,Chapter 3.

18 The fact that Communist leaders, the nomenklatura of the post-Stalinist period,came increasingly to adopt bourgeois life-styles themselves no doubt contrib-uted, in part, to the "legitimation crisis" of Communist regimes and ulti-mately, therefore, to their demise.

181

The general will

these claims cannot be sustained. It is appropriate, therefore, toturn to the role of justice - and also virtue, especially the virtue ofjustice - under communism. With this focus, it will become pos-sible to recognize what genuinely is insightful in traditional ac-counts of communism. We will find that the received doctrine canbe reconstructed in a way that is faithful to its fundamental con-tentions but that does not burden communism with unrealistic andundesirable properties.

FROM VIRTUE TO JUSTICE

We have seen that the central objective component of the circum-stances of justice, relative scarcity, can never be eliminated en-tirely - because some (generally desired) goods are likely toremain scarce, as a matter of fact, no matter how massively devel-oped productive forces become, and because others, like honors orprivileges, are necessarily scarce in the sense that they would bedevalued the more generally available they become. In addition,the fact that people, even under communism, would be bound tobudget their time, if only to decide which (virtually) free goodsthey will deploy at particular moments, implies that a preemi-nently important good, time itself, will always remain in shortsupply in any imaginable human future.

On the other hand, there is no reason in principle why the sub-jective component of the circumstances of justice, self-regard, can-not be entirely overcome. Human beings could become full-timealtruists or general willers or even grow indifferent altogetherto distributional concerns. However, such far-reaching changesin human beings' characters are unlikely. More importantly, thewholesale abandonment of private volition would be unappealing,even if it could somehow be achieved.

Thus a society entirely beyond the circumstances of justice is im-possible and undesirable in any case. But if the historical materi-alist picture is sound, scarcity will diminish markedly beforecommunism is established and then under its aegis, and humanbeings will become less egoistic - in roughly the way that repub-lican theorists supposed. Thus it might seem that the moral ur-gency of justice will wither away, even if it will never disappearcompletely. However, this conclusion is importantly misleading.Paradoxically, the ever-receding salience of the circumstances ofjustice, far from making justice unimportant, will actually makejustice paramount in communist societies.

182

Communism

With diminishing scarcity and egoism abating, struggles overdistributive shares are likely to diminish in frequency and urgency.But if justice as a property of institutions withers away (withoutever vanishing entirely) as a normative concern, justice as a prop-erty of persons is sure to become more important. As a dispositioninternalized in the men and women of communist societies, jus-tice is indispensable for communism in just the way that Rousseaudeclared virtue indispensable for the de jure state. Allowing fordifferences in precision and discounting the eighteenth-centurycast of Rousseau's republicanism, these contentions represent thesame idea. In brief, communism presupposes the virtue of justice.19

In contrast to the political culture of bourgeois societies, politicsin communism's higher phases will no longer be a contest of pri-vate wills, a struggle over who gets what. With diminishing scar-city and increasingly transformed individuals, people will beincreasingly disposed to look to the general interest in making col-lective choices. They will be general willers in those aspects oftheir lives - the political sphere, above all - in which communalconcerns are their true interests; otherwise, they will be just menand women. Similarly, Rousseau believed that in a state based onthe social contract individuals would be citizens in the assembliesof the people and virtuous independent producers otherwise. Thereason is plain: In order to achieve the uncoerced equality of com-munism, people must be disposed to do what just social institu-tions would compel them to do were private interests still drivingthem apart.

Formally, justice requires that like cases be treated alike. Thisprinciple is transhistorical and universal but without content be-cause it says nothing about what constitutes like cases or equaltreatments. In different times and places, different views of thesematters have gained wide acceptance. The idea that human beingsgenerally constitute a category that counts as a like case is a dis-tinctly modern conception; more inegalitarian understandingswere commonplace in earlier times. Even today, hardly anyonewould place all individuals in the same category with respect toevery distributive concern. Thus it is widely believed that somegoods should be distributed for reasons of merit or need or pro-ductive contribution - in other words, that people can be sorted outinto morally relevant categories with respect to how meritorious or

19 Cf. G. A. Cohen, "Self-Ownership, Equality, and Communism," Supplemen-tary Proceedings of the Aristotelian Society (1990).

183

The general will

needy they are or how much they have contributed. The idea thatall human beings merit equal treatment is only a presumption.As such, it is almost universally entertained. Inequalities may ac-cord with justice, according to contemporary understandings,but when they do they must be justified; there must be defen-sible grounds for treating persons differently. If no acceptable casefor unequal treatment can be made, everyone should be treatedthe same.20

The principle "to each the same" is therefore fundamental tomodern conceptions of justice. But "to each the same" is not byitself equivalent to the modern conception. Thus, in monastic or-ders or in the radical sects of antiquity, egalitarian distributionsserved to foster a sense of in-group identification and differencefrom the outside world, not an awareness of unity with human be-ings generally. Historically, the egalitarian principle has probablyfunctioned more often to foster differences among people than tounite them. It is only when joined with a belief in the moral equal-ity of persons, a conviction liberals and Marxists share, that theprevailing sense of justice exhibits the egalitarian cast of contem-porary liberalism and Marxism.

Understood as a character trait, justice denotes a disposition onthe part of each individual to treat everyone in the same way - ex-cept insofar as a case can be made, compatible with a commitmentto the moral equality of persons and to their equal treatment, fordifferentiating between them. A just disposition mitigates privatevolition without eliminating it, and it renders persons more capa-ble of general willing in those instances in which it genuinely is inindividuals' interests to place themselves under the supreme di-rection of a general will. But the virtue of justice does not functionin quite the same way that virtue does in Rousseau's political phi-losophy. Rousseau's invocation of virtue was an act of desperation.He needed a "corrective" in the face of an economic system thatthreatened to generate social divisions inimical to the exercise ofthe general will. For want of an alternative to markets and privateproperty, virtue was the straw he grasped. However, under com-munism, even in its earliest (socialist) phases, the economic or-der - because it works against the reproduction of class divisions -

Since human beings differ in their preferences for resources and in their situ-ations, equal treatment does not entail providing everyone with exactly thesame things. See Ronald Dworkin, "Equality of Resources/' Philosophy and Pub-lic Affairs, vol. 10, no. 4 (1981), pp. 283-345.

184

Communism

promotes equality of citizenship. Therefore the virtue of justice isa complement, not a corrective, to the economic system of socialistand later communist societies.

EQUALITY AND "BOURGEOIS RIGHT"

Modern views of justice exhibit a tendency toward equality but al-low for inequalities of various kinds. Thus there are accounts ofjustice, liberal in intention, that countenance inequalities for thesake of property rights; others that acknowledge claims on re-sources based on merit or effort or whatever else might render in-dividuals deserving; and others that permit inequalities for reasonsof efficiency (broadly conceived). Rawls's theory of justice is a the-ory of this sort. According to Rawls, unequal distributions are justif they satisfy the Difference Principle - that is, if they benefit rep-resentative members of the least well-off group. Inequalities canhave this effect only if they help to improve productivity enough tomake the worst position better than it would have been withoutthose inequalities. Rawls's theory is among the most egalitarian ofextant liberal conceptions and is certainly the most theoreticallydeveloped. It also bears on distributive principles that give expres-sion to what Marx called "bourgeois right," the idea that distribu-tions should correlate directly with (productive) contributions.This idea is pervasive in the political cultures of capitalist societiesand appropriate, in Marx's view, for socialist societies too. It isworth reflecting on these interconnected issues in order to probethe implications for equality - and for the general will - of societ-ies populated by individuals imbued with a sense of justice.

In The Critique of the Gotha Program, Marx maintained that undersocialism, communism's lowest phase, what individuals get shouldbe in proportion to their productive inputs, denominated in so-cially necessary labor time, after various "social deductions" havebeen exacted to support nonproductive sectors of the populationand to provide revenue for the state.21 This concession to "bour-geois right," like Rawls's Difference Principle, was plainly moti-vated by efficiency concerns. Individuals emerging from capitalistsocieties would not be inclined to develop their talents or to deploythem efficiently without material incentives for doing so. In con-sequence (nearly) everyone, including the least well-off, would

21 Cf. Karl Marx and Friedrich Engels, Selected Works, vol. 2 (Moscow: ForeignLanguages Publishing House, 1962), pp. 13-48. Also, see Chapter 7, herein.

185

The general will

likely suffer. Hence the need to provide differential remunerations -based no longer on ownership of external things, as under capi-talism, but on the quality of workers' skills and their expendituresof effort. It is only as socialism evolves into communism - in par-ticular, after human beings have been weaned away from una-bashed self-interest - that differential rewards for direct producersand other economic agents can be relegated to a superseded pastand human communities can finally inscribe upon their banners"From each according to his ability, to each according to his need."

This well-known slogan of Marx's is easily misunderstood. Bythe time that human societies will have moved into the higherphases of communism, a truly massive economic surplus will ex-ist. We can therefore assume that basic needs - insofar as they canbe distinguished from wants - will have been met long before.Thus the "need" Marx speaks of comes to the same thing as"wants." The idea would then be that, under communism, indi-viduals will take from the common stock as they please, regard-less of their contributions to it, and contribute to the commonstock according to their different abilities. Marx does not quitesay so, but we can infer as well that, with productive capacitiesmassively developed, contributions too will be provided more orless as producers please. In any case, under communism, it wouldno longer serve a purpose for distributive shares to be propor-tional to the duration or intensity of the labor power recipients ex-pend. The Difference Principle would therefore be otiose too -inequalities never would enlarge the distributive share of the leastwell-off. In short, under communism, the Difference Principlewould mandate equal distributions; the presumption for strictequality would generally prevail. Then if, following Rawls, rights-based or desert-based arguments for inequality are taken to haveno normative force - if only the Difference Principle or, more gen-erally, incentive-based arguments can justify inequalities - undercommunism there would be no way to defend inequalities ongrounds of justice.

Since people do not value resources in the same way, a distri-bution of resources that implements the goal of treating personsequally can involve people getting very different things. It cantherefore look inegalitarian.22 In addition, production and trade

22 Despite their many differences, virtually all accounts of equal treatment agreeon this point. See, among others, Dworkin, "Equality of Resources"; Richard

186

Communism

are bound to upset distributional patterns and therefore to per-turb distributions that implement the (complex) ideal of resourceequality. Finally, individuals' free choices or, more generally, their"option luck" will upset distributional patterns.23 For egalitar-ian liberals, the resulting distributions would not violate the idealof resource equality. Quite the contrary. For them, distributiveshares that arise in consequence of individuals' free choices wouldactually implement resource equality - even if, as a result of thesechoices, individuals have very different holdings. Therefore, egal-itarians can defend differences that arise for reasons distinct fromthe efficiency considerations the Difference Principle articulatesand still maintain that only the Difference Principle justifies ine-qualities. They need only argue - in a principled way - that thedifferences they allow are not, in fact, inequalities. Thus, even ifthe Difference Principle, having become otiose, were to mandatestrictly equal distributions, it would be compatible with its view ofjustice that people hold different goods.

I would venture that Rousseauean Marxists can accept the ideathe Difference Principle articulates. They differ from liberals, I willsuggest, only in the kinds of differences they would allow underthe egalitarian ideal. Liberals want egalitarian distributions to besensitive both to individuals' preferences for different resourcesand to their "ambitions," as registered in the choices they freelymake. I shall argue presently that Rousseauean Marxists are evenmore radically egalitarian. They would have distributions be sen-sitive only to individuals' preferences and would therefore inter-vene to rectify (some) of the consequences of individuals' freechoices. Before we turn to this contention, however, a few remarkson Rawlsian justice are in order.

The idea that motivates the Difference Principle, that inequalitiesare justified only if they work to the advantage of the least well-off, is congenial to both liberal and Rousseauean-Marxist posi-tions. Marxists can also agree with Rawlsian liberals that material

Arneson, "Equality and Equality of Opportunity for Welfare," PhilosophicalStudies, vol. 55 (1989), pp. 77-93; and G. A. Cohen, "On the Currency of Egal-itarian Justice," Ethics, vol. 99 (1989), pp. 906-944.

23 The term owes to Dworkin. It designates an individual's luck in deliberatelyundertaken gambles - that is, in choices made in conditions of uncertainty orrisk. It is distinguished from "brute luck," which designates an individual'sgood or bad fortune in "the natural lottery."

187

The general will

rewards can be legitimately used as incentives to productivity. ButRousseauean Marxists need not accede to Rawls's way of introject-ing the latter idea into the Difference Principle itself. The Differ-ence Principle, as Rawls conceives it, accords normative standingto the advantages talented or industrious people are able to pro-cure for themselves. Thus it incorporates incentive effects into thetheory of justice itself. However, it is not at all clear how claims ofjustice can arise out of individuals' differential abilities to take ad-vantage of their respective situations.24 It is one thing to concedethat effort and talent must be rewarded differentially under pre-communist conditions. It is something else to depict this conces-sion as a requirement of justice. Thus the distributional principleproposed for socialism in The Critique of the Gotha Program has pre-cisely the rationale Marx imputed; its justification depends on anappeal to efficiency (broadly conceived). In contrast to Rawls's ex-press view, this rationale does not implicate justice directly. Marxwas surely wrong to eschew justice as an evaluative category.However, this mistake may actually work to the advantage of hisposition in this case. For Marx, inequalities under socialism areunfortunate though unavoidable concessions to a human naturethat only communism in its higher stages can sufficiently trans-form. They are not requirements of justice.

BASIC LIBERTIES

If distribution and contribution are dissociated, individuals can-not claim, as a matter of right, that they are entitled to the wealththeir labor produces simply because they expended their own la-bor. If they are accorded such a right, it must be for other reasons.But if there are no entitlements to the revenues labor generates,must individuals also relinquish control over their labor itself? Ifso, radical egalitarianism would look less attractive than it other-wise might.

Fortunately, Rawlsians and Rousseauean Marxists, insofar astheir position implicitly accords with Rawls's, need not draw thisconclusion. But it is easy to see how it might suggest itself. In gen-eral, ownership is a combination of income rights and controlrights, and claims on the income a resource generates typically docome bundled together with control rights over this resource. But

24 See G. A. Cohen, "Incentives, Inequality, and Community," Tanner Lectures onHuman Values, May 1991, pp. 263-329.

188

Communism

it is plain that these rights can be distinguished conceptually anddisaggregated normatively. Pooling income generated from labordoes not entail pooling labor inputs and vice versa. If Rawls andMarx are right, justice tends to support the former, even if it doesnot strictly require it. But it is hard to imagine any normative prin-ciple that could justify the latter. So pervasive, however, is the pre-vailing view of ownership - joining income rights with controlrights - that procommunists must nevertheless confront issuesabout labor pooling and labor conscription.

Arguably, under communism, the problem would solve itselfautomatically. If a society could reproduce itself with hardly anylabor expenditures, it could safely rely on individuals to expendlabor according to their uncoerced desires. Diminishing scarcitywould therefore obviate any need for labor conscription. But be-fore we resort to this "technological fix," it is well to recall the pri-ority Rawls accords to the equal distribution of basic rights andliberties in all well-ordered and (moderately) developed societies.25

Rawls's principles of justice are intended to regulate the distri-bution of "primary goods" - basic rights and liberties, income andwealth, powers and offices, and "the bases for self-respect." InRawls's view, justice requires that basic rights and liberties be dis-tributed equally and to the greatest extent possible, and that allother primary goods be distributed as the Difference Principle re-quires (so long as opportunities for benefiting differentially arethemselves equally distributed). These principles are lexically or-dered; the first must be satisfied before the second goes into effect.It is, of course, a matter of controversy what the basic rights andliberties are. Rawls, like most liberals, understands the term to re-fer in the main to the political rights that liberal polities tradition-ally accord their subjects. Control rights over one's own body andpowers are assumed and therefore not directly acknowledged. But

25 Rawls distinguishes the "general" from the "special" theory of justice: Theformer is given by the Difference Principle under conditions of equal oppor-tunity; the special theory, in addition, accords lexical priority to basic rightsand liberties. In Rawls's view, individuals in the "original position" wouldadopt the special theory whenever it would be "irrational [for them] to ac-knowledge a lesser liberty for the sake of greater material means and amenitiesof office." According to Rawls, this option would always be irrational except inconditions of extreme material deprivation. Rawls expressly holds that existingliberal (capitalist) societies exceed this threshold. Thus he thinks the specialtheory already applicable at levels of development vastly inferior to those Marxdeemed necessary for socialism - and a fortiori communism; cf. A Theory ofJustice, pp. 60-64 and esp. pp. 541-548.

The general will

it is well to be explicit - especially if prevailing views of propertythat join income rights with control rights are challenged. It istherefore fair to assert what Rawlsian liberalism implicitly under-stands: that, in specifying the basic rights and liberties they woulddistribute equally, parties in the original position would instituteprotections against the appropriation of their bodies and powersby the whole community. They would do so whether or not thecommunity pools the wealth their bodies and powers generate.

Thus there is a "rational kernel" implicit in those (non-Rawlsian)strains of liberalism that endorse the idea of self-ownership.26 Self-ownership confers rights against coerced deployments of individ-uals' bodies and powers. But the same result can be obtainedwithout invoking a notion steeped in a form of discourse appro-priate for neo-Lockean libertarians but hardly for liberal or radicalegalitarians. We need only recognize that control rights over one'sown body and powers, as distinct from rights to the income these"assets" generate, are basic rights in Rawls's sense and that, likeother basic rights, they should be distributed equally and to thegreatest extent possible. If, under communist conditions, the Dif-ference Principle warrants income pooling, the lexical priority ac-corded these control rights continues to guard against laborconscription or other unsavory and unwarranted features thatmight be thought to afflict a social order in which contribution anddistribution are radically dissociated.

JUSTICE INTERNALIZED

Under communist conditions, justice and equality converge, and ajust - or equal - society can be sustained without a state, thoughnot entirely without some public coercive force to overcome collec-tive weaknesses of will. These claims capture most of what Marxenvisioned for communism. Therefore communism is not quite be-yond justice. It is, however, beyond the deleterious effects of thecircumstances of justice.

When the conditions that make justice necessary and possible,relative scarcity and general self-regard, are most acute, justice isdifficult to attain. Justice is then in tension with peoples' inclina-tions to treat others unfairly - in those all-too-common cases in

26 Cf. my "Capitalist Persons," in Ellen Frankel Paul, Fred D. Miller, Jr., JeffreyPaul, and John Ahrens (eds.), Capitalism (Oxford: Basil Blackwell, 1989), pp.39-59. See also Richard J. Arneson, "Lockean Self-Ownership: Towards a Dem-olition," Political Studies, vol. 29, no. 1 (1991), pp. 36-54.

190

Communism

which, by so doing, they can accrue a larger share of the relativelyscarce resources they covet. Then it is all the more in each of ourinterests that institutions and practices be just. Therefore thestruggle to make society conform to what justice requires is an ur-gent task. On the other hand, when scarcity and self-regard areless overwhelming, justice is more readily internalized. In thosecircumstances, just institutions would emerge as by-products ofindividuals spontaneously being just. Like the de jure state of TheSocial Contract, social justice will be realized fully only when it isno longer needed.

But why would people be disposed to achieve justice when theyno longer need it? Because justice or, at least, modern egalitarianjustice articulates a commitment to the moral equality of persons,a commitment that democratic participation under socialist andlater communist conditions nurtures and expands. In materialconditions this side of communism, equal treatment would notgenerally mandate equal provision. Justice and equality would fre-quently diverge. However, under communist conditions, the ten-dency toward equality implicit in all modern conceptions of justicewould be realized almost perfectly; except for those (few) goodsthat in one way or another remain scarce, justice and equal provi-sion would come to the same thing. People committed to equaltreatment, as people nowadays already are and as they likelywould be all the more under communism, would therefore seek toreproduce the social order that allows for its greatest possible re-alization. The disposition to be just would then be tantamount to acompelling preference, internalized in communist men andwomen, to reproduce the social order in which they live - and inwhich their most fundamental impulses toward self-realizationand sociability are realized to a degree unprecedented in societiesplagued by material scarcity and riven by class divisions.

RESPONSIBILITY

From an egalitarian liberal perspective, individuals ought not to beheld responsible for their good or bad fortune in "the naturallottery" or for its consequences. Mental and physical endow-ments and other "accidents of birth" do not generate entitlementsto distributive shares. However, it may be just to accommodate todifferences that these contingencies help to produce. Thus eventhe most radically egalitarian liberals maintain that capitalist mar-kets, in which outcomes depend in part on happenstance and

191

The general will

other morally arbitrary factors including luck in the natural lot-tery, can be just. But no claims of justice arise from these factorsthemselves. No one has, as it were, a presocial entitlement to theassets that they own in consequence of their good or bad fortune.If capitalist markets are just, it is in spite of - not because of -the morally arbitrary factors that determine the outcomes theyproduce.

However, egalitarian liberals do not exclude the consequencesof luck entirely from the requirements of justice. When individ-uals engage in production and trade, they typically make choicesin conditions of risk or uncertainty. Then, whether they choosewisely or poorly, luck affects the outcomes of their choices signif-icantly, perhaps even decisively. Characteristically, egalitarian lib-erals allow fortune its due in these cases. Thus no liberal endorsesa theory of justice that nullifies the consequences of choices indi-viduals freely make. Justice, in the liberal view, does not demandthat there be compensation for freely made choices that turn outbadly, or that individuals be deprived of benefits generated by"gambles" that turn out well.

Of course, for luck to accord with justice, choices must be gen-uinely free. Whenever they are, respect for autonomy requiresthat individuals be held responsible for what they autonomouslychoose. But individuals' choices are affected by their circum-stances and their psychological propensities - including their atti-tudes toward risk - and these factors would seem to be morallyarbitrary too. Nevertheless, I shall, without argument, accept theidea, defended by Kant, that autonomy can coexist with an empir-ical order governed by strict causality.27 Following Kant, let ustherefore suppose that genuinely autonomous choices do exist -because they involve some further volitional fact beyond empiricaland morally arbitrary contingencies - and concede as well that theacts of will that make autonomous choices autonomous sustainclaims of justice. Even so, actual choices are so confounded withwhat is unequivocally arbitrary that their outcomes, even whenconsonant with justice, can hardly be defended solely on thegrounds that they are just.

It would therefore follow that the distinction liberals draw be-tween "option luck" and "brute" fortune, whatever its analyticalcogency, cannot generally be made in practice. But if it cannot,

27 Cf. Immanuel Kant, The Critique of Pure Reason, trans. Norman Kemp Smith(London: Macmillan, 1963), pp. 409-415.

192

Communism

how can the idea that institutions ought to hold individuals re-sponsible for the choices they (freely) make be defended? Auton-omy may be real at the level of abstraction that Kant intended, butactual choices would seem to be too confounded by morally arbi-trary factors to count as autonomous in a way that could sustainclaims of justice.

What is at issue is not the reality of autonomous choice but itsimplications for social justice. Liberals defend distributions gener-ated by the use individuals make of the resources they control,insofar as they result from free choices made in conditions of un-certainty, but their doing so seems to be in conflict with theirprincipled determination to discount - and wherever possible tonullify - the effects of morally arbitrary contingencies. In otherwords, two core liberal intuitions about justice - that luck does notgenerate morally binding entitlements, and that individuals mustbear the fruits of the choices they freely make - are, to say theleast, in tension. To what extent do similar tensions afflict the po-sitions of other partisans of autonomy who, like Rousseau andMarx, eschew the idea, dear to liberals and nearly everyone else,that all interests are private? To pursue this question is to rethinkthe connection between justice and autonomy and, in so doing, toelaborate some implications of general will coordination.

JUSTICE AND THE GENERAL WILL

It has been recognized at least since Aristotle that the core ideathat justice articulates is that "like cases be treated alike." Then dif-ferent substantive principles - for instance, "to each according toneed" - supply content to the underlying formal principle.28 Ihave suggested that, under communism, with the circumstancesof justice severely mitigated, the principle "to each according toneed" is not strictly a substantive principle of justice at all. It doesnot hold that, under communism, differences of need are morallyrelevant and therefore provide grounds for identifying like casesdeserving of like treatment. It is instead a (speculative) descriptiveprinciple, an account of how distributions under communism willoccur. Under communism, on this understanding, individuals willadd to social wealth according to their abilities, and take from it asthey please, according to their needs.

28 Cf. Chaim Perleman The Idea of Justice and the Problem of Argument, trans. JohnPetrie (New York: Humanities Press, 1963), Chapter 1.

193

The general will

When scarcity is salient, it is of paramount importance that in-dividuals be treated fairly. But what counts as fair treatment,according to intuitions nearly universally entertained, dependsupon facts about these same individuals - on what they have doneor, on some accounts, on what is true of them regardless of whatthey have done. In other words, it is supposed that there exists anintrinsic connection between facts about individuals and their de-fensible claims on societal benefits and burdens. That some rela-tion of this sort should exist is to be expected. If the individual,conceived atomistically, is the proper point of departure for justi-fying social and political arrangements, facts about individualsshould determine the justice of distributions to these individuals.

However, insofar as individuals are not radically independent"atoms" bearing private wills but indivisible members of "wholecommunities" bearing general wills, the facts that determine theirdistributive shares will not be facts about atomic individuals. Pro-ponents of one or another substantive principle of justice do ef-fectively conceive individuals and their properties atomistically.These principles do not therefore lend themselves to articulating anotion of justice for societies of the kind Rousseau and Marx en-visioned. The sense of justice they express can coexist with thenew order, but it is not peculiar to it. It is the old justice that hasnot and, in all likelihood, cannot entirely "wither away." If there isa distinctive notion of justice associated with general will coordi-nation, the properties of individuals that will be pertinent for de-termining their distributive shares must concern individuals quamembers of the "moral and collective bodies" they constitute. Butthe only pertinent fact of this kind is that individuals are free, thatthey are autonomous beings.

Insofar as equality in autonomy is all that matters for justice - asit is when individuals appropriately "place themselves under thesupreme direction of the general will" - justice entails equalitytout court, not inequality in consequence of the choices individuals(autonomously) make. As essentially autonomous beings, individ-uals are equal qua constituents of the moral and collective bodiesthey constitute. Therefore considerations of justice, to the extentthat they remain normatively pertinent, become considerations ofequality. It remains problematic how societies coordinated througha general will, communist societies, would implement this ideal atany given moment and over time - since production, trade, and thevagaries of fortune will occur under communism as well as in pre-communist economies. It is likely that, even under communism,

194

Communism

equality cannot be realized perfectly. It is also likely, as we shall seepresently, that, for reasons extrinsic to but not incompatible withjustice, inequalities of various kinds might persist indefinitely.But "the tendency towards equality," to borrow an expression ofRawls's, would be vastly more pronounced under communismthan in even the most egalitarian liberal societies.29 Where thegeneral will is in control, justice requires that, so far as possible,individuals ought at all times to have equal shares - irrespective oftheir option luck.3°

Liberal conceptions of justice will probably remain appropriatein some contexts for an indefinite period if not forever. As we haveseen, the role of general will coordination in communist societiesis anything but exclusive. And even in those circumstances whereit is appropriate for individuals to be governed by the general will,there will probably remain efficiency reasons, even as scarcityradically diminishes, for rewarding individuals differentially ac-cording to their good or bad fortune in gambles they have freelyundertaken. Thus, under communism, competing intuitions aboutwhat justice requires are likely to coexist - and so too are compet-ing practices. Exactly how this tension will be resolved cannot bespecified in advance. All we can say for now is that, as the cir-cumstances of justice "wither away" and the moral urgency of jus-tice pales, the tensions will become less acute. There may alwaysbe a tendency to countenance inequalities that arise in conse-quence of individuals' free choices. But the balance will surelyshift in the direction of equality irrespective of option luck. Liberalegalitarians have put the link between what individuals do or areand what they get into question. Under communism, the sever-ance of this link will be carried to its limit. If merit or contributionor needs are to be rewarded differentially, it will not be for reasonsof justice.

Marx himself was famously ambivalent about radical egalitarian-ism. In large part, he was drawn to non- and even antiegalitarian

29 For Rawls's very different construal of the tendency he named, see A Theory ofJustice, pp. 100-108.

30 The idea that option luck should be countered rather than embraced, as in lib-eral egalitarian theories, does not imply that there should be compensation forliterally anything individuals do with the resources they control. For example,individuals who squander their resources outright will usually be acting in op-position to the general will. There is therefore no reason of Rousseauean (orMarxist) provenance to seek to equalize outcomes that have been produced inthis way.

The general will

positions for polemical reasons - to distinguish his views of social-ism and communism from those of other socialist and anarchistthinkers. But it also seems that he was insufficiently sensitive tosome of the implications of the model of rational cooperation thathe effectively deployed. But, as generations of Marx's followershave believed, sometimes despite express pronouncements to thecontrary, the struggle for communism is a struggle for equality. Itis, in fact, a struggle for an equality more radical than most ofMarx's opponents envisioned. Equality under communism woulddisconcert anyone still in the thrall of atomic individualism. Forthe egalitarianism that would reign under communism tends toobliterate individual responsibility in the domain of distributivejustice.31 In contrast to the liberal view, for partisans of general willcoordination, no fact about anyone except membership in themoral community itself matters, so far as justice is concerned, fordetermining an individual's distributive share.

THE EXERCISE OF THE GENERAL WILL

In Rousseau's de jure state, virtue facilitates general will coordi-nation even as it exemplifies a volitional principle at odds with it.This tension is evident in the difference Rousseau's political phi-losophy implicitly identified between the just state and civil soci-ety, the difference that led Marx to fault "political emancipation"for intensifying unfreedom (in some respects) at the same that itadvanced it (in others). Under communism, this tension would beseverely mitigated. Just individuals (in the liberal sense) wouldstill be private willers. But, under communist conditions, privatevolition would no longer work to undermine uncoerced coopera-tion between individuals. It would instead work in tandem withgeneral will coordination. Then, necessarily in cases where thegeneral will directs individuals and tendentially in cases where

31 Absolution from responsibility in this sense in no way mitigates an individual'smoral responsibility for freely chosen actions. It has been recognized, at leastsince Aristotle, that distributive and "retributive" or "rectificatory" justiceconstitute distinct domains, linked only by the formal principle that "like casesbe treated alike." Thus the circumstances that make distributive justice morallyurgent can pale without in any way lessening the urgency of retributive justice.And the criteria for counting as a like case can differ markedly from one do-main to another. For Aristotle, retributive justice was largely an affair of insti-tutions, tantamount to a (normative) theory of punishment. But similarconsiderations pertain beyond institutional "retribution" to moral judgmentsgenerally. Cf. Aristotle, Nicomachean Ethics, Book V, section H, Part 1.

196

Communism

private volition remains in force, the freedom and self-realizationof everyone would be everyone's concern.

It would be foolish, from the vantage point of a world in whichuntrammeled capitalism seems (temporarily) triumphant, to try tosurmise which aspects of communist societies would be regulatedby an internalized sense of justice and which by general will co-ordination. However, some very general speculations are feasibleeven now. In this instance, as in so many others, Rousseau's leadis instructive - to a point.

For Rousseau, the general will was to be exercised in the domainof collective choice - that is, in individuals' capacities as compo-nents of the sovereign power. Sovereignty, on his construal, is "theexercise of the general will" because, for (potentially) autonomousagents, the existence of a moral community is the preeminent con-cern and general will coordination is indispensable for the realiza-tion of this end. This contention is insightful and crucial forconceiving collective decision making under communism. ButRousseau's idea can be understood in a number of ways, andRousseau himself provides little help in adjudicating among theseunderstandings.

At the very least, the claim might be that autonomy can onlybe realized within communities that recognize and implementthe moral equality of persons. Then, insofar as each individual'sprincipal interest is the maintenance and enhancement of his orher own autonomy, individuals have an overriding interest in at-tending to the interests of the community that make autonomypossible. For matters in which the community itself is directly im-plicated, therefore, it would be reasonable for individuals to delib-erate as indivisible members of it. But in a well-functioningcommunist polity, as in a de jure state, it is unlikely that issues willarise very often in which the interests of the whole community areat stake. General will deliberation would then seldom impinge oncollective decision making. On this understanding, communistmen and women would be like the Rousseauean citizens who, ac-cording to one plausible reading of Rousseau's intention, delegatevirtually all political tasks to "magistrates," while themselves com-ing together to legislate as indivisible components of "the wholecommunity" only at periodic intervals - and then only to recon-firm (or deny) that the government is executing the sovereign'swill. At the other extreme, it could be held that there is a generalinterest on a wide range of issues and that, as a matter of fact,these general interests advance the autonomy of the individuals

197

The general will

who constitute the whole community. On this view, "the wholepeople would legislate concerning the whole people" regularlyand extensively.32 This claim is plausible to the extent that there isa determinate general interest on many of the issues - or on manyaspects of the issues - that arise for collective choice. This is themost likely understanding of Rousseau's view. Whether or not it iscorrect is of course an empirical question that philosophical spec-ulation is helpless to decide.

In any case, the claim is that, under communist conditions (or intheir unworldly anticipation, the state of The Social Contract) when-ever individuals do appropriately legislate as indivisible parts ofthe whole communities they constitute, they ought to choose op-tions that maximize overall autonomy. This position is not so im-plausible as may at first appear. To be sure, there are interestsbesides autonomy that in general voters ought to heed - welfareinterests, in particular. However, under communism, welfare in-terests are already largely (though not entirely) addressed. Withscarcity substantially overcome and self-interest rendered benign,people can therefore set about to become the free (autonomous)beings that, according to Rousseau's vision, they always essen-tially were. They can make the enhancement of autonomy theiroverwhelmingly preeminent concern.

AGGREGATING AUTONOMY

Insofar as the virtue of (liberal) justice is instilled in communistmen and women, they will contend with each other fairly when-ever their (private) interests put them at odds. Then the likelihoodthat they might seek to advance their own autonomy at the ex-pense of the autonomy of others will be severely mitigated. But itwill still be appropriate to regard aggregate autonomy as the log-ical sum of individuals' autonomies. Whoever would propose themaximization of autonomy would therefore hold a view analogousto the position of utilitarians - provided, of course, that degrees ofautonomy can be regarded quantitatively the way utility is.33 Util-

32 Cf. The Social Contract, Book II, Chapter 6.33 It is not necessary to suppose that there is actually an autonomy metric but

only that, in a very general sense, autonomy can be conceived as if it could berepresented as an additive magnitude. Despite all the attention paid to mea-suring utility intra- and interpersonally, most utilitarians and their critics ef-fectively assume a similar position on utility. It is enough, for them, in

198

Communism

itarians would maximize social utility, Us, where Us is the logicalsum of the utilities of individuals 1, . . . , n.

(i) us = u, + u2 + un.Autonomy maximizers would maximize social autonomy, As,where As is the logical sum of the autonomies of individuals1, . . . , n.

(2) As = A, + A2/ . . . , +An.

Therefore the possibility that some individuals' autonomy may bediminished in order to maximize As remains at least a theoreticalpossibility.

On the other hand, when individuals' interests are internally re-lated, as they are when their actions are directed by a general will,the general interest, G - insofar as it is tantamount to an interest inenhancing individual autonomy - is not the logical sum of indi-viduals' autonomies but more nearly the effect of their interac-tions. Rousseau's idea does not lend itself well to formalrepresentation, but a reasonable approximation would be:

(3) G = A1 x A2 x . . . x An.

Whoever is concerned to maximize autonomy and who acknowl-edges a place for both justice and general will coordination wouldtherefore aim to combine (2) and (3). The global maximind, totalautonomy, At, would then be the sum of individual autonomiesplus their interactive effects.

(4) At = As + G,

or equivalently,

(5) At = A, + A2 + . . . + An + (A, x A2 x . . . An).

In these terms, Rousseau and Marx after him, insofar as his viewof communism is Rousseauean, implicitly claimed: (a) that, in a dejure state or under communist conditions, At is indeed the propermaximind, and (b) that the way to maximize At is to be just in pri-vate interactions and In the public arena to seek to maximize G di-rectly. The rationale for (a) has already been discussed. In order to

assessing utilitarianism, that we can think of utility as an additive magnitudewhether or not there is a good way - in practice or even in principle - to con-struct an appropriate metric.

199

The general will

defend (b), an additional claim, also of Rousseauean provenance,must be brought to bear.

As noted, a very weak construal of Rousseau's core idea wouldhave autonomy be realized only within communities that recog-nize and implement the moral equality of persons, and would thenconclude that, since autonomy enhancement is everyone's princi-pal interest, individuals have an overriding interest in maintainingthe viability of the community that makes autonomy possible. Butsince the existence of the political community itself is seldom likelyto be implicated in collective decision making, the preeminence ofgeneral will coordination would properly become manifest only inrare instances. I have suggested that Rousseau probably intendedmore than this in grounding the de jure state on "the exercise ofthe general will" but that the reasonableness of this claim cannotbe settled a priori. The issue can however be represented perspic-uously.

Schematically, what Rousseau can be taken to have claimed isthat individual autonomy itself is a function, at least in part, of to-tal social autonomy. Then, if A{ designates the autonomy of the ithindividual and x designates causal factors distinct from total socialautonomy that affect i's autonomy,

(6) A = f(At) + g(x).Rousseau provides scarce help in specifying the form of the func-tions / and g, although it is evident that he does assume them toincrease monotonically. In these terms, questions about the scopeof general will coordination are questions about the importance ofthe function g(x). At one extreme, it could be held that the value ofA{ is mainly determined by this function; at the other, that g(x) isset at zero or nearly zero. As remarked, the relative importance ofAt and x respectively is ultimately an empirical question.

From the timeless perspective of The Social Contract, Rousseauappears to be committed to the view that in a de jure state indi-vidual autonomy, Aif is largely a function of total social autonomy,At. Rousseauean Marxists, equipped with a theory of history's tra-jectory, would draw the same conclusion, but relativized to a par-ticular historical context. For reasons already marshaled, theirconviction is that in the radically egalitarian world of communismthe enhancement of individual autonomy would increasingly be-come a collective endeavor in which "the condition for the free de-velopment of each" is indeed "the free development of all."Insofar as individuals imbued with the virtue of justice and living

200

Communism

in a world with massive economic surpluses would need to rendercollective choices, and insofar as there is a determinate general in-terest implicated in the choices they must make, they would be in-clined, increasingly, to make this general interest their own and toplace themselves, accordingly, "under the supreme direction ofthe general will."

It should be noted, finally, that although the diminution of someindividuals' autonomies will seldom, if ever, be required for theenhancement of total autonomy under communist conditions it islikely, even under communism, that in order to maximize At someindividuals' autonomies will be enhanced more than others. Un-dertaking some public work - constructing a new highway, for in-stance - will almost certainly affect individuals differently, withdifferent implications for their autonomy. But insofar as formula(6), joined with formula (4) or (5), captures a sound idea, when-ever anyone's autonomy is enhanced, everyone becomes, to somedegree, more free. Egalitarians cannot hope for more. Inasmuch asautonomy depends on access to resources, and since even social-ized resources are bound to be differentially available to individ-uals, it is unlikely that autonomy can ever be equally distributed.But under communism - a society without compelling scarcities,directed by a general will, populated by just individuals, and withdistributive shares equalized to a degree uncontemplated even byliberal egalitarians - autonomy will be, as it were, sufficientlyabundant that, like most other things, its distribution will nolonger be of particular normative concern.

As remarked in Chapter 1, the most likely arena for general willcoordination this side of communism is in small groups and inti-mate associations. There is no reason to doubt that this situationwill remain true after communism too. If anything, a multiplica-tion of "moral and collective bodies," of various sizes and (over-lapping) constituencies, will be a likely feature of communistsocieties. In the (decentralized) administrative organs throughwhich individuals' behaviors will be coordinated, institutions for"the administration of things," people may often find it reasonableto subordinate their private wills to the general will of these col-lective entities for just the reason Rousseau identified: in order tofurther their own most fundamental concerns.

Will "the exercise of the general will" then supplant its comple-ment, the virtue of justice? Not entirely, since justice internalizedwill continue to coordinate individuals' behaviors in those spheresof life in which they set about to discover and realize their own

201

The general will

distinctive potentialities, and since "the realm of freedom" is thebest imaginable, materially feasible social order conducive to thisend. Paradoxically, the recognition and implementation of the so-cial dimension of autonomy are the conditions for the fullest pos-sible realization of the core values, autonomy and self-realization,that liberals share with Rousseauean Marxists. The difference isthat liberal construals of these ideas are flawed by their allegianceto the governing metaphor of the atomic individual. FollowingRousseau, Marxists can overcome this allegiance as they grope,bereft of "recipes for the cookshops of the future," toward discov-ering what can now only be grasped schematically: the preciserole of general will coordination under communism, the earthly -that is, materially possible - "form of association" in which, fol-lowing Rousseau's prescription, "each one, while uniting with all,nevertheless obeys only himself."34


2 0 2

Chapter 9

After Communism, communism?

We are still too close to a yet unfolding story to draw more thantentative conclusions about (big-C) Communism's demise. Buteven without the benefit of a long historical perspective, it is plainthat 1989, the year Communism ended in Eastern Europe and im-ploded dramatically in the Soviet Union itself, and 1991, whenCommunism officially ended even in the land of its origin, werewatershed years for socialist theory and practice. They marked theend of the historical project begun in 1917 and prepared, in part,in the three or four decades preceding the Bolshevik Revolution -during Marxism's classical phase, the Age of the Second Interna-tional. As of this writing, China remains officially Communist,accounting for nearly a quarter of the world's population, andthere are other survivors too at the peripheries of the old Commu-nist system - in North Korea, Vietnam, and, of course, Cuba. Butthe exceptional status of these regimes is unlikely to last. Alreadyin the Chinese case, it is mainly the political system that remainsCommunist (and even Stalinist). Vast sections of China have beenopened up to infusions of capital and technology from Japan, EastAsia, and the West - a fact that helps explain the persistence of theregime in power. Even if China remains officially Communist forsome time, it is rapidly becoming transformed from an economicsystem built on the Soviet model into a source of cheap labor forforeign capital, a workshop for industries developed and orga-nized elsewhere.

The causes and significance of these transformations will be de-bated for years. It has become commonplace for celebrants ofcapitalism in the West to gloat over the poor economic perfor-mance of the erstwhile socialist countries and to take their demiseas proof of capitalism's superiority. But, in fact, the Soviet Unionand Eastern Europe, before glasnost and perestroika, had been doing

203

The general will

rather well by their own historical standards and in comparisonwith most Third World countries. In any case, it does not appearthat any precipitous economic crisis brought the Second World toits demise. It is true that, for more than twenty years, the Easternbloc countries had performed poorly when measured against lev-els of economic growth in the West and in some newly industri-alized countries in the Third World.1 No doubt, the perception of(comparative) stagnation, the sense that the gap in economic per-formance between the East and the West was no longer closing butactually intensifying, delegitimated Communist economic systemsand helped lead to their eventual unraveling. More important,however, were the accumulated effects of the chronic shortcom-ings that had plagued the Soviet system from its inception: noto-rious inefficiencies in production and distribution, a generalfailure to meet consumer demands, and a glaring inability to pro-duce at international standards, to innovate, or to marshal entre-preneurial forces.

But Communism almost certainly declined and fell more for so-cial and political than strictly economic reasons. The rise of ahighly educated, technocratic class and increasing cultural levelsthroughout the general population came into conflict with the oldcommand system of economic management and its political super-structure. Ironically, in light of the future classical Marxism im-puted to capitalism, the very success of the state bureaucraticsystem in the Eastern countries undid the conditions for its ownreproducibility; the Soviet model of socialism dug its own grave. Ithas been famously argued by Herbert Marcuse that, despite thebest intentions of liberal social philosophers, free speech and tol-erance generally - in conditions of economic inequality and in thepresence of mechanisms that militate against critical thinking -can work to ensure conformity and even function "repressively."2

In the Soviet Union and Eastern Europe, the phenomenon of re-pressive tolerance seems hardly to have existed. There only repres-sion functioned repressively. Thus glasnost, however necessary itmight have been for economic development, was deadly for thepolitical order of the erstwhile socialist countries.

It is impossible, after only a few years, to assess these changes.It is plain, however, that the optimism of 1989 has quickly soured

1 They had also done remarkably well in comparison with most of the ThirdWorld, a not insignificant sector of the world capitalist system.

2 See Chapter 3, note 15, herein.

204


and that the short- and middle-term prospects for the countries ofthe former Soviet bloc seem bleak - particularly on the economicfront where the introduction of "free markets" was supposed towork magic. Perhaps someday parts of Eastern Europe will reachWestern European standards of economic performance. However,for the foreseeable future, unemployment, stagnation, and con-tinuing environmental hazards lie ahead. And devastating eco-nomic consequences for third parties are already evident too.Thus, in even the most optimistic of scenarios, Western and Jap-anese investments will be channeled away from Third Worldcountries, especially from Africa and Latin America, exacerbatingthe already severe hardships the peoples of those continents face.With the end of Communism, the worst-off will therefore becomeworse off still, even allowing, as is far from assured, that the lot ofthe majority of people in what was once the Second World willeventually improve.

It is also plain that the end of Communism has unleashed longrepressed nationalist and chauvinist passions along with religiousand other reactionary and obscurantist convictions. By 1992, theconsequences of political and moral regression have alreadycaused incalculable harm. In addition, the collapse of the SovietUnion has removed the one constraint on the exercise of Americanmilitary power in the Third World - with devastating conse-quences evident in the Gulf War, in the "moderation" of statesthat might otherwise seek a measure of national independencefrom the world capitalist system, and in the diminished programsand expectations of the remaining national liberation movements.Finally, even assuming economic success and discounting geopo-litical consequences, the populations of the formerly socialistcountries will have succeeded merely in trading one set of evils foranother. The trade-off will seem appealing only to those too lack-ing in imagination to realize that the systems in place prior to 1989hardly exhaust the historical possibilities.

If the experience of liberal democracies in the West provides anindication, the regimes likely to succeed the formerly Communiststates will enjoy a greater measure of de facto legitimacy if theyprovide a (largely illusory) measure of popular control over publicofficials and if they are able to offer historically high standards ofliving to some significant portion of the population and the hopeof higher standards of living to another significant portion. SomeWestern countries - the United States, for example - have at-tained extraordinary levels of political stability in so doing, even as

205

The general will

nearly a third of the population is left politically and economicallymarginalized. It is far from clear, however, that this sense of legit-imacy will continue indefinitely. In time, as in all the Westerncountries, even the relatively well-off will come to wonderwhether a consumer society, with its perils to human well-beingand to the world in which human beings live, is a prize worthstruggling for after all.3

But this is not the place to speculate on the future of the erst-while Second World or even on the causes of the fall of the systemin place there. The question instead is what, if anything, the end ofCommunism has to do with small-c communism. The immediatebut too easy answer is Nothing at all. This response cannot standwithout modification, but there is substantial merit to it nonethe-less. It is therefore worth exploring this answer before broachingthe more subtle issues that make it too facile.

SMALL-C COMMUNISM

The idea of communism arose within radical and revolutionarycurrents in late-eighteenth- and early-nineteenth-century Frenchpolitics. But its connection - or lack of it - with Communism hasto do with the appropriation of that idea by Marx and Marxistsafter him. The too facile reason why the end of Communismhas nothing to do with communism is that Communism has noth-ing - or only very little - to do with the communism of Marx-ist theory.

The term "Marxist theory" actually designates a constellation oftheories that are not nearly so tightly integrated as was once as-sumed. Two components of this constellation are of particular rel-evance: historical materialism (Marx's theory of history) andMarxist political theory.

Historical materialism is a theory of history's structure and di-rection. It purports to provide an account of necessary and suffi-cient conditions for epochal historical change. In the historicalmaterialist scheme, Communism is the final epochal stage, the his-torical successor of capitalism. However, communism admits of

3 Twenty-five years ago, this consciousness was sufficiently widespread in Franceand then in Italy to promote a sense of impending social revolution. Elsewhere,at least among students and youth generally, it became the basis for a large andinsurgent social movement. Today, social movements inspired by a genuinelyoppositional political vision are in eclipse, but awareness of the evils of (late)capitalism persists unabated.

2O6


phases. Socialism, still a form of class society in which the work-ing class is politically dominant, is its lowest phase. It has becomeclear in recent years that the orthodox theory can be weakenedand still remain "Marxist"; it can even be recast as an account onlyof necessary material conditions for the historical transformationsthe orthodox theory describes.4 But any account of history's struc-ture and direction that remains continuous conceptually withMarx's theory would maintain that socialism, communism's low-est phase, can only come about after capitalism has exhausted itspotential for developing productive forces.5 In other words, capi-talism must first be "mature," even rotten (or rotting), before thesocialist project can hope to succeed.

Now Communism was launched not where capitalism was al-ready rotting but in one of Europe's least developed regions. Thusthe Bolsheviks, because they were historical materialists, as-sumed, at the outset, that they were only attacking capitalism at its"weakest link" - the first step in a revolution of global or at leastEuropean dimensions. The idea that socialism could be built inone country was a later innovation - still framed, significantly,within a theoretical perspective shaped by classical historicalmaterialist conceptions. But the doctrine of "socialism in onecountry," introduced under Stalin's aegis, represented a desper-ate - and barely coherent - attempt to legitimate Soviet reality inMarxist terms. Classical Marxists have always been wary of it, ifnot implacably opposed.

From a historical materialist vantage point, then, Communismwas bound to fail. If anything, what embarrasses the theory is not1989 or 1991 but the fact that it took more than seventy years afterthe Bolshevik project began for it to run its course in defeat.

It is also plain that Communism had little, if anything, to dowith the strain of Marxist political theory or practice sketched inChapter 5. Far from undermining the conditions for the reproduc-ibility of the state form of political governance, the Communistspermanently installed a repressive and ideological state appara-tus - varying, over the years, in severity but not in form - that

4 See Wright, Levine, and Sober, Reconstructing Marxism, Chapter 5,5 In this respect, the theory and practice of Communism after Communist power

became entrenched in the Soviet Union are not conceptually continuous withMarx's theory. In view of the enormous influence of Communism on Marxistand socialist thought throughout most of this century, especially in the ThirdWorld, a similar conclusion would have to be drawn for many, if not most,twentieth-century Marxists.

207

The general will

worked, as ought to have been anticipated, to reinforce aspects ofclass society. For decades, socialists in the West and elsewherepondered over the nature of Soviet society and shaped their polit-ical orientations around competing analyses of it. Many socialistshave argued, with some plausibility, that the Soviet state evenbrought a "new class" of state and party bureaucrats to power. Butwhatever the true nature of Soviet and Eastern European societiesbefore 1989, the one hypothesis that seems least defensible to allbut the most ideologically blinded is the idea that the political sys-tem in place in those countries represented a form of workers'power. Again, in the Marxist scheme, it is precisely workers'power that socialist societies in transition to communism are sup-posed to implement.

Insofar as there was a Marxist warrant for neglecting to trans-form political institutions - to "smash" the old forms, as Leninwould have it, and introduce political structures of a new kind intheir stead - it was drawn ironically from a crude historical mate-rialist world view. Historical materialism invests the dynamic ofhistorical change in the development of productive forces. The Bol-sheviks accordingly took it upon themselves to develop productiveforces by any means necessary - above all, by brutally enforced hi-erarchical controls and the subordination of all social struggles tothe paramount task of increasing productivity. Social relations andtheir political and ideological superstructures could then take careof themselves - evolving into appropriate forms as productiveforces advanced. It need hardly be said that this attitude is not con-sonant with Marx's express views, nor is it implied by historicalmaterialist positions, properly understood.

In short, Communism neglected historical materialism when itshould have taken it seriously and then misappropriated it, turn-ing it into an excuse for not attempting the radical transformationof the social order Communists ruled.6 One cannot help thinkingthat this manifest misuse of historical materialist positions wasmotivated, at least in part, by an attempt to reinforce the already

6 Without successful revolutionary endeavors in more advanced capitalist coun-tries, it is unclear that the Bolsheviks or their successors could have done muchbetter. Nor is it clear, except perhaps retrospectively, that they ought to haveabstained from taking power in 1917. What is clear is just that historical mate-rialism provides strong grounds for predicting the deformation and eventual de-mise of their revolutionary project - so long as socialism remained in place onlyin (relatively) undeveloped parts of the globe.

208


considerable power of state and party bureaucracies.7 Thus, inso-far as communism is an idea proper to Marxist theories, Commu-nism had nothing to do with it. Therefore it would appear that theend of Communism does not in any way challenge procommunistconvictions. But, of course, this response is too quick.

SOCIALISM

Since even a meretricious link between Communism and commu-nism long ago dropped out of Communist thinking, and sincecommunism has played little role of late either in the popularimagination or in political theory, the implosion of Communismand then its hasty termination in Eastern Europe and the SovietUnion and its decline and transformation in Western Europe andin most of the Third World have not done much harm to commu-nism as an idea. But these events plainly have damaged the ideaof socialism.

The damage has been exacerbated and perhaps even impelledby a parallel decline and transformation of social democracy. Thisphenomenon, which has been more subtle but nevertheless signif-icant for the socialist project, stems from different causes: amongothers, the decline of labor movements in the advanced capitalistcountries and a general inability on the part of the older SocialDemocratic parties to join forces constructively with ecologists,with feminists, or indeed with any of "the new social movements."In consequence, Social Democrats have become increasingly dis-inclined to imagine alternatives to capitalism or to propose evenmodest "structural reforms" of the regimes they manage or serveas "loyal oppositions."

In the First and Third Worlds, socialism has been hurt too by theapparent economic successes of capitalism in the 1980s. Self-representations to the contrary, the advanced capitalist countrieshave hardly succeeded in solving their fundamental social prob-lems. But they have succeeded in enriching larger segments of theirpopulations than heretofore, without too seriously diminishing

7 In this regard, it is well to recall that an important and compelling (partial) ex-planation for Communism's demise is that a large number of the new genera-tion of state and party functionaries thought that they could do even better forthemselves if they could take advantage of their expertise by becoming capital-ists or at least by modifying the old command structures of the economies theygoverned.

209

The general will

the consumption possibilities of many - but by no means all - ofthose who have become worse off. The even more distorted eco-nomic successes registered by export-driven capitalist develop-ment in a few Third World countries in Asia have also redoundedto socialism's disrepute.

Why these phenomena should be taken to justify traditional(nonegalitarian) economic liberalism is something for historians ofthe future, with the advantages of hindsight, to try to explain. Butthere can be no doubt that even laissez-faire capitalism is nowwidely thought to have been vindicated. Until quite recently, itseemed politically appropriate to debate the relative merits of cap-italism and socialism. I have argued that, in that debate, socialismhad the upper hand.8 Nothing of substance has changed, but soabruptly has the political and intellectual climate been trans-formed that nowadays the idea that socialism might triumph overor even rival capitalism seems anachronistic even in academic dis-cussions. Indeed, in many circles, the very idea of socialism hascome to seem increasingly quaint.

To account for this turn of events fully, these and other vicissi-tudes of Communism, social democracy, and (late) capitalismwould have to be taken into account. I would venture that it isCommunism's fate that has had the greatest effect. But I will notpursue the issue further here. Instead, I shall turn to a problemthat this change in intellectual fashion betokens. This problem isbest approached not by reacting to the situations of political move-ments that in varying degrees identified with or opposed Marxismbut by considering the vicissitudes of Marxism itself. Small-c com-munism is an idea proper to Marxist theory. If there is somethingabout the end of Communism that has made it (more) problematic,we can best see what it is by focusing on the theory in which itis embedded.

MARXISM TODAY

For more than two decades, "analytical" Marxists have pursuedthe reconstruction, defense, and criticism of distinctively Marxistpositions - in a way that has finally made many Marxist claimsamenable to rational adjudication and revision.9 The result has

8 See Levine, Arguing for Socialism, passim; on the political urgency of the capi-talism-socialism debate, see esp. pp. 1-4.

9 Cf. Wright, Levine, and Sober, Reconstructing Marxism, Chapter 1.

2 1 0


been damaging to traditional understandings, but not nearly sodevastating as is widely believed. It will be instructive to sketchhow matters now stand before returning to communism in light ofthat understanding.

Historically, there were three kinds of projects that Marxists ef-fectively combined - either because they recognized them as dis-tinct and asserted a connection between them or because theyfailed even to acknowledge their differences.10 There was first ofall "scientific socialism," an account of "the laws of motion" ofmodes of production generally and, in particular, of the capitalistmode of production. Then there was class analysis, a view aboutwhat is explanatory of a range of phenomena of special interest toscientific socialists. Finally, there was the theory and practice ofclass emancipation, an attempt to identify and alleviate a particu-lar form of oppression. Marxism as a theory of class emancipationidentified the disease afflicting the existing world; Marxism asclass analysis provided a diagnosis of its (rectifiable) causes; Marx-ism as scientific socialism identified the cure. Without class anal-ysis and scientific socialism, the emancipatory critique would onlybe a moral condemnation; without the emancipatory objective,class analysis and scientific socialism would be of scientific inter-est only.

Until recently, the appeal of Marxism came, in large part, fromthe purported unity of these elements. Together they provided abasis for thinking that the principal oppressions of the existing or-der could be overcome; that human miseries could be understoodscientifically and that their elimination could become the target ofa feasible - indeed, of an ongoing - political project.

But in light of recent work, Marxism as scientific socialism andMarxism as class analysis have had their explanatory pretensionsreduced. I have already suggested that, at best, historical materi-alism explains much less than was thought. It almost certainlydoes not explain the real course of historical change. At most, itidentifies an endogenous process that establishes a historicalagenda, an account of the economic structures that become mate-rially possible or impossible as productive forces develop. It is alsoan open question how explanatory class analysis is. There is rea-son to think that it explains more than many people nowadays, in-cluding some who once thought it explained (nearly) everything,would be inclined to suppose. But it surely does not explain all

10 Many of the claims that follow are elaborated and defended in ibid., Chapter 8.

211

The general will

that Marxists used to think. More importantly, it may not connectwith scientific socialism as tightly as was formerly thought. Thusclass analysis may not figure as prominently as was once assumedin accounting for capitalism and its futures - or socialism and itspossibilities.

Even more tellingly, the link between class emancipation andMarx's explanatory objectives is now less clear than it once seemedto be. The problem is not just that new social movements havebrought to the fore the idea that there are oppressions that are notclass oppressions. Then we would only be left with a diffuse plu-ralist emancipatory project - as some "post-Marxists/7 includingmany feminists, keep suggesting. The problem is rather that so-cialism itself, or at least its role in any emancipatory project, nolonger seems as obvious as it formerly did. It is at least evidentthat not just any socialism is preferable to any capitalism. Thewholesale rejection of Communism by the masses of people livingunder Communist regimes makes this conclusion indisputable.

An autobiographical digression will illustrate what I have inmind. In Arguing for Socialism, written less than a decade ago whenthe socialism-capitalism debate still seemed timely, I posed thefollowing question: Given prevailing views of such values as free-dom, justice, welfare, and democracy, can we say that socialism isbetter (for realizing these values) than capitalism or vice versa? Notsurprisingly, I found that nothing very sure could be concludedwithout more specification of the particular institutional arrange-ments through which these economic structures were maintained.What was then called "actually existing socialism" was plainly lesssatisfactory with respect to many of the values in considerationthan some existing (capitalist) liberal democracies. But I arguedthat many of the procapitalist arguments floated at the time car-ried little force and that, even at the level of abstraction at whichmy question was posed, a weak case could be made for socialismover capitalism with respect to all of the values in question, exceptperhaps welfare (or more exactly economic efficiency).

Since the terms "socialism" and "capitalism" are used veryloosely in our political culture, it was a desideratum to definethem, for my purpose, in a way that encompassed a wide range oflegitimate uses. I also wanted definitions that were not tenden-tious - that did not, for instance, make socialism win the democ-racy competition by definition, as many prosocialist writers do.Finally, I wanted my definitions to concur with the historical ma-

2 1 2


terialist purchase on the subject. I therefore defined capitalismminimally, in the historical materialist way, as an economic struc-ture without private ownership of other persons but with privateproperty in external productive assets. Then socialism (or, better,postcapitalism) would, like capitalism, deny ownership of (other)persons but also exclude private property in (productively signif-icant) external things. This is not the place to resume why, despitemassive indeterminacies, socialism, so defined, fares better thancapitalism. But it should be noted that this conclusion, and theproblem it addresses, reflected precisely what the received Marx-ist view assumed: that the deprivatization of external things wasinstrumental for realizing fundamental normative objectives. Itthen followed that actually existing socialism, for all its faults andfor all the respects in which it was worse (by consensus standards)than some existing capitalist regimes, was nevertheless a better ormore advanced social system, because of its economic structure.

Strictly speaking, this position is not incompatible with whathappened in 1989 and 1991. Nor is it a view that I would now re-pudiate. But it can no longer be assumed, in light of recent events,that socialism (postcapitalism) and capitalism are the right catego-ries for comparison in light of the overwhelming popular repudi-ation of the (allegedly) more advanced form of society by themasses of people living in its grip.

To be sure, Communism and communism have little in common.But there is one powerful understanding they share inasmuch asboth arose within the socialist movement and both, even Commu-nism, have roots in the Marxist tradition. For both Communistsand communists, capitalism's (likely) future is socialism, and so-cialism (and eventually communism) is indispensable for humanemancipation. The events of 1989 and 1991 in no way prove theseviews wrong. But the massive popular rejection of "actually exist-ing socialism" does show that even the most dedicated proponentof communism cannot simply assume them any longer. The rela-tion among capitalism, socialism, and human emancipation mustbe rethought, even if, as I believe, the old understandings will fi-nally be vindicated.

Of course one could argue that the Soviet Union and the re-gimes modeled on it, their laws and self-representations notwith-standing, were not genuinely socialist (postcapitalist). This viewmay seem disingenuous, but it is worth taking seriously as a the-oretical claim. It has been advanced in different ways since the

213

The general will

1930s by dissident Trotskyists11 and since the 1960s also by someMaoists.12 In the final analysis, the nature of these societies isan empirical question. I will not pursue the issue here except toregister the opinion that the popular perception of the socialistcountries as socialist is almost certainly correct in the sense thattheir principal productive resources were in fact deprivatized.Of course, the means of production in these countries were not inany meaningful way "socialized." Nothing that has happened inrecent years therefore impugns the idea that the socialization -indeed, the democratization - of society's principal means ofproduction is instrumental for communism. But the more funda-mental Marxist idea - that deprivatization is in itself emancipa-tory - has indeed been put in question.

Specifically, what needs to be reassessed is the connection be-tween socialism, in the historical materialist sense, and the eman-cipatory interests that have made socialism and communismattractive to those with more to lose than their chains. The chal-lenge is to rethink the connection between Marxism as scientificsocialism and Marxism as a theory of human emancipation. Theevents of 1989 and 1991 are relevant to efforts to address this chal-lenge if only because they put historical materialism in question inpreviously unforeseen ways.

HISTORICAL MATERIALISM

Historical materialism, the theoretical basis for scientific socialism,postulates the existence of discrete economic structures and of anendogenous dynamic connection between them. In its orthodoxversion, it proposes, as it were, a map of historical possibilities andan account of necessary movements along that map. Its weakerversion, which only purports to show what economic structuresare materially possible, retains the orthodox theory's map andeven its notion of directionality. It only denies the inexorability ofmovement along the depicted trajectory. The difference in histor-ical materialisms therefore does not touch on the features of themap. Historical materialists, however, have disagreed on what ep-ochal structures the theory recognizes. Marx sometimes populated

11 See, for example, James Burnham, The Managerial Revolution (London, 1942).12 See, for example, Charles Bettelheim, Calcul economique et formes de propriete

(Paris: Maspero, 1971), La transition vers Veconomie socialiste (Paris: Maspero,1970), and Class Struggles in the USSR, vol. 1 (New York: Monthly Review Press,1976), vol. 2 (New York: Monthly Review Press, 1978).

214


the historical trajectory with a host of precapitalist economicformations.13 G. A. Cohen, in reconstructing Marx's theory with aview of defending it, identified only four epochal structures.14 Butall historical materialists agree that capitalism and socialism are onthe list and that, at the level of abstraction at which the theory ispitched, these modes of production are mutually exclusive histor-ical alternatives.

At that level of abstraction, Cohen's or Marx's periodization ofhistory is as defensible (or indefensible) as it ever was. However,what has become more clear is that this level of abstraction is lessuseful for plotting the course of real history than even proponentsof weak historical materialism used to think. Again, an autobio-graphical digression will illustrate the point.

In Arguing for Socialism, I supposed, following Marx, that polit-ical, ideological, and cultural factors were of secondary importancefor understanding and assessing historical transformations of anepochal stature; what mainly mattered was "the economic base." Idid maintain that forms of the state and even ideology could bedecisive for understanding the future(s) of existing socialist andcapitalist societies. But I assumed that what was most important -for implementing the normative values in consideration - wasthe underlying set of production relations. Again, this position isnot refuted by recent events. But these events suggest that thehistorical materialist scheme, at most, charts only very long-rangestructural possibilities. Recent history forces theoreticians to ques-tion whether these possibilities may not be too remote for thepolitical (and emancipatory) tasks to which they had been previ-ously put.

According to the traditional historical materialist understand-ing, particular events, no matter how they are individuated, canbe explained causally at least in principle, and historical trendscan also be explained. However, these explananda are distinct;hence the frequently remarked observation that Marx's historicalwritings bear little evident connection to his theorizing about his-tory. In any case, for Marxist politics, the account of historicaltrends is central. For Marxism since its inception, the struggle toadvance to the next epochal stage - to socialism - has been theparamount emancipatory objective.

13 See, for example, The German Ideology, Part 1; and, more in point, the portionsof the Grundrisse published in Karl Marx, Pre-Capitalist Economic Formations, ed.E. J. Hobsbawm (New York: International Publishers, 1964).

14 Cohen, Karl Marx's Theory of History, pp. 197-201.

215

The general will

However, it is now beginning to appear, especially after 1989,that the Marxist emancipatory project - whether or not it is con-tinued in a Marxist idiom15 - will need to identify economic struc-tures at lower levels of abstraction than those historicalmaterialism acknowledges. It may therefore be necessary to re-think the notions of capitalism and socialism that have for so longshaped Marxist political practice. I am not suggesting that the tra-ditional focus on economic structures, on sets of production rela-tions, be abandoned in favor of a theoretical and political programthat treats "superstructural" phenomena on a par with the "eco-nomic base." The suggestion rather is that, for the Marxist eman-cipatory project, the capitalism-socialism distinction may be tooabstract, that intermediate economic structures may be more politi-cally pertinent.

In other words, the historical materialist "map," even if it doesrepresent a materially possible course for human history, may betoo abstract to illuminate emancipatory possibilities. If this sugges-tion is sound, it is not enough to know, in general, where we arecoming from and where we are going. We also need a clear ac-count of the shape of the land in order to know how to get fromhere to there. Socialism and ultimately communism may be hu-manity's (possible) future, but to reach that destination we mayneed to think more carefully about the feasibility and desirabilityof economic structures at lower levels of abstraction or, in more tra-ditional terms, about "mixed" economic systems.

Historical materialism will be altered, but not overthrown, if itturns out that there are "intermediate" economic structures, nei-ther capitalist nor socialist, that are in some sense "transitional" tothe economic structures Marxists have traditionally acknowledgedand that have a role to play in advancing class and general humanemancipation. On the other hand, if it becomes clear, in light ofsufficient evidence, that the deprivatization of alienable produc-tive assets is irrelevant to the emancipatory interests it was sup-posed to serve, or even that it is detrimental, then the link betweenMarxism as scientific socialism and Marxism as class analysis willhave been so profoundly severed that "Marxism" will indeed havecome to an end.

15 I discuss the possibility that Marxism might be continued without allegiance tothe traditional trappings of Marxist discourse in "What Is a Marxist Today?"Canadian Journal of Philosophy, supp. vol. 15, Analyzing Marxism (1989), pp.29-58.

2 l 6


MUST WE BRING SOCIAL DEMOCRACY BACK IN?

Neither recent events nor recent work in Marxist theory necessi-tate this conclusion. But in the present theoretical and politicalconjuncture it is prudent nevertheless for procommunists to lookbeyond Marxist sources. This conviction has motivated the atten-tion paid to egalitarian liberalism in preceding chapters. If the cen-tral argument of those chapters is sound, liberal social philosophyis indispensable for forging a defensible normative theory of com-munism. At a more practical political level, the ecumenical attitudeI am suggesting should also motivate taking non-Marxist socialismand the anarchist tradition seriously. It is worth recalling that inmany European countries, and even in the United States, thereonce existed strains of socialist theory and practice distinct fromMarxism that aimed at overcoming capitalism, not at humanizingit or managing it more efficiently. In recent decades, non-Marxistrevolutionary socialism has all but disappeared. However, at atime when Marxism's core theoretical intuitions are in doubt, it iswell for those who would continue the emancipatory project Marxundertook to look more carefully at these alternative traditions.

Perhaps the most pressing set of questions underscored by theevents of 1989 and 1991 have to do with the role of markets in theorganization of national and world economies. After Communism,it seems to many commentators in both the East and the West thatthe once raging contest between markets and plans has been wondecisively by the former. Needless to say, it has been known inthe capitalist West for decades that pure market economies arenot sustainable. Existing capitalisms, even at the Thatcherite-Reaganite extremes, are state-regulated market exchange econo-mies, not market economies in the strict sense. Nevertheless, thereis now a virtual consensus - probably temporary, possibly not -that market incentives are indispensable for promoting the effi-cient utilization of productive capacities (including labor power).In addition, some measure of market competition or its functionalequivalent now seems useful too for rendering bureaucrats andother state functionaries accountable. For many years, a number ofsocialist opponents of Communism, many of them declared Marx-ists, have insisted on the viability of market socialism16 - that is, on

16 See, among others, Oskar Lange and Fred M. Taylor, On the Economic Theory ofSocialism, ed. Benjamin Lippincott (Minneapolis: University of Minnesota

217

The general will

the possibility and desirability of market or marketlike arrange-ments without private ownership of (external) productive assets.17

But no defender of market socialism committed to a view of social-ism even remotely connected with Marx's can coherently proposethat there be (substantial) capital markets in a socialist economy.Private control over investments and claims on the revenues thatinvestments generate define private ownership. In the historicalmaterialist scheme, therefore, the private control of investment ef-fectively marks an economy as capitalist.18

If capital markets turn out to be indispensable for efficiency andtherefore for creating the massive economic surpluses necessaryfor communism, then pace Marx and Marxism, it may be necessaryfor procommunists to propose roads to communism that bypasssocialism (ostensibly, communism's lowest phase) and even towonder if capitalism itself, suitably reformed, may not be a via-ble - perhaps the only viable - path to this end.19 These are ques-tions that, to date, have barely been broached - in large partthanks to the influence of Communism and the perceived need ofnon-Communist socialists to contrast their positions with Commu-nist views.20 But if these issues are to be raised and explored, asnow seems urgent, strains of socialist theory and practice that have

Press, 1938); David Schweickart, Capitalism or Worker Control? An Ethical and Eco-nomic Appraisal (New York: Praeger, 1980); and John Roemer, "The Morality andEfficiency of Market Socialism," Ethics, vol. 102, no. 3 (1992), pp. 448-464.

17 A major shortcoming of the market socialist literature - and, more generally,of all attempts to conceive alternatives to capitalism distinct from the Sovietmodel - is a deep unclarity about "public ownership," the system of propertyrights that would supersede private ownership. For the beginnings of a theoryof public ownership, see John Roemer, Free to Lose (Cambridge, Mass.: HarvardUniversity Press, 1988), Chapter 10.

18 More precisely, individuals in societies that Marxists would regard as socialistprobably can capitalize some of their assets and even introduce markets inthese capitalized assets without destabilizing the basic economic structure.What socialism proscribes therefore is not capital markets per se, but capitalmarkets that control the general contours of social investment.

19 Cf. Philippe Van Parijs and Robert Van der Veen, "The Capitalist Road to Com-munism," Theory and Society, vol. 15 (1986), pp. 635-655; and Philippe VanParijs, "Basic Income Capitalism," Ethics, vol. 102, no. 3 (1992), pp. 465-484.

20 In so doing, Social Democrats have staked out a place for themselves to theright of the Communists, tending therefore to relinquish even a tenuous con-nection with socialism and communism. Those who have opposed the Com-munists from the left, particularly in the 1960s and 1970s, tended to wrapthemselves even more resolutely than the Communists in the mantle of Marx-ist and Leninist orthodoxy, much to the detriment of the kind of rethinkingthat is now so plainly needed.

2l8


evinced more tolerance than Communism has for market arrange-ments - including tendencies that are identified with the SocialDemocratic tradition - are likely to be valuable resources.

There is also a Social Democratic political legacy that procom-munists used to despise but that now appears more worth takingseriously. For decades, Communists played a generally conserva-tive role in the maintenance of world order, their support forsome Third World liberation movements notwithstanding. But theCommunists, in view of their historical link with Bolshevism, wereat least officially committed to the revolutionary overthrow ofcapitalist societies. Social Democrats, in contrast, favored moregradual structural transformations and continuity with liberaldemocratic political forms. If, as I have argued, liberal constraintsare crucial for societies in transition to communism, political con-tinuity with liberal democracy, wherever possible, may be thesoundest way to achieve this desideratum. In the face of currenteconomic and political realities, gradual structural transformationsmay be the only way fundamentally to alter the system in place.

I do not mean to suggest that, in the long-standing dispute be-tween Social Democrats and Communists, the Social Democra-tic position has been vindicated. My point is rather that, with theend of Communism and the devolution of social democracy intoa centrist, procapitalist political formation, the old debate hasbeen superseded by events in a way that finally makes it possiblefor communists to look to both sides - and elsewhere too - forwhatever guidance past theory and practice can afford. In a verydifferent time and place, Lenin insisted on drawing "lines of de-marcation" in theory and practice, even on abstruse matters ofphilosophical doctrine.21 Now, in contrast, a broad ecumenicism,drawing even on theoretical currents traditionally opposed toMarxism and therefore ostensibly to communism, seems much thewiser stance.

What remains of the Left after "the crisis of Marxism" and the col-lapse of Communism is a motley of good causes, free of any co-herent social vision. Good will and energetic social activism abide.But in all too many cases what has replaced the old Marxistproject, grounded in scientific socialism, is a moralizing politicsbereft of any guiding principle. The idea of communism, of a re-

21 See, for example, Materialism and Empirio-Criticism (1908), in Lenin, CollectedWorks, vol. 14.

219

The general will

public of ends directed by the general will of each individual, wasa lodestar pointing in a different direction. Perhaps it can servethis function again; perhaps communism genuinely can be the endof history. In any case, I hope to have shown that the idea is notnearly so Utopian or indefensible as many, including many on theleft, nowadays assume.

But, in the end, theoretical speculation by itself can never estab-lish the possibility or desirability of moving from "the realm of ne-cessity" into "the realm of freedom." Therefore, insofar as I havebeen able to make a case for communism, my conclusions are ten-tative and provisional. What is beyond serious dispute is just thatthe traditional resources in the communist arsenal are not ade-quate to the task of again making communism a viable politicalideal. For anyone who would take communism seriously - indeed,for anyone concerned with the emancipatory objectives that havespurred Marxist theory since its inception - now is a time inwhich many things that once seemed secure must be rethought.

It is fair to suppose that the end of Communism, harmful as ithas been for this endeavor in the short run, can only help even-tually in this increasingly urgent project. It is far from assured thata genuine communism will result from Communism's demise.But, with Communism daily discrediting communism, it couldhardly be expected to emerge without it. In any case, after Com-munism, communism is not defunct. Indeed, the communistproject remains much as it has always been. The principle differ-ence after the events of recent years, surely a change for the better,is that communists must now face the future with fewer illusions,and with a sharper determination to rethink the convictions thatsustain them.

220

Index of names

Adorno, Theodor, 14811Althusser, Louis, xii, 3311, 115-122, 14811Anderson, Perry, i48nAristotle, 128, 131-132, 136, 139, 142,

193, i96nArneson, Richard, i87n, i9onArrow, Kenneth, 78n, 84-85Augustine, St., 48nAxelrod, Robert, 64n

Balibar, Etienne, i48nBayle, Pierre, 9Bentham, Jeremy, i37nBerlin, Isaiah, i33n, 16mBettelheim, Charles, 2i4nBossanquet, Bernard, 33Bossuet, Jacques Benigne, 9Bradley, Francis Herbert, 33Brandt, Richard, 22nBurnheim, James, 2i4nBurnheim, John, 82n, 99n

Cohen, G. A., i04n, i48n, i49n, i7on,i83n, i87n, i88n, 215

Cohen, Joshua, 57nColletti, Lucio, i69nCromwell, Oliver, 180

Delia Volpe, Galvanno,Diderot, Dennis, 9Dworkin, Ronald, i5on, i84n,

Elster, Jon, 64n, i28nEngels, Friedrich 5,

Fannon, Frantz, 72nFenelon, Francois de Salignac de la

Mothe, 9Feuerbach, Ludwig, 129Fralin, Richard, 9o~9in

Gauthier, David, 37n, 5on,Gilbert, Alan, ii9nGramsci, Antonio, 115-120,Grice, Russell,Gutman, Amy,

Hacking, Ian, ionHampton, Jean, 5onHardin, Russell, i54nHart, Herbert L. A., 17mHartman, Heidi, i3nHegel, Georg Wilhelm Friedrich, 105,

110, 129, 135Hillel, Rabbi, 60-63, 74Hindess, Barry, i3nHirst, Paul Q. i3nHobbes, Thomas, 7, 9, 11, 19, 27,

36-59, passim, 66-68, 69n,75, 103, 105, 163, 17m, 174—175. 179

Horkheimer, Max, i48nHume, David, ii5n, i66n, 17m

Kant, Immanuel, 5, 2m, 32, 55n, 68-72, 160-167, 192-193

Kavka, Gregory S., 5onKorsch, Karl,

Lange, Oskar,Larmore, Charles E.,Lenin, Vladimir Illyich, 93, 101-122

passim, 137, 180, 208, 219Locke, John, 7, 9, 154, 190Lukacs, Georg, i48nLukes, Steven, 7m, i48n,

MacCallum, Gerald,Macchiavelli, Nicolo,Macpherson, Charles B.,Malebranche, Nicolas de, 9

221

Index of names

Mao Ze-dong, 120-122Marcuse, Herbert, 72-73, 14811, 204Merleau-Ponty, Maurice, 14811Mill, John Stuart 7, 9, 11, 112, 123-146

passim, 152Miller, Richard, i48nMontesquieu, Charles Louis de

Secondat, 9, 98n

Nozick, Robert, 15 m

Okin Susan Moller, i3onOkun, Arthur M., 15mOilman, Bertell, 33

Parfit, Derek, 57nPascal, Blaise, 9, 66-68Pashukanis, E. B., 173-175, 178Peffer, Rodney G.,Perelman, Chaim,Plato, 113Popper, Karl R., 16m

Rawls, John, 38n, 125-126,132, 15m, 153, i62n, i74n,185-190, 195

Resnick, Stephen, i3nRiley, Patrick, 9nRobespierre, Maxmillien, 181Roemer, John, 2i8nRoyce, Josiah, 33

Saint-Just, Louis Antoine, 181Saint-Simon, Claude Henri, 5Sandel, Michael J., i24nSartre, Jean-Paul, 72-73, i48nScanlon, Thomas M.,Schofield, Norman }., j&Schumpeter, Joseph, 77nSchwartz, Thomas, 78nSchweikart, David, 2i8nSen, Amartya K., 22nShklar, Judith N., 9nSmith, Adam, i57nStalin, Joseph, 172

Talmon, Jacob, 16mTaylor, Fred M.,Taylor, Michael, 64nTeitelman, Michael,Thompson, Dennis F, ii2n,Trotsky, Leon, 214Tucker, Robert,

Van der Veen, Robert, 2i8nVan Parijs, Philippe, 2i8n

Walzer, Michael, 92n, i24n, i8onWeber, Max, 71-72Wilde, Oscar, 92Williams, Bernard, i53nWolff, Richard, i3nWolff, Robert Paul, xi-xii, i26n

222

andrew levine - the general will, rousseau, marx, communism (1993)

Documents