sign in sign up
<<
Home , Amy Gutmann, Communitarianism, Michael Sandel, Michael Walzer, Original position

The Communitarian Critique of Author(s): Reviewed work(s): Source: Political Theory, Vol. 18, No. 1 (Feb., 1990), pp. 6-23 Published by: Publications, Inc. Stable URL: http://www.jstor.org/stable/191477 . Accessed: 24/08/2012 12:14

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

.

Sage Publications, Inc. is collaborating with JSTOR to digitize, preserve and extend access to Political Theory.

http://www.jstor.org THE COMMUNITARIAN CRITIQUE OF LIBERALISM

MICHAEL WALZER Institutefor A dvanced Study

1.

Intellectualfashions are notoriously short-lived, very much like fashions in popularmusic, art, or dress.But thereare certainfashions that seem regularlyto reappear. Like pleated trousers or short skirts, they are inconstant featuresof a largerand more steadily prevailing phenomenon - in this case, a certainway of dressing. They have brief but recurrent lives; we knowtheir transienceand excepttheir return. Needless to say,there is no afterlifein whichtrousers be permanentlypleated or skirtsforever short. Recur- renceis all. Althoughit operatesat a muchhigher level (an infinitelyhigher level?) of culturalsignificance, the communitarian critique of liberalismis likethe pleatingof trousers:transient but certainto return.It is a consistently intermittentfeature of liberalpolitics and social organization.No liberal successwill make it permanently unattractive. At thesame time, no commu- nitariancritique, however penetrating, will everbe anythingmore than an inconstantfeature of liberalism.Someday, perhaps, there will be a larger transformation,like the shiftfrom aristocratic knee-breeches to plebian pants,rendering liberalism and its critics alike irrelevant. But I see nopresent signsof anythinglike that, nor am I surethat we shouldlook forward to it. Fornow, there is muchto be saidfor a recurrentcritique, whose protagonists hope onlyfor small victories,partial incorporations, and whenthey are rebuffedor dismissedor coopted,fade away for a timeonly to return. Communitarianismis usefully contrasted with social ,which has succeededin establishinga permanent presence alongside of andsome-

A UTHOR'SNOTE: Thisessay wasfirstgiven as theJohnDewey lecture at HarvardLawSchool intSeptember 1989.

POLITICAL THEORY,Vol. 18 No. 1, February1990 6-23 ?c 1990 Sage Publications,Inc. 6 Walzer/ COMMUNITARIAN CRITIQUE 7 timesconjoined with liberal . Social democracyhas itsown intermit- tentlyfashionable critics, largely anarchist and libertarian in character.Since it sponsorscertain sorts of communalidentification, it is less subjectto communitariancriticism than liberalism is. But it can neverescape such criticismentirely, for liberals and social democrats alike share a commitment to economicgrowth and cope (althoughin differentways) with the deraci- natedsocial formsthat growth produces. itself is largelyan ideologicalpresence in modern ; it has no recurrentcritics of its own. It is intermittentlyfashionable only because it no longerexists in anything likefull strength, and itis criticizedonly when it is fashionable. The communitariancritique is nonethelessa powerful one; itwould not recurif it were not capable of engaging our minds and feelings. In thisessay, I wantto investigate the power of its current American versions and then offer a versionof myown -less powerful,perhaps, than the ones withwhich I shall begin,but more available for incorporation within liberal (or social democratic)politics. I do not mean (I hardlyhave the capacity)to lay communitarianismto rest, although I wouldwillingly wait for its reappear- ance in a formmore coherent and incisivethan that in whichit currently appears.The problem with communitarian criticism today - I amnot the first to noticethis - is thatit suggeststwo different, and deeplycontradictory, argumentsagainst liberalism. One of thesearguments is aimedprimarily at liberalpractice, the other primarily at liberaltheory, but they cannot both be right.It is possiblethat each one is partlyright -indeed, I shallinsist on just this partialvalidity -but each of the argumentsis rightin a way that undercutsthe value of the other.

[I.

The firstargument holds that liberal political theory accurately represents liberalsocial practice. As ifthe Marxist account of ideological reflection were literallytrue, and exemplified here, contemporary Western (Amer- ican societyespecially) are takento be the home of radicallyisolated ,rational egotists, and existentialagents, men and womenpro- tectedand divided by their inalienable . Liberalism tells the about theasocial society that liberals create -not, in fact,ex nihiloas theirtheory suggests,but in a struggleagainst traditions and and authorities thatare forgotten as soon as theyare escaped, so thatliberal practices seem tohave no history. The struggle itself is rituallycelebrated but rarely reflected on. The membersof liberalsociety share no politicalor religious traditions; 8 POLITICAL THEORY /FEBRUARY 1990 theycan tellonly one story about themselves and that is thestory of ex nihilo creation,which begins in thestate of natureor theoriginal position. Each individualimagines himself absolutely free, unencumbered, and on his own- andenters society, accepting its obligations, only in order to minimize hisrisks. His goal is security,and security is, as Marxwrote, "the assurance of hisegoism." And as he imagineshimself, so he reallyis,

thatis, an individualseparated from the community, withdrawn into himself, wholly preoccupiedwith his privateinterest and actingin accordancewith his privateca- price.... The onlybond between men is naturalnecessity, need, and private interest.1

(I haveused masculine pronouns in orderto fit my sentences to Marx's.But itis an interestingquestion, not addressed here, whether this first communi- tariancritique speaks to the of women: Are necessity and private interesttheir only bonds with one another?) Thewritings of the young Marx represent one of the early appearances of communitariancriticism, and his argument,first made in the 1840s, is powerfullypresent today. Alastair Maclntyre's of theincoher- enceof modem and cultural life and the loss of narrative capacity makesa similarpoint in updated, state-of-the-art, theoretical language.2 But theonly theory that is necessaryto thecommunitarian critique of liberalism is liberalismitself. All thatthe critics have to do, so theysay, is totake liberal theoryseriously. The self-portraitof the constituted only by his willfulness,liberated from all connection,without common values, binding ties,customs, or traditions- sans eyes,sans teeth,sans taste,sans every- thing- needonly be evokedin order to be devalued:It is alreadythe concrete absenceof value. Whatcan thereal life of sucha personbe like?Imagine himmaximizing his utilities,and societyis turnedinto a warof all against all, thefamiliar rat race, in which, as Hobbeswrote, there is "no othergoal, norother garland, but foremost."3 Imagine him enjoying his rights, and societyis reducedto thecoexistence of isolatedselves, for liberal rights, accordingto this first critique, have more to do with "exit" than with "voice."4 Theyare concretely expressed in separation,divorce, withdrawal, solitude, ,and politicalapathy. And finally,the very fact that individual life can be describedin thesetwo philosophicallanguages, the languageof utilitiesand the language of rights, is a furthermark, says Maclntyre, of its incoherence:Men and womenin liberalsociety no longerhave access to a singlemoral culture within which they can learnhow theyought to live.' Thereis noconsensus, no public meeting-of-minds, on the nature of the life,hence the triumph of private caprice, revealed, for example, in Sartrean ,the ideological reflection of everyday capriciousness. Walzer/ COMMUNITARIAN CRITIQUE 9

We liberalsare free to choose, and we havea rightto chose,but we have no criteriato govern our choices except our own wayward understanding of ourwayward interests and desires.And so ourchoices lack thequalities of cohesionand consecutiveness. We can hardlyremember what we didyester- day;we cannotwith any assurance predict what we will do tomorrow.We cannotgive a properaccount of ourselves.We cannotsit together and tell comprehensiblestories, and we recognizeourselves in thestories we read onlywhen these are fragmented narratives, without plots, the literary equiv- alentof atonalmusic and nonrepresentational art. Liberalsociety, seen in thelight of thisfirst communitarian critique, is fragmentationinpractice; and community is the exact opposite, the home of coherence,connection, and narrative capacity. But I am lessconcerned here withthe different accounts that might be providedof this lost Eden than I am withthe repeated insistence on the of fragmentation after the loss. This is thecommon theme of all contemporarycommunitarianisms: neoconserva- tive lamentation,neo-Marxist indictment, and neoclassicalor republican hand-wringing.(The need for the prefix "neo" suggests again the intermittent or recurrentcharacter of communitariancriticism.) I should think it would be an awkwardtheme, for if the sociological argument of liberaltheory is right,if society is actuallydecomposed, without residue, into the problematic coexistenceof individuals, then we mightwell assume that liberal politics is thebest way to deal with the problems of decomposition. Ifwe haveto create an artificialand ahistorical union out of a multitudeof isolatedselves, why nottake the or the as ourconceptual starting point?Why not accept, in standard liberal fashion, the priority of procedural justiceover substantive conceptions of the good, since we canhardly expect, givenour fragmentation,to agree aboutthe good? asks whethera communityof those who put first can everbe morethan a communityof strangers.6The questionis a goodone, but its reverse form is moreimmediately relevant: If we reallyare a communityof strangers, how can we do anythingelse butput justice first?

Ill.

We aresaved from this entirely plausible line of argument by the second communitariancritique of liberalism.The secondcritique holds that liberal theoryradically misrepresents real life. The worldis notlike that nor could itbe. Menand women cut loose fromall socialties, literally unencumbered, eachone theone andonly inventor of his or her own life, with no criteria,no 10 POLITICAL THEORY /FEBRUARY 1990 commonstandards, to guide the invention -these aremythical figures. How can anygroup of peoplebe strangersto one anotherwhen each memberof thegroup is bornwith parents, and whenthese parents have friends,rel- atives,neighbors, comrades at work,coreligionists, and fellowcitizens - connections,in fact,which are not so muchchosen as passed on and inherited?Liberalism may well enhancethe significance of purely contrac- tualties, but it is obviouslyfalse to suggest,as Hobbessometimes seemed to do, thatall ourconnections are mere"market friendships," voluntarist and self-interestedincharacter, which cannot outlast the advantages they bring.7 It is inthe very nature of a humansociety that individuals bred within it will findthemselves caught up inpatterns of relationship, networks of power, and communitiesof meaning. That quality of being caught up is whatmakes them personsof a certainsort. And only then can theymake themselves persons of a (marginally)different sort by reflectingon whatthey are and by acting in moreor less distinctiveways within the patterns, networks, and commu- nitiesthat are willy-nilly theirs. Theburden of the second critique is thatthe deep structure even of liberal societyis in factcommunitarian. Liberal theory distorts this reality and, insofaras we adoptthe theory, deprives us of anyready access to ourown experienceof communal embeddedness. The rhetoricof liberalism -this is theargument of the authors of Habits of the Heart - limitsour understanding ofour own heart's habits, and gives us no wayto formulatethe convictions thathold us togetheras personsand thatbind personstogether into a community.The assumptionhere is thatwe arein factpersons and that we are in factbound together. The liberalideology of separatismcannot take personhoodand bondednessaway fromus. Whatit does takeaway is the sense of our personhoodand bondedness,and thisdeprivation is then reflectedin liberal politics. It explains our inability to form cohesive solidar- ities,stable movements and parties,that might make our deep convictions visibleand effectivein theworld. It also explainsour radical dependence (brilliantlyforeshadowed in Hobbes's Leviathan) on thecentral state. But how are we to understandthis extraordinary disjunction between communalexperience and liberal , between personal conviction and publicrhetoric, and between social bondedness and political isolation? That questionis notaddressed by communitarian critics of the second sort. If the firstcritique depends on a vulgarMarxist theory of reflection,the second critiquerequires an equallyvulgar . Liberal theory now seemsto havea powerover and against real life that has been granted to few theories in humanhistory. Plainly, it has notbeen granted to communitariantheory, whichcannot, on thefirst argument, overcome the reality of liberal separat- Walzer/ COMMUNITARIAN CRITIQUE 11 ismand cannot, on the second argument, evoke the already existing structures of social connection.In anycase, thetwo critical arguments are mutually inconsistent;they cannot both be true.Liberal separatism either represents ormisrepresents the conditions of everyday life. It might, of course, do a little ofeach - theusual muddle - butthat is nota satisfactoryconclusion from a communitarianstandpoint. For if the account of dissociation and separatism is evenpartly right, then we have to raisequestions about the depth, so to speak,of the deep structure. And if we areall tosome degree communitarians underthe skin, then the portrait of social incoherence loses itscritical force.

IV

Buteach of the two critical arguments is partly right. I will try to say what is rightabout each, and then ask ifsomething plausible can be madeof the parts.First, then, there cannot be muchdoubt that we (in theUnited States) live in a societywhere individuals are relativelydissociated and separated fromone another,or better, where they are continually separating from one another-continuallyin motion,often in solitaryand apparentlyrandom motion,as ifin imitation of what physicists call Brownianmovement. Hence we live in a profoundlyunsettled society. We can best see the formsof unsettlementifwe trackthe most important moves. So, consider(imitating theChinese style) the Four Mobilities:

1. Geographicmobility. Americans apparently change their residence more often than any peoplein history, at leastsince the barbarian migrations, excluding only nomadic tribes and familiescaught up in civilor foreignwars. Moving people and theirpossessions fromone cityor townto another is a majorindustry in theUnited States, even though manypeople manageto move themselves.In anothersense, of course,we are all self-moved,not refugees but voluntary migrants. The senseof place mustbe greatly weakenedby this extensive geographic mobility, although I find it hard to say whether it is supersededby mereinsensitivity or by a new senseof manyplaces. Either way, communitarianfeeling seems likely to declinein importance.Communities are more thanjust locations,but theyare mostoften successful when they are permanently located. 2. Social mobility.This article will not address the arguments about how best to describe social standingor how to measurechanges, whether by income,education, class membership,or rankin thestatus hierarchy. It is enoughto say thatfewer Americans standexactly where their parents stood or do whatthey did than in any society for which we havecomparable . Americans may inherit many things from their parents, butthe extent to whichthey make a differentlife, if only by makinga differentliving, meansthat the inheritance of community, that is, the passing on of beliefs and customary 12 POLITICAL THEORY /FEBRUARY 1990

ways,is uncertainat best.Whether or notchildren are therebyrobbed of narrative capacity,they seem likely to tell different stories than their parents told. 3. Maritalmobility. Rates of separation, divorce, and remarriage are higher today than they haveever been in our own society and probably higher than they have ever been in any other(except perhaps among Roman aristocrats, although I knowof no statisticsfrom thattime, only anecdotes). The firsttwo mobilities, geographic and social, also disrupt familylife, so thatsiblings, for example, often live at great distances from one another, andlater as unclesand aunts, they are far removed from nephews and nieces. But what we call "brokenhomes" are the product of marital breaks, of husbands or wives moving out- and then,commonly, moving on to new partners.Insofar as homeis thefirst communityand the first school of ethnic identity and religious conviction, this kind of breakagemust have countercommunitarian consequences. It meansthat children often do nothear continuous or identicalstories from the adults with whom they live. (Did thegreater number of children ever hear such stories? The deathof one spouseand the remarriageof the other may once have been as commonas divorceand remarriage are today.But, then, other sorts of mobilityhave to be considered:Both men and women aremore likely today to marryacross class, ethnic, and religious lines; remarriage will thereforeoften produce extraordinarily complex and socially diverse -which probablyare without historical precedent.) 4. Politicalmobility. Loyalty to leaders,movements, parties, clubs, and urban machines seemsto declinerapidly as place and social standingand familymembership become lesscentral in the shaping of personal identity. Liberal citizens stand outside all political organizationsand then choose the one that best serves their ideals or interests. They are, ideally,independent voters, that is, people who move around; they choose for themselves ratherthan voting as theirparents did, and theychoose freshly each timerather than repeatingthemselves. As theirnumbers increase, they make for a volatileelectorate and hencefor institutional instability, particularly at thelocal levelwhere political organi- zationonce served to reinforcecommunal ties.

The effectsof the Four Mobilities are intensified in a varietyof ways by othersocial developmentswhich we arelikely to talk about in thecommon metaphorof movement: the advance of knowledge, technological , and so on. But I am concernedhere only with the actualmovement of individuals.Liberalism is, mostsimply, the theoreticalendorsement and justificationof this movement.9 In theliberal view, then, the Four Mobilities representthe enactment of ,and thepursuit of (privateor personal) happiness.And it has to be said that,conceived in thisway, liberalism is a genuinelypopular creed. Any effortto curtailmobility in thefour areas describedhere would require a massiveand harsh application of state power. Nevertheless,this popularity has an undersideof sadness and discontent that are intermittentlyarticulated, and communitarianismis, most simply, the intermittentarticulation of these feelings. It reflectsa senseof loss,and the loss is real.People do notalways leave theirold neighborhoodsor home- townswillingly or happily.Moving may be a personaladventure in our standardcultural mythologies, but it is as oftena familytrauma in reallife. WalzerI COMMUNITARIAN CRITIQUE 13

The samething is trueof social mobility, which carries people down as well as up andrequires adjustments that are never easy to manage. Marital breaks maysometimes give riseto newand strongerunions, but they also pile up whatwe mightthink of as familyfragments: single-parent households, separatedand lonelymen and women, and abandonedchildren. And inde- pendencein politicsis oftena not-so-splendidisolation: Individuals with opinionsare cut loose fromgroups with programs. The resultis a declinein "the sense of efficacy,"with accompanying effects on commitmentand morale. All inall, we liberalsprobably know one anotherless well, and with less assurance,than people once did,although we maysee moreaspects of the otherthan they saw, and recognize in him or her a widerrange of possibilities (includingthe possibility of moving on). We are more often alone than people once were,being without neighbors we can counton, relativeswho live nearbyor with whom we areclose, or comrades at work or in the movement. This is thetruth of thefirst communitarian argument. We mustnow fix the limitsof this truth by seeking what is truein thesecond argument. In itseasiest version, the second argument - that we arereally, at bottom, creaturesof community - is certainlytrue but of uncertain significance. The ties of place, class or status,, and even politicssurvive the Four Mobilitiesto a remarkableextent. To takejust one example,from the last of the Four: It remainstrue, even todayin thismost liberal and mobileof societies,that the best predictor of how people will vote is ourknowledge of how theirparents voted."0 All thosedutifully imitative young Republicans and Democratstestify to thefailure of liberalismto makeindependence or waywardnessof mindthe distinctive mark of itsadherents. The predictive value of parentalbehavior holds even forindependent voters: They are simplythe heirsof independence.But we do not knowto what extent inheritancesof thissort are a dwindlingcommunal resource; it may be that eachgeneration passes on less thanit received. The fullliberalization of the social order,the production and reproductionof self-inventingindividuals, maytake a longtime, much longer, indeed, than liberals themselves expected. Thereis notmuch comfort here for communitarian critics, however; while theycan recognizeand value the survival of olderways of life, they cannot counton, and they must have anxieties about, the vitality of those ways. Butthere is anotherapproach to the truth of the second critical argument. Whateverthe extent of the Four Mobilities, they do notseem to moveus so farapart that we can no longertalk with one another.We oftendisagree, of course,but we disagreein mutuallycomprehensible ways. I shouldthink it fairlyobvious that the philosophical controversies that Maclntyre laments 14 POLITICAL THEORY /FEBRUARY 1990 are notin facta markof social incoherence.Where there are philosophers, therewill be controversies,just as wherethere are knights,there will be tournaments.But these are highly ritualized activities, which bear witness to theconnection, not the disconnection, of theirprotagonists. Even political conflictin liberalsocieties rarely takes formsso extremeas to its protagonistsbeyond negotiation and compromise, procedural justice, and the verypossibility of speech.The Americancivil rightsstruggle is a nice exampleof a conflictfor which our moral/politicallanguage was and is entirelyadequate. The factthat the struggle has had only partial success does notreflect linguistic inadequacy but rather political failures and defeats. MartinLuther King's speechesevoked a palpable tradition,a set of commonvalues such that public disagreement could focus only on how(or how quickly)they might best be realized.1But thisis not,so to speak,a traditionalisttradition, a Gemeinschaft tradition, a survival of the preliberal past.It is a liberaltradition modified, no doubt, by survivals of different sorts. Themodifications are most obviously Protestant and republican in character, thoughby no meansexclusively so: The yearsof mass immigrationhave broughta great variety of ethnic and religious memories to bear on American politics.What all of thembear on, however, is liberalism.The languageof individualrights-voluntary association, pluralism, , separation, privacy,free speech, the careeropen to talents,and so on-is simply inescapable.Who amongus seriouslyattempts to escape?If we reallyare situatedselves, as thesecond communitarian critique holds, then our situa- tionis largelycaptured by thatvocabulary. This is thetruth of thesecond critique.Does it makeany sense then to arguethat liberalism prevents us fromunderstanding or maintaining the ties that bind us together? Itmakes some sense, because liberalism is a strangedoctrine, which seems continuallyto undercut itself, to disdain its own traditions, and to produce in each generationrenewed hopes for a moreabsolute from history and societyalike. Much of liberalpolitical theory, from Locke to Rawls,is an effortto fixand stabilizethe doctrine in orderto end theendlessness of liberalliberation. But beyondevery current version of liberalism,there is alwaysa superliberalism, which, as RobertoUnger says of his own doctrine, "pushesthe liberal premises about state and society,about freedom from dependenceand govemanceof social relationsby thewill, to thepoint at whichthey merge into a largeambition: the building of a socialworld less aliento a selfthat can alwaysviolate the generative rules of its own mental or social constructs."'2Although Unger was once identifiedas a communi- tarian,this ambition - largeindeed! - seemsdesigned to preventnot only any stabilizationof liberaldoctrine but also any recoveryor creationof Walzer/ COMMUNITARIAN CRITIQUE 15 community.For thereis no imaginablecommunity that would not be alien tothe etemally transgressive self. If the ties that bind us togetherdo notbind us, therecan be no such thingas a community.If it is anythingat all, communitarianismis antithetical totransgression. And the transgressive self is antitheticaleven to the liberal community which is itscreator and sponsor.'3 Liberalismis a self-subvertingdoctrine; for that reason, it reallydoes requireperiodic communitarian correction. But it is nota particularlyhelpful formof correction to suggestthat liberalism is literallyincoherent or thatit can be replacedby somepreliberal or antiliberalcommunity waiting some- howjust beneath the surface or just beyond the horizon. Nothing is waiting; Americancommunitarians have to recognizethat there is no one outthere butseparated, rights-bearing, voluntarily associating, freely speaking, liberal selves.It wouldbe a good thing,though, if we couldteach those selves to knowthemselves as socialbeings, the historical products of, and in partthe embodimentsof, liberal values. For the communitarian correction of liberal- ismcannot be anythingother than a selectivereinforcement of those same values or,to appropriatethe well-known phrase of MichaelOakeshott, a pursuitof the intimations of community within them.

V

The place to begin the pursuitis withthe liberalidea of voluntary association,which is notwell-understood, it seems to me, eitheramong liberalsor amongtheir communitarian critics. In bothits theoryand its practice,liberalism expresses strong associative tendencies alongside its dissociativetendencies: Its protagonists form groups as wellas splitoff from thegroups they form; they join up andresign, marry and divorce. Neverthe- less, it is a mistake,and a charactericallyliberal mistake, to thinkthat the existingpatterns of associationare entirelyor even largelyvoluntary and contractual,that is, the product of will alone.In a liberalsociety, as inevery othersociety, people are bom intovery important sorts of groups, born with identities,male or female,for example, working class, Catholic or Jewish, black,democrat, and so on. Manyof their subsequent associations (like their subsequentcareers) merely express these underlying identities, which, again, arenot so muchchosen as enacted."4Liberalism is distinguishedless bythe freedomto formgroups on thebasis of theseidentities than the freedom to leave thegroups and sometimeseven theidentities behind. Association is alwaysat risk in a liberalsociety. The boundaries of the group are not policed; peoplecome and go, or theyjust fade into the distance without ever quite 16 POLITICAL THEORY /FEBRUARY 1990 acknowledgingthat they have left.That is whyliberalism is plaguedby free-riderproblems - bypeople who continue to enjoythe benefits of mem- bershipand identitywhile no longerparticipating in the activitiesthat producethose benefits." , by contrast, is thedream of a perfectfree-riderlessness. Atits best, the liberal society is thesocial union of social unions that John Rawlsdescribed: a pluralismof groups bonded by shared ideas of toleration anddemocracy.'6 But if all the groups are precarious, continually on the brink ofdissolution or abandonment,then the larger union must also be weakand vulnerable.Or, alternatively,its leaders and officialswill be drivento compensatefor the failures of associationelsewhere by strengtheningtheir own union,that is, thecentral state, beyond the limits that liberalism has established.These limits are best expressed in terms of individual rights and civilliberties, but they also includea prescriptionfor state neutrality. The good lifeis pursuedby individuals,sponsored by groups;the state presides over the pursuitand the sponsorshipbut does not participatein either. Presidingis singularin character; pursuing and sponsoring are plural. Hence itis a criticalquestion for liberal theory and practice whether the associative passionsand energiesof ordinarypeople are likelyover the long haul to survivethe Four Mobilities and provethemselves sufficient to therequire- mentsof pluralism. There is at leastsome evidence that they will not prove sufficient-withouta little help. But, to repeatan old question,whence comethour help? A fewof the existing social unions live in theexpectation of divineassistance. For therest, we can onlyhelp one another,and the agencythrough which help of that sort comes most expeditiously is the state. Butwhat kind of a stateis itthat fosters associative activities? What kind of a socialunion is itthat includes without incorporating a great and discordant varietyof social unions? Obviously,it is a liberalstate and social union;any otherkind is too dangerousfor communities and individuals alike. It would be an odd enter- priseto arguein thename of communitarianism foran alternativestate, for thatwould be to argueagainst our own political traditions and to repudiate whatevercommunity we alreadyhave. But thecommunitarian correction does requirea liberalstate of a certainsort, conceptually though not histori- cally unusual:a statethat is, at least over some partof the terrainof sovereignty,deliberately nonneutral. The standardliberal argument for neu- tralityis an inductionfrom social fragmentation. Since dissociatedindivid- uals will neveragree on thegood life,the state must allow themto live as theythink best, subject only to JohnStuart Mill's harmprinciple, without endorsingor sponsoring any particular understanding ofwhat "best" means. Walzer/ COMMUNITARIAN CRITIQUE 17

Butthere is a problemhere: The more dissociated individuals are, the stronger thestate is likelyto be, sinceit will be theonly or the most important social union.And then membership in thestate, the only good that is sharedby all individuals,may well come to seemthe good that is "best." This is onlyto repeatthe first communitarian critique, and it invitesa responselike the second critique: that the state is notin fact the only or even, forordinary people in their everyday lives, the most important social union. All sortsof othergroups continue to existand to giveshape and purpose to the lives of theirmembers, despite the triumph of individualrights, the FourMobilities in which that triumph is manifest,and the free-riding that it makespossible. But these groups are continually at risk.And so thestate, if itis toremain a liberalstate, must endorse and sponsor some of them, namely, thosethat seem most likely to provide shapes and purposes congenial to the sharedvalues of a liberalsociety.'7 No doubt,there are problems here too, and I do notmean to denytheir difficulty. But I see no way to avoid some suchformulation - and not only for theoretical reasons. The actualhistory of thebest liberal states, as ofthe best social democratic states (and thesetend increasinglyto be thesame states), suggest that they behave in exactlythis way,although often very inadequately. Let me give threerelatively familiar examples of statebehavior of this kind.First, the Wagner Act of the 1930s: This was nota standardliberal , hinderingthe hindrances to union organization, for it actively fostered union organization,and it did so preciselyby solvingthe free-rider problem. By requiringcollective bargaining whenever there was majoritysupport (but not necessarilyunanimous support) for the union, and thenby allowingunion shops,the Wagner Act sponsoredthe creation of strongunions capable, at leastto somedegree, of determiningthe shape of industrialrelations."8 Of course,there could notbe strongunions without working class ; unionizationis parasiticon underlyingcommunities of feelingand . But thoseunderlying communities were already being eroded by theFour Mobilitieswhen the Wagner Act was passed,and so theAct served to counter thedissociative tendencies of liberal society. It was neverthelessa liberal law, forthe unions that it helped create enhanced the lives of individualworkers andwere subject to dissolution and abandonment in accordance with liberal principlesshould they ever cease to do that. The secondexample is theuse oftax exemptions and matching grants of tax moneyto enabledifferent religious groups to runextensive systems of day-carecenters, nursing homes, hospitals, and so on- societies insidethe . I do not pretendthat these private and pluralist societiescompensate for the shoddiness of theAmerican welfare state. But 18 POLITICAL THEORY /FEBRUARY 1990 theydo improvethe delivery of servicesby makingit a moreimmediate functionof communalsolidarity. The state'srole here, beside establishing minimalstandards, is to abate,since in thiscase itcannot entirely solve the free-riderproblem. If some number of men and women end up in a Catholic nursinghome, even though they never contributed to a Catholiccharity, they will at leasthave paid their taxes. But why not nationalize the entire welfare systemand endfree-ridership? The liberalresponse is thatthe social union of social unionsmust always operate at twolevels: A welfaresystem run entirelyby private, nonprofit associations would be dangerouslyinadequate andinequitable in its coverage; and a totallynationalized system would deny expressionto local andparticularist .'9 The thirdexample is thepassage of plant-closing designed to afford someprotection to local communities of work and residence. Inhabitants are insulated,although only for a time,against market pressure to moveout of theirold neighborhoodsand search for work elsewhere. Although the market "needs"a highlymobile work force, the state takes other needs into account, not only in a welfaristway (throughunemployment insurance and job retrainingprograms) but also in a communitarianway. But thestate is not similarlycommitted to thepreservation of every neighborhood community. It is entirelyneutral toward communities of ethnicity and residence, offering no protectionagainst strangers who wantto move in. Here,geographic mobilityremains a positivevalue, one ofthe rights of citizens. Unions,religious organizations, and neighborhoods each draw on feelings andbeliefs that, in principleif not always in history,predate the emergence ofthe liberal state. How strongthese feelings and beliefs are, and what their survivalvalue is, I cannotsay. Have theunions established such a gripon the imaginationsof theirmembers as to makefor good stories? There are some good stories,first told, then retold, and sometimes even re-enacted. But the narrativeline does notseem sufficiently compelling to youngerworkers to sustainanything like the old workingclass solidarity.Nor is itsufficient for a religiousorganization to provide life cycle services for its members if they areno longer interested in its religious services. Nor are neighborhoods proof forlong againstmarket pressure. Still, communal feeling and beliefseem considerablymore stable than we once thoughtthey would be, and the proliferationof secondaryassociations in liberalsociety is remarkable- evenif many of them have short lives and transient memberships. One has a senseof peopleworking together and tryingto cope, and not,as thefirst communitariancritique suggests, just getting by on their own, by themselves, one byone. Walzer/ COMMUNITARIAN CRITIQUE 19

Vl.

A good liberal(or socialdemocratic) state enhances the possibilities for cooperativecoping. provided a usefulaccount of such a statein ThePublic and Its Problems.Published in 1927,the book is a commentary on anda partialendorsement of an earlierround of communitarian criticism. Deweyshared with the critics of his time, who called themselves "pluralists," an uneasinesswith the sovereign state, but he was notquite as uneasyas most of themwere. He also sharedan admirationfor what he called "primary groupings"within the state, but he was moreinclined than the pluralists were to qualifyhis admiration. Primary groupings, he wrote,are "good, bad, and indifferent,"and theycannot by theirmere fix the limits of state activity.The stateis not"only an umpireto avertand remedy trespasses of one groupupon another." It has a largerfunction: "It rendersthe desirable associationsolider and more coherent.... It places a discountupon injurious groupingsand renderstheir tenure of life precarious. . . [and] it gives the individualmembers of valued associationsgreater liberty and security;it relievesthem of hampering conditions.... It enables individual members to countwith reasonable upon what others will do."20 These may seem like taskstoo extensivefor a liberalstate, but they are constrainedby the constitutionalestablishment ofindividual rights -which arethemselves (on thepragmatic understanding) not so muchrecognitions of what individuals bynature are or have as expressionsof hope about what they will be anddo. Unlessindividuals act togetherin certainways, state action of thesort that Deweyrecommended cannot get started. When we recognizethe "right of thecitizens peacefully to assemble,"for example, we arehoping for assem- bliesof citizens.If we thendiscriminate among such assemblies, we do so on limitedgrounds, fostering only those that really do expresscommunities offeeling and belief and do notviolate liberal principles of association. It is oftenargued these days that the nonneutral state, whose activities I havemade some attempt to justify, is bestunderstood in republican terms. A revivalof neoclassicalrepublicanism provides much of thesubstance of contemporarycommunitarian politics. The revival,I haveto say,is largely academic;unlike other versions of communitarianism in Dewey's time and ours,it has no externalreference. There really are unions,churches, and neighborhoodsin Americansociety, but there are virtually no examplesof republicanassociation and no movementor partyaimed at promotingsuch association.Dewey would probably not recognize his "public,"nor Rawls his"social union," as versionsof , ifonly because in both these 20 POLITICAL THEORY /FEBRUARY 1990 cases, energyand commitmenthave been drainedfrom the singularand narrowlypolitical association to the morevarious associations of .Republicanism by contrastis an integratedand unitarydoctrine in whichenergy and commitment are focused primarily on thepolitical realm. It is a doctrineadapted (in bothits classical and neoclassicalforms) to the needsof small,homogeneous communities, where civil society is radically undifferentiated.Perhaps the doctrinecan be extendedto accountfor a "republicof ," a decentralizedand participatory revision of .A considerablestrengthening of local governmentswould then be requiredin the hope of encouraging the development and display of civic virtuein a pluralistvariety of social settings.This indeed is a pursuitof the intimationsof community within liberalism, for it has moreto do withJohn StuartMill thanwith Rousseau. Now we areto imaginethe nonneutral state empoweringcities, towns, and boroughs;fostering neighborhood commit- teesand review boards; and always on the look-out for bands of citizens ready to takeresponsibility for local affairs.2' None of thisis any guaranteeagainst the erosionof the underlying communitiesor thedeath of local loyalties.It is a matterof principlethat communitiesmust always be atrisk. And the great paradox of a liberalsociety is thatone cannot set oneself against this principle without also setting oneself againstthe traditional practices and sharedunderstandings of thesociety. Here,respect for tradition requires the precariousness of .If thefirst communitarian critique were true in its entirety,if therewere no communitiesand no traditions,then we couldjust proceedto inventnew ones. Insofaras the secondcritique is even partlytrue, and thework of communalinvention is wellbegun and continually in progress, we mustrest contentwith the kinds of corrections and enhancements - they would be, in fact,more radical than these terms suggest -that Deweydescribed.

VI.

I haveavoided until now what is oftentaken to be thecentral issue between liberalsand theircommunitarian critics -the constitutionof theself.22 Lib- eralism,it is commonlysaid, is foundedon theidea of a presocialself, a solitaryand sometimesheroic individual confronting society, who is fully formedbefore the confrontation begins. Communitarian critics then argue, first,that instability and dissociationare the actual and disheartening achievementof individualsof this sort and, second, that there really cannot be individualsof thissort. The criticsare commonly said in turnto believe Walzer/ COMMUNITARTAN CRITIQUE 21 in a radicallysocialized self that can never"confront" society because it is, fromthe beginning, entangled in society,itself the embodiment of social values.The disagreementseems sharp enough, but in fact,in practice,it is not sharpat all -for neitherof theseviews can be sustainedfor long by anyonewho goes beyondstaking out a positionand triesto elaboratean argument.23Nor does liberalor communitariantheory require views of this sort.Contemporary liberals are not committed to a presocialself, but only to a selfcapable of reflectingcritically on thevalues thathave governedits ;and communitariancritics, who are doingexactly that, can hardlygo on to claimthat socialization is everything.The philosophical and psychologicalissues here go verydeep, but so faras politicsis concerned, thereis littleto be won on thisbattlefield; concessions from the other side cometoo easily to countas victories. The centralissue for political theory is notthe constitution ofthe self but theconnection of constituted selves, the pattern of social relations. Liberal- ism is best understoodas a theoryof relationship,which has voluntary associationat itscenter and which understands voluntariness as theright of ruptureor withdrawal.What makes a marriagevoluntary is thepermanent possibilityof divorce. What makes any identity or affiliation voluntary is the easy availabilityof alternativeidentities and affiliations.But the easier this easinessis, the less stable all ourrelationships are likely to become. The Four Mobilitiestake hold and society seems to be inperpetual motion, so thatthe actualsubject of liberalpractice, it mightbe said, is nota presocialbut a postsocialself, free at last fromall but the mosttemporary and limited alliances.Now, the liberal self reflects the fragmentation of liberal society: It is radicallyunderdetermined and divided,forced to inventitself anew for everypublic occasion. Some liberals celebrate this freedom and self- inven- tion;all communitarianslament its arrival, even while insisting that it is not a possiblehuman condition. I have arguedthat insofar as liberalismtends toward instability and dissociation,it requires periodic communitarian correction. Rawls's "social unionof social unions"reflects and buildson an earliercorrection of this kind,the work of Americanwriters like Dewey,Randolph Bourne, and HoraceKallen. Rawls has givenus a generalizedversion of Kallen'sargu- mentthat America, after the great immigration, was and shouldremain a "nationof nationalities."24 In fact, however, the erosion of nationality seems to be a featureof liberalsocial life,despite intermittent ethnic revivals like thatof the late 1960sand .We can generalizefrom this to themore or less steadyattenuation of all theunderlying bonds that make social unions possible.There is no strongor permanent remedy for communal attenuation 22 POLITICAL THEORY /FEBRUARY 1990

shortof an antiliberalcurtailment of theFour Mobilities and therights of ruptureand divorceon whichthey rest. Communitarians sometimes dream of sucha curtailment,but they rarely advocate it. The onlycommunity that mostof themactually know, after all, is just thisliberal union of unions, alwaysprecarious and always at risk.They cannot triumph over this liberal- ism;they can only,sometimes, reinforce its internal associative capacities. The reinforcementis only temporary, because the capacity for dissociation is also stronglyinternalized and highly valued. That is whycommunitarian- ism criticismis doomed- it probablyis not a terriblefate - to eternal recurrence.

NOTES

1. ,"On the JewishQuestion," in Early Writings,ed. by T. B. Bottomore (London:C. A. Watts,1963), p. 26. 2. AlasdairMacintyre, After (Notre Dame: Universityof Notre Dame Press,1981). 3. ThomasHobbes, The Elements of Law, Part1, ch. 9, para.21. 1 havenoticed that the twofavorite writers of communitarian critics of this first kind are Hobbes and Sartre. Is itpossible thatthe essence of liberalismis bestrevealed by these two, who were not, in theusual sense of theterm, liberals at all? 4. See AlbertHirschman's Exit, Voice, and Loyalty (Cambridge, MA: HarvardUniversity Press,1970). 5. Maclntyre,, chs. 2, 17. 6. Thisis RichardRorty's summary of Sandel's argument: "The Priority of Democracy to Philosophy,"inThe VirginiaStatue forReligiousFreedom, ed. by Merrill D. Petersonand Robert C. Vaughan(Cambridge: Cambridge University Press, 1988), p. 273; see Sandel,Liberalism and theLimits of Justice (Cambridge: Cambridge University Press, 1982). 7. ThomasHobbes, De Cive,ed. byHoward Warrender (Oxford: , 1983),Part I, ch. l. 8. RobertBellah et al., Habitsof theHeart (Berkeley:University of CaliforniaPress, 1985),pp. 21, 290; see Rorty'scomment, "Priority," p. 275, n. 12. 9. And also its practicalworking out, in the careeropen to talents,the rightof free movement,legal divorce, and so on. 10. See A. Campbellet al., TheAmerican Voter (:Wiley, 1960), pp. 147-148. 11. See theevocation of Kingin Habits of the Heart, pp. 249, 252. 12. RobertoMangabeira Unger, The Critical Legal StudiesMovement (Cambridge, MA: HarvardUniversity Press, 1986), p. 41. 13. Cf. Buff-Coat(Robert Everard) in thePutney debates: "Whatsoever ... obligationsI shouldbe boundunto, if afterwards God shouldreveal himself, I would break it speedily, if it werean hundreda day."In Puritanismand Liberty,ed. byA.S.P. Woodhouse(London: J. M. Dent,1938), p. 34. Is Buff-Coatthe first superliberal or Unger a latterdayPuritan saint? 14. I do notintend a deterministargument here. We mostlymove around within inherited worldsbecause we findsuch worlds comfortable and even life-enhancing; but we also moveout Walzer/ COMMUNITARIAN CRITIQUE 23 whenwe findthem cramped-and liberalism makes the escape mucheasier than it was in preliberalsocieties. 15. 1describe how free-ridership works in ethnic groups in "Pluralism: A PoliticalPerspec- tive,"in theHarvard Encyclopedia of American Ethnic Groups, ed. by StephanThernstrom (Cambridge,MA: HarvardUniversity Press, 1980), pp. 781-787. 16. JohnRawls, A Theoryof Justice (Cambridge, MA: HarvardUniversity Press, 1971), pp. 527ff. 17. See theargument for a modest"perfectionism" (rather than neutrality) in JosephRaz, TheMorality of Freedom (Oxford: Clarendon Press, 1986), chs. 5 and6. 18. IrvingBernstein, Turbulent Years: A Historyof theAmerican Worker, 1933-1941 (Boston:Houghton Mifflin, 1970), ch. 7. 19. See myessay on "Socializingthe Welfare State" in Democracy and theWelfare State, ed. byAmy Gutmann (Princeton, NJ: Press, 1988), pp. 13-26. 20. Dewey,The Public and Its Problems (Athens, OH: SwallowPress, 1985), pp. 71-72. 21. This kindof pluralistrepublicanism is also likelyto advancethe prospects of whatI called"complex equality" in Spheres ofJustice (New York:Basic Books,1983). 1 cannotpursue thisquestion here, but it is worthnoting that both liberalism and communitarianismcan take egalitarianand non- or antiegalitarianforms. Similarly, the communitarian correction of liber- alismcan strengthenthe old inequalitiesof traditionalistways of lifeor it can counteractthe newinequalities of theliberal market and the bureaucratic state. The "republicof republics" is likely,though by no meanscertain, to have effects of thesecond sort. 22. The issueis starklyposed in Sandel,Liberalism and theLimits of Justice; much of the recentdiscussion is a commentaryon or argumentwith Sandel's book. 23. See Will Kymlicka,"Liberalism and Communitarianism,"in Canadian Journal of Philosophy(June, 1988), pp. 181-204. 24. Kallen,Culture and Democracyin the UnitedStates (New York:Boni & Liveright, 1924).

MichaelWalzer is Professorat the nstituteforAdvancedStudy in Princeton, NewJersey. He is theauthor of numerous books and articleson political theory.