1

DiversityandtheSocialContract
RyanMuldoon
April2009

Introduction

Motivation
Economicandpoliticalglobalizationhasencouragedwavesofimmigration
thathavechangedthesocialcompositionofcountries,leavingoncehomogenous
culturestodealwiththeirnewfounddiversity.Unfortunately,Westerncountries
havenothandledthisongoingtransitionwithmuchsuccess.AsWesternEurope
hasseenimmigrationrise,ithasseensupportforliberalpoliciesdecline,andcivil
unrestincrease.LiberalattitudesinAmsterdamaregivingwaytomorerestrictive
policies.Parishasseenlargescaleriots,firstintheethnicminoritybanlieue,then
amongthemoreprivilegedstudentswhoresistedpoliciesaimedatresolvingthe
originaltensions.IntheUnitedStates,evenwithitslongerhistoryofethnic
diversity,immigrationandaffirmativeactionpolarizepoliticaldiscourse.
Icontendthatthesepoorresponsestonewdemographicrealitiesare
understandable,andevenentirelyunsurprisingconsequencesoftheclassicliberal
accountsofsocialcontracts.ThankstoRawls,socialcontracttheoryhasbecomethe
dominantframeworkforconsideringquestionsofjustice,butasithasbeen
practicedsofar,itassumesfairlyhomogeneoussocieties.Thiswas,formuchof
Westernhistory,averyreasonableassumption.WhenRawlsATheoryofJustice
waspublished,Westernsocietieshadmanufacturingeconomies,andwererelatively


3
homogenous,atleastculturallyifnotalwaysethnically.Aswegobacktoprevious
accountsofsocialcontracttheory,ethnichomogeneityonlyincreases.
Thoughthecruxofsocialcontracttheoryhasalwaysbeenabouthowweare
toreconcileourcompetingpoliticalinterests,IsubmitthatfromLockeforward,the
implicitassumptionhasbeenthatthedepthofpoliticaldisagreementwasnotthat
great.1Likewise,accountsoftolerance,likeLockesEssayonToleration,presented
argumentsforsecondbestsolutionstoreligiousdisagreement:namely,thatwe
cannotimposeourviewsonothersthroughthepowerofthestatebecausethey
mightonedaybeabletoimposetheirviewsonus.Andsincewedonotwantto
fightreligiouswars,weoughttotoleratetheexistenceofotherviews.2Thisview
assumestwothingsthatIthinkarerelevanttoourunderstandingofcurrent
problems.Thefirstthingtonoteisthatdiversityisaproblemtobemanaged.
Therearenobenefitsseen.Thesecondthingtorecognizeisthatmanagingthis
diversityimposesaburdenonindividualsinsociety.Eachpersonhastoputupwith
thosewithwhomtheydisagree.
Today,especiallyinEurope,thedemographicstoryismuchdifferent.Not
onlyhasthenatureoftheeconomychangeddrastically,butthecompositionof
societyhasalsochanged.Thesecountries,mostwithlowbirthrates,soughtto

1
In Hobbes, disagreement was essentially absolute: without a contract, we would be in a
state of war. Locke dramatically weakened this assumption, and his account of the state
of nature included individuals with recognized property rights, and people living
successfully together. Rawls explicitly states that he is not going to argue with nonliberals.
2
Though not all views. Locke claimed that atheists could not be tolerated.


bringinworkersfromothercountriestomaintaintheirwelfarestates.Those

workershavestayed,andhavenowbecomesizableportionsofthepopulation,who
expectrightslikeothercitizens,includingthefreedomtocontinuetheirown
culturalpractices.AndthisiswhereweseetheLockeanaccountoftolerationfail.If
wearetounderstanddiversityasaburden,anddisagreementbecomeslargerthan
thedifferencebetweenCatholicsandProtestants,theneventuallypeoplesimply
stopthinkingthattheburdenisworthit.Whenthishappens,wecanexpecttosee
policiesenactedthataimtopreventnewimmigrationinflows,andpolicesthat
eitherisolateminoritygroupsorforcethemtotakeonmajorityculture.Thisis
exactlywhatweseehappeningthroughouttheWest.
Thispathisnotsustainable.Economiclogicdictatesthatmoreimmigration
willhappensolongasrichercountrieshavelowbirthrates,oreconomic
globalizationismaintained.Butcurrentpoliticalthinkinghasdictatedthatthe
resultantdiversitycreatespoliticaltensionandeventuallysocialunrest.Andso,
withthestatusquo,wereachanimpasse:eitherwealthycountriesdrastically
changetheirsocialcontractstocontendwithanaging,increasinglypoorpopulation,
ortheyfacecivilunrest.Iwanttoresistthisfalsechoice.
Theaimofthisdissertationistoprovideanalternativeaccountofsocial
contracttheorythatexplicitlyaimstoaddresstheuniqueproblemsand
opportunitiesthatarefoundindiversesocieties.Inparticular,itestablishesthree
keyideasthatdifferfromtraditionalsocialcontracttheory.Firstistheideaof
perspectiveinpoliticalreasoning.Ratherthanassumethatallcitizensviewmoral


andpoliticalproblemsinthesameway,itarticulatesawaytoproductively

aggregatedisparateviewstofindpotentialareasofagreement.Thesecondkeyidea
ismydevelopmentofabargainingmodelthatrequiresvirtuallynoagreement
amongstitsparticipants.Andfinallyitpresentsamodelofhowagentscanbe
motivatedtoseekoutmorediversesocieties,eveninthefaceofdistributive
disputes.Thesecomponentscombinetocreateatheoryofsocialcontractsthat
embracesMillsconceptofexperimentsinliving,bypresentingasystemof
constantsocialexperimentationandrefinement.Ratherthanarticulateanidealend
state,myfocusisonthetrajectoryofchange.Letusnowbrieflyconsidereachof
theseideasinturn.

Perspectives
Politicaldisagreementisoftencastastwoormorepartieshavingdifferent
preferencesaboutwhichoutcomeoutofarangeofpossiblealternativestheyare
mostinterestedin.Differentpeopleoftenhavedifferentpoliticalinterests,andso
thisisanimportantideatocapture.Butitdoesnotcaptureevendeepersourcesof
disagreement:thecasesinwhichthepartiesdonotagreeonhowtocharacterize
thesituationinthefirstplace.Beforewecanevendiscusswhatourcompeting
preferencesmightbe,wemustfirstestablishwhetherweagreewithhowto
representtheproblem.Thisissueofrepresentationcanbeunderstoodintermsof
ourperspectives.Perspectivesdotwothingsforus:theyprovideuswitha


6
categorizationoftheworld,andtheyprovideuswithamethodofnavigatingthat
categorization.Associetiesbecomemorediverse,weshouldexpectthattherange
ofperspectivesfoundinthepopulationwillonlygrow.
Onthefaceofit,thiswouldseemlikeyetanotherkindofburdenthat
diversityplacesonsociety.ButIarguethatdifferentperspectiveshaveepistemic
advantagesindifferentkindsofsituations.Thereisnosingleperspectivethatis
alwaysbest.Havingarangeofperspectivesinthepopulationcanhelpasociety
solvemoreproblemsmoreeffectively.Inparticular,wecantakeadvantageofthe
factofourdiversityofperspectivestobetterassessourmoralbeliefstohelpcreate
theinitialfoundationofagreementinasocialcontract.

Bargaining
Thenotionofperspectiveistheconceptaroundwhichtherestofthe
dissertationisorganized.Ifwearetotakeperspectivesseriously,thereareseveral
areasofsocialcontracttheorythatmustbeadjusted.Inparticular,afocuson
perspectivesshedslightonwhyapublicreasonmodelofpoliticaldeliberation
cannotbeusedinadiversesociety.Publicreasonrequirestheexistenceand
adoptionofaneutralperspective.Icontendthatsuchaperspectivedoesnotexist.
Evenworse,thepublicreasonmodellimitswhatcountsasacceptableinthepublic
sphere.Thismodelhasgoodintentions,asitaimstoallowonlysharedorneutral
arguments,butIarguethatitunfairlyconstrainsindividualswhohaveperspectives


thatarenotinthemainstream.Theirviewsmustbetranslatedsomehowinto

anotherframeworkofideasbeforetheycanevenbeexpressed.Thisalmost
assuredlyreducestheirpoliticalvoice.
Inplaceofpublicreason,Iarguethatabargainingmodelofrightsallocation
inasocialcontractismostappropriate.Bargaininghasseveraladvantages.Most
importantamongthoseisthatbargainsdonotrequireauniversallyagreedupon
perspective.Infact,Ishowthatabargaincantakeplaceevenwhereindividualsin
thebargaindonotagreeonwhatitistheyarebargainingabout.Alleachagent
needstoconcernherselfwithisherperceptiononwhethersheisgainingorlosing.
Onthismodel,allperspectivesaretreatedequally,asanagentsperspectiveis
irrelevanttohowotheragentsengagewithher.Buteachagentisabletoexpress
herinterestsbymakingdemandsandconcessionsoverthecourseofthebargain,
evenifagentswithdifferentperspectivesmightinterpretthoseinterestsdifferently.
Inthisway,wecanarriveatanallocationofrightsthatleaveseveryonewiththe
sameleveloffranchise,regardlessoftheirperspective.

NoSubsidyPrinciple
Thefinalcomponentoftheviewofsocialcontractsdevelopedinthis
dissertationillustratestheconditionsunderwhichdiversitycanbeactively
embracedinsociety.Justasperspectivediversityisepistemicallybeneficial,itis
alsoeconomicallybeneficial.Inparticular,Iarguethatperspectivediversity


encouragesincreaseddivisionoflabor,andtheresultantgainstotradecreatea

largersocialsurplus.Thisfactoughttomotivateincreaseddiversityinsociety.In
particular,agentshaveincentivestoseekoutpeoplewhoaredifferentfromthemto
getlargergainstotrade.
Thispositivestoryaboutthebenefitsofdiversitytoeconomicproductionis
constrainedbytheveryrealdisagreementoverthedistributionofthesocialsurplus
createdbytheincreasedproduction.Here,weoughttoexpectthatasdiversity
increases,disagreementwillalsoincrease.Iarguethatthewayaroundthissource
oftensionistointroducethenosubsidyprinciple,whichisadistributiverulethat
assuresthateachcoalitioninsocietybenefitsfromthesocialsurpluscreatedin
proportiontotheircontributionstoproduction.Thegoalofthisprincipleisto
assuresocialstability:whenitissatisfied,nocoalitionisabletoclaimthatthey
couldbemadebetteroffiftheylefttoformtheirownsociety.Itassuresthat
everyonesharesfromthesystemofsocialcooperationthatismadepossibleby
livinginsociety.Thoughnoonegetstodictatehowtheentiresocialsurplusis
distributed,eachisassuredthattheygetmoreofwhattheywantthantheywould
haveinalessdiversesociety.Itisthisguaranteethatassuresthatagentsare
rationallymotivatedtoembracediversity.

PuttingitTogether:JusticeasaTrajectory
TheaccountofsocialcontracttheorythatIhavedevelopedexplicitlyavoids
makingclaimsaboutwhattheendresultoftheprocedureIoutlinewouldlooklike.
Thisisbecauseultimatelymyviewisthatwedonotyetknowwhatanoptimal
socialcontractwouldbe.Notonlythat,butevenifwefoundasocialcontractthat
wasoptimalforagivensetofeconomicconditionsandsocialdemographics,that
contractmaybecomelessoptimalastheunderlyingsocialconditionschange.
Ratherthanendowagivenendstatewiththestatusofaregulativeideal,Ithinkitis
moreimportanttofocusontheproceduresbywhichasocietycanchangeits
contract.
Thethreepositivechaptersofthedissertationshouldbeseenasoutliningan
iterativeprocedure.Wefirstattempttoassessourmostrobustlysupportedmoral
beliefs.Onceweagreeonwhatweviewassimplyunquestionable,wemoveonto
whatwecannegotiateabout.Thisnegotiation,constrainedbyourpreviousstep,
determinesanallocationofrightsthatweallcanendorse.Finally,wefindthe
implicationsoftheserightsbyactuallylivingwiththemandattemptingsmallscale
experimentsinliving.Aswegathermoreinformationabouthowaccurateour
assessmentswereofoursupportofourmoralbeliefs,andhowwellwedidin
determiningaparticularrightsallocation,wecandecidetogobackandrefineour
socialcontractwithournewinformation.Eachstageoftheprocessisdesignedto


10
ensurethatnooneisknowinglymadeworseoffthantheywerebefore.Ifsome
changeisfoundtonotwork,wesimplytryagain.Everychangethatweasasociety
makeisarefinementofhowwethinkweoughttolivewitheachother.Ratherthan
stickwithasocialcontractthatwasdesignedforadifferentera,orplungeinan
entirelynewdirection,weshouldexperiment.Wecanfindwhatworks,and
abandonwhatdoesnot.

11

Chapter1:
DiversityandthePublicSphereintheLiberalTradition
Introduction

AtriumphoftheWesternLiberalpoliticaltraditionistheenshrinementof
thedivisionofthepublicandprivatespheres.Thisdistinction,aidedbythe
developmentofindividualrights,specifiesthelimitsofthestate,andtheextentof
thefreedomofindividuals.Thatis,itdeterminestheareasinwhichthestatehasa
legitimateclaimtoaninterestincontrollingorregulatingthebehaviorof
individualsorassociationsforthepurposesofpromotingthepublicinterest.These
areasareconsideredtobethepublicspheretherealmofthepolitical.Everything
elseismeanttobelargelyoutsideofgovernmentmeddling.Heretheinterestsof
privatecitizensaresupposedtofreelymanifestthemselves:peoplecanassociate
withwhomevertheyplease,saywhattheywant,andforthemostpartpursue
whateverpersonallifeprojectsthattheymighthave.Theprivatesphereisthusthe
realmofindividualinterests.
Intheprivatesphere,individualdiversityistolerated,andinanumberof
respects,consideredtobeagoodthing.Wearefreetohaveourownmusicaltastes,
ourownpreferenceforgroupassociation,andwearefreetochooseourprofession
andourhobbies.Theonlylimitsonourbehavioraregenerallyprohibitionson
thoseactionsthatharmothers.Asthecommonexamplegoes,myrighttofreely


punchstopsatsomeoneelsesface.Butthisisnotaparticularlyonerous

12

restriction.AsRawlsputsit,wearefreetodevelopourownconceptionofthegood,
andpursueitwithoutinterference.
Microeconomicscanbeseenasaframeworkforstudyingourprivatesphere
interactions.Asrationalagents,wearefreetomakecontractsandconducttrades
withotherstoeveryonesbenefit.Individualdifferencesinpreferences,skillsand
knowledgecreateopportunitiesformarketstoemergeandwealthtobecreated.
Rationalagentsdonothavetoagreeonmuchinamarket:itdoesnotmatterwhat
raceorreligiononeis,orwhatonespoliticalviewsarefortheretobeasuccessful
transaction.Theonlyrelevantfeatureinamarketinteractioniswhetherornotone
canagreetothepriceofagoodortothetermsofatrade.
Inthepublicsphere,however,thereisamuchgreaterneedforwide
agreement.Sincethepublicsphereencapsulatesthosethingsthatthestatehasan
interestinregulating,atheoryofparticipatorygovernmentmayhavedifficultywith
thekindofindividualdiversitythatisencouragedintheprivatesphere.Inthe
WesternLiberalpoliticaltradition,however,thinkerslikeLocke,Rousseauand
Rawlshavecontendedthatrationalagentswillcometoagreeonpublicsphere
issuesbymeansofreasonalone.Thereisanapparenttensionhereafterall,both
LockeandRawlspromoteindividualfreedomintheprivatesphere,andsoone
wouldexpectthattheseprivatespheredifferenceswouldhaveaneffectonthe
publicsphere.Insteadofarrivingatthisconclusion,however,theyarguethatones
privatepreferencesshouldnotmanifestthemselvesinthepublicsphere.Onesrole


aspubliccitizenisdistinctfromonesroleasprivateagent.Itisfurther

13

contendedthatrationalitywillleadcitizenstoagreeontheshapeofthepublic
sphere.So,theidealpublicsphere,whichtheyclaimisachievable,isonethatall
agreeupon.ThisidealofuniversalagreementcanbeseeninRousseausdistinction
betweentheGeneralWillandtheWillofAll.TheGeneralWillistheagreementof
individualsthinkingofthemselvesascitizensonmattersofthepublicsphere,
whereastheWillofAllissimplytheunionofthepopulationsprivatesphere
preferences.3TheWillofAllmighthaveagreatdealofinconsistentpreferences,
reflectingagreatdealofdisagreement,buttheGeneralWillisexplicitlydefinedto
bethatwhichweallagreeuponascitizens.Rawlsperformsasimilarkindof
exerciseintheOriginalPosition,whereoneishypotheticallybehindathickveilof
ignorance.Oneismeanttobeabletocharacterizeonespreferencesforthebasic
structureofsocietywithouthavinganyknowledgeofonesprivatesphereself.
Becauseofthissetup,Rawlsarguesthatthereasoningbehindtheveilofignorance
willbethesameforallindividualstherefore,regardlessofprivatesphere
diversity,rationalityleadstouniversalagreementinthepublicsphere.
Thissharpboundarybetweenpublicandprivateinterests,andthelackof
influenceofprivateinterestsonpublicinterestsseemstobedescriptivelymistaken
andtheoreticallyworrisome.Forinstance,ourprivatereligiousbeliefscanhavea

3
Rousseaus social contract, unlike Lockes or Rawls, does not protect private-sphere
interests. Rather, the private sphere is overwhelmed by the public sphere as guided by
the General Will. I raise this as an even more extreme example of the general
phenomenon of establishing the public sphere with little regard to the private sphere.
Mill is perhaps the archetype of private-sphere liberty, but I save that discussion for the
next chapter.


directinfluenceonourpublicattitudes,ascanbeseenbyAmishandMormon

14

rejectionsofsomeAmericanpoliticalauthority.Lessextremecasesincludenature
loversfightingfornationalparks,orpacifistsarguingagainstalargemilitary.Not
onlyareweinfactinfluencedbyourprivatesphereversionsofourselves,butIalso
contendthatthisinfluencecanbeanavenueforsocialimprovement,onethatthe
rigidpublicprivatesplitfailstorecognize.
Twoexamplesoftheoriesthatrequireaformofstateinterventioninthe
privatesphere,andthussupportfromthepublicsphere,areperfectionistor
communitariancomprehensivemoraldoctrines.4Suchtheoriescannotbe
circumscribedtotheprivatesphere:whiletheymayshapeindividualvalues,they
havepublicsphererequirements.Forexample,acommunitariandoctrinerequires
thatsocialinstitutionsbedesignedinsuchawayastosupportcommunity
connectedness,andasubordinationoftheindividualtothecommunity.
Communityhistoryandculturalstandardsmustbemaintained,potentiallytothe
restrictionofoptionsoftheindividualsinthecommunity.Thesearenecessaryfor
maintainingprivatebeliefsincommunitarianism,butrequirecooperationfromthe
publicsphere.
Thepossibilitythatprivatemoralbeliefscouldhavepublicconsequencesis
ignoredbytheattemptstomaintainthesharpboundarybetweenspheres.Yet,we

4
Here I take Perfectionism to be a doctrine that makes claims as to what the objective
human good is, along with a framework for determining social and political mechanisms
for optimally promoting the good. Communitarianism I take to be a rejection of
liberalisms emphasis on the individual in place of an emphasis on community structures.


cannotmeaningfullydiscussthevalueofsuchtheorieswithoutreferenceto

15

featuresofthepublicsphere.Thisdividebetweenpublicandprivatespheres
representsnotonlyarestrictionofthepublicsphere,butimplicitly,italso
representsarestrictiononwhatisallowedintheprivatesphere.Thismovegoes
againstoneoftheavowedgoalsofthepublicprivatesplit:thatbeliefsandactivities
intheprivatesphereareunrestricted,solongastheyfollowtheharmprinciple.As
doctrinessuchasperfectionismorcommunitarianismarenotobviouslyharmfulor
otherwiseillegitimate,theycaneasilybeacceptedintheprivatesphere.Butwe
knowthattheywouldhavesignificanteffectsonthepublicsphere.Thisveryfact
violatestheseparationofthepublicspherefromtheprivatespheresocentraltothe
liberalWesterntradition.
Onemaychoosetoargueagainstthelegitimacyofcertainmoraldoctrines
thathavepublicsphererequirements,andthustrytosavetheboundary.However,
eventhismoveisproblematic.Indeed,shiftingmoraldoctrinesthatcastdoubton
thepublicprivatecleavagetothepublicsphereisaverydangerousmoveit
deprivesusofawholesetofvaluesthatwemaywanttoguideourlives.Thevalue
oftheprivatesphereispreciselyitsbeinglikealaboratorywherewecantestvalues
andtheirviability.Unlesssomeofthesevaluesareharmfultoothers,whyshould
wecreateartificialboundariesthatexcludeentirerealmsofpossibilities?
Forexample,whyshouldthevalueofsocialconnectedness,socentralto
communitariandoctrine,beleftoutoftheprivatespherejustbecauseithaspolitical
consequences?Ournoninterventionthenbecomesakindofinterventionby


prohibitingcertainkindsofsocialarrangements,whileleavingothersopen.It

16

shouldbenotedthatamajorconsequenceofcreatingsharpboundariesbetween
spheresandtransferringallthoseprivatevaluesthatinvolveapubliccommitment
tothepublicsphereisthatweseverelycurtailourchancesofsucceedinginbuilding
satisfactorytheoriesofpublicinstitutions.Withouthavinganopportunityfor
privatesphereexperimentation,wecannotdeterminewhatinstitutionsmost
effectivelyservetohelppeopleachievetheirlifeprojects.
Oneverydifferent,andradicalapproachtosolvingtheproblemsthatthe
publicprivatesplitcreatesistoeliminatethenotionofthepublicsphere
altogether,andattempttoscaleupprivatesphereinteractionstodealwithlarge
scaleproblems.Oneexampleofthisstrategyistouseapurelymarketbased
solutiontosocialneedsandproblems.Thismovecollapsesthepublicintothe
private,but,asIshalldiscussfurther,thiscollapsecannotpossiblybecompletely
successful.ThereareelementsofthiskindofsolutioninLibertarianapproachesto
socialorganization.Inthiscasetoo,theproblemofprovidingpublicgoodsand
dealingwithexternalities5raisesitsuglyheadhavingsomemeansofproviding
nationaldefenseorasystemofjusticeisabasicchallengeforfullLibertarianor

5
To make my point clear, let us briefly define public good and externality. A public
good is a good that is non-rival and non-excludable. That is, once the good is provided,
no one can be kept from enjoying it, nor does anyones consumption of the good reduce
the amount left for everyone else. National defense is a classic example of a public good.
An externality is a cost or benefit related to some good that is not captured in the price of
that good.


17
anarchocapitalisttheories.Nozick,forexample,reliesonanotionofaminimal
governmentinordertosatisfytheprovisioningofthesebasicpublicgoods.6
ItisnotstrictlythefaultofLibertarianoranarchocapitalistthinkersthat
theyhavenotbeenabletodevelopmeansofpublicgoodsprovisioningwithout
somegovernmentinterventionitisabasicprobleminmicroeconomics.Mancur
OlsoninTheLogicofCollectiveActionandRussellHardininCollectiveActionoutline
thesignificantchallengestopublicgoodsprovisioninginlargegroups.Oneofthe
maindifficultiesisthatmarketsdonotnecessarilyensurethatallthecostsand
benefitsassociatedwithproductionofsomegoodareincorporatedintoitsprice
thatis,notallcostsandbenefitscanbeinternalized.Forexample,factoriestendnot
topayforthesocialcostsofpollutingriversortheairwithoutsomekindof
additionalgovernmentregulation.Thisisduetothefactthatthecostsarespread
widely,whilethebenefitsofpollutingarekeptbythefactoryintheformoflower
manufacturingcosts.Soeventhoughtheindividualbenefitsaregreatlyoutweighed
bythecollectivecosts,thepollutersshareofthecostsisoverwhelmedbythe
benefits.Thisproblemofincentivecompatibilityarisesforallproblemsof
collectiveaction.Infact,itissoseverethatevenapopulationofaltruistswillrun
intothesameproblems,asanyindividualscontributionisnotsufficienttoprovide
thegood,andsothealtruistwillnotgivewithoutsomeassurancethatotherswilldo
thesame.

6
See Nozicks Anarchy, State, and Utopia (1974) for an elaboration of his account of the
minimal state in a Libertarian framework.

18
Astrategythatsupportscollapsingthepublicsphereintotheprivate

sphereistointernalizetheexternalitiesbycreatingnewpropertyrights,along
withmarketsinwhichwecanbuyandsellthesenewrights.TheCoaseTheorem
detailshowthisincentivecompatibilityissuecouldberesolvedifwewerewillingto
deviseasystemofcompleteindividualrights,suchthatindividualsowned,say,
vouchersforbreathingfreshair.Ifthetransactioncostswerenearnegligible,then
theproblemofcollectiveactionwouldgoaway,astherewouldbeample
opportunitiesforthefactorytosimplypurchasethefreshairrightsofthose
affected.7However,wedonothavesuchfinegrainedallotmentsofindividual
rights.Sowefindourselveswithnomeanstomakeindividualincentives
compatiblewithpublicgoodsprovisioning.Becauseofthisproblem,thestandard
economicconclusionisthatindividualeffortstoprovidepublicgoodswillfail.
Individualsarealwaysmotivatedtofreerideonthecontributionsofothers,and
notcontributetopublicgoodsthemselves.
Sowefindourselvesinapositionwheretakingapurelyprivatesphere
approachtoorganizingsocietyhasfundamentaldifficulties,andallowingthepublic
spheretoignorethefeaturesoftheprivatesphereisunmanageable.Insteadof
eitherapproach,wecanaimforanovelpathbetweentheseextremes.Weretaina
publicprivatedistinction,butinsteadofallowingthepublicspheretobe
determinedinisolationfromtheprivatesphere,weshowthatitcanbeactively

7
For a more detailed discussion of the Coase Theorem and its implications for allocations
of rights, see Coase's The Problem of Social Cost, Journal of Law and Economics
(October 1960) or The Firm, The Market, and The Law (1990)


19
shapedbyit.Themostimportant,andindeedmotivating,consequenceofthisis
thatthediversitythatisfosteredattheprivatelevelisretainedatthepubliclevelas
well.Ratherthanbeingaproblemtobemitigated,thisdiversitycanbeusedto
betterrefineasystemofsocialorganizationandallowittoadapttothechanging
needsofasociety.
Thegoalofthischapteristoexaminetheviewofthepublicsphereasfound
insocialcontracttheoryandcompareittotheanalysisofdiversitythatisfoundin
economicthought.Icontendthattheapproachtakenbysocialcontracttheoristsis
abletocaptureonlysomenegativeaspects,andnotanybenefits,ofdiversity,asitis
analyzedineconomics.IfirstclarifywhatImeanwhenIusetheworddiversity.
UponestablishingthedefinitionsIrequire,Ibeginthesubstantiveanalysis.Iuse
RousseausnotionoftheGeneralWill,andRawlsnotionoftheOriginalPosition,as
majorexamplesofthestyleofreasoningcommontotheWesternLiberaltradition
withregardstothepublicsphere.Ithenexamineseveralideasfromeconomic
theorytoarguethatthereareanumberofprivatespherebenefitstodiversity.I
leavetolaterchapterstheworkoffullyaccountingforhowthesebenefitsmanifest
themselves.FinallyIconcludebyarguingthattheeconomicanalysispresents
problemsfortheapproachtakenbytheWesternLiberaltradition.

DiversityofWhat?
Atermlikediversityisinsomesenseillsuitedforthetaskahead.Byitself,
itdoesnotidentifyapreciseconcept,asitisarelationaltermwithoutanyparticular


restrictionsonitsdomain.Assuch,weareleftwithatermwhosesenseisan

20

amorphousmixtureofrelatedconcepts.Inmyviewthisfuzzymeaningledtoause
ofdiversityinpublicdiscoursegermanetoanotionofculturaloridentity
diversity.Icontendthatidentitydiversityisnotabasicconcept,andmustbe
brokendownintoitsconstituentparts,tobetterunderstandculturaldiversity,but
alsotoseehowthebasicpartsinteractwitheachother.
Todothis,wemustconsiderthewaysinwhichindividualsinasocietycanbe
diverse.Oncethesedistinctionshavebeenmade,wecanthenseethemultifaceted
waysinwhichthesekindsofdiversitycaninteractwitheachother.Itisthroughan
understandingoftheseinteractionsthatmanyoftheproblemsthatthesocial
contractfacescanbesolved.

IdentityDiversityandWhyGroupsarenottheUnitofAnalysis
MostreadilyapparenttomanyofusisIdentityDiversity.Peopleidentify
withdifferentracial,social,andgendergroups,andthesedistinctionsarefrequently
usedtodivideapopulation.However,identitydiversityisalsoashiftingnotion
overtime,someidentitiesfallaway,andnewonesappear.Race,ethnicity,and
socialgroupingsarenotnaturalkindstherearenoprincipled,objectivewaysof
definingbordersbetweenmostgroups.Onemightattempttocreatedistinctions
betweensocialgroupsbymeansofsharedhistories:individualswhosefamiliesare
fromthesamegeographicallocation,andhavesignificantratesofintermarriage


21
8
mightbegroupedtogether. Thoughatechniquelikethisisnotgoingtoproduce
sharpboundaries,letussupposethatitcould,atleast,provideroughgroupings.
Thefuzzybordersbetweengroupsmaycauseproblems,butwecanatleastassume
thatformanygroups,therearesharpboundariesbetweenalmostanytwoofthem.
So,letussupposethatforethnicgroupingswecanfindanobjectivestandardby
whichwecandifferentiatebetweenmostpairsofgroups.Evenwiththis,wearenot
yetoutofthewoods:whilethistechniquemaybesuccessfulatdifferentiatingIrish
fromItalians,itdoesnotseemtoworkforseparatingCapitalistsfromCommunists,
CatholicsfromProtestants,UnionistsfromLoyalists,orPunksfromGoths.Social
groupingsmayoftenhavesomepositivecorrelationwithgenealogy,butformany
significantgroups,itwasthespreadofideasthatcreateddistinctions.Assuch,it
seemsunlikelythatthereexistsanobjectivedistinctionthatcanbeusedtoset
boundariesbetweengroups.
Withoutsomethinglikeanobjectivemethodofdistinguishinggroups,weare
unabletospeakpreciselyaboutidentitygroupsasaclassanddescribetheir
propertiesanddynamics.Forexample,theredoesnotseemtobemuchincommon
amongthenaturesoftheRedSoxNation,Republicans,lesbians,academics,andthe
British.Eachprovidessomeculturalidentity,buttoagreaterorlesserdegree.Red

8
We already know from biology, however, that this is going to be a fruitless task. There
is significant literature arguing against the concept of a species (or at least reducing a
species to an individual). For a sample of this literature, see Mishler, B. and Donoghue,
M., 1982, "Species Concepts: A Case for Pluralism", Systematic Zoology, 31: 491-503
and Ereshefsky, M.,1998, "Species Pluralism and Anti-Realism", Philosophy of Science,
65:103-120. If this cannot be done for species, identifying particular cultural groups, a
much finer-grained distinction, has very little chance of success.


22
Soxfansareunlikelytopushforpoliticalconcessions,beyondperhapsprotecting
FenwayParkfromdestruction,whereasothergroupsaremuchmoreinclinedto
pushforpoliticalconcessions.Ontheotherhand,RedSoxfanshaveagreatdealof
sharedcultureanddemonstrateingroupbehaviorevenwithstrangers.RedSox
fansprobablypubliclygatheratgreaterfrequenciesthan,say,Republicansor
academics.Butevenso,beingaRedSoxfanhasonlylimited,ifany,spilloverinto
onespublicsphereattitudes,whereasthepoliticalgoalsoflesbians,Republicansor
academicsaremuchmorepronounced.Thisexamplemightencourageustothink
thatwecanatleastmakedistinctionsbasedonhowmuchgroupsaddressthepublic
sphere,butthisalsofails.ConsiderthecaseoftheSierraClub.TheSierraClubwas
originallyanorganizationwhosepurposewastoorganizeoutingsforpeoplewho
lovedtheoutdoors.Itprovidedameanstoplantrips,meetotherpeopleinterested
intheoutdoors,andothersimilarfunctions.However,itevolvedintoan
organizationwhoseprimaryaimistolobbythegovernmenttoprotectthe
environmentbymeansofregulation.Onecanseehowthiscouldhavehappened:
themembershipwasinterestedinoutdooractivity,andfromthat,theygainedan
interestinprotectingtheoutdoorstoensuretheircontinuedenjoymentofit.So
theywereabletojointogethertolobbythegovernmenttosupplyapublicgoodfor
whichtheyhadprivatemotivations.ItisnotinconceivabletoimaginethatRedSox
fanscouldeventuallyturntopoliticalactivismtoperhapsliftimmigration
restrictionsonCubanorDominicannationals,giventheirtendencytobegood
baseballplayers.

23
Groupscandomorethanshiftfromaprivatefocustoapublicfocus.

Groupscandisappearaltogether,liketheWhigs,ortheycanmerge,asinsomeways
theEuropeancountriesarebeginningtodo.Groupscanalsosplinterinto
subgroups,oncethegroupsmembershipstartsrecognizingadditionaldistinctions
thattheyfindincreasinglysalient.ThevariousstrainsofChristianity(orevenjust
RomanCatholicism)areexamplesofthiskindofsplinteringovertime.This
inconstancyofgroupsovertime,notjustofmembership,butoffocusand
distinguishingfeaturesfurthersuggeststhattreatinggroupsthemselves,andthus
culturalidentity,astherelevantunitofanalysisinadiscussionofdiversityinthe
publicspherewouldbeamistake.Whilegroupssurelyexist,theyarebothdifficult
todifferentiateatanabstractlevel,andinconstantthroughtime.
Thoughweareabletodismisstheconceptofgroupsbeingnaturalkinds,this
objectivedistinctiondoesnotprecludeasubjectiveperceptionofgroupsasnatural
kinds.Thatis,itispossiblethatMuslimsorItalianscouldviewthemselvesas
naturalkinds.Theycouldholdthisviewregardlessoftheobjectivefactsofthe
matter.Itislikelythatthemembersofatleastsomegroupholdthisview,andsowe
needtoconsiderthisandseewhatconsequencesithasfortheanalysisIsuggest.
Fortunately,thereisastraightforwardresponsetothisworry.Recallthatwe
aretryingtodeterminewhattheappropriateunitofanalysisisinasocialand
politicalcontext.Eveniftheindividualshavethemistakenbeliefthattheirgroupis
insomesensenatural,theabstractanalysisofthesocialdynamicstilloughttofocus
onindividuals,forthereasonselaboratedabove.However,thereshouldbesome


mechanismbywhichwecandescribegroupsthatrecognizesthissubjective

24

stance.Somegroupswillexistforlongerperiodsoftime,orperhapshave
heightenedattitudestowardsingroupindividuals,andthisshouldberecognized.
Thatis,individualswhoconsiderthemselvesmembersofgroupsandfurthersee
thesegroupsasfundamentalaregoingbehaveinwaysthataredifferentthan
individualswhoseeindividualsasfundamental.But,thisdoesnotrequirethe
grouptobetheunitofanalysis.Wecaninsteadprovideananalysisofthis
differencepurelyattheindividuallevel,asitistheindividualswhoseethemselves
inthisway.Whatisneeded,then,isthemeanstodescribethedifferencesin
individualperspective,andtocharacterizesomeindividualfeaturesasthosethat
areindicativeofmembershipinacertaingroup.Inthefollowingsection,Idevelop
anaccountofwhatpropertiesofindividualsweoughttoconsiderthebasicbuilding
blocksofdiversity.Itiswiththesesamebuildingblocksthatwecanconstruct
groups,bytreatingthemasportfoliosofindividualproperties.Theseportfolios,as
willbeexplainedinmoredetaillater,canbethoughtofastheconstraintsonan
individualsattributessuchthattheindividualcouldbeconsideredamemberofa
particulargroup.AtrivialexampleisthatRedSoxNationmemberswouldhavethe
constraintofhavingtohatetheYankeesandlovetheRedSox.Withoutthis
constraint,onecouldnothavemuchclaimtogroupmembership.
Itisbymeansoftheportfolioaccountofgroupsthatwecanstilldiscuss
groups,andstrongindividualfeelingsaboutgroups,whilestillrelyingon
individualsastheappropriateunitofanalysis.

25

IndividuallevelDiversity
Thoughwehavealreadyestablishedthatwearenotlookingatgrouplevel
diversity,andinsteadarefocusingonindividualleveldiversity,wehavenotyet
examinedtherelevantkindsofindividualdiversity.Ofcourse,therearemanyways
inwhichindividualsaredifferentthechallengeinthiskindofanalysisisto
identifythedifferencesthatcanaccountformuchofthedynamicsofboththe
privateandpublicspheres.Wehavediversityofourpreferencesorgoalsinbotha
politicalandpersonalcontext.Further,evenifweagreeonthesamegoalsthe
desiretolowercrime,forinstancewecanoftendisagreeovermeanstoourends.
Finally,wehavediversityinourabilities,perspectives,andourknowledge.We
bringdifferentskillstothesameproblems,andaproblemcanbeeasierorharder
dependingonhowweapproachit.Letusconsidereachoftheserelevantkindsof
diversityinturn.
Preferencediversityisfundamentaltobotheconomicandpoliticalmodelsof
individuals.Inthesemodels,individualshavewhatiscalledapreferenceordering
acompletelistofthegoodsoroutcomesthatanindividualisinterestedin,inorder
ofhowstronglythatpersonisinterested.So,forinstance,onecanimaginea
(partial)preferenceorderingforfruits,inwhichIpreferorangestoapples,applesto
bananas,andbananastopeaches.Preferencesareassumedtobetransitive,which
istosaythat,giventhisordering,Imustpreferorangestopeaches.Itisclearhow


26
diversitycanenterintotheframework:whileImightpreferorangestoapples,it
isveryeasyforsomeoneelsetopreferapplestooranges.Whileadiversityof
preferencesoverfruitsisfairlyinnocuous,whenpreferencesareoversocial
outcomes,likewhichpublicgoodstoprovision,orwhichsocialpoliciestopromote,
preferencediversitycanbeextremelyworrisome.
Theliteratureonsocialchoicedemonstrateshowmuchofaproblem
preferencediversitycanbe,stemmingfromArrowsimpossibilitytheorem.Arrows
impossibilitytheoremassumesatleastthreeindividualswanttovoteoveratleast
threepossiblechoices.Wefurtherassumefourbasicconstraintsonthetypeofvote
thatweareallowedtouse.Thefirstclaimsthatthereisnodictatorwhodecidesthe
outcomeirrespectiveofothersinterests.Thesecondconstraintisthatthevoting
mechanismshouldbeuniversalitshouldtakeintoaccountallofeachvoters
preferences,andprovidethesamevotingresultgiventhesamesetofpreferences.
Thisjustistosaythatthereisadeterministicvotingmechanismthatincludesallof
thevoterspreferences.Thesefirsttwoconditionsjustmakeitclearthatweare
interestedindemocraticvoting.Theotherconditionsaremeanttoimposesome
basicconsistencyrequirements.ThefirstisthatifeveryoneprefersAtoB,thenthe
collectivevoteshouldalsopreferAtoB,asthevoteshouldrecognizethedominance
ofAtoB.Ifeveryoneprefersapplestooranges,thenifwevoteonwhichfruitto
buy,thevoteshouldnttellustobyoranges.ThesecondisthatifthevoteprefersA
toB,ifwethenexpandtheoptionsettoincludeC,thevotedoesnotthenrankB


aboveA.Thatis,ifweaddbananastothemix,thatshouldntcauseusto

27

suddenlypreferorangesoverapples.9
EvenpriortoArrow,politicaltheoristssawthatpreferencediversitywasa
problemtobesolvedHobbessolveditwiththeintroductionoftheLeviathan,but
mostpoliticalphilosopherssincehaveattemptedtodealwithitbymeansofthe
publicprivatedistinction.Individualscandifferalltheywantinwhatfruitsthey
prefer(andinfactitisgoodforthemtodiffer,insofarasitenablesmorepeopleto
satisfytheirpreferencesduetolessresourcecompetition),butinthepublicsphere,
theclaimhasbeenmadebysocialcontracttheoriststhatourrationalityandpublic
spiritednesswillleadustoconsensusinwhatpublicgoodsweoughttopursue.
Becauseofitscentralitytotheproblemsthatpoliticalphilosophyhastriedto
address,preferencediversitywillalsoholdaprevalentpositioninthepresent
analysis.
Preferencediversityissometimesconfusedwithadistinctnotion:diversity
overthechoiceofmeanstoends.WhilepreferencediversityasIvedefinedit
rangesoveroutcomes,thereisalsodiversityoverthemeanstoachievethose
outcomes.Thisisoftenoverlookedintheanalysisofpoliticalconflict,butis

9
For a more detailed discussion, see Arrow, A Difficulty in the Concept of Social
Welfare, Journal of Political Economy 58(4) (August, 1950), pp. 328-346 Gibbard and
Satterthwaite have shown similarly damning problems wherein voting systems are either
dictatorial, have a choice that cannot succeed, or are open to strategic manipulation. See
Allan Gibbard, "Manipulation of voting schemes: a general result", Econometrica, Vol.
41, No. 4 (1973), pp. 587601 and Mark A. Satterthwaite, "Strategy-proofness and
Arrow's Conditions: Existence and Correspondence Theorems for Voting Procedures and
Social Welfare Functions", Journal of Economic Theory 10 (April 1975), 187-217.


28
essentialtounderstandingwhenapparentlyconflictingpreferencescaninfactbe
reconciled.Forinstance,ifonepartyisinterestedininvestingresourcesinbigger
prisonsandmorestringentlaws,andtheotherpartyisinterestedinbetter
educationandmoresocialworkers,itmaybethattheiroutcomepreferencesarein
line,buttheirpreferencesovermeansdiverge.Inthiscase,bothsidesmight
ultimatelybeinterestedinreducingcrime.Itisrarethatsomeonehasaninterestin
prisonconstructionforthesakeofit.Ifanalysiscanhelpdemonstratethatthegoals
areinfactthesame,theconflictcantransitionfromanideologicalone,whichcan
oftenbeintractable,intoanempiricalissueofwhichmethodhasgreatersuccessin
efficientlyreducingcrimerates.Whengoalsagree,andpreferencesovermeans
differ,theroleforsocialexperimentationbecomesclear:itallowsasocietyto
resolvedisagreementsthatareovermeans.10
Socialexperimentationfindssupportfromanareaotherthanpreference
diversity:italsofindssupportfromskilldiversity.Individualshavenotable
differencesinskills,andtheseskillswillleadthemtoengageindifferentprofessions
andprojects.Economictheoryhashadmuchtosayaboutthebenefitsofskill

10Achallengefortreatinginstrumentalmethodsasdistinctfromendsisthatthe
familiarAristotelianargumentcouldbeusedtoarguethatallbutoneofourends
areactuallybestunderstoodasinstrumentalmeanstoafinalendofhappiness.In
practice,Idonotconsiderthistobeagraveconcern.Aristotlehimselfrecognized
thedistinctionbetweenagoodthatwasmerelyameanstowardssomeotherend,
andagoodthatisalsogoodinitself.Thedifficulty,then,istoidentifythesedistinct
cases.Whilemanygoodscouldbeeitheranendorameans,thegoodinthecontext
ofitsadvocacyprovidessufficientinformationtodistinguishbetweenthetwocases.
Forinstance,aprisonwouldnotbearguedforonthebasisofitsaestheticappeal,
whereasamuseumwouldbe.


29
diversity,aswewillseeinlatersections.Akeyquestiontoinvestigate,however,
iswhattherelationshipisbetweenanoftenbeneficialdiversityofskills,andan
oftencontentiousdiversityofpreferences.
Justasskilldiversityismoreclearlybeneficialtosocietythanpreference
diversity,thediversityofperspectivesandknowledgeisalsoanareawherewecan
morereadilyexpectbenefitstoaccrue.ScottPageandLuHonghavearguedthat
perspectivediversitythedifferenceinhowindividualscategorizetheworldand
analyzeproblemsisbeneficialtosolvingmanykindsofdifficultproblems.Infact,
theyarguethatdiversityofthiskindcantrumpindividualabilityintermsofits
epistemicvalue.11Acommonexampleofperspectivediversityishowonewouldgo
aboutcategorizingavarietyoffoods.Differentpeopleoftenchoosedifferent
methodsofcategorization.Forinstance,somemayorganizethembywhetheror
nottheyareorganic,somebycountryorregionoforigin,somebywhetheritis
frozen,canned,orfresh,somebycolor,somebynutritionalcontent,someby
whetheritisafruitoravegetable,somebyprice.Manyothercategoriesare
possible.Mostlikely,noonewouldoptforallofthesecategoriesatonce,butforany
givencategory,somesubsetofthepopulationwouldfinditsalient.Forinstance,a
poorerindividualmightbeinclinedtofocusonprice,whileawealthierperson
mightfocusonwhetherfoodisfreshorfrozen.Thesecategorizationsfitintoa

11Forafurtherelaborationofthisargument,seeLuHongandScottPage,Groupsof
diverseproblemsolverscanoutperformgroupsofhighabilityproblemsolvers
ProceedingsoftheNationalAcademyofSciences,November16,2004,andPages
bookTheDifference(2007).


30
largerbackgroundofthekindsofproblemsthattheseindividualstypicallyhave
tosolve.Ifoneisblind,forexample,colordistinctionsarenotparticularlyrelevant.
Somewhattiedupwiththenotionofperspectivesisindividualknowledge
diversity.Ourbackgroundknowledgeshapesandisshapedbyourperspectives,but
knowledgeisstilldistinct.Somespecializedknowledgeisalsoanability,suchas
knowledgeofmathematicsorphysics,butmuchspecializedknowledgeisjust
knowledgeofparticularfacts.Forinstance,knowingaboutthebattingaverageofa
particularbaseballplayer,orknowingaboutalocalshortageofmachineparts.
Knowledgecanbeofafairlytrivialnature,likebaseballstatistics,orhavegrave
consequences,likethepredictionsofclimatemodels.Itcanalsosimplybeculturally
enriching,liketheknowledgeofMozartorDostoevskysworks.Thiskindof
knowledgeaidsinshapingperspectives,eventhoseatafairlyhighlevel.Ones
exposuretoliteratureandmusic,forinstance,islikelytohelpshapewhetherone
considersoneselfahumanist.Greaterawarenessofhistoryislikelytoresultina
greatertendencytoframecurrenteventsinthecontextofpastevents.Knowledge
oftheprinciplesofeconomicsmightencourageaparticularviewofhowindividuals
makedecisions.Thuswefindthatknowledgediversitycanhaveasignificantrolein
understandingindividualdiversity.
Nowthatwehavediscussedthedifferentaxesonwhichonecanconsider
individualdiversity,wecanbrieflyreturntothecommonnotionofidentityor
culturaldiversity.Thoughwehaveseenthatculturaldistinctionsarenotbasicin
thesensethattheyarenotappropriateunitsofanalysis,itdoesnotmeanthatwe


havenopossiblemeansofdiscussingcultureandculturaldiversity.Wecan

31

insteaddescribeaculturalidentityasaportfolioofpreferences,skills,perspectives
andknowledge.Underthismodel,culturescanbecharacterizedbyapartial
orderingofpreferences,asetofcommonlyassociatedskillsandperspectives,and
whatculturalknowledgeisessentialtomembership.Thesecanbeseenas
constraintsonindividualsinsofarastheirindividualidentitymustsuitablycoincide
withtheserequirementsinordertobeabletolayclaimtotherelevantcultural
identity.Eventhoughconstraintsareimposed,agreatdealofflexibilityremains.
Thisportfolioapproachallowsoneindividualtosimultaneouslybelongtomultiple
groups,asgroupsarenotcompletedeterminantsofindividualcharacteristics.It
alsoallowsustoinvestigateculturalchange,aswecaneasilyexaminethe
dimensionsacrosswhichindividualsmightchangeovertime,includingwhat
incentivesmightinducethemtodoso.Mostimportant,however,isourabilityto
promotediversitywithoutmakingthemistakethatdoingthatentailsossifying
currentsocialcategories.

TheSocialContractApproachToDiversityinThePublicSphere
Socialcontracttheoryhasalongandmultifacetedtradition,buthereIam
focusingonlyontheroleofdiversityinthepublicsphere.Furthermore,Iam
consideringonlyRawlsandRousseau,andIwilltreatthemasrepresentativeofthe


largertradition.12Thisisheavilyrestrictingthesubjectmatter,butasthegoal

32

hereisnothistoricalanalysis,butinsteadconceptualanalysis,focusingon
representativeclaimsoughttobesufficient.
Diversityassuchisnotaprimarypointofdiscussioninthesocialcontract
tradition.IfoneweretoclassifyHobbesasthefirstsocialcontracttheorist,atleast
intheModernperiod,theclaimcouldbemadethatHobbesisanexception.He
definespeopleasequalsonlyinthesensethattheweakestcanalwayskillthe
strongest,andtheleastintelligentcanstillconspireagainstthemostintelligent.In
otherrespects,peopleareratherdifferentfromeachother,andimportantly,they
arelikelygoingtohavedifferinginterests.Thisdifferenceofinterestsisso
fundamentalandsoproblematicthatheshapesmuchofhistheoryaroundtryingto
mitigatethechallengesitcreates.Namely,theinstitutionofaLeviathanmustbe
imposedsoastoofferpeoplethemeansofofferingenforceableassurancesthatthey
willnotviolentlytrytoimposetheirwillonothers.Inhisanalysisoftheformsof
government,amonarchyisdeterminedtobethebestkindofgovernmentbecauseit
istheonlyoneinthereisnopossibilityforcivilstrifecausedbydisagreements
amongpoliticalleaders.13Soacertainkindofdiversitydriveshisview,andforces
himtocreatepoliticalinstitutionswhosefocusistoregulateandmitigatethe

12
I do not focus on Locke here because his method of argumentation, both in the Treatise
and in A Letter Concerning Toleration, relies on ones relationship to God and religion.
This is not found in other social contract theories, and as such, there is less to be gained
by using it as an example.
13
See Leviathan Book 2, Chapter 19 for a detailed discussion of how only a monarch can
lead, and not an assembly. The focus of the argument is on the potential for disagreement
and different interests.


negativepowerofdiversity.AfterHobbes,however,thiskindofargumentfell

33

outoffavorwithsocialcontracttheorists.Theunderlyingassumptionofakindof
moralegoismwasreplacedwithanotionoftheReasonable,whichwasfoundin
Lockeandallsubsequenttheorists.Oncethismovewasmade,thethreatthat
diversityposedwaslargelyeliminated.Becauseofthis,Iclaimthattherewasno
longeranyneedtofocusonitasatheoreticalworryinsteadfocuscouldbeturned
tothewaysinwhichindividualsaresimilar.Thisisnottosaythatmostthinkers
endedupclaimingthateveryonewasidentical.Rather,myclaimisthatpost
Hobbes,diversitywasviewedaslargelyirrelevantfromthepointofviewofthe
publicsphere.WithHobbes,diverseinterestscouldclashwithoneanotherunlessa
socialcontractprovidedsomemethodofadjudicatingbetweenthem.AfterHobbes,
asocialcontractbecameadeliberationonthecommongood.Thisisasignificant
shiftintheperceptionofthepublicsphere.
Onemightsupposethatthestorycanbreakdownwhenweconsider
somethingliketheliberalvalueoftolerance,whichisexplicitlyaboutwhatposition
publicindividualsoughttotakeinthefaceofdiversity.However,argumentsfor
tolerance,likeLockes,donotargueforthevalueofadiversityofviews,nordothey
arguethatdiversityisnegative.Insteadtheyarguethatattemptingtoforce
someoneto,say,converttooneschoiceofreligionunderminesonesreligion.
Statesaretostayoutofreligionforsimilarreasonsthethreatthatonespreferred
religionwouldbeoutlawed,ordilutedbyimposingitonnonbelievers.Thiskindof
argumenthasnothingtodowithdiversityperse,asitisnotanobjectofanalysis.


34
Instead,itfocusesonindividualstrategiesformaintainingonesownbeliefs.As
such,accountsoftolerationthattakethisformarenotthemselvesacounter
exampletomyclaimthatsocialcontracttheoristsconsiderdiversitytobeirrelevant
toanyconsiderationsinthepublicsphere.
Toprovidefurthersupporttothisclaimaboutthenatureofthepublicsphere
inthesocialcontracttraditionafterHobbes,letusfirstexamineRousseaus
distinctionbetweentheGeneralWillandtheWillofAll,alongwithRousseaus
accountoftheroleoftheGeneralWillinthestate.Usingthisasabaseline,wecan
thenmoveontoananalysisofRawlsOriginalPosition.Iclaimthatbothofthese
areexamplesofanattempttoestablishapublicspherethatisnotgroundedinthe
privatesphere.

RousseauandtheGeneralWill
Thewillofall,accordingtoRousseau,istheunionofalloftheprivate
interestsofthepopulace.Thisis,then,justanagglomerationofindividualwants
anddesires,andnotsomethingthatRousseauconsidersrelevanttothepublic
sphere.Thereasonforthisisthattheperspectiveundertakenbyindividualssoas
tocreatethewillofallisnotpublicminded.Thewillofallistheunionofprivate
goods,notthepublicgood.ThisisacrucialfeatureofthewillofallforRousseau,as
theindividualwillsthatcomprisethewillofallneednotconsideranyothersin


35
society,andmayevenwishgoodsthatcouldbedisadvantageoustoothers.The
generalwill,ontheotherhand,requiresadifferentperspective.
Thegeneralwill,unlikethewillofall,ismadefromtheperspectiveof
individualsaspublicmindedcitizens,notprivateindividuals.Itisfromthispublic
perspectivethatindividualsattempttodeterminewhatisinthepublicgood,
separatingthemselvesfromtheirprivateinterests.Instead,asRawlsarguesinhis
lecturesonpoliticalphilosophy,theyconsiderwhattheirfundamentalinterestsare,
derivedfromtheirbasichumanity.Aseveryonehastheseinterests,thegeneralwill
issomethingthatoneoughttobeabletoendorseduetothefactthatthepublic
goodisthenalsoindividualgood.Ratherthanthegeneralwillbeingaunionof
individualinterestslikethewillofall,thegeneralwillcanbeunderstoodasthe
intersectionofpublicinterests.14Notehowever,thatthedifferenceisnotmerelythe
substitutionoftheintersectionoperationfortheunionoperationinhowwecollect
individualviews.Theindividualwillsbeingcollectedareofadifferentnature.The
generalwillisoperatingwithpublicmindedviewsofthepublicgoodbasedonour
fundamentalinterests.

14RousseauhereislookingtotakeadvantageofsomethingliketheCondorcetJury
Theorem,or(asRawlsdescribesit)Bernoulli'slawoflargenumbers:solongasany
givenindividualisatleast50%likelytobecorrectastothetruenatureofwhatthe
publicgoodis,thenthegeneralwillwillprescribeit,providedthatthereisalarge
enoughpopulation.Butforthistobeapplicable,therecannotbeanylargeinterest
groups,butmoreimportantly,therecanbenocommunicationbetweenindividuals.
ThisstrangerequirementisduetothefactthatBernoulliandCondorcetrelyonthe
independenceofstatisticalsamples.Ifindividualscommunicate,theirassessments
ofthepublicgoodwillbecomecorrelated,andthenthetheoremsarenotapplicable.
ForRousseaususageofthesetheorems,seeTheSocialContract,book2,chapter3.

36
ItisherethatwecanseetheshiftmadebyRousseauthatHobbesdoesnot

makebutRawlswillalsomake:individualsaretodeterminethegeneralwillnotby
theirknowledgeoftheirprivateinterests,butbytheirknowledgeofour
fundamentalsharedinterestsashumans.Buttosomedegree,thatiswhatweare
tryingtodetermineinthefirstplace.Thepublicstancetakenbytheindividualscuts
individualsofffromtheirpersonalconceptionsofthegood,whichistheonlyreadily
availablesourceofinformationinunderstandingwhatoursharedneedsmightbe.
Afurtherquestionisatwhatlevelofanalysisdowefindthegeneralwill.Inhis
lecturesonRousseauspoliticalphilosophy,Rawlsclaimsthatthegeneralwillisa
methodofpublicreason.15Becauseofthis,eachindividualhasthegeneralwill,as
wellasherownprivatewill.Butthisdoesnotseemtobeconsistentwith
Rousseausdiscussionofthegeneralwill.Ifthegeneralwillistobedeterminedby
theintersectionofindividualpublicwills,thenitissomethingdifferentfromany
individualwill,oratleastthevastmajorityofindividualwills.Thegeneralwillis
alsodescribedbyRousseauasbeingcorporealitissomethingthatbecomes
independentoftheindividualsthatgeneratedit:Thebodypolitic,therefore,isalso
acorporatebeingpossessedofthisgeneralwillwhichisnolongerthatofany
individual.16Thisdoesnotseemconsistentwithananalysisofthegeneralwillasa
methodofdiscourse.

15
See Rawls, Lectures on the History of Political Philosophy, Samuel Freeman, ed.
(2007)
16
Discourse on Political Economy, pp.132, 140

37
However,thegeneralwillismeanttobeusedasthebasisofthepublic

sphere.17Itdefinestheconditionsofjustice,freedomandequality,andgovernsthe
relationshipsbetweenindividuals.Butaswehavenoticed,thegeneralwilldoesnot
utilizeanyaccountofindividualdifferencesitservesonlytostripthemaway.The
privatesphereisexplicitlyrejectedasasourceofthepublicsphere,asthegeneral
willrequiresapublicstancethatrejectsindividualinterests.Thegeneralwillitself
isthenmadetoexistindependentlyoftheparticularsoftheindividualsinsociety
whocreated(orfound)it.18Herewefindparallelstoourdiscussionoftherelevant
unitsofanalysisinourreasoning.Therewerejectedanotionofgrouplevel
analysis,andinsteadembracedindividualsastheappropriateunitofanalysis.But
herewefindthegeneralwillasitselfoperatingatagrouplevelofanalysis:itisan
entityindependentofitsmembers,anditisformedbydeducingtheirfundamental
sharedinterests.19Itsindependencebyitselfwouldbesufficienttosuggestthat
thereisaproblem,butletusalsoconsiderthenotionoffundamentalshared
interests,andshowwhyitalsowouldbeacauseforconcern.
Thechallengeoffundamentalsharedinterestsistheassumptionthatthereis
apropersetoffundamentalsharedintereststhatcoverallindividualsastheyare,
doesnotleaveanyimportantinterestsofindividualsout,andissufficientlywell
definedsuchthatthesetofinterestsisenoughtoaccountforthepublicgoodsthat

17
ibid, p. 132
18
ibid, p. 140, The Social Contract, book 2, chapter 3
19
As long as several men in assembly regard themselves as a single body, they have
only a single will (The Social Contract, book 4, chapter 1) Also relevant are book 4,
chapter 2, as well as book 2, chapter 4.


38
thestateoughttoprovide.ButtoclaimasRousseaudoesthatsuchasetexistsis
justtobegthequestion.Thepublicsphere,afterall,ismeanttostructuresocietyin
suchawayastoaccommodateourneeds.Butiftherearefundamentaldifferences
inwhatourneedsare,oreveninwhatourunderstandingisofthesedifferences,the
publicsphereiswherethisdiscussionneedstotakeplace.Thegeneralwill,which
underliesthepublicsphere,reliesontheideathatthereisasingleunchangingsetof
fundamentalhumanneeds.Asthefundamentalinterestsunderdiscussionarethose
intereststhatwetakeonasmembersofasociety,andnotasthepresocialbrutes
thatRousseauimaginesinthestateofnature20,itisevenmoreplausiblethat
differentindividuals,orevendifferentsubgroups,havedifferentfundamental
needsinsocietytosatisfytheirconceptionofthepublicgood.21Butifthisisthe
case,thenweatleasthaveplausiblegroundsforquestioningwhetherthereisaset
offundamentalsharedinterests.Giventheexistenceofthisquestion,thenthe

20
The passage from the state of nature to the civil state produces a very remarkable
change in man, by substituting justice for instinct in his conduct, and giving his actions
the morality they had formerly lacked. Then only, when the voice of duty takes the place
of physical impulses and right of appetite, does man, who so far had considered only
himself, find that he is forced to act on different principles, and to consult his reason
before listening to his inclinations. Although, in this state, he deprives himself of some
advantages which he got from nature, he gains in return others so great, his faculties are
so stimulated and developed, his ideas so extended, his feelings so ennobled, and his
whole soul so uplifted, that, did not the abuses of this new condition often degrade him
below that which he left, he would be bound to bless continually the happy moment
which took him from it for ever, and, instead of a stupid and unimaginative animal, made
him an intelligent being and a man. The Social Contract, Book I, Section 8, Paragraph 1
21
Since these needs are not food and shelter, but rather social needs, one would expect
that these would at least track social, economic and technological changes. Rosseaus
doubt about this is understandable, as at the time he was writing, there had yet to be
significant technological changes that fundamentally affected the quality of life. These
did not arrive until after the industrial revolution and subsequent advances in farming.
For evidence of this lack of social change, see Gregory Clark, A Farewell to Alms. (2007)


constructionofthepublicsphereoughttothenpayattentiontothediversityof

39

basicneedsandintereststhatarerepresentedinsociety.
SoweseewithRousseauthattheformulationofthegeneralwillis
problematiconafewgrounds.First,wefindthatthegeneralwillisdetachedfrom
theindividualswhowillit,aswesawinRosseauscorporealdescription.Thisleads
todifficultiesverysimilartothosefoundinthenotionoftreatinggroupsastheunit
ofanalysis.Insofarasthegeneralwillaimstodescribethepublicmindedconcepts
ofjusticeforthegroupatlarge,wehavethebasicproblemofestablishingwhatthe
groupis,andwhetheritiswelldefinedovertime.Thiscanbemadeevenmore
problematicifdifferentindividualshavedifferentconceptionsofthegroup,such
thateachmemberhasadifferentbasisfordetermininggroupmembership.For
example,someRepublicansmightclaimthatonehastobeinfavorofsmall
governmenttocountasaRepublican,whileothersmightthinkthatRepublicanism
isinsteadbetterdefinedinvirtueofonesconservativesocialattitudes.Thesetwo
criteriawillcreateoverlapping,butnonidenticalgroups.Neithergroupcanrelyon
beingtherealgroup,astheyexistonlyinvirtueofindividualsjoininguptocreate
them.Thiscanberesolvedbydecidingonesgroupiseveryone,sincetherecannot
beanymisunderstandingovergroupcomposition.Butwhiletheextensionofthe
groupisthenclear,theintensionofthegroupisnotmanygroupsaredefined
contrastively,andifthereisnothingtocontrastagainst,itishardtohaveanin
groupwithoutanoutgroup.

40
Evenifwesupposethatwecanresolvetheissueoftreatinggroupsasthe

appropriateunitofanalysis,wefindthattheprocessofwillingisalsoproblematic.
Thiscanbestbeunderstoodintermsofcontemporaryworkinsocialpsychologyon
socialidentity.Socialidentitytheorists,suchasMarilynnBrewer,havearguedthat
ouridentityismorethanjustourpersonalidentity,butseveralsocialidentitiesas
well.Onemightidentifyoneselfasaphilosopher,andseeonesidentitytiedupin
thesuccessesandfailuresofotherphilosophers.Similarly,onecouldgeneralize
furtherout,andseeoneselfasanacademic.Thisgeneralizationcouldkeepgoing,or
onecouldstartfromanewbasisforasocialidentity,suchasbeinganamateur
singer.22Inthistheory,wehavemanydifferentversionsofourselves,andwithany
ofthesocialselves,ouridentityistiedupinthegroup.Assuch,thegroups
interestsareourowninterests,becauseweseeourselvesasintermsofourgroup
membership.ItisherethatwefindsimilaritywithRousseau:thepublicperspective
takeninconstructingthegeneralwillisakintotakingonasocialidentity.However,
wehavemanysocialidentities.Howdoweallchoosethesameone?Brewerclaims
thatcertainidentitiescanbemademoresalientthanothers:goingtothevoting
booth,forinstance,willtendtoencourageonesidentityasacitizen,whereasgoing
tothestockexchangewillencourageonespersonalidentity.ButBreweralso
claimsthatoneschoiceofsocialidentitywillbesuchthatonesatisfiescompeting
interestsforindividualityandaneedtofitin.Ifthisweretobetrue,thenviewing

22
For a more detailed discussion, see Brewer, Marilynn, The Social Self: On Being the
Same and Different at the Same Time. Personality and Social Psychology Bulletin, 17,
475-482. (1991)


oneselfsimplyasacitizenisnotgoingtobesuccessful,asthereisnoroomfor

41

anyexpressionofindividuality.
Insocialidentitytheory,thereareusuallytwowaysinwhichasocialidentity
canbeestablishedorused:oneisthatthereisacommonsetofinterests,orthat
thereisanoutgroupthatdefinestheingroup.Inthecaseofcommoninterests,one
findsthatthegroupssuccesscorrelateswithonesownsuccess.Anincreasein
prestigeforphilosophyingeneralincreasestheprestigeofindividualphilosophers.
Butthishassomelimits:incommonsdilemmasituations,socialidentificationcan
helpincreasecooperation,butnotasmuchasitwouldifpeopletrulysawthe
groupsinterestsasbeingtheirowninterests.Inthecaseoftheexistenceofanout
group,individualswilltendtodothingsthatincreasestheadvantageoftheirgroup
membersevenifthegroupdivisionisessentiallymeaningless,likewhetheroneis
anoverestimatororunderestimatorofthenumberofdotsonapieceofpaper.23
Giventhesuccessofgroupdivisioneveninthesemeaninglesscontexts,itdoesnot
seemasifgroupidentityassuchiswhatisdrivingtheresults.Afterall,nooneis
goingtoclaimthatoverestimatorsofdotssharecommonintereststhatare
differentfromtherestofthepopulation.24

23SeeTajfel,H.,Billig,M.,Bundy,R.P.,&Flament,C.Socialcategorizationand
intergroupbehavior.EuropeanJournalofSocialPsychology,1,149177.(1971)
24Analternativeaccountforexplainingcooperationandsocialcoordinationcanbe

foundintheoriesofsocialnorms.Insteadofclaimsaboutsocialidentity,anorms
basedapproachreliesontheideathatindividualsdeterminethatsomebehaviors
areexpectedoftheminthepopulation,andtheycareaboutmeetingthose
expectations,potentiallybecauseofafearofsocialsanction.Thisapproachismore

42
Sincethegeneralwillismeanttoencompassanation,ifnotallof

humanity,itisdifficulttoseehowthisprocessofwillingcanbedone.Thegroupis
toolargeforsocialidentitytobestrictlymeaningful,unlessthereisanoutgroupto
contrastagainst.Anditisnotclearthateachindividualsconceptionofthegroupis
thesame.Butmorebasically,withsuchalargegroup,thesocialidentitymaynot
evenexistinsuchawaythatencouragesstrictidentificationofonesinterestswith
thegroupsinterests.
Finally,thegeneralwillreliesonanotionoffundamentalsharedinterests
thatgoeswithoutelaborationordiscussion.Asthegeneralwilltriestodiscover
whatthefundamentalsharedinterestsareinthefirstplace,itisillegitimateto
assumeawaythepossibilitythatthereisnosinglesetoffundamentalhuman
interests.Particularlyasthesizeofthepopulationgrows,thelackofsocial
identificationwillmakecommoninterestsincreasinglyunlikely.
Itispossiblethatthislineofcritiqueleveragesthefactthatthegeneralwillis
afairlycomplicatedtheoreticalentitythatreliesonanequallycomplicatedsocial
process,ratherthangettingatsomethingmorefundamentalaboutthemethodology
employedbysocialcontracttheoristsintheconstructionofthepublicsphere.To

abletoaccountfortheempiricalevidence,aswellasavoidtheconceptualissues
raisedabove.Foracritiqueofsocialidentitytheoryfromanormsbasedapproach,
seeBicchieri,CooperationandCommunication:GroupidentityorSocialNorms?in
N.Gold[ed],Teamwork:MultiDisciplinaryPerspectives(2005).Foradetailed
elaborationofatheoryofsocialnorms,seeBicchieri,TheGrammarofSociety(2006)


resolvethis,letusturnourattentiontoRawlssocialcontracttheory,andthe

43

mechanismoftheOriginalPositioninparticular.

RawlsandtheOriginalPosition
Rawls,likeRousseau,facesthechallengeofdevelopinganaccountofthe
publicsphere.Ratherthantryingtousesomethinglikethegeneralwill,which
requiresindividualstotakeapublicperspectiveratherthananindividualist
perspective,Rawlsimaginesinsteadthattheparametersofthepublicspherecanbe
developedbymeansofanidealized,hypotheticalbargain:theOriginalPosition.
TheOriginalPositioncanbethoughtofastheculminationofthebasicideas
inthehistoryofthesocialcontracttradition.Thatis,wefindanotionofbargaining
thatisprevalentinHobbesandLocke,butfromRousseauwehaveanaccountof
impartialityandindependencefromthestatusquo.TheOriginalPositionsidesteps
thefictionofahistoricalstateofnaturethatsuppliesthejustificationforcertain
assumptionsaboutthenatureofhumansandtheinfluenceofsocietyonthisbasic
nature,andinitsplaceitexplicitlyprovidesanaccountoftheidealsituationfrom
whichtoreasonaboutjustice.
IntheOriginalPosition,wearetoimagineindividualswhohave
comprehensiveinterestsandindividualnotionsofthegoodthathavegonebehind
thethickveilofignorance.Thisveilradicallyalterstheirepistemicsituation.
Thoughthesearemeanttobeactualpeoplewholiveinaparticulartimeandplace,


44
havecertainskills,areofaparticulardemographic,andhaveparticularinterests
andaffiliations,behindtheveilofignorance,theyarenotawareofanyofthesefacts
aboutthemselves.Thisignoranceextendseventoknowledgeofwhatgeneration
oneisin:individualsliterallyknownothingabouttheirparticularsituation.Their
ignoranceextendsbeyondthis,however.Notonlydotheynotknowanythingabout
themselvesinparticular,theydonotknowanythingaboutthegeneralshapeoftheir
society.Theydonotknowtheproportionofrichtopoor,orwhatthedistributionof
jobsis,oranythingofevenastatisticalnatureaboutthedemographicsofsociety.
Theirignorance,however,doesnotextendtoalldomainsofknowledge:theyknow
somethingabouteconomicprinciples,naturallaws,andothertheoretical
knowledgethatisnotdependentuponparticulars.
RawlsalsoassumesthattheindividualsintheOriginalPositionarerational
inthesensethattheyobeytherulesofrationalchoiceasdevelopedineconomic
theory.Thisisanimportantassumption:Rawlsaimstoshowthatinstrumental
rationalityaloneisenoughtogethisprinciplesofjustice.25Theaxiomsofrational
choiceherearesignificantlyconstrained,astheOriginalPositionisapositionof
massiveuncertainty.Whatisstrikingaboutthismodelisthattheagentsare
engagedinabargainoverhowtodeterminethebasicstructureofsociety,which

25
It should be noted here that Harsanyi developed an account of the original position
independently of Rawls at approximately the same time as Rawls, though saw it
justifying a form of utilitarianism. There is some difference, however, in the thickness
of the veil of ignorance in Harsanyis account, as he allows some demographic
information. See Harsanyi, J. (1953) "Cardinal Utility in Welfare Economics and in the
Theory of Risk-Taking", Journal of Political Economy 61(5):434-5


45
26
shapesthedistributionofprimarysocialgoods, withouthavinganyknowledge
ofwhattheirbargainingpositionis.Inthisway,itisaverypeculiarbargainthe
partiesinvolvednotonlyhavenoideaofwhattheircurrentpositioninsocietyis,
buttheyalsohavenoknowledgeofwhattheconsequencesofthebargainarefor
themasindividualstakeholders.27ThiskindofuncertaintyissuchthatRawls
arguesarationalagentisforcedtoadoptanattitudeofheavyriskaversion.Since
onecannottellwhooneis,oneoughttomaximallyhedgeonesrisksbyensuring
thattheworstoffinsocietyisatleastaboveacertainthresholdofacceptableliving
conditions.28Thebargaincanthenbeseenasoneoveralternativeconceptionsof
justice,whichshapethedistributionofprimarysocialgoods.Sincethereisan
infinitepossibilityofalternativeconceptions,heassumesthattheagentsaregivena
shortlistofthemostplausiblealternativestohistheory,withutilitarianismbeingof
primaryinterest.29

26
Primary social goods, which Rawls articulates in section 15 of Theory of Justice, can
be thought of as liberties and opportunities, income and wealth, and how they are
distributed.
27
For a further elaboration on this point, see Rawls Some Reasons for the Maximin
Criterion, American Economic Review (1974)
28
See the article referenced in the previous footnote on the details of this reasoning.
Since risk aversion is a psychological trait, and one is supposed to be ignorant of such
traits behind the veil of ignorance, I view this as a point of contention in Rawls theory.
Arrow has several powerful critiques of Rawls analysis of the veil of ignorance in his
review Some Ordinalist-Utilitarian Notes on Rawls Theory of Justice, Journal of
Philososphy, 1973. However, I grant the assumption for the purposes of this discussion.
29
I focus on the Rawls of A Theory of Justice here, and not the later Rawls. Rawls
abandoned the Original Position as a basis for justification in the 1980s, which I
interpret as a recognition of the serious challenges he faced from economists. He later
moved to a theory of overlapping consensus that I claim still suffers from the same
basic trouble, though instead of having a basis in decision theory, it shifted to Kantian
constructivism. One can see this trajectory of his thought in Kantian Constructivism in

46
WhatIwouldliketodiscussisnotwhetherhisanalysisofrationalchoice

underuncertaintyiscorrect:insteadIwouldliketoarguethatthesetupforthis
thoughtexperimentisonethat,likeRousseausaccountofthegeneralwill,
precludesthepossibilityofdiversityinthepublicsphere.Thefirstthingtonoticeis
thatmultiplepartiesaresuperfluoustothebargainintheOriginalPosition.Since
eachindividualhasnoknowledgeofwhosheis(orevenwhethersheisashe),or
whereshewillendupinanygivendistributionofprimarysocialgoods,she(by
design)hasnothingtodifferentiateherfromanyoftheotherindividuals.OnRawls
conception,theonlyfeaturesofherselfthatshehasaccesstoarethatsheisrational,
andthatsheisriskaverseinthefaceofmassiveuncertainty.So,thereisnothingto
begainedbybargainingwithanotherindividualratherthanjustbargainingwith
herself.Neitherpartyrepresentsanyparticularinterests,nordotheyhaveany
privateinformationthatcaninformthebargain.Assuch,asingleindividualcanbe
anadvocateforallsidesofthebargainatthesametime.Sotrivially,therecanbeno
diversityintheoriginalposition.Noonehasanopposingview,specialknowledge,
differentcognitiveabilities,oranythingthatdifferentiatesthemfromtheother
parties.TherecanbeplentyofdiversityintheprivatesphereRawlsprinciplesof
justicearedefinedsoastoallowforagreatdealoftoleranceofdiversityinthe
privatesphere.Butatnopointisthereroomforanydiversityatthelevelofthe
publicsphere.

Moral Theory, Journal of Philosophy (September 1980), 77 (9): 515-572, then "The
Idea of an Overlapping Consensus Oxford Journal for Legal Studies (Spring 1987), 7
(1): 1-25, and culminating in Political Liberalism (1993).

47
Wehaveyettosee,however,whydiversitymightbesomethingthatwe

oughttopaycloseattentionto.Tobetterunderstandwhatdiversitycando,wenow
turntoaneconomicunderstandingoftheroleofdiversity.

EconomicPerspectivesonDiversity
Relyingoneconomicexamplestomotivatethepotentialdesirabilityofagent
diversityhastwomainbenefits:first,ashasalreadybeensuggested,therearea
numberofsuchexamplesattheverycoreofeconomictheory,andsecond,
economicsmighthelppointusinthedirectionofanunderstandingofhowtoallow
diversitytothrivewithoutthedownsideofincreasedconflictduetodifferencesin
individualpreferences.Marketmechanisms,whentheyareworking,havean
impressiveabilitytocoordinatetheactionsoflargenumbersofpeoplewhoagree
onverylittle.Infact,theearliestargumentsforacapitalistsystemwerenotthe
nowstandardargumentsfromefficiency,butrathertheyhadamoralperspective.
AsHirshmanndiscussesinThePassionsandtheInterests,commerceandits
mechanismswereseenasagentsofpositivechange.Smithclaimedthatmarkets
madeagentsmoreempathetic,asunderstandingwhatone'stradingpartnerswant
andneedcanbetterfacilitateexchange.Giventhatthereisamonetaryincentivefor
encouragingasmanyexchangesaspossible,thereisthenacorrespondingincentive
forcomingtobetterunderstandotherpeople.Otherkindsofinteractionsthenare
madeeasierbecauseoftheagents'moredevelopedmoralsentiments.Hirshmann


alsoidentifiesone'sinterestsinmonetarygainasaconstraintonone's

48

passionsformoreephemeralgoals,suchaspersonalhonororglory.Theselatter
goalstendedtoinvolvebloodshedandalsoalargeamountofunpredictability.
Commerce,andtheinterestsitgenerated,encouragedbothpredictabilityand
cooperation.Agentsgoingaftertheirinterestsmaynotagreeanymorethanagents
followingtheirpassions,buttheirdiverseinterestscanbecoordinatedbyamarket
withoutanyrealrequirementsforagreement.Althoughitwillbeaddressedmore
directlyinlaterchapters,asweconsiderthecasesineconomictheoryinwhich
diverseagentsareeitherrecommendedorrequired,wecanbrieflyconsiderhow
theagentsareabletocoordinate.
Aswesawearlier,thereareseveralwaysinwhichagentscanbediverse.In
thissection,wewillfocusonthree:preferencediversity,skilldiversity,and
knowledgediversity.Theseplayalargeroleineconomictheory.Thisisnot
withoutcause:whatwewillnowturntoiswhythesecasesofagentheterogeneity
areofparticularimportance.Firstwewilllookatthegeneralstructureofmarkets,
butwewillthendelveinwithamorecarefullookatRicardo'stheoryoftradeand
Smith'sdivisionoflabor.Fromthere,wewillturntoHayek'sunderstandingofthe
roleoflocalknowledgeinorganizingasociety.Finally,wewillbrieflyturntothe
rolediverseagentscanplayintheprovisioningofpublicgoods.
Amarketcanbeconsideredacollectionofrationalbuyersandsellers,and
thetransactionsthattakeplacebetweenthem.Buyershaveaparticulardemandfor
agoodorbundleofgoods,andsellershaveacertainsupply.Whatistobenoticedin


thisdescriptionisthatitimmediatelypresupposesanasymmetrybetween

49

buyersandsellers.Thatis,sellershavesomethingthatthebuyersdonot,andvice
versa.Abasicassumptionofanymarketmodelisthattransactionsbetween
rationalagentstakeplaceonlywhenthereisatleasttheexpectationofmutual
advantage.Thatis,bothpartiesmustexpecttogainfromthetransaction,otherwise
theagentwhoanticipatesgettinghurtinthedealwouldsimplynotagreetoitinthe
firstplace.Becauseofthisassumptionofrationality,ifamarketexists,itmustbe
thecasethatindividualsaredifferentintermsoftheirendowments,preferences,or
skills.
Sofarthisisatrivialclaimbuyershavetowantwhatsellerssell.Butfrom
thiswecanseemoresubstantialclaimsarepossible.Ifwearetofindfurther
conclusions,weneedtoaddinmorespecificsaboutbasiceconomictheory.Tostart
with,wewillconsiderthebasiccontributionofAdamSmith,inthefirstchapterof
hisWealthofNation:thedivisionoflabor.

TheDivisionofLabor
Smithmotivatesthenotionofthedivisionoflaborwithhisfamousexample
ofthepinfactory.Theproblemhewastryingtoaddressisbasic:insmaller
societies,theneedstobeprovisionedarefairlysmall.Becauseofthis,anygiven
individualcantakeonanumberoftasksanddothemadequatelyenoughtoensure
thatneedswillbemet.However,associetiesgetlarger,eventuallythereareso


50
manydifferentneeds,andinsuchgreatamounts,thatthepreviousmethodisno
longertenable.Inasmallsociety,theneedsforpinsarenotsogreat,andsoasingle
individualcanbeanartisan,andtakeontheentiremanufacturingprocess.Butina
largesociety,thereissuchalargedemandthattheresimplyarenotenough
individualstoprovisionpinsinthismanner.Soinstead,pinmakingisbrokendown
intoseparatetasksdrawingthewire,straighteningit,cuttingthewire,and
grindingtheends.Anygivenindividualisresponsibleonlyforlaboringonasingle
task.Soratherthanmakeanentirepin,anindividualwillonlycutwire,while
otherswillperformtheotherelementsofthemanufacturingprocess.30
Thisdivisionofthetasksallowsindividualstobecomespecialized:thatis,
theywillgainadditionalproficiencyatperformingtheirgiventask.Smithdescribed
thisasgainingadditionaldexteritywiththeirspecialization.Anotheradvantageis
thattimeissaved,asworkersdonothavetoconstantlyswitchtasksworkerscan
spendmoretimebeingproductive.Afinaladvantage,whichisnotrelevanttothis
discussion,isthatspecializationalsoencouragesspecializedmachinerythatcanaid
inalaborer'stask.
Itisthenotionofspecializationthatisofparticularinteresthere:thefactory
isabletobecomemoreproductivebecausedifferentgroupsofagentstakeon
differenttasks,andbecauseonegroupdedicatesitselftoitstask,itcanmoreeasily
developbettertechniques.Animportantconsequenceofthisisthat,asmore
opportunitiesformorefinegraineddivisionsoflaborbecomepossible,thereare

30
See The Wealth of Nations, Book 1, chapter 1 for the full discussion.


51
31
economicrewardstoutilizingthem. Thedivisionoflabor,then,canbeseenas
anengineforthepromotionofskilldiversity.Thishastwocomponents:first,it
demonstrateshowtotakeadvantageofanexistingsetofskilldiverseagents,and
second,itshowshoweconomicpressuresexistforincreasingapopulations
diversityovertime.
SmithsDivisionofLaborisseenasamajorengineforgrowth.Thematerial
benefitsofcontinuingthedivisionareverylargeatasociallevelitenablesmore
goodstobeproducedatmuchlowercosts,andsothewellbeingofeveryindividual
isindirectlyimprovedthroughtheavailabilityofmoreconsumergoodsatlower
prices.Itisuseful,however,atthispoint,tonotethattheadvantagesofthedivision
oflaborcanbeviewedeitherfromtheperspectiveofthesocietyortheindividual.
Socially,itistheoreticallyclearthatthedivisionoflaborcreateshugegains,and
therearealwayssocialinterestsinpromotingfurtherdivisionsanddistinctions
amongagentsandprocesses.However,fromanindividualperspective,thereisa
tradeoffthatisnotrepresentedatthesociallevel.Whileindividualstendtobe
mademoreeconomicallywelloffbecauseoftheirincreasedproductivity,Smith
noteswellbeforeMarxdoesthatthiscanoccasionallyresultinalossofwell
roundedness.Anartisanorcraftsmanmakesanentireproduct,andisabletohavea
morewellroundedsetofskills.Aspecialist,ontheotherhand,isnotabletomake
thewholeproduct,andassuch,ishighlydependentonothers.Marxpushesthis
ideafurthertoargueforthealienationoftheworkerfromtheproductofhislabor.

31
Smith points this out in Book I, chapter 2.


Itisforthisreasonthatheseemstoendorseanidealofanindividualasa

52

generalist,ratherthanaspecialist.Theseproblemsareimportant,andwillbe
consideredinlaterchapters.Forpresentpurposes,itissufficienttoshowthatthe
divisionoflaborbothsupportsexistingdiversityandpromotesgreaterdiversityin
thefuture.

ComparativeAdvantage
DavidRicardoextendedthedivisionoflaborthemeinhistheoryof
internationaltrade.InRicardostradetheory,hearguedthatthemereexistenceof
differentialproductioninasituationinwhichatleasttwogoodsarebeingtraded
ledtotheopportunityforwealthcreation.Toshowhowthisworks,letusfirst
considertwoterms:absoluteadvantage,andcomparativeadvantage.Imaginetwo
individuals,AliceandBob,bothofwhomproduceshirtsandpants.Letussuppose
thatinadayAlicecanmake12shirtsor9pairsofpants.IfBobmadeonly4shirts
or8pairsofpantsaday,thenAlicewouldhaveabsoluteadvantageoverBobin
makingbothshirtsandpants.IfBobcouldmake5shirtsinadayor15pairsof
pantsinaday,thenAlicewouldhaveabsoluteadvantageinshirtmaking,whileBob
wouldhaveabsoluteadvantageinpantmaking.
ThecaseofabsoluteadvantagewasaccountedforbySmith.Absolute
advantagemerelytakesintoaccountwhoismoreabletoproduceagivengood
giventhecosts(inthiscase,time).Tradetheoriesbasedonabsoluteadvantage


wouldclaimthatBobandAlicewouldtradeonlyinthesecondsituation.

53

Comparativeadvantage,ontheotherhand,arguesforconsideringtheopportunity
costsofproduction,ratherthantheactualcosts.So,evenifAlicehasabsolute
advantageoverBobintheproductionofbothshirtsandpants,Bobhascomparative
advantageintheproductionofpants,asitcostshimlessintermsoftheproduction
ofshirtstomakeapairofpantsthanitdoesAlice.Inthiskindofsituation,Aliceand
Bobcanbenefitfromatradethatgeneratesmorewealth,evenifAliceisbetterable
toproducebothgoods.ThiscomparativeadvantageshouldcauseBobtoproduce
onlypants,whileAliceproducesonlyshirts.Anexchangeratewouldbedetermined
toensurethatbothpartiesaremadebetteroff.Theexchangeratewouldfavor
Alice,butthisisbecauseshealreadyhadanabsoluteadvantage,andisableto
producemore,absentanytrade.ButeventhoughAlicewouldexperiencealittle
morebenefitthanBob,botharestillmadebetteroffthanhadtheynotmadeany
exchange.
Thenotionofcomparativeadvantageisimportantforseveralreasons.First,
justaswithSmith,itdemonstratesthatmaterialbenefitscanbefoundfromthe
existenceofagentdiversity.Italsohelpsfurthertheaccountofhowincentivesexist
tocreatemorediversityandspecialization.Whatisuniqueaboutcomparative
advantage,however,isitsimplicitargumentthatdiversityisinsomesensemore
importantthanabsoluteadvantage.Thatis,itisaremarkabletestamenttothe
powerofthedynamicsofadiversepopulationthatonecanbemadebetteroffby
tradingwithsomeonewhoislessskilledateverything.Agentdiversitybothcreates


theopportunitiesforandencouragesspecialization,andspecializationcreates

54

opportunitiesformaterialgrowth.32
Soweseethattherearealwaysincentivesforagentswhoaredifferentto
takeadvantageofthefactthattheyaredifferent,andmoreimportantly,thereare
incentivesforagentstobecomemoredifferent.Atpresenttime,thiscanbe
understoodtobetakinganexistingcomparativeadvantageandincreasingones
specialization.Thishasaninterestingconsequenceaswell:whilethereareclearly
individualbenefitstoimprovementsinonesspecialization,thereareincentivesfor
individualstohelpothersimprovetheirspecialization,thoughtoalesserdegree.
Thisisbecausetheoverallskillincreasetranslatesintoalargersocialproductfrom
whichallparticipantsgain.Whiletheincentivesaresmallerthanwhattheyarefor
thedevelopmentofonesowntalents,iftheyareviewedintheaggregate,theyare
almostalwayslargerthanthevalueofimprovingonesownskills.Imaginea
talentedbaker,Anne,whoisalsoatalentedbrewer.NextdoortoAnneisCarol,who
enjoysbrewing,butisntverygoodatityet.SinceAnnewishestoenjoyboththe
breadthatshebakesandthebeerthatshebrews,shehasaconundrum:the
opportunitycostofbrewingisveryhigh,asitisdifficulttoacquirebreadofthe
qualitysheiscapableofelsewhere.Ratherthansimplygrowdespondentoverher
predicament,AnnedecidestoinvestinCarolsbrewingendeavors.ItisinAnnes
interesttoinvestherownresourcesinCarolsabilities,ratherthanherown,
becauseheropportunitycostsaremuchhigherthanCarols.IfAnneinvestedinher

32
I say material growth here given the economic context. However, I intend to argue for
the cultural benefits of diversity and specialization later using similar arguments.


ownbrewingskill,shewouldhavetoforegoallthebakingthatshecouldhave

55

doneduringthattime.Soinsteadofhavingtochoosebetweenbreadandbeer,Anne
choosestoinvestinCarolsothatAnnecanendupenjoyingmoreofboth,bymeans
oftrade.Comparativeadvantage,then,canprovideaselfinterestedmotivationfor
investingintheeducationandskillsofothers.
Outsideofmaterialbenefits,however,doeseconomicshaveanyothermeans
ofsupportingtheclaimthattherearebenefitstoagentdiversity?Theepistemic
situationthatagentsfindthemselvesinprovidesanotablesourceofpotential
benefits.Thatis,individualsareexposedtoagreatdealofinformation,muchof
whichissuperfluous,andverylittleofwhichiscomplete.Insomesenses,wehave
fartoomuchinformationtofullyutilizeitall,butinothers,welackagreatdealof
informationthatwouldbeofgreatusetousifweweretohaveaccesstoit.This
epistemicproblemstemsfromboththefactthatwearecognitivelylimited,andthe
factthatourenvironmentisverycomplexandvaried.

EpistemologicalChallengesofComplexEnvironments
Hayek,inhisTheUseofKnowledgeinSocietyarguesthatacentralplanner
wouldbeunabletoproperlymanageaneconomy,becausetheepistemicfailuresof
anyindividualwouldbetoogreat.Whileheframesthediscussionintermsofa
centralplanner,wecanthinkoftheargumentintermsofagenthomogeneity.That
is,ratherthanoneagentmakingdecisions,wehavealotofagentswhohave


homogeneouspreferences,usethesamedecisionmakingprocedure,andwho

56

haveaccesstothesameinformation.Inthiscase,theywillalwaysallmakethe
samedecision.33
Thereareseveraldimensionsalongwhichthishomogeneitycausesa
problem.Iwillarguethatacrosseachdimension,agentdiversityisincentivized.
Hayekfocusesononedimension:knowledgeandinformation.Giventhatthe
environmentweareinisextremelyvaried,knowledgeoflocalconditionscouldbe
extremelyimportant.Forexample,Europeansattemptingtousefarming
techniquesdevelopedinWesternEuropeinsubSaharanAfricafailedduetothefact
thatthelocalgrowingconditionsweresignificantlydifferentfromthoseinEurope.34
Lessseverecasesofthisaretrueingeneralwithsomethingasbasicandimportant
asfarming:rainlevels,soilconditions,accesstotechnologyandinfrastructure,the
localpopulation,localfloraandlocalfaunacanallimpactwhatanappropriate
farmingdecisionis.Individualfarmershaveverygoodreasonstobecomeexperts
intheconditionsoftheirlocalarea,buthavelittletoreasontodevelopsuch
specializedknowledgeaboutotherareas.Eveniftheydidhavesuchaninterest,it
wouldbeanimpossibletask,asthereisfartoomuchinformationtolearn.
Similarly,amayorhasgoodreasontoknowallaboutthepoliticalplayersinhercity
andtosomeextentherstate,buthaslittlereasontomakesucheffortstolearn

33
Of course, this would be false if there were some method of randomization in the
decision-making process.
34
Jared Diamonds Guns, Germs and Steel (1997) provides a great deal of evidence for
the importance of the transferability of knowledge and skills across environments in
determining larger socio-economic outcomes.


57
aboutthepoliticalplayersinothermunicipalities.Atrivialexampleisknowledge
oftheroadsinonestown.Therearegoodreasonstoinvesttimeandenergyin
learningshortcuts,trafficstops,andotherpeculiaritiesoftheroadsthatonetravels
oneveryday,butthereisnoreasontodothatforsomewhereelse.Simplyinvirtue
oftheirphysicallocation,agentsneedtodevelopspecializedlocalknowledge.
Hayekcorrectlyarguesthattherearewaysinwhichinformationcanbe
aggregatedforthebenefitofsocietyasawhole.Inamarketeconomy,theprimary
meansbywhichthishappensisthepricesystem.IfIweretofindoutthatorange
treeshadseverefrosting,IwouldbuyasmuchorangejuiceasIcould,in
anticipationthatthereducedsupplyoforangeswoulddriveupprices.Thisactin
itselfwouldhelpdrivepricesup.Theprice,then,functionsasanaggregate
statementabouttherelativeprosperityoftheorangecropincomparisonwithother
goods.Predictionmarketscandothisforavarietyofareasinwhichwelacksolid
information.Theclassiccaseofthisissportsbetting,butnowthesamemethodsare
appliedtoanumberofareas,includingtheoutcomesofpoliticalelections.People
canprofitfromtheirspecializedknowledgebybettingbeforeothersdo.Bythis
means,themarketincentivesmotivatepeopletomaketheirprivateknowledge
public,andweallbenefitepistemically.
Asecondmajoravenuebywhichacomplexenvironmentincentivizesdiverse
agentpopulationsisbymeansofencouragingavariedapproachtodecisionmaking
procedures.Inamarketexample,ifweconsiderthehugeamountofcomplexity
involvedinthechangesinpricesacrossdifferentmarketsectors,itisinfact


58
impossibleforasingleinvestmentapproachtobesuccessful.Ifeveryonebought
andsoldthesamestocksatthesametime,priceswouldbesubjecttofluctuations
thatwouldresultinhugeextremespricesofsomestockswouldbecomezero,and
otherswouldapproachinfinity.35Modernfinancialmarketshavebeguntoshow
somefeaturesofthis,duetowhatisknownasprogramtrading.Programtradingis
thepracticeofprogrammingacomputertomakeautomatedtradesofseveral
financialinstrumentsatoncebasedoncertainthresholdvalues.Thisstrategyis
heavilyemployedbyhedgefunds,whichoftenhavefairlysimilarthresholdsfor
trades.Thisautomatedprocess,whichmoveshugeamountsofmoneyaroundat
once,createsagreatdealofvolatilityinfinancialmarkets.
Supposeweassumeamuchlessextremeversionofthis,whereonlythe
majorityofinvestorsapplythesamedecisionprocedure,andtheremainingpartof
thepopulationhavetheirowninvestmentstrategies.Inthiscase,theresponsefrom
theremainingportionofthepopulationwouldbetoarbitrageawaythegainsthat
mightbepossibleforthem.Asanexample,imaginethattheChinesestockmarket
werecompletelyclosedtoexternalinvestment.Thoughsomemoreaffluentcitizens
mightbeinvestinginthemarket,thedominantinvestoristhestate.Ifthestates
investmentstrategywerediscernabletootherinvestors,theycouldjust,for
instance,purchasethestocksthatthestatewouldsoonbuy,andsellthestocksthat
thestatewouldsoonsell.Stepsliketheseallowtheotherinvestorstotreatthestate

35
Of course there are practical limitations to infinite prices, as there is not an infinite
money supply. However, if everyone were to buy the same stock at the same time, with
no one willing to sell, the price would be in effect infinity, as there is no market-clearing
price possible.


59
asamoneypump,arbitragingawayanygainsthatthestatemighthavemade.In
general,ifoneemploysaknowninvestmentstrategy,oronethatiscommon,one
opensoneselfuptosucharbitrageopportunity.Marketscannotmovethatis,
maketradesunlesstherearesomanyinvestmentstrategiesbeingemployedat
oncethatanygivenstrategycannotbedistinguishedfromwhitenoise.Thisisa
consequenceoftheNoTradeTheoremineconomics.
WhiletheNoTradeTheoremestablishesthatdiversityisrequiredforbasic
(stock)marketfunction,thereisamoreinterestingsecondclaimthatcanbemade,
thatismoreinlinewithHayeksobservationaboutknowledge.Becausetherearea
greatmanyunknownsaboutthestructureoftheinformationwereceivefromthe
world,itisdifficulttoknowwhetherthedecisionproceduresweareusingareany
good,letaloneoptimal.Thesimplestwayofgettingmoredataonthequalityof
onesdecisionstrategyisbymeansofperformancecomparison.Soifapopulation
employsarangeofdecisionstrategies,itiseasiertoseewhichofthemhavebetter
outcomes.Thepopulationcanthenshifttowardsusingthestrongerdecision
strategies,andthenattemptfurtherrefinements.Thisprocessofimitationand
refinementcanbefruitfullycomparedtoanevolutionarymechanism.Wecansee
thatimitationcanworkonlyinsofarasthereexistsvariationinthepopulation.
Whatsmore,variationmustbecontinuallypresent.Thesevariedstrategieswill
haveanassociatedfitness,basedonhowwelltheydealwiththeunderlying
epistemicsituation.Betterstrategieswillbeselectedforinthepopulation,as


peoplecanthensimplycopywhattheyseeasthemoresuccessfulstrategiesin

60

thepopulation.
Itmayfurtherbethecasethatsomedecisionstrategiesaresimplybetterin
certaincontexts.Thatis,localvariationsintheinformationenvironmentwillfavor
adecisionprocedurethatisnotgloballyfavored.Forexample,environmentsthat
canbecharacterizedashavingmanymicroclimates,suchastheSanFranciscoBay
area,Seattle,orPortlandmightbeonesinwhichitmakesmoresensetoinvestmore
heavilyintoregularlygatheringweatherinformation,whereasthiswouldnotmake
muchsenseinalocationwithafairlystableclimate,suchasLosAngeles.Assuch,
theenvironmentitselfcanbeastrongmotivatorofdecisionmakingdiversity.
Thusfarwehavefoundthatwithineconomicsbothfromamaterial
perspectiveandfromanepistemicperspective,therearesignificantclaimstoboth
thefactofdiversityandthevalueofdiversityevenamongideallyrationalagents.
Thefinalareawewillconsiderfromtheeconomicperspectiveisthatofcollective
action.Inthiscase,wewilltakeintoconsiderationHardinsworkinthisarea.

PublicGoodsandCollectiveAction
InhisCollectiveAction,Hardinprovidesaframeworkforjudginghow
differentgroupssucceedinprovidingpublicgoodsorpreventingpublicbads.For
ourpresentpurposes,wewillonlyneedtoconsiderthebasictenetsofhisaccount.
WecanimagineapopulationofsizePwhopotentiallyhaveaninterestin


provisioningagoodG.Thisgoodhasavaluegitoeachindividualiinthe

61

population.Thereisalsoasinglecostforprovisioningthegood,C.Thiscostisa
fixedamount,independentofthesizeofthepopulation.Duetothenatureofpublic
goods,onecannotpreventothersfromhavingaccesstothegooditiseither
provisionedtoeveryone,oritisprovisionedtonoone.Thegoodisprovisionedif
thereisaminimalsufficientsubgroupKofthepopulationP,forwhomthevaluegiis
greaterthanC/K.36Ifthisisthecase,theneachoftheindividualsinKwouldbe
willingtopayforprovisioningthegoodG.Ifthisisnotthecase,thenthegood
cannotbeprovisionedwithoutsomekindofexternalauthoritythatcouldcoerce
peopleintoprovidingthepublicgood.
Hardinconsidersafewextensionstothisbasicmodelrelatingto
asymmetriesintheagentpopulation.37Heconceivesofasymmetriesasdifferential
benefitstoapublicgood,ordifferentialcosts.Thesetranslatewelltopreference
diversityandskilldiversity,respectively.38Animportantfeatureofthese
asymmetriesisthattheycanmaketheprovisioningofpublicgoodseasierbecause

36
This further depends on the size of K itself. I refer to K here as the minimal subgroup
of P that meets the constraints described. Hardin has a detailed account of the importance
of smaller Ks for successful provisioning of public goods. Here all we need to rely on is
that smaller Ks are better able to successfully provision public goods than are larger Ks.
This argument is developed in detail in Olsens The Logic of Collective Action, and is
also refined in Hardins Collective Action, particularly in chapter 3.
37
See pages 68, 111, and 119.
38
Note that there is a difficulty in directly translating differences in costs to differences in
skills, if we assume that individuals have a declining marginal utility for money. In this
case, costs can be lowered either by means of the possession of some skill or by
possessing a large endowment of money. Insofar as money can be turned into a
technological advantage, or just a direct purchase of the good in question, it can lower
costs the same way a skill can.


ofhowtheycanaffecttheminimalsizeofK.Takethecaseofpreference

62

diversity.Imaginetwopopulations,PandP.InpopulationP,everyonehasthe
samesetofpreferences,andassucheveryonehasthesamevaluationforthe
provisioningofagoodG.InpopulationP,however,agentshavedifferent
preferences,butwewillassumethatthemean39ofthesepreferencesinrelationto
goodGisthesameasthemeaninP.Sincepreferencesarediverse,weknowthat
theremustbeindividualswhohavevaluationsgreaterthanthemean.Justwiththis
information,wecanestablishthat,foranygivenKwhereCard(K)>2,thesizeofK
ofPcanalwaysbesmallerthanK.Ifwehaveadditionalinformationaboutthe
distributionofpreferences,wecanmakemoreaccuratestatementsabouthow
muchsmallerKwillbe.40
Hardinnotesthatasingleasymmetricpopulationislikelytobeabletofind
supportforanumberofpublicgoodssimultaneously,sinceindividualscanbe
highlymotivatedacrossmanydifferentdimensionsofapreferencespace.41Italso
createsopportunitiesforsomeamountoflogrolling,inwhichdifferentsubgroups
cantradesupportfortheircausesinordertoachieveasufficientnumberof
contributors.Itshouldbenotedthatwecanapplysimilarargumentstothecaseof
skilldiversityinsofarassomeskillscanlowertheeffectivecostofprovisioninga
givenpublicgood,evenifweholdpreferencesconstantinapopulation,the
differencesinskillscreateopportunitiesforsmallerKs,sincetherewillexist

39
This argument would also work if we employed the median rather than the mean.
40
On page 68, Hardin speculates that the nature of, say, political views is such that the
distribution will be sufficiently skewed to encourage the creation of smaller Ks.
41
See page 119 for his discussion.


subgroupsforwhomtheeffectivecostofprovisioningthegoodislowerthanit

63

wouldhavebeenhadeveryonesharedthesameskillset.Theselowercosts
increasethelikelihoodthatthegoodcanbeprovisioned.
Amoredifficultquestiontoaskiswhathappenswhenapopulationhas
diversityofbothskillsandpreferencesatthesametime.Thiswill,toalargeextent,
dependonthecovarianceofskillsandpreferences.Therearethreeclassesofcases
thatareinterestingtoconsider:positivecorrelation,independence,andnegative
correlation.Inthecaseofpositivecorrelation,wefindthatindividualswhohavea
greaterpreferenceforagoodGalsohaveskillsthatwouldlowerthecostof
providingthegood.Herewecanimaginethatpeoplewholikepublicgardensmight
alsohaveagreaterabilitytogardenthemselves,orthosewhowouldlikea
communityorchestraarealsothosethatarethemselvesmusicians.Wecanseehow
otherpublicgoodscanhavesimilarlymotivatedindividuals,likeinthecaseofa
firehouse,acommunitywatchgroup,andlifeguardsatthebeach.Becauseofthis
positivecovariance,weshouldexpecttheminimalKtobeevensmaller,since
perceivedbenefitsarelargerandcostsarelower.
Inthecaseofindependence,weeffectivelyreduceourselvestotheprevious
considerationofvaryingonlyoneparameteratatime.Sinceskillsandpreferences
areindependent,weshouldexpectminimalKstoberoughlythesamesizeasthe
casesinwhichonlyoneparameterisvaried,thoughperhapsmorelikely.Thereare
threewaysonecanconsiderthecompositionofthemembersofasubgroupK:they
couldallbehigherskilled,theycouldallbemoreinterestedinthegood,orthey


64
couldbeamixtureofthetwo.Inthefirstcase,wecanimaginethatengineersare
morecapableofprovidingpublicgoodslikelevees,eveniftheydonthaveany
specialinterestintheconstructionofanyparticularlevee.Aswithanycaseofthose
whoaremoreskilledbutperhapslessinterestedinprovisioningthegood,
opportunitiesforlogrollingareavailable.Thatis,thosemaytradetheireffortsto
supportonepublicgoodinordertogethelponanothergood.Inthisexample,
perhapsengineersasagrouphaveagreaterinterestinhavingrecyclingprograms.
Theycantradetheirtalentsinleveeconstructionforsupportofrecyclingcenters.
Thesecondcasetoconsiderhereisagroupwithskillsnodifferentfromtherestof
thepopulation,butwithagreatinterestinprovidingthegood.Anationalarmy
seemstobeacandidateforthis.Soldiersappeartoplaceagreatdealofvalueonthe
kindofservicetheyperformfortheirnation.Eventhoughthereareplentyofable
bodied,andevenathleticpeoplethatarenotsoldiersthemselves,theydonotplace
thesameamountofvalueontheservice.Thefinalcasetoconsiderhereisthatofa
Kwithamixedpopulation.Thislikelycoversabroadrangeofcases,butone
importantexampleisthatofthesystemofpublicdefendersandprosecutorsinthe
legalsystem.Somepublicdefendersdoitbecauseitisthevirtuousthingtodo,and
somemightdoitasacareersteppingstoneitisameanstodemonstratetheirlegal
skills,andbetterpositionthemforeitherlucrativejobsinprivatepracticeor
considerationforajudgeship.Thoughtheindividualsmighthavedifferent
motivations,asawholetheyenablethepublicgoodtoexistandtopersistthrough
time.

65
Inthefinalcasethatwecanconsider,wecouldhavenegativecorrelation

ofskillsandpreferences.Inthiscase,weareforcedintothelogrollingscenario:
individualswhohavethemostabilitytoachievesomeendX,butdonotwantto
achieveit,cantradetheirabilitiestoagroupwhowantsXbutcannotprovisionit
themselves,inexchangefortheirabilitytoprovisionY.Howwellthisscenario
worksisdependentontheparticularutilitycalculationsofthedifferentcohortsof
individuals.Individualswillbewillingtotradeskillsonlysolongasthetrade
remainsutilityincreasing.Thatis,thepositiveutilitygainedfromgettingtheir
publicgoodmustoverwhelmthepotentiallynegativeutilityofworkingtoachieve
someotherpublicgoodthattheirskillsethappenstobesuitedto.Thisisfurther
complicatedbywhetherthereexistopportunitiesforshirkingifthetradeis
sequential,oronegoodrequiressomeamountofmaintenance,whiletheotherhas
onlyonetimestartupcosts.Analysisofthesesituationsislefttolaterchapters.
Thisanalysisofasymmetryincollectiveactionmakesasignificantclaim
aboutthevalueofdiversityineithereconomicorpoliticalcontexts.Agentsinmore
diversepopulationsaremorelikelytohavemorepublicgoodsprovisioned,evenif
theirmeanpreferencesandtheirmeanskillsareidenticaltoahomogeneous
population.Morepublicgoodsprovisioningresultsinhigheraverageutility,andso
providesmaterialbenefitstotheagents.Thelastcaseconsidered,inwhicha
diversepopulationhadanegativecorrelationbetweenskillsandpreferences,isby
farthemostinterestingtoconsider,butittellsapotentiallymixedstory.Inthis
case,itisnolongerclearthateachindividualsutilityhasincreased,ifthereare


manypublicgoodsandmanyskills,assomepublicgoodsmaybeviewedas

66

publicbadsbyportionsofthepopulation.Anexampleofthiscouldbethe
provisioningofatownbelltoannouncethefiveprayertimesforaMuslim
community,whichmaybeviewedasanuisancebyanonMuslimportionofthe
population.Sincelogrollingandtradescouldbesequencesofbilateralagreements
inapopulationwithmorethantwogroupsofindividuals,itcouldbethecasethat
somebilateraltradesmighthavenegativeexternalitiesforthirdparties.Atthe
sametime,however,itisequallypossiblethattherearepositiveexternalities.
However,thecomplexityofthisparticularcaseshouldnotdistractfromthegeneral
resultthatmorepublicgoodscanbeprovisioned,andinmostcases,thiswill
generatematerialbenefitsforthewholepopulation,thoughtheamountofbenefit
willvaryacrossthepopulation.

SummaryofEconomicPerspectives
Whatwehaveseensofaristhatinfourdifferentcasesineconomictheory,
wehaveargumentsthatsupportthevalueofdiversity.Inourbeginning
investigationofeconomics,wenotedthatdiversityisrequiredfortheexistenceofa
market.FromSmith,wefindthatthedivisionoflaborbothcreatesadditional
materialbenefits,andprovidesincentivesforagentstodifferentiatetheirskillsets.
Thisprovidesdirectsupportforthevalueofskilldiversity,aswellassomeevidence
foritsinevitability.WefoundthatRicardofurtherdevelopsthisthreadbymeansof


67
thenotionofcomparativeadvantage.Combiningcomparativeadvantagewiththe
abilitytotradeincreasesthevalueofskilldiversityevenfurther,aswefindthat
evenifoneagentsskillsarecompletelydominatedbyanothersskills,therearestill
opportunitiesfortradethatmakebothbetteroff.OurdiscussionofHayekshowed
thatthereareepistemicbenefitstoskilldiversity,knowledgediversityand
perspectivediversity.Finally,ananalysisofcollectiveactiondemonstratedthat
preferencediversityandskilldiversitycanprovidematerialbenefitstoboth
individualsandapopulationasawhole.
Thisinvestigationallowsustoseethatapopulationofrationalagentswill
notonlyalwaysbediverse,butwillfrequentlyhavereasonstobecomemore
diverse.Insofarasthesekindsofdiversityarecorrelatedwithdifferencesin
perspectivesonthenatureofthepublicsphere,ananalysisofthepublicspherewill
havetorecognizetheseperspectives,orheavilyrestricttheprivatespheresothatit
toobecomeshomogeneous.Ifwearetoretainatleastsomeofthegoalsofthe
liberalpoliticaltradition,andwanttomaintainarelativelyfreeprivatesphere,then
weareforcedtoacknowledgethelegitimacyofdiversityinthepublicsphereas
well.Ajustsocialorderwillrelyonunderstandingthenatureoftheworldthatwe
findourselvesin,understandingofournatureashumans,understandinghowwe
areabletointeractwitheachother,anddiscoveringmechanismstoachievethe
endsthatwesetforourselves.Thisiscontingentontoomanyfactorsthatare
discoverableonlybyrationaldeliberationinformedbysocialexperimentation.As


68
thischapterdemonstrates,thepresenceofandthoughtfulleveragingofdiversity
helpsmakethesechallengesmoremanageable.

69

Chapter2:
TheViewFromEverywhere

Introduction

Afundamentalquestiontoask,andindeedthefirstquestiontoask,inmoral
andpoliticalreasoningisfromwhatperspectiveweoughttobereasoning.There
aretwoprominentmoralstancesthatareadoptedinanswertothisquestion.Nagel
hasfamouslycalledthemoreprevalentstancetheviewfromnowhere.This
positionsupposesthatmoralreasoningoughttobedonefromacompletelyneutral
position,whereparticularfeaturesofindividuals,includingtheirparticularwants
anddesires,areeliminatedfromtherealmofreasoning.Thisstancewasadopted
byRawls,andIwouldclaimthatitisthedominantapproachincontemporarymoral
andpoliticalphilosophy.Thelesscommon,butstillhistoricallyimportantstanceis
whatIwillcallTheviewfromsomewhere.Thisviewsupposesthatratherthan
eliminateallperspectivesinordertoachieveneutralityinourreasoning,weought
tochooseaparticularperspective,andreasonfromthere.Prominenttraditionsof
theviewfromsomewherearefoundinAristotle,Hobbes,andneoHobbesianssuch
asGauthier.IntheneoHobbesiantradition,thechosenperspectiveisfirst
personal:eachindividualoughttoreasonaboutwhatwouldmakethatpersonbest
off,recognizingthatothersaretryingtodothesamething.Inthisperspective,
individualsareindifferenttootherpeopleswellbeing,andcareaboutothersonly
insofarastheyhelporobstructthemintryingtoachievetheirends.

70
Thesepositionshavedominatedthedebateoverwhatourmoral

perspectiveoughttobe,buttheydonotexhaustthepossibilities.Notablyabsent
fromthisdebateistheotherpossibleperspective,Theviewfromeverywhere.Itis
thisviewthatIaimtodefendhere.Theviewfromeverywhereisbotharejectionof
thetwoalternativeviewsandanattempttoincorporatetheirkeyvirtues:namely,
thegoalofneutralityfoundintheviewfromnowhere,andtherecognitionofthe
importanceofparticularinterestsanddesiresfoundintheviewfromsomewhere.
ItisbecauseofthiscombinationoftheirvirtuesthatIwillarguethattheviewfrom
everywhereisinanepistemicallysuperiorpositioncomparedtotheotherchoices
ofperspective.
Beforewecanseewhattheviewfromeverywhereis,however,wemustfirst
seewhatitisnt.Tothatend,wewillfirstlookattheviewfromnowhereasRawls
andNageluseit,andthenlookattheviewfromsomewhereasGauthierusesit.The
goalineachsectionistounderstandbothwhateachperspectiverequiresandwhat
eachperspectiveprovidestothereasoner.Oncewehaveseentheoutlinesofeach,
wearetheninapositiontoseewhatisdifferentabouttheviewfromeverywhere.
Armedwiththisknowledge,wecanthenseehowtheviewfromeverywhere
providesasuperiorperspectiveformoralreasoning.

71

TheViewFromNowhere
Theviewfromnowhereisanattempttoobtainapurelyobjective
perspective.Thisobjectiveperspectiveisobtainedonlybysheddingparticularities
ofoursubjectiveselvesourparticularcapacities,perspectives,andrelationships.
AsNagelsaysinhisopeningparagraphstoTheViewFromNowhere,wemaythink
ofrealityasasetofconcentricspheres,progressivelyrevealedaswedetach
graduallyfromthecontingenciesoftheself.42Thisgetsattwocrucialfeaturesof
hissystem.Namely,heisamoralrealist,andhebelievesthatifwearetodiscover
objectivereality,wemustlooktoourobjectiveselvesratherthanoursubjective
selves.
Nageldoesindeedholdthatwehavebothobjectiveselvesandsubjective
selvesoursubjectiveselveshavethefeaturesthatwepretheoreticallyassociate
withourselves,suchasparticularwantsanddesires,viewsoftheworld,and
interpersonalrelationships.Ourobjectiveselves,ontheotherhand,arewithout
individualperspectives.Weeachhaveourownobjectiveself,whichismeanttobe
us,butstrippedofallofoursubjectivefeatures,includingasenseofwhatitisliketo
beourselves.Theobjectiveselfthusseestheworldfromaviewthatisoutsideof
theworldnowhere.Oursubjectiveselvesareinthisworld,butourobjective

42
T. Nagel, The View From Nowhere, Oxford University Press, 1986 , p. 5


72
selvesintheviewfromnowherehavenoparticularattachmenttooursubjective
selves.
Togetatmoraltruths,Nagelbelievesthatweoughttoreasonfromtheview
fromnowhere.Afterall,inthisview,wehaveeliminatedthesubjectivitythatcan
cloudourvision.Fromthisobjectivepointofview,wenotonlyensurethatthe
outcomeofourmoraldeliberationwillbeobjective,butthatitwillberesponsiveto
objectivefeaturesoftheworld.Thisisnottosay,however,thatthisobjective
perspectivedoesnotseeandattempttorespondtodifferencesandparticularitiesin
individuals.Itissimplythattheseparticularitiesdonotalsocomewiththeforceof
asubjectiveexperience.
Thispositioniscontrastedwithadifferentviewfromnowhere,which
Rawls'sworkembodies.Rawlsmovesawayfromanymetaphysicalclaims,and
insteadoptsforaprocedurethatsimilarlyaimstostripawaythosefeaturesof
ourselvesthatmightbiasusinourprocessofmoralreasoning.43Theprocedure
RawlshasinmindisdeliberationintheOriginalPosition,behindtheveilof
ignorance.
Behindtheveilofignorance,one(orone'sdelegate)isselfinterested,but
doesnotknowanyparticularfeaturesofone'sself.Reasoningbehindtheveilof
ignoranceisinasenseamorestringentconditionthanwhatNagelproposes.This

43
Indeed, while Nagel's view from nowhere is theoretical philosophy, Rawls' Original
Position is meant as a piece of practical philosophy. While these views have substantive
differences between them, I lump them together to show that they have important
common features and suffer from the same epistemic problems.


reasoningisfromapositionofcompleteignorance:ignoranceofoneself,ones

73

placeinsociety,andindeed,themakeupandnatureofthatsociety.Individualsin
thispositionarerational,butdonotknowwhattheirendsmightbe.Asthe
individualsintheOriginalPositionattempttodeterminewhichrulesoughtto
governthebasicstructureofsociety,theirrationalityinthefaceofignoranceleads
themtochoosethoserulesthatensuretheworstoffinsocietyismadeasbestoffas
possible.
NowthatwehaveseenbriefaccountsofbothNagel'sviewfromnowhere
andRawls'originalposition,wecanconsiderwhatthecommonfeaturesarethatare
desirableforamoralstancetohave.Mostnotably,theybothembodyaperspective
ofneutrality.Reasoningfromtheseperspectivesdoesnotallowonetoprivilege
oneselfasmoreimportantthanothersinmoraldeliberation.Thisisdone
procedurallytheresimplyisn'tanythingthatagentsinthesemoralstancescould
usetobecomebiased.Theproceduresthatareintroducedbothaimtoremoveany
informationthatismorallyirrelevant.
Whilethispositionofneutralityisaverydesirablequalityinamoralstance,
theirapproachtoachievingitpresentstwopotentialepistemicchallenges.Thefirst
worryisthatitisunclearthatweknowhowtoreasonfromsuchaperspectiveless
perspective.Sincewehavealwaysviewedtheworldfromourownperspectives,we
mustnotonlybeabletoevaluateanoveldecisionscenario,butdosofromanalien
perspective.Thecommontechniqueforattemptingtoreasonfromsucha
perspectiveistousecounterfactuals.Thatis,weengageinthoughtexperiments


74
wherewetakeourselvestohaveadoptedthedesiredperspectivelessstance.Itis
withintheconfinesofthesethoughtexperimentsthatweattempttodeterminethe
consequencesoftheseperspectivesthatwecannotourselvestakeon.
JustasNagelthinkswecannotknowwhatitisliketobeabat,andRawls
thinksthattheUtilitarianneedforaninterpersonalcomparisonofutilitiesis
doomed,Isuspectthatwecannotknowwhatitisliketoreasonwithoutourown
perspective.44Thisisdifferentinkindthansomeotherthoughtexperiments:inthe
famousTrolleyProblem,peopletendtohaveclearreactionstothescenarios.
Thoughfewpeoplehaveeverhadtodecidewhethertokillonepersontosavefive
others,thescenariosaskustoreasonaswenormallywould.Sothoughthe
situationpresentedisnovel,themethodsbywhichwearemeanttoevaluateitare
thesameaswhatweuseinoureverydaylives.Itisthisfeaturethatmakesthe
trolleycase,andotherscenarioslikeit,sousefulinmoralreasoning.Theexamples
canbedesignedtoputusinnewsituationsthatmightrevealsomethingabouthow
weoughttothinkaboutsituationsthataremoretypical.Buttheyallassumethat

44
It might be contended that mathematics or the hard sciences embody precisely this kind
of objective, perspectiveless perspective. Features of individual scientists do not
matter in their work, and so they just drop out, and all that is left is the science itself. I
will deal with this more substantively later in this chapter, but I will note here that this
seems to be a false characterization of the practice of science and mathematics. Feminist
critiques have amply demonstrated that even the hard, data-driven sciences show
significant evidence of bias. Even mathematics is in an important sense deeply
perspective-laden: there are often multiple ways to prove the same theorem the way that
is chosen can be based on notions of elegance, utility, concision, explanatory power, or
any number of other criteria. This is readily evident when one compares how a human
proves theorems to how a computer proves theorems the approaches are drastically
different, even though humans programmed the computer. Computers are not reliant on
the semantics of axiom systems, and so consider portions of the solution space that
humans ignore, and as a result often find novel proofs.


wereasonasourselves.Thisisthekeydifferencebetweensituational

75

counterfactualsandperspectivalcounterfactuals.Situationalcounterfactuals
remaingroundedinourexperiencepreciselybecauseweusethesameperspective
thatwenormallywouldinanovelsituation.Perspectivalcounterfactuals,onthe
otherhand,offernosuchgrounding.Evenifthesituationhasmanyexternal
featureswithwhichwearefamiliar,wecannotaccountforhowtorelateour
experiencestoanovelperspective.Thisisduetothefactthatourexperiencesare
mediatedbyourperspectives.ThoughIhavespentagreatdealoftimein
Philadelphia,Ihavenosenseofhowasquirrelwouldexperiencethesameplace.I
maybeabletoimaginemyselfasasquirrel,andevengetagreatdealofthekey
featuresofsuchanexperiencecorrect.Itcouldevenbeaveryvividthought
experiment.Theworryis,however,thatthesekindsofthoughtexperiments
generatevividfalsepositives:thesethoughtexperimentsmightfeelreallyright,but
thisfeelingwouldbethesamewhetherornotthethoughtexperimentwascarried
outwell.Imayhaveleftoutsomekeyfeatureofbeingasquirrel,andwouldnever
knowit,becausemythoughtexperimentappearedtobesuccessful.Thisposesa
basicchallengetotheclaimthatweareabletoconductthethoughtexperiments
thatNagelandRawlspropose,whichwouldinhibitourabilitytoadopttheirmoral
stancesatall.
Letussupposethatwecanknowwhatitisliketoreasonwithouta
perspective(oratleast,fromaforeignperspective),whichwouldputourfirst
worrytorest.Westillhavetodealwithoursecondepistemicworry.Namely,how


doweknowwhenwehaveridourselvesofourownperspectivesand

76

successfullyadoptedthisperspectivelessperspective?Thestartingpointisclear
enoughwestartwithourselvesasweare.Thenweshedthosefeaturesof
ourselvesthataresubjective,untilallweareleftwitharetheobjective,universal
aspectsofourselves.Itisthissecondstepthatposesachallenge.Imaginethatyou
aretryingtoeliminatethosefeaturesthataresubjective(ormoregenerally,morally
irrelevant).WhatNagelproposesistorunthroughallofyourpersonal
commitments,relationships,beliefsaboutyourselfandyoursurroundings,and
decide,foreachone,whetherornotitismorallyrelevant.Ifitis,youkeepthis
attribute.Ifitisnot,youthrowitaway.Buttocarryoutthisprocedure,youmust
alreadyknowwhatyouwanttokeep.45Moretroubling,however,isthatyoudonot
haveawayofdeterminingifyouhavecarriedouttheprocedureappropriately.
Peoplemakemistakes,thereneedstosomewayofcheckingtoseehowwellyoudid
incarryingouttheprocedure.Butthereisnoindependentstandardavailableto
checkagainst.OnemaysuggestturningtoReflectiveEquilibriumtosolvethis
problem.ButReflectiveEquilibriumiscalibratedagainstone'sowntheoretical
commitmentsandmoraljudgmentsthesearepreciselythethingsinquestion.So
itcan'tprovideanindependentstandard.Thus,itseemsthatwhileyoucantryto
carryouttheprocedureforeliminatingmorallyirrelevantfeaturesofone's
perspective,thereisnowaytoknowiftheprocedurehasbeensuccessful.This

45
A procedure of this kind would rely on the idea that the view from nowhere is just a
restriction on what information one has access to. But the view from nowhere is not
merely a restriction of information, but an alternative perspective. (See upcoming
discussion of perspectives.) In this sense, then, the problem is much more difficult than is
being expressed.


suggeststhatwecannotrelyonreasoningthatusessuchaprocedureasits

77

standardforjustification.
HabermasraisesasimilarchallengetoRawlsinthefamousRawlsHabermas
debateintheJournalofPhilosophy:
Thingsaredifferentwhentheveilofignoranceconstrainsfromthe
beginningthefieldofvisionofpartiesintheoriginalpositiontothebasic
principlesonwhichpresumptivelyfreeandequalcitizenswouldagree,
notwithstandingtheirdivergentunderstandingsofselfandworld.Itis
importanttoseethatwiththisinitialabstractionRawlsacceptsadouble
burdenofproof.Theveilofignorancemustextendtoallparticular
viewpointsandintereststhatcouldimpairanimpartialjudgment;atthe
sametime,itmayextendonlytosuchnormativemattersascanbe
disqualifiedwithoutfurtheradoascandidatesforthecommongoodtobe
acceptedbyfreeandequalcitizens.Thissecondconditionmakesademand
onthetheorythatisdifficulttomeet,asisshownbyabriefreflection.
Followingthejustificationoftheprinciplesofjustice,theveilofignoranceis
graduallyraisedatthesuccessivestepsofframingtheconstitution,of
legislation,andofapplyinglaw.Sincethenewinformationthatthereby
flowsinmustharmonizewiththebasicprinciplesalreadyselectedunder
conditionsofinformationalconstraint,unpleasantsurprisesmustbe
avoided.Ifwearetoensurethatnodiscrepanciesarise,wemustconstruct
theoriginalpositionalreadywithknowledge,andevenforesight,ofallthe
normativecontentsthatcouldpotentiallynourishthesharedself
understandingoffreeandequalcitizensinthefuture.Inotherwords,the
theoreticianhimselfwouldhavetoshouldertheburdenofanticipatingat
leastpartsoftheinformationofwhichhepreviouslyrelievedthepartiesin
theoriginalposition!Theimpartialityofjudgmentwouldbeguaranteedin
theoriginalpositiononlyifthebasicnormativeconceptsemployedinits
constructionthoseofthepoliticallyautonomouscitizen,offaircooperation,
andofawellorderedsociety,inthespecificsenseRawlsattachestothese
termscouldwithstandrevisioninlightofmorallysignificantfuture
experiencesandlearningprocesses.(Habermas,p.118)

WhatIwanttofocusonwiththispassageishispointthatthetheoretician
musthavealreadybuiltinallofthenormativecontentsintothemoralstanceinthe


78
46
firstplace. Thesechallengesechomysecondepistemicchallenge:thatwehave
toalreadyknowwhatallofthemorallyrelevantfactsarebeforeweevenbeginto
engageinourmoraldeliberation.TheOriginalPosition,then,isnotaveryneutral
positiontotakeinstead,itplacesheavyconstraintsonwhattheoutcomescould
be,independentofthequestionsofjusticeatstake.Andwhiletheoutcomeisagood
one,thissuggeststhattheprocedureitselfisn'tdoingthework,butratherwhat
wentintosettinguptheinitialconditionsfortheprocedurecarryasignificant
burden.Iraisethisasaconcernpreciselybecauseitreliesonthetheoretician
alreadyknowingthemorallyrelevantfactstoderivetheresult.Itisnotaprocessof
discovery,butratherakindofjustification.47
ThebasicworrywiththeviewfromnowhereapproachthatItaketobe
embodiedbybothNagelandRawlsisthatthispresupposesthatwe(whetherwe
aretheoreticiansortheagentsintheirthoughtexperiments)wouldhavetoknowso
muchaboutnormativeconceptsandfactsbeforeweevenbeginthattheviewitself
doesnotdoanywork.Whilethegoalofneutrality(andinthecaseofNagel,

46
In his Rawls on Justice, Nagel makes a similar point: namely, that the Original
Position is far from neutral or even objective in its construction, as it has a heavy
individualistic bias. The allocation of primary goods themselves can also be non-neutral,
depending on the actual life plans of citizens. Primary goods are not equally valuable
across various conceptions of the good. (Nagel, 228) Rawls would contest the final point
by noting that the set of primary goods are meant to enable any conception of the good,
rather than necessarily be taken advantage of by all conceptions of the good.
47
I do not think that Rawls would disagree with the claim that the original position is
meant to be a justification rather than a discovery process, but I do think it is relevant to
how useful it is as a moral stance. Habermas has another critique of this process in
(Habermas, p. 128).


79
objectivity)islaudable,themethodbywhichtheyattempttoachieveitishighly
questionable.
Nowthatwehavelookedattheviewfromnowhere,wecanturntoitsmost
significantrival,whichIwillcalltheviewfromsomewhere.

TheViewFromSomewhere
Thecharacteristicdifferencebetweentheviewfromnowhereandtheview
fromsomewhereisthat,whiletheformeraimstoachieveneutralitybetween
individualpreferencesbyadoptingastancethatisoutsidetheworldof
preferences,thelattermakesnoparticularattemptatachievingneutrality,and
insteadadoptsaperspectivethatrespondstoparticularpreferencesandbeliefs.
Forthepurposesofdiscussionhere,Iwillconsideraparticularstrainoftheview
fromsomewhere,whichisrationalegoism.Ichooserationalegoismherebecause
withinsocialcontracttheory,itistheprimaryalternativetoaRawlsianproject.48
Withinthewiderrealmofethics,however,theviewfromsomewhereappliesto
muchmorethanjustrationalegoism.Aristotelianvirtueethics,forexample,would
fallundertheclassificationofaviewfromsomewhere.NothingIsayhererelieson
featuresfoundonlyinrationalegoism,andshouldbereadasaddressingalltheories
thatfallundertheviewfromsomewhere.

48
Utilitarianism does not neatly fit into the divisions I suggest. Of the extant moral
theories, it most closely shares some characteristics with my own view in the sense that it
is both impartial between agents and interested in the subjective interests of the agents
themselves. But it differs from my theory in how it aggregates those subjective interests.

80
Theprimaryfeatureofaviewfromsomewhereisthatitabandonsthegoal

ofneutrality.Ratherthantrytoshedallofone'sparticularbeliefsandattitudesin
ordertoattainanimpartialperspective,theviewfromsomewhereinsteadargues
thatoneshouldviewmoralityfromone'sownperspective.Afterall,onecan
reasonablysupposethattheirownexperiences,beliefsanddesireshavehelped
shapetheirmoralattitudes.Thosetheoriesthatputcharacterasacentralfeatureof
whatitistobemoralmusteliminateimpartiality,preciselybecauseonesown
motivationsanddispositionsarecentraltosuchaconceptionofmorality.Ineither
thecaseofarationalegoistorofavirtuetheorist,onesowninterestsand
dispositionsaretakentobeinformativeofthecorrectthingtodo.Sonotonlyis
impartialitynotachieved,butitisactivelydismissed,duetothefactthatitwould
requireasignificantlossofmorallyrelevantinformation.Thefirstperson
perspectiveisthusisshapedbyone'sowninterests,relationships,andbeliefs.This
perspectiveiswellknowntotheagent,andassuch,itisclearhowonereasonsfrom
it.Thus,thereisatradeoffbetweenneutralityandepistemicfeasibility.
ThistradeoffcanbeseeninGauthier'sMoralsbyAgreement.Themoral
stancedevelopedthereiscompletelyfromtheperspectiveoftheindividualagent.
Gauthierarguesthatmoralityfollowsfromindividualrationality.Neutralityis
completelyabandoned.Infact,Gauthier'stheoryrequiresthatagentsbeindifferent
towardseachother:theirutilityiscompletelydeterminedbythesatisfactionof
theirowndesires.49Thisconstraintdoesnotstemfromamoralpointofview,but

49
The conception of non-Tuism is developed in (Gauthier, p.311).


81
moreofanepistemicone:neitherthesadistnorthealtruisthasreliableepistemic
accesstotheexperiencesofpleasureandpainofothers,sotheyarenotableto
countthoseinaccessibleexperiencesaspartoftheirutility.Thislimitationdoesnot
meanthatotherindividualsdonotweighonone'smoralconsiderations,however.
Whilewemaynotbeabletoascertaintheparticularutilitiesthatothersderivefrom
particularactionsofours,weareabletoseehowtheyrespondtothoseactions.So
thesadistcantakepleasureinpeople'sscreams,orthealtruistcantakepleasurein
expressionsofhappiness.Thisexternalmanifestationofapparentinternalattitudes
isreadilyavailabletoanyone,andsocanbeusedinutilitycalculations.Agentsare
thusabletoreasonasiftheywereRobinsonCrusoe,whetherornottheyarealone.
Thisisconsistentwiththeviewfromsomewhere'sinterestinepistemicfeasibility.
WefindasimilarsituationwithAristotelianvirtueethics.Insofarasour
friendshipswithothersareimportantforthedevelopmentofourownvirtue,we
mustlooktoseewhichofthosefriendsactinwaysthatdemonstratetheirown
virtue.Associatingwithpeoplewhohavenoticeablecharacterdeficienciesmakesit
moredifficultforustomaintainourowncharacter.Similarly,onemustbegenerous
totherightkindsofpeople,andshowrespecttothosethataredeservingofit.
Onthisview,neutralityisnotonlylost,butwholeheartedlyabandoned:
agentsactonlyfromapositionoftheirowninterests(whethertheyaredefinedby
utilityorattainmentofvirtues),andcareaboutothersonlyinsofarastheyhelpor
hindertheirabilitytoattainthoseinterests.Inneutrality'splace,wehaveamuch
moreplausiblesetofepistemicconstraints.Theviewisinsomesenselessdesirable


becauseofthelackoftraditionalmoralconcernsthatmightbeembodiedby

82

neutrality,butitisfarmorefeasibletoachieveinthesensethatitdoesnotplaceany
unrealisticepistemicburdensoneitherthetheoreticianortheagentsthemselves.
Inoursurveyofthetwomostprominentchoicesformoralstances,wefind
thattheyhaverepresentthetwowaysofrespondingtoatradeoffbetween
neutralityandepistemicfeasibility.Whiletheviewfromnowherecanmakeaclaim
toneutrality,itcanonlydosobymeansofheavyepistemicdemands.Theview
fromsomewhereisastancewithfairlystraightforwardepistemicrequirements,but
itlacksneutrality.Whilethistradeoffseemsinescapable,giventhatneutrality
seemstorequireadoptingaviewpointthatisnotnaturallyavailabletous,Ipropose
analternativemoralstancethatsidestepsthistradeoff,andallowsustoattainboth
neutralityandepistemicfeasibility.

TheViewFromEverywhere
Theviewfromeverywhereisanattempttoreconcilethepositivefeaturesof
boththeviewfromnowhereandtheviewfromsomewhere.Inparticular,itaimsto
achievetheneutralityoftheviewfromeverywherewhilemaintainingtheepistemic
feasibilityoftheviewfromsomewhere.Todothisrequiresasignificantdeparture
frombothviews:mostnotablyinthatthemoralpoliticalstanceisnotsomethingto
betakenonbyaloneindividual.Instead,itisfundamentallyanaggregativesocial
viewpoint.Inaddition,theviewaimstobeaprojectofdiscoveryratherthan


83
justification.Ratherthanbuildinandjustifysubstantivemoralclaims,theviewis
designedtoallowtheseclaimstoarisefromcollectivedeliberation.
Thebasisofthisviewistheconvictionthatindividualbeliefsandvalues
matterformoralreasoning,butonlyinsofarastheyprovidesupportforwhatmoral
principlesweoughttoadopt.Payingattentiontoindividualbeliefsratherthan
universalsmakesmoraldisagreementmuchmorelikely,whichonthefaceofit
wouldsuggestthatthisisaproblem.However,itisimportanttonotethat
individualsnotonlycanhavedifferentbeliefs,butalsodifferentperspectives.These
differentperspectivesareanessentialaspectoftheviewfromeverywhere:where
diversebeliefsmightcausedisagreement,diverseperspectivesprovidethe
opportunitytounderstandwherethesebeliefscomefrom,andhowtheymightbe
combinedtodetermineananswertothatdisagreement.
Thusfarwehaveseenthattheviewfromeverywherecanbeseenasasearch
processthathasthreebasicelements.First,wehaveacollectionofeveryones
beliefsanddesires.Theseserveasthesupportforourprocedure.Thereisaworry,
however,thatbeliefsareonmuchweakergroundthanthekindofevidence
gatheredinthehardsciences.Iclaimthatbeliefs,thoughpotentiallyunreliable,are
thebestindirectmeasureforevaluatingparticularmoralprinciples.Consideracase
offairdivision.AliceandBobaretryingtodeterminehowtodivide12avocados
and12grapefruit.Wemayhaveseveralevaluativebeliefsaboutthiscase,suchas
whatwouldconstituteanequitabledivision,butthesebeliefsdonotstandontheir
own.Theyareinsteadinformedbyfactualbeliefs.Forinstance,perhapsBobhasa


84
nutritionaldeficiencythatwouldbebetterservedbyavocadosthangrapefruit,or
Alicedoesn'tlikeavocados.Thesefactswillshapeourevaluativebeliefseven
thoughtheyarenotthemselvesvalueladen.50
However,beliefsalonewouldnotsuffice,soweneedtointroduceasecond
elementinthetheory.Thissecondelementisthecollectionofeveryones
perspectives.51Theseperspectivesallowustoevaluatethestatusofgivenbeliefsby
meansofhowmanyindependentlinesofsupporttheyhave,whichhelpstosolve
someoftheperceivedproblemswithotheraggregationmeasureslikethe
CondorcetJuryTheorem.Thus,insteadofconsideringrawbeliefsandthenumber
ofpeoplethatholdthem,wewanttoevaluatebeliefsintermsoftheirsupportfrom
differentperspectives.52Finally,wehaveamethodfortakingthesebeliefs,desires,
andperspectivesasinput,andoutputimpartialstatementsabouttheevidentiary
statusofparticularmoralprinciples.Thismethodaggregatesbeliefsintermsof
theirsupportfromperspectives,andperformsaversionoftheCondorcetJury
Theoremoverbeliefperspectivepairs,asopposedtobeliefpersonpairs.

50
See M. E Yaari and M. Bar-Hillel, On Dividing Justly. Social Choice and Welfare
(1984) 1:1-24 for more examples of this nature, and a detailed discussion of how our
factual beliefs shape our evaluative beliefs.
51
I later define perspectives as having two components: an imposed ontology, and an
evaluative heuristic.
52
For example, belief B1 might be consistent with perspective P1, whereas belief B2
might be consistent with perspectives P1 and P2. In this case, we would say that B2 is
better supported than B1.

85

TheRoleofBeliefs
Aswasthecaseintheviewfromsomewhere,weareinterestedinfirst
personinformation,includingonesownbeliefsanddesires.Individualbeliefs
matterinvirtueofthefactthattheymattertotheindividualswhoholdthem.The
differencethatwehaveinthepresentcaseisthatweareinterestedinthebeliefsof
notjustoneindividual,butallindividuals.53Thecontentandnatureofthosebeliefs
isnotrestricted:wearenotinterestedinonlysocalledimpartialreasons,but
ratheranybeliefsanddesiresthatindividualsmighthave.Theaimisforallofthese
beliefsanddesirestobepotentiallyusedinmoraldeliberation,notjustsomepre
determinedset.54
Thebeliefsofindividualsmatteronthisviewbecausebeliefsareaprimary
sourceofsupportaboutthestatusofthemoralprinciplesthatwearetryingtolearn
somethingabout.55Thislearningissomethingthatmusttakeplaceoverthecourse
ofalifetime:oneformsworkingtheoriesabouttheworld,andnewexperiencesmay

53
The scope of all individuals is a relevant concern, as it could be fairly limited or
rather extensive. For the purposes of this chapter I take it to be referring to the relevant
community that is seeking to engage in a social contract.
54
I say potentially used rather than used here because the point is that there is no a
priori restriction on what can be used in moral deliberation. I do not wish to hold the
view that everything will be used, as that is almost surely false: my beliefs that my shoes
are tied and I am sitting down are unlikely to be productive contributions to moral
discourse.
55
Other relevant sources of information are scientific, such as biological facts about
humans.


86
causeustoupdatethesetheories.Onthisviewthen,anapriorideductionofthe
principlesofmoralityisasopentomodificationasanyempiricalhypothesis.
Apotentialchallengetotheclaimthatweshouldusepeoplesbeliefsand
interestsassourcesofsupportisthatthisisunlikethehardsciences.Biologists
dontattempttoconfirmthetheoryofevolutionbymeansofmarshalingevidence
aboutpeoplesbeliefsaboutthefactsofevolution(atleastnotbeforesaidfactsare
discovered):itisfoolishonthefaceofittoattemptconfirmationthroughsuch
means.Similarly,thebeliefsofindividualsdonotcarrymuchweightina
justificatoryprocessforphysicsorchemistry.Beliefsofscientistsmatteronly
insofarastheyarebasedonempiricalandtheoreticalfindings.So,whyshouldwe
payattentiontobeliefsandinterestsinthecaseofmoraltheory?56
Theanswertothischallengeliesinthefactthatwehavenotyetfoundan
agreeduponmethodofdirectlymeasuringgoodnessorbadness,orrightnessor
wrongness,orevenresolvedwhethersuchathingcouldbepossible.Thereisnotan
identifiablephysicalpropertyforwhichwecancreatemeasurement
instrumentation.Whatsworse,notonlydowenothavesomephysicalpropertyto
pointto,butwehavenoagreeduponnonphysicalpropertythatwehavesome
extraphysicalaccessto.So,aswedonothaveatargettodirectlymeasure,weare

56
This is true both of laypeople and moral philosophers. Though much theoretical work
is done in the hard sciences, empirical testing remains an essential standard against which
theories are judged. String theory, for instance, is under increasing levels of attack from
other areas of physics precisely because it remains without any empirical validation. In
philosophy, no such checks exist against the intuitions of philosophers except
competing intuitions precisely because the intuitions cover those things for which we
are unable to get scientific data.


forcedtofindalternativemeansofmeasurement.Onesuchapproachisby

87

attemptingindirectmeasurements:findproxiesforthepropertiesthatweareafter,
andinferringfromthoseproxiestoclaimsaboutourtargetpropertiesgoodness
andrightness.
Themosteasilyavailablesuchproxyisourevaluativebeliefsandinterests.
Oneatleasthopesthatourbeliefs,attitudesandinterestsarecorrelatedwithwhat
wethinkisrightandgood.Onefeelsitwhentheyhavebeenwronged,hassome
senseofwhentheyareflourishing,andcanfindqualitiesinothersthatthey
admire.57Itisnotcleariftherearephysicalormaterialcorrelatesthatarereliable:
goodpeoplearentmoreorlessattractivethanthegeneralpopulation,anditis
doubtfulthattheyarenotablyricherorpoorer,tallerorshorter,ormanifestany
otherphysicaldifferentiator.Becausethereisadearthofphysicalproxiesforour
attemptsatmeasurement,anditseemsthatwehaveatleastsometractionwith
beliefsanddesires,wearethuspushedtowardstheuseofevaluativebeliefsand
desiresasoursourceofevidenceinmoraltheory.58

57
Of course, this need not be (and is not!) a perfect correlation, and some may be more
reliably sensitive to these issues than others. There may well be psychological tendencies
or limitations that help determine such sensitivities.
58
As previously discussed, beliefs can be either factual or evaluative. Here I take it that
factual beliefs can be derived by empirical or scientific means, which support, but do not
necessarily logically entail, evaluative beliefs. More formally, I am treating factual
beliefs as generated signals, and evaluative beliefs as interpreted signals. The world
generates signals that we can gather, and after some processing, we produce interpreted
signals. Thus, we all receive the same noisy generated signal, but there may be a
diversity of interpreted signal output.

88
Nowthatwehaveestablishedthattherelevantsourcesofevidencein

moraltheoryareourevaluativebeliefsanddesires,weshouldtakestockofhowthis
mightaligntheviewwithotherexistingtheories.Inparticular,becausetheview
seekstoaggregatethebeliefsandinterestsofindividuals,onemightfindsome
similaritytoomniscientobserverviews,suchasthosefoundinHume,Smithand
Bentham.Ontheseaccounts,thereisanidealobserverwhocanseewhateveryone
wantsandwhatwouldmaximizetheutilityofthewholepopulation,andonthis
basisdeterminewhattherightthingtodowouldbe.Thoughtherearesome
similaritiesbetweenmyviewandtheomniscientobservertradition,suchasthefact
thatbothaimtoaggregateindividualinterests,theviewbeingpresentedhereis
importantlydifferent.Theviewfromeverywhereseekstoavoidthefictionofan
individualwhocanknowandseeall:inthisview,thebeliefsofindividualsare
aggregated,butthereisnoadditionalinformationavailable.Wearenottoimagine
someobserverwhocanseeallofthemorallyrelevantfacts,andthendetermine
whatweoughttodobasedonwhatthisobserverwouldcalculateasbestfor
everyone.Justaswehavenoabilitytodirectlymeasuregoodnessorrightness,we
cannotassumethatsomefictitiousentitycanmakethosemeasurementsforus.We
arelimitedbywhatwecollectivelyknow,andanyinferencesthatwemightbeable
tomakeonthatbasis.Eveniftherearepertinentfactsthatcouldbeknown,ifno
onehasanybeliefsaboutthem,wedonothaveanyaccesstothem.

89
Nowthatwehaveabetterunderstandingofwhattheevidentiaryroleof

beliefsisintheviewfromeverywhere,wecannowturntothenextcomponent:the
roleofperspectives.

TheRoleofPerspectives
Thattheviewfromeverywhereaimstotakebeliefsandinterestsinto
accountisnotstartlinglynovel:Rousseauoutlinedabeliefaggregationprocedurein
TheSocialContract,andsincethentheconcepthasatleastbeenknown,ifnot
alwaysembraced.Andsincethen,aregularworryabouttryingtotakethesebeliefs
andinterestsintoaccounthasbeenthattherearenojudgmentaggregation
proceduresthatcanguaranteetoarriveatasingleresultthewillofthepeople
thatsatisfiesverybasicconstraints.Diversebeliefsandinterestscreateproblems
preciselybecauseitissodifficulttofindameanstoaggregatethemintosomething
coherent.Weareinterestedinbeliefsandinterestsbecausetheyaresupposedto
provideuswithsomeevidenceofwhatweoughttobelieve.59Butjustthebeliefs

59
This can be seen most clearly in a market setting. Consider the spot price of a
commodity like wheat. This price is determined based on hundreds of thousands of
small-scale exchanges that determine the future value of wheat. This includes
information about how large people think the harvests will be, what demand will look
like, whether there are any weather or pest concerns, and a host of other relevant
considerations. Though any single individual has a more or less reliable belief about
some of these factors, the final price represents what one ought to think about the future
availability of wheat. Similarly, a jurys aim is to determine what we ought to think
about the innocence or guilt of the defendant. The aggregation of the views of the
individual jurors has greater weight than their individual views.


90
andintereststhemselvesleaveuswithamorass.Tobegintoworkthroughthis
mess,wemusttakeindividualperspectivesintoaccount.
Perspectivesaresimplythefiltersthatweusetoviewtheworld.This
supposesthatwedonottakeintheworldasitis,butinstead(consciouslyor
unconsciously)choosetogroupcertainfeaturestogether,choosetoignorecertain
informationwhilefocusingonotherinformation,andchoosesystemsof
representationandinterpretation.Perspectivesarethusmentalschematathat
embodythosechoices.60Aperspectivecanbeastrivialasstandingononesideof
theroomratherthananother,oritcanbeascomplicatedastakingonsubstantive
theoreticalcommitments.Cultureswilloftenshareagreatmanyperspectives.For
instance,anindividualistperspectiveiscentraltomuchsocialreasoninginthe

60
The notion of mental schemata is found in the psychology literature. Schemata are
understood as representations of general knowledge. A schema is a basic theory about
the world. This theory is meant to apply to many situations, and includes an associated
web of beliefs, attitudes, emotions and expectations for the situations to which it applies.
On a beach in the summer one would be unnerved if there were a group of people in
business suits having a meeting it does not fit into ones schema for beach attire or
behavior. The thought of a beach conjures up an image of a generic beachscape, with
people swimming, sunbathing, making sandcastles, and enjoying leisure time. One might
also think of sunscreen or a Frisbee, as these concepts are associated with beaches. When
one encounters a beach, one immediately tries to categorize the activity one sees in terms
of these beliefs and expectations. This might even result in people discounting, ignoring,
or misinterpreting objects or activities that are inconsistent with this schema. People who
share these schemata will then have similar beliefs and attitudes about beaches, and
largely interpret a beach scene or beach experience in the same way. If, however, people
have different schemata relevant to their interpretation of, say, a philosophy classroom,
they can interpret the same scene in very different ways. While some might expect to
find a rewarding intellectual exercise, and believe that the room will be filled with
interesting and thoughtful people, others might expect to be bored, tired, or otherwise
disengaged. Though they are exposed to the same stimuli, they expect, and wind up
experiencing, very different things.


UnitedStates,wherethisismuchlesstrueinsocialistorcommunistcountries.

91

Groupmembershipmightevenentailtheadoptionofcertainperspectivesthatbind
thegrouptogetherforinstance,SierraClubmembersshareparticular
perspectivesabouttheimportanceandsalienceofenvironmentalissues,
Republicanstendtousesmallgovernmentandlowtaxesasframesforhowthey
interpretpoliticalquestions,andAmericanstendtoviewinternationalaffairsfrom
theperspectiveofhoweventsaffecttheUnitedStatesanditsstandingintheworld.
Sharedperspectivesareasourceofsocialcohesiontheyprovideaframeworkfor
mutualunderstandingandinterpretationofsharedevents.
Inordertobetterhonetheaccountofperspectivesthatisrequiredformoral
reasoning,wewillfirstconsiderexamplesofperspectivesinsimplercontexts.Once
wehaveseenhowperspectivesoperateinthosecases,wecanthencompletethe
analysisforwhattherelevantconceptionofperspectivesis.
Letusfirstconsiderhowwecometolearnabouttheshapesandspatial
locationofthreedimensionalobjects.Imagineacuponthetable.Onecancometo
knowtheshapeofthecupandwhereitislocatedonthetablebylookingatitwith
twoeyes,andpossiblymovingoneshead,ifnotoneswholebody,toexaminethe
cupfromdifferentangles.Wedothissonaturallywedonotnotice.Eachdifferent
angleweexaminethecupfromimpartsnewspatialinformation.Ifforexample,we
hadaneyepatchon,andcouldnotmove,wewouldbelimitedtoasingleviewing
angle.Inthisincapacitatedcondition,wewouldnotnecessarilybeabletotellifit
wereasmallcupthatwasupclose,oralargecupthatwasfartheraway,orevenjust


atwodimensionalimageofacup.Eachviewingangleofthecupisadifferent

92

perspectivethatwecouldtakeon,andeachoneprovidesuswithinformationthat
anyothersingleperspectivedoesnot.Sincecamerashaveonlyoneeye,
photographersandcinematographerscancreatevisualillusionsusingatechnique
knownasforcedperspective.Itcanmakeobjectsseemlargerorsmaller,orcloser
orfartherawaythantheyactuallyare.Ifwecouldlookatthesamesetupwitha
slightlydifferentperspective,theillusionwouldimmediatelygoaway,andwe
wouldbeabletoseethingsforwhattheyreallyare.
Thisexampleisinmanywaysanaptanalogyforwhattheviewfrom
everywhereaimstodo.Weoftenhavenuancedanddisagreeingviewsonwhatthe
appropriatesetofmoralprinciplesforsocietywouldbe.Inthefaceofthis,weneed
togatherasmanyperspectivesaspossiblesoastobeabletogetabetterpictureof
whatthesenuancesare.Unfortunately,however,visionisinsomesenseaninapt
analogy:afterall,wehavealreadyseenthatourbeliefsandinterestsarentdirect
reflectionsofourmoralprinciples.Wedonthavesomespecialabilitytoseemoral
principlesasbareconcepts.Insteadwecanonlyseetheireffectsonotherthings
thatwecansee.Platoseemstoberightwereonlyseeingshadowsonawall.If
thisisthecase,howcanwethinkofperspectivesinawaythatmightbeabletogive
usthesamekindoftractionthatwefoundinthecaseofvision?
Forsuchanexample,weshallturntotheoreticalcommitments.Eveninthe
hardsciences,whicharemeanttobeanarchetypeoftheviewfromnowhere,we
canfindexamplesofperspectives(inthiscasetheoreticalcommitments)shaping


ourbeliefs.Considertheexampleofunderstandingtheevolutionaryhistoryof

93

thefemaleorgasm.Standardaccountsofthefemaleorgasmassumedthatitmustbe
adaptivethatis,itwasselectedforonthebasisofsomefunctionithadthat
increasedevolutionaryfitness.However,Lloydhasdemonstratedthatallofthe
proposedadaptationistaccountslacksufficientempiricalsupport,andinfactthe
evidencepointstowardsthebyproductorFantasticBonustheory,which
suggeststhatthefemaleorgasmcamealongfortheridefromthestrongselective
pressureforthemaleorgasm.61Thisdemonstrationwasnotmadeonthebasisof
new,crucialevidencethathadnotyetbeenuncovered,butinsteadwasmadeonthe
basisofadoptingafeminist(andantiadaptationist)perspective.Mosttheorists
assumedthatthefemaleorgasmwassimilartothemaleorgasm,andthus
encouragedtheproductionofoffspring.Itisnotsurprisingthatthesetheoristswere
alsomen.Adaptationistsarguemoregenerallythatanytraitmusthavebeen
selectedfor:eachtraitexistsbecauseitindependentlyenhancesthefitnessofthe
organism.Thetaskismerelytodeterminewhatfunctionitservessowecan
understandwhyitwasselected.Traitslikefemaleorgasmsormalenipples,
however,donotseemtoserveanyevolutionaryfunction,butaremerelyharmless
byproductsofothertraitsthatwereinfactselectedfor.
Itisonthebasisofanewperspectivethatabreakthroughwasmade.Itwas
notfromamaximallyneutralposition,norfromknowledgeofuniversaltruths,nor
wasitonthebasisofnewevidence.Itwasbasedontheexaminationofthesame

61
For a detailed discussion of this, see Lloyd, Lisa. The Case of the Female Orgasm:
Bias in the Science of Evolution. (Harvard University Press, 2005)


evidencethateveryoneelsehadaccessto,butthatevidencewasseeninanew

94

light.Thisisnotevennecessarilytosaythatafeministorantiadaptationist
perspectiveisalwaysgoingtobemostuseful:anadapatationistwouldbeableto
provideaverygoodexplanationoftheevolutionaryhistoryoffemalenipplesor
maleorgasms.Further,itisentirelylikelythatneitherofthesetwoperspectives
atleastbythemselveswouldbeadequateforunderstandingsomeother
phenomenon.Aswiththeexampleofvision,itisnottheclaimthatthereisasingle
privilegedperspective,butratherthatanysingleperspectiveisgoingtobe
inadequatetofullyunderstandsomephenomenon.Additionalperspectivescan
oftenrevealnewinformation,eveniftheyrelyonthesameevidence.
Thisideathatthereisnosinglebestperspectiveforapproachingall
problemsiswellknowninthecomputerscienceliterature.TheNoFreeLunch
Theoreminsearchandoptimizationdemonstratestheimpossibilityofasinglebest
procedureforproblemsolvingacrossallproblemdomains.Infact,thereisno
procedurethatcandobetterthananyotherprocedureonaverage,when
consideringallproblemdomains.62,63Thisisabasicconstraintthathasnotbeen
fullyappreciatedinlesstechnicalcontexts.Importantly,however,ithelpsusmore

62
We show that all algorithms that search for an extremum of a cost function perform
exactly the same, when averaged over all possible cost functions. In particular, if
algorithm A outperforms algorithm B on some cost functions, then loosely speaking there
must exist exactly as many other functions where B outperforms A. Wolpert and
Macready, No Free Lunch Theorems for Search (Technical Report SFI-TR-95-02-010).
Santa Fe Institute, 1995
63
Hong and Pages Problem Solving by Heterogeneous Teams Journal of Economic
Theory 2001 demonstrates that there are equivalences between encodings of problems
and heuristics. They show that across all problems, there is no best encoding, given a
fixed heuristic.


fullyseetherolethatperspectivesplay.Aswehaveseen,perspectivesare

95

mentalschematatheyarebundlesofexpectations,judgmentsofsalience,
interpretivenorms,andemotionsforclassesofsituations.Onecanseethisasa
combinationoftwobasicparts:firstistheimpositionofanontologytoaclassof
situations,andsecondistheheuristicsorproceduresoneusestoevaluatethose
situations.64Theimpositionofanontologyisjustthedeterminationofhowone
individuatesandgroupsobjectsandphenomena,andthebasicclassificationsthat
areset.Asanexample,foodscouldbeclassifiedbasedonsize,shape,color,texture,
nutrition,freshness,price,oranynumberofotherdimensions.65Ontologiescanbe
fineorcoarsegrained,andmaybeamixtureofthetwo.Forinstance,anontology
ofpeoplemightincludeveryfinegraineddistinctionsformembersoftheingroup,
butverycoarsegraineddistinctionsforeveryoneelse.66Theremaybemanycases
oftwoontologiesAandB,whereBincludesallthedistinctionsofAbutalsoincludes
additionaldistinctions.Inthesecases,theperspectivesthatincludethese

64
The first systematic discussion of categorization and schemata in the context of social
and moral philosophy is in chapter 2 of Biccheiri's The Grammar of Society. There she
develops a detailed account of how these two mental processes pre-consciously shape our
behaviors, particularly in relation to norm adherence.
65
Note that there is a possibility for one ontology to be a superset of another that is, for
one ontology to contain all of the distinctions and categories that the first makes, but also
additional finer-grained distinctions. These two should be viewed as distinct ontologies,
and thus the basis for distinct perspectives. There is the potential for considerable
overlap between these two perspectives, which has consequences for their independence.
I deal with this in the next section.
66
This provides a framework for looking at stereotypes. Stereotypes involve very coarsegrained generalizations about classes of other people. A simple example is how the
media discusses the Hispanic vote as if it were a single bloc of voters with identical
interests, when often the subgroups have conflicts with each other and recognize a great
number of distinctions. These individuals dont typically define themselves as
Hispanic, but instead as, say, Dominican-American, Puerto Rican, or MexicanAmerican.


ontologiesmaywelldisagreewhenthosefinerdistinctionsarerelevanttothe

96

questionathandevenifthesameevaluativefunctionisinuse.67
Onceanontologyisestablished,heuristicscanbeemployedfordetermining,
say,whichfruitstobuy,whichtopaint,orwhichtoproduce.Wemightnoticethata
classificationbasedonnutritionwouldbefairlyuselessindecidingwhichfoodto
paint,butmightbeveryusefulwhendeterminingwhattoeat.Whatsmore,even
whenasingleontologyisappropriatefortwodifferentcasessay,priceisan
importantfactorindecidingwhattoeatorwhattoproducedifferentheuristics
becomerelevant.Whenconsuming,cheaperfoodmightbemoreappealing,
whereasmoreexpensivefoodmightbemoreappealingwhenchoosingwhatto
produce.Thusinasenseourperspectivesfunctionverymuchlikesearch
algorithms:theyarespecializedforuseinparticularcontextsandfromwithinthe
ontologiesthattheyarebundledwith.Asaresult,theycannotbeseenasgeneral
purpose:otherperspectiveswilloutperformtheminotherproblemdomains.

67
For example, even within a single religion, there are different levels of commitment
that can create finer or coarser ontological distinctions. A category of food might be
further refined to include acceptable food and unclean food, marriage customs might
have finer distinctions, etc. In this scenario, there is no reason to make the claim that one
or the other ontology is the correct ontology, such that one perspective is subsumed in
the other. The perspectives are surely related, but there are evidently important
differences between them. There are no a priori lower or upper limits to the number of
possible distinctions or categories in a given perspective's ontology, outside what our
cognitive limitations are at the upper end, and some sense of utility at the lower end. As
such, it is unlikely that perspectives would be useful with ontologies that have only a
single category, and it is unlikely that humans can cognitively handle perspectives that
make millions of distinctions.

97
Aperspective,then,canbeseenasamentalschemathathastwo

identifiablecomponents:theclassificationstagewherethegeneralontologyis
imposed,andtheheuristicorjudgmentstagewherechoicesandevaluationsare
made.Wehaveseenthatthereisnosinglebestperspective,norcantherebe,dueto
thelimitationsimposedonusbyNoFreeLunchtheorems.Sonowwecanseethat
wecanbenefitfrommultipleperspectivesbecauseoftheneedtocovermultiple
problemdomains.Butdoesthismeanthatweoughttoonlyhaveasinglebest
perspectiveineachdomain?Thiswasnottrueinthecaseofvisualperception,and
itisnottruegenerally.Whilewemaybeabletofindthesinglebestperspectivein
agivenproblemdomain,itisnottosaythatwecannotdobetterbyaugmentingthat
perspectivewithalternativeperspectivesthatrelyondifferentontologiesand
differentheuristics.Theseother,lesser,perspectiveswillnotfullyoverlapwith
thebestperspective,andthuspotentiallybeabletoimproveouroverall
judgmentsbycoveringareasmissedbythebestperspective.Wecanconsiderthe
exampleofdiagnostictestingtoseewhythisisthecase.Imaginethattherearetwo
diagnostictestsforadangerousdisease.TestAhas80%accuracy,andTestBhas
only40%accuracy.ButifTestBworksondifferentprinciplesthanTestA,taking
bothtestscanboosttheaccuracyofdiagnosisbecausetheyeachcaptureadifferent
aspectofthediagnosis.Soperspectivesworkingtogethercanprovidegreater
epistemicbenefitsthanjustrelyingonasingleperspective.68

68
In fact, it is also true that combining many poor performers together can outperform the
single best performer. As Scott Page and Lu Hong have put it, Diversity trumps
ability.

98
Nowthatwehaveseenhowperspectivescanimpactourdailylivesand

thescientificenterprise,wecanturntohowtheyaffectourmoralreasoning.Asin
thescientificcase,wearetryingtomarshalevidencetounderstandwhatmoral
principlesweoughttoendorse.Andjustaswecannotdirectlyobserveatrait's
evolutionaryhistory,wecannotdirectlyobservemoralprinciples.Instead,inboth
instanceswemustfindproxiesforwhatweareinterestedin.Inthecaseofthe
femaleorgasm,iftheadaptationiststoryweretrue,weoughttofindwomenwho
orgasmmoreeasilywouldhavemorechildren,butthisisnotthecase.Andaswe
haveseeninthelastsection,ourproxiesinthecaseofmoralreasoningareour
beliefsandinterests.
So,wemustrelyonourperspectivestohelpmakesenseofourmoralbeliefs
andinterests.Thiscanprovetobeadifficultthingtodo,asourperspectiveshelpto
shapeourmoralbeliefsinthefirstplace.Itisverydifficultforasingleindividualto
besufficientlyselfreflectivetobeabletotellthattheyhavecaughtthemselvesina
mutuallyreinforcingrelationshipbetweentheirmoralbeliefsandtheir
perspectivesespeciallywhenthoseperspectivesaretheoreticalcommitments.If
ourtheoryhelpsusinterpretourevidence,andwetakeourevidencetoconfirmour
theory,thenhowarewetobreakoutofthiscycle?69Theonlypathoutisbymeans
ofintroducingnewperspectives.Thiscanbedoneeitherbyanindividualorbya

69
This is characteristic of the worry about the reliability of reflective equilibrium as a
method of determining or justifying ones moral beliefs. There is nothing in the process
of reflective equilibrium that can provide a check against potential biases imposed on
ones thought process by the perspective one has. The challenge is being able to identify
those cognitive tendencies and work around them if they are liabilities. But this cannot
be done without some external check that determines whether such biases are there.


group.Whileindividualslearntotakeonnewperspectivesoverthecourseof

99

theirlifetimes,Iwouldclaimthatmostoften,theseperspectivescomefrom
exogenousfactors.Goingtocollege,takingaclassinanunfamiliarsubject,traveling,
andmeetingnewpeoplearewhatIsuspectarethetypicalwaysonelearnstofoster
newperspectives.Whileitissurelypossibleforonetodevelopperspectiveson
one'sown,thisisusuallyasignificantbreakthrough,andnotaneveryday
phenomenon.70SowhileIdothinkthatanygivenindividualhasatoolkitofmany
differentperspectivestodrawfrom,Idonotthinkthatitispossibleforany
individual'stoolkittobesolargesoastoincorporateeverypossible(oreven
currentlyexisting)perspective.Asaresult,wemustrelyongroupstobeableto
generateasmanyperspectivesaswecansowecanimproveourepistemicabilities.
Evenifwehavesuccessfullymanagedtobreakoutofourownnarrow
perspective,andarenolongerstuckinarutofhavingourperspectivescreateour
beliefswhichthenconfirmourperspectives,westillhavenotyetdealtwithhowthe
viewfromeverywheretakesthesebeliefsandperspectivesandoutputsimpartial
moralclaims.Forthis,wemustnowturntoadiscussionoftheaggregation
procedure.

70
Being able to conceive of, say, how the universe looks to a photon, as Einstein did, is
not a common exercise. At least for some, meeting someone from a different background
and learning about how they see the world is.


TheAggregationProcedure

100

Itisintheaggregationprocedurethatwecanseehowtheviewfrom
everywherecanbetakentobeamoralstance,inthatitprovidesaframeworkfrom
whichtoreasonaboutmoralquestions.Thusfarwehaveseenwhybeliefsarethe
evidencethatwehaveaccesstoinmoralreasoning,andperspectiveshavealarge
roleinbothgeneratingandinterpretingthosebeliefs.Unlikeintheviewfrom
nowhere,whichonlytakesinobjective,impartialinformationaspartofthe
reasoningprocess,sothattheoutputisalsoimpartial,herewetakeinbiased,
subjectiveinformation.Thetaskaheadofusistooutlineaprocedurethatcantake
thissubjectiveinformationanduseittocreateimpartialmoralclaims.
Forsimplicityssake,whatIwilldescribeassumesthatthereisonlyone
issueindiscussionatagiventime,andthatpotentialbeliefsaboutthisissuecanbe
mappedontoalinearscale.Forexample,anactionthatisundermoral
considerationcouldbedownloadingmusicwithoutpaying.Wecanthenimaginea
scalefromtheworstpossiblemoraloffensetosupererogatoryactionontowhich
ourbeliefsabouttheactivitycanbemapped.
IfweweretofollowRousseauandemploysomethingliketheCondorcetJury
Theorem,whatwewoulddonextiscollecteveryonesbeliefs,anddeterminethe
valuetowhichtheyconverge.Theideabehindthisisthatifweassumethat
everyoneisatleastslightlybetterthanchanceatknowingthetruth,wewillfind
thatasweaddmorepeopletoourjury,wewillbecomeincreasinglymorelikelyto


arriveatthetrueanswer.71Whileprobabilitytheoryconfirmstheveracityof

101

thetheorem,thechallengeisdeterminingwhenwehavetheappropriateconditions
forapplyingit.Namely,thejurytheoremrequirestwothingsoftheagents
participating:first,thattheywillbebetterthanchanceatpickingthetrueanswer,
andsecond,thateachindividualschoiceisindependent.
Formanyquestionsatleast(thoughcertainlynotforall),wecansafelyclaim
thatindividualsarebetterthanchanceatpickingthetrueanswer.Butonthe
secondconditionthatindividualchoicesmustbeindependentfromeachother
wearemuchlessfortunate.Therearetwosourcesofdifficultyhere.First,aswe
haveseen,peoplewhoshareperspectiveswillhavecorrelatedbeliefs.Buteven
morebasically,peopletalktoeachother.Eveninjuries,jurorsarerequiredtotalk
itoutandtrytoconvinceeachotherthattheyareright.Thiswilltendtoviolatethe
independencecondition.Nordoesitseemparticularlydesirabletotakesteps
towardstryingtoenforcesuchacondition.Ifpeoplearenotallowedtotalktoeach
otherorsubstantivelyinteract,theredoesntseemtobetoomuchofapointto
havingasocietyinthefirstplace.Soitseemsthat,eventhoughtheJuryTheoremis
epistemicallyaverydesirableprocedure,wecannotemployitandcontinuetoenjoy
livinginsociety.

71
The Condorcet Jury Theorem was meant to be used for binary choices about matters of
fact. Here, however, we are dealing with evaluative beliefs, which make no truth claims
themselves. As a result, we are forced to shift to an alternative standard, wellsupportedness. The aggregation procedures discussed in this section thus make claims
about which evaluative belief has the most support, which is independent of a truth claim.

102
ThoughwecannotadopttheCondorcetJuryTheorem,wecanlearnfrom

whereitfallsdowntodevelopaprocedurethatcancapturemostofitsepistemic
virtueswhilealsohavinglessburdensomeassumptions.Mostimportantly,this
meansgivinguptheindependenceassumption.Indoingthis,however,weopena
fairlylargecanofworms.Namely,evenifeveryoneinsocietyagreesonthesame
moralprinciple,wemaynotbeabletoendorsethatprinciple.Toseewhythisis,
recalltheconceptofforcedperspective.Ifeveryonelooksatsomethingfromthe
sameperspective,theywillarriveatthesamefalseconclusionaboutwhatisgoing
on.EveryonewatchingLordoftheRingsthoughtGandalfwasmuchtallerthan
Frodo,butthiswasjusttheeffectoftheactorplayingGandalfstandingclosertothe
camera.
Forthesereasons,wecannottakejustbeliefsintoconsideration:insteadwe
mustlookatbeliefperspectivepairs,andfurther,thedistributionsofthesepairs.
Aswehaveseen,perspectivesprovideboththeontologyandthemethodof
evaluationforagivensituation.Thus,wecanthinkofperspectivesaslinesof
supportforbeliefs.Theyprovideaframeworkforunderstandingandjustifyingthe
beliefstheyproduce.Millprovidesaprecedenthereforwhyrelyingonbeliefsalone
isnotenough:hisdefenseofthefreedomofspeechinchaptertwoofOnLiberty
arguedthattheroleofspeechwastohelpuslearntheargumentsforourbeliefs.
Weoughtnotjustholddogmaticbeliefseventruebeliefsbutmustalsoknow
whyweshouldholdthesebeliefs.Perspectivesplayasimilarrole:theytelluswhy
weholdcertainpositions.Whatsmore,theyplayaroleinourjudgmentsof


epistemicnormativity:giventhatoneholdssuchaperspective,oneshouldbe

103

ledtocertainkindsofbeliefsandjudgments.Sowhenweseektoaggregateour
beliefs,wemusttiethataggregationtotheperspectivesthathelpedtogenerate
thosebeliefs.
TheconceptualshiftIaimtomakeherewhendepartingfromtheCondorcet
JuryTheoremisthat,wheretheJuryTheoremprovidesonevotetoeachindividual,
Iaimtoprovideonevotetoeachperspective.72Becauseaperspectiveisbothan
ontologyandanevaluativeprocedure,eachperspectiveiscompatiblewithonlyone
setofbeliefs.Eachperspectivereceivesequalweight,regardlessofhowmany
peopleshareit.Ifanentirepopulationsharedonlyoneperspective,forexample,
therewouldbeonlyonevotetocount.Thislooksvehementlyantidemocraticon
thefaceofit,butrecallthisprocedureismeanttobeepistemic,andnotpolitical.
Thegoalistodeterminethesetofbeliefsthathavethemostindependent73linesof

72
This move is meant to account for the problem of independence with individual voting.
By counting perspectives, which essentially represent lines of argument, we can
guarantee that we are counting independent judgments. This continues to capture the
spirit of the Condorcet Jury Theorem, but instead identifies perspectives as the carriers of
beliefs instead of individuals.
73
This claim of independence relies on our ability to identify different perspectives as
being different, and further, that different perspectives are non-overlapping. The first
challenge to this that we are able to identify different perspectives, is an empirical
assumption. Even if one cannot identify one's own perspective, it is likely that someone
with a different perspective will point it out. The more difficult question is whether or
not any two given perspectives are always independent of each other. One way of
assessing this question is by recalling that perspectives have both an ontological
component and an evaluative component. If two perspectives have different ontologies,
then they are independent. If they share ontologies but have different methods of
evaluation, then statistical comparison across numerous cases will determine their level
of correlation. In these cases of dependence, then the value of the vote for each
dependent perspective should be reduced to match the increase in information it provides


104
argumentationsupportingthem,notthosethataremostwidelyheld.Theseare
themoralprinciplesthatwecanthentakeforgrantedatthebeginningofthe
politicalprocess.Partofthetaskofthepoliticalprocessistodeterminewhetherto
actonthoseprinciplesforwhichthereisnotmuchepistemicsupport.Thetaskof
thenextchapteristodevelopthisprocess.
Nowwecanturntoadiscussionofthedifferentpossiblecasesofthe
distributionsofperspectivesandbeliefs.First,letusconsideracaseofagreement.
Asdiscussedearlier,thismaybeduetothefactthateveryonehasthesame
perspective.Sothoughthereisagreement,whenwehaveagreementbasedonthe
samesingleperspective,weoughtnottoplacemuchcredenceonthefactof
agreement.Afterall,weessentiallyonlyhaveonevote.JuryTheoremsrequire
largenumbers.Asaresult,weshouldhavenofaiththatasinglevoteisreliable.As
such,ratherthanendorsingtheagreeduponmoralclaim,agentsintheviewfrom
everywherewouldhavetoseekoutalternativeperspectivestoprovide
confirmationthattheiragreementwouldcontinueinlightofnovelperspectives.
Withonlyoneperspective,wehavenowayoftestingwhetherourbeliefsarerobust
againstalternativewaysofthinkingaboutandevaluatingthesituationinquestion.
Robustnessagainstalternativesisthesolechecktheviewhasagainstmoral
relativism.Amoralbeliefneedstobeabletowithstandthejudgmentsofmultiple
perspectivestohaveanyweight.Justasscientificpracticefindsvalueinrobust

over those perspectives on which it depends. As this is a statistical phenomenon, it is

likely that these values will have to be adjusted over time, as better information allows
for more refinement.


resultswhicharesupportedbymultiplemodels,sotoodoestheviewfrom

105

everywherebuttressmoralbeliefsbymeansofmultipleperspectives.Amoral
beliefsupportedbyasingleperspectiveisonlyasstrongastheweakestassumption
intheperspective.Amoralbeliefsupportedbymultipleperspectivesisstronger
thanthestatusofanyindividualperspective.74Arobustmoralbelief,then,remains
stableagainsttheintroductionofnovelperspectives,andtheeliminationoffaulty
perspectives.75
Alternatively,wecouldfindagreementamongstindividualswithseveral
differentperspectives.Thedifferentperspectiveswouldthenprovideaconsilience
ofevidenceinsupportofthesamebelief.Inthiscase,theagentsintheviewfrom
everywherecanclaimthattheiragreementisevidenceofarobustmoralbelief.Now
thatwehaveseveralactiveperspectives,wehaveajurythatisinagreement.Our
beliefsarerobustagainstperturbationinhowweclassifythesituationandhowwe
trytoevaluateit.Thissuggeststhatwehavefoundprinciplesthatareandwill
remainwellsupported.Thesecasesarestraightforward,butthisislargelybecause

74
This idea was developed in the context of scientific models in Wimsatt's Robustness,
Reliability and Overdetermination, found in his Re-Engineering Philosophy for Limited
Beings (2007).
75
It is important to emphasize that the moral beliefs at issue here are token beliefs, not
type beliefs. A belief like everyone should be treated equally needs to have
accompanying definitions of everybody and equally for it to count as a moral belief.
Sen importantly raised this issue in chapter 1 of his Inequality Reexamined (1992). All
mainline political frameworks have some notion of equality, but they pick different
substantive theories on which they define equality. So while they are structurally similar
and may agree on a type belief, they will disagree on token beliefs, since those token
beliefs will have fleshed out the concepts involved. Thus type beliefs cannot be robust
because they are not substantive only token beliefs can be subject to robustness checks.


agreementiseasytodealwith.Themoreinterestingcasesarethoseinwhich

106

agentsdisagreewitheachother.
Forsimplicityssake,Iwillconsiderdisagreementtoberepresentedasa
bimodaldistributioninbeliefs.Thisisofcoursebutonewayinwhichsubstantive
disagreementcouldmanifestitself,butitisafairlycommoncase:itrepresentstwo
opposingpositionsthateachhavealargeportionofthepopulationassupporters.
AnempiricalexampleofthiscanbefoundwithattitudesoffairnessintheDictator
Game.Inthisgame,oneoftwoplayersthedictatorisgiven$10,andtoldthat
theycangiveasmuchastheywanttothesecondplayer,andkeeptherest.The
dictatorsdecisionisfinalanddeterminesthedistributionofmoney.Cristina
BicchieriandJasonDanaaskedstudentsenrolledinintroductiontophilosophyat
CarnegieMellonUniversitywhattheythoughtwasafairamountofmoneyforthe
dictatortogiveinaDictatorGame.Theresultswereabimodal:68.3%ofstudents
thoughtthatequalsplitwasthefairallocation,while21.4%thoughtthatnotgiving
anythingwasperfectlyfair.Slightlylessthan20%thoughtthatnotgivinganything
wasunfair.76Giventhisbimodaldistributionofbeliefsaboutfairness,wecan
imaginetwopossiblescenarios:first,thatthesebeliefsarecorrelatedwith
particularperspectives,orthatthereisnocorrelationbetweenbeliefsand
perspectives.
Inthefirstcase,wehavesubstantivemoraldisagreement,buteachside
marshalsadifferentsetofarguments.Thisispreciselythecaseinwhichmoral

76 Details of this study can be found on pp. 137-139 of Biccheiris The Grammar of Society (2006).


deliberationoughttobemostproductive.Theagentscometodifferent

107

conclusions,butthismayverywellbebecausetheyseetheworlddifferently.Just
asourcasewithperspectivesinscience,thereisanopportunityhereforan
exchangeofperspectivesthatmighthelptosettlethemoraldispute.Itcouldbethat
onesideiscorrect,oritcouldbethatbothsidesarecorrectlycapturingsome
portionofthesituation,withoutrealizingthatthereareothercomponents.Inthe
firstinstance,wefindourselvesinasituationnotunlikethecaseofthefemale
orgasmthatweconsideredabove.Onesidesperspectivesareinfactillsuitedfor
theparticularsituationathand,whiletheothersidesaremuchbettercalibrated.In
thesecondinstance,wefindourselvesinacasesimilartothatofmedical
diagnostics:eachsetofperspectivesisusefulforsomeportionoftheproblem,but
onlybycombiningthemtogetherdowegetamorecompletepicture.Itisthus
possiblethattheexposuretotheotherperspectivescanenablethesidewithless
robustmoralbeliefstoseehowfragiletheirbeliefsareinthefaceofnew
perspectives,orforbothsidestousetheinformationgleanedfromtheothers
perspectivestofindthatthereisabetter,thirdpositionthattheybothcanagreeon.
Assuch,deliberationamongstpeoplewiththesediverseviews,notoverwhich
perspectiveisbesttotakeineverycase,77butinaparticularinstance,hasgreat
potentialforgettingtothecorrectview.Thisdoesnotmeanthattherewillbe
agreementineveryinstance,butfromapurelyformalepistemicperspective,thisis
thecaseforwhichthereisthemostusefulevidencetoleverage.Asthenumbersof

77 After all, as we have seen already, the No Free Lunch Theorem shows us that there is no such
perspective.


activeperspectivesincreases,weinfactfallbackintoourmodifiedversionof

108

theCondorcetJuryTheorem.Assuch,weoughttohaveincreasingconfidenceinthe
factthatastraightforwardaggregationmechanismwillyieldthemostreliable
result.
Inourfinalcase,wefindthatweagainhaveabimodaldistribution,butthis
time,beliefsdonotcorrelatewithperspectives.Bothpositions,then,havewhat
appearstobeaconsilienceofevidenceintheirfavor,exceptforthefactthatthe
othersidehassimilarlevelsofsupport.Thatis,bothsidesmarshalthelinesof
argumentationtodefendtheirbeliefs.Theybothappeartobeendorsedbythe
samesetofperspectives.Hereiswheretheviewfromeverywherediffersmost
drasticallyfromtheviewfromnowhere.Sinceeachperspectiveappearstoendorse
multiplebeliefs,noJuryTheoremcanoperate.78Itwouldbelikeallowing
individualsintheCondorcetJuryTheoremtocasttwooppositeballotstherecould
benoconvergenceintheaggregation.Asthereisnoformalproceduretousethat
canleverageperspectivestohelpustoinformanunderstandingofwhereour
beliefscamefrom,andthusenlightenthediscussion,wearestuckatasituationof
substantivemoraldisagreement.Theagentsintheviewfromeverywherehaveno
logicalmeanstoconvinceeachother,sincetheyshareperspectivesalready,andso
furthermoraldeliberationwillprovetobefruitless.

78
Note that this is not a case of a finer-grained perspective disagreeing with it's coarsergrained counterpart. Instead, these are identical perspectives that ostensibly say two
different things.

109
Inthissituation,wefindadeviationinthesupposedepistemic

relationshipbetweenbeliefsandperspectives.Inthethreeothercaseswe
considered,wesupposedthataperspective,combinedwithepistemicrationality,
constrainedtheindividualswhotookonthoseperspectivesinwhattheycould
believe.Thatis,aperspective,astakentobeamentalschemathatprovidesan
ontologyandamethodofevaluationforagivenclassofsituations,willtendtocause
individualstohaveacertainsetofbeliefs.Thosebeliefscanbethoughofasbeing
entailedbytheperspective.79Inthiscase,however,theconstraintsonbeliefsthat
perspectivestypicallyprovidearesomehowviolated.Thus,perspectiveslosetheir
roleasguarantorsofevaluativeindependence,asthisentailmentrelationisbroken.
Assuch,alltheviewcandoisrepresentwhattheviewsthatareunderactive
considerationare,anddescribethenatureofthedisagreement.Thesepositionsare
insomesensereasonindependentgiventhatthesamestrainsofevidencesupport
bothpositions,itsuggeststhatthedisagreementisnotactuallybasedonthe
evidenceitself,butjustsomekindoffundamentalarationaldisagreement.

79
The relationship must actually be something other than entailment, as a perspective is
not a set of propositions. Further, there is likely to be at least some natural variation
amongst individuals who share the same perspective. I see perspectives as constraining
what that variation can be, and providing the justificatory framework for a given set of
beliefs. Epistemically rational agents will then tend to have those beliefs for which the
perspective provides justification.


Conclusion

110

Aswehaveseen,theviewfromeverywhereaimstotakeeveryonesbiased
viewsabouttheworld,andcreatefromthatanimpartialmoralclaim.Bytaking
advantageofbothourbeliefsandourperspectives,theviewaimstouncoverthe
wholepictureoutofthepiecesofitthatweallhave.Indoingso,however,theaim
doesnotsupposethatwehaveanswersforeverythinginfact,different
instantiationsofthisviewwillhavevaryingabilities,dependingonthemakeupof
theindividualsparticipating.Whatsmore,theviewcansuggesttousthatweeither
donotknowenoughtomakecertainmoralclaims,orthatoursupportforsome
positionsmightnotactuallyrelyonanyevidenceorrationalsupport.Thisisan
unpleasantresult,butthevalueinthiskindofresultisthatitcanpointtothe
limitationsinourmoralepistemology.Rationaldeliberationcannotalwaysensure
thatwewillendupagreeingwitheachother,norcanitassurethatwewillalways
honeinonthebestprinciples.Theviewfromeverywheremerelyidentifiesforus
wherewecancometoknowwhatmoralclaimsarewellsupported,andwherewe
arecurrentlyunabletoproceed.

111

Chapter3:
JusticeWithoutAgreement

Introduction

Inthepreviouschapter,weexaminedtheepistemicbenefitsofleveraging
multipleperspectivesinoureffortstofindwellsupportedmoralprinciples.This
formsthebasisforourtaskinthischapter,whichistofindaprocedurefor
determiningthesetofrightsaffordedtoindividualsinsociety.Whileitispossible
thatindividualsparticipatingintheviewfromeverywherewillfindthemselvesin
broadagreementoverthebasicstructureoftheirsociety,thischapterdoesnot
assumethatsuchagreementhasbeenfound.Infact,theaimofthischapteristo
developaprocedurefordeterminingthesetofrightsaffordedtoindividualswhen
thosesameindividualsagreeonverylittle.Infact,Iwillconsiderthecasewhere
peoplenotonlydisagreeaboutwhattheirpreferredoutcomesare,buttheyalso
disagreeaboutwhatthesetofoutcomesevenlookslike.

SensInsight
InthefirstchapterofInequalityReexamined,Sennoticedsomethingquite
important.Whenoneconsidersthemajorconceptionsofjustice,everyoneofthem
hassomenotionofequality.Itisjustthateachaccountofequalityisdifferenteach


112
theoryismeasuringsomethingdifferent.Libertariansareconcernedwithequal
negativeliberties.Egalitariansareconcernedwithequaloutcomes.Utilitariansare
concernedwithequalconsiderationinutilitymaximization.IfanEgalitarianwere
tolookateitherautilitarianorlibertarianregime,shewouldlikelyfindsignificant
inequalitiesinindividualoutcomes,butwhenthosesocietiesareviewedbytheir
ownmembers,theymightseethemasmodelsofequality.Itisnotthatthe
egalitarianismistakeninherview,butsimplythatsheisusingadifferentstandard
ofequalitythantheutilitarian,whoisinturnusingadifferentstandardthanthe
libertarian.
Withthisinsight,Senshiftsthedebate.Ratherthanarguingaboutwhether
equalitymatters,wefindthateveryoneisinfactarguingaboutwhichpolitical
valuesmatter.Libertariansassertthatfreedomisthemostimportantvalue.
Egalitariansthinkthatsocioeconomicoutcomesmatterthemost.Utilitariansclaim
thathappinessisthemostimportantvalue.Eachtheory,then,developsawayof
lookingattheworldwhichaccordswiththevaluesthatmatterthemosttothat
theory,includingstandardsofcomparison.Unfortunately,onceSenmakesthis
remarkableinsight,hegoesontosimplyproposeanewmeasurementofequality
capabilities.Whilehiscapabilitiesapproachisinterestinginitsownright,much
morecanbeminedfromtheoriginalinsight.
AwaytolookatwhatSennoticedisthateachpoliticaltheoryisa
representationofaparticularperspective.Thisperspectivecategorizestheworldin
termsofthevaluesthatthetheoryholdsdear.Italsoprovidesajudgment


apparatusbuiltaroundthiscategorization.Assuch,thesepoliticaltheories

113

carrywiththemperspectivesthatshapeourreasoning.Notonlydothese
perspectivesshapeourpreferencesoverpotentialpoliticaloutcomes,buttheyalso
determinewhatweseeastheoutcomes.Forexample,aLibertarianmightseea
politicaldebateintermsoftwooutcomes:lossoflibertybecauseofincreased
governmentinterventionandincreasedlibertybecauseoflessgovernment
interventionandthuspreferthelatteroption.Meanwhile,anEgalitarianmightsee
freegovernmentprovidedhealthcareforallandeliminationofsocialsafetynet,
andthenprefertheformeroption.Bothpreferencesareinformedbypolitical
values,butmoreimportantly,thosepreferencesappeartobeoverdifferent
outcomes.Thelibertarianandtheegalitariannotonlydisagreeaboutwhatsa
betterchoice,buthowtheyseethosechoicesinthefirstplace.
Thepreviouscaseisnotevenaparticularlyproblematicexample:both
partieshereagreeonwherethecontoursofthetwooutcomesare,butseethem
underdifferentdescriptions.Thiswillnotalwaysbetrue.Forinstance,considera
vegan,avegetarian,anomnivore,andanomnivorethatkeepskosher.These
individualswouldntsharethesamepreferencesaboutwhattoeatfordinner,but
moreimportantly,theywouldhavedifferentcategorizationsofthepossibilities.A
veganwouldlargelyseetheworldintermsofthingsthatareveganandthings
thatarenotvegan.Vegetarianswoulddrawthelinedifferently.Omnivoreshave
noparticularneedforsuchdistinctions,whilesomeonekeepingkosherwouldlook
todistinguishbetweenmeats,basedonwhicharepreparedandserved


114
appropriately.Avegandoesntevenhavetothinkaboutwhethersomethingis
kosherornoteverythingthatisntkosherisalreadynotvegan.These
categorizationswillnaturallyaffectpreferencesoverchoices,buttheyalsocausethe
differentpartiestoseedifferentchoicesaltogether.Whereastheomnivorecanhave
developedviewsaboutwhetherporkbellyisbetterthanbaconandporkchops,the
othereaterswouldsimplylumpallofthoseoptionsunderacanteatcategory.
Likewise,ifsomeonewantedtorequirethatallbutchershopsinthecountrymust
behalal,noonebuttheomnivorewouldhavereasontocare,asitwouldonlyaffect
foodsthattheyalreadywouldnteat.
Animplicationofthisviewofperspectivesisthatoncewehavechosena
particularpoliticaltheoryasthebasisforourpoliticaldiscussion,wehavecutoff
theabilityofpeoplewhohaveadifferentperspectivetoevenproperlyrepresent
theirownviews.Itisnotmerelyamatterofwhattheirpreferencesmightbe,but
howtheyseetheproblem.Choosingtoprivilegeasingleperspectiveabovethe
othersmeansthatthevaluesembeddedinthoseotherperspectivescannotbefully
articulated.80Thismeansthatonceaperspectiveischosen,muchofthesubstance

80
Donald Davidson argues against this position in On The Very Idea of a Conceptual
Scheme Proceedings and Addresses of the American Philosophical Association, Vol.
47, (1973 - 1974), pp. 5-20. He claims that if we think of conceptual frameworks as
languages, then the claim is either trivial or incoherent. If we merely mean that we can
only assign truth claims within a linguistic framework, then this is a trivial claim.
However, going more in the Kuhnian direction is incoherent, since he claims that we can
only make sense of different points of view by reference to a common coordinate
system which allows us to map the differences. But then, given such a coordinate
system, we can establish a translation from one conceptual scheme to another. I would
respond, however, by suggesting that Davidson is himself stuck within his own
perspective, that of linguistic-turn philosophy. Instead of considering conceptual


115
ofpoliticaldisagreementhasalreadybeenshutdown.Oncethishasbeendone,
thewayweconsiderwhattheoutcomesareandwhatouroptionslooklikehave
alreadybeendetermined.FollowingSen,thismeansthatthedebateoverwhich
valuesmatterthemosthasalreadylargelybeencompleted.Itisnodoubtafactof
psychologythatindividualsaregoingtobemorepronetoviewtheworldinterms
ofoneperspectiveoranother.Ifanything,thisshouldbolsterourinterestin
preventingasingleperspectivefrombeingtheonlyappropriateframeforpolitical
discussioninthepublicsphere.Whilemanymaysharethesameperspective,it
wouldpreventotherswhohaveadifferentperspectivefrombeingabletofully
participateinpoliticaldiscourse.
Whatthissuggestsisthatinsteadofchoosinganothermeasureofequality,or
anothersetofvaluestoprioritizeforsocialorganization,weneedtoturntoamodel
thatallowsforpoliticalagreementamongstindividualswhoholdanyarbitraryset
ofperspectives.Itisimportanttonotethatthisisnottosaythatdevelopingnew
perspectivesisnotvaluable:theviewIpresentcommitsmetotheideathatwecan
benefitfromthedevelopmentandpresentationofnovelperspectives.However,
anointingasingleperspectiveassomehowneutralorprimaryisproblematicforthe
reasonsIhavedescribed.Assuch,wemustthenmovetowardsamodelofpolitical

schemes as synonymous with languages, we could imagine them as mathematical objects

known as projections. A projection takes a higher-dimensional object down to a lower
dimension, and in the process, some information is necessarily lost. Even if we start with
the same high-dimensional object, and then take two different projections of it, we cannot
translate between them because even though they started from the same source, they each
differ in what information they retained. So no direct mapping between them is possible.
And thus, had Davidson had an alternative conceptual scheme to work from, he could
have seen that they are very much possible.


116
agreementthatattemptstocombinethesedisparateperspectivesinawaythat
allowsallparticipantstobeenfranchised.

WhatisaSocialContract?
Beforeweproceed,itisimportanttoarticulatewhatexactlyitisthatweare
aimingforinamodelofpoliticalagreement.Inparticular,weareinterestedin
developingasocialcontract.Sowemustfirstaskourselveswhatthegoalsofa
socialcontractare,andthenwecanseehowtogoaboutsatisfyingthem.Indoing
so,wecanidentifythenecessarycomponentsofasocialcontracttheory,andthen
bestdeterminehowtosatisfythem,givenourcommitmenttoenfranchisingall
perspectives.
Ifweconsiderwhatismostfundamentalaboutasocialcontract,wecanseeit
asatminimumanattempttodesignasetofrulesandassurancestofurtherthegoal
ofallowingindividualstosuccessfullylivetogetherinsociety.Inparticular,wecan
imagineourselvesasseeingthepotentialvalueoflivingtogether,butweneedto
establishthebasisforoursocialcooperation.Itisthisneedtoestablishthebasisfor
socialcooperationthatledtheclassicalsocialcontracttheoristssuchasHobbes,
LockeandRousseautopostulateastateofnaturewherehumansareprecontract
andinturn,presocialinthesensethatthereisnocooperation.81These

81
Note that in Lockes conception, the state of nature is not entirely pre-social. In fact,
individuals have basic property rights and other things that we would normally associate
with a social contract.


117
philosophersweredifferentinhowtheypicturedthestateofnature,buttheyall
arguedthatindividualswouldallbemotivatedtoenterintoasocialcontractwith
theirfellows.Fromthis,wecandrawourfirstconclusionaboutthenatureofthe
socialcontract:thatwecanviewthesocialcontractasamutuallyadvantageous
agreementdesignedtofacilitatesocialcooperation.
Mutualadvantagemightinitiallyseemlikealoadedtermaimedatskewing
thediscussion,butIwanttoargueitismerelyaminimal,necessaryconditionofa
socialcontract.Ifthisbasicconditionwerenotbeingmet,itwouldimplythatsome
subsetofthepopulationisbeingharmedbyremaininginsociety.Thissuggeststhat
eitherthosepeoplewouldnothaveenteredintothecontractinthefirstplace,or
wouldleaveifgivenachance.Arationalagentwouldnotjoinasocietyifitharmed
herunless,likeaslave,shewereforcedtojoinsociety.ForHobbes,theabsenceof
mutualadvantagemakessocialcontractsimpossible,asitwouldunderminehisown
argument.ThekeylegitimatingfeatureoftheLeviathanisthatsheenables
cooperationamongsttheruled,andanescapefromthenasty,brutishandshort
lifeinthestateofnature.EvenRawlsneedsmutualadvantage,asindividualsinthe
OriginalPositionarestillrationalagents,andwouldnotchooseundertheveilof
ignoranceasocialcontractthatleavestheworstoffinaworsepositionthansome
alternative.Eventherichwouldwanttobeinsomesocialcontract,asitassures
themprotectionoftheirpropertyandtheirpersonsagainstattacksofthoseless
welloff,aswellasinsuranceagainstdramaticreversalsoffortune.

118
Thesocialcontractmustenablerationalagentstocooperatewitheach

otherandindoingsomakethemselvesbetteroffthantheywouldhavebeen
otherwise.Thisisapurelyindividualisticrequirement.Itfollowsfromagentsbeing
rationalthattheywillnotacceptadealthatmakesthemworseoffthanbefore.82
Mutualadvantageisthusreallyacollectionofindividualadvantages,sincenoagent
needstobeconcernedwithwhethertheotheragentsviewthemselvesasbetteroff
thanbeforeeachagentwatchesoutforherowninterests.83
Mutualadvantage,thusminimallydefined,alsoimpliesthatindividualswill
abidebythecontractonlysolongasitistotheiradvantagetodoso.Ifsubsetsof
thesocietywouldbemadetobebetteroffifthesocietysplitupintomultiple
societies,thensocietywillsplit.Theclassicalsocialcontracttheoristsdonotdiscuss
this,butitisaconsequenceoftheideaofmutualadvantage,andonethatIwill
returntoatgreaterlengthlater.
Mutualadvantageisaniceprinciple,butinasenseitiswithoutcontentfor
thepurposesofdevelopingsocialcontracts.Afterall,weneedtospecifyhowwecan
enablecooperation.Socialcontractsenablecooperationbymeansofprovidingaset
ofrightstocitizens.EvenintheminimalHobbesianmodel,theLeviathanensures

82
In the following chapter we will see this principle, which I call the no-subsidy
principle, elaborated in greater detail.
83
It should be noted that this individual focus is independent of whether agents view
advantage in a relative or absolute sense. That is, whether individuals compare
themselves to their neighbors to assess how well they are doing. Even if an agent has a
relative sense of advantage, she does not have to care about whether her neighbor
considers himself to be better off just that she sees herself as better off. So even if she
is concerned with relative losses or gains, she does not have to care about her neighbors
perceptions of loss and gain.


119
propertyrightsandrightstobodilyintegrity.Rawlshasamuchmoreexpansive
accountofrights,butwhatiscommonisthatineachsocialcontractaccount,wesee
thatrightsaretheguarantorsandenablersofsocialcooperation.Socialcontracts,
then,alwaysembedsetsofrights.Theyspecifythelimitsandscopesofrights,
includingbothwhatthegovernmentcannotinterferewithandwhatthe
governmentowesitscitizens.Itisbymeansoftheserightsthatindividualscan
assesswhethertheyareworseofforbetteroffwiththemcomparedtosomeother
bundleofrights.Forexample,individualscanassesswhetheraschemeof
individualpropertyrightswouldbebetterorworseforthemthanaschemeof
communalproperty.Individualpropertyrightshaveanumberofincentives
orientedadvantagesovercommunalproperty,thoughthatcouldbebalanced
againstreductionsininequalityandamorecommunitarianspiritengenderedby
communalproperty.Individualscanassesswhichsystemofpropertybettersuits
them.
Nowthatwecanconceiveofsocialcontractsasembeddingbundlesofrights,
wehavetwoquestionstoanswer.First,wemustthinkofaprocedureforallocation.
Thisdetermineswhatrightsweallocateaswellaswhogetsthem.Itisherewhere
wecanfindthemostsignificantsourceofdifferencesintheoriesofthesocial
contract.Iwillconsidertwoalternativeapproachestothisquestion:deliberation
modelsandbargainingmodels.Butwearenotdonethere:eventhoughwemaybe
abletosaysomethingabouttheprocessforallocatingrights,westillneedtoclearly


120
84
specifywhatwemeanbyrights,andhowtointerpretthem. Iwillproceedby
firstcomparingdeliberationmodelsofthesocialcontracttobargainingapproaches,
andshowthatbargainingapproachesaresuperior.Fromthere,Iwilldevelopa
conceptionofrightstobebargainedover.Finally,Iwillturntotheparticular
bargainingmodelthatIwishtoadvocate.

BargainingversusDeliberationapproachestotheSocialContract
ThesocialcontracttheoriesofRousseauandRawlsemphasizeaconception
ofpublicreason.Inthismodel,moralagentshavesubstantialmoralagreement.In
particular,theGeneralWillarticulatesthewillofcitizensquacitizens.Itismeantto
beaneutralstatementofwhatthegoodofsocietyis.Rawlsdevelopsasimilar
conceptionintheoriginalposition,whereinpeopleidentifythosegoodsthat
everyoneneeds.WhileRawlstriestoavoidtherequirementthatagentsshareinthe
samecomprehensivemoraldoctrine,hedoesbelievethatthereareonlycertain
formsofpublicjustificationthatareappropriate.Publicreasonisthenthose
justificatoryargumentsthatcanbemadesuchthattheycanbemadefroma
commonpoliticalpointofview.Publicreason,liketheGeneralWill,isacommon
politicalconceptionofthegood,andasystemofpoliticaljustification.

84
Consider the example of free speech. We can say that we have a right to free speech,
but this right must be fleshed out by specifying the particular instances of speech that are
deemed protected. I will elaborate this in detail in a later section.

121
Rawlssawpublicreasonasthepoliticalsolutiontopluralism.

Comprehensivemoraldoctrinescouldnotbeusedforpublicjustification,becausein
apluralistsocietythereisnoassurancethatotherswillshareonesown
comprehensivemoraldoctrine.Assuch,theywouldbeveryunlikelytoacceptthe
formofjustification,andthentherewouldbenopoliticalagreement.Byturningto
acommonpoliticalratherthanmoralconceptionofjustification,Rawlswasaiming
tosuppresstheprivatespheredifferencesinasocietyandfocussolelyoncommon
politicalcommitments.ThisisnotunlikeRousseausgoalofformingtheGeneral
Willbyhavingcitizensreasonascitizensratherthanasselfinterestedindividuals.
Thisisaproposalthathasmanymerits,butultimately,ithastoconfrontthe
problemsthatwehaveconsideredabove:namely,thatthisreasoningascitizenis
notneutral,butinsteadisprivilegingacertainconceptionofvaluesthatmaynotbe
universallyheld,particularlyassocietiesbecomemorediverse.Thisproblemwill
arisewithanyconceptionofpublicreason,preciselybecausethereisnoassurance
thattherewillbeasufficientsetofoverlappingconsensusaboutwhatmaybe
describedasacommonpoliticalpointofview.Ourcomprehensivemoraldoctrines
andourotherprivatesphereinterestsplayasignificantroleinshapingour
perspectives,andinturnourperspectivesshapeourpoliticalinterests.Though
takingpartintheviewfromeverywheremighthelptoestablishsuchapublicpoint
ofview,wehavenoguaranteethattheoverlappingconsensusissufficientforthe
kindofrobustpoliticalagreementnecessaryforpublicreason.

122
Thisproblemofdiverseperspectivesposesasignificantchallengetothe

goalofpoliticalagreement.Eventhoughwemaybeabletouncoversomeareasof
consensus,therewillbemuchthatwillhavenoconsensus.Therearesimplytoo
manydifferentareasofpoliticallifethatrequireindependentconsiderationto
expectthatajustificatoryframeworkthatcanbeestablishedinoneareacanbethen
extendedtoanotherarea.Evenifweallagreethatmurderiswrong,thatdoes
nothingtohelpusagreeonquestionsofdistributivejustice.Likewise,evenifwe
agreedontheprinciplesofdistributivejustice,thatdoesnothingforestablishing
freedomsofspeechorassembly.Thisiswhyacommonpoliticalpointofview
wouldbesovaluableifwecouldhaveit:itresolvesalloftheseissues,andgrantsus
asystematicframeworkforpoliticaljustification.
Thuswearestuckwithaproblem.Weneedsomemethodforreconciling
competingpoliticalgoalswithouthavingrecoursetoaneutralpoliticalperspective
thateveryoneendorsesasasystemofpublicjustification.Iproposethatthis
problemcanberesolved,buttosolveitwemustmoveawayfromaconceptionof
publicreason,andinsteadmovetowardsamodelofbargaining.Bargaininghasthe
advantageofrequiringverylittleagreementbetweenthepartiesinvolved.WhenI
negotiatethepriceofashirtinaChinesemarket,Idonothavetosharemany
politicalvalueswithmycounterpart:wejusthavetofindapricesuchthatshefinds
ittobeatleastequaltohervaluationoftheshirt,andIfinditnomorethanequalto


123
85
myvaluationoftheshirt. Infact,mycounterpartandIbarelyneedtohavea
commonlanguagebeyondnumeracyandsomedisplayofagreementor
disagreement.WecertainlydontneedtoagreeonthefinerpointsoftheCultural
Revolution,theGreatLeapForward,orforthatmatter,theGreatSociety.Norneed
weevendiscussourviewsonfairness:theexchangeeitherhappensoritdoesnt.
Thisisasignificantdeviationfromthepublicreasonmodelinseveralrespects.
Mostimportantly,wedonothaveacommonsystemofjustification,eventhoughwe
aretypicallyabletofindagreement.Neitherofushastomakeaclaimaboutsome
neutralpositionfromwhichwecanevaluatethebargain.Icanseektobetter
understandtheneedsofmycounterpartsoImightbeabletobetterdetermineher
reserveprice,andshelikewisecandothesamewithme,butinthesecaseswe
arentfindinganoutsidestandardinsteadweareattemptingtobetterunderstand
eachothersownstandards.Eachpartymustbeconvincedonhisorherownterms.
Thisbargainingviewhasseveralconsequences.Firstamongthemisthat
bargainingoutcomesaregoingtobesensitivetotheparticularsoftheperspectives
oftheagentsinvolved.Thisisanaturalconsequenceoftheeliminationofaneutral
standardthatbothpartiesshare,aseachbargainisgoingtobeaffectedbythe

85
It may be argued that there are some other implicit agreements that we both have: that
money is a reasonable unit of exchange, that you dont just kill the other person and take
what they have, etc. These assumptions can be eliminated without much difficulty we
could barter and trade goods instead of using money, and we can both be armed with
guns or other weapons that make unilateral attack not terribly likely to succeed. Even in
US prisons, where individuals are more likely to be violent and prone to disagreement
than the general public, market systems emerge with agreed-upon units of exchange.
Originally the monetary unit was a cigarette until they were banned, and now it is canned
mackerel. See Justin Schecks Mackerel Economics in Prison Leads to Appreciation for
Oily Fillets, http://online.wsj.com/article/SB122290720439096481.html


pairingsofparticularperspectivesratherthanasingleneutralshared

124

perspective.Sincedifferentperspectivesviewoutcomesdifferently,thereisno
reasontosuspectthattheywillallidentifythesameoutcomeastheappropriate
solution.Thisiseasytoseeinnormalfleamarketcases:thesamegoodisoftensold
atdifferentpricesevenbythesamevendor.InthecasethatIlooktoemploya
bargainingmodel,thismeansthattheparticularsofthesocialcontractdepend
heavilyonthemakeupoftheparticularsocietyinwhichitisbeingenacted.More
specifically,thesocialcontractdependsontheperspectivesoftheagentscreating
thecontract.Toreturntothefoodexample,asocietyofpeoplewhoallkeepkosher
couldverysensiblybanporkandshellfishsoastopreventaccidentalpollutionof
theirfood.Butthismaychangeifpeoplewithnoeatingrestrictionsarealsointhe
population,astheircategorizationofthosefoodswouldnowbeincludedinthe
bargainingprocess.Anotherexamplemightbethecaseofafeministperspectivein
articulatingwhatworkplaceequalitymightentail:ratherthanjustmakingjobs
availabletobothmenandwomen,asmightbesuggestedbyamaleperspective,
governmentpoliciesrequiringmaternityandpaternityleave,andsomesubsidyon
childcaremightberequiredtoensureequalaccesstoemployment.Inlater
sections,thisperspectivesensitivitywillbedevelopedatgreaterlength.
Supportersofpublicreasonanddeliberationalmodelsofpoliticalagreement
mightwanttochallengeabargainingpositiononthegroundthatbargainingdoes
notpromoteasystemofreciprocityandmutualunderstandingandrespect,butis
insteademblematicofamodusvivendi.Onthiswayofthinking,abargainingmodel,


presuminganagreementcouldbereached,wouldmerelyallowindividualsto

125

coexist,butwouldnotdoanythingtopromotedevelopingstrongersocialties.In
thisway,abargainingbasedsocialcontractwouldbetheshallowestformofsocial
union:itwouldmerelyallowindividualstopromotetheirownselfinterestina
societyofotherrationalmaximizers.Thiswouldbeafarcryfromtheidealofpublic
reason,inwhichpublicdebaterelyingoncommonlyheldpoliticaljustifications
wouldservetodrawpeopleclosertoeachotherandfosteragreatersenseofshared
identitythatwouldpromoteasystemofsocialreciprocation.Iwanttochallenge
thisobjectionontwofronts.First,thatbargainingjustpromotesamodusvivendi.
Thebargainitselfisindependentofthedepthofsocialties.Whileitistruethat
bargainscanbereachedbetweenpeoplewhowillinteractonlyonceandhaveno
particularinterestineachother,bargainsdonothaveanyfeaturesthatrequire
individualstohavenosocialties.Itistruethatabargainingmodelwouldpermita
modusvivendiarrangement,butthisseemstobeanadvantage:abargainingmodel
mightfindsuccesswherenonewouldbepossibleinapublicreasonmodel.Apublic
reasonmodelrequiresgreatercommonalitiesamongindividuals.AndasIquestion
thepremiseastowhetherthesecommonalitiesoftenobtain,abargainingmodel
seemsmoreappropriate.
Thereisthefurtherquestionofwhetherabargainingmodeldoesanythingto
promotestrongersocialties.Publicreasonmodelshaveagreatdealoftraction
here:theydoencouragefosteringasharedpoliticalidentity,andthiswouldseemto
encouragereciprocity.Iwouldcontendthatabargainingmodeldoesmuchto


promotesocialties,butinamannerthatisquitedifferentfromthepublic

126

reasonmodel.Inparticular,abargainingmodelencouragesindividualstobetter
understandtheinterestsoftheircounterparts.Betternegotiationoutcomescanbe
hadifallpartiescanbetterunderstandwhattheothersidewants,sotheycanbein
abetterpositiontogiveittothem,andinturnbeabletodemandotherconcessions.
Inotherwords,bargainingencouragesempathy.86Whatsmore,thisempathyis
basednotonsomesharedpoliticalviewpoint,butinsteadanattemptat
understandingtheperspectiveandvaluesofonescounterpart.Forexample,cities
thathavehistoricallybeenmoretradeoriented,suchasportcities,tendtobemore
tolerantthancitieswithlittleactivetrade.87Ratherthansuppressaprivate
sphereversionofoneselfforacommonpublicsphereidentity,citizenscanbring
theirfullselvestobearonsocialconcerns,andbeengagedwithonthosegrounds.
Thisengagementlegitimateseachothersperspectives,asthebargainingprocess
acknowledgesthemasarelevantsetofconsiderationstobereckonedwith.This
validationhasnocounterpartinapublicreasonmodel,asonlythesharedpolitical
perspectiveisconsidered.Inabargainingmodel,alternativeviewpointsare

86
It should be noted that bargaining encourages empathy in the long run. If two people
are randomly paired to bargain only once (as we might imagine in a flea market case),
there is only so much that one party can do to come to understand the other. However, in
cases of repeated bargains, one can get a better sense of what the other party responds to,
and has an incentive to investigate this further. Taking on a cost to learn more about the
other party may not make sense in a single-shot case, but it will often pay dividends in a
repeated interaction. This argument has a long tradition, most notably in Hume and
Smith. A detailed discussion of this view can be found in Hirschmans The Passions and
the Interests. (Princeton University Press, 1997).
87
A recent case study of this is found in Jha, Saumitra, Trade, Institutions and Religious
Tolerance: Evidence from India (January 10, 2008). Stanford University Graduate
School of Business Research Paper No. 2004 Available at SSRN:
http://ssrn.com/abstract=948734.


embracedasrelevant,eveniftheyarenotassuredofcarryingtheday.This

127

promotionofempathyandtheproceduralrespectshowntoalternativeperspectives
woulddomuchtopromotesocialcohesion,preciselybecauseitservesto
enfranchisethoseindividualswhomightotherwisehavehadtheirperspectives
excludedfromthepoliticalprocess.Thisisnotunlikethedouxcommercetheory
advocatedbymanyoftheearlyeconomicthinkers.Smitharguedthatmarket
playersdobestwhentheyarehonestandempathetic.Selfinterestdrivesagentsto
developreputationsforbeingfairandconsiderate,foriftheyarenot,noonewould
wanttodobusinesswiththem.88
Thoughthesourceofjustificationofabargainingoutcomeisultimately
individualratherthancommunal,thisformofjustificationassuresusthat
individualsfromeveryperspectivearewillingtoendorsetheoutcome.Thisisnot
tosaythatthiswillassureeveryonethebestpossibleoutcome,butaswewillseein
latersections,itwillensurethateveryonewillgetthebestoutcomethatcanbe
gottengiventheconstraintsthateveryoneplacesoneachother.

AConceptionofRights
CitizensofboththeUnitedStatesandFranceenjoyreligiousfreedom,which
includestheseparationofChurchandState.Theserights,however,areinterpreted
inaquitedifferentmannerinthetwosocieties.IntheUnitedStates,thestatedoes

88
Hirschman traces the history of the doux commerce thesis in his The Passions and the
Interests (Princeton University Press, 1977). Adam Smith develops his account in The
Theory of Moral Sentiments


muchtoprotecttherightofindividualsreligiousexpression.Whenwe

128

interprettherightsofchurchstateseparation,weseeitintermsofanti
establishment.Thestatecannotdoanythingtopromoteonereligionoverany
other.Butweallowforamarketofreligionscitizensarefreetochoosetheir
religionandlevelofreligiosity,includingtheiroutwardexpressionsoftheirfaith.
France,however,hasadifferentapproachtointerpretingtheserights.In
part,thisisduetothewaythattheFrenchinterprettheconceptofthecitizen.
Namely,individualsareFrenchfirstandeveryotherculturalidentitycomes
second.Individualsareaskedtosubsumeeveryotheraspectofthemselvestotheir
Frenchness,andinreturn,thestategrantscitizenstherightofchoosingwhatever
religiontheywant.Thishashadsomesuccessinparticular,PostNapoleanFrance
washistoricallymorewelcomingofJewishcitizensthanmanyotherEuropean
countriessolongastheysawthemselvesasprimarilyFrench.However,this
interpretationofreligiousrightsdiffersinanimportantwayfromtheUnitedStates.
Namely,itisnotacceptabletohaveoutwarddisplaysofreligiosity,particularlyin
publicinstitutionssuchasschools.Thishasinrecentyearsbecomeasignificant
issuewithFrancesgrowingMuslimpopulation.Thoughstudentscanwear
necklaceswithreligioussymbolsonthemundertheirshirts,theyarenotallowed
moreoutwardexpressionsoffaith,suchashijabs.Muslimgirlshavebeenexpelled
fromschoolforwearingahijabtoclass.Teachershaverefusedtoteachsolongas
studentswearthem.TheMuslimcommunityclaimsthattheserestrictionsarea
violationoftheirrights,butthestaterespondsthatitisapplyingthesamestandard


toeveryone.Ifanything,FranceistryingtoclingtoitsFrenchfirst

129

interpretationofrightsmoreaggressivelynow,asitfacesincreasingethnic
diversity.Thestate,andmanyinit,seethisinterpretationofrightsasawayof
promotingnationalunity.ManyMuslimsseethisinterpretationofrightsas
discriminatory.
WhatmakestheFrenchcaseparticularlyusefulisthatwecanseehowone
rightcanaffectmanyothers.ArighttoreligiousexpressionforMuslimgirlscanbe
seenbyotherFrenchcitizensasaviolationofgenderequality.Havingthisviolation
ofsuchanimportantrightoccurinschoolcanhaveanunderminingeffectonother
nonMuslimgirlswhomighttakeinaconceptionofwomenassubmissiveoras
objectsofsexualdesirethatmustbecovered.Thisrightofexpressioncanbeseen
aspromotionthatthestateisnowendorsingthepracticesofaparticularreligion.
ThiscanmakeotherswhoarenotMuslimuncomfortableorevenfeelpressureto
conformwithMuslimpractices,whichwouldviolatetheirownreligiousautonomy.
Whatisimportanttonoticehereisthatourchoicesaboutinterpretationsofcertain
rightscanimposeexternalcoststhosecostsarelimitationsonhowwemightbe
abletointerpretotherrights.Ifwewanttotakeastrongstanceonrightsofgender
equality,wehavetoscalebackourinterpretationofarighttoreligiousexpression.
Thegoalofthissectionistosetthegroundworkforunderstandinghowwe
canapplyabargainingmodeltothesocialcontract.Mostbasically,thebargainis
abouttherightsandprivilegesthatformthebasicstructureofsociety.Theserights,
aswecansee,cancomeintoconflictwitheachother,andweneedtofindawayto


cometoacommonagreementaboutwhichrightsweoughttoendorseover

130

othersinparticularcases.Butweneedtounderstandwhatexactlyarightis.Aswe
canseefromtheexample,notonlycanweinterpretwhatappearstobethesame
rightintwoverydifferentways,butalsoeachinterpretivechoicehasmany
externalitiesthatcanaffecthowweinterpretotherrights.Thus,wearehere
concernedwiththecorefreedomsanddutiesthatcitizensenjoyinsociety.89
Letusconsiderwhatarightis.Rightsareapportionmentsoffreedomsand
limitationsonactions.Theydeterminewhatindividualsareallowedtodo,howthey
areallowedtodoit,andwhatconstraintsofactiontheycandemandofeachother.
Importantly,rightsalsodeterminewhatactionscannotbeconstrainedbyothers.It
willherebeusefultointroduceastandarddistinctionbetweenkindsofrights:there
arenegativerightsandpositiverights.Negativerightsassureyourfreedomfrom
restraint:therighttofreespeechsaysthatIcansaywhatIwantwithoutanyone
beingabletosilenceme,andtherighttohabeascorpussaysthatIcantbe
imprisonedwithoutcause.Positiverightsincurobligationsonotherstoassureyour
wellbeing:ThelasttwoofFDRsFourFreedomsareanexampleofthis.The
freedomfromwantandthefreedomfromfearareobligationsthatweallhaveto
eachother.Positivefreedomsrequiresomemeasureofgovernmentorsocial
interventionafreedomfromwantimpliesasocialminimum,forexample.So

89
Though I am only discussing rights allocation here, this discussion can easily be
extended to an analysis of economic allocation. I do not develop this here as I am
primarily concerned with questions of political power in this chapter. In the final chapter,
material considerations will play a much larger role in my discussion. Very loosely, this
chapter is addressing what Rawls saw to be the domain of the first principle of justice,
and the next chapter addresses the domain of the second principle of justice.


whilenegativerightstendtobeprohibitionsofgovernmentintervention,

131

positiverightsoutlinegovernmentobligationstowardsitscitizens.
Thedistinctionbetweenpositiveandnegativerightsisanextremelyuseful
one,butitisnottheonlydistinctionthatwemightmake.Importantly,wecanalso
categorizerightsbasedonwhethertheyarerivalrous90ornonrivalrous.Wecan
alsocategorizebasedontheorthogonalcriteriaofwhetherrightsareexcludable91
ornot.Combiningthesetwodistinctions,wecandeterminewhetherrightscanbe
understoodaspublicgoods92,oriftheycomeshortinsomerespect.Rightsthatare
rivalrousimplythatwhenonepersonconsumestheright,ittakesawayresources
fromanother.Manypositiverightshavethischaracter.Forexample,guaranteesof
asocialminimum,orarighttohealthcarewouldberivalrousgoods,simplybecause
theyinvolvetheconsumptionofothermaterial(rivalrous)goods.93Otherrightsdo

90
Rivalrous goods are goods where consumption reduces availability of the good to
others. Normal physical goods, like chairs or food are examples of rivalrous goods. If I
eat a sandwich, you cant eat it, as its now gone. Non-rivalrous goods do not have this
character. When I breathe, I am not depriving you of any air in any meaningful sense.
Likewise, if I swim in the ocean, I am still leaving plenty of ocean for you to swim in.
91
Excludable goods are goods that can be access-controlled. I can let some people into
my home, and keep others out. However, non-excludable goods dont have this
character: it would be impossible to limit the light from street lamps to only those people
who drive American cars, for instance.
92
Public goods are those goods that are neither rivalrous nor excludable. This means that
there is zero marginal cost to providing another person with the good (this condition is
non-rivalry), and that if some people have access to the good, there is no way to restrict
others from having the same access (this condition is non-excludability). A lighthouse is
a classic example of a public good it does not matter how many ships use its light, nor
can the lighthouse prevent ships from seeing its light if some ships can see it. A normal
manufactured good is the opposite of a public good it is both rivalrous and excludable.
93
One could try to make the argument that the cost incurred in providing social minima
or health care is offset by the savings in other areas of government spending, such as
emergency services or prison expenditure. This may very well be the case, but it would


nothavethischaracter,however.Freespeechisnonrivalrous:myabilityto

132

speakmyminddoesnotinanywayinhibityourspeech.Infact,ifanything,myfree
speechcanpromotespeechfromyou,whetheryouwanttoagreeordisagree.
Aninterestingquestioniswhetheranyrightsarenonexcludable.Havinga
righttocleanairisnonexcludable,inthesensethatifyouvesuppliedittoone
group,othersinthesameareacantbedeniedthesameright.Arighttoprotection
fromforeigninvasionissimilar:onecantprotectonlysomepeopleinacountryand
notothersfromsomeforeigninvasion.Eitheryouhaveanarmyoryoudont.
However,mostotherrightsdonotseemtohavethissamecharacter.Police
protectioncouldbeselectivelygranted,forinstance.Evenarighttospeechcouldbe
selectivelygranted:oneclassofindividualscouldhavethatright,whileotherscould
bedeniedit.Thisistrueofrightstopracticeagivenreligion,righttoassembly,ora
righttohealthcareorasocialminimum.Nonexcludablerights,then,areonlythose
rightsthatarealsopublicgoods,intheeconomicsenseoftheterm.Sowehaveseen
thatwhilemostrightsareexcludable,manyofthoseexcludablerightsarenon
rivalrous.Thismeansthatthoughtheserightscanbeextendedtomanyorfew
peoplewithnodirecttradeoffs,therearewaysofpreventingtheserightsfrombeing
extendedwidely.
Whymightsomeonewanttonotextendrightstoagroup,especiallywhen
therearenodirectcostsassociatedwithsuchanextension?Therearethree

have to be a perfect offset in every case for such an argument to work. Being that we are
in the end dealing with material goods, which are always rivalrous, this argument is not
going to be successful.


133
plausiblereasonsthatIwillconsider.First,onemightgetsatisfactionfromthe
denialofrightstoothers.Second,onemighthavethepaternalistinclinationthatit
isbestforthegroupinquestiontohavearestrictedsetofrights.Third,onemight
findsomeaspectoftheserightsinconflictwithsomeothersetofrightsthatone
takestobemorevaluable.Thefirstreasonrequiresaparticularpsychological
disposition:notonlydoesonecareaboutothers,butonecaresabouttheminsofar
asonecanhurtthem.Thiscannotbeapreferenceofanegoistoranaltruist.So
whilethispreferenceappliesonlytoalimitedportionofthepopulation,itisa
possiblepreference.Inparticular,membersofhategroupscouldbeclassifiedas
havingthiskindofpreference.94Theyoppresspeoplesimplybecausetheyget
satisfactionfromtheoppressionitself.Otherwise,therearenorelevantbenefits.
Thecaseofpaternalismissimplyoneinwhichthepaternalistmustassume
thattheyknowbetterthanothersabouthowtobesthelpthem.Arestrictionin
rightsstemsfromthebeliefthatacertainclassofpeople,lefttotheirowndevices,
willmakepoorchoices.Thiscanhaveuglyconsequences,ascanbeseenbysome
paternalistargumentsinfavorofracialsubjugation.95However,inotherinstances,
therecanbecasesofpaternalismthathavemorejustification,suchasmandatory

94
Though this classification is possible, it may be incorrect. Another plausible reason for
being a member of a hate group is that one believes that once you grant rights to the
group whom you are oppressing, these newly-granted rights will inhibit or diminish your
own rights. Which would then move us two our other reason for not extending rights.
95
A striking example of this is found in Thomas Carlyles infamous essay Occasional
Discourse on the Negro Question Fraser's Magazine for Town and Country, London,
Vol. XL., (February 1849). Mill famously responded by challenging the paternalist
approach to a denial of rights in his essay The Negro Question in 1850 in the same
magazine.


134
seatbeltlawsandsmokingbans,whereinminorrightsrestrictionscomewitha
fairlysizablepublicbenefit.
Themoreinterestingcasetoconsideriswhererightsarenotextendedfully
becausetheycomeinconflictwithotherrights.Thisisthecasethatrequiresamore
nuancedunderstandingofwhatrightsthemselvesare.Therearetwocasesto
consider:restrictingthenumberofpeoplewhoshareinaccesstoaparticularright,
andrestrictingthescopeoftherightitself.Toillustratethefirstcase,consider
votingrights.IntheUnitedStates,wehaveasignificanthistoryofrestrictingvoting
rightstoonlycertainclassesofpeople.Originally,onlylandowningwhitemales
couldvote.Whilewomenandminoritiesarenowgrantedvotingrights,westill
significantlyrestrictvotingrights.Thismainlyhappensintwopopulations:minors
andfelons.Individualsunder18arenotallowedtovote,nomatterhowpolitically
activeandinformedtheymaybe,simplybecausetheyarenotyetoldenough.This
istrueinallstates.Insomestates,felonswhoareoutofprisonarenotallowedto
voteeither.Werestrictthevotesofminorsandfelonsinpartbecause(rightlyor
wrongly)webelievethattheycannotbetrustedtoputthepublicgoodfirstintheir
vote.96Inthisinstance,rightsarerestrictedeventhoughtheyarenonrivalrous
becausetherestofsocietydeemsthatthedisenfranchisedpopulationhasnot
earnedthoserights.

96
In the case of felons, another likely source of justification is that their loss of suffrage
is part of their punishment for committing serious crimes.

135
InBelgium,however,rightsarerestrictedtoparticulargroupsnot

becausetherestdonotdeservethem,butratherdifferentculturalgroupshave
interestsindifferentsetsofrights.Inparticular,culturallinguisticrightsare
managedbyCommunityratherthanbythefederalgovernment.Sothedifferential
distributionofrightsismeanttoreflectunderlyingdifferencesinthepopulations,
notastatementaboutpowerdifferencesorequaltreatment.However,theBelgian
casecanbeexplainedintermsofrivalry.Ifarightissuchthattherearecosts
incurredforprovidingittothemarginalindividual,thenitisreasonabletosupply
therightonlytothosewhowantitatleastasmuchasthemarginalcost.Providing
therighttoeveryoneonlyservestoreducethenumberofotherrivalrousrightsthat
onecouldotherwiseprovidethatpeoplewouldvaluemore.So,whenoneis
consideringrightsthatarerivalrousinnature,itisentirelyconsistenttohaveboth
differentialdistributionofrightsandequaltreatmentofallcitizens.Butwhen
rightsarenonrivalrous,differentialdistributionofrightsneedsfurtherexplanation.
Letusnowconsiderthesecondcaseoflimitingrights,wherewelimitthe
rightsthemselves,ratherthanthescopeofpeoplewhohavethem.Toillustrate,
letsconsidertherighttofreespeech,andtherighttopersonalbodilyintegrity.We
havefreespeechrights,whichshouldentailthatwecanmakewhateverclaimswe
maywant.Butifweshoutfire!inacrowdedroom,wewillcausepeopleto
trampleovereachother.Similarly,ifweshoutTerrorist!Hehasabomb!ona
plane,thepersonwepointtowilllikelybetackledandbeaten.Inbothofthese
circumstances,allowingforthefullestversionofarighttofreespeechwould


136
adverselyaffecttherightsofothersinparticular,theirrighttomaintainbodily
integrity.Sowehaveaconflictbetweenfreespeechandbodilyintegrity.This
conflictraisesanimportantissueaboutthenatureofrights.Weneednotnow
decidethatthereisnorighttofreespeech,orthatofbodilyintegrity.Instead,we
canchoosetoreinterpretoneofthoserightssoastoeliminatetheconflict.
Reinterpretingrightssuggeststhattheycanbebrokenupintocomponent
parts.Soratherthanhavingasingleentity,freedomofspeech,wecaninstead
thinkofthisrightasbeingabundleofmanyallowances.Onesuchallowancemight
betherighttorunapress,anothermightbetogiveaspeechatastreetcorner,and
anothermightbeinducingpanicsincrowds.Thoughtheseactivitiesareallrelated
toeachother,theyarenotmutualrequirementsforoneanother:itisperfectly
consistentthatonemightbeallowedtospeakatastreetcorner,andnotallowedto
inducepanicsincrowds.Similarly,propertyrightscanbeabsolute,orwecanlimit
thoserightsinsuchawaythatmaketaxationacceptablewhilestillprohibiting
stealing.Sotheconceptofarightmustshiftawayfrombeinganabsolute,andmore
towardsabundleofmanyaffiliatedallowancesforactionorguaranteesthatmaybe
consideredatleastsomewhatindependentlyofeachother.
Thisconceptionofrightsasbundlesofallowancesandguaranteesmakes
themfirmlyasocialconception,asopposedtoametaphysicalone.Weagreetowhat
ofeachbundlewetaketobeimportant,andwhatwecandiscard.Thisisevidentin
thedifferentialallocationofrightsacrosscountries.EvenWesternliberalcountries
varyamongsteachotherintheirrightsallocations.Ifweviewedrightsas


137
somethingotherthananagreementamongstcitizens,wewouldexpecttohave
activedebatesbetweenfriendlycountriessoastodeterminethecorrectsetof
rights,justaswedowithscientificdisputes.ThereisnoFrenchconsensusonthe
atomthatdiffersfromtheBritishorAmericanonethereisjustthescientific
consensus.97
Whatwewillturntonextishowwecandevelopanaccountformakingthese
kindsofdecisions.Itiseasilyseen,however,thatthisrequiresthatweviewrights
assubjecttorevision,andmoreimportantly,subjecttotradeoffs.Thoughwemight
caredeeplyforthefreedomofspeechifwethinkaboutitingeneral,notmany
amonguswouldsupporttheideathattherighttoinducepanicincrowdsismore
importantthanbodilyintegrity.So,wearewillingtomodifyourconceptionofthe
freedomofspeechtoeliminatesucharightsoastogetridoftheconflict.Thus,the
conceptionofrightsIdevelophererequiresthatrightsbebundlesofassociated
allowancesandguarantees,andtheseareallsubjecttotradeoffsamongstother
bundles,sincewecannothavetworightsthatconflictwitheachother.98
Atthispoint,wehaveseenthatabargainingmodelhassomeadvantages
overapublicreasonmodelofasocialcontract.Wehavealsooutlinedaconception

97
This is not to say that evidence of debate shows that there is no underlying truth.
Rather, I suggest that a lack of debate in the face of differences suggests that these
differences are not caused by different attempts to uncover hidden truths, but instead a
recognition that the differences represent different social agreements.
98
Having conflicting rights is incoherent: if rights are to be respected or even just always
followed, then if two rights require conflicting sets of actions, at least one will not be
followed. But then rights would not be notably different than just sets of guidelines that
we can choose to follow if we so wish. For rights to have the inviolable character that we
think they have, they must be mutually consistent.


138
ofrightsthatallowsforcategorizationalongtheeconomicconceptionsofrivalry
andexcludability.Employingthesedistinctions,wecanseehowrightscanbe
tradedoffamongsteachotherandinfactthattheymustbetradedoffsoasto
assureaconsistentsetofrights.Withtheseconceptsinmind,wecannowturntoa
fullelaborationofabargainingmodelofjustice.

TheBargainingModel
Aswehaveseen,thestartingassumptionofapublicreasonmodelisthat
thereissomesuitableperspectivethateveryonecanadopttoreasonascitizens.
Thisisproblematicpreciselybecauseindividualsdonotallsharethisperspective,
andsincedifferentperspectivescanhavedifferentassociatedcategorizations,
individualsmightnotevenbeabletoaccuratelyrepresenttheirownpolitical
interestsinanonnativeperspective.Ourgoalhereistocorrectthismain
deficiencyofthepublicreasonmodelbyallowingindividualstoengageeachother
withtheirownperspectives.Inparticular,inthissectionIwilllayoutabargaining
procedurethatallowsindividualstoagreeonaparticulardistributionofrights
withouthavingtoagreeonacommonperspectivefromwhichtodebate.Todothis,
Iwillrelyoneconomicaccountsofbargaining,thoughIwillextendthemtosatisfy
multipleperspectives.
Thefirststepistoexaminetheconsequencesofperspectivismforan
economicdescriptionofbargaining.Intheprevioussection,weestablishedthat


139
rightscanbetreatedasbundlesofassociatedguarantees.Innormaleconomic
theory,wewouldfirstdefinestatesoftheworldbypartitioningstatespace:we
divideupthestateoftheworldintomanypieces,andeachpartitionidentifiesa
particularscenario.Bundlesofrightsarethesetofactionsthatwehavethatcan
applytoparticularpartitionsoftheworld.Thescopeofagivenrightcanbethought
ofashowmanypartitionsofthestatespaceitappliestowhenweareallowedto
performthegivenaction.Anassumptionhereisthateveryoneseesthesame
partitionsofstatespace,andcanseehowabundleofrightsapplytotheworld.99
Whentwobundlesofrightsaffectoverlappingportionsofstatespace,wecanhave
conflictsbetweenthetwobundles.Thatis,rightscancomeintoconflictwitheach
otherwhentheyapplytothesamestateoftheworld,anddescribedifferent,
incommensuratesetsofallowableactions.100InthecaseofhijabsinFrance,an
accountofreligiousfreedominterfereswithgenderequality.Soarighthastwo
components:first,alistoftheallowableactions,andsecond,thelistofpartitionsof
thestatespacethattheseactionsareallowedin.Inthosecaseswherethereare
conflictsbetweensetsofallowableactions,theagentcanthenhavepreferences
overtheserightsintheparticularpartition.Forexample,onecanratefreespeechin

99
This is slightly incorrect. In fact, the standard axioms require that if Alice and Bob are
playing a game, Alices partition set must be a subset of Bobs. This means that one
agent may make finer distinctions than the other player, but he recognizes all the
distinctions that she does. An example of this would be that Alice recognizes sunny days
and cloudy days, but Bob recognizes sunny days, cloudy days and partially cloudy days.
Bob can make finer distinctions since recognizes the difference between sunny days and
partially cloudy days, and thus make more careful choices about his actions than Alice
can, but he does not lack any categories that Alice has.
100
In standard economic language, this would be ones strategy set. A strategy set
defines the set of actions (also known as strategies) available to an agent.


crowdedroomsaslessimportantthanbodilyintegrity.Forpartitionsofthe

140

statespacewherethereissomepotentialconflict,agentshavetodecideifthey
preferrightAtoB,BtoA,orareindifferentbetweenAandB.Finally,onecanhave
preferencesoverthepartitionsthemselves:thatis,onecandeterminewhichcases
theydeemtobemostimportant,andwhatcasestheycareabouttheleast.Onemay
stronglycareaboutfreespeechwhenitcomestotheabilitytohaveaprintingpress,
butnotcaremuchwhenitcomestotheabilitytosaywhateveryouwanttoyour
elementaryschoolteacherinclass.
Undernormaleconomictheory,wehavesufficientpiecestodescribea
bargain.Individualshavecompletepreferenceorderingsoverallthecasesofrights
conflict.Agentscanthenattempttoconcedeonsomecasesthattheycarelessabout
towinconcessionsonthosecasesthattheycarethemostabout.Eachagentsaimis
tofindthebargainingoutcomethatmaximallyincreasesherownutilitywiththe
constraintthatherbargainingpartnerwillagreetothearrangement.Atthebare
minimum,thismeansthatneitherpartycancomeoutoftheagreementworseoff
thanhowtheystarted.Butthisleavesopenahugerangeofpotentialbargaining
outcomes.Ifwetakeseriouslytherequirementthatbothpartiesseektomaximize
theirreturnssubjecttobeingabletogetagreementfromtheotherparty,wearein
factleftwithasinglesolutiontheNashBargainingSolution.TheNashsolutionis
thebargainingoutcomewheretheproductoftheutilitygainsismaximized.
Maximizingtheproductratherthan,say,thesum,helpstoensurefairness:each
agentisgainingatasclosetothesamerateastheotheragentsasispossible.While


thismightseemlikeaconvenientmathematicaltrick,itrepresentsadeep

141

thought:itcapturesthewillingnessofagentstoacceptaparticularbargain.The
Nashsolutionistheuniqueefficient101solutionthatsatisfiessomebasicaxiomsthat
supportthisconditionofmaximizationsubjecttoagreement.102Soeachagentis
maximizingtheirutilityasmuchastheycan,subjecttotheconstraintthatothers
havetoagreetothebargainaswell.Ifanyagentproposesabargainthatincreases
theirownutility,atleastoneotheragentwouldbelosingrelativetotheNash
solution,andthusrejectthebargain.Thesolutiontreatsallagentsequally,andso
servesasafocalpointforthebargain.103
Thisbargainingapproachhassomenicefeatures.Inparticular,theNash
Solutionprovidesthebestdealthatanyindividualcangetsubjecttobeing
constrainedbyothers.Butitisdonewithouthavingtodoanyinterpersonal
comparisonsofutility.Eachagentcanhaveutilityonwhateverscaletheywant,and
atnopointdotheyhavetotrytointerpretsomeoneelsesutility.Theyjusthaveto
worryaboutsomeoneacceptingtheoutcome.Aswesawintheearlierdiscussionof
mutualbenefitmodels,thisistakencareofinvirtueoftheinteractionofrational

101
In this context, efficient means that there are no pareto moves left to make. That is,
any gain must be associated with a loss by someone else.
102
The first axiom is invariance to change in how one represents their utility (invariance
to affine transformations), which means that it does not matter how you describe your
utility function, just what it is. The second axiom is that it is pareto optimal. The third
axiom is that it is independent of irrelevant alternatives. This means that if you prefer
Apples to Oranges when those are all thats available, you wont all of the sudden prefer
Oranges to Apples when you can also choose Bananas. The fourth axiom is symmetry.
103
For a much more extensive justification of the Nash Bargaining Solution, see chapter
7 of H. Peyton Youngs, Equity: In Theory and Practice, Princeton University Press,
Princeton, NJ 1995


142
agentsitself.Butwestillrunintoproblems.Inparticular,alltheagentshaveto
sharethesamepartitionsets.Ineffect,thisisthesameassayingthattheymust
havethesamecategorization,whichseriouslyconstrainstheirabilitiestohave
differentperspectives.Sowemustfindawaytomodifythisbargainingprocedure
toaccountformultipleperspectives.
Therearetwochangesthatbecomenecessary.First,wemustfindawayto
dealwiththefactthatdifferentindividualscanhavedifferentcategorizations.
Second,wemustweakenourclaimsaboutthecompletenessofpreference
orderings.
Iftwoagentshavedifferentcategorizations,thismeansthattheywillnot
dividestatespaceupinthesameway.Thisissimplybecausetheycategorizethe
worlddifferently,andsotheyhavedifferentconceptionsofsalience.Aswesawin
thepreviouschapter,theyareseeingthesamething,butinterpretitdifferently.As
thereisnoneutralrepresentationofthestatespace,wecannottakeonepartitionas
betterthananother,andsowecannotjustpickoneandrequireeveryonetouseit.
Thismuchisclear.Butwealsocannotjusttaketheunionofeveryonespartition
setsandexpecteveryonetosharethislargerpartitionset.Thisisbecausethereis
noassurancethatpeoplecanseethedistinctionsthatareinotherperspectivesif
theydonotthemselvesholdthem.Sojustaswedonothavethesameaccountsof
value(andhenceutilityfunctions),wedonothavethesamepartitionsets.Buteven
thoughwecannotexpectthatindividualscanadoptthisunionofallpartitionsets,


wecandescribetheoverallbargainbymakingreferencetothisunionset,so

143

longaswedonotassumethattheagentsinthebargainthemselvesareawareofit.
Thoughwecanusetheunionofpartitionsetsinourmodelofthebargain,it
maybeusefultoconsideranalternativerepresentationofthesameidea.In
particular,wecanthinkofthecategorizationsofaperspectiveasdimensionsof
somehighdimensionalobject.Eachdimensionisacriterionbywhichonecanmake
distinctionsaboutstatesoftheworld.So,onedimensionmightbeweather,another
mightbedayoftheweek,athirdmightbehowmanypeoplearepresent,andsoon.
Wecanaggregateallagentsperspectivesbysimplyaddingdimensionsforevery
wayofcategorizingthattheseperspectiveshave.So,wehaveasinglehigh
dimensionalspacethateachindividualcansee,butonlybytakingalower
dimensionalprojection,asrepresentedbytheirparticularperspective.Thishigh
dimensionalspaceisnotametaphysicallyrealentityitisjusttheagglomerationof
alltheagentsperspectives.Thisspaceallowsustoseehoweveryoneisstillinthe
endtalkingaboutsomethingshared,eveniftheydonotagreeonhowtodescribeit.
Whatisimportanttonoticeisthat,asweeachcanseeonlyaprojectionof
thislargerspace,whenAlicewalksintothestoreandseesjustfood;Bobcanwalkin
andseemeats,grains,fruits,vegetablesanddairy;andCarolcanseecannedfood,
frozenfoodandfreshfood.Theyareallstilllookingatthesameobjects,andthey
areeachmakinglegitimatecategorizationstheyjustdontrecognizeeachothers
categories.Thiscapturessomethingimportantaboutpoliticaldiscourse.
Disagreementdoesnotstemonlyfromdifferencesinpreferences.Itcanalsocome


144
fromdifferencesinhowpeopleinterprettheworld.Soitisessentialtocapture
thissourceofdisagreementifwearetocreateanaccuratemodelofpolitical
negotiation.
Thus,thebargainitselftakesplaceoverdisagreementsinpreferencesabout
rightsallocationacrossthisspace,eventhougheachagentcanseeonlytheir
particularcategorizationofthedisagreement.Agentsdonotseethemselvesas
engaginginthesamebargaintheywoulddescribethebargainratherdifferently
fromeachother,basedontheirownperspective.Further,becausethese
categorizationsaredifferent,preferencesaregoingtobeoverdifferentthingsas
well.Thougheachagentcanhaveacompletepreferenceorderingofrightsacross
statesoftheworldontheirownperspective,theymusthaveonlyapartialorder104
acrossthewholehighdimensionalstateoftheworld,sinceitincludesmany
distinctionsthattheagentdoesnotrecognize.Itisimpossibletostatepreferences
aboutobjectsthatonedoesnotrecognizeasexisting.Oneconsequenceofthisis
thatifAlicedoesnotrecognizesomedistinctionthatBobmakes,Aliceisgoingtobe
morethanhappytoletBobhavewhatheprefersinthatcase,sinceAliceliterally
cantseewhatBobwants.105
Withthesetwomodifications,wenowhaveabargainingmodelthatallows
formultipleperspectives.Eachagenthasapartialorderoverrightsintheunionof

104
A partial order is one where an agent cannot say whether they prefer A to B, or B to
A, or are indifferent between the two of them, for some A and B in their choice set.
105
This is true in those cases where Bobs preference does not affect any outcomes that
Alice can recognize.


allpartitionsets,whichdefinestheirutilityfunction.Theyalsoeachhavea

145

partitionsetthatisastrictsubsetoftheunionset.Evenifagentsdonotagreeabout
whattheyarebargainingabout,theycancarryoutabargain.
Whatenablestheseagentstocarryoutabargainisthatthereisalwaysa
pricetoallocatingagivenright.Evenifdifferentagentsinterpretrightsdifferently,
apricecanbeestablished,sinceeveryonehasadefinedutilityfunction.In
particular,thisanalysisofasocialcontractintermsofabargainallowsthesociety
engaginginthebargaintodiscoverexactlyhowmuchcertainrightsareworthto
peoplewhoholdagivenperspective.Todiscussthisimportantidea,Iwouldliketo
examinetwoissues.First,howwefindaprice,andsecondwhatthisprice
mechanismmeansforrightsallocation.
Thatrightshaveapriceisaperhapsuncomfortableclaimtomake.Rawls
andothercontracttheoristsdenycitizenstherighttoselltheirrightsandfreedoms
forotherbenefits.Oneisnotallowedtomakeanytradeoffsatall.Itoowouldlike
torestrictthesaleofrights.However,thatdoesnotmeanthattherecannotbe
tradeoffsbetweenrights.Asmentionedearlierinthedescriptionofthemodel,I
assumethatagentsplacevalueonparticularinterpretationsoftheirrights.Ifurther
assumethattheutilityonegainsfromrightsisfungible,sothattradeoffsbetween
rightsarepossible.ThisisjusttosaythatImightbewillingtoacceptalimitationon
myrighttobeararmsifIcanhaveincreasedrightstofreespeech.

146
Withjustthesetwoassumptions,wecanestablishapricemechanismfor

rights.Thepriceofarightisjustitssocialcost.Thesocialcostofarightisdefined
astheutilitylossincurredbytheconflictswithotherrightsbundles.Thesocialcost
ofastrongrighttogenderequalityinFranceisthentheutilitylosssufferedfrom
theperspectiveofthosewhowantgirlstowearhijabstoschool.Somerightsmay
havenosocialcost,andarethusfree.Averybasicversionoffreespeechislikelyto
imposenosocialcosts,asitdoesnotimpingeonanyrights.Itisevenpossiblethat
itcouldhaveanegativecostifithelpedenableotherrights.Buttherighttohate
speechcanhavehighsocialcosts,asitimpingesonvariousrightstoequality.
Thisaccountofapriceforrightsrequiresnothingmorethananormal
marketdiscoverymechanism.Theagentsbargainingwillacceptorobjecttovarious
rightsbundlesbasedontheirpersonalutilityassessment,andassuch,theywill
revealtheirvaluationsoftheserightsoverthecourseofthebargain.Themarket
priceofagivenrightisjustthatpriceatwhichthesocialcostsoftherightare
covered.Itisonlyatthispointthatotheragentsinthebargainwillacceptthe
impositionofthatright,sincetheyareotherwiselosingmoreutilitythantheyare
gaining.Thepriceofarightisthepointofindifferencebetweenhavingandnot
havingtheright.Bundlesofrightscanthenbeusedtohaverightsthatoffseteach
other.Aweakerrighttopropertymightbebundledwithastrongerrighttohealth
care,forexample.Soratherthanspendmoneyorsomephysicalresourcetopay


foragivenright,wecanthinkaboutpayingforarightwiththeallocationor

147

removalofotherrightsinaparticularbundleofrights.106
Thispricemechanism,unlikethepublicreasonmodelofsocialcontracts,
allowsagentswithdifferentperspectivestoengagewitheachotherinawaythat
doesntprivilegeanygivenperspective.Sincepriceisjustthesocialcostofaright,
thiscanbeevaluatedfromtheperspectiveofanygivenagentiftheagentislosing
morethanshegains,thenshepullsoutoftheagreement.Butsheneverhastoagree
onwhatitisthatsheisbargainingabout.Thisisnotpossibleunderapublicreason
model.Inaddition,asopposedtojusttalkingaboutthebalanceofreasons
supportingaparticularright,wecandeterminehowmuchaparticular
interpretationofarightisvaluedbylookingatwhatasocietymightbewillingto
giveuptogetit.
Additionally,thepricemechanismembeddedinabargainingmodelallowsus
totreatdifferentkindsofrightsdifferently.Inparticular,wecanidentifythose
rightsthatarepublicgoodsandthosethatareexcludableandrivalrous.Rightsthat
havethecharacterofpublicgoods,orareevenjustnonrivalrousarethosetobe
allocatedtoeveryoneinsocietyatonce.Publicgoodsbydefinitioncanonlybe
providedtoeveryone,andnonrivalrousgoodscostthesametoprovideto

106
This particular form of payment is known as logrolling in the political science
literature. Logrolling occurs when legislators agree on voting patterns across a series of
bills. An example of this is when Alice agrees to support Bobs budget bill if Bob will
then support Alices climate change bill. Of course, logrolling can involve many more
than two people and two bills. The idea is that legislators need to get enough votes on
their bills, so they agree to exchange votes that they would not normally make so as to get
what they want at a price they are willing to pay.


everyone,sothereisnoreasontonotallocatethemfully.Butrivalrousgoods

148

arealsoexcludablegoods,andthoserightsthathavethischaractercanbemore
selectivelyallocated.So,onlythoseindividualswhovaluerivalrousrightswill
choosetopayforthem.Thiswillresultindifferentpeoplehavingdifferentrights,
butasdiscussedbefore,thisisnotanindicationofdiscriminationorinequality.
Insteaditisarecognitionofdiversity.Whatassuresthisclaimistheprice
mechanismitself.Thosethatwantadditionalrightsmustpayforthesocialcostof
theserights.Whilerightsthatarepublicgoodsmustbesupportedbyeveryone,
rivalrousrightscanbeexclusive,andsowecanallowforthoseagentswhowant
themtohavethem,andthosewhodonottonothavetopayforthem.

Conclusion
Buildingontheaccountofperspectiveslaidoutinthepreviouschapter,we
canseethatapublicreasonaccountofthesocialcontractisunabletoaccommodate
diversityinawaythatisneutralacrossperspectives.Asasocialcontractismost
importantlyanembeddingofasetofrights,weexaminewhatwemeanbyrights.
Onceweadoptaconceptionofrightsthatconsidersrightstobebundlesof
associatedinterpretationsoftheright,andaclassificationofrightsbasedonrivalry
andexcludability,wecandefineabargainingprocedurethatcanfullyaccommodate
perspectivaldiversity.TheresultingsocialcontractisaNashBargainingSolution,
whichensuresthatitistheoutcomethatensuresthateveryonereceivedascloseas


theycouldgettothesetofrightsthattheymostwanted,giventhateveryone

149

elsewasouttodothesamething.Thiskindofjustificationisdifferentinkindthan
thejustificationfoundinpublicreasonmodels,asitisindividualratherthansocial.
Inthepasttwochapters,wehavelookedatatwostageprocessfor
determiningmoralprinciplesandestablishingasetofrights.Thesetwostagesare
thefoundationforaniteratedsearchprocessforsocialimprovement.Butweneed
onefinalstagetoprovidethenecessaryinformationtofeedbackintothebeginning
oftheprocess.Namely,weneedtoexaminehowwetakeourgivenmoralprinciples
andrightsallocationsandthenapplytheminactualsocialcooperation.Thenext
chapteristhusconcernedwithissuesofsocialcooperationanddistributivejustice
indiversesocieties.Itakethisnextstagetobewhereourthoughtexperiments
becomeactualexperimentsinliving.Itisfromtheirexperiencesinactuallytrying
tofollowwhattheyhavedeterminedinthepreviouschapters,alongwiththe
distributiveissuesthattheyfacethatagentscandeterminewhethertheywanttore
evaluatethesocialcontract,orcontinuetoabidebyit.

150

Chapter4:
ExperimentsinDistributiveJustice

Introduction
Inthepreviouschapter,weexaminedamechanismforachievingagreement
abouttheallocationofrightsinsociety.Thisbargainingmechanismisdesignedto
workevenifdifferentpartiesdisagreeonthetermsofthebargainitself.The
allocationofrightsthatresultsfromthisbargainingmechanismshouldbethought
ofasanagreeduponandlargescaleexperimentinlivingitissomethingthatwill
betried,butisexpectedtobesubjecttorefinementorevenlargerrevision.The
particularallocationofrightsmayturnouttobeclosetoidealforeveryone,orsome
aspectofitmayturnouttohaveunintendedconsequencesthatareunacceptableto
atleastsomepartofthepopulation.Ifthesystemofrightsisfoundtobewanting,
thenthedifferentgroupsinsocietycancollectivelydeterminewhetheritisbestto
returntothestagesofdeterminingasocialcontractfoundinchapters2and3,or
splitthesocietyintomultipleindependentunits.Asweareepistemicallylimited
beings,theonlywaythatthesuitabilityoftheallocationofrightstotheparticular
individualsinsocietycanbedeterminedisbymeansofactuallyimplementingitand
examiningtheconsequences.
WhatIaimtodointhischapterisdemonstratehowthisprocedureof
evaluationmightlookbyexamininghowdistributivequestionscanbenegotiated
onceaschemeofrightshasbeenestablished.Aspropertyrightsaresocial


151
constructions,therangeofpossiblepropertyregimesisquitebroad.However,
giventhatthestructureofthebargainissuchthatthereisfullinformationavailable
toallagentsabouttheircurrentstatus,andthatthebargainsupposesthatallrights
arepricedbasedontheharmsimposedonothers,wecanassumethattheactual
choicesthatcouldbemadeaboutaschemeofpropertyrightsismorelimitedthan
thesetofalllogicalpossibilities.Inparticular,moreextremeversionsofproperty
rights,bywhichImeancompletecollectiveownership,orabsoluteprivate
ownership,wouldnotbeselected.Thisisfortworeasons.First,itisveryunlikely
forapopulationtochooseeitherschemeofrights.Second,itwouldbeextremely
difficultforasocietytoshiftfromsomeothermoremoderateschemetoanextreme
systemofpropertyrights.Toseewhythisisthecase,letusconsiderthebargaining
implicationsofsuchrightsregimes.
ALockeanorNozickeanschemeofabsolutepropertyrights107wouldlikely
befavoredbytherichestinsociety,asitwoulddothemosttoassuretheprotection
oftheirownwealth.However,suchaschemeofrightswouldimposeagreatdealof
harmuponthepoorestinsociety108,andatleastsomeharmonthoseinthemiddle,
becauseofthelossofthestatesabilitytofundsocialprograms.Sotheonlyway
thatthispropertyschemecouldbeawardedisiftherichsacrificedagreatdealin
otherwayseitherbygivingintoallocatingrightstoothersthattheywould

107
By absolute I mean that the state does not have rights to tax or expropriate property.
Under such a property rights regime, market outcomes are final.
108
The poor are harmed in this scheme because the state has no wealth to redistribute
toward them. They are left with market outcomes, and the state is unable to guide
additional resources to provide a safety net.


otherwisereject,orgivinguponrightsthattheythemselveswouldotherwise

152

advocatefor.109Bythesametoken,astrongversionofcollectiveownershipof
propertywouldalsobeverydifficulttoachieve.Itisconceivablethatthepoorcould
viewthisschemeaspreferabletomanyoftheotherpossiblerightsallocations,and
thushavesomeinterestinpushingforit.However,therichwouldlosesomuchthat
theonlywaytheywouldagreetothebargainingoutcomeisiftheyearned
concessionsonawidevarietyofotherrights.Assuch,thestructureofthebargain
wouldgenerallysupportonlymoremoderateschemesofpropertyrights,which
wouldinvolve(limited)privateproperty.110Theseschemesofpropertyrights
wouldinvolvefewerharmsimposedonagivenclassofindividuals,andsorequire
fewerconcessionsonsomeotheraspectoftheoverallbargain.
Moreimportantthantheparticularschemeofproperty,however,isthatfor
thepurposesofthisdiscussion,wecantakeforgrantedthatthereexistsacurrent
distributionofpropertyrightsinthepopulation.Theprocedureoutlinedinthe
previouschapterresultsinafullsystemofrights,andthediscussioninthischapter

109
For example, the rich could lose their ability to gain political office, or lose some
speech rights, like being able to donate money to campaigns or purchase political
advertising. So while their property rights would be increased, the value of those rights
would decrease in that their economic power could not be converted into political power.
110
It is possible, of course, that in particular societies, we could find instances of wide
agreement on more extreme versions of property rights. For example, a society made up
of Ayn Rand adherents would likely want absolute property rights, regardless of current
wealth and income. I submit that this would be a very uncommon phenomenon, if we
assume that states have even moderately sized populations. However, on those occasions
when the particular makeup of the population would support these extreme versions of
property rights, the remaining analysis is unaffected.


assumesthatsystemofrights,andexaminestheconstraintsonasystemof

153

distributivejusticeinadiversesociety.

TwoCompetingImagesofSocialBonds
Iwouldliketoopenwithtwoimages.First,wehavetheOriginalPosition
easilyoneofthemostcompellingthoughtexperimentsinpoliticalphilosophy.In
thisthoughtexperiment,Rawlsasksustoimagineourselvessteppingbehindthe
thickveilofignorance,whereweloseknowledgeofnotonlycurrentfactsabout
thesocietywelivein,butwhoweareinthissocietyourupbringing,our
education,ourrace,sex,gender,ourinterests,oursocialandeconomicstatus,and
evenourconceptionofthegood.Wearethenaskedtoimaginehowwewouldpick
therulesgoverningthebasicstructureofsociety.Rawlsarguesthat,givenour
massiveignorance,ourrationalitywouldcompelustochosethatbasicstructure
thatwouldmaximizethepositionoftheleastwelloffmembersinsociety,sincewe
wouldberighttofearthatwewouldstepoutoftheveilofignoranceandfind
ourselvesstuckinanintolerablelife.TheOriginalPositionisthususedtojustifythe
choiceofthedifferenceprincipleasbeinganessentialpartofRawlssecond
principleofjustice.
Forthisthoughtexperimenttowork,Rawlsmustmakethefurtherclaimthat
whenwereflectonouractualpositioninsociety,andonboththosewhoseposition
isworsethanoursandthosewhosepositionisbetterthanours,werecognizethat


muchofthisdifferenceisduetofactorsthataremorallyirrelevant.In

154

particular,wedonotdoanythingtodeserveourparents,orhowwearebroughtup,
whatresourcesandopportunitiesweremadeavailabletous,ournaturaltalents.
Thatwebecomebetterorworseoffinourlivesislargelyduetothenaturallottery
allthosethingsoverwhichwehavenocontrol,butnonethelessshapeourlives.
Alongsidethisclaim,Rawlsclaimsthatwearefreeandequalmoralagents.Thatis,
despitethedifferenceofwherewehaveendeduponthescaleofsocioeconomic
status,eachofushasequalrightsandclaimsovertheprimarygoodsthatenableus
toleadthekindsoflivesthatwevalue.Critically,notonlymustweinfactbefree
andequalmoralagents,butalsowemustrecognizethatweallarefreeandequal.
Thus,aconsequenceofthethoughtexperimentisthatwemusttreateachotheras
equallydeservingoftheresourcesneededtolivethelivesthatwefindvaluable.
ThesecondimageIwouldliketoconsiderisReaganscondemnationof
CadillacdrivingWelfareQueensasanargumentforwhytheUnitedStatesought
todrasticallycutbackonitssocialsafetynet.Leftimplicitinthisimage,andmany
otherslikeitofferedbyconservativepoliticians,isthatthesocalledwelfarequeens
areminorities.
Reaganwasverysuccessfulwiththisrhetoric,andthisstyleofargumentcan
beseenthroughoutmanypolicydiscussions,particularlywhenthereisan
identifiableOtherthatcanbeattacked.111Incurrentpoliticaldiscussions,

111
This rhetorical strategy is often more effective if the social safety net is not universal,
but rather targeted at the poor. If the comparatively well off only lose money on welfare,


immigrantsbringdiseases,aremorecriminal,arelazier,orareevenpotential

155

terrorists.InEuropeancountries,whichhavemoregeneroussocialsafetynets,
welfaristpoliciesarebeingtrimmedasimmigrationincreases.Theargumentfor
cuttingthemisusuallythatimmigrantstakeadvantageoftheseprogramsinaway
thatmajoritycultureneverwould.Inthisimagery,theOtherhasaseriesof
negativequalitiesforwhichtheyaremorallyblameworthy.Thiscannotonlyserve
toremoveourmoralobligationtoinsurethattheyhaveadequateresourcesto
conductalifethattheyfindvaluable,butitcanservetoinstigateadesiretoview
resourceallocationasazerosumgameinwhichwhatevertheyreceiveisalossto
therestofus.Wethenfindourselveswiththeviewthatwemustprotectourselves
fromthisparasiticgroup,andworktodenythemanyresourcesthatwemight
otherwisehavebeenabletoallocatetoourselves.
Thesetwoimagesconflictwitheachotherinimportantways.The
psychologicalattitudesondisplaytowardsoutgroupsthatweseeinpolitical
discoursepushadoptionofpoliciesthatmaketheleastwelloffinsocietyeven
worseoff.Whatsmore,thesepoliciesoftenmaketherestofsocietyworseoffas
well,asIwilldemonstratelater.Butourmostcompellingphilosophicalmodelof
howtothinkaboutarrangingthebasicstructureofsocietyleavesthese
psychologicalattitudesout,preciselybecausetheyaredeemedmorallyirrelevant.
TherealpoliticalactorsareunmovedbytheRawlsianargumentpreciselybecause
theydonotseethemselvesasmorallyequaltomembersofthegroupsthatthey

and see no personal benefits, it is easier to be convinced that those who do receive
welfare benefits are different in a morally relevant way.


156
disparage.Inpart,thisisbecausemorallyequaltendstomeansomethinglike
issimilartomeincommonpracticalusage.112Aswecanidentifysalient
differencesbetweenourselvesandmembersofothergroups,wecanquicklydecide
thattheyarenotourmoralequals.113Thisisonlymadeworsewhenweconsider
regionsoftheworldwithevenmoreintergroupconflict,suchastheMiddleEastor
Africa.Butevenifwelimitourselvestoreasonablywellorderedsocieties,suchas
thosefoundintheWest,wefindthesebasicconflictseitherpreventingor
eliminatingtherealizationofbasicstructuresofsocietythataredesignedto
promoteamoreegalitariandistributionofwealthandincome.
WhatIwanttoargue,then,isthatthequestionofdistributivejusticecannot
beconsideredinisolationfromthequestionoftolerance,particularlywhenweare
facedwithapluralisticsociety.Wecannotbegintotrysocialexperimentationin
goodfaithifwecannotassureourselvesthatweareaimingforimprovingoutcomes
foreveryone.Thisbecomesincreasinglydifficultifminoritiesdonotfeelasthough
theirinterestsarebeingfairlyrepresented.Animportantpremiseintheoriginal
positionisthatwecaninfactseeourselvesandotherswithoutourgroupidentities,

112
As has been argued in earlier chapters, there is a great deal of psychological literature
supporting the claim that people treat members of their own in-group preferentially, even
if the in-group is entirely arbitrary. An example of this literature can be found in J
Sidanius, F Pratto, M Mitchell, In-Group Identification, Social Dominance Orientation,
and Differential Intergroup Social Allocation. Journal of Social Psychology, 134(2),
151-167, 1994
113
A non-standard source of evidence for this claim can be seen in political cartoons.
Representatives of other groups are often depicted as sub-human. This often takes the
form of some simian features being introduced, or simply wildly exaggerated features
that are stereotypically associated with that group. A collection of such political cartoons
from Britain during the Irish Potato Famine can be seen here:
http://www.nde.state.ne.us/SS/irish/unit_2.html


157
andinsteadseeourselvesasbaremoralagents.WhatIwanttosuggestisthat,
atleastinitially,pluralsocietiesarenotgoingtohavemembersthatarewillingor
abletodothis.ThisisbecausewemaynotseeanOtherasaPerson,inthesense
thattheyareentitledtothesamemoralconsiderationsasmembersofourownin
group.Inthesecircumstances,wecannotexpectanoriginalpositionstyleargument
toevengetofftheground.Becauseofthis,itbecomesnecessarytodesignthebasic
structureofsocietyinawaythataimstofostertheattitudethatallshouldbe
treatedasmoralequals,ratherthandesigninstitutionsthatalreadyassumethatthis
attitudeexists.Thisbearsimportantlyonthestabilityofaparticularchoiceforthe
basicstructureofsociety:ifweneedtocreatethebondsthattiediverseindividuals
together,weneedtoinsurethatwehavedonenothingthatundulyincreasesthe
strainsofcommitment.Inmakingthisargument,Ihopetomakeclearthat
tolerationitselfisaninsufficient,andifanything,dangerouslylimiting,conceptin
liberalthought.Whatweneedinplaceoftolerationisademonstrationthat
diversitypromotesanindividualsownabilitytoconductherlifeasshewants.
Toproceed,Iwillfirstargueforamorelimitedconceptofequalitytoground
therestofthechapter.Fromthere,Iwilldemonstratethateconomicmatterswithin
astatetendtohaveapositivesumnature,andfromthere,showthatamorediverse
populationisbetteratcreatingwealththanahomogeneousone.Withthese
productionorientedquestionsoutoftheway,wecanturntothequestionofwhat
theconstraintsareonmakingdistributionaldecisionsinadiversesociety.In
particular,Iwillarguethatdistributionistightlyalignedwithcontributionsin


158
allocation,butthatthisalignmentisdifferentthantraditionallythought.These
distributionalconstraintslimittheoptionsfornotionsofdistributivejustice,butdo
notthemselvespresentamodelofdistributivejustice.Instead,theyprovidethe
necessaryconditionsforsocialexperimentationthatdoesnotunderminesocial
stability.Finally,Iwillshowthatwhileontheallocationside,morediversityis
alwaysbetter,thedistributionalrequirementswillplacelimitsonhowdiversea
societycanbecomebeforeitgetsunstable.

MinimalEquality
Wehavealreadyseenthatdifferentperspectivesaregoingtoprovide
differentaccountsofequality,buttherewewereconsideringequalityintermsof
distributionsofgoods.Herewemustfirstaddresswhatwemeanbyequalityof
persons.Rawlsprovidesathickaccountofequalitywhenheusesthetermfreeand
equalpersonsinparticular,weareallequallyvaluablemoralagentswhotreat
eachotherassuch.Whilewemighthavespecialrelationshipswithsomepeople
andnotothers,westillviewpeopleequivalentlyquapersons.Thisistruewhether
ornottheysharethesamegroupidentificationsorotherindividualtraits.
Thisclaimisnormative,ratherthandescriptive.Andinsofarasitisa
regulativeideal,Ifinditcompelling.Iwouldprefertoliveinasocietyinwhichwe
treateachotherasfreeandequalcitizenswhoareengagedinasystemofsocial
cooperation.However,Idisagreethatwecanworkfromthisregulativeidealinour


159
designofthebasicstructureofsociety,sinceitappearstobeabundantlyclear
thatdescriptively,theclaimisfalse.Asthewelfarequeenexamplewasmeantto
demonstrate,wesystematicallytreatmembersofothergroupsworsethanwetreat
membersofourowngroup.Thisiseitherinconsistentwithabeliefinmoral
equality,orademonstrationthattheconceptdoesnotgetusmuch.114Whatis
worse,ourpreviousdiscussionofequalitydemonstratedthattherearemany
differentwaystodefineequal,manyofwhichareactivelyemployedtothe
exclusionofothersbysomesubsetofthepopulation.Thissuggeststhatweneedto
insteadfindsomemoreminimalaccountoftheequalityofpersonsthatcanbe
utilizedacrossthespectrumofperspectives.
Ourfirststepwillbetoabandonattemptstofindanaccountofequalitythat
isdevelopedintheabstract.Aswehaveseen,thesekindsofaccountsdonot
connectwelltoouractualconcernsaboutassuringprosocialbehavior.Instead,we
willturntoamoreempiricalaccount,wherewecandirectlyengageinthequestion
ofwhetherindividualshaveaneffectivevoiceinpoliticaldiscourse.

114
The Spanish Inquisition, for example, could be interpreted as believing in moral
equality in that they wanted to assure everyones entry into heaven. Which, they
believed, required a great deal of torture and suffering on earth. Moral equality is also
consistent with arguments for reducing the claims scope. Slave owners could very well
have held that all persons are moral equals, but their slaves didnt count as persons.
Finally, we might find that our strong idea of moral equality leads us to conclude that if a
minority group is systematically behind socio-economically, it must be due to personal
traits, like laziness or lack of intelligence. So then, given that we are all equal, those
individuals do not deserve assistance because they have chosen their position in life.
Moral equality and feelings of superiority are not mutually inconsistent.

160
Tofindaconceptofminimalequality,wewillstartwithHobbes.Hobbes

arguedthatallmenwereequalbecausetherewasnomanwhowassostrongthat
noonecouldkillhim,andtherewasnomanwhowassointelligentthatnoone
couldoutwithim.115Noticethatthisdefinitionofequalityisdescriptive,andnot
normative.Itjustisthecasethatnopersonissopowerfulsoastobeimmunefrom
attacksfromothers,whethertheyareviolentormerelyclever.Hobbesdoesnot
arguethatthisshouldbethecase;hejustarguesthatitisthecase.IthinkHobbesis
correctinthisapproach.
WhatissoappealingaboutHobbessapproachisthatherecognizesthat
humanshaveabasicneedforsociality.Inordertoprotectthemselves,theymust
joinsociety.Otherwise,theywillneverbeabletoassuretheirownsafetythere
willalwaysbesomecoalitionthatcandestroythem.Whileacommonreadingof
thisisthatitisjustoneofmanyexamplesofHobbessdimviewofhumanity,Ifindit
tobeakeyelementinwhatwewantoutofanaccountofequalityofpersons.Itake
Hobbestohaveadequatelyarguedforthisaccountforselfdefensereasons,butI
wouldnowliketoextendtheaccountintoanareamorerelevanttoourcurrent
inquiry.Inparticular,thereisntanindividualwhocanbeeconomicallyself
sufficient.Evenifyouarethemosttalentedpersonintheworldateachandevery
humanendeavor,thelimitsonyourowntimeimpedewhatyoucanachieve.In
particular,ifyouwereaveritableRobinsonCrusoe,andneverengagedinany
economicactivitywithanyone,andtrulyprovidedeverythingforyourself,you

115
In particular, no one is so powerful that they can resist any coalition of others.


wouldalmostassuredlybeverypoororverydead,dependingonyour

161

circumstances.Havingtostartfromscratch,withnotools,noaccesstooutside
learning,andnofoodreserveswouldbenearly,ifnotcompletely,impossible.
Sothissetsusupfortheclaimthatwemayhavesomeinterdependence,but
wearenotyetataclaimofequality.HerewecanverynaturallyextendHobbesto
ourmoreeconomiccase.Inparticular,noindividualisgoingtobemore
economicallyproductivethananyothercompetingeconomiccoalition.116BillGates
mightbeveryrich,butifyouaddupthetwonextrichestpeopleswealth,heis
overwhelmed.Moreimportantly,however,isthathiswealthwasnotofjusthis
creationtogetit,hehadtocreateacompanywithtensofthousandsofemployees,
buildupanentireindustryaroundhiscompanysproducts,andsellthemtoavery
largepieceoftheworldspopulation.ThoughGatesmighthavehadakeyingredient
inthecreationofallofthiswealth,hecouldnothavedoneitonhisown.117
Whatwefindisthateveryindividualisequalinsofarashereconomic
bargainingpowercanbeequaledbysomecompetingcoalition.Itisimportantto
recognizetwofeaturesofthis.One,moreproductiveagentscreatemorewealth,

116
This claim is restricted to be about individuals. It is possible that one could form an
economically dominant coalition of individuals, but this is a separate issue from whether
there is an individual who is singularly economically dominant.
117
The example of Bill Gates is complicated by the fact that his wealth is largely derived
from the peculiarities of our legal systems treatment of intellectual property. While I do
not want to make claims about what intellectual property rights might be chosen in as a
result of the bargaining process outlined in chapter 3 (other than perhaps claim that it
would be unlikely to match the current system in the US), I do want to suggest that what
is relevant in this example is just that Gates incredible wealth is not something solely of
his own making, but is in fact a result of a great deal of social cooperation.


andassuchhaveasuperiorbargainingpositionrelativetolessproductive

162

agents.Buttwo,thesemoreproductiveagentshavealotmoretoloserelativetothe
medianagentiftheyareunabletofindabargainingpartner,sorelativetothesetof
possiblepartnersactingasacoalition,theyareinaweakerposition,especiallyin
thelongrun.118Asimilarclaimcanbemadeaboutcurrentwealthendowments.
Onecanwieldonesexistingwealthasabargainingadvantageonlyuptothepoint
whentheothersidedecidesthattheywouldbebetteroffiftheyjusttookitaway.119
Whilewealthmaybeusedinaoneshotgameasasignificantbargainingadvantage,
asitshiftsthedisagreementpointinfavorofthewealthierplayer,inlongrun
gamesinitialwealthendowmentsareoverwhelmedbythepotentialfutureearnings
ofbothparties,andsotheshadowofthefuturebecomesmoreimportantthanthe
disagreementpointofthefirstroundofbargaining.
Anydiscussionofcomparativebargainingpowermusttakeplaceinthe
contextofwhatkindofbargainweenvision.Asisoftenthecaseinstrategic
situations,oneshotgameshavedifferentconsiderationsthanlongrun,indefinitely

118
This basic bargaining insight explains why trade unions are good for employees. As
an individual worker who is easily replaced by another worker, the individual has very
little bargaining power. But the class of workers has a great deal of bargaining power, as
without any of them, the companies in which they work cannot function. This is why
union solidarity is so important: the unions power is reliant on maintaining a complete
coalition of workers in order to match the power of the employer. Absent a complete
coalition, scabs can undercut wages, and the union is undermined.
119
Seizing someones assets ought to be no more difficult than killing a person, and as
weve seen already, no person is immune from some coalition attacking her. So we can
assume then that some coalition is capable of taking away someones money. In less
violent terms, this happens regularly when democracies decide that inequality has
become too extreme, and create a more progressive tax structure. In essence, the wealthy
can have short-term bargaining advantages, but these advantages can be corrected for.


repeatedgames.Inparticular,inoneshotgames,itismucheasiertouse

163

perceivedbargainingadvantagestoonesbenefit.However,inrepeated
interactions,thisbecomesmoredifficult,becausethenatureofrepeated
interactionsallowsforbothpartiestousedefection120fromcooperationasameans
oftrainingtheotherpartyovertimetobehaveinawaythatismorebeneficialto
thedefector.Aslivinginsocietyisclearlyalongrun,repeatedinteractionofmany
parties,weshouldunderstandourselvestobethinkingaboutbargainsinthe
repeatedcontext.Becauseofthis,wecansafelyfocusonthelongrunbargaining
poweroftheagentsinvolved,ratherthanpowerintheoneshotcase.
Bydefiningequalityintermsofbargainingpower,wecannowalsoseehow
thisaccountofequalitycanaccommodatemultipleperspectives.Wehaveleftthe
conceptofequalityasanextremelyminimalconcept,andassuch,havenotincluded
anythingthatwouldconflictwithathickeraccountthataperspectivemightprovide.
Instead,theconceptofequalitysimplydescribestherelationsofindividualstoeach
otherinthecontextofdecisionsaboutdistributingresources.Notunlikeinthe
previouschaptersdiscussionofabargainofrights,herewefindthatbargainsover
theproductofsocialcooperationcanbetreatedinthesameway.Individualsmay

120
When I use the term defection, I refer to choosing to not bargain, or holding out for
an outcome that the other would not accept. Importantly, I am also lumping together
actual defection with the threat of defection. Often the threat of defection (or retaliation
for the other partys possible defection) is sufficient to enable cooperative behavior. The
classic example of this is the strategy of Mutually Assured Destruction. Neither party
intends to use their weaponry, but both sides must have sufficient weapons to completely
destroy the other. With this explicit threat of complete destruction, neither party will
engage in even small-scale attacks on the other, for fear of massive retaliation.


164
differentiallyvaluegoodsandservices,orevencurrencyitself,butthatdoesnot
affecthowweaccountfortheirrelativebargainingpower.
Minimalequality,then,canbeunderstoodasadescriptiveclaimaboutthe
statusofagentsinalongrunbargain.Noindividualcanoptoutofthebargain,and
noindividualcandominatethebargain.Assuch,everyindividualisfacedwitha
similarsetofconstraints,andcanbeforcedtorespectthoseconstraintsbytheother
agentsinthebargain.
Onemaybesuspiciousofthisclaim,andsimplyassumethatthesocial
bargainisgoingtobemadeupofcoalitionsprimarilyorganizedaroundeconomic
class.Onemightfurtherthinkthatthecapitalistclasswouldbeinapositionto
crushtheworkerclass,andthuswewouldfindourselveswithnoworkingconcept
ofequality.Ithinkthatsuchsuspicionsaremisplaced.First,aswehaveseenin
Chapter1,therearenonaturalgroupsitmaywellbethecasethatwehavesome
coalitionsthatformaroundeconomicinterests,butjustaseasily,coalitionscan
formaroundregionalinterests,religiousinterests,sharedpoliticalgoals,orany
numberofotherfocalpoints.Butletsassumethateconomicinterestswill
dominategroupformation,andwefindourselveswithacoalitionofcapitalistsand
acoalitionofworkers.Inaoneshotinteraction,absolutelyitwouldbethecasethat
thecapitalistswouldbeabletoforceanextremelyunfairbargainontotheworkers.
Butaswehaveseen,inalongrunbargainingsituation,thisdoesnotappeartobe
likely.Unionsarethenaturalresponsetoexcessivecorporatepower,andinthe
longrun,theworkerclasscanuseitsabilitytodefecttodenythecapitaliststheuse


165
oftheirproductivecapacity.Thisdefectioncandrivethecapitaliststoaccepta
moreequalbargain.So,eveninthisworstcasescenario,wherealloftheeconomic
capitalisinonecoalition,thelongrundynamicsofthebargaincreateopportunities
fordefectiontobeusedtoshiftbargainingpowertowardstheweakerparty.
IfoneiscomingfromaRawlsian,Kantian,orRousseauviancontext,this
accountofequalitymightcomeacrossashollow.Theaccountofequalityofpersons
heremakesnoclaimabouttheirmoralvalue,andinsteadsubstitutesineconomic
argumentsabouttheirbargainingpower.ButIwanttoreiteratethatthereasonfor
thismoveisthat,evenifIgrantthemthatwewantsocietytobemadeupofpeople
whorecognizethatallarefreeandequalpersonsengagedinasystemofsocial
cooperation,weneedtofindameanstoleadpeopletherefromhere.Soifour
institutionsassumethatwevealreadyachievedthis,andfurtherassumethat
peopleseethemselvesasequalsinthisthickersense,thenwewillnotbeableto
deliveronsuchasociety,particularlyinapluralenvironment.Withthisverystrong
assumption,ourinstitutionaldesignmightbeinsufficientlyrobusttodealwith
individualswhodonotbelieveinthisthicknotionofequality.Thiswoulddecrease
socialstability,whichcouldfurtherundermineafosteringofthethicksenseof
equality.Whenweknowalreadythatpeopledonotviewthemselvesthisway,we
shouldnotassumethattheydo,evenifitseemslikeanidealatwhichweoughtto
aim.However,Idowanttocontendthatthismoreminimalaccountofequalityis
fullycompatiblewithachievingthosekindsofends.Inparticular,Iwillargueinthe
nextsectionsofthischapterthatindividualswhoholdaviewofminimalequality


166
willrationallylearntobecomemoreacceptingofothersandmorewillingtosee
agentswhoaredifferentfromthemasvaluablemembersofsociety.Aswewillsee,
however,therearelimitstothisapproachinhowacceptingwecanreliablybecome.

CreatingProsperity
Thefirstchallengewefaceinunderstandingdistributivejusticeinaplural
societyisdeterminingwhatkindofsituationweface.Inrealworldpolitical
discourse,peopleoftenunderstandthenational(andworld)economyasazerosum
game,likepoker.Inthesekindsofsocialinteractions,everygainmadebyone
personmustbeaccompaniedbylossesbysomeoneelse.Inaparlorgamelike
poker,thisisveryclear.Thereisafixedamountofmoneyatstake,andsoifIwina
hand,atleastoneotherpersonmusthavelost.Wealthonlychangeshands;itis
nevercreatedordestroyed.Asmostofushaveplayedpokerorsomesimilargame,
thisisoftentimesourmodelforlargereconomicphenomena.Thankfully,itisthe
wrongmodel.Iwillfirstpresentsomereasonsforthinkingitmightbeacompelling
model,andthenshowwhyitisthewrongone.
Thebasicargumentfortheeconomyasonelargepokergameisthatwedo
have,toafirstapproximation,afixedpoolofmoney.Centralbanksarenotinthe
habitofprintingmoney,andindividualscannotmaketheirownmoney.So,then,if
thereisafixedamountofmoney,thenitmustbethecasethatifImakemore
money,someoneelseislosingmoney.Further,weoftenhearabouthow


167
immigrantshavetakenthejobsthatusedtobelongtonaturalborncitizens,or
moreprominentlylately,thatjobshavebeenmovedtoanothercountry.Onthese
descriptions,jobsaretakenfromonesetofindividualsandgiventoanotherset.So
wecanneatlypointtothewinnersandlosers,andseehowthemoneychanged
hands,notunlikeapokergame.
Thereisaproblemwiththisdescription,however.Namely,thatwealth
regularlyexpands(andoccasionallycontracts).Whilecentralbanksdonotprint
money,themoneysupplydoesexpandandcontractbasedonhowbankslend
money.Evenmoreimportant,though,isthatmoneyandwealtharenotthesame
thing.Wecanunderstandmoneyjusttobethenavenotionofwhatonekeepsina
bankaccount.Butwealthisbetterunderstoodastherangeofgoodsandservices
onehasaccessto.Toseewhythisisanimportantdistinction,letsimaginea
scenarioinwhichonecouldhavelessmoneybutmorewealth.
Imagineimmigrantscomeintoacountry,andexertdownwardwage
pressurebyincreasingthelaborsupply,andhavingawillingnesstoworkat
substantiallylowerpay.Someofthepreviouslyemployedworkerswilllosetheir
jobs,whilestillmoreofthemwillkeeptheirjobsbutatapaycut.Theseworkers
willhavelessmoneythantheyhadbefore.But,ifthisdownwardwagepressureis
occurringinmorethanjusttheirindustrysector,thentheyverywellmightfindthat
theirwealthhasincreased,becausethepricesofthegoodsandservicesthatthey
areinterestedinhavegonedown.Sowhiletheyhavelessmoney,eachdollarthey
dohavehasmuchgreaterpurchasingpower.

168
Evenmorefundamental,however,isthatunlikeinpoker,inthereal

economy,wemakethings.Everycarthatismadeisacarthatwasntthere
yesterday,andthathasvalue.Allthesethingsthataremade,orservicesthatare
provided,havevalueeveniftheythemselvesarenotmoney.Itisforthisreason
thatitisamistaketothinkofoureconomicinteractionsasazerosumgame.The
economyisproductive,andsowemustbeinapositivesumgame.121
Itisimportantthatwetakeourselvestohaveestablishedthattheproblem
wefaceispositivesum.Ifwefacedazeroornegativesumgame,thenaplural
societywouldfail.Insuchgames,theonlywaytogainistotakefromothers,and
eliminatethemfromthegame.Undersuchconditions,itwouldbepsychologically
impossibleforagentstodevelopfellowfeeling.Onlypositivesumgamesprovide
theappropriateincentives.122Itisaveryfortunatefactthatwedoliveinapositive
sumworld.Ourfirsttaskissimplytorecognizethatwedo.
Beinginapositivesumgamedoesnotguaranteethateveryonealways
benefits.Afterall,alleconomiesarepositivesumgames,andthereareplentyof

121
It is tempting to want to claim that the existence of a positive sum game should be one
of the circumstances of justice, beyond what Hume specified. Positive sum games
provide the opportunity for shared gains, rather than just merely agreeing to giving
something up in an equitable manner.
122
This follows from the fact that in positive sum games, there is at least one cooperative
option that increases social utility, at least in comparison to relevant alternatives. Even in
the pre-industrial revolution period, when economic growth was safely under 1%, social
utility would have decreased if something other than the cooperative option had been
chosen. So too with the stationary state envisioned by Ricardo and Mill: though they
imagine an upper limit of productive capacity, it can only be maintained through
cooperation. If one party chooses to defect, social utility would decrease. This is unlike
a zero sum game like poker if I bluff and win a hand, no money is lost. It is just moved
amongst the players.


169
peoplewhoclearlylose.Butapositivesumgamedoesguaranteethatthereis
somewayofarrangingthebenefitssuchthateveryonedoesgain.Thisiswhywe
canrepresentthequestionofdistributivejusticeasabargain.Indoingso,whatwe
arerepresentingisthefullrangeofpossibleallocationanddistributiondecisions
thatcouldbemadeinaneconomy.Theparticularbargainingsolutionwechooseis
thenourstatementonwhatkindofdifferentialgainsandlosseswearewillingto
accept.
Adownsideofthebargainingrepresentationisthatitlumpstogethertwo
importantanddistinctprocessesproductionanddistribution.Sincewewantto
lookathowtheseprocessesfareinapluralsociety,wemustlookatthem
individuallyandnotasasinglebargain.Soletusnowturntothequestionof
production.

PluralProduction
Beforewecandeterminehowweoughttodistributewealthandincome,we
mustfirsthavesomewealthandincometodistribute.Assuch,wemustfirstturn
ourattentiontotheproductionofwealth.Whilewearenotinterestedinparticular
industriesoreventhedistributionofindustry,weareinterestedintheincentives
thatexistinthelaborforcetoeitherbewelcomingoforhostiletothosewhoare
different.Thisdiscussionfollowstheeconomicargumentsfordiversitythatwe
examinedinChapter1.Inparticular,wefindthatdiversityleadstogreater


economicproductivity.Themainreasonforthisisthattheeconomybenefits

170

fromthedivisionoflabor,andadiversesocietyoffersmoreopportunitiesfor
furtherdivision.Individualscanspecializeattasksbestsuitedtotheirskillset,
whichallowthemtoreachtheirproductivecapacity.Asanexample,Alice,anexpert
mechanic,canfocusonfixingcarsallday,whileBob,abaker,canfocusonbaking
bread.SinceAliceisstillinterestedinconsumingbread,andBobneedshiscarfixed,
theycantradeamongstthemselvestomakesurethattheybothgetwhattheywant.
Thisisofcourseasimpleillustration,butitisneverthelessusefulforustoconsider.
Whatweseeinthisexampleisthatdiverseperspectivesbringwiththemdiverse
problemsolvingabilities.Whetheritisonamicroscalesuchasourexampleofa
bakerandamechanic,oronamuchlargerscalewherewehaveethniccommunities
whotakeoncertainlaborfunctions123,thisprincipleofspecializationhassignificant
implications.Thisisimportantintwoways.First,jobcompetitionisntdirect.
Second,thegainstotradearelarger.
Thefirstconcernisaboutlaborcompetition.Considertwosocieties,onethat
ishomogenous,andtheotherthatisdiversewithrespecttoperspectives.Imagine
thatbothAliceandBobarelookingforjobs,andaretryingtodeterminewhich
societytheywouldratherbein.Iftheyareinthemorehomogenoussociety,itis
morelikelythattheirskillsarecorrelatedwitheachother.Inasocietythathasa

123
It is not unusual to find immigrant communities specialize with certain businesses,
like Korean groceries or Chinese laundries. Many immigrant communities carve out
market niches for themselves in a similar manner. This standardization makes it easier
for the communities to support business development efforts, since the business
knowledge is shared in the community, and there are collective bargaining opportunities
with suppliers to keep costs low.


171
sharedculture,broadlysharedvalues,andstandardizededucation,weshould
expectthatindividualswouldhavealargelyoverlappingsetofskills.Skilldiversity
needstocomefromsomewhere,andifindividualsupbringingislargelythesame,
thereislessroomfornovelty.
Wecanseethiscasefordiversitybringingnoveltywithintheframeworkof
perspectives.Iftwopeoplesharethesameperspective,itmustmeanthatmanyof
thoseculturalordevelopmentalfactorsweresimilar,foritisremarkablyunlikely
thattwoindividualswhogrewupincompletelydifferentenvironmentswillsettle
onidenticalwaysofcategorizingtheworld.124Andsoiftwoindividualsdoshare
thesameperspective,theylikelyalsoshareahugeculturalandeducational
background.125 Iftheyhavedifferentperspectives,ontheotherhand,itislikelythat
theyhavedissimilarbackgroundsinatleastsomerespect.Thisintroducesan
opportunityforadifferencethatisrelevanttothesortofproblemsolvingthatis

124
To press on how unlikely this is, remember that a categorization scheme is a full
description of how a person organizes concepts. For, say, an Amish person with
schooling through age 16 to have the same set of concepts as an African villager and a
Ivy League educated hedge fund manager from a wealthy family seems radically
implausible. At the very least, the Amish person would have more fine-grained
distinctions of crafts and farming methods than the hedge fund manager, and likewise the
hedge fund manager would have far more distinctions to make about modern technology.
125
To press the point further, let us look at how shared cultural heritage can lead to
shared skills. The traditional jobs that European Jewish people take on revolve around
fields that do not require property ownership, such as medicine and banking. This is in
large part a result of the history of Jewish persecution in Europe. Jews were not allowed
to own property, and Christians for a long time were not allowed to lend money with
interest or study anatomy on cadavers. As such, Christians needed doctors and bankers,
but were religiously prohibited from performing such tasks. Jews were politically
restricted from property ownership, but did not have religious restrictions on medicine or
banking. So Jewish people now have a long history of specialization in such fields, even
though the restrictions on both sides no longer exist.


relevanttoonesjobskills.Soiftwopeoplesharethesameperspective,then

172

theirskillsetsarealsolikelytobeshared.Whenthisisthecase,AliceandBobhave
moretofearfromeachotheraslabormarketcompetition.Iftheydonothave
uniqueskills,butinsteadaremerelytwoamongmanywhosharelargelythesame
rangeoftalentsandknowledge,thentheyhavereasontoattempttoerectbarriers
toenterthelabormarket.Iftheywereinthemorediversesociety,however,it
wouldbemorelikelythatAliceandBobhavedifferentperspectives,andthus
differentskillsets.Thiswouldreducethechancethattheyarefungiblewitheach
other,andwouldmeanthattheyarelesslikelytobeindirectcompetitionforthe
sameclassofjobs.Inthiscase,thereisnoreasontofearnewlabormarketentrants.
Aplumberdoesntcareifanumberofmechanicsmovetotown,astheyarenot
goingtobedrivingdownherprices.Ifanything,theplumberhasreasontobe
excited,becauseitshouldlowerhercarrepaircosts,andfreeupmorehouseholds
tospendmoneyonhiringplumbers.Themorethelabormarketisdiversified,the
morereasonweoughttohaveforcheeringitsexpansion.Whenthemarketis
homogenous,however,wehavegoodreasontotryandplaceasmanyrestrictions
onitaspossible.
Toillustratethiscase,imaginethatyouaretryingtolandajobasa
philosophyprofessor.Thequestionathandiswhetheryouwouldratherbeina
societythatisfamedforitsproductionofphilosophyprofessors,orifyouwould
ratherbeinasocietyofthesamesizethatproducessomephilosophyprofessors,
butalotmoreprofessionalsinother,reasonablydisparate,fields.Whileitmaybe


173
funtohavesomanyphilosophyprofessorstotalktointhefirstinstance,italso
wouldmeanthattherewouldbeamuchhigherlevelofjobcompetitionthanwould
befoundinthesecondinstance.So,economicrationalitywouldpointoneto
wantingtobeinthesecondsociety.
Sofarthishasjustbeenapurelynegativecase.Wewelcomediversity
becauseitmeansweavoidunwantedcompetition.Butthisisnotsufficient.We
shouldalsobeabletoarticulateapositivereasonforwantingadiversesociety.And
ifwecontinuewithourexample,itmaybefuntobeaphilosopherinaseaofother
philosophers,butevenphilosophersneedtoeatsomething,livesomewhere,and
wearsomething.Andtodothat,weneedfarmers,grocers,chefs,architects,
engineers,tailorsandshopkeepers,amongstothers.Thisisacomponentofbeing
abletoengageinasystemofsocialcooperation.Weneedtohavemanydifferent
kindsoftasksperformed,andforthatwerequiresomedivisionoflabor.
Theprimaryreasonforwantingasmuchdiversityaspossibleinproduction
isnotforreasonsofavoidingcompetition,butthehugegainstotradethataremade
possiblebyhavingmanydifferentkindsofspecialists.AsIhavearguedindetailin
Chapter1,Ricardiantradetheoryshowsthattradeamongstdiversespecialists
increasessocialproductmorethanproductionbygeneralistsorhomogenous
specialists.Theaverageindividualbenefitsintwoimportantwaysfromthis.First,
foranygivenclassofgood,thereiseithermoreofitoritisofhigherquality(or
both),duetothespecializationofproduction.Sonomatterwhatmytasteis,
whateveritisthatIlikethereismoreformetoconsume.Bydefinition,rational


agentsmustpreferthistoanyalternative.Thesecondwayisperhapsmore

174

interesting.Thereisalsoalargerclassofgoodstosample.Thistranslatesnotjust
intothirtydifferentkindsofcansofsoupatthegrocerystore,andlotsofnewethnic
restaurantstotry,butitalsomeansnewcinema,newartwork,newmusic,new
literature,andnewclothing.Theselatterchoicesarenotjustconsumergoodsthat
onecanhaveapreferenceabout:theyalsorepresententrypointsintonewwaysof
livingandexperiencingtheworld.Whileagentsdonotneedtoopttoexperiment
withhowtheyliveintheworld,somedowantsuchoptions,andtheyareonlymade
possibleinanenvironmentwheredifferentperspectivescanbemixedand
reinterpreted.126Thisgoesbeyondpreferencesatisfaction,andcreatesthe
possibilityfordiscoveringnewtastesandnewselfconceptions.127
Itshouldbenotedthatindividualsmaywellremainstubbornlyagainst
diversity.Theycouldsimplyhavedeepseatedfearsorprejudicesagainstother
kindsofpeople,andthiswouldleadthemtodowhatevertheycantoblockothers
fromentryintosociety.Thisisaveryplausibleattitude,andonethatseemsto
manifestitselfquiteoftenintherealworld.ThoughIcannotprovideanargument
thateachofthesepeoplewillcometodecidethattheyarewrong,andbegintosee
thevalueofhumandifferences,Icanarguethattheincentivesarealignedagainst

126
It is possible that an agent could have a preference for stasis, and find new options to
be a disutility. While I register the possibility of such curmudgeon preferences, I do
not seriously consider them in my analysis.
127
This opportunity for smaller-scale experiments in living is essential for the overall
procedure of social contract formation that has been developed in the past two chapters.
In particular, new experiments in living provide societies with new evidence for our
moral beliefs.


175
theirviews.Ifwethinkoftoleranceasameasureforhowmuchdifferenceone
iswillingtoacceptinotherpeoplesviewsandcustoms,thenitiseasytoseethat
themoretolerantoneis,thegreaterthepotentialformaterialrewards.Themore
differentothersare,themoregainstotradethereare.Ifoneisonlywillingto
interactwithotherslikeher,thenlittleeconomicgrowthispossible.128Evenif
theseindividualsnevercorrecttheirintolerantways,theywillbeeconomically
marginalized.129Thosethatembracetolerantattitudesthrivebecauseofthegreater
economicopportunityopeninfrontofthem.Aswesawinthepreviouschapter,this
canhelpexplainwhytradeorientedcitiesareverytolerant.Toleranceallowsan
individualtotakeadvantageofeconomicopportunitiesthatwouldotherwisebe
closedoff.Tolerancecanbethoughtofhereasideologicalleewayitistherangeof
positionsthataredifferentfromonesownthatoneiswillingtoputupwith.Asone
increasesthatrange,moreeconomicinteractionsbecomefeasible.
Letusnowtakestockofwhatwehavefoundinthecaseofpluralproduction.
Wehaveseenthatitreduceslabormarketcompetition,sincepeoplewhohold
differentperspectivesarelesslikelytohavethesameskillsetsandthusbe

128
Milton Friedman makes a similar case for how capitalism eliminates discrimination in
chapter 7 of his Capitalism and Freedom (1982). While I do not endorse the argument
insofar as it should be taken to convince every individual, I do think it should be taken
seriously to illustrate the general push of incentives at the macro level.
129
This argument relies on the assumption that not everyone in society is intolerant. If
there is some variation of levels of tolerance in society, then there would be cases of
some doing better than others, and incentives would be made clear. However, if society
is uniformly intolerant, than there is no symmetry breaker to see that being more tolerant
makes one better off, and so no one would become more tolerant without some external
factor, such as the 1964 Civil Rights Act in the United States. Thanks to Samuel
Freeman for pointing this out.


176
competingforthesamekindofjob.Becauseofthis,thereshouldbelowerlevels
ofcomplaint(andevenhappiness)aboutnewentrantstothejobmarket.Itiseasier
toseethatthenewentrantsarenotdirectcompetition,andsoitiseasiertotolerate
theirentry.Moreimportant,however,isthefactthatdiverseperspectivescreate
morespecialization,whichinturncreatesgreaterreturnstotrade.130Infact,the
morediversethesociety,themoresocialsurpluswecancreate.Andinsofaras
diversitycreatesopportunitiesfornewinnovations,itfostersevenmorediversity.
Inproduction,thereareonlyincentivesforencouragingmorediversity.Thisistrue
fromasocialperspective,butitisalsotruefromanindividualperspective:each
individualcanonlybemadebetteroffasproductionismademorediverse.
Whilewehaveshownthatindividualsalwayshavereasontowantmore
diversitywhenweconsiderproduction,wehavenotyetconsideredthecruxofour
challenge:thequestionofdistribution.Itistothisquestionthatwenowturn.

130
One objection to this view is that there is an upper limit to how much specialization
can be useful. Eventually, more specialization would reduce production, because it does
not create any new efficiencies, and instead might reduce them. This objection is true if
we consider a single manufacturing process, such as Adam Smiths pin factory, but is not
true if we consider an economy as a whole. It is certainly the case that in any given
factory process, there are only so many tasks to perform. But, part of what specialization
enables in general is new market niches and new products and services. The bundle of
goods and services available to consumers today is far more extensive and varied
compared to consumers 20 years ago, let alone 100 years ago. Part of what enabled that
is new areas of specialization.

177

DistributiveJusticeinaDiverseSociety
Thusfarwehaveseenthateconomiesarepositivesumgames,andthat
diversesocietiesarethemostproductivesocieties.Butwhatwehavenotyet
addressedisthatdiversesocietiesarealsogoingtofacethemostdisagreementin
howwedistributethegainsinproduction.Thisisundoubtedlyachallenge,butthis
challengeismutedtosomedegreebythefactthatasweincreasediversity,wealso
increasetheresourcestobedistributed.131Thisfactwillhelpuswithour
distributiveproblem.
Evenso,wefaceachallengeofhowtodistributetheproductsofsocial
cooperation.Aswehaveseen,Rawlsproposesthedifferenceprincipleasan
appropriatemechanismtoorganizethebasicstructureofsociety.Butwecontinue
tofacetheproblemthatthisprincipleisunlikelytosurvivethepsychological
dispositionsofagentsinadiversesociety.WhileIwillnotaimtodefenda
particulardistributionalrule,Iwilltrytoarguethatrulesthatfollowtwoprinciples
willbemorestableinapluralsociety.Inparticular,thechoiceofdistribution
shouldbeontheParetofrontierofthesetofpossibleoutcomes,andthegains

131
The reason for this is straightforward: as we have seen in the previous section, as we
increase diversity in a society, holding everything else constant, we also increase the
productivity of the society. This ought to, in turn, lead to greater social surplus. We now
consider the fact that, even though increased diversity creates more social surplus, it at
the same time increases disagreements about how to divide that social surplus. So even
though this society has more disagreements because of more diversity, it also has more
surplus than it would have were less diverse, and thus less prone to disagreement.


shouldbedistributedinproportiontocontribution.132Theseprinciplesare

178

meanttoensurethatnoindividualorgroupisbeingtakenadvantageofbyanother
group.
Letsconsiderthefirst,lesscontroversial,principle.Thisprinciplestatesthat
weoughttoensurethatwenotonlymakeParetomoves,butthatwemake
maximallyefficientdistributions.TheParetofrontierofabargaindefinesthesetof
outcomesofthebargainatwhichtherearenoexchangesofgoodspossiblethat
wouldbenefitonepartywithoutharmingtheother.Theseoutcomesareallefficient
sincethereisnowealthorincomethatcouldhavebeendistributedbutwasnot.All
resourcesarebeingconsumed.Thisfirstprinciple,then,justassuresusthat
distributionsinsocietyarenotwasteful.Italsoassuresusthatnooneloses.Every
movetotheParetofrontierguaranteesthateveryagentinvolvedinthebargainis
madenoworseoffforhavingparticipatedinthebargain.Itispossiblethatoneor
moreagentswouldnotbenefit,buttheywillnotbeharmed.Thisprincipleis
essentialforstability.Ifitwerepossibleforsomearbitrarygrouptolosegroundin
theirabsolutewealthandincome,theywouldhavenoreasonforsupportingthe
institutionsthatcausedthischange.Infact,theywouldhaveverygoodreasonfor
opposingthoseinstitutions.Theywouldbeinessencesubsidizingthegainsof
othersinthepopulationwithoutreceivinganybenefitinkind.Thiswouldbe

132
The idea of contribution needs further clarification. As I will explain in some detail
later, I determine ones contribution counter-factually. Ones contribution is the
difference in production one makes by participating in the productive process. An
alternative approach that is a little less abstract is that contribution is what the agent or
agents could have produced absent the cooperation of others.


179
irrationaltoendorse.Theagentswhoaretakinglossesmustthenbeeitherin
factirrational,orbeingcoercedintoparticipation.Ineithercase,thejustification
forthisisuntenable.Assuch,thisfirstprincipleismeanttolimitthespaceof
possibledistributionalrulesdowntothosethatareincentivecompatible.This
principleassuresthatanydistributionthatischosenisonewherenoagentisworse
offforhavingagreedtoit.133
Thesecondprinciplewewillconsiderismoresubstantivethanthefirst.It
statesthatthegainsfromsocialcooperationshouldbedistributedinproportionto
contributionstoproduction.Thisprincipleisbasedontwoideas.First,thatinsofar
aswearetreatingsocietyasasystemofsocialcooperation,wemustensurethat
everyoneengagedinthatcooperationbenefitsfromthefruitsofthateffort.The
secondideaisthatnogroupshouldberequiredtosubsidizeanothergroupsuch
thattheywouldhavebeenbetteroffontheirown.Thesetwoconditionstogether
imposesignificantconstraintsonhowwealthandincomecanbedistributed,but
bothareimportantlybasedontheproductionstep.

133
It should be noted that this principle would almost certainly eliminate the difference
principle from consideration, or at least place significant restrictions on it. This is for
several reasons. First, the difference principle, as it is not properly a bargaining solution
(in that it lacks a disagreement point), cannot be evaluated for whether no agent is made
worse off. It is possible that the best off in society would be made less well off under this
distributional rule. Second, the difference principle does not speak to any efficiency
questions, and in fact may run counter to them. If inequality can only be justified by
making the worst off best off, it may well be the case that there are Pareto moves left to
make to get to the Pareto frontier, but the difference principle ignores them because the
gains could not be transferred to the worst-off in society. This can happen if there is
significant deadweight loss in social transfers. It is possible to imagine a reformulated
difference principle that avoids these problems, but it would not be as Rawls presents it.

180
Iwouldliketofocusonthissecondidea,whichIcallthenosubsidy

principle.Iaimtoarguethatthenosubsidyprinciplemustbeattheheartofany
attemptsatdistributivejusticeinadiversesocietyifmaintainingstabilityisan
activegoal.Unlikeinmorehomogenoussocieties,weareunabletorelyona
broadlysharedcultureorsharedhistorytoprovidesocialcohesion.Thismight
seemtobeasmallmatter,butItakeittobeofgreatimportance.Societies,
particularlythosefacingsomelevelofdisagreementinthepublicsphere,needsome
reasontomotivatecitizenstowanttoresolvetheissuesandbewillingto
compromise.
Sharedcultureisapowerfulmotivator.Itprovidesareasonforwhya
societyshouldstaywhole,andnotsplitintoseparateentities.Italsooftenmeans
thatdisagreementsinthepublicspherearelimitedintheirseverityashared
cultureoftenimpliesabaselinesetofvaluesthatarealsoshared.WhenDemocrats
andRepublicansargueintheUnitedStates,theymaydisagreeoverafew
percentagepointsonthetopmarginaltaxrate,orthegenerosityofthesocialsafety
net,butneithersidequestionsmorebasicprinciples,likewhetherweendorse
marketcapitalismasanefficientmethodofallocatingresources.Theinstancein
whichtheUnitedStatesdidfacethepotentialofsplittingintotwocountrieswas
largelyanareainwhichdifferencesincultureplayedamajorrole.Thoughitis
undoubtedlythecasethattheNorthhadastrongermoralpositioninregardsto
slavery,itwasmadeeasiertoadoptinthattheNorthhadbecomelargelyan
industrialregion,whiletheSouthwasfirmlyagrarian.Sharedculturedoesnot


guaranteestabilityofcoursetheRevolutionaryWarintheUnitedStates

181

showedthatbutitdoesgoalongwayindampeningconflict.
Withoutanyassuranceofsharedculture,wemustfindsomealternative
socialglue.Rationalselfinterestisapowerfulsocialglue,particularlyasthe
incentivesforproductionallpointtowardgreaterlevelsofsocialcooperationwith
morediversepopulations.Whileitmaybethecasethat,overtime,amorerobust
senseofcountryandcitizenshipcoulddevelop,thisisnotsomethingthatcanbe
assumed.Someindividuals,whoseperspectiveinvolvessomekindofnationalist
identity,maywellrecognizeothersascitizens,andseektotreatthempreferentially
onthosegrounds.Othersmaybemorecosmopolitaninnature,andjustlookto
embraceothercitizensasfellowmoralagentswhoareworthyofconcern.Butthis
doesnotchangethefactthatmanyothers,whomaysortpeoplebasedonraceor
someotheridentifier,willseektotreatmanyfellowcitizensasmembersofout
groups.Becauseofthis,wecannotrelyonanassumptionofgoodwillbetween
citizens.Whilewedonotnecessarilyhavetosupposethattherewouldbean
adversarialrelationshipbetweencitizens,wecannotsupposethattherelationship
wouldbealtruistic.
Themainconcernthatadiversesocietymustfaceissomeformof
factionalization.Ifthevariouspartiesinasocietyhavenonaturalalignmentwith
eachother,theonusisonthepoliticalapparatusofthestatetoshowwhythey
shouldtietheirfatestoeachother.Ifthestateisunabletodoso,thosepartiesmay
findittobeintheirbestintereststosplitwitheachotherandformsmallerunions.


182
Undertheseconditions,thereasonpartieshavetostaytogetheristhattheyare
moreabletoprovidethemselveswithmaterialresourcesasalargerandmore
diverseeconomicunit.Butthisiscompellingtoanygivenpartyonlysolongaseach
partygetsmorethantheycouldhaveachievedontheirown.Itisherethatwecan
seetheimportanceofacontributionbasedaccountofdistribution.Ifwethink
aboutcontributionsintermsofwhatanagentoragentscouldhaveproducedon
theirown,wegetabaselinemeasureforwhattheymustgetinordertobe
rationallymotivatedtoremaininsociety.Ifthisgoesunsatisfied,itsuggeststhat
theirproductivecapacityissubsidizingotherpartsofsocietyaneffectivetaking
oftheirwealthandincome.ThisisnottoexpressaNozickeannotionofabsolute
propertyrights:itisnotthatinpractice,thisprincipleimpliesnosystemoftaxation.
Itjustimpliesthatratesoftaxationorotherformsoftransfermustbelimitedby
thisbaselineaccountofcontribution.Individualsmusthavearationalreasonto
remainintheirparticularsocietyratherthansomeothersociety,andifweareto
supposenoculturalconnections,wemustthensupposethatitisbecausethesociety
isprovidingthemwithbenefitsgreaterthantheycanfindwithsomeothersocial
arrangement(thatisheldtothesamesetofconstraints).134Ifweadheretoano
subsidyapproach,weassurethatthisisthecase.

134
This last condition is just meant to prevent objections that the wealthy could always
just claim that their alternative society is some kind of slave-owning kleptocracy where
they have unlimited control of resources and political power. The counter-factual
alternatives to be considered must be held to the same constraints that the actual society
is held to, otherwise the counterfactual societies would have the same stability problems
that we are aiming to resolve here.

183
Wecannowseethatdisagreementsabouthowtodivideresources,

whetheritisinactualdistributionofwealthandincome,orthechoicesoverwhich
publicgoods135toprovide,mustberesolvedinsuchawaysoastorespectthe
conditionthateachpartygetsmoreofwhattheywantthanwhattheycouldhave
providedforthemselves.Thisismadeeasierbytheearliernotedfactthatthemore
diversitycauseddisagreementthereis,themorethereistodistribute.Butthisfact
canonlyhelpussomuch.Whatwemustbemostcarefulofisthat,nomatterwhat
setofdistributivegoalswemighthave136,wedonotcauseanyonegrouptohave
reasontobreakofffromsocietytoprovisiontheirowngoods.
Asanexample,considerIraq.Wecansimplifytheexamplebyassuming
threecoalitions:Kurds,SunniandShia.LetsfurtherassumethatKurdscontrol
mostoftheoilwealth,butSunnisandShiitesprovideusefullaboringettingthatoil

135
Recall that public goods are defined as those goods that are both non-excludable and
non-rivalrous. A good is non-excludable if, in providing it to some people, you are
unable to prevent others from accessing it as well. An example of this is a lighthouse.
Even if one shipping company builds a lighthouse for the benefit of their own ships, they
cannot prevent the light from being seen by other boats. A good is non-rivalrous if it
does not cost anything to provide it to additional people. While, say, providing chairs to
people costs a fixed amount per new chair, operating the lighthouse does not cost extra if
more boats pass by.
136
Recall that we are being agnostic to particular goals of distributional justice. Some
factions of society may want to distribute money towards the poor, some might want to
distribute it towards the most productive members of society, or some might want to
distribute it towards religious organizations. The question I am attempting to address
here is largely independent of these concerns. I contend that a significant component of
my framework is that societies can experiment with different distributional goals to see
which ones best achieve the ends that society is interested in. As such, I do not wish to
make any arguments for which distributional scheme is the best. However, I am
interested in apportioning wealth and resources between competing factions in society.
These factions can do with the money what they will, but the claim that I am pushing is
that each faction must receive its fair portion of the resources in order for stability to be
maintained.


184
tomarket.Kurdsmaybewillingtoacceptdistributingwealthinthedirectionof
theothertwogroups,butonlyinsofarastheirabsoluteshareofwealthremains
largerthanwhatitwouldbeiftheywereautonomousandreceivednohelp.Iftheir
wealthshareeverdropsbelowthislevel,thentheywouldbesubsidizingtheother
coalitionsandreceivingnothinginreturn.Thereisnorationalreasonforgoing
alongwiththissubsidy,andsoKurdswouldsimplyexitfromthesocialcontract,
andformasmalleroneamongstthemselves.
Thegoalofproportionaldistributionistopurchaseaunitedcitizenry.Inthe
absenceofastrongculturalbond,weareleftonlywitheconomicties.Theseties
canhelptofacilitateculturalconnectionsinthefuture,buttheymustatleastinthe
beginningserveastheprimarymeansbywhichasocietystaystogether.By
distributingthegainsofaunioninsuchawaysoastomakesurethateveryfaction
ismadebetteroffthantheywouldhavebeenotherwise,eachfactionhasareason
forremaininginsociety.Whenproportionaldistributionbasedonproductionis
violated,variousfactionsmightfindtheirinterestsbetterservedbyforminga
separateunion.137

137
In some ways, similar issues sparked the Revolutionary War. Britain taxed imports at
high rates, demanded that colonies buy only British goods, and instituted a number of
other taxes that were excessive. Economic self-interest demanded that colonists have
some governmental representation to resolve these taxation issues, and it was denied.
Early instances of rebellion, like the Boston Tea Party, were explicitly economic
complaints.


TheDistributionofPublicGoods

185

Upuntilthispoint,wehavediscussedeconomicdistributionwithoutregard
towhetherwemeanthedistributionofwealthandincome,ortheprocurementof
publicgoods.Iwouldliketonowturntoashortdiscussionofpublicgoodsasthey
raisesomeconsiderationsthatdonotcomeupinthedistributionofwealthand
income.
Whenwedistributewealthandincome,theeffectsofthesedistributionscan
belargelyignored:ifpeoplewhomyouconsiderundeservingreceivemoney,you
canatleaststoppayingattention.Money,afterallisaprivategoodthat(atthe
margin)doesnothaveanyexternalities.Andonceonechoosestoignoreit,their
livescanlargelygoonwithouthavingtothinkaboutitagain.Thisisnotnecessarily
thecasewithpublicgoods.Publicparks,trains,andabortionclinics138canallbe
hardtoignoreifonehasadeepseatedobjectiontothestateprovidingthese
services.Whenonesimplydistributesmoney,itiseasytoimaginethatmuchofthe
moneywillbespentlargelyonfairlyinnocuouspurchases:food,clothingand
sheltertendtotakeupalargeportionofpeoplesexpenses.Butthisdoesnot
happenwhenpublicgoodsareprovided.Youcaneitherbehappywiththegoodor
not.Ifthestateprovidesfreeabortionclinics,manypeoplemaybegladtoknow
thatsuchservicesareavailableforthosethatneedthem,butothersmightdecide

138
Trains and abortion clinics do not meet the precise definition of public goods, but are
relevant in that providing them at a public scale involves collective action, and they have
externalities.


186
thatthestateisforcingthemtoendorsesomethingthattheywouldotherwise
vociferouslyobjectto.
Theprovisioningofpublicgoodsiswherewecanseethelimitsoftoleration,
andthemostlikelysourceoffactionalization.Ifweassumethatweareina
positivesumgame,thereisalwaysadistributionofwealthandincomesuchthat
everyoneismadebetteroffthanbefore.Butwithgoodsthathaveexternalities,itis
veryfeasiblethatagentscanbenegativelyimpactedbyhavingthemprovidedto
others.Thiscanbeseenevenwithrelativelyharmlessgoods,likepublicschools.
Ifoneregionorgroupdecidestoprovisionextremelyhighqualitypubliceducation,
inpartbypayinghighsalariestogetthebestteachers,theirinvestmenteffectively
crowdsouttheteacherinvestmentmadebyotherregionsorgroupswithlower
levelsofexpenditures.Soallthebestteacherswillflowtothehighsalaryschools,
andtheotherschoolsareleftwithlowerqualityteachers.Thiscreatesanegative
externalityonthestudentsinotherschooldistricts,byloweringthequalityoftheir
education.139Manygoods,however,canbecomeevenmorecontentiousthan
simplecasesinvolvingpublicschools.ConservativeChristiansarenotgoingtobe
happywithabortionclinics,eveniftherestofsocietywantsthemavailable.
Similarly,ifChristiansdecidetoputupcrossesortheTenCommandmentsinthe

139
This sort of case is best captured by the theory of local public goods. Local public
goods are close to public goods in that they exhibit non-excludability, but they are
partially rivalrous. That is, they face a congestion problem: town pools and public
schools can both become crowded. So this encourages localities providing local public
goods for their own communities. This was first developed in Tiebout, Charles M. "A
Pure Theory of Local Expenditures." Journal of Political Economy. 64 (October 1956),
416-424. It was extended to a model of clubs in: Buchanan, J.M. "An Economic Theory
of Clubs." Economica. 32 (1965), 1-14.


publicpark,nonChristianswouldhaveeveryreasontobecomeupsetatthe

187

apparentevangelism.Moregenerally,everyfactionislikelytohavesomepublic
goodthatitwouldliketoprovideforitsmembersthatotherfactionswouldfind
objectionable.Thesetofpublicgoodsthateveryonewouldlikeismostlikelygoing
tobefairlynarrow.
Therearetwopotentialsolutionstothisproblem.Thefirstistolimitpublic
goodstothesetofgoodsthateveryonecancometoagreeto.Thiswouldensure
thatnooneisupsetwithanyperceptionsofbeingforcedtoendorsesomethingthat
theywouldotherwiseobjectto.Butitwouldhavetheconsequenceofdrastically
limitingtheabilityofthestatetoprovisiongoodsthatmanypeoplewant.This
attempttolimitharmsmightinfactcreatemoreharmbydenyingsomanypeople
accesstothepublicgoodsthattheywant.Additionally,itcouldwellencourage
factionalizationonthebasisofagreementoverpublicgoods.Ifpeoplefeel
constrainedbytheperspectivesofothers,theywillfindasmallergroupthatshares
theirviewsonthedesirabilityofparticularpublicgoods,andformasocialcontract
withthem.
Forthesereasons,wemustturntothesecondpossibleavenue,whichisto
allowthedevelopmentofanygoodsthatpeoplearewillingtoproduce,solongas
thegoodsarepaidforonlybythosethatseekthem.140Aswesawinour

140
As money is a fungible good, it may be the case that the best a state can do is assure
that public expenditure on given goods is proportionate to interests. It would be nearly
impossible to guarantee that a particular dollar meant for parks was spent on parks as
opposed to the military.


188
examinationofrightsinthepreviouschapter,itispossibletobargainoverthese
goods,solongaseveryoneabidesbytheprinciplethattheypayforthegoodsthat
theyseek,anddonotattempttoforceotherstopayforthem.Itshouldbenoted
thatthiswouldlikelyconstrainspendinginanumberofarenas.Forexample,itis
unlikelythatsomethinglikemilitaryspendingcouldbemaintainedatthecurrentUS
level,ifpeoplecouldindicatehowtheirmoneyistobespent.141Thisbargaining
approachwouldencourageeithergoodsthatcanbeuniversalandthusgarner
reasonablyhighfundinglevels,orthosegoodsthatareoflimitedappeal,which
wouldsecurecommensuratelylowerlevelsoffinancialsupport.
Aconsequenceofthissystemofprovidingpublicgoodsisthatthosegoods
thatareseenlargelyaswealthtransferswouldeitherhavetobelimitedinsize,or
universalinnature.Asystemofwelfarepayments,then,wouldeithermeanfairly
smallcheckstopoorpeople,orlargercheckstoeveryone.142Onthisapproach,we
shouldexpecttoseeanumberofsmallergoodsbethedominantmodeofpublic
production.Thisispurelyafunctionofthenatureofcollectiveactionproblems.

141
While it is clear that the state has an interest in requiring a certain funding level for a
defensive force, offensive force is less clearly essential. It may be that, in times like
WWII, the public would be willing to provide money for an offensive action. But this
approach to appropriating money would make it very difficult to fund unpopular wars,
like Vietnam. In terms of the system we are outlining, unpopular wars must be seen as an
unwanted subsidy. The productive capacity of the state could have been redirected
towards other goods that more people want. Popular wars, on the other hand, would not
suffer from this problem.
142
This assumes that some portion of the population is not interested in wealth transfers,
like those that are compelled by the Cadillac driving Welfare Queens argument that we
examined earlier. If everyone finds wealth transfers appealing, then the circumstances
change, and it would be possible for larger, targeted welfare systems. But this is a fairly
unlikely scenario in a diverse society.


189
Lessexpensivegoodsrequirefeweragentstopayforthem,andasitiseasierto
arrangeacoalitionofsmallersizesthanlargersizes,smallergoodsaremorelikely
tobepaidforthanlargergoods,evenifthelargergoodsareevenmoredesirable.
Thoughthisbiastowardtheproductionofsmallerscalegoodsmayseem
problematic,itinfacthasanumberofvirtues.Firstamongthemisthatasociety
cancomparetheeffectivenessofanumberofsmallerprojectswithsimilargoals,
andovertimeincreasesupportfortheprojectorprojectsthathaveproventobethe
mostsuccessful.Andforthoseprojectsthatmaybemetwithskepticisminitiallyby
thepublicatlarge,itprovidesanopportunitytodemonstratetheirusefulnessona
smallscaletobuildlargersupport.Inasocietyfilledwithdiverseperspectives,we
mustassumethatmanydifferentgroupswilladvocatealternativeapproaches.
Whatthisbiastowardssmallerprojectsdoesisallowallgroupstohavethe
opportunitytotesttheirideas,andtrytoprovetoothersthattheirapproachhasthe
mostmerit.Itallowsallgroupstobeincludedinongoingsocialexperimentation,
andcanorganicallyrepresentmountingsupportforagivenapproachwhenthat
goodisabletoappropriatemorefundingthanotherapproaches.Ratherthantryto
resolvethesedifferenceswithaprioridebatesandtheimpositionofauniversal
solution,thisbiasallowsactualexperiencetodictatewhatapproachesshouldwin
thedebate.Followingthisprinciple,thestatecanbeseenintheroleofa
clearinghouseofsortsfortheprovisioningofrightsandpublicgoods.Individuals


190
canstatetheirinterests,andtheirwillingnesstopay,andonceacoalitionexists
tosupporttheproductionofagood,thestatesanctionsitscreation.143
Thereisstilloneremainingchallengeinhowwetreattheprovisionofpublic
goods.EvenifChristianspayfortheircrosses,Muslimspayfortheirmosques,and
womensrightsactivistspayforabortionclinics,theyareimposingexternalitieson
otherssimplybecauseoftheexistenceofthesegoods.Peoplewhodonotsubscribe
toChristianityorIslamcouldeasilybeoffendedbyblatantreligiosityifitviolates
theirownviews.Antiabortionadvocatescouldfeelharmbythepresenceof
abortionclinicsiftheyseethemassitesofmurder.Thirdpartiescansufferharms
fromthetransactionorcreationofgoods,eveniftheyhavenopartinthem.Thisis
easilyseeninthecaseofsecondhandsmoke:breathinginsmokeharmspeople
evenwhentheydidnotchoosetosmokeacigarette.Likewise,forcedexposureto
goodsthatrepresentvaluesthatareseenasunderminingonesownispotentially
harmful.
Localpublicgoodsarenottheonlygoodsthathavethisproblemwith
externalities.Anotherpotentialsourceofexternalitiesthatmustbeaccountedfor
arepositionalgoods.Positionalgoodsarethosegoodsthatderivetheirvaluefrom
theirrelationtoeachother.Inparticular,positionalgoodssupposesomerankingof
statusordesirability.Peopleusepositionalgoodstocompeteforstatuswitheach
other.Becauseofthisquality,competitionsoverpositionalgoodsarezerosumin

143
This view has several implications about the broader function of the state, as well as
the dynamics of goods production. However, I will leave developing this for future
work.


191
nature.Gainingpositionmustmeanthatsomeoneelsehaslostinposition.This
zerosumcharacteristicofpositionalgoodsimpliesthattheyhavenegative
externalities.Examplesofpositionalgoodsarehavinganexclusiveaddress,ora
biggerthanneededhouse,oralargediamondononesengagementring.Theseare
soughtafternotfortheirintrinsicqualities,buttodemonstratestatus.Likewise,
peoplespendmoneyoncarsthatgowellaboveanyspeedlimitstocompeteover
whohasthefastestcar,eventhoughthatspeedcanneverbelegallyattainedon
publicroads.144Thesegoodsexistprimarilyfortheirexternaleffects.Assuch,they
oughttobesubjecttothesameprinciplesthatgovernpublicgoods.
Thepreviouschaptersetupamechanismfordealingwithbalancingoutthe
positiveandnegativeexternalitiescreatedbyrightsallocations.Thismechanism
canalsobeappliedtopublicandprivateproperty,butherewewillseesome
additionalcaveats.Inparticular,inthepreviousprocedure,allrightscouldbe
pricedintermsofthelimitationsonotherrightsthattheycreated.Whenwehave
goodsthathaveamonetaryvalue,thispricingmechanismislessstraightforward.
Therootoftheproblemisthatwehavetwodifferenttypesofgoodsthatweare
attemptingtodistribute.First,wehavethewealthandresourcesthatwehave
alreadycommittedtodistributeinproportiontocontribution,soastonotviolatea

144
Some positional goods are also Veblen goods, which are commodities for which the
demand curve is the inverse of what it is for normal goods. That is, as the price
increases, demand increases. This is because the price is part of the value of the good,
insofar as it can be used as a signaling device for status and quality. Expensive wine,
chocolates, cars and jewelry can fall into this category.


nosubsidyprinciple.Butwhenwefactorinexternalitiesaswell,wehaveto

192

relyonthelanguageofharms,aseconomicexternalitiesarebytheirnaturenot
priced.145
Assuringthattheinitialdistributionofwealthandincomesatisfiestheno
subsidyprincipleandsimultaneouslythattheharmscausedbytheexternalities
frompublicandpositionalgoodsareequallydistributedacrosssocietywould
requireveryspecialcircumstances.Itisineffectamutualmaximizationproblem,
whichcannotbesolvedexceptinveryunusualandconstrainedcircumstances.
Becauseofthis,oneoftheseprincipleswillhavetobetreatedassubordinatetothe
other.However,nomatterwhichprincipleischosentobesubordinate,wewill
necessarilyweakenthecaseforstability.Ifwechoosethenosubsidyprinciple,
thenitmeansthatsomegroupwillbesubsidizingothergroups,whichisnotintheir
rationalselfinterest.Ifwechoosetheprincipleofequalburdensharing,thensome
groupwouldhavetotakeonmoreburdensthanothers,whichtheymightconsider
tobeunjustifiable.Theimportantquestionisthennotwhetheroneofthese
principleshastobeviolated,becauseonewill,buthowmuchitisviolated.Thiswill
dependontheparticularcircumstances:themixofgoodsthatpeopleareinterested
in,therangeofperspectivesandutilityfunctions,andtheproductivecapacityof
variousgroups.Smalldeviationsfromtheseprinciplesmightbeacceptable,butas
theviolationsbecomelarger,socialstabilitywillbecomemoreandmoretenuous,as

145
If externalities were priced, that would mean that the social costs of the good would be
included in the price of the good. But then the outside harms would not be externalities:
they would just be costs. And those costs would be included in the price, and just made
internal. So there would no longer be any externalities.


193
thesocialglueofrationalselfinterestwillweakenandeventuallydisappear,at
leastforsomeportionofsociety.
Whatthismeansisthattherewillbecaseswherediversesocietieswill
simplynotbestableenoughtomaintainthemselvesasasinglepoliticalunit.Inthe
casewheredisagreementoutstripsthegainsinsocialsurplus,societieswilland
shouldbreakapart.However,giventhatdiversityisabletogeneratesomuch
socialsurplus,andourdistributiveapproachisdesignedtopromoteasocial
experimentalapproachtoresolvingdisagreementsaboutpublicgoodsprovisioning,
whileattemptingtoassurethatnogroupissufferingincomparisontorealistic
alternativesocialarrangements,socialstabilityshouldbethenorm.However,itis
importanttorealizethat,absentextensivesocialculturalbonds,wemustexpect
thatnoteverycollectionofpeoplecansuccessfullyformandmaintainapolitical
union.Thoughtherearemanyincentivesforpromotingdiversityinsociety,there
arealsolimitsonhowfarthoseincentivescantakeus.

Conclusion:TheLimitsofDiversity
Thetraditionalliberalaccountoftolerance,asexemplifiedinLockesEssay
OnToleration,isoneofduties.Weoughttotolerate(religious)differencesinothers,
andnotforceourviewsontheminpartbecausewewouldnotwantthemtobeable
todothesametousifthetableseverturn.Butthisviewdoesnotpresentanycase
forwhyonemightwantamorediversesociety.Instead,theassumptionisthat


diversityisaproblem,anditneedstobemanaged.Thismaybeasatisfactory

194

accountinrelativelyhomogenoussocieties,butassocietiesbecomemorediverse,
wecannotdependonpeoplefollowingobligationsthattheyperceivetobeonly
burdensandrestrictionsontheiractivity.AstheReaganexamplefromthe
beginningofthischapterdemonstrates,peoplearequicktoseepeopledifferent
fromthemselvesasaburden,andfromthere,peopleseektorelievethemselvesof
theperceivedburden.
PartofwhatIhavearguedforinthischapterisanalternativeconceptionof
tolerance.Insteadofrelyingonanotionofobligation,Ihavearguedthatindividuals
haveaselfinterestedreasonforwantingtopromotediversityintheirsociety.
Diversitycreatessocialsurplus.Diversityalsocreatesnewopportunitiesforsocial
innovation,openinguppossibilitiesforlivingoneslifethatwerenotpreviously
available.Thispositivecasefordiversitywilltypicallyoverwhelmtheperceptionof
burdensimposedonindividualsfrombeinginadiversesociety.Whilethisself
interestedembraceofdiversitywillnotalwaysultimatelywinout,itwillhelpto
promoteandsustainsocietiesmuchmorediversethanwhataLockeanaccountof
tolerationcould.
Thoughitisafactthatsocietiesarebecomingmorediverse,onecould
potentiallyarguethatratherthanabandontheliberalaccountoftolerance,along
withotheraspectsoftheliberalframework,inordertoadapttothesemorediverse
societies,weshouldaimtopullbackandreturntomorehomogenoussocieties.
However,indoingso,wewouldlosesomethingthatonlyreallybecomespossiblein


adiversesociety.Withanembraceofdiversitycomesrecognitionofthe

195

importanceofexperimentsinliving.Theseexperimentsprovidetheevidence
neededtofurtherrefineourmoralconcepts,ourideasaboutwhatrightsweought
toguarantee,andhowweoughttoorganizeourlivesasconstrainedbyormoral
beliefsandourrightsandprivileges.Differentideasofsocialimportance,like
provisioningpublicgoods,canbetriedandsuccessfulideaswillslowlybecome
moredominantaspeoplelearnabouttheirsuccess.
Asthebenefitsofdiversitymustbeweighedagainstitscosts,Ihaveargued
fortheconstraintsnecessaryforthepromotionofbothdiversityandtheframework
ofsocialexperimentation.Paretooptimality,thenosubsidyprinciple,andequal
burdensharingaimtoassurethateveryindividualismadebetteroffintheir
currentsociety,ascomparedtopossiblerelevantalternatives.Theseconstraints
aimtoassurethestabilitynecessaryforthebenefitstodiversitytomanifest
themselves.Ifpeoplefeelliketheyarebeingtreatedfairly,andnotbeingtaken
advantageofbyothers,theybecomemorewillingtoembracediversity,andthe
constantflowofchangesandopportunitiesitbrings.

196

Conclusion

Letusnowconsiderwhathasbeenshown,andthenlooktowhatworklies
ahead.Whatwehaveseenisthatifdiversityistakenseriously,muchofsocial
contracttheoryissubjecttorevision.Inparticular,Ihavearguedforthecentral
roleofperspectivesinthedesignoftheprocedurestogenerateasocialcontract.
Thiscentralroleofperspectivesleadsustoabandonthemodelofpublicreasonfor
politicaldiscourse,andreplaceitwithabargainingmodel.Thisisfurther
supplementedbyanosubsidyprincipleforthedeterminationofthedistributionof
wealthandincomeinsociety.
Thoughthesecomponentshavebeenmotivatedfromaninterestin
developinganaccountofsocialcontracttheoryfordiversestates,theyalsooffera
distinctepistemologicalapproachtosocialcontracttheory.Thisismostexplicitin
theaccountoftheViewfromEverywhere,inwhichIarguethatamoralstancemust
properlybeunderstoodasasocialstance.Butthisapproachcanbeseen
throughout.Thebargainingmodelarticulatesamethodofpoliticalnegotiationthat
doesnotundulyburdeneithersidewithhavingtotakeonperspectivesthattheyare
notabletotakeon.Andthisbargainingmodelisimportantlycheckedbythereal
worldresultsthatitproduces.Inthiscontractframework,individualsinsocietyare
askedtotryoutparticularrightsallocations,andthentryoutdifferentwaysof
livingwithintheconfinesofthatrightsframework.Astheybetterunderstandwhat
optionsforlifetheyhavewithinaparticularrightsregime,individualscanthen


assesswhetherornottheywouldliketokeep,refine,orabandonthecurrent

197

contractandfollowtheproceduresagainarmedwithnewinformation.Itisthis
abilitytoreturntotheproceduresforformingasocialcontractthatisstarkly
opposedtootheraccountsofsocialcontracttheory.However,Itakethistobeakey
improvement.Wecanonlyassessthewisdomofparticularpoliticaldecisionsonce
wehavesomeinformationaboutthem.Often,thismeansthatwemustseewhatit
isliketolivewiththem.Butifwearetotakeamoreexperimentalapproachtothe
choicesthatframethebasicstructureofsociety,wemustnotbeforcedtostickwith
choicesthathaveproventobemistakes.Revisingourcontractallowsustobeable
totakemoreriskswithouthavingtobepermanentlystuckwiththenegative
consequencesiftheyarise.
ThoughItakethisprojecttohavedevelopedthekeycomponentsofasocial
contract,throughthearticulationofainitialmoralstandpoint,aprocedureforthe
distributionofrights,andanaccountofthedistributionofwealthandincome,there
ismuchthatIwishtodevelopinfuturework.Thisprojecthasledmetobegin
thinkingaboutanalternativeaccountofthestateanditsroleinpublicdecision
making.AsIbrieflynotedinthefinalchapter,Ihavethebeginningsofaviewofthe
stateasaclearinghouseforrightsandgoods.ThiscanalsobeseenintheView
FromEverywhereandinthebargainingmodel.Thestatesmainfunctionistoserve
asafacilitatorfortheviewsandinterestsofthepeople.Onthisview,legislators
wouldnothaveclearpowersbeyondbeingproxiesfortheirconstituents.

198
Ibelievethatthishaswiderimplicationsforanaccountofthestate.The

politicalphilosophyliteraturegenerallytakesacountrysbordersasfixed,anddoes
notoftenaddresstheroleofstatesorregionsversusacentralgovernment.Inlarge
part,thisisbecausepoliticaldisagreementisaddressedatahigherlevel,and
regionaldifferencesareabstractedaway.Ifweextendtheaccountofstability
developedinthefinalchapter,however,wecanfindthatmorediversestatescan
remainmorestableifweallowforgreaterregionalautonomy.Thatis,ratherthan
requiringeveryonetoagreetoallofthesamerules,wecanallowasmallersubsetof
agreementtoestablishacentralgovernment,andtheneachregioncanbetterfocus
onlargerlevelsofagreementonmoremattersofjustice.Thisparallelsa
confederationmodelofgovernmentasopposedtoafederationmodel,which
requiresmuchmorecentralcontrol.
Ibelievethatadoptingaconfederationmodelleadstotwoimportantideas.
First,thisanalysispullsusawayfromthemodelofthenationstateasan
appropriatepoliticalunit,andtowardsamodeloffederationsofcitystates.This
increaseinregionalismallowsforlargerscalesocialcooperationwhereitcanbe
agreedupon,butreducesconflictbyallowingforgreaterlocaldifferentiation.
Second,itwillallowmetoextendthisanalysistodevelopamorerobustaccountof
Milleanexperimentsinliving.Onthisaccount,notonlydoindividualshavethe
freedomtotrysmallscaleexperimentsinliving,butstatesarethenencouragedto
conductlargerscalesocialexperimentation.Onthisview,citystatesmustineffect
competeforcitizens,andarethusdriventoexperimentinwhatpublicgoodsthey


199
provide,andwhatlifeplanstheycanbestsupport.Successfulcitystatesdraw
inmoremembers,whichinturnpushesothercitystatestoadoptsimilarpoliciesas
theyattempttoinnovateontheirown.
Thiscompetitivemodelofregionalconfederationsallowsmetopresentthe
entirepoliticalframeworkasadynamicprocess.Whilemypreviousworkhas
establishedtheepistemologicalbenefitsofleveragingdiverseperspectives,ithas
notconsideredlongrunsocialdynamics.Thesedynamicsnotonlyallowusto
developaricheraccountofstability,butalsodeepenourunderstandingofthetrue
roleofsocialexperimentationinshapingliberalsocieties.Inthecompetitivemodel
thatIlooktodevelop,theevolutionarydynamicsofthesocialcontractarebrought
intotheforefront.Thisisessentialtounderstandingtheeffectsofaprocedural
accountofjustice.Inthissense,Icanclaimthatjusticeshouldbeunderstoodasa
trajectory,andnotastateofbeing.Ratherthanclaimingtoknowwhattheidealsof
justiceare,thisviewarticulatestheprocedurebywhichwecanfindthem.

200

Bibliography

KennethArrow,ADifficultyintheConceptofSocialWelfare,JournalofPolitical
Economy58(4)(August,1950),pp.328346
KennethArrow,SomeOrdinalistUtilitarianNotesonRawlsTheoryofJustice,
JournalofPhilososphy,1973.
CristinaBicchieri,CooperationandCommunication:GroupidentityorSocial
Norms?inN.Gold[ed],Teamwork:MultiDisciplinaryPerspectives(London:
PalgravesMacmillan,2005).
CristinaBicchieri,TheGrammarofSociety(Cambridge:CambridgeUniversityPress,
2006)
MarilynnBrewer,TheSocialSelf:OnBeingtheSameandDifferentattheSame
Time.PersonalityandSocialPsychologyBulletin,17,(1991)475482.
James M. Buchanan, "An Economic Theory of Clubs." Economica. 32 (1965), 1-14.
ThomasCarlyle,OccasionalDiscourseontheNegroQuestionFraser'sMagazinefor
TownandCountry,London,Vol.XL.,(February1849).
GregoryClark,AFarewelltoAlms.(Princeton:PrincetonUniversityPress,2007)
RonaldCoase,TheProblemofSocialCost,JournalofLawandEconomics(October
1960)
RonaldCoase,TheFirm,TheMarket,andTheLaw(Chicago:UniversityofChicago
Press,1990)
DonaldDavidson,OnTheVeryIdeaofaConceptualSchemeProceedingsand
AddressesoftheAmericanPhilosophicalAssociation,Vol.47,(19731974),pp.520.
JaredDiamond,Guns,GermsandSteel(NewYork:W.W.Norton&Co.,1997)
MarcEreshefsky,"SpeciesPluralismandAntiRealism",PhilosophyofScience,65
(March1998)103120
MiltonFriedman,CapitalismandFreedom(Chicago:UniversityofChicagoPress,
1982).
DavidGauthier,MoralsbyAgreement(NewYork:OxfordUniversityPress,1999)
AllanGibbard,"Manipulationofvotingschemes:ageneralresult",Econometrica,
Vol.41,No.4(1973),pp.58760


201
Jurgen Habermas, Reconciliation through the Public Use of Reason: Remarks on
John Rawlss Political Liberalism, The Journal of Philosophy XCII(3), 1995. 10931.
RussellHardin,CollectiveAction(NewYork:RFFPress,1982)
JohnHarsanyi,"CardinalUtilityinWelfareEconomicsandintheTheoryofRisk
Taking",JournalofPoliticalEconomy61(5)(1953)4345
FriedrichHayek,TheUseofKnowledgeinSociety,AmericanEconomicReview.
XXXV,No.4.(1945)pp.51930.
AlbertHirschman,ThePassionsandtheInterests.(Princeton:PrincetonUniversity
Press,1997)
ThomasHobbes,Leviathan,(NewYork:Hackett,1994)
SaumitraJha,Trade,InstitutionsandReligiousTolerance:EvidencefromIndia
(January10,2008).StanfordUniversityGraduateSchoolofBusinessResearchPaper
No.2004AvailableatSSRN:http://ssrn.com/abstract=948734.
JohnLocke,Locke:PoliticalEssays.MarkGoldie[ed](Cambridge:Cambridge
UniversityPress,1997)
LuHongandScottPage,Groupsofdiverseproblemsolverscanoutperformgroups
ofhighabilityproblemsolversProceedingsoftheNationalAcademyofSciences,
November16,2004
HongandPage,ProblemSolvingbyHeterogeneousTeamsJournalofEconomic
Theory2001
Lloyd,Lisa.TheCaseoftheFemaleOrgasm:BiasintheScienceofEvolution.(Boston:
HarvardUniversityPress,2005)
J.S.Mill,TheNegroQuestion,Fraser'sMagazineforTownandCountry,London,Vol.
XL.,(February1849).
Mishler,B.andDonoghue,M.,1982,"SpeciesConcepts:ACaseforPluralism",
SystematicZoology,31:491503
ThomasNagel,TheViewFromNowhere(Oxford:OxfordUniversityPress,1986)
ThomasNagel,RawlsonJustice.ThePhilosophicalReview,Vol.82,No.2.(Apr.,
1973),pp.220234
RobertNozick,Anarchy,State,andUtopia(NewYork:BasicBooks,1974)
MarkA.Satterthwaite,"StrategyproofnessandArrow'sConditions:Existenceand
CorrespondenceTheoremsforVotingProceduresandSocialWelfareFunctions",
JournalofEconomicTheory10(April1975),187217


202
MancurOlsen,TheLogicofCollectiveAction(Boston:HarvardUniversityPress,
1971)
ScottE.Page,bookTheDifference(Princeton,PrincetonUniversityPress,2007)
JohnRawls,ATheoryofJustice(NewYork,BelknapPress,1971)
JohnRawls,"TheIdeaofanOverlappingConsensusOxfordJournalforLegalStudies
(Spring1987),7(1):125
JohnRawls,KantianConstructivisminMoralTheory,JournalofPhilosophy
(September1980),77(9):515572
JohnRawls,LecturesontheHistoryofPoliticalPhilosophy,SamuelFreeman,ed.
(NewYork,BelknapPress,2007)
JohnRawls,PoliticalLiberalism(NewYork,ColumbiaUniversityPress,1993)
JohnRawls,SomeReasonsfortheMaximinCriterion,AmericanEconomicReview
(1974)
Rousseau,DiscourseonPoliticalEconomyandTheSocialContract.ChristopherBetts,
[ed](Oxford:OxfordUniversityPresss,2009)
JustinScheck,MackerelEconomicsinPrisonLeadstoAppreciationforOilyFillets,
http://online.wsj.com/article/SB122290720439096481.html
Sen,InequalityReexamined(Boston:HarvardUniversityPress,1992)
JSidanius,FPratto,MMitchell,InGroupIdentification,SocialDominance
Orientation,andDifferentialIntergroupSocialAllocation.JournalofSocial
Psychology,134(2),151167,1994
AdamSmith,TheTheoryofMoralSentiments(NewYork,LibertyFundInc,1984)
AdamSmith,TheWealthofNations(NewYork,BantamClassics,2003)
Tajfel,H.,Billig,M.,Bundy,R.P.,&Flament,C.Socialcategorizationandintergroup
behavior.EuropeanJournalofSocialPsychology,1,149177.(1971)
CharlesM.Tiebout,"APureTheoryofLocalExpenditures."JournalofPolitical
Economy.64(October1956),416424.
WilliamWimsatt,Robustness,ReliabilityandOverdetermination,inRe
EngineeringPhilosophyforLimitedBeings(Boston,HarvardUniversityPress,2007)
WolpertandMacready,NoFreeLunchTheoremsforSearch(TechnicalReport
SFITR9502010).SantaFeInstitute,1995


M.EYaariandM.BarHillel,OnDividingJustly.SocialChoiceandWelfare
(1984)1:124

203

H.PeytonYoungs,Equity:InTheoryandPractice,(Princeton:PrincetonUniversity
Press,1995)