Está en la página 1de 31
An Egalitarian Law of Peoples Author(s): Thomas W. Pogge Source: Philosophy & Public Affairs, Vol. 23,03/201 4 16:00 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . . JSTOR is a not-for-profit service that helps scholars, re searchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact . Wiley is collaborating with JSTOR to digitize, preserve and extend access to Philosophy &Public Affairs. " id="pdf-obj-0-2" src="pdf-obj-0-2.jpg">

An Egalitarian Law of Peoples Author(s): Thomas W. Pogge Source: Philosophy & Public Affairs, Vol. 23, No. 3 (Summer, 1994), pp. 195-224 Published by: Wiley

Stable URL: .

Accessed: 04/03/2014 16:00

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .


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


An Egalitarian Law of Peoples Author(s): Thomas W. Pogge Source: Philosophy & Public Affairs, Vol. 23,03/201 4 16:00 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . . JSTOR is a not-for-profit service that helps scholars, re searchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact . Wiley is collaborating with JSTOR to digitize, preserve and extend access to Philosophy &Public Affairs. " id="pdf-obj-0-33" src="pdf-obj-0-33.jpg">

Wiley is collaborating with JSTOR to digitize, preserve and extend access to Philosophy &Public Affairs.


An EgalitarianLaw of Peoples

Expanding on a brief sketch of over twenty years ago, Rawls has recently

offereda moredetailed extension of his theoryof justice

to the international

domain.' Like that first sketch, the "lawof peoples"he now proposeshas no egalitariandistributive component. In my own extension of Rawls'sframe- work, I had argued that a criterionof globaljustice must be sensitive to internationalsocial and economicinequalities.2 Here I take anotherlook at this issue in light of Rawls'snew and more elaboratedeliberations about it.

There are three components of Rawls'sconception of domestic justice that, in his view (LP, p. 51), qualifyit for the predicate"egalitarian":

  • (i) His first principleof justice requires that institutionsmaintain the

fair value of the politicalliberties, so that persons similarlymotivated and endowed have, irrespective of their economic and social class, roughly

equal chances to gain politicaloffice and to influence that shape their lives (cf. TJ, p. 225).

the politicaldecisions


His second principle of justice requires that

institutions maintain

Workon this paperwas supportedby a LauranceS. Rockefellerfellowship at the Princeton UniversityCenter for HumanValues. Written for a conferenceon "theEthics of Nationalism" at the Universityof Illinois at Urbana-Champaign,it was also presented to audiences at Princeton, Harvard,New York,Stanford, and Oxford Universities. I am grateful for the barrageof forceful criticisms I have received-in particularfrom Alyssa Bernstein, Brian Barry,Lea Brilmayer,Anthony Coady,John Cooper,Roger Crisp, Peter de Marneffe,Alan Houston,Frances Kamm, Elizabeth Kiss, Christine Korsgaard, Ling Tong, Stuart White, and the editors of Philosophy& PublicAffairs.

  • i. John Rawls,"The Law of Peoples,"in On HumanRights, ed. Stephen Shute and Susan Hurley(New York:Basic Books, i993), pp. 4 i-82, 220-30. Pagenumbers preceded by "LP" refer to this lecture. The earliersketch is on pp. 378f of A Theoryof Justice (Cambridge, Mass.: HarvardUniversity Press, 1971); henceforthTJ.

    • 2. See Chapter6 of RealizingRawls (Ithaca:Cornell University Press i989); henceforth



Philosophy & Public Affairs

fair equalityof opportunity,so that equallytalented and motivatedpersons have roughly equal chances to obtain a good education and professional

position irrespective of their initial social class (cp. TJ, pp. 73, 301).


His second principle also requires that, insofar as they generate

social or economic inequalities, social institutionsmust be designed to the maximum benefit of those at the bottomof these inequalities (the differ-

ence principle-cf.

TJ, pp. 76f).

Each of these egalitarian components furnishes separate grounds on which the currentbasic structureof the United States can be criticizedfor producingexcessive inequalities. Analogouspoints can be made aboutour current worldorder:

  • (i) It fails to give membersof differentpeoples roughly equal chances to

influence the transnationalpolitical decisions that shape their lives.


It fails to give equallytalented and motivatedpersons roughlyequal

chances to obtaina goodeducation and professionalposition irrespective of the society into which they were born.


It also generatesinternational social and economic inequalities that

are not to the maximum benefit of the world'sworst-off persons.

These observationsare certainly true. The question is: Do they show faults in the existing global order?

Rawls'slaw of peoples contains no egalitariandistributive principle of any sort; and he seems then to be committedto the view that none of the

three analogous criticismsis valid, even though the analogue to his third egalitarian concern:

he explicitly attacks only the proposal of a global

difference principle. My own view still is that all three of the analogous

egalitarianconcerns are valid in a world characterizedby the significant

political and economic interdependencies that exist today and will in all likelihoodpersist into the indefinite future. Here I will, however, defend against Rawls a much weaker claim: A plausible conception of global justice must be sensitive to internationalsocial and economic inequalities.

My focus on this

agree with much in

one disagreement should not obscure the fact that I this Amnesty Lecture-both substantivelyand meth-

odologically.Substantively, I agree with his view that a just

worldorder can

contain societies governedby a conception of justice that differs from his own political liberalism by being nonpolitical, nonliberal, or both (LP, pp. 42f, 46); and that a main demand to make upon how their institutions

  • 197 An Egalitarian Law of Peoples

work domesticallyis that they secure human rights (LP, pp. 6i-63, 68-

71). Methodologically,I agree that it is too early to tell how his idea of the original position-initially devised to deal with a closed, self-contained

society (TJ, pp. 4, 8, 457)-should best be our interdependentworld (LP, pp. 50, 6sf).

adaptedto the complexities of Variouspossibilities should be

worked out in some detail. One main strategy is Rawls's:Apply the two principles to the basic structureof a national society,and then reconvene the parties for a second session to deal with the relations among such societies. Anothermain strategyis to start with a global originalposition

that deals with the worldat large, even asking, as Rawlsputs it (somewhat

incredulously?),"whether, and in what

form, there should be states, or

peoples, at all" (LP, p. 5o). Variantsof this second strategy have been entertained by David Richards, Thomas Scanlon, Brian Barry,Charles Beitz, and myself. I can leave aside this second strategy here, because internationalegalitarian concerns can easily be accommodatedwithin the first strategy;as we shall see, Rawls simply decides against doing so. Myfocus on one disagreementshould also not obscurethe fact that there are others. Two of these are relevanthere. First, I do not believe that the notion of "apeople" is clear enough and significant enough in the human

world to play the conceptual role and to have the moral significance that Rawls assigns to it. In many parts of the globe, official borders do not correlatewith the main characteristicsthat are normallyheld to identify a people or a nation-such as a common ethnicity,language, culture, his- tory,tradition. Moreover, whether some groupdoes or does not constitutea people would seem, in importantways, to be a matterof more-or-lessrather than either-or.I have suggested that these complexities might be better accommodatedby a multilayeredinstitutional scheme in which the powers

of sovereigntyare verticallydispersed rather than heavilyconcentrated on the single level of states.3 But I will set aside this topic as well. Let us

assume that there really is a clear-cut distinction

between peoples and

other kinds of groupings,that every person belongs to exactly one people, and that each nationalterritory really does, nearlyenough, contain all and only the members of a single people. In this highly idealizedcase, egalitar- ian concerns would seem to be least pressing. Hence, if I can make them

plausiblefor this case, they should be plausiblefor morerealistic scenarios as well.


"Cosmopolitanismand Sovereignty,"Ethics 103



Philosophy & Public Affairs

Second, I do not believe that Rawls has an adequate response to the historicalarbitrariness of national borders-to the fact that most borders have come about through violence and coercion. He writes:

From the fact that boundariesare historicallyarbitrary it does not follow that their role in the law of peoples cannot be justified. To wit: that the boundariesbetween the several states of the United States are histori-

cally arbitrarydoes not argue to the elimination of our federal system, one way or the other. To fix on their arbitrarinessis to fix on the wrong

thing. The right question concerns the political values served

by the

severalstates in

a federalsystem as comparedwith the values servedby a

central system. The answer is given by states'function and role: by the

political values they serve as subunits, and whether their boundaries

can be, or need to be, redrawn,and much else. (LP, p. 223


Let us suppose that the mere fact of historicalarbitrariness is indeed no argument against the status quo, that a forwardlookingjustification suf- fices. What such a justification should be able to justify is threefold:that there should be boundariesat all, that they should be where they are now, and that they should have the institutional significance they currently have. I am not interested in the first two issues: Let there be national bordersand let them be just where they are today.The issue I am raising is the third:How can Rawlsjustify the enormousdistributional significance nationalborders now have, and in a Rawlsianideal worldwould continue to have, for determining the life prospects of persons born into different states? How can he justify that boundariesare, and would continue to be, associatedwith ownershipof, full controlover, and exclusive entitlementto

all benefits from,land, naturalresources, and capitalstock? It is


that, in the midst of discussing nationalborders, Rawls switches

to consid-

ering state borderswithin the U.S., which have virtuallyno distributional significance. It does not reallymatter whether one is bornin Kansasor in Iowa, and so there is not much to justify, as it were. On the other hand, it mattersa greatdeal whether one is borna Mexicanor a U. S. citizen, and so we do need to justify to a Mexican why we should be entitled to life pros- pects that are so much superiorto hers merely because we were born on the other side of some line-a difference that, on the face of it, is no less morallyarbitrary than differencesin sex, in skin color,or in the affluence of one's parents. Justifying this is more difficult when national bordersare

  • 199 An Egalitarian Law of Peoples

historically arbitraryor, to put it more descriptively,when the present distributionof nationalterritories is indeliblytainted with past unjust con-

quest, genocide, difficulty as well

colonialism,and enslavement. But let me set aside this and focus on moral rather than historical arbitrariness.

Let us assume that peoples have come to be matched up with territoriesin

the morallymost benign way one can conceive. My defense, against Rawls, of an egalitarianlaw of peoples labors then under a self-imposed triple handicap: I accept Rawls's stipulation that global justice is addressed in a second session of the original position, featuringrepresentatives of peoples who take the nation state system as a given; I accept Rawls'sfantasy that the world'spopulation neatly divides into peoples cleanly separatedby nationalborders; and I waive any support my egalitarianview could drawfrom the role that massive past crimes have playedin the emergence of currentnational borders. I make these conces- sions strictlyfor the sake of the argument of Sections I-V and otherwise stand by my earliercontrary positions.


Some of the arguments Rawlsadvances against incorporatingan egalitar- ian componentinto the law of peoples are pragmatic,mainly having to do with inadequate administrativecapabilities and the dangers of a world government.To make it easier for you to assess these worries,I want to put before you a reasonably clear and specific institutional proposal and thereby give our central disagreementa concrete institutionalform. I lack the space, however, to develop and defend a complete criterionof global justice and to show what specific institutional arrangements would be favoredby this criterion.I will thereforeemploy a little shortcut.I will make

an institutionalproposal that virtuallyany plausibleegalitarian conception of global justice would judge to be at least a step in the right direction.

Rawls'slaw of peoples, by contrast,would not call for such a

step. It would

permit the step among consenting peoples, but would not view it as re- quired or suggested by justice. When sketching how a property-owningdemocracy might satisfy the difference principle,Rawls entertains a proportionalincome or consump- tion tax with a fixed exemption.The tax rate and exempt amount are to be

set so as maximallyto benefit the lowest economic positionin the present

  • 200 Philosophy & Public Affairs

and future generations. Focusing on one such (as he says) instrument


fromhaving to considerthe differenceprinciple on everyquestion



  • I have proposeda similarinstrument to controlinternational inequality:

a global resources tax, or GRT.5The basic idea is that, while each people

owns and fully controlsall resources within its national territory,6it must pay a tax on any resources it chooses to extract. The Saudi people, for example, would not be requiredto extractcrude oil or to allow others to do so. But if they chose to do so nonetheless, they would be requiredto pay a proportionaltax on any crude extracted,whether it be for their own use or for sale abroad.This tax could be extended, along the same lines, to reus- able resources:to land used in agricultureand ranching,for example, and, especially,to air and water used for the dischargingof pollutants. The burdens of the GRTwould not be borne by the owners of resources alone. The tax would lead to higher prices for crude oil, minerals, and so

forth. Therefore, some Japanese (who have no

of the GRT on oil would ultimately fall upon the oil of their own, but importa good bit), even while

the tax would be actually paid by the peoples who own oil reserves and choose to extractthem. This point significantlymitigates the concern that the GRTproposal might be arbitrarilybiased against some rich peoples, the resource-rich,and in favorof others. This concern is furthermitigated by the GRT'spollution component. The GRTis then a tax on consumption. But it taxes different kinds of consumptiondifferentially. The cost of gasolinewill containa much higher portionof GRTthan the cost of a ticket to an art museum. The tax falls on goods and services roughly in proportionto their resource content: in proportionto how much value each takes from our planet. The GRTcan thereforebe motivatednot only forwardlookingly,in consequentialistand contractualistterms, but also backwardlookingly:as a provisoon unilateral appropriation,which requires compensation to those excluded thereby. Nations (or persons) may appropriateand use resources, but humankind at large still retains a kind of minority stake, which, somewhat like pre-

  • 4. John Rawls,"Justice as Fairness:Revisited, Revised, Recast," 1992,

typescript,p. 136.

  • 5. RR, pp. 256nI 8, 264f. See also "AnInstitutional Approach to HumanitarianInterven-

tion," Public Affairs Quarterly 6 (1992):89-1o3,

p. 96.

  • 6. This accommodatesRawls's remark that "unlessa definiteagent is given responsibility

formaintaining an asset and bearsthe loss of not doing so, thatasset tends to deteriorate"(LP,

P. 57).

  • 201 An Egalitarian Law of Peoples

ferredstock, confers no controlbut a share of the materialbenefits. In this picture, my proposal can be presented as a global resources dividend,

which operates as a provisoby giving up

modern Lockeanproviso. It differs from Locke'sown the vague and unwieldy7condition of "leavingenough

and as good for others":One may use unlimited amounts, but one must

share some of the economic benefit. It is nevertheless similar enough to

the originalso that even such

notoriouslyantiegalitarian thinkers as Locke

and Nozick might find it plausible.8 National governmentswould be responsible for paying the GRT,and, with each societyfree to raise the requisitefunds in any way it likes, no new

administrativecapabilities would need to be developed. Since extraction and pollutionactivities are relativelyeasy to quantify,the assurance prob-


would be manageable and total collection costs could be negligible.

Proceeds from the GRTare to be used towardthe emancipationof the

present and future global poor: towardassuring that all have access to education, health care, means of production(land) and/or jobs to a suffi-

cient extent to be

able to meet their own basic needs with dignity and to

represent their rights and interests effectivelyagainst the rest of human-

kind: compatriotsand foreigners.In an ideal worldof reasonablyjust and

well-orderedsocieties, GRTpayments could

be made directly to the gov-

ernments of the poorest societies, based on their per capita income (con-

vertedthrough purchasing power parities) and populationsize. These data arereadily available and easy to monitor-reliable and comprehensivedata are currentlybeing collected by the United Nations, the WorldBank, the IMF, and variousother organizations.9 GRTpayments would enable the governmentsof the poorerpeoples to maintain lower tax rates, higher tax exemptions and/or higher domestic spending for education, health care, microloans,infrastructure, etc. than

  • 7. Consider:Must we leaveenough and as goodfor future generations? For how many? Are

air and waterpollution ruled out entirelybecause the airand waterleft behindis not as good?

  • 8. Cf. John Locke,Second Treatise, ? ?27,

(New York:Basic Books I974),

pp. 175-77

33; RobertNozick, Anarchy,State, and Utopia and Chapter4.

  • 9. One may think thatdomestic income distributionshould be

taken into accountas well.

Even if two states have the same per capitaincome, the poorin the one may still be much worseoff than the poorin the other.The problemwith takingaccount of this factis thatit may providea perverseincentive to governmentsto neglect theirdomestic poor in orderto receive

largerGRT payments. This incentive is bad, because governmentsmight act on it, and also

because governmentsmight, wrongly,be thoughtto act or be accused of actingon it ance and assuranceproblems).


  • 202 Philosophy & Public Affairs

would otherwise be whole GRT scheme

possible. Insofar as they would actually do this, the would require no central bureaucracyand certainly

nothing like a world government,as governmentswould simply transfer the GRTamounts to one another through some facilitatingorganization,

such as the WorldBank, perhaps, or the development aid are: Payments would

UN. The differences to traditional

be a

matter of entitlement rather

than charity and-there being no matching of "donors"and recipients

would not be conditionalupon renderingpolitical or economic favorsto a


or upon adopting a donor'sfavored political or economic institu-

tions.'0Acceptance of GRTpayments would of course be voluntary:A just

society may certainlyshun greateraffluence if it, democratically,chooses to do so.

In a

nonideal worldlike ours, corruptgovernments in the poorerstates

pose a

significantproblem. Such governmentsmay be inclined, for exam-

ple, to use GRTfunds to underwriteindispensable services while diverting


domestic tax revenue saved to the rulers'personal use. A government

that behavesin this way may be cut off from GRTfunds." In such cases it

may still be possible through existing UN

to administer meaningful development programs agencies (WorldFood Program,WHO, UNICEF,

etc.) or through suitablenongovernmental organizations (Oxfam). If GRT funds cannot be used effectively to improvethe position of the poor in a particular country, then there is no reason to spend them there. They should ratherbe spent where they can make more of a differencein reduc- ing povertyand disadvantage. There are then three possibilitieswith regardto any countrythat is poor enough in aggregate to be eligible for GRTfunds: Its poorercitizens may

benefit through their government, they may benefit

from development

programsrun by some other agency, or they may not benefit at all. Mix- tures are, of course, also possible. (A countrymight receive 6o percent of the GRTfunds it is eligible for, one third of this through the government

and two thirds of

it through other channels.) How are these matters to be

decided? And by whom? The decisions are to be made by the facilitating

io. Fora detailedaccount of how the latterfeature renders current aid highly inefficient, if not useless, see the coverstory "Why Aid Is an Empty Promise,"The Economist 331/7862

(May 7, 1994),






Fora contraryconception, see BrianBarry, "Humanity and Justice in GlobalPerspec-

tive," in NOMOS XXIV, Ethics, Economics, and the Law, ed. J. R. Pennock and J. W Chap-

man (New York:New YorkUniv. Press, I982),



Barryholds that the governments

of poorsocieties should receive funds regardlessof theirdomestic policies.

  • 203 An Egalitarian Law of Peoples

organization, but pursuant to clear and straightforwardgeneral rules. These rules are to be designed, and possibly revised, by an international groupof economistsand internationallawyers. Its taskis to devise the rules so that the entire GRTscheme has the maximum possible positiveimpact on the world'spoorest persons-the poorestquintile, say-in the long run.

The qualification"in the long run"indicates that incentive effects must be taken into account. Governmentsand also the wealthierstrata of a people stand to gain from GRTspending in variousways ("trickle-up")and there- fore have an incentive to ensure that GRTfunds are not cut off. The rules should be designed to take advantageof this incentive. They must make it clear to members of the politicaland economic elite of GRT-eligiblecoun- tries that, if they want theirsociety to receive GRTfunds, they must cooper- ate in making these funds effective towardenhancing the opportunities and the standardof living of the domestic poor.'2

Specifying how GRTfunds should best

be raised poses some complex

problems,among them the followingfour: First, setting tax rates too high may significantlydampen economic activity-in extreme cases so much that revenues overallwould decline. It must be noted, however, that the

funds raised throughthe and for the benefit of, the

GRTscheme do not disappear:They are spent by, globalpoor and therebygenerate effective market

demand that spurs economic activity.Second, imposing any GRTon land use for cultivationof basic commodities(grains, beans, cotton, etc.) might increase theirprices and therebyhave a deleteriouseffect on the positionof the globallyworst-off. Hence it may make sense to confine any GRT on land to land used in other ways (e. g. to raise cattle or to grow tobacco, coffee, cocoa, orflowers). Third, the setting of tax rates shouldalso takeinto

account the interests of the future globally worst-off. The GRT should target the extractionof nonrenewableresources liable to run out within a few decades in preference to that of resources of which we have an abun- dant supply;it should target the dischargingof pollutantsthat will persist for centuriesin preferenceto the dischargingof pollutantsthat decay more quickly.Finally, while designing the GRTis inevitablydifficult and compli- cated, the tax itself should be easy to understandand to apply.It should, for example, be based on resources and pollutantswhose extraction or dis-

  • 12. In some GRT-eligiblecountries there may well be factionsof the ruling elite for whom these incentives would be outweighedby their interest in keeping the poor uneducated, impotent,and dependent. Still, the incentiveswill shift the balanceof forcesin the directionof reform.

  • 204 Philosophy & Public Affairs

charge is reasonablyeasy to monitor or estimate, in order to ensure that everypeople is payingits fair share and also to assure everypeople that this

is so.

The general point behind these brief remarks is that GRT liabilities should be targetedso as to optimizetheir collateral effects. Whatis perhaps surprisingis that these effects may on the whole be positive,on account of the GRT'sconsiderable benefits for environmentalprotection and conser- vation.These benefits are hardto secure in a less concertedway because of familiarcollective-action problems ("tragedy of the commons"). What about the overallmagnitude of the GRT?In light of today'svast globalsocial and economic inequalities,one may think that a massive GRT scheme would be necessary to supportglobal background justice. But I do

not think this is so. Currentinequalities are the cumulative result of


cades and centuries in which the more-developedpeoples used their ad- vantages in capital and knowledge to expand these advantagesever fur- ther. They show the power of long-term compounding rather than overwhelminglypowerful centrifugal tendencies of our globalmarket sys- tem. Even a rather small GRT may then be sufficient continuously to

balance these ordinarycentrifugal tendencies of market systems enough to prevent the developmentof excessive inequalities and to maintain in equilibriuma rough globaldistributional profile that preserves global back- groundjustice. I cannot here workthrough all the complexitiesinvolved in determining the appropriatemagnitude of the GRTscheme. To achieve some concrete- ness nevertheless, let us, somewhat arbitrarily,settle for a GRTof up to i percent of worldproduct-less than i percent if a smaller amount would betteradvance the interests of the globallyworst-off in the long run. Almost any egalitarianconception of globaljustice would probablyrecognize this

proposalas an improvementover the status quo. A i percent GRTwould currentlyraise revenues of roughly $270 billionper annum. This amount is quite large relativeto the total income of the world'spoorest one billion

persons and, if

well targetedand effectivelyspent, wouldmake a phenome-

nal difference to them even within a few years. On the other hand, the amount is rather small for the rest of us: not only less than the annual defense budget of the U.S. alone, but also a good bit less than the market

price of the currentannual crude oil production,which is in the neighbor- hood of $400 billion(ca. 6o million barrelsper day at about$i8 per barrel). Thus the entire revenue targetcould be raisedby taxing a small numberof

  • 205 An Egalitarian Law of Peoples

resource uses-ones

whose discouragement seems especially desirable

for the sake of future generations.A $2-per-barrelGRT on crude oil extrac- tion, for example, would raise about one sixth of the overall revenue


increasing the price of petroleum products by about 5? a

gallon. It would have some substitutioneffects, welcome in terms of con-

servation and environmentalprotection; and, if it had any dampening effect on overalleconomic activityat all, this effect would be quite slight.

Having tried to show that introducing a I percent GRT would be an instantlyfeasible and morallyattractive institutional reform of the existing globalorder,'3 let me now focus on its plausibilityas a piece of ideal theory.

To do this, we append my GRTproposal to Rawls'slaw of


resulting alternativeto Rawlsis not my consideredposition on globaljus- tice. Its point is rather to allow us to focus sharplyon the topic of interna- tionalinequality. Egalitarian concerns will be vindicated,if it can be shown that the amended law of peoples is morallymore plausible than Rawls's original-and especially so, if this can be shown on Rawlsiangrounds.



In his initial sketch, Rawls's brief discussion of international justice was characterized by a tension between three views:

  • (i) He speaks of the second session of the original position as featuring

persons from the various societies who make a rational choice of principles so as best to protect their interests while "they know nothing about the

particular circumstances of their own society, its power and strength in

  • 13. This is a bit of an exaggeration:I have not yet given you any reasonnot to dismiss my

GRTproposal as unfeasiblein the politicalsense. This I hope to do in the final section.

  • 14. Rawls characterizesthis law of peoples by the followinglist of principles(LP, p. 55):

"(i) Peoples(as organizedby their government)are free and independentand


and independenceis to be respected by other peoples. (2) Peoplesare equal and parties to

their own agreements. (3) Peoples have the right of self-defense but

no right to war. (4)

Peoples are to observe a duty of nonintervention.(5) Peoples are to observe treaties and undertakings.(6) Peoplesare to observecertain specified restrictions on the conduct of war (assumed to be in self-defense).(7) Peoplesare to honorhuman rights. " Though this list is not meant to be complete (ibid.), the completelist would not contain an egalitariandistributive principle(LP, pp. 7sf). ThroughoutRawls makes no attemptto show that representativesof peoples would, in his second session of the originalposition, adopt these principles.The presentationis farless rigorousthan the one he had offeredin supportof his two principlesof domestic justice. My response in this essay is then not so much a critique of Rawls as a detailedand, I hope, constructiveinvitation to defend his conclusions.

  • 206 Philosophy & Public Affairs

comparisonwith other nations, nor society"(TJ, p. 378, my emphasis).I

do they know their place in their own called this readingRI (RR, pp. 242ff).


On the same page, Rawls also speaks of this second session as

featuring "representativesof protect their interests" (TJ,

states [who are] to make a rationalchoice to p. 378, my emphasis). Here, "the national

interest of a just state is defined by the principles of justice that have

alreadybeen acknowledged.Therefore such a nation will aim above all to maintain and to preserveits just institutionsand the conditionsthat make them possible"(TJ, p. 379). I called this reading R2 (RR, pp. 243ff). (3) Rawls also wanted to endorse the traditional(pre-World War II) principles of internationallaw as outlinedby James Brierly.

I have tried to show (RR, ?21) that no two of these views are compatible. Rawls has now fully resolved the tension by clearly and consistently endorsingthe second view,R2-without, however,offering any reasons for favoringit over RI. He stipulates that the parties "arerepresentatives of

peoples"(LP, p.

48) and "subjectto a veil of ignorance:They do not know,

for example, the size of the territory,or the population, or the relative

strength of the people whose fundamental interests they represent. Al- though they know that reasonablyfavorable conditions obtain that make democracypossible, they do not know the extent of theirnatural resources, or level of their economic development,or other such relatedinformation"

(LP, p. 54). And what are those fundamentalinterests of a people? As in his initial

account of R2,Rawls takes each people to have only one such fundamental interest: that its domestic institutionssatisfy its conception of justice (LP,

pp. 54,

64). And while the partiesto the first session of

the originalposition

do not know the particularconceptions of the good of the persons they represent, Rawls assumes, without justifying the disanalogy,that each partyto the second session does knowwhat conceptionof domesticjustice "her"people subscribes to. It would seem that the variousdelegates would then favor different versions of the law of peoples, each one especially hospitable to a particularconception of domestic justice.'5 Rawls claims, however,that within a certainrange of conceptionsof domesticjustice the

i5. I say that any global(session of the) originalposition features delegatesrather than partiesor representatives. This is my expression,not Rawls's.Its sole purposeis to makemore perspicuous that the reference is to deliberatorsabout global rather than domestic institutions.

  • 207 An Egalitarian Law of Peoples

interests of peoples regarding the law of peoples coincide: Delegates of peoples whose conceptionof domesticjustice is either liberalor hierarchi- cal would all favorexactly the same law of peoples (LP, p. 6o). This is then the law of peoples that we, as membersof a societywith a liberalconception of justice, should endorse: It is hospitable to liberal regimes and to the more palatablenonliberal regimes as well. The regimes it does not accom- modateare "outlaw regimes" of varioussorts or else committedto an expan- sionist foreign policy (LP, pp. 72f). Rawls'slaw of peoples cannot be justi-


to them as being (behind the veil of ignorance)in theirinterest as well.

But this fact cannot count against it from our point of view-which, after all, is the one to which we seek to give systematic expression.

Given this structure of his account, Rawls decides to run the second session twice: Once to show that delegates of peoples with any liberal conception of domestic justice would favorhis law of peoples and then to show that delegates of peoples with any hierarchicalconception of domes-

tic justice woulddo so as well.'6He does not actuallyperform either of these two runs in any detail, and I am quite unclear as to how the second is

supposed to go.

In the next two sections, I shall focus exclusively on the liberalrun, in

which "the parties deliberate among availableprinciples for the law


peoples by referenceto the fundamentalinterests of democraticsocieties in accordancewith, or as presupposedby, the liberalprinciples of domestic justice" (LP, p. 54). A liberalconception of justice is defined (LP, p. 5I) as one that


that certainrights, liberties,and opportunitiesbe secure for

all citizens


this demand a high priorityvis-'a-vis other values and interests,



that all citizens should have adequatemeans to take advan-

tage of their rights, freedoms, and opportunities.

Liberalconceptions ofjustice may differfromRawls's by being comprehen- sive ratherthan political,for example, or by lacking some or all of the three egalitariancomponents he incorporates.

  • i 6. See LP,pp. 52 and 6o for Rawls'sdistinction of these two steps, as account of internationalideal theory.

he calls them, of his

  • 208 Philosophy & Public Affairs

Rawls makes each delegate assume that her people is interested exclu- sively in being constituted as a just liberal society and he asserts that delegates with this sole interestwould adopt his law of peoples, which lacks any egalitariancomponent. I will now argueagainst his stipulationthat the delegates have only this one interest (Section III) and then against his claim that delegates with this sole interest would adopthis law of peoples

(Section IV).If only one of my two argumentssucceeds, in trouble.

Rawls'saccount is


An obvious alternativeto the stipulationis this: Each delegate assumes that her people has an ultimateinterest not only in the j ustice of its domes- tic institutions, but also in the well-being of its members (beyond the minimum necessary for just domestic institutions). Each delegate as- sumes, that is, that her people would, other things equal, prefer to have a higher rather than a lower averagestandard of living.'7 Delegates so describedwould favorthe GRTamendment. This is clearly true if, like the partiesto the domestic session, they deliberateaccording to

the maximin rule. But it is also true if they focus on average expectations:

The GRTamendment would benefit all peoples by reducing pollutionand

environmental degradation. It is unclear whether it would have a positive or negative effect on per capita income for the world at large. But it would keep national per capita incomes closer together and thereby, given the decreasing marginal significance of income for well-being, raise the aver- age standard of living as anticipated in the original position. An increase in national per capita income at the bottom matters more than an equal decrease in national per capita income at the top-in terms of a people's ability to structure its social world and national territoryin accordance with its collective values and preferences, for example, and also in terms of its members' quality of life. We need not stipulate that a people's interest in well-being is strong

relativeto its interest in domesticjustice. For suppose we have each dele- gate assume that her people's interest in well-being is very slight and

subordinate to the interest in domestic justice. Then the delegate will care

  • 17. Rawlsmakes the analogousstipulation for the domesticsession of the originalposition,

remarkingthat it cannothurt a personto have greatermeans at her disposal:One can always

give them awayor forego their use

(TJ, pp. 142f).

  • 209 An Egalitarian Law of Peoples

relativelylittle aboutwhat her people would gain through the amendment in case it would otherwisebe poorer.But then she will also care little about what it would lose through the amendment in case it would otherwise be more affluent. She would care little both ways, and thereforewould still

have reason to

adopt the amendment if, as I have argued, the gains out-

weigh the losses. I conclude that, if a delegate assumes her people to have an interest in well-being, and be it ever so slight, then she will favormy amendment- regardlessof whether she seeks to maximize her people'saverage or worst-

case expectations. Rawlsmust thereforeposit the opposite:Each delegate assumes that her people has no interest at all in its standard of living (beyond its interest in the minimum necessary for just domestic institu- tions). This is, of course, precisely what his stipulationentails. But why should we find this stipulationplausible once we see what it excludes? There are severalreasons to find it implausible.There are, for one thing, variantsof liberalismthat-unlike Rawls'sown-are committedto contin- ued economic growthand progress;and a people committedto one of them should be presumed to want to avoid economic stagnation and decline. There are also cosmopolitanvariants of liberalismwhich extend the egali- tarian concerns that Rawls confines to the domestic case to all human beings worldwide;and a people committedto one of them should be pre- sumed to want to avoidrelative deprivation for itself as well as for others. The stipulationalso has implausibleside effects. In explicating the out- come of the liberalrun of his second session, Rawls writes: "Thereshould be certain provisionsfor mutual assistance between peoples in times of famine and drought and, were it feasible, as it should be, provisionsfor ensuring that in all reasonablydeveloped liberal societies people's basic needs are met" (LP, p. 56). Does he really mean what this sentence sug- gests: that provisionsare called for to meet basic needs only in reasonably developed societies? His account may well leave him no other choice. In his second session, each delegate cares solely abouther people'sachieving domestic justice. However,helping a people meet their basic needs may not enable them to achieve domestic justice, if their society is still quite

undeveloped.Hence aid to members of such societies is not

a requirement

of globaljustice on Rawls'sstipulation. His law

of peoples requires basic

food aid, say, only to peoples who but for their povertywould be able to maintainjust domestic institutions.

Now it would be outrageousto suggest that Rawls deems it a matter of

  • 210 Philosophy & Public Affairs

moralindifference whether members of undevelopedsocieties are starving

or not. But, given his stipulation,he would have to say that such aid is

an ethical duty, which we might discharge individually,or


through our government. Internationaljustice requires institutions de- signed to meet basic needs in societies where this contributesto domestic justice, but not in societies where it does not. Yetthis looks counterintui- tive:Why, after all, do liberalswant the law of peoples to be supportiveof the internal justice of all societies, if not for the sake of the persons living in them? Andif our concernfor the domesticjustice of societies is ultimatelya concern for their individualmembers, then why should we focus so nar- rowly on how well a law of peoples accommodatestheir interest in living

under just domestic institutionsand not also, more broadly,on how well it accommodatestheir underlyingand indisputableinterest in secure access

to food, clothing, shelter,education, and health care, even where a reason- ably developedliberal society is still out of the question? The danger here is not merely moralimplausibility, but also philosophi- cal incoherence between Rawls'sconceptions of domestic and of global justice. Accordingto the latter, a just domestic regime is an end in itself. According to the former, however, it is not an end in itself, but rather something we ought to realize for the sake of individualhuman persons,

who are the upholdjust

ultimateunits of moralconcern. Ournatural duty to create and domesticinstitutions is a dutyowed to them (TJ, p. I I 5). Their

well-beingis the point of socialinstitutions and therefore,through the first session of the original position, gives content to Rawls's conception of domestic justice. The incoherence might be displayedas follows. Suppose the parties to the first, domestic session knew that the persons they represent are the members of one society among a pluralityof interdependentsocieties; and suppose they also knew that a delegate will represent this society in a subsequent international session, in which a law of peoples is to be adopted.How would they describe to this delegate the fundamentalinter-

ests of their society? Of course they would want her to push for a law of

peoples that is supportiveof the two principlesof justice which,

kind of nationalinstitutions favored by the accordingto Rawls, they have adoptedfor

the domestic case. But their concern for such domestic institutions is derivativeon their concern for the higher-orderinterests of the individual human persons they themselves represent in the domestic originalposi- tion. Therefore,they wouldwant the delegate to push for the law of peoples

  • 211 An Egalitarian Law of Peoples

that best accommodates, on the whole, those higher-orderinterests of individuals."8They would want her to consider not only how alternative proposalsfor a law of peoples would affect their clients' prospects to live underjust domesticinstitutions, but also how these proposalswould affect

their clients' life prospectsin other ways-for example through the


ence of their society. This point, by the way, strongly suggests that those committed to a Rawlsian(or, indeed, any other liberal) conception of do-

mestic justice should want the delegates to any

globaloriginal position to

be conceived as representativesof persons rather than peoples. I suspect that Rawlswants his second session of the originalposition to be informed by the interests of peoples, conceived as irreducible to the interests of persons, because the latter would inject an individualisticele- ment that he deems unacceptableto hierarchicalsocieties. The problem he sees is real enough, but his solutionaccommodates the hierarchicalsat the expense of not being able to accommodatethe liberals.I will return to

this point in Section V.



The foregoing arguments notwithstanding, let us now allow the stipula- tion. Let us assume that each people really does have only the one interest in the justice of its own domestic institutions, and that its delegate to the second session of the original position is instructed accordingly. Would such delegates prefer Rawls's law of peoples over my more egalitarian

alternative? The answer, clearly, is NO: They

would at most be indifferent

between the two proposals. I don't know why Rawls thinks otherwise. But he may have been misled by an unrecognized presumption that a laissez- faire global economic order is the natural or neutral benchmark which the delegates would endorse unless they have definite reasons to depart from it. This presumption would explain his discussion of a global difference principle, which is peculiar in two respects. First, Rawls considers such a principle only in regard to one part of nonideal theory: coping with unfavor- able conditions (LP, p. 75), although it has generally, if not always, been proposed as an analogue to the domestic difference principle, which is

I8. ForRawls's account of the threehigher-order interests of the personswhom the parties to the first, domestic session represent,see his PoliticalLiberalism (New York:Columbia

Univ. Press I993),

pp. 74f, io6.

  • 212 Philosophy & Public Affairs

used primarilyto design the ideal basic structure.'9Second, the tenorof his remarksthroughout is that a globaldifference principle is too strongfor the internationalcase, that it demands too much from hierarchicalsocieties (e.g., LP, p. 75). This suggests a view of the difference principle as a principle of redistribution,which takes from some to give to others: The more it redistributes,the more demandingis the principle.But this view of the difference principle loses an insight that is crucial to understanding Rawls'sown, domestic differenceprinciple: There is no priordistribution, no natural baseline or neutral way of arrangingthe economy, relative to which the difference principlecould be seen to make redistributivemod- ifications. Rather,there are countless ways of designing economic institu- tions, none initially privileged,of which one and only one will be imple- mented. The difference principle selects the scheme that ought to be chosen. The selected economic groundrules, whatever their content, do not redistribute,but rather govern how economic benefits and burdens get distributedin the first place. This point is crucialfor Rawls'sreply to Nozick'scritique. Nozick wants to make it appearthat laissez-faireinstitutions are naturaland define the baseline distributionwhich Rawls then seeks to revise ex post through redistributivetransfers. Nozick views the first option as natural and the second as making great demands upon the diligent and the gifted. He allows that, with unanimous consent, people can make the switch to the second scheme; but, if some object, we must stick to the first.20Rawls can respond that a libertarianbasic structure and his own more egalitarian liberal-democraticalternative are options on the same footing: the second

is, in a sense, demanding on


gifted, if they would do better under the

first-but then the first is, in the same sense and symmetrically,demand- ing on the less gifted, who would do much better under the second

scheme. In his discussion of the globaldifference principle, Rawls's presentation of the issue is the analogue to Nozick'sin the domestic case. It is somehow natural or neutral to arrangethe world economy so that each society has absolute control over, and unlimited ownership of, all natural resources within its territory.Any departuresfrom this baseline, such as my GRT

proposal,are demanding and, it turns out, too demanding on some


Ig. In the penultimatedraft of "The Law of Peoples,"Rawls did argue also against the globaldifference principle as a proposalfor ideal theory,but he has deletedthose arguments.


See Anarchy, State, and Utopia, pp. I67-74,



  • 213 An Egalitarian Law of Peoples

eties. I want to give the analogue to

the Rawlsiandomestic response: Yes,

egalitarianinstitutions are demanding upon naturallyand historicallyfa- voredsocieties, as they would do betterin a scheme with unlimitedowner- ship rights. But then, symmetrically,a scheme with unlimited ownership rights is at least equally demanding upon naturallyand historicallydis- favored societies, since they and their members would do much better under a more egalitarianglobal basic structure.

  • I have arguedthat Rawlshas given no reasonwhy the delegates-even if

each of them cares solely about her people'sprospects to live under just domestic institutions-should prefer his inegalitarianlaw of peoples over more egalitarianalternatives. Might they have a reason for the opposite preference? I believe that they do. In a world with large international inequalities,the domesticinstitutions of the poorersocieties arevulnerable to being corruptedby powerful political and economic interests abroad. This is something we see all around us: politicians and business people from the rich nations self-servinglymanipulating and interferingwith the internalpolitical, judicial, and economicprocesses of third-worldsocieties. Rawls is presumablyaware of this phenomenon, but he fails to see its roots in gross internationalinequality: In poorersocieties, he writes, "the problem is commonly the nature of the public political culture and the religiousand philosophicaltraditions that underlieits institutions.The great social evils in poorersocieties are likely to be oppressivegovernment

and corruptelites" (LP, p. 77). Now Rawls is surely right that many poor

countrieshave corruptinstitutions and ruling elites, which

do not serve the

interests of the people and contributeto their poverty.But the inverse is

certainly true as well: Relativepoverty breeds corruptibilityand corrup- tion. Powerfulforeign governments support their favorite faction of the local elite and often manage to keep orinstall it in power-through financialand organizationalhelp for winning elections, if possible, or through support

for security forces,

coups d'etat, or "revolutions"otherwise. Third-world

politiciansare bribedor pressuredby firms from the rich societies to cater to their sex tourismbusiness, to accept their hazardouswastes and indus- trial facilities, and to buy useless products at governmentexpense. Agri- businesses, promising foreign exchange earnings that can be used for luxury imports, manage to get land use converted from staple foods to export crops: Wealthyforeigners get coffee and flowersyear-round, while many locals cannot affordthe higher prices for basic foodstuffs.Examples

could be multiplied;but I think it is indisputablethat the oppressionand

  • 214 Philosophy & Public Affairs

corruptionin the poorercountries, which Rawlsrightly deplores,is by no means entirely homegrown. So it is true, but not the whole truth, that governmentsand institutionsof poorcountries are often corrupt:They are activelybeing corrupted,continually and very significantly,by privateand official agents fromvastly more wealthy societies. It is entirely unrealistic to expect that such foreign-sponsoredcorruption can be eradicatedwith- out reducing the enormous differentialsin per capita GNP. So long as the delegates to Rawls'ssecond session are merely presumed to know that large internationalinequalities may have a negative impact upon the domesticjustice of the poorersocieties, they have a tie-breaking reason to favora more egalitarianlaw of peoples over Rawls's.2"


If only one of my two arguments is sound, then delegates of liberal societies

would prefer a more egalitarian law of peoples over Rawls's inegalitarian

alternative. I suppose Rawls would regret this fact, if it destroys the desired

coincidence between the law of peoples adopted by delegates of liberal societies (at step I of his second session) and the law of peoples adoptedby

delegates of hierarchical societies (at step 2). But this coincidence fails, in

any case, on account of human rights. Rawls claims that both sets of delegates would adoptprecisely the same

law of peoples (LP, p. 6o), which includes a list of human rights (LP,

pp. 62f, 68, 70) featuringminimum rights to life (means of subsistence and

security), to liberty (freedom from slavery, serfdom, and forced occupa-

tions), to personal property, to "a measure" (LP, p. 63) of liberty of con-

science and freedom of thought and "a certain" (LP, p. 68) freedom of

association (compatible with an established religion), to emigration, and to

the rule of law and formal equality as expressed by the rules of natural

justice (for example that similar cases be treated similarly). He gives no

reason, and I can see none, historicalor philosophical,for believing that


While the delegates to an R2-typesecond session would view my modificationas an

improvement,they wouldpresumably like even betterthe additionof

a more statistegalitar-

ian component, such as Brian Barry'sproposal cited in note i I. This does not worryme, because it was only for the sake of the argumentthat I have here acceptedRawls's R2 set-up, which treatspeoples, not persons, as ultimateunits of moralconcem. I am confidentthat a more plausible construalof a global originalposition-G, for example, as defended in RR ? ?22-23-would supportsomething very much like the GRTas an essential partof a fully just law of peoples.

  • 215 An Egalitarian Law of Peoples

hierarchicalsocieties, as such, wouldincorporate these human rights into their favoredlaw of peoples. Perhapsmany such societies can honor these rights while retaining their hierarchical,nonliberal character,as Rawls suggests (LP, p. 70); but this hardlyshows that they would choose to be bound by them. Human rights are not essential to hierarchicalsocieties, as they are essential to liberal ones. Not only is it highly doubtful that delegates of hierarchical societies

would choose to commit themselves to so much; it is also quite unclear

why delegates of

liberalsocieties would not want to

incorporatemore than

Rawls'slist, which specifically excludes freedom of speech (LP, p. 62),

democraticpolitical rights (LP,pp. 62, 69f), and equalliberty of conscience and freedom of thought (LP, pp. 63, 65).

Rawls'squest for a "politicallyneutral" (LP, p. 69) law of peoples-one that liberals and hierarchicalswould independentlyfavor on the basis of their respective values and interests-thus holds little promise: Those

who are really committed to

a liberalconception of justice will envision a

law of peoples which demands that persons everywhereenjoy the protec- tion of the full list of human rights as well as adequate opportunitiesand

materialmeans that are not radicallyunequal. The

friends of hierarchical

societies will prefera worldorder that is

much less protectiveof the basic

interests of persons as individuals.The formerwill want the interests of persons to be representedin the second session of the original position. The latter will care only about the interests of peoples. Occasionally,Rawls suggests a different picture, which jettisons the claim to politicalneutrality. On this picture, the law of peoples he proposes is not what liberals would ideally want, but ratheris affected by the exis- tence of hierarchicalsocieties. The allegedcoincidence of the results of the two runs of the second session is then not luck, but design. It comes about because good liberals seek to accommodatehierarchical societies by ad-

justing theirideal of globaljustice so as to "expressliberalism's own princi-

ple of toleration for other reasonable ways of ordering society" (LP, p. 43).22

Just as Rawls himself may be expressing this desire by conceiving the second session of the originalposition in nonindividualisticterms, he may conceive of his liberaldelegates as having a similardesire to adopta law of peoples acceptable to hierarchical societies. This could explain their-

  • 22. Another,related reason might be, as Rawls remarksin another context, that "all principlesand standardsproposed for the law of peoplesmust, to be feasible,prove acceptable to the consideredand reflectivepublic opinion of peoplesand their governments"(LP, p. 50).


Philosophy & Public Affairs

otherwise incredible-decisions against certain human rights (precisely those most offensive to the hierarchicals) and against any egalitarian principle. This picture is not at all that of a negotiated compromisein which the liberal delegates agree to surrendertheir egalitarianconcerns and some human rights in exchange for the hierarchicaldelegates accepting the


a bargainingmodel is quite un-Rawlsianand also does

not fit with his account, on which the two groupsof delegates deliberatein

mutual isolation.The tolerationmodel is more noble than

this: The liberal

delegates, informedthat their societies share a worldwith many hierarchi- cal societies, seek to design a law of peoples that hierarchicalsocieties, on

the basis of their values and interests as such, can reasonablyaccept. Yet, for all its nobility,the tolerationmodel has a drawbackthat the bargaining model avoids:It is ratherone-sided. The hierarchicals,unencumbered by

any principle of

toleration,get their favoritelaw of peoples, while the lib-

erals, "toexpress liberalism'sown principleof toleration,"surrender their egalitarian concerns and some important human rights.23This fits the witty definition of a liberal as someone who will not take her own side in

any disagreement. What goes wrong here is that Rawls, insofar as he is committed to this picture, does not clearly distinguish two views, and hence is prone to accept the second with the first:

  • (i) Liberalisminvolves a commitment to tolerance and diversity that

extends beyondthe familyof liberalconceptions: A liberalworld order will thereforeleave room for certain kinds of nonliberalnational regimes.


Liberalisminvolves a commitment to tolerance and diversity that

extends beyondthe family of liberalconceptions: It would thus be illiberal to impose a liberalglobal order on a worldthat contains many peoples who do not share our liberalvalues.

By acknowledging(i),

we are not compromisingour liberal convictions. To

the contrary:We would be compromisingour liberalconvictions if we did

not envision a liberal worldorder in this way. A world orderwould not be genuinely liberal if it did not leave room for certain nonliberal national

regimes. Those who acknowledge(2),

by contrast,are compromisingtheir

  • 23. And probablysome additionalhuman rights as well, if I was right to argue that the

hierarchicaldelegates would not adopteven the truncatedlist thatRawls incorporates into his law of peoples.

  • 217 An Egalitarian Law of Peoples

liberal convictionsfor the sake of accommodatingthose who do not share

them. Liberalsshould then accept (i)

and reject (2).

This reasoning is the analogue to what Rawls himself would say about the domestic case. Consider:


A liberal society must leave room

for certain nonliberal commu-

nities and lifestyles.


It would be illiberalto impose liberalinstitutions on a society that

contains many persons who do not share our liberalvalues.

Rawlswould clearlyaccept (I') and reject (2'). He could give the following

rationalefor this: While our society can contain many different kinds


communities, associations, and conceptions of the good, some liberal in

characterand others not, it can be structuredor organizedin only one way.

If my neighborwants to be

a Catholicand I an atheist, we can bothhave our

way, can both lead the life each deems best. But if my neighbor wants the U.S. to be organizedlike the CatholicChurch and I want it to be a liberal state, we can not both have our way.There is no room for accommodation

here, and, if I

reallybelieve in egalitarianliberal principles, I should politi-

cally supportthem and the institutionsthey favoragainst their opponents. These institutions will not vary with the shifting political strength of groups advocatingvarious religious, moral,or philosophicaldoctrines.24 My rationaleis the analogue to this: While the world can contain soci-

eties that are structuredin a varietyof ways, some liberaland some not, it cannot itself be structuredin a varietyof ways. If the Algerianswant their

society to be organized as a religious state consistent with a just global

orderand we want ours to be a liberaldemocracy, we can

bothhave ourway.

But if the Algerianswant the worldto be organizedaccording to the Koran, and we want it to accordwith liberalprinciples, then we can not both have our way.There is no roomfor accommodationhere, and, if we reallybelieve in egalitarianliberal principles-in everyperson's equal claim to freedom and dignity-then we should politicallysupport these principles,and the global institutions they favor,against their opponents. These institutions

  • 24. In supportinghis view that our conceptionof justice shouldnot, in the mannerof (2'), be sensitive to what competingviews happen to be prevalentamong our compatriots,Rawls also stresses that such sensitivitywould render this conception"political in the wrong way,"

thus leadingto some of

the problemsassociated with institutionsthat reflect a modusvivendi.

(See PoliticalLiberalism, Lecture IV, esp. pp. I4 I-48.) the globalplane (see RR, Chapter5).

This concern,too, has an analogueon


Philosophy & Public Affairs

will not vary with the shifting political strength of states committed to various conceptions of domesticjustice.

  • I conclude that Rawls has failed to show that the law of peoples liberals would favorand the law of peoples favoredby hierarchicalseither coincide

by sheer luck or can be

made to coincide by morallyplausible design. We

should then work towarda global order that-though tolerantof certain nonliberalregimes, just as a liberalsociety is tolerantof certainnonliberal sects and movements-is itself decidedlyliberal in character,for example by conceiving of individualpersons and of them alone as ultimate units of

equal moralconcern. This quest will put us at odds with many hierarchical

societies whose ideal of a

fully just worldorder will be differentfrom ours.

It may seem then that my more assertiveliberalism will lead to greater internationalconflict. And this may well be so in the areaof human rights.

But it may not be so in the area here at issue: internationalinequality. Rawls rejects all egalitariandistributive principles of internationaljustice

on the

ground(among others) that they areinseparable from liberal values

and thereforeunacceptable to hierarchicalsocieties (LP,p. 75).25 Butin the real world,the chief opponentsof proposalsalong the lines of my GRTare the affluent liberal societies. We are, after all, also the wealthy ones and accountfor a vastlydisproportionate share of globalresource depletion and

pollution. If we submitted the GRT proposalto the rest of the world, I

believe it would be accepted by most

societies with some enthusiasm.26

Given that institutionalprogress is politicallypossible, it would be per-

verse to oppose it by saying to the rest of the world:"We care deeply about

equality,and we would very much like it

to be the case that you are not so

much worse off than we are. But, unfortunately,we do not believe that you

ultimatelycare aboutequality the way we do. Thereforewe feel entitled to

refuse any globalinstitutional reforms that

would lead to greaterinterna-

tional equality."One reason this would be perverse is that those touting

hierarchicalvalues and those suffering most from global inequality are rarelythe same. Those whose lot a GRTwould do most to improve-poor women and rurallaborers in the thirdworld, for example-rarely give the

  • 25. I believe, to the contrary,that rathera lot could be said to supportthe GRTscheme in

terms of nonliberalvalues prevalentin many hierarchicalsocieties today,though I cannot

undertakethis task here.

  • 26. Witness the debatesduring the I970s,

in UNCTADand the GeneralAssembly of the

United Nations, abouta new intemationaleconomic order.

  • 219 An Egalitarian Law of Peoples

hierarchical values of their rulers and oppressors their considered and reflective endorsement.27


Delegates of liberalsocieties might preferan egalitarianlaw of peoples and yet adopt Rawls'sinegalitarian alternative.28 For they might believe that a scheme like the GRTwould simplynot work:The moralmotives ("sense of justice") that a just world order would engender in peoples and their governments would not be strong enough to ensure compliance. There would alwaysbe some wealthypeoples refusing to pay theirfair share, and this in turn would undermine others'willingness to participate.In short:

The GRT scheme is practicableonly if backed by sanctions.29And sanc- tions presupposea worldgovernment, which the delegates have abundant

reasons to reject.

In response, I accept the claim that the GRTscheme would have to be

backed by sanctions. But sanctions do not require a world government.

They could work as follows:Once the agency facilitatingthe flow of GRT

payments reports that a country has not met its obligations under the

scheme, all other countries are required to impose duties on imports from,

and perhaps also similar levies on exports to, this country to raise funds equivalent to its GRTobligations plus the cost of these enforcement mea-

sures. Such decentralizedsanctions stand a very good chance

of discour-

aging small-scale defections. Our world is now, and is likely

to remain,

highly interdependent economically; most countries export and import between io percent and 50 percent of their gross domestic product.None of them would benefit from shutting down foreign trade for the sake of avoiding a GRTobligation of around I percent of GDP. And each would have reasons to meet its GRTobligation voluntarily: to retain full control

  • 27. ShouldI apologizefor my liberalbias here, for being concemed with endorsementby

individualpersons rather than by whole peoples(as expressed,presumably, by their govem-

ments and "elites")?

  • 28. The problemI tryto deal with in this final sectionis not one raisedby Rawls,who holds

that the delegates would even prefer his law of peoples. So nothing I say in response to problemis meant to be criticalof him.


  • 29. One might justify including this claim among the generalknowledge available to the delegates by pointing to how lax many states have been aboutpaying their much smaller membershipdues to the UN.

  • 220 Philosophy & Public Affairs

overhow the funds areraised, to avoidpaying for enforcement measures in addition,and to avoidthe negativepublicity associated with noncompliance. This leaves the problemof large-scaledefections, and the related prob- lem of getting most of the more affluentsocieties to agree to something like

the GRTscheme in

the firstplace. This scheme could not workin our world

withoutthe willing cooperationof most of the wealthiercountries. You may

be tempted to look at the worldas it is and conclude that the hope for such

willing cooperationis not realistic. So you