Jump to ContentJump to Main Navigation
The Quality of Life$

Martha Nussbaum and Amartya Sen

Print publication date: 1993

Print ISBN-13: 9780198287971

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/0198287976.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2021. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use. date: 03 August 2021

Justice, Gender, and International Boundaries

Justice, Gender, and International Boundaries

(p.303) Justice, Gender, and International Boundaries
The Quality of Life

Onora O'Neill

Oxford University Press

Abstract and Keywords

Questions about justice to women and about international justice are often raised in discussions of development. Yet the most influential theories of justice have difficulty in handling either topic. I shall first compare some theoretical difficulties that have arisen in these two domains, and then sketch an account of justice that may be better suited to handling questions of gender and international justice....

1 Justice for Impoverished Providers

Questions about justice to women and about international justice are often raised in discussions of development. Yet the most influential theories of justice have difficulty in handling either topic. I shall first compare some theoretical difficulties that have arisen in these two domains, and then sketch an account of justice that may be better suited to handling questions of gender and international justice.

I begin by distinguishing idealized from relativized theories of justice. Idealized accounts of justice stress the need to abstract from the particularities of persons. They paint justice as blind to gender and nationality. Its principles are those that would regulate the action of idealized ‘abstract individuals’. They take no account of differences between men and women; they transcend international boundaries. Relativized accounts of justice acknowledge the variety and differences among humankind; they ground principles of justice in the discourse and traditions of actual communities. Since nearly all of these relegate (varying portions of) women's lives to a ‘private’ sphere, within which the political virtue of justice has no place, and see national boundaries as the limits of justice, appeals to actual traditions tend both to endorse institutions that exclude women from the ‘public’ sphere, where justice is properly an issue, and to insulate one ‘public’ sphere from another.

Both idealized and relativized accounts of justice look inadequate from the perspective of those whom they marginalize. Women, in particular poor women, will find that neither approach takes account of the reality of performing both reproductive and productive tasks, while having relatively little control over the circumstances of one's life. Women's lives are not well conceived just as those of abstract individuals. A world of abstract individuals assumes away relations of dependence and interdependence; yet these are central to most lives actually available to women. Nor are women's lives well conceived solely in terms of traditions that relegate them to a ‘private’ sphere. The productive contributions and the cognitive and practical independence of actual women are too extensive, evident, and economically significant to be eclipsed by ideologies of total domesticity and dependence.

The awkward fit of theory to actuality is most vivid for poor women in poor economies. These women may depend on others, but lack the supposed securities (p.304) of dependence. They are impoverished, but are often providers. They are powerless, yet others who are yet more vulnerable depend on them for protection.1 Their vulnerability reflects heavy demands as much as slender resources. They may find that they are relegated to and subordinated within a domestic sphere, whose separate and distinctive existence is legitimated not by appeals to justice but by entrenched views of family life and honour. They may also find that this domestic sphere is embedded in an economy that is subordinate to distant and richer economies. They not only raise children in poverty; they raise crops and do ill‐paid and insecure work whose rewards fluctuate to the beat of distant economic forces. This second subordination, too, is legitimated in varied discourses which endorse an internationalized economic order but only national regimes of taxation and welfare. A serious account of justice cannot gloss over the predicaments of impoverished providers in marginalized and developing economies.

2 Preview: Abstraction and Contextualization

Both idealized and relativized approaches to justice make seemingly legitimate demands. Idealized approaches insist that justice must abstract from the particularities of persons. Blindness to difference is a traditional image of justice and guarantees impartiality. Yet principles of justice that are supposedly blind to differences of power and resources often seem to endorse practices and policies that suit the privileged. Hence a demand that justice take account of context can seem equally reasonable. Justice, it is argued, needs more than abstract principles: it must guide judgements that take account of actual contexts and predicaments and of the differences among human beings. Relativized principles of justice meet this demand: but since they are rooted in history, tradition, or local context, they will endorse traditional sexism or nationalism. Any relativism tends to prejudice the position of the weak, whose weakness is mirrored and partly constituted by their marginalization in received ways of thought and by their subordination and oppression in established orders. Yet idealizing approaches do no better. Where relativist approaches are uncritical of established privilege, idealized approaches are uncritical of the privileges from which they abstract.

If idealized and relativized accounts of justice were the only possibilities we would have to choose between demands for abstraction from difference and for sensitivity to difference. If there are other possibilities, an account of justice may be able to meet demands both for abstract principles and for contextualized judgements. I shall try to sketch a third possibility, which gives both abstraction (p.305) and contextualization their due—but only their due. This can be done by meeting the demands for abstract and contextual reasoning in two distinct, successive moves.

The first move is to argue for abstract principles of universal scope, while rejecting the supposed link between abstraction and positions that not merely abstract but (in a sense to be explained) idealize. Much contemporary moral reasoning, and in particular ‘abstract liberalism’ (whether ‘deontological’ or utilitarian), handles issues of gender and international justice badly not because it abstracts (e.g. from sex, race, nationality), but because it almost always idealizes specific conceptions of the human agent and of national sovereignty which are often more admired and more (nearly) feasible for developed rather than developing societies and in men rather than in women. Genuine abstraction, without idealization, is, however, the route rather than the obstacle to broad scope and is unobjectionable in principles of justice.

The second move answers demands that we take account of the context and particularities of lives and societies, but does not endorse established ideals of gender and of national sovereignty. Abstract principles of justice are not rejected, but viewed as intrinsically incomplete and indeterminate, a guide rather than an algorithm for judging cases. The second move insists that justice can take account of certain differences by applying abstract principles to determinate cases without either tacitly reintroducing restricted ideals (e.g. by privileging certain views of gender and sovereignty) or relativizing principles of justice to accepted beliefs, traditions, or practices. Abstract principles can guide contextualized judgements without lapsing into relativism.

3 Feminist Critiques of Abstract Justice

Discussions of gender justice have been structured by disagreements over the extent and import of differences between men and women. For liberals who defend abstract principles of justice it has been embarrassing that the Rights of Man were taken for so long and by so many of their predecessors as the rights of men, and that liberal practice failed for so long to end male privilege.2 (Socialist feminists suffer analogous embarrassments.) Starting with Wollstonecraft and J.S. Mill, liberal feminists argued against the different treatment of women, and claimed that women's rationality entitled them to equal rights.

Later liberal feminists noted that even when women had equal political and legal rights, their political participation and economic rewards remained less than those of men—and less than those of men whose qualifications and labour force participation women matched. Supposedly gender‐neutral and neutralizing institutions, such as democratic political structures and markets, turned out (p.306) to be remarkably bad at reducing gender differentials.3 Approximations to political and legal justice in various domains of life evidently will not close the radical gap between men's and women's paths and prospects.4

In response many liberal feminists argued that justice demands more thorough equal treatment. It may, for example, require forms of affirmative action and reverse discrimination in education and employment, as well as welfare rights to social support for the poor and those with heavy family responsibilities. Some differences are to be acknowledged in principles of justice. This move has two difficulties. First, many liberals deny that justice demands compensatory redistribution, especially of positional goods. They think these should be allocated by competitive and meritocratic procedures. This debate is of particular importance in the developed world.

The second problem arises even where the goods to be distributed are not positional, and is particularly significant in the Third World. Where resources are scarce, non‐positional goods such as basic health care, income support, children's allowances, or unemployment insurance will not be fundable out of a slender national tax base. If social justice demands basic welfare provision, justice has to reach across boundaries. An account of gender justice would then have to be linked to one of international distributive justice.5

This liberal debate continues, but its terms have been increasingly questioned by feminists in the last decade, many of whom claim that, despite its aspirations, gender bias is integral to liberal justice.6 Their suspicions focus on the very abstraction from difference and diversity which is central to liberal justice. Some feminist critics of abstract liberalism have highlighted respects in which particular supposedly gender‐neutral theories covertly assume or endorse gendered accounts of the human subject and of rationality. Many aspects of this critique are convincing.

However, the most fundamental feminist challenge to abstract liberalism impugns reliance on abstraction itself. Gilligan's influential work claims that an emphasis on justice excludes and marginalizes the ‘other voice’ of ethical thought. ‘Abstract liberalism’ simply and unacceptably devalues care and concern for particular others, which are the core of women's moral life and thought, seeing them as moral immaturity.7 The voice of justice is intrinsically ‘male’ (p.307) in its refusal to grasp the actualities of human difference, in its supposed agnosticism about the good for man, and in its resulting disregard of the virtues, and specifically of love and care. On this account the problem is not to secure like treatment for women, but to secure differentiated treatment for all.

In locating the distinction between justice and care (and other virtues) in a disagreement over the legitimacy of relying on abstract principles, some feminist critics of abstract liberalism construe concern for care as opposing concern for justice. They can end up endorsing rather than challenging social and economic structures that marginalize women and confine them to a private sphere. Separatism at the level of ethical theory can march with acceptance of the powers and traditions that be. The cost of a stress on caring and relationships to the exclusion of abstract justice may be relegation to the nursery and the kitchen, to purdah and to poverty. In rejecting ‘abstract liberalism’ such feminists converge with traditions that exclude and marginalize women. Even when they appeal to ‘women's experience’, rather than to established traditions and discourses, as the clue to understanding the other ‘voice’, they agree that differences are taken seriously only when actual differences are endorsed.8

The disputes that divide liberal feminists and their contextualist critics pose an unwelcome dilemma about gender justice. If we adopt an abstract account of justice, which is blind to differences between people—and so to the ways in which women's lives in the developed and in the undeveloped world differ from men's lives—we commit ourselves (it is said) to uniform treatment regardless of difference. If we acknowledge the ethical importance of differences, we are likely to endorse traditional social forms that sustain those differences, including those that subordinate and oppress women.

4 The Communitarian Critique of Abstract Justice

This dilemma recurs in certain discussions of international justice. Abstract liberalism proclaims the Rights of Man. As Burke was quick to complain, this is quite a different matter from proclaiming the rights of Englishmen, or of Frenchmen, or of any coherent group. Abstraction was the price to be paid for ethical discourse that could cross the boundaries of states and nations and have universal appeal; and Burke found the price unacceptable. The internationalist, cosmopolitan commitments that were implicit in the ideals of liberalism have repeatedly been targets of conservative and communitarian criticism.

Liberal practice has, however, once again been embarrassingly different. It (p.308) has not been universalistic, but clearly subordinated to the boundaries and demands of nation states. This is evident in relations between rich and poor states. Like treatment for like cases is partially secured by laws and practices within many democratic states; only a few enthusiasts argue for world government, or think that rights of residence, work, and welfare, as well as burdens of taxation, should be global. Such enthusiasm is often dismissed by practical people who hold that a plurality of national jurisdictions provides the framework within which liberal ideals can be pursued. Liberals may not be generally willing to take differences seriously; but they have taken differences between sovereign states remarkably seriously.

Their communitarian critics take differences and boundaries seriously in theory as well as in practice.9 When boundaries are taken wholly seriously, however, international justice is not just played down, but wiped off the ethical map. Walzer's work is a good case in point. He holds that the largest sphere of justice is the political community and that the only issues not internal to such communities are about membership in them and conflicts between them. The issues of membership concern the admission of individual aliens; rights and duties do not go beyond borders.10 A commitment to community is a commitment to the historical boundaries of political communities, whatever these happen to be and whatever injustices their constitution and their preservation entail. Communitarians cannot easily take any wider view of ethical boundaries since their critique of abstraction is in part a demand for ethical discourse that takes ‘our’ language, ‘our’ culture, and ‘our’ traditions seriously.11

Like current debates on gender justice, discussions of international justice apparently pose an unwelcome choice. Either we can abstract from the reality of boundaries, and think about principles of justice that assume an ideal cosmopolitan world, in which justice and human rights do not stop at the boundaries of states. Or we can acknowledge the reality of boundaries and construe the principles of justice as subordinate to those of national sovereignty. Cosmopolitan ideals are evident in the discourse of much of the human rights movement; but recent liberal theorists have shifted towards the relativism of their communitarian critics, and even come to regard liberal principles of justice as no more than the principles of liberal societies. Rawls in particular now12 (p.309) hinges his theory of justice not on an abstract and idealized construction of an original position but on the actual ideals of citizens of liberal democratic societies. Here we see a surprising and perhaps unstable convergence between abstract liberal theorists and their communitarian critics.

5 Abstraction With and Without Idealization

Debates about gender and international justice are not merely similar in that each is structured by a confrontation between advocates of abstract and of contextualized justice. In each debate many advocates of supposedly abstract approaches to justice go far beyond abstraction. What these debates term ‘abstraction’ is often a set of specific, unargued idealizations of human agency, rationality, and life, and of the sovereignty and independence of states. And in each debate what is said to be attention to actual situations and contexts in judging is in fact relativism about principles. These conflations are avoidable.

Abstraction, taken strictly, is simply a matter of detaching certain claims from others. Abstract reasoning hinges nothing on the satisfaction or non‐satisfaction of predicates from which it abstracts. All uses of language must abstract more or less: the most detailed describing cannot dent the indeterminacy of language. Indeed it is not obvious that there is anything to object to in very abstract principles of justice. Highly abstract ways of reasoning are often admired (mathematics, physics), and frequently well paid (accountancy, law). What is different about abstract ethical reasoning? When we look at objections to ‘abstract’ ethical principles and reasoning in detail, it appears that they are often objections not to detachment from certain predicates, but to the inclusion of predicates that are false of the objects of the domain to which a theory is then applied. Reasoning that abstracts from some predicate makes claims that do not hinge on the objects to which the reasoning is applied satisfying that predicate. Reasoning that idealizes does make claims that hinge on the objects to which it is applied satisfying certain predicates. Where those predicates are unsatisfied the reasoning simply does not apply.

The principles and theories of justice to which the critics of ‘abstract liberalism’ object are indeed abstract. They take no account of many features of agents and societies. However, these principles and theories not only abstract but idealize. They assume, for example, accounts of rational choice whose claims about information, coherence, capacities to calculate, and the like are not merely not satisfied by some deficient or backward agents, but are actually satisfied by no human agents (perhaps they are approximated, or at least admired, in restricted shopping and gambling contexts!). They also assume idealized accounts of the mutual independence of persons and their opportunities to pursue their individual ‘conceptions of the good’, and of the sovereignty and independence of states, that are false of all human beings and all states. Such idealizations no doubt have theoretical advantages: above all they allow us to (p.310) construct models that can readily be manipulated. However, they fail to apply to most, if not all, practical problems of human choice and foreign policy.

If idealized descriptions are not abstracted from those that are true of actual agents, they are not innocuous ways of extending the scope of reasoning. Each idealization posits ‘enhanced’ versions of the objects of the domain to which the model is applied. They may privilege certain sorts of human agent and life and certain sorts of society by covertly presenting (enhanced versions of) their specific characteristics as the ideal for all human action and life. In this way covert gender chauvinism and exaggerated respect for state power can be combined with liberal principles. Idealization masquerading as abstraction produces theories that appear to apply widely, but which covertly exclude those who do not match a certain ideal, or match it less well than others. Those who are excluded are then seen as defective or inadequate. A review of the debates about gender and international justice shows that the feminist and communitarian critique of liberal justice can legitimately attack spurious idealizations without impugning abstraction that eschews idealization.

6 Gender and Idealized Agents

Liberal discussions of justice ostensibly hinge nothing on gender differences. They apply to individuals, considered in abstraction from specific identities, commitments, and circumstances. Recent critics insist that liberal theories of justice are far from being as gender‐blind as their advocates claim. An instructive example is Rawls's A Theory of Justice. Rawls was particularly concerned not to rely on an extravagant model of rational choice. His principles of justice are those that would be chosen by agents in an ‘original position’ in which they know less rather than more than actual human agents. He conceives his work as carrying the social contract tradition to ‘a higher level of abstraction’. In particular, agents in the original position do not know their social and economic position, their natural assets, or their conceptions of the good.13 The original position operationalizes the image of justice as blind to difference.

However, Rawls has at a certain point to introduce grounds for those in the original position to care about their successors. He suggests that we may think of them as heads or at other times as representatives of families, ‘as being so to speak deputies for an everlasting moral agent or institution’14 and that some form of family would be just. Yet in doing so he pre‐empts the question of intrafamilial justice. He pre‐empts the question not by crude insistence that heads of families must be men, but by taking it as read that there is some just form of family which allows the interests of some to be justly represented by others. The shift from individuals to heads of families as agents of construction is not an innocent abstraction; it assumes a family structure which secures identity (p.311) of interests between distinct individuals. It takes for granted that there is some just ‘sexual contract’,15 that justice can presuppose a legitimate separation of ‘private’ from ‘public’ domains. This is idealization indeed: it buries the question of gender justice rather than resolving it. Rawls's text leaves it surprisingly obscure whether women are to be relegated to a ‘private’ sphere and represented by men in the construction of justice, whether both ‘public’ and ‘private’ realms are to be shared by all on equal terms, or whether women alone are to carry the burdens of both spheres.16

The more radical feminist critique of abstract liberalism rejects not merely the suppressed gendering of the subject which Pateman and Okin detect in classical and contemporary liberal writers, but abstraction itself. In advocating an ethic of care these critics, we have seen, come close both to traditional misogynist positions and to ethical relativism. When the ‘voices’ of justice and of care are presented as alternatives between which we must choose, each is viewed as a complete approach to moral issues. However, the two in fact focus on different aspects of life. Justice is concerned with institutions, care and other virtues with character, which is vital in unmediated relationships with particular others. The central difference between the ‘voices’ of justice and care is not that they reason in different ways. Justice requires judgements about cases as well as abstract principles; care is principled as well as responsive to differences. Justice matters for impoverished providers because their predicament is one of institutionally structured poverty which cannot be banished by idealizing an ethic of care.

7 Idealized Boundaries

A comparable slide from abstraction to idealization can be found in discussions of international justice. Discussions of global economic and political issues often take it for granted that the principal actors are states. Traditionally the main divide in these discussions has been between realists, who contend that states, although agents, are exempt from moral obligations and criticism, and idealists, who insist that states are not merely agents but accountable agents who must meet the demands of justice.17

In discussions of distributive justice, however, the salient issue has not been the conflict between idealists and realists, but their agreement that state (p.312) boundaries define the main actors in international affairs. These shared terms of debate endorse an exaggerated, idealized view of the agency and mutual independence of sovereign states, which is now often criticized as obsolete. The common ground on which realists and idealists traditionally debated international relations is being eroded as other actors, including international agencies, regional associations, and above all transnational corporations, play a more and more significant role in world affairs.18 A world that is partitioned into discrete and mutually impervious sovereign states is not an abstraction from our world, but an idealized version of it, or perhaps an idealized version of what it once was. Realists as well as idealists idealize the sovereignty of states.

Idealized conceptions both of state sovereignty and of state boundaries limit discussions of international distributive justice. Although long subject to theoretical questioning from advocates of human rights, who deny that states can be sovereign in determining the fates of individuals, many liberals are coy about criticizing rights violations elsewhere. They limit criticism to violations of liberty rights, and offer little account of the agency or responsibilities of institutions; they find it hard to see how justice could require that state boundaries be breached to reduce the poverty that lies beyond them. Even those liberals who defend welfare rights are often concerned with welfare in one (rich) country. It is commonplace to view economic development of poorer regions as optional ‘aid’, not obligatory justice. Those who have tried to argue for global welfare rights within a liberal framework have to show who bears the obligations that correspond to these rights, and this has proved uphill work.19 Meanwhile liberals, like communitarians, confine justice within national boundaries. Liberals do so self‐consciously and provisionally; communitarians on principle and unapologetically; others tacitly and without discussion.

8 Abstraction Without Idealization

The only way to find theories that have wide scope is to abstract from the particularities of agents; but when abstraction is displaced by idealization we are not led to theories with wide scope, but to theories that apply only to idealized agents.

This suggests that if we are interested in international or gender justice we should resist the temptation to rely on idealizing models of human agency or national sovereignty. We should instead consider what sort of theory of justice we would have if we abstracted but refused to idealize any one conception of rationality or independence, and so avoided marginalizing or excluding those who do not live up to specific ideals of rationality or of independence from others. Abstraction without idealization may allow us to consider a wide range of human (p.313) agents and institutional arrangements without hinging anything on the specific features of agents' traditions, ideologies, and capacities to act. If we could do this we might avoid idealized accounts of agency and sovereignty without following feminist and communitarian critics of abstract liberalism into relativism.

Recent discussions may simply have been mistaken in treating appeals to idealized and relativized standards of rationality and agency as the only options. There are other possibilities. We do not have to hinge liberal arguments for rights or for the limits of government power either on the hypothetical consent of those who meet some ideal standard of rationality and mutual independence, or on the actual acceptance of an outlook and its categories that relativizes consent to an established order. We could instead begin simply by abstracting from existing social orders. We could consider what principles of action must be adopted by agents who are numerous, diverse, and neither ideally rational nor ideally independent of one another, and yet avoid specific assumptions about these agents. We can bracket both idealizations and the status quo. The issue then becomes: how powerful and convincing an account of justice can we offer if we appeal neither to fictions of ideal rationality and independence nor to the contingencies of actual agents and institutions? What happens if we abstract without idealizing?

9 Plurality and Justice: Who Counts?

Let us begin with the thought of a plurality of potentially interacting and diverse agents. This rules out two cases. First, it rules out the case where justice is not a problem because there is no plurality, or no genuine plurality, of agents, hence no potential for conflict between agents. (The action of agents in such a degenerate plurality would be automatically or necessarily co‐ordinated, e.g. by instinct or by a pre‐established harmony.) Second, it rules out hinging an account of justice on an assumed, contingent, and determinate limit to the diversity of its members, which provides a common ground between them and permits a contingent, socially guaranteed convergence and co‐ordination. The two cases that are ruled out are once again those which would base principles of justice on an assumed ideal convergence or an assumed actual historical or social convergence.

What does justice require of such a plurality? At least we can claim that their most basic principles must be ones that could be adopted by all. If they were not, at least some agents would have to be excluded from the plurality for whom the principles can hold, whose boundaries would have to be drawn more narrowly.

Such a redrawing of boundaries is, of course, the very move often used to exclude women and foreigners, let alone foreign women, from the domain of justice. Those who exclude simply refuse to count certain others as members of a plurality of potentially interacting agents. An account of justice which hinges on the sharability of principles can be pre‐empted by excluding some from the domain of justice without argument. So it is important to see the move for what (p.314) it is. This can best be done by asking who makes the move.

The move is not made by idealized genderless theorists who live outside state and society. It is made by people who generally expect women to interact with them, to follow language and reason, to understand and take part in elaborate traditions and institutions, perhaps even to love, honour, and obey. It is made by people who expect ordinary processes of translation, trade, and negotiation to work with foreigners. To deny the agency of others with whom we plan to interact in complex ways reeks of bad faith. Bad faith can be avoided only by counting as members of the plurality for whom principles of justice are to hold anybody with whom interaction is to be attempted or held possible. The question then becomes: are there any principles which must be adopted by all members of a plurality of potentially interacting agents? We cannot simply stipulate that such principles are irrelevant for interactions with certain others on whose (no doubt imperfect) capacities to reason and (no doubt limited) abilities to act independently we know we depend.

If women were all transported to Betelgeuse, and so beyond all interaction with the remaining men on Earth, neither men nor women would have to see the other as falling within the domain of justice. Less fancifully, since the ancient inhabitants of the Andes and their contemporaries in Anglo‐Saxon England could not and did not interact, neither would have acted in bad faith if they had excluded the other from the domain of justice. Neither could practise either justice or injustice towards the other. Things are different for the actual men and women who inhabit the earth now: the potential for interaction cannot be assumed away, and others cannot be arbitrarily excluded from the domain of justice. We rely on global economic and political processes, so cannot consistently insist that justice (conveniently for the developed world) stops at state frontiers, any more than we can rely on women's rationality and their productive contribution and then argue that justice (conveniently for some men) stops at the edge of a supposed ‘private’ sphere, whose existence and demarcation is in fact presupposed in defining a ‘public’ sphere.

10 Plurality and Justice: What Principles?

Justice is then in the first place a matter of keeping to principles that can be adopted by any plurality of potentially interacting beings. But if we eschew both idealization and relativism, and rely on mere abstraction, will we have strong enough premisses to identify those principles? Does a universalizability test cut any ice? Granted that universalizability is not uniformity (as some critics of abstract liberalism suppose), is it not too weak a demand to ground an account of justice? In particular, will not any internally coherent principle for individual action be a universalizable principle?20

(p.315) We have, however, to remember that we are considering the case of a plurality of potentially interacting beings, that is of beings who share a world. Any principle of action that is adopted by all members of such pluralities alters the world that they share and becomes a background condition of their action. This is why certain principles of action which can coherently be held by some cannot be coherently held by all. Examples of non‐universalizable principles can illustrate the point. A principle of deception, which undermines trust, would, if universally adopted, make all trusting, hence all projects of deception, incoherent. Selective deception is on the cards: universal deception is impossible. Since a principle of deception cannot be a fundamental one for any plurality, justice requires that it be rejected. Equally, a policy of coercion which seeks to destroy or undercut the agency and independence of at least some others for at least some time cannot be universally held. Those who are victims of coercion cannot (while victims) also act on the principles on which their coercers act.21 Equally, a principle of violence which damages the agency of some others cannot be universally acted on. Put quite generally, principles of action that hinge on victimizing some, so on destroying, paralysing, or undercutting their capacities for action for at least some time and in some ways, can be adopted by some but cannot be adopted as fundamental principles by any plurality.22

To keep matters under control, let us imagine only that justice demands (at least) that action and institutions should not be based on principles of deception and victimization. (There may be other principles of justice.) We are still far from showing just what justice demands, since we do not know what refusing to deceive or to coerce may demand in specific circumstances. These guidelines are highly indeterminate. We seem to have paid the classic price of abstraction. Highly abstract principles do not tell us what to do in a specific context.

Abstract principles are only part of practical, or specifically of ethical, reasoning, however. Principles never determine their own applications; even the culturally specific principles that relativists favour do not determine their own applications. All practical reasoning requires judgement and deliberation by which principles are applied to particular cases. An account of gender and international justice is no exception. We need in particular to be able to judge what specific institutions and action are needed if poor women in poor economies are to be accorded justice.

11 Plurality and Justice: Deliberation without Relativism

Two background issues must be dealt with summarily before considering moves from abstract basic principles to determinate judgements. First, we have no (p.316) reason to expect that principles of justice will provide any algorithm of rational choice. Nor do we need any algorithm for principles to be important. Even principles that provide only a set of side constraints on action may exert a powerful influence. Second, we have no reason to think that principles of justice are relevant only to the action of individuals. A full account of the agency of institutions would be a complex matter. I shall not go into it here, but will assume that it can be given and that institutions and practices, like individuals, must meet the demands of justice.

These moves, however, are preliminary to the main task of giving a more determinate account of what may be required if principles of deception or victimization are rejected. How, for example, can we judge whether specific types of family or economic activity are based on deception or victimization? Are all forms of hierarchy and subordination coercive? If not, how do we discern the boundaries of deceit and coercion in actual contexts? It is not hard to see that certain categories of individual action — for example, fraud or wife burning or battering — deceive or victimize, but other cases of deception and coercion by individuals are hard to adjudicate. It may also be hard to judge whether social traditions that isolate or exclude women, or economic and familial arrangements that ensure their acute economic vulnerability, amount to modes of deceit and coercion.

In this paper the task cannot be to reach determinate judgements about particular cases, but only to see that reasoned moves from very abstract principles towards more specific principles, whose relevance and application to particular cases may be easier to assess, may be possible. It will not be enough to lean on the received criteria by which ‘our’ tradition or nation picks out ethically significant ‘cases’ or ‘options’ for approaching them. We beg questions if we assume that categories of thought that have been hospitable to male dominance and imperialism can be decisive for discerning or judging justice to those whose problems have been marginalized and whose agency and capacities have been formed, perhaps deformed, by unjust institutions. We cannot rely uncritically on the categories of established discourse, including the discourse of social scientists and the ‘helping’ professions, to pick out the significant problems. These categories are themselves matters for ethical concern and criticism.23 We have, after all, no more reason to trust relativized discussions of justice, gender, or boundaries than to trust idealized approaches unequivocally. Those discussions are no more free of theory and ideology than are idealized discussions of justice. Their ways of individuating typical problem cases may be familiar; but familiarity may mask contentious and unjust delimitations. If the received views of a society or tradition are taken as defining the domain of problems to which abstract principles of justice are applied, unvindicated ideals will be introduced and privileged, just as they are in idealized approaches to justice.

Some confirmation of the ways in which received descriptions of social relations (p.317) reflect larger and disputed ideals is suggestive. Consider, for example, how issues of gender are passed over as if invisible. We find an enormous amount of shifting around in the choice of basic units of social analysis. In the shifts between descriptions that focus on individuals, wage‐earners, and heads of families, there is enough flexibility for the blunt facts of economic and other subordination of women to be veiled. Women's low wages can seem unworrying if they are wives for whom others provide; their dependence on husbands and fathers can seem acceptable if they are after all wage‐earning individuals, so not invidiously dependent. Reproductive labour may (with convenient ambiguity!) be thought of as priceless.24 Wage‐earning women's low pay can be seen as fitting their low skills and vindicating their domestic subordination to wage‐earning men, who as ‘heads of families’ are entitled to discretionary expenditure and leisure which wage‐earning women must do without because they (unlike men!) have family commitments. The gloomy evidence of social structures that classify women's contributions as less valuable even when more onerous or more skilled is evident enough. We continually find ourselves ‘thinking about men as individuals who direct households and about women as family members’.25

There are equally serious reasons to mistrust the move from abstract principles to determinate judgements in discussions of individual motivation. These too are shaped by received views, and in milieux which are strongly individualist are easily diverted into attempts to pin blame for injustices on individuals. Women, after all, commonly acquiesce in their social and economic subordination. Are they then to be blamed for servility? Or are men to be blamed for oppressing or exploiting women?26 Or do these individualist approaches to assigning blame lead no further than the higher bickering? It can seem that we have reasons to mistrust not only relativist approaches to gender justice but even the attempt to apply an abstract, non‐idealized principle of justice. But we do not inhabit an ideal world. Idealized conceptions of justice simply do not apply to international relations, social relations, or individual acts in a world in which states, men, and women always lack the capacities and the opportunities of idealized agents. States are not really sovereign — even superpowers have limited powers; and men and women are always more or less vulnerable, ignorant, insecure, lacking in confidence or means to challenge or oppose the status quo. In a world of agents with finite capacities and opportunities, poor women in poor economies differ not in kind but in degree in their dependence on others and in others' demands on them.

12 Just Deliberation in a World of Vulnerable Agents

If we are to apply principles of justice that are neither idealized, nor merely relative to actual societies, to vulnerable lives and their predicaments, we must (p.318) see how to move towards determinate judgements about actual cases. The principles of justice for which I have argued take us in this direction because they focus neither on the arrangements to which ideally rational and mutually independent beings would consent, nor on the arrangements to which others in possibly oppressive situations do consent. Rather they ask which arrangements a plurality of interacting agents with finite capacities could consent to. I have suggested, provisionally, that this non‐idealizing construction identifies the rejection of deception, coercion, and other ways of victimizing others as principles of justice.

But principles are not enough. Non‐idealizing abstraction avoids some problems, but not others. If we are to move from abstract principles to determinate judgements we need to operationalize the idea of avoiding acting on unsharable principles, without subordinating it to the categories and views of the status quo. One reasonable way of doing so might be to ask to what extent the arrangements that structure vulnerable lives are ones that could have been refused or renegotiated by those whom they actually constrain. If those affected by a given set of arrangements could have refused or renegotiated them, their consent is no mere formality, but genuine, legitimating consent. If they could not but ‘accept’ those institutions, their ‘consent’ will not legitimate. The point of this way of operationalizing the notion of possible consent is that it neither ascribes ideal reasoning capacities and ideal independence from others nor hinges legitimation on an actual ‘consent’ that may reflect injustice. On this account justice requires that institutions, like acts, allow those on the receiving end, even if frail and dependent, to refuse or renegotiate the roles and tasks assigned to them.

Dissent becomes harder when capacities to act are less developed and more vulnerable, and when opportunities for independent action are restricted. Capacities to act are constrained both by lack of abilities and by commitments to others. Institutional arrangements can disable agency both by limiting capacities to reason and act independently and by increasing the demands to meet the needs and satisfy the desires of others. Apparent consent to such arrangements does not show that they are just. Whenever ‘consent’ reflects lack of capacity or opportunity to do anything else, it does not legitimate. Thinking in this way about justice, we can see that it demands more, not less, to be just to the vulnerable. The vulnerable are much easier to deceive and to victimize than the strong. If we are to judge proposals for action by seeing whether they involve serious deception or victimization (coercion or violence), more will be demanded when others are vulnerable than when they are secure, and most when they are most vulnerable.27 By contrast both idealized and relativized accounts of justice tend to conceal the fact that justice to the weak demands more than justice to the (p.319) strong. Idealized accounts of justice tend to ignore vulnerability and relativized accounts to legitimate it.

13 Achieving Justice for Impoverished Providers

The lives of poor women in poor economies illustrate these points well. Consider, for example, daily commercial transactions and practices. Their justice, it is usually said, lies in the fact that arrangements are mutually agreed. But where there are great disparities of knowledge and vulnerability between agents, the ‘agreement’ of the weak may be spurious. They may have been duped by offers they did not understand or overwhelmed by ‘offers’ they dared not refuse. Within national jurisdictions these facts are well recognized, and commercial practice is regulated to prevent pressure and fraud. Contracts can be voided for fraud; there are ‘truth in lending’ provisions; debt and bankruptcy lead not to loss of liberty but to loss of property; those with dependants can rely on a safety net of welfare rights. International economic transactions take place in a far less regulated space, yet link agents with far greater disparities in power and resources. The weak can suffer both from particular others who take advantage of their ignorance and vulnerability, and because nothing informs them about or shields them from the intended or unintended consequences either of distant or of local economic forces. The poor, and above all women who are impoverished providers, cannot refuse or renegotiate economic structures or transactions which hurt them. They are vulnerable not only to low wages, low standards of industrial safety, endemic debt, and disadvantageous dependence on those who provide credit, but also to disadvantageous patterns of entitlement within the family. Debtors who need further loans for survival, for example, cannot make much fuss about the terms creditors offer for purchasing their crops.28 Market ‘imperfections’ are neither avoidable nor trivial for vulnerable agents with many dependants; equally, ‘perfect’ markets can magnify vulnerability to distant economic forces.

Idealized pictures of justice have tended to overlook the import of economic power: by idealizing the capacities and the mutual independence of those involved in market transactions they obscure the reasons why the weak may be unable to dissent from arrangements proposed by the strong. They also tend to distinguish sharply between intended and unintended consequences, and to view the latter as unavoidable ‘forces’. Yet these forces are themselves the outcome of institutional arrangements and could be changed or modified, as they have been within many jurisdictions. The problem of shielding the weak from these forces has nothing to do with ‘natural’ processes, and everything to do with the weakness of the voices that call for change. This is hardly surprising. Market institutions magnify the security and so the voices of the haves. Formal democracy provides only slender and partial redress for the weak, and is often lacking.

(p.320) Typical family structures illustrate the gulf between ideally independent agents (whom ‘ideal’ market structures might suit) and actual powerlessness. These structures often draw a boundary between ‘public’ and ‘private’ domains, assign women (wives and daughters) to the ‘private’ domain, and leave them with slender control of resources, but heavy commitments to meet others' needs. They may lack adequate economic entitlements, effective enfranchisement, or access to sources of information or debate by which to check or challenge the proposals and plans of the more powerful. Women in this predicament lack security, and must meet the demands of others (often fathers and husbands) who dominate them. Family structures can enable, even impose, forms of deception and domination. Where women are isolated, secluded, barred from education or wage earning, or have access to information only via the filter of more powerful family members, their judgement is weakened, and their independence stunted. Often this vulnerability may be shielded by matching concern and restraint; often it will not. A rhetoric of familial concern and protective paternalism can easily camouflage a callous lack of concern and legitimate deceptive acts and practices.

Similar points can be made about victimization. A principle of non‐coercion, for example, basically demands that action should not undercut others' agency. If agents were all ideally independent of one another, they might find little difficulty in dissenting from many forms of attempted domination. However, family structures always limit independence, and usually limit women's independence more. A woman who has no adequate entitlements of her own, and insecure rights to a share in family property or income, will not always be coerced, but is always vulnerable to coercion.29 When her independence is also restricted by family responsibilities she will be even easier to coerce. In these circumstances ostensible consent reveals little; it certainly does not legitimate forms of domination and subordination. Relations of dependence are not always or overtly coercive; but they provide structures of subordination within which it is all too easy to silence or trivialize the articulation of dissent. To guarantee that action is not based on principles which others cannot share, it is necessary to ensure that proposals that affect others are ones from which they can dissent. Institutionalized dependence tends to make dissent hard or impossible. Those who cannot secure economic independence or who cannot rely on others to take a share in caring for genuine dependants (children, the elderly) cannot easily say ‘no’ or set their own terms. They must go along with the proposals of the more powerful.

Genuine, legitimating consent is undermined by the very institutions which most readily secure an appearance of consent. Institutionalized dependence may ensure that the weak provide a spurious ‘consent’ to the action of the strong, while remaining at their mercy. If the strong reliably show restraint, there may in (p.321) fact be no injustice within relationships which institutionalize dependence; but institutions that rely too heavily on the self‐restraint of the stronger cannot reliably avoid injustice. Whether the proposals of the strong are economic or sexual, whether they rely on the ignorance and isolation of the weak to deceive them, or on their diminished opportunities for independent action, or on the habits of deference and appeasement which become second nature for the weak, they ride on unjust social practices. The weak risk recurrent injustice unless institutions are structured to secure the option of refusal or renegotiation for those whose capacities and opportunities are limited.

A woman who has no entitlements of her own lives at the discretion of other family members who have them, so is likely to have to go along even with proposals she greatly dislikes or judges imprudent. If she were an ideally independent agent, or even had the ordinary independence and opportunities of those who have entitlements adequate for themselves and their dependants, she could risk dissent from or at least renegotiate proposals put by those who control her means of life. Being powerless and vulnerable she cannot readily do either. Hence any consent that she offers is compromised and does not legitimate others' proposals. Just as we would find it absurd to hinge legitimating consent to medical treatment on procedures geared to the cognitive capacities and independence of a notional ‘ideal rational patient’, so we should find it absurd to hinge legitimating consent to others' plans on the cognitive capacities and independence of a notional ideal rational impoverished provider for others.

This is not to say that impoverished providers are irrational or wholly dependent or cannot consent. It is, however, a matter of taking seriously the ways in which their capacities and opportunities for action constrain their possibilities for refusal and negotiation. If they are to be treated with justice, others who interact with them must not rely on these reduced capacities and opportunities to impose their will. Those who do so rely on unjust institutional structures that enable deceit, coercion, and forms of victimization.

In applying abstract, non‐idealizing principles we have to take account not indeed of the actual beliefs, ideals, or categories of others, which may reflect unjust traditions, but of others' actual capacities and opportunities to act — and their incapacities and lack of opportunities. This move does not lead back to relativism: no principle is endorsed because it is actually accepted. Put in general terms we can use modal notions to identify principles, but indicative ones to apply them. The principles of justice can be determined for any possible plurality: for they demand only the rejection of principles that cannot be shared by all members of a plurality. Judgements about the justice of actual situations are regulated but not entailed by these principles. The most significant features of actual situations that must be taken into account in judgements about justice are the security or vulnerability that allow actual others to dissent from and to seek change in the arrangements which structure their lives.

(p.322) Bibliography

Bibliography references:

Alston, P. (1984). ‘International Law and the Human Right to Food’, in P. Alston and K. Tomasevski (eds.), The Right to Food. Dordrecht: Nijhoff.

—— and Tomasevski, K. (eds.) (1984). The Right to Food. Dordrecht: Nijhoff.

Beitz, Charles (1979). Political Theory and International Relations. Princeton, NJ: Princeton University Press.

Bittner, Rüdiger (1974). ‘Maximen’, in G. Funke (ed.), Akten des 4. Internationalen Kant‐Kongresses. Berlin: De Gruyter.

Brown, Peter, and Shue, Henry (eds.) (1981). Boundaries: National Autonomy and its Limits. New Jersey: Rowman & Littlefield.

Charvet, John (1982). Feminism. London: Dent.

Chodorow, Nancy (1978). The Reproduction of Mothering. Berkeley, Calif.: University of California Press.

Edelman, Murray (1984). ‘The Political Language of the Helping Professions’, in Michael J. Shapiro (ed.), Language and Politics. New York: NYU Press.

Gewirth, Alan (1982). ‘Starvation and Human Rights’, in his Human Rights: Essays on Justification and Applications. Chicago: Chicago University Press.

Gilligan, Carol (1982). In a Different Voice: Psychological Theory and Women's Dependence. Cambridge, Mass.: Harvard University Press.

Harriss, Barbara (1987). ‘Merchants and Markets of Grain in South Asia’, in Teodor Shanin (ed.), Peasants and Peasant Societies. Oxford: Blackwell.

—— (1988). Differential Female Mortality and Health Care in South Asia, Working Paper 13. Oxford: Queen Elizabeth House, and Journal of Social Studies, 41, Dhaka University.

—— (1991). ‘Intrafamily Distribution of Hunger in South Asia’, in J. Drèze and A. K. Sen (eds.), The Political Economy of Hunger, i. Oxford: Clarendon Press.

Hill, Thomas (1973). ‘Servility and Self Respect’, Monist, 57, 87–104.

Höffe, Otfried (1977). ‘Kants kategorischer Imperativ als Kriterium des Sittlichen’, Zeitschrift für Philosophische Forschung, 31, 354–84.

Hoffman, Stanley (1981). Duties Beyond Borders: On the Limits and Possibilities of Ethical International Politics. Syracuse, New York: Syracuse University Press.

Jaggar, Alison M. (1983). Feminist Politics and Human Nature. Brighton: Harvester Press.

Keohane, Robert O., and Nye, Joseph S. (eds.) (1970). Transnational Relations and World Politics. Cambridge, Mass.: Harvard University Press.

Kittay, Eva Feders, and Meyers, Diane T. (eds.) (1987). Women and Moral Theory. New York: Rowman and Littlefield.

Lloyd, Genevieve (1984). The Man of Reason: ‘Male’ and ‘Female’ in Western Philosophy. London: Methuen.

Luper‐Foy, Stephen (ed.) (1988). Problems of International Justice. Boulder and London: Westview Press.

MacIntyre, Alasdair (1981). After Virtue. London: Duckworth.

—— (1984). Is Patriotism a Virtue? University of Kansas, Lawrence: Philosophy Department.

MacMillan, Carol (1982). Women, Reason and Nature. Oxford: Blackwell.

Nicholson, Linda (1987). ‘Feminism and Marx: Integrating Kinship with the Economic’, in Seyla Benhabib and Drucilla Cornell (eds.), Feminism as Critique. Cambridge: Polity Press.

(p.323) Noddings, Nell (1984). Caring. Berkeley, Calif.: University of California Press.

Okin, Susan Miller (1979). Women in Political Thought. Princeton, NJ: Princeton University Press.

—— (1987). ‘Justice and Gender’, Philosophy and Public Affairs, 16, 42–72.

O'Neill, Onora (1986). Faces of Hunger: An Essay on Poverty, Justice and Development. London: George Allen and Unwin.

—— (1988a). ‘Children's Rights and Children's Lives’, Ethics, 98, 445–63.

—— (1988b). ‘Ethical Reasoning and Ideological Pluralism’, Ethics, 98, 705–22.

—— (1989). Constructions of Reason: Explorations of Kant's Practical Philosophy. Cambridge: Cambridge University Press.

Pateman, Carole (1988). The Sexual Contract. Cambridge: Polity Press.

Pfeffer, Raymond, (1985). ‘The Responsibility of Men for the Oppression of Women’, Journal of Applied Philosophy, 2, 217–29.

Postow, B. C. (1978–9). ‘Economic Dependence and Self‐respect’, Philosophical Forum, 10, 181–201.

Rawls, John (1970). A Theory of Justice. Cambridge, Mass.: Harvard University Press.

—— (1985) ‘Justice as Fairness: Political not Metaphysical’, Philosophy and Public Affairs, 14, 223–51.

Ruddick, Sara (1987). ‘Remarks on the Sexual Politics of Reason’, in Kittay and Meyers (1987).

—— (1989). ‘Maternal Thinking’, in her Maternal Thinking: Towards a Politics of Peace. Boston, Mass: Beacon Press.

Sandel, Michael (1982). Liberalism and the Limits of Justice. Cambridge: Cambridge University Press.

Scott, Alison MacEwen (1986). ‘Industrialization, Gender Segregation and Stratification Theory’, in Rosemary Crompton and Michael Mann (eds.), Gender and Stratification. Cambridge: Polity Press.

Sen, Amartya K. (1981). Poverty and Famines: An Essay on Entitlement and Deprivation. Oxford: Clarendon Press.

—— (1987). Gender and Cooperative Conflicts. WIDER Working Paper. Helsinki: World Institute for Development Economics Research.

Shue, Henry (1980). Basic Rights: Subsistence, Affluence and U.S. Foreign Policy. Princeton, NJ: Princeton University Press.

—— (1981). ‘Exporting Hazards’, in Brown and Shue (1991).

—— (1984). ‘The Interdependence of Duties’, in Alston and Tomasevski (1984).

Singer, Peter (1972). ‘Famine, Affluence and Morality’, Philosophy and Public Affairs, 3, 229–43.

Stiehm, Judith Hicks (1983). ‘The Unit of Political Analysis: Our Aristotelian Hangover’, in Sandra Harding and Merrill B. Hintikka (eds.), Discovering Reality: Feminist Perspectives on Epistemology, Metaphysics, Methodology and Philosophy of Science. Dordrecht: Reidel.

Walzer, Michael (1983). Spheres of Justice: A Defence of Pluralism and Equality. Oxford: Martin Robertson.

Williams, Bernard (1985). Ethics and the Limits of Philosophy. London: Fontana.


I would particularly like to thank Deborah Fitzmaurice, James Griffin, Barbara Harriss, Martha Nussbaum, and Sara Ruddick for help with various problems that arose in writing this paper.

(1) Cf. Ruddick (1989). Her account of women's predicament stresses that it reflects heavy demands as much as meagre resources. To be preferred, I think, because it does not take for granted that the lack of resources is significant because ‘public’ while the press of others' demands is less so because merely ‘private’.

(2) Okin (1979); Charvet (1982); Pateman (1988); Jaggar (1983).

(3) Scott (1986).

(4) The differences run the gamut of social indicators. Most dramatically in some Third World countries women and girls do worse on a constellation of very basic social indicators: they die earlier, have worse health, eat less than other family members, earn less, and go to school less. See Sen (1987); Harriss (1988 and 1991).

(5) The problem is not merely one of resources. Where funds have been adequate for publicly funded welfare provision, this too has been inadequate to eliminate the differences between the economic and political prospects of men and of women. Many women in the formerly socialist countries, for example, find that they have secured greater equality in productive labour with no reduction in reproductive tasks. This is a reason for doubting that arguments establishing welfare rights—e.g. a right to food—take a broad enough view of disparities between men's and women's prospects.

(6) E.g. Pateman (1988); Okin (1987).

(7) Gilligan (1982); Kittay and Meyers (1987); Lloyd (1984); MacMillan (1982); Ruddick (1987); Noddings (1984); Chodorow (1978).

(8) Many of those who urge respect for the ‘other’ voice insist that they do not reject the demands of justice, and that they see the two ‘voices’ as complementary rather than alternative. The positions taken by different writers, and by the same writers at different times, vary. The protests must be taken in context: those who appeal to ‘women's experience’ or ‘women's thinking’ appeal to a source that mirrors the traditional relegation of women to a ‘private’ sphere, and cannot readily shed those commitments. It is important to remember that those who care have traditionally been thought to have many cares.

(9) Such approaches can be found in Walzer (1983); Sandel (1982); MacIntyre (1981 and 1984); Williams (1985), and, perhaps most surprisingly, Rawls (1985). For some discussion of the implications of these works for international justice see O'Neill (1988b).

(10) Walzer acknowledges that this means that he can ‘only begin to address the problems raised by mass poverty in many parts of the globe’ (1983: 30). Critics may think that his approach in fact pre‐empts answers to questions of global justice.

(11) Communitarians can, however, take lesser loyalties seriously: where a state is divided into distinct national and ethical communities, those distinct traditions may in fact be the widest boundaries within which issues of justice can be debated and determined. They could argue for secession from a multinational state; but they cannot say anything about what goes on beyond the boundaries of ‘our’ community. Cf. Walzer (1983: 319).

(12) Rawls (1985).

(13) Rawls (1970: 11–12).

(14) Rawls (1970: 128).

(15) Cf. Pateman (1988); Nicholson (1987).

(16) See Okin (1987: 46–7). She considers whether the original position abstracts from knowledge of one's sex. Even if she is right in thinking that Rawls relies on a covertly gendered account of the subject, this idealization may have little effect on his theory of justice if the thought experiment of the original position has so relentlessly suppressed difference that the supposed plurality of voices is a fiction. In that case we should read the work as taking an idealized rather than a merely abstract view of rational choice from the very start, and as appealing to a single ideally informed and dispassionate figure as the generator of the principles of justice.

(17) See Beitz (1979) for an account of debates between realists and idealists.

(18) Keohane and Nye (1970); Luper‐Foy (1988).

(19) See Shue (1980, 1984); Alston and Tomasevski (1984); Brown and Shue (1977); Gewirth (1982); Luper‐Foy (1988); O'Neill (1986).

(20) This is the hoary problem of formalism in Kantian ethics. For recent discussions of aspects of the problem see Bittner (1974); Höffe (1977); O'Neill (1989: Part II).

(21) It does not follow that every coercive act is unjust; some coercion, e.g. the use of sanctions to enforce law, may be the condition of any reliable space for uncoerced action. In such cases the appropriate expression of an underlying principle of rejecting coercion is, suprisingly, and crucially for political argument, one that, taken out of context, might express an underlying principle of coercion.

(22) I have put these matters briefly. For more extended treatment see the references given in n. 20 and O'Neill (1988a).

(23) Edelman (1984).

(24) Nicholson (1987).

(25) Stiehm (1983); Scott (1986); Sen (1987).

(26) Postow (1978–9); Hill (1979); Pfeffer (1985); Sen (1987).

(27) I focus here on the obligations of the strong rather than the rights of the weak. This is not to deny that agitation and resistance by the weak can help remind and persuade the strong of their obligations and make it more difficult for them to repudiate them. However, to focus primarily on rights falsifies the predicament of the weak, who are in no position to ensure that others meet their obligations.

(28) Shue (1984); Harriss (1987 and 1991).

(29) See Sen (1987) for a fuller account of entitlements. While I have chosen to stress vulnerabilities that must not be exploited, rather than capabilities that ought to be secured, I believe this account of justice to the powerless is fully compatible with Sen's.