Jump to ContentJump to Main Navigation
Women, Culture, and DevelopmentA Study of Human Capabilities$

Martha C. Nussbaum and Jonathan Glover

Print publication date: 1995

Print ISBN-13: 9780198289647

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/0198289642.001.0001

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: 22 June 2021

Gender, Caste, and Law

Gender, Caste, and Law

(p.332) Gender, Caste, and Law
Women, Culture, and Development

Cass R. Sunstein

Oxford University Press

Abstract and Keywords

Sunstein explores the issue of law and its impact on women's quality of life, scrutinizing how law sustains and supports discrimination against women and how it might embody a commitment to sex equality. Sunstein proposes what he coins an ‘anticaste principle’ which forbids law from turning a morally irrelevant characteristic such as sex into a systematic source of social disadvantage. Focusing on the situation in which women's sexual and reproductive capacities are turned into objects for the use and control of others, Sunstein critiques American sex discrimination law, makes proposals for national and international legal change, and comments on the limitations of market mechanisms in ending discrimination.

Keywords:   American law, anticaste principle, antidiscrimination principle, market economy, reproduction, sex discrimination, sex equality, sexuality

The study of the law of sex discrimination requires an inquiry into law of two different kinds. The first area of law consists of legal practices that sustain and support sex discrimination. When there is sex discrimination, what contribution does law make? The second area consists of the law, constitutional or otherwise, that might operate against social practices or laws that produce sex discrimination. Suppose that we were committed to sex equality. How would we use the law to bring about this result? What form would the antidiscrimination principle take?

The questions will not receive the same answers in all nations. In developing countries, for example, the norms that produce sex discrimination are in significant part reflected in custom rather than law. Indeed, in developing countries the line between custom and law is extremely thin, even artificial. The norms that make sex discrimination possible have a large customary place in developed countries as well; but legal practices are far easier to identify and criticize as such.

Moreover, the use of law to eliminate sex discrimination creates special puzzles and confronts special obstacles in developing countries. Is there any tradition of judicial invalidation of laws that offend basic rights? Is there a tradition of judicial invalidation at all? Does some centralized authority have power to eliminate practices or law intruding on human rights? What role is reserved to courts, the Constitution, and laws purporting to protect such rights? Sometimes the answers to these questions will make it exceptionally hazardous to rely on Western‐style models, which consist of judicially enforced legal rights, brought to bear on legislative and administrative practices.

Notwithstanding these difficulties, I believe that it is indeed possible to sketch the kinds of legal practices that produce sex discrimination, and also to outline the kind of antidiscrimination norm that is entailed by a commitment to sex equality. I also believe that this norm has a fair claim to universality. It is clear that there are powerful commonalities, with respect to sex discrimination, in seemingly diverse systems, and those commonalities often have a great deal to do with law. And in describing and implementing a legal norm of sex equality, other nations would do well to learn from the conceptual mistakes of American law.

In this essay I identify a legal norm for opposing sex discrimination, which I call an anticaste principle. I will describe the anticaste principle in some detail below, but for the moment the principle should be taken to forbid social and legal practices from turning sex, a morally irrelevant characteristic, into a systemic source of social disadvantage. A pervasive (p.333) aspect of a caste system based on gender, uniting many seemingly diverse nations, is that women's sexual and reproductive capacities are turned into objects for the use and control of others.

I claim that the anticaste principle should be used as the basis for a legal assault on a number of social practices. These include above all the exclusion of women from places where political power is exercised; unequal access to education; subjection of women to public and private violence; and unequal access to nutrition and health care. The anticaste principle might well be made the basis for a movement in international human rights law, an area that is quite primitive on issues of sex equality.1

I do not argue that it is law that is responsible for most sex discrimination in the world, nor do I suggest that law can operate as a kind of deus ex machina, bringing justice where there is now oppression. But often law can accomplish a good deal. It can set forth aspirations. It can give voice and content to injuries and assaults that are indeed felt, sometimes very deeply, but that have been buried and hence rarely articulated. In this way it can energize people and make them feel that in important ways, they are not alone. It can produce real victories in the real world. Usually it cannot produce massive social change by itself; but it can bring about some improvements in people's lives.

This essay is divided into three parts. Part 1 briefly outlines the American law of sex equality. My purpose here is to describe an especially interesting body of law, and also to suggest the lessons to be learned from one nation's concrete experience with using law to address problems of sex discrimination. I draw particular attention to the crucial and troublesome issue of ‘difference’, an issue that has haunted the movement for sex equality through law in many nations. Women and men are ‘different’, it is frequently claimed, and the difference is said, quite plausibly, both to explain and to justify social and legal differences in the treatment of the sexes. I attempt to respond to the resulting conundrums.

Part 2 explains where American law has gone wrong, introduces the anticaste principle, and tries to explain why that principle is superior to plausible alternatives. In Part 2, I also discuss the idea that legal intervention is unacceptable because it involves an unacceptably paternalistic interference with women's current preferences. The point bears especially on efforts to impose international norms of human rights on diverse countries. I conclude with a brief discussion of the increasingly prominent view that the appropriate approach for law is to rely on free markets, which, it is said, will eliminate invidious sex discrimination and overcome gender caste.

Part 3 begins with some disclaimers about the limitations of law in creating and dismantling sex discrimination. It then applies the anticaste (p.334) principle to a number of disputed issues of sex equality. The applications consist of no more than a brief outline. My purpose here is not to set out a full programme for legal reform, but instead to offer some notations on the role of law in creating and maintaining gender caste, and to provide a few ideas about how law might serve as a corrective.

1 American Law

I will introduce the matter of gender difference and law through the lens of American constitutional law. It turns out that this seemingly parochial subject offers large and very general lessons. An understanding of American law should go at least some distance toward providing an understanding of some of the problems posed by sex equality law in other nations, even those that are or seem quite different from America. After discussing the experience of one nation, I will try to develop some principles for general use.

1.1. An Example

It will be useful to begin with Muller v. Oregon,2 one of the most important sex discrimination cases in all of American law. In the Muller case, the Supreme Court upheld a law limiting the number of hours that women could work per week or per day. The law did not apply to men. The law was attacked on the ground that it interfered with freedom of contract, an attack that drew strength from a previous decision in which the Court had struck down a sex‐neutral maximum hour law.3 Despite that precedent, the Court held that the sex‐based maximum hour law was permissible.

The Court's reasoning was simple. According to the Court, the law was justified because of ‘the difference between the sexes’. These differences included ‘woman's physical structure and the performance of maternal functions’. The Court emphasized empirical work apparently showing that a reduction of the working day was necessary in light of ‘(a) the physical organization of women, (b) her maternal functions, (c) the rearing and education of the children, (d) the maintenance of the home.’ The legally relevant differences between men and women therefore included not merely physical ones, but also a set of social roles unique to women. For the Court, a ‘difference justifies a difference in legislation’; and for the Court there is an ‘inherent difference’ between men and women.4

(p.335) The key question in the Muller case was therefore whether men and women were ‘different’. If they were, they could be treated differently. If they were not, they had to be treated the same. Because men and women were obviously different, different treatment did not offend the equality principle.

For the modern observer, there are several striking features in the Muller opinion. The first is the Court's assumption—supported in the case by an elaborate but highly anecdotal factual brief from Louis Brandeis—that the differences between men and women are sufficiently real and sufficiently large to justify a maximum hour law targeted at women alone. Perhaps these differences do not exist in sufficient scope to justify such differences in law. This is in fact the general answer offered by current constitutional law in America. The Supreme Court says that measures of the sort upheld in Muller reflect ‘overbroad stereotyes’, and these are unacceptable under the equal protection clause of the American Constitution (see Part 1, section 1. 2. below).

Notwithstanding this change, it is striking that current sex equality law, in America and generally elsewhere, asks the same question asked in Muller: are men and women the same? The difference in legal outcomes stems not from a different approach to the subject, but from the fact that men and women are usually thought to be the same. Women are no longer said to be ‘different’, at least not very often. It is for this reason that different treatment, through law, is generally unacceptable—a point to which I will return.

We might therefore conclude that Muller was wrong because it treated women as different when they are really the same. (The word ‘really’ disguises some complex issues of fact and value.) But the problems in Muller go much deeper than this. The Court treated the differences between men and women as ‘inherent’ when in fact some of these differences were a creation of social customs, and indeed in part of the legal system itself. Of course there are physical differences between the sexes, and the law is not responsible for all of these differences. Law cannot make men into women, or vice versa, at least not in the strict biological sense.5 But consider the rearing and education of children, or ‘maternal functions’, or the maintenance of the home. With respect to these, the Court attributed to ‘nature’ a set of tasks that are socially produced and in part a product of law.

Indeed, those tasks are in part a product of laws of the very sort at issue in Muller. Such laws help freeze women out of the workforce, by making women employees more costly or less remunerative to employers. This (p.336) form of discrimination will of course contribute to a division between the social roles of men and women. It will influence women to occupy the domestic sphere and encourage men to leave that sphere.

Thus far the problem with the approach in Muller is that the Court saw differences as inherent when in fact they were a product of society and law. But a separate and perhaps even more serious problem is that the Court treated the differences between the sexes as a sufficient justification for laws disadvantaging women. Recall here the Court's claim: ‘Difference justifies a difference in legislation.’ Let us suppose that men and women are different. Even if this is so, the decision to turn any difference into something with social consequences, or into a social disadvantage, is a legal and hence social one. That decision must be justified. By itself, the mere fact of difference is insufficient to justify disadvantage. Even if we are dealing with natural differences, the normative claims of nature are quite weak.6

The point is often overlooked. But there are innumerable differences among human beings—height, eye colour, strength, hormones, capacity to smell, gender—and those differences are made meaningful only through social and legal decisions. Such decisions turn differences into advantages and disadvantages. Sometimes, such decisions make differences into something that is interesting or even noticed. The translation of a difference into a disadvantage requires a reason; it is not, standing alone, a reason at all.7 A law that would impose a minimum wage for people with blue eyes could not plausibly be justified on the ground that such people are ‘different’ (even though they are).

Perhaps more fundamentally, differences between men and women may well be insufficient even to explain disadvantage, or to help us to understand how it came about. Sometimes lawyers, politicians, and others act as if there is a natural or tight connection between the ‘real’ differences between men and women and the different social roles of men and women.8 (p.337) But this is at best speculative. We lack anything like a good account of how sex differences relate to social roles and of the complex causal connections between the two.

In Muller, the Court relied on differences to justify inequality, as if gender differences came first and inequality second. In fact the opposite may be true.9 It is inequality, through social norms and law, that creates many of the relevant differences between the genders: the different social roles of men and women, with men being concentrated in the public sphere and women in the private sphere. It is inequality that helps make the differences into something socially and legally relevant.

I do not claim that there are no differences between men and women, either before or after society has acted. Surely there are important biological differences. What those differences must mean for society is not likely to be a question that any of us is in a position to answer. Nor have I shown that every social or legal difference between men and women is impermissible. But it should by now be clear that differences, even if real, do not justify laws that treat women unequally, or that turn women's differences into a social disadvantage.

1.2. Notes on American Sex Equality Law

Muller is no longer the law. Under the current understanding of the American Constitution, most legal distinctions between men and women are invalid. Indeed, many people think that there have been extra‐ordinary advances in sex equality law in America and the West in general. It is unquestionably remarkable to find that since the 1970s, the Supreme Court has invoked the Constitution to strike down a number of laws discriminating on the basis of sex. But how much has the Court contributed to sex equality in America? What might the Court have done differently? Here we will find some surprising answers. Indeed, we will find some modest but intriguing continuity between contemporary law and Muller.

In this section I offer a brief survey of American law, and then say something about the consequences of the law for sex equality.

1.2.1. A Brief Survey

A key early moment in American law came in Bradwell v. Illinois,10 in which the Court upheld Illinois' refusal to licence a woman to practise law. The revealing and much‐quoted concurring opinion by Justice Bradley reads in part as follows:

The natural and proper timidity and delicacy which belongs to the female sex evidently unfits it for many of the occupations of civil life. The constitution of the family organization, which is founded in the divine ordinance, as well as in the (p.338) nature of things, indicates the domestic sphere as that which properly belongs to the domain and functions of womanhood. The harmony, not to say identity, of interests and views which belong or should belong to the family institution, is repugnant to the idea of a woman adopting a distinct and independent career from that of her husband . . .

It is true that many women are unmarried and not affected by any of the duties, complications, and incapacities arising out of the married state but these are exceptions to the general rule. The paramount destiny and mission of woman are to fulfill the noble and benign offices of wife and mother. This is the law of the Creator. And the rules of civil society must be adapted to the general constitution of things, and cannot be based upon exceptional cases.

It is striking that some version of these views continues to capture the opinions of many people throughout the world and indeed in the United States itself.

The law continued in this vein for an extremely long time. It was not until 1971—nearly a century later—that sex discrimination through law came under serious constitutional challenge. In a seemingly minor case decided in that year, the Court struck down a statute giving a preference to men over women in establishing the hierarchy of persons entitled to administer the estate of a decedent who died intestate.11 The Court said that the state must show that ‘a difference in the sex of competing applicants’ bears a ‘rational relationship to a state objective’. It concluded that here the gender classification was ‘arbitrary’.

In this case, the Court indicated that the ordinary ‘rational relationship’ test would continue to apply to sex discrimination. This test is extremely deferential to the legislature. Under the ‘rational relationship’ test, the state must not be completely arbitrary; but it may usually adopt crude devices for categorizing people. The result of the ‘rational relationship’ test would be that in practice, almost all sex discrimination would be upheld. But two years later, in Frontiero v. Richardson,12 the Court indicated that legal rules discriminating on the basis of sex would henceforth be treated with great scepticism.

In the Frontiero case, the Court struck down a law allowing men in military service automatically to claim their spouses as dependents for purposes of receiving medical and other benefits, but requiring women to prove actual dependency. The case was important because the discriminatory law made the package of employment benefits larger for men than for women. Men automatically obtained protection of their spouses; women did not. In this way, the law was a direct successor of that in Muller, providing a comparatively greater incentive for men to enter the workforce. Laws of this kind make into a self‐fulfilling prophecy their assumption that men are more frequently workers than women.

(p.339) Four of the nine justices in Frontiero wrote that sex discrimination should be treated like race discrimination. This revolutionary step would mean that courts would give so‐called ‘strict scrutiny’ to all discrimination on the basis of sex. If ‘strict scrutiny’ were applied, government could rarely, if ever, draw lines between men and women. The opinion's important passages are worth quoting at length:

There can be no doubt that our Nation has had a long and unfortunate history of sex discrimination. Traditionally, such discrimination was rationalized by an attitude of romantic paternalism which, in practical effect, put women, not on a pedestal, but in a cage . . . As a result of notions such as these, our statute books gradually became laden with gross, stereotyped distictions between the sexes and, indeed, throughout much of the 19th century the position of women in our society was, in many respects, comparable to that of blacks under the pre‐Civil War slave codes. Neither slaves nor women could hold office, serve on juries, or bring suit in their own names, and married women traditionally were denied the legal capacity to hold or convey property or to serve as legal guardians of their own child . . .

It is true, of course, that the position of women in America has improved markedly in recent decades. Nevertheless, it can hardly be doubted that, in part because of the high visibility of the sex characteristic, women still face pervasive, although at times more subtle, discrimination in our educational institutions, in the job market and, perhaps most conspicuously, in the political arena.

Moreover, since sex, like race and national origin, is an immutable characteristic determined solely by the accident of birth, the imposition of special disabilities upon the members of a particular sex because of their sex would seem to violate ‘the basic concept of our system that legal burdens should bear some relationship to individual responsibility . . . The sex characteristic frequently bears no relation to ability to perform or contribute to society.

As noted, only four justices joined this opinion. Five of the nine justices refused to do so and hence ‘strict scrutiny’ has not been applied to sex discrimination. In the next key case, Craig v. Boren,13 the Court announced that laws distinguishing between men and women would not face strict scrutiny, but nonetheless ‘must serve important governmental objectives and must be substantially related to achievement of those objectives’. In principle, this standard is an intermediate one, somewhere between ‘strict scrutiny’ and ‘rational basis’ review. In practice, the standard has meant that the Court will almost always invalidate laws that contain explicit sex discrimination. This has been an extraordinary development and an enormous victory for the lawyers attempting to eliminate laws discriminating on the basis of sex.

Thus, for example, the Court has invalidated laws setting out different drinking ages for males and females;14 providing for all‐female nursing (p.340) schools;15 making a husband the ‘head and master’ of property jointly owned with his wife, and thus allowing him to dispose of the property without his wife's consent;16 requiring the consent of the mother, but not the father, for adoption of a child born out of wedlock;17 allowing payment of social security benefits to a widow under all circumstances, but to a widower only if he can prove that he was receiving at least one‐half of his support from his wife;18 and providing that husbands, and not wives, could be required to pay alimony on divorce.19

In a very few cases, the Court has upheld laws discriminating explicitly on the basis of sex. When it has done so, it has invoked ‘real differences’ between males and females. (Recall Muller.) Three cases are especially important in this regard. In Michael M. v. Sonoma County Superior Court,20 the Court upheld a law defining statutory rape as sexual intercourse with females under the age of 18 years. The Court acknowledged that the law treated males differently from females, but said that ‘young men and young women are not similarly situated with respect to the problems and the risks of sexual intercourse. Only women may become pregnant, and they suffer disproportionately the profound physical, emotional, and psychological consequences of sexual activity . . . A legislature acts well within its authority when it elects to punish only the participant who, by nature, suffers few of the consequences.’ Thus the Court emphasized that women were naturally deterred from engaging in sexual intercourse; that males were under no ‘similar natural sanctions’; and that the discriminatory law could therefore serve the interest of equality.

In Rostker v. Goldberg,21 the Court upheld a law requiring men, but not women, to register for the military draft. The Court reasoned that only men would be needed in combat, for which women were ineligible. ‘Men and women, because of the combat restrictions on women, are simply not similarly situated for purposes of a draft or registration for a draft.’ Discrimination was therefore acceptable. The combat restriction was not challenged in the case, and most people agree that the Court would have rejected any such challenge.

In Califano v. Webster,22 the Court upheld a law giving retired female workers higher monthly old‐age benefits than similarly situated retired male workers. Under the law, women could exclude more lower earning years than men, for purposes of developing the computation formula from which old‐age benefits would be calculated. The Court emphasized the compensatory purposes of the law, discriminating (that is, treating differently) (p.341) not to subordinate women, but to restore them to a position of equality. It said:

Reduction of the disparity in economic condition between men and women caused by the long history of discrimination against women has been recognized as . . . an important governmental objective . . . The more favorable treatment of the female wage earner enacted here was not a result of ‘archaic and overbroad generalizations’ about women or of ‘the role‐typing society has long imposed’ upon women . . . The challenged statute operates directly to compensate women for past economic discrimination.

Thus far I have discussed laws that discriminate explicitly on the basis of sex. What if a law has large discriminatory effects, but does not expressly treat women differently from men? What if a law harms women, but does not explicitly discriminate against them? In a key case, the Court concluded that such a law would be permissible so long as the law was minimally ‘rational’. Personnel Administrator v. Feeney23 involved a law providing an employment preference for veterans in state government. Because very few women were veterans, the law operated overwhelmingly to the advantage of men.

In upholding the law, the Court relied on the fact that some women were veterans, and many men were not. Since the law did not single out sex as a classifying factor, and merely had a discriminatory effect, it would be upheld unless it could be shown that it had been motivated by a discriminatory purpose. Discriminatory purpose, said the Court in the key passage, implies ‘more than intent as volition or intent as awareness of consequences. It implies that the decisionmaker selected or reaffirmed a particular course of action at least in part “because of”, not merely “in spite of” its adverse effects upon an identifiable group.’

In practice, this means that almost all laws that have a discriminatory effect on women will be upheld. It is exceptionally difficult to prove that a law was adopted ‘because of’ its adverse effects on women. It follows that employment practices that harm women—by, for example, imposing height and weight requirements, or refusing to adjust for childcare responsibilities—are fully acceptable.

This brief survey should be sufficient to provide a picture of current American law under the Constitution. Laws that discriminate explicitly on the basis of sex are seen as raising issues of sex discrimination. The governing equality principle forbids those laws in the vast majority of circumstances. A discriminatory staute is acceptable only if it responds to ‘real differences’ between men and women, or (what may be a subcategory of the same point) if it is plausibly compensatory for past discrimination.

(p.342) The category of ‘real differences’ is in turn read narrowly. It generally does not include differences in what we might call the social situation of men and women: the fact that women care disproportionately for children, are less likely to be wage workers, are more likely to be financially dependent on their spouses, are shorter and less heavy. Thus, for example, the fact—and it is a fact—that women are more likely to be financially dependent on men does not justify sex discrimination in the social security law. But laws that discriminate in their effects are almost always permissible.

We might describe contemporary American law most generally as a prohibition on unreasonable sexual differentiation. Women must be treated the same as men when they are the same as men. Since women and men generally are the same, explicitly discriminatory laws generally are invalid. Laws that are neutral, in the sense that they do not explicitly refer to sex, are acceptable, since they do not embody sexual differentiation at all.

I have emphasized the law of the American Constitution; but it is important to note that there are prohibitions on sex discrimination in ordinary, nonconstitutional law as well. The basic American civil rights law (enacted in 1964) forbids sex discrimination in employment, including hiring, firing, promotion, pay, and working conditions.24 It is worthwhile to observe that this provision—the most important legal prohibition on sex discrimination in the United States—was added to the bill not by its supporters, but by Southern Senators hoping to defeat the bill in its entirety. The Southern Senators believed that a prohibition on sex discrimination was so clearly ridiculous that its inclusion would doom the bill as a whole. Remarkably, the authors and advocates of the bill had originally excluded sex discrimination from the set of prohibitions, and it appears that they were singularly unenthusiatic about the idea of adding sex to the list of forbidden discriminations.

Nonetheless, the bill passed with the new addition, and it has been extremely important. Going beyond the Constitution, which applies only to government and which requires explicit discrimination or discriminatory purpose, the civil rights law applies to the private sector, and it requires employers to offer persuasive justifications for all measures that have a discriminatory effect on women. Height and weight requirements will therefore be invalidated unless they can be firmly justified. This basic civil rights law has also been held to forbid sexual harassment on the job.25

A separate legal provision forbids discrimination by institutions receiving federal funds, including colleges and universities. American law also requires equal pay for equal work,26 though equal is defined narrowly. Despite the urgings of some, American law does not guarantee equal pay for ‘comparable work’. Discrimination on the basis of pregnancy is also banned.27

(p.343) 1.2.2. A Note on Consequences

Many people think that American law, thus understood, reveals a highly developed understanding of the principle of sex equality. But what has American law actually accomplished for women?

The results are ambiguous. Professional women have probably been benefited a good deal, and it is possible that in employment, there have been very general improvements.28 No employer can reserve jobs to men, and this has in all likelihood made a major difference. The ban on sexual harassment has probably improved working conditions for thousands and perhaps millions of women. It may well have contributed to changes in male and female attitudes as well, imposing social stigma on a practice that is a continuing source of discrimination for women throughout the world.29

It is far from clear, however, that American women are very much closer to equality as a result of the apparently major changes in American law.30 American women earn about 65 cents for every dollar earned by American men, and there is no evidence that antidiscrimination law has had an important effect on this disparity.31 It is hard to connect the legal decisions to general improvements in the labour market status of women. There is no good evidence that these decisions have materially improved women's income or women's access to jobs.32

Several areas of special importance seem unaffected by the changes in law. For example, a serious problem for American women consists of what happens after divorce. A California study showed that men can expect their standard of living to increase by 42 per cent, whereas women can expect theirs to fall by 73 per cent.33 This is typical of a pattern experienced internationally, in which a high financial cost is exacted from women who leave their marriages. The high costs of ‘exit’ have a large range of effects during marriage, with respect to the allocation of power within the marriage and with respect to the practise of physical abuse and battery. These effects are a product of family law, not of nature.

The recent change in antidiscrimination law has not remedied this situation. In some cases, it may even have made things worse. If courts cannot treat women better than men in awarding alimony, they may neglect the sacrifices women have made for men's careers, and they may devalue the domestic contributions made by women. In the area of social insurance and welfare, the legal decisions require similarly situated men and women to be treated ‘the same’. But these decisions have accomplished very little in (p.344) improving women's lives. This is partly because with respect to poverty, women are simply not similarly situated to men.34

It is ironic but true that many of the key sex discrimination cases in America were brought by men, seeking to overturn protectionist laws that perhaps ‘stigmatized’ women in theory but may have benefited them in practice. If the law had sought to make a real difference, could it have done so? What would it have done?

2 The Anticaste Principle Defined

2.1. What's Wrong With American Law

I suggest that American equality law has gone wrong in two ways. First, it has misidentified the class of laws that raise issues of sex discrimination. Secondly, it has mischaracterized the relevant equality principle. The two points are related.

More concretely, American law sees sex discrimination always and only in laws that explicitly treat women differently from men. This understanding has caused two problems. First, a few laws that treat women differently from men are acceptable, and indeed promote the goal of equality, rightly understood. Califano v. Webster is an explicit and unusual reflection of this point. For example, the law of alimony and child custody might work in the direction of equality if it actually took sex into account. Sex‐neutral rules might well harm the cause of equality. As I have noted, if courts cannot take sex into account, there is pressure for them not to take into account the disproportionate contributions of women to domestic life. The result is less money at time of divorce and less sex equality.35

Secondly, and more important, some laws raise issues of sex discrimination even if they do not treat women differently from men, as that concept is conventionally understood.36 The reason is that existing social practices, and indeed men, are used as the baseline from which to decide whether women can make out a claim of inequality. In particular, women can be treated the same as men only insofar as they are the same as men. But sometimes women are not the same as men, and they should nonetheless be allowed to make an inequality claim.

Let me be more concrete. Suppose that the law forbids women from having an abortion, or excludes pregnancy from a disability programme. Under current American law, there is no issue of sex discrimination.37 Men cannot get pregnant; women and men are to that extent not similarly (p.345) situated. A law that restricts abortion or excludes pregnancy therefore raises no equality problem. But this is an odd way to think about the equality issue. If the law takes a characteristic limited to one group of citizens, and turns that characteristic into a systemic source of social disadvantage, surely there is a problem of equality.38 It would take only a minor extension of this point to suggest that laws that disproportionately burden women—consider a veterans' preference for employment—might well be subject to legal doubt, at least where there is no firm sex‐neutral justification for such laws.

We can connect this issue to the broader failure of American law to do as much as it might have about existing inequalities. Ironically, many of those existing inequalities are a product of contemporary law—a point that applies in numerous countries other than America. Consider a few areas. The criminal justice system treats men and women differently, so that women are disproportionately subject to criminal violence. The failure to provide adequate protection against rape, sexual harassment, and other forms of sexual assault and abuse might well be seen to raise equality issues. Indeed, the fact that rape within marriage is generally not a crime, and that domestic violence is often not the object of real police attention, might well be treated as an embodiment of sex discrimination.

Or return to the fact that after divorce, women's economic welfare goes sharply down, whereas men's goes sharply up. This is not the result of nature, but instead of legal rules that assure this result. The relevant rules might well be subject to legal attack. Or consider again the veterans' preferences laws, and more important the existence of a social security system that was designed for and that benefits male breadwinners, while helping women much less because they do not follow conventional male career paths.39

Now we are in a position to make some general observations about the question of sex ‘differences’. It is often said that women and men are different, and that the differences help to explain and to justify social and legal inequality. Indeed, differences are usually invoked as the justification for disadvantage. It is often said, for example, that women are different from men and that different treatment in social practice and in law is therefore perfectly appropriate. Here the reasoning of the old Muller Court continues to capture a good deal of current thinking about law, not simply in America, but throughout the world.

The claim will not do. As we have seen, the question for decision is not whether there is a difference—often there certainly is—but whether the legal and social treatment of that difference can be adequately justified. Differences need not imply inequality, and only some differences have that (p.346) implication. When differences do have that implication, it is a result of legal and social practices, not the result of differences alone. Since they are legal and social, these practices might be altered even if the differences remain.

An analogy may be helpful here. The problems faced by handicapped people are not a function of handicap ‘alone’ (an almost impenetrable idea—what would current handicaps even mean in a different world?), but instead of the interaction between physical and mental capacities on the one hand and a set of human obstacles made by and for the able‐bodied on the other. It is those obstacles, rather than the capacities taken as brute facts, that create a large part of what it means to be handicapped.

It would be implausible, for example, to defend the construction of a building with stairs, and without means of access for those in wheelchairs, on the ground that those who need wheelchairs are ‘different’. The question is whether it is acceptable, or just, to construct a building that excludes people who need an unusual means of entry. That question may not be a simple one, but it cannot be answered simply by pointing to a difference. The same is true for sex.

We can go further. Differences between men and women are often said to explain sex inequality, indeed to be the origin of inequality. But as the Muller case reveals, it might be equally right to think that differences are an outcome of inequality, or its product.40 Certainly some and perhaps many of the ‘real differences’ between men and women exist only because of sex inequality. Differences in physical strength, for example, undoubtedly have a good deal to do with differences in expectations, nutrition, and training. The nature and degree of difference between men and women is notoriously variable across time and space. The variations are sufficient to show that what we attribute to nature is often a social product.

Even differences in desires, preferences, aspirations, and values are in significant part a function of society and even law. The point suggests that it is wrong to base sex discrimination policy entirely on what women currently ‘want’,41 a subject taken up below. Many of the differences that are said to justify inequality are really a product of inequality.

(p.347) We can go even further. It is possible that many of the differences between men and women are noticed, or have anything like their current social meaning, only because of inequality. It is at least possible that the differences between men and women have such foundational status only because of the ways in which inequality and social practice make gender crucial. I do not claim that women are ‘the same’ as men, or that law should try to make them ‘the same’. I claim only that the differences are noticed and have consequences in significant part because of sexual inequality.

Some people go so far as to argue that there is nothing in the brute biological facts to establish that there are just two sexes.42 On this view, the biology of the matter could mean that there is one sex, or three, or five, or ten. As counterintuitive and even bizarre as this may seem, I think that it is not entirely implausible. The fact that men and women really are different—and this is indeed a fact—does not mean that the division of human beings into two, and only two, categories is compelled by biology. That division is social and sometimes legal.43

Whether or not this last point seems at all reasonable, I hope that I have said enough to suggest the enormous difficulties in the effort to approach the law of sex equality through the lens of ‘differences’. It is especially odd (p.348) to attribute social and legal practices involving gender to ‘nature’ or ‘divinity’. (Recall Justice Bradwell.) If the law is to do something about sex inequality, it should look elsewhere.

2.2. An Alternative: The Antidiscrimination Principle as an Anticaste Principle

There are striking and significant international commonalities in sex inequality as between men and women.44 If we take these commonalities as a whole, we might well describe them as amounting to the creation of a system of caste based on gender. That system, like so many others, is then attributed to ‘nature’ and ‘natural differences’. A principal feature of the caste system consists in the translation of women's sexual and reproductive capacities into a source of second‐class citizenship. Those capacities are often made into objects for the use and control of others.

In these circumstances, the appropriate equality principle is not an outgrowth of the question of differences, but an opposition to caste. The legal objection should be understood as an effort to eliminate, in places large and small, the caste system rooted in gender. A law is therefore objectionable on grounds of sex equality if it contributes to a caste system in this way. The controlling principle, to be vindicated through law, is not that women must be treated ‘the same’ as men, but that women must not be second‐class citizens.

This principle might operate not only against laws, but also against social customs and practices that have law‐like effects for social organization. Instead of asking, ‘are women similarly situated to men, and if so have they been treated differently’, we should ask, ‘does the law or practice in question contribute to the maintenance of a caste based on gender?’

The concept of caste is by no means self‐defining, and I will have to be a bit tentative about it here. I do not suggest that the caste‐like features of all societies with sex inequality are the same. I do claim that the similarities are what make those features a reason for social and legal concern.

The motivating idea behind an anticaste principle is that without very good reasons, social and legal structures ought not to turn differences that are highly visible and irrelevant from the moral point of view into social disadvantages. They certainly should not be permitted to do so if the (p.349) disadvantage is systemic. A difference is morally irrelevant if it has no relationship to individual entitlement or desert. Sex is certainly a morally irrelevant characteristic in this sense.45 A systemic disadvantage is one that operates along standard and predictable lines in multiple important spheres of life, and that applies in realms that relate to basic participation as a citizen in a democracy. These realms include education, health care, freedom from private and public violence, wealth, political representation, and political influence. The anticaste principle suggests that with respect to basic human capabilities and functionings, one group ought not to be systematically below another.46

In the areas of sex discrimination, the problem is precisely this sort of systemic disadvantage. A social or biological difference has the effect of systematically subordinating the relevant group: not because of ‘nature’, but because of social and legal practices. It does so in multiple spheres and along multiple indices of social welfare: poverty, education, health, political power, employment, susceptibility to violence and crime, and so forth. That is the caste system to which the legal system should be attempting to respond.

The anticaste principle is grounded in a familiar conception of equality. It was set out long ago by John Stuart Mill: ‘The principle which regulates the existing social relations between the two sexes—the legal subordination of one sex to the other—is wrong in itself, and now one of the chief hindrances to human improvement; and . . . it ought to be replaced by a principle of perfect equality, admitting no power or privilege on one side, nor disability on the other.’47 As I understand it here, the anticaste principle is not ‘egalitarian’. It is perfectly comfortable with significant disparities in the resources of different citizens. Its target is far more narrow: the creation of second‐class citizenship, based on a highly visible and morally irrelevant characteristic. And while I cannot defend this claim here, I believe that the anticaste principle is universal in its scope.48 It applies even in areas in which it would (for example) collide with religious convictions and firmly entrenched traditions.

(p.350) Of course there are difficult issues of strategy, timing, and implementation. Some legal interventions may not be fruitful; they may even be counterproductive. Some may breed confusion and resentment. Others may be unintelligible. Still others may disrupt a society's basic organizing frameworks, in a way that does great harm and little good. Outsiders (and insiders too) will often know too little, and they must be exceptionally careful about introducing legal principles that do not cohere with cultural norms. Context will therefore matter a great deal. All this is important; but it is quite a different matter from the issue of principle that I am now discussing.

2.3. Women's Preferences?

I want to deal now with an influential claim, bearing directly on the anticaste principle, about the relationship between sex discrimination and law. The claim is that in many different nations, women are frequently content with the sexual status quo, and that legal efforts therefore represent an unacceptable form of paternalism. If women themselves are content, on what basis can the legal system intervene? Is not legal intervention an illegitimate interference with women's right to liberty or autonomy?

These questions raise some complex issues; I deal with them briefly here.49 The chief response is that the satisfaction of private preferences, whatever their content and origins, does not respond to a persuasive conception of liberty or autonomy. The notion of autonomy should refer instead to decisions reached with a full and vivid awareness of available opportunities, with all relevant information, or without illegitimate or excessive constraints on the process of preference formation. When there is inadequate information or opportunities, decisions and even preferences should be described as unfree or nonautonomous.

Private preferences often do adjust to limitations in current practices and opportunities. People may well adapt their conduct and even their desires to what is now available. Consider here the story of the fox and the sour grapes.50 The fox does not want the grapes because he considers them to be sour; but his belief to this effect is based on the fact that the grapes are unavailable. It is therefore hard to justify their unavailability by reference of his preferences. Mary Wollstonecraft's A Vindication of the Rights of Women51 applies this basic idea to the area of discrimination on the basis of sex. The book can well be seen as an extended discussion of the social formation of preferences and the phenomenon of the adaptation of preferences, beliefs, and desires to an unjust status quo. Thus Wollstonecraft (p.351) writes: ‘I will venture to affirm, that a girl, whose spirits have not been damped by inactivity, or innocence tainted by false shame, will always be a romp, and the doll will never excite attention unless confinement allows her no alternative.’52 Mill makes the same points in his work on sex equality.53

Amartya Sen offers an especially vivid real‐world example from India. In 1944, the All‐India Institute of Hygiene and Public Health surveyed widows and widowers about their health. About 48.5 per cent of the widowers said that they were ‘ill’ or in ‘indifferent’ health, compared to 2.5 per cent of widows so describing their condition. In fact the widows were in worse condition than the widowers.54 In these circumstances it would seem odd to base health policy on subjectively held views about health conditions. Such an approach would ensure that existing discrimination would be severely aggravated.

When an adaptation of this sort is at work, respect for preferences is unjustified on grounds of autonomy and welfare as well. A social or legal system that has produced preferences, and done so by limiting opportunities unjustly, can hardly justify itself by reference to existing preferences.

There is suggestive evidence in the psychological literature to this effect, and the evidence bears directly on the claim that law should respect existing preferences with respect to sex discrimination. Most generally, the beliefs of both beneficiaries and victims of existing injustice are affected by efforts to reduce the cognitive dissonance produced by such injustice.55 The strategy of blaming the victim, or assuming that an injury or an inequality was deserved or inevitable, tends to permit nonvictims or members of advantaged groups to reduce dissonance by assuming that the world is just—a pervasive, insistent, and sometimes irrationally held belief.56

Victims also participate in dissonance‐reducing strategies, including the lowering of self‐esteem to accommodate both the fact of victimization and the belief that the world is essentially just. Sometimes it appears easier to assume that one's suffering is warranted than that it has been imposed cruelly or by mere chance.57 The phenomenon of blaming the victim also (p.352) reflects the ‘hindsight effect’, through which people unjustifiably perceive events as more predictable than they in fact were, and therefore suggest that victims or disadvantaged groups should have been able to prevent the negative outcome. All this bears directly on sex discrimination.

Other work in this vein reveals that people who engage in cruel behaviour change their attitudes toward the objects of their cruelty and thus devalue them. Observers tend to do the same.58 Such evidence relates to sex equality law in general.

Of course poverty itself is the most severe obstacle to the free development of preferences and beliefs. Programmes that attempt to respond to the multiple deprivations faced by poor people, of whom women are a disproportionately high number, might well be approached in this light. Indeed, poverty and similar forms of intense social disability can impair the formation of goals themselves, breeding instead a combination of frustration and resignation.59 Severe deprivation influences and even closes off the development of desires. An important reason to respond to the deprivation is to promote freer and better processes of desire formation.

One goal of a legal system, in short, is to ensure autonomy not merely by allowing satisfaction of preferences, but also and more fundamentally in the processes of preference formation. John Stuart Mill himself was emphatic on this point, going so far as to suggest that government itself should be evaluated in large measure by its effects on the character of the citizenry.60 The view that freedom requires an opportunity to choose among alternatives finds a natural supplement in the view that people should not face unjustifiable constraints on the free development of their preferences and beliefs.

A crucial feature of a system of caste based on gender consists of profound effects on the preferences and aspirations of women. The point is sufficient to show that existing preferences should be no obstacle, in principle, to an effort to eliminate a system of caste through law.

2.4. Markets?

With the extraordinary recent outburst of international enthusiasm for free markets, it should not be surprising to find the view that all invidious discrimination on the basis of race and sex will be eliminated by laissez‐faire.61 On this view, the appropriate approach for law is to eliminate (p.353) constraints on market ordering, and to rely solely on property rights and freedom of contract.

In many ways, free markets are indeed an ally of sex equality. Legal barriers to female employment are a form of government intervention in the market, and they have often been an effective and severe hindrance to equality. In a free market, women will do well to the extent that they are able to carry out the relevant tasks. Frequently women do perform as well as or better than men. Once discriminatory laws are eliminated, free markets may well accomplish a great deal in breaking down gender caste.

Suppose, for example, that an employer prefers to hire only men. Suppose that he believes that women belong in the home. In the end, he should be driven out of the market. An employer who discriminates may well fail. As experience has shown in many nations, competitive pressures have worked powerfully against sex discrimination.

As a complete solution, however, free markets will be inadequate. In this section I want to outline some of the reasons.62 There are four major problems.

The first problem is that in a market system, third parties might be able to impose costs on people who agree to treat women equally with men.63 Customers and others sometimes withdraw patronage and services from nondiscriminatory employers. A law firm that hires female lawyers might find itself punished in the market‐place. There are many parallels in developing countries, in which certain services are traditionally provided on a single‐sex basis. In these circumstances, market pressures do not check discrimination, but instead guarantee that it will continue.

Secondly, sex discrimination can be perhaps an ordinary market response to generalizations or stereotypes that, although overbroad and perhaps even invidious, provide an economically rational basis for market decisions. Stereotypes and generalizations are of course a common ingredient in market decisions. There are information costs in making distinctions within categories, and sometimes people make the category do the work of a more individualized and sometimes more costly examination into the merits of the particular employee. Such categorical judgements are not only pervasive, but usually legitimate. But categorical judgements might well dis‐serve the cause of equality.

In the area of sex discrimination, for example, an employer might discriminate against women not because he hates or devalues them, but because he has found from experience that women devote more time to childcare than do men, or that they are more likely to take leave for (p.354) domestic duties. This form of ‘statistical discrimination’64—judgements based on statistically reasonable stereotyping—can ensure that second‐class citizenship will persist for women, even or perhaps especially in free markets.

The third problem is that women, acting rationally in response to market signals, may fail to attempt to overcome their second‐class status, or fail to invest in ‘human capital’ (the economists' term for production of economically valued characteristics), simply because of the current social practices and the gender status quo. Suppose, for example, that there is current discrimination for any number of reasons—because employers prefer women, or because third parties impose pressures in discriminatory directions, or because employers engage in statistical discrimination. If so, there will be harmful effects on women's decisions about education or training, and indeed on their aspirations in general. As market participants, women might well invest less than men in training to be (say) doctors or technicians if these professions discriminate against women and thus reward their investment less than that of men.

Finally, markets will in some ways incorporate the norms and practices of men, at least if men are in a position to rule the society. Consider the many ways in which employment expectations and requirements are structured for traditional male career patterns. In these circumstances, a legal system committed to an anticaste principle might in some cases attempt to restructure market arrangements so as to put women on a plane of greater equality. This is so not by allowing women to be ‘like’ men, but sometimes by changing the criteria themselves, at least when those criteria do not have a firm independent justification. In American law, the prohibition on discrimination on the basis of pregnancy, and provisions for parental leave, are prominent examples.

These considerations suggest that while free markets can often help further the cause of gender equality, they are not a panacea. Supplemental legal controls are often necessary.

3 What the Law Might Do and Undo: Disclaimers, Examples

In this section, I offer some examples of how the law might create or oppose sex equality. 3.1. is a disclaimer, outlining some of the limits of law. 3.2. briefly discusses what I see as the most fundamental issues for a legal norm of sex equality. 3.3. offers some notes on questions of sexuality and reproduction.

(p.355) 3.1. The Limits of Law: A Cautionary Note

Western lawyers tend to think that much of the problem of sex discrimination lies in law. The distinctive history of sex inequality in the United States and elsewhere might well justify this view. When women have been subordinated, it has frequently been through law—not merely through old enactments depriving women of the power to vote, to hold property, and to enter into contracts, but also through current rules involving family law, welfare law, and the criminal justice system.65 Moreover, many of the attitudes and customs of Western men and women have plausibly been influenced by legal practices.

But much of sex discrimination lies elsewhere than in law, and this is conspicuously so in the many nations in which the line between law and custom is elusive. Often no formal law is responsible for the sexual subordination of women. Sometimes subordination is the result of custom.66 Sometimes subordination is produced by attitudes that are extremely deeply engrained, indeed that attribute existing norms and practices to nature or to divinity. Sometimes efforts to dislodge those attitudes will make no sense to people; sometimes such efforts will be close to unintelligible. Sometimes sexual subordination is partly the result of the physical power of men. In such cases, it is wrong to think that law, as Westerners understand that term, is an important causal agent in women's subordination.

Moreover, there are limits to what can be done through law, even with a properly defined equality principle. For example, most of the formerly communist countries had wonderful constitutions, with powerful and explicit guarantees of individual rights, including rights against sex discrimination. But these documents were worth less than the paper on which they were written. In practice, the relevant rights, including rights against sex discrimination, were not protected at all. The legal guarantees were, in James Madison's words, mere ‘parchment barriers’,67 of no use against public and private power.

Currently, we can find powerful legal guarantees of sex equality in such places as Albania, Czechoslovakia, Greece, Hungary, Portugal, Romania, China, and Spain.68 Do women have equality in all of those nations? Do they have equality in any of them? Certainly not.

(p.356) For legal guarantees to be effective, a number of conditions must obtain. It helps to have an independent judiciary, willing and able to enforce legal rights against public and private institutions. With or without an independent judiciary, it is important to ensure that government officials are genuinely respectful of legal guarantees. It helps to have public and private lawyers who are willing to take on cases of sex discrimination.

The point helps explain an anomaly in many Western nations. In such nations, there are usually powerful guarantees of freedom from discrimination on the basis of sex. But in many Western nations, women face frequent sex discrimination. Consider, for example, the prohibition on sexual harassment, a prohibition that is frequently violated in practice. Legal rights mean little, even with an independent judiciary, if people are without real access to lawyers or others who can make credible threats against violators.

All this suggests that the law can operate against sex discrimination only if the relevant nation already has in place many of the features of a legal culture. If those features do not exist, a legal right to be free from sex discrimination can mean very little. Perhaps international law can serve as a catalyst here.

3.2. Citizenship, Education, Physical Security, Health, Others

In this section I outline the possible consequences of an anticaste principle, operating through law. I will be extremely brief and cursory; the issues deserve much more attention than I can offer here. My goal is simply to give a sense of the legal and social practices against which an anticaste principle might operate.

The first task of such an anticaste principle is to ensure nondiscrimination with respect to the most fundamental matters. Among these, citizenship belongs first. No one is a citizen unless entitled to vote and to hold political office. A system in which women may not do these things runs afoul of the equality principle.

The right to equal education is closely connected with citizenship. To exercise the functions of a citizen, women need to be educated on equal terms with men. Violations of this principle are omnipresent. In no fewer than 76 poor countries, less than half of eligible girls are even enrolled in secondary schools.69 There is an especially large gap between men and women in terms of literacy. The gap is widening.

In Western nations, a core area for the law of sex discrimination has involved violations of the right to own property, the right to enter into contracts, and the right to sue and be sued. The last of these is often denominated ‘the right to protect one's rights’. The right to litigate is an indispensable political right; it is a natural corollary of the right to vote. The (p.357) rights to own property and to enter into contracts are necessary if women are to have basic civil capacity.

In many developing countries, current law continues to forbid women from owning land, and when property is held jointly, it is controlled by the husband. The law deprives women of equal rights over land in Brazil, Chile, Bolivia, South Africa, Zimbabwe, Libya, Nigeria, and Saudi Arabia. In most of these countries, women are also deprived of equal rights of inheritance.70 An anticaste principle should be directed against these restrictions.

The connection between physical security and citizenship is perhaps indirect. But in order to be citizens, people need a degree of independence and immunity from private and public violence. Otherwise they are entirely subject to the whim of others. Protection against sex‐based public and private violence is therefore a precondition for the status of citizenship. This is an important task in eliminating caste‐like features of current regimes. To dismantle a caste system through law, it is generally necessary to extend government protections against violence to groups formerly oppressed in this way.

For women, the particular problem is extremely severe. In Peru, 70 per cent of all crimes reported to the police involve women being beaten by their partners.71 In Japan, wife‐beating has been the second‐most frequent reason cited by women for divorce.72 About 31 per cent of all female murder victims in America in 1988 were killed by husbands and boyfriends. Women were more than twice as likely to be killed by husbands and boyfriends than were husbands and boyfriends to be murdered by wives and girlfriends.73 Battery of women by men is exceedingly common, with estimates ranging as high as four million women per year and including a large number of married women.74

An anticaste principle should be directed against this form of subjection. In some circumstances, at least, it is easy to see what form the principle would take as a matter of law. Rape within marriage should be criminalized. Sexual assault should be taken as seriously as other crimes. The failure to provide protection against violence, on the street or in the home, should be seen as a denial of equal protection of the laws. International human rights policy might well be harnessed to these ends.

Of the most fundamental legal rights entailed by the anticaste principle, equal access to adequate nutrition and medical care deserves an especially prominent place. A key element of second‐class status consists of group‐based subjection to inferior health. Nutrition and medical care are therefore central ingredients in the provision of sex equality. If women are deprived of adequate food and medical care, and men are not similarly deprived, the (p.358) anticaste principle has been violated. Any legal equality norm should be directed against this outcome.

In the same category are laws and practices that deprive women of equal employment opportunity. No law should exclude women from jobs open to men. If taken seriously, this principle would of course have large consequences in many nations.

3.3. Sexuality and Reproduction

I have said that a principal locus for sex discrimination consists of social and legal practices that turn women's sexual and reproductive capacities into objects for the control and use of third parties. If this is so, a number of issues treated as sex‐neutral under American and much of Western law should be approached not through the lens of ‘difference’, but under the anticaste principle. Laws and norms that treat women's sexual and reproductive capacities in this way often do not run afoul of the idea that the similarly situated must be treated similarly. Often, at least, some of the relevant practices appear sex‐neutral. Often men and women are not similarly situated.

There are many examples. These include access to contraception, sometimes foreclosed through law; legal restrictions on abortion; forced prostitution, which creates a kind of international traffic in women; female genital mutilation; marriage of young girls and teenagers, recognized in law and often against their will; sexual violence; sexual harassment; the making and distribution of pornography, especially that involving girls and young women; and perhaps surrogacy arrangements. In all of these areas, the principal problem is that women's sexual and reproductive capacities are made into commodities for the use and control of others.

A legal anticaste principle might well play a major role here. For example, sexual harassment could be recognized as a violation of equality under law. Forced prostitution could be forbidden. The age of marriage could be raised to 18, and the prohibition on marriages below that age could be taken seriously by the relevant officials. Access to contraception could be seen as a legal right operating as part of the principle of sex equality.75

I have only started to sketch these points here.76 But the elimination (p.359) of legal and social controls on women's sexuality and reproductive capacities should be a central part of the effort to combat gender caste through law.

4 Conclusion

Women are second‐class citizens throughout the world. For all the differences among nations, there are striking commonalities as well. Much of this is a product of law. Exclusion from political participation; inequality in the ownership of land and in rights of inheritance; restrictions on contraception and abortion—all these, and many more, are legal controls. I have suggested that at a minimum, a legal norm of sex equality should counteract legal contributions to the creation and maintenance of a caste system based on gender. Even when law is not the source of caste, a legal commitment to gender equality, deployed against discriminatory social customs, could do a considerable amount of good.77 This commitment should, I suggest, operate through the international law of human rights, as it currently does not.

It could be important to put a principle of sex equality into the laws of developed and developing countries alike, preferably in the relevant constitutions. This is only a first step. Enforcement of any legal guarantee requires the existence of something like a legal culture, with rights, citizen awareness of rights, lawyers, and access to independent tribunals. But social norms and legal practices should not be seen as altogether distinct spheres. A movement for sex equality could help energize the law; and an effort to inculcate legal principles of sex equality could have pervasive social effects as well.


(1) See, for example, the United Nations Convention on the Elimination of All Forms of Discrimination Against Women (1980), reprinted in Senate Executive Documents, 96th Cong., 2nd Sess. 9–19 (1980).

(2) 208 US 412 (1908).

(3) Lochner v. New York, 198 US 45 (1905).

(4) Ibid. at 421, 420 n. 1, 422–3. Compare this description of attitudes in prerevolutionary America: ‘So distinctive and so separated was the aristocracy from ordinary folk that many still thought the two groups represented two orders of being . . . Ordinary people were thought to be different physically, and because of varying diets and living conditions, no doubt in many cases they were different. People often assumed that a handsome child, though apparently a commoner, had to be some gentleman's bastard offspring.’ Gordon Wood, The Radicalism of the American Revolution (New York: Knopf, 1991), 27.

(5) Though this issue is more complex than it appears. See T. Laqueur, Making Sex (Cambridge, Mass.: Harvard University Press, 1990); J. Butler, Gender Trouble (New York: Routledge, 1990).

(6) See, for the classic discussion, John Stuart Mill, Nature, in J. M. Robson ed., The Collected Works of John Stuart Mill (Toronto: University of Toronto Press, 1967): ‘If the artificial is not better than the natural, to what end are all the arts of life? To dig, to plough, to build, to wear clothes, are direct infringements on the injunction to follow nature . . . All praise of Civilization, or Art, or Contrivance, is so much dispraise of Nature; an admission of imperfection, which it is man's business, and merit, to be always endeavoring to correct or mitigate . . . In sober truth, nearly all the things which men are hanged or imprisoned for doing to one another, are nature's every day performances . . . It remains true that nearly every respectable attribute of humanity is the result not of instinct, but of a victory of instinct; and that there is hardly anything valuable in the natural man except capacities—a whole world of possibilities, all of them dependent upon eminently artificial discipline for being realized . . . The duty of man is the same in respect to his own nature as in respect to the nature of all other things, namely not to follow but to amend it . . . Conformity to nature, has no connection whatever with right and wrong . . . That a thing is unnatural, in any precise meaning which can be attached to the word, is no argument for its being blamable.’

(7) See C. MacKinnon, Feminism Unmodified (Cambridge, Mass.: Harvard University Press, 1987), ch. 2.

(8) See C. MacKinnon, Sexual Harassment of Working Women (Cambridge, Mass.: Harvard University Press, 1979), ch. 2.

(9) An argument of this kind is made ibid.

(10) 83 US 130 (1873).

(11) Reed v. Reed, 404 US 71 (1971).

(12) 411 US 677 (1973).

(13) 429 US 190 (1976).

(14) Ibid.

(15) Mississippi University for Women v. Hogan, 458 US 718 (1982).

(16) Kirchberg v. Feenstra, 450 US 455 (1981).

(17) 441 US 380 (1979).

(18) Califano v. Goldfarb, 430 US 199 (1977).

(19) Orr v. Orr, 440 US 268 (1979).

(20) 450 US 464 (1981).

(21) 453 US 57 (1981).

(22) 430 US 313 (1977).

(23) 442 US 256 (1977).

(24) 42 USC 2000e et seq.

(25) Meritor Savings Bank v. Vinson, 477 US 57 (1986).

(26) 29 USC 206(d).

(27) 42 USC 2000e(k).

(28) Evidence of benefits is offered in Donohue, ‘Prohibiting Sex Discrimination in the Workplace: An Economic Perspective’, University of Chicago Law Review 54 (1989), 628.

(29) C. Mackinnon, Sexual Harassment of Working Women (1979), is the classic discussion.

(30) See Becker, ‘Prince Charming: Abstract Equality’, Supreme Court Review (1987), 201.

(31) Becker, ‘Politics, Differences and Economic Rights’, University of Chicago Legal Forum (1989), 169, 172.

(32) See G. Rosenberg, The Hollow Hope (Chicago, Ill.: University of Chicago Press, 1991).

(33) See L. J. Weitzman, The Divorce Revolution (New York, The Free Press, 1985), 338–9.

(34) Becker, ‘Politics, Differences and Economic Rights’, supra note 31: 176–8.

(35) The point is made nicely in Becker, supra note 30, and Becker, supra note 31.

(36) The reference to convention is necessary because by itself, the ‘similarly situated’ test is purely formal, and could be used to reach a very wide variety of results.

(37) See Geduldig v. Aiello, 417 US 484 (1974).

(38) The point is further developed in Sunstein, The Partial Constitution (Cambridge, Mass.: Harvard University Press, 1993).

(39) See Becker, supra note 31: 176–8.

(40) See C. MacKinnon, supra note 7.

(41) M. Wollstonecraft, ‘A Vindication of the Rights of Women’, in C. Poston (ed.) (New York: Norton, 1975; originally published in 1792); J. S. Mill, ‘The Subjection of Women’, reprinted in J. S. Mill, On Liberty and Other Essays (J. Gray, ed., New York: Oxford University Press, 1991), 471. See also Sen, ‘Rational Fools’, Philosophy and Public Affairs, 6 (1977), 317; Nussbaum, ‘Shame, Separateness, and Political Unity’, in Essays on Aristotle's Ethics (A. Rorty ed., 1988), 395; Sunstein, ‘Preferences and Politics’, Philosophy and Public Affairs, 20 (1991), 3. Consider as well Tocqueville's observations: ‘Should I call it a blessing of God, or a last malediction of his anger, this disposition of the soul that makes men insensible to extreme misery and often gives them a sort of depraved taste for the cause of their afflictions? Plunged in this abyss of wretchedness, the Negro hardly notices his ill fortune; he was reduced to slavery by violence, and the habit of servitude has given him the thoughts and ambitions of a slave; he admires his tyrants even more than he hates them and finds his joy and pride in servile imitation of his oppressors.’ Alexis de Tocqueville, Democracy in America (New York: Arlington House, 1987), 317. On the related phenomenon of cognitive dissonance, see Leon Festinger, A Theory of Cognitive Dissonance (Stanford, Calif.: Stanford University Press, 1957); on some of its implications for social theory, welfare, and autonomy, see Jon Elster, Sour Grapes (Cambridge: Cambridge University Press, 1983).

(42) See sources cited in supra note 5.

(43) There are important and revealing connections between discrimination on the basis of sex and discrimination on the basis of sexual orientation. I believe that the two forms of discrimination are closely related, indeed in some ways the same thing; they are both connected with maintenance of a caste system based on gender. This is not the place for a discussion of the point, but the treatment in text of the issue of ‘differences’ seems to raise it, and I therefore offer a few tentative speculations along these lines.

Consider the social and legal ban on same‐sex marriages, a ban that can be found in one or another form in almost all nations. The ban on same‐sex marriages is not thought to raise a problem of sex inequality in most legal systems. But might the legal ban (and the social taboo) not be crucially a product of a desire to maintain a system of gender hierarchy, a system that same‐sex marriages tend to undermine by complicating traditional and still‐influential ideas about ‘natural difference’ between men and women? There could well be differences here between the reasons for stigmatizing/outlawing male homosexual relations on the one hand and female homosexual relations on the other. Might not the ban on male homosexual relations be an effort to insist on and to rigidify ‘natural difference’, in part by ensuring firm and clear lines, defined in terms of gender, about sexual (and social) activity as opposed to sexual (and social) receptivity or passivity? Might not the ban on lesbian relations be at least in part an effort to ensure that women are sexual available to men? I speculate that considerations of this sort help to maintain the legal and social taboo on homosexuality, in a way that might well be damaging to both men and women, and to both heterosexual and homosexual alike, though of course in very different ways and to quite different degrees.

If these points seem exotic, we might think about a close analogy: legal and social bans on racial intermarriage. It is certainly not exotic to insist that such bans are typically (though not always) an effort to maintain a system of racial caste. For example, the American legal system has come to see such bans as an effort to maintain white supremacy, or racial caste, by keeping racial lines firm and distinct. This is so even though bans on racial marriage are formally equal: whites and blacks are treated ‘the same’ by such bans. Bans on same‐sex marriages might similarly seem to treat men and women ‘the same’, and thus to involve discrimination on the basis of sexual orientation rather than discrimination on the basis of sex. But here is my hypothesis: in terms of their purposes and effects, bans on same‐sex marriage have very much the same connection to gender caste as bans on racial intermarriage have to racial caste. Related points are discussed in Law, ‘Homosexuality and the Social Meaning of Gender’, Wisconsin Law Review (1988), 187; Rich, ‘Compulsory Heterosexuality and Lesbian Existence’, in Blood, Bread and Poetry (New York: Norton, 1986), 23.

(44) See Okin, in this volume.

(45) I realize that there are some complexities here. My claim is that social advantage or disadvantage cannot plausibly be based on the sex characteristic, even if that characteristic might sometimes be relevant to social roles.

(46) On capabilities and functionings, see Sen, in this volume; Nussbaum, in this volume. I am adding to these discussions a suggestion that in the context of gender, the problem lies in the particular fact that one group is systematically below another along the relevant dimensions, and an explanation of how this situation is produced by social practices and law.

(47) See J. S. Mill, ‘The Subjection of Women’, supra note 41. Mill's comments on his own argument remain highly relevant: ‘And there are so many causes tending to make the feelings connected with this subject the most intense and deeply‐rooted of all those which gather round and protect old institutions and customs, that we need not wonder to find them as yet undermined and loosed less than any of the rest by the progress of the great modern spiritual and social transition; nor suppose that the barbarisms to which men cling longest must be less barbarisms than those which they earlier shake off.’

(48) See Nussbaum, in this volume, on this point.

(49) See also Sen, in this volume; sources cited in supra note 41.

(50) Jon Elster, Sour Grapes (1983), which is an extended argument on the point; see also Sunstein, The Partial Constitution, ch. 6.

(51) M. Wollstonecraft, A Vindication of the Rights of Women, ed. C. Poston (New York: Norton, 1975) (originally published in 1792).

(52) Ibid.

(53) See Mill, The Subjection of Women, supra note 41, writing against the claim that the existing desires of women are a product of consent.

(54) See Amartya Sen, Commodities and Capabilities (Amsterdam: North‐Holland, 1985), 82.

(55) See infra note 56. Consider also the discussion of women's illiteracy in Bangladesh in Nussbaum, ‘Aristotelian Social Democracy’. Drawing on Martha Chen, A Quiet Revolution: Women in Transition in Rural Bangladesh (Cambridge, Mass.: Schenkman, 1983), Nussbaum explores the fact that many women in Bangladesh did not demand or even want greater education or literacy, and indeed expressed satisfaction with their current educational status. Of course desires of this sort were a product of a lack of available opportunities and of social and cultural pressures.

(56) See M. Lerner, The Belief in a Just World: A Fundamental Delusion (New York: Plenum Press, 1980).

(57) Consider here the astonishing fact that after a draft lottery, those with both favourable and unfavourable results decided that the outcomes of the purely random process were deserved. Rubin and Pepau, ‘Belief in a Just World and Reaction to Another's Lot’, Journal of Social Issues, 29 (1973), 73–93.

(58) See M. Lerner, supra note 56.

(59) See Eckstein, ‘Rationality and Frustration in Political Behavior’, in K. Monroe (ed.), The Economic Approach to Politics (New York: Harper Collins, 1991).

(60) See Mill, Considerations on Representative Government (New York: Liberal Arts Press, 1861).

(61) See, e.g., R. Epstein, Forbidden Grounds (Cambridge, Mass.: Harvard University Press, 1992).

(62) A more detailed discussion, with many citations, can be found in Sunstein, ‘Why Markets Don't Stop Discrimination’, Social Philosophy & Policy, 8 (1990), 22.

(63) See Akerloff, ‘The Economics of the Caste System’, in G. Akerloff, An Economist's Book of Tales (Cambridge: Cambridge University Press, 1979).

(64) See Phelps, ‘The Statistical Theory of Racism and Sexism’, American Economic Review, 62 (1972), 659.

(65) Some of this is well discussed in Becker, supra note 31.

(66) I will not deal here with the complex issue of what counts as law and what counts as custom. In many developing countries, the line between the two is extremely thin. I follow convention in understanding law, roughly, as the product of the will of some sovereign constituted with law‐making power, and custom as norms or practices that do not have this sort of authority. The best discussion of this is H. L. A. Hart, The Concept of Law (Oxford: Oxford University Press, 1961).

(67) The Federalist, No. 37.

(68) See J. Seager and A. Olson, Women in the World: An International Atlas (New York: Simon & Schuster, 1986), 12.

(69) See J. Seager and A. Olson, Women in the World: An International Atlas (New York: Simon & Schuster, 1986), 12.

(70) Ibid.

(71) International Atlas, 102.

(72) Ibid.

(73) See G. Rosenberg, supra note 32, 214.

(74) Ibid. 215.

(75) See World Health Organization, Reproductive Health (Geneva, WHO, 1992): ‘Without fertility regulation, women's rights are mere words. A woman who has no control over her fertility cannot complete her education, cannot maintain gainful employment . . . and has very few real choices open to her’; K. Bhate et al., In Search of Our Bodies: A Feminist Look at Women, Health and Reproduction in India (Bombay: Skakti, 1987), 1: ‘Women's oppression is intimately connected with the control of women's reproductive potential.’

(76) Some of these problems are discussed in more detail in Sunstein, ‘Neutrality in Constitutional Law (With Special Reference to Pornography, Abortion, and Surrogacy)’, Columbia Law Review, 92 (1992), 1.

(77) Subject to the qualifications involving timing and strategy, noted above.