… the Regulation permits an AH to marry at the age of 23 if she has joined the service at the age of 19 which is by all standards a very sound and salutary provision. Apart from improving the health of the employee, it helps a good deal in the promotion and boosting up of our family planning programme. Secondly, if a woman marries near about the age of 20 to 23 years, she becomes fully mature and there is every chance of such a marriage proving a success, all things being equal. Thirdly … if the bar of marriage within four years of service is removed then the Corporation will have to incur huge expenditure in recruiting additional AHs either on a temporary or on ad-hoc basis to replace the working AHs if they conceive …32
Later in the judgement, while the Court refused to accept compulsory retirement on first pregnancy, as a ‘reasonable compromise’ it endorsed a proposal by Air India to amend the Regulation and replace ‘first pregnancy’ with ‘third pregnancy’, along with a series of requirements such as unpaid pregnancy leave and regular physical check-ups. This, the Court observed, was reasonable because:
[it] would be in the larger interest of the health of the AH concerned as also for the good upbringing of the children. Secondly … when the entire world is faced with the problem of population explosion it will not only be desirable but absolutely essential for every country to see that the family planning programme is not only whipped up but maintained at sufficient levels so as to meet the danger of overpopulation …33
Read together, these two paragraphs constitute the foundation of the stereotype approach to sex discrimination, which, in turn, is based on a social vision known as the ‘separate-spheres theory’. Broadly, the separate-spheres theory holds that because of certain natural differences between the sexes, while the public sphere is the appropriate arena of action for men, the private sphere—the home, the family, and domestic life—is the appropriate arena for women. In Nargeesh Meerza, the separate-spheres theory formed the backbone of the Court’s sex discrimination analysis: differential treatment between men and women was held not to be discriminatory, and was justified by invoking ‘family planning’, ‘successful marriage’, ‘upbringing of children’, and ‘control of population explosion’, each of which was deemed to be the specific responsibility of women.34 Hence, the structure of the Regulations, aimed at achieving this by disincentivizing women from early marriage and too many pregnancies, was held to be constitutional.35Nargeesh Meerza represents, therefore, the apogee of the formalist approach to Article 15(1), combining all the elements that we have discussed above.36
Furthermore, this reading of Article 15(1) has also spilt over into the interpretation of Article 15(3). In Yusuf Abdul Aziz v. State of Bombay,37 the constitutional validity of Section 497 of the Indian Penal Code was challenged. Under Section 497, in an extramarital affair, only the man could be held guilty of the offence of adultery. When this section was challenged as being discriminatory against men, the Supreme Court upheld it in a very short judgement, holding that since Article 15(3) authorized the State to make ‘any special provisions for women’,38 a criminal provision exempting women from liability for precisely the same act that a man could be punished for was nonetheless constitutional. However, this surely cannot be right: if Article 15(3) is to be read as providing a carte blanche to the State in this fashion, then the prohibition upon ‘sex discrimination’ under Article 15(1) becomes redundant (the provision might as well have read,‘The State shall not accord favourable treatment to men’). More importantly, however, what the Supreme Court failed to analyse was that Section 497—while ostensibly for the benefit of women—was nonetheless based upon gendered stereotypes: in this case, the stereotype that, when it comes to sexual relations, women are passive and devoid of agency, always the seduced, and never the seducers.39 In fact, three decades later, in another challenge to the adultery provisions, the Court articulated the stereotype and accepted it as a legitimate basis for the law: ‘it is commonly accepted that it is the man who is the seducer and not the woman’.40 Yusuf Abdul Aziz, therefore, highlighted the important insight that Articles 15(1) and 15(3) cannot be separated. Our approach to understanding what constitutes ‘discrimination on grounds only of sex’ will also structure our approach towards what ‘special provisions’ the State can make, ostensibly for the benefit of women.
B. The Transformative Reading
In 1954, just three years after the judgement in B.B. Sen, the High Court of Allahabad was faced with a similar situation. The Uttar Pradesh Courts of Wards Act allowed the local government to declare that persons suffering from certain specific conditions, such as physical or mental ‘defects’, conviction for non-bailable offences, failure to discharge debts, and so on, were ‘incapable of managing their property’. The Act also, however, allowed the Government to declare any woman incapable of managing her property, whether or not any of the above conditions existed. In Rani Raj Rajeshwari Devi v. State of UP,41 the Act was defended on the basis that, as in B.B. Sen, the legislative classification was not drawn solely along the lines of sex, but along the lines of sex and property. And like Yusuf Abdul Aziz (which was decided the same year), it was argued that the provision was a ‘benign’ one, designed to protect women from unscrupulous predators of their property (‘… women generally are not such competent managers of property as men and are much more liable to be led astray …’42).
The Allahabad High Court, however, was having none of it. It swiftly held that whatever the manner in which the legislative classification had been drawn, the fact remained that it treated (otherwise) identically placed men and women differently. In other words, if the legislation’s effect was to treat men and women differently, its form (i.e., the precise character of the classification) or reason was irrelevant. Equally brusque was the Court’s rejection of the Government’s plea that because women were more liable to be led astray, they could be treated as ‘a class by themselves’ for the purpose of drawing legislative distinctions:
… evasion of the Constitution can[not] be permitted merely by calling an act classification and not discrimination … A classification which the Constitution forbids cannot possibly be said to be reasonable.43
This observation is crucial. The Allahabad High Court effectively rejected the argument that legislative differentiation founded upon ‘natural differences’ was not discrimination but only ‘classification’. On the contrary, in the teeth of Article 15(1), it was not open to the State to treat men and women as separate ‘classes’ by themselves by attributing certain stereotypical or generalized characteristics to all men or all women (competent management of property, physical weakness, maternal functions, etc.)
The Allahabad High Court’s analysis was followed by the High Court of Orissa in 1969. The Orissa Civil Service Rules allowed the state government to disqualify married women from employment if the ‘efficiency of the service’ required it.44 Men were placed under no similar disqualification. The justification for this was provided in the Indian Administrative Rules, which stated that ‘marriage brings about certain disabilities and obligations which may affect the efficiency or suitability for employment’.45 Once again, the State defended its law by arguing that the legislative classification was not based ‘only’ on grounds of sex, but that it had a ‘reasonable nexus in relation to … the maintenance of the efficiency of the service’.46 The role played by ‘sex plus physical weakness’ in R.S. Singh and ‘sex plus obligations of motherhood’ in Nargesh Meerza was now played by ‘sex plus efficiency’. The Orissa High Court was as swift in rejecting this line of reasoning as the Allahabad High Court had been, observing that:
… marriage does not operate as a disqualification for appointment … in the case of men, whereas in the case of married women, by Rule 6(2), they are being excluded from appointment. Such a disqualification being thus based on sex is unconstitutional.47
Both the Allahabad and the Orissa High Courts intuitively grasped that what was being defended as ‘sex plus another ground’ was little more than ste
reotype-based justification of laws that discriminated against women. However, neither of those cases examined the stereotypes in any detail. A move towards this was first made by the Delhi High Court, in a case where the claim of sex discrimination was brought by a man. In Walter Alfred Baid v. Union of India,48 Lady Irwin Hospital’s Recruitment Rules permitted only women to be appointed as senior nursing tutors. Striking this down, the High Court held that:
… it is difficult to accept the position that a discrimination based on sex is nevertheless not a discrimination based on sex ‘alone’ because it is based on ‘other considerations’ even though these other considerations have their genesis in the sex itself … while it is true that there are patent physical disparities between the two sexes, yet it is not possible to justify a conclusion … that all women or all men, as the case may be, would be unfit … for a particular class of work.49
Taking one step further along the road, the Delhi High Court correctly identified the fact that the State’s ‘sex plus’ justification meant nothing more than assumptions about what sex entailed. The Court made it clear that existing physical differences between men and women could not be made legislatively salient in this manner. In other words, the State could not take differences between men and women that existed in some, or even most cases, and frame laws as if that difference was what defined the sexes. To put it another way: while in the Sen-Singh line of cases, the Courts had (implicitly) held that the word ‘sex’ under Article 15(1) carried with it an entire constellation of physical and social characteristics that attached themselves to the sexes, the Delhi High Court rejected any reading of ‘sex’ that went beyond the biological.
This emerging jurisprudence was consolidated by the Kerala High Court soon afterwards. In A.N. Rajamma v. State of Kerala,50 women were denied appointment in the Last Grade Service by invoking ‘the arduous and special nature’51 of the duties (including late office hours and touring). Invalidating the disqualification on the basis that what mattered was the effect of the law, here, for the first time, a court explicitly identified what had been at play in the Sen-Singh line of cases and, drawing from American jurisprudence, gave it a name—‘romantic paternalism’:
[Under romantic paternalism] the social thinking and the approach to the question of the role women had to play in society was … tradition bound.52
As opposed to the formalist approach to Article 15(1) and sex equality, I characterize this as the transformative approach. The hallmarks of the transformative approach are: first, that it focuses not on the particular manner in which the legislative classification is drawn (its object or its form), but upon its impact, or effects, upon men and women (or sub-classes of men and women); second, it rejects any State justification that relies upon the trope of ‘natural differences’ between the genders;53 and third, as a corollary, it no longer allows the word ‘only’ to be used to limit the scope of Article 15(1) by granting safe harbour to laws that have not formally disadvantaged women, or have done so on the presumption of natural differences. And it was in Anuj Garg v. Hotel Association that this second judicial tradition found its culmination.54
Recall that in Anuj Garg, the State justified the ban on women’s employment in the liquor industry on the basis that it was a benign legislation (perhaps even a ‘special provision’ under Article 15[3]) for their own security. Rejecting this argument and observing that the burden of providing adequate security lay upon the State (and not upon women), the Court went on to note:
The impugned legislation suffers from incurable fixations of stereotype morality and conception of sexual role. The perspective thus arrived at is outmoded in content and stifling in means.55
In this way, the Court rejected the stereotype-based justification that would have upheld the law as either a non-discriminatory benign classification, or saved by Article 15(3). But in any event, the State’s reasoning could not be accepted, because:
Legislation should not be only assessed on its proposed aims but rather on the implications and the effects … No law in its ultimate effect should end up perpetuating the oppression of women.56
I want to focus closely on this, because both parts of the Court’s formulation are important. In the first sentence, the Court affirmed the effects-oriented test (as opposed to the formal, classification-based test) of discrimination under Article 15(1). But ‘effects’ upon what? In the second sentence, the Court answered this question by highlighting that it would scrutinize whether laws perpetuated the oppression of women. In other words, the Court endorsed a view of discrimination that understood it as something both ongoing (perpetuation) and institutional (oppression).
A significant consequence of adopting an institutional, effects-based view of discrimination is the impact of the role of the word ‘only’ in Article 15(1). Recall that previous judgements had invoked the word ‘only’ to deny discrimination claims in two contexts: first, where the State invoked reasons other than sex to justify a straightforward discriminatory classification that impacted women; and second, where the classification itself was drawn along the lines of ‘grounds’ other than sex, and therefore did not result in burdening only and all women, but only some women in an indirect manner. A straightforward example is pregnancy; but consider also, for example, a law that set educational barriers for contesting elections in a society where women are disproportionately denied access to education.57
A simple effects test, with nothing more, would be able to deal with the first kind of claim, because the reasons for State action would now drop out of the picture. However, only an institutional view of discrimination could deal with the second kind of claim, where legislative classification did not single out ‘women’ qua women for unfavourable treatment, but nonetheless, in a broader institutional context, contributed to ‘perpetuating the oppression of women’.58 This is particularly important because, in modern times, it will be rare to find the State expressly invoking stereotypes to justify discriminatory laws. What will be far more common, as we shall see, is that legislative form, which will appear to be worded neutrally, will nonetheless be founded on stereotypical assumptions that have disadvantaged women and continue to do so at a structural and institutional level. The importance of Anuj Garg was to decisively shift the enquiry from the former to the latter.
Anuj Garg, therefore, brought together the two disparate strands of the transformative approach to Article 15(1) that had been present to varying and scattered degrees in previous decisions. Under Article 15(1), legislation would have to be tested on the basis of its systematic and institutional effects, and not on the basis of the goals or aims that the State sought to invoke; and secondly, in no circumstances could discriminatory legislation be justified by resorting to stereotypes about the roles, abilities, or capacities of the sexes. These two strands, when brought together, yielded a powerful jurisprudence of sex equality, which not only adopted the perspective of those whose actions, identities, and lives were actually impacted by the law, but also ruled the entire justificatory logic of ‘natural differences’ constitutionally invalid.
To understand what is at stake, let us briefly return to our discussion in Section I. The disagreement between the formal and the transformative approaches, it should now be clear, is a disagreement over the meanings of the four words at the heart of Article 15(1): ‘grounds’, ‘discriminate’, ‘only’, and ‘sex’. The formal approach reads ‘grounds’ to mean the bases, or objects, of legislation, and equates it with the manner in which the legislature has gone about drawing the line of classification along which benefits or burdens are distributed. The transformative approach, on the other hand, understands ‘grounds’ to refer to the five indices of discrimination set out under Article 15(1), and consequently assesses legislation for its impact, or effect, upon those indices (in our case, the index of sex).
Secondly, the formal approach holds that if legislation tracks supposed or assumed ‘natural differences’ between men and women, then it isn’t discriminatory at
all, but merely a ‘reasonable classification’. The transformative approach, on the other hand, holds that basing legislation on assumed ‘natural differences’ is precisely what it means to discriminate. In framing a guarantee against discrimination, the Constitution was ruling out the use of ‘natural differences’ as a factor in classification groups of individuals for the allocation of differential benefits and burdens.
Thirdly, the formal approach holds that the word ‘only’ in Article 15(1) exempts State action that can be justified by reasons other than sex, or legislative classifications that are not directly and frontally drawn along the lines of sex. On the other hand, the transformative reading, which focuses on discrimination as involving institutional patterns, does not accord any such limiting role to the word ‘only’.
The Transformative Constitution Page 5