Affirmative Action Under Article 15(3): Reassessing The Meaning of “Special Provisions” for Women


– Unnati Ghia[1]
[Note: This is an advanced version of the article that will be published in Volume 32(2)]


Article 15(3) of the Constitution of India is an affirmative action provision intended to remedy the disadvantage faced by women, yet its application has been predominantly based in paternalism. In interpreting the phrase “special provisions for women”, Indian courts have often reinforced oppression through an extensive reliance on gender stereotypes and constructed notions of protective discrimination. In light of the shortcomings of the existing approach, this paper illustrates the need for a reconceptualisation of Article 15(3) within a substantive equality framework. Through an exploration of legal theory on affirmative action, as well as the text and history of Article 15(3), this paper suggests an alternative three-part framework that redirects the application of the provision at specific manifestations of historical and structural disadvantage.


The Constitution of India, 1950 is an aspirational document which firmly acknowledges the promise of a just, equal and free society in the Preamble itself.[2] This aspiration is equally entrenched in Part III of the Constitution, in particular, in the Constitution’s “equality code”,[3] i.e., Articles 14, 15, and 16. The equality code aspires to a consummate relationship between three stakeholders – the State, the individual and society, by accommodating both fundamental rights and the exercise of sovereign power. In this sense, the equality code is a powerful legal tool in the arsenal of both the State and the individual.

In recent years, Indian courts have expanded on the scope of the equality code, particularly to assure rights to women, as well as sexual and gender minorities. The Supreme Court of India (‘SCI’) has interpreted the code towards ends such as the decriminalisation of queer sex and adultery, and the grant of permanent commissions to women in the defence forces. These developments may be a catalyst for future claims from gender minorities, but their impact travels beyond constitutional litigation. Notably, the recent judicial recognition of concepts such as substantive equality, indirect discrimination and structural disadvantage contributes to a corpus of constitutional principles which influence State policy, including affirmative action.[4]

This paper explores the judicial treatment of Article 15(3) on the touchstone of this new corpus of principles. The objective is to draw attention to the assumptions made by courts in justifying differential treatment for women under the Constitution. I attempt this exploration in two broad segments: first, by delving into the existing jurisprudence of Indian courts under Article 15(3) and explaining the gendered assumptions therein, and second, by suggesting an alternative three-part framework for the application of Article 15(3), that is drawn from legal theory as well as the text and history of the provision. Through this exploration, I seek to arrive at an interpretation of Article 15(3) that is in consonance with the principle of substantive equality.

Within Article 15, clause (1) contains a right to non-discrimination,[5] while clause (2) is targeted at specific forms of discrimination. Both clauses provide concrete rights of non-discrimination to individuals on specified grounds. Article 15(3) follows these provisions by stating that “[n]othing in this article shall prevent the State from making any special provisions for women and children”. The focus here shifts from the individual to the State, in terms of the invocation of the provision and its intended addressee. Article 15(4), (5) and (6) similarly define the scope of permissible State action with respect to socially and educationally backward classes, Scheduled Castes and Scheduled Tribes, and economically weaker sections. Therefore, from a plain reading of Article 15, it is clear that the provision is not limited to individual claims to equality. Rather, it possesses considerable potential for reform by political and governmental institutions as well.

The scope of this paper is limited to Article 15(3) alone. However, given the frame of Article 15, theoretical discussions on the purpose of affirmative action and its position vis-à-vis non-discrimination guarantees are also helpful.[6] Further, while Article 15(3) is concerned with special provisions for both women and children, this paper restricts itself to commenting on its application to “women”.

In the first section, I cull out the opinion of the Indian judiciary on the rationale and application of Article 15(3) and highlight the harmful stereotypes that are implicit in this approach. The second section then discusses the judicial treatment of affirmative action for women qua the non-discrimination guarantee in Article 15(1). In the third section, I conduct a conceptual inquiry into the meaning and purpose of affirmative action, as well as its role in the achievement of substantive gender equality. This section demonstrates why the current interpretations of Article 15(3) are inadequate. The fourth section then delves into the history of Article 15(3) in the Constituent Assembly and the colonial era opposition to special measures for women. Finally, the fifth section breaks down the text of the provision in order to propose a framework that is better suited to the goals of substantive equality, clarity and the aspiration of Article 15.

The Judicial Protection of Women under Article 15(3)

In this section, I am primarily concerned with what the courts perceive to be the rationale behind Article 15(3). Towards this end, I examine both the decisions of various High Courts as well as those of the SCI. The interpretation of Article 15(3) depends on the meaning of “special provisions”, and the purpose with which it is invoked by the State and justified by the court. Prima facie, one may argue that the clear purpose of clause (3) is simply to benefit women. However, in her book on non-discrimination, Kalpana Kannabiran argues that while special provisions were introduced to address gender inequality, the State utilises constructed ideas of femininity in exercising its discretion under these provisions.[7] Building on this observation, I argue that the “protective discrimination” under Article 15(3) often relies on two key gendered concepts— suitability and vulnerability.


Social norms determine behaviour by demanding compliance with unspoken/implicit notions of propriety and decorum. In such systems, deviations from the norm are deemed to be outliers and castigated as “unnatural” or “abnormal”. In the same vein, the claim of certain tasks, occupations or activities being more appropriate for or more natural to women is a gendered assertion meant to pigeon-hole women into socially acceptable roles. Even while such rhetoric may seem to recognise a woman’s abilities, it is often only to the extent that it maintains patriarchal control over her. For instance, a compliment that women are generally more efficient as secretaries may be both a positive recognition of abilities and simultaneously a disqualification from more assertive leadership roles. These prejudices have also permeated into the judicial analysis under Article 15(3).

For example, the Gujarat High Court has upheld recruitment rules that prefer female officers for the position of superintendents in charge of district shelters/homes under Article 15(3), on the grounds that these homes were exclusively for women and the duties were better suited to women.[8] In the same vein, the Delhi High Court, in Charan Singh v Union of India,[9] upheld the appointment of only women as Enquiry and Reservation clerks in the Railways under Article 15(3) on two grounds— first, that women were barred from field jobs, which was an exclusion to be compensated for.[10] Second, that women were more courteous and patient, and therefore more suitable for dealing with long queues of people.[11]

These decisions, while advantaging women in terms of access to opportunities, rely on traditional gender roles to buttress the outcome. They reinforce the idea that women are necessarily more polite, sensitive and submissive as a gender, and thus better suited to professions requiring these characteristics. Through such opinions, the judiciary implicitly entrenches problematic stereotypes and roles into law. This also impacts how both the State and litigants then view the application of special provisions under Article 15(3). This idea of “suitability” must therefore be scrutinised carefully by courts.


Another approach characterises women as helpless victims in need of rescue, essentially reducing them to objects without agency. This assumption of vulnerability has been codified into provisions such as Section 66(1)(b) of the Factories Act, 1948, which prohibits women from working beyond 7 p.m. These provisions exclude women from the workplace and reinforce traditional gender roles.[12]

This assumption of vulnerability formed the basis of the SCI’s decision in EMC Steel Ltd. v Union of India,[13] where the Court upheld Section 14-D of the Delhi Rent Control Act, 1958. Section 14-D gave widowed landladies the ability to recover immediate possession of premises. Here, the Court specifically noted that “a widow is undoubtedly a vulnerable person in our society and requires special protection”.[14] This observation is based on a stereotype of widows being needy and vulnerable.[15]

The Karnataka High Court employed similar logic in order to uphold a rule permitting only women to be recruited as heads of all-girls institutions.[16] However, it went further to note that classifications allowing only female doctors in maternity hospitals, or disallowing women from working beyond 8 p.m. in a telephone exchange, would be in conformity with the Constitution as well.[17] In these cases, the observations suggest that certain measures must be taken by the State to “protect and save women”. This appears to be one understanding of “special” under Article 15(3). The premise of this approach is that women are passive victims due to their gender alone.

The prejudicial notions of “vulnerability” are also seen in cases involving penal provisions concerning women.[18] Section 497 of the erstwhile Code of Criminal Procedure, 1898, which granted preferential treatment to women in non-bailable offences, was upheld by the Rajasthan High Court as a special provision in favour of women under Article 15(3).[19] Similarly, the Kerala High Court upheld Section 56 of the Code of Civil Procedure, 1908, which protects women from arrest and detention in the execution of a money decree.[20] Despite the appellants arguing that women could be arrested and detained under other provisions of the Code, the court still invoked Article 15(3) in this case as a “limited protection” for women.[21]

This thread of reasoning also permeates into family law and the differential rights of the sexes therein. The Kerala High Court has observed that the provisions conferring additional grounds for divorce on women (such as desertion and cruelty), in comparison to men, can be justified under Article 15(3) as a special provision for the benefit of women.[22] Similarly, a Full Bench of the Calcutta High Court observed that Section 11 of the Indian Divorce Act, 1869, which permits only a wife to sue her husband for adultery without joining the adulterer as a party is permissible as a favourable provision for the “softer sex”.[23] In another case, the Calcutta High Court has also upheld the validity of Section 36 of the Special Marriage Act, 1954, which provides for alimony for the wife alone in cases where she has insufficient independent income.[24] Here, the Court noted that even if this was discrimination under Article 15(1), it was protected as a special provision under clause (3).

An important caveat is necessary at this juncture. A degree of State intervention or “protection” may be necessary for a minority, in this case, women, due to the gendered realities in which we live. For instance, a rule prohibiting women from being summoned to the police station at night may possibly be a “special provision”. I do not argue that a rejection of stereotypical ideas of vulnerability must equate to a rejection of such provisions. However, the approach of the court in such cases must be to acknowledge the reality of systemic gendered violence privately and at the hands of the State, as opposed to basing the application of a “special provision” in the vulnerability/inherent weakness of women. This clarity is often absent in the judicial reasoning under Article 15(3), which is often focused on the characteristics of the oppressed community (here, women) instead of the structures that oppress them.

The logic of protectionism

Across these cases, even though there is a benefit per se being conferred on women through these measures (for instance, access to opportunity), the rationale behind the benefit remains gendered. In the conceptions of both suitability and vulnerability, the premise remains that women are inherently different from men (physically weaker or more submissive by nature) thereby warranting special protection under the law. Protective discrimination based on the inherent differences between the sexes is not restricted to the jurisprudence under Article 15(3) but extends to sex discrimination cases under Articles 15 and 16 as well.[25] Therefore, these conceptions of “protective discrimination” and “inherent differences” warrants a closer look.

Interestingly, a seminal case on protective discrimination in favour of women and the standards of scrutiny applicable to such measures does not even mention Article 15(3). In Anuj Garg v Hotel Assn. of India,[26] while adjudicating the validity of Section 30 of the Punjab Excise Act, 1914, which prohibited the employment of women in establishments serving liquor and intoxicants, the SCI criticised the use of measures that victimised women in the name of protection and security.[27] It went on to hold:

It is to be borne in mind that legislations with pronounced “protective discrimination” aims, such as this one, potentially serve as double-edged swords. Strict scrutiny test should be employed while assessing the implications of this variety of legislations. Legislation should not be only assessed on its proposed aims but rather on the implications and the effects. The impugned legislation suffers from incurable fixations of stereotype morality and conception of sexual role”.[28]

(emphasis supplied)

The SCI’s decision in Anuj Garg, though not adjudicated under Article 15(3), has two implications for the State actions that seek to “protect” women under the Constitution. First, the reasoning prohibits reliance on gender stereotypes in the justification of protective provisions for women, setting in motion a line of cases under Article 15 that seeks to dispel the treatment of women as victims with no agency.[29] Second, if this reasoning is applied to Article 15(3), it would subject the State’s justification to a higher standard of scrutiny, thereby increasing the difficulty in defending special measures for women, particularly in cases where a classification relies on any difference between the sexes.

An important question here is whether Article 15(3) is relevant to cases of stereotyping in the first place, given that in cases such as Anuj Garg, the SCI does not even consider it. In fact, in Girdhar Gopal v State,[30] the Madhya Pradesh High Court justified Section 354 of the Indian Penal Code, 1860 (‘IPC’), pertaining to an offence outraging the modesty of women, under Article 15(1) and not 15(3).[31] The failure to consider Article 15(3) may be attributable to several reasons, such as the unfamiliarity of the counsels or the presumption that Article 15(3) only applies to reservations. Even so, this inconsistency in application does not detract from Article 15(3)’s potential to justify stereotypes, nor the established jurisprudence under Article 15(3) that is reliant on stereotypical notions of suitability and vulnerability, as discussed above.

The risk of stereotypes being employed under the guise of security under Article 15(3) has now been explicitly noted by the SCI in Joseph Shine v Union of India,[32] which declared Section 497 of the IPC to be ultra vires the Constitution. In his opinion, Chandrachud J. observed that Article 15(3)’s protective discrimination cannot employ paternalistic notions of protection so as to limit women’s sexual agency.[33] In entrenching the narrative of women as victims who are seduced by men, there was no benefit being conferred by Section 497, even though it exempted women from being punished as abettors in cases of adultery.[34] Malhotra J. similarly observed that a penal provision perpetuating the oppression of women cannot be termed as “beneficial legislation” and protective discrimination.[35]

The SCI’s observations in Joseph Shine have significant implications for this existing body of law on Article 15(3),[36] particularly statutory provisions employing stereotypical reasoning that were heretofore declared valid or remain unchallenged, for reasons of security. For instance, Section 360 of the Code of Criminal Procedure, 1973 allows courts to release men under the age of 21 and women of any age on probation, if not convicted of an offence punishable with death or life imprisonment. While extending the benefit of early release to women of all ages, the Code equates them to young and presumably immature men thereby perpetuating the traditional assumption of women being the ‘weaker sex’. In light of the emerging jurisprudence in this field, this form of protective discrimination and the precedents under Article 15(3) may no longer be constitutionally sound.

From the above-discussed case law, it is also clear that Article 15(3) is still considered to be “discrimination” per se — in that it is protective as per the courts, as opposed to a prohibited form of discrimination. Given that there is no mention of ‘protective discrimination’ or any such equivalent phrase in the Constitution, the immediate concern is how courts exempt certain prohibited classifications from the force of the equality code, i.e., how have courts reached the conclusion that protective discrimination can be justified despite being acknowledged as ‘discrimination per se’? Arguably, this conclusion is based on a clear conflation between “protective discrimination” and “special provisions”, which in my opinion are distinct concepts.

Accordingly, I now examine the judicial opinions on the nature and scope of Article 15(3) in order to understand why it is considered to be a justified form of discrimination.

The Relation between Clauses (1) and (3) of Article 15

There are two categories of precedents pertaining to the nature of Article 15 qua its different components: first, that Article 15(3) is an exception to Article 15(1), and second, the nature and application of clause (3) itself.

Article 15(3) is an exception to Article 15(1)

The SCI and various High Courts have presumed Article 15(3) to be an exception to Article 15(1). For instance, in EMC Steel, the SCI simply states that clause (3) is an express exception to clauses (1) and (2) of Article 15, and then proceeds to the facts of the case.[37] Similarly, in Om Narain Agarwal v Nagar Palika, the SCI states that clause (3) of Article 15 is an exception not just to clauses (1) and (2) of Article 15, but also to Article 14.[38] These decisions present no clarity on the scope and nature of Article 15(3) because there is no scrutiny of the text and possible interpretations of the provision in the first place.

The clearest exposition of Article 15(3) was carried out by the SCI in Govt. of A.P. v P.B. Vijayakumar.[39] In order to enable better access to and representation of women in public employment, the government of Andhra Pradesh introduced Rule 22-A in their public service regime in 1984.[40] Rule 22-A gave a preference to women in direct recruitment when they were better or equally suited to male applicants, and reserved posts exclusively for them.[41] Speaking for the court, Manohar J. held that although Article 15(1) would prohibit classifications on the ground of sex alone:

the State, by virtue of Article 15(3), is permitted, despite Article 15(1), to make special provisions for women, thus clearly carving out a permissible departure from the rigours of Article 15(1)”.[42]

(emphasis supplied)

In opining that Article 15(3) is a departure from the rigours of Article 15(1), the SCI clearly seems to suggest it is an exception. However, in describing the relationship between Articles 15 and 16, the Court later notes that Articles 15(1) and 15(3) in fact “go together”,[43] creating ambiguity on the relation between the two.

The SCI then goes on to state that the power under Article 15(3) is wide enough to cover the entire range of State activity, including employment under Article 16(1). This reasoning seems to suggest that Article 15(3) is not an exception to simply Article 15(1) but also Article 16. The Court’s logic was that because Article 16 does not contain a provision concerning women, it cannot derogate from the State’s power under Article 15(3).[44]

According to Manohar J. in this case, the purpose of Article 15(3) is to “strengthen and improve the status of women”, in recognition of their socio-economic backwardness.[45] Creating job opportunities is one facet of this commitment. On this basis, the Court defines “special provisions” under Article 15(3) to mean as follows:

What then is meant by “any special provision for women” in Article 15(3)? This “special provision”, which the State may make to improve women’s participation in all activities under the supervision and control of the State can be in the form of either affirmative action or reservation ……… to the effect that a special provision contemplated by Article 15(4) like reservation of posts and appointments contemplated by Article 16(4), must be within reasonable limits. These limits of reservation have been broadly fixed at 50% at the maximum. The same reasoning would apply to Article 15(3) which is worded similarly.”[46]

 (emphasis supplied)

Rule 22-A was therefore upheld through a harmonious reading of Articles 15(1), 15(3) and 16.[47] This decision is important in that it explains how reservations for women would fit within the mandate of Article 15(3),[48] but differentiates Rule 22-A from a reservation because it does not create a separate category for women per se. In fact, the Court argues that both kinds of measures would be justifiable under Article 15(3).[49]

More importantly, the meaning of “special provisions” as defined by Manohar J. is concerned with creating access, i.e., improving the participation of women in various spheres. However, Manohar J. subsequently subjects this equality to a standard of reasonability. This standard is explained by the Court in the context of reservations but there is little clarity as to what would be a reasonable limit in the case of other measures under Article 15(3).For instance, in the case of maternity benefit laws, where must the State draw a line? Thus, the definition of “special provisions” remains to be articulated clearly.

In P.B. Vijayakumar, the Court also attempts a harmonious interpretation of Article 15 and 16, but only for the purposes of making Article 15(3) an exception both to Article 15(1) as well as Article 16.[50] In this manner, it expands on the “logic of exception”, i.e., the tendency of courts to presume clause (3) is an exception, without examining whether Article 15(3) can be reconciled with its surrounding provisions in any other manner. In doing so, it entrenches this reasoning into the jurisprudence on Article 15(3).[51]

The word ‘exception’ may not be expressly used by courts. In Madhu Kishwar v State of Bihar, Ramaswamy J. in his dissenting opinion refers to the relation between clauses (1) and (3) without using the term “exception” but implying it all the same. While dealing with the question of whether tribal women are entitled to parity with tribal men in intestate succession, Ramaswamy J. held that Article 15(1) prohibits gender discrimination but Article 15(3) “lifts that rigor” and allows the State to positively discriminate in favour of women to accord them parity.[52] It was further held that Article 15(3) relieves the State from the “bondages of Article 14 and 15(1)” to accord socio-economic equality to women.[53]

This opinion echoes the reasoning in Ramaswamy J.’s earlier concurring opinion in Thota Sesharathamma v Thota Manikyamma, where Section 14(1) of the Hindu Succession Act, 1956 was justified on an identical reading of Article 15(3).[54] The constitutional validity of Section 14(1) had, in fact, earlier been upheld by the SCI in Partap Singh v Union of India, where the SCI similarly observed that beneficial provisions for women under the Hindu Succession Act could not be challenged by Hindu males as hostile discrimination because Article 15(3) overrides Article 15(1).[55] Thus, the logic of exception may be articulated differently by different courts. However, the consensus appears to be that clause (3) exists despite the rest of Articles 15 and 16, and not along with them.

Another question which arises out of the logic of exception is which classifications become permissible under Article 15(3)? Classifications only on the basis of sex are prohibited under Article 15(1) but ostensibly can be saved if they are considered to be special provisions under clause (3).[56] For instance, in a case before the Allahabad High Court,[57] boys were restricted from appearing for intermediate examinations if they had been admitted into a girls’ college, based on the fact that there was separate recognition for girls’ colleges. Here, the         Court read Article 15(3) with Article 29(2) of the Constitution[58] to argue that a separate institution for girls is a permitted classification and would not amount to discrimination under Article 15(1).[59] The purpose of this classification was held to be to encourage more girls into education. Even so, there was no analysis by the Court on how restricting boys from appearing in the intermediate examinations achieved that purpose under Article 15(3).

My focus here is not whether such classifications are legitimate. My point is that a classification in violation of Article 15(1) cannot be saved under Article 15(3) for the sole reason that it provides a benefit to women simpliciter. As per Manohar J. in the P.B. Vijayakumar case, Article 15(3) is an affirmative action provision. The objective of such provisions, according to the SCI, is to remedy the oppression faced by women.[60] Given this purpose, a measure must bear a nexus to or operate to remedy some form of disadvantage or oppression faced by women, in order to fulfil the objective of Article 15(3). The above decisions presume that a benefit to women necessarily remedies the disadvantage they face in a field, without examining whether the benefit has any bearing on the disadvantage. This leaves open the possibility that a measure may pay lip service to equality while creating no impact on subordination.

Why have Indian courts had no occasion to examine whether a measure satisfies the objective of Article 15(3)? In my opinion, it is because courts now presume that it is an exception. This provides courts with an easy route to justify a classification on the basis of sex without delving into the contours of Article 15 as a whole. By labelling a measure as an exception to the general rule, the need to balance competing considerations and principles within Article 15 becomes redundant. Although the logic of exception is certainly the less tiresome approach, I will later argue that it does not achieve the guarantee of equality under Article 15.

The nature of Article 15(3)

The relationship between clauses (1) and (3) also raises a question as to the substance and application of clause (3) itself, primarily because the text affords no clarity on how and when it is to be applied.

The first confusion arises with respect to the construction of Article 15(3), i.e., whether the provision must be construed strictly or liberally. In Independent Thought v Union of India, the SCI held that clause (3) cannot be interpreted restrictively and must be given “full play”, particularly in legislation involving the girl child.[61] On the other hand, the SCI on other occasions has held that exceptions to a provision must be construed strictly in order to ensure that the main provision is not defeated.[62] One may certainly argue that Article 15(3) is an empowering provision and that it must not be interpreted as a traditional exception would. However, the SCI has not ventured into this question as yet, and hence, the construction remains unarticulated.

This confusion on construction also exists in how courts view the impact of clause (3). An interesting aspect of the SCI’s opinion in Partap Singh (with respect to beneficial provisions under the Hindu Succession Act) is that even though the court considers Article 15(3) to be an over-riding provision, it refers to it as a “benign constitutional provision”.[63] If Article 15(3) possess the potential to excuse invidious forms of discrimination under clause (1), as long as they benefit women, it can hardly be termed as benign.

The second cause of ambiguity exists with respect to the kinds of measures that may be justified under Article 15(3). For instance, in Neelam Rani v State of Punjab, the Punjab and Haryana High Court held that only horizontal reservations are in line with the constitutional mandate under Article 15(3).[64] In this case, the vertical reservation that was disallowed referred to the creation of separate posts for male and female teachers in an all-girls school, which was construed to mean reservation in favour of men.[65] However, the Court simultaneously stated that separate cadres for women or, for instance, separate schools were permissible,[66] but vertical reservations for women were not. From this decision, it is unclear how separate cadres constitute vertical reservations which are impermissible under Article 15(3).

Interestingly, the SCI has taken a different view in a similar factual scenario. The Supreme Court upheld the reservation of 50% of teaching posts for female candidates and 50% for men for appointment as Assistant Teachers in schools, under Article 15(3).[67] The rationale of the Court here was that given a large number of girls under the age of 10 were taught in primary schools, it was preferable that young girls were taught by women.[68] In this case, even though the effective result of the reservation meant that there were separate categories for men and women, the SCI did not classify the measure as a reservation in favour of men. These cases indicate a lack of consensus on the ambit and application of the provision.[69]

The final question here would be the extent to which clause (3) can derogate from the guarantee under Article 15(1). Given that it is considered to be an exception, can it completely defeat the guarantee under Article 15(1)? According to the oft-cited dictum of Chagla J. in Dattatraya v State of Bombay,[70] Article 15(3) is a proviso to Article 15(1), but a proviso cannot nullify the parent provision itself.[71] This reasoning was relied on in Samsher Singh v Punjab State, where a Full Bench of the Punjab and Haryana High Court noted that while specific instances under Article 15(3) may not be possible to enumerate, an unreasonable benefit or protection should not be given to women at the cost of men.[72] In order to support this reasoning, the Court similarly rejected the nullification of the main provision.[73]

There are two important observations to make here. First, the idea of an ‘unreasonable benefit’ echoes the reasoning of Manohar J. in P.B. Vijayakumar, where the SCI spoke of reasonable limits being placed on the powers under Articles 15 and 16. However, an undefined standard of reasonability places significant discretion in the hands of the judiciary in balancing competing rights and interests under Articles 15(1) and (3). There is also no express mention of reasonability in the text of clause (3), in contrast to other rights under Part III such as Article 19(2) which explicitly use the term “reasonable”. Therefore, this standard of reasonability may lead to further inconsistency in this body of law.

Moreover, there is little clarity on how to read Article 15(3) so as to ensure that it does not function to nullify protections under Article 15(1). The concern of nullification is important, because not only are Articles 15(1) and (2) important parts of the equality code, but they have also been declared to be parts of the basic structure of the Indian Constitution.[74] Hence, there must be a clear and defined legal framework within which essential entitlements under Articles 15 and 16 can be balanced against the objective of Article 15(3), to ensure that such entitlements are not rendered redundant.

From the above discussion, two conclusions can be drawn with respect to the doctrine on Article 15(3). First, the judiciary has extensively relied on stereotypical assumptions of suitability and vulnerability in its understanding of special provisions. Second, the relationship between clauses (1) and (3), and the scope of clause (3) remains insufficiently explored by the courts. These tendencies impact the force of clause (3), in that it becomes a sanctuary for paternal State action, and diminish the weight and value of the rest of the equality code.

Article 15 from the perspective of Affirmative Action theory

In analysing the doctrine on this point, theoretical works may offer new perspectives or meanings through which the nature and purpose of affirmative action provisions may be better understood. Hence, I draw from legal theory on affirmative action and equality, in order to suggest interpretations alternative to those put forth by the judiciary.

What is affirmative action?

What do judges mean when they say that Article 15(3) is an affirmative action measure? The response to this must begin with defining affirmative action.

Affirmative action pertains to measures directed at the redressal of “historic and lingering” deprivations of the right to equality.[75] How would such actions be defined? For the purposes of this paper, I rely on Tarunabh Khaitan’s definition of affirmative action as measures “designed to benefit any members of one or more protected group(s) qua such membership”.[76] There are three phrases/features of this definition as discussed by Khaitan that are relevant for this paper — “design”, “benefit” and “protected group”.

Khaitan suggests two aspects with respect to the requirement of “design”. First, these measures must be designed to or must have their primary objective as providing a benefit to a disadvantaged group. Second, the design must entail a reasonable prospect of the measure successfully providing a benefit, i.e., the costs or harms associated with the measure should not outweigh the advantage conferred by it.[77]

Khaitan’s idea of a “benefit” is an advantage conferred on relatively disadvantage groups, as opposed to the comparative benefit that would accrue when the State “levels down”, i.e., imposes burdens on privileged groups. The benefits in this sense may be distributive (conferring direct tangible or expressive benefits) or facilitative (increasing access, transparency or accommodating difference).[78]

Finally, affirmative action measures are directed at members of “protected groups”, which would mean that there has to be some identification or analysis of existing or historical relative group disadvantage, as opposed to a simple numerical minority.[79] This form of disadvantage arises from a ‘morally irrelevant but seemingly immutable’ characteristic, such as caste and gender, and operates at a societal level.[80] In this sense, affirmative action measures are often distinct from traditional welfare or distributive equality measures because they derive their force from a corrective justice ideal.[81]

This explains the contours of an affirmative action measure, but why is it necessary? Khaitan argues that anti-discrimination legislation, even if effectively implemented, is insufficient because it is concerned with specific acts of discrimination. Further, it depends on compliance by or claims of victims, i.e., the victim or institution must actively challenge discrimination. This may be inadequate in addressing the more substantial and pervasive inequalities, and at best, it may maintain status quo and prevent exacerbation.[82] Hence, positive measures to redress historical and structural disadvantage become necessary.

Khaitan’s argument on the design of affirmative action measures is important in the context of “protective discrimination” under Article 15(3). The reliance on and entrenchment of gender stereotypes is a significant harm in its impact on women, both in legal and social experience, because it reinforces patriarchal norms, while limiting the agency of women themselves.[83] If there is a net harm arising out of the measure, then the measure cannot be qualified or justified as affirmative action.[84] Therefore, there is no place for measures assuming the vulnerability of women from this perspective. On this basis, the application of Article 15(3) must necessarily involve some degree of analysis on the “design” of the measure, i.e., courts should acknowledge and weigh the benefits and the harms that accrue from a measure. This approach may also assist in identifying instances of paternalism and stereotyping on part of the State, which are often overlooked.

Affirmative action and equality

Having considered the definition of affirmative action, I now examine the purpose of such measures. Indian courts have held that the purpose of affirmative action under Articles 15 and 16 is to further the goals of substantive equality.[85] However, a question often raised in the context of such programmes is that even if these measures are necessary for the upliftment of a group, can preferential treatment for one group result in equality of all?

Catherine MacKinnon would argue that such questions are premised on the traditional Aristotelian approach of ‘treating like cases alike’ or ‘on par’, which is aimed at achieving sameness. This is formal equality. The legal protections under this form of equality assist largely in dilemmas of misclassification.[86] For instance, the test for a reasonable classification under Article 14 (an intelligible differentia between the groups and rational nexus to the object sought to be achieved) can only address issues of dissimilar treatment, but cannot factor in social hierarchies and structural disadvantage.[87]

MacKinnon puts forth two reasons why affirmative action programmes cannot be justified or founded under a formal equality approach. First, affirmative action treats “unlikes alike on the basis of their unlikeness”, i.e., it is directed at maintaining diversity, as opposed to formal equality’s emphasis on sameness.[88] Second, in the context of gender inequality, even when formal equality acknowledges difference, the solution therein is to superficially address the difference without addressing the underlying inequality. In such cases, differential treatment is based on the rhetoric of “inherent differences” between the sexes, which becomes a legitimate classification.[89] Within this framework, subordination is automatically justified as “equal treatment for equal differences”.[90]

This framework often relies on stereotypical ideas of gender to create these “inherent differences”. This in turn perpetuates the subordination of women by denying them access and opportunity on the grounds of these inherent differences. Therefore, protective discrimination that is based on this idea of “inherent differences” [for instance, that women are inherently weaker and need protection] is the result of grounding special provisions in a formal equality framework. This is because such provisions are aimed at parity, notwithstanding the underlying subordination.

In contrast to a formal equality approach, in B.K. Pavitra v Union of India,[91] the SCI has held that:

“For equality to be truly effective or substantive, the principle must recognise existing inequalities in society to overcome them. Reservations are thus not an exception to the rule of equality of opportunity. They are rather the true fulfilment of effective and substantive equality by accounting for the structural conditions into which people are born”.[92]

(emphasis supplied)

The SCI clearly defines another form of equality, that is concerned with the “equality of outcomes”. This is substantive equality. This alternative equality framework is premised on dismantling group disadvantage and hierarchy. [93] According to this framework, equality is not sameness, but rather the elimination of the systemic social subordination of groups, created by historical hierarchies.[94] It is under this approach to equality that affirmative action finds its place, because substantive equality permits the treatment of unlikes based on their unlikeness, in order to address subordination. Hence, the concept of “unlikeness” or differences between the sexes cannot be ignored as a whole, despite its problematic usage by formal equality proponents to perpetuate unequal status.

Ratna Kapur suggests that equality and differences can be reconciled in a substantive equality framework, by viewing gender differences from the perspective of historical disadvantage and inequality.[95] This is what a formal equality framework cannot do. Kapur argues that measures that are aimed at gender differences can be of three forms — sameness, protectionist and compensatory.[96] The first approach presumes women are equal to men and treats them the same. As already discussed, this is formal equality. The protectionist approach presumes women are weaker than men and need to be protected.[97] This mirrors the approach taken by the Indian judiciary; however, it strips women of their agency and autonomy. The compensatory approach accounts for the historical disadvantage faced by women, by treating the genders differently to correct existing inequality.[98] This is the only approach that addresses disadvantage and autonomy, and, in my opinion, is the approach that must be taken under Article 15(3).

How would subordination and disadvantage be identified under such a framework? MacKinnon suggests that this is a factual inquiry that depends on an examination of the historical context and the reality of social hierarchies.[99] The identification cannot depend on a single factor, because different forms of disadvantage may intersect with each other. Khaitan argues that the threshold for group disadvantage warranting concern under discrimination law should be high — the discrimination must be “substantial, abiding and pervasive”.[100] Such disadvantage occurs across multiple sectors and aspects of life, over a length of time, and manifests in forms that are more than a mere inconvenience.[101]

In fact, the SCI has defined affirmative action as measures directed at “the existing patterns of discrimination, disadvantage and disempowerment among the different sections of society”.[102] Therefore, in accordance with this definition, Article 15(3) cannot justifiably lend protection to measures that either are in ignorance of existing patterns of disadvantage or do not function to remedy such disadvantage. Hence, the identification of disadvantage is a key component to any analysis under Article 15(3).

This is also where the approach taken by the Indian courts is sorely lacking — there is no requirement to identify disadvantage. Nor is the State required to prove that the measure actually impacts or remedies such disadvantage, i.e., there is no requirement of proving a nexus between the measure and disadvantage. In essence, therefore, there is no test or framework for applying Article 15(3), thereby placing significant discretion in the hands of both the State and the judiciary.

Affirmative Action and Non-discrimination

As explained in Section II, one of the evident ambiguities in the decisions of Indian courts is the treatment of clause (3) of Article 15 as an exception to clause (1), wherein clause (3) often renders the right to non-discrimination otiose by “overriding” it.

In fact, the Indian judiciary is cognizant of this nullification of Article 15(1). For instance, Manohar J. in P.B. Vijayakumar attempts to introduce the standard of reasonability into Article 15(3), possibly to prevent the nullification of other guarantees. However, as already argued, reasonability as a standard places significant discretion with the judiciary. Further, the understanding of reasonability may itself be gendered.[103] Instead, I would argue that substantive equality would provide a more coherent framework under which nullification may be prevented. Importantly, substantive equality would not support an interpretation of Articles 15(1) and (3) that renders the operation of a right to non-discrimination entirely redundant. This is because discrimination, if permitted, functions to diminish an individual’s liberty and dignity, which is a value at the core of substantive equality.[104] Therefore, a substantive equality framework would necessarily attempt to balance measures to accommodate difference (such as affirmative action) with other equality goals (such as non-discrimination).[105]

Even so, can affirmative action be termed as discrimination against the members of a non-preferred group? This question is important, because it suggests an inherent tension between a general non-discrimination guarantee and affirmative action, leading to concerns about who is to be regarded as ‘disadvantaged’ in different contexts.

Renee Leon argues that although the impact of affirmative measures is often borne by individuals who were not responsible for the creation of disadvantage in the first place (in this case, men today), such classes benefit from the existence of the system that causes the disadvantage, and cannot be categorized as “victims” facing “disadvantage” per se.[106]

While that may be true in terms of lived experiences, the immediate legal concern remains that this is in fact differential treatment on the basis of a prohibited marker (sex), and may thus violate a non-discrimination guarantee.[107] In this sense, clauses (1) and (3) are two prima facie opposing values, as the practical outcomes of the application of each principle may be at odds.

In response, Leon argues that the premise of this concern is itself problematic, in that such measures are still considered to be discrimination in the first place.[108] This approach is not alien to discrimination law. For instance, Article 4.1 of the Convention on the Elimination of All Forms of Discrimination Against Women (CEDAW) 1979, states that “temporary special measures” aimed at de facto gender equality would not be considered as discrimination under the Convention. However, this may simply be a matter of semantics, because affirmative action measures would still prima facie satisfy a definition of sex discrimination. In response to this dilemma, Khaitan suggests that the forms of discrimination are different. Affirmative action may result in what Khaitan calls “collateral discrimination” against the dominant group, but may still not be a concern for discrimination law (outside of its expressive social consequences) because it is not against a group with relative disadvantage, i.e., a “protected group”.[109]

Given this, are Indian courts correct in ruling that affirmative action is a “special exception” to the general principles of non-discrimination? At the outset, I would disagree. Following Khaitan’s reasoning above, Article 15(1) and (3) address entirely different forms and manifestations of discrimination, hence clause (3) cannot be treated as an exception to any “general principle”.

Further, Leon criticises the concept of “special exceptions” in anti-discrimination laws, on two grounds. First, such exceptions function largely to provide equality of opportunity by removing specific instances of barriers to access, but this does not guarantee equality of outcomes or accommodate differences.[110] For instance, an exception reserving positions for female delegates in an assembly, in case of similarly placed candidates, removes the immediate barrier but does not account for the structural reasons why women may not stand in the first place. Second, the term “special” masks the legitimacy of these claims and is likely to be interpreted in a paternalistic fashion that diminishes the liberty of women, much in the way Indian courts have.[111] Ginsburg and Merritt, for instance, argue that “special provisions” for motherhood and childhood sequester women at home and maintain traditional gender roles because the understanding of “special” here is influenced by patriarchal stereotypes.[112] Therefore, treating affirmative action as an exception to the principle of equality is undesirable in theory and practice. This is why a reinterpretation of Article 15(3) becomes necessary.

However, in the context of affirmative action for women, legislation must tread a careful path in formulating provisions. Kimberlé Williams Crenshaw, best known for her work on intersectionality and equality, revises the often used ‘track metaphor’[113] to explain an ideal model of affirmative action. In an ideal race, the competitors begin from the same points, allowing for the best runner to reap the accolades. However, a few runners are placed at a head start, not because the runners themselves lack the capability to participate, but because their lanes possess more obstacles in comparison to the lanes of other runners. Crenshaw argues that from this perspective, affirmative action is not about preferences or unequal treatment but rather about providing a head start that neutralizes the existence of these obstacles, in order to make the race fair and equal. These obstacles may be structural disadvantage, bias, systemic exclusion, or other forms of discrimination that create unequal conditions of life.[114]

Crenshaw’s metaphor is useful here because the rationale of affirmative action measures is not to focus on the characteristics of who the measure is targeted at. The metaphor is directed at the lane, as opposed to the runner. An approach that focuses on the runner and her circumstances would in all likelihood disparage the individual or community itself through the reliance on prejudicial, socially constructed identities, thereby increasing resentment and stigmatization, because the programmes are rarely created by the minority themselves. Rather, affirmative action measures must be focused on the obstacle (i.e., historical and/or structural disadvantage) that the individual is faced with. This metaphor provides the basis for my argument — the objective of a special provision for women must be to counter a specific structural and systemic disadvantage and discrimination.

From these theoretical perspectives in sub-sections (A), (B), and (C), three observations are relevant to the application of Article 15(3): First, measures must be aimed at a specific disadvantage which must be identified by the courts in each case. Second, there must be a nexus between the measure and disadvantage, or the measure must remedy/attempt to remedy this disadvantage. Third, the impact of the measure must be weighed against the other goals of substantive equality.

The current interpretation of Article 15(3) may hence be reconstructed along these lines. However, the impulse for reinterpretation need not be derived from theory alone. Rather, this impulse is equally supported by arguments based in the history and text of the provision itself, to which I now turn my attention.

The History of “Special Provisions” for Women

Examining the context in which a provision was drafted may provide valuable guidance in interpreting the text, although it is by no means authoritative.[115] The purpose of a historical analysis is not to suggest that an originalist interpretation of Article 15(3) is necessarily the correct one. The discussion of historical sources here is not to determine the original intent of the drafters, which is near impossible to definitively ascertain, but rather to present cues towards original meaning, i.e., the meaning of the phrases at the moment of drafting.

Colonial-era antecedents of equality and their relevance today

One of the earliest recognitions of a right to equality in India was the Constitution of India Bill, 1895, which stated that the law shall be equal to all.[116] This was expanded on in the Commonwealth of India Bill, 1925, which was drafted by the National Convention, including by members of the Women’s Indian Association (WIA).[117] Article 7, titled ‘Fundamental Rights’, refers to “no disqualification or disability on the basis of sex”.[118] These documents subsequently played an influential role in the deliberations and report of the Nehru Committee, which was set up by the All Parties Conference in 1928 as a response to the Simon Commission.[119] The Committee Report subsequently acknowledged that men and women must have equal rights as citizens in its section on Fundamental Rights,[120] in addition to a general guarantee of equality.

In the aftermath of the Second Round Table Conference and the discourse on fundamental rights there, certain legal principles were incorporated into the Government of India Act, 1935.[121] One of the relevant provisions introduced pertained to non-discrimination on the grounds of sex in any appointment to the civil service, unless otherwise provided for by the British Government.[122]

Although organisations such as the WIA were represented within the Nehru Committee, it is important to recall the position of these organisations on special measures such as reservations for women. One of the earliest national platforms for women, the All India Women’s Conference (AIWC), created in 1927, in fact, opposed the concept of special provisions for women.[123]A resolution passed by the AIWC in 1936, for instance, explicitly states that reservations for women in assemblies are not approved of.[124] In the memorandums submitted to the First Round Table Conference by WIA and AIWC, they argued that women must be permitted to contest elections on equal terms with no reservations or nominations.[125]Unfortunately, reservations for women in Central and State legislative bodies remains a contested political issue even today.[126]

The AIWC’s notions of fairness and equality were centred on sameness or the similar treatment of the sexes (formal equality), as opposed to addressing systemic disadvantage. The motto of the AIWC was “equality and no privileges, a fair field and no favour”. Thus, differential treatment in favour of women constituted an admission of the inferiority and vulnerability of women according as per the AIWC.[127] There was in fact a split in the women’s movement on this issue, between women aligned with the Congress (through AIWC and WIA) and those outside the nationalist congregation seeking  upliftment and reservations.[128]

In this sense, the Constitution constitutes a repudiation of formal equality, in that it explicitly recognises the need for affirmative action for women. The prescience exhibited by the Constitution in recognising women as a disadvantaged category under Article 15(3), in and of themselves and independent of their other identities, is important. If women did not constitute an independent category in terms of the unique disadvantage they face, it is unclear why the need for special provisions for their advancement would be explicitly provided for within the Constitution itself.

Admittedly, it can also be validly argued that Article 15(3) was created out of paternalism as opposed to a prescient commitment to substantive equality. Does Article 15(3) therefore subscribe to the judiciary’s gendered paternalism as discussed in Section I, or substantive equality between the sexes? In order to further explore this question, I now turn to the more immediate constitutional history of the provision by examining its passage in the Constituent Assembly (‘CA’).

Affirmative action for women within the Constitution-making process

The CA’s Drafting Committee framed the language of “special provisions” in one of its earlier drafts from February 1948.[129] While the State was barred from creating any law that abridged fundamental rights, the draft provision stated that the State was not prevented from “making any law for the removal of any inequality, disparity or discrimination arising out of any existing law”.[130] This early drafting history is important because it indicates an awareness of the disadvantage faced by minorities and the role of the State in remedying such inequalities. This proviso was maintained and revised subsequently, and placed before the CA later that year.

The CA’s Sub-Committee on Fundamental Rights also discussed multiple drafts proposed by different members, prior to formulating a list of rights.[131] For instance, K.M. Munshi’s draft on the prohibition of non-discrimination stated that women were the equal of men in all spheres of political, economic, social and cultural life and entitled to the same rights and duties, except when otherwise provided by the Union.[132] On the other hand, Dr B.R. Ambedkar’s draft possessed the same principle of non-discrimination, but no specific principle on gender equality. In its report, the Sub-Committee ultimately formulated a provision that possessed traces of Articles 14, 15(1) and 15(2) of the final Constitution.[133] The Constitutional Adviser, BN Rau, observed in his explanatory note that the draft provision on non-discrimination followed the Nehru Report and the 1933 Congress declaration.[134] However, he also cautioned that a simple prohibition of discrimination as in the draft report would negatively impact the creation of separate schools and hospitals for women.[135]

When this clause was placed before the Advisory Committee, several members, including Rajkumari Amrit Kaur, argued that the principle of social equality required that in addition to a general prohibition of discrimination, a clause enabling the provision of special facilities for women and children was also required.[136] This is an explicit rejection of the position taken by the AIWC. The Committee subsequently then redrafted the clause to include a proviso allowing for “separate provisions” made for women and children.

However, the text of Article 15(3) as it stands was specifically introduced by BN Rau, presented in the Draft Constitution dated October 1947.[137] B. Shiva Rao observes that the Adviser found support for this language in his discussions with US Supreme Court Justice Felix Frankfurter, who suggested that legal provisions may be necessary for women, for instance to provide maternity relief. This clause was accepted by the Drafting Committee on October 30, 1947, and introduced as Article 9 of the Draft Constitution.[138]

These examples of maternity relief, or of separate schools and hospitals, indicate that the purpose of Article 15(3) was enabling State intervention where women were being materially disadvantaged in society. Such instances can be easily distinguished from the vague ideas of vulnerability and suitability that Indian courts have subsequently applied to Article 15(3). This is buttressed by the discussion of Article 9 in the CA Debates.

The Draft Constitution was formally introduced into the CA on  November 4, 1948,[139] while the discussions on Draft Article 9 took place on  November 29, 1948. Article 9(2) mirrored the text of Article 15(3) as it exists today. While proposing an amendment that provided the same benefit (i.e., special provisions) to Scheduled Castes and Tribes,[140] KT Shah argued that such classes face handicaps or disabilities due to their historical treatment in society, and, therefore, should be permitted special facilities in order to achieve the “real equality of citizens”.[141] He defined this equality as not simply of name or on paper but equality of fact.[142] This statement is indicative of some consciousness regarding the importance of substantive equality, even though it was not made in the context of the subjugation of women.[143]

KT Shah also noted that there are indeed special provisions that “exclude women from certain dangerous occupations”, intended to “safeguard, protect or lead to their betterment”.[144]This statement manifests the same paternalism that is evident in the judicial treatment of Article 15(3). While KT Shah’s statement alone is not indicative of the CA’s opinion, it admittedly lends credence to the judiciary’s position on Article 15(3). However, the CA’s response to this paternalism is not found in the deliberations on Draft Article 9, but interestingly in a related discussion on another right under Part III of the Constitution.

While discussing draft Article 18 (Article 24 of the 1950 Constitution), which pertains to a prohibition on the employment of children, an amendment was moved by Damodar S. Seth requesting a similar provision for women.[145] He argued that women should not be employed at night, in mines or industries detrimental to their health because such occupations are not suited to women’s “delicate health and position in society”.[146] This amendment was rejected in voting by the CA.

Admittedly, the reliance on CA debates has its limitations. The statements of a few members cannot provide a decisive interpretation of constitutional provisions, because they are not representative of the other coalitions in the assembly or the politics that resulted in the final outcome. Further, the CA itself was representative primarily of upper caste, upper class male views on the status of women.[147] However, these historical sources can still be successfully utilised to preclude certain interpretations that have no basis in constitutional history.[148]

From the discussions on draft Article 18, it is clear that the CA did not accept provisions excluding women from “dangerous” occupations in order to protect them. This discussion is relevant to historical arguments under Article 15(3) because it provides an instance of a “special provision” that the drafters explicitly rejected. Therefore, the Indian precedents which excluded women from certain jobs for their “protection and security” are in direct juxtaposition to the stand taken by the CA as a historical body, and the original meaning of Article 15(3) certainly cannot be said to encapsulate such protective discrimination.

These historical antecedents indicate that interpretations based in either the AIWC’s formal equality approach or paternalism, may not meet fulfil the purpose of Article 15(3). The drafting history in particular suggests an intent to address concrete instances of historical disadvantage faced by women. These antecedents present useful cues for the interpretation of clause (3) today which, I will now argue, are supported by the text of the clause as well.

Textual Analysis of Article 15(3)

A textual analysis focuses on the language of the provision itself alongside its contemporaneous meaning, excluding any collateral sources.[149] In this section, I attempt a textual analysis of Article 15(3) and argue that a principle for re-interpreting the provision from the perspective of substantive equality can be culled out from within the Constitution itself.

“Nothing in this article shall prevent the State”

Article 15 exists within Part III of the Constitution, under the broad category of the ‘Right to Equality’. The provision is concerned with non-discrimination, as evidenced by its title — “Prohibition of discrimination on grounds of religion, race, caste, sex or place of birth”. Clause (1) contains a right to non-discrimination in the form of a restriction on state action, while clause (2) is targeted at specific forms of socio-economic discrimination within or outside of state action.[150] A straightforward reading of the text indicates that Article 15(1) and (2) are similar in that they can be read to support two applications, i.e., as restrictions on State action and as a panacea for individuals fighting State discrimination. Clause (3) is constructed differently in that it is an enabling provision for State action and possesses no rights and entitlements per se.

Clause (3) reads “nothing in this article shall prevent the State from making any special provision for women and children”. The language utilised in the first part of the clause is ubiquitous across Articles 15 and 16 and has also been utilised by the drafters outside of the affirmative action provisions.

For instance, Article 23(2) relies on the same terminology in stating “nothing in this shall prevent the State from imposing compulsory service for public purposes” (emphasis supplied).[151] It is also important to note that these are not the only enabling provisions in Part III. Articles 33 and 22(7) are also enabling provisions that positively grant Parliament the ability to create specific laws in relation to those rights under Part III.[152] The differentiation in language indicates that the former set of provisions are not simply a positive conferment of power, but rather they tie the State’s enabling power to the function of the other clauses within the provision.

In the context of Article 15(3), this language has been interpreted to mean that it is an exception to clauses (1) and (2), which are understood as the general or main provisions. The primary issue with this reasoning is that the functions of clauses (1), (2) and (3) are different and they address different forms of discrimination. Here, it is important to recall Khaitan and Leon’s distinctions between discrimination simpliciter and affirmative action (or “collateral discrimination”). If so, then it is unclear why clause (3) is necessarily an exception, and, to what? Article 15(1) and (2) cannot be understood as the main provisions because they serve different purposes, as opposed to merely laying down a central principle.

The obvious response to this reasoning is that “nothing in this article shall prevent the State” can clearly “save” prohibited forms of classifications on the basis of sex under clauses (1) and (2).

This necessitates a closer look at the construction and interpretation of provisos. According to the principles of interpretation laid down by the SCI, a proviso may serve one of four functions — it may qualify or create an exception to the provision, it may insist on mandatory conditions to make the provision workable, it may be embedded in or an integral part of the enactment itself, or it may be an added optional addenda to explain the provision itself.[153] Moreover, provisos are not generally interpreted to subsume the provision and nullify the right granted by it.[154] Given these principles, the Indian judiciary’s conclusion that Article 15(3) is necessarily an absolute exception is not incontrovertible simply based on the text of the provision. Arguably, it could be interpreted to qualify or clarify the nature of Article 15(1) and (2), or to limit the scope of these provisions, as opposed to being a “departure from the rigours of Article 15(1)” as held by the SCI.

The language of clause (3) alternatively indicates that it is possibly a non-obstante clause as opposed to an exception proviso. The SCI has previously interpreted the phrase “nothing contained in the main provision shall” to mean that the proviso is a non-obstante clause.[155] One could argue that if Article 15(3) is interpreted to be a non-obstante clause, it would have an overriding effect over the other parts of Article 15. However, the SCI has also held that in case the enacting (main) provision is clear and capable of only one interpretation, then its scope and operation cannot be cut down; rather, in such cases, the non-obstante clause must be read as clarifying the whole provision.[156] The language of Article 15(1) and (2) are clear in their prohibitions and grounds, therefore, any valid interpretation of clause (3) must be clarificatory.

With respect to similarly worded provisions, the SCI has interpreted Articles 16(4) and 16(4-A) to be non-obstante clauses that confer a discretion on the State to create reservations.[157] Clearly then, the textual interpretation of clause (3) has differed substantially from other affirmative action provisions with the same language within Articles 15 and 16 as well, in that Articles 16(4) and (4A) have not been held to be exceptions.[158] These alternative interpretations may possess a textual key to reconcile the tension between Articles 15(1) and 15(3). For instance, in Gulshan Prakash v State of Haryana, the SCI held that Article 15(4) is not an exception to clause (1), but rather a “special application” of the principle of reasonable classification.[159] However, a principle for reconciling affirmative action with a right to non-discrimination and equality is best articulated in Indra Sawhney v Union of India.[160] Here, the majority held that Article 16(4) is not an exception to Article 16(1), but rather an “emphatic way of stating a principle implicit in clause (1)”.[161] Sahai J. further went on to argue that both clauses are part of the same scheme and represent different forms of equality.[162]In the same manner, Article 15(3) can be understood as another form of commitment to the same general principle of substantive equality, as in Article 15(1).

Therefore, the words “nothing in this article shall prevent” do not mean it is an exception which can justify violations of Article 15(1), but rather a non-obstante clause that clarifies the scope of Article 15(1) and limits its application in certain cases. This interpretation of Article 15 would ensure that special provisions are only permissible when they do not entirely nullify the principles enshrined in clause (1). For instance, special provisions for women that rely on stereotypes would then be prohibited because such provisions militate against the guarantee of non-discrimination and the anti-stereotyping principle articulated under Article 15(1).[163] This is important because it creates a framework wherein the court must conduct the enquiry in two stages: First, the identification of disadvantage, and, second, the implications of the measure must be weighed or balanced against the competing principles under Article 15(1). This enquiry mirrors the conclusions drawn from the theory on affirmative action in Section III of this paper. Therefore, a textual interpretation of clause (3) to some extent resolves the concerns around protective/ stereotypical discrimination raised in Section I of this article.

“Any special provision for women”

The text of Article 15 is therefore vital in reconciling clauses (1) and (3). However, what it does not resolve is the judicial indecision that accompanies different kinds of special measures and their validity under clause (3), i.e., what do “special provisions” mean and include? In order to answer this question, I extend my textual analysis in two ways – first, understanding “special” within the text of the equality code itself, and, second, venturing beyond the equality code to consider the relevance of measures under Part IV.

The immediate textual aid to understanding “special provisions” is the equality code itself. Indian courts have recognised time and again that Articles 14, 15, and 16 present a single code of equality.[164] In Akhil Bharatiya v Union of India, the SCI confirms what we now know about “special provisions” – they are not an exception to the “soul” of equality. The Court in this case held as follows:

Articles 14 to 16 form a code by themselves and embody the distilled essence of the Constitution’s casteless and classless egalitarianism.

 Nevertheless, our founding fathers were realists, and so did not declare the proposition of equality in its bald universality but subjected it to certain special provisions, not contradicting the soul of equality, but adapting that never-changing principle to the ever-changing social milieu….Article 16(4) imparts to the seemingly static equality embedded in Article 16(1) a dynamic quality by importing equalisation strategies geared to the eventual achievement of equality as permissible State action, viewed as an amplification of Article 16(1) or as an exception to it. The same observation will hold good for the sub-articles of Article 15.”[165]

(emphasis supplied)

The dictum of the SCI in this case clearly indicates that “special provisions” are a manifestation of the guarantee of equality; but more importantly, they are also “equalising strategies” in the SCI’s words. The SCI seems to suggest in Akhil Bharatiya that clause (4) of Article 16 alters the nature of equality through certain “equalisation strategies”, towards the eventual achievement of equality. My emphasis here is on the concept of an “equalising strategy”. I would argue that as strategies, they must necessarily be targeted at some goal or objective, or in this case, some manifestation of inequality/disadvantage, in order to qualify as a “special provision”. This interpretation of “special provisions” is buttressed by the DPSPs under Part IV, which provide useful examples of equalising strategies.

The text of two DPSPs under Part IV is particularly instructive in understanding the forms of special measures that may be implemented and justified under Article 15(3), outside of reservations in political bodies.[166] The first is Article 39(d), which states that:

The State shall, in particular, direct its policy towards securing (d) that there is equal pay for equal work for both men and women”.

The second DPSP that is relevant is Article 42, which states that:

The State shall make provision for securing just and humane conditions of work and for maternity relief”.

Both provisions require the State to “secure” or “make provisions” in order to reach a certain goal, or provide a benefit respectively. Admittedly, there are parallels between the language of these provisions and Article 15(3), and both depend on the exercise of the State’s discretion, albeit to different degrees. Even so, Article 15(3) ostensibly stands on a higher footing of importance given its presence within Part III and can be relied on effectively in a court of law. However, in my opinion, these Part IV principles actually provide instantiations of policies that are covered by Article 15(3). For instance, in order to achieve the goal of ‘equal pay for equal work’, the State would necessarily have to take positive measures to remedy the gendered wage gap and its effects. Similarly, legislation is required in order to provide for maternity relief and ensure employment benefits for pregnant women specifically.[167] These are instances of “special provisions” in favour of women that genuinely attempt to remedy or bear a nexus to historical and/or structural disadvantage, but would also be classifications based on sex prohibited under Article 15(1).[168] It is inconceivable that the DPSPs would direct the State to take measures that could be in violation of a right under Part III (i.e., Article 15(1)) because that would create a conflict within the text itself. These principles must thus necessarily be read alongside and as instantiations of Article 15(3).

Articles 39(d) and 42, therefore, assist in the interpretation of “any special provision” by providing illustrative forms of structural disadvantage which the court may utilise as reference points for other forms of disadvantage. Notably, utilising these principles as reference points would necessarily require both the State and the court to justify how and why instances of “protecting the security” of women similarly remedies disadvantage or subordination.

Based on this reading of Article 15(3) alongside the rest of the equality code and Part IV, a third prong may be added to the discussed framework of enquiry under Article 15(3) – that of a nexus between the measure and the disadvantage. This three-stage enquiry, i.e., the identification of disadvantage, the requirement of a nexus between the measure and disadvantage and the balancing of substantive equality considerations, may present a cogent framework for the application of Article 15(3), or at the very least, generate greater clarity on its scope and interpretation.[169]


Despite the recent SCI jurisprudence championing the cause of gender equality under Article 15, the application of clause (3) has remained chaotic and gendered. The object of this article has been to underscore the need for a clear, defined test under Article 15(3). The provision as it stands has largely been relied on by Indian courts with little scrutiny into its potential and purpose. As elucidated above, the existing applications also defeat the purpose of affirmative action for women by entrenching harmful stereotypes, which in turn limit female agency and perpetuate inequality. In contrast, I argue that the exclusive purpose of clause (3) must be to remedy manifestations of disadvantage — a purpose that is well-founded in constitutional history and the text of Part III and IV, and in the interests of substantive equality.

[1]The author has completed her BA LLB (Honours) from National Law School of India University, Bangalore, and is currently practicing at the Bombay High Court. The author would like to thank Dr. Sudhir Krishnaswamy and Ms. Malavika Prasad for their comments on earlier drafts of this paper, as well as Mr. Dhruva Gandhi and Mr. John Simte for their valuable feedback. All oversights are those of the author.

[2]The aspiration is clearly identifiable in parts that set forth an objective or ideal for the State, such as the Directive Principles of State Policy (‘DPSP’) encapsulated in pt IV. It is equally present in provisions that aspire to do away with certain social practices, such as art 17, which prohibits untouchability, and art 23, which prohibits trafficking and begar. The “social revolutionary” potential of art 17 is best encapsulated in Chandrachud J.’s opinion in Indian Young Lawyer’s Assn. v State of Kerala,(2019) 11 SCC 1, where he expands on the understanding of untouchability from a singularly caste-based phenomenon to one that is forbidden in any form, including gendered discrimination based in ideas of purity and pollution. For an analysis of the transformative potential of art 23, see Gautam Bhatia, The Transformative Constitution: A Radical Biography in Nine Acts (2019).

[3] The reference to arts 14, 15 and 16 together as a code of equality finds its place in several judicial opinions of the SCI. For instance, see State of Kerala v N.M. Thomas, (1976) 2 SCC 310, [139]— “It is platitudinous constitutional law that arts 14 to 16 are a common code of guaranteed equality, the first laying down the broad doctrine, the other two applying it to sensitive areas historically important and politically polemical in a climate of communalism and jobbery”.

[4] The phrase “affirmative action provision” in this paper refers to those provisions that enable the State to remedy structural disadvantage or prejudice through the introduction of “special provisions” aimed at the advancement of a disadvantaged category of people. While I use affirmative action and preferential treatment interchangeably in this paper, the use of “special provisions” is in direct reference to its legal meaning or interpretation under Article 15. The phrases “special provisions” and “advancement” are used in the language of arts 15(3), (4), (5), placing considerable discretion with the State with regards to the kinds of measures which may introduced to remedy disadvantage. On the other hand, arts 16 (3), (4), (4-A) and (4-B) are focused on the equality of opportunity in employment specifically through reservations.

[5] The Constitution of India 1950, art 15. It reads: “(1) The State shall not discriminate against any citizen on grounds only of religion, race, caste, sex, place of birth or any of them.

(2) No citizen shall, on grounds only of religion, race, caste, sex, place of birth or any of them, be subject to any disability, liability, restriction or condition with regard to—

(a) access to shops, public restaurants, hotels and places of public entertainment; or

(b) the use of wells, tanks, bathing ghats, roads and places of public resort maintained wholly or partly out of State funds or dedicated to the use of the general public”.

[6] Academic materials are relied on to the extent that they assist in understanding the ramifications of these possible constructions.

[7] Kalpana Kannabiran, Tools of Justice: Non-Discrimination and the Indian Constitution (1st edn, Routledge 2012) 361.

[8] Acharya (B.R.) v State of Gujarat, 1987 SCC OnLine Guj 85 [5].

[9] Charan Singh v Union of India, 1978 SCC OnLine Del 204.

[10] ibid [5].

[11] ibid [4].

[12] The assumption that women are vulnerable and need protection is not unique to Indian legislation and doctrine. For an analysis of similar “benign and protective” gender discrimination in the United States, see Rosalie Berger Levinson, ‘Gender-Based Affirmative Action and Reverse Gender Bias: Beyond Gratz, Parents Involved, and Ricci’ (2011) 34(1) Harvard Journal of Law & Gender 1, 4.

[13] EMC Steel Ltd v Union of India, (1991) 2 SCC 101 (‘EMC Steel Ltd’). 

[14] ibid [6].

[15] Rosalie Berger Levinson (n 12) 7.

[16] Srinivasaiah v State of Karnataka, 1986 SCC Online Kar 297. The SCI upheld an identical rule for the principal of a Government women’s college under art 15(3) in Vijay Lakshmi v Punjab University, (2003) 8 SCC 440 [10].

[17] ibid [3].

[18] S 497 of the Penal Code, 1860 and s 198(2) of the Code of Criminal Procedure, 1973 were examples of this rationale before they were declared unconstitutional by the SCI in Joseph Shine v Union of India, (2019) 3 SCC 39.

[19]Chokhi v State, 1953 SCC OnLine Raj 52 [4]. A similar provision exists in the first proviso to s 437(1) of the Code of Criminal Procedure, 1973.

[20]Cyril Britto v Union of India, 2003 SCC OnLine Ker 190.

[21] ibid [8].

[22] Mary Sonia Zachariah v Union of India, 1995 SCC OnLine Ker 288 [57].

[23] Ramish Francis Toppo v Violet Francis Toppo, 1988 SCC OnLine Cal 206 [9].

[24] Purnananda Banerjee v Swapna Banerjee, 1980 SCC OnLine Cal 202 [9].

[25] This logic also exists in cases where the outcome is ultimately in favour of complainant, such as the oft-cited decision of the SCI in C.B. Muthamma v Union of India, (1979) 4 SCC 260 [7], where the Court acknowledged the discriminatory and misogynistic practices of the Indian Foreign Services, but refused to lay down a principle prohibiting all differentiation between the sexes in order to “pragmatise where the requirements of particular employment, the sensitivities of sex or the peculiarities of societal sectors or the handicaps of either sex may compel selectivity”. Such reasoning permits invidious discrimination to masquerade as “natural” or inherent differences. See Gautam Bhatia (n 2).

[26] Anuj Garg v Hotel Assn. of India (2008) 3 SCC 1 (‘Anuj Garg’).

[27] ibid [36].

[28] ibid [46].

[29] The recent grant of Permanent Commissions to female officers in the Indian army and navy is a continuation in many ways of the reasoning present in Anuj Garg, indicating an increasing sensitivity to gender roles and stereotypes in the Indian judiciary (See Ministry of Defence v Babita Puniya, (2020) 7 SCC 469: 2020 SCC Online SC 200 and Union of India v Annie Nagaraja, (2020) 13 SCC 1: 2020 SCC Online SC 326).

[30] Girdhar Gopal v State,1952 SCC OnLine MP 202: AIR 1953 MB 147.

[31] Penal Code 1860, s 354. It reads: “Whoever assaults or uses criminal force to any woman, intending to outrage or knowing it to be likely that he will there by outrage her modesty, [shall be punished with imprisonment of either description for a term which shall not be less than one year but which may extend to five years, and shall also be liable to fine]”.

[32] Joseph Shine v Union of India, (2019) 3 SCC 39 (‘Joseph Shine’).

[33] ibid [185]-[189].

[34] ibid [118].

[35] ibid [274]-[277].

[36] Nariman J.’s opinion in this case is also significant, albeit for a different reason. In overruling the SCI’s decision in Yusuf Abdul Aziz v State of Bombay, AIR 1954 SC 321: 1954 SCR 930 (which had earlier upheld the validity of s 497), Nariman J. notes that art 15(3) only refers to the State’s ability to “make provisions” in the present — as opposed to the SCI in Yusuf Abdul Aziz, which noted that art 15(3) applies to future and existing laws.  Reading art 15(3) alongside art 366(10), Nariman J. argues that art 15(3) would not apply to existing laws, i.e., those passed prior to the commencement of the Constitution (Joseph Shine (n 18) [90]-[91]). This reasoning would have considerable ramifications for several pre-constitutional general statutes that grant preferences to women or afford affirmative action to women, but which the State would not be able to justify as a “special provision” under art 15(3).

[37] EMC Steel Ltd (n 13) [6].

[38] Om Narain Agarwal v Nagar Palika, Shahjahanpur, (1993) 2 SCC 242 [12].

[39] Govt. of A.P.  v P.B.  Vijayakumar, (1995) 4 SCC 520 (‘P.B. Vijaykumar’).

[40] ibid [1].

[41] The Andhra Pradesh State and Subordinate Service Rules 1996, r 22-A.

[42] P.B. Vijayakumar (n 39) [3].

[43] P.B. Vijayakumar (n 39) [6].

[44] ibid.

[45] P.B.  Vijayakumar (n 39) [7].

[46] P.B.  Vijayakumar (n 39) [8].

[47] P.B.  Vijayakumar (n 39) [12].

[48] The question of reservations under art 15(3) for women and the maximum permissible limit for the same was earlier superficially considered by the SCI in Toguru Sudhakar Reddy v Govt. of A.P., 1993 Supp (4) SCC 439, where the SCI upheld a reservation in excess of 50% by stating that the limit only applied to arts 15(4) and 16(4).

[49] P.B.  Vijayakumar (n 39) [9]-[10].

[50] Despite the SCI’s reasoning on this point, there remains ambiguity in the position of art 15(3) vis-à-vis art 16. Earlier, in Indra Sawhney v Union of India,(1992) Supp (3) SCC 217 [514], a nine-judges bench of the SCI had observed that art 15(3) cannot save a reservation in services that discriminates on the grounds of sex, because it falls under the ambit of art 16(2). This reasoning has been followed in recent cases such as Laxmi Kanwar v State, 2013 SCC OnLine Raj 1039: (2015) 1 RLW 744 to limit the forms of reservations to be given to women in public employment. On the other hand, the SCI’s reasoning on arts 15(3) and 16 has been followed by another High Court in G.K. Pushpa v State of Karnataka, 2012 SCC OnLine Kar 8725, leading to inconsistencies in outcomes.

[51] The SCI’s reasoning in this case was subsequently followed by another Division Bench of the SCI in Union of India v K.P. Prabhakaran, (1997) 11 SCC 638.

[52] Madhu Kishwar v State of Bihar,(1996) 5 SCC 125 [20].

[53] ibid [35].

[54] Thota Sesharathamma v Thota Manikyamma, (1991) 4 SCC 312 [19] and [21].

[55] Partap Singh v Union of India, (1985) 4 SCC 197 [6].

[56] See Shreya Atrey, ‘Through the Looking Glass of Intersectionality: Making Sense of Indian Discrimination Jurisprudence under Article 15’ (2016) 16 Equal Rights Review 160 for a discussion of whether the classification is prohibited when it is only on the basis of sex, and the gendered nature of the sex-plus criteria or “other considerations” adopted by the SCI in Air India v Nergesh Meerza,(1981) 4 SCC 335.

[57]  Anglo Vaidyik Balika Inter College v Board of High School and Intermediate Exam, 2002 SCC OnLine All 1676 (‘Committee of Management’).

[58] Constitution of India 1950, art 29(2) – “No citizen shall be denied admission into any educational institution maintained by the State or receiving aid out of State funds on grounds only of religion, race, caste, language or any of them”.

A similar reading of arts 15(3) and 29(2) was taken up in University of Madras v Shantha Bai, 1953 SCC OnLine Mad 182, where the Madras High Court held that while men could not take admission in women’s colleges under art 15(3), women could be admitted into men’s colleges depending on the regulations of the college.

[59] Committee of Management (n 57) [26].

[60] Independent Thought v Union of India, (2017) 10 SCC 800 [55]-[56].

[61] ibid [60]. Here, the SCI went as far as to say that affirmative action in regards the girl child must not just be liberally construed, but would override any other legislation limiting such benefits, in the spirit of art 15(3).

[62] State of A.P. v P.  Sagar, AIR 1968 SC 1379: (1968) 3 SCR 595 [9] (‘P. Sagar’).

[63] Partap Singh (n 55) [6].

[64] Neelam Rani v State of Punjab, 2010 SCC OnLine P&H 204 [34].

[65] ibid [36].

[66] ibid [32]-[33].

[67] Rajesh Kumar Gupta v State of U.P., (2005) 5 SCC 172.

[68] ibid [14].

[69] While this confusion leads to inconsistent positions across States, some courts do not even enter into a discussion of what measures for women would be justifiable in the first place — the validity of a specific reservation in favour of women under art 15(3) is often assumed in light of existing case law in such cases. See M.J. Simon v Union of India, 2010 SCC OnLine Ker 131 [12]; P. Sagar v State of A.P., 1966 SCC OnLine AP 35 [32].

[70] Dattatraya Motiram More v State of Bombay, 1952 SCC OnLine Bom 120.

[71] ibid. Despite this reasoning, the Bombay High Court read both provisions together to mean that the State can discriminate against men in favour of women, but not vice versa.

[72] Shamsher Singh v Punjab State, 1970 SCC OnLine P&H 67 [17].

[73] ibid [18], relying on Devadasan v Union of India, AIR 1964 SC 179.

[74] Ashoka Kumar Thakur v Union of India, (2008) 6 SCC 1 [118] (per Balakrishnan C.J.). In their concurring opinion, Pasayat J. and Thakker J. held that affirmative action cannot violate the basic structure of the Constitution or infringe the principles of equality [325].

[75] Ruth Bader Ginsburg and Deborah Jones Merritt, ‘Affirmative Action: An International Human Rights Dialogue’ (1999) 21(1) Cardozo Law Review 253, 254-255.

[76] Tarunabh Khaitan, A Theory of Discrimination Law (OUP, 2015) 216.

[77] ibid 218.

[78] ibid 219.

[79] ibid 220.

[80] Sean Pager, ‘Anti-subordination of Whom? What India’s Answer Tells Us About the Meaning of Equality in Affirmative Action’ (2007) 41 University of California at Davis Law Review 289, 336.

[81] ibid.

[82] Tarunabh Khaitan(n 76) 216-217.

[83] Alexandra Timmer, ‘Toward an Anti-Stereotyping Approach for the European Court of Human Rights’ (2011) 11(4) Human Rights Law Review 707, 708-715.

[84] Tarunabh Khaitan(n 76) 218.

[85] Food Corpn.  of India v Jagdish Balaram Bahira, (2017) 8 SCC 670 [56] (‘Food Corporation of India’).

[86] Catherine A. MacKinnon, ‘Sex equality under the Constitution of India: Problems, Prospects, and

‘‘Personal Laws’’’ (2006) 4(2) International Journal of Constitutional Law 181, 182.

[87] Martha C. Nussbaum, ‘India: Implementing Sex Equality Through Law’ (2001) 2(1) Chicago Journal of International Law 35, 47.

[88] Catherine A. MacKinnon (n 86).

[89] Catherine A. MacKinnon (n 86) 183. MacKinnon argues that “when inequality is socially institutionalised, it creates distinctions among people that can themselves serve as reasons for treating people worse not only will appear to be, but indeed will be, reasonable and not arbitrary at all” (Catherine A. MacKinnon (n 86) 187). This argument also reveals the issue with using “reasonability” as a standard for justification in cases of discrimination.

[90] Catherine A. MacKinnon (n 86)184.

[91] B.K. Pavitra v Union of India, (2019) 16 SCC 129.

[92] ibid [112].

[93] Catherine A. MacKinnon (n 86)186.

[94] Catherine. MacKinnon (n 86) 187.

[95] Ratna Kapur, ‘Gender Equality: Constitutional Challenges and Competing Discourses’ in Sujit Choudhry, Madhav Khosla and Pratap Bhanu Mehta (eds), The Oxford Handbook of the Indian Constitution (OUP 2016) 772.

[96] ibid 774.

[97] ibid 775. This is also the issue with the term “protective discrimination” for affirmative action measures, as frequently used by the Indian judiciary.

[98] ibid 775.

[99] Catherine A. MacKinnon (n 86) 187.

[100] Tarunabh Khaitan(n 76) 220.

[101] Tarunabh Khaitan(n 76) 35-6.

[102] K. Krishna Murthy v Union of India, (2010) 7 SCC 202 [40].

[103] Catherine A. MacKinnon (n 86) 187

[104] Food Corpn. of India (n 85) [2].

[105] See Sandra Fredman, ‘Substantive Equality Revisited’ (2016) 14(3) International Journal of Constitutional Law 712.

Fredman rejects a formal, legalistic understanding of equality in favour of a multidimensional substantive approach that accounts for 4 key markers:

i. Redressing disadvantage

ii. Addressing stigma, stereotypes, prejudice and violence

iii. Enhancing voice and participation

iv. Accommodating difference to achieve structural change. (Sandra Fredman, 713).

[106] Renee Leon, ‘W(h)ither Special Measures? How Affirmative Action for Women can Survive Sex Discrimination Legislation’ (1993) 1 Australian Feminist Law Journal 89, 105.

[107] Luc Tremblay raises the same concern with respect to the Canadian jurisprudence on sex discrimination and affirmative action, in that placing non-discrimination and affirmative action within the paradigm of substantive equality does not resolve the tension between these norms. See Luc B. Tremblay, ‘Promoting Equality and Combating Discrimination Through Affirmative Action: The Same Challenge? Questioning the Canadian Substantive Equality Paradigm’ (2012) 60(1) American Journal of Comparative Law 181, 196-200.

[108] Renee Leon (n 106) 93.

[109] Tarunabh Khaitan (n 76) 231.

[110] Renee Leon(n 106) 94-6.

[111] Renee Leon(n 106) 110.

[112] Ruth Bader Ginsburg and Deborah Jones Merritt (n 75) 258 and 269.

[113] Reva B. Siegel, ‘From Colorblindness to Antibalkanization: An Emerging Ground of Decision in Race Equality Cases’ (2011) 120(6) Yale Law Journal 1278, 1280.

[114] Kimberle Williams Crenshaw, ‘Framing Affirmative Action’ (2006) 105 Michigan Law Review First Impressions 123, 131-132.

[115] Philip Bobbitt, Constitutional Fate: Theory of the Constitution (OUP 1982) 14-21.

[116] Constitution of India Bill 1895, art 20 <> accessed 18th April 2021.

[117] K.P. Singh, ‘Role of the Congress in the Framing of India’s Constitution’(1990) 51(2) Indian Journal of Political Science 153, 154.

[118] Commonwealth of India Bill 1925 art 7(g) <> accessed 18th April 2021.

[119] Vivek Prahladan, ‘Emergence of the Indian Constitution: Affirmative Action and Cultural Fault Lines’ (2012) 47(7) Economic and Political Weekly 45, 47; See generally Niraja Gopal Jayal, Citizenship and its Discontents: An Indian History (Harvard University Press 2013).

[120] The Nehru Committee Report 1929, art 4 (xix) <> accessed18th April 2021.

[121] B. Shiva Rao (ed),The Framing of India’s Constitution: Select  Documents, vol [II]  ([Indian Institute of Public Administration] [1968]) 173-174.

[122] Government of India Act 1935 art 275 <> accessed 18th April 2021.

[123]All India Women’s Conference, ‘AIWC at a Glance: The First Twenty-five Years 1927-1952’ <> accessed 18th April 2021.

[124] Resolution on ‘Women and Government’, Eleventh Session of the All-India Women’s Conference, Ahmedabad (December 1936), as referred to in the Report of the Twelfth Session of the All India Women’s Conference (Nagpur, December 1937), 177.

[125] These organisations were closely associated with the Indian National Congress than with minority parties and groups, and, therefore, represented largely upper class and upper caste views. See Laura Dudley Jenkins, ‘Competing Inequalities: The Struggle Over Reserved Legislative Seats for Women in India’ (1999) 44 (S7) International Review of Social History 53, 59-61.

[126] Notably, this hesitation towards reservations for women has continued to exist in political spaces post-independence. See ‘Towards Equality: Report of the Committee on Status of Women in India’ (Department of Social Welfare, Ministry of Education and Social Welfare, Government of India 1974).

The Committee on the Status of Women was constituted by the Ministry of Education and Social Welfare [Resolution dated 22 September 1971] to analyse the socio-economic status of women. In their ground-breaking report ‘Towards Equality’, the Committee observed that a large number of women in India were unaffectedby the rights provided under the Constitution, indicating a need to make these rights more “real and meaningful”. In lieu of this goal, the Committee made recommendations across a broad spectrum of economic, social, legal and political issues. With respect to political status and representation specifically, the Committee observed that even if women may not numerically constitute a minority, they possess the features of one in terms of economic, social and political inequalities. In cognizance of this, regional women’s organisations had made extensive suggestions for reservations in political bodies to increase representation and mobilisation. On the other hand, political parties and female legislators opposed this because special representation would detract from the equality conferred by the Constitution. After considering the matter, the Committee refused to recommend reservations for women in legislatures for two reasons — first, that women did not constitute a community independent of other socio-economic group identities that they shared with men, and, second, women are capable of competing with men as equals and reservations would perpetuate their unequal status.  However, it did recommend special provisions for women in local government [301-304].

[127] Presidential Address of Shrimati Sarojini Naidu, All India Women’s Conference, Fourth Session, Bombay (1930).

[128] Jana Matson Everett, Women and Social Change in India (Heritage, 1981) 195.

[129] B. Shiva Rao (n 121) 178.

[130] ibid.

[131] B. Shiva Rao (n 121) 176.

[132] B.  Shiva Rao (n 121) 183.

[133] Cl 5, Draft Report of the Constituent Assembly Sub-Committee on Fundamental Rights (April 1947).

[134] B.  Shiva Rao (n 121) 184.

[135] ibid.

[136] B. Shiva Rao (n 121) 185.

[137]B. Shiva Rao (n 121) 187; Draft Constitution 1947, cl 11(2).

[138] ibid. Interestingly, the draft constitution prepared by the Socialist Party also contained an identical provision – Draft Constitution of the Republic of India (Socialist Party) 1948, art 19 <>.

[139] Draft Constitution 1948, as introduced into the Constituent Assembly on 4 November 1948 <> accessed 18th April 2021.

[140]Added as art 15(4) by the Constitution (First Amendment) Act, 1951.

[141]K.T. Shah, Constituent Assembly Debates, vol VII (29 November 1948)7.62.89.

[142]K.T. Shah, Constituent Assembly Debates, vol VII (29 November 1948)7.62.93.

[143]K.T. Shah, Constituent Assembly Debates, vol VII (29 November 1948)7.62.124-126; Although this amendment was rejected by Dr B.R. Ambedkar, it was because Dr Ambedkar argued such measures would not achieve the purpose of assimilation of the lower castes into mainstream society, not because this approach to equality was fallacious.

[144]The statement made was as follows:

The rage for equality which has led to provide equal citizenship and equal rights for women has sometimes found exception in regard to special provisions that, in the long range, in the interest of the country or of the race, exclude women from certain dangerous occupations, certain types of work”.

See KT Shah, Constituent Assembly Debates,vol VII (29 November 1948)7.62.90.

[145] Damodar S. Seth, Constituent Assembly Debates, vol VII (3 December 1948) 7.66.214.

[146] Damodar S. Seth, Constituent Assembly Debates, vol VII (3 December 1948) 7.66.215. A similar provision featured in the Socialist Party’s draft – Draft Constitution of the Republic of India (Socialist Party) 1948, art 47.

[147] Martha C. Nussbaum (n 87) 39.

[148] Philip Bobbitt (n 115) 12.

[149] Philip Bobbitt (n 115) 34-36.

[150] For an expansive reading of art 15(2) and its application to private acts of discrimination, see Gautam Bhatia, ‘Horizontal Discrimination and art 15(2) of the Indian Constitution: A Transformative Approach’ (2016) 11(1) Asian Journal of Comparative Law 87.

[151] Constitution of India 1950, art 23(3). It reads: “Nothing in this article shall prevent the State from imposing compulsory service for public purposes, and in imposing such service the State shall not make any discrimination on grounds only of religion, race, caste or class or any of them”.

[152] The Constitution of India 1950 art 33. It reads: “Parliament may, by law, determine to what extent any of the rights conferred by this Part shall, in their application to, —

(a) the members of the Armed Forces; or

(b) the members of the Forces charged with the maintenance of public order; or

(c) persons employed in any bureau or other organisation established by the State for purposes of intelligence or counter intelligence; or

(d) person employed in, or in connection with, the telecommunication systems set up for the purposes of any Force, bureau or organisation referred to in clauses (a) to (c), be restricted or abrogated so as to ensure the proper discharge of their duties and the maintenance of discipline among them”.

[153] S. Sundaram Pillai v V.R.  Pattabiraman, (1985) 1 SCC 591 [43].

Here, the use of SCI case law is not to present a binding precedent or a doctrinal argument, but to present principles of textual interpretation or the legal meanings of terms such as “proviso” and “non-obstante clause” as ascertained by courts.

[154] Director of Education (Secondary) v Pushpendra Kumar, (1998) 5 SCC 192 [8]. In the context of art 15, this argument has been upheld by the SCI in the context of art 15(4) nullifying art 15(1)’s guarantee in P. Sagar (n 62) [7].

[155]Yogendra Pratap Singh v Savitri Pandey, (2014) 14 SCC 812 [16], relying on an earlier decision of the SCI in Sarav Investment & Financial Consultancy (P) Ltd v Llyods, Register of Shipping, Indian Office, Staff Provident Fund, (2007) 14 SCC 753[16].

[156] Dominion of India v Shrinbai A. Irani, AIR 1954 SC 596: (1955) 1 SCR 206 [11]; R.S. Raghunath v State of Karnataka, (1992) 1 SCC 335: AІR 1992 SC 81 [10]-[11]. See generally G.P. Singh, Principles of Statutory Interpretation:  Including the General Clauses Act, 1897 with Notes (Lexis Nexis, 2016).

[157]Ajit Singh (2) v State of Punjab, (1999) 7 SCC 209 [28] and [29].

[158]A similar position was taken by Mathew State of Kerala v N.M. Thomas, (1976) 2 SCC 310, where he observed as follows in his concurring opinion:

“I agree that art 16(4) is capable of being interpreted as an exception to art 16(1) if the equality of opportunity visualized in art 16(1) is a sterile one, geared to the concept of numerical equality which takes no account of the social, economic, educational background of the members of Scheduled Castes and scheduled tribes. If equality of opportunity guaranteed under art 16(1) means effective material equality, then art 16(4) is not an exception to art 16(1). It is only an emphatic way of putting the extent to which equality of opportunity could be carried viz., even up to the point of making reservation” [78].

[159] Gulshan Prakash v State of Haryana, (2010) 1 SCC 477 [16]. Another Bench of the SCI took a different view in P. Sagar (n 62) and held that art 15(4) is an exception, therefore this question remains unresolved.

[160] Indra Sawhney v Union of India, 1992 Supp (3) SCC 217.

[161]ibid [859].

[162]ibid [563].

[163] See Gautam Bhatia (n 2). See also Unnati Ghia and Dhruva Gandhi, ‘The Anti-Stereotyping Principle: A Conundrum in Comparative Constitutional Law’ (IACL-AIDC Blog, 5 May 2020) <> accessed 18th April 2021.

[164] The most recent recognition of this reading of Part III is in B.K.  Pavitra v Union of India, (2019) 16 SCC 129 [120] and [148].

[165]Akhil Bharatiya Soshit Karamchari Sangh (Railway) v Union of India, (1981) 1 SCC 246 [36].

[166]Reservations as a special measure for women already possess a degree of constitutional recognition due to arts 243-D(2) and (3), though only at the level of local governments.

[167] The SCI has already read art 15(3) alongside arts 39 and 42 in MCD v Female Workers (Muster Roll), (2000) 3 SCC 224 [7]-[11].

[168] Maternity benefit laws may still violate art 15(1) in that they often rely on a stereotype of women being primary caregivers, and hinder the equal distribution of parental labor and care. However, to the extent that they are targeted at alleviating the disadvantages pregnant women face at the workplace, I would argue that it is still an instance of remedying structural disadvantage.

[169] The emphasis on ‘disadvantage’ within the suggested framework also draws from the test outlined by the Canadian Supreme Court in Fraser v Canada (Attorney General), 2020 SCC 28. As per the majority in Fraser, in assessing discrimination under s 15(1) of the Canadian Charter of Rights and Freedoms (1982), courts must look at “whether the law has the effect of reinforcing, perpetuating, or exacerbating disadvantage”, which is “viewed in light of any systemic or historical disadvantages faced by the claimant group” [76]. This test was recently followed by Chandrachud J. in the context of indirect discrimination in Nitisha v Union of India 2021 SCC OnLine SC 261.


Please enter your comment!
Please enter your name here