WHEN THE SUPREME COURT released the Handbook on Combating Gender Stereotypes (‘the Handbook’) in 2023, the then Chief Justice D.Y. Chandrachud explained the need to combat the “inadvertent” perpetuation of gender stereotypes, and in particular “judgments which have referred to a woman as a ‘concubine’ when she is in a relationship.” In family law, the term ‘concubine’ is more than an insult; it precisely identifies a cause of action. It may be used to decide a wife’s entitlement to maintenance or to establish grounds for desertion, cruelty, and adultery in divorce proceedings. The 2023 Handbook set out to retire the word, and so does the new 2026 Report, Judgments and Gender (Sensitivity and Compassion in Writing Judgments) (‘the Report’).
While reform that sanitises language and nomenclature works when the words it retires only describe a person, can it do the same when those words define legal status and entitlement? This piece follows that gap into the Report’s treatment of non-marital relationships. Stripping judgments of degrading language is simple, but excising that language from the status-based legal rights it defines is not. The Report seems to assume that the first accomplishes the second.
Changes from the 2023 Handbook
The Report is not a sequel to the Handbook. While the latter was an administrative product of the Supreme Court’s e-Committee, published without any judicial order requiring its use, the Report sprang from an adjudicated case. In March 2025, the Allahabad High Court reduced charges against two accused persons who had “committed sexually offensive acts” against a minor girl, finding that their actions only reached the stage of “preparation,” not attempt. The Supreme Court set the order aside in February 2026, finding the case for graver charges “clearly and inevitably” made out on the facts alleged. In the same order, the Court directed the National Judicial Academy, through director Justice Aniruddha Bose, to constitute a Committee of Experts to develop “Guidelines to Inculcate Sensitivity and Compassion into Judges and Judicial Processes …”. During the proceedings in February, Chief Justice Surya Kant reportedly remarked that the Handbook was “too Harvard-oriented” for a rape survivor, her family, or the ordinary litigant to use.
The Committee’s report was approved by an oral order on July 14, 2026, directing that it “shall serve as a mandatory guiding resource in matters involving sexual offences and other cases concerning vulnerable persons,” to be uploaded on every court website and circulated to police for use in FIRs and chargesheets.
Whereas the Handbook addressed gender stereotyping broadly, reaching into marital and domestic roles, the Report’s mandate extends only to sexual offences and vulnerable victims and witnesses. This narrower focus arguably reflects a broader judicial pushback against the Handbook’s seemingly elitist framing. Beyond the oral ‘Harvard-oriented’ remark, the July order explicitly faulted “earlier guidelines” for being “framed with reference to a predominantly anglicised model of court functioning.” Despite this critique, neither any order, nor the Report itself, says what becomes of the Handbook outside sexual-offence matters.
The Handbook as well as the Report seek to replace the use of ‘concubine’, ‘keep’ and ‘mistress’, but struggle to provide a replacement that courts can use instead.
Words as the cause of action
The Handbook as well as the Report seek to replace the use of ‘concubine’, ‘keep’ and ‘mistress’, but struggle to provide a replacement that courts can use instead. The Handbook proposed to simply refer to a “woman with whom a man has had romantic or sexual relations outside of marriage,” a description standing in for a definition. The Report arguably goes further, proposing the general label “partner” as the “ideal substitute” for all three terms, which would also extend to same-sex relationships. In ordinary usage, replacing a demeaning word with a neutral one is enough, because the word is only describing a person. However, in family law, these terms are imbued with legal force, enabling them to trigger or defeat a right and neither replacement can do that.
Section 18(2)(e) of the Hindu Adoptions and Maintenance Act, 1956 (‘HAMA’) lets a Hindu wife live apart from her husband without losing her right to maintenance if he “keeps a concubine in the same house in which his wife is residing, or habitually resides with a concubine elsewhere.” The term ‘concubine’ here is not mere colour, but the fact a court must find before her claim can succeed. A court applying this section has to be able to name, as a matter of fact, the kind of relationship the husband is in. This gap showed up on the bench in March 2025. On the Madurai Bench of the Madras High Court, Justice G.R. Swaminathan accepted that “concubine” was now a “forbidden expression” under the Handbook. However, in his words, the alternative was “a description and not a definitional term,” thus failing to do the statutory work ‘concubine’ had done. Unable to find an equivalent term, the court used the retired word anyway, asserting that “We need definitions for descriptions.” Even ‘partner’, as suggested in the new Report cannot perform this status-defining function because it does not distinguish marriage from its absence. It fails to address exclusivity, permanence, or second households, and without precise findings on these elements, a wife’s maintenance claim collapses.
Justice Swaminathan traced the term beyond HAMA to broader Supreme Court precedent, most notably Rajnesh v. Neha (2020), the landmark ruling on spousal maintenance frameworks, where the Court noted:
“The law presumes in favour of marriage, and against concubinage, when a man and woman cohabit continuously for a number of years. Unlike matrimonial proceedings where strict proof of marriage is essential, in proceedings Under Section 125 Code of Criminal Procedure such strict standard of proof is not necessary.”
Here, ‘concubinage’ serves as the negative pole of a strict legal binary: continuous cohabitation presumes a legitimate, protected marriage, while ‘concubinage’ represents the unprotected, illegitimate alternative.
This binary can act as a legal shield for the ‘lawfully wedded’ spouse. In February 2022 for instance, the Chhattisgarh High Court held that a wife who left her marital home after her husband brought a ‘concubine’ into the house and fathered a child with her, had not deserted him under Section 13(1)(ib) of the Hindu Marriage Act, 1955. In this case, the husband sought divorce on grounds of desertion (presumably to stop paying spousal maintenance), claiming she had lived apart from him for twenty-five years without cause. The Court found that his conduct (in ‘keeping’ a concubine) had given her reasonable cause to leave, and upheld the lower court’s order of spousal maintenance. The same fact, a husband having a ‘concubine’, grants a maintenance claim under one section, and defeats a desertion charge under another.
The term also performs an exclusionary function. In 2024, when considering whether a girlfriend could be prosecuted as a “relative of the husband” under Section 498A of the Indian Penal Code, 1860, the Supreme Court affirmed that “by no stretch of imagination would a girlfriend or even a concubine in an etymological sense be a relative,” which is defined by blood, marriage, or adoption. Liability does not attach precisely because the relationship falls outside marriage. Terms that broadly refer to any romantic partner cannot perform this function.
The preoccupation with vocabulary leaves the Report just as detached from the material needs of litigants as any ‘anglicised’ model of court functioning.
The Handbook’s substitute is a description doing the work of a definition. The Report’s substitute is broader still. Both successive fixes widen the word at the exact point the statute needs narrowed. The term ‘concubine’ has never simply named a woman, but drawn a line: who gets maintenance, whose departure counts as desertion, who can be criminally charged. Retiring the word does not eliminate the need to draw these lines; it merely deprives courts of the language used to draw them precisely, without offering an adequate replacement.
Classifying non-marital relationships
The sorting function performed by terms like ‘concubine’ is most visible in the law’s treatment of live-in relationships. Indian family courts are tasked with classifying non-marital relationships for several purposes that “partner” cannot perform. Section 2(f) of the Protection of Women from Domestic Violence Act, 2005 protects a “relationship in the nature of marriage” leaving courts to decide which relationships qualify. The Supreme Court supplied the test in D. Velusamy (2010), and refined it in Indra Sarma (2013), requiring fact specific findings on inter alia duration, shared household, pooling of resources and public representation as a couple, before a woman gets any relief at all. Indra Sarma notably sorted live-in relationships into four kinds: between (a) two unmarried adults; (b) an unmarried woman and a married man she knowingly entered into a relationship with; (c) a married woman and an unmarried man; and (d) between an unmarried woman and a married man she did not know was married. Only the first and the last can attract the Act’s protection.
A woman in the second category is functionally in the position ‘concubine’ used to name. In both rulings, courts declined to recognise a “relationship in the nature of marriage” for lack of exclusivity and monogamy. Consequently, the word lives on in judgments that reject claims, even as reform projects erase it from those that allow them. The proposed substitute, “partner,” fails to explain a denial under the PWDVA since a “partner” is exactly what the claimant claims to be. Admittedly, this is also perhaps the one place the Report can claim real credit over the Handbook. The Handbook’s substitute for ‘concubine’ spoke only of “a woman” in relation to “a man”, while the Report's ‘partner’ is offered as extending “to same-sex relationship also.” Yet, a word that includes more people cannot entitle them to more than what the law allows, especially under the PWDVA and HAMA’s restrictive coverage. The Report's gesture toward inclusion might be sincere, but stops at the line between describing a person and defining their legal position.
This omission is unavoidable, given the Report’s limited mandate. Under the Court’s approach, gender sensitivity concerns how a judgment refers to a person, rather than how it decides their legal standing. Replacing problematic language is passed off as reform, while the rules governing relief remain tethered to the old terms.
Ultimately, both the Handbook and Report treat harmful stereotypical language as mere descriptive words, without considering the legal outcomes attached to their application. This slippage recurs in the Report’s treatment of gender identity, flattening socio-cultural identities like hijra and thirunangai into the biological category of “intersex” alongside “pejorative” slurs. The preoccupation with vocabulary leaves the Report just as detached from the material needs of litigants as any ‘anglicised’ model of court functioning. Despite the dubious reliability of Report’s self-reported survey data, it captures the disconnect. When 86.4 percent of respondents, as per the Report’s survey, already understand court proceedings, but 80.7 percent have never heard of a witness protection scheme, the real crisis is not ‘Harvard-oriented’ vocabulary.
This slippage recurs in the Report’s treatment of gender identity, flattening socio-cultural identities like hijra and thirunangai into the biological category of “intersex” alongside “pejorative” slurs.
Even a judge who faithfully replaces ‘concubine’ with ‘partner’ is constrained to find, in fact if not in name, whether the husband had a ‘concubine’. There is no other route to that finding in law. The crucial question is therefore not whether courts will use the Report’s vocabulary, but whether these new terms can carry the specific legal consequences the law demands.
Reform in this regard requires far more than updating vocabulary, and faces two separate challenges: preserving a person’s dignity in descriptive language and also relieving courts of the need to specifically characterise relationships to determine protections. While better wording can do the first job, the second is not a matter of vocabulary, but of the statutory threshold required for relief. A euphemism can soften a story, but it cannot carry a cause of action. Compassion in judgment writing means little if its language cannot survive contact with the realities of status-based legal entitlements and family proceedings.