The Prohibition of Child Marriage Act carries a puzzle in its own title. A statute enacted to prohibit a practice does not declare the prohibited act a nullity. A child marriage in India is an offence committed by nearly everyone who arranges, performs or permits it, yet the marriage itself is, in the ordinary case, a valid and subsisting marriage in the eyes of the law. It remains valid the day after the ceremony, it remains valid when children are born of it, and it remains valid for years afterwards, unless one particular person, the party who was a child when the marriage was solemnised, goes to a district court and asks for it to be annulled inside a window that is quietly closing. The prohibition falls on the adults. The burden of undoing falls on the child.

That single design choice organises everything else in the Act, and it is the choice that almost every popular account of the statute gets wrong. Pages that describe child marriage as illegal and therefore void collapse two different legal ideas into one. Illegality describes the conduct of the people who brought the marriage about, and the Act punishes that conduct with rigorous imprisonment. Validity describes the status of the relationship the ceremony created, and on that question the Act deliberately withholds the automatic answer. Between those two ideas sits the whole working life of the statute: the annulment petition and its deadline, the maintenance and residence orders that accompany a decree, the legitimacy of children born of the marriage, the small class of cases in which the marriage is void from the start, the officers appointed to intervene before a ceremony happens, and the injunctions by which a magistrate can forbid one.

The statute also carries a second, quieter story about how Indian law changes and fails to change. The Act of 2006 was itself a reform statute, passed to replace a restraint law of 1929 whose penalties had been too small to matter. In December 2021 the government introduced a bill to amend it, raising the minimum age of marriage for women from eighteen to twenty-one and giving the Act an express overriding force over personal law. That bill was sent to a parliamentary standing committee, which invited public views and received an extraordinary volume of responses, the overwhelming majority of them opposed. The committee took extension after extension for over two years and never presented a report. When the seventeenth Lok Sabha was dissolved in June 2024, the bill lapsed with it, unpassed, unvoted on, and formally undecided. A series that has criticised legislation passed without scrutiny meets here the opposite failure: scrutiny without decision.

This article is the statute pillar for the Act. It states the text accurately, works through the validity scheme provision by provision, explains the offences and the prevention machinery, places the Act beside the criminal law that surrounds a married child, gives both sides of the marriage age argument at full strength, records what the Supreme Court directed in 2024, and ends with a defended verdict on the design.

The law before prohibition: restraint without consequences

The statute the 2006 Act replaced had been on the book for seventy-seven years, and its history explains why Parliament eventually concluded that a new architecture, rather than another amendment, was required.

The Child Marriage Restraint Act, 1929 (Act 19 of 1929), known as the Sarda Act after its sponsor Harbilas Sarda, was passed on 28 September 1929 and came into force on 1 April 1930. It fixed the minimum age of marriage at fourteen for girls and eighteen for boys. Its very name recorded its modesty. The Act restrained; it did not prohibit, and it said nothing at all about the validity of a marriage solemnised in breach of it. A marriage of a girl of thirteen was an offence by the adults involved and a perfectly effective marriage at the same time, and the penalties matched the era’s ambivalence: simple imprisonment measured in days or a few months, and fines measured in hundreds or a thousand rupees. Prosecution was hedged with procedural conditions, and the offences were neither cognizable nor structured for prevention. There was no officer charged with stopping a marriage before it happened, and no power in any court to forbid one.

The restraint Act’s penalties, in their amended form, record the era’s priorities with embarrassing precision. A male adult below twenty-one who married a child faced simple imprisonment of up to fifteen days or a fine of up to one thousand rupees. A male adult above twenty-one faced simple imprisonment of up to three months and a fine. The person who performed or directed the ceremony, and the parent or guardian who promoted or negligently permitted it, faced the same three-month scale. The offences were not cognizable in the ordinary way, complaints were channelled and time-limited, and no woman could be imprisoned under the Act at all. A legislature that prices an offence at fifteen days for the groom has not prohibited the practice; it has taxed it at a rate the practice could easily pay.

The ages moved in steps, each step trailing the social argument by decades. In 1949 the Child Marriage Restraint (Amendment) Act (Act 41 of 1949) raised the age for girls from fourteen to fifteen. In 1978 the Child Marriage Restraint (Amendment) Act (Act 2 of 1978), with effect from 1 October 1978, substituted the definition that the modern law still uses, fixing a child as a male who has not completed twenty-one years and a female who has not completed eighteen. The restraint framework itself was left untouched by that amendment. The ages rose; the machinery did not.

Three failures of the 1929 design mattered by the time reform was seriously taken up. The first was penal triviality: a maximum of three months of simple imprisonment for a man above twenty-one who married a child, under the old Section 4, could not deter a practice sustained by family and community decision. The second was the absence of any preventive instrument. The restraint Act operated, if it operated at all, after the ceremony, when the social fact the law objected to had already been created and when prosecution meant prosecuting the family of the very child the law claimed to protect. The third was silence on consequences. The old Act had nothing to say about the position of the girl afterwards, no maintenance, no residence, no statement of the legitimacy of her children, no route by which she could exit the marriage as a statutory entitlement rather than through the general divorce law of her personal law, which a minor could rarely invoke and which treated the marriage as valid for the purpose of dissolving it.

The reform pressure of the 1990s and the following decade pushed in a consistent direction. The National Commission for Women, in its recommendations of 1995-96, proposed that child marriages be declared void and that registration of marriages be made compulsory. The government of the day introduced the Prevention of Child Marriage Bill, 2004, and the departmentally related Standing Committee on Personnel, Public Grievances, Law and Justice, which examined it, recommended a common marriageable age of eighteen for both sexes and recommended that a child marriage be void from the beginning rather than voidable at the option of a party, precisely because of the inhibitions that keep the parties from approaching a court. The enacted statute accepted neither recommendation. The ages stayed split at twenty-one and eighteen, and validity was left with the child. The Law Commission returned to the subject in its Report No. 205 of 2008, a proposal to amend the Prohibition of Child Marriage Act, 2006 and other allied laws, which kept the case for strengthening the validity and registration provisions alive after the new Act had commenced. The enacted design, in other words, was a compromise that its own examining bodies had warned against, and the warnings are part of the record against which the design has to be judged.

The Act of 2006 and the year trap in its name

The statute that emerged is cited as the Prohibition of Child Marriage Act, 2006, and it is Act 6 of 2007. Both halves of that sentence are correct, and the gap between them is the first accuracy trap the subject sets. The bill was passed in the closing part of 2006 and took its short title from that year. It received the assent of the President on 10 January 2007 and was published in the Gazette of India on 11 January 2007 as Act 6 of 2007. Commencement came later still: by notification S.O. 1850(E) dated 30 October 2007, the Act was brought into force on 1 November 2007. A reader who dates the statute’s operation from its short title is therefore wrong by the better part of a year, and a reader who searches for it among the Acts of 2006 will not find it. The short title is a label Parliament gave it; the Act number, the assent date and the commencement date are the operative facts, and they belong to 2007.

Section 1(3) left commencement to notification and permitted different dates for different states, the standard device for a social statute whose machinery is state-built; in the event the Central Government appointed a single date, 1 November 2007, and the Act’s operative life runs from then. Section 21 repealed the Child Marriage Restraint Act, 1929, with a savings clause under which cases and proceedings pending under the old Act continued to be disposed of under it. Section 1 extends the Act to the whole of India and applies it to all citizens of India within and beyond the country, with a single stated exception for the Renoncants of the Union territory of Pondicherry. As enacted, the Act did not extend to the State of Jammu and Kashmir; those words were omitted from Section 1 by the Jammu and Kashmir Reorganisation Act, 2019 (Act 34 of 2019) with effect from 31 October 2019, which is why the Act as amended is printed with that amendment noted against it. Section 19 leaves rule-making to the State Governments, a federal allocation that matters in practice because so much of the Act’s machinery, beginning with the appointment of its officers, is state action.

The Act has twenty-one sections, and they fall into four working groups. The first group, Sections 2 and 3 with the void provisions in Sections 12 and 14, governs status: what a child marriage is and what it does to the marriage. The second group, Sections 4 to 8, governs the consequences when a marriage is annulled: maintenance, residence, custody, legitimacy, modification of orders and the forum. The third group, Sections 9 to 11 with Section 15, is the penal code of the Act: who is punished, for what, and with what procedure. The fourth group, Sections 13 and 16 to 18, is the preventive machinery: injunctions and the Child Marriage Prohibition Officers. Section 20 reaches outside the Act to raise the penalty under the Hindu Marriage Act, 1955 for marrying in breach of its age condition, and Section 21 performs the repeal. A statute pillar earns its length by taking the groups in that functional order rather than reciting the sections in numerical sequence, and that is the order followed here, after the definition on which every group depends.

The definition: a child is twenty-one or eighteen, depending on sex

Section 2(a) defines a child, for the purposes of this Act, as a person who, if a male, has not completed twenty-one years of age, and if a female, has not completed eighteen years of age. Section 2(b) defines a child marriage as a marriage to which either of the contracting parties is a child. The definition is asymmetric on its face, and the asymmetry is inherited rather than chosen. The 2006 Act did not sit down and decide that a husband of twenty and a wife of nineteen stand in different relations to childhood. It adopted the marriageable ages that the general law had settled on by 1978, twenty-one for a bridegroom and eighteen for a bride, and defined childhood as everything below the permitted line. The definition is therefore a mirror of the marriage ages, not an independent judgment about maturity.

That inheritance produces the statute’s most frequently examined oddity: the same person is a child for one purpose and an adult for another, inside the same statute’s family. The Indian Majority Act, 1875 fixes majority at eighteen, and Section 2(f) of this Act borrows that statute for its own separate definition of a minor. A man of nineteen is accordingly an adult of full legal capacity under the Majority Act, competent to contract, to vote having been competent since eighteen, and to be tried as an adult for any offence, while under Section 2(a) of the child marriage statute he is a child whose marriage is a child marriage, voidable at his option, and whose bride, if she is eighteen or above, commits no offence under this Act by marrying him although an adult man who marries a girl of seventeen commits a serious one. The definitional divergence across the child statutes is mapped in full in the guide to child protection laws in India, where the eighteen of the juvenile justice and sexual offences statutes, the fourteen and eighteen of the labour statute and the six to fourteen of the education statute are set side by side; this Act’s split line is the outlier in that company, the only one of the major child statutes that does not use a single age for both sexes.

Why does Indian law set the marriage age at twenty-one for men and eighteen for women?

The two ages are the marriageable ages fixed in 1978 for the earlier restraint statute and carried into the Act of 2006, which defined a child as anyone below the age at which marriage was permitted. The split was not a fresh judgment in 2006 about maturity; it was the inherited settlement of the general marriage law, adopted wholesale.

The criticism of the definition runs in two directions that do not agree with each other, which is why the definition has survived attempts from both. One line of criticism, pressed by the Standing Committee that examined the 2004 bill and by many child rights bodies, is that childhood in Indian law ends at eighteen for every other protective purpose, so a definition that treats a man of twenty as a child for marriage is defensible but a scheme that leaves the female line at eighteen while the protective statutes also stop at eighteen invites the charge that the marriage age for girls was set by marriageability rather than by protection. The opposite criticism, which became the government case in 2021, is that equality requires the same age for both sexes and that the gap itself entrenches the assumption that a wife should be younger than her husband. Both criticisms were available to Parliament in 2006. It answered neither, and the 2021 bill, which would have answered the second by lifting the female age to twenty-one, lapsed unanswered as well.

One further feature of the definition deserves to be stated plainly because it controls the offences. The definition looks at the parties, not at the ceremony. A marriage is a child marriage if either party is a child, whichever sex the child is and whoever arranged the event. A marriage between a man of twenty-five and a girl of seventeen is a child marriage. So is a marriage between a man of twenty and a woman of thirty. The penal provisions then distribute liability unevenly across that single definition, and the distribution, examined in the offence sections below, is where the Act’s real judgments about blame are found.

One further question about the definition is asked constantly and answered badly: what counts as a marriage for the Act’s purposes. The Act does not require a registered marriage, a religious ceremony of any particular kind, or a form valid under the parties’ personal law. Section 12 speaks of a minor being “made to go through a form of marriage”, and the scheme of the Act treats the social fact of a marriage ceremony, by whatever rite and before whatever witnesses, as the event that engages it. A customary ceremony in a courtyard engages the Act exactly as a registered ceremony in a court does, because the harms the Act addresses flow from the social fact and not from the paperwork. This is why the offences can be committed at a wedding nobody recorded, why the injunction can be sought against a ceremony nobody has registered, and why the absence of a marriage certificate is no answer to a Section 3 petition, although it makes the proof of the marriage, like the proof of age, a matter for evidence. Registration matters to this Act as evidence and as prevention, as the discussion of Seema v. Ashwani Kumar below shows. It is not a condition of the Act’s application, and a couple married by rite alone are as fully inside the statute as a couple with a certificate in hand.

Voidable, not void: the central provision

Section 3 is the heart of the Act and the source of its namable claim. It provides that every child marriage, whether solemnised before or after the commencement of the Act, shall be voidable at the option of the contracting party who was a child at the time of the marriage. A petition for annulling the marriage by a decree of nullity may be filed in the district court only by that party. No one else holds the option. Not the parents of the other side, not the state, not a prohibition officer in the officer’s own name, and not the adult party to the marriage. The adult who married a child cannot use Section 3 to exit the marriage, and a family that arranged a marriage it has come to regret cannot invoke it either. The option is a personal right of the person the statute identifies as the victim of the event, and it is exercisable or not as that person chooses.

Voidable is a term of art and its consequences are concrete. A voidable marriage is a valid marriage for every purpose until a decree of nullity is passed. The parties are husband and wife in law. Their children are legitimate, and Section 6 confirms that position even after annulment. Property, maintenance and succession rights that depend on the existence of a marriage operate while the marriage stands. If the party who was a child never petitions, the marriage simply continues, ripening into an ordinary adult marriage that no court will disturb on the ground of the parties’ ages at solemnisation. The ceremony the Act prohibited thus produces, by default, the very status it was enacted to prevent, and the default is displaced only by litigation begun by the person with the least power in the arrangement.

The temporal reach of Section 3 is wider than readers expect. The words “whether solemnised before or after the commencement of this Act” make the voidability available for marriages performed under the restraint regime, so a girl married in 2005, when the old Act punished her family and said nothing about her status, acquired on 1 November 2007 a statutory option to annul that no law had given her when she was married. The window, however, runs from majority and not from commencement, so for an older cohort the option could expire soon after it was created, an interaction between the section’s generosity and its limitation that the drafting does not acknowledge.

The annulment window and how it closes

Section 3(3) provides that the petition may be filed at any time but before the child filing it completes two years of attaining majority. Majority under the Majority Act, 1875 arrives at eighteen. For a girl married at fifteen or sixteen or seventeen, the petition must therefore be filed before she turns twenty. For a girl married at thirteen, the same deadline applies, since the clock runs from majority and not from the marriage: she has from the wedding day until the eve of her twentieth birthday, but the years in which she could realistically act are the years in which she is a minor living inside the arrangement, and Section 3(2) meets that problem part of the way by allowing a petitioner who is still a minor to file through a guardian or next friend along with the Child Marriage Prohibition Officer. The officer’s presence in that sub-section is one of the few places where the preventive machinery and the status provisions touch, and it matters: a minor petitioner needs a next friend willing to litigate against, very often, her own family, and the statute pairs that fragile figure with a public officer.

For the male party the arithmetic of the same sub-section produces a drafting friction that commentary has repeatedly noticed. A man is a child under the Act until twenty-one but attains majority at eighteen. A man married at twenty has, on the words of Section 3(3), a window that closes when he completes two years of attaining majority, which happened at twenty, at or about the time of the marriage itself. Read on its own words, his option can be stillborn. The section was plainly drafted around the paradigm case of the under-age bride, for whom majority at eighteen and marriageability at eighteen coincide, and the male case was fitted into words designed for her. The 2021 bill proposed to extend the window from two years to five years after attaining majority, which would have eased the female deadline considerably and given the male case a workable margin as well; with the bill’s lapse the two-year text remains the law.

Two features of the petition deserve emphasis because they are routinely missed. The first is that the decree, when granted, is a decree of nullity and not a divorce. The distinction is developed in the consequences sections below, but its root is here: annulment under Section 3 declares the marriage avoided at the petitioner’s election, whereas divorce presupposes a valid marriage and dissolves it. The second is Section 3(4). While granting a decree of nullity, the district court shall make an order directing both parties and their parents or guardians to return to the other side the money, valuables, ornaments and other gifts received on the occasion of the marriage, or an amount equal to their value, after notice and an opportunity to show cause. The provision treats the transfers that accompanied the wedding as reversible along with the status, and it binds parents and guardians as well as the parties, which recognises where the property in a child marriage actually sits.

Does a child marriage end on its own when both parties grow up?

No. Nothing in the Act dissolves a child marriage by efflux of time, by the parties reaching the permitted ages, or by the birth of children. The marriage continues in force until the child party petitions a district court and obtains a decree of nullity, and the petition must be filed before two years run from attaining majority.

The design choice embedded in Section 3 has a defence, and the stance charter requires it to be stated at full strength before any verdict. The defence is that automatic voidness is a blunt instrument against the person it claims to protect. If every child marriage were void from the start, the girl would hold no status as a wife at any point, and with that status would go claims that depend on it: maintenance as a wife, residence in the matrimonial home, the legitimacy framework for her children in personal laws less careful than Section 6, and inheritance on the husband’s death. Jurisdictions beyond India have met the same difficulty and several have chosen voidability, or voidness softened by judicial discretion, for the same reason. A girl who wishes to stay, and there are marriages contracted at seventeen that both parties, as adults, affirm, is not forced by the state into a nullity she does not want. The option is hers, and an autonomy-respecting statute can defend putting it in her hands.

The objection answers the defence on its own ground. The statute could have made child marriages void while preserving the protective consequences by express provision, exactly as the void category it did enact preserves nothing and relies on other laws, and exactly as Section 6 preserves legitimacy notwithstanding annulment. The protective consequences are creatures of drafting, not of validity. What voidability adds to voidness-plus-protection is not protection; it is inertia. It places the burden of escaping the marriage on the person least able to bear it, inside a two-year window that runs while she is most dependent on the family that arranged the marriage, and it guarantees that the great majority of child marriages, the ones nobody petitions about, stand untouched. The Standing Committee that examined the 2004 bill made this point in terms, recommending voidness because of the inhibitions against approaching a court. Parliament heard the objection and chose the option model anyway.

The void category: where Parliament did refuse validity

Section 12 carves out the class of cases in which the Act does what its title promises. The marriage of a minor child is null and void where the child is taken or enticed out of the keeping of the lawful guardian; or is by force compelled, or by any deceitful means induced, to go from any place; or is sold for the purpose of marriage and made to go through a form of marriage; or, having been married, is afterwards sold or trafficked or used for immoral purposes. Three features of the section repay attention.

The first is that the section speaks of a minor child, borrowing the Majority Act line of eighteen through the definition in Section 2(f), rather than of a child in the Section 2(a) sense. The void category is therefore narrower than the child marriage category in a second way as well: a man of twenty who is enticed from his guardian’s keeping is a child under the Act but not a minor, and his marriage is voidable only. The second feature is the character of the listed circumstances. Each involves the child being moved, taken, deceived, coerced or traded. The void cases are, in substance, the abduction and trafficking cases, in which the ceremony is an incident of a crime against the child’s custody rather than a family arrangement the community endorsed. The third feature is the consequence. A marriage within Section 12 is null and void without any petition, any decree or any window. It never had legal existence, and a court asked about it declares a fact rather than exercising an option.

Section 14 adds the second void route, and it is procedural rather than circumstantial. Any child marriage solemnised in contravention of an injunction order issued under Section 13, whether interim or final, shall be void from the beginning. The logic is contemptuous in the strict sense: a marriage performed in defiance of a court’s prohibition cannot claim the shelter of the option model, because the parties and their families were warned by a court and proceeded anyway. Section 14 is the Act’s only bridge between its preventive machinery and its validity scheme, and it gives the injunction a consequence far beyond the penalty for disobedience in Section 13(10). A family that defies an injunction does not merely expose the adults to imprisonment; it destroys the status the ceremony was meant to create.

Why is a trafficked child’s marriage void when an ordinary child marriage is only voidable?

Because Parliament reserved voidness for the cases in which the ceremony is an incident of abduction, force, deceit or sale, where no family or community decision stands behind the marriage that a court might hesitate to undo. The ordinary child marriage, arranged by families, is left valid until annulled; the trafficked child’s marriage is denied legal existence from the start.

The resulting map is the article’s findable artifact. Read it from the circumstances of the marriage, which are the facts a prohibition officer, a magistrate or a family member will actually know, to the status that follows.

Circumstances of the marriage	Status under the Act	Who can act	Time limit	Consequences that follow
Ordinary child marriage, arranged without force, deceit, abduction or sale	Voidable	Only the party who was a child, by petition to the district court	Before completing two years after attaining majority	On annulment, maintenance and residence for the female party until remarriage, legitimacy of children preserved, custody by best interests, gifts and valuables returned
Minor taken or enticed out of the keeping of a lawful guardian	Void	No petition needed; nullity may be declared on the fact	None	The marriage is treated as never having existed; protection and support must be found under other laws, since the maintenance scheme of the Act is built around the annulment decree
Minor compelled by force or induced by deceitful means to go from any place	Void	No petition needed	None	As above; the adults involved also face the Act’s offences and the general law on abduction and trafficking
Minor sold for the purpose of marriage, or sold or trafficked after the marriage	Void	No petition needed	None	As above; the transaction itself attracts the trafficking provisions of the general criminal law
Child marriage solemnised in defiance of an injunction under Section 13, interim or final	Void from the beginning	The injunction is obtained by a prohibition officer, a complainant, or the magistrate acting on information	The injunction must precede the ceremony	The marriage is a nullity and disobedience of the injunction is separately punishable with imprisonment of up to two years or a fine of up to one lakh rupees or both
Marriage in which neither party is a child under Section 2(a)	Unaffected by the Act	Not applicable	Not applicable	The marriage stands or falls under the personal or general law that governs it; the Act’s offences and machinery are not engaged

The table exposes the Act’s structural secret. Validity under this statute is not a function of the parties’ ages alone. It is a function of ages plus circumstances plus procedure, and two marriages identical in the ages of the parties can occupy different rows because one was preceded by an injunction or accompanied by deceit. That is why the single most common error in writing about this Act, the flat statement that a child marriage is void, is not a simplification but a falsehood, and why the second most common error, that a child marriage is fully valid, is a half-truth that misses the option, the void rows and the decree’s consequences.

What annulment carries with it: maintenance, residence, custody, legitimacy

A decree under Section 3 does not stand alone. Sections 4 to 7 build a small protective code around it, and the code is directed almost entirely at the female contracting party, which is a further sign of the paradigm case the drafters had in mind.

Section 4 empowers the district court, while granting the decree, to make an interim or final order directing the male contracting party, or, where he is himself a minor, his parent or guardian, to pay maintenance to the female contracting party until her remarriage. The quantum is to be fixed having regard to the needs of the child, the lifestyle enjoyed by her during the marriage and the means of the paying party, and the sum may be ordered monthly or in a lump sum. Where the petitioner is the female party, the court may also make a suitable order as to her residence until her remarriage. The remarriage boundary is doing deliberate work in both limbs. The Act treats annulment as restoring the woman to the position of an unmarried person for the future while charging the dissolved marriage with her support in the interval, and it withdraws that support when a new marriage begins. The design is neither alimony for life nor a clean break; it is a bridge, and its length is set by her remarriage rather than by any period the court chooses.

Section 5 addresses the children of the marriage. Where there are children born of a child marriage, the district court shall make an appropriate order for their custody, and the welfare and best interests of the child are declared to be the paramount consideration. The custody order may include access for the other party and may be accompanied by an order for the child’s maintenance against a party to the marriage or the parties’ parents or guardians. Section 6 then removes the oldest cruelty of nullity law: notwithstanding that a child marriage has been annulled by a decree of nullity, every child begotten or conceived of the marriage before the decree is made, whether born before or after the commencement of the Act, shall be deemed a legitimate child for all purposes. The phrase “for all purposes” is as wide as statutory language gets, covering status, maintenance and succession. Section 7 completes the scheme by empowering the district court to add to, modify or revoke an order under Section 4 or Section 5 on a change of circumstances, at any time during the pendency of the petition and even after its final disposal, which keeps the protective orders adjustable in a way the decree itself is not. Section 8 fixes the forum generously, giving jurisdiction to the district court of the place where the defendant or the child resides, where the marriage was solemnised, where the parties last resided together, or where the petitioner resides on the date the petition is presented.

Does an annulment under this Act work like a divorce?

No. A divorce dissolves a marriage the law accepts as valid. A Section 3 decree of nullity instead avoids the marriage at the child party’s election, because the law will not hold anyone to a status entered in childhood. She is not a divorcee, and her Section 4 maintenance and residence orders run only until remarriage.

The limits of this protective code deserve to be stated as plainly as its contents. Every rupee of it depends on the petition. A girl who does not annul, because she does not know, cannot reach a court, or is pressed by both families to stay, receives nothing from Sections 4 to 7, and her maintenance position is thrown back on the general law. The code also runs against the male party or his family, so its real value in a poor household, where the families are often equally poor and the marriage joined two households of similar means, can be small. And the residence order protects her until remarriage in a house that may be the marital home she is leaving, a remedy easier to state in a section than to inhabit in a village. None of this makes the provisions worthless. It makes them conditional in a way that tracks the central design: the Act’s protections are the price of litigating, and litigating is the one thing the protected person is least placed to do.

The general position set out in this article is legal information about a statute, not advice on any individual marriage; the annulment window in Section 3 closes on a fixed date in each case, and a person inside that window should take advice on their own facts well before it does, because a petition filed late is not a petition a court can entertain. Anyone assembling the papers for such a petition, the proof of age, the marriage record, the notices and the correspondence, faces a document problem as much as a legal one, and keeping that file complete, ordered and private in a single encrypted workspace, the working discipline VaultBook sets out for journalists protecting their sources, is the closest model for how such a file should be held: one controlled copy, assembled carefully, shared on instruction rather than scattered across phones and inboxes.

The offences: who pays for the ceremony

The penal group of the Act is built in three concentric rings, moving outward from the groom.

Section 9 is the innermost ring. A male adult above eighteen years of age who contracts a child marriage is punishable with rigorous imprisonment which may extend to two years, or with a fine which may extend to one lakh rupees, or with both. The section is aimed at the adult bridegroom, the man who marries a girl below eighteen or, given the definition, a woman who is a child because the marriage joins her to a child party. Its severity should be measured against the statute it replaced: the restraint Act’s maximum for the corresponding conduct was three months of simple imprisonment. The move from simple to rigorous imprisonment, from three months to two years, and from a fine of a thousand rupees to one lakh, is the clearest numerical statement of the 2006 Parliament’s intentions available anywhere in the Act.

Section 10 is the second ring, aimed at the ceremony itself. Whoever performs, conducts, directs or abets a child marriage is punishable with rigorous imprisonment which may extend to two years and shall also be liable to a fine which may extend to one lakh rupees, unless that person proves that he had reason to believe the marriage was not a child marriage. Two features distinguish it from Section 9. The fine is cumulative rather than alternative, and the section carries a built-in exculpation: the priest, registrar or elder who reasonably believed the parties to be of age has a defence the bridegroom does not need because the bridegroom’s own age and knowledge are the substance of his offence. The word “abets” pulls the organisers into the same ring as the officiant.

Section 11 is the widest ring and the one that does the Act’s social work. Where a child contracts a child marriage, any person having charge of the child, whether as parent or guardian or in any other capacity, lawful or unlawful, including a member of an organisation or association of persons, who does any act to promote the marriage, or permits it to be solemnised, or negligently fails to prevent it from being solemnised, including attending or participating in it, is punishable with rigorous imprisonment which may extend to two years and a fine which may extend to one lakh rupees. Attendance is itself within the section, which means the guest list of a child wedding is, in law, a list of offenders. Section 11(2) then reverses the evidential burden in the way the brief for any prosecution under the section must reckon with: it shall be presumed, unless and until the contrary is proved, that where a minor child has contracted a marriage, the person having charge of that child has negligently failed to prevent it. A parent of a married minor does not wait to be shown negligent; the parent must prove the negative.

Three qualifications complete the penal picture honestly. The proviso to Section 11(1) states that no woman shall be punishable with imprisonment under the section, a limitation repeated in the proviso to Section 13(10) for disobedience of an injunction. A mother who promotes her daughter’s marriage is liable to the fine and not to prison, a gendered mercy that sits oddly beside the Act’s equality rhetoric and that the Karnataka amendment of 2017 deleted within that state. Section 15 makes every offence under the Act cognizable and non-bailable notwithstanding the Code of Criminal Procedure, 1973, which arms the police to arrest without warrant and removes bail as an entitlement, a procedural severity that becomes the centre of the enforcement argument examined later in this article. And Section 20 reaches into the Hindu Marriage Act, 1955, substituting clause (a) of its Section 18 so that breach of the age condition in Section 5(iii) of that Act is punishable with rigorous imprisonment of up to two years or a fine of up to one lakh rupees or both, aligning the personal law’s penalty with the prohibition Act’s scale. A marriage can thus be valid under the Hindu Marriage Act, which does not make breach of the age condition a ground of nullity by itself, punishable under its Section 18 as substituted, and voidable under Section 3 of this Act, all at once, which is as neat an illustration as the Indian statute book offers of the difference between validity, liability and option.

Does the consent of the child’s parents make a child marriage lawful?

No. Parental consent is no defence under the Act; it is the usual route into liability. The persons whose consent arranged the marriage are the persons Section 11 charges with promoting or permitting it, and Section 11(2) presumes their negligence once a minor is shown to have married. Consent explains how such marriages happen; it also identifies the accused.

Why does the Act punish the adults and never the child who was married?

Because the child party is the person the statute exists to protect, and criminalising the victim would destroy the annulment scheme built around her choices. The Act’s punishments therefore run outward through the adult groom, the officiant and the family, while the child party receives rights instead of liabilities: the option to annul, maintenance, residence and custody orders.

The asymmetry between Section 9 and the rest is worth a paragraph of its own, because aspirants and commentators both trip on it. Section 9 punishes a male adult who contracts a child marriage. It does not punish a female adult who marries a male child, although such a marriage is a child marriage within Section 2(b) and although she may be caught by Section 10 if she directs or abets, or by other laws. Nor does it punish either child party. The Act’s penal imagination is fixed on the standard case, an adult man and a girl, and its drafting follows the paradigm even where the definition does not. Whether that is a defect depends on the frequency of the non-paradigm cases, which the survey data discussed below suggests are not trivial: a substantial minority of child marriages involve grooms below twenty-one, many of them married to brides also below the line, in matches between two families and two children, where the only adults available to punish are the parents on both sides. In those cases Sections 10 and 11 carry the whole penal weight of the Act, and the presumption in Section 11(2) is not an ornament but the prosecution’s principal tool.

The preventive machinery: prohibition officers and injunctions

The restraint Act of 1929 had no machinery at all. The 2006 Act creates two preventive instruments, an officer and an order, and the 2024 judgment of the Supreme Court, examined later in this article, is largely an account of how far the first of them remained a paper institution.

Section 16 requires the State Government, by notification in the Official Gazette, to appoint Child Marriage Prohibition Officers for the whole state or for such part of it as the notification specifies. The state may request a respectable member of the locality with a record of social service, a panchayat or municipal officer, a government or public sector officer, or an office bearer of a non-governmental organisation to assist the officer, and such a person is bound to act accordingly. Section 16(3) sets out the duties as a connected job description. The officer is to prevent the solemnisation of child marriages by taking such action as the officer deems fit; to collect evidence for the effective prosecution of persons contravening the Act; to advise individuals and counsel the residents of the locality generally against promoting, helping, aiding or allowing child marriages; to create awareness of the evil which results from child marriages; to sensitise the community on the issue; to furnish periodical returns and statistics as the State Government directs; and to discharge such other functions as the state assigns. The state may invest the officer with specified powers of a police officer by notification, and Section 16(5) gives the officer standing to move the court for orders under Sections 4 and 5, to seek an injunction under Section 13, and to accompany the child in a petition under Section 3. Section 17 deems prohibition officers to be public servants, and Section 18 protects action taken by them in good faith from suit and prosecution.

The design is a prevention officer with a prosecution file, a counselling brief and a statistics duty, and its weakness is written into Section 16 itself. The Act nowhere requires the office to be a person’s whole charge. In state after state the notification appoints an officer who already holds a full-time post in the district administration, the education department or the women and child development machinery, and the prohibition work is an additional duty layered onto a diary that has no room for it. The Supreme Court’s 2024 directions, which insist that states appoint officers solely responsible for the district and not burdened with additional duties that impede the work, are directed at exactly this gap between the section’s text and its staffing. The gap is not a drafting error. Section 16 creates the office and leaves its resourcing to the same state governments whose local establishments were failing to prevent the marriages, and the statute provides no consequence when an appointment is nominal.

The second instrument is sharper. Section 13 empowers a Judicial Magistrate of the first class or a Metropolitan Magistrate, on the application of a prohibition officer, or on receipt of information through a complaint or otherwise from any person, to issue an injunction prohibiting a child marriage that has been arranged or is about to be solemnised, where the magistrate is satisfied that such a marriage in contravention of the Act has been arranged or is imminent. A complaint may be made by any person having personal knowledge or reason to believe, and by a non-governmental organisation having reasonable information. The court may also take suo motu cognizance on the basis of any reliable report or information. For the days on which mass child marriages are traditionally solemnised, the section names Akshaya Tritiya expressly: for preventing solemnisation of mass child marriages on certain days such as Akshaya Tritiya, the District Magistrate is deemed to be the Child Marriage Prohibition Officer with all the officer’s powers, and the District Magistrate has additional powers to stop or prevent such marriages, using the minimum force required. Procedure is guarded: no injunction is to issue without prior notice and an opportunity to show cause, except that in a case of urgency the court may issue an interim injunction without notice, and an injunction may be confirmed or vacated after hearing, or rescinded or altered on the court’s own motion or on the application of a person aggrieved. If an application is rejected in whole or part, reasons must be recorded in writing. Disobedience, by a person who knows an injunction has been issued against him, is punishable under Section 13(10) with imprisonment of either description for up to two years or a fine of up to one lakh rupees or both, subject to the same proviso sparing women imprisonment.

Section 19 leaves the rule-making power under the Act to the State Governments, and the rules are where the officer on the ground actually lives. State rules under this Act typically prescribe the manner of appointment and the area of a prohibition officer, the forms of the periodical returns the officer must furnish, the procedure for assisting officers drawn from panchayats, municipalities and non-governmental organisations, and the coordination arrangements with the police and the child protection machinery. The quality of those rules varies as state drafting always varies, but their structural significance does not: the Act’s Centre-written text is a shell of duties and powers, and the state rules decide whether the shell contains a functioning office with a defined beat and a reporting line or a notification pinned to an officer whose real duties lie elsewhere. When the 2024 judgment directed that prohibition officers be appointed as a sole district charge, it was directing states to use the appointment and rule powers they had held since 2007 in the manner the Act’s design always assumed and their own notifications had not delivered. The federal allocation in Section 19 is accordingly not a footnote to the Act. It is the place where the Act succeeds or fails, state by state, in a pattern no central amendment of the validity provisions can cure.

The injunction is the best-designed provision in the Act, and it is worth saying why. It operates before the harm, at the moment when intervention costs least and achieves most. It is triggered by information rather than by a victim’s complaint, so it does not depend on the child. It can be obtained urgently and without notice where the ceremony is imminent. It carries its own teeth, in the disobedience offence, and through Section 14 it carries a consequence no other preventive order in Indian social legislation carries: the marriage held in its defiance is void from the beginning. A prohibition officer who learns of a wedding on a Thursday and obtains an interim injunction on the Friday has not merely created a prosecutable event for the Saturday; the officer has determined the status of anything that happens on the Saturday. The whole article could be reorganised around the observation that the Act works best where it is used earliest, and that its annulment scheme, the part everyone reads, is the machinery for its failures.

The marriage under three laws at once

A child marriage in India does not live under one statute. It lives, at a minimum, under three, and each of the three answers a different question.

This Act answers the question of prohibition. It declares the solemnisation punishable, builds the prevention machinery and gives the child party the option of annulment. The personal law that governs the parties, very often the Hindu Marriage Act, 1955, answers the question of validity in its own terms. Section 5(iii) of that Act makes eighteen for the bride and twenty-one for the bridegroom a condition of a valid Hindu marriage, but breach of that condition is not among the grounds on which the marriage is void under Section 11 or voidable under Section 12 of the same Act. The marriage is therefore neither void nor voidable under the Hindu Marriage Act for the age defect alone; it is a marriage that satisfies the validity provisions while carrying a penal consequence, since Section 18 of that Act, as substituted by Section 20 of the prohibition Act, punishes breach of the age condition with rigorous imprisonment of up to two years or a fine of up to one lakh rupees or both. The Special Marriage Act, 1954 takes the same structure with its own condition in Section 4(c), eighteen for the bride and twenty-one for the bridegroom, and the Indian Christian Marriage Act, 1872 works at the same ages through its own provisions. The prohibition Act then answers the option question that none of the personal laws asks, and the criminal statutes answer the liability question around the ceremony and the conjugal relationship. The result is a legal position that has to be assembled statute by statute and cannot be read off any one of them, which is why a family that has checked the position under its personal law may still be in breach of the prohibition Act, and a couple that has secured the prohibition Act’s position, by neither party being a child within its definition, may still be in breach of the criminal law if the bride is below eighteen.

Marriage registration adds the fourth layer and the prevention system’s best ally. In Seema v. Ashwani Kumar, (2006) 2 SCC 578, decided in February 2006, the Supreme Court directed the states and Union territories to make the registration of marriages compulsory, observing that registration would furnish the best evidence of marriage and would deter, among other things, the solemnisation of marriages involving minors. The prohibition Act does not itself require registration, and Section 19 leaves rules to the states, so the registration position varies across the country. Where registration is compulsory and enforced, a registrar can demand age proof before recording the marriage, which converts the preventive moment into an administrative checkpoint rather than a criminal prosecution. Where it is not, the Act must wait for an informant or a prohibition officer to learn of a private ceremony. The difference between those two situations is the difference between preventing the ceremony and prosecuting it, and the states that take registration seriously report the injunctions that other states have to obtain from magistrates after the drums are already sounding.

How the annulment petition works in practice

The Section 3 option is easy to state and hard to use, and the difficulty deserves a paragraph of its own because it is the practical test of the entire design.

The petition goes to the district court, which includes the Family Court where one exists and otherwise the principal civil court of original jurisdiction, and Section 8 gives a generous choice of forum: the court where the defendant or the child resides, where the marriage was solemnised, where the parties last resided together, or where the petitioner resides when the petition is presented. A petitioner who is still a minor files through a guardian or next friend along with the Child Marriage Prohibition Officer, a provision that pairs a child litigant with a public officer against, in the typical case, both sets of parents. The proof that matters most is age, and age in India is often the fact least reliably documented. The petition must establish that the petitioner was a child at the time of the marriage, and where birth records, school admission registers and the parties’ assertions conflict, the court must determine age on the evidence, a forensic exercise in which ossification tests, electoral rolls and admission forms pull in different directions and in which the family’s interest in the outcome is usually visible. The Supreme Court has, in other statutes where age determines jurisdiction or liability, emphasised the primacy of documentary proof of age, and a district court trying a Section 3 petition works under the same discipline with none of the institutional support of a criminal trial.

The window compounds the difficulty. The petition must be filed before the petitioner completes two years of attaining majority. For the girl married at sixteen, the litigation must be conceived, funded and filed before her twentieth birthday, in a setting where she typically lives in the household that arranged the marriage, where her mobility, her literacy and her knowledge of the remedy are all controlled by the respondents, and where the family’s response to the prospect of litigation may range from persuasion to coercion. The statute gives her no state counsel as an entitlement and no legal aid trigger in its own text, and no interim protection order between filing and decree except what the district court can fashion under its general powers. The prohibition officer who is supposed to accompany a minor petitioner is, as the 2024 judgment found, often a nominal appointment with a full-time other job. Against those facts, the generosity of Section 3(1) in covering marriages solemnised before the Act’s commencement reads as an irony: the statute reaches back to marriages the restraint Act left untouched and offers them a remedy that must be claimed by the person with the least capacity to claim it.

Where the petition succeeds, the decree is of nullity and the court turns, under Sections 4 and 5, to maintenance, residence, custody and the return of gifts, as this article’s earlier sections set out. Where it fails on limitation, it fails completely; the section has no condonation provision and no power in the court to extend the window. A petition filed a day after the deadline is a petition the court cannot entertain, and the marriage that survived to that day is a marriage for all purposes thereafter. The voidable design thus ends, in the cases that matter most, not in a judicial decision on the merits but in the quiet expiry of a deadline that the person it protects may never have known existed.

The criminal law around the married child

The Act does not operate alone around a married child, and two collisions with the general criminal law define the practical position more sharply than anything in the Act itself.

The first collision is with the age of consent. The Protection of Children from Sexual Offences Act, 2012 defines a child as any person below eighteen and contains no exception for marriage, so a husband’s sexual relations with his wife of sixteen or seventeen fall within that statute’s offences whatever the validity of the marriage under this Act; the POCSO Act’s own architecture, its offences, its presumptions and its reporting duties, is the subject of its own pillar in this series and is compressed here deliberately. The second statute in the collision was older. Exception 2 to Section 375 of the Indian Penal Code excluded from rape sexual intercourse by a man with his own wife where the wife was not under fifteen, a line that allowed the penal code to treat as lawful within marriage what the protective statutes treated as an offence against a child.

In Independent Thought v. Union of India, (2017) 10 SCC 800, decided on 11 October 2017 by a bench of Madan B. Lokur and Deepak Gupta JJ., the Supreme Court read the exception down, holding that it applies only where the wife is not under eighteen, and thereby harmonised the penal code with the child marriage Act, the juvenile justice statute, the sexual offences statute and the domestic violence statute, all of which treat a girl below eighteen as a child. The Court was careful about the limits of what it did. It did not strike the exception down. It did not address marital rape of an adult wife, a question it expressly left alone. It performed a definitional repair, refusing to let one surviving figure in an old exception contradict the considered age line of the whole protective field. The marital exception’s passage into the Bharatiya Nyaya Sanhita, whose own exception to the rape provision in Section 63 adopts the age of eighteen in terms, converted the reading down into enacted text, so the position the Court reached by interpretation is the position the new code states on its face.

The practical consequence is a striking layering of liabilities around a single wedding. A man of twenty-four who marries a girl of sixteen commits the offence in Section 9 of this Act. His sexual relations with her are offences under the sexual offences statute and, after Independent Thought and under the new code, rape under the general law, the marriage notwithstanding. Her parents are presumed under Section 11(2) to have negligently failed to prevent the marriage. The officiant is liable under Section 10 unless he proves a reasonable belief in her age. And through all of it the marriage itself is not void: it is voidable at her option under Section 3, and unless she petitions before she completes two years of attaining majority, it subsists. Indian law can therefore describe one household, without contradiction, as a marriage, a rape, a trafficking risk and a maintenance obligation, depending on which statute is asking the question. That layering is the true subject of the proposition that child marriage law is incoherent, and the honest answer is that the incoherence is not in the drafting of any single Act but in the sequence: each statute was written against the failure of the last, and none was given the job of reconciling the set.

The second collision is with personal law, and on this one the law remains genuinely unsettled. The Act’s text does not say that it overrides personal law. Section 1 applies it to all citizens of India, and the offences and machinery plainly operate across communities, but the question whether the validity scheme of Sections 3 and 12 displaces a personal law under which a marriage involving a party below the statutory ages may be recognised has divided the High Courts. One line of decisions, visible in rulings of the Punjab and Haryana High Court among others, has treated marriages under Muslim personal law as governed by that law’s own conditions where the parties have attained puberty, and has protected such marriages against interference while leaving the penal provisions to operate. Another line, in other High Courts, has treated the prohibition Act as a secular statute of general application whose scheme governs the marriage itself. The conflict is real, it is documented, and it produces different answers to the same question in different states. In the 2024 judgment the Supreme Court was asked by the Union government to declare that the Act prevails over personal law, and it declined, noting that the Act does not itself address the matter squarely and that the 2021 bill, which contained an express overriding clause, was the vehicle Parliament had chosen for that clarification. The Court left the question to Parliament. Parliament, by then, had let the bill lapse. The honest statement of the position is therefore a refusal stated twice over: the question is open, the Court has declined to close it, and the legislative instrument drafted to close it expired without a decision.

The bill that expired: raising the age for women to twenty-one

The centrepiece of the Act’s later history is a bill that never became law, and its record has to be stated precisely because a remarkable number of published accounts state it wrongly, describing the Prohibition of Child Marriage (Amendment) Bill, 2021 as pending, as passed, or as having raised the marriage age already.

The bill’s origin was a task force. In June 2020 the Ministry of Women and Child Development constituted a task force to examine matters pertaining to the age of motherhood, the imperatives of lowering maternal mortality, the improvement of nutritional levels and related issues, chaired by Jaya Jaitly and including, among others, V. K. Paul of NITI Aayog and officials of the concerned ministries. The task force reported in December 2020 and recommended that the minimum age of marriage for women be raised to twenty-one, alongside recommendations on education, including the formal introduction of sex education in the school curriculum. On 15 December 2021 the Union Cabinet approved the proposal. On 21 December 2021 the Minister of Women and Child Development, Smriti Irani, introduced the Prohibition of Child Marriage (Amendment) Bill, 2021 (Bill No. 163 of 2021) in the Lok Sabha, against opposition protests about the speed of the introduction and the absence of prior consultation, and the government itself proposed the reference that followed.

The bill’s contents were fourfold. It would have amended the definition in Section 2(a) so that a child means a person, male or female, who has not completed twenty-one years, equalising the marriage age at twenty-one for both sexes. It would have extended the annulment window in Section 3 from two years to five years after attaining majority. It would have provided that the Act as amended shall have effect notwithstanding anything inconsistent contained in any other law for the time being in force, or any custom, usage or practice governing the parties, the overriding clause whose absence the Supreme Court later noticed. And it would have made consequential amendments across the marriage statutes, the Indian Christian Marriage Act, 1872, the Parsi Marriage and Divorce Act, 1936, the Muslim Personal Law (Shariat) Application Act, 1937, the Special Marriage Act, 1954, the Hindu Marriage Act, 1955 and the Foreign Marriage Act, 1969, so that the age of twenty-one ran uniformly through the law of marriage. The stated object was the constitutional mandate of gender equality in marriageable age, the bill’s preface observing that women are disadvantaged in higher education, vocational training and psychological maturity by early marriage and that the disadvantage perpetuates dependence.

The bill was referred to the departmentally related Standing Committee on Education, Women, Children, Youth and Sports, chaired by Vinay Sahasrabuddhe. The committee’s composition itself became part of the argument: of its thirty-one members drawn from both Houses, Sushmita Dev was, at the time of its constitution for this examination, the sole woman, a fact she and other women parliamentarians pressed publicly, with the chair responding that the committee would examine the bill thoroughly and invite stakeholders widely. The committee invited public views, and the public answered at a volume with few precedents in committee practice. By April 2022, PTI reported from committee sources that the panel had received around 95,000 email representations on the bill, that nearly 95 per cent of them opposed raising the age, and that some 90,000 of the opposing emails carried identical content and appeared to have been drafted from a single source, a circumstance that led sources in the committee to suspect a coordinated campaign to defeat the bill. Each element of that report matters and none should be dropped. The volume was extraordinary. The opposition share was overwhelming. And the identical content of the great bulk of the opposition is a fact of the record that the bill’s supporters used to discount the consultation and its critics answered by observing that form responses are the ordinary currency of organised public mobilisation, on every side of every question, and that a committee counts arguments rather than weighing envelopes.

The committee heard evidence. Members of the Jaya Jaitly task force deposed before it. Senior officials of the Health Ministry, including the Health Secretary, Rajesh Bhushan, supported the proposal, telling the panel that early marriage and early pregnancy adversely affect the health of mother and child and that raising the age could assist in reducing infant and maternal mortality and child stunting and wasting. Non-governmental organisations divided in their depositions in much the same proportions as the wider debate, with bodies such as the Kailash Satyarthi Children’s Foundation and Lado Panchayat supporting the bill before the panel. And then the committee’s examination simply continued. It received a succession of extensions of time, running over two years, through 2022, 2023 and into 2024, without presenting a report to either House. The examination of a five-clause amendment bill outlasted the remainder of the Parliament that had referred it.

The end came by operation of the Constitution and not by any decision. Article 107(5) provides that a bill pending in the House of the People, or which having been passed by the House of the People is pending in the Council of States, shall lapse on a dissolution of the House of the People. The seventeenth Lok Sabha was dissolved in June 2024 following the general election. The bill, pending before a committee of that House, lapsed with it. The former Secretary-General of the Lok Sabha and constitutional authority P. D. T. Acharya confirmed the lapse publicly at the time. The Ministry of Women and Child Development, answering a question in the Lok Sabha after the lapse, confirmed the sequence from the government’s side: the bill had been introduced on 21 December 2021, referred to the committee, and had lapsed on the dissolution, and no report was received from the committee. Accounts that continue to describe the bill as under examination by the standing committee are therefore describing a committee that no longer holds it and a bill that no longer exists in the parliamentary system. To revive the proposal, a government must introduce a fresh bill in the eighteenth Lok Sabha and begin again. The episode is the cleanest recent case study in this series of how parliamentary committees scrutinise bills, and of what happens to scrutiny when the calendar, rather than a conclusion, ends it.

The series thesis deserves a paragraph of application. This series has criticised, in cluster after cluster, the habit of passing consequential legislation without committee scrutiny, and the criticism stands. The marriage age bill is the mirror image and the caution against reading the first criticism too simply. Here scrutiny was not skipped. A task force consulted and reported. A bill was introduced and referred. A committee invited the public and was answered by the public at a scale that committees rarely achieve. Evidence was taken from ministries, experts and civil society on both sides. Every institutional virtue the scrutiny model asks for was present except one: an ending. The process generated a record and no outcome, neither enactment nor rejection, and the underlying questions, the age, the window, the override, returned to exactly the position they occupied in December 2021, with the addition of a public consultation whose contents nobody in authority ever had to answer. A Parliament that cannot finish examining a five-clause bill in two and a half years has not scrutinised it more deeply than a Parliament that passes a bill in a week. It has merely declined, at greater length and with better documentation.

The age argument, both cases at full strength

Because the bill lapsed without a report, neither side of the age argument was ever formally adjudicated, and this article will not pretend otherwise. What it can do is state each case as its serious holders make it.

The case for twenty-one, as made by the Jaya Jaitly task force and by the government that adopted its recommendation, rests on three legs. The first is health. Early marriage is followed, in the population the survey data describes, by early pregnancy, and early pregnancy carries elevated risks of maternal death, infant death, stunting and wasting; the Health Ministry’s senior officials put this evidence to the standing committee in terms, linking the marriage age to infant mortality, maternal mortality and child nutrition. The second leg is education and economic independence. A legal age of eighteen cuts across the years in which a woman might complete higher secondary and tertiary education, and the bill’s own statement of objects argued that early marriage disadvantages women in education, vocational training and the acquisition of skills and maturity, perpetuating dependence on men. The third leg is equality. A law that permits a woman to marry at eighteen and requires a man to wait until twenty-one writes an assumption about the proper shape of a marriage into the statute book, and Smriti Irani, introducing the bill, framed it as giving women the right to equality in marriage across all religions and castes. On this view the age gap is not a protection but a relic, and the fact that the gap originated in assumptions about a younger bride is a reason to remove it rather than a reason to keep it.

The case against, as made by the overwhelming majority of the consultation respondents and by many women’s organisations and researchers, also rests on three legs, and they are not the legs of an argument for child marriage, which nobody in the debate defended. The first is criminalisation. Raising the female age to twenty-one converts every marriage of a woman of eighteen, nineteen or twenty into a child marriage, exposing her husband to Section 9, her parents to Section 11 and its presumption, and the couple to the cognizable and non-bailable machinery of Section 15, in a country where the survey data records that marriages in the eighteen to twenty band are common and frequently chosen. Researchers such as Madhu Mehra of Partners for Law in Development have argued, in the cognate debate on the age of consent, that protective criminal statutes in India are used most energetically not against predators but against couples who have married or eloped against family wishes, with the girl’s family deploying the law to recover a daughter who chose her partner. A higher marriage age enlarges the class of marriages families can attack on that pattern. The second leg is evasion. A marriage pushed below the legal age does not disappear; it is solemnised without registration, outside the systems through which the state distributes maternity benefits, institutional delivery support and legal protection, leaving the woman worse documented and worse defended than a registered wife of nineteen. The third leg is sequence. The opponents argued that the drivers of early marriage are poverty, the absence of secondary schooling within reach, and the economics of dowry and household labour, and that a state which has not universalised girls’ secondary education is legislating the symptom. Asaduddin Owaisi put the political form of the objection in the introduction debate: an eighteen-year-old can vote, so the claim that she is a child for the single purpose of marriage requires more justification than the bill supplied. Supriya Sule and other members pressed the procedural form, that a change of this reach was introduced without consulting the stakeholders it governed.

Two features of the consultation require honest handling because they cut against easy morals on both sides. The 95 per cent opposition figure is a fact of the committee’s inbox and not a survey of India; it measures mobilisation, and the identical drafting of some 90,000 emails shows the mobilisation was organised, which the bill’s supporters in the committee read as a campaign to manufacture a verdict and the bill’s opponents read as citizens using the only channel the committee offered them. Both readings are available on the record and neither cancels the other. What cannot be said is that the public was not consulted, or that those who answered were in favour. They were consulted, they answered in very large numbers, and they were opposed, and a Parliament that invites views and then never reports has left those citizens with a particular ground for cynicism about the next invitation.

The verdict on the age question that the evidence supports is narrow. The health and education evidence against marriage in the teens is strong, and it is evidence against the marriages the Act already prohibits below eighteen. The case for moving the line to twenty-one was not shown, in the committee record that became public, to survive its own enforcement consequences, and the strongest single datum against the proposal is the one its supporters least liked to discuss: the law would have criminalised, in the first instance, the marriages of adult women of eighteen to twenty who under the Majority Act, the franchise and every other civil statute are adults. A reform whose principal new class of offenders is the husbands of adult women carries a burden of justification that committee evidence on maternal health, directed at pregnancies in the middle teens, does not discharge. That is a judgment about the bill that lapsed. It is not a judgment that the present split ages are principled; they are inherited, as the definition section of this article showed, and a Parliament that wished to equalise at eighteen, the course the Standing Committee of 2004 recommended, would be answering the equality argument without enlarging the criminal law. That possibility was never the subject of the lapsed bill, and it remains unexamined in any enacted text.

The personal law conflict was predictable from the text

The conflict that the Supreme Court declined to resolve in 2024 was not an accident of litigation; it was built into the enactment of 2006, and understanding why requires seeing what the Act did not say.

The Act defines a child marriage and attaches penal and procedural consequences to it, but it does not declare that it overrides any other law. It contains no non-obstante clause of general application. The marriage statutes it touched were left exactly as they stood except for the single amendment made to Section 18 of the Hindu Marriage Act. The Hindu Marriage Act’s own conditions for a valid marriage, including the age condition in Section 5(iii), do not treat a breach of the age condition as making the marriage void, and classical Muslim personal law has long recognised the capacity of a person who has attained puberty to marry, which generates the friction: a marriage of a girl of sixteen may be treated as valid under the personal law applicable to the parties, punishable under the prohibition Act, annullable under the prohibition Act, and an offence, in its consummation, under the sexual offences statute. Each proposition is defensible within its own statute. The system as a whole has never been made to answer for their combination.

Before the prohibition Act, Hindu law itself had moved away from validity. The Hindu Marriage Act penalises rather than voids the under-age marriage, a position that survived from an era when the same legislature was raising ages in the restraint Act and declined to visit nullity on marriages its own communities treated as sacramental and indissoluble. The 2006 Parliament had every opportunity to sever the arrangement. The Standing Committee that examined the 2004 bill had recommended both a common age and voidness; the government that introduced the final bill retained the split ages and the option model; and on the personal law interface it chose silence. Silence, in a federal system where personal law is politically defended and where the Concurrent List lets both levels of legislature occupy the field, is itself a legislative position, and it has held for nearly two decades: the offences run against everyone, the annulment option runs for everyone, and the anterior question, whether the marriage is to be recognised at all where personal law validates it, is answered differently by different High Courts and awaits a Parliament that has already drafted the answer once and let the drafting lapse.

The 2021 bill’s overriding clause was therefore not a drafting flourish. It was the only part of the bill that addressed a question no amendment of Section 2 could reach. Its lapse did not merely postpone the age change; it removed from the parliamentary pipeline the only instrument that would have told the High Courts, litigants and registrars of India whether the prohibition Act is a penal statute riding on top of the personal laws or a law of marriage validity superior to them. The Supreme Court’s 2024 refusal to supply the answer by judgment is best read as the Court returning the question to the institution that drafted it, drafted it correctly, and then failed to vote on it.

The agency defence of voidability, stated fully

One further defence of the option model must be recorded, because it comes not from defenders of child marriage but from within the women’s movement, and because it qualifies the verdict this article reaches.

The defence is that many marriages involving a party below eighteen are not, in social fact, transactions imposed on a passive victim. A significant number are unions the young couple chose themselves, against the wishes of one or both families, and the criminal and status machinery of protective law has become one of the instruments by which families retrieve daughters who have eloped. Researchers who study the working of the sexual offences statute have documented, in case after case, complaints filed by parents against the partner their daughter chose, with the daughter’s own testimony running against the prosecution. A voidness rule gives those families a further weapon: the marriage their daughter chose is a nullity, and she holds no status in it at all, no maintenance, no residence, no widow’s claim, unless the statute builds the protections back in. The agency defence does not deny that girls married at fourteen need the law’s protection. It insists that the population of the under-eighteen married is not one population, and that a single status rule, whether void or valid, will injure some of the people it governs. Voidability, on this account, is the only rule that lets the woman herself decide which injury she prefers, and the burden objection answers it only by pointing to the failures of the machinery supposed to help her exercise the choice.

The answer available to the critics is the one this article has already given: protections can be carried across to voidness by express drafting, and other legal systems that void child marriages commonly preserve the woman’s financial claims and the children’s legitimacy in the same text. The agency defence survives that answer in one narrowed form. It is a defence not of voidability but of a judicial discretion: a power in the court, where a marriage within the prohibited ages was the free choice of parties close to the line, to decline nullity or to mould the consequences, so that the law distinguishes the coerced child bride of fourteen from the eloping woman of seventeen and a half without pretending they are the same case. Neither the statute of 2006 nor the bill of 2021 contains such a discretion, and the debate about this Act is poorer for arguing between two rigidities, the marriage that cannot be escaped except by petition and the marriage that never existed at all, when the question the cases actually present is which consequences should follow for this marriage, on these facts, at the instance of this woman.

The Supreme Court in 2024: prevention over prosecution

On 18 October 2024 a bench of the Supreme Court consisting of the Chief Justice, D. Y. Chandrachud, and Justices J. B. Pardiwala and Manoj Misra delivered judgment in Society for Enlightenment and Voluntary Action v. Union of India, Writ Petition (Civil) No. 1234 of 2017, reported at 2024 INSC 790. The petition, brought by a non-governmental organisation, complained that child marriages continued at scale despite the Act and sought effective implementation. The judgment, authored by the Chief Justice, is the most consequential judicial treatment of the statute since Independent Thought, and its centre of gravity is the opposite of the penal one. The Court held that child marriage violates the rights of the child, including the right to life and personal liberty, to education and to choose a partner, and it stated the governing proposition of its directions in terms: the aim of the enforcement machinery must not be solely to increase prosecutions without making the best efforts to prevent and prohibit child marriage, because a penalisation approach waits for the harm to occur before acting and has proved ineffective at bringing social change. The Court was equally careful to add that it should not be understood as discouraging prosecution of those who commit offences under the Act.

The directions run across the whole machinery. On legal enforcement, the Court directed state governments and Union territories to appoint Child Marriage Prohibition Officers solely responsible for discharging the functions of the office at the district level, not burdened with additional duties that impede the focus on prevention, with accountability for default; it fixed district-level responsibility for active prevention on the district administration, and directed the establishment of specialised police capacity, including a Special Child Marriage Prohibition Unit, for cases under the Act. On judicial measures, it empowered and encouraged magistrates to take suo motu action and issue preventive injunctions, directed exploration of special fast-track courts for child marriage cases, and required mandatory action against public servants who neglect their duties. On community involvement, it directed annual action plans, community-centric capacity building, and adoption of the Child Marriage Free Village initiative, under which panchayats and community leaders take responsibility for prevention in their area. On awareness, it directed campaigns led by prohibition officers in schools, religious institutions and panchayats, and attention to sexuality and relationship education. On training, it required capacity building for law enforcement, judicial officers, teachers and health workers. A copy of the judgment was directed to be circulated through the Ministry of Women and Child Development to the chief secretaries of all states and Union territories, the National Legal Services Authority and the National Commission for Protection of Child Rights for compliance.

Two further holdings of the judgment matter for the statute’s open questions, and both are holdings about what the Court would not do. The Court observed that the Act is silent on the status of a child betrothal, a marriage fixed during minority for celebration later, and suggested that Parliament consider prohibiting child betrothals expressly, treating a betrothed child as a child in need of care and protection under the juvenile justice statute in the interim analysis. And it declined the Union government’s request for a declaration that the Act prevails over personal law, noting the confusion in the High Court case law, noting that the Act does not itself contain an overriding provision, and leaving the question to Parliament, whose 2021 bill with its express override clause had by then lapsed. A reader who reports the 2024 judgment as having settled the personal law question has misread it in the direction of the government’s submission. The Court settled implementation. On validity against personal law it settled nothing, deliberately.

Read alongside Independent Thought, the 2024 judgment completes a judicial pattern worth naming. The courts have repaired the Act’s edges where repair was possible by interpretation, aligning the rape exception with the protective age line in 2017, and have governed its administration by direction where the failure was administrative, in 2024. The two questions that define the Act’s architecture, voidability and the personal law interface, are questions interpretation cannot reach, because the first is the enacted text’s plain choice and the second is a conflict the text does not address. Both are waiting in the same place, in a Parliament that has had the drafting for one of them on its table and let the table be cleared.

The betrothal gap the Court named

One boundary of the Act deserves its own treatment because the Supreme Court marked it expressly in 2024: the Act prohibits the solemnisation of a child marriage and says nothing about betrothal, the engagement or fixing of a marriage during childhood for celebration at a later date.

The gap is not technical. In communities where marriages are arranged years ahead, the betrothal is the moment at which the child’s future is settled, schooling decisions are taken in its shadow, and the social fact the Act opposes is created, while the ceremony the Act prohibits may be scheduled for after the parties reach the statutory ages and therefore never engage the Act at all. A girl betrothed at twelve and married at eighteen and a half has never been the subject of a child marriage within Section 2(b), yet the substance of the harm, the foreclosure of choice during childhood, is complete. The Court in Society for Enlightenment and Voluntary Action treated betrothal as the statute’s conspicuous silence, observed that marriages fixed during the minority of a child violate free choice, suggested that Parliament consider prohibiting child betrothals, and indicated that a betrothed child may fall within the category of a child in need of care and protection under the juvenile justice statute, where the protection system can act on the child’s situation without waiting for a ceremony.

The suggestion exposes a design limit of the entire prohibition model. A betrothal prohibition would have to define the act it forbids, an agreement, a ceremony of engagement, a public announcement, and each definition invites the evasion of informality, since the families concerned can settle a match by word alone and leave nothing for a prohibition officer to find. The Court’s suggestion is therefore best understood as directional rather than drafted: it tells Parliament where the practice lives, in the years before the ceremony, and it tells the protection system to treat the betrothed child as within its care jurisdiction in the meanwhile. A legislature that takes the suggestion up will have to decide whether betrothal is to be an offence, a ground for protective intervention, or a factor in schooling and welfare decisions, and the choice among those is a choice about how far the criminal law should reach into family arrangement, the same choice the Act of 2006 made conservatively and the bill of 2021 made expansively and neither time with a parliamentary decision that survived. The betrothal question, the personal law question and the validity question stand together as the Act’s three unfinished items of business, all three identified, all three drafted or draftable, and all three awaiting a House that stays in session on the subject long enough to vote.

The wedding calendar and the mass marriage problem

The Act’s preventive machinery was designed for a particular social fact: most child marriages are not secret. They are public ceremonies, announced in advance, performed before the community, and concentrated, in several regions, on days the calendar itself advertises. Akshaya Tritiya is the named example in Section 13(4), a day regarded as auspicious for weddings across western and northern India, on which mass marriage ceremonies involving hundreds of couples have historically been held, some of them organised by community associations and some attended by public representatives. The section’s answer is to deem the District Magistrate to be the Child Marriage Prohibition Officer for the purpose of preventing mass solemnisation on such days, with additional powers to stop or prevent the marriages and to use the minimum force required.

The mass marriage setting reverses every assumption of the ordinary case. The families are not concealing an offence; they are performing a norm. The officiants are not hiding; they are booked. The evidence problem that Section 11’s presumption exists to solve does not arise, because nobody disputes what is happening. What arises instead is a legitimacy problem: the district administration that arrives to stop a hundred weddings on the community’s most auspicious morning is not enforcing a settled consensus against a deviant few but imposing the statute’s norm on a community that has not internalised it, in public, at the moment of maximum emotional investment. This is the setting in which the Supreme Court’s 2024 insistence on prevention, awareness and community engagement, and on the Child Marriage Free Village idea, has its sharpest application. A district that has worked through the year with its panchayats, schools and religious institutions does not need to choose, on the morning of the festival, between the statute and a confrontation. A district that has not done the work has no good options left by then, and the injunction, obtainable urgently and without notice, is the least bad of them.

The single marriage in a courtyard and the hundred marriages on a maidan are the same offence under Sections 10 and 11, but they are different problems for the state, and the Act knows it: the injunction and the District Magistrate provisions are addressed to the public case, the annulment and the offences to the private one. The enforcement record, where it exists in public reporting, reflects the distinction. Prevention claims in state after state speak of marriages stopped, which in the mass setting means ceremonies negotiated down, ages verified at registration desks and couples turned away before the event, while prosecution claims speak of the Assam pattern, arrests after ceremonies performed years earlier. The first record is the statute working as designed. The second is the statute being used as the nearest available instrument for a political commitment, and the difference between them is the difference the 2024 judgment asked the enforcement machinery to learn.

The comparative frame: how other systems answer the validity question

The void or voidable question is not an Indian eccentricity. Every system that prohibits marriage below an age must decide what the prohibition does to a marriage that happens anyway, and the answers cluster in the same places.

England and Wales, whose law long permitted marriage at sixteen with parental consent, settled the question by the Marriage and Civil Partnership (Minimum Age) Act 2022, which raised the minimum age to eighteen without exception and provided that a marriage involving a person below eighteen is void, alongside an offence of arranging the marriage of a child. The reform was driven by campaigners who argued, in terms that map directly onto the Indian debate, that the consent exception had become the route by which coerced marriages were laundered into lawful ones, and that a bright line at eighteen with voidness attached was the only rule families and registrars could not negotiate around. The Indian legislature considered the same bright line in 2021 and chose twenty-one rather than eighteen, then did not enact it. The comparison is instructive in both directions: it shows that the voidness half of the English package is severable from the age half, and that a system can adopt voidness at its existing ages, which is the course the 2004 Standing Committee in India had already recommended on domestic reasoning alone.

Nepal’s law treats a marriage below the statutory ages as voidable in some bands and void in others, and Bangladesh’s Child Marriage Restraint Act 2017 retained voidability for the general case while adding a much-criticised special circumstances exception that permits marriage below the age with court and parental approval in what the statute calls the best interest of the minor, a provision its critics describe as legislating the loophole the Act exists to close. The United States offers the federal version of the problem: marriage ages are state law, most states set eighteen with exceptions historically available for younger teenagers with parental or judicial consent, and reform campaigns have moved state after state toward an unqualified eighteen by removing the exceptions one legislature at a time. The pattern across jurisdictions is consistent enough to name. The exceptions collapse first, because each exception becomes the standard route. The validity question is decided next, and systems that reach it tend to voidness with protected consequences, because voidability’s burden problem presents identically in every legal culture that has examined it. India has collapsed no exceptions, because its Act contains none to collapse, and has not reached the validity question since 2006. It is, on the comparative record, a debate behind rather than a debate apart.

The international instruments point in the same direction without deciding the domestic question. The Convention on the Rights of the Child fixes childhood at eighteen, and the joint general recommendation of the Committees on the Rights of the Child and on the Elimination of Discrimination against Women treats eighteen as the minimum marriage age for both sexes and child marriage as a harmful practice to be eliminated, which supplies the standard against which a split line of twenty-one and eighteen is awkward and a line of twenty-one for both is defensible on equality grounds even where its enforcement is not. India acceded to the Convention in 1992 with declarations on other articles. The instruments do not answer the criminalisation objection, which is a question about Indian enforcement rather than international standards, and honest reliance on them stops at the age line.

Education: the variable the statute cannot supply

Every serious account of why child marriage declines, including the government’s own evidence to the standing committee, assigns the largest share to a variable this Act does not control: the schooling of girls. The survey rounds that record the fall from 47.4 per cent to 20.1 per cent record, alongside it, the rise in the proportion of girls completing secondary education, and the mechanism is not mysterious. A girl in school is visible to an institution that notices absence, is deferred from the marriage market by the timetable of examinations, and acquires the earning prospect that changes her family’s calculation of what an early marriage costs. The prohibition Act interacts with that mechanism at two points only. Its prohibition officers have awareness duties that run through schools, and the 2024 directions assign schools a role in identifying girls at risk of being withdrawn for marriage. At neither point does the Act create an entitlement, a seat, a scholarship or a transport route. Those live in the education statute, which confers a right from six to fourteen and stops, leaving the decisive years from fourteen to eighteen, the years in which the marriage decision is actually taken, outside the enforceable entitlement.

This is the structural answer to the question why a well-drafted prohibition statute coexists with a fifth of young women married as children. The Act governs the ceremony. It does not govern the conditions that produce the ceremony, and its own best provisions, the injunction and the prohibition officer, are attempts to reach those conditions with instruments designed for events. A legislature serious about the survey line rather than the statute book would fund the fourteen to eighteen schooling bridge and the registration checkpoint, and would treat the criminal provisions of this Act as the backstop for the coerced and trafficked cases rather than as the policy. The 2021 debate, conducted almost entirely about the number twenty-one, illustrates the displacement precisely: the age in the definition is the one variable in the entire field that the evidence suggests matters least, because marriages occur below whatever line is drawn when the conditions favour them, and are deferred above any line when the conditions change. Parliament spent two and a half years not deciding the variable that matters least. That is the deepest criticism the lapsed bill sustains, and it belongs to the bill’s sponsors and its opponents alike, since neither side’s public campaign was organised around the schooling bridge.

The enforcement record: decline on the survey, force on the ground

The prevalence evidence is the National Family Health Survey, and it should be quoted with its rounds and periods attached. Among women aged twenty to twenty-four, the proportion married before the age of eighteen stood at 47.4 per cent in the third round of the survey (2005-06), fell to 26.8 per cent in the fourth round (2015-16), fell again to 23.3 per cent in the fifth round (2019-21), and fell again to 20.1 per cent in the sixth round (2023-24). The trend is long, steep and continuous across the life of the Act, and it coincides with the expansion of girls’ schooling, the spread of conditional support schemes, urbanisation and the general rise in the age of marriage across the developing economy. Attribution is the contested part. The Act’s defenders point to the decline as the statute working through deterrence, awareness and the prohibition machinery. Its critics answer that the decline tracks the education and income variables far more closely than it tracks prosecutions, that the steepest falls came in periods and places where the Act’s machinery was barely staffed, and that the Supreme Court’s 2024 observation, that prosecution has not proved an effective deterrent, is the judiciary’s own reading of the same numbers. The survey cannot settle that argument because it measures outcomes and not causes, and honest writing on this statute reports the decline and labels the attribution as argument rather than fact.

State action of a harder kind supplies the live test of the penal model. In Assam, whose government under Chief Minister Himanta Biswa Sarma announced a drive against child marriage on a public health rationale, citing the state’s teenage pregnancy figures, the police in February 2023 arrested 2,278 people in the first three days of the drive on the basis of 4,074 first information reports registered across the state, including priests who had officiated. A second phase in October 2023 arrested over 800 more, and the two phases together produced nearly 5,500 arrests. The state cabinet had resolved that men who married girls below fourteen would be proceeded against under the sexual offences statute and those who married girls between fourteen and eighteen under this Act. The Chief Minister stated that the drive would continue until the state assembly elections of 2026 and that his objective was the complete elimination of the practice. The criticism came from the opposition parties in the state and from the districts where the arrests concentrated. The Assam Congress, through its state president Bhupen Kumar Borah, accused the government of misusing the law for political gain and of creating panic; opposition speakers argued that a social practice is not ended by police intervention and that the arrests fell overwhelmingly on minority-dominated districts; women protested outside police stations seeking the release of arrested husbands, which is the voidability problem wearing a crowd’s face, since the wives in whose protection the arrests were made were asking for the marriages, and the men, back. The Gauhati High Court, in bail proceedings arising out of the drive, was drawn into the management of its consequences. The drive is the strongest available evidence for both sides of the enforcement argument. For the penal school, it demonstrates that a state which decides to use Sections 9, 10, 11 and 15 at scale can disrupt the practice’s public performance within weeks. For the prevention school of the 2024 judgment, it demonstrates the cost: thousands of households prosecuted years after the ceremonies, breadwinners in custody, wives petitioning for the release of the men the law calls their offenders, and the marriages themselves, voidable at the wives’ option and not void, mostly undisturbed in status by the entire operation.

One state chose the other lever, the validity lever. The Prohibition of Child Marriage (Karnataka Amendment) Act, 2016, enacted as Karnataka Act 26 of 2017, inserted sub-section (1-A) into Section 3, providing that every child marriage solemnised on or after the amendment’s commencement in Karnataka shall be void from the beginning. The same amendment hardened the penal provisions within the state, substituting a minimum rigorous imprisonment of one year in Sections 9, 10 and 11, deleting the provisos that spare women imprisonment under Sections 11 and 13, and inserting a Section 15-A obliging police officers to take cognizance of offences under the Act of their own motion. Karnataka’s experiment is the natural control case for the national debate. Within its territory, the defence of voidability was rejected by a legislature that had the same statute in front of it, and the consequences the defence predicts, the stripping of status and support from the female party, fall to be managed in that state through the general law rather than through Sections 4 to 7, whose scheme is built around the annulment decree. The experiment also illustrates the federal texture of the subject: marriage sits in the Concurrent List, a state amendment of a central Act required and received the President’s assent, and the law of child marriage validity is therefore not one law but a national scheme with a state exception, a fact that the examination candidate and the practitioner both forget at their cost.

Measured against its purposes, the record is therefore mixed in a specific pattern. As a statement of public policy and as the legal foundation for prevention work, injunctions, awareness and the survey-era decline in prevalence, the Act functions. As a penal code, it functions where a state chooses spectacle and idles where it does not. As a law of status, which is what a marriage statute finally is, it functions almost entirely on paper, because the annulment petition on which its whole protective scheme depends is filed in a small fraction of the marriages it covers. The statute’s institutions exist and are unevenly staffed; its offences exist and are unevenly invoked; its central option exists and is rarely exercised. The distance between the text and the record is not a mystery requiring research. It is the predictable output of a design that prohibits loudly and undoes quietly, at the instance of the party with the quietest voice.

The married child’s remedies under the neighbouring statutes

A woman married as a child who does not annul is not remediless, and a complete account of her legal position must assemble the remedies the neighbouring statutes supply, because in practice they are used far more often than Section 3.

The Protection of Women from Domestic Violence Act, 2005 protects a woman in a domestic relationship, and its definition reaches a relationship in the nature of marriage, so a wife in a subsisting child marriage may seek protection orders, residence orders and monetary relief under that Act without touching the validity of the marriage at all. The general criminal law’s cruelty provision, Section 498A of the Indian Penal Code and its successor provisions in the Bharatiya Nyaya Sanhita, reaches cruelty by a husband or his relatives whatever the wife’s age at marriage. Maintenance is available under the summary maintenance jurisdiction, Section 125 of the Code of Criminal Procedure and its successor in the Bharatiya Nagarik Suraksha Sanhita, which obliges a person of sufficient means to maintain a wife unable to maintain herself, and the courts have read the maintenance provisions with the benevolence appropriate to their purpose. The juvenile justice statute treats a married girl below eighteen who is in need of care and protection as within the Child Welfare Committee’s jurisdiction, a proposition the child protection guide in this series develops and one the 2024 judgment’s betrothal observations reinforce: marriage does not emancipate a minor, and the protection system does not lose her because a ceremony occurred.

The practical significance of the assembly is that the Indian legal system protects the married child around the marriage rather than through it. She may be protected from violence, maintained, sheltered and treated as a child in need of care, while the marriage itself, the source of her exposure, stands untouched unless she petitions under Section 3. Lawyers advising in such cases therefore work the neighbouring statutes first, because they deliver protection without requiring the client to litigate her status against her family, and reserve the annulment petition for the client who affirmatively wants the marriage ended. That professional practice is rational case by case and corrosive in aggregate: every matter resolved around the marriage is a matter in which the Act’s central provision went unused, and the statistics of its disuse are then cited as evidence that the provision is not needed. The voidability design thus generates its own defence in the data, a feedback loop that any honest assessment of the Act has to discount.

The wider frame and the verdict

This Act sits inside a wider body of women’s and children’s legislation whose map, the maternity, violence, dowry, workplace and succession statutes alongside the child statutes, is drawn in the guide to women’s rights legislation in India, and the candidate who reads the two together will see this statute’s peculiarity in relief: of all the major protective statutes in that company, this is the one whose central provision declines to determine the legal status of the act it prohibits.

The three readers of this statute take different things from it, and a pillar should say so before it judges. The person married as a child, and anyone advising her, needs the operational core: the marriage is voidable and not void, the option is hers alone, the deadline is two years from majority, the petition lies to the district court of her residence among other forums, a minor petitioner acts through a guardian or next friend with the prohibition officer, and the decree carries maintenance, residence, custody and the return of gifts with it, while the neighbouring statutes protect her in the meanwhile whether she petitions or not. The prohibition officer and the district official need the preventive core: the duties in Section 16, the injunction in Section 13 with its urgent interim form, the District Magistrate’s deemed authority on the mass marriage days, the offences in Sections 9 to 11 with the presumption that does the prosecution’s work, and the 2024 directions that made the office a sole charge and fixed accountability for default. The aspirant needs the examinable skeleton and its traps: Act 6 of 2007 under a 2006 title, assent on 10 January 2007, commencement on 1 November 2007, the definition split at twenty-one and eighteen, voidable in Section 3, void in Sections 12 and 14, cognizable and non-bailable in Section 15, the 2021 bill’s four proposals and its lapse under Article 107(5) without a committee report, Independent Thought’s reading down of the marital exception in 2017, and the 2024 judgment’s prevention directions with the personal law question expressly left open. Each of those statements has been established in the body of this article from the text or from the named record, and each is a statement that published summaries of this Act can be found getting wrong.

The verdict follows from the architecture. The namable claim with which this article began can be stated in its full form: prohibited but not void is not a slogan about the Act but a description of its load-bearing choice, and the choice fails on the Act’s own terms. A statute whose purpose is the prohibition of child marriage, and whose machinery is organised around preventing ceremonies and punishing celebrants, has no coherent reason to guarantee the product of a successful ceremony a presumptive validity that only its victim can displace, inside a window that runs during her maximum dependence, under a burden of litigation that the examining committee of 2004 identified before the Act was even passed. The defence of voidability, that voidness would strip the female party of status and support, is a real argument answered by drafting: legitimacy is already preserved by Section 6 regardless of annulment, and maintenance, residence and custody could have been attached to voidness expressly, as Parliament preserves consequences in other nullity contexts, leaving voidness-plus-protection available to any legislature that wants both. What voidability uniquely supplies is not protection. It is the quiet continuation of nearly every marriage the Act prohibits.

The reforms the record supports are therefore identifiable and modest beside the rhetoric that surrounds this subject. Make the ordinary child marriage void, with the Section 4 to 7 consequences expressly preserved and with a judicial discretion to recognise consequences where a party in good faith has ordered her life on the marriage. Extend the annulment and claims window well beyond two years wherever voidability survives. Settle the personal law interface by enacted words, in one direction or the other, rather than leaving litigants to a lottery of High Courts. Staff the prohibition office as the Supreme Court directed, as a sole district charge rather than an additional duty. And if the marriage age is to be revisited, revisit it at eighteen for both sexes or justify twenty-one against its enforcement record before a committee that is required to report. None of these requires the criminalisation of a single additional marriage. All of them require a Parliament to finish something. The bill that lapsed in June 2024 is the standing evidence of the price of not finishing: a consultation answered by nearly a hundred thousand citizens, an evidence record built over two years, a clarification the Supreme Court has since asked for, and a file closed by the calendar with nothing decided at all.

Frequently Asked Questions

Q: What is the legal age of marriage for women and men in India?

The minimum age of marriage is eighteen for a woman and twenty-one for a man. The Prohibition of Child Marriage Act, 2006 defines a child in Section 2(a) as a person who, if a male, has not completed twenty-one years of age and, if a female, has not completed eighteen years of age, and a marriage involving either party below those lines is a child marriage. The same ages appear as conditions in the Hindu Marriage Act, 1955, the Special Marriage Act, 1954 and the other marriage statutes. A bill introduced in 2021 proposed raising the age for women to twenty-one, but it lapsed without being passed, so the ages of eighteen and twenty-one remain the law.

Q: Is a child marriage void or valid?

In the ordinary case, neither void nor fully secure: it is voidable. Section 3 of the Act makes every child marriage voidable at the option of the party who was a child at the time of the marriage, which means the marriage is valid and subsisting for all purposes until that party petitions a district court and obtains a decree of nullity. If the party never petitions, the marriage continues untouched. Two classes of case are void instead. Under Section 12, the marriage of a minor who was taken or enticed from a lawful guardian, compelled by force or induced by deceit, or sold or trafficked for marriage is null and void, and under Section 14 a marriage solemnised in defiance of a court injunction is void from the beginning.

Q: Who can be punished for a child marriage?

Three rings of adults. Under Section 9, a male adult above eighteen who contracts a child marriage faces rigorous imprisonment of up to two years, or a fine of up to one lakh rupees, or both. Under Section 10, whoever performs, conducts, directs or abets the marriage faces rigorous imprisonment of up to two years and a fine of up to one lakh rupees, unless they prove they had reason to believe the marriage was not a child marriage. Under Section 11, any person having charge of the child, including a parent or guardian, who promotes the marriage, permits it, or negligently fails to prevent it, including by attending it, faces the same scale, and Section 11(2) presumes that person negligently failed to prevent the marriage. The child party is never punished. All offences under the Act are cognizable and non-bailable under Section 15.

Q: Can a child marriage be annulled and within what time?

Yes, by the person who was a child when the marriage was solemnised, and by no one else. Section 3 gives that party the option to petition the district court for a decree of nullity. The petition may be filed at any time before the petitioner completes two years of attaining majority. Majority arrives at eighteen under the Indian Majority Act, 1875, so for a bride married below eighteen the deadline is her twentieth birthday. There is no power in the court to extend the window. If the petitioner is still a minor, the petition may be filed through a guardian or next friend along with the Child Marriage Prohibition Officer. A marriage within Section 12 or Section 14 needs no annulment, being void already.

Q: Does the child marriage law override personal law?

The Act does not say that it does, and the question remains open. Its text contains no general clause giving it effect over personal law, custom or usage, although its offences and machinery apply to all citizens. High Courts have divided on whether a marriage recognised by the personal law of the parties, particularly where that law treats puberty as the marriage threshold, is governed by the Act’s validity scheme. The Union government asked the Supreme Court in 2024 to declare that the Act prevails over personal law, and the Court declined, leaving the question to Parliament. The 2021 amendment bill contained an express overriding clause, but the bill lapsed, so no enacted answer exists and the position varies with the forum.

Q: What happened to the bill raising the marriage age to 21?

The Prohibition of Child Marriage (Amendment) Bill, 2021 was introduced in the Lok Sabha on 21 December 2021 by the Minister of Women and Child Development, Smriti Irani. It proposed defining a child as any person below twenty-one for both sexes, extending the annulment window from two years to five, and giving the Act overriding effect over other laws, customs and practices. It was referred to the Standing Committee on Education, Women, Children, Youth and Sports, which received around 95,000 emailed representations, nearly 95 per cent of them opposed according to PTI reporting of committee sources, took evidence, and received extensions of time over more than two years. The committee never presented a report. When the seventeenth Lok Sabha was dissolved in June 2024, the bill lapsed under Article 107(5) of the Constitution.

Q: What did the Supreme Court direct on child marriage in 2024?

In Society for Enlightenment and Voluntary Action v. Union of India, decided on 18 October 2024 by a bench of Chief Justice D. Y. Chandrachud and Justices J. B. Pardiwala and Manoj Misra, the Supreme Court held that child marriage violates the child’s rights, including the rights to life, education and choice of partner, and directed that enforcement prioritise prevention over prosecution, observing that a purely penal approach waits for the harm and has proved ineffective. The directions required states to appoint Child Marriage Prohibition Officers as a sole district charge, fixed district administration responsibility for prevention, called for specialised police capacity and fast-track courts, empowered magistrates to act on their own motion with injunctions, and directed community programmes including the Child Marriage Free Village initiative, awareness campaigns and training. The Court also suggested Parliament consider prohibiting child betrothals.

Q: What maintenance and custody rights follow annulment?

When a district court grants a decree of nullity under Section 3, it may order the male contracting party, or his parent or guardian where he is a minor, to pay maintenance to the female contracting party until her remarriage, fixed by her needs, the lifestyle she enjoyed during the marriage and the payer’s means, payable monthly or in a lump sum. Where she is the petitioner, the court may also order her residence until remarriage. Under Section 5 the court makes custody orders for children of the marriage, with the child’s welfare and best interests as the paramount consideration, and may order maintenance for the children as well. Under Section 6, children born of the marriage remain legitimate for all purposes despite the annulment. Orders under Sections 4 and 5 may be modified later on changed circumstances.

Q: What does a Child Marriage Prohibition Officer do?

The Child Marriage Prohibition Officer is the Act’s field officer, appointed by the State Government by notification under Section 16 for the whole state or a specified part of it. The duties in Section 16(3) are to prevent solemnisation of child marriages by such action as the officer deems fit, to collect evidence for the effective prosecution of offenders, to advise individuals and counsel residents of the locality against promoting or allowing child marriages, to create awareness of the harm child marriages cause, to sensitise the community, to furnish periodical returns and statistics to the State Government, and to perform other assigned functions. The officer may be invested with specified police powers and may move the court for maintenance, custody and injunction orders, and accompany a minor petitioner in an annulment case. Officers are public servants, and the Supreme Court’s 2024 directions required them to hold the district charge solely, without additional duties.

Q: How can a child marriage be stopped before it is solemnised?

Through an injunction under Section 13. A Judicial Magistrate of the first class or a Metropolitan Magistrate who is satisfied, on the application of a Child Marriage Prohibition Officer or on information from any person, that a child marriage has been arranged or is about to be solemnised may issue an injunction prohibiting it against any person, including a member of an organisation. A complaint may be made by anyone with personal knowledge or reason to believe, and by a non-governmental organisation with reasonable information, and the magistrate may also act on the court’s own motion on a reliable report. Notice and a hearing are normally required first, but in a case of urgency the court may issue an interim injunction without notice. For mass marriage days such as Akshaya Tritiya, the District Magistrate is deemed to be the prohibition officer with power to stop such marriages. Disobeying an injunction is punishable, and a marriage held in defiance of one is void.

Q: Is a marriage performed in defiance of a court injunction valid?

No. Section 14 provides that any child marriage solemnised in contravention of an injunction order issued under Section 13, whether interim or final, shall be void from the beginning. The injunction route is therefore the one procedural step in the Act that changes the status of a marriage rather than merely penalising the people involved. A family warned by a court that proceeds with the ceremony anyway loses the shelter of the ordinary rule under which a child marriage is merely voidable at the child party’s option. Disobedience of the injunction is separately an offence under Section 13(10), punishable with imprisonment of up to two years or a fine of up to one lakh rupees or both, subject to the proviso that no woman shall be punishable with imprisonment under that provision.

Q: Can a woman be imprisoned under the Prohibition of Child Marriage Act?

Under the central Act, in two important situations, no. The proviso to Section 11(1), which punishes a person having charge of a child who promotes or permits the marriage, states that no woman shall be punishable with imprisonment, so a mother or female guardian convicted under that section is liable to the fine but not to prison. The same proviso appears in Section 13(10) for disobedience of an injunction. A woman is not shielded in every capacity: a woman who performs, conducts, directs or abets a child marriage falls under Section 10, which carries no such proviso. Karnataka removed the position within that state by its amendment of 2017, which deleted the provisos that spare women imprisonment under Sections 11 and 13 and substituted minimum terms of rigorous imprisonment in the principal offence sections.

Q: Who can file an annulment petition when the person married as a child is still a minor?

The petition under Section 3 belongs to the contracting party who was a child, and where that petitioner is still a minor, Section 3(2) permits the petition to be filed through the petitioner’s guardian or next friend along with the Child Marriage Prohibition Officer. A next friend is a person who acts for a minor in litigation because the minor cannot sue alone, and the statute’s pairing of that figure with a public officer recognises how difficult the step is: the minor’s family is usually the respondent side, so the support of an officer of the state is the difference between a theoretical remedy and a filed petition. The adult party to the marriage cannot file under Section 3 at all, and neither can the state in its own name. The deadline still applies: the petition must be presented before the petitioner completes two years of attaining majority.

Q: What happens to the gifts and valuables exchanged at a child marriage when it is annulled?

Section 3(4) requires the district court, while granting a decree of nullity, to order both parties to the marriage and their parents or guardians to return to the other side the money, valuables, ornaments and other gifts received on the occasion of the marriage, or an amount equal to their value. The order follows notice to the persons concerned and an opportunity to show cause. The provision treats the property transfers that accompanied the wedding as reversible along with the marital status, and it deliberately binds parents and guardians as well as the parties themselves, because in a child marriage the gifts and valuables are typically held by the families rather than by the children whose marriage is being annulled. The return order is part of the decree machinery, so it is available to a petitioner who obtains annulment; it has no operation where the marriage simply continues unchallenged.

Q: Why did Karnataka make every child marriage void in that state?

Karnataka amended the Act by the Prohibition of Child Marriage (Karnataka Amendment) Act, 2016, enacted as Karnataka Act 26 of 2017, inserting Section 3(1-A) to provide that every child marriage solemnised on or after the amendment’s commencement in Karnataka shall be void from the beginning. The state legislature thereby adopted, for its own territory, the position the parliamentary Standing Committee had recommended for the whole country in 2004 and Parliament had declined: that the burden of undoing a prohibited marriage should not rest on the child. The same amendment hardened the penal side, substituting minimum rigorous imprisonment of one year in Sections 9, 10 and 11, deleting the provisos sparing women imprisonment, and obliging police officers to take cognizance of offences of their own motion. Because marriage is a Concurrent List subject, the state amendment required and received the President’s assent, and the validity law of child marriage in Karnataka differs from the rest of India as a result.

Q: How did the Child Marriage Restraint Act 1929 differ from the 2006 Act?

The Child Marriage Restraint Act, 1929, known as the Sarda Act after its sponsor Harbilas Sarda, restrained rather than prohibited. It came into force on 1 April 1930, fixing marriage ages at fourteen for girls and eighteen for boys, raised for girls to fifteen in 1949 and to eighteen for girls and twenty-one for boys by amendment in 1978. Its punishments were trivial beside the later Act, reaching at most three months of simple imprisonment and modest fines, and it said nothing about the validity of a marriage performed in breach of it, made no provision for maintenance, residence or legitimacy on any annulment, and created neither prohibition officers nor an injunction power. The Act of 2006 replaced it by repeal under Section 21, raised the punishments to rigorous imprisonment of up to two years and fines of up to one lakh rupees, made offences cognizable and non-bailable, and added the validity scheme, the annulment consequences and the preventive machinery the restraint Act had lacked.

Q: How common is child marriage in India according to the National Family Health Survey?

The National Family Health Survey measures the proportion of women aged twenty to twenty-four who were married before the age of eighteen. That proportion was 47.4 per cent in the third round of the survey, conducted in 2005-06, fell to 26.8 per cent in the fourth round in 2015-16, fell to 23.3 per cent in the fifth round in 2019-21, and fell to 20.1 per cent in the sixth round in 2023-24. The trend is a long and continuous decline across the life of the Act of 2006, coinciding with rising girls’ schooling and incomes. The survey measures outcomes rather than causes, so how much of the decline the Act itself produced, through deterrence, prevention work and awareness, and how much belongs to education and economic change, remains a matter of argument between the Act’s defenders and its critics.

Q: What did the Assam enforcement drive of 2023 reveal about using arrests to end child marriage?

In February 2023 the Assam government, under Chief Minister Himanta Biswa Sarma, launched a statewide drive justified on public health grounds, citing teenage pregnancy. Police arrested 2,278 people in the first three days on 4,074 registered cases, including officiating priests, and a second phase in October 2023 arrested over 800 more, bringing the two phases to nearly 5,500 arrests. The state proceeded against men who married girls below fourteen under the sexual offences statute and those who married girls between fourteen and eighteen under this Act. Supporters read the drive as proof that the Act’s offences, used at scale, disrupt the practice’s public performance within weeks. Critics, including the state Congress leadership, objected that the arrests fell heavily on minority-dominated districts, prosecuted marriages years after the ceremonies, and left wives petitioning for the release of husbands, while the marriages themselves, being voidable rather than void, mostly stood. The episode is the strongest recent evidence for both the penal and the prevention schools.

Q: Does the Act apply to an Indian citizen who goes through a child marriage outside India?

Yes. Section 1(2) extends the Act to the whole of India and provides that it applies also to all citizens of India without and beyond India, so an Indian citizen who participates in a child marriage outside the country is within the Act’s reach in the same way as a participant in a ceremony at home. The single exception the section states is for the Renoncants of the Union territory of Pondicherry, to whom the Act does not apply. Extraterritorial application of the offences depends in practice on the general rules governing offences committed by Indian citizens abroad, including the sanction requirements that apply to prosecution for conduct outside India, but the statute’s own statement of reach is express. The validity consequences in Sections 3, 12 and 14 are framed around the marriage and the parties, and a marriage to which an Indian citizen is a party is a child marriage within the definition wherever it was solemnised, if either party was a child within Section 2(a).

Q: Is a child marriage solemnised before the 2006 Act commenced still voidable under it?

Yes. Section 3(1) states that every child marriage, whether solemnised before or after the commencement of this Act, shall be voidable at the option of the contracting party who was a child at the time of the marriage. The Act commenced on 1 November 2007, and the option therefore extends to marriages performed under the Child Marriage Restraint Act, 1929, which punished the adults involved in such marriages and said nothing about their validity. The limit is the same closing window: the petition must be filed before the petitioner completes two years of attaining majority, and majority under the Indian Majority Act, 1875 arrives at eighteen regardless of when the marriage took place. For the older cohort the window could therefore close soon after the option was created, since the limitation runs from majority and not from the commencement of the Act, an interaction the section’s text does not soften.